Upload
ana-maria-rozzi
View
214
Download
2
Embed Size (px)
Citation preview
LEARNING FROM LEARNERS’ ERRORS
ANA MARIA ROZZI DE BERGEL
2008
Contents
INTRODUCTION .................................................................................................................... 6
Part 1: Understanding the learner’s personal curriculum. ................................................ 11
1. Language and language learning. ................................................................................. 11
1.a Language ....................................................................................................................... 12
1.a.1 Language as a system ........................................................................................... 12
1.a.2 Early views of language as a system .................................................................... 15
1.a.3 The social function of the language system ......................................................... 17
1.a.4 Conceptualisation, metaphor and metonymy in cognitive linguistics ................. 20
1.a.5 The functional description of language – using the system ................................. 24
1.b Language learning ......................................................................................................... 26
1.b.1 Discussing learning and acquisition .................................................................... 26
1.b.2 Connectionism ...................................................................................................... 28
1.b.3 Cognitivism .......................................................................................................... 30
1.b.4 The Gestalt theory ................................................................................................ 31
1.b.5 Integrative theories ............................................................................................... 32
1.b.6 Constructivism and related theories ..................................................................... 34
1.b.7 Humanistic, situational and social views of learning ........................................... 40
1.b.8 Concept of learning in this book .......................................................................... 44
EXPLORATION AND REFLECTION ............................................................................... 48
2. From interlanguage to the personal curriculum ............................................................. 49
2.a Interlanguage re-defined ............................................................................................... 50
2.b Universals and progression .......................................................................................... 54
2.c The notion of mastery ................................................................................................... 57
2.d Defining the personal curriculum ................................................................................. 62
EXPLORATION AND REFLECTION ............................................................................... 66
3. Main components of the personal curriculum ................................................................. 67
3.a The learners‟ needs and expectations ........................................................................... 68
3.b The learner‟s communication style and strategies. ...................................................... 69
Learning from learners‟ errors Ana María Rozzi de Bergel
iii
3.b.1 Discussion and definition. .................................................................................... 70
3.b.2 Favourable and unfavourable communication styles and strategies .................... 75
3.b.3 Communication styles and strategies analysis .................................................... 83
3.c The learner‟s learning hypotheses, styles and strategies .............................................. 86
3.c.1 Learning hypotheses ............................................................................................ 86
3.c.2 Learning styles ..................................................................................................... 88
3.d The learner‟s method .................................................................................................... 94
3.e The learner‟s questions and their role in the personal curriculum ................................ 95
3.f Compensation: a re-definition and a description of its role .......................................... 97
3.g Errors as manifestations of the personal curriculum .................................................. 101
EXPLORATION AND REFLECTION ............................................................................. 108
4. Final level of attainment in the personal curriculum ................................................... 109
4.a A re-definition of fossilisation .................................................................................... 110
4.b The impact of feedback on fossilisation ..................................................................... 112
4.c The role of acculturation – in whose culture?............................................................. 114
4.d Biological factors ........................................................................................................ 117
4.e From fossilisation to language ceiling ........................................................................ 119
EXPLORATION AND REFLECTION ............................................................................. 123
PART 2: Exploring errors in the EFL classroom .............................................................. 124
1. Error analysis as a teaching tool ................................................................................. 124
EXPLORATION AND REFLECTION ............................................................................. 130
2. Methods and instruments for error analysis ................................................................ 131
2.a Valid samples for error analysis ................................................................................. 131
2.b Descriptive and diagnostic analyses ........................................................................... 139
2.c Error classification in the descriptive analysis ........................................................... 144
2.c.1 Error Rating ....................................................................................................... 148
2.c.2 Error Type .......................................................................................................... 152
2.c.3 Error Class ......................................................................................................... 157
2.c.4 Causes of intralingual errors .............................................................................. 161
Learning from learners‟ errors Ana María Rozzi de Bergel
iv
2.c.5 Compensatory utterances/errors ......................................................................... 175
Final comments ..................................................................................................................... 176
EXPLORATION AND REFLECTION ............................................................................. 179
3. Completing the classification of errors .......................................................................... 180
3.a Productive and unproductive errors ............................................................................ 180
3.b Pre-systematic, systematic, post-systematic stages. ................................................... 186
3.c Developmental, fossilisable errors. ............................................................................ 187
3.d Teaching-induced errors .............................................................................................. 189
3.d.1 Specific causes of teaching-induced errors ....................................................... 198
3.d.2 Compensation-provoking teaching ................................................................... 211
EXPLORATION AND REFLECTION ............................................................................. 214
PART 3: Exploiting errors .................................................................................................. 215
1. Error analysis for completing progress evaluation ...................................................... 215
EXPLORATION AND REFLECTION ............................................................................. 220
2. Error exploitation vs. error correction .......................................................................... 221
2.a Problems with correction ............................................................................................. 222
2.b Error exploitation ....................................................................................................... 228
2.c Comments on error exploitation ................................................................................. 239
EXPLORATION AND REFLECTION ............................................................................. 242
3. Teaching learners at their language ceiling .................................................................. 243
3.a Recognising language ceiling ..................................................................................... 243
3.b Unproductive errors and language ceiling .................................................................. 247
3.b.1 The significance of learners‟ comments on learning and teaching. .................. 256
3.c Horizontal development. ............................................................................................ 261
EXPLORATION AND REFLECTION ............................................................................. 266
4. The correspondence between errors and learning hypotheses .................................... 267
4.a Errors and learning hypotheses: finding the match ..................................................... 267
4.b Error prognosis ............................................................................................................. 269
Learning from learners‟ errors Ana María Rozzi de Bergel
v
4.c Error prognosis and teaching ........................................................................................ 275
4.d Ceiling prognosis .......................................................................................................... 276
4.e The desirability of prognoses ....................................................................................... 277
EXPLORATION AND REFLECTION ............................................................................. 285
Summing up .......................................................................................................................... 286
APPENDIX 1 ........................................................................................................................ 288
APPENDIX 2 ........................................................................................................................ 290
A - Text of the test. ....................................................................................................... 291
B - Description of the testing items. ............................................................................. 293
References.............................................................................................................................. 301
Figures
Figure 1 – Measuring the development of learners‟ interlanguage ........................................... 61
Figure 2 – Main components of the personal curriculum .......................................................... 65
Figure 3 – Kolb‟s learning styles in his experiential learning model ........................................ 92
Figure 4 – The processes of error production and error analysis .............................................. 94
Figure 5 – Summary: the personal curriculum ........................................................................ 107
Figure 6– Valid sample for error analysis ............................................................................... 133
Figure 7 – Practical applications of error analysis .................................................................. 139
Figure 8 – Error classification ................................................................................................. 146
Figure 9 – Compensation ......................................................................................................... 176
Figure 10 – Errors of learners with scores between 65% and 80% in language tests ............. 183
Figure 11 – Errors of learners who scored below the passing mark........................................ 184
Figure 12 – The teacher‟s leadership ....................................................................................... 194
Figure 13 – Experiential learning model ................................................................................. 228
Figure 14 – The process of error exploitation. ........................................................................ 238
Figure 15 – Students who had probably reached their ceiling ................................................ 247
Figure 16 – Ms XXX may have reached her language ceiling. .............................................. 250
Learning from learners‟ errors Ana María Rozzi de Bergel
6
INTRODUCTION
We have been reminded recently of Von Humboldt's statement that we cannot
really teach language, we can only create conditions in which it will develop
spontaneously in the mind in its own way. We shall never improve our ability to
create such favourable conditions until we learn more about the way a learner
learns and what his built-in syllabus is. When we do know this (and the learner's
errors will, if systematically studied, tell us something about this) we may begin to
be more critical of our cherished notions. We may be able to allow the learner's
innate strategies to dictate our practice and determine our syllabus; we may learn
to adapt ourselves to HIS needs rather than impose upon him OUR preconceptions
of HOW he ought to learn, WHAT he ought to learn and WHEN he ought to learn
it. (Pit Corder in Richards, 1974, p.27)
Pit Corder‟s visionary statement, which so clearly defines learner-centredness in foreign
language teaching, is consistent with the now over thirty-year-old trend towards the adaptation
of curricula and methods to the learners‟ learning styles and language needs, which places the
focus on the learner rather than on prescribed methodology. Informed eclecticism, a
combination of methods and techniques taken from different approaches, to construct tailor-
made courses, created by a teacher who is well-grounded on theoretical knowledge and hands-
on experience, is thought to guarantee fully learner-centred teaching and learning. Although
teachers have turned their attention to the analysis and diagnosis of learners‟ needs,
preferences and cognitive structure, this has not always included a deeper exploration of
learners‟ interlanguage as a source of information on inner learning processes and their
Learning from learners‟ errors Ana María Rozzi de Bergel
7
characteristics. The way learners construct their interlanguage system and the errors they make
provide valuable insight into how, when and what they are able or willing to learn and might
truly lead teachers to re-consider their “cherished notions” and to fully understand and
implement learner-centredness.
The classroom teacher is often well prepared to analyse interlanguage from the point of view
of its accuracy and appropriacy, but in teacher education, errors have seldom been presented
as concrete, observable manifestations of language learning processes, to be exploited rather
than simply corrected.
I will attempt a discussion and interpretation of errors as manifestations of the learner‟s
personal curriculum, which runs parallel to the course curriculum and reflects a way of
structuring the language system and approaching learning. I will also provide examples of
teaching methods and activities which seem to enhance learning through the exploitation of
errors, at courses partially based on the learners‟ personal curricula, and we will also reflect
upon the phenomenon of fossilisation, which I prefer to call language ceiling.
This work is based on research I undertook at a language school for adults in Buenos Aires,
Argentina, and then continued privately. It lasted for over twenty-seven years, in several
stages, and I was assisted by a team of professionals consisting of a psychologist, a linguist, a
mathematician, language teachers and course supervisors. The data came from recordings or
transcriptions of the learners‟ linguistic performance during lessons and at language tests,
from interviews with teachers and learners, reports on guided lesson observations and the
administration of a learning hypotheses test, as appropriate for each phase of the research
project. Also depending on the aims of each stage, different quantitative, qualitative and
interpretative methods of data analysis were used, as I will explain later on. The phases of the
project sometimes overlapped, so it is not possible to present them sequentially. I will prefer to
Learning from learners‟ errors Ana María Rozzi de Bergel
8
discuss each manifestation of the personal curriculum and the explored point of
fossilisation/realisation and outline the method used in each case, with a focus on error
analysis, as well as the practical pedagogic implementations of error exploitation.
The subjects under study were learners of English as a foreign language, older than eighteen,
the age when people leave secondary school in Argentina. This was arbitrarily chosen as the
beginning of “adulthood” and the moment when learners start taking English lessons at private
language schools or at extra-curricular courses at universities voluntarily, rather than as part of
the obligatory school curriculum. Learners were, then, university students, company
employees or independent professionals who took lessons at the rate of three to four hours per
week, nine months per year (March through December). Their native language was Spanish
and they had started learning English after puberty, that is, after the end of the critical period,
first defined by Penfield and Roberts (1959), further described by Lenneberg (1967), under
constant scrutiny since then and re-named the sensitive period (Locke, 1997).
Learners who had previously learned English at home, at bilingual schools, at a bilingual
workplace or during long stays in English-speaking countries were excluded from the project,
to focus the study on a specific population and limit the number of variables to control, which
was already very large. Previous language instruction at state-run secondary schools was
sometimes considered insignificant when the results of placement testing showed that its
results had been very poor – a frequent situation in the country.
The learners‟ language production was analysed within the teaching-learning situation: no
data was gathered during spontaneous conversations outside the classroom or in informal
environments. Research was centred on observable behaviour and performance which was
deemed useful to account for errors as manifestations of the personal curricula and on the
results of classroom practices which supposedly addressed them.
Learning from learners‟ errors Ana María Rozzi de Bergel
9
Though based on research, this book is not a collection of research reports, but a teacher-to-
teacher account of the knowledge gained, both formally and informally, during the research
process, and the questions that remain unanswered – probably, the most interesting part of this
exploration. These questions may be related to the deep, hidden, psychological, sociological or
biological roots of the personal curricula, a study which was beyond the scope of our work and
remains an exciting topic of research.
Part 1, UNDERSTANDING THE LEARNER‟S PERSONAL CURRICULUM, is an
overview of the theoretical framework and aims of this book.
Part 2, EXPLORING ERRORS IN THE EFL CLASSROOM, loosely describes part of the
research carried out and its tentative findings, with plenty of examples of learners‟ production
and accounts of lesson observations. It also outlines the insight as well as the doubts derived
from the process.
Part 3, EXPLOITING ERRORS, is more teaching-oriented and has three important aims: to
discuss some accepted teaching practices and materials which may actually go against
learning, to describe activities and methods to exploit errors rather than correct them, in order
to enhance learning and accelerate progress, and to invite discussion and further study.
I will also attempt a description of learners‟ errors at the point of language ceiling, which I
consider the moment when the learner‟s personal curriculum is realised, when the end
performance in that curriculum is reached. It also contains advice to teachers on how to teach
learners at their ceiling.
Learning from learners‟ errors Ana María Rozzi de Bergel
10
Finally, the reader will find a summary of the main ideas outlined in the book and their
implications for teaching and course management, as well as suggestions for possible future
research.
The EXPLORATION AND REFLECTION sections invite the readers to review their
“cherished notions” about their teaching practice and explore new possibilities. They may also
be used to trigger discussion or as activities in teacher education courses or study groups.
For the sake of clarity, I will refer to the teacher as “she” and to the learner as “he”. This will
avoid sentences such as “When the learner makes a mistake he/she must be made aware of
his/her error and given the tools to correct it himself/herself”. No sexism involved.
Learning from learners‟ errors Ana María Rozzi de Bergel
11
Part 1: Understanding the learner’s personal curriculum.
1. Language and language learning.
The terms language and language learning have changed their meaning through history,
depending on the development of different theories and scientific discoveries arising from
empirical research and advances in different fields, such as the neurosciences and cognitive
psychology. In order to clarify some terms and describe the background to our study, I will
briefly discuss several theories of language belonging to the second half of the twentieth
century, with a focus on some notions which seem central to all of them, and which we used as
the basis for our exploration of interlanguage, errors and learning hypotheses:
The concept of language as an integrated system.
The value of conceptual and non-conceptual categorisation styles for building
the language system.
The role of metonymic and metaphorical association styles in language
production and use.
The role of imagination and creativity in the construction of the language
system.
I will then give an equally brief account of the theories of cognitive psychology which
explain how these processes are set in motion in the individual‟s cognitive structure when
learning takes place.
Learning from learners‟ errors Ana María Rozzi de Bergel
12
1.a Language
Language: a system of conventional spoken or written symbols by means of which
human beings, as members of a social group and participants in its culture,
communicate. (Encyclopædia Britannica, 2008).
This clear, simple definition is far from expressing the variety of theories of the nature and
use of language which have developed through time. Language is, to teachers and learners, the
object of study, the thing to learn, the skill to acquire, the knowledge to transmit or the code to
access, depending on the definition they adhere to, which bears close correspondence with the
foreign language methodology they put into practice, for how we teach and learn a language
depends largely on how we define this “object of learning”.
1.a.1 Language as a system
I will refer to language as a highly integrated and interactive symbolic system of systems
used for human communication, the apprehension of culture and the creation of knowledge,
with a grammar code governed by generative rules, a highly flexible lexical code, a phonetic
code, and conditioned or shaped by social, psychological, genetic, cultural, neurological and
physiological factors.
This conception of language comes from a comparison of the main late 20th
century
linguistic theories about the nature of language carried out by Akmajian, Demers, Farmer &
Harnish (1995), which revealed that some concepts are shared by practically all theories, the
clearest exception being behaviourism. Their conclusions are summarised below:
a) Human languages are always governed by rules. All known languages follow
systematic rules of pronunciation, word formation and grammatical structure. The way
in which meanings are associated to utterances in a language is also governed by fairly
Learning from learners‟ errors Ana María Rozzi de Bergel
13
stable rules; that is, the use of language for communication is ruled by important
generalisations which can be stated as rules. These rules are descriptive and not
prescriptive and do not lock the language system but always allow for language to
continue changing and developing.
b) The different human languages are a unified phenomenon. The comparative study of
several languages will disclose common, universal characteristics. There is little which
is expressed in one language which would not have an equivalent in another. At the
abstract level, beyond the surface, languages are notably similar in form and function
and ruled by universal principles.
c) Like Chomsky (1975), most linguists view language as a complex system. The idea of
a language system refers to interrelated elements, ordered in a hierarchical fashion (the
broader concepts including the more specific or restricted ones), which are relevant
only because they belong to the system and if separated from the system, lose their
sense and function. For example, we would have no use for a letter if it did not belong
to the language system and were not used to form words.
How are these elements pictured in the mind? Representations of meanings are
concepts. Concepts allow us to categorise, to group elements into categories, to leave
aside circumstantial or accessory characteristics and see the essential properties of
things. The concept verb allows us to form a category where we place auxiliaries,
modals, etc. as sub-categories, for example. The difficulty seems to lie in describing
the internal structure of a concept.
Traditional views advocated the existence of simple concepts, derived from
perception, such as the concept of red and complex concepts resulting from the
combination of several simple concepts, such as red hat. This is not compatible with
Learning from learners‟ errors Ana María Rozzi de Bergel
14
modern conceptions of language which attach great importance to contexts, culture and
idiosyncratic factors and according to which, red may be a different concept to
different language users or in different contexts. Zadeh‟s (1965) fuzzy view, used to
construct artificial intelligence nowadays, may help to define concepts. Concepts
would be mental categories applied to objects in the outside world which belong to
these mental categories, because of their characteristics, properties and functions. The
mental category triangle would be closed with the definition of “geometric figure,
consisting of three sides”, but it is not so easy to define the category happy, for
example. Zadeh holds that each object belongs to a category to some extent and the
notion of belonging to that category has different degrees, which he defines on a
numerical scale.
However, not even this theory fully explains the relationships between concepts,
which appear to be influenced by several rules, for example, the rules of prototypes: it
is more prototypical to combine the concepts of dear and sister, in dear sister than the
concepts of dear and penguin, in dear penguin. Therefore, the combination of two
fairly typical simple concepts would not result in an equally typical complex concept.
Even with these limitations, the theory of the equivalence between meanings
and concepts is still valid, as is the idea of the necessity to combine concepts and the
possibility to do so creatively. To overcome part of these limitations, we might add
cultural, psychological and social sub-systems to the complex language system.
d) Language is an unbounded system allowing for an infinite number of combinations to
express an infinite number of meanings. Chomsky (Ibid.) attributes the creative power
of language to the reactions that utterances cause in the interlocutor, who is stimulated
to create new ways of expressing his ideas or feelings. These social and creative
Learning from learners‟ errors Ana María Rozzi de Bergel
15
properties of language depend, of course, on the associations which people are able to
create between the different elements in the system, combining them in personal,
original ways, to express new meanings. These associations and combinations are
influenced by the language user‟s creativity and imagination and by the scope and
degree of mastery of his language: people who have recourse to higher levels of
language use have more possibilities to produce innovative or more precise
expressions of meanings.
Although this description of language was produced as a synthesis of concepts which are
common to theories of language, I am going to review some of those theories, to better define
the notion of language as a system and the role of the three central aspects of learning outlined
above in the process of system-formation: abstract conceptualisation, association styles and
the use of imagination and creativity to advance in the construction of the system.
1.a.2 Early views of language as a system
The idea of language as a system in which all elements are interrelated and where the value
of a given element depends exclusively on the simultaneous presence of all the others dates
back to de Saussure‟s (1916/1959) definition of language as “un système dont tous les termes
sont solidaires et où la valeur de l'un ne résulte que de la présence simultané des autres "
1(p.9). It is a structuralist concept, in which each element in the system acquires a contrastive
value deriving from the fact that it belongs to the system. (Allen, 1977) The concept has been
expanded to include the social nature of language and its value to construct meaning and
create a categorization of reality, but the central idea of an interrelated system is still valid.
1 “…a system of interdependent elements, in which the value of the different parts is derived from the
simultaneous presence of all the others”.
Learning from learners‟ errors Ana María Rozzi de Bergel
16
Systems are basically constructed by creating associations, discriminating between similar
and different elements and ordering categories hierarchically. Two predominant association
styles, association by proximity (metonymy) and association by analogy (metaphor) are placed
at the service of building the language system. Associating root to origin is a metaphor,
associating root to tree or soil is metonymy.
These association styles are present in de Saussure‟s paradigmatic (having to do with
synonyms, coherence with stimuli, etc.) and syntagmatic (related to word order, syntax, etc.)
axes of language. Associations are primarily metaphoric along the paradigmatic axis and
metonymy prevails in the syntagm. Language learners also set their creativity and imagination
in motion to advance in the creation of this system by finding new associations between
concepts.
It is interesting that de Saussure‟s ideas should have influenced Lacan (1959), who holds that
the unconscious is structured like a language and claims that its two central mechanisms,
condensation and displacement, are basically two linguistic phenomena: metaphor and
metonymy. Metaphor condenses meaning and metonymy displaces it. The unconscious
possesses language awareness.
Behaviourism advocated the idea of language as verbal behaviour (Skinner, Ibid.), a group
of behaviours conditioned by a social process of stimulus-response-reinforcement by which
users learn progressively more complex chains of language units. Chomsky (1957) challenged
this view and postulated that in any syntactic description of language the surface structure has
to relate to a deep structure from which it derives. According to his transformational grammar,
each utterance “contains” others, as part of its deep structure. He also stated the existence of a
finite number of rules used to generate an infinite number of utterances, a process which is
Learning from learners‟ errors Ana María Rozzi de Bergel
17
made possible by the explicit knowledge of such rules. Similarly to de Saussure before him,
Chomsky talked about language universals: abstract concepts which rule the structuring of any
human language.
In the second half of the 20th century, more importance was given to the study of the
influence of the social medium in the structuring, transformation and development of
language. Hymes (1979) criticized Chomsky for having defined linguistic competence as the
ideal command of abstract rules, thus not giving enough importance to communicative
competence, that is, the use of language for communication in certain contexts, for specific
purposes and within various cultures. He also pointed out that language use should be marked
by errors, hesitations and idiosyncratic elements.
Pragmatics was later incorporated to linguistic study, to go beyond syntax and semantics and
explore the value of utterances as carriers of the users‟ intentions and purposes. (Dascal, 1999)
has even held that without an element of intentionality, there is no communication.
1.a.3 The social function of the language system
In the 1970s, language began to be regarded as a social construction, including the notions of
contexts of use and cultural components of discourse. One of the starting points in this trend
was Firth‟s (1977) criticism of Bloomfield‟s (1977) structuralism and his disagreement with
the American school. He considered that language is fundamentally meaning and that
linguistics should concentrate on the analysis of meaning and not of grammatical systems. He
rejected the distinction between langue and parole made by de Saussure (Ibid.) as well as
Chomsky‟s (1957) concepts of performance and competence. He saw language as a social
function, a way of “doing things”, each language act deriving its meaning from the context
where it occurred, which included many non-linguistic elements, and had an effect on this
Learning from learners‟ errors Ana María Rozzi de Bergel
18
same environment. Language appeared as an integrated system of linguistic and social
elements, with a primarily social function.
This theory was called contextual theory of meaning or context of situation. The latter name
was actually taken from Malinowski (1923), who had coined the word years before. Its
implications for language teaching were extremely important, as it highlighted the importance
of contexts over and above the teaching of grammar rules. At the same time, it denied all
possible generalizations about language and the notion of universals, by making meaning
depend entirely on contexts. Malinowski regarded language as “a mode of action and not an
instrument of reflection” (p.296), that is, as human behaviour. This view emphasised the role
of language in practical action and as a link in concerted human activity, as a piece of human
behaviour (p.296). Note that these ideas have a behaviouristic component: language as
actions, events, social behaviour, but they also lay an emphasis on communication and its
cultural aspects, by making language use rely heavily on contexts.
We find the origin of functional and systemic descriptions of language in Halliday (1976).
He conceived of grammar as a non-arbitrary network or system of potential options. It is
functional because it tries to fulfill communicative functions by making choices and
combining elements within the system. His position was different from the idea of context of
situation in that although he attached more importance to communicative aspects of language
than to structural features, he did not discard them for language analysis. Halliday‟s basic
three functions of language could well be considered metafunctions: experiential or ideational
(to represent processes, objects, things, developments in the world); interpersonal (to transmit
information to other human beings) and textual (the composition of text so that language will
be operational and perform a social role). Later, Halliday expanded and supplemented this
early classification of functions, but this extremely brief description will be enough here, but I
Learning from learners‟ errors Ana María Rozzi de Bergel
19
would like to point out that he then added the logical function, to express a set of relations
between ideas (Halliday, 1989).
Halliday (1976) warned us that we should not equate functions of language with use of
language:
There are indefinitely many uses of language, which no linguistic theory has
attempted to systematize; but the fact that language can serve such a variety of
purposes is precisely because the language system is organized into this small set
of highly generalized functional components. Whatever we are using language for,
we need to make some reference to the categories of our experience; we need to
take on some role in the interpersonal situation; and we need to embody these in
the form of text. (I think there may also be a „logical‟ component to be brought in,
but this need not concern us here.) We draw on all these areas of linguistic
potential at the same time (pp.29-30).
Language users have to consider the role-relationships, the purposes of the utterances and
their coherence and cohesion (Widdowson, Ibid.) with the interlocutor‟s utterances – their
place in the broader context where they occur, among other elements. The user will also have
to select the adequate verb tense, adjectives and other notions and combine them in a syntagm;
that is, put the words in the correct order. To carry out horizontal associations, such as those in
the syntagm, the person has to associate by proximity, using metonymy, to know “what comes
next”. To find the elements of cohesion and coherence, the vertical associations such as those
in the paradigm, he needs to associate by analogy, using metaphorical associations.
According to Halliday, then, to use language we draw from all three metafunctions at the
same time, so it is not too bold to assume that in order to actually become proficient language
users, EFL adult learners need to construct the system of the foreign language according to a
Learning from learners‟ errors Ana María Rozzi de Bergel
20
set of logical principles, by association and differentiation, and by making connections among
realms of the system to the degree that their imagination and creativity permit.
Errors are manifest in language use, but if to use language we resort to the three
metafunctions simultaneously, we may claim that errors will lead us to gain insight into the
way a person performs these metafunctions and thus enable us to learn more about the
learner‟s hypotheses about the construction of a language system.
1.a.4 Conceptualisation, metaphor and metonymy in cognitive linguistics
Cognitive linguistics advocated the study and description of language with reference to its
underlying mental processes, an idea that Firth had rejected. Two fields were studied:
cognitive semantics and cognitive grammar. Johnson and Lakoff (1980), among others,
centred language study around several key processes:
a) The conceptual metaphor, defined as the comprehension of a conceptual
domain in terms of another conceptual domain, for example, understanding a person‟s
life experience by comparing it to somebody else‟s, which is taken as a model or
paragon. A conceptual domain is any organization of experience and the theory is
based on metaphors joining concrete to abstract concepts or thoughts, sometimes
previous to language, in the brain.
b) Categorisation, based on the Aristotelian concept of categories as individual
entities whose members share certain properties which are necessary and sufficient to
define the group. This notion was then expanded on by Rosch (1981) on the basis that
- people do not categorise according to objective characteristics of things;
Learning from learners‟ errors Ana María Rozzi de Bergel
21
- when they categorise, people find that certain members of a given category are
more typical or preferable, for example, a humming bird is a better candidate for
the category “bird” than an ostrich, and
- because of this, it is easier for people to place more typical members within
categories than those who are less typical. Her research also disclosed the
existence of more comprehensive categories in smaller children, which then
become more restricted or specific with the passing of time, as well as the
presence of networks of categories and sub-categories associated to them.
The presence of these categories created by children on the basis of over-
generalisations seemed to contradict Piaget‟s idea of the absence of abstraction in
children, as pointed out by Wohlwill (1970), who also contributed the idea that
categorisation is taught. When children call all men “Dad”, they are corrected by
adults, who teach them to restrict the category. In the initial categorisation, the child
has worked on the basis of wrong indicators – for example, that all the members of the
category are men – to make the abstraction, but the abstraction has existed.
c) Metonymy: attaching the concrete meaning of a word to another, when the two
have a known similarity in their meaning. For example, “Let‟s buy a Monet” (It is
actually a painting by Monet), or “These lands belong to the crown”. (They belong to
the monarchy or to a monarchic state)
d) Conceptual organization: the way in which we organize the categories into
which we classify linguistic concepts. The frames of reference, the schemata, the
relationships and meanings which develop in the mind and which we access, for
example, to answer questions, give information or make judgements.
Learning from learners‟ errors Ana María Rozzi de Bergel
22
e) The structure of mental schemata. A schema is a recurring structure of
cognitive processes which establishes a pattern of comprehension or reasoning. It
arises from interactions with the social environment, experience, linguistic knowledge
and the historical and cultural factors. (Johnson & Lakoff, Ibid.)
f) Social stereotypes. They are, actually, metonymies. A sub-category within a
category has such a well-known status within a social group that its characteristics are
taken as valid for the whole category. This simplifies judgements and definitions of the
individual members of the group. Schools are often full of these stereotypes: the nerd,
the class clown, the cheerleaders, the barbies, the apple-polishers, etc.
g) Iconicity. The similarity or correspondence between a linguistic form and its
meaning, as well as the analogy between two concepts and the fact that one can stand
for the other. For example, two rock‟n roll singers have iconic representation because
they belong to the same profession, but if one of them becomes extremely successful,
people may refer to him as “a Paul McCarthy”.
h) Gestures and body expression, often conveyors of involuntary but powerful
messages. In pragmatics, communications which are not ruled by volition are often
left out of analyses and in linguistics, non-verbal forms of communication are not
studied. However, body language can only be encoded and decoded according to the
native language or the shared language of the interlocutors involved. Gestures, facial
expressions, body attitudes stand for words and concepts.
Body language would not be perceptible or comprehensible if the user were unable to
“translate” it into other “languages” or adapt it to reach different audiences. Besides, it
should be possible to verbalise the possible meanings of a gesture, a certain body
posture or any other movement made in the representation-interpretation (Austin,
Learning from learners‟ errors Ana María Rozzi de Bergel
23
1962) game, for this body language to be meaningful, because it relies on verbal
language to be comprehended. Because of this, gestures are often causes of social
blunders, when they are addressed at users of different languages where these gestures
“translate” into words or phrases not shared by the interlocutors.
i) Linguistic relativity: language is not a scientific notion; it is a social, political
and cultural notion. Carroll (1973), among others, has claimed that a particular
language
………exerts an influence on the minds of those who use it, channelling their
thoughts in special and distinct ways and perhaps even causing them to experience
their world differently from those who speak other languages. […..] The theory
that languages have special effects on the mental activities of their users has been
called the theory of linguistic relativity, because it asserts that mental activity is to
at least some degree relative to, and dependent on, the language in which it takes
place. (p.126)
The most characteristic statement of this theory is known as the Whorf (1973)
hypothesis or the Whorf-Sapir hypothesis. They held that
“…we cut up nature – organise it into concepts – and ascribe significances as we
do … largely because of the absolutely obligatory patterns of our [own] language.
[…….] Grammars are the cement out of which we fashion experience…” (p.p.134)
The theory was hard to prove, because despite offering many examples of the
correspondence between language and thought, Whorf was unable to demonstrate that
this correspondence was conditioned by language (Gumperz & Levinson, 1996).
Learning from learners‟ errors Ana María Rozzi de Bergel
24
Language relativity analised categorisation most specifically. Whorf pointed out that it was
almost impossible to define categories “on the basis of nature”, and that they were not there
“staring us in the face” but their definition had to be attempted with reference to the necessary
and obligatory categories in the language of the definer, containing a vision of the world
shared only by the users of that language.
Carroll (Ibid.), on a more moderate key, held that as languages differ in the way they encode
subjective and objective experience, their users tend to classify and categorise experience in
different ways, according to the categories which exist in their languages. These differences
should necessarily affect behaviour.
We should note the influence of these theories on foreign language teaching and learning:
learning a language would also entail observing and codifying experience in a manner
resembling that of native speakers of that language. Whether that is possible or even desirable,
in the case of a foreign language, remains a moot point.
1.a.5 The functional description of language – using the system
Austin (1962) and Searle (1980) contributed to a more precise functional description of
language, particularly with the concept of speech acts, which acknowledges the difference
between structure and function, joined by intent. Language users choose the elements they
need to use in order to perform a certain speech act. Austin (Ibid.) made a distinction between
what we say (locutionary act), what we mean (illocutionary act) and what we achieve when
we perform a speech act (perlocutionary act). Searle (Ibid.) has been credited with
participating in the development of the theory of speech acts and having introduced novel
elements into it, like the role of the interlocutors‟ intentions in the construction of meaning in
speech acts. His research, however, was more addressed at defining the use of language than
Learning from learners‟ errors Ana María Rozzi de Bergel
25
its intrinsic characteristics, better described by Wilkins (1976) and used since then as the basis
for all language programs with a functional-notional language syllabus.
Notional-functional descriptions do not question the existence of paradigmatic and
syntagmtic axes in language or language structuring as a system of interrelated elements,
deriving their significance from the whole. The functions, similar to Austin and Searle‟s
speech acts, are constructed with recourse to general notions (Duration, time, etc.) and specific
notions, which are semantic units (Within the general notion of time, the word yesterday is a
specific notion). This conception also includes, as we can see, a hierarchical-logical
organisation of language categories.
Going back to the concept of language as a system, it is quite clear that de Saussure‟s idea
has been expanded to include social, cultural and personal factors in the system, but the
systemic nature of language remains practically unchallenged. It seems possible, then, to
conceive of language learning as a process of system formation ruled by the learner‟s style for
placing items into categories, organizing these categories hierarchically and making
associations among them, either by proximity or analogy. Creativity would entail opening up
possibilities for more novel associations or transferring concepts from one context to many
others. We are going to analyse the role of these concepts in different theories of learning.
Learning from learners‟ errors Ana María Rozzi de Bergel
26
1.b Language learning
When do we claim that a person has “learnt” a language? How are languages learnt? These
and many other questions have haunted researchers for years, and will continue being the
object of scrutiny. Finding partial answers to these queries was one of the main objectives of
our work. Errors can help us gain more insight into learning processes, but the information
about these processes is by and large still incomplete.
In the previous section, I defined the view of language used in our research and then
expanded on it. Here, I am going to get to our conception of learning through an analysis of
some relevant schools of thought, thus reversing the procedure, as I have already explained the
three basic cognitive processes central to our research: conceptualization, association, use of
imagination and creativity, within system-formation. I will now point out the importance
attached by cognitive psychology to these processes.
1.b.1 Discussing learning and acquisition
When talking about learning I mean learning in order to acquire. Learning and acquisition
have been often described as two different processes, not necessarily related (Krashen &
Terrell, 1983). I will consider that learning leads to acquisition and always precedes it, even
when it takes place in very informal environments and at an extremely fast pace which makes
the learner think there has been no formal learning. In Ellis‟ (1985) definition, acquisition is
considered the internalisation of rules and formulae which are then used to communicate in the
foreign language. In this sense, the word acquisition seems to be a synonym of learning, but if
we consider that communicative activities are the link which leads to the acquisition of what
has been learnt, we should conclude that learning occurs before acquisition.
Learning from learners‟ errors Ana María Rozzi de Bergel
27
A similar view was held by O‟Neill (2002), who even suggested that Krashen‟s ideas would
be better understood if he stopped using the word acquisition, particularly when referring to
adult language learning. Harmer (2001) also saw conscious learning as a methodological
device which encourages and favours acquisition through the provision of a considerable
amount of finely-tuned input and an emphasis on communicative tasks and activities which
facilitate internalisation of what has been learnt.
I will also refer to Harmer‟s (Ibid.) criterion for defining the role of age in the
correspondence between learning and acquisition, that is, that adults need to learn about the
language to learn the language. This approach allows for the explicit statement of rules, the
exploration of verbal logic, learning activities and structured learning methods in general, a
conception which departs from the basic tenets of the natural approach.
At this point, it seems appropriate to reflect upon the word natural, and its different
implications, as what is natural for some may go against somebody else‟s nature. Moreover, at
this stage of psychological research, it is a household truth that there are as many cognitive
structures as human beings, which may make a generalised definition of “natural” learning
rather complex, not to say undesirable, for it might lead to useless and even harmful over-
generalisations.
Considering that many of the arguments to demonstrate the existence of acquisition as
defined by Krashen and Terrell come from observation of first-language development or of
second-language development by immigrants in English-speaking communities, we should be
careful when trying to transfer this notion to the EFL learning situation of adults, where it
hardly ever applies. It is extremely unlikely that a learner will “acquire” language by osmosis
in a classroom, following a course based on a syllabus, and surrounded by other non-English
speaking learners. In these situations, formal teaching, however flexible and adaptable,
Learning from learners‟ errors Ana María Rozzi de Bergel
28
coupled with conscious learning and communicative practice, leads to acquisition in the sense
Ellis uses the word. This follows the natural process for these teaching-learning situations.
It should be noted that I am making this claim only on the basis of my life-long experience
in teaching and not based on formal research. To make up for this feebleness in my arguments,
I will offer, in the following sections, some brief reviews of cognitive theories which shed
more light on learning mechanisms and describe conscious as well as unconscious processes,
their significance and role in the apprehension of different forms of knowledge.
1.b.2 Connectionism
Early connectionism of the first sixty years of the 20th century explained learning as a series
of stimulus-response associations. The exploration of learning processes was largely carried
out by means of laboratory experiments, such as Pavlov‟s (1927) work with conditioned
reflexes in animals.
Psychologists researched concepts such as the frequency of occurrence of an event and how
it could be modified, the effect of positive and negative reinforcement on behaviour, the
extinction of behaviours when their reinforcement disappeared, their spontaneous recovery,
generalization (the tendency to produce the same response to similar, but different stimuli) and
discrimination (the tendency to distinguish and discriminate, through reinforcement, similar
stimuli).
These scientists were the heirs of positivism and believed that their work had to be centred
on the study of observable phenomena, leaving the explanation of their causes in the hands of
philosophers. Positivism was against inferring, assuming or exploring non-observable entities
or processes. Milder versions of this school of thought accept that human behaviour is not
Learning from learners‟ errors Ana María Rozzi de Bergel
29
always the response to direct stimuli and Skinner (1957) recognized changes in behaviour
depending on the operant conditions; that is, the circumstances under which it was emitted.
Since 1986, we have known modern connectionism, which conceives of the mind as a
system with a series of processing units or nodes resembling neurons, interrelated by multiple
connections and forming networks. Between the points of input and output of information,
there are a series of hidden layers through which information spreads until an outcome is
reached in the output units.
Learning takes place when two units of a neural network are excited in a simultaneous
fashion and the connecting force between them increases. When a pattern arrives at the output
layer, it finds a system there which compares the ideal output to the actual output, and this
reinforces, retrospectively, the path followed by the activation pattern which best resembles
the ideal output. After several repetitions of this trial and error process, learning occurs.
(Díaz-Benjumea, 2002)
Connectionism has explored conceptualization. The placement of elements in conceptual
categories with necessary and sufficient characteristics seems to pose certain problems,
mainly, the variations and exceptions within categories. Connectionism seems to clearly
define the characteristics of categories and handle categorization with great ease, as networks
can learn to detect subtle patterns of repetition of special characteristics which would be
difficult to explain applying general rules. This tries to account for the flexibility and insight
of human intelligence with recourse to methods which are not exception-free, so as to avoid
rigid forms of symbolic representations. Cognition needs a language of thought, but its rules
have to be flexible, not rigid as those which govern algorithms or as the rules of physics
(Horgan & Tienson, 1996).
Learning from learners‟ errors Ana María Rozzi de Bergel
30
1.b.3 Cognitivism
Cognitive theories of learning differ from connectionism in that they concentrate on human
perception, attitudes, beliefs, motives and values and the way in which these forms of
cognition interact with experience, shaping each other in the process. They deal with the
causes of phenomena, claiming that a mere objective description of behaviour is by no means
enough for the study of human learning, for if we lose sight of people‟s motivation, feelings
and life experience, we may be treating them like laboratory rats. Humans do not always learn
because there is a piece of cheese at the end of the maze and sometimes, they learn without the
stimulus of a visible, tangible reward – with no cheese. Even more, they often learn even by
putting up with a lot of negative reinforcement, criticism or failure. Cognitive psychologists
challenge connectionists to explain these phenomena, which the modern equations ruling
neural networks do not seem to clearly map out.
According to cognitive theories, people learn when they are involved in situations they find
meaningful and significant, to the point that there are no generalisable meanings, because the
meaning of a situation is inside the individual, and only by understanding the person
empathically, by talking to him, can we reach this meaning. Please note that this is
intrinsically related to the idea of the personal curriculum in this book, which I will explain in
the following section.
These positions derive from Descartes‟ philosophical idealism and have found a place in
phenomenology. Rogers (1972) held that the individual exists within a phenomenological
field. The important thing, in this study of human cognition, is not an event or an object per se,
but how the person perceives and understands it.
Learning from learners‟ errors Ana María Rozzi de Bergel
31
All the learner-centred approaches of the 1980s derive from this position. The teacher can
only understand the processes developing in the learner‟s experience by “stepping into his
shoes”, by understanding the learner empathically. This is what our research attempted and the
main reason for embracing an exploration of the personal curriculum. Perhaps this exploration
and the desire to empathise with the learner contributed more to raise the quality of our work
than the answers we might have found.
1.b.4 The Gestalt theory
The first questioning of connectionism came, however, from the Gestalt theory advocated by
three German theoreticians living in the United States: Wetheimer, Köhler & Koffka.
They considered learning as the perception of a complete object which cannot be divided
into components of the stimulus-response type without damaging its meaning. The learner‟s
conscience perceives a scene as a meaningful whole and this is what makes this whole
relevant to produce changes in the person‟s behaviour. Our thought is composed of whole,
meaningful perceptions, scenes, rather than by elements associated by conditioning or
otherwise, or by isolated images. (Köhler, Koffka & Sander, 1963) When we see a giraffe, the
concept our mind perceives in the first place is that of the whole animal, a giraffe, not an
ensemble of four legs, two ears, a body, a neck, a tail and all the other “components” which,
put together, would make a giraffe.
According to this, learning does not entail erasing tracks and producing others, but changing
the person‟s perceptive configuration of whole fields of stimuli (meaningful gestalten)
Learning occurs when this restructuring of the field takes place, which may happen all of a
sudden, through insight, or after a period when the learner seems not to understand or care
about the problem, although he is exposed to it.
Learning from learners‟ errors Ana María Rozzi de Bergel
32
Insight calls for presenting different aspects of a situation as a whole or at least, related in
such a way that they may stand out as a Gestalt on a less clear-cut background. The laws of
Gestalt state that items which are closer in time, space or in their logical association are
always grouped together and that closed areas tend to be perceived as units more quickly. The
importance of the learning activity does not depend on reaching the reward at the end of the
road but on completing an activity the learner may find meaningful. In industry, this principle
was used to explain that it is frustrating for a worker to labour at an assembly line without ever
seeing the finished product. In class, we always try to round-up lessons so that learners will
experience a feeling of achievement.
Memory tends to close open areas, so if teachers do not round-up topics or lessons, learners
will, following their own cognitive paths. They will tend to turn the unknown into known,
which explains that over-generalisations are not accidents, but they will necessarily occur
when teaching does not help learners to discriminate between similar grammatical structures.
According to the Gestalt theory, discrimination is not spontaneous but must be guided and
stimulated. Generalisation, on the contrary, is a general tendency towards forming closed,
coherent systems.
The Gestalt theory also claims that comprehension aids learning more than mechanical
exercises or, in any case, it is the best possible reinforcement of learning.
1.b.5 Integrative theories
Edward Chace Tolman tried to bring together the tenets of connectionism and cognitivism.
His theory has been called purposive behaviourism. Tolman (1932) believed that human
beings do not simply respond to stimuli in a conditioned fashion, but they act on the basis of
their beliefs and interests, responding in complex ways to fields of stimuli but never to
Learning from learners‟ errors Ana María Rozzi de Bergel
33
individual, isolated stimuli, in a one-to-one stimulus-response relationship, at least, outside the
experimental laboratory. Tolman considered many “variables” of the stimulus-response
process, mostly: knowledge, beliefs, demands, interests which affect the individual‟s response.
When an element of the field of stimuli makes an impact on the learner, it triggers the wish
to respond according to the person‟s view of the world, his beliefs about causes and
consequences, how things are sequenced in time and other concepts belonging to his particular
perception of the world. Predicting behaviour becomes more difficult than in the simple, direct
stimulus-response paradigm.
The notion of mediating response was also a move towards reconciling connectionist and
cognitive theories. A mediating response to a field of stimuli is one which produces further
stimuli in the same person. For example: when a learner receives a stimulus he finds
interesting, this produces in him the response of paying attention, beyond the actual contents
of the original stimulus, which of course, he will feel more ready to explore, thanks to the
mediating response. (Slamecka & Ceraso, 2000)
These are the typical learning situations in which, rather than respond to a stimulus, learners
show interest in finding knowledge or answers by themselves. We might conclude it is better
to present a problem than to send a stimulus, but the problem remains of how to frame this
problem so that it is close-ended enough to avoid over-generalisation and open-ended enough
to induce discrimination.
We may claim there is a mediating response for each concept in each person, which is a
word associated to it which evokes another concept, paving the way for the individual to
complete the meaning of the thing in question. When a word is associated to a thing, a third
concept, associated to both, is evoked. This completes the meaning of the word.
Learning from learners‟ errors Ana María Rozzi de Bergel
34
In the classroom, we accept semi-gramatical utterances as long as they convey their meaning
and achieve communication: they maintain their connection with the mediating response they
evoke – the utterance which would have been fully correct.
This concept is shared by pragmatics (Dascal, Ibid.) and it is also found in day-to-day
communication among native speakers of a language, where a lot of imperfections are
tolerated.
1.b.6 Constructivism and related theories
In 1947, Jean Piaget published his theory of human cognitive development, which
acknowledged four stages: sensory-motor, pre-operational, of concrete operations and of
formal operations, with the following characteristics:
a) They always occur in the same order.
b) They are universal and are not affected by cultural factors.
c) They are related to cognitive development.
d) The sequence is organised hierarchically, that is, each stage contains elements of the
previous ones, which are integrated but also stand alone.
e) The represent qualitative as well as quantitative differences in thinking styles.
In the last stage, which appears in adolescence, Piaget places the capability for conceptual
thinking, which enables the individual to group objects or events according to their essential
and permanent characteristics and associate or differentiate abstract concepts with no need for
empirical or motor elements, by means of conservation: the possibility to retain and remember
abstract or invariable characteristics.
In each stage, cognitive structures adapt to the development process by assimilation and
accommodation, an idea which Bartlett (1932, 1958) had already put forward. In assimilation,
Learning from learners‟ errors Ana María Rozzi de Bergel
35
the person interprets events with reference to his previously acquired structure, but in
accommodation, this structure changes to adapt to the medium. Learning may be described as
a constant challenge to the existing structures, which have to make room for others, either
more advanced or requiring higher mental processes – a concept related to the pedagogic idea
of creating cognitive conflicts to trigger curiosity and the eagerness to learn.
Piaget grouped mental operations into systems. Each group is a class, but classes are placed
into a classification or system of classes, within which they make sense. These groupings must
be liable to several operations: classification, categorization, substitution, multiplication,
relations; in other words: hierarchical classification systems must be related to each other for
intelligent thinking to exist.
In this theory, Language is not only the vehicle for the apprehension of culture but also an
agent for the construction of knowledge. These ideas bear great resemblance with Vygotsky‟s,
whose work Piaget came to know after the former‟s death, for political and historical reasons.
They are the foundations of constructivism. In Vygotsky‟s view, however, there are no clear-
cut genetically determined stages, but the moment when children go from one stage to the next
depends on the stimulation they receive and is different for each individual. The social
medium heavily influences the duration of each developmental stage. (Vygotsky, 1934/1962).
It is interesting to note that Vygotsky considers that language emerges as a social tool, then
becomes egocentric and finally reaches the stage of inner talk, one of the higher mental
processes. Piaget also talks about egocentric speech in children before the stage of abstraction,
not meaning that children are self-centred, but that they are not able to comprehend events
except through the filter of their own experience, which they do not differentiate from that of
others (Piaget, 1962). He also describes egocentric speech as a succession of monologues in
conversation, without strategic adaptation of the messages or a genuine attempt at following
Learning from learners‟ errors Ana María Rozzi de Bergel
36
the interlocutor‟s line of thinking. He places this egocentric speech in the period of pre-logical
thinking, when the person groups elements by their function, external characteristics, uses, and
not by their essential characteristics.
These concepts are related to the discussion of communication strategies and learning
hypotheses I will undertake further on. Our research disclosed egocentric speech and the use
of non-conceptual or pre-conceptual categories for grouping elements in adult learners, very
often, in those who seemed to have a lower language ceiling in the foreign language.
Egocentric speech often materialized in a communication strategy I have termed soliloquy, by
which conversation is just two monologues, one interrupted by the other to let the interlocutor
deliver a portion of his speech, but where no real exchange of ideas occurs.
On the basis of Vygotsky‟s ideas and the main tenets of constructivism, Bruner (1966)
developed the idea of learning as an active process by which learners build new knowledge on
the basis of already acquired knowledge, by a process of selection, hypotheses-forming and
testing, decision-making and others, all of them relying on cognitive structures which allow
them to go “beyond the information given” (Bruner, Ibid.).
His theory of instruction rests on four main concepts:
a) The predisposition for learning.
b) The structuring of materials or any form of didactic presentation of new knowledge, so
that learners will find it accessible.
c) The efficient sequencing of items to be learnt.
d) The way in which reward and punishment are progressively provided. Bruner (1986,
1990, 1996)
Learning from learners‟ errors Ana María Rozzi de Bergel
37
Bruner (1960) attached special significance to the categorisation of information and the
construction of cognitive maps and most specially, to the processes of perception,
conceptualisation, decision-making and inference.
In his view, elements are grouped into categories which possess objectively necessary and
sufficient characteristic to be in a particular category; however, these classifications are always
influenced by subjective factors, as we have seen.
Bruner specifies four basic attributes objects must possess to be placed in a certain category:
a) Concrete, necessary and sufficient characteristics for an object to be included in a
particular category.
b) The way in which these characteristics are combined in the object.
c) The weight that the different characteristics or properties of objects have for their
definition and inclusion in a given category. (A bicycle is still a bicycle if it lacks a
klaxon, but not, perhaps, if it has three wheels instead of two).
d) The recognition of the boundaries of these characteristics or properties, distinguishing
between variable and fixed properties. In the case of a bicycle, its colour is acceptably
a variable characteristics, but not the fact that it must have two wheels and a seat, for
example.
Bruner describes two basic types of categories:
a) Identity categories: they include objects which share distinctive traits.
b) Equivalent categories: they provide rules to combine objects according to affective,
functional or formal criteria.
Learning from learners‟ errors Ana María Rozzi de Bergel
38
For example, a banana is still a banana (identity) although it may be ripe or green, large or
small, etc. and it is food or nourishment (function) as well as belonging to the fruit category,
according to a botanical classification (formal).
The systems for relating categories form a coding system, a hierarchical structuring of
categories which are related or differentiated. According to Bruner, people interpret the world
in terms of differences and similarities and thus build their personal conceptions of the social
and natural environment.
