NEWCASTLE AND NORTHUMBRIA VOLUME 2 ISSN: 2041-1057

i

ii

NEWCASTLE AND NORTHUMBRIA

WORKING PAPERS

IN LINGUISTICS

VOLUME 22 (2016)

ISSN: 2041-1057

iii

GENERAL EDITORS

Marwan Jarrah

School of English Literature, Language and

Linguistics, Newcastle University,

Newcastle upon Tyne, NE1 7RU, United

Kingdom

Abdulkareem Yaseen

School of Education, Communication &

Language Sciences, Newcastle University,

Newcastle Upon Tyne, NE1 7RU, United

Kingdom

EDITORIAL SUPPORT TEAM

Hadi Alsamdani

Murdhy Alshamari

Eisa Alrasheedi

PUBLISHED BY

Centre for Research in Linguistics and Language Sciences (CRiLLS)

Newcastle University

COVER DESIGN

Abdulkareem Yaseen

FOR INQUIRIES

CRiLLS,

Room 3.01 & 3.02, Level 3,

Old Library Building

Newcastle University,

Newcastle upon Tyne NE1 7RU

United Kingdom, Telephone: 0191 208 8790

Fax: 0191 208 6592, E-mail: [email protected]

mailto:[email protected]

iv

CONTENTS

The syntax of temporal adverbial clauses in Najdi Arabic 1

AHMAD ALSHAMMARI

In favor of contrastive topic phrase in North Hail Arabic left periphery: Evidence

from the discourse particle tara

23

MURDY ALSHAMARI

An exploratory study of first exposure to Urdu: Discovering what matters 40

EVA MALESSA and REBECCA LIM

Weightlessness preservation in Jordanian Arabic varieties: A stratal OT analysis 71

MARWAN JARRAH

v

EDITORIAL NOTE

Newcastle and Northumbria Working Papers in Linguistics aims to publish the research of the

Centre for Research in Linguistics and Language Sciences (CRiLLS) as well as external

contributions. The current volume focuses on some syntactic and phonological aspects related

to some Arabic varieties and one exploratory study of first exposure to Urdu. Postgraduate

students at Newcastle University contributed the papers. The general editors and the editorial

team support would like to thank all anonymous reviewers for their invaluable comments.

Newcastle and Northumbria Working Papers in Linguistics 22 (2016) Alshammari

1

THE SYNTAX OF TEMPORAL ADVERBIAL CLAUSES IN NAJDI ARABIC

AHMAD ALSHAMMARI

(Newcastle University)

Abstract

This research investigates the derivation of temporal adverbial clauses in Najdi Arabic (NA)

with special reference to argument and adjunct fronting. It argues that NA exhibits the

dichotomy between central adverbial clauses (CAC) and peripheral adverbial clauses (PAC) in

the sense of Haegeman (2003, 2010). The adverbial clauses introduced by gablla, gablma

‘before’, and baʕdma ‘after’ are CAC, whereas the adverbial clauses introduced by yoom and

lamma ‘when’ are PAC. Both negation and epistemic modality are used as diagnostics of

bearing out this assumption. Additionally, the current paper argues that CAC has a truncated

left periphery in that no Force Phrase, Topic Phrase, nor Focus Phrase are projected. It is

impossible for arguments and adjuncts thus to be fronted in such clauses since there are no

dedicated projections for them. The current research advocates for the truncation approach,

instead of the operator movement approach, to account for the reason why arguments and

adjuncts resist fronting. Unlike CAC, PAC in NA has a fully-articulated CP where both

argument and adjuncts can be displaced to the left periphery without imposing any restriction.