The stages of development Bruner outlines are not similar to Piaget‟s but gradually progress
towards independence from the concrete and reliance on the abstract. Learning may begin at
the concrete level and move to conceptual levels.
Bruner has distinguished between paradigmatic or logical-scientific thinking and narrative
thinking, attaching different functions to each. The former deals with categorization and seeks
to classify and explain reality. The latter produces coherent and reliable narrations which place
human experience in time and space, chronicle historical events and also tell stories which are
not concerned with the truth or the verification of facts but with their credibility within certain
conventions: a fairy tale is credible if frogs turn into princes and marry the poorest girl in
town, who happens to be the sweetest one.
Paradigmatic thinking is the instrument of science, whereas the narrative modality preserves
collective memory, culture, beliefs and traditions. It is the foundation of our life experience,
which we also conceive of as a tale. The capability for relating events or inventing credible
stories is closely related to second or foreign language learning ability, as is the learner‟s
recognition of different genres and their conventions. The capability for generating, accepting
or understanding coherent fiction is also central to language development. These capabilities
Learning from learners‟ errors Ana María Rozzi de Bergel
39
are usually acquired in the native language and transferred to the foreign or second language
(Bruner, Ibid.).
Within the same school of thought and heavily influenced by Piaget, Ausubel (1963, 1978)
developed his theories through classroom observation and field experiments, not in the
research laboratory, and postulated that learning processes are juxtaposed, representational and
combinational. Knowledge merges with relevant ideas already present in the learner‟s
cognitive structure, though not explicitly verbalized. He describes the use of advanced
organizers, present in the mind before learning occurs and which exist at the highest level of
abstraction, generalisation and categorisation, selected according to their potential for
explaining, integrating and interrelating the material being learnt. These organisers act as a
nexus between new knowledge and the previous extant ideas which relate to this knowledge.
There is a strong connection between these principles and Bartlett‟s (Ibid.) concept of schema
stating that memory is composed of schemata which provide framework for understanding and
remembering information.
These ideas are related to Bruner‟s concept of spiralled learning and with Kolb‟s (1984)
experiential learning model. They have been the foundations of materials design; Ausubel‟s
idea that learners have to discover knowledge through problem-solving is one of the main
premises of task-based learning. The main difference between Ausubel and the constructivists
is, however, that his subsumption theory states that new knowledge is merged with previous
knowledge in a process which does not add cognitive structures, but modifies them.
Learning from learners‟ errors Ana María Rozzi de Bergel
40
1.b.7 Humanistic, situational and social views of learning
The so-called humanistic schools in cognitive psychology point to motivation as the engine
of learning. Maslow and Rogers are two of the most representative thinkers in this domain,
with a strong influence from Dewey (1897) and his experiential pedagogy. Dewey had
advocated that education had both social and individual aspects, had described the role
interests in the advancement of education and generally favoured experience and
experimentation over lecturing or rigid, authoritarian teaching methods. Maslow (1954) sees
learning as an activity which humans carry out to achieve their utmost development and
satisfy affective and cognitive needs. He places this achievement at the top of his well-known
pyramid of needs. According to Maslow, human beings seek to satisfy their physical needs
first, then their needs for safety and security; for love, recognition and esteem; for knowing
and understanding and ultimately, their need for self-actualisation, which leads to satisfying
the need for transcendence. In this paradigm, the individual may move to higher levels of need
satisfaction only when the lower-level needs have been met.
These theories consider the teacher a facilitator and they have greatly influenced teaching in
the late 20th century, but the pillar of learner-centredness was Rogers‟ (Ibid.) notion that
learning is based on the interpersonal relationship between learners and teachers, the latter
being responsible for detecting learners‟ needs and catering for them, developing rapport and
meting out reward and criticism. His idea of the need for a non-threatening learning
environment is similar to Krashen & Terrell‟s (Ibid.) recommendation of lowering the
affective filter to foster learning, by reducing correction, criticism and the degree of demand
placed on the learner.
Learning from learners‟ errors Ana María Rozzi de Bergel
41
Bandura (1977) and Lave & Wenger (1991) focus on the social aspects of learning, claiming
that the learner should find his identity within a community whose knowledge and practices
are acquired through observation, trial and apprehension. In their idea of situated learning,
Lave & Wenger (Ibid.), place learning within social relations or in participatory situations, so
that it always takes place in communities where the learner may act as an observer or as an
active member of the group. Note the affinity with Dewey‟s principles of social, experiential
learning.
These researchers do not focus on cognitive processes as much as on the type of social
interaction which will foster or facilitate learning. Learning, in this conception, is not
acquiring structures or schema to understand the world, but participating in communities of
practice which are already structured in a particular way. Intelligence, in this view, is
distributed in the world and does not belong to each individual, as Gardner (1991, 1993) has
also stated. He claims that “knowledge does not reside with a single individual” (¶8) but
belongs to a community of knowledge, as is the case with an office full of employees who
depend on the skills and understanding of the group to carry out their work. If they work on
computer programming, for example, their intelligence about this topic is widely distributed
across these individuals, in that situation.
If we take these principles in their most radical form, we may assume that, if all learning has
to take place through experimentation, these humanistic approaches underestimate the value of
abstract knowledge. This kind of knowledge may not be the result of concrete experimentation
but of deductive, creative or associative thinking. We might also point out that, if learning
depends on community action, it might never reach higher individual levels than those already
achieved by the learner‟s community, or go beyond the community‟s resources. However, the
relationship individual-social group also includes changes brought about by those individuals
Learning from learners‟ errors Ana María Rozzi de Bergel
42
who act upon their community to transform it, sometimes radically. We could also question
the validity of the communities of practice for learning a foreign language, since in the EFL
situation there is no native-speaking community available to the learners, nor will they build
their own identity in the process of learning a foreign language, particularly, at an adult age.
Lesser & Storck (2001) seem to define a more sensible role for the communities of practice,
for they view them as valuable contributions to the learning process, but not as central to
learning.
These situational or social theories offer very little insight into inner cognitive processes but
have made an important pedagogic contribution to language teaching by emphasising the
relevance of group work and the connections between language and society.
Within humanistic views of learning, the study of the characteristics of different age groups
has also received attention and it is extremely relevant in this work, because we will focus on
adult learners of a foreign language.
Andragogy is a concept which has been used for more than two centuries to denote adult
self-directed learning, sometimes taken as the opposite of pedagogy. I will discuss some of the
claims made by Knowles (1984), one of the most important representatives of andragogy in
the 20th
century:
a) Self-directed learning: as people mature, they stop considering themselves dependent
and begin to act as independent human beings. They need to be involved in the process
of planning and carrying out their own learning.
b) Experience: as people mature, they gain experience and this is a rich learning resource.
However, it is also flawed with biases and preconceptions. Mistakes are part of this
experience and therefore, a part of learning.
Learning from learners‟ errors Ana María Rozzi de Bergel
43
c) Readiness to learn: as people mature, their interest for learning is based on their
willingness to cope effectively with life situations.
d) Learning orientation: as people mature, they begin trying to acquire knowledge which
they will apply immediately and not in a future time. Thus, they are more ready to
learn if they know this will enable them to perform real life tasks more effectively.
Their focus of attention goes from theory to practical or potential problems and in
general, adults learn better through problem-solving.
e) Motivation: as people mature, their motivation for learning becomes more intrinsic.
Knowles‟ ideas have been criticised as non-scientific and largely based on assumptions, but
most of them refer to processes an experienced teacher has observed in her classes. Some
other points are highly arguable, of course: If adults were so independent and self-directed, the
world would have no room for fanaticism, fundamentalism and totalitarianism sustained
through brain-washing of whole populations. As for motivation, any classroom teacher might
argue that a lot of adults study English because they want to obtain a better job, move to an
English-speaking country or obtain a scholarship. Some of them even explicitly state they
would have never taken the course if they had not been driven by external stimuli of this type.
We should also consider that the boundaries between external and internal motivation are so
flimsy that the real problem might not be trying to establish if adults possess one or the other,
but how to define them.
It is also odd to attempt to generalise about how adults learn while at the same time
postulating that their learning is self-directed, because if it is, then it should necessarily escape
any generalisation or categorisation, as each individual is unique and will have a personal way
of learning. It does not seem possible for this self-direction to be realised in the same way for
everybody. However, the concept of self-directed learning is relevant for this discussion, from
Learning from learners‟ errors Ana María Rozzi de Bergel
44
a slightly different point of view: I will set to explore the processes underlying the personal
curriculum, as manifest in errors, as a step to create a classroom medium where each learner
might best use his particular learning hypotheses. Self-direction will be taken not only as
related to motivation, but also as the inevitable result of operations the learner is bound to
perform, because they are already part of his cognitive structure, which he may be unable or
unwilling to modify. Errors will open windows into this cognitive structure.
1.b.8 Concept of learning in this book
I will draw the concept of learning from Bruner‟s (1966,1986, 1990, 1996) theories of
conceptualisation and association, influenced by Piaget‟s previous work and Vygotsky‟s
constructivism, summarised as follows:
a) Learning is an internal mental process involving perception, information processing,
conceptualisation, experimentation and memorisation, among other operations.
Individuals create their own construction of knowledge rather than incorporate a
construction already present in the outside world.
b) Constructing knowledge entails constructing a system of inter-related hierarchical
categories. Categories of a higher class, like verbs, include lower-class categories, such
as modals or phrasals.
c) Formal education provides capabilities and skills which allow people to learn better, as
their higher mental processes develop. All forms of education make us “better
learners”.
d) The central role of the teacher or the materials is to structure contents to be learnt so
that they may be accessible to the learners.
Learning from learners‟ errors Ana María Rozzi de Bergel
45
e) Adults learn within their acquired cognitive structures, creating knowledge on the basis
of previous knowledge and with reference to their schemata. The influence of the
native language on the process of learning a foreign language will necessarily be very
strong. This does not refer so much to translation, but to using the same devices to
construct the system of the foreign language which were used to build up the system of
the native language.
f) Associations, either by proximity (metonymy) or by analogy (metaphor) are natural
and necessary processes.
g) There is a natural tendency towards generalisation. Learners tend to generalise in order
to construct a rule-governed system. The correctness of these generalisations depends
on their cognitive capabilities and on external influences.
h) The transfer of an element from one context into another is carried out by a process of
transposition and similitude, when the item is given a unique identity which makes it
independent from a single context and applicable in various contexts.
i) Imagination and creativity are present in narrative thinking, which Bruner (Ibid.)
considers the means for the transmission of culture and values. Knowledge and respect
for the conventions of narrative texts help the apprehension of culture and are vital for
language learning. The role of education is crucial here, as we are born narrators, but
we need to learn how to structure and develop our narrations.
j) Reward and punishment do not come from a parent-child relationship with the teacher,
but from the learners‟ appraisal of their achievements and failures. The learner is an
intelligent, autonomous being, capable of self-evaluation, as advocated by social-
situational cognitive theory (Lave & Wenger, Ibid.)
Learning from learners‟ errors Ana María Rozzi de Bergel
46
The process of native language acquisition, often used as a paradigm to explain foreign
language learning processes, provides misleading foundations for classroom practices. First
language acquisition takes place in social contexts which help the learner to decode meanings
and assimilate them to his life experience. Language acquisition is associated with identity-
building and acculturation. These conditions are impossible to recreate in the foreign language
classroom, where learners will be putting known concepts into new words.
As Vygotsky (Ibid.) points out:
.... It is known that learning a language at school and learning the native language
entail two completely different processes. When we learn a foreign language, we use
words that we already master in the native language and we only translate. Foreign
language acquisition differs from native language acquisition in that the former uses the
semantics of the latter as its foundations.(p.54)
Vygotsky does not talk about translation in a literal sense, but refers to the constant use, by
adult learners, of the systems of the native language as a source of reference, as a conceptual
framework for the construction of the system of the foreign language.
In his Fundamental Difference Hypothesis, Bley-Vroman (1988) describes the differences
between child language acquisition and adult language learning. He lists nine factors which
produce these differences, of which two are particularly relevant for our discussion: the fact
that in adulthood, the need for formal instruction through a graded and structured programme
is taken for granted, and the advisability of exploiting and formally correcting learners‟ errors,
which will otherwise not take care of themselves, as it seems to happen in the development of
the first language.
We will be discussing learning a foreign language through formal instruction, in a country
where this language is not the official one. We will distinguish this from learning a second
Learning from learners‟ errors Ana María Rozzi de Bergel
47
language, which we will define as a process taking place in a country where this language is
spoken by the largest part of the population. In these latter situations, part of the input for
learning comes from the social medium.
Learning from learners‟ errors Ana María Rozzi de Bergel
48
EXPLORATION AND REFLECTION
1. Think about your motivation for learning something. Did your situation match the
principles of andragogy?
2. Have you ever learnt/acquired a foreign language? How did that happen? Which of the
theories of learning discussed in this section might explain the process you went
through?
3. Think of three instances when you could say you “acquired” knowledge of a foreign
language, for example, while walking along a street in a foreign country.
4. Take one of the instances of acquisition you have listed and map it out into steps: what
happened? What was the process like?
Example: a) I was walking in Rome and I saw several signs which read “Fermata”; b)
I thought that the word had something to do with the French verb “fermer”, which
means “close” or “shut”; c) I tried to see what was happening around the places with
the “Fermata” sign; d) I did not see anything “closed”; e) I noticed that a bus stopped
at one of these spots; f) I began to suspect that “Fermata” meant “bus stop”; g) I
observed what happened at other places and noticed that buses were stopping near
those signs; h) I thought I had learnt that “Fermata” meant “bus stop”, but when I went
back to the hotel, I asked the reception clerk about the meaning of the word, to make
sure I had come to the right conclusion.
5. Analyse what cognitive processes were involved in the example above and in your
examples.
Learning from learners‟ errors Ana María Rozzi de Bergel
49
2. From interlanguage to the personal curriculum
Language teachers have shown great interest, in the last few years, in the role that factors
such learners‟ learning styles, motivation, personality and culture play in learning. Learners‟
behaviour, feelings, opinions and strategies have been carefully scrutinised and continue under
scrutiny. The data has come from lesson observation, surveys, personality and learning styles
tests and video-taping of lessons and interviews.
Sometimes, it seems as if researchers had somehow neglected the main source of
information about learners‟ cognitive processes: their interlanguage. It provides ostensible,
measurable and recordable data. It contains all the tangible and analysable manifestations of
the mechanisms learners set into motion within the classroom situation and is a reflection of
teaching methods and procedures, the teacher‟s leadership style, the results achieved through
the use of certain materials and the learners‟ expectations. When interlanguage has been
analysed, the analysis has mostly remained at the linguistic level, concentrating on
correspondences between linguistic forms in L1 and L2, or the way the verb system develops,
or whether speakers of a certain language show more difficulty learning the English
prepositions than others. The value of interlanguage as a mirror of learning processes and
styles, however, should take its exploration beyond the linguistic level into the learner‟s
personal methods, needs and type of expected end performance. To do this, we need to
broaden the concept of interlanguage to include these further levels of analysis.
Learning from learners‟ errors Ana María Rozzi de Bergel
50
2.a Interlanguage re-defined
Selinker (1972) defined interlanguage as a developing system modelled partly on
successfully acquired target language forms, partly on elements erroneously transferred from
the native language and partly on self-originated rules and semantic content, which sooner or
later reaches fossilisation; a point beyond which learners would keep their interlanguage
unchanged, regardless of further instruction, input or personal efforts.
Without challenging Selinker‟s definition, I will re-define interlanguage for the classroom
teacher as the observable manifestation of the learner‟s personal curriculum, which is shaped
up by a variety of social, psychological, biological and cultural factors, and fossilisation as its
end performance objective, the point when learners consider, consciously or subconsciously,
that they have completed their personal curricula, often different from the teacher‟s or the
course‟s.
The concept of a personal curriculum is similar to the idea of a built-in syllabus, (Pit Corder,
1974; Ellis, 1989) which acknowledges the existence of a personal order of acquisition of
grammar items in foreign language learning, not necessarily matching the teaching
curriculum. I prefer the name “personal” because the term “built-in” might suggest that this
curriculum is innate, a claim Pit Corder (Ibid.) makes, in fact. In the light of modern
conceptions of cognitive development, it may be more appropriate to assume that the personal
curriculum may have genetic components but it should have been influenced by the
individual‟s life experience.
On the other hand, I am not simply talking about a way of building up the grammar system,
but about a style for constructing the language system, according to a person‟s learning
hypotheses, involving the satisfaction of communicative needs, concepts which were
incorporated into curriculum design long after the language learners‟ built-in syllabus was
Learning from learners‟ errors Ana María Rozzi de Bergel
51
detected and explored but which were already present in Pit Corder‟s (1972) conception of a
personal order of acquisition and learning style. I will refer to curriculum in the sense Candlin
suggests: a combination of general statements about language learning, purpose and
experience, goals and objectives, evaluation, the role of teachers and learners, also containing
banks of learning items and how they may be used and exploited in class, according to an
approach and a method. Within curricula, syllabuses are more detailed accounts of what to
teach. (Candlin in Nunan, 1988)
The interlanguage of an adult EFL learner contains some elements modelled on the learners‟
native language, plus standard features of the target language which the learner has
successfully acquired, plus elements of a personal grammar containing errors with reference to
the ideal system of the language being learnt; in Selinker‟s (Ibid.) words, a “separate
linguistic system based on observable output which results from a learner's attempted
production of target language norms” (p.31)). Interlanguage is in constant change and
evolution, but as stated, it tends to fossilise and become relatively stable at a certain stage of
its development.
In first language acquisition, this interlanguage system is partly formed by accessing the
latent language structure (Lenneberg, Ibid.), a kind of in-built brain device which realizes the
universal grammar into the grammar of a particular language. However, Selinker (1975,
1988), attributed the emergence of interlanguage to the activation of a latent psychological
structure in the brain, accessed and activated for the learning of a second or foreign language.
The differences between a learner‟s interlanguage and the language of native speakers should
then be visible by contrast, when both try to convey similar meaning through utterances which
have obviously been constructed according to different linguistic systems. In this view,
interlanguage description and analysis would entail a comparative study of the learner‟s
Learning from learners‟ errors Ana María Rozzi de Bergel
52
utterance with reference to the same utterance in his native language and an equal or
equivalent utterance by a native speaker of the target language.
The latent language structure and the latent psychological structure are in the brain but the
latter is considered less reliable, independent of the universal grammar and generally
overlapping with other learning areas. According to Selinker (Ibid.), 95% of foreign language
learners have access to this second device only, not to the latent language structure described
by Lenneberg, and therefore never achieve mastery of the new language.
Selinker (Ibid.) lists five factors affecting the formation of the interlanguage system:
a) Positive language transfer or negative language interference. Utterances and words
are sometimes modelled on the systems of the native language. When the two systems
are equal, positive transfer facilitates the learning process, but when the systems differ
and one is used to form the other, interference occurs and results in erronous
utterances.
b) Transfer of training – language items resulting from particular approaches, methods,
materials, features of group dynamics or the tutor‟s leadership.
c) Strategies of second language learning - identifiable techniques and methods used by
the learner to process the material to be learned. In this work, I have made a distinction
between strategies and hypotheses and I will suggest that error analysis sheds light on
hypotheses, not only on strategies, which are sometimes observable without too much
analysis.
d) Strategies of second-language communication - identifiable techniques and methods
used by the learner to communicate with native speakers of the target language. I have
also drawn a distinction between communication styles, general approaches to
communication, and communication strategies as defined by Selinker.
Learning from learners‟ errors Ana María Rozzi de Bergel
53
e) Overgeneralization of target language rules.
By analysing learners‟ interlanguage, Selinker (Ibid.) suggests we can see the type of
psychological devices being used for learning. Error analysis should shed light on learning
processes, in particular, those aspects related to fossilisation, special “linguistic items, rules,
and subsystems” that foreign language learners will
…tend to keep in their interlanguage relative to a particular target language no
matter what the age of the learner or amount of explanation and instruction in the
target language received (p.35).
In this perspective, some changes to a fossilised language system may still occur, through
unconscious acquisition or formal correction. Ellis (1985), on the other hand, does not believe
in possible changes to a person‟s fossilised interlanguage, but adds to the definition of
fossilisation the concept of correct fossilised language forms. This idea would be more akin
with the concept of fossilisation as the end performance in a personal curriculum.
I will use the terms language ceiling to refer to this fossilised stage, on the model of
Vygotsky‟s (1962) related concept regarding first language development, which he deems
closely associated to the development of thinking, one acting as a resource for the other, so
much so that thought cannot develop beyond the person‟s language ceiling. I will then add a
few more processes to Selinker‟s list, particularly, sub-categories of intralingual errors,
namely: adherence to first form or meaning learnt, types of simplification and cue-copying,
while also making a distinction between errors which represent an attempt at learning and
those which reflect processes that seem to hamper progress or sidestep learning.
Learning from learners‟ errors Ana María Rozzi de Bergel
54
2.b Universals and progression
Pit Corder (1981) considers learners‟ errors as components of a systematic, regular and
consistent system, probably built according to the rules of a personal grammar. Interacting
with the data they are exposed to, learners create a personalised version of the system of the
target language, partly based on the structural properties of this language, partly on those of
their native language, and partly on their version of the target system. In this view, the system
is also liable to be modified, either by unconscious or conscious processes.
Pit Coder (Ibid..) points out that interlanguages have features in common but show personal
variations as well, depending on each learner‟s personality and situation and probably, the
learning context. He explores the order of development of interlanguage, making some
important considerations:
a) Interlanguage structures are more similar in younger learners, probably due to a more
limited need for communication at an early age. It is interesting to note, in this case,
that Pit Corder does not base this claim on complexities or characteristics of language
systems, but on communication needs, more likely to be found in the latent
psychological structure than in the latent language structure.
b) The properties of the learner‟s interlanguage become more similar in more
communicative learning contexts. Again, communication is seen as shaping up
language and not vice-versa.
c) Considering that there is a universal grammar and there are universal properties of
language, there should be similarities between interlanguages.
Pit Corder holds that interlanguage is a process of transitional competence and not of
restructuring and re-organising. He also acknowledges a point of fossilisation as the learner
advances towards full command of the foreign language. While the latter concept remains
Learning from learners‟ errors Ana María Rozzi de Bergel
55
largely undisputed, it is not very clear how transitional competence might reach higher stages
of mastery without restructuring and re-organising its system. Even if we conceived of
language learning as a linear process, each element added to the previous ones would have an
impact on the whole of the language system.
The theory of a universal grammar and its associated concept of a natural order of
acquisition of both the native and the foreign language seem to contradict Selinker‟s idea of a
latent psychological structure which sometimes works independently from the latent language
structure and therefore, from universal grammar. If universal grammar may only occasionally
be accessed in foreign language learning, how is it that interlanguages show similarities in the
order of acquisition of language items? Is there a universal grammar of interlanguage? There
may be, and this would not necessarily deny the existence of a latent psychological structure
but point to psychological universals within it causing certain items to appear before others.
These psychological universals might be the basic communicative elements of human
interaction which, according to life experience and socialisation, human beings regard as
essential. The properties of interlanguages listed by Pit Corder, which I have reproduced
above, refer more to communicative requirements than to language universals. Psychological
universals might actually be communication universals, fairly independent from cultural
elements. They are perhaps realised by the learner in the system of a particular foreign
language according to an order of priorities determined by basic human communication needs,
drawing from the foreign language only those elements which appear to satisfy those needs
more readily. When these elements are not available or are difficult to access, parts of the
system may be creatively devised by the language user and parts can be modelled on the
native language, but the communicative need is never ignored, set aside, or placed in a
different slot in the order of acquisition.
Learning from learners‟ errors Ana María Rozzi de Bergel
56
Thus, learners may be permanently selecting language elements from the course curriculum
or other input which matches their needs, discarding those which appear superfluous at a
certain stage, and teachers may not know enough about these processes even to realise they are
taking place.
Nemser (1974) describes interlanguage as a succession of approximative systems, directed to
the goal of attaining command of the target language. This view is similar to Rutherford‟s (As
cited in Gas & Selinker, 1983), as he considers interlanguage a structured system constructed
by the learner, which gradually approximates the system of the language being learnt. Its
evolution is achieved by testing the hypotheses the individual makes about the differences and
similarities between the two languages and the specific features of the foreign one. This view,
akin to contrastive analysis, has proved insufficient, as other processes have come to light
through research in error production. Hypotheses are tested by trial and error, according to
Rutherford, when the person attempts to communicate and is rewarded with comprehension or
punished with lack of it, with the subsequent frustration on the part of the unsuccessful
communicator.
Hypothesis-testing, however, is not a collection of independent processes. As Piaget (Ibid.)
pointed out, trial and error is a method which produces a re-accommodation of the whole
system involved, either to incorporate a new feature or to correct an erroneous assumption.
The newcomer causes chaos and the system re-arranges all its elements to accommodate it
and regain stability, so an interlanguage system develops and is transformed as a whole with
each hypothesis which stands or falls. In this process, there are moments of stagnation when
hypothesis testing seems to stop, or a lot of information is apparently held in store for later
use, until suddenly, a significant re-accommodation of the system occurs, producing a leap
Learning from learners‟ errors Ana María Rozzi de Bergel
57
forward in the learner‟s mastery of the language, a process which has been called learning in
plateaux or U-shaped behaviour (Kellerman, 1983).
2.c The notion of mastery
All these theories seem to consider that learners master a language and address their efforts
at mastering another, but this assumption may need further consideration, as mastery does not
seem to be a clear-cut concept. We might define it as the almost perfect command of an
ideally correct system of grammatical, phonological, pragmatic and semantic rules, effectively
used for the apprehension of knowledge and culture, for socialising and communicating within
a community where these rules are generally accepted and applied; that is, a relevant
communication system for the attainment of personal and social goals. We should also accept
that this description fits the system developed by all the native speakers of a language.
However, the native speaker‟s “mastery” will necessarily have heavy idiolectal components,
as Pit Corder (Ibid.) himself acknowledged, because each person‟s perception of mastery will
vary according to life experience, education, personality, goals in life and other factors. It will
also vary in degrees of approximation to the ideally correct system, which in turn makes it
necessary to further define the concept of “native speaker command of a language” within a
spectrum, in different degrees of linguistic and communicative competence. We may find that
comparing the systems of two languages will not shed enough light on the interlanguage
between them; rather, what we should compare is the learner‟s personal mastery of the native
language to a similar level of performance in the foreign language, assuming that this is the
goal he will set for his learning, as this is the level of communication within which he attains
his personal and social goals.
Learning from learners‟ errors Ana María Rozzi de Bergel
58
Learners may not be trying to master an abstraction called “the foreign language”, but a
concrete, personal representation of this language, which they have mentally described,
measured and assessed according to their hypotheses of how much they should learn, for what
purpose, and how they should learn it. Their interlanguage will approximate this self-
originated goal more than the actual system of the foreign language; the learner will have a
personal curriculum modelled on his communicative needs, logical structuring of his native
language, compensation and communication strategies, culture and personal history, to
mention the most important factors. Interlanguage analysis and description, instead of marking
a point between the native and the foreign language, should shed light on the progress of the
personal curriculum towards the end performance, which will be similar to the end
performance in L1, and additionally, on the gap between the end performance in the personal
curriculum and the ideally complete and correct system of the target language.
The degree to which the end performance in the learner‟s curriculum resembles mastery of
the foreign language of native speakers with similar communication needs and goals and at
almost the same level of linguistic and communicative competence might determine the real
height of the person‟s language ceiling with reference to an ideal “mastery” of language. This
mastery, though related to grammatical and lexical accuracy, refers primarily to the
individual‟s ability for using language effectively to cater for his communicative purposes and
needs, as suggested by Bialystok & Hakuda (1994), who defined mastery with reference to
language use rather than usage (Widdowson, 1978).
It seems that learners with better access to the latent language structure should have higher
language ceilings. The key may be the possession of language awareness, defined as conscious
knowledge and appreciation of the linguistic system, its functions, registers and possibilities
Learning from learners‟ errors Ana María Rozzi de Bergel
59
for the expression of ideas, the apprehension and creation of knowledge and its manifestations
in literature of various genres.
This does not refer only to grammar, vocabulary and functions, but also to conscious
handling and understanding of humour, satire, beauty, rhythm, metaphor, symbols and more
subtle aspects of language, such as its tactical, aesthetic and artistic possibilities. This type of
awareness includes appreciation and understanding of foreign languages and the openness to
communicate with other cultures; according to the Association for Language Awareness, it is
the “explicit knowledge about language, and conscious perception and sensitivity in language
learning, language teaching and language use.” (¶ 1)
However, language awareness is both explicit and implicit, sometimes because the implicit
has become explicit through instruction; sometimes because conscious knowledge and
reflection on language and its role in human life have increased and sharpened intuitive,
implicit processes. The interplay of these processes resembles the relationship between
cognition and metacognition outlined by James (1999) in connection with language
awareness:
There are however two versions of LA. In fact these two versions of LA have been
with us for some time. The first kind, LA as cognition, works from the outside in, so to
speak: one first learns about language or something about a language that one did not
know before. You can stop here, in which case you have done some linguistics. Or you
can go on and turn this „objective‟ knowledge towards your own language proficiency,
making comparisons and adjustments.
This is to personalise the objective knowledge gained. The second variant, LA as
metacognition, works in the opposite direction: one starts with one‟s own intuitions
and through reflection relates these to what one knows about language as an object
Learning from learners‟ errors Ana María Rozzi de Bergel
60
outside of oneself. It is a case now of objectivising personal knowledge, not
subjectivising outside knowledge. This distinction can be reformulated as a distinction
between working at what one does not know and working on what one knows. I shall
refer to the first as Consciousness-Raising and to the second as Language Awareness
proper”. (p.102)
James limits access to true language awareness (LA) to native speakers, as he claims that to
develop LA, a person needs to have certain control of a language and have developed a set of
intuitions about it. He believes it is more appropriate to talk about Consciousness Raising, in
the case of foreign language learning:
Consciousness Raising is, unlike LA, for learners, who are not yet in command of
certain desirable knowledge, skills, and intuitions. I define CR as becoming able to
locate and identify the discrepancies between one‟s present state of knowledge or
control and a goal state of knowledge or control. CR gives the learner an equally
important but different insight into what he does not know and therefore needs to learn
if he is to achieve objectives. (p.103)
The distinction does not refer to the whole of two languages. As soon as the learner‟s
proficiency in a foreign language or in certain aspects of a foreign language reaches
satisfactory levels of fluency, appropriacy and correctness, the user can treat it as a native
language and begin developing awareness.
The fact remains that making knowledge about language rules and mechanisms explicit to
learners has always been acknowledged as an important learning resource. Ellis & Gaies
(1998) regard it as a tool for effective grammar learning and the issue seems to be whether we
should proceed from declarative to procedural knowledge (Johnson, 1996) or vice-versa, not
whether declarative knowledge has a place in the foreign language curriculum. Despite its
Learning from learners‟ errors Ana María Rozzi de Bergel
61
value, there is little evidence that declarative knowledge of language could actually help
learners overcome fossilization. Butler-Takana (2000) conducted an experiment with Japanese
learners of English, to explore if consciousness-raising could remedy fossilization. The
experiment was successful, but the learners participated in singling out the errors they wanted
to de-fossilize and knew the researcher‟s objectives. Without this degree of willingness on the
part of the learner, it is not clear whether consciousness-raising would actually stop or correct
fossilization.
With no conclusive proof that fossilization of interlanguage may be overcome, we may
claim that the personal curriculum has an end performance which constitutes the final
objective, the desired level of attainment of the curriculum owner and is his language ceiling.
Figure 1 – Measuring the development of learners’ interlanguage
as compared to
an ideal level of mastery,
in both languages.
a similar native
speaker‟s linguistic and
communicative
competence. (The end
performance in the
learner‟s L2 curriculum)
resembles
the extent to which the
foreign language
learner‟s linguistic and
communicative
competence in L2.
The stage of development of a
learner‟s interlanguage
is
resembling
the learner‟s level of linguistic and
communicative competence in L1
Hei
gh
t o
f th
eir
lang
uag
e ce
ilin
g
Learning from learners‟ errors Ana María Rozzi de Bergel
62
2.d Defining the personal curriculum
In Candlin‟s (Ibid.) definition, then, a foreign language course curriculum includes language
contents, a final objective and is taught according to an approach, with special materials. An
approach is the combination of a theory of language and a theory of learning, used as the
theoretical background to a method, which is a series of practical steps or procedures to learn
something (Richards & Rodgers, 1998). Theories of learning and theories of language,
approaches, methods and curriculum content are not the expert‟s exclusive domain, as each
learner seems to have a personal approach to language learning.
The learner knows, partly intuitively and partly consciously and rationally, what he needs to
learn in terms of lexis, grammar and function or genres and this determines the personalised
contents of his curriculum. We may also find a non-personalised part, comprising
communication universals all learners seem to find useful and worth learning, such as saying
the alphabet, saying numbers, telling the time and other fairly culture-independent functions.
The personalised contents will be more heavily influenced by cultural and psychological
factors.
Regarding the approach, its principles are largely unconscious to the learner and only
partially revealed through research, but each person has a personal conception of language
which works as his theory of language and a cognitive structure which determines his learning
hypotheses and therefore acts as his theory of learning. This is the learner‟s personal approach,
which provides the foundations for his learning method: the observable strategies and
procedures the learner‟s cognitive structure sets into motion.
The personal theory of language has to do with the learner‟s knowledge of his L1 – his
degree of language awareness and general level of proficiency, on the one hand, and on the
other, the role of language in his life: the value he attaches to precise and appropriate
Learning from learners‟ errors Ana María Rozzi de Bergel
63
communication, and his appreciation of language as a vehicle for humour, poetry, subtleties,
fiction, rhetoric, that is, whether he cares for going beyond straightforward and factual
communication.
The theory of learning has to do with the three central processes I have already explained
and will discuss in more detail further on: the learner‟s style for structuring his own language,
using conceptual or pre-conceptual classification and categorisation systems; his association
style: by proximity or by analogy, and the degree to which he uses his creativity and
imagination to build the system of the foreign language.
Thus defined, the notion of personal curriculum relates to Selinker‟s idea of a latent
psychological structure in the brain and the existence of a universal grammar of interlanguage,
with basic features belonging to a another universal system, which I have tentatively called
psychological or communication universals, plus elements common to a particular social
group and completely personal elements determined by the individual‟s construction of reality.
Learners may be unable to access the latent language structure, but they may access universal,
almost instinctively needed features of language, and use the more flexible part of the latent
psychological structure for the rest, the reason being that their construction of language and
mental representation of its uses has been affected and transformed by life experience and they
can no longer resort solely to a universal grammar to build up the system of a particular
language, as was the case with their mother tongue. They also need to use their socialised and
personalised language behaviour.
The meagre 5% of learners mentioned by Selinker (Ibid.) as having recourse to the latent
language structure might well be individuals whose education, culture and life experience has
enabled them to keep this access relatively intact, mostly by enabling them to acquire a high
degree of language awareness in L1 to which they have recourse when learning L2.
Learning from learners‟ errors Ana María Rozzi de Bergel
64
A curriculum is taught with teaching materials, usually selected or designed by the teacher or
the school authorities. Many schools tailor-make materials to cater for learners‟ special needs
in specific professional fields or use supplementary texts, videos or other sources of input,
sometimes provided by the learners themselves. These adaptations refer mostly to content and
are highly desirable, but an in-depth exploration of learners‟ learning hypotheses, styles and
strategies should be added to tailor-making, to best fulfil the learners‟ needs.
The management of a course is also part of its curriculum. The initial placement of the
learners, the total number of hours needed to cover the detected training gap, the number of
weekly contact periods, the size of the group, the time-table and the type of evaluation should
be tailored to the learners‟ needs but also to their possibilities and expectations. We should not
forget that learners also have a management style and expect to handle their learning
experience in particular ways: they have an idea of their placement which may or may not
agree with the results of the school‟s placement test, they normally know how many hours
they are ready to devote to reaching the level they aspire to; they have preferences regarding
learning in groups or in one-to-one lessons and they make choices regarding the frequency of
their lessons, enrolling in intensive or non-intensive courses. The exploration of the learner‟s
curriculum should take all these aspects into consideration.
Learning from learners‟ errors Ana María Rozzi de Bergel
65
Figure 2 – Main components of the personal curriculum
In the following section, I am going to discuss some of the components of the personal
curriculum and explain how, during our research, we attempted to unveil them.
Communication universals.
Personal communication needs.
Special language areas (ESP)
Language
content
Learner‟s conception of language.
Learner‟s appreciation of language.
Learner‟s acculturation in his own
culture.
Learner‟s proficiency and awareness
in L1.
Learner‟s style for structuring L1,
using conceptual or pre-conceptual
categories.
Learner‟s association style: by
proximity or by analogy.
Learner‟s capability for using
imagination and creativity to
advance into the construction of the
L2 system.
Learner‟s
“theory of
language”.
Learner‟s
“theory of
learning”. Lea
rner
‟s a
pp
roac
h
Chosen and/or designed by the
teacher.
Partly provided by the learner.
Materials
Learner‟s expectations.
Learner‟s fulfilled needs.
Learner‟s view of desired end
performance, if ceiling is reached.
Expected end
performance.
Total number of hours.
Time-table.
Size of the group/one-to-one.
Evaluation.
Course
management
Infl
uen
ced b
y p
sych
olo
gic
al, so
ciolo
gic
al, poli
tica
l an
d e
conom
ic f
acto
rs.
Learning from learners‟ errors Ana María Rozzi de Bergel
66
EXPLORATION AND REFLECTION
1. Selinker identifies positive language transfer or negative language interference as one
of the five factors affecting the formation of the interlanguage system. This may mean
that teachers who do not master their learners‟ native language have access to less
information about their interlanguage. How could they cope with this problem?
2. Consider the language proficiency of five native speakers of your native language. List
them and place the height of their communicative competence in L1 on a scale from 1
to 10. Account for your ranking: what factors have you considered for their placement?
3. Consider the case of a “good” learner and a “poor” learner. Is there any
correspondence between their level of language awareness in L1 and their achievement
in L2?
4. List three situations where you felt that the learners‟ curriculum did not match yours,
and that they wanted something else. What did they want and what were they getting?
5. In the three cases listed in 4, what did you do? Did you consider you knew better than
they did, and go on with your curriculum? Did you introduce changes to the course,
method or syllabus? What were the results?
Learning from learners‟ errors Ana María Rozzi de Bergel
67
3. Main components of the personal curriculum
Considering that a large part of the personal curriculum is not visible or even conscious,
teachers should be concerned with looking for its observable manifestations. When and how
do learners send us signals that they are focusing on their curricula or performing operations
which are personal-curriculum specific? Here, I will briefly discuss some of the observable
manifestations of the personal curriculum which we deemed available to us: the learners‟
communication strategies, compensation strategies, language needs, objectives and learning
hypotheses. In Parts 2 and 3, I will concentrate on a discussion of errors as the most revealing
manifestations of the personal curriculum, their interpretation and the classroom practices and
activities which may exploit them to accelerate progress and enhance learning.
The learner‟s personal curriculum sometimes resembles the course curriculum and
sometimes differs from it, originating a struggle between the two. The course curriculum is
available to the public and teacher and learners may consult, review or supplement it as
needed, but how can each person‟s curriculum be unveiled and followed? The external
manifestations of the personal curriculum are so many and so complex, its structure so
intricate and fluctuating, that we may reasonably expect never to answer this question in full.
We found it wiser to limit our exploration to some of its possible manifestations, rather than
attempt a comprehensive study of the phenomenon.
The learner‟s manifest needs and objectives, his communication style and strategies, his
learning hypotheses and style constitute the necessary background to the description of his
personal syllabus, and they can be partly disclosed through questionnaires or tests. The task of
deciding what types of observable classroom performance may be manifestations of the
Learning from learners‟ errors Ana María Rozzi de Bergel
68
personal curriculum, however, will necessarily be based on arbitrary choices, trial-and-error or
intuition, as not enough is known about this phenomenon. The problem does not concern areas
in the learners‟ curricula which match the course curriculum, but those which depart from it.
We may attempt to make their identification more reliable by using a general criterion, partly
derived from lesson observation, partly based on informed assumptions: we may define the
manifestations of the personal curriculum in the classroom as learners‟ contributions during
lessons which do not address the contents of the course, or the focal point in the lesson, or
follow the prescribed approach or method. They are expressed through questions, comments
or requests for explanations or clarification which seem to engage the learner‟s deepest
interest. However, the most powerful, useful and clearest manifestations of learners‟ personal
curricula is to be found in errors.
The general background information about the personal curriculum and a description of its
possible manifestations in the classroom or even before a learner joins a course are discussed
below.
3.a The learners’ needs and expectations
This is not a particular manifestation of the personal curriculum in the classroom, but its
general background, similar to the overall aims of a course or set of materials. The learner‟s
needs and expectations are normally explored before starting a course, at an interview or by
asking the learner to complete a form or answer a questionnaire. In Appendix 1, there is an
example of a protocol for such an interview and its accompanying questionnaire. The results
should be analysed vis-a-vis the school‟s or the teacher‟s appraisal of the learner‟s entry level
within the programme of study (Beginner, pre-intermediate, intermediate, post-intermediate,
Learning from learners‟ errors Ana María Rozzi de Bergel
69
advanced) and the degree of feasibility of the learner‟s goals, particularly in what concerns the
time-learning gap relationship, with reference to five main questions:
a) Is the learner realistic in his appraisal of his present level? In his expectations of
achievement? In his measurement of the time he will need to reach his desired level?
In other words, is he able to measure his training gap?
b) What is his previous history as a learner of foreign languages/English?
c) To what factors does he attribute success or failure? How much responsibility is he
willing to assume for his learning?
d) What is his attitude towards learning this language?
e) Does he hold the culture he attributes to this language in high regard? Does he reject
it?
If the answers to these questions show that the learner is realistic in his expectations, can
describe his needs in a concrete fashion and is willing to commit himself to learning, his
personal curriculum may or may not agree with the course curriculum, but is correctly
addressed at the desired objectives and will most probably lead the learner to achieve his
goals.
3.b The learner’s communication style and strategies.
Are errors communication strategies, or manifestations of communication styles? Although
some communication styles and strategies favour the appearance of certain types of errors or
underlie them, they are not errors and in fact, appear in perfectly correct utterances as often as
in erroneous ones. This departs from the conception of communication strategies as resources
for overcoming limitations in the mastery of a foreign language, which equates
communication strategies with error styles, two concepts which are best kept separate, the
Learning from learners‟ errors Ana María Rozzi de Bergel
70
former depending more on the person‟s way of encoding and transmitting messages and the
latter relating to methods for system-formation.
3.b.1 Discussion and definition.
We will refer to communication styles as the ways and means by which people issue and
interpret messages using language, how they interact with the social medium through language
and generally perform as communicators. We will define communication strategies as
individual, identifiable, observable plans or techniques for conveying and negotiating
meaning, arising from particular communication styles.
From this perspective, styles and strategies reveal people‟s particular way of making contact
with their fellow human beings, their degree of reliance on language for communication and
the type and range of concepts they use in their interactions. This is mainly conditioned by
their cultural level, schooling and degree of language awareness, as well as by cultural,
psychological and even age factors on which human beings base their conceptual
representation of the world and their relative position in society.
Communication styles and strategies thus defined are not acquired by adults at the foreign
language class, but have developed through the person's life experience. At the EFL class, it
has been suggested that the problem of adult language learners is more textual than discoursal
(Widdowson, 1979). Their problem is to textualise discourse into the foreign language, and
rather than change the characteristics of their communication strategies, they will simply try to
acquire the necessary language elements to use them. I will claim that the scope and type of
their personal discourse and the communication strategies which characterise it are identical in
any language for any one person, so adult learners will not be likely to modify them at the
Learning from learners‟ errors Ana María Rozzi de Bergel
71
EFL course. Besides, they ought to re-live their lives, change their view of the world and of
interpersonal relations, or engage in specific instruction, to bring about this change.
It is not too bold to suppose that there may be communication styles and strategies which
correspond to a high language ceiling and others which are typical of low ceilings. By the
same token, we may expect a certain correspondence between communication styles and
strategies and error types. It is very common to see adult learners fail at EFL courses because
they are limited as communicators, although this situation is regarded as failure by language
teachers more often than by learners. Poor communicators are frequently content with learning
just enough to satisfy their limited communication needs, an issue we will address when we
discuss the pedagogical implications of a low language ceiling.
If communication strategies exist in adults before they begin learning a foreign language,
they can hardly be considered attempts at using this particular language or at coping with
problems of interlanguage communication, as has been suggested by Tarone, Cohen & Dumas
(1983). They produced a novel analysis of errors as the product of communication strategies
used by language learners to cope with limitations in their command of the foreign language.
They claimed that learners' communication strategies were only observable when errors were
made, a highly debatable concept which might suggest that there is practically no distinction
between how messages are emitted and received, and how language limitations are overcome.
During our exploration of learners‟ communication strategies and styles, we observed that,
with errors or correctly, the person who uses the strategy of message abandonment (Tarone,
Cohen & Dumas, Ibid.), for example, will leave his sentences unfinished in any language, in
utterances such as: "Well, I did not like the film and ..... well, you know." If his English is poor,
he will probably say " Well, I not liked the film and ........ well, you know ", but his
Learning from learners‟ errors Ana María Rozzi de Bergel
72
communication strategy will not change: it consists of leaving the final part of many messages
in the interlocutor‟s hands, as if everybody was supposed to know what he meant and this
would make it unnecessary for him to say it or as if completing the meaning was the listener‟s
responsibility.
My disagreement with Tarone, Cohen & Dumas (Ibid.) arises from results of our research
which suggested that messages are not abandoned because the learner does not know how to
continue or complete them, but because this is the way in which he usually frames his
messages – a communication strategy he will use in any language, including the mother
tongue.
Cohen (1998) insists on the approach to communication strategies already discussed, when
he defines them as
“…….those processes which are consciously selected by learners and which may
result in action taken to enhance the learning or use of a second or foreign
language, though storage, retention, recall, and application of information about the
language” (p.4)
Færch & Kasper (1983) establish a distinction between strategic and non-strategic
interlanguage use and hold that
“describing communication strategies in interlanguage communication is
the same as describing the strategic use of interlanguage systems"(p.xviii)
It is not very clear how language use might not be strategic. A broad definition of strategy
would consider it an operation performed by the language user in order to attain his goals, and
no instance of language use is purposeless, so we might conclude that language is a strategic
instrument per se.
Learning from learners‟ errors Ana María Rozzi de Bergel
73
Therefore, the person‟s strategic use of the first language system can also be described, and
this description would not be related to error production. Would the concept of “strategic use”
be different for the two languages? The processes defined by these authors, however, provide
insight into important aspects of interlanguage and learning, such as the distinction between
production strategies, learning strategies dominated by "the desire to learn the target language
more than the desire to communicate" (Tarone, Cohen & Dumas, Ibid.) and the so-called
communication strategies, a concept which may, however, lead to the more pedagogically
appropriate idea of "compensation", meaning the output of language which the learner,
because of the amount of instruction or input received, cannot possibly produce accurately and
appropriately. The mechanisms used to make up for this lack of resources reflect the learner‟s
compensation strategies.
Later, research into learning strategies undertaken by authors such as O'Malley & Chamot
(1990) and, among others, drew a better distinction between learning strategies and
communication strategies, and although they did not include data from error analysis in their
studies beyond the provision of examples, it was obvious that the analysis of the learners'
production and behaviour had provided the basis for their work. O‟Malley & Chamot (Ibid.)
define a learning strategy as marked by the desire to learn the target language rather than the
desire to communicate, but agree that communication strategies emerge “from failure to
realise a language production goal”. Their comprehensive description of learning strategies,
grouped into metacognitive, cognitive and social-affective, was not aimed at differentiating
them from other processes, but did so by showing that these strategies had their own domain.
The problem of separating error production from communication strategies, however,
remained unsolved and in fact, the O‟Malley and Chamot agree with Faersch & Kasper (Ibid.)
that
Learning from learners‟ errors Ana María Rozzi de Bergel
74
…communication strategies may be a „psycholinguistic‟ solution to the
communication problem instead of one which relies upon negotiation of
meaning (p.45).
In Oxford‟s (1990) classification of learning strategies, we also find “social language
learning strategies”, which are not attempts to develop linguistic competence but negotiations
of meaning such as asking for clarification or information. According to this, a learning
strategy is used to construct the language system and a communication strategy, to convey
meaning. The distinction is rather complex, not clear enough and we might even find that the
two categories overlap, as a communication strategy such as asking for clarification could well
be used to better grasp certain parts of the language system. We are faced, then, with two
aspects of communication strategies: the cognitive and the interactional. Are they cognitive
processes of the speaker or an exchange among interlocutors to get their meaning across? If
we agree on a distinction between learning and communication strategies we will find that
only in the second case can we come across processes we may legitimately call
communication strategies.
The definition of communication strategies I am going to use is based on Bialystok &
Kellerman (1987):
...the manipulation of either a semantic concept or a language or both in
order to express particular intentions” (p.169)
I will also adhere to their idea that there are no differences in the processes used by native
and non-native speakers when using communication strategies, beyond language proficiency
and age.
Learning from learners‟ errors Ana María Rozzi de Bergel
75
The exploration of communication styles and strategies we are going to undertake is based,
then, on the following tenets:
a) There is no distinction between interlanguage communication styles and strategies of
adults and their native language communication styles and strategies.
b) Communication styles and strategies develop in the native language, and learners carry
them over into any foreign language regardless of their degree of proficiency in this
language. They are the basis of the personal curriculum and often mark the height of
language ceiling. In a rather over-generalised fashion, we might claim that people will
learn as much as they need in order to use their communication styles and strategies in
those areas and genres where they wish to use them.
c) Styles and strategies which do not exist in the native language do not develop in
the foreign language, unless the learner makes a conscious effort in this sense. When
this happens, styles and strategies should change in both languages, with the foreign
one influencing the native tongue. (Cook, 2003)
3.b.2 Favourable and unfavourable communication styles and strategies
Communication strategies may have an influence on the height of language ceiling. We may
assume that learners will tend to lose interest in learning when they begin communicating in a
manner they consider satisfactory. This degree of satisfaction varies according to how
sophisticated a communicator a person is, so in a way the quality, of his communications
determines the degree of his ambitions and his needs. In the learner‟s personal curriculum,
communication styles and strategies account for a large part of the functional content. A poor
communicator will be satisfied with less and reach his ceiling at an earlier stage in his mastery
of a foreign language.
Learning from learners‟ errors Ana María Rozzi de Bergel
76
On the other hand, despite their role in determining the kind of end performance in the
personal curriculum, communication styles and strategies do not account for error production,
neither can error analysis be used to detect them. An error is a sign that, when placed in a
communicative situation, the learner will display his command of the language to cope with
the demands of communication and that in so doing, will also resort to rules, concepts or
compensation mechanisms present in his interlanguage. This will shed light on inner processes
used to construct the language system, which will be the same he uses to learn, as they derive
from his learning hypotheses. Communication strategies belong, in this view, to a different
domain, having more to do with relational and psychological aspects of the individual as a
social being.
Communication styles are mostly personal, and any classification is bound to be incomplete.
It is not my aim, here, to produce such a classification, neither will I set out to duplicate those
already available. In the process of studying the possible bearing of communication styles and
strategies on the height of language ceiling, my research identified two broad categories,
which we termed “favourable” or “unfavourable” for lack of better names. Our exploration
was by no means complete or conclusive, but enough evidence was gathered as to attempt a
tentative definition of these communication styles and strategies.
a) Styles which reflect an awareness of social, aesthetic and tactical uses of language and
a reliance on linguistic codes as means of communication, such as
a.1) Use of various degrees of formality, according to the perceived
role-relationship.
a.2) Predominance of linguistic over non-linguistic forms of
communication. More is communicated with words than with
gestures, for example.
Learning from learners‟ errors Ana María Rozzi de Bergel
77
a.3) Conciseness.
a.4) Precision.
a.5) Humorous and/or artistic-aesthetic use of language.
a.6) Message adjustment to respond to ongoing stimuli during
conversational exchanges.
These styles usually aid language learning and may be considered “favourable”.
b) Styles which make use of a very narrow range of strategies, language, registers or
aesthetic possibilities or tend to replace language by other, non-verbal codes of
communication. We found these styles in learners with very low language ceilings, not
reaching beyond pre-intermediate or intermediate level. Because of this, we will
consider them “unfavourable”:
b.1) Soliloquy
The person cannot listen to his interlocutor and adapt his messages to the
ongoing conversation. When the others talk to him, he waits politely until they
finish and then goes on with the idea he has interrupted. These learners often
have problems of listening comprehension and their phonological capabilities
may not be very high. They may be, however, good readers.
b.2) Egocentrism
The communicator only talks about personal experiences or passes all topics
through the filter of his personal experience. He cannot discuss any subject that is
not within his day-to-day life experience, and conceives of the world as a highly
limited space where he includes only what is within his eye span. He often has no
make-believe capabilities or sense of humour, and tends to be a bad actor or
impersonator. A variation on this is the learner who does not listen or interact
Learning from learners‟ errors Ana María Rozzi de Bergel
78
because he is extremely self-centred and vane. This is sometimes the case with
successful people who have lost their humility.
It is interesting that Vygotsky (Ibid.) considered that language appears as an
external, social manifestation, then becomes ego-centric and finally reaches the
stage of inner speech, which is one of the higher mental processes. In his
conception of ego-centric speech he disagrees with Piaget, although the latter
held that Vygotsky had misunderstood him, clarifying that he did not think that
cognitive egocentrism arose from the individualism which precedes socialisation
but from a lack of differentiation between the personal point of view and those of
others (Piaget, 1962). Piaget describes egocentric speech as a succession of
monologues in conversation, with no strategic adaptation to the interlocutor‟s
stimuli.
This egocentric speech is typical of the period of concrete operations,
dominated by pre-logical or pre-conceptual thinking, where groupings and
categorisations of concepts are carried out according to functions, external
properties and uses of objects rather than by their essential characteristics.
Piaget‟s descriptions of stages of development have not been seriously
questioned, but the boundaries between them have been found not to be so clear
as he set them (Vygotsky, Ibid.) and further studies have demonstrated the
enormous influence of the social medium on human development, also
acknowledged by Bruner (1990).
Another important questioning of Piaget‟s stages of development has come
from the discovery of broad categories in children‟s conception of language, like
calling all four-legged mammals “dog”. These too comprehensive categories
Learning from learners‟ errors Ana María Rozzi de Bergel
79
point to conceptual generalisation and would show that children are capable of
conceptual thinking from an early age. However, despite all these new
developments in our knowledge of cognitive evolution, the main concepts
outlined by Piaget remain valid and in the case of egocentric communication
styles, the fact remains that this style is common in learners whose language
ceiling appears to be low and seems to point to lacks in the conceptual system.
c) Literality
Nothing may have a fictional or humorous meaning. The person cannot use
words figuratively, and does not understand metaphors or similes. If told that
"every cloud has a silver lining", he will probably say he cannot confirm this
because he does not know very much about natural phenomena. If he hears that
somebody has "piles of money" he will ask exactly how high these piles are and
where they are. These people often react angrily to humour and irony, which they
find insulting. This points to limitations in the field of metaphor, a vital domain
for mastering a language. As for humour, it has been listed by Vygotsky
(1930/2000) as one of the higher mental processes.
More than a communication style, literality represents the impossibility to
move in the domain of imaginary or figurative meanings. The learner is relatively
unable to role-play situations, to impersonate characters or to enjoy fiction, let
alone move in the metaphoric domain, which accounts for a large part of
language use.
d) Duplication of the messages
The person verbalises his message and mimes or draws it in the air at the same
time. "Susan was here (points at a napkin in front of him) and Tom was there
Learning from learners‟ errors Ana María Rozzi de Bergel
80
(points at a jar of water on the table) Susan approached him (moves the napkin
towards the water jar)......", etc. "My book was this thick" (gestures with his hand)
A lot of these communicators constantly "speak with their hands" when they talk,
not clarifying but duplicating the message. Note that in the classroom, this is
sometimes favoured by teachers who overuse the technique of joining language
to body or facial expression to be better understood, a practice they often pass on
to their students, particularly, if they are small children. While observing lessons
in the first years of primary school, it is common to witness storytelling by
teachers who might well not say a word to convey the meaning of the story they
are telling, as they clearly duplicate all their messages with mime or gesture.
e) Use of only one or two levels of formality
The person uses semi-formal and informal language, but never other registers,
which they ignore as useless or not very operational. The disappearance of more
formal levels is a narrowing of the linguistic field and consequently, of the
conceptual domain.
f) Preambling.
Nothing is said without an introduction, such as “Let me tell you …”, “I would
like to say…”, “It is important for me to say ….”. The person permanently
announces what language function he is going to perform, or asks for permission
to perform it, or tells his audience what type of reaction he expects. “This is
going to surprise you …”, ”Let me give you another example..", "I would like to
point out that ..", "Let us discuss this .....", "I am going to tell you an anecdote
.....", "I know I am going to bore you, but at the risk of being repetitive I would
Learning from learners‟ errors Ana María Rozzi de Bergel
81
like to mention what happened the other day ....", “I would like to ask you a
question”, etc.
This seems to show lack of confidence on the interlocutor‟s capability for
understanding what the speaker is going to say, or his intentions, or perhaps lack
of reliance on language as a means of expression, which seems to be so limited
that it has to be permanently explained.
g) Overuse of detail.
The person does not distinguish main from additional information, has no
power of synthesis and no priorities. If this communicator wants to tell his
audience that he has bought a new car, he will also include a detailed and totally
irrelevant description of the dealer's shop, the clothes the employee was wearing,
how he got to the dealer‟s, what the weather was like, etc. This seems to point out
problems with organising elements hierarchically or finding boundaries between
semantic domains.
h) Vagueness
Messages are not precise, either because they are not syntactically clear or
because they contain umbrella words such as things, stuff, folks used to avoid the
use of words belonging to more restricted categories: "Bring me that stuff", "I
think things are sort of wrong here", "Folks like that kind of thing", "When I like
somebody I feel something inside, you know, that kind of thing", “Shit happens”,
etc.
I will not attempt a classification of communication strategies for the same reasons I
explained in connection with communication styles. I cannot give a comprehensive list of
“favourable” strategies, either, as they are related to the expression of favourable learning
Learning from learners‟ errors Ana María Rozzi de Bergel
82
styles and easy to infer from their description. I will simply describe two of the strategies we
more often found in learners with low language ceilings, which can be considered
“unfavourable”:
a. Message abandonment
"I got up this morning, looked out of the window and ............" The person drops the
message altogether. It seems as if he had lost interest in the idea or as if he took it for
granted that his interlocutor will understand what he means and mentally provide the
missing part. He does not take responsibility for conveying the whole message or try to
replace it for mime or gesture.
b. Message substitution
The same as before, but the last part of the message is replaced by a gesture, an
onomatopoeia, an exclamation: "I slipped on a banana peel in the yard, and ......paf!"
It is important for the teacher to detect the learner's communication styles and strategies in
his native language and exploit them at the language course. The learner‟s performance during
lessons will reveal them, but by the time the teacher is in a position to carry out an analysis of
the learner‟s communication strategies he will have been attending the language course for
several weeks and decisions about the curriculum and the materials will have been made.
Therefore, we are proposing an analysis of the learner's communication style and strategies at
the placement interview, in the manner outlined below, provided that the interviewer speaks
the learner‟s mother tongue. Unfortunately, if this is not the case, the teacher will simply have
to observe the learner during lessons and infer his communication style and strategies.
Learning from learners‟ errors Ana María Rozzi de Bergel
83
3.b.3 Communication styles and strategies analysis
The following procedure was used during our research and is suggested for a quick, informal
analysis of a prospective learner's communication strategies:
a) The candidate is asked to tell a story, an anecdote, an accident, etc. and to report a
conversation. There should also be a brief discussion of local and world events, and he
should be asked to give opinions. This gives the interviewer some insight into the
person's cultural awareness and general knowledge and exposes his general
communication style and communication strategies in several domains: narrative,
reporting, giving opinions. Care should be taken to discuss just the news of the day or
matters known by any informed citizen in a casual, natural fashion, not as a question-
and-answer session or a general knowledge quiz.
b) Beginners carry out the interview in their native language, but with intermediate or
advanced learners the recommended procedure is as follows:
b.1) The conversation is held in the learner‟s native language and in English,
alternatively.
b.2) The candidate is allowed to answer in either language, but the interviewer
should make sure the learner uses both in the course of the interview.
b.3) Some topics dealt with in one language, are brought up again in the other a
bit later on in the interview. This is done to check whether the same strategy
is used in both languages.
We will discuss an example.
Learning from learners‟ errors Ana María Rozzi de Bergel
84
Example (Transcription of a fragment of an actual interview):
Candidate: My dog had sick. (Meaning: My dog was sick)
Interviewer: What happened to your dog?
Candidate: I had that take him to the vet, because he had sick. (Meaning: I had to take him
to the vet, because he was sick. )
Interviewer: But what was wrong with him?
Candidate: I had that pay a large bill. (Meaning: I had to pay a large bill)
Interviewer: Is your dog all right now?
Candidate: Yes. My wife was very angry with me.
Interviewer: Bueno, pero no era culpa suya. (Translation: Well, but it wasn‟t your fault)
Candidate: Ella detesta los animales y siempre dice que no quiere tener perro. (Translation:
She hates animals and always says that she doesn‟t want to have a dog)
Interviewer: Qué raro.... Los perros son muy lindos. ¿Qué le pasaba al suyo? (Translation:
How strange ... Dogs are very nice. What happened to yours?
Candidate: Lo llevé al veterinario y me cobró un montón de plata y mi mujer se puso
furiosa. (Translation:I took him to the vet and he charged me a lot of money
and my wife was furious)
Our candidate never actually answers his interlocutor – a style I have called soliloquy, and
on the other hand, he shows traces of egocentrism, as he finds it difficult to talk about a topic
unless it is through the filter of his personal experience or feelings regarding the issue, which
even take precedence over the topic, as in:
Learning from learners‟ errors Ana María Rozzi de Bergel
85
Interviewer: But what was wrong with him?
Candidate: I had that pay a large bill. (Meaning: I had to pay a large bill) (He does not
answer the question about the dog but concentrates on talking about himself)
The question about the dog‟s condition is repeated four times and never acknowledged, let
alone answered, and this happens in both languages, at different levels of proficiency.
The error in "I had that pay a large bill" is a feature of the learner's grammar, not a
communication strategy. He structures this utterance on the model of Spanish, probably
because he has not been able to grasp the difference between different uses of the verb to have
and have to.
The other error, "My dog had sick", is a confusion between to be and to have. The learner
provided a very interesting authoritative explanation of this error, which we would not take for
a communication strategy: To be, in English, is often used in expressions where tener (To
have) is used in Spanish. ("I am twenty years old" is equivalent to the Spanish "Yo tengo
veinte años", and in expressions such as "I am cold", the translation is also "Tengo frío" and "I
have a cold" is "Tengo un resfrío" or "Estoy resfriado") The learner had come to confuse to be
and to have because of his overuse of translation as a learning tool.
Language errors occur in the foreign language and are related to the learner‟s degree of
mastery, whereas the communication style depends on personality factors and remains the
same in both languages. Errors tell us how and what he is learning, his communication style
tells us what and how he is prepared to communicate.
Learning from learners‟ errors Ana María Rozzi de Bergel
86
3.c The learner’s learning hypotheses, styles and strategies
When reading the many inventories of learning strategies, we feel there must be some
general, unconscious processes ruling them. It is not possible to describe strategies as
behaviour without wondering what produces this behaviour. I am going to propose three levels
of analysis: observable learning strategies which are part of general learning styles determined
by the learner‟s learning hypotheses. Error analysis discloses strategies; by grouping and
describing them we can have access to learners‟ learning styles but most importantly, we can
get a glimpse of their learning hypotheses.
3.c.1 Learning hypotheses
The concept has already been introduced, but we I clarify it further here. We will call
learning hypotheses the processes and operations which the mind sets in motion for
constructing the system of the foreign language during the learning process. The specific
characteristics of these mental processes are personal, as each human mind chooses what it
considers “has to be done” to achieve the final goal of structuring the system. Learning
hypotheses depend largely on each individual‟s cognitive development and on his structuring
of the native language.
These learning hypotheses, which could be compared to the process of selecting tools and
procedures to carry out a task, are not independent from social, affective or idiosyncratic
factors and are also influenced by education and the social context.
If we accept that learning a language entails constructing a system, as we have discussed,
then the central processes should be categorisation, association and distinctions between
Learning from learners‟ errors Ana María Rozzi de Bergel
87
elements and categories, as well as the generation of novel connections or creations within the
system, so we will centre our attention on how learners
a) Place concepts into categories, using conceptual or pre-conceptual systems of
classification. For example, to place the concept chair into the furniture category, a
learner may define it conceptually, with a dictionary-style definition, such as “A seat
which usually has four legs and a back and is only for one person”, or may define it
pre-conceptually, “The things we put around the dining-room table”, “An object
where we sit”, “My favourite kind of seat”, “A comfortable seat”, etc. Note that the
pre-conceptual definitions are based on external characteristics, uses or location of the
object and do not attempt to define what the object is.
Conceptual categorisation has been discussed by Bruner (Ibid..) as essential for learning
and although fuzzy logic (Zahed, Ibid.) has questioned that categories might be clearly and
objectively defined, it remains a central concept in all cognitive models. In language
learning, it is the capability needed, for example, to distinguish nouns from adjectives or
auxiliaries from main verbs. Is does a third person singular auxiliary for the present
simple, interrogative or negative, or is it “the word we put when we talk about „he‟ or
„she‟?” Note the implications for learning of using a conceptual or a pre-conceptual
category for categorising this word. If the learner uses the pre-conceptual category, he will
tend to use does whenever he refers to he or she, independently from time or aspect
references. This could lead to errors such as “He does is American”, “She does liked the
film”, and similar utterances.
b) Associate concepts, either by analogy, (metaphor) or by proximity (metonymy).
Metaphor is associated to the paradigmatic axis of language (De Saussure, Ibid.) and
Learning from learners‟ errors Ana María Rozzi de Bergel
88
has to do with synonyms, different ways of expressing similar ideas, etc., whereas
metonymy is useful to move along the syntagmtic axis: understanding that after a subject
we need a verb, then perhaps a direct object, that adjectives precede nouns in attribution,
etc.
c) Use their imagination and creativity to advance in the structuring of the system,
opening up a series of creative associations, or just using closed, self-contained
systems. For example, if asked to use several related words found in a particular
context to create a new text, does the learner simply seek to produce a variation of the
original context, or does he use the words in a new one, connecting them to previously
acquired knowledge, thus exploring more possibilities to exploit his resources?
During our research into the role of errors in the learner‟s personal curriculum, the
psychologist in our team developed a Learning Hypotheses Test which appears in Appendix 2.
This test is currently used by a number of teachers and has proved extremely reliable for
diagnosing learning hypotheses, with a margin of error of 0.6%.
I will frequently refer to the correspondence between error styles and the different learning
hypotheses the test helped us to disclose and understand, in order to clarify the role of
conceptualisation, association styles and creativity in foreign language learning.
3.c.2 Learning styles
Although the concepts of learning hypotheses, learning styles and learning strategies are
connected, they should be clearly differentiated. Ausubel (1968) defines cognitive styles or
learning hypotheses as “self-consistent and durable individual differences on general
principles of cognitive organisation(p.171), while the learning hypotheses explored by Piaget
& Inhelder (As cited in McCarthy Gallagher, J.& Reid, D.K. 2002) are structures and
Learning from learners‟ errors Ana María Rozzi de Bergel
89
cognitive processes present in the mind, which represent ways of approaching the processing,
apprehension and structuring of knowledge. Garger and Guild (1984) describe learning styles
as “…stable and pervasive characteristics of an individual, expressed through the interaction
of one‟s behaviour and personality as one approaches a learning task” (p.11). A learning
style is a general preference for certain operations or ways of learning whose origin or internal
roots have more to do with personality aspects which are also a part of the learning
hypotheses. According to Kinsella (1995) they refer to an individual‟s natural, habitual,
preferred ways of absorbing, processing and retaining new information and skills, which
persist regardless of teaching methods and content areas. A learning strategy is a conscious
act, a procedure or action taken to achieve an objective, usually having to do with problem
solving: how to remember vocabulary, how to develop reading comprehension techniques, etc.
Different learning styles, which complement learning hypotheses, lead the individual to
choose certain learning strategies. Learners will not be able to identify their learning
hypotheses or learning styles without some guidance from a professional, but they will always
know what they DO or what they PREFER to DO, what strategies they use for learning, and
their counterpart: what they prefer the teacher to do. Questionnaires which help the teacher
find out the learners‟ learning styles and strategies also shed light on their inner learning
hypotheses. Reid (1995) provides interesting models of such questionnaires and surveys. We
are excluding from our discussion the diagnosis of auditory, kinaesthetic, spatial, visual
learning styles, because that classification is only concerned with the way information is
captured and retained, not with how it is processed. That classification complements the
exploration of learning styles and it is very useful for the teacher.
Are errors learning strategies? The British Encyclopaedia (2002) defines a strategy as “the
art of devising or employing plans or stratagems to achieve a goal”. This definition implies
Learning from learners‟ errors Ana María Rozzi de Bergel
90
that a strategy is a conscious act and also that, as it is created according to the objective, there
must me as many as objectives are pursued. Tarone, Cohen & Dumas (Ibid.) define learning
strategies as operations that learners carry out to develop linguistic and socio-linguistic
competence. The motivation for the creation of the strategy comes from the desire to learn, not
from the desire to communicate, which would turn it into a communication strategy.
O‟Malley & Chamot (1990) have classified learning strategies into three categories:
cognitive, having to do with the processing, manipulating and storing knowledge; socio-
affective, concerning the learner‟s interaction with other members of his group, the way he
asks for clarification and his “inner speech” (considered by Piaget and Vygotsky as a higher
mental process and not conscious at all, however); and metacognitive, which relate how the
learner plans, monitors and assesses his own learning. Further classifications have been added
in recent years, mostly as a result of research into brain functions and of a renewed interest in
the role of the social medium in learning processes. I would like to discuss one of them as an
example: Kolb‟s (1984) experiential learning diagram, expanded by MacCarthy (1995) to
explain learning styles, which relies heavily on Piaget‟s (Ibid.) theories, the Gestalt theory and
more specifically, the experiential learning cycle by Lewin (1951) and Dewey‟s (Ibid.)
concept of learning by doing.
Kolb holds that the unique and particular characteristics of individual learning styles arise
from each person‟s way of perceiving and processing information. His research into learning
styles starts with and examination of the degree of emphasis people place on the four steps in
the learning process: concrete experience, observation and reflection, abstract
conceptualisation and active experimentation. He combines two dimensions of perception:
concrete experience and reflective observation; with two dimensions of processing: abstract
conceptualisation and active experimentation. With each stage of active experimentation a
Learning from learners‟ errors Ana María Rozzi de Bergel
91
new concrete, enriched experience occurs, which originates a new cycle of continuous
spiralled learning, an idea also shared by Bruner (Ibid.)
Figure 3 – Kolb’s learning styles in his experiential learning model
Accommodator
Learning from feeling
(Concrete Experience)
Diverger
HO
W W
E
P
ER
CE
IVE
Learning by doing
(Active experimentation)
Learning by watching
(Reflective observation) HOW WE PROCESS
Learning by thinking
(Abstract
conceptualisation)
Assimilator Converger
Learning from experiences
Relating to people.
Sensitivity to feelings.
Ability to get things
done.
Risk taking.
Influencing people and
events through action.
Careful observation before
making a judgement.
Viewing things from different
perspectives.
Looking for the meaning
of things.
Logical analysis of ideas.
Systematic planning.
Acting on an intellectual
understanding of the situation.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
93
Kolb‟s classification, like many others, clearly shows that a learning style is the description
of external, observable manifestations of deep processes (Reid, 1995). These classifications
are the map, not the territory.
Errors are also external manifestations of inner processes. They are the result of certain
learning strategies but they are not, in themselves, learning strategies. They do not only expose
strategies but also shed light on learning styles and hypotheses, which makes error analysis a
valuable tool for exploring the learners‟ unconscious cognitive structure. The learner
producing “He cans goes with me”, due to an overgeneralisation of the rule of the third person
singular verbs in the present simple, is perhaps showing a strategy of drawing general rules by
examining specific cases, which reflects the learning style of the assimilator, in Kolb‟s
classification, and shows he can place concepts into categories using conceptual classification
systems, but the categories being used are not sufficiently restricted2.
In this case, the teacher ought to remember that this learner needs help in examining and
remembering rule restrictions, which has to do with his learning hypotheses, not with the
strategy applied. Thus, insightful error analysis may provide information about the deepest
layers of the learner‟s cognitive processes and cognitive structure.
2 This interpretation, of course, is only valid as an example here, as we ought to know a lot more about the learner
and his interlanguage to venture such a diagnosis.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
94
Figure 4 – The processes of error production and error analysis
3.d The learner’s method
The learner‟s observable strategies and procedures, plus his attitude towards learning, lead
him to develop a method for learning which does not always agree with the procedures and
methods proposed by the teacher. Learners‟ behaviour during lessons and their remarks about
the course, the pedagogy, the teacher and the materials should be recorded and analysed, as
supplementary information to understand the processes observed through communication
strategies analysis, error analysis and the diagnosis of learning hypotheses. Standard methods
Error:
He cans goes with you.
resulting from
drawing general rules
from concrete cases. derived from
learning by watching
and thinking.
(Assimilator)
showing
an ability for using conceptual classification systems in which the categories,
however, are too broad and lead the learner to over-generalisation.
Implications for teaching: the learner needs help to find rule restrictions and
exceptions to rules, by being exposed to many contexts and communicative
situations where the rule does not apply, as well as those where it applies.
However, he will profit from inductive processes and can be trusting for
accessing rules on his own.
Full arrows show the
process of error analysis.
Dotted arrows show the
process of error production
Learning from Learners‟ Errors Ana María Rozzi de Bergel
95
of detecting learning styles and strategies, such as the questionnaires already suggested, may
be used here to complete the description of personal “methods” or understand them better.
Chapelle (1995) describes the procedures and methods used by several learners to support her
classification of learners into field dependent and field independent. Perhaps the analysis
could be taken further, to describe their learning hypotheses. Field-independent learners seem
to reach abstract conceptualisation more easily and consequently, may be faster learners.
When there is an overt disagreement between the teacher‟s and the learner‟s method, it is
usually the latter which prevails, unless the teacher manages to thwart it to corset the learner
into the course curriculum, usually with poor results. In this case, more attention is often paid
to training learners in the “correct” methods and procedures than to learning, and achievement
is usually measured by how well the learner adjusts to classroom conventions, developing
competencies which are very valuable for surviving …. in the classroom. Acknowledging the
existence of the learners‟ methods and respecting them usually means doing away with many
of our “cherished notions” (Pid Corder, Ibid.) about what should be done in order to learn. It
also means correcting teachers‟ wrong assumption that being professionals means being able
to tell people what to do, rather than interpret what they want to do and support these
processes with advice and encouragement.
3.e The learner’s questions and their role in the personal curriculum
During our research, we carried out an experiment with a group of learners, which was then
replicated with other groups, to explore whether learners‟ questions may be considered
manifestations of their personal curricula. The research team hypothesised that if they were,
they should be answered as they were asked, without delays or re- formulation, and this
Learning from Learners‟ Errors Ana María Rozzi de Bergel
96
treatment of questions would accelerate progress, because it would address the personal
curricula, which we regarded as more important than the course curriculum. Some
experimental groups were chosen, where teachers were instructed to answer questions in this
fashion. Groups of students at similar levels of attainment were also set as control groups, and
after six months, all the groups were given the same tests, to assess their progress, which
turned out to be dramatically higher for the experimental groups.
In the experiment, some variables were only formally controlled; for example, all the groups
were supposed to be at the same level of communicative competence, but of course, individual
differences persisted. This may have influenced the reliability of the results, but the teachings
derived from the experience were extremely useful and could well be used as the basis for
further, deeper research:
a) Learners‟ questions which seem to be above and beyond the level of the course or
which are asked in class “out of the blue” may be manifestations of their personal
curricula and they also show that these curricula are not identical to the course
syllabus. It is common for teachers to avoid answering these questions or to tell the
learner that the topic will be dealt with at a later stage. The personal syllabus may well
contain a completely different order of presentation of language items, which is usually
determined in course design by giving priority to highly productive and frequent
language forms or functions. The degree of productivity and frequency of language
forms has been determined by linguistic research and is the product of statistical
studies, but some learners may depart from the profile of the average language user
and have their own criteria for the order of presentation of language items.
b) Questions about broad language areas which are usually taught gradually, in
progressive degrees of difficulty. "How many auxiliaries are there in English? How do
Learning from Learners‟ Errors Ana María Rozzi de Bergel
97
they work?", "How can I connect sentences in English?" Gradualism is a pedagogical
concept created to make subject matter more “teachable” and does not necessarily
reflect real cognitive processes. Together with the order of presentation of language
items, it is perhaps the “cherished notion” which requires more reviewing in
curriculum design. Gradualism may simply mean teachers are corseting learners into
classroom procedures instead of facilitating their learning. In class, learners not always
ask questions about one particular language point but about a whole area: how to talk
about the past, how to use the word for in a variety of situations, how to use the bare
infinitive, how to join sentences, etc. When their curiosity is satisfied with an equally
comprehensive answer, they may select the parts they can reasonably handle and
incorporate them, leaving the rest in stock for later use. They can be trusted for
sensibly dosing their learning without following the textbook writer's or the tutor's
programme, thus following their own idea of gradualism.
3.f Compensation: a re-definition and a description of its role
Compensation has been listed as a learning strategy, for example, in cases when the learner
infers the meaning of a word from a context, thus enriching his knowledge. In our research,
however, we used the word in a different sense, dividing learners‟ performance into
compensatory and non-compensatory, a distinction we found relevant for the discussion of
learner performance in the classroom. We will define a compensatory utterance as one coined
by the learner in an attempt at expressing an idea requiring language forms he has not yet been
exposed to. We shall call compensation strategies the processes and resources learners use to
bridge the gap between their communication needs and their communication possibilities as
determined by their level, and which reflect their style for problem-solving, also derived from
Learning from Learners‟ Errors Ana María Rozzi de Bergel
98
learning hypotheses. When coining utterances, the learner often produces wrong language
forms, but it is not very appropriate to say he makes an error; rather, he creates a form to fill
the gap.
It might be argued that all errors are compensation strategies, but when we talk about foreign
language classes we can assess when the learner is making errors in items that have been
taught or he has been exposed to but he has not yet learnt, and when he is making
overambitious attempts at using language that is above the course level and beyond his means.
It is within this context that compensation strategies can be detected: when the input is well-
known. The relevance of this distinction is purely pedagogic and marks the difference between
failing to use what has already been taught and not learnt and applying one‟s creativity to
solve a communication problem, and how this is done.
Example: A learner produces, “The down old get jobs easily”, meaning that it is easy for
younger people to get jobs. This learner has never been exposed to the comparison of
adjectives. The production of The down old" to compensate for Younger people is showing
great ingenuity and a very positive trend towards trying to solve his communication problems
with elements well within the code of English. He is also combining two conceptual categories
in a personal fashion. He does not take the easy way out, which would be modelling the
sentence on the code of Spanish, his native language, to produce something like The more
youngs. This is resourceful compensation disclosing conceptual learning hypotheses, although
it results in an erroneous utterance.
If this happens at an intermediate level course and the learner producing The down old would
have been supposedly able to say Younger people, because of the input received, it is showing
the learner has a gap in his learning and fills it in with elements to which he is attaching novel
uses and meanings. He should not only know younger but the uses and meanings of down, too.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
99
This is an error in a non-compensatory utterance and although the strategies, styles and
hypotheses for solving the problem remain the same, its origin and implications are different.
The teacher cannot equally interpret errors which involve elements already taught but not yet
learnt and errors produced by a learner who is over-ambitious or takes risks. Both learners use
their creativity for turning out an innovative combination of concepts, but the first is engaged
in an exploration of boundaries of the code of English, whereas the second tries to make do
with the little knowledge he has managed to acquire. The utterance is a sign of
experimentation in one case, of a gap in learning in the other. Of course, this latter definition
does not imply that all classroom input should be learnt, but that the input has existed and has
been discarded or ignored.
Compensatory production of a highly monitored nature may be one of the key manifestations
of the learner‟s personal curriculum, when the learner seeks to fill the gaps and asks, "How
can I say .......?" or "What is the English for ......?" The missing elements may well be part of
the contents of the learner‟s curriculum and the teacher should be well advised to supply them.
Again, most learners can be trusted with deciding whether they are ready to incorporate these
forms.
Compensation is sometimes the result of teaching practices or materials-induced, as we will
discuss later on, but when it is learner-originated, it sometimes reflects problem-solving
strategies or learning hypotheses. In those cases, it may take three main forms:
a) Learners effect compensation because they are very ambitious and want to explore the
language. They take risks and use their imagination and creativity to structure the
language system. If their personal curriculum is respected, they will sooner or later
tailor their ambitions to their possibilities.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
100
b) Learners effect compensation because they do not regard the language they are
learning as a reliable tool for communication. They do not try to exploit what they are
exposed to, which they ignore completely at the time of performing in class. This
would be the case of the raw beginner who has just learnt how to say "Thank you", but
insists on trying to say the equivalent of "I am thankful to you" and asks the teacher for
help to frame his utterance. The result will probably be "I am thanks to you", or a
similar compensatory sentence. The difference between these learners and those in the
previous category is that they do not explore more advanced language forms out of
curiosity or to learn more quickly: they just try to ape the production of native
speakers, because they are unaware of gradualism. They probably have difficulty
organising language into hierarchical categories, and do not rank language forms from
simple to complex or organise their learning with some central criterion, e.g. going
from the general to the particular, or from the concrete to the abstract, etc.
Care has to be taken not to confuse these learners with those who ask ambitious
questions as part of their personal curriculum. The distinction is not easy and has more
to do with how they process their requests: whether they ask for specific language all
the time, as if they were repeating the lines in a script, “How can I say …….?”, but
then never use these elements again, or whether they ask for more advanced language
forms which are then incorporated to their language, at least partially. The language
they request is not assimilated to the language being taught but used at the moment it is
needed and then discarded, while the contents of the lesson are ignored. Thus, the
learner acquires neither the items in his curriculum nor those in the course curriculum.
c) Learners are creative and imaginative but they cannot structure the system
conceptually. Their production is hap-hazard, their conclusions erratic and their
Learning from Learners‟ Errors Ana María Rozzi de Bergel
101
associations unpredictable. This type of completely unstructured creativity does not
help for the creation of a language system or even a personal grammar, as changes are
constant and do not follow patterns. Compensation occurs because of this deep lack of
rule-governed language structuring.
3.g Errors as manifestations of the personal curriculum
In the 20th
century, audio-lingual approaches of the sixties tried to prevent learners‟ errors as
undesirable; later, cognitive code learning stated that errors were features of experimentation
and necessary tools for learning, but also problems to overcome, and communicative
approaches of the 80‟s and 90‟s sometimes advocated that all errors were developmental and
would disappear as learners‟ command of the target language increased. (Richards & Rodgers,
1998)
As we have explained, the well-known paradigm of learning by trial-and-error was discussed
by Piaget (Ibid.), who pointed out that the trial of a single element produced a re-
accommodation of the whole system this element belonged to, not simply a correction or
confirmation of an individual hypothesis.
We have already explained the theories of modern connectionism and the statement that
nodules or network units in the brain are connected in a unidirectional fashion and the
connections can be activators or inhibitors and have an assigned value, which is the
connection force. Learning occurs when two units of the neural network are simultaneously
activated, which raises the connection force between them. We have also described the role
assigned to trial an error in this process, which causes a pattern to activate a system before the
exit which compares the real output to the ideal output and this reinforces, in retrospect, the
path followed by the activation pattern which resembles the ideal output more closely.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
102
Learning takes place by repeating this trial and error process several times (Díaz-Benjumea,
Ibid.). As we can see, the value of errors is also recognised by more biologically-oriented
theories of learning.
The clearest manifestations of the learner‟s personal syllabus are the errors found in the
learners‟ interlanguage and the underlying cognitive processes they seem to reflect, namely,
whether learners are categorising conceptually and what type of associations and creations
they are making. The term error will be used to refer to a language form in learners‟
interlanguage which does not comply with the standards of correction and/or appropriacy of
the type of language being taught at their EFL course, observable in open-ended oral or
written production and which is a manifestation of learning hypotheses, styles and strategies.
Errors can be classified with respect to the language category where they belong: an error in
the use of a verb tense, a plural form, a referential lexical error, a phonological error involving
a particular sound, etc. This is essential in learner profiling and also when assessing progress.
This type of classification is also vital in the description of a learner's personal syllabus,
particularly for detecting the learner‟s conception of progression and gradation of difficulty,
but we will concentrate on error types, classifying them according to the cognitive processes
which appear to have originated them and their implications, either placing them in Selinker‟s
(Ibid.) categories or in the new categories we have incorporated through our research.
It is a complex task to describe the desirable levels of correctness or appropriacy of the
language of a social group, given all the national, cultural and generational variants, and the
range of criteria available for such a description, but the language taught at an EFL course is
fairly identifiable and describable. Besides, all courses have an end performance level, stated
by the curriculum or examination designer, which is used as a parameter for measuring the
learner‟s degree of achievement.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
103
The term appropriacy is used here in the sense Widdowson gave it (1978) but it should be
remembered that the appropriacy and correctness of the utterances analysed will be measured
with respect to communicative classroom activities, not in natural social situations.
We will consider an utterance appropriate if
a) It is in the right register, according to the context and to role-relationship.
b) It performs the function for which it was intended.
c) It conveys meaning in a manner satisfactory to the speaker, which helps him achieve
his communicative goals.
Appropriate language may contain lesser errors, which should not affect any of the
characteristics listed above. Conversely, a fully correct utterance may be inappropriate if it is
not in the right register or departs from the topic of the conversation, for example:
A: What‟s the time?
B: You‟re welcome.
We will not use the traditional distinction between an error = a systematic breach of the
code, and a mistake = an occasional slip of the tongue or the pen (Pit Corder, Ibid.). This
analysis is not so much related to the external manifestation of errors but to the underlying
processes they may disclose. If a learner makes mistakes or errors, according to the traditional
classification, we should analyse whether they point to overgeneralisation, interference from
the mother tongue, lack of proper conceptualisation, etc. The learner‟s error styles are present
in all his errors, be they systematic or occasional, because they are reflections of deeper mental
processes – learning hypotheses – which do not change even when the learner produces wrong
Learning from Learners‟ Errors Ana María Rozzi de Bergel
104
utterances on purpose, for instance, when he is joking, or when he goofs because he is tired, ill
or under emotional strain.
Tired, confused, trying to be funny, erring on purpose, the learner will not be able to change
his cognitive structure and if he uses pre-conceptual categories for grouping concepts, he will
have no recourse to any other form of categories for joking; neither will he take his confusion
or doubts to a conceptual level. This study will pay more attention to the systematicity of the
error style, not of the systematicity of the erroneous form. For example, we will discuss to
what extent the learner effects simplification, not how many times he simplifies the rule for
the use of the "s" in the third person singular present.
Although all errors are manifestations of the learner‟s personal curriculum, there are some
which seem to expose it more clearly, as was observed in the experimental courses:
a) Highly frequent errors which are becoming a definitive element of the learner‟s
language system.
b) Errors common to a group of learners. They point to error-provoking materials or
teaching.
c) "Master errors" – errors that affect many subsystems, errors that have important
ramifications and, when addressed, have a domino effect on other errors which are also
corrected as a result of the master error disappearing. This is a name we coined when
we observed mysterious effects that exploiting a particular error had on others which
had not been addressed.
For example, if a learner consistently uses a double comparison, like “My car is
more better than yours”, this error will affect just the structures related to the
comparison of adjectives and adverbs, but if the learner consistently says “I no come
at seven” or “I no like apples”, we are probably in the presence of a master error: if
Learning from Learners‟ Errors Ana María Rozzi de Bergel
105
the learner has not grasped the sub-system of auxiliaries and their role in negative
statements with respect to the auxiliary do, he will also have problems with other
auxiliaries, including modals.
If we address another error we have discussed, “He cans goes", this may improve the
syntax of the modals and the syntax of the present simple and probably of the past
simple or produce an improvement in the learner‟s ability for properly restricting
conceptual categories.
Sometimes master errors are personal and not easily recognisable through a purely
linguistic analysis. During our research, we found learners who made unexpected
associations between language domains. For example, a learner under observation
corrected the word order of his sentences when he was trained to stress words properly,
a phenomenon nobody could explain and nobody could have anticipated.
To detect master errors, great attention has to be paid to the errors that worry THE
LEARNER, even if the teacher considers them irrelevant. In the case we have just
mentioned, the teacher was trying to teach word order when the learner asked to be
taught word stress, showing complete lack of interest in word order.
The personal curriculum also contains items, methods and approaches which agree with the
course curriculum, but it is not a proven fact that learners who do not seem to depart from the
course curriculum learn more or faster. They are just less troublesome, because they do not
make unexpected errors, they do not ask odd questions, they do not propose novel activities or
methods. They are the “good pupils” that teachers like, but their adaptive behaviour may be a
sign of passivity, a refusal to take risks or a reluctance to explore new paths in language use.
Error avoidance is not a good learning style, but neither is continuous exploration by learners
who are pushed beyond their level of mastery.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
106
Lewis (1993) gives the example of a learner who, when asked how he had travelled to
Britain, answered “I came by plane” because he did not remember the past of the verb fly.
Lewis suggests that it would have been better for this learner to answer, “No, …er… I flied”,
because then, his teacher could have re-formulated the utterance, “Oh, you flew, didn‟t you?”.
The basic idea here is that experimentation leads to learning whereas error avoidance stunts it.
Why should learners have to choose, however? The answer could have been, “I came by
plane” and the learner could have asked about the verb later. Encouraging learners to make
mistakes by departing from language forms they already master seems just as negative as
favouring error avoidance. Learners should be well advised to put the language they know to
full communicative use and explore new language.
This permanent swing of the pendulum seems to have caused great harm to EFL teaching, by
prescribing extreme positions: to correct or not to correct, to translate or not to translate, to
repeat or not to repeat, to make grammar rules explicit or not. Incorporating the learners‟
personal curricula to the course curriculum puts an end to these unproductive dichotomies: the
teacher follows both curricula, according to need, allowing learners to choose part of the work
to be done or topics to be learnt, in the order and to the level they decide, but the teacher does
not relinquish her role as a leader of the learning group.
In his report on an experiment to teach a course on the basis of a process syllabus, Budd &
Wright (As cited in Nunan, 1993), explain the learners‟ demands for more guidance and their
wish that the teacher assumed responsibility for professional decisions regarding content and
method. This reaction forced the teacher, who had been negotiating the syllabus practically
before each lesson, to reconsider the advisability of such practices. Their experiment of
putting a process syllabus into effect had to incorporate more guidance and a firmer grip on
the general conduct of the course, allowing learners to make contributions but not crucial
Learning from Learners‟ Errors Ana María Rozzi de Bergel
107
decisions. They wanted to be taken into consideration, but they also wanted the teacher to
carry out his professional work and lead the way.
Error analysis as one of the main explorations of the learners‟ curricula may contribute to
implement this balanced combination in an informed manner.
Figure 5 – Summary: the personal curriculum
END PERFORMANCE (Language ceiling)
The learner’s personal curriculum
Culture, life experience, psychology, personality,
cognitive structure, logical structuring of native
language, communication needs and strategies and
other factors.
determine
The learner‟s degree of command of his
native language and communicative range
shape up
Nee
ds
and e
xpec
tati
ons.
Com
munic
atio
n s
trat
egie
s.
Lea
rnin
g h
ypoth
eses
, st
yle
s, s
trat
egie
s.
Lea
rnin
g m
ethod
Lea
rner
‟s q
ues
tions.
Com
pen
sati
on.
Err
or
types
consisting of
evidenced
by
grammar lexis language
content
learning
hypotheses
method
leading to
Learning from Learners‟ Errors Ana María Rozzi de Bergel
108
EXPLORATION AND REFLECTION
1. Make a list of the manifestations of the learner‟s personal curriculum we have
described in this chapter. How many of them are you addressing in your lessons or for
course planning?
2. Carry out a communication strategies interview in the manner recommended in this
chapter. Do it with a learner you consider either “fast” or “slow”. See if you detect
“favourable” or “unfavourable” communication strategies in this person and if they
match your assessment of this learner.
3. Think about it: what do you do when a learner asks you a question “out of the blue”?
4. Draw up a list of master errors.
5. Think about a situation when you felt your students were not interested in what you
were teaching. Were they interested in another topic, structure, activity? What did you
do?
Learning from Learners‟ Errors Ana María Rozzi de Bergel
109
4. Final level of attainment in the personal curriculum
It is not always very easy to define the expected level of attainment for a particular language
programme with precision, but this is usually done in terms of performance and through a
rubric. For example, the end performance of an advanced business English course could be:
To be able to participate in meetings, discussions and negotiations. To deliver
presentations and field questions from the audience. To produce written
communications such as e-mails, memos, letters and reports. To perform all these
tasks with lesser language errors which would not seriously hamper communication, in
the appropriate register and using the style, vocabulary and format proper of Business
English, thus communicating efficiently and effectively.
When a learner decides to enrol in such a course and this description of the end performance
is made available to him, he may consciously accept to work towards this level of attainment,
but there are still two levels of interpretation of this goal: the learner‟s and the teacher‟s.
What does “participating in meetings, discussions and negotiations” mean to the learner?
The teacher is dreaming of pertinent, clever and fluent contributions to discussions and the
learner may be a person who just drops a few words here and there on such occasions.
How does the learner understand the concept of “errors which would not seriously hamper
communication”? We have all known people whose language is error-ridden and yet manage
to make themselves perfectly understood with the aid of gestures, mime and even visuals.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
110
We may even wonder what the learner understands by “Business English”, as he is likely to
be familiar with the style and jargon used by his company and maybe a couple of other
companies in the same line of business, and will probably know his needs better than his
teacher. He will, for sure, know his clients and their culture and will know, for example, than
in some countries, being a heavy drinker is more valuable for doing business than being a
great talker. So, exactly how much of this pre-established end-performance is the learner ready
to attain….. if any?
4.a A re-definition of fossilisation
The personal curriculum is not linear but spiralled, developing in plateaux towards the
desired end performance. When this final goal is attained, the learner reaches his language
ceiling and changes still take place, but only in certain areas. We will define the manifestation
of language ceiling as a point beyond which learners refuse to incorporate, mainly, new
grammatical items, but also lexical or functional elements and prefer to “make do” with the
language they have acquired, however flawed. Rather than a process of fossilisation, we might
consider it a sub-conscious refusal or impossibility to move forward. Motivational,
physiological, sociological and cultural factors always play a role and sometimes the decisive
one; we will discuss them briefly, but they are better studied by psychologists, neurologists
and sociologists.
Refusal may occur when the learner has achieved to the measure of his needs and
expectations or has completed the system he attributes to the foreign language, according to
his cognitive structure and learning hypotheses. He has reached the desired end performance
in his personal curriculum. Selinker and Lakshmanan (1992) define fossilization as
interlanguage containing persistent non-target-like structures, but in 1996, Seliker talks about
Learning from Learners‟ Errors Ana María Rozzi de Bergel
111
cessation of learning and a permanently fossilized competence and insists that adult learners
are not able to reach native-like competence. Since 1972, then, Selinker‟s idea of fossilized
structures has evolved to an idea of global fossilization, which is not cessation of learning but
a level of ultimate attainment, similar to the idea of end performance in the learner‟s personal
curriculum and to the concept outlined by Ellis (Ibid.), which we have already discussed.
We may assume that some features of this end-performance fossilise because the learner‟s
motivation is low or he is not very ambitious and in his personal curriculum, the satisfactory
end level has not been set too high. His satisfaction is achieved fairly early in the learning
process, but the learner does not always seem disturbed by his lack of progress, which is often
evident only to the teacher or other observers. In these cases, we may find transient
fossilisation, and the learner may continue improving his command of the foreign language if
he makes a conscious effort or his motivation changes.
The impossibility to move forward becomes apparent when learners reach their language
ceiling even while striving to go on learning, which brings about frustration. This may be
related to biological or psychological explanations of permanent fossilisation. In these cases,
motivation and need do not seem to help; rather, the desperate attempt to move forward
creates more obstacles. The situation may be dramatic when a promotion, a scholarship or a
new job depends on the attainment of a certain level of command of the foreign language or
when huge sums of money are spent on intensive courses, to no avail. When opportunities
may be lost or money is wasted – particularly if it is the employer‟s money – anxiety and
stress reinforce the presence of language ceiling and learners find themselves beyond help. In
these cases, we might find that the content of the personal curriculum has not been learnt and
the satisfaction of communicative needs has not been attained because the method and
learning hypotheses in the personal curriculum were not appropriate and they are so firmly
Learning from Learners‟ Errors Ana María Rozzi de Bergel
112
lodged in the brain that they cannot be altered. The end performance may have been set at a
higher level than the personal learning approach can lead the person to attain.
4.b The impact of feedback on fossilisation
Vigil & Oller (1976) outlined a model of fossilisation which took extrinsic feedback
(Selinker & Lamendella, 1979) as the determining factor. In their view, positive cognitive
feedback combined with negative affective feedback favoured fossilisation, whereas negative
cognitive feedback combined with positive affective feedback would cause learners to modify
their linguistic knowledge and move forward in their learning. The model has several weak
points, however, as already discussed by Selinker & Lamendella (Ibid.) who argue that the
beginning of fossilisation is controlled by internal factors, although extrinsic feedback plays a
role in reinforcing certain parts of the learner‟s interlanguage; the lower boundary of
fossilisation is set by the learner‟s communicative needs; reinforcement of communicative
competence is not necessarily combined with reinforcement of correctness and finally, there is
no fossilisation in the native language. The latter argument is a moot point. At any social
event, we will be surrounded by native speakers who display different degrees of mastery of
their native language, depending on their background and personality. Their language ceilings
are not at the same height.
We would also need to characterise “positive” and “negative” feedback, both cognitive and
affective, and find a way to describe the participation of affect in cognitive processes, to
outline the boundaries of each type of feedback. There is no feedback which can be
objectively classed as “positive” or “negative”, as this would ultimately depend on the
learner‟s perception of the nature of feedback; something which he himself may find
impossible to define or classify. We would not know, either, how much feedback of each type
Learning from Learners‟ Errors Ana María Rozzi de Bergel
113
a learner would need to avoid fossilisation or fall into it. If we were talking about feedback in
a limited way, similar to the notions of positive or negative reinforcement (Skinner, 1977), it
should be relatively easy to answer all these questions, but in the broader context of holistic
communication, feedback may take so many forms and it may be interpreted in so many ways
that its categorisation appears quite elusive. A typical example of the impossibility to class
feedback as “positive” or “negative” may be the case of a highly motivated and optimistic
person who, when told that his utterances are incomprehensible, or faced with a complete
breakdown of communication with his interlocutors, sees this as an interesting challenge to
learn more rather than as a source of frustration. On the other hand, learners who communicate
fairly well but are extremely competitive and self-demanding, may feel frustrated when they
make a lesser, irrelevant error, even if the interlocutor does not seem to notice it and it does
not affect their performance.
We might even consider cases of learners who respond to aggression, criticism or
punishment better than to any other form of stimulation, because they come from
dysfunctional environments or are in need of psychological counselling. Professional teachers
would never ill-treat a learner even if he required it, but their democratic, professional, non-
inhibiting feedback may well be taken by these learners as weak and unreliable and cause lack
of motivation. This is, of course, an extreme example, but it shows the relativity of the
concepts of “positive” and “negative” feedback. The case of interactive computer-based
programmes for learning English should be an interesting field of research. The same learners
who would find a teacher authoritarian if she constantly said “You are wrong” and “You are
right”, without further explanations, gladly sit in front of a computer and learn with programs
which just sanction their production as right or wrong, providing no real feedback.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
114
Brown (1994) adheres to Vigil and Oller‟s (1976) theory of the importance of feedback and
claims that fossilization may be reverted if learners are given positive affective feedback (“Try
again, you can do it”, e.g.) together with neutral or negative cognitive feedback, i.e., mistakes
are pointed out, but the learner is encouraged to reconsider his hypotheses. The method is
rather naïve and it is not clear how the teacher would distinguish a fossilized error from a
developmental one and whether the same procedure would be effective with both.
The influence of extrinsic factors on the height of a person‟s ceiling is undeniable, as man is
a social being and many mental processes originate in the environment. Measuring and
assessing this influence reliably is almost impossible, unless we are satisfied with describing
models of feedback in the classroom in order to correlate their use with observable features of
the learners‟ interlanguage. We might then be able to see the correspondence between types of
feedback thus classed and the learners‟ production, but not to make generalisations about their
affective value. We would know if saying “I think I see what you mean, but could you clarify
it, please”, rather than, “I don‟t understand”, produces better re-formulations of messages,
but we would be hard put to establish if this happens because the remarks are taken as positive
or negative feedback.
4.c The role of acculturation – in whose culture?
Tollefson & Firn (1983) also stressed the role played in the appearance of fossilisation by
negative cognitive feedback combined with an overemphasis on communicative tasks in the
classroom, placing interaction at the core of the phenomenon. They also discussed the role of
acculturation, holding that fossilisation occurs when the learner‟s acculturation to the target
language culture ceases. We will refer to acculturation in foreign language learning as the
acceptance and adoption, by the learner, of cultural features of the foreign language speech
Learning from Learners‟ Errors Ana María Rozzi de Bergel
115
communities, without significantly changing his own culture or losing awareness of his
identity. We should note that the acculturation agents in a language course are not the real
members of the foreign community, as would be the case with second language acquisition,
but the teacher, the school and the materials. Fossilisation would appear when the
acculturation process has reached its feasible limits, when it cannot continue because of
insufficient input; for example, because learners are not in contact with a foreign community,
or when the learner rejects further acculturation.
Although learning a language inevitably entails apprehending its culture, it should be noted
that English as a lingua franca is so widely used nowadays that the teacher or the school would
be hard put to decide which English-speaking community their learners should acculturate in;
perhaps, they would prefer to concentrate on teaching a fairly universal, culturally unmarked
variety of English which might be understood by other non-native speakers or to expose
learners to different language varieties. The issue is extremely relevant for the consideration of
language ceiling, as acculturation involves power struggles, ideologies, supremacy, resistance
and survival as well as tolerance, personal enrichment and maturity, depending on how both
parties approach the issue. When two communities are in touch and have to learn from each
other, these processes are very clear and usually result in the “dominant” culture prevailing on
the other or in the weaker community adapting to a culture it regards as “superior”, but in the
foreign language class, it may well happen that neither the teacher nor the learners will be
natives of an English-speaking community and will not be willing to acculturate. They will
seek to learn and teach the language while at the same time defending their cultural identity. In
this case, acculturation may come exclusively from the materials.