1 Introduction and study aims

Much recent work in cross-linguistic syntax has addressed syntax of adverbial clauses

(Hengeveld 1996, Haegeman 2003, 2006, 2010, and Lahousse et al., 2014). The main reason

for this cross-linguistic investigation of such clauses has been ascribed to the assumption that

certain types of adverbial clauses resist argument fronting, hence the possibility that these types

of adverbial clauses exhibit reduced left periphery (Danckaert 2011). For instance, while

arguments, e.g. the direct object, cannot be topicalized in certain English adverbial clauses,

these arguments can be fronted in Romance languages such as Italian and French. In addition,

asymmetry between arguments and adjuncts is held with regard to fronting in adverbial clauses

(Haegeman 2010: 632). Unlike the case in the matrix clauses, arguments and adjuncts do not

behave similarly concerning fronting in adverbial clauses. These observations have attracted

much attention from scholars, arguing for the notion that the left periphery of some types of

adverbial clauses is different from that of the root clauses (see, Demirdache and Uribe-

Etxebarria 2004, Nilsen 2004, and Munaro 2005).

On the other hand, less agreement has been gained on the actual structure of adverbial

clauses in natural languages, which are, in turn, different in placing certain restrictions on

argument and/or adjunct fronting within adverbial clauses (Haegeman 2010). Several calls have

reportedly been made for a wider cross-linguistic investigation in order to determine the syntax

of adverbial clauses (cf. Cinque 2004, Frey 2012, and Endo & Haegeman 2014, among others).

Along these lines, the current research looks into the syntax of adverbial clauses in one Arabic

dialect, Najdi Arabic (henceforth, NA), which has never been investigated in respect of this

particular point.1 Generally speaking, the main aim of the current research is twofold. Firstly,

it provides a full description of one type of adverbial clauses, namely the temporal adverbial

clauses, in terms of their distribution and external syntax. Secondly, it attempts to account for

the internal syntax of such clauses with emphasis placed on the issues related to their left

periphery. To achieve these aims, the main assumptions of the Minimalist Program (Chomsky

1993, 1995, 2000, 2001) will be used.

1 NA is spoken in Najd region, a part of what is today known as ‘Kingdom of Saudi Arabic’ (Ingham 1994).

Alshammari The syntax of temporal adverbial clauses in Najdi Arabic

2

2 The structure of temporal adverbial clauses in NA:

Temporal adverbial clauses must begin with a subordinator. Table 1 shows the subordinators

introducing the temporal adverbial clause in NA:

Table 1: The subordinators introducing the temporal adverbial clause in NA

The subordinator The meaning

yoom when

lam-ma when

gabl-la before

gabl-ma before

baʕd-ma after

As can be seen from Table 1, all subordinators with the exclusion of yoom ‘when’ consist of

two parts. It should be stressed that nothing can intervene between the two parts of the

compound subordinators whatsoever. See the following sentences where the temporal adverbial

clauses are bracketed (S= subject; O= object):

(1) a. *[gabl Fahd (S) ma yswe al-ʕamliah

before Fahd Particle do.3SM.PRES DEF-operation

bi-l-mistaʃfa] ʔaxδ al-ʕladʒ

in-DEF-hospital take.3SM.PAST DEF-medicine

Intended: ‘Before Fahd had the operation in the hospital, he had taken the

medicine.’

b. *[gabl al-ʕamliah(O) ma yswe Fahd

before DEF-operation Particle do.3SM.PRES Fahd




medicine.’

c. *[gabl bi-l-mistaʃfa (Adjunct) ma yswe

before in-DEF- hospital Particle do.3SM.PRES

Fahd al-ʕamliah] ʔaxδ al-ʕladʒ

Fahd DEF-operation take.3SM.PAST DEF-medicine

Intended: ‘Before Fahd had the operation in the hospital, he took the medicine.’

The same observation holds true of the other three compound subordinators gabl-la ‘before’,

lam-ma ‘when’, and baʕd-ma ‘after’. Furthermore, the subordinator must appear at the

beginning of the adverbial clauses. No elements are permitted to appear to the left of the

temporal subordinator.