Will this result in a lower language ceiling for the learners? The case of non-native teachers
who have had very little or no direct experience of living in an English-speaking community
Learning from Learners‟ Errors Ana María Rozzi de Bergel
116
and whose command of the language is extremely high, might invite reflection on the relative
value of acculturation. On the other hand, it is a fact that first-hand contact with a native-
speaking community enhances acquisition of a foreign language, particularly if the non-native
speaker feels comfortable in this social group, as the two cultures come together in an
experience from which both profit. In this case, we should be talking about interculturation, a
more enriching phenomenon.
Further research may be needed into the influence that the degree of identification and
appropriation of the learners‟ native culture has on language ceiling and on the learner‟s
willingness and readiness to regard the contact with a foreign culture as a way of better
understanding other human beings and a contribution to his appreciation of his own cultural
identity. Informal observations and conversations with learners who show a low language
ceiling very often disclose particular attitudes in the cultural domain, for example:
a) A complete and usually unfounded rejection of the target language culture, usually
because of a single traumatic event related to it or due to very specific actions of one or
two prominent members of this community, or simply due to clinging to stereotyped
concepts of its users‟ idiosincracy. This points to a cultural background where reasons,
concepts and ideas seem to have been replaced by prejudice, clichés and over-
generalisations.
b) A refusal to learn about events or people belonging to other cultures, which is seen as
completely alien and useless for learning the language, compounded with a lack of
interest in culturally relevant events in the native culture. The learner does not want to
hear about the history of the foreign culture or that of his own. He only reads the
football page of the local newspaper and is totally unconcerned about local music,
politics or economy, let alone history or art, which are considered a waste of time
Learning from Learners‟ Errors Ana María Rozzi de Bergel
117
because they are not immediately useful. The impossibility to relate another culture to
one‟s own is due to an incomplete acculturation in the native community. The foreign
culture finds no links in the native culture and learning is not meaningful because it
cannot be associated to previous knowledge. This may cause a low language ceiling,
besides limiting the topics of conversation and areas of interest.
c) The complete willingness to acculturate in the foreign language, due to poor
appreciation or ignorance of the value of one‟s own. Learners look up to the foreign
culture, but trying to become good Americans or good Canadians is an impossible
goal, because new knowledge is usually constructed on the basis of previous
knowledge and in these cases, the foreign culture finds no links to the native one,
which has been rejected. Knowledge thus acquired is never significant or meaningful
and tends to be easily forgotten or to become totally context-bound, as if the learner
were trying to learn the contents of a phrase-book for tourists in the country of his
dreams.
There may be cultural universals where language anchors itself, such as the notions of
identity, roots in a historical past, appreciation of traditions and cultural heritage, to mention
but a few. Perhaps, an exploration of language universals and probable communication
universals should also include a study of learners‟ awareness and appreciation of their culture
and their language, to estimate if the foreign language and culture will be anchored on solid
ground or sand dunes.
4.d Biological factors
Studies in the field of neurolinguistics have disclosed the presence of permanent fossilisation
brought about by biological factors, mostly genetic and age-related, and transient fossilisation
Learning from Learners‟ Errors Ana María Rozzi de Bergel
118
produced by acculturation limits or other socio-affective influences on the learning process.
(Selinker & Lamendella, 1978; Tollefson & Firn, Ibid.) and more recent experiments, using
the event-related potentials (ERP) method of brain scanning seem to back this theory. This
method studies consistent patterns that accompany the registration and evaluation of each
discrete piece of sensory information in the brain, as opposed to general electric changes and
fluctuations. Vos et al (2001) used ERP techniques to research how individuals process
language input and achieve comprehension and found that the subjects of the experiment
possessing a high working memory capacity were more accurate in their interpretation of
words and sentences and were more efficient in their processing of concurrent information.
These experiments have explored comprehension rather than production, and at a very limited
level, but they are indicative that biological factors play a crucial role in determining the
height of a person‟s language ceiling.
The assumption that age brings about the loss of an executive component of the capacity to
adapt existing systems or construct new systems (Lamendella, Ibid.) has been seriously
challenged by electrophysiological investigation, which has pointed out that the brain
performs its executive functions at any age, so fossilisation should be attributed to other
factors: processing efficiency, for example, rather than executive capacity, or the location in
the brain of native language and foreign language systems. If they are separated in the brain,
the former should be language-specific and the latter might be of a different nature. (Gunter et
al, 1997) The distinction may support the theory of a latent language structure and a latent
psychological structure (Selinker, Ibid.).
The efficiency to process information and to form associations and systems has been pointed
out by Purdy (2001) as a brain capacity potentially traceable through ERP techniques. A
working hypothesis might be that the processing of fossilised syntactic forms of a foreign
Learning from Learners‟ Errors Ana María Rozzi de Bergel
119
language shares the same characteristics as the processing of native language forms and their
representations are collocated together, whereas non-fossilised forms are found in the brain in
another area and are not easily accessible under communication constraints, when the
fossilised forms tend to prevail. The difference here is that we are not talking about fossilised
errors, but about fossilised forms constituting a solid system, such as the first language system,
either correct or incorrect. An important factor to consider in these biological or genetic
explanations is that the neurosciences do not ignore the role of motivation and social stimuli in
shaping up neural maturation, so the limits between genetic and developmental or extrinsic
factors remain unclear when attempting to explore the causes of language ceiling.
4.e From fossilisation to language ceiling
As we can see, research into the factors which determine the height of a person‟s ceiling
would call for collaborative, multi-disciplinary efforts, as they may be psychological, social,
physiological and even economic or political and probably originating in several of these
domains simultaneously. It is even unclear whether the causes determining the height of this
ceiling might be identified at all, as research ought to be, necessarily, ex-post facto: once
learners have reached their ceiling, we should explore the whole of their previous life
experience and physical/psychological traits in depth to find traces of possible causes of high
and low ceilings. The data would be extremely ambiguous, incomplete and subjective, when
not impossible to obtain.
Teachers are unable to detect the causes which set the height of a learner‟s ceiling and are
not professionally prepared to act upon them – a job which might be undertaken by a
psychologist or a neuroscientist. However, they can be trained to detect the manifestations of
language ceiling and address this stage in learners‟ progress appropriately.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
120
As explained, teachers can greatly enhance the results of their courses if they diagnose the
learners‟ expectations, needs, background, learning hypotheses and communication styles
before the start of a language programme, then describe their learners‟ linguistic performance
and behaviour during lessons, particularly at ceiling level, use the information to interpret the
external manifestations of personal curricula and language ceiling and tailor their teaching to
them.
Is there a way to predict the height of a person‟s ceiling? This was a recurrent question
during our research, and one which gave rise to a lot of misgivings: What if you can, but in
doing so, actually condition the teacher‟s expectations, and learners who are supposed to have
low ceilings are, in fact, led by the tutor in that direction? This is the famous Rosenthal effect,
by which the researcher conditions the subjects‟ behaviour so that a hypothesis may be proven
right.
If we analysed the diagnoses of the Learning Hypotheses Test included in Appendix 2, we
might presume that resorting to pre-conceptual categories for grouping concepts, having a low
degree of imagination and creativity, poor acculturation in the native language and culture, a
predominantly metonymic association style, and a poor level of command of the native
language, might point to a low language ceiling. We may also suppose, without using the test,
that poor motivation, lack of ambition and rejection of the language or the culture which is
associated to it might also determine a low language ceiling. However, humans are very
fluctuating beings, the mind is a liable entity, the social and affective domains have unforeseen
effects on cognitive capabilities, so making a prognosis of the height of a person‟s ceiling does
not seem an easy or even feasible task.
Many of my research colleagues would say that much as they do not believe in witches, but
they are there, nothing is a sure indicator of a potentially low or high language ceiling, but the
Learning from Learners‟ Errors Ana María Rozzi de Bergel
121
indicators are there. In twenty years, we did not find a single case of a learner with a pre-
conceptual style for structuring language and either very high or very low imagination and
creativity (Diagnosed using the Learning Hypotheses Test) who would reach a high language
ceiling, and this included some who were highly motivated. Of course, there were others who
had a low language ceiling despite their conceptual style for structuring language, because
they lacked interest in learning or had personal problems.
Therefore, we could not be so bold as to claim that the height of a learner‟s language ceiling
can be predicted by analysing his style for structuring language and his level of imagination
and creativity, but we can report that in our longitudinal studies, 100% of students with a pre-
conceptual style for structuring language and either a very high or very low level of
imagination and creativity proved to have low language ceilings, reached at pre-intermediate
or intermediate levels or even earlier, regardless of their degree of motivation or the approach
and methods used to teach them. These longitudinal studies were carried out with 12 cases,
over a period of about five years, initially, and then continued with all the cases which were
diagnosed as pre-conceptual, with either a very high or very low level of imagination and
creativity, as a matter of routine. The possible prognosis of the height of an adult learner‟s
language ceiling is, undoubtedly, a line of research to be pursued much further, but the
diagnosis of the Learning Strategies Test already provides valuable indicators.
Another interesting issue is whether learners at their language ceiling show a particular
pattern of error production. Can we diagnose ceiling through error analysis? The answer
depends on the causes of this ceiling. When learners reach their ceiling due to an impossibility
to progress any further, caused by the nature of their learning hypotheses or other factors
already described, there is a considerable increase in the number of “unproductive” errors: too
Learning from Learners‟ Errors Ana María Rozzi de Bergel
122
many semantic errors or errors in appropriacy, interlingual errors or a considerable number of
compensatory utterances, for example.
When ceiling is caused by unwillingness to learn or other motivational problems, error
analysis will not disclose it. The composition of the learner‟s errors will show fairly balanced
values for all the types of errors involved. Thus, we may conclude that whereas error analysis
is not enough to diagnose language ceiling, it will reveal language ceiling which has been
reached due to an impossibility to continue learning. Correlating the results of the Learning
Hypotheses Tests with the results of error analysis can shed more light on these assumptions.
We will come back to the question of diagnosing and making a prognosis on language ceiling
later, after a description of the categories used in our analysis of errors.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
123
EXPLORATION AND REFLECTION
1. Record yourself delivering a short speech in your native language. Listen to the
recording. How many fossilized errors have you made? If you can‟t detect them,
consult with a specialist in that language.
2. Have you ever had students who were at their language ceiling? How did you know
this? Make a list of the indicators you paid attention to.
3. Have you ever had students who reached their language ceiling only because they were
not interested in moving forward? Or do you think they had lost interest because they
had reached their desired end performance? Would you re-consider your opinion of
these learners now?
4. Are there structures, functions, lexis or other language items adult learners always
seem to learn? Make a list. Why do you suppose they learn this easily?
5. Do you know whether your students share your idea of a level of end performance for
the course you are teaching? In other words, do you have the same expectations?
Devise a questionnaire and carry out an informal survey, asking them what they expect
to learn.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
124
PART 2: Exploring errors in the EFL classroom
1. Error analysis as a teaching tool
Traditionally, teachers have considered errors as problems to overcome and have taken
corrective action in order to help their students eliminate them from their interlanguage.
Approaches of the early 20th
century regarded correction as a process by which teachers
pointed out and corrected errors, offering the correct form, and students were expected to
study or even memorise this correction in order not to make recurrent mistakes.
Behaviourism advocated that if students were not exposed to erroneous language; for
example, their classmates‟ interlanguage or even their own, they would not produce mistaken
forms. Therefore, care was taken to reduce production to the repetition of correct sentences; a
procedure in line with the idea that learning a language was practically a question of
developing the right habits. Unfortunately, these practices seemed to deprive the student of the
possibility of making hypotheses about the language they were learning and trying them out in
order to correct or corroborate them.
The communicative approach re-defined the teacher‟s role as that of a moderator, co-
ordinator, facilitator or guide. From this position, teachers were not supposed to use their red
pen to cross out errors and provide correct forms, or to interrupt a learner while he was
speaking, to point out a mistake. Correction techniques changed, but correction was still the
only treatment of errors, although the concept of error acquired new meanings and new facets.
Communication became of paramount importance and utterances containing lesser errors
began to be tolerated, reflecting real language use, and learners‟ interlanguage began to be
analysed not only from the point of view of accuracy but also according to degrees of
Learning from Learners‟ Errors Ana María Rozzi de Bergel
125
appropriacy, which aroused researchers‟ curiosity about the processes behind error production
beyond interference from L1.
In the late 20th
century and early 21st century, methods and approaches continued focussing
on error correction, albeit with innovative techniques based mostly on feedback and re-
formulation (Krashen & Terrell, Ibid.; Johnson, 1988; Lewis, Ibid.). Correction became less
and less threatening and less teacher-based, as techniques such as peer or group correction
sought to strip correction of any resemblance with criticism or punishment.
Some approaches that came into fashion in the nineteen eighties, namely Krashen and
Terrell's Natural Approach (1983), the constant preaching against teacher-centred classroom
practices and the focus on learner autonomy that prevailed in the nineteen nineties were
widely misinterpreted as banning error-treatment and correction altogether. The idea that
mistakes should not be regarded as problems to overcome but as manifestations of the learners'
needs gained widespread acceptance but in most methodologies the only concrete
implementation of this principle referred to methods of correction that might encourage
reflection, analysis and self-correction or metacognition, rather than to concrete exploitation of
errors for learning purposes.
In some of the so-called "natural" approaches, correction was discouraged on the assumption
that learners should learn the foreign language as they had learnt their first language:
presumably, spontaneously and without formal correction. It is arguable whether lack of
correction or treatment of errors is natural or desirable. It may not be natural, because even a
little baby is punished and corrected when he makes a language error. He is punished with
lack of communication and corrected by the social environment. When faced with the anguish
of isolation and the impossibility to satisfy his most basic needs, such as nourishment, pleasure
Learning from Learners‟ Errors Ana María Rozzi de Bergel
126
and comfort, the baby earnestly tries to pick up the words he needs to communicate and
survive.
The baby's plight is heart-breaking: he has to communicate with people the size of a three-
storey building who manipulate him, feed him, clothe and unclothe him at will and even force
him to sleep, wake up or dive into an unwanted bath. These giants seem to get everything
wrong: when he consistently throws objects to the floor to learn how things fall, and the notion
of "down", they think he has dropped his toys and tie them up to his pram or cot, thus
interrupting his learning game. Baby cries, and instead of untying his toys, his gorilla-size
mother thinks he is hungry, or cold, or hot, or maybe teething; so baby gets bottles, medicines,
a fresh set of nappies - anything but his toys. One day, he manages to utter a word, but it is not
recognisable in the giants' language, so he still gets the wrong results. The relief comes the day
when he manages to say "Toys". Who wants to be in this situation at the language class?
Methods had better avoid all resemblance to this early period and find more humane methods
of dealing with language limitations.
To hold that babies are not formally taught – again an implication made by Krashen and
Terrell (Ibid.), seems a moot point as well. A simple walk around a toy-shop or a playground
will provide enough evidence to counteract this assertion. We will hear a lot of adults telling
very small babies: "Say „Hello‟ to Santa Claus", "Say „Thank you‟ to this nice lady", " Oh!
What has baby got? An ice-cream. Ice-cream. Ice-cream. Come on, say ice-cream.", "Good
Lord, he can say „Mum‟!” Ok. Can you say "Dad", too? Come on: Dad, Dad, Dad.", not to
mention poems, songs and lullabies which are sung and recited with mime and gesture to
convey the linguistic message to the tiny, not proficient user of the language in question. The
world is a language class, and parents and adults are nothing but diligent teachers.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
127
Approaches which deny the value of error treatment have left learners fairly helpless and
teachers with a weaker role to play in their learners' language acquisition process. In spite of
the appearance of task-oriented and experience-based methodologies, the teacher's work is still
not centred on the interpretation of errors as much as it should be. As discussed before,
descriptions and classifications of learning strategies and styles have finally succeeded in
changing the emphasis of teaching from concern with methods and approaches to an in-depth
study of the learner and an ongoing, pragmatic response to feedback. Describing learning
styles and strategies, as well as analysing errors from a purely linguistic point of view, may be
simply the map, not the territory. It may mean describing the manifestation of processes, not
their roots. The roots should be sought in the type of learning hypotheses these features of the
learners‟ personal curricula seem to reflect: conceptualisation or lack of it, association by
proximity or analogy, use of creativity and imagination.
The study of errors as manifestations of the learning hypotheses in the learner‟s personal
curriculum is based on the assumption that teachers cannot teach adult learners how to
structure their learning. Rather, they should seek to unveil and follow the learner‟s personal
curriculum and incorporate it to the course curriculum.
When we look at present-day approaches and methods, at the amount of research into
learning styles and strategies, motivation, brain structure and functions, we can clearly see that
an in-depth study of the learner takes central place in modern conceptions of learning and
teaching. Nevertheless, it is not very clear whether teachers always attach so much importance
to these aspects in their actual practice; in fact, they often seem to be simply concerned about
results, measured by the degree of accuracy of their learners‟ interlanguage. Error analysis, in
the classroom, frequently stops at the linguistic level and the following anecdotes are
illustrative of this situation:
Learning from Learners‟ Errors Ana María Rozzi de Bergel
128
Several years ago, I heard a colleague complain, in the staff room, that her students made a
lot of errors and she seemed to be unable to eradicate them from their interlanguage. The
situation was serious, because these students were trainee teachers who would, upon
graduation, teach English, and my colleague happened to be their Language teacher.
I offered to analyse the errors of that class if she gave me some samples of their production,
collected according to my instructions. Although she knew I had been engaged into this kind
of study for years, she answered, “Analyse their errors? What for? They are always the same!
I can find them myself!” That answer was disappointing to say the least. Persistent errors of a
group of learners may well be teaching-induced, on the one hand, and on the other, the teacher
did not understand the purpose of error analysis at all and thought I wanted to help her detect
what part of the language system the errors affected: verbs, sentence structure, etc.
My worries did not end there. Another Language teacher decided to accept my offer and
brought me a list of errors for me to analyse, which read, for example: “The „s‟ in the third
person singular affirmative, present simple”; “I suggest you to come”; “We use to have ice-
cream in summer”, and similar transcriptions of learners‟ errors. She was very surprised when
I told her the list was useless, because it did not tell me anything about the contexts of the
errors, the kind of learners producing them or the type of teaching they were exposed to.
Besides, she remarked that if she had to provide all this information, error analysis was not
worth the effort.
These episodes reflect a deeply-engrained attitude towards error analysis in the teaching
profession, whereby analysis stops at the linguistic level, probably because teachers are more
concerned with the observable manifestations of interlanguage than with the inner processes
they reflect, as overt linguistic performance provides data for student evaluation and is scored
at examinations. When these examinations are administered and marked by third parties, like
Learning from Learners‟ Errors Ana María Rozzi de Bergel
129
international examination boards, they even act as auditing instruments for the teacher‟s work,
so it is not surprising that teachers should concentrate more on correcting language errors than
on addressing the messages they send – a procedure that would correct them more readily,
however.
I am going to talk about error exploitation and not about error correction, because if we
consider errors as items in the learners‟ personal curricular, we cannot talk about correcting
them, but about using them to foster learning and to complete the course curriculum, to which
they make important contributions. We should study the extent, manner and system of
exploitation in great depth, to avoid inhibiting learners or penalising errors. Errors should even
be celebrated as productive experiments and ways that learners have of validating their
hypotheses about the language being learnt (Strevens, 1969).
Errors provide a lot of information about the process of learning but it should be
remembered that the reason for a learner's progress or lack of it cannot be found exclusively
through error analysis. Learners whose errors disclose highly favourable learning hypotheses
often fail at their courses because they are not interested in learning or because their teachers
or bosses do not motivate them properly. Sometimes, motivation that is apparently positive
works against adult learners, as is the case with company employees who want to learn
English in order to get a promotion. Getting the promotion, not learning the language, often
becomes the priority, and the anguish this produces hampers learning. I will concentrate on the
interpretation and exploitation of errors, but understanding learners means using several types
of analyses and approaching the issue with an open mind.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
130
EXPLORATION AND REFLECTION
1. Give your students a communicative oral task and record their errors while they are
performing it. How many of these errors could you attribute to interference from their
native language?
2. Have you ever heard your students come to personal conclusions about the language
items they were learning? Think about an example of a learner-generated rule of
grammar.
3. What methods of error correction have you tried out? Make a list of those which
worked and those which did not work. Can you think of reasons for their success or
failure?
4. Would you say that learners who make few errors always communicate better?
5. Have you tried not correcting your learners‟ errors? What were the results? If you
haven‟t tried, ask other teachers or some students about their experience with this
approach to error treatment.
6. When you were a very small child, did people correct you when you made language
errors? If they did, how did they do it?
7. Was there any teaching involved in your acquisition of your native language? Any
formal learning?
Learning from Learners‟ Errors Ana María Rozzi de Bergel
131
2. Methods and instruments for error analysis
The colleague who thought that error analysis was not “worth it” because it entailed a lot of
work and consideration for many factors had, in fact, grasped one of the aspects of data
collection in this field with great precision: the fact that so much information about an error is
needed before it can be analysed.
I am going to describe how to obtain valid data for error analysis and how to use certain
criteria to analyse it, but before I do that, I wish to express my deepest appreciation for the
course supervisors who observed hundreds of lessons to record errors for our research project
and conducted dozens of interviews with learners to obtain either authoritative explanation or
more insight into the processes behind errors. Without these reliable records, nothing would
have been possible.
2.a Valid samples for error analysis
In error analysis, the samples to be analysed have to be drawn from the object under
scrutiny: language. Therefore, they have to contain all the elements that make them instances
of language use (Widdowson, Ibid.): text, context, intention, expression, and the user's
purpose. One of the key issues in error analysis is, then, how to collect valid samples of
learners' production. James (1998) recommends two procedures: broad trawl elicitation
through role-plays and simulations, where the researcher‟s intervention is minimal, and
targeted elicitation, where the researcher selects specific samples from learner diaries, cloze
passages or specific sources, to focus the analysis on particular traits of error production. In
our case, the procedure resembled broad trawl.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
132
Utterances lose part of their meaning when they appear out of context or divorced from other
communication codes which usually accompany them, so a simple list of errors made in
writing is of little value unless the whole text produced by the learner is available. A
transcription of errors made in oral performance hardly ever reflects non-verbal codes of
communication or even intonation, prominence and tone, so a complete register of the error
should be available. The analyst should also know the type of activity where the error was
produced, and have access to basic information about the school, the course, the materials, the
teacher and the learner. Additionally, it is essential to discuss the error with the learner in most
cases, to obtain an authoritative explanation and avoid mistakes in the interpretation and
diagnosis of the error. This shows that the researcher‟s task is extremely complex and perhaps
this explains why the research which serves as the basis for this work lasted twenty-seven
years.
The classroom teacher, however, will find error analysis much easier, as a lot of the
necessary information is already available to her. Besides, the teacher can correct and clarify
her assumptions as needed, as she is in daily contact with the class, which in turn gives her a
lot of insight into the problems under analysis. Also, the teacher‟s objective is instrumental,
whereas the researcher‟s is scientific. If the teacher finds that error analysis is a positive
teaching tool, she will not be bothered with keeping records, controlling variables and carrying
out statistical analyses: the learners‟ achievements will be her only concerns and to achieve
these good results, her task will be much easier than the researcher‟s.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
133
Figure 6– Valid sample for error analysis
The learners' oral and/or written production should be recorded in fairly open-ended and
creative activities such as communicative writing tasks, conversations about a classroom topic,
role-playing, etc.. Errors produced in close-ended activities such as fill-in-the-blanks or
transformation exercises will not be very revealing, for when a person concentrates on only
one language item, his errors will be related to this item, and when the operation to perform is
too mechanical, the only information we may get is whether the learner can perform the
operation requested by the exercise. This may be of interest to the language teacher and useful
What shall we analyse?
ERRORS PRODUCED IN FAIRLY
OPEN-ENDED ACTIVITIES,
The learner's
personal
data.
The level
of the
course.
The
activity.
The group
dynamics.
What the learner
meant or
should have
produced.
Age,
background,
occupation,
objectives,
history as a
foreign
language
learner, needs,
expectations, learning
hypotheses/styl
e/strategies,
communication
style, and any
other relevant
information.
Level
includes
placement
system,
materials,
curriculum,
approach,
method and
resources as
well as
background
knowledge
of the school
or institute.
Language
objectives,
pedagogic
aims, type of
interaction,
learner roles
and teacher
roles.
Suitability
for course
level.
Previous
preparation
provided.
Teacher‟s
leadership
style.
Learner-
learner and
teacher-
learner
relationships
Dominant
interaction
style
(circular,
radial, by
sectors)
Reconstruction of
the utterance, by the
learner and/or the
teacher.
Distinguishing
between
compensatory or
non-compensatory
utterances.
Interpretation of
non-verbal
communication.
if we know
Learning from Learners‟ Errors Ana María Rozzi de Bergel
134
for passing certain examinations, but is not so useful for gaining insight into the learner's
hypotheses or strategies, basically because he is not selecting what operation to make.
The analyst of an error should know the context in which it is produced, the level of the
course, the level of the activity in which the learner is performing, the role he is performing in
the activity and all the pertinent personal data of the learner: age, occupation, education,
nationality, native language, reason for learning English, needs, expectations, etc. Wright‟s
(1987) classification of activities is very useful in this domain, as is Tomlison‟s (1998),
because they map out the processes that learners are expected to carry out, as well as their role
in the activity.
The information related to the textbook, the approach, the method and the teacher's
leadership style is also relevant. The context of errors made in writing is relatively easier to
preserve than the context of errors made in oral production. In this latter case, a reliable
method is to analyse the error on the spot, the moment it is produced, before the context is
forgotten, even if the researcher uses video cameras or audio tapes to record the learners'
production.
When we speak of roles, we do not refer only to the role outlined for the learners by the
activity designer, but also to the psychological role he decides to play in the interaction: his
intentions, his aims and how he sets about achieving them (Wright, Ibid.; Pichon Riviére,
1984).
It is vital to know exactly what the learner meant in order to determine what he should or
could have produced. We will call authoritative explanation the procedure by which the
learner explains the structuring and meaning of his utterance. The concept is partly borrowed
from Pit Corder (Ibid.), who called it authoritative reconstruction. I will attach further
pedagogical implications to Pit Corder‟s idea, too, because I will present it not only as a
Learning from Learners‟ Errors Ana María Rozzi de Bergel
135
research method for data collection, as a necessary step in the analysis of an error, providing
the basic raw data, but also as a teaching tool for accurately understanding an error to be
exploited. We will call expert reconstruction the procedure by which the teacher reconstructs
and often re-phrases the learner‟s utterance according to her interpretation of the learner’s
intentions and before or instead of obtaining authoritative explanation.
Inviting authoritative explanation - that is, asking the learner "What do you mean?", or "Why
did you say that?", or just “So you mean that ….”, when he makes an error is not such a
standard procedure as it should be and very often, telling people what they should have said
precedes asking people what they wanted to say. The result is that many errors go unexploited
whereas a lot of errors that learners have not made are carefully corrected and discussed.
Let us analyse some examples recorded at EFL lessons:
Example 1:
Learner: (Describing a picture) There are some books on the floor and there are /tu:/ books on
the table.
Teacher: (Expert reconstruction) There are four books on the table.
Learner: Yes, there are three books on the floor and there are /tu:/ four books on the table.
Teacher: (Expert reconstruction) Oh! I see. You mean there are some books on the table, too.
Learner: Two? There are four.
In the utterance "There are some books on the floor and there are /tu:/ books on the table",
/tu:/ stood for too, not for two, as the teacher thought.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
136
Example 2:
Learner: This weekend, will paint my house.
Teacher: (Expert reconstruction and on-the-spot correction) No, no: "my house will be
painted". You need a verb in the passive voice: "to be" plus a "past participle".
Learner: OK. This weekend, I will be painted my house.
Teacher: (Authoritative explanation. The teacher begins to suspect her interpretation is wrong)
Will you paint it yourself?
Learner: (Authoritative explanation) Yes, I will be painted it.
Teacher: Oh! Then, you meant "This weekend, I will paint my house"!
Learner: Yes. What did I say?
In "This weekend, will paint my house", the learner was simply omitting the subject. His
error had nothing to do with the correct use of the Passive Voice. This example also illustrates
that there are ways of asking for authoritative explanation which do not necessarily entail
asking “What do you mean?” or any other direct question about the learner‟s error.
Example 3:
Learner's written production:
"Dear John,
I am very happy what you are getting married ...............etc."
Teacher's corrected version:
"Dear John,
I am very happy to hear that you are getting married ............ etc."
Learning from Learners‟ Errors Ana María Rozzi de Bergel
137
Expert reconstruction by the teacher: The learner tries to use what as the relative pronoun
that. This is interference from Spanish, but he is unaware of the difference between qué and
que in his native language, too.
Authoritative explanation by the learner: "My previous teacher told me that, so as not to
confuse „what‟ with „that‟ I had to translate „what‟ as „lo que‟ and „that‟ as „que‟. For
example: „This is what I want‟ means „Esto es lo que quiero‟. In this letter, I meant „Estoy
feliz, lo que vas a casarte‟. I forgot to write a comma, that's why you did not understand."
This expert reconstruction is wrong on two counts: a) The teacher has hastened to produce a
bold interpretation of an error which she did not understand well enough; b) She has not
effected a reconstruction of the utterance but has edited the learner's letter, providing a
sentence that is "better" than the learner's but does not address the error. We will discuss
editing again later on.
It should be noted that the Spanish sentence on which the learner modelled his utterance
contains a sub-standard use of lo que which the teacher who recommended translation of the
relative what may not have thought about.
Example 4:
Teacher: Can you swim?
Learner: Yes, I possibility swim.
Teacher: (Expert reconstruction) You mean you CAN swim.
Learner: Yes, I can swim.
Teacher: And can you swim well?
Learner: Yes, I possibility swim well.
Teacher: (Expert reconstruction) You CAN swim well, you mean ...
Learner: Yes. And I possibility teach children to swim.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
138
Teacher: (Authoritative explanation) Why do you say "possibility" and not "can"?
Learner: (Authoritative explanation) Because it's the same. My previous teacher told me that
“can" means "possibility".
An inspection of this learner's copybook showed that his previous teacher had made him
copy a list of modals from the board, which read:
can = possibility
must = obligation, assumption
may = permission, etc.
When describing errors, the problem does not lie so much in making the right inferences or
deductions as in getting information from the learners and interpreting it. Error analysis that is
not based on authoritative explanation often tells the researcher more about the teacher than
about the learner.
It would not be possible to show or describe all these aspects of utterances in any type of
evidence that might back up research into errors or include it in a research report which
covered a relevant number of cases. Analyses and discussions are extremely difficult unless all
the participants have had first-hand contact with the errors being discussed and are familiar
with the complete data, as outlined above. Audio or video recordings of lessons, transcriptions
of errors, researchers‟ logs and reports on classroom observation are some of the instruments
which attempt to capture the error and its circumstances and all these records ought to be
available to the readers for them to judge the analyses of errors we are going to present.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
139
Due to these obvious constraints, the examples of errors in this work will not include all the
information about the social, institutional, pedagogical and psychological components of each
error, but the reader can trust that the diagnoses we will offer as examples were made only
after examining all the pertinent data.
2.b Descriptive and diagnostic analyses
Once we have a reliable sample for analysis, we can undertake the study of learners' errors
both for a diagnostic, teaching purpose and for a descriptive purpose, to gain insight into
learning hypotheses, compensation strategies and communication strategies. We will not
propose doing this from a purely scientific point of view, but for the practical purpose of
facilitating teaching and learning. The process is summarised in Figure 7.
Figure 7 – Practical applications of error analysis
ERROR
ANALYSIS
Diagnostic analysis complements
consists of consists of
Descriptive analysis
Evaluation of results
Better results brings about
Description of processes underlying errors
leads to leads to
Adjustments to teaching and
materials.
Exploitation of the learner‟s
curriculum.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
140
The diagnostic aspect of error analysis simply addresses what parts of the language code
have been breached: grammar, lexis, rules of appropriacy or register. It is the kind of analysis
teachers carry out when they are correcting compositions, for example. This enables teachers
to measure their learners' progress, evaluate the materials they are using, reflect upon their
teaching and correct or adapt methods and procedures to obtain better results. It is directly
related to the classroom situation, to examination-taking and to the general results of the
language course. When carrying out diagnostic analysis, the teacher will find out what
language forms have been wrongly used: verb tenses, the wrong exponent of a function, a
mispronounced word, etc. This type of analysis also sheds light on the learner‟s personal
curriculum, by showing what he seems ready to learn and where he appears to depart from the
course curriculum. Both Norrish (Ibid.) and George (1969) have produced important
inventories of errors from a diagnostic point of view and all the materials of common errors in
English are valuable resources in this field.
The other aspect of error analysis, that we have called descriptive, provides insight into the
learners' learning hypotheses, compensation and communication strategies, among other
processes, by unveiling operations used to incorporate and process knowledge, to cope with
problems of learning and to use newly acquired language for communicative purposes in
personal creations. When carrying out descriptive analysis, the teacher will find out if the
errors detected during the diagnostic analysis are due to interference, overgeneralisation,
simplification, or various other processes which tell her not what people are learning, but how
they are doing it. Teachers may also pay attention to the number and frequency of
manifestations of these processes, to establish error trends.
More often than not, descriptive analyses will bear out error-provoking teaching or expose
flaws in the materials, particularly when error types are common to a whole group of learners
Learning from Learners‟ Errors Ana María Rozzi de Bergel
141
or when an error can be clearly traced back to some features of the materials or to something
said in class, as was the case in our example of the learner who produced “I possibility swim”.
We will deal with materials or teacher-induced errors in greater detail later on.
I am going to discuss an example of Diagnostic and Descriptive analyses, to show how they
complement each other, even though I have not fully explained the terminology I am going to
use:
Example 1:
Sample of a learner‟s written production:
Dear Fred,
Thank you for your letter. Peter also send me a letter yesterday.
Is very well, and may be come the next week. He bought a cycle and ride it to work every
day. I would be like that you seen him .................
DIAGNOSTIC ANALYSIS (After expert reconstruction)
SEND for SENT: He has not internalised the use of the Past Simple Tense.
IS for HE IS: Has not internalised the necessity of the subject.
MAY BE COME for MAY COME: Has concluded that modals must be followed by "be".
THE NEXT WEEK/NEXT WEEK: Does not know when to omit the definite article.
CYCLE for BIKE: He knew he could shorten the word "bicycle", but could not remember
exactly how to do it.
RIDE for RIDES: He has not internalised the use of the "s" in the third person singular,
Present Simple Tense.
WOULD BE LIKE for WOULD LIKE: Same as "may be come".
Learning from Learners‟ Errors Ana María Rozzi de Bergel
142
THAT YOU SEEN HIM for "QUE LO VIERAS": He does not know how to say it and effects
compensation by creating an utterance as
best he can.
From this analysis, the teacher knows this learner needs further practice in the use of
certain verb tenses, needs to be reminded of the need to put a subject in most sentences and
must be re-taught the use of articles.
DESCRIPTIVE ANALYSIS (After authoritative explanation)
SEND for SENT:Appropriate, semi-correct. Syntactic, intralingual error, due to simplification
(only one form of the verb for all cases). Non-compensatory.
IS for HE IS: Appropriate, semi-correct. Syntactic, interlingual error (interference). Non-
compensatory.
MAY BE COME for MAY COME: Appropriate, semi-correct. Syntactic, intralingual error due
to adherence to first form learnt. (The lesson in the book where modals were introduced
presented them in connection with "be": will be, may be, should be. He now keeps “be” as a
necessary part of the construction, refusing to flexibilise it). Non-compensatory.
THE NEXT WEEK for NEXT WEEK: Appropriate, semi-correct . Syntactic, interlingual error.
Non-compensatory.
CYCLE for BIKE: Semi-appropriate, correct. Semantic, intralingual error due to confusion.
Non-compensatory.
RIDE for RIDES: Appropriate, semi-correct. Syntactic, intralingual error due to simplification
(The same verb form for all persons). Non-compensatory.
WOULD BE LIKE for WOULD LIKE: same as "may be come".
THAT YOU SEEN HIM for "QUE LO VIERAS": Semi-appropriate, semi-correct, semantic-
syntactic, compensatory utterance created on the model of the Spanish pattern, but the learner
Learning from Learners‟ Errors Ana María Rozzi de Bergel
143
is aware of the fact that he needs a verb form he does not know, and chooses "seen", hoping
this is close to the subjunctive, because he identifies past participles with complex forms. Half
interlingual and half intralingual. The construction is modelled on the Spanish "que lo vieras"
but the choice of verb is an overgeneralisation of the use and meaning of the past participle.
There are actually two errors.
N.B.: The sample of the learner‟s production analysed is too small to venture
any deeper interpretation of the processes behind his interlanguage. At least
fifty errors would be needed, recorded at different moments and in different
contexts, to attempt a preliminary in-depth analysis. However, we will offer a
mock interpretation: the learner relies on the code of English for most of his
utterances, resorting to the code of his native language only when there are
gaps in his learning. His errors are mostly syntactic, not semantic. Both
trends are productive, as we will discuss later on, but some of his intralingual
errors are due to non-productive processes, like adherence to first form
learnt or simplification, which are attempts at simplifying learning by taking
shortcuts. The materials need special treatment in those areas where they
present several analogous items together. Activities should encourage
inductive processes and language rules should be discussed with the learner
when he accesses them, to point out exceptions and special applications.
Although the example is extremely limited, it shows the wealth of information
to be obtained by combining the two types of analyses.
The purpose of the descriptive study is not to grade or score the learner's production, but to
learn more about him and view his progress in terms of whether certain error types tend to be
replaced by others, which we may consider more conducive to learning. In the example, if a
Learning from Learners‟ Errors Ana María Rozzi de Bergel
144
second analysis of the learner‟s written production showed the same number of mistakes, even
some affecting the same language forms, but no adherence to first form learnt and no
simplification, which are, in general, operations used to simplify the process of learning or the
object of learning, we might assume that the learner has made progress – not in his command
of the language, but in the mechanisms he is setting in motion for learning. This type of
progress will later produce an improvement in his language, as well.
To detect learning hypotheses, styles or strategies through error analysis, about fifty errors,
belonging to the same learner, should be collected and recorded in the manner we have
explained, over a period of at least three months. This avoids focusing on a learner‟s
production at a very specific point in time and making generalisations on limited data.
2.c Error classification in the descriptive analysis
Errors do not fall neatly into any one category, but should be classified using criteria which
describe their different aspects. Classifications of errors by Selinker (Ibid.), Richards (Ibid.)
and Pit Corder (Ibid.), attempt to classify errors according to the element which the author of
the classification considers crucial: the semantic meanings, the grammar rules breached, the
degree of appropriacy of the utterance or even its strategic value. James (1998) offers a
comprehensive review of error taxonomies, ranging from those produced by grammarians who
have listed types of deviances, to those who make allowances for ambiguity by describing the
type of foreign language targets the error approximates: when an error occurs, the linguist
considers several possibilities for what the learner intended to say (Legenhausen in James,
1998)
James (Ibid.) discusses comparative taxonomies and communicative effect taxonomies, the
former dealing with error causes in terms of language deviances and the latter with error
Learning from Learners‟ Errors Ana María Rozzi de Bergel
145
gravities in terms of whether the communicative purpose of the utterance has been achieved
(Dulay, Burt and Krashen in James, 1998) He goes on to describe these taxonomies separately,
including breaches of the grammar and lexical codes in comparative taxonomies and lists a
number of alterations made by learners to surface structure features. However, he then goes on
the describe and discuss mental or linguistic processes leading to error production, such as
omission, overgeneralisation, ignorance of rule restriction, and others already described by
Richards (Ibid.). He disagrees with considering these processes learning strategies, making an
important point: how can ignorance of a system be considered a strategy for learning that
system, as in, for example, ignorance of rule restrictions? James (Ibid.) further questions the
identification of error-producing processes with learning strategies with reference to “system
simplification”, holding that rather than a strategy for learning, it is a strategy for simplifying
the object of learning – a point we also make when talking about productive and unproductive
errors.
Dulay, Burt and Krashen (Ibid.) propose a blend of the comparative and effect taxonomies as
a means of describing all the aspects of errors: the formal and the communicational. James
(Ibid.) suggests that a third dimension should be added to this description: the number of
errors, a quantitative dimension.
I will use a method of classification which
a) takes pedagogic considerations into account, by rating errors by the degree of
achievement they reflect, according to the criteria of the communicative approach,
b) measures form and meaning as two aspects of language use, closely linked but having
an independent value as well,
c) distinguishes between errors produced by interference from the native language
(interlingual) and those made within the code of the foreign language (intralingual),
Learning from Learners‟ Errors Ana María Rozzi de Bergel
146
d) adds new categories to the types of intralingual errors, which we will consider
extremely important for unveiling inner learning hypotheses and needed to define
intralingual errors better and more precisely,
e) distinguishes between compensatory and non-compensatory utterances and errors, for
pedagogical purposes.
This set of criteria was created with the focus on the research context, subjects and purpose:
the errors to be analysed in the framework of the classroom, to describe the performance of
foreign language learners, in order to improve the quality of instruction and by doing so, we
hoped to improve the quality of learning. Within these parameters, the resulting taxonomy was
more than three-dimensional, as proposed by James (Ibid.); perhaps multi-dimensional, as it is
quantitative, descriptive, communication-oriented, pedagogic and qualitative.
The table used for classifying errors is shown below.
Figure 8 – Error classification
CATEGORIES SUB-CATEGORIES
ERROR
RATING
Appropriate, semi-correct
Semi-appropriate, correct
Semi-appropriate, semi-correct
Inappropriate, incorrect
ERROR
TYPE
Semantic
Syntactic
Semantic-syntactic
ERROR
CLASS
Interlingual
Intralingual
CAUSES
OF
INTRALINGUAL
ERRORS
Accessing attempts
(Overgeneralisation, confusion,
misunderstanding)
Simplification (Of the system,
of the message, of the syllabus)
Shortcutting (Cue-copying,
conditioning, adherence to first
form or meaning learnt)
Compensatory utterances/errors
Learning from Learners‟ Errors Ana María Rozzi de Bergel
147
The classification of an error should include all the pertinent categories and sub-categories.
In the descriptive analysis shown in the previous section, for example, the use of RIDE for
RIDES in the example of descriptive and diagnostic analyses I have presented was classified
as an appropriate, semi-correct, syntactic, intralingual error due to simplification (The same
verb form for all persons). Non-compensatory.
To avoid writing these lengthy classifications of errors, we devised a code. In our research,
the utterance we have just presented would have been coded 30-020-10b. (3 = appropriate,
semi-correct; 0 = non-compensatory; 020 = syntactic; 10 = intralingual; b = simplification)
This is just an example of how codes can help the teacher classify errors. When describing our
classification table, I will also provide the suggested code for each category and sub-category,
in case the reader wishes to learn how to use this coding system. My experience in this field
shows that if the teacher needs to record and classify an error produced in an ongoing lesson, it
saves time to write a code rather than all the labels. The context of the error is as important as
the error itself or more, so even if the utterance is written down, a classification attempted
after the lesson may be unreliable, as details of the context or the authoritative explanation
may have already been forgotten. On-the-spot classifications are often necessary and codes
facilitate them.
In the following section, I will now describe all the categories and sub-categories in the
classification table and the criteria used for creating them. These criteria were chosen for this
particular research and are not necessarily applicable to other studies of errors. They reflect the
decisions which were made along the process, on the basis of the theory available at the
moment.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
148
2.c.1 Error Rating
The first category is called RATING, because it defines the communicative value of the
utterance, which is taken into consideration for scoring learners‟ production, for example, at
tests, within a predominantly communicative approach where fluency and appropriacy are
ranked higher than correctness. Please note that in this category, we are analysing whole
utterances, not specific language items. An utterance, here, is a group of words constituting a
unit of meaning, a sentence with a finite verb or a verbless clause such as “Good luck!”
The parameters in this category are correctness and appropriacy. (Widdowson, Ibid.) As
explained, we will consider that an utterance is appropriate when it is produced in the right
register, respecting role relationship, when it fits the context in which it occurs, performs the
intended function and carries its meaning across satisfactorily, thus catering for the language
user's communicative needs. If this utterance has been correctly constructed according to the
rules of English grammar and syntax, and lexis has been accurately used, we say it is also
correct.
In this communicative view, it is possible for an utterance to be fully appropriate and semi-
correct, but it is not possible for an utterance to be totally correct and inappropriate, for
grammar is not considered as a pattern for putting words together, but as a part of meaning,
subordinated to appropriate language use. The sub-categories in Error Rating are as follows:
a) Appropriate, semi-correct. (Code 30) The utterance conveys its meaning satisfactorily,
is suitable to the context, is in the right register, but contains an error in grammar.
Example: (John and Mary are having dinner. Mary wants the salt)
Mary: Would you pass me the salt?
John: Certainly. Here you is.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
149
Although semi-correct, the utterance is appropriate, for it conveys its meaning, it is in the
right register, it performs the function for which it was intended and thus caters satisfactorily
to the user's communicative needs. It is the type of utterance a teacher would not correct on the
spot, not to interrupt communication during an activity, or perhaps would rather not correct at
all, but jot down for future exploitation. At an examination, it would obtain a higher mark than
the utterances in the next category, if communication were ranked above correctness in the
teaching-learning approach being used.
b) Semi-appropriate, correct. (Code 20) The utterance has been correctly constructed, but
it does not fully fit the context, or it contains lexical or other errors that make it
partially inappropriate according to the definition above.
Example:
Mary: Would you pass me the salt?
John: Certainly. It is here. (Meaning "Here you are")
Note that only the physical action of giving Mary the salt would denote that John has
understood the request and is willing to comply with it. In the classroom, this would depend
on the degree of realism of this kind of practice in situations. The textual meaning of his
utterance is not clear. Again, if the error occurred during a communicative activity, the teacher
might take it down for future exploitation, as it does not significantly hamper communication,
and store for future exploitation.
c) Semi-appropriate, semi-correct. (Code 10) An utterance that contains errors in
grammar and does not convey meaning satisfactorily, but is not totally
incomprehensible and to some extent caters for the user's communicative needs, for it
partially fits the context in which it occurs and the language user‟s objectives:
Learning from Learners‟ Errors Ana María Rozzi de Bergel
150
Example:
Mary: Would you pass me the salt?
John: Welcome. Here are you. (Meaning. "Certainly. Here you are")
As in the previous example, only the physical action and the intention, the connotation,
makes his meaning clear, but in this case, there are also mistakes in grammar and lexis.
d) Complete failure to communicate, or inappropriate, incorrect utterances:
(Code 00)
Communication may be interrupted due to lack of comprehension, a combination of
inappropriate, incorrect elements or an utterance rendered incomprehensible by serious lexical
or grammatical errors.
Examples:
Mary: Would you pass me the salt?
John: It's half past ten.
Although superficially well-formed, this utterance interrupts communication and cannot be
considered correct, if we adhere to the concept that grammar must carry meaning and this
meaning must serve communicative objectives. Special consideration should be given to the
fact that John may CHOOSE to say "It's half past ten", to interrupt communication on
purpose, because he is angry with Mary, or because he wants to remind her that it is time to
stop eating and go home, for instance. In this case, his utterance would be fully appropriate
and correct.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
151
Example:
Mary: Would you pass me the salt?
John: (Before performing any action) Excuse me salt give you.
Mary cannot know, until he performs some physical action, whether she will be getting the
salt or not. Communication has been totally interrupted by grammatical and lexical errors.
Even if the action were performed, the rating of this error would not change, as we are not
analysing communication as a whole but language within communication.
e) There is a category of error that we shall not include in this analysis because it is very
difficult to detect and extremely rare: the error that does not look like an error, because
the utterance is correct by chance.
Example:
Mary: Would you pass me the salt?
John: (Giving her the salt) Certainly. Here.
John has not said "Here" because he knows this is correct, but because he meant to say
"Here you are" and could not remember the rest of the expression. John's utterance seems
correct, but it is an error, because he is not aware of the fact that he has produced an
appropriate and correct utterance, and will probably never repeat it. The utterance actually
shows a gap in his command of English which is difficult to perceive.
The concept of right by chance was first stated by Pit Corder (1974) and he even
recommended that all learner utterances be considered erroneous until proven correct. This is a
rather apocalyptic view of learner interlanguage, but teachers should train themselves to be
excellent listeners and ask for clarification when learners produce sentences with structures
Learning from Learners‟ Errors Ana María Rozzi de Bergel
152
they have not yet been exposed to and presumably do not master. There may be two
explanations: they have learned the form elsewhere, or it is correct by chance.
In one of the courses under observation during our research, a learner at a fairly elementary
level said: "My father is to work here", when he meant "My father works here". He had
misinterpreted the rules for the formation of the present simple affirmative, but the resulting
utterance was superficially well-formed and seemed correct, semi-appropriate, when it was
actually semi-correct, semi-appropriate. It could not be concluded that this learner new the use
of the verb “to be” to denote a future event which has been programmed and he would not use
it with that meaning – in fact, he was surprised when the teacher explained to him what he had
actually said.
2.c.2 Error Type
The second category is TYPE, where we classify erroneous utterances according to the
language area they affect the most, into Syntactic, Semantic and Semantic-syntactic. The
parameter here is: does the error affect meaning at face value or does it affect form, not
harming meaning seriously? If it affects meaning primarily, the error will be classified as
semantic. If form is affected but meaning is conveyed in a satisfactory manner, the error will
be considered syntactic. I talk about affecting meaning or form at face value because what was
considered important here was the result of the learner‟s production; for example, a
phonological error such as “The sheep (ship) sailed at dawn” will be considered lexical
because the effect of the error is that the listener heard the wrong concept, even if a
sympathetic listener would probably understand that the speaker meant “ship”.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
153
a) A semantic error (Code 200), here, is an error in lexis, such as using one word for
another, the use of an expression in the wrong context, the use of the wrong register,
the wrong pronunciation or intonation.
Examples:
The learner said ...................................... when he meant ..................................
Reporters questioned the car racers. Reporters interviewed the car racers.
Frank works in a great factory. Frank works in a big factory.
Please, madam, may I help you? Excuse me, madam, may I help you?
Dad, behold the fair fish I've caught!
(The learner loved reading Sir Walter Scott)
Dad, look what a fish I've caught!
The following are errors in pronunciation and stress which were considered semantic,
according to the criterion I have explained. Some of the errors in the examples are so serious
they may bring about complete failure to communicate. The diagnostic analysis will have to
describe them as phonological, however, so that classroom exploitation of the error may be
correctly focused.
Examples:
What the learner said sounded like…….. when he actually meant ………..
The king lost the battle on the bitch. The king lost the battle on the beach.
He bought a ship. He bought a sheep.
I am in the semen industry. I am in the cement industry.
The male man has come. The mailman has come.
The beds in our hotel have double shitting. The beds in our hotel have double
sheeting.
He is the major in our town. He is the mayor in our town.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
154
b) A syntactic error (Code 020) is, in this classification, an error in grammar or syntax,
affecting the structure of the language, having to do with the correct superficial
formation of the utterance. It affects form at face value.
Examples:
The learner said ............................. when he meant....................................
A lot of people was watching the race. A lot of people were watching the race.
He likes the fish. He likes fish.
At what time leaves the train? What time does the train leave?
When he leave the hotel, he walk down.... When he leaves the hotel, he walks down ....
What did do you on Saturday evening? What did you do on Saturday evening?
Then I am going to read newspaper. Then, I am going to read the newspaper.
We included spelling and punctuation in this category, unless they affected meaning. "They
gave her a price" for "They gave her a prize", would have been considered semantic even if
the learner had explained it as an error in spelling, whereas "They gave her a prise" for "They
gave her a prize", would have been considered syntactic. A typical phonological-syntactic
error was the absence of the final /d/ or /t/ sound in regular past tenses, which made the word
sound as if the learner had produced a bare infinitive.
Authoritative explanation often disclosed that the learner thought he had pronounced this
final sound when he had not, simply because Spanish, with few exceptions, does not have /d/
or /t/ sounds at the end of words, and Spanish speaking learners of English never produce
those sounds audibly enough. The phonological difficulty resulted in a semi-correct utterance.
c) A semantic-syntactic error (Code 220) is at the same time a breach of the grammar
code and an error in lexis or in appropriacy. Both form and meaning are affected.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
155
Examples:
The learner said.................................. when he meant ...................................
In general cases, people says .............. In general terms, people say ....................
I things it is important ....................... I think it is important .........................
What did house? Where was the party?
I couldn't think so. I don't know.
Exist ignorance of the price. We don't know the price.
Can you indication me how to do it? Can you show/teach me how to do it?
It might seem that syntactic and grammatical errors are the same thing, and that semantic
errors always result in semi-appropriate or inappropriate production, but there are a few
exceptions, so it is not superfluous or redundant to make this distinction.
The distinction between appropriacy and correctness versus semantic and syntactic accuracy
is also relevant in that appropriacy and correctness are results, not causes, of errors.
Example:
Mary: Would you pass me the salt?
John: Welcome. Here you are.
John has confused the exponents of two language functions. He says "Welcome" when he
means "Certainly". This confusion is a semantic error that results in an inappropriate
utterance.
Appropriacy is breached as a result of a semantic problem, and it could equally be breached
as a result of a syntactic problem:
Mary: Would you pass me the salt?
John: Yes, I would pass you the salt.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
156
Mary will be totally puzzled by this answer, or she will think John is trying to be funny, or
that he is half-deranged, but what actually happens is that John does not know what to say to
agree to Mary's request, and thinks that if he answers her question literally he will be
expressing this agreement. Besides, he is not very proficient in the use of the modal would, but
knows "I would like some tea", which he associates with "Would you pass me the salt" and
other polite expressions, so he feels that he is more or less safe if he frames his answer as he
does; on the other hand, his teacher is constantly forcing him to answer questions in
"complete" sentences ("Did you go to the club on Saturday?" "Yes, I went to the club on
Saturday") so he sees nothing wrong with repeating what Mary said.
This classification of errors into semantic and syntactic, done according to the manifest
results of the utterance, is debatable, and the two categories thus described are rather broad.
However, the distinction serves the purpose of focusing on observable language behaviour and
on communicative language use. An in-depth revision of these criteria may prove relatively
futile, as the truly interesting information about learning hypotheses is to be found elsewhere;
for example, in the sources of intralingual errors. For pedagogical purposes, the relative
number of semantic and syntactic errors may show a learner‟s potential or give information
about the course: if most errors are semantic, the learner is facing more communication
problems than if his errors are mostly syntactic, but this may be due to his learning hypotheses
or to flaws in his placement, in the approach chosen to teach him or in the teaching style he is
being exposed to, so when semantic errors outnumber syntactic ones, the situation should be
carefully analysed.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
157
2.c.3 Error Class
The third category is CLASS, where we group errors according to their code of origin:
whether they were originated within the code of the language being learnt or whether they
were caused by the learner's attempt at trying to structure the foreign language on the model of
his native language. Errors, here, fall into two sub-categories:
a) Interlingual, (Code 01) with utterances in the foreign language constructed according
to the rules of the learner‟s native language, in cases where these rules are not identical
in both languages and therefore cannot be transferred from one to the other.
Examples: (Interlingual errors of Spanish-speaking learners of English)
The learner said ................................. when he meant ................................
I have twenty years.(Tengo veinte años) I am twenty years old.
We saw to Mary.(Vimos a Mary) We saw Mary.
The important is to study the problem.
(Lo importante es estudiar el problema)
The important thing is to study the
problem.
Interlingual errors are usually called errors due to negative transfer or language interference
(Selinker, Ibid.). The learner is thought to resort to the code of his native language either
because he does not master the necessary rules of the foreign language or because he still finds
the code acquired in the first place more reliable for effective communication. This rather
simplistic view of the relationships between L1 and L2 was replaced, in the 1970s, by the
more dynamic and humanistic idea of an interaction between two languages based on the
learner‟s perceived distance between the two, in a series of papers by Kellerman, who rejected
the notions of interference or transfer, and preferred to talk about cross-linguistic influences,
ruled by a process he called psychotypology.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
158
Our research has many points in common with Kellerman‟s psychotypology and the
influence of the native language on the foreign language, which is based on the learners‟
judgements, strategies, and views of language and the construction of a language system.
Analysing the two languages would not be enough, then, to gauge the influence one can have
on the other. It would be more enriching to analyse the person‟s perception of language and
how it should be structured. (Kellerman & Sharwood Smith, 1986)This is the kind of
information we have sought to obtain through the Learning Hypothesis Test and descriptive
error analysis.
The second factor influencing cross-linguistic influence is learners‟ perception of the
distance between the two languages and the third one is their actual knowledge of the foreign
language. (Kellerman, as cited in Gass and Selinker, Ibid.). Why and how learners use
psychotypology to effect language transfer is still a topic of research. Gass & Selinker (1992)
and Ringbom (1987) consider the learner as an active and conscious participant in the learning
process, making comparisons between the two languages as a default learning strategy.
Beginners are supposed to use this strategy more often, as they tend to believe that the
distance between the two languages is very short, but as they learn more about L2, they begin
to realize it is longer, and do not rely on transfer so often. Conversely, some learners tend to
consider the distance between the two languages much longer than it actually is, and reject
cross-linguistic influences altogether. This leads them to make mistakes in areas where the two
languages are identical and transfer would have led them to acquire the correct form.
A very peculiar form of interference occurs when a person who speaks several languages
attempts learning one more. In this case, interference hardly ever comes from the learner‟s
native language, but rather from the language whose grammar the learner has internalised the
least, or monitors the most. Interference comes from the language whose system the learner
Learning from Learners‟ Errors Ana María Rozzi de Bergel
159
uses most consciously. It has been claimed that this type of cross-linguistic influence occurs
when a learner feels that a language rather than his own is a better candidate for transfer when
learning a third language (Ringbom, Ibid.) This may happen because declarative knowledge is
more readily available in the second language than in the first, if we are to adhere to the view
that transfer is a conscious process.
During our research, we often discovered what foreign language a learner had acquired
before English by analysing his errors. The following examples were recorded during lesson
observation:
Examples:
(Native speakers of Spanish who speak French and then learn English)
The learner said .................... when he meant ................................
Give me of coffee. (Donne moi du café) Give me some coffee
I attend the bus. (J'attend l'autobus) I wait for the bus.
It is not clear whether cross-linguistic influences are always the result of conscious
strategies. If they vary according to psychotypology, depending on the learner‟s perception of
the distance between the two languages, they seem to have a sub-conscious origin. The first
factor in Kellerman‟s explanation of psychotypological factors, the learner‟s style for
structuring L1, is not at all conscious. We may then talk about conscious strategies which
work on the basis of unconscious processes leading the learner to rely on cross-linguistic
differences and similarities to a greater or lesser extent, depending on his learning hypotheses.
b) Intralingual , (Code 10) with utterances which denote an effort to adhere to the code
of the foreign language, at grasping the rules of the system being learnt.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
160
Examples:
The learner said ............................ when he meant ............................
I go to school yesterday. I went to school yesterday.
He cans goes. He can go.
I have two book. I have two books.3
He is come at two. He comes at two.
Some errors have an intralingual and an interlingual component. When this happens the
teacher should take the root problem as the source, or record the error twice. A typical case
was the statement "My dog had sick", discussed in the sample Communication Strategies
interview. The real origin of the error was interlingual but it had resulted in an intralingual
problem. The learner was trying to model the utterance on the code of the native language, but
he had confused "to have" and "to be". Note that a researcher not resorting to authoritative
explanation might have even thought that if the learner had resorted to Spanish to construct the
equivalent of "Mi perro estuvo enfermo" in English, he would not have made an error.
Knowing its sources, this error should be considered interlingual.
As we can see, in this category we are no longer concerned with the superficial effects of the
utterance but with the processes that produce it. We are moving into a different facet of the
analysis.
3 The use of a Spanish-based construction would have prevented this error.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
161
2.c.4 Causes of intralingual errors
The fourth category is actually a sub-category, but it is of paramount importance in revealing
learning hypotheses and diagnosing language ceiling. It comprises the causes of intralingual
errors.
a) The most common cause of intralingual errors is the learner's attempt at grasping the
new language code, at coping with the difficulties of textualizing his messages in the
foreign language. We will call these operations Accessing Attempts, which are affected
by overgeneralisation, unawareness of rule restrictions, misinterpretation or confusion
(Code a). These categories were outlined by Richards (Ibid.) as separate types of
errors, but we will group them together because they are an attempt at accessing the
code of the foreign language and reflect highly productive operations or processes.
Examples:
The learner said .... when he meant .................................
He cans speaks English. He can speak English. (The learner is trying to
grasp the rule for the third person inflection of
the present simple, but he places all verbs in the
same category, including the modals)
He should have came. He should have come. (The learner mistakes the
past form of the verb for its past participle)
We did saw it. We did see it. (The learner remembers he can use
the auxiliary verb in the affirmative sentence for
the sake of emphasis, but not that he should not
use the main verb in its past simple form)
How many children does he
has?
How many children does he have? (The learner
extends the use of the third person singular
inflection in the present to the two verbs
involved)
Learning from Learners‟ Errors Ana María Rozzi de Bergel
162
Why are you telling me all
these?
Why are you telling me all this? (He is being told
many things, so he thinks he has to use a plural.
Had he structured this utterance on the model of
the Spanish equivalent, he would not have made
an error.)
The boys are in the living-
room are watching TV.
The boys are in the living-room ,watching TV.
(He has learnt the use of "to be" in utterances
such as "The boys are in the living-room", and he
has grasped the formation of the present
progressive as a chunk of "to be+present
participle". He cannot yet combine the two
concepts in this utterance successfully. Note that
if he had constructed this utterance on the model
of the Spanish construction, he would not have
gone wrong.)
What did they happen with? What happened with them? (The learner thinks
that all prepositions must be at the end of
questions, overgeneralising the rule for questions
such as "What did you do it with?" or "What are
you talking about?" He would not have gone
wrong if he had resorted to Spanish.)
Who can coming? We must
decide.
Who can come? We must decide. (Incomplete
application of rules. He uses the bare infinitive
after “must” but not after “can” because he has
not yet grasped that the bare infinitive is to be
used after all modals. Maybe the category of
“modals” is not clear to him, either)
The learner who overgeneralises a rule or uses it where it does not apply, or confuses two
rules, is anyway trying to grasp the foreign language system, he is trying to get to the
Learning from Learners‟ Errors Ana María Rozzi de Bergel
163
generative rules on which to base his language production - in other words, he is trying to
form conceptual categories and organise them into a system. All he needs is time, guidance
and communicative, purposeful practice.
Overgeneralisation, Unawareness of Rule Restrictions and Incomplete Application of Rules
(Richards, Ibid.) are closely related concepts, and the extent to which they affect the learner‟s
production is determined by the learner‟s conceptual classification style, on which they often
shed light. We have already stated our assumption that learning and subsequent acquisition
depend largely on the learner‟s style for placing concepts into conceptual or pre-conceptual
categories, for ordering these categories into hierarchical systems and for finding or creating
associations between different language forms. Ordering categories hierarchically means being
aware of the fact that some categories include others, for example, within the category verb,
we can find the sub-categories auxiliary, irregular, non finites, etc. Learners have to identify a
word as a verb before including it into any of the other sub-categories.
Conceptual categories can be precise, (A chair is a kind of seat, usually with four legs and a
back – dictionary-like definition); insufficiently restricted (It is a seat – Stools, benches and
other objects fall into the same category); insufficiently comprehensive (It is a seat with four
legs and a padded back – Chairs with no padding or with fewer legs are left out of the
category); or vague (It is a piece of furniture – The category is so broad that it would be
impossible to know it includes only chairs).
The following examples and their analyses illustrate how conceptual categorisation relates to
over-generalisation, unawareness of rule restrictions, incomplete application of rules and how
our analysis changes if we carry it out considering categorisation styles as learning
hypotheses. Once again, I would like to reassure the reader that the sentences in these
Learning from Learners‟ Errors Ana María Rozzi de Bergel
164
examples were analysed after authoritative explanation and taking all the elements of valid
data for error analysis into consideration.
Examples:
Utterance Learner’s personal grammar or structuring of the foreign
language.
My brother cans goes
with you.
(Learner-generated rule: Third person singular verbs in the
present take an "s" - "can" and "go" belong to a broad category:
"verb". The category is insufficiently restricted, and the learner is
relative unaware of rule restrictions, due to his particular style for
creating broad categories.)
I have spoken to Mary.
She says she had never
saw that car until she
saw it parked in front
of her house yesterday.
(Learner-generated rule: The present perfect is formed with the
auxiliary "have" in the present and the past participle of the main
verb. Category "had" belongs to: past verb. The category is
insufficiently comprehensive, for it does not include the concept
of "perfect auxiliary" or "perfective aspect", so the learner does
not believe that the rules for the formation of the present perfect
apply to the formation of the past perfect, and resorts to "saw"
because it is also a past verb, which he places in the same
category as "had".)
He informationed me
of that.
(Learner-generated rule: All words that refer to an action take
"ed" in the past tense. Category "information" belongs to: word
that refers to an action. The category is vague. Any word can
"refer to an action". The learner needs to build up the categories
of noun, verb, adjective, etc.,)
The use of insufficiently restricted and insufficiently comprehensive categories is not
uncommon, for very few people handle precise logical categories and clear concepts all the
time and even the possibility of defining categories clearly has been questioned, but frequent
recourse to vague concepts or categories may hamper learning if the teacher does not lead the
Learning from Learners‟ Errors Ana María Rozzi de Bergel
165
learner to define categories better and make his concepts more precise. Unfortunately, when
the use of vague logical categories is coupled with great imagination and creativity, the learner
will show a tendency to jump to hectic conclusions and to make wild associations. In these
cases, teaching will have to provide guidance and organisation.
Creative, imaginative individuals often make very poor language learners when their
imagination and creativity do not rest on a strong conceptual system, and the fashion of
stimulating creativity and imagination in the language class at all costs is actually a double-
edged weapon that should be used with great care and together with activities for the
development of the learners' conceptual thinking.
b) Intralingual errors due to simplification (Code b) occur when learners do away with
annoying redundancies, nuances of meaning, exceptions to rules, second meanings to
words; in brief, when they suppress all the elements of the linguistic code they find
"unnecessary" or difficult to cope with.
The most important learning hypothesis of simplifiers is to close systems before they
are complete. The system of "plural of nouns", for instance, is complete with the
addition of a quantifier to the noun, after a process of reasoning which resembles the
following:
I. "two" is a plural.
II. "Two book" is a plural.
III. "This two book" is a plural.
This is simplification of the language code, but sometimes we find simplification of the
message: the learner keeps only the semantic part of the system, and syntax is
neglected, shortened or left out altogether.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
166
Examples:
The learner said ................... when he meant ........................................
Yesterday, we go to the theatre. Yesterday, we went to the theatre. (There are two
markers of past time in the fully correct version of this
utterance: "yesterday" and "went". The learner eliminates
the most troublesome one.) The language code is
simplified.
He come every day. He comes every day. (Only one marker of frequency
"every day" and one marker of third person singular are
left.) Simplification of the language code.
Which platform is the platform? Which platform is it? (The learner suppresses an
operation: the substitution of the second noun for a
pronoun) Simplification of the language code.
One room, single room, please. A single room, please. (The learner has resorted to the
minimum possible language operation - "Tharzan
English" - to get his message across. We may call this
Simplification of the Message. Language is practically
stripped of the complexities of syntax )
Ill, you? Are you ill? (Another Simplification of the Message)
Old? How old are you? (Simplification of the Message is taken
to such an extreme that it may prevent communication
altogether)
Alongside simplification of the language code and simplification of the messages, that we
have already discussed, we can recognise a third type: simplification of the syllabus, which is
frequent in cases of unsuitable materials. The learner finds it impossible to cope with the
avalanche of structures and vocabulary, either because the design of the materials is poor or
because the course is too advanced for him, and he simplifies the contents of the materials.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
167
Example: After a lesson in which five Modals were introduced and practised together, all the
learners in a class of four started using "have to" instead of all five,
Learners said…. when they meant …..
I have to swim I can swim.
She has to go. She should go.
We have to come. We must come.
You have to pass the exam if you studied You would pass the exam if you studied.
He has to work here tomorrow. He will work here tomorrow.
In some cases, the simplification process was not very easy to detect, for only through
authoritative explanation could we know that the learners who were saying "She has to go", an
apparently correct utterance, were using have to to avoid using should, and thus incurred
simplification of the syllabus.
This type of simplification is not so much an error trend as a defence strategy, and the
learner will stop using it if the materials are changed or taught in a more adequate fashion.
Although it is part of the descriptive analysis, it is heavily diagnostic.
When simplification is a result of learning hypotheses or a feature of the learner‟s personal
curriculum, it produces a highly unproductive type of error, for it shows learners are trying to
simplify the system to be learnt. Eliminating, rather than incorporating, and making do with
primitive forms of communication point to limitations in the learner's language awareness and
may forecast a low language ceiling for the person effecting these operations consistently. It
resembles the process of Shortcutting that we explain below, but the difference is that whereas
in Simplification the learner is trying to modify the object of learning in Shortcutting he is
trying to devise ways and means of simplifying the process of learning.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
168
c) Shortcutting: Cue-copying, Conditioning, Adherence to First Form or Meaning Learnt
(Code c). Although these are different processes, we have placed them in the same sub-
category because we are mostly interested in the learning hypotheses and attitudes they
reflect, which are common to all three: stagnation, passivity, reluctance to develop a
system of rules, and sometimes associating by proximity. The language system is not
misused or simplified; rather, there is an attempt at finding formulas and recipes for
side-stepping the effort of building up a language system. Sometimes, the learner does
not try to cope with the system of the foreign language because he is relatively
unaware that language IS a system. All these process entail taking shortcuts at
mastering the foreign language, hence the name we have given to this category.
Examples of Cue-copying:
The learner said............. when he meant ......................
- Who wants some coffee?
- I want.
- I do. (The learner explained he could not use a word in
the answer that did not appear in the question)
- What do you want to do?
- I want to do go to a movie.
- I want to go to a movie. (The learner explained his
teacher had asked him what he wanted TO DO)
- How much does George earn?
-George earns how much
$ 3,000.
- George earns $ 3,000. (The learner explained he had
been asked HOW MUCH George earned)
- What gate the flight
departure?
(The learner was looking at the chart of ARRIVALS and
DEPARTURES at an imaginary airport, which he was
supposed to complete with information he had to get
from another learner)
- What does he have to do?
- He have to study more.
- He has to study more. (The learner explained that the
question said "he have")
- Where may those who are
successful work?
- Those who are successful may work abroad.(This was
an answer to a question in a written test. Learners were
Learning from Learners‟ Errors Ana María Rozzi de Bergel
169
- Those who are successful
work may work abroad.
mislead by the distance between the modal and the bare
infinitive, and tended to attach the latter to the subject,
that was closer to it.)
In cue-copying, then, the learner tends to reproduce the stimuli rather than respond to them
in a creative way. Certain elements are included in the utterance "because they are there". A
typical case of teacher-induced cue-copying occurs when the teacher prompts the learner the
correct form when he makes an error, as in the following case:
Learner: Tomorrow, he going to study for his exam.
Teacher: is ....
Learner: Yes, he going to study for his exam is.
Cue-copying, even when it is training-induced, will develop more often in those learners
who associate by proximity and not by analogy. This is greatly favoured by teaching, methods
and approaches that rely too heavily on repetition, imitation, rote learning or mechanical
stimulus-response activities. The learner's production will be aimed at imitating rather than
exploring and generating, at following classroom procedures rather than coping with
communication requirements. Because of this, cue-copying was very high in audio-lingual
approaches, where it produced an "illusion of progress" that misled teachers into believing that
learners were learning a lot of English, when they had actually become proficient at going
through the steps of their highly programmed and standard lessons, but were at a loss to
perform in open-ended situations. The word “cue-copying” was borrowed from Selinker
(Ibid.), but we gave it a completely different meaning. Selinker used it to refer to sounds or
words students pick up from their teachers‟ particular regional or idiosyncratic accents, a
concept similar to transfer of training, but referred to a more involuntary process.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
170
Adherence to First Form or Meaning Learnt occurs when the learner clings to elements
learnt - or taught - at early stages of the course and refuses to attach further meanings or uses
to them, or even carries them over into new structures. It is an unproductive type of error, for
the learner refuses to expand or modify his early knowledge, or to gain access to more
complex language forms.
One wonders if this does not happen because the order of presentation of language items in
textbooks is always based on linguistic analyses and not on a serious study of the learners'
personal curriculum, where these items may appear in a different order. Gradualism and the
assumptions regarding grading of difficulty may account for most of these errors, too.
Examples:
- Learners who are taught the verb to be at an elementary stage tend to carry it over to the
formation of the present simple, and produce "I am go to school every day". This is
Adherence to First Form Learnt.
- First Meaning Learnt is often fossilised in words such as room. The learner learns room as
in "There is a table in the room", and then refuses to incorporate "There is no room for so
many people on this bus" and prefers to say "There is no space" or "There is no place".
Because he adheres to the first meaning learnt and refuses to attach further meanings to the
word, the learner has to resort to interference from his native language to construct his
utterance.
Further examples:
The learner says ...... when .....................
Where is you from? "is" was taught before "are"
Can you typing letters? "can" was introduced after the Present Continuous.
Look at she! "she" was taught before "her"
Learning from Learners‟ Errors Ana María Rozzi de Bergel
171
He is come at nine. "is " is the marker of Present Tense he knows, and he carries it
over to other verbs.
He was write a letter
yesterday.
the past of "to be" was taught first, as a marker of past tense,
and the learner carries it over to other verbs.
I love you going to the
theatre
(I love going to the theatre.) the learner picked up "I love you"
as a language unit, a chunk, in the first place. When it came to
learning "I love (doing something)" he refused to drop the
"you".
Have you been written the
note?
(Have you written the note?) The present perfect was
introduced and practised with the verb "to be". The learner
took this to be a unit, a chunk.
These errors are normal at the moment of incorporating a new form within a sub-system that
the learner thought was closed. We can only talk about Adherence of First Form or Meaning
Learnt as a trend in the learner‟s production when the first form or meaning learnt does not
disappear as the learner advances - or the course advances - into more complex language
forms.
Some of the most frequent assumptions of those learners who adhere to first forms or
meanings learnt in a consistent fashion, are:
I) There is a one-to-one correspondence between all the words in the native language and
those in the foreign language. Each item in one language has one, and only one, equivalent
in the other. Thus, for the Spanish speaking adherent to first forms or meanings learnt,
llegar is arrive and cannot be get to, which should have a different translation into
Spanish. They learn one of these words and then refuse to incorporate the other. This is
one of the reasons why Adherence to First Form or Meaning Learnt is often compounded
with a large number of interlingual errors. It should be noted that in these cases the
Learning from Learners‟ Errors Ana María Rozzi de Bergel
172
learner's vocabulary is usually limited in the vernacular language, and he hardly ever uses
synonyms or metaphors. Association is often by proximity or pre-conceptual.
II) Language items are context-bound. This is a way of conceiving of language that adherents
to first form or meaning learnt share with cue-copiers, because both types of learners are
likely to use a pre-conceptual system for structuring language, based on attributes,
functions or uses of items (A chair is what I sit on; It is a useful piece of furniture; It is
what I put in front of my desk, etc).
This makes it difficult for them to use language flexibly, to generate language from
general rules and to get to abstract rules from an analysis of concrete cases. Induction and
deduction are equally difficult. These learners find it hard to accept that the "is" in "there
is" is the same as in "he is"; when they find the expression "table of contents", they will
never accept that "table" is formally the same word as in "The book is on the table", with a
different meaning - for them, there has to be a new word to fit the new context, so they
will tend to call the table of contents index, for example, unless they learn table of contents
as a unit. If they do this, they will have to be re-taught the word contents in expressions
such as the contents of my pockets and the word table in expressions such as time-table.
Learning a language, then, becomes a process by which each word has to be learnt in
relation to all the possible contexts where it may appear as if it were a new word each time,
for the learner will not tend to make transfers. It calls for a memorisation of all the possible
contexts for each item - an impossible task that the learner simplifies by sticking to the first
form or meaning he can grasp, or by simply imitating the stimuli.
A curious side of the particular relationship that adherents to first form or meaning learnt
and cue-copiers have with contexts is that when a word is found in isolation it is taken as
having only one meaning in L1 and presumed to fit any context where that L1 meaning is
Learning from Learners‟ Errors Ana María Rozzi de Bergel
173
required, without regard for connotations or specificity of use. Small, limited bilingual
dictionaries have a role to play in this phenomenon. For example, if the English-Spanish
dictionary entry says: Isolate = Aislar with no further explanation, this learner will say "I
have to isolate the cable" meaning insulate, and he will argue that the dictionary says aislar
and will not be persuaded to accept that he is wrong, since both insulate and isolate
translate as aislar in Spanish. He will sooner conclude that the dictionary is wrong.
III. Many flexible structures are transformed into rigid language chunks, mostly as a
consequence of the tendency towards binding language to contexts and associating by
proximity. In the statement "I want to get rid of him", the adherent to first form or meaning
learnt or the cue-copier will ask for the meaning of to get rid of him, and not for the
meaning of to get rid of, because he does not consider that him can be a variable element
in this set of items that he has transformed into a chunk.
A further problem with these self-originated chunks is that they are so arbitrary that they
may contain an unpredictable number of elements and variations may occur within the
same chunk, according to where the learner focalises each time the “chunk” is used. For
example, when he comes across "I want to get rid of them", after having asked about to get
rid of him, he may ask for the meaning of rid of them, because the pronoun is the
unfamiliar element this time and he focalises more on that than on the rest of the
expression.
Arbitrary chunk-formation points to difficulty for organising the syntagm and
unawareness of certain basic language concepts such as the distinction between verbs,
nouns, adjectives and adverbs and is impossible to correct unless these concepts develop in
the learner's mind. This would call for the learner‟s acceptance of the problem and a
Learning from Learners‟ Errors Ana María Rozzi de Bergel
174
considerable conscious effort, but leading the learner in this process of developing new
hypotheses about learning is often beyond the teacher of adults.
Conditioning is almost always training-induced, and occurs in classes where teachers
rely heavily on programmed, stimulus-response activities or over-simplified language
exchanges. It is part of the diagnostic analysis of errors more than of the descriptive
analysis, for it is not usual to find learners who will show conditioning as a result of their
learning hypotheses unless they are absolutely formulaic in their conception of language.
Example:
Whenever he is asked a question, the learner answers with the same words used in the
stimulus. He does not replace or manipulate them.
Question: Do you know Mary Brown, my teacher?
Answer: No, I don't know Mary Brown, your teacher.
Question: How many sandwiches must we send to that restaurant?
Answer: We must send five hundred sandwiches to that restaurant.
These examples are typical of learners whose teachers demand that they answer "in complete
sentences" - a very old-fashioned and unsuitable practice that dies hard. Although it can't be
said that the answers are wrong, they are certainly not fully appropriate and would sound odd
to native speakers.
Other examples of conditioning are:
- Do you speak French?
- Yes, I do. I speak French. (The teacher wants him to practice the "short" and the "long"
answers, so he systematically produces them together)
Learning from Learners‟ Errors Ana María Rozzi de Bergel
175
-Will you have tea or coffee?
-Coffee, please. I'd like some coffee, but I don't want any tea. (Learnt from an audio-lingual
drill, and produced within the wrong context, where the natural answer would have been just
"Coffee, please.")
2.c.5 Compensatory utterances/errors
We have so far discussed cases of errors produced on the basis of language the learner has
been exposed to, language he is "supposed to master", or according to his teacher, language
"he has been taught" and could have learnt. We have already explained the concept of
Compensatory Production (Code: change the 0 for a 5 in error Rating: 05, 15, 25, 35) in this
research and its relevance for the classroom teacher. The following example is a reminder of
the distinction between compensatory and non-compensatory production:
The learner says .............. when he means ....................
I wanted that he came with me. I wanted him to come with me.
If this learner is attending an elementary course and he does not know the structure
want+somebody+to do something, because he has not yet been exposed to it, his utterance is
Compensatory. If this learner is attending an advanced course, and he has not learnt the
structure although he has been repeatedly exposed to it, the utterance is Non-Compensatory.
The following diagram illustrates the process of deciding if an error is compensatory or non-
compensatory.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
176
Figure 9 – Compensation
3. Final comments
Teaching focuses more on diagnostic error analysis and the trend towards correction, rather
than exploitation of errors, dies hard. In view of recent developments in the neurosciences and
cognitive psychology, we might expect to see a change in the near future. Methods which
mistakenly place communication above all other concerns, including accuracy or appropriacy,
have also misled teachers into considering that anything learners produce should be celebrated
as an attempt at communicating, or even worse, as a learning attempt. I usually find evidence
of teachers‟ prime concern for obtaining linguistic production of sorts when I conduct
seminars and workshops on error analysis. I like to begin by asking teachers to look at two
errors and say whether it would make them happy to receive these answers from their beginner
students, after at least five lessons:
Teacher: What‟s his name?
Learner: What‟s his name is Pedro.
Teacher: What‟s his name?
Learner: Name Pedro.
ERROR
Could the learner have produced the correct form? Yes No
Non-Compensatory
Compensatory
Learning from Learners‟ Errors Ana María Rozzi de Bergel
177
The answer is always affirmative, and teachers explain to me they would feel satisfied to see
that these beginners are trying to communicate and producing a comprehensible answer. When
I insist that there is, perhaps, something a bit worrisome in these responses, they claim that
errors are learning attempts and should be tolerated, suggesting that I belong to the old school
that would seek to correct every deviant form.
It is clear that teachers are concerned with developing strategic and communicative
capabilities in their learners and that they are also aware of the role of errors in the learning
process, but it is also clear that they have not often considered errors as manifestations of
cognitive processes, let alone as items in the learners‟ curricula. It is also clear that most of the
theory teachers learn and preach is often never put into practice in the classroom. When I point
out, on these occasions, that “What‟s his name is Pedro” shows that the learner is not
effecting any transformation but just copying the stimulus, which is not a very profitable
learning operation, and that “Name Pedro” is a simplification of the structure of the answer,
teachers begin to understand the purpose of descriptive analysis and how it can complement
their diagnostic analysis. At this point, I even claim, quite boldly, that the mechanisms behind
these errors cannot be considered learning attempts. Of course the teacher should feel happy
that learners are trying to communicate as best they can, but that is only one aspect of the
value of learners‟ interlanguage, which can give us much more information for us to carry out
truly learner-centred teaching.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
178
Note:
Examples of coded errors:
Using the codes we have been supplying, the following errors would have been coded and
classified as follows:
“He goed to the park yesterday” (Meaning: He went …) = Appropriate, semi-correct; non-
compensatory; syntactic; intralingual, due to overgeneralisation: 30-020-10a; the same error,
but compensatory: 35-020-10a.
“I met her by casualty” (Meaning: by chance) = Semi-appropriate, correct; compensatory;
semantic; interlingual: 25-200-01. The same error, non-compensatory: 20-200-01.
It should be remembered that these diagnoses depend on the teacher‟s or the researcher‟s
use of a reliable sample for error analysis and on the availability of authoritative explanation.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
179
EXPLORATION AND REFLECTION
1. Choose a “good” and a “poor” student from one of your courses. For two or three
lessons, record the errors that they make in relatively open-ended activities. Analyse
them and find out whether the “poor” learner makes more semantic errors than the
“good” one. If this is not the case, review your opinion of these learners or look
elsewhere for the “poor” learner‟s lack of progress. His motivation, perhaps?
2. Record four errors which you would classify as appropriate, semi-correct, syntactic,
intralingual: wrong verb forms, wrong word order, etc. Make a quick, on-the-spot
diagnostic analysis of these errors and then ask for authoritative explanation. Had you
made the right diagnosis or did the students‟ explanations reveal aspects of the errors
you had not discovered?
3. Record four interlingual errors and ask for authoritative explanation. Were they really
interlingual or fully interlingual?
4. Analyse your students‟ production to find examples of Simplification. If there are too
many, reflect upon your teaching practice or the level of the course. Learners may be
finding it too difficult.
5. Record three instances of compensation and ask yourself: why did it happen? Were
learners too ambitious? Were they trying to tell you something important, but lacked
the language for doing so? Was the topic of conversation or the written assignment too
difficult for them? Did you ask them to undertake a task without enough preparation?
Learning from Learners‟ Errors Ana María Rozzi de Bergel
180
3. Completing the classification of errors
There are other criteria for classifying errors, besides those described in the categories and
sub-categories I have presented. They supplement the classification and aid the teacher to
tailor the curriculum and approach to the learners‟ personal curricula.
3.a Productive and unproductive errors
We have already previewed these sub-categories and defined them according to the following
criteria:
a) A productive error denotes an attempt at learning the code of the foreign language.
b) An unproductive error denotes an attempt at taking shortcuts at learning or at
simplifying the object or the process of learning.
Actually, what is productive or unproductive is not the error, but the learning hypothesis it
reflects through the strategy originating the error. However, we will use the term loosely to
denote both the error and the learning hypothesis.
This classification of errors into productive and unproductive contradicts the generally
accepted idea that all errors are learning attempts and means for learners to try out their
hypotheses about the language being learnt. There is no attempt at learning in simplification or
in shortcutting, as we have seen, but rather, an attempt at sidestepping it with some kind of
gimmick or stratagem. This invites us to take a less sympathetic view of some errors, which
should be addressed as problems to overcome. We are not proposing criticism or penalisation
of these errors, but changes in methods and teaching to favour conceptualisation and
metaphorical associations to counteract these processes, perhaps coupled with a different
organisation or better dosing of the input.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
181
As Lewis (Ibid.) pointed out, language is grammaticalised lexis and not lexicalised grammar.
The value of lexis and meaning, the higher value of appropriacy and communication over
correctness has been acknowledged by linguists and language teachers. Again according to
Lewis, we can talk with words but we cannot talk with structures (Lewis, Ibid.).
Appropriate, semi-correct utterances tend to be more productive than correct, semi-
appropriate ones, for they show the learner‟s achievements in the fields of meaning and of
social and strategic language use. The proportion may be: around 65% of the learner‟s errors
should be appropriate, semi-correct utterances and 35% correct, semi-appropriate utterances,
with a remaining 10% of more flawed utterances. Roughly the same values apply to the
balance between syntactic and semantic utterances. Syntactic errors should almost double
semantic errors to constitute a productive trend. Please remember that these values should be
calculated over at least 50 errors, collected in the prescribed manner, over at least three months
and that these figures are always tentative.
Interlingual errors are unproductive, for their presence shows that the learner is not fully
aware of the existence of the code of the foreign language, but they should be present in a
learner‟s production, accounting for approximately 30% of the errors made. This shows that
the learner is using the code of his native language as a resource to a reasonable degree.
Intralingual errors, usually 70% of the learner‟s erroneous production, tend to show a
productive trend when they are due to Accessing Attempts operations: overgeneralisation,
confusion, misunderstanding, etc., but they are unproductive if they are mostly due to
Adherence to First Form or Meaning Learnt, Cue-copying or Conditioning. In these cases,
they hamper learning more than Interlingual errors.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
182
If a learner's interlingual errors are dramatically above 30% and his intralingual errors are
mostly due to Accessing Attempts, we may suppose that his production will improve if he is
guided into detecting differences and similarities between the two languages concerned.
If a learner's interlingual errors are dramatically above or below 30% and his intralingual
errors are mostly due to Shortcutting, he may well be near his language ceiling and trying to
reduce interlingual errors to the supposedly desired 30% will not raise this ceiling unless the
causes of his intralingual errors become more productive. No change is impossible, but it
should be noted that in twenty-seven years, our research did not produce a single documented
case of significant changes in any one learner‟s causes of intralingual errors.
Compensation is unproductive if it is very high, accounting for more than 10% of a learner‟s
erroneous production. Otherwise, it is a normal process and can often be exploited to help the
learner develop his personal curriculum.
All these percentages have to be taken as indicative and in a very broad sense. These figures
were produced by Selinker in his research, and they appeared to apply to the learners we
studied.
After analysing and classifying 5,312 errors made by learners at different levels of
proficiency, who had all passed the same kind of tests with marks between 65% and 80%, we
found the average distribution of errors to be as shown below and to be the same for all levels.
The degree of mastery of the foreign language produced differences in the diagnostic analysis
of errors and in their number, but the descriptive analysis revealed practically the same picture
for learners at different levels.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
183
CATEGORIES SUB-CATEGORIES VALUES
ERROR RATING Appropriate, semi-grammatical 63 %
Semi-appropriate, grammatical 24 %
Semi-appropriate, semi-
grammatical
8.50 %
Inappropriate, ungrammatical 4.50 %
ERROR
DESCRIPTION
Semantic 24 %
Syntactic 64 %
Semantic-syntactic 12 %
ERROR
CLASSIFICATION
Interlingual 30 %
Intralingual 70 %
CAUSES
OF
INTRALINGUAL
ERRORS
Accessing attempts 70 %
Simplification 22 %
Shortcutting 8 %
Compensatory errors 6 %
Figure 10 – Errors of learners with scores between 65% and 80% in language tests
The fact that proficiency levels did not make a difference in this picture of error seems to
question the idea that cross-linguistic transfer is more usual when people begin learning a
foreign language, or even that it is a default strategy, and that it decreases when more
proficiency is attained. In our analysis, about 30% of learners‟ errors were due to interlingual
problems, at any level.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
184
When 1,236 errors belonging to learners who had not passed the tests or who were supposed
to have reached their ceiling were classified and the results averaged, the values showed
considerable increases in the number of the errors we have classed as unproductive.
The size of the sample was, in this case, considerably smaller, because fewer learners usually
scored below standard or just above the passing mark.
The table below shows the results.
Figure 11 – Errors of learners who scored below the passing mark.
CATEGORIES SUB-CATEGORIES VALUES
ERROR RATING Appropriate, semi-grammatical 53 %
Semi-appropriate, grammatical 25 %
Semi-appropriate, semi-
grammatical
10 %
Inappropriate, ungrammatical 12 %
ERROR
DESCRIPTION
Semantic 46 %
Syntactic 41 %
Semantic-syntactic 13 %
ERROR
CLASSIFICATION
Interlingual 47 %
Intralingual 53 %
CAUSES
OF
INTRALINGUAL
ERRORS
Accessing attempts 52 %
Simplification 31 %
Shortcutting 17 %
Compensatory errors 12 %
Learning from Learners‟ Errors Ana María Rozzi de Bergel
185
The tests were as communicative as possible and their productive, relatively open-ended
sections, Writing and Oral Test, were scored with criteria which resembled those used for
error analysis:
a) Appropriate, correct answer: 5 marks.
b) Appropriate, semi-correct answer: 3 marks.
c) Semi-appropriate, correct answer: 2 marks
d) Semi-appropriate, semi-correct answer: 1 mark
e) Inappropriate, incorrect answer: 0 marks
These more open-ended and communicative sections of the tests were given more weight
than the close-ended sections, such as cloze passages or multiple-choice questions and the
errors made in the close-ended sections were not analysed.
We fully realised that the criterion of using test scores to group learners into average
achievers and under-achievers was artificial and ambiguous, but a clear-cut criterion was
needed and this was the clearest we could find. The test results were supplemented with
teachers‟ reports on the development of the courses and the learners‟ progress. Attendance
records were also examined, to determine the possible influence of learners‟ absences on their
progress. Supervision reports provided information on possible problems regarding the
teacher-student relationship or the group dynamics, which could have led learners to under-
achieve. Learners whose performance at the tests was not consistent with their usual
performance in class or whose test results were probably influenced by known external events,
such as the death of a close relative, were not included in our general analysis of errors, but
were treated separately.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
186
Further research was carried out with learners who were thought to be at their language
ceiling, and I will discuss the findings in Part 3.
3.b Pre-systematic, systematic, post-systematic stages.
These are the categories defined by Pit Corder (Ibid.):
It has already been noted that learners often appear inconsistent in their
production of errors. They often seem to alternate between getting something wrong
and getting it right. We may be able to distinguish three steps in learning as
evidenced by the nature and degree of systematicity: (a) the presystematic stage: the
learner is unaware of the existence of a particular system or rule in the target
language. His errors are quite random. He may even occasionally produce a correct
form. When asked to correct his sentences he cannot do so nor give any account of
why he chose the particular form he did. (b) The systematic stage: his errors are
regular. He has discovered and is operating a rule of some sort, but the wrong one.
When asked to correct his error he cannot do so, but he can give some coherent
account of the rule he is following. (c) The post-systematic stage: the learner
produces correct forms but inconsistently. He has learned the rule but fails through
lack of attention or lapse of memory to apply it consistently. This is the practice
stage of learning a particular bit of the language. When asked to correct his error he
can do so and give a more or less coherent account of the rule. Learners will, of
course, be at different stages in respect of any particular system of the language, e.g.
postsystematic in the number system, systematically erroneous in the use of articles
and presystematic in the use of the perfective aspect. (p. 131)
Learning from Learners‟ Errors Ana María Rozzi de Bergel
187
Edge (1989) proposes a similar classification, with different criteria. He calls “slips” the
errors which the teacher suspects the learner could self-correct; “errors” are mistakes the
learner can‟t correct even if they are pointed out to him, but the teacher thinks the class should
be familiar with the correct form and is therefore called upon to provide it, and “attempts” are
mistakes made in forms that the class have not yet learnt. The third would be a case of
compensatory production, in our classification, resulting in an erroneous utterance.
Both categorisations of error are consistent with the theory that describes learning as taking
place in stages: from unconscious ignorance to conscious ignorance and then from conscious
knowledge to unconscious knowledge (Taylor, 1988). This is also related to the idea of the
realisation of the learner's personal curriculum: items move from one stage into another as the
course curriculum comes closer to the learner's personal curriculum and it is impossible to
predict how long it is going to take for a particular item to go through the three stages, or how
long it will remain at a certain stage.
Neither the diagnostic nor the descriptive analysis of a learner's production, nor test scores,
will show the placement of the errors at any one of Pit Corder's stages, and yet, changing his
errors from one category into the next may be a remarkable achievement for a learner and the
teacher should pay attention to these processes to measure his progress.
3.c Developmental, fossilisable errors.
Developmental errors appear when a new item is incorporated to a system the learner
thought closed and are signs that learning is taking place. We should remember, in connection
with these phenomena, Piaget‟s (Ibid.) and Ausubel‟s (Ibid) learning theories and the idea of
learning in plateaux or U-shaped learning. The items in the closed system will have to "move"
to make room for the newcomer, and for a while this will cause a disturbance leading to error
Learning from Learners‟ Errors Ana María Rozzi de Bergel
188
production. When the new item has finally found its place in the system, the "stir" will quiet
down and errors in the use of that system will disappear.
Example:
a) Learners incorporate the use of "to be" in the present simple and close the verb system.
b) Learners learn the present continuous and have to open the verb system to include the
present participles. For a while, they say "I going" and "You is coming", until they
acquire the sub-system of the present continuous and close the verb system again.
c) Learners learn the present simple of verbs expressing likes and dislikes and they have
to open up the verb system again. For a while, they say "I is like" or "You disliking",
etc., until they acquire the new item, which does not happen before they have properly
associated and/or differentiated it from the rest of the sub-system in the system, and
established its relative value.
Adherence to first form or meaning learnt is often present in developmental errors, as in “I
am like coffee”, for “I like coffee”, produced when the learner clings to previous forms of the
present while incorporating the syntax of the present simple, a very natural phenomenon
which then corrects itself after a few lessons.
Developmental errors sometimes affect correct forms which were already stable (Harmer,
Ibid.), for example, when the past simple is introduced, learners who have been using different
present tenses accurately and appropriately suddenly begin to confuse them or use them
wrongly. The newcomer has shaken the foundations of the complete sub-system. If teachers
do not know the role and manifestations of developmental errors, they believe that learners are
no longer making progress, or what is worse, they feel that the new item has been introduced
too soon. The usual reaction is to go back and “consolidate” present tenses, seeking to “get
them to learn each point properly before moving on to the next”. This is not only useless, as
Learning from Learners‟ Errors Ana María Rozzi de Bergel
189
the developmental errors will re-occur when a new element is introduced, but pedagogically
wrong. Knowledge, when we are talking about learning a system, means knowing how the
whole of the system works and the relative value of each element with respect to all the others.
Inter-relations are so intricate that there is no way anybody can learn “one thing at a time”.
When errors remain in the learner's production long after the expected developmental period,
they become fossilisable or fossilised and the learner develops a personal grammar. The
teacher is faced with the problem of distinguishing between developmental and fossilising or
fossilised errors, which is no easy task, because the learning pace is different for all items and
for all individuals. Things are complicated even further by the fact that the order of
presentation of items in the course curriculum is hardly ever the order in the learner's personal
curriculum, nor is there valid evidence to determine if this personal curriculum is more or less
standard for all the individuals in a particular category or age-group. Only experience,
observation and common sense can help the teacher determine if an error is developmental or
fossilised.
3.d Teaching-induced errors
The role of teaching and materials in error production can never be overstated. Selinker
(Ibid.) and Richards (Ibid.) called it transfer of training and listed it as one of the main causes
of errors. The word “training” is appropriate to describe instances in which learners simply
reproduce language which has been drilled into the code of their interlanguage by teachers
who are often equally conditioned to use certain practices and procedures, with little or no
awareness of their suitability or significance. It may not be appropriate to talk about “training”
when we refer to inferences, deductions, associations or conclusions induced by the materials
or the teachers, which do not entail literal repetitions of language elements. Because of this
Learning from Learners‟ Errors Ana María Rozzi de Bergel
190
distinction, we prefer to talk about teaching-induced errors, including materials, learning
contexts, course curricula and teaching proper within this concept.
We will briefly describe some error-provoking features of methods, materials and teaching
and discuss some concrete examples, hoping to induce reflection upon some classroom
practices and a revision of some of our “cherished notions” (Pit Corder, Ibid.) regarding
curriculum and materials design.
There is a difference between scientific knowledge and pedagogic knowledge. Scientific
knowledge is the formulation of general rules or principles, theories and descriptions based on
empirical research or on an analysis of academic sources. It is usually transmitted through
publications, papers, academic books and lectures.
Pedagogic knowledge is that part of scientific knowledge that reaches the learner through
didactic mediation – that is, knowledge which has been made easier, simpler for the learner, or
which is expressed by means the learner can reasonably grasp, maybe using analogies,
examples, movement or visual aids. Didactic mediation creates the bridge between science and
the classroom (Gibaja, 1993), for example, between grammar and linguistics and the learner.
The mediator per se is the teacher, aided by the materials, the method and other teaching
resources and it is a pre-requisite that teachers should possess sound scientific knowledge if
they are to engage in effective pedagogic mediation, as the simplification of incomplete or
already simplified concepts can only lead to erroneous conclusions. In the case of pedagogic
mediators, a “sound knowledge” does not mean just good grounding in grammar, phonology
and linguistics, but also in all aspects of the teaching profession.
Mediation starts through a process of selection, when the mediator decides which part of
scientific knowledge will be useful for the learner at a certain stage, or which part he can
reasonably be expected to understand, considering his previous knowledge and general
Learning from Learners‟ Errors Ana María Rozzi de Bergel
191
background. The next step is determining how to transmit this portion of scientific knowledge:
an explanation, an experiment, a visual organiser or various other means may be selected, on
the basis of the teacher‟s appraisal of which will be more suitable for different learning styles
or hypotheses.
As we can see, the decisions are quite complex and they sometimes have to be made on the
spot, in the presence of a particular query or when a problem is detected and they ought to be
based on deep knowledge of the learner‟s cognitive structure, which is usually beyond the
teacher. It is not surprising, then, that teachers should make mistakes when carrying out their
pedagogic mediation and that they should tend to cling to recommended procedures in their
wish to step on more solid ground.
The nature of the decisions regarding how much to teach and how, is closely linked to the
teachers‟ styles for leading a group of learners. The concept of the teacher as a leader is related
to the classification of groups into primary and secondary, according to Cooley‟s (1909)
traditional classification. A primary group is small and its members are engaged in close,
personal, affective relations. They share activities and culture and care for one another. A
typical example is the family. A secondary group is often larger and its members interact on a
less personal and affective level. Their relationship is often temporary and they usually have a
group objective related to a function to perform, work to do or concrete goals to achieve. A
typical example is a group of learners, a class.
To achieve their goals, the members of a secondary group need a leader who may possess
the professional knowledge to guide them, the personal qualities to keep the group together
and empathise with its members and the character to represent a role model or at least, an
example of fairness and integrity. Learners will, initially, deposit the leader‟s role in their
teacher, assuming that this professional possesses the desired qualities to guide them
Learning from Learners‟ Errors Ana María Rozzi de Bergel
192
successfully in their learning process. Should the leader fail them in any way, they may
conclude that the person performing the role is not apt and ask for another teacher or go to
another school, but the role will never be challenged or eliminated: a class needs a teacher.
How will a particular teacher play the leader‟s role? There are social aspects of the role, such
as behaviours, dress codes, activities, which have been designed by the social group and the
teacher will respect to be identified as such. The word role comes form roll and it was used to
refer to the roll of paper from which mediaeval actors read their parts on stage. The roll
contained words and patterns of behaviour written by an author or the company‟s director –
so, actors had to behave as expected and speak words which were not their own. Together with
the roll containing the text of the play, actors were often given a mask and of course, they
always wore special costumes. Of course, actors also expressed their personalities and styles
while reading from the rolls. Here we have the basic elements of the concept in modern
psychology: a role is an organised model of behaviour pertaining to a certain position of the
individual in an interaction web, defined by his/her own expectations and those of others.
There is interplay between the inner world and the outside world, that is, the social and
cultural dimension of the role and the role player‟s individual conception of the role, partly
conditioned by his personal history. So we can talk about social roles and psychological roles.
Roles have other characteristics, as well:
a) They are complementary, each role having its counter-role.
b) They generate expectations in the social group, as each player is expected to behave
according to an accepted model.
c) They are hierarchical, as different roles have different status and enjoy various degrees
of respect or consideration.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
193
d) They depend for their existence on general agreement as to how the role should be
played.(Wright, 1987)
The tutor may display different leadership styles, which will have particular effects on
students‟ progress. Although many and more specific types of leadership have been identified
since Lewin (1939) produced his seminal classification, they may all be considered variations
on his main three categories or subdivisions of the original groups. The most interesting aspect
of Lewin‟s classification is, to us, that the research leading to them was undertaken at schools
and the leaders being observed were all teachers.