Certain restrictions are placed on the word order utilized in the adverbial clause when

the subordinators gabl-la, gabl-ma ‘before’, and baʕd-ma ‘after’ are used, whereas no such

restrictions are noticed when yoom ‘when’ and lamma ‘when’ are used. When gabl-la, gabl-

ma ‘before’, and baʕd-ma ‘after’ are used as subordinators, VS(O) is invariably used; SV(O) is

ungrammatical. Consider the following sentences:


3

(2) a. Fahd ʔaxδ al-ʕladʒ [gablma

Fahd take.3SM.PAST DEF-medicine before

yidʒi (V) ad-doctoor (S)]

come.3SM.PRES DEF-doctor

‘Fahd had taken the medicine before the doctor came.’

b. *Fahd ʔaxδ al-ʕladʒ [gablma


ʔad-doctoor (S) yidʒi (V)]

DEF-doctor come.3SM.PRES

Intended: ‘Fahd had taken the medicine before the doctor came.’

In sentence (2a), VS is the word order in the bracketed adverbial clause. In sentence (2b), SV

is used instead; hence its ungrammaticality. The same observation holds of gabl-la ‘before’ and

baʕd-ma ‘after’. Consider the contrasts in the following sentences:

(3) a. Fahd ʔaxδ al-ʕladʒ [gablla


yidʒi (V) ad-doctoor (S)]



b. *Fahd ʔaxδ al-ʕladʒ [gablla


ʔad-doctoor (S) yidʒi (V)]

DEF-doctor come.3SM.PRES

Intended: ‘Fahd had taken the medicine before the doctor came.’

c. [baʕdma kala (V) Fahd (S) wadʒbah bi-l-mațʕam]

After eat.3SM.PAST Fahd meal in-DEF-restaurant

raḥ li-l-mistaʃfa

go.3SM.PAST to-DEF-hospital

‘After Fahd had eaten a meal in the restaurant, he went to the hospital.’

d. *[baʕdma Fahd (S) kala (V) wadʒbah bi-l-mațʕam]

After Fahd eat.3SM.PAST meal in-DEF-restaurant

raḥ li- l-mistaʃfa

go.3SM.PAST to-DEF- hospital

Intended: ‘After Fahd had eaten a meal in the restaurant, he went to the hospital.’

In addition, other marked word orders including OVS or OSV are also prohibited in temporal

adverbial clauses introduced by gablla, gablma ‘before’, and baʕdma ‘after’, i.e., no object

fronting is allowed. Consider the following sentences:

(4) a. *[gabl-ma/la al-ʕamliah (O) yswe-ha (V) Fahd (S)

before DEF-operation do.3SM.PRES-it Fahd




medicine.’


4

b. *[ baʕdma al-ʕamliah (O) sawa-ha (V) Fahd (S) bi-l-mistaʃfa]

After DEF-operation do.3SM.PAST-it Fahd in-DEF-hospital

ʔaxδ al-ʕladʒ

take.3SM.PAST DEF-medicine

Intended: ‘After Fahd had had the operation in the hospital, he took the

medicine.’

This restriction on the position that the direct object can occupy indicates strongly the

assumption that with such subordinators, the direct object cannot be fronted. Similarly, no

adjunct fronting is allowed either. Summing up, neither arguments nor adjuncts can be fronted

in the temporal adverbial clauses introduced by gablla, gablma ‘before’, and baʕdma ‘after’,

and VS(O) is invariably used in such clauses. These observations will be accounted for below,

arguing that the temporal adverbial clauses introduced by gablla, gablma ‘before’, and baʕdma

‘after’ have their left periphery truncated.

On the other hand, no such restrictions are placed when the subordinators used are yoom

and lamma ‘when’. Both unmarked word orders SVO and VSO are freely used. Also, other

marked orders (e.g., OVS) can be used. Consider the following sentences for yoom ‘when’:

(5) a. [yoom yʃtiri (V) Fahd (S) as-sayarah (O) min al-ma ʕra ḍ]

when buy.3SM.PRES Fahd DEF-car from DEF-car show

kant dʒadi:dah

be.3SF.PAST new

‘When Fahd bought the car from the car show, it was new.’

B. [yoom Fahd (S) yʃtiri (V) as-sayarah (O) min al-ma ʕra ḍ]

when Fahd buy.3SM.PRES DEF-car from DEF-car show

kant dʒadi:dah

be.3SF.PAST new


c. [yoom as-sayrah (O) yʃtiri-ha (V) Fahd (S) min al-maʕraḍ]

when DEF-car buy.3SM.PRES-it Fahd from DEF-car show

kant dʒadi:dah

be.3SF.PAST new


VSO, SVO, and OVS are all accepted, regardless of the interpretive differences between them.