According to Lewin and his team, leadership can be authoritarian or autocratic, participative
or democratic and delegative or laissez-faire. The authoritarian leader does not take the
group‟s needs and interests into consideration when making decisions, and is relatively
insensitive to feedback or input. However, this type of leadership is sometimes necessary in
certain situations when quick, effective decisions are required or the leader is the only member
of the group who is knowledgeable in a certain topic.
Authoritarian leadership does not seem so widespread at schools as in the past. The
traditional image of the teacher as a kind or army sergeant, armed with a red pen and the
power to punish or reward is rapidly disappearing. The problem is that an authoritarian leader
may be despotic, or overprotective and the latter is a very dangerous type of leadership, as it
creates dependent learners who do not realise very easily that they are being harmed.
Participative, democratic leaders offer guidance, advice and help to the group. They listen
to input and feedback from group members and encourage contributions and participation.
They set the pace, but taking everybody‟s needs and preferences into consideration. However,
these leaders retain the final say in most decision-making processes, particularly those
requiring knowledge and expertise. Group members gladly rely on their leader‟s decisions,
Learning from Learners‟ Errors Ana María Rozzi de Bergel
194
because they feel taken into consideration and valued. Democratic leadership means leading a
group to the achievement of their goals by empowering all the members to use their potential
and helping them overcome their shortcomings.
Delegative, laissez-faire leaders offer little guidance or help and leave decision-making in
the group‟s hands. This often leads to lack of motivation, particularly when no member of the
group is properly qualified to make decisions regarding their work. If a member of the group
is capable of actually emerging as a leader, then further discomfort will ensue, as the group
will have a natural leader who is not performing as such, and a parallel one. These situations
often lead to discouragement, confusion, anger and whenever possible, a request for a change
of teacher. Sometimes, these secondary groups with a laissez-faire leader tend to become
primary groups, for example, when all the members end up becoming good friends but drop
out of the English course.
Figure 12 – The teacher’s leadership
It is important to note that, although most teachers have a predominant leadership style, they
never use only one. In fact, situational leaders, as defined by modern management theories,
tend to use a predominantly democratic style, but are ready to switch into other styles
according to need.
When faced with the task of pedagogic mediation, the basic questions tutors ask themselves
are, “How much can these learners grasp at this stage? Are they ready for this? How should I
THE TEACHER’S LEADERSHIP
COGNITIVE AFFECTIVE ETHICAL
Authoritarian
Democratic
Laissez-faire
Learning from Learners‟ Errors Ana María Rozzi de Bergel
195
lead them to access this knowledge” Authoritarian teachers are likely to underestimate
learners and spoon-feed them some rules of thumb, or just give them the grammar book
explanation to assert their superiority on them and make the class beg for more clarification.
Democratic teachers will probably attempt a more realistic mediation, giving a slightly
simplified version of the rules but using several means, to give learners choices. They will also
act on information about what seems to work for the group, or about their learning styles and
hypotheses.
Laissez-faire teachers may decide not to do anything, present the lesson and let the class
draw their own conclusions or send them to look up the grammar in a book.
Avoiding the problem may not be so easy, as the following example shows:
Imagine a lesson where there were several sentences like these:
It is important to book the hotel well ahead of time.
It is easy to understand why he didn‟t come.
The teacher knows that the anticipatory it is an empty carrier of a subject, used to comply
with the rule that a verb must be preceded by a subject, when the real subject of the sentence,
in this case, an infinitival phrase, is placed at the end of the utterance. This is scientific
knowledge, taken from grammar books and should be transformed into pedagogic knowledge.
Many teachers will believe that learners of English as a foreign language may not be able to
understand this explanation, as some technical knowledge of grammar is necessary and trying
to make sense of the explanation might be harder than mastering the structure by simply using
it in communicative situations. This reasoning may be prompted by the teacher‟s inability to
carry out effective pedagogic mediation, by a laissez-faire leadership style, or by blind
adherence to certain prescribed teaching practices.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
196
The teacher, then, engages the class into communicative practice without giving them a
technical explanation of the structure of the sentence or attempting to convey it in any way,
but suddenly, a learner asks: “What does “it” mean in that sentence?” or many learners in the
group start saying “Is important to remember the restaurant‟s name” or “Is clear that you
don‟t love me”, omitting the anticipatory “it”, and the teacher‟s problems begin. How does she
mediate between the grammar book and the learners? Does she give them the complete
explanation anyway? Does she give them a simplified explanation of the rules? Does she give
them a pattern for them to copy – it+to be+adjective+to .....? Does she contrast the native
language and the foreign language? What other ways and means can she devise for this
mediation?
The teacher‟s choice will then be mainly influenced by her leadership style, her actual
knowledge of grammar, her assessment of the learners‟ limitations and possibilities, and her
methodological orientation. As for the learners, their comprehension will be influenced by
personality factors, their cognitive structure, the way they have structured their native
language, their previous knowledge, the stage of development of their interlanguage and their
communicative goals but it will also depend on the type of pedagogic mediation they are
exposed to. Of course, social and cultural factors will permeate the whole process.
Lesson observation and the process of obtaining authoritative explanation of errors usually
show that most learners not only believe anything the teacher or the book says, but also take it
literally. Another observable phenomenon is that teachers are relatively deaf-and-blind to
errors caused by classroom practices which have been sanctioned as correct by other teachers
or methodologists. Like learners, teachers believe anything that books say, and take it literally.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
197
Precious few teachers question accepted practices or the profession‟s “cherished notions”
and this is one of the reasons why approaches and methods soon become ritualised in the
classroom: the theory and general aim of the approach or teaching philosophy are soon
forgotten, and just the procedures remain. Unfortunately, without theoretical substantiation
procedures become rituals applied in a uniform and systematic fashion to all teaching-learning
situations, when it should be obvious that no procedure can serve all conceivable contexts.
The quest for understanding the learner, for stepping into his shoes, moves good-quality
teaching. To understand how the learner thinks and acts, besides having a solid professional
background, it seems that the teacher needs to be a fairly good mind-reader. But is it possible
to read somebody else‟s mind, to think along the same lines, to empathise with a learner so
closely that we may claim that we understand how he learns? According to Humphrey (1986),
human beings are born psychologists and they possess an innate capability as mind-readers, an
inner eye, which they use to detect intentions, attitudes, motivations in others and conduct
their strategic interactions according to this information.
Baron Cohen (1995) explains that the inner eye operates through the Intentionality Detector,
the Eye Direction Detector, which tries to identify what a person is looking at and why, and
the Shared Attention Mechanism, which detects the connections between the self, an agent and
an object and the Theory of the Mind Mechanism, used to infer mental states through the
observation of behaviour. In this process of mind-reading, the inner eye (Humphrey, Ibid.)
enables people to create common narrations of experience and humans seek to relate to those
with matching narrations, or with narrations which they can build and develop together.
Because of this, the problem with this inner eye is that it works according to parameters
belonging to the mind-reader‟s own mind, so that he shapes his perception of other minds on
the basis of the attitudes and beliefs present in his own. This would account for the fact that
Learning from Learners‟ Errors Ana María Rozzi de Bergel
198
thieves live, sometimes, in fear of being robbed, because they attribute a thief‟s mind to
everybody else, as they are unconsciously trying to find akin minds.
Humphrey carried out most of his research with chimpanzees, and when he was once asked,
tongue-in-cheek, if his chimpanzees were so clever that they could be teachers, he answered,
“Yes, indeed, but they would expect their students to have chimpanzee minds”.
Our research may have sought to explore the learners‟ cognitive structure, to perfect learner-
centredness as a teaching philosophy and learn more about personal curricula, but any insight
that is gained in these domains cannot translate into enhanced teaching unless teachers expect
to find responsive minds in the classroom.
3.d.1 Specific causes of teaching-induced errors
I will discuss the most outstanding and clearest cases of error-provoking elements in
methods, books or teaching practices found during our research, but it should be remembered
that this is an area where much depends on the personal, teacher-learner relationship and it is
always dangerous to present human relations as stereotyped cause-effect processes, so no
generalisable conclusions should be drawn from the experiences we will describe.
The data for the following examples of teaching-induced errors came from:
- Authoritative explanations of errors, in cases where learners referred to something they
had seen in their books or the teacher had said.
- Common errors of a group of learners using the same book or being taught by the same
tutor, which were traced back to features of the book or distinctive characteristics of the
teacher's methods and procedures. The identification of the error-provoking element was
done through lesson observation, sometimes supplemented by an analysis of the teacher's
leadership style, carried out by our psychologist.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
199
Examples of error-provoking practises, teaching and materials:
a) Insistence on literal answers or "complete" answers that reproduce most of the
questions:
Learners answer questions on a passage by copying sentences from the text, because
the answers do not require the least elaboration.
Example:
The passage says: “.... and at night, we slept under the stars, in the vast green field.”
Teacher‟s question: “Where did you sleep at night?”
Expected answer: At night, we slept under the stars, in the vast green field.
This encourages cue-copying and results in intralingual errors due to cue-copying,
such as:
Example:
Teacher: Where may those who graduate work?
Learner: Those who graduate work, may work abroad.
Teacher: What do you do?
Learner: I do you do I am a doctor.
b) Learners are not allowed to give one-word or short answers. The teacher constantly
tells them to produce "complete" sentences even when the brief answer is also
"complete".
Learning from Learners‟ Errors Ana María Rozzi de Bergel
200
Example:
Teacher: Where were you born?
Learner: In Buenos Aires.
Teacher: A complete sentence, please.
Learner: I was born in Buenos Aires.
Besides cue-copying, these practices encourage conditioning.
c) The provision of lists of lexical items and their translation into the native language, out
of context. This practice reinforces the idea of "one-form-one-meaning", which
produces Adherence to First Form or Meaning Learnt and runs contrary to the
formation of conceptual categories.
Example:
A teacher writes on the board, for the class to copy:
table = mesa
board = tabla
room = habitación
It is unlikely that learners will be willing to incorporate further meanings of these words,
such as “the board of directors”, for example.
d) Fictitious contrasts. Some of them have been created by course writers in their wish to
demonstrate differences and similarities between items which could very well have
been taught separately. Establishing a contrast between them does not seem to add
clarity but to confuse the learner.
- Present perfect and past simple contrasted.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
201
This often leads to utterances such as “I have wrote two letters today”. The
authoritative explanation of this error by a pre-intermediate learner was that have was
necessary because the day had not finished yet and wrote, because he had already
finished writing the letters.
The items had been taught contrastively and the learner was trying to combine, rather
than distinguish, the rules for both.
- The "contrast" between for and since, associated to the present perfect.
This often produces:
I have worked here since two months. (Meaning: I have worked here for two months)
Authoritative explanation: I am mentioning when I started working here, so I use “since”,
and as I am still working here, I use the present perfect.
We have come to Paris for two years. (We have been in Paris for two years)
The authoritative explanation was: I use present perfect, because I am still in Paris. I
came two years ago.
They have had this car for last month. (Meaning: They have had this car since last month)
Authoritative explanation: “last month” is a period and I should use “for”.
- The "contrast" between "very" and "too".
This may produce:
This exercise is very hard for me to do.
Authoritative explanation: I can‟t do it. It‟s impossible.
This exercise is too hard, but I know how to do it.
Authoritative explanation: It is very, very hard, but I can do it.
- Present simple and present continuous contrasted.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
202
This often leads to Adherence to First Form or Meaning Learnt, because the contrast
places great emphasis on the difference between a repeated action and an action in
progress. Learners strive to learn this and then refuse to incorporate:
She leaves the house, gets into her car, starts the engine and leaves.
We are leaving tomorrow.
Besides, the contrast often produces:
I am wanting a cup of tea. (Meaning: I want a cup of tea)
Authoritative explanation: I am wanting a cup of tea now.
Items presented contrastively confront the learner with three problems: learning the two
items and the contrast or comparison. Those who tend to associate by proximity and whose
categorisation style is predominantly pre-conceptual, will be hard put to grasp the contrast.
e) Oversimplified rules, taught as absolutes, or wrong statement of rules.
- Frequently heard statement of a rule: "The present simple is used to
express habitual actions or eternal truths", similar to “To express routine”.
The rule seems to leave out other uses of the present simple, as in
I need a pen.
When Columbus sees America, he thinks he has arrived in Asia.
She likes your new dress.
Although it is not possible, perhaps, to teach all the uses of a tense at the same time, learners
should be made aware of the existence of several uses, so that they do not refuse to flexibilise
the structure later on.
Teaching partial concepts as general rules results in lack of flexibility in the use of the
foreign language. In the area of conditionals, for example, a lot of problems arise due to their
Learning from Learners‟ Errors Ana María Rozzi de Bergel
203
classification into four types, taught as formulae, at the moment of introducing the fifth, fuzzy
category of “mixed conditionals”. All too often, the observer sees that learners either think the
teacher introducing mixed conditionals is misleading them or that previous teachers have lied
to them or exposed them to books which contained mistakes. There is a feeling of betrayal, of
“And now they tell me” in the learners.
- Rules replaced by formulae. Often heard in the classroom: "Whenever you talk about
actions happening "now", you must use the present continuous." Even worse: “Whenever
you say “now” you have to use the present continuous”. And even worse: the teacher
writes on the board: “now = - ing”.
These over-simplified formulae, besides encouraging shortcutting or simply confusing
the learner, omit other cases or the use of “now”:
I‟ve been working here for three days now.
What do you want now?
Where is she now?
Oversimplified semantic values. For example: "The modal "must" expresses obligation."
Some of the results of that statement may be:
When learners come across “They must have left”, they understand some
people left because they were forced to, or because they had to.
A statement such as “You must be joking” is interpreted as “It is your
obligation/duty to tell jokes now.”
“He mustn‟t have done it” is erroneously produced to express that somebody
did something wrong, which he shouldn‟t have done.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
204
While observing a lesson, I witnessed a heated argument between a teacher and a group
of young learners, about twelve years old, when she told them that the modal must
expressed “moral, self-imposed obligation”. The children were completely unable to
understand the concept of self-imposed obligation, arguing that all their obligations were
imposed on them by adults. They were also puzzled by the concept of moral obligation,
which they associated with religious principles or accepted social practices. They argued
that moral obligations, taken from that point of view, were also set by adults.
The situation was very interesting for the observer and anguishing for the teacher, who
really had no arguments, as she was simply repeating something which had probably been
handed down to her by her own teachers, in a ritualised fashion. She finally settled the
matter by taking back the “moral” or “self-imposed” qualifiers and leaving just, “It
expresses obligation” in her definition. Her attitude was authoritarian all through; the final
resolution of the conflict was reached in the spirit of “We had better not discuss it any
further, because you will never understand it anyway” and the class were not satisfied with
the outcome.
Teacher-made, erroneous rules often heard in the classrooms, such as: “The Present
perfect is used to denote a past action, the results of which continue into the present.”
Learners do not see why they cannot say “I have broken my leg yesterday”. If someone
broke his leg yesterday, the results of that continue into the present: the person still has his
leg in a plaster cast. This “rule” also leads to “Shakespeare has written many plays”, as the
results of his past actions are available in the present.
f) Analogy and proximity.
These practices have been listed by Richards (Ibid.) as error-provoking. They consist in
putting analogous items next to each other in the same text, with no explanation of the
Learning from Learners‟ Errors Ana María Rozzi de Bergel
205
differences and when learners are trying to acquire both. The result is similar to that of the
"contrasts", when it does not result in simplification of the syllabus.
For example, teaching "She said she was coming" together with "She told me she was
coming" will probably produce "She said me she was coming".
g) Teaching that lays particular stress on accuracy or on communication is equally error-
provoking. A balanced diet is always the best option. An analysis of long-term results in
courses where there has been excessive emphasis on either aspect shows very paradoxical
results: usually, the area that has been stressed is the one the learners have failed to grasp.
Completely grammar-based courses where a high degree of accuracy is demanded do not
succeed in teaching even grammar; on the other hand, courses where fluency is placed
well above accuracy do not produce learners who communicate effectively.
In the first case, learners do not necessarily transform the grammar in the textbook
exercises into meaningful language; in the second, they make so many errors they do not
communicate. It would not be very bold to state that a language course will fail in the areas
that have been artificially and disproportionately highlighted by the tutor or the book. In the
learners' personal curricula, language is a balanced whole.
Examples of this problem, recorded during lesson observation:
a.
Teacher: (Highly controlled practice, out of context aimed at teaching question
forms) Ask me where I live.
Learner: Where you live? (He cue-copies the teacher, because he does not associate
question forms with meaningful communication, as he is practising out of context)
Learning from Learners‟ Errors Ana María Rozzi de Bergel
206
b.
Learner: (Highly "communicative" lesson) Where you live?
Teacher: I live at 45 Suffolk Road. (Does not correct the question because communication
has been achieved anyway. The learner concludes the question is correct and will probably
produce "Where you go?" or “Where you work?”.
c.
Teacher: ("Balanced diet" approach) You are the Principal. You are hiring me as a teacher.
Please fill in this form with my personal data. Get all the necessary information from me. Ask
me my name, address, etc.
Learner 1: (To teacher) Where you live?
Teacher: Pardon? (Shows through gesture there is something missing).
Learner 1: Where .... do you live?
The effects of heavily communicative or non-communicative teaching have been discussed
by Lightbown and Spada (1999). They cite studies carried out in Canada (Savignon, 1972;
Montgomery & Eisenstein, 1986) which showed that learners who attended courses with a
communicative component combined with formal grammar teaching learnt better than those
whose instruction had been heavily grammar-based, thus making the case for a balanced
approach.
On the other hand, we would argue that a fully communicative approach would probably
reduce correction to a minimum and favour early fossilization of many errors.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
207
h) Arbitrary order of presentation of grammatical items. The sequencing of items in the
syllabus is determined by the curriculum designer‟s perception and the theoretical
framework. Many criteria have been outlined for ordering items in a syllabus, but they
may be grouped into two broad categories: learner-based or teaching-based. In a
learner-based syllabus, the order of presentation will be determined by the utility
principle (Kaur, 1990) and priority will be given to
which is needed most immediately by the learner,
which has high surrender value, that is, of most use to the learner,
which is necessary to avoid a communication breakdown,
which is flexible, that is, can be used most widely, and
which is most frequently used by the learner.
However, this order still needs a sequence of functions, lexis and grammar items, besides
relying on fairly ambiguous criteria: it is impossible to know what would avoid
communication breakdown for a particular learner, or what a learner will use most
frequently, beyond a few functions and lexis.
Kaur (Ibid.) lists the sequencing principles of a pedagogically-based syllabus, defining it
as one which
can be taught most effectively and efficiently given in the classroom situation,
can be used in teaching other languages,
is needed for classroom purposes,
is simpler in form or meaning.
As we can see, the criteria become more ambiguous as the focus of syllabus design
moves from the learner to the teacher. The recommended design should be a combination
of learner-based and teaching-based elements, as proposed by Brumfit (1981). He
Learning from Learners‟ Errors Ana María Rozzi de Bergel
208
distinguishes between intrinsic criteria, which derives from language itself, and extrinsic
criteria, usually referred to motivational factors, coherence, usefulness and other traits
associated with learners‟ needs and interests.
Despite advances in this field, we have to acknowledge the fact that learners‟ personal
curricula will defy any order of presentation of syllabus items.
Some of the accepted sequences seem to be particularly error-provoking:
- Teaching the Past Simple of the verb "to be" before the Past Simple forms of all
the other verbs produces "I was went to the cinema yesterday" or "I was go home
early last Monday". We have already discussed that this is a normal process when
it is a developmental error, but not when the two forms tend to remain together for
a long time…. or permanently.
- Teaching the Present Continuous before the Present Simple produces "I am
work here."
Research into the learners' personal curricula would probably disclose that any order of
presentation is arbitrary and that such an order is necessary only for the sake of teachability.
The teacher should be prepared to incorporate whichever items her learners are interested in
or seem to need. She should not worry if the class do not pay attention to the course
curriculum but work very earnestly on the items they consider important. Their personal
curricula will sooner or later lead them back to the course curriculum or co-exist with it.
i) Gradation of difficulty is always relatively arbitrary, as there is no way of knowing
what will be easier or more difficult to learn. Contents are usually graded in terms of
what needs to be taught first in order to move on with the syllabus, but there are
arbitrary criteria for grading difficulty which place artificial restrictions on the
Learning from Learners‟ Errors Ana María Rozzi de Bergel
209
language available to the learners, particularly when they induce learners to avoid large
language areas for long periods:
- Forcing learners to use only present tenses throughout beginner or elementary
courses creates a reluctance to incorporate past forms, which appear as second-rate
items, as learners have been making do without them for a long time. They have
also made great efforts to avoid them and the teacher has probably been helping
them to do so. Consequently, they tend to cling to their achievements. This
produces adherence to first form learnt.
- The use of gerunds is formally taught when people have already been saying "I
like swim" for quite a while. The wrong concept here, on the teacher‟s part, is that
nothing can be taught before it appears in the curriculum or the materials as formal
instruction. Books and materials often include language for passive recognition
which may be transformed into active if it seems that learners could profit from
them or if learners ask questions regarding those language elements.
- Teaching modals at a late stage, when learners have been using substitute
expressions for years, creates the feeling that modals are optional language
elements, easily replaced by utterances such as "Maybe I'll come", "It is possible to
come tomorrow", "It is your obligation to do it", etc.
j) Tutors' contradictions and erroneous statements about language, sometimes produced
by insecurity or the wish to make things easier for the learner.
Examples:
Teacher: "When we mean "now", we use "is". (Meaning: When we talk about an
action in progress at the moment of speaking, we often use the present continuous)
Possible result: Learners say "I is need a pencil now".
Learning from Learners‟ Errors Ana María Rozzi de Bergel
210
Teacher: "We use the Present Simple to describe our daily routine. Now tell me
what you do every morning, beginning like this: It is 7 AM. I am getting up".
Possible result: Learners will confuse the tenses involved.
Teacher: "We use "there is" when we want to say where things are. For example:
There is a book on the chair".
Possible result: Learners may well say "There is my pencil on the floor", or
"There is the bookstore on the corner."
Passive voice taught by transforming sentences from active into passive is usually
error-provoking. As an instance of language use, the operation exists only in limited
contexts, so learners do not find it meaningful. As a teaching-learning device, it results
in a lot of semi-appropriate, correct utterances, as it focuses on form but not on
meaning or appropriacy. The learners‟ conclusion is often that it is the same to say “I
closed the door” or “The door was closed by me” and that the utterance is fully correct
and appropriate just as long as the syntactic rules have been respected.
It is not possible to foresee all the misunderstandings and erroneous conclusions which may
arise in a classroom. It would not be suitable, either, for teachers to become obsessed with
everything they say or do, to the point of losing spontaneity during lessons. It would be useless
to try to eliminate communication problems between teachers and learners, which are a normal
part of any human relationship. The benefit of reflecting upon teaching-induced errors lies in
the awareness it creates of the need to question accepted practices and to consider learners
intelligent individuals who can understand a simple but complete explanation. Teachers should
Learning from Learners‟ Errors Ana María Rozzi de Bergel
211
not feel constrained or inhibited by the possible mistakes they can make, but neither should
they work in the blind, completely unaware of the consequences of their words and actions in
the classroom.
3.d.2 Compensation-provoking teaching
When analysing the results of language courses through error analysis, we singled out some
groups where compensation was exceptionally high and devised an informal research project
to gain some insight into its causes. We found that too much compensation may be an
indication that the course is being taught above the learners' level, or that the exigency placed
on communication is too high, or that the teacher is overambitious, or that learners are asked
to perform in open-ended activities without previous practice and preparation.
Another important factor that seems to induce compensation is the teacher's attitude towards
the language being taught. This language has to be presented as a tool for communication, and
used as naturally as the learners' native language both in life-like situations designed for
classroom practice and spontaneously, according to need. When language is presented as a
subject of study, or deprived of its communicative nature, the learners' compensatory
production in more open-ended situations is usually very high.
Regarding this point, some compensation-inducing teaching practices we may point out, are:
a) The use of native language for classroom communications not directly related to the
lesson, clearly indicating that English is the language to use for classroom practice, and
the native language for communication. This does not refer to the use of the native
language as a resource. It refers to making a clear distinction between the use of
English for lesson-related communications and the use of the native language for
personal communications, even stressed by remarks such as "Lo voy a decir en
Learning from Learners‟ Errors Ana María Rozzi de Bergel
212
castellano para que lo entiendan" or "Ahora que estamos fuera del aula, hablemos en
castellano". ("I am going to say it in Spanish for you to understand it", or "Now that
we are not in the classroom, let's speak Spanish")
b) The teacher's permanent request for examples or strings of examples of questions,
commands or other exponents of language functions that should elicit a response, when
this response is not requested. Questions are left unanswered, invitations are neither
accepted nor turned down and exclamations of surprise or anger are met with dead
silence. Textbooks often provide tables, matching activities or other forms of eliciting
questions, commands, instructions, etc. but it is not the materials writer‟s duty to
remind the teacher that these activities have to be supplemented interaction and
inserted in contexts. A similar effect is achieved by having the learner answer his own
questions, in a dialogue with himself.
c) Non-contextualized grammar practice, which causes a divorce between structure and
language use. Learners effect compensation because they do not recognise the
communicative value of the language they master and go for original, self-originated
forms.
d) Teaching above the course level or placing the communicative demands of the course
above the learners‟ level of mastery. Compensation becomes indispensable. A clear
example of this is the teacher's attempt at discussing a text that is only meant for
reading comprehension and not for an oral debate. The skill of reading comprehension
can be developed above and beyond the level of oral production and it is not
uncommon to find that these texts contain lexis and structures that learners are
supposed to recognise but not to use. Discussing the text also spoils the purpose of the
reading comprehension tasks, which is to train the learner to focus on the familiar to
Learning from Learners‟ Errors Ana María Rozzi de Bergel
213
get to the unfamiliar, grasp global meanings, make associations, inferences, skim, scan
the text, etc. When placed in the position of having to talk about the text, the learner
wants to use the language in it, something for which the reading task has not prepared
him, so he resorts to compensation.
e) Overcorrection. The learner is interrupted so often that, when he really wants to
communicate, he resorts to his native language. This produces, in the long run, the
situation described in point a) on this list and results in insufficient reliance on the
foreign language, leading to compensation.
f) Lack of correction. Learners enjoy such blissful overconfidence that they do not
monitor their language at all, and compensation becomes normal. This produces highly
remedial learners who are sometimes difficult to handle, because they are convinced
that their English is excellent, and contradicting them may erode their self-confidence.
g) Lack of transfer of language items taught in class to topics and situations within the
learners' personal experience. Conversation and writing are always centred on the
textbook topics and materials. Learners are never encouraged or allowed to talk about
themselves and their interests, to give opinions or relate anecdotes or incidents. Neither
are they asked to carry out tasks or research, write real letters or reports, etc.
In general, we may consider that approaches which present language in natural
environments, with the objective of centring the class on the learners, their level and their
needs, and where the teacher acts as a responsible monitor and co-ordinator of tasks, may be
the antidote to compensation.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
214
EXPLORATION AND REFLECTION
1. Think of the teachers you have had. Identify three examples of teachers‟
authoritarian, democratic and laissez-faire leadership. What effect did these
leadership styles have on you or on your classmates? Think about it.
2. Teach a new grammar point and record your lesson. You may use a tape-recorder
or a video camera. Listen to the recording or watch it. Evaluate your explanations
and procedures. Could any part of the lesson be error-provoking? You may find the
answer to this question if you analyse your students‟ errors.
3. Pay attention to class performance and answer: How much compensation do your
students make? If it is too much, could it be that you are teaching above their level?
Review the probable causes of compensation and reach a conclusion.
4. Explore the materials and the syllabus you are using and see if you can find
examples of error-provoking grading of difficulty, order or manner of presentation
of language items. If this is the case, consider how you can deal with this problem.
5. Do your students regard English as a means of communication or as a subject of
study? Do they speak English informally? Outside the classroom? Do they watch
films or read in English for pleasure? Get information about this from them and
describe their attitude. Is it favourable or unfavourable? If it is unfavourable, think
what you can do to change it.
6. How do you correct? List three correction methods that you use. Have they been
effective so far? Why/Why not? Analyse them.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
215
PART 3: Exploiting errors
1. Error analysis for completing progress evaluation
The usual way of measuring progress is through language tests. It is assumed that the marks
obtained will reflect the learners' progress or lack of it. This is usually a valid criterion, but
assessment carried out on the basis of test scores appears insufficient when a more
comprehensive evaluation of the learner's progress is attempted through error analysis. While
test scores may reflect achievement in terms of language performance, error analysis can tell
us whether the learner has improved his learning methods and procedures, by looking at
whether he is making more productive or unproductive errors.
The following examples show how error analysis can complement test scores in progress
evaluation:
CASE A: Ms BBB
An Intermediate Level learner taking lessons in periods of 90 minutes, twice a week. Her
course started in March, and finished in November. She took two language tests which
assessed the four skills, and the overall results were:
We might conclude that her performance at the course remained stable, requiring no major
changes in teaching or method for the following year. However, error analysis showed a
slightly different picture:
Test 1 (June 1998) Score: 79 %
Test 2 (November 1998) Score: 78 %
Learning from Learners‟ Errors Ana María Rozzi de Bergel
216
From this picture of error, we may derive some tentative conclusions:
a) Whereas her command of syntax has improved or at least remained fairly stable, she is
making a large number of errors in lexis and appropriacy. These errors hamper
communication more than purely grammatical errors, as already discussed.
b) The absence of Inappropriate, Incorrect errors shows that her grammar has improved.
c) The increase in Interlingual errors seems to point out that the learner does not rely so
heavily on the code of the foreign language for communicating. As this happens,
mainly, within the areas of lexis and appropriacy, it may show a certain divorce
between form and meaning.
d) The absence of Intralingual Errors due to Shortcutting would be a change for the better
if her interlingual errors had not increased. It might also be assumed that, as her
grammar has improved, she has departed from her trend towards Shortcutting and
resorts more to Accessing Attempts. This hypothesis would be fairly accurate if her
CATEGORIES Sub-categories 1st
TEST
2nd
TEST
ERROR
RATING
Appropriate, semi-correct 63.33 % 52.94 %
Semi-appropriate, correct 23.33 % 32.24 %
Semi-appropriate, semi-
correct
6.66 % 8.82 %
Inappropriate, uncorrect 6.66 % -
ERROR TYPES Syntactic 64 % 52.94 %
Semantic 23 % 38.24 %
Semantic-syntactic 13 % 8.82 %
ERROR CLASS Intralingual errors 67 % 47.06 %
Interlingual errors 33 % 52.94 %
CAUSES OF
INTRALINGUAL
ERRORS
Accessing Attempts 65 % 87.50 %
Simplification 25 % 12.50 %
Shortcutting 10 % -
COMPENSATORY
ERRORS/UTTERANCES
5.88 % 6.66 %
Learning from Learners‟ Errors Ana María Rozzi de Bergel
217
Interligual errors were all Appropriate, Semi-correct utterances, which is not the case.
She does not resort to Shortcutting for structuring language syntactically, but copies
meanings and lexis from the native language.
In view of this, we may suspect the learner is not copying cues or adhering to first forms
or meanings learnt in the foreign language because the "first forms or meanings learnt" she
is clinging to, are those of her native language.
Comments:
Ms BBB needs to acquire greater fluency, precision and lexical ability. She should not be
taken into more complex grammar before she has managed to put her acquired knowledge to
communicative use. She should not be promoted to the following level, even though she has
passed this test, without a period of consolidation and adjustment. Her progress in the
syntactic area does not represent much progress in the communicative area, and if we give
priority to communication over correctness we may consider she has not made much progress
or even that her performance has deteriorated.
CASE B: Miss CCC
An adult learner taking lessons in 90 minute periods, twice a week, from March to
November. This is an analysis of three tests, each taken at the end of a nine-month language
course. Miss CCC went through three levels of instruction: Elementary, Pre-intermediate,
Intermediate. The results were:
Test 1 (November 2003) Score: 84 %
Test 2 (November 2004) Score: 84 %
Test 3 (November 2005) Score: 73 %
It seems that Miss CCC's performance remained stable for two years and then deteriorated.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
218
Error analysis of her production shows that this is not exactly so:
CATEGORIES Sub-categories Test
1
Test
2
Test
3
ERROR
RATING
Appropriate, semi-correct 74 % 60 % 72.22 %
Semi-appropriate, correct 26 % 36.67 % 22.22 %
Semi-appropriate, semi-
correct
- - -
Inappropriate, incorrect - 3.33 % 5.56 %
ERROR TYPES Syntactic 74 % 56.67 % 72.22 %
Semantic 26 % 36.67 % 22.22 %
Semantic-syntactic - 6.66 % 5.56 %
ERROR CLASS Intralingual errors 52.18 % 63.33 % 72.22 %
Interlingual errors 47.82 % 36.67 % 27.78 %
CAUSES OF
INTRALINGUA
L
ERRORS
Accessing Attempts 66.66 % 63.20 % 84.62 %
Simplification 33.33 % 31.58 % -
Shortcutting - 5.26 % 11.11 %
COMPENSATORY ERRORS/UTTERANCES 4.53 % 3.33 % 5.56 %
Tentative conclusions:
1. In the last test, there is an increase in the number of Inappropriate, Incorrect errors but
it is not accompanied by an increase in the number of Semantic-Syntactic errors.
2. The third test shows more achievement than the second in the area of Semi-
appropriate, Grammatical errors and Semantic errors.
3. There is significant and steady improvement in the percentages of Interlingual and
Intralingual errors. In this area, the third test is the most balanced.
4. She has stopped simplifying but resorts to shortcutting more than before. This is
perhaps the result of the reduction of Interlingual production and simplification: the
effort to adhere exclusively to the code of L2 is too great, and sometimes she cannot
cope.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
219
Comments:
The learner makes more errors than before, but not very serious ones. She may need to slow
down, to go through a period of consolidation.
These analyses offer some insight into the rules and processes governing the learner's
interlanguage. However, when attempting error analysis of a learner's production at a single
test, several considerations have to be made:
a) The qualitative analysis of a learner's production at a language test cannot override the
value of the quantitative appraisal. It is unquestionable that a learner who makes very
few errors is doing better than one who makes a lot, even if they are of a productive
type.
b) The result of analysing errors made in one test will merely show the state of the
learner's performance in that test, not the state of his interlanguage as a whole.
Taking the result of a single test as an indication of a learner's level of achievement is
just as misleading as taking the analysis of the errors made at a single test as the
description of the learner's interlanguage.
c) Even though error analysis can disclose new information about the learner‟s progress
when contrasted with mere test scores, it has to be supplemented with process
evaluation by the teacher and self-evaluation by the learner.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
220
EXPLORATION AND REFLECTION
1. Sometimes, teachers have a high opinion of a learner‟s performance and then feel
surprised when that learner does not pass a test. Has this ever happened to you? Reflect
upon the experience. Why do you think it happened? Was there anything you did not
take into account? Anything that misled you?
2. Give a writing test to your students. Analyse the errors they make. Are all the “good”
learners making productive errors? If they are not, review your assessment of your
learners.
3. Given two students with identical test scores, do you think that if one of these learners
had made more productive errors than the other, he/she would deserve higher marks?
Make a list of reasons for your answer.
4. Compare the written production of two learners, in the same task. Does error analysis
show that the learner who has made fewer errors has also made more productive
errors? Is there any correspondence between the quality and the quantity of their
errors?
Learning from Learners‟ Errors Ana María Rozzi de Bergel
221
2. Error exploitation vs. error correction
Norrish (Ibid.) proposes an experiment which I have often carried out at courses and
seminars on error analysis. He asks teachers to define errors, and the most frequent answers,
are:
They are experiments to confirm hypotheses.
They are learning attempts.
They are features of experimentation on the part of the learner.
Then, he asks teachers to write down what they do about errors in class, and the most
frequent answers are of the following type:
I always correct errors made in writing, but I don‟t interrupt students to correct
them when they are speaking.
I write down errors and then correct them with the class.
I point out errors and ask the students to self-correct them.
I signal to students that they have made an error and wait for them to correct it.
Norrish then asks teachers to substitute the word “error” in these statements for the
definition they have produced, so the results are, for example:
I always correct learning attempts made in writing.
I write down experiments to confirm hypotheses and then correct them with the
class.
I point out features of experimentation on the part of the learners and ask the
students to self-correct them.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
222
I signal to the students that they have made a learning attempt and wait for
them to correct it.
These bizarre statements create awareness that teachers do not always walk their talk or that
there is a gap between what we know and the actual implementation of this knowledge in our
teaching practice. I have tried the experiment and found it rather aggressive but extremely
effective. Teachers feel exposed, but if the procedure is spiced with a bit of humour, the initial
embarrassment is overcome and everybody gets down to work with a no-nonsense attitude.
The experiment provides the foundations for delving into the question of how to treat errors as
items in the learner‟s curriculum rather than focus on correcting them. This is the essential
change of mindset we need in order to begin discussing methods and activities for error
exploitation.
2.a Problems with correction
Lack of correction leaves learners helpless and makes teachers passive witnesses of learning
processes of mixed results; a laissez-faire situation in which some learners will develop
fossilised errors and reach their language ceiling sooner than if they have the teacher‟s
professional support and guidance (Scarcella & Oxford, 1992)
James (Ibid.) proposes subject-sensitive correction, that is to say, centred on the learner and
respecting individual learning styles and preferences. To achieve this, he advocates creating a
non-threatening learning environment and exploring learners‟ preferences regarding how they
want to be corrected. Among other aspects of corrections, he highlights the role of the learner
as an error analyst, when he is given the opportunity to contrast an interlanguage form with the
correct standard form and draw conclusions which might apparently increase his language
awareness, but he makes a distinction between language awareness and consciousness-raising.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
223
Language awareness would be the learner‟s implicit knowledge turned explicit and
consciousness-raising is getting explicit insight into what one does not know implicitly of the
foreign language.
I will suggest that error treatment, though not necessarily correction, is an integral part of the
teacher‟s work and it is widely accepted nowadays that it should not entail criticism,
punishment or aggression, but should encourage reflection, discussion and contribution among
learners. Although new and creative techniques for error treatment have been developed,
correction remains highly ineffective in many cases:
a) Teachers do not ask learners to discuss their errors or to clarify the meaning of
erroneous utterances (authoritative explanation) but diagnose and describe errors on
the basis of their informed assumptions (expert reconstruction), which may be wrong.
They correct errors which have never been made. We have already discussed this
problem, but we will provide one more example here:
Example:
Learner: I crashed into another car and I saw the driver, I was joke, because it was a friend
of mine I was trying to contact for several days.
Teacher: (Authoritative explanation) Do you mean you started joking with him? Or that it was
funny?
Learner: Yes, it was funny, and I was joke.
Teacher: (Expert reconstruction) …. and I joked about it…. (Writes the sentence on the board)
Learner: (Authoritative explanation) No, no, I was surprised. Very surprised, it was a “jock”.
Teacher: (Authoritative explanation) Oh! You mean “shocked”!
Learner: Yes. Yes, I was shocked.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
224
The example shows, once again, how the root of the error is not detected unless the teacher
asks for authoritative explanation, without jumping to conclusions.
b) Teachers encourage self-correction, but errors may be at a pre-systematic stage (Pit
Corder, Ibid.), in which the learner is unaware of a particular system or rule in the target
language, makes random errors and cannot correct himself or account for the forms he has
chosen. Under these circumstances, self-correction is impossible.
This problem renders two of the most accepted correction techniques, feedback and re-
formulation (Lewis, Ibid., Krashen & Terrell, Ibid.), practically useless. Among advocates
or re-formulation, dialogues like the following, taken from Lewis (Ibid.) are considered
enough to attract learners‟ attention to their errors:
Teacher: See you all on Thursday.
Learner: No, we don‟t are here on Thursday.
Teacher: You won‟t be here on Thursday? Why not?
Learner: We don‟t must be here.
Teacher: You don‟t have to be here? I don‟t understand. Why not?
Learner: It is sports day. We are to the stadium. So we don‟t are in school.
Teacher: Oh, I see. It‟s sports day so you won‟t be here on Thursday. Well, let‟s hope the
weather is good.(p.175)
The teacher has acted, in fact, as an interpreter, and the learner has remained totally
unaware of his errors, because they are at pre-systematic level. Had they been at
systematic level, he would have profited from re-formulation and corrected his utterances.
It is not an exaggeration to suppose that the learner whose errors are at pre-systematic
stage will find the teacher‟s constant paraphrasing of his utterances slightly bizarre. In the
Learning from Learners‟ Errors Ana María Rozzi de Bergel
225
case in hand, a highly productive master error, the overgeneralised use of “don‟t” to denote
a negative form, which could have helped the learner advance in his command of all
negative forms of verbs, has been left unexploited. The only justification for the procedure
suggested by Lewis is that the learner and the teacher were having an informal
conversation, probably outside the classroom, and the teacher may have preferred not to
hamper fluency in that context. As a method of error correction, it can be termed a method
of non-correction.
Lewis‟s and Krashen & Terrell‟s idea that teachers should simply provide enriched input
and hope that learners will acquire it is not even effective in second language learning
environments, as we all know immigrants who retain their accent after forty years of
residence in the host country, that is, of forty years of exposure to high-quality input.
Krashen & Terrell (Ibid.) themselves state that it is not clear whether re-formulation
actually encourages learners to speak more accurately and correctly.
The issue of stages of systematicity in errors is also relevant for giving feedback. Writing
codes beside errors, as proposed by Bartram and Walton (1991), will work if the error is at
systematic or post-systematic stage and the learner is capable of self-correction. An error
marked GR, in their proposal, is an error in grammar. The pre-systematic learner who is
not even aware of the existence of an error in that utterance will probably be unable to take
any action on the basis of the teacher‟s coding.
c) The teacher encourages one-to-one peer correction and does not monitor it. Learners
develop personal grammars which they gradually and partially adjust to the ideal system of
the target language (Selinker & Lamendella, 1981) and the person in charge of correcting a
peer may be using a rule or concept from his personal grammar. The learner being
corrected will simply change one mistake for another.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
226
d) Teachers edit, rather than correct, learners‟ written production:
Example:
Learner‟s production: “.... I was surprised when I saw the look on the face of Mary.”
Teacher‟s correction: “I was surprised to see the look on Mary‟s face.”
The learner is confronted with an “improved” version of his sentence. The focus of the
correction, which should have been placed on “the face of Mary”, for example, by simply
underlining the phrase and discussing it with the learner, is shifted to the whole sentence,
so that the learner cannot really know where he went wrong.
Lewis (Ibid.) advocates editing as a feedback technique, stating that, instead of
improving slightly on the learner‟s utterance, the teacher should provide a sample of
correct, appropriate language. The suggestion seems mistaken on two counts: I) the learner
may not be ready for a competent native speaker-like utterance and fail to grasp it
altogether; or feel that what he attempted to say was completely beyond him, with the
subsequent feeling of frustration; II) the item in his personal curriculum, the error he did
make, is not addressed, consequently, he is denied the opportunity of learning from it.
e) The teacher prompts words or phrases to a learner who does not command the structure of
an utterance and finds no place for the prompted elements, or who does not understand
where the mistake is because he has not been led to analyse his production or because his
personal curriculum leads him to focalise on something else at the moment.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
227
Example:
Learner: I not like to go to shops with my mother.
Teacher: DO not like .....
Learner: Yes, do not like to go to shops with my mother.
Teacher: "go shopping ....."
Learner: .... to go to shops with my mother go shopping.
Teacher: No. "I do not like to go shopping with my mother" Don't say "go to shops", say
"go shopping".
Learner: I do not like go shopping with my mother.
In this particular case, the teacher should have realised that addressing “I not like” was a
priority, as it is obviously a master error. She could have continued the conversation, so as not
to interrupt communication, to go back to the error later and exploit it. Trying to correct it on
the spot was not possible, as the learner was not monitoring his syntactic production but
concentrating on meaning.
This takes the discussion to the issue of whether to interrupt or not to interrupt learners when
they make mistakes. In general, it is not advisable to interrupt learners repeatedly and teachers
should prefer to keep records of mistakes and deal with them at a later stage. However, there
are also items which should receive immediate attention at the moment when the learner‟s
mind is ready for them. We will deal with these problems in the next section.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
228
2.b Error exploitation
Error exploitation, rather than correction, can enhance learning because it aims at exploring
the learner‟s previous knowledge and assumptions about the language, encouraging reflection,
providing authoritative input from the tutor or other sources and opportunities for the learner
to try out the resulting revised hypotheses about the target language. It is aligned with Kolb‟s
Experiential Learning Model (1997):
Figure 13 – Experiential learning model
The stage of observation and reflection will be enriched by vicarious experience: readings,
explanations by tutors or any other forms of input and contributions from peers. In this
process, the teacher‟s role is that of a consultant, facilitator and monitor and learning takes
place in a continuous spiralling of experiences, each loop reaching higher levels of
comprehension and apprehension of knowledge, with reflection acting as the vehicle for
further learning. It should also be noted that conceptualisation is at the heart of the learning
process, so learners whose hypotheses are predominantly pre-conceptual will need further
guidance.
Concrete experience
Observation and
reflection
Active
experimentation
Abstract
conceptualisation
New concrete experience
Learning from Learners‟ Errors Ana María Rozzi de Bergel
229
In their discussion of effective grammar teaching, Ellis and Gaies (1998), advocate the
development of four skills: noticing, by which learners analyse the input and focalise on what
is to be learnt; training in being able to listen to grammatical features of spoken language; a
grammar-discovery approach to texts and finally, monitoring, where learners are asked to use
their explicit knowledge of grammar to identify their errors and correct them, particularly
those they typically make.
Although their procedure did not refer exclusively to error exploitation, it is interesting to
note the role they assign to awareness, or the capability to use explicit knowledge for language
acquisition. Ellis and Gaies see this awareness as a possible means of dealing with the
requirements of the built-in syllabus, which we have called personal curriculum.
The starting point for error exploitation is to determine what errors are more significant and
typical as personal curriculum items. Our research identified three main types, which we have
already discussed in Part 1:
a) Errors that worry the learners, even when the teacher does not consider them serious or
when they are not within course level.
b) Errors which are common to a group of learners or to the whole class, particularly if
they are recurrent.
c) “Master errors” which affect vast language areas because they have ramifications into
several sub-systems.
This list is by no means complete and other types of errors may be important in the personal
curricula of different learners, according to individual needs. These three types were singled
out as most typical manifestations of personal curricula because their exploitation seemed to
have a particularly strong impact on learners‟ progress.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
230
In a study carried out in 1999, we asked a group of fifty five teachers from all over
Argentina what errors they would consider important, and the two most chosen categories
were “Errors which hamper communication” and “Errors which have to do with the items
being taught”. The survey was repeated in 2005, with a group of 123 teachers, with the same
results. Note that most errors hamper communication, although not all of them block it, and
that the items being taught are the most likely to produce errors, as learners experiment with
the new forms in order to internalise them. Errors in these items are simply necessary features
of experimentation and probably developmental, not “serious”. Focusing exclusively on the
topic of the lesson may cause the teacher to overlook errors which are manifestations of the
learners‟ personal curricula or to those which are becoming fossilised – that is, the really
“serious” ones.
The notion of errors which hamper communication, when further explored by asking
teachers to provide examples, turned out to be more related to instances of compensation,
when communication was interrupted because learners were at a loss for words, or to cases
when a misinterpretation on the part of a particular learner caused short circuits in the flow of
a group activity.
In both surveys, teachers regarded errors in grammar as less serious than semantic errors or
errors in appropriacy, which is perhaps a result of the present awareness of the value of lexis
in communication. The same question was posed by Norrish (Ibid.), in 1983, and the answer
he obtained was that grammar was more important than communication.
How to treat errors? As items in the learner‟s personal curriculum. First, the stigma of
embarrassment and criticism has to be removed from error production. Items in the course
syllabus are not undesirable. Why should items in the learners‟ curricula be objectionable?
Learners have to be explained that their errors show the teacher a parallel curriculum, that they
Learning from Learners‟ Errors Ana María Rozzi de Bergel
231
are positive signs of learning and valuable contributions to the course. The teacher should state
in no uncertain terms that the course has a pre-set curriculum which needs to be supplemented.
As a curriculum is a public document (Ur, 1996), a list of the learners‟ contributions should be
available to the class and they should know this is their curriculum, so they should be proud of
participating in its design.
They should also know they are expected to learn the curriculum they build up, as it is not
accessory or optional but will become an integral part of the course. The class will then exploit
errors as curriculum items, following Kolb‟s cycle, in the following way:
a) Concrete experience: The teacher records a number of master errors, systematic
errors of the group or errors that worry the learners, produced during oral work or in
writing, in fairly open-ended activities, in the manner we have outlined. Authoritative
explanation should be subtly sought, by inviting re-formulation, confirming
interlocutor hypotheses or making remarks such as, “Did you say XXX because you
thought XXXXX?”. This brief discussion may take place during the activities, if the
teacher is participating in them, or afterwards, if the teacher has acted as an observer.
When we talk about open-ended activities, we mean interactive situations with an
information gap, role-plays, mock negotiations, presentations, telling anecdotes or
stories, giving opinions, carrying out group or pair work to solve a problem,
summarising, reporting and generally performing within an activity where the
language outcome is not fully predictable, and language is a means to carry out the
task. In close-ended activities, such as blank-filling, transforming sentences from direct
to indirect speech or multiple choice questionnaires, the expected outcome is 100%
predictable and can be written in an answer key. The choices the learner has to face are
limited and may not reflect his learning hypotheses, whereas in open-ended tasks he
Learning from Learners‟ Errors Ana María Rozzi de Bergel
232
can choose from the whole of his inventory of language forms and even create new
ones, thus offering the teacher a glimpse at the contents of his personal curriculum.
b) Observation and reflection, plus abstract conceptualisation aided by vicarious
experience and carried out through collaborative learning: The following lesson, the
teacher presents a list of erroneous utterances containing a selection of the errors
recorded during the previous lesson and asks the class to correct them, acting as a
coordinator of the discussion. It may not be advisable to transcribe learners‟ utterances
literally in these activities, but to produce a slightly modified version, so that learners
will recognise their errors without feeling exposed.
The selection should include not more than ten errors, so as not to create the
impression that the class are producing nothing but wrong language and besides,
working on a long list of errors may prove tedious and go against learning. It is also
important to select errors belonging to different learners even if two or three of them
have produced the most part. The idea is not to focus on the slower learners or to create
the conditions for the faster ones to show off their achievements.
A fundamental part of this stage of error exploitation is to thank learners for their
contributions to the curriculum and remind them that they are designing part of the
course. The most usual reaction to this is that, while the teacher tries to keep the
providers of the errors anonymous, they are delighted to recognise their contributions
and call out, “That‟s my error! I said that!” when they identify their utterances.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
233
Examples of errors, their discussion and exploration:
1 . Error: I want going to the party.
Discussion and exploration: a multiple-choice exercise including in each question:
a. the correct form of the utterance.
b. the error that a particular learner has made.
c. another kind of error referred to the same utterance.
Choose the correct form:
1. I want ............... to the park.
a. go
b. to go
c. going
2. Error: I suggest you to come back tomorrow.
Discussion and exploration:
Choose the correct form:
1. I suggest that you write this letter again.
2. I suggest you to study more.
3. I suggest that you waiting for her here.
The following is an example of a guided discussion presenting two conversations. Here,
the learner‟s utterance has been transcribed verbatim:
Learning from Learners‟ Errors Ana María Rozzi de Bergel
234
3. Error: Like you this cake?
Discussion and exploration:
Where is the error?
1
- Do you like this cake?
- Yes, I do. May I have a piece?
2
- Like you this cake?
- Yes, I like. May I have a piece?
More creative activities for discussing and exploring errors include:
Error auction. Learners are given a fictitious sum of money to spend and errors are
auctioned. They should seek to buy those that interest them, making a good
investment. Once they are in possession of the purchased errors, they have to analyse
them and produce the correct forms. This discussion has to be shared with the class
or done in groups.
Error hunt. Several texts containing the errors to be dealt with are presented to
groups of learners who have to discuss them, find the errors and correct them. The
activity may be presented as a game and a prize may be given to the learners who
find and correct the most errors.
Error hangman. In case of spelling errors, learners play hangman to give the correct
version of the erroneous words.
Error domino: wrong form+correct form. The errors are written on cards and
there is another set of cards containing the correct forms or sentences. Learners play
Learning from Learners‟ Errors Ana María Rozzi de Bergel
235
domino, finding the matches. Then, they discuss the matches with the class and the
teacher.
Error draw. Errors are written on pieces of paper and put into an envelope or box.
Learners take turns to take a piece of paper out of the box, read the error
and correct it, discussing it with the class.
Error basket. Learners keep a “basket of errors”. It is a box or an envelope where the
most important errors discussed in class are kept, written on pieces of paper
or cards, with their correct form. This is practical for revision work.
During this discussion, learners complete a personal error log where they record the
errors they find more interesting in the activities proposed by the teacher.
c) Active experimentation and new concrete experience: The teacher devises a
communicative activity and instructs the class to use items from the corrected list of
errors while carrying it out, presenting them as “useful language”.
The activity should provide a suitable context for the use of this useful language. If
we referred to the errors discussed in “a”, it could be, for example, a task for pairs of
learners where one is the travel agent and the other is a client looking for a package
holiday. The activity might include an information gap, with two role cards: one with
information for the travel agent about available holidays and the other with hints to the
client as to his/her preferences and means.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
236
The list of useful language should not be spoon-fed to the class or copied from the
previous activities but elicited from the learners at a preliminary discussion of the task, for
example, by asking them, “Think about your ideal place for a holiday. Where do you want
to go? What do you like about that place? Listen to what Peter says. Do you have any
suggestions for him?” and so on. The answers or comments arising from this discussion
which contain the desired items make up the list of useful language and are written on the
board. Each learner is free to choose which items to use but it should be clear that they
have to be included in the activity.
Role-play, discussions of topics, presentations, negotiations, interviews and other oral or
written communicative activities are useful at this stage. The teacher‟s role during the
development of the task or while correcting written work thus produced is twofold:
Contrary to prescribed practices, the teacher will interrupt the learners to correct errors in
the forms which they have just practised or which appear on the list of useful language,
that is, recurrent errors in their personal curricula. At the same time, the teacher will record
new errors and store them for future exploitation. This is the “new concrete experience” of
producing the right form or the error. If the error re-appears and the teacher does not
interrupt the learner to correct him, she will have lost “the” opportunity for incorporating
the correct form to the learner‟s language: the moment when the item is relevant and
meaningful to the learner, because he is using it to express his ideas in a communicative
situation.
d) Follow-up: The items thus dealt with should be practised during subsequent lessons, in
properly contextualised and communicative activities, and included in tests. Learners
should review their logs periodically, so the teacher should provide contexts to allow
them to do this, similar to the interactive tasks in the previous step. A review of the
Learning from Learners‟ Errors Ana María Rozzi de Bergel
237
error log could consist, for example, of analysing the errors made in a writing activity
and checking whether some of them were already in the error log. If they were, the
learner should know that he has not grasped the correct form yet and may need more
experimentation with that language item. An autonomous learner will engage in this
experimentation on his own, others may need encouragement or reminding. However,
learners who keep error logs and review them periodically tend to strengthen their
autonomy for learning by doing this.
The teacher also keeps an error log, containing not only the texts of all the activities
devised to exploit errors, but also a record of those which seem more productive and
should be reviewed periodically or those which are recurrent and need further analysis.
This method turns the teacher into a course designer, but this is a natural consequence
of following two curricula, one of which is built up as the course develops.
At courses where these methods were implemented, teachers asked whether listing
everybody‟s errors and having the class discuss them would not be meaningless for those
learners who could claim authorship of only one or two errors and might not be interested in
the rest. Experience showed that learners found it useful to review and confirm their
hypotheses about the items they seemed to master, and were glad to be given the chance to
enlighten other learners about them. In their error logs, they discarded the elements which did
not need special attention.
The issue is not only what errors to use or what procedures to follow but basically, the type
of team spirit and group dynamics the teacher creates: non-judgemental, cooperative,
encouraging, and how seriously the personal curricula are addressed. If the practice of
exploiting errors is dropped, or done only occasionally, learners will feel discouraged or lose
interest in errors or faith in the teacher.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
238
The method of error exploitation I have outlined relies so heavily on motivational factors
that it is inadvisable in cases where the group dynamics is determined by external factors, for
example, if an authoritarian boss is taking lessons with three of his most submissive
employees. In all probability, the boss will not want to share his errors with the class and the
employees will not want to discuss them. As usual in the classroom and in life, common sense
has to prevail and teachers should base their practice on the learners, not on recommended
methods.
Figure 14 – The process of error exploitation.
1.SIGNIFICANT ERRORS
2. RECORD
3. DISCUSSION AND
CORRECTION
4. COMMUNICATIVE
PRACTICE
5. COURSE SYLLABUS:
REVIEW AND TEST
Observation - Reflection
Abstract conceptualisation
Concrete experience
Active Experimentation
New concrete experience
Learning from Learners‟ Errors Ana María Rozzi de Bergel
239
2.c Comments on error exploitation
My team‟s first experience with error exploitation as part of a way of addressing the
learners‟ personal curricula was an informal experiment carried out at a company where we
had a group of eight learners taking an intermediate level course.
The company asked us to divide the course into two, because four of those learners needed
to receive specific instruction in language related to some new production methods. In a few
months, they would be sent to Japan to take a course in English about these innovations.
We were left with two courses, then, which I will call Course A and Course B. Course A
continued using the materials and curriculum which had been chosen for them, which seemed
to be yielding good results. Course B was taught with materials dealing with the production
processes in question and I decided to experiment with error exploitation and teaching which
would address their personal curricula. The teacher received instructions to
answer questions which seemed to be “out of the blue”;
satisfy the learners‟ curiosity about language which seemed to be above their
level of mastery or was unrelated to the lesson;
give complete explanations of grammar rules, including their exceptions,
leaving grading of difficulty aside;
exploit errors in the manner I have outlined in the previous section.
The teacher soon reported a remarkable acceleration in the learners‟ progress and the
elimination of most of their errors from their interlanguage. A learner from Course A was
transferred to Course B because he was included in the group going to Japan and remained in
this course for two months. At that point, he was withdrawn from the travelling team, so we
were asked to send him back to Course A. His level was, by that time, considerably higher
Learning from Learners‟ Errors Ana María Rozzi de Bergel
240
than that of the other members of the group, his former classmates. It should be noted that the
learner in question had always been “the slowest” in the original class.
Besides the treatment of errors and the efforts to address the learners‟ curricula, the main
differences between Course A and Course B were the learners‟ motivation and the specificity
of the materials. One could assume that the learners in Course B learnt faster because they
were more interested in the topics dealt with and eager to go to Japan and take their training
course.
To further explore these phenomena, the methods used at Course B were implemented in
other courses, where learners‟ motivation was not particularly high and who had no distinctive
characteristics. Experiments were also made at one-to-one courses.
In these new implementations, teachers observed a radical reduction in the number of errors,
to the point that their exploitation gradually became unnecessary. Improvement was coupled
with more fluency and an increase in appropriacy, as learners‟ accuracy ceased to be a cause
of hesitation or insecurity.
The acceleration of progress was dramatic, with learners passing on to the following levels
after at least 25% fewer hours of class than were normally needed for the same type of learners
and the same programme of instruction. In three groups and at a one-to-one course, the
textbook had to be discarded and changed for a more advanced one when learners had gone
through little more than half the units.
Although progress was fast, there was no evidence that the learners‟ language ceilings were
raised and we might consider that learners simply reached the end performance in their
curricula at a faster pace. The picture of error was the standard one for all the learners in the
error exploitation programmes, with unproductive intralingual errors increasing as they
reached their ceiling. The only significant difference was the number of errors and the fact that
Learning from Learners‟ Errors Ana María Rozzi de Bergel
241
the percentages of semi-appropriate or inappropriate utterances and semantic errors remained
fairly stable, so we might say they reached language ceilings of a higher communicative value.
Whether this means raising their language ceilings remains a moot point.
When the method of error exploitation and the concept of addressing the learners‟ curricula
were recommended to other language schools and their teachers were trained to implement
these principles, the reported results were the same as for the initial experiments. A colleague
reported the same effects with a group of children, but no further information was received
about experiences with young learners.
The extent to which a personal-curriculum-based approach can accelerate and enhance
learning remains an interesting research topic, as the experiments and experiences were rather
informal and more concrete information about the possibilities of these implementations might
shed more light into the significance of allowing the learner‟s personal curriculum determine
our practice, as well as to how many of “our cherished notions” (Pit Corder, Ibid.) we should
review.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
242
EXPLORATION AND REFLECTION
a) What errors worry your learners? Have they asked you questions which were
related to their errors but not necessarily to the topic of the lesson? Make a list
of these errors and devise ways of exploiting them
b) Which of these are more likely to be “master errors”? (A very tentative opinion,
as you do not know the contexts of the errors)
I say she.
He came fastly.
My brother didn‟t seen the accident.
Susan musts works harder.
Remember: a classification is not possible, as master errors may be personal,
but in general, they tend to have more ramifications and affect more sub-
systems of the language than ordinary errors.
c) Think of communicative situations in which your students might need to ask
questions. What errors could you address with these activities?
d) Try out the method of error exploitation outlined in this chapter for two
months. Reflect upon the results.
e) Teach a course according to the personal-curriculum-based approach described
here. Observe whether progress seems to be faster than what you consider
“average”.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
243
3. Teaching learners at their language ceiling
The title begs the question; can learners at their language ceiling be taught? We might well
suspect nobody can ever be taught, and in fact, we should add this concept to the list of our
“cherished notions”. The issue may be approached from a different perspective if we conceive
of teaching as guidance, advice, empathy with students and continuing professional
development. Then, we will not think about teaching as the jar and glass metaphor but rather
as a process of shared growth. In this latter paradigm, learning will also be taken as having
many facets and not just as a linear process of adding so many items to the learners‟ repertoire
of language forms. It is within this frame of mind that we should approach learners at their
language ceiling.
3.a Recognising language ceiling
We have discussed the definition of language ceiling as a point beyond which learners refuse
or an unable to incorporate new and more complex language forms and remain at the level
they have reached. Interlanguage is relatively fossilised and some errors have become a
permanent part of the learner‟s grammar. I have re-defined language ceiling as the point where
learners attain the objectives or end performance in their personal curricula, or find themselves
unable to move any further. I have also explained that low language ceilings determined
primarily by the learner‟s cognitive structuring of language are usually marked by an increase
in unproductive errors due to short-cutting, simplification or severe interference, resulting in a
large number of semantic errors and utterances which are semi-appropriate, semi-correct.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
244
It is not completely correct to say that learners at their ceiling cannot go on learning; it is
better to analyse in what areas their progress seems to stop. Grammar and structure are usually
the most rigidly fossilised areas, but some lexis can still be incorporated and learners can gain
more fluency within their language level. It may seem that learning a few more words without
having access to more complex grammar is not much of an improvement, but this is a relative
view which depends on the relevance of that vocabulary in the learner‟s personal curriculum.
In the area of skills development, reading comprehension can be developed above and beyond
listening comprehension, which tends to fossilise at the same level as grammar and lexis,
whereas writing seems to remain at an even lower level than talking.
Detecting the beginning of language ceiling is of paramount importance, but not easy. It is a
crucial assessment of a learner, which can make all the difference between success and failure.
If ceiling is not detected, and the unaware teacher tries to push the learner beyond it, the
learner will become a remedial case. If ceiling is wrongly detected and a learner is treated as if
he had reached it when he has not, progress may be stunted or at least slowed down due to
lack of appropriate input and challenging tasks. In both cases, learners will feel discouraged
and frustrated, because the course will not address their needs.
As learning takes place in plateaux, it is extremely difficult to distinguish between a stage of
stabilisation (Selinker, Ibid.) or the lower part of a curve in U shaped-behaviour (Kellerman,
Ibid.) (a period of great achievement, followed by a period of relative confusion represented
by the dip in the “U”, followed by another leap forward in progress) and the onset of language
ceiling. Selinker (2000) reviewed all the manifestations of stabilisation and pointed out the
difficulty of distinguishing transient stabilisation from permanent fossilisation.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
245
He also noticed the apparent dichotomies in fossilised or stabilised language systems,
described as follows:
- systematicity and variability in the same learner‟s interlanguage.
- permeability and impermeability in the same learner‟s interlanguage.
- stability and instability in the same learner‟s interlanguage.
- transition and non-transition in the same learner‟s interlanguage.
(Selinker, 2000, p. 281)
In brief, the phenomenon we have called language ceiling has two faces: there is a part of the
learner‟s language which remains stagnant and another part which is still flexible. This makes
the diagnosis of language ceiling quite elusive and the teacher may be hard put to find the
answers to two central questions.
a) How much of the learner‟s overt production (what he says and writes) reflects inner
processes of stagnation? Or are there processes of learning and progress which
remain invisible, until a leap forward occurs?
b) How can we know that the phenomena we are witnessing are signs of language
ceiling and not of a longer stage of stabilisation?
These two questions are also addressed by Selinker in the paper we are referring to and he
suggests that longitudinal studies of learners‟ production could be useful for finding some
answers. There is an aspect which these studies need to consider, however, and that has been
the focus of my research: the cognitive processes which seem to govern error production.
Studying learners‟ production would help us to describe their interlanguage better, but it
would not shed light on inner processes determining the structuring of this interlanguage.
Our classification and description of errors, coupled with the Learning Hypotheses Test,
look like better instruments to contribute to a diagnosis of a learner‟s ceiling. They do offer
Learning from Learners‟ Errors Ana María Rozzi de Bergel
246
information on inner cognitive processes and some of our longitudinal studies point to some
correspondence between certain learning hypotheses, error types and styles, and low language
ceilings.
Ceiling is initially diagnosed intuitively, by teachers who get anxious due to some learners‟
apparent lack of progress and state that these people “have stopped learning”. The reliability
of such diagnoses is very relative and depends on the teacher‟s notion of progress, learning,
and her experience and knowledge. It also depends on the teacher‟s leadership style and level
of anxiety and self-esteem. Sometimes, the diagnosis reveals more about the teacher than
about the learner.
When we began performing error analysis as a means to confirm such intuitive diagnoses, it
was often found that the learner in question was not likely to have reached his ceiling, but
perhaps other factors were holding back his progress, even health problems or trouble at work
or at home. Sometimes, error analysis also disclosed teacher or materials-induced negative
processes. In one case, a teacher who was extremely keen on getting grammatical correctness
from his learners grew desperate over the case of a learner whose written production was
deteriorating instead of improving and concluded that this learner was “at his ceiling”. Error
analysis showed that practically all the learner‟s errors were due to accessing attempts, mainly
overgeneralisation, a situation which does not point at language ceiling. Lesson observation
revealed that the learner could not cope with the multitude of grammar rules he was constantly
been fed or with the in-depth analysis of language that the teacher demanded from him. The
teacher was given advice and help and the learner‟s errors began to disappear, showing that
they were not fossilised forms.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
247
3.b Unproductive errors and language ceiling
The descriptive analysis of errors should help us determine whether a learner is at his
ceiling, especially when this ceiling has been reached too early, due to the particular
characteristics of the learner‟s learning hypotheses. It may also show that ceiling is not related
to learning hypotheses. When this is the case, the descriptive analysis may reveal a picture of
error similar to that in Figure 10, and yet, the teacher will report no progress for that learner.
Let us examine a few cases of learners whose picture of error seemed to indicate that they
had reached their language ceiling due to the nature of their learning hypotheses:
Figure 15 – Students who had probably reached their ceiling
CATEGORIES SUB-
CATEGORIES
CASE
A
CASE
B
CASE
C
CASE
D
ERROR RATING
Appropriate,
semi-correct
44.06 % 62.50 % 62.50 % 89.67%
Semi-appropriate,
correct
37.29 % 7.50 % 15 % 3.44 %
Semi-appropriate,
semi-correct
6.78 % 2.50 % 10 % 6.89 %
Inappropriate,
incorrect
11.86 % 27.50 % 20 % 0
ERROR
DESCRIPTION
Semantic 40.68 % 62.50 % 62.50% 89.67 %
Syntactic 37.29 % 7.50 % 15 % 3.44 %
Semantic-
syntactic
22.03 % 30 % 30 % 6.89 %
ERROR
CLASSIFICATION
Interlingual 30.51 % 20 % 17.50 % 41.37 %
Intralingual 69.49 % 80 % 82.50 % 58.63 %
CAUSES OF
INTRALINGUAL
ERRORS
Accessing
Attempts
21.87 % 21.87 % 48.48 % 58.82 %
Simplification 43.75 % 43.75 % 42.42 % 29.41 %
Shortcutting 27.50 % 27.50 % 9.10 % 11.76 %
Compensatory errors/utterances 10.17 % 7.50 7.50 % 0
Learning from Learners‟ Errors Ana María Rozzi de Bergel
248
CASE A: The learner produces a large number of semi-appropriate, semi-
correct utterances, and compensatory production is high. This is combined with
a high percentage of intralingual errors due to simplification and shortcutting.
CASE B: An extreme case of a learner who reached his ceiling at the end
of the elementary stage. Subsequent efforts by various teachers and different
tailor-made programmes of study did not succeed in taking him to a higher
level. However, the learner was extremely keen on learning English, as this was
important for his career. The diagnosis of the Learning Hypotheses Test
showed he would probably cling to the first systems learnt and reject the
incorporation of new knowledge that might call for a readjustment or
flexibilisation of these systems. We thought this meant he would structure L2
on the basis of L1, trying to simply reproduce the system he knew, but we were
proven wrong: He stuck to the very simple elements of the code of L2
(Adherence to First Form or Meaning Learnt) and also tried to discard what he
could not grasp (Simplification) He did not use L1 as a source of reference, for
he had structured it as a pre-conceptual system and this was not very helpful.
Most of his production was Incorrect, Inappropriate. Note the low percentage
of problems due to Accessing Attempts, showing he is not experimenting with
the rules of L2 or trying to grasp them.
CASE C: Again, the learner prefers to simplify the code of L2 and discards
L1 as a source of reference.
CASE D: In this case, there is an overuse of the code of L1 as a model for
structuring L2. This is also shown by the low percentage of Accessing
Attempts problems.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
249
As we can see, when language ceiling seems to be low because of the characteristics of the
learner‟s hypotheses for structuring the language system, the point where learners reach their
ceiling shows a particular composition of their picture of error. We have said that error
analysis alone cannot diagnose ceiling, but if unproductive errors account for most of a
learner‟s error production, he has probably reached his language ceiling and the teacher should
be warned that an adaptation of methods and course contents is necessary.
Below you will find the results of a longitudinal study of a learner who reached her language
ceiling after only two years of instruction. The data came from 350 errors per year, made
during fairly open-ended and communicative oral and written tests, given in July and
November of each year, plus errors recorded at ten lessons, during open-ended production, by
a researcher. All the errors were authoritatively explained. The learner took one-to-one
lessons, ten contact hours per week. No important changes in teaching or the materials were
made from one year to the other, except the standard adaptations to her needs and interests,
and there were no significant events in her private life.
Ms XX was a raw beginner when she came to us, and the diagnosis of her Learning
Hypotheses Test said, "Associates by proximity. Clings to the first systems learnt. Finds it
difficult to infer general rules from concrete cases and uses pre-conceptual classification
systems. Very low degree of imagination and creativity”. In view of the poor results Ms XX
achieved after a year of learning, we asked our psychologist to give her a more in-depth LHT,
where her motivation was also analysed. Some of our psychologist's comments were:
"Unfamiliar situations, where she has to solve problems on her own, paralyse her. She needs
to be led and does not show initiative or creativity. She is passive and prefers rituals or pre-set
processes to open-ended situations. She expects THE TEACHER to achieve results. She
Learning from Learners‟ Errors Ana María Rozzi de Bergel
250
expects to be taught, not to learn. Her ambitions are unrealistic: on a scale from 1 to 10, she
places her level after 1 year of taking lessons somewhere between 7 and 10."
We worked together with the tutor preparing materials for Ms XXX, monitoring her errors
and suggesting courses of action, but she seemed to reach her ceiling despite our efforts. She
put the blame on us (to be expected, as she wanted US to achieve results) and hired a teacher
from another school. After a few months this colleague, who knew about our research, phoned
us to ask about Ms XXX‟s record and seek help.
Many variables remain out of the researcher‟s control in this analysis, of course, but the
figures are still relevant to show the picture of error which might be found at the point of
language ceiling:
Figure 16 – Ms XXX may have reached her language ceiling.
CATEGORIES SUB-CATEGORIES YEAR
1
YEAR
2
ERROR RATING
Appropriate, semi-correct 42.10 % 42.50 %
Semi-appropriate, correct 26.31 % 17.50 %
Semi-appropriate, semi-
correct
7.89 % 12.50 %
Inappropriate, incorrect 23.68 % 27.50 %
ERROR
TYPE
Semantic 42.21 % 52.50 %
Syntactic 26.31 % 17.50 %
Semantic-syntactic 31.57 % 30 %
ERROR
CLASS
Interlingual 7.90 % 20 %
Intralingual 92.10 % 80 %
CAUSES
OF
INTRALINGUAL
ERRORS
Accessing attempts
(Overgeneralisation,
confusion,
misunderstanding)
45.71 % 21.87 %
Simplification (Of the
system, of the message, of
the syllabus)
40 % 43.75 %
Shortcutting (Cue-copying,
conditioning, adherence to
first form or meaning learnt)
8.57 % 27.50 %
Compensatory utterances/errors 10.57 % 7,50%
Learning from Learners‟ Errors Ana María Rozzi de Bergel
251
There is an interesting phenomenon in the ratio between interlingual and intralingual errors,
which we observed many times: when learners seem to make progress towards the supposedly
desirable 30%-70% ratio, but are not able to structure language conceptually, the percentage of
unproductive intralingual errors rises significantly, sometimes together with the number of
compensatory utterances. What is gained in one domain is lost in the other and these learners
reach their ceiling faster than if they had kept modeling the foreign language on the code of
L1, which is, after all, a learning resource. The following case study, undertaken during the
first stages of our research into the probable causes of language ceiling, clearly shows that a
large number of interlingual errors is not necessarily an indication of language ceiling. It also
shows that avoidance of translation as a learning resource does not result in improvement of
the learner‟s production when his learning hypotheses are of a pre-conceptual nature and he
associates by proximity:
Learning from Learners‟ Errors Ana María Rozzi de Bergel
252
Learner: Roberto
Age: 28 Course: Group. 3 hours per week.
Level: Elementary First analysis: March 1992 - Second analysis: July 1992
ERROR ANALYSIS
CATEGORIES SUB-CATEGORIES March
1992
ERROR RATING
Appropriate, semi-grammatical 67%
Semi-appropriate, grammatical 23%
Semi-appropriate, semi-
grammatical
5%
Inappropriate, ungrammatical 5%
ERROR
DESCRIPTION
Semantic 22%
Syntactic 66%
Semantic-syntactic 12%
ERROR
CLASSIFICATION
Interlingual 52%
Intralingual 48%
CAUSES OF
INTRALINGUAL
ERRORS
Overgeneralisation 65%
Simplification 29%
Shortcutting 6%
Compensatory errors 5%
Roberto was the slowest learner in his group. He constructed English mostly on the basis of
the structure of Spanish.
Examples:
The learner said……… When he meant ……………
"I need a pencil for write" "I need a pencil to write with."
"My sister no go more to the school” "My sister does not go to school any more"
"She likes play tennis" "She likes to play tennis"
"We go to the work in bus" "We go to work by bus"
Learning from Learners‟ Errors Ana María Rozzi de Bergel
253
The teacher thought that the problem was clearly described: the learner was resorting to the
code of his native language too much. Therefore, she took steps to prevent this learner from
using translation to structure his utterances in L2. She spoke to him and explained that this
tendency to effect word-for-word translation was not going to yield good results, and asked
him to make a conscious effort to stop using this strategy. Roberto agreed. During lessons, his
teacher constantly discouraged him from translating and he constantly struggled to oblige.
However, he did not seem to be making much progress as a user of English. He eventually
stopped producing an excessive number of errors due to interference from the native language,
but his English was almost incomprehensible.
Examples: In the structures analysed in the first instance, he would now say
The learner said …….. When he meant. ……………..
"I need pencil write" "I need a pencil to write with"
"My sister is not go more to school” "My sister does not go to school any more"
"She like is playing tennis" "She likes to play tennis."
"We are go to work bus" "We go to work by bus"
Unfortunately, resorting to translation was the symptom, not the disease, and when this
process was blocked nothing was solved, for the second analysis of his errors was as follows:
Learning from Learners‟ Errors Ana María Rozzi de Bergel
254
ERROR ANALYSIS
CATEGORIES SUB-CATEGORIES March
1992
July
1992
ERROR RATING
Appropriate, semi-grammatical 67% 60%
Semi-appropriate, grammatical 23% 21%
Semi-appropriate, semi-
grammatical 5% 12%
Inappropriate, ungrammatical 5% 7%
ERROR
DESCRIPTION
Semantic 22% 22%
Syntactic 66% 60%
Semantic-syntactic 12% 18%
ERROR
CLASSIFICATION
Interlingual 52% 34%
Intralingual 48% 66%
CAUSES OF
INTRALINGUAL
ERRORS
Accessing Attempts 65% 56%
Simplification 29% 38%
Shortcutting 6% 16%
Compensatory errors 5% 9%
There was a dramatic reduction in the number of Interlingual errors, which was close to the
supposedly desirable 30%, but his Intralingual errors were mostly due to Shortcutting and
Simplification. As the percentage of Semantic errors and the percentage of Semi-appropriate
utterances were also very high, we suspected that Roberto was reaching his language ceiling
at the end of the elementary level. Preventing him from translating had only accelerated the
process, because Roberto had problems for conceptualising and integrating concepts into a
logical system.
The diagnosis of his Learning Hypotheses Test stated that he could not be expected to
compare, infer, organise or group elements according to logical hierarchies without guidance.
He would tend to copy, imitate, memorise, describe, relate by proximity. When the process of
transfer - positive and negative - was blocked, the learner resorted more and more to
Simplification and Shortcutting, all processes that do not seem useful for structuring of the
Learning from Learners‟ Errors Ana María Rozzi de Bergel
255
language system. The increase in Compensatory Errors showed that his problems became
greater as he advanced into the course. He lost awareness of the scope of his language
possibilities as he was exposed to more words and more advanced language than he could
possibly process and organise.
He would have obtained better results if he had been trained to translate properly. A
contrastive analysis of the two languages would have been relatively easy for him to grasp,
because he was ready to deal with facts, not with abstract concepts. As regards his
communicative capabilities, his initial attempt at constructing English on the basis of Spanish
structure had resulted in a much more intelligible production. "She like play tennis" is
obviously better understood than "She like is playing tennis".
This example should invite reflection on the approach to error analysis that considers L1 as
the main source of errors, the relative value of contrastive analysis of L1 and L2 as a way of
predicting errors and the damaging effect of not allowing or inducing learners to use the code
of their own language as a learning tool.
Despite the knowledge we gained from diagnosing learning hypotheses and analysing errors,
the question of how to determine if a learner had actually reached his language ceiling was not
conclusively answered and perhaps such an answer was and remains impossible. The most
important aspect of the problem is that ceiling exists and learners reach it, no matter how
difficult teachers may find its diagnosis, so further indicators are needed in order to define it
with more precision.
In our search for such indicators, we decided to collect students‟ comments about their
learning and their courses, teachers and materials, to find out if there were some typical
remarks that learners would make when they had reached or were about to reach their ceiling.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
256
Below is a description of some of the findings of this study, which can be considered
indicative of the value of students‟ remarks on their learning and an expression of their needs,
sometimes not even clear to themselves, let alone to the researcher.
3.b.1 The significance of learners’ comments on learning and teaching.
We observed that the onset of language ceiling was accompanied by changes in the learner's
motivation and attitude towards the course. For four years, our follow-up of learners who
were believed to be reaching or to have reached their ceiling included a record of the remarks
they made about the course or the materials at that point, which they had not made before.
What these comments had in common was:
a) They expressed the learners‟ awareness of having reached their language
ceiling,
b) They contained demands for stopping vertical development, for not learning
more advanced grammar,
c) They suggested changes in methodology, such as more communicative
practice, narrowing down topic fields, or not using higher level materials.
We will discuss the most frequent remarks and our interpretation:
I want to get more practice in conversation, and less grammar. (This remark expresses the
aims and characteristics of pedagogic implementations for learners at their language
ceiling so accurately that it could have been made by a teacher. It is also a very clear
diagnosis of language ceiling, where grammar is the most stagnated area)
I know enough grammar, but I lack vocabulary. (It was often found that learners claiming
they lacked vocabulary actually lacked fluency and appropriacy. When they reached their
Learning from Learners‟ Errors Ana María Rozzi de Bergel
257
language ceiling, they seemed to be asking for grammar teaching to stop, but they wanted
to continue developing other areas).
I need something more useful, something related to my profession/career/studies. (The
learner is asking us to narrow the scope of the language he is exposed to – one of the
recommended features of programmes for learners at their ceiling).
Why should I learn so much grammar? I'm not going to teach English. (Learners run
down what they cannot learn or cannot learn it because they are not aware of its
importance. In either case, they are asking us to stop vertical development and in a way,
send a message: "Can't you see I've had enough?")
Can't we drop the book, and use videos or articles? (Again, this is a request to stop
vertical development, to stay where they are and improve without incorporating much new
language).
Do I need all this to talk to native speakers of English? ( The learner is probably a limited
communicator even in the native language and is telling us he considers he has reached the
end performance in his personal curriculum)
I know as much as I need. (Sometimes, it is not a statement of linguistic limitations, but
simply the expressions of the learner's motivation or ambition, which will not take him
further.)
They‟ll understand me anyway. (This remark is often heard when, for example, more
advanced forms of the exponent of a function are taught. The learner can say “I want a
cup of coffee” and refuses to learn “I‟d like a cup of coffee, please”, claiming that it is
useless because those unidentified people he calls “they” will understand him anyway.
Being understood is a basic objective which can sometimes be achieved without words
Learning from Learners‟ Errors Ana María Rozzi de Bergel
258
and points to limitations in the learner‟s curriculum goals. The statement is also a sign of
“enough”.)
In general, we were greatly impressed by the learners‟ degree of awareness of the state of
their learning and the expression of their needs, and also with their specific recommendations
for changes to methods and content. Sometimes, their diagnoses were better than their
teachers‟.
3.b.2 – Monitoring linguistic production and learning.
Another interesting research question was what correspondence there could be between the
learners' attitude towards monitoring their learning and their language, and the height of their
language ceiling.
Perhaps the most famous theory of monitoring is Krashen & Terrell‟s (Ibid.):
Our fluency in production is thus hypothesized to come from what we have 'picked
up', what we have acquired, in natural communicative situations. Our 'formal
knowledge' of a second language, the rules we learned in class and from texts, is not
responsible for fluency, but only has the function of checking and making repairs on
the output of the acquired system. ( p. 30)
According to this, to use the monitor, learners need time, they have to be thinking about
correctness and have to know the rule. Thus, monitoring seems to be accessory and not central
to learning, and but we have already discussed the importance that further research attached to
consciousness-raising, where monitoring was associated to developing the capabilities for
“noticing” and “appreciating” language, processes which may be considered central and not
peripheral to language learning: a basic learning strategy.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
259
Even early studies of learning strategies acknowledged that “good learners” actively monitor
their language and their progress, use an active task approach, pay attention to form, initiate
communication, seeking to understand and make themselves understood, and used
clarification/verification strategies, as well as comparing L1 and L2 in an analytical fashion.
(Rubin, 1978, 1981; Reiss, 1985; Naiman et al, 1978) They all agreed that self-monitoring was
important, including looking up information in grammar books or dictionaries and paying
attention to corrections. Further studies into the strategies of “good learners” have confirmed
these initial findings, but a serious limitation of these studies is that they are based on
questionnaires asked of students, with direct questions, such as, “Do you ask your teacher for
clarification when you do not understand something?” or on academic success, measured by
test results. It may be that learners give researchers the answers they want to receive, in some
cases, and that the successful learner in the classroom is a failure as a language user in the
outside world.
Our experiment consisted in introducing monitoring possibilities and formal language
analysis for students during lessons, seeking to create awareness, in the presence of an
observer. The learners‟ errors had already been analysed over a period of at least two years, so
there were indicators of the probable height of their language ceilings. Further data came from
informal interviews with the learners, who had been divided into two groups for the purpose of
data analysis: those who had apparently reached their ceiling and those who were still
progressing towards it. At these interviews, we discussed their errors with the learners and also
offered them some advice on how they might improve their performance. Their reactions and
comments were recorded and analysed, with some interesting, though tentative, results:
Learning from Learners‟ Errors Ana María Rozzi de Bergel
260
a) Learners making a majority of intralingual errors due to accessing attempts
affected by overgeneralisation, incomplete application of rules or confusion
were interested in analysing and discussing their production in order to
improve it. Sometimes, they would initiate this discussion. In most cases,
these learners‟ ceiling seemed fairly high. We found these learners among
those who had reached their upper-intermediate level successfully or those
who were still progressing towards the end performance in their curricula.
b) Learners whose intralingual errors were mostly due to simplification looked
slightly bored and shrugged off most of their errors with remarks like:
"Come on, is it really necessary to learn all that?”, "Do English people
actually say this?", or would refuse to incorporate new language or to
discuss their errors, typically saying: "Never mind, people understand me,
anyway." or “Well, I don‟t want to speak like Shakespeare, you know…”
Many of these learners had already reached low ceilings or were not doing
very well.
c) Learners resorting to shortcutting refused to discuss or monitor their
production and sometimes became aggressive and blamed the tutor or the
method for what they considered their failure. It was very difficult to
address the analysis of their errors from a constructive point of view. A
favourite attitude of these learners was: "My problem is I have no time to
study at home", or "The trouble is I did not learn English as a child." Other
excuses included "I need a native speaker teacher", "This method is very
slow. I will try using CD's" or "I will never learn English unless I take a
total immersion course in London". They often refused to assume
Learning from Learners‟ Errors Ana María Rozzi de Bergel
261
responsibility for their progress, but placed it on external factors. These
learners also seemed to be reaching their ceilings at fairly low levels of
attainment.
These interviews shed some light on the learners‟ degree of awareness of their needs and the
importance of listening to them very attentively.
The learners‟ remarks about the course or the curriculum and the degree to which learners
seem willing to monitor their performance and assume responsibility for their progress may
help the teacher to assess the learners‟ closeness to the end performance in their personal
curricula – their language ceiling. The diagnosis is still difficult and perhaps inaccurate, so
great care has to be taken not to jump to conclusions and close observation of many factors is
necessary to detect language ceiling, including an analysis of teaching practices, the teacher‟s
leadership, the materials and the course curriculum.
It is a serious and dangerous error to diagnose ceiling using only one indicator or on data
collected over two or three lessons. In what concerns error analysis, the picture of error which
supposedly describes language ceiling has to remain stable for at least six months, with proper
monitoring and follow-up, before a tentative diagnosis is made.
3.c Horizontal development.
Can teachers help learners raise their language ceiling? It is not very likely, because this task
should be accompanied by changes in the learner‟s cognitive structure – a rare and painstaking
process which might only be undertaken with the learner's full awareness and cooperation, and
with the assistance of a psychologist. It would resemble therapy more than language learning.
In twenty-seven years of research, we cannot report a single case of a learner whose language
Learning from Learners‟ Errors Ana María Rozzi de Bergel
262
ceiling was raised at a normal course of instruction, by whichever method, once we had
diagnosed it by the described means.
A basic consideration about language ceiling is that teachers are more worried about it than
learners, unless there are very heavy pressures on the learner to attain a higher level of
mastery. As suggested by the examples of the learners‟ remarks we have discussed, the adult
who has reached his ceiling will reject being pushed beyond it, probably because he has
already covered his personal curriculum and attained its end performance. He may even regard
attempts at exposing him to more complex language as a threat, as a bombardment of
information he cannot classify, categorise or process and which greatly endangers his present
level of proficiency by introducing doubts and questions into a system he thought already
stable.
The teacher‟s attempts at pushing learners beyond their ceilings have disastrous results: the
systems are practically closed, and the new items the teacher feeds the learner sneak into them
like viruses into a computer, they are not processed by logical means and the learner falls into
what our team termed Total Confusion. This is a state of "unlearning" where we witness the
destructive process we may call Indiscriminate Loss, by which learners forget things in
random hierarchical order and arbitrary degrees of complexity. They may keep a fairly
advanced item, like the conditionals, and drop the past of the verb "to be", or the use of "going
to". They will probably be perfectly able to argue a point at a meeting, but fail to spell their
names on the phone. Learners in Total Confusion do not go from Upper Intermediate level
back to Intermediate, for example, but remain at an undetermined level which has elements of
several levels. The usual type of performance is characterised by the communicative features
and possibilities of their level of attainment, with errors belonging to much lower levels – if
Learning from Learners‟ Errors Ana María Rozzi de Bergel
263
we are to judge it by what is usually taught at conventional levels: elementary, intermediate,
post-intermediate, advanced.
The phenomenon is similar to Selinker‟s (Ibid.) idea of backsliding, by which learners focus
on meaning but produce a previously learnt interlanguage form rather than the desired one.
For learners who are pushed beyond their ceiling, meaning is a crucial asset to keep, whereas
form is underrated, but the loss is indiscriminate because it is not a case of erasure of a whole
level. The state of Total Confusion differs from backsliding in that it is generalised for the
whole of the person‟s language and usually attained after having reached language ceiling,
whereas backsliding is not necessarily related to fossilisation. It is a phenomenon produced by
giving priority to meaning over form.
These learners become remedial cases, which teachers and schools often treat by placing
them two or more levels below their last level of attainment and re-teaching them. This
produces no results, because learners continue performing at their actual level even when
retaking a lower-level course, as their communicative level within their personal curricula
remains stable. This approach to remedial teaching seems to be based on the assumption that
parts of the learners‟ experience can be erased, the clock set back and learners placed in the
same situation as when they were beginners. By the same token, learners who do not do well
at university ought to go back to primary school and re-live their lives. The result is
discouragement and loss of trust in the teacher or the school, as learners find that they are
basically familiar with all the materials in the lower level courses, but they still want to
perform at a higher level. Their questions and doubts are also above course level and cause
disruption during lessons.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
264
The possible solution is a Horizontal Development programme of the following
characteristics:
a) Learners who have reached their ceiling, whether they have fallen into Total
Confusion or not, should not be exposed to new grammar or moved into
more advanced textbooks. This is sometimes regarded as a problem by
teachers who believe that the learner will object to using textbooks which
are always at the same level. Learners in this situation usually feel thankful
not to be promoted to a more advanced course and agree to join a Horizontal
Development programme if this is properly negotiated with them. As usual,
learners know what they need.
b) The course should focus on the topics learners find useful and interesting.
Very often, this amounts to an ESP course. Exposing learners to a variety of
topics, seeking to broaden their vocabulary and probably their cultural base
is not very helpful in these situations. On the contrary; the learner‟s
immediate needs should be addressed, to insure the yield of the language he
has acquired.
c) The grammar syllabus should seek to recover the elements he has lost rather
than incorporate new ones. Formal grammar exercises should not practise
new elements, but recycle those already learnt. However, they should not
take the learner to a lower level. New grammar may be introduced through
reading comprehension, for “passive recognition”, a term which has lost part
of its meaning, as we now know that comprehension is actually a dialogue
between the reader and the writer and not passive at all, for the reader is
actively constructing meanings The general aims of the Horizontal
Learning from Learners‟ Errors Ana María Rozzi de Bergel
265
Development programme are: more precision, accuracy, appropriacy,
fluency in the use of the language already acquired, improved reading
comprehension. Progress should be measured accordingly. We must not
think a learner has raised his language ceiling because he improves in these
areas. It only means that his personal curriculum is better realised, and he
has to continue practising within its end level of attainment.
d) Reading comprehension may be further developed but it should be based on
texts learners are likely to encounter in their professional or personal lives.
Real texts are optimal.
e) The error exploitation method we have outlined should be implemented. It
may be found, then, that errors have not really fossilised but many can
disappear. This will not raise the person‟s ceiling to permit access to more
complex language forms. It will just improve what the learner has acquired.
The characteristics of a horizontal development programme match the learners‟ requests and
comments at the point of language ceiling which we have analysed.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
266
EXPLORATION AND REFLECTION
1. Write down four reasons why you think one of your students has reached his/her
language ceiling. Compare this list to the possible manifestations of language ceiling
discussed in this chapter.
2. Talk to your students about their progress. Do they think they are learning enough?
Record their answers and compare them to the students‟ comments we have analysed
in this chapter. Is there any correspondence?
3. If you have students who may have reached their language ceiling, how are you
teaching them? Compare what you are doing to our recommendations for horizontal
development. Do you think you should introduce changes to your methodology and
approach?
4. How do you give feedback to students who may be at their ceiling?
5. Is any of your students asking for methods or procedures which might be part of a
horizontal development programme, e.g. “no grammar” or “more reading”?
Learning from Learners‟ Errors Ana María Rozzi de Bergel
267
4. The correspondence between errors and learning hypotheses
Error analysis, carried out for the purpose of disclosing learning operations and the learner‟s
style for structuring the language system, inevitably leads to an exploration of the possible
correspondence between learning hypotheses and error profiles. It remains an interesting
experiment to diagnose learning hypotheses through error analysis, but it is even more
interesting and perhaps more useful to attempt a prognosis of kinds of errors for a particular
learner through a diagnosis of his learning hypotheses. Were this prognosis possible, it might
lead us to reflect upon more of our “cherished notions” about materials, approaches, methods
and even learning and teaching styles.
4.a Errors and learning hypotheses: finding the match
When discussing learning hypotheses, I explained that our exploration had been centred on
three basic aspects; namely:
a) The learner‟s personal way of placing concepts into categories, using
conceptual or pre-conceptual systems of classification.
b) The learner‟s association style, by proximity or by analogy (Using
metonymy or metaphor)
c) The learner‟s degree of imagination and creativity, placed at the service of
structuring the language system.
These features of learning hypotheses would define the type of learning style, which in turn
would determine the observable learning strategies used by learners in the classroom. I often
referred to these concepts when analysing errors and attempted to account for the mechanisms
Learning from Learners‟ Errors Ana María Rozzi de Bergel
268
in the learner‟s cognitive structure which seemed to have generated them or which the errors
seemed to disclose.
When talking about language ceiling, I have repeatedly explained that it may or may not be
associated with these learning hypotheses in the learner‟s cognitive structure and that the
factors outside this domain which determine the height of language ceiling could well be the
topic of research for a multi-disciplinary team – which was not our case.
Both in the discussion of errors and in the analysis of the phenomenon of language ceiling,
there was an obvious connection between error analysis and the learning hypotheses listed
above, which we diagnosed by using the Learning Hypotheses Test you find in Appendix 2.
This was a two-way connection: through error analysis, we gained insight into learning
hypotheses and through the diagnosis of learning hypotheses, we gained insight into the causes
of errors.
Which process came first? When we started analysing errors, we began to develop
awareness that their causes and significance needed further research. We also regretted that to
know whether the errors a student was making were productive or unproductive, and whether
teaching and materials seemed appropriate for his learning structure, we had to wait until that
person had made a reasonable number of errors and had been attending a course for at least a
year. Sometimes, certain discoveries came too late, when the learner had already failed or
complained about the syllabus, the materials or the teacher. A question we often asked
ourselves, was, Can we know in advance what type of errors a learner is likely to make? We
felt that the answer to that question would enable us to further develop a learner-centred and
more effective approach to teaching, by helping us to prevent mismatches between learners‟
learning hypotheses, teaching and materials.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
269
We began diagnosing learning hypotheses and then correlating the results with error
analyses of the learners who had taken the Learning Strategies Test. The results were
interesting though not highly revealing because, obviously, the causes of errors are not
determined exclusively by the learners‟ learning hypotheses. Even considering these
limitations, the study yielded some tentative conclusions, which should be read more as
theoretical inferences than as findings. We will later on discuss how the knowledge gained
through these experiments may be put at the service of learner-centred approaches.
4.b Error prognosis
The word is odd, when used to refer to possible ways of predicting the kind of errors a
learner is more likely to make. At least, we can be sure it has not been used very frequently, if
at all, simply because it is not a popular line of research and because it seems extremely bold
and even dangerous to attempt such predictions, considering the changing nature of human
beings and their potential to adapt to circumstances or even act upon circumstances. We may
suspect that any such prognosis would most probably be proven wrong, or that it might
succeed in leading the teacher to create the necessary conditions for the learner to develop
along the lines in the prognosis, so that it can be proven right. In this section, I will attempt a
discussion of whether error prognosis is possible and desirable.
Is error prognosis possible? Not completely, but a learner‟s hypotheses may lead him more
specifically into certain error types and styles. We will outline these probabilities with
reference to the three domains of learning hypotheses described above:
a) Conceptual/pre-conceptual styles for placing concepts into categories: when learners
use predominantly conceptual, sufficiently restricted and sufficiently comprehensive
categories, the picture of error may resemble that of the learners in Figure 10.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
270
Both the use of predominantly insufficiently comprehensive or insufficiently
restricted conceptual categories may lead learners into excessive simplification of the
language system. When they cannot very well see the limits of certain categories, they
prefer to reduce the number of elements which might fit into them. When they see the
conceptual categories as too limited, on the other hand, they tend to exclude the
elements which do not fit this narrow category. As these elements are often not
included in any other category, they are just erased or discarded.
The presence of too many self-originated categories may cause the learner to come to
hectic conclusions about the language system and its use and often results into very
abundant compensatory production of the type which tends to exasperate teachers, who
often do not know what to do with these learners who seem to be always trying new
ways of expression instead of concentrating on what they are being exposed to and
coming to the wrong conclusions about everything.
The use of predominantly pre-conceptual categories usually causes a large number of
errors due to cue-copying and adherence to first form or meaning learnt. Elements are
included into pre-conceptual categories figuratively (With reference to the group where
the element belongs), attributively (By its attributes or qualities), anecdotally or by
their use. Therefore, errors occur by proximity, with learners clinging to the first form
or meaning learnt because they adhere to the one-form-one-meaning correspondence,
or they occur by cue-copying, because this is the closest figurative association they can
make in order to remember and use an item which is placed in a category we might call
“being there”. These learners usually make more interlingual errors than those who
have access to conceptual categories, as the association by proximity is often to the
Learning from Learners‟ Errors Ana María Rozzi de Bergel
271
corresponding word or form in the native language. When this correspondence does
not exist, the learner may strive to learn the item within the system of English, discard
it or create a personal translation or equivalent form.
b) Association style: the effects of one association style or the other have to be appraised
together with the degree to which learners use conceptual categories.
Learners who associate mostly or exclusively by proximity and do not use conceptual
categories do not make errors due to overgeneralisation, but rather simplify, cue-copy
and try to construct the system of English on the blueprint of their native language.
Those who associate mostly through metonymy but have recourse to conceptual
categories incur overgeneralisation more often than the other types of intralingual
errors and keep their interlingual errors within reasonable limits.
Learners who associate mostly by analogy, metaphorically, and who make use of
conceptual categories, transfer positively from L1 into L2 and can transfer concepts to
various contexts. Their errors are mostly due to overgeneralisation or confusion and
not significantly to interlingual problems.
Learners who associate mostly metaphorically but have little access to conceptual
categorisation, may make incorrect inferences, come to mistaken conclusions, produce
a great number of errors due to interlingual problems and display large compensatory
production: their creativity prevails over their conceptualisation and makes their minds
fly into domains the teacher not always accesses. They require guidance and some
limits.
c) Imagination and creativity: as with the association style, these features of the learner‟s
learning hypotheses have to be correlated with his use of conceptual categories.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
272
A highly imaginative and creative learner who does not use conceptual categorisation
may use cue-copying, simplification, adherence to first form or meaning learnt and
compensatory production. Again, imagination will prevail over logic. On the other
hand, if high creativity and imagination are coupled with access to conceptualisation,
the learner will identify rule restrictions, transfer items from one context to others, find
differences and similarities and generally use his imagination and creativity to further
advance into the construction of the language system without too much guidance.
Errors may be mostly due to overgeneralisations or confusion and some interlingual
problems will be present, but not in large numbers.
The previous analysis suggests two clear general tendencies in learners, which should also
be taken into consideration when designing courses and choosing materials: the need for either
well-structured or loosely structured teaching-learning situations. Gauging the correct degree
of guidance needed by a learner or a group of learners appears as a key ability in teaching. Too
much guidance will suffocate imaginative, conceptual learners and open-endedness will
discourage and frustrate unimaginative, pre-conceptual learners.
To a point, this concern for degrees of guidance appears in Scarcella & Oxford (Ibid.) when
they discuss “orientation to closure” as a learning style:
“Students oriented toward closure have a strong need for clarity in all aspects of
language learning. They want lesson directions and grammar rules to be spelled out
and are unable to cope with much slack in the system. Not for such students are
spontaneous, rollicking conversations and games in the language classroom, un less,
of course, they have had adequate time to prepare their vocabulary lists and
understand the rules involved in any given interaction.” (p.62)
Learning from Learners‟ Errors Ana María Rozzi de Bergel
273
They go on to cite a study by Ehrman and Oxford (1989) in which they found that
“Sometimes their desire for closure and control can short-circuit their ability to
participate in open-ended communication necessary for developing fluency.”(p.62)
According to the same authors, “open learners” care more for communication and “learn by
osmosis”, in a highly relaxed fashion, but do not function well in “highly structured and
traditional classroom settings”. The closure Scarcella & Oxford refer to may well be similar
to the early closure of categories I explained when describing our criteria for analysing errors
and it might go hand in hand with a trend towards simplification and cue-copying.
There are two important considerations to make at this point. The reader must have noticed
that I have not included error rating or error types in these descriptions of tentative prognoses.
The reason for this is that, although it is not very difficult to infer that the closer the
correspondence between conceptual categories-high imagination and creativity, the more
appropriate a learner‟s production will be, or that this correspondence will cause more
syntactic than semantic errors, as learners seem to use the language more appropriately, this is
not necessarily the case. The problem here has other implications, mostly, personality traits
and cultural background. A learner may resort to conceptual categories to structure the
language being learnt and aid the process with creative and imaginative efforts, and yet fail to
use English appropriately because he does not understand the culture associated to the
language. Learners may even understand but reject the culture or the ideology they associate
with the foreign language and not care very much about using it appropriately in what
concerns social customs, degrees of formality or even respect. They may refuse to be
acculturated. Some learners‟ errors in appropriacy stem from their own upbringing and will
pass on to any language. A person who stated, “I never say thank you to porters and waiters”,
Learning from Learners‟ Errors Ana María Rozzi de Bergel
274
would not learn how to thank, even if his conceptual categorisation capabilities enabled him to
do so.
In my experience, the cases I have just discussed are rare, as most individuals who have
developed conceptual thinking and display a solid degree of imagination and creativity usually
come from rich educational backgrounds, are inquisitive, seem eager to learn and display not
only tolerance of other cultures but also interest in learning more about them. The point I am
making is that the diagnosis of learning hypotheses cannot shed light on these cultural, social
and educational features of personality which will then influence a learner‟s language use,
therefore, I prefer not to include error rating and error types in the description of the general
prognosis for a particular learner.
The other important consideration I would like to make is that we should not fall into the
temptation of considering the picture of error in Figure 10 as depicting “normal values” versus
which we might evaluate other pictures of error. Those values simply show the error rating,
types and classes for learners at a particular point in time, at a particular place and under
certain circumstances and are not in any way generalisable. Within our research, they showed
the picture of error of apparently successful learners, who had passed our written and oral
tests. We might have used these results as parameters, and in fact we often did, but there is no
evidence that they reflect a “normal” picture of error for other adult learners in different
situations or cultures.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
275
4.c Error prognosis and teaching
We should now consider the desirability, even the advisability of attempting a prognosis of a
learner‟s errors. What would be the use of making an error prognosis for a particular learner or
group of learners? How could the results translate into concrete pedagogic applications?
The objective should not be to prevent errors but to take steps to avoid error-provoking
teaching or resorting to practices which may reinforce unproductive error trends.
Knowing what type of errors learners may make enables teachers to:
Verify if learners who are very imaginative and creative but do not categorise
conceptually have come to the right conclusions about the material being learnt,
without taking this for granted.
Avoid proposing too creative and imaginative activities to learners who are not capable
of performing them or to those who will carry them out beautifully, but with sizeable
compensatory production and no regard for accuracy.
Lead learners in their process of transferring items from the context where they first
found them, to other contexts, when they cannot do this spontaneously.
Provide proper guidance and fairly structured lessons to groups of learners who will
not tend to categorise conceptually and open-ended, fairly unstructured lessons to
groups of learners who do not need so much guidance.
Select clearly structured or loosely structured materials, according to need, avoiding
the use of the same materials for all groups.
Avoid error-provoking procedures, such as teaching similar items together to groups of
learners who associate mostly by proximity.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
276
Language schools might use error prognoses to select materials and train teachers to tailor
their methods and procedures to the learners‟ personal curriculum.
Although much has been said and written about learning styles and strategies, about focusing
on the learner and not on methods, error prognosis has not received much attention, or rather,
it has been ignored as an option for tailoring courses to the learners‟ real needs. The root of the
problem is that the study of errors has been neglected, so teachers would have very little use
for a prognosis they might be relatively unable to understand, let alone use as the basis for
their teaching practice. As for detecting learning hypotheses, teaching would greatly profit
from input from cognitive psychologists who devised instruments like our Learning
Hypotheses Test or which served a similar purpose.
4.d Ceiling prognosis
If error prognosis is possible, we might ask, Is ceiling prognosis possible, too? When the
question is put to me, I always remember the old joke, I don‟t believe in witches, but they are
there. A responsible researcher should answer there is no proof or evidence that the height of a
person‟s language ceiling can be predicted, but it would be irresponsible not to claim it might
be. It would be irresponsible on my part, because although no formal research was carried out
by my team, throughout these almost thirty years we have seen many cases of learners with
low language ceilings who seemed to share certain learning hypotheses. This remains one of
the most interesting and important research questions deriving from our work: is there a
profile of learning hypotheses which corresponds to a low language ceiling in the foreign
language?
Learners whose structuring of L1 is predominantly conceptual, and who have a reasonable
degree of imagination and creativity seem to have fairly high ceilings or the potential for high
Learning from Learners‟ Errors Ana María Rozzi de Bergel
277
ceilings. It should be remembered once again that the height of language ceiling is not
determined exclusively by learning hypotheses.
We may suspect that learners whose structuring of L1 is predominantly pre-conceptual and
who have an average degree of imagination and creativity may have a potentially high
language ceiling if they are taught with clear and simple materials and receive proper guidance
from the teacher.
Learners whose structuring of L1 is predominantly pre-conceptual, and who have very little
imagination or creativity, show trends towards cue-copying, shortcutting, simplification and
perhaps interference from L1. We may assume that their language ceilings may be relatively
low, but of course, this will depend not only on their motivation and endeavour, but on the
kind of teaching and materials they are exposed to.
Learners whose structuring of L1 is predominantly pre-conceptual, and who have a very
high degree of imagination and creativity, show a trend towards compensation, simplification
and interference form L1. Their conclusions about L2 may be hap-hazard or hectic. We may
suspect that their language ceilings will be low, maybe reached at the end of the pre-
intermediate level or before.
4.e The desirability of prognoses
The temptation to forecast success or failure at learning a foreign language has always been
the researcher‟s trap, even in cases such as Carroll & Sapon‟s (2002), with their much-
publicised and much-criticised Modern Language Aptitude Test. The use given to that test, for
example, to decide who, among a company‟s staff, might deserve to be included in a foreign
language learning programme, reeks of discrimination and ignorance of other factors which
greatly influence learning, besides being perverse. The test was also heavily biased towards
Learning from Learners‟ Errors Ana María Rozzi de Bergel
278
behaviourism and tends to find out if people can learn according to the methods prescribed by
the behaviourists. It seems to have been, however, highly successful.
The criticism levelled at the MLAT in the sense that it did not take motivation into
consideration prompted the development of the Pimsleur Language Aptitude Battery (1966)
which tested motivation, verbal intelligence and auditory ability. Again, the exploration was
incomplete, as it did not take such factors as influence of the social medium or cultural
implications into account. Again, it proved highly successful but was used for what we might
even term unethical purposes, such as denying an employee the benefit of a language
programme because of a low result in the PLAB.
Modern researchers are more careful and discourage their students and readers from making
success or failure prognoses, despite pointing out which learning styles or strategies seem to
be more favourable for language learning. We have now gained more insight into the
unpredictability and flexibility of human beings and their potential for turning their
weaknesses into strengths. We now more about the power of social influences, the drive
people may find in self-determination or the value of alternative methods and learning
strategies, as well as having learnt more about the brain. Despite all this wisdom, prognoses of
the type we are discussing are still a valid and largely unexplored field of research.
Chapelle, in Reid (1995), describes the learning strategies of field dependent and field
independent learners and explains that research tends to show, though not conclusively, an
advantage for learners who have abilities associated with field independence; that is, having
analytical possibilities, being autonomous, not relying on guidance or prescribed methods,
setting self-imposed goals, being task oriented, all characteristics which obviously need to be
supported by conceptual capabilities and a reasonable degree of creativity and imagination.
Despite these claims, Chapelle clearly states that studies of field dependent or field
Learning from Learners‟ Errors Ana María Rozzi de Bergel
279
independent learning styles should not be considered predictive…“…. but they are there”. Are
researchers afraid of making statements which are not “politically correct” in academia or is it
true that there are no reliable predictions in this matter?
I believe the latter to be the case and that academia has grown wiser. Error and ceiling
prognoses seem to be potentially dangerous weapons rather than potentially useful resources,
at first sight, and this probably accounts for the fact that those of us who have taken a glimpse
at the problem do not dare make any claims about our observations. There is great danger that
these prognoses might be wrongly used, particularly, when they are part of selection
procedures for staff recruitment or admission at educational establishments.
Another possible danger is biasing the teacher, labelling the student and thus conditioning
the student's progress to the teacher‟s or the researcher‟s expectations. The famous experiment
by Rosenthal, who prompted different probable results to two groups of researchers
performing the same experiment, and got exactly the results he had suggested, demonstrated
that the researchers‟ expectations had conditioned the results of experiments so that they
would be proven right. Rosenthal & Jacobson (1992) took their research to the classroom and
showed that students often perform to the measure of their teachers‟ expectations. This has
also been called the Pygmalion effect (after the play by George Bernard Shaw, who took it
from the ancient Greek myth) or, more explicitly, the teacher-expectancy effect. Its threat
looms in the horizon of prognoses of all types: could they possibly become self-fulfilling
prophecies?
According to Merton (1948) a self-fulfilling prophecy begins with a person‟s or a society‟s
false belief about a future development or event, which then changes behaviour in such a way
as to finally produce the predicted event or behaviour. The initial belief conditions the result of
the process. Similar situations are acknowledged by various schools of psychology, which
Learning from Learners‟ Errors Ana María Rozzi de Bergel
280
claim that parental expectations, for example, result in their children‟s accomplishments or
failures. Neuro-linguistic programming stresses the importance of positive feedback, reduction
of self-criticism, elimination of negative comments about oneself as a way of fostering self-
confidence. Here, self-fulfilling prophecies are not even treated as influences from others or
from the medium, but created by ourselves to condition our own lives.
Because of all these considerations about the risk of hypothesising about the kind or errors a
learner was likely to make, and even worse, about the probable height of his ceiling, when we
moved into the slippery area of error and ceiling prognosis we were faced with a dilemma:
whether to share the results with the teacher or to keep them to ourselves and just tell the
teacher what materials and method to use (without giving reasons for our advice and choices),
and monitor the courses. We opted for the latter procedure, but events proved Rosenthal
wrong, as our approach had negative effects:
a) Teachers did not fully understand the reasons for the choice of materials, methods
and procedures they were instructed to use with certain learners, and used them half-
heartedly. They were simply obeying orders, a staff management policy which is the
passport to failure.
b) Teachers saw that some learners were not making much progress, or that they
consistently tended to make the same type of errors, and became very anguished
because they could not account for these problems. Some teachers even developed a
sense of guilt.
c) Some learners who had reached or were about to reach their ceiling, or who disliked
a particular type of activity, asked their teachers for changes in the syllabus or the
method and got inadequate responses: teachers either agreed to whatever they were
asked to do or tried to "convince" the learners that no changes were needed.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
281
d) Some teachers chose to disobey our instructions and failed. They came under heavy
criticism from their students, and some had to be removed from their courses. This
was sometimes the case with teachers who did not “like” some highly structured
books they were told to use and decided to supplement them heavily; however, the
books had been chosen on the basis of the diagnosis of learning hypotheses and
students rejected the teacher-supplied extra materials and the teacher‟s policy.
e) Some teachers of students with a low language ceiling followed our instructions to
the letter with excellent results and took their beginner students well into the
Intermediate level, only to find that this was their point of language ceiling. The
question these puzzled teachers asked, was: "What happened to them? They were
progressing so steadily .....". When we told them we had always known this would
happen, they felt terribly cheated.
It is not very sensible or respectful for the management of a language school to withhold
information from the tutors not to bias them in their appraisal of results. Teachers are
professionals, and they should have access to all the data about their learners. Also, this bias
is very difficult to produce, as further experiments demonstrated: we gave some teachers the
wrong information about their learners, and after some time they came to tell us we were
wrong. Telling a teacher that her students are geniuses does not turn them into geniuses.
Considering the assertions that have been made to the contrary, one wonders under what
circumstances the experiments were carried out, and what variables were not controlled.
Perhaps the Rosenthal, Pygmalion and teacher-expectancy effect did not work in our case
because we were dealing with adults, who are not so likely to be affected by their teacher‟s
expectations. Perhaps the Rosenthal experiment itself was a self-fulfilling prophecy. Be it as it
Learning from Learners‟ Errors Ana María Rozzi de Bergel
282
may, the experience taught us to tailor research to the context where it is carried out and
pointed out the dangers of generalising research results beyond the boundaries of that context.
So we began sharing tentative prognoses with teachers, together with the pertinent
information about all the other aspects of the personal curriculum we had researched:
communication strategies, needs, interests, etc.
We may claim that teachers with a very strong tendency towards overprotection or over-
criticism can turn adult students into slow learners, because a totalitarian-overprotective or
totalitarian-despotic leader needs victims to save or slaves to command. It may also be true
that a democratic teacher can get the best out of groups of relatively average learners. We
should remember that roles are always complementary: there is no teacher without a pupil, no
master without a servant (Pichon Riviére, Ibid.). The teacher's leadership has an
unquestionable influence on the outcome of a language course, but to assume that
communicating error and ceiling prognosis to the teacher will bias the results of a course is a
naive idea. Democratic, reliable teachers should be given all the information they need. The
key to this problem seems to lie with staff selection, support and training, rather than with
disclosing or withholding information.
Another question that arose with respect to the advisability of reporting diagnoses and
prognoses to the teachers referred to how much responsibility the tutor would be willing to
assume for students who had a low language ceiling. Wouldn't this teacher just give up the
fight from the start? Wouldn't she think : "Why worry? The Head already knows these people
are not going to learn very much"? Would she wonder why she had been chosen to teach
those learners? Again, we found that the problem was personal, not institutional. Responsible
tutors always work well and take pride in their achievements. Many tutors faced classes of
Learning from Learners‟ Errors Ana María Rozzi de Bergel
283
potential low-ceiling learners as a challenge, and in fact managed to raise their ceiling to the
highest possible point for those learners.
When our tutors started getting their students‟ prognoses, the results were:
a) Teachers understood the rationale of the methods and procedures they were
instructed to use with certain learners, and used them confidently. Their relationship
with the management improved, because they felt taken into consideration. They not
only obeyed instructions, but contributed useful suggestions as well.
b) Teachers who saw that some learners were not making much progress, or that they
consistently tended to make the same type of errors were fully prepared to deal with
these problems and did not panic when they arose.
c) Some learners who had reached or were about to reach their ceiling, or who disliked
a particular type of activity, asked their teachers for changes in the syllabus or the
method and got highly encouraging responses, because all teachers were prepared to
negotiate the contents of the syllabus and the activities in the course.
d) Some teachers chose to disobey our instructions and failed. They immediately
backtracked on what they were doing and came to us for help, admitting the initial
advice had been correct.
e) Some teachers of students with a low language ceiling followed our instructions to
the letter with excellent results and enthusiastically took their students well into the
Intermediate level, only to find that this was, after all, their point of language ceiling.
They were a little disappointed, but they realised our prognosis had been correct, and
accepted the facts.
f) Group dynamics improved considerably. Cases of students who complained against
the method, their classmates, the book or the teacher became very rare. We had no
Learning from Learners‟ Errors Ana María Rozzi de Bergel
284
obnoxious learners, parallel leaders or "difficult" groups. Tension and insecurity
seemed to have disappeared from most classrooms.
We learnt two invaluable lessons in staff management: hire first-class professionals, and then
trust them for using information wisely; create a non-threatening atmosphere of trust, a
community of learning in your school and you will serve your students better.
Let us not forget, however, that a person‟s performance as a foreign language learner
depends on so many factors that error and ceiling prognoses should be taken just as useful
information and not as language aptitude indicators or as predictions of failure or success.
Teachers can use them as tools to better adapt their practice to their learners‟ needs, but they
should approach each class with an open mind and great faith in their learners‟ potential. In
any case, prognoses should be supplemented with a study of the other elements of the learner‟s
personal curriculum which we outlined in Figure 5.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
285
EXPLORATION AND REFLECTION
1. Think about two cases of students who have surprised you by performing above or
below your expectations for them. Why did you expect them to behave in a different
way?
2. Talk to one of your students‟ previous teacher or look at the school records. In what
way does your evaluation of these learners differ from the previous teacher‟s opinion?
3. Think of yourself as a student. Did you perform differently according to what each
teacher expected of you?
4. When you face a new class, do you rely on the previous teachers‟ evaluations of those
learners or examine their records? How useful is this?
5. Do you often find that learners‟ self-appraisal is more optimistic/pessimistic than
yours? What considerations does this prompt you?
6. What would you do if somebody told you that a group of learners you are going to
teach is “absolutely hopeless”?
7. What would you do if the school told you that the learners who obtain low marks in
your class are excellent in all the other subjects?
8. Follow up a learner you consider “very good/poor” at first sight. After three or four
months, evaluate the accuracy of this first impression.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
286
Summing up
I learnt four fundamental lessons from this long process of study and research.
The exploration of learners‟ interlanguage and the study of errors took me into worlds I had
never meant to explore. It seemed, at many points, that events led me astray or made me divert
from the main focus of my study and venture into apparently unrelated areas such as the
teacher‟s leadership, learning hypotheses or materials design. After some time, I decided not
to question these digressions and to undertake them as an integral part of the study of errors,
their origin, role and significance, because the first lesson I learnt was that the process of
learning-teaching is also a system and as such, all its components are interrelated and each
component acquires significance because it belongs to the system.
Due to this systemic characteristic of the teaching-learning process, it was impossible to
study errors in isolation from the factors influencing the learners‟ interlanguage, although
some of them would never be explored, let alone disclosed. I was able to focus on errors, but I
discovered they were only a tiny part of a sophisticated whole I would never get to know in
depth.
The second lesson was that the learner‟s curriculum would always prevail and that teachers
do not succeed in imposing methods and contents on adult learners. Success seems better
achieved when teaching is tailored to learners‟ needs, preferences and cognitive structure,
even to their limitations. The concepts of gradation of difficulty and order of presentation, the
methodological prescriptions of theoreticians, even the linguists‟ conceptions of language and
the psychologists‟ ideas on learning are some of the “cherished notions” Pit Corder invited us
to re-consider. We can do that if we accept that learners know just as much as we do about
Learning from Learners‟ Errors Ana María Rozzi de Bergel
287
when to learn, what to learn and how to learn it – they just need professional help from their
teachers to realise their personal curricula.
The third and perhaps the most dramatic lesson learnt from this process was that teachers of
children are largely responsible for creating good language learners, by paying attention,
among other things, to four central aspects of a person‟s education: conceptual thinking,
imagination, creativity and a broad cultural base, coupled with an excellent command of the
native language.
Most of the adult learners I met during my exploration who seemed to have a very low
language ceiling were lacking in one of those aspects and were relatively unable to make up
for that lack or even interested in doing so.
In what concerns education management, I learnt the enormous value of a school which may
be a true community of learning and to achieve this through professional advice, research,
democratic leadership and an orientation towards continuous improvement, involving all the
staff in a process of successive action-research cycles. Our study was possible because we
created these conditions in the language school where we worked.
As regards the future, further research could be related to learning more about the
characteristics and components of the learner‟s personal curriculum, knowing that along the
way, the most interesting part of our exploration will be the new questions which may arise,
not the answers we may find.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
288
APPENDIX 1
Protocol: Needs and expectations interview.
Part 1 – Personal information.
Ask questions in English, if possible, to obtain the following information:
Name:
Telephone: e-mail:
Age: Occupation: Seniority:
Education:
1a) Primary:
2a) Secondary:
3a) University: (Graduate and post-graduate)
Previous English courses – total number of months, years, etc.: Interrupted/consecutive
Part 2 – Needs and expectations
Ask the questions in English, if possible.
1) Why do you need to learn English? (The question may be rephrased to “Why do you
wish to learn English?”)
2) Would you like to use English for other purposes? Which?
3) What is your evaluation of the courses you have taken? What did you like and dislike?
4) How long do you think it can take an adult to achieve perfect command of a foreign
language? (NB: The expression “perfect command” has been chosen to assess the
learner‟s reaction to this concept – does he think this is possible?)
5) How long do you think it could take you to learn as much as you wish to learn?
6) If you placed your desired command of the language on a scale from 0 to 10, how
much would you say you have already achieved and how much do you still have to
learn?
7) What do you think is more important, and why: a good textbook or a good teacher?
Learning from Learners‟ Errors Ana María Rozzi de Bergel
289
8) Do you think you would feel more comfortable with a male or a female teacher?
9) What professional and personal characteristics would you expect to find in a teacher?
10) What kind of materials would you like to work with?
11) Would you like to suggest or contribute materials? Which?
12) Is there any cultural, artistic or historical aspect of an English-speaking country you
might be particularly interested in, or that you know very well?
13) Are there any characteristics of English-speaking countries or people that you find
either particularly attractive or particularly disagreeable?
14) What do you do in your free time?
Part 3 – Interviewer‟s evaluation
Evaluate the answers according to these criteria:
a. Is the learner driven by a personal interest in the language or by concrete
needs?
b. Is the learner realistic in I) his assessment of his level; II) his training gap?
c. Is the learner realistic in his assessment of learning processes and the time
needed to achieve results?
d. How does he see the teacher-student relationship?
e. How does he envisage his role in the course?
f. What is the learner‟s attitude towards English-speaking cultures?
g. Does the learner have a broad range of interests?
NB: This interview should be supplemented with the results of the Placement Test, the
diagnosis of the Learning Strategies Test and the transcript or recording of the Communication
Strategies Interview.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
290
APPENDIX 2
THE LEARNING HYPOTHESES TEST
It tests the main abilities a person needs for developing systematic patterns of system-
formation:
1. The ability to give definitions, and how he does this. Does the learner have access to
concepts or does he use anecdotal, utilitarian, etc., pre-conceptual classificatory
systems?
2. The ability to associate, either by creating a closed-in system or by opening up a
number of related or unrelated sub-systems.
3. The ability to relate systems, by proximity (metonymy) or by analogy (metaphor).
4. The ability to use his creativity and imagination to advance into the structuring of the
system.
The test is for adult learners of English as a foreign language and must be undertaken in the
learners‟ native language. The text herein is in Spanish because the experiment was carried out
in Argentina, and we warn our colleagues against simply translating it to suit the needs of
different nationalities.
Versions into other languages must be produced by specialised psychologists.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
291
A - Text of the test.
The following is the original test in Spanish. The translation into English is not the
recommended version in this language. It has been included here only for the sake of
enabling English-speaking readers to become acquainted with the text we used.
EVALUACION DE HIPOTESIS DE APRENDIZAJE (Learning Hypotheses Test)
1. Defina por escrito las siguientes palabras:
manzana - almohadón - arsenal - borde - estrofa - inminente – retroactivo - aflicción
Explique el significado de cada término. No dé sinónimos.
2. Escriba, o bien un párrafo, o bien tantas oraciones independientes como desee,
con la condición de que se incluyan las palabras siguientes:
vestido - pasto - charco - sobre - habilidad - enojo.
3. Escriba una palabra relacionada con cada una de las siguientes:
raíz - baldosa - casa - maestro - templo - noche - ventana – capital
4. Diga en qué se parecen:
naranja y banana - vagón y bicicleta - periódico y radio - aire y agua - madera y
alcohol - ojo y oído - poema y estatua - elogio y castigo.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
292
(The Learning Hypotheses Test:
1. Define the following words in writing: apple - cushion - arsenal - edge - stanza -
imminent - retroactive - grief.
Explain the meaning of each word. Do not give synonyms.
2. Write either one paragraph or as many isolated sentences as you wish, provided you
use the following words: dress - grass - puddle - on - ability - anger.
3. For each of the following words, write a word that is related to it: root - floor tile -
house - teacher - temple - night - window - capital.
4. State the likeness between: orange and banana - wagon and bicycle - newspaper and
radio - air and water - wood and alcohol - eye and ear - poem and statue.)
NB: The word "borde" has a more comprehensive meaning in Spanish than "edge" in
English.
The word "almohadón" is liable to be defined as "almohada grande" ("Almohada
grande" = "A large pillow") by the student using pre-conceptual categories, because of the
–ón augmentative ending in Spanish. This does not happen with the word "cushion". The
word "sobre" means "on", "over", "about", etc. but also "envelope" in Spanish.
These are some of the reasons why this English language version should be regarded as a
mere translation and never be used as a test.)
Learning from Learners‟ Errors Ana María Rozzi de Bergel
293
B - Description of the testing items.
Exercises 1 and 4 measure the degree of conceptual and logical hierarchical
organisation of the answers. These are the key exercises in this test, for they reveal the
testee's personal style for language structuring.
Testees' renderings are rated as:
- Fully conceptual: The placement of a word within the language is carried out through
the use of a logical-hierarchical system of categories. The "fully conceptual" placement
makes use of categories that are
necessary (not eventual or temporary characteristics)
objective (as independent as possible from the speaker's personal points of view or
experience)
adequately restricted in their scope, so that the word defined will fall within the
category used and it will not allow the inclusion of a number of other elements
within the same definition.
Examples:
arsenal: establecimiento para almacenar y fabricar armas de guerra, conjunto de estas
armas.
ojo y oído: órganos de los sentidos.
poema y estatua: obras de arte
(arsenal: an establishment for the storage and manufacture of weapons; collection of
weapons.
eye and ear: sense organs
poem and statue: works of art)
Learning from Learners‟ Errors Ana María Rozzi de Bergel
294
- Conceptual, insufficiently restricted: The placement of a word within categories that
DO allow the inclusion of a lot of elements pertaining to subcategories thereof, within
the same definition.
Examples:
manzana: es un fruto.
ojo y oído: son órganos
(apple: it is a fruit
eye and ear: they are organs)
- Conceptual, insufficiently comprehensive: The placement of a word through the use
of categories that do not permit the word defined to fall completely within the
definition.
Examples:
almohadón: objeto de cuero relleno de material mullido.
arsenal: conjunto de pistolas
(cushion: a padded, leather object
arsenal: a collection of pistols)
- Vague and approximate: The testee is resorting to a concept, but it is so diffuse and
imprecise that it cannot be considered a conceptually valid response to the stimulus
provided. Rather, it allows the inclusion of an indefinite number of heterogenous
elements within the same definition.
Examples:
retroactivo: referido al pasado.
aire y agua: elementos físicos.
ojo y oído: órganos vivos
Learning from Learners‟ Errors Ana María Rozzi de Bergel
295
aflicción: estado de ánimo.
(retroactive: referring to the past
air and water: physical elements
eye and ear: living organs
grief: a state of aim)
- Pre-conceptual. The placement of a word within the language is carried out without
recourse to a logical-hierarchical system of categories, but through the mention of a
particular use or function of the word defined (Utilitarian); and action it can perform or
soon can perform with it (Anecdotal); a graphic or emblematic representation of its
meaning (Figurative); an attribute or quality (Attributive); or just the testee's personal
and/or imaginary creation concerning the meaning of the word defined (Self-originated)
Examples:
Utilitarian: almohadón: objeto para decoración y confort.
elogio y castigo: se usan para premiar a las personas.
(cushion: an object for decoration and comfort.
praise and punishment: they are used for rewarding people)
Anecdotal: retroactivo: ir hacia atrás en el tiempo.
aflicción: sentirse triste y dolorido.
(retroactive: to go back in time
grief: to feel sad and pained)
Figurative: ojo y oído: ambos están en la cabeza.
almohadón: almohada grande.
(eye and ear: they are both in the head.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
296
almohadón: big pillow - the suffix "on" denotes larger size, in
other cases)
Attributive: arsenal: lugar grande y peligroso.
vagón y bicicleta: ambos tienen ruedas.
(arsenal: big, dangerous place.
wagon and bicycle: they both have wheels)
Self-originated: retroactivo: algo que produce ciertos efectos en el pasado.
poema y estatua: son poesía
(retroactive: something that produces certain effects in the past.
poem and statue: they are poetry.)
Synonym = no answer
Exercise 1 measures the testee's ability to use logical-hierarchical categories in formulating
statements within a productive language task. The testee has at his disposal (at least,
theoretically speaking) the whole edifice of his native language for performing in this task. His
logical structuring of language will prompt him to choose certain elements of L1 for
formulating his answers. This choice will be effected within different logical-hierarchical
categories.
Exercise 4 has the same objective, but the use of logical categories is measured with respect
to recognition rather than production. The testee does not have an ample range of possibilities
as in Exercise 1, but is made to perform within a restricted, close-ended context.
Recognition and production are based on the same ability to formulate statements within
logical-hierarchical categories, but they are clearly differentiated as two particular skills by the
course designer and the language teacher. Widdowson calls them "Interpreting" and
"Emitting", other authors separate Reading and Listening Comprehension from Speaking and
Learning from Learners‟ Errors Ana María Rozzi de Bergel
297
Writing as Receptive and Productive Skills respectively. Even integrated approaches make this
difference. Because of this, this test measures the same logical-linguistic ability from two
different angles.
Exercise 2 measures the testee's ability to compose, from isolated stimuli, a meaningful
peace of written discourse.
The accepted forms will be: one sentence for each word, several sentences containing one or
more of the given words each, or one paragraph where all the words have been included.
The testee's ability for composing his statements will be judged by the type of statements
produced, regardless of whether he has opted for independent sentences or a single paragraph.
However, given the same level of logical ability, the production of only one paragraph
containing all the given words will be ranked higher than the production of independent
sentences.
The statements are rated as:
Creative
Coherent "A"
Coherent "B" (A poor, barely sufficient, though correct contextualization)
Puerile
Nonsensical
Incoherent
The parameters for rating this exercise are:
a. Soundness of the dramatic situations or scripts where the words have been contextualized.
b. Style.
c. Editing.
d. Grammatical correctness.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
298
Examples:
Creative: En nuestra Puna, crece el pasto en las vegas, donde se aprovecha con suma
habilidad cuanto charco llega a formarse sobre la tierra, por lo general seca y
resquebrajada.
Los rostros de los coyas no muestran enojo; enfrentan la dureza del clima
llevando sobre sus cuerpos un tipo de vestido que parece demasiado pesado a
nuestros ojos.
(The following is not a translation, but rather an English language equivalent:
In our Puna, grass grows only in the lowlands. The natives have developed great
ability for exploiting every puddle that forms on the otherwise arid soil.
In spite of their daily struggle against the harsh climate, these coya people look
relaxed and happy. They dress in colourful clothes, and their music is extremely
lively. They never show any anger at their plight.)
Coherent "A": Elisa salió a pasear con hermoso vestido blanco. Cruzó la plaza distraídamente,
caminando sobre el pasto. Se detuvo al encontrar frente a sí un profundo charco,
luego pensó en probar su habilidad para saltarlo. Se impulsó corriendo, pero,
torpemente y para su gran enojo, cayó en las aguas barrosas.
(Eliza went for a walk wearing a beautiful white dress. She crossed the square
absent-mindedly, walking on the grass. She stopped when she came across a
deep puddle and then thought of testing her ability for jumping over it. She ran to
gain momentum, leapt forward but, clumsily and much to her anger, fell into the
muddy waters.)
Learning from Learners‟ Errors Ana María Rozzi de Bergel
299
Coherent "B": Al corretear sobre el pasto, la niña cayó en un charco y se ensució el vestido. La
madre descargó su enojo sobre la pequeña, demostrando muy poca habilidad para
manejar una situación sin importancia.
(As she was running on the grass, the little girl fell into a puddle and stained
her dress. Her mother made her feel the full weight of her anger, showing very
little ability for handling an unimportant problem)
Puerile: La niña mojó su vestido en el charco, y para no producir el enojo de su madre, lo
lavó con habilidad y lo puso a secar sobre el pasto.
(The girl made her dress wet in the puddle and, so as not to arouse her mother's
anger, she washed her dress with great ability and put it on the grass to dry.
Nonsensical: Sin ningún enojo, y mostrando gran habilidad, se puso el vestido sobre el que ya
tenía puesto, cuidando de no mancharse en el charco que había sobre el pasto.
(Showing no anger, but great ability, she put the dress on over the one she was
already wearing, trying not to stain it with the water from the puddle on the
grass)
Incoherent: Luego del enojo por haber manchado su vestido en el charco, demostró su
habilidad para bailar sobre el pasto.
(After the outburst of anger at having stained her dress on the puddle, she
showed her ability for dancing on the grass.)
Exercise 3 measures the testee's predominant verbal association style.
For this purpose, we have used the two basic associative lines described by de Saussure as
the paradigmatic and syntagmatic axes of language, that is; metonymic and metaphoric
association. This diagnosis of the associative style is evaluated by taking into account, as
regards the general style and the creative productive ability of the testee, Jacobson's
Learning from Learners‟ Errors Ana María Rozzi de Bergel
300
discoveries on metonymy and metaphor in aphasias and other organic-psychic and
psychological pathologies and those of the psychoanalytic school of J. Lacan related to
metaphor and metonymy as being the two ways in which the symbolic net is created in the
shaping of the unconscious and, particularly, in the hysterical and obsessive structures.
Examples of verbal association styles:
Metonymic: raíz - suelo maestro - libro
(root - soil) (teacher - book)
The testee relates words by proximity.
Metaphoric: raíz - origen maestro - guía
(root - origin) (teacher - leader)
The testee relates words by analogy. He has drawn fundamental characteristics from
different areas and has created a conceptual link between two sectors of reality.
People who show a strong metaphorical trend in their verbal production are often more
creative and have a greater potential in the field of signifiers. This must, however, be
correlated with the conceptual-logical level in each case: with a clearly pre-conceptual
production, a high creative potential may well produce a lot of self-originated production and
circumlocutions, used to make up for the lack of concepts. A highly metonymic production
coupled with a low conceptual level, on the other hand, usually results into vague, imprecise
and primitive linguistic production.
When evaluating Exercise 3, allowances must be made for the fact that the presence of
metaphor is usually higher in women due to their predominantly hysterical structure, whereas
there is a predominance of metonymy in men because of their mainly obsessive structure.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
301
References
Akmajian, A., Demers, R.A., Farmer, A.K. & Harnish, R. M. (1995). Linguistics: An
Introduction to Language and Communication. Cambridge, MA: The Massachusetts
Institute of Technology Press.
Allen, J.P.B (1977). Some Basic Concepts in Linguistics, The Edinburgh Course in Applied
Linguistics, (Vol.2, 16-43).Oxford: Oxford University Press.
Association for Language Awareness. http://www.languageawareness.info/ Visited in
December 2007.
Austin, J.L & Urmson, J.O. (Ed.). (1962). How to do things with Words: The William James
Lectures delivered at Harvard University in 1955. Oxford: Clarendon.
Ausubel, D. (1963). The Psychology of Meaningful Verbal Learning. New York: Grune &
Stratton.
Ausubel D. (1968). Educational Psychology: A Cognitive View. New York: Holdt, Rinehart &
Wilson.
Bandura, A. (1977). Social Learning Theory, Englewood Cliffs, NJ: Prentice Hall.
Baron Cohen, S. (1995). Mindblindness. An essay on autism and theory of mind. Cambridge,
MA: Cambridge University Press.
Bartlett, F.C. (1932). Remembering: An Experimental and Social Study. Cambridge (UK):
Cambridge University Press.
Bartlett, F.C. (1958). Thinking. New York: Basic Books.
Bialystok, E. & Hakuda, K. (1994). The Science of psychology of second language
acquisition. New York: Basic Books.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
302
Bley-Vroman, R. (1988). The fundamental character of foreign language learning. In W.
Rutherford & M. Sharwood Smith (Eds.). Grammar and second language teaching: A
book of readings (19-30). Rowley, MA: Newbury House.
Bartram, M. & Walton, R. (1991). Correction. Language Teaching Publications. 87 - 91.
Bloomfield, L. (1977) Objectivity in Linguistics. The Edinburgh Course in Applied
Linguistics, (Vol.1, 194 – 199).Oxford (UK): Oxford University Press.
Brown, H. D. (2006). Principles of Language Learning and Teaching. (5th edition.) London:
Pearson Longman.
Bruner, J. (1960). The Process of Education. Cambridge, MA: Harvard University Press.
Bruner, J. (1966). Toward a Theory of Instruction. Cambridge, MA: Harvard University Press.
Bruner, J. (1986). Child's talk: learning to use language. New York: Norton.
Bruner, J. (1990). Actos de significado: más allá de la revolución cognitiva. Madrid: Alianza.
Bruner, J. (1996). The culture of education. Cambridge, USA: Harvard University Press.
Brumfit, C. J. (1981) Notional syllabuses revisited: a response. Applied Linguistics 1981 II:
90-92.
Butler-Takana, P. (2000). Fossilization: a chronic condition or is consciousness-raising the
cure? Unpublished dissertation, Faculty of Arts of the University of Birmingham,
Birmingham, United Kingdom. Retrieved in June 2007 from
www.cels.bham.ac.uk/resources/essays/ButlerTdiss.pdf
Cook, V.J. (2003). Effects of the second language on the first. Clevedon: Multilingual Matters.
Carroll, J.B. (1973) Linguistic Relativity and Language Learning. The Edinburgh Course in
Applied Linguistics, (Vol.1, 126 – 145).Oxford: Oxford University Press.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
303
Carroll, J. (1981). Twenty-five years of research on foreign language aptitude. In K. C. Diller
(Ed.), Individual differences and universals in language learning aptitude (83-117).
Rowley, MA: Newbury House.
Carroll, J. B. & Sapon, S. (2002) Modern Language Aptitude Test: Manual 2002 Edition.
Bethesda, MD: Second Language Testing, Inc.
Chomsky (1975). Reflections on Language. New York: Pantheon Books.
Chomsky, N. (1957). Syntactic Structures. The Hague:Mouton & Co.
Cooley, Ch.H. (1909). Social Organization: A Study of the Large Mind. New York: Scribner.
Dascal, M. (1999). Filosofía del Lenguaje II – Pragmática. Madrid: Editorial Trotta.
De Saussure, F. (1959). Course in General Linguistics. New York:McGraw-Hill.
Díaz-Benjumea, M.D.J. (2002). Lo inconsciente psiconalítico y la psicología cognitiva: una
revisión interdisciplinar. Aperturas psicoanalíticas. No 11. Retrieved in March 2008 from
http://www.aperturas.org/11diazbenjumea.html.
Dewey, John (1897) My pedagogic creed. The School Journal, Vol. LIV, N. 3 (January 16,
1897), 77-80. Also available in the informal education archives. Retrieved in July 2007
from http://www.infed.org/archives/e-texts/e-dew-pc.htm.
Edge, J. (1989). Mistakes and correction. London:Longman.
Ellis, R. (1985). Understanding Second Language Acquisition. Oxford (UK): Oxford
University Press.
Ellis, R. (1989). Are classroom and naturalistic acquisition the same? A study of the classroom
acquisition of German word order rules. Studies in second language acquisition, 11, 305-
328.
Ellis, R. and Gaies, S. (1998). Impact Grammar. Hong Kong: Longman.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
304
Encyclopædia Britannica Online: http://www.britannica.com/eb/article-9108460 Consulted in
January 2008.
Firth, J.R. (1977). Personality and Language in Society. J.P.B. Allen & A. Davies (Eds.) The
Edinburgh Course in Applied Linguistics (Vol.1, 15-22). Oxford, UK: Oxford University
Press.
Gardner, H. (1991). Intelligence in seven steps. D. Dickinson (Ed.). Creating the future.
Retrieved in July 2007 from
http://www.newhorizons.org/future/Creating_the_Future/crfut_gardner.html.
Gardner, H. (1993). Intelligence Reframed. Multiple intelligences for the 21st century, New
York: Basic Books.
Garger, S., & Guild, P. (1984). Learning styles: The crucial differences. Curriculum Review,
23(1), 9-12.
Gass, S & Selinker,L. (1983). Language Learning. Ann Arbor: University of Michigan Press.
Gass, S. & Selinker, L. (1994). Second Language Acquisition: An Introductory Course.
Hillsdale, NJ.: Laurence Erlbaum Associates.
Gibaja, R. (1993). El tiempo instructivo. Buenos Aires: Aique
Gumperz, J. J.& Levinson, S.C.(Eds.). (1996). Rethinking Linguistics Relativity. New York:
Cambridge University Press.
Gunter, T., Stowe, L., & Mulder, G. (1997). When syntax meets semantics. Psychophysiology,
34, 660-676.
Halliday, M.A.K. (1976) System and function in language: Selected Papers. G. Kress (Ed).
London: Oxford University Press.
Halliday, M. A. K., & Hasan, R. (1989). Language, Context, and Text. Oxford (UK): Oxford
University Press.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
305
Harmer, J. (1997).The practice of English language teaching. London:Longman.
Harmer, J. (2001). The practice of English language teaching. Essex: Pearson Education
Limited.
Horgan, T. & Tienson, J. (1996). Connectionism and the Philosophy of Psychology.
Cambridge, MA: MIT Press.
Humphrey, N. (1986). The Inner Eye. Madrid: Alianza.
Hymes, J. (1979) D. On Communicative Competence. In C.J. Brumfit, & K. Johnson (Eds.),
The Communicative Approach to Language Teaching (Ch. 1, 5-26) Oxford (UK): Oxford
University Press.
James, C. (1998). Errors in language learning and use. London: Longman
James, C. (1999). Language awareness: implications for the national curriculum. Language,
Culture and Curriculum. Vol.12 No 1. Pp. 94-115. Retrieved in August 2007 from
http://www.multilingual-matters.net/lcc/default.htm.
Johnson, K. (1996). Language Teaching and Skill Learning. Oxford (UK): Basil Blackwell.
Johnson, M. & Lakoff, G. (1980) Metaphors We Live By. Chicago: University of Chicago.
Kaur, A. (1990). Considerations in language syllabus design. The English Teacher Vol XIX.
Retrieved in December 2007 from http://www.melta.org.my/ET/1990/main1.html.
Kellerman, E. & Sharwood Smith, M. (1986). Crosslinguistic influence in second language
acquisition. London: Prentice Hall.
Knowles, M. S. et al (1984). Andragogy in Action. Applying modern principles of adult
education, San Francisco: Jossey Bass.
Kohler, W., Koffka, K. y Sander, F. (1963). Psicología de la forma. Buenos Aires: Paidós.
Kolb, D. A. (1984). Experiential Learning: Experience as the Source of Learning and
Development. Englewood Cliffs, N.J.: Prentice-Hall, Inc.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
306
Krashen, S. & Terrell, T. (1983). The natural approach. New York: Pergamon/Alemany.
Lacan, J. & Wilden, A. (1968) The Language of the Self: The Function of Language in
Psychoanalysis. Baltimore: Johns Hopkins Press.
Lado, R. (1957). Linguistics across cultures. Ann Arbor:University of Michigan Press.
Lave, J. & Wenger, E. (1991). Situated Learning. Legitimate peripheral participation,
Cambridge (UK): University of Cambridge Press.
Lenneberg, E. H. (1967). Biological foundations of language. New York: Wiley.
Lesser, E. & Storck, J. (2001). Communities of practice and organizational performance.
IBM Systems journal.40:4
Lewin, K. (1951). Field theory in social science. New York: Harper Torch books.
Lewin, K., Llippit, R. & White, R.K. (1939). Patterns of aggressive behavior in
experimentally created social climates. Journal of Social Psychology, 10, 271-301
Lewis, M. (1993). The Lexical Approach. Hove, UK: Language Teaching Publications.
Lightbown, P. M. & Spada, N. (1999). (2nd edition). How Languages are learnt. Oxford
(UK): Oxford University Press
Locke, J.L. (1997). A theory of neurolinguistic development. Brain and Languages 58, 265–
326. Article No BL971791
McCarthy Gallagher, J.& Reid, D.K. (Eds.) (2002) The Learning Theory of Piaget and
Inhelder. New York: Authors Choice Press.
Malinowski B (1923). The problem of meaning in primitive languages. Supplement to C. K.
Ogden and I. A. Richards (Eds.) The meaning of meaning: A study of the influence of
language upon thought and the science of symbolism (pp. 451-510). London: Routledge
Maslow, A. (1954). Motivation and personality. New York: Harper.
Merton, R.K. (1948). The self-fulfilling prophecy. Antioch Review, pp. 193-210.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
307
Montgomery, C. & Eisenstein, M. (1986). Reality revisited: An experimental
communicative course in ESL. TESOL Quarterly 19(2), 317-333.
Naiman, N., Frohlich, M., Stern, H.H. & Todesco, A. (1978). The Good Language Learner.
Toronto: Ontario Institute for Studies in Education.
Nemser, W. (1974), Approximative systems of foreign language learners. In J. Richards (Ed.)
Error Analysis. (Ch.4, 55-64) London: Longman
Norrish, J. (1995). Language learners and their errors. London:Macmillan.
Nunan, D. (1988). Curriculum design. Oxford (UK): Oxford University Press.
Nunan, D. (Ed.) (1993). Collaborative language learning and teaching. Cambridge (UK):
Cambridge University Press.
O'Malley, M. & Chamot, A. U.(1990) Learning strategies in second language acquisition.
Oxford (UK): Oxford University Press.
O‟Neill, R. (2002) My debt to - and argument with - Stephen Krashen Robert O'Neill offers a
rebuttal to Krashen's reply. Paper presented at the ESL Miniconference online. Retrieved
in August 2006 from http://www.eslminiconf.net/april/oneillreply.html.
Pavlov, I.P. (1927). Conditioned reflexes: an investigation of the physiological activity of the
cerebral cortex. Oxford, (UK): Oxford University Press.
Penfield, W. & Roberts, L. (1959). Speech and brain mechanisms. Princeton: Princeton
University Press.
Piaget, J. (1960). The psychology of intelligence. New York: Humanities Press Inc.
Piaget, J. (1962) Comments on Vygotsky‟s critical remarks concerning „The Language and
Thought of the Child‟, and „Judgment and Reasoning in the Child‟. Cambridge, MA.: The
MIT Press.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
308
Pichon Rivière, E.(1984). El proceso grupal. Del psicoanálisis a la psicología social. Buenos
Aires: Nueva Visión.
Pimsleur, P., Reed, D.J. & Stansfield, C.W. (2004). Pimsleur Language Aptitude Battery:
Manual 2004 Edition. Bethesda, MD: Second Language Testing, Inc.
Pit Corder, S. (1972). Errata. Papers in error analysis. In J. Svartvik (Ed.). The Elicitation of
Interlanguage. (Ch.5, 36-48). Lund: CWK Gleerup
Pit Corder, S. (1974). Error analysis. In J.P.B. Allen & S.Pit Corder (Eds.). The Edinburgh
course in applied linguistics, (Vol. 3,122-154). Oxford (UK): Oxford University Press.
Pit Corder, S. (1981). Error analysis and interlanguage. Oxford, (UK): Oxford University
Press.
Purdy, J.D. (2001) An investigation into the neurolinguistic basis of fossilization. Retrieved in
November 2006 from http://www.tc.columbia.edu/Academic/TESOL/Han/JD.html.
Reid, J. M. (Ed.) (1995). Learning Styles in the ESL/EFL Classroom. New York: Heinle &
Heinle Publishers.
Reiss, M.A. (1985). The good language learner: Another look. The Canadian Modern
Language Review 4, 511 - 23.
Richards, J. (Ed.) (1974). Error Analysis. London:Longman.
Ringbom, H. (1987). The role of the first language in foreign language learning. Clevedon:
Multilingual Matters.
Rogers, C. (1972) Psicoterapia centrada en el cliente. Buenos Aires: Paidós.
Rosch, E. (1981). Prototype Classification and Logical Classification: The Two Systems. In E.
Scholnick (Ed), New Trends in Conceptual Representation. (73-85). New Jersey: Erlbaum.
Rosch, E.H. (1981) Principles of categorization. In E. H. Rosch & B. Lloyd, (Eds.). Cognition
and Categorization, (27-48). New York: Lawrence Erlbaum Associates.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
309
Rubin, J. (1975). What the 'good language learner' can teach us. TESOL Quarterly 9/1, 41- 51.
Slamecka, N.J. & Ceraso, J.(2000). Retroactive and Proactive Inhibition of Verbal Learning.
Retrieved in June 2007, from
http://www.uwf.edu/psych/bmikulas/Webpage/learning/section%20four.htm
Savignon, S. (1972). Communicative Competence: An Experiment in Foreign Language
Teaching. Philadelphia: Center for Curriculum Development.
Scarcella, R.C. & Oxford, R.L. (1992). The tapestry of language learning. Boston, MA.:
Heinle & Heinle.
Searle, J. (1980). Actos de habla. Madrid: Ediciones Cátedra.
Selinker, L. (1972). Interlanguage. International Review of Applied Linguistics in Language
Teaching, 10, 219-231.
Selinker, L., Swain, M. and Dumas, G. (1975). The interlanguage hypothesis extended to
children, Language Learning 25, 139-52.
Selinker, L. & Lamendella, J. (1978). Two perspectives on fossilization in interlanguage
learning. Interlanguage Studies Bulletin, 3, 143-191.
Selinker, L. & Lamendella, J. (1979). The role of extrinsic feedback in interlanguage
fossilization: A discussion of "Rule fossilization: A tentative model". Language Learning,
29, 363-375.
Selinker, L. & Lamendella, J.T (1981). Updating the interlanguage hypothesis. Studies in
Second Language Acquisition, 3 (2), 201-220.
Selinker, L. (1988). Papers in interlanguage. Occasional papers No.44, Southeast Asian
Ministers of Education Organization (Singapore): Regional Language Centre. ED321549.
Selinker, L (1992).Rediscovering interlanguage. New York: Longman
Learning from Learners‟ Errors Ana María Rozzi de Bergel
310
Selinker, L. and Gass, S. (Eds.) (1992). Language transfer in language learning. Amsterdam:
John Benjamins.
Selinker, L. (2001). Fossilisation: moving the concept into empirical longitudinal study. In C.
Elder, A. Brown, E. Grove, K. Hill, N.Iwashita, T. Lumley, T. McNamara & K.
O‟Loughlin, (Eds). Experimenting with uncertainty: Essays in honour of Alan Davies. (pp.
266-288). Cambridge (UK): Cambridge University Press.
Skinner, B.F. (1957). Verbal Learning. New York: Appleton-Century-Crofts.
Skinner, B.F. (1977). Verbal behaviour. The Edinburgh course in applied linguistics, (Vol.2,
16-43). Oxford (UK): Oxford University Press
Strevens, P. (1969). Two ways of looking at error analysis, ERIC ED037 714.
Tarone, E. Cohen, A.D. & Dumas, G. (1983) A closer look at some interlanguage
terminology: a framework for communication strategies. In C.Færch & G. Kasper (Eds.)
Strategies in interlanguage communication. London: Longman.
Tollefson, J. & Firn, J., (1983). Fossilization in second language acquisition: An inter-model
view. RELC Journal, 14, 19-34.
Tolman, E.C. (1932). Purposive Behavior in Animals and Men. New York: Appleton-Century-
Crofts.
Ur, P. (1996). A course in language teaching: from practice to theory. Oxford (UK): Oxford
University Press.
Vigil, N. & Oller, J. (1976). Rule fossilization: A tentative model. Language Learning, 26,
281-295
Vos, S., Gunter, T., Schriefers, H., & Friederici, A. (2001) Syntactic parsing and working
memory: The effects of syntactic complexity, reading span, and concurrent load. Language
and Cognitive Processes, 16, 65-103.
Learning from Learners‟ Errors Ana María Rozzi de Bergel
311
Vygotsky, L. (1962). Thought and language. In E. Hanfmann & G.Vakar, (Eds.). Cambridge,
MA: MIT Press.
Vygotsky, L. S. (1978). Mind in society. Cambridge, MA: Harvard University Press
Whorf, B.L. (1973) Linguistic Relativity. In J.P.B. Allen & A. Davis (Eds.). The Edinburgh
Course in Applied Linguistics, (Vol.2, 103-114). Oxford (UK): Oxford University Press.
Widdowson, H.G.(1978). Teaching language as communication. Oxford (UK):Oxford
University Press.
Wilkins, D.A. (1976) Notional Syllabuses. Oxford: Oxford University Press.
Wohlwill, J. (1970) Hacia una reformulación del papel de la experiencia en el desarrollo
cognitivo. En Ajurriaguerra, Bresson, Fraisse, et al. Psicología y epistemología genéticas.
Temas piagetianos. (167-177) Buenos Aires: Proteo.
Wright, T. (1987). Roles of teachers and learners. Oxford (UK): Oxford University Press
Zadeh, L. (1965) Fuzzy Sets. Information and Control, 338-353 . Retrieved in January 2005
from: http://theory.lcs.mit.edu/~iandc/ic65.html