By the same token, adjunct fronting is allowed with these two subordinators. Consider the

following sentences (fronted adjuncts are in italics for ease of explanation):

(6) a. [yoom min al-maʕraḍ yʃtiri (V) Fahd (S) as-sayarah (O)]

when from DEF-car show buy.3SM.PRES Fahd DEF-car

kant dʒadi:dah

be.3SF.PAST new



5

b. [lamma bi-l-mazraʕah katb (V) Fahd (S) al-qaṣi:dah (O)]

when in-DEF-farm write.3SM.PAST Fahd DEF-poem

ʕrsal-ha li:

send.3SM.PAST-it to me

‘When Fahd wrote a poem in the farm, he sent it to me.’

Following this, we are inevitably led to the conclusion that unlike the temporal adverbial clauses

introduced by gablla, gablma ‘before’, and baʕdma ‘after’, the temporal adverbial clauses

introduced by yoom and lamma ‘when’ permit both adjunct and argument fronting. These

differences can be represented in Table 2:

Table 2: differences between adverbial clauses

Adverbial clauses

introduced by:

SVO VSO Argument fronting

(e.g. OVS)

Adjunct fronting

gablla,

gablma

baʕdma

NO YES NO NO

yoom

lamma

YES YES YES YES

The current study aims primarily at providing an account of these differences within the recent

relevant syntactic approaches. Prior working out these questions in detail, it is important to

introduce the major assumptions held of adverbial clauses and sketching out the main proposals

suggested for these clauses in cross-linguistic syntax, namely: the truncation approach and the

operator movement approach.

3 Peripheral versus central adverbial clauses

In her extensive study of adverbial clauses, Haegeman (2002) distinguishes between two types

of adverbial clauses, namely: central adverbial clauses and peripheral adverbial clauses. In

addition to their differences with regard to the degree of syntactic integration with the matrix

clause they are associated with, both types are assumed to be different in terms of their internal

syntax. In the following discussion, these differences: the degree of integration as well as the

internal syntax will be explored.

3.1 External syntax: degrees of integration

For Haegeman (2002, 2003), adverbial clauses differ with regard to their integration with the

associated clause (i.e., matrix clause). She observes that certain adverbial clauses are more

syntactically integrated with the associated clauses than other adverbial clauses. She claims that

this difference in syntactic integration with the associated clauses affects their external syntax,

in that those adverbial clauses with more syntactic integration are assumed to be merged with

the matrix clause at an earlier point in the derivation than those with less syntactic integration

with the associated clause (Haegeman 2003: 71). Haegeman calls the former type of adverbial

clauses, with more syntactic integration with the associated clause, ‘Central Adverbial Clauses

(henceforth, CAC)’, whilst the latter, with less syntactic integration with the associated clause,

are ‘the Peripheral Adverbial Clauses’ (henceforth, PAC).

Accordingly, CACs and PACs are different in terms of their (semantic) interpretation

and relation with the event of the matrix clause. For Haegeman (2002) and depending on

English data, the CACs main semantic function is to structure the event, which is expressed in

the associated clause, whereas the PACs main function is to structure the discourse. PACs

express propositions, which are to be processed as part of the discourse background for the


6

proposition, which is expressed in the associated clause. In order to appreciate this point,

consider the following examples in (7a) and (7b) (both adapted from Haegeman 2002: 62):

(7) a. According to Smith, a group of Arkansas state troopers who worked for Clinton

while he was a governor wanted to go public with tales of Clinton’s womanising.

(event time: 'during the time that')

b. While [Dr Williams’] support for women priests and gay partnerships might

label him as liberal, this would be a misleading way of depicting his

uncompromisingly orthodox espousal of Christian belief. (background assumption:

'whereas',)

In (7a), the adverbial clause introduced by while provides a temporal specification of the event,

whereas in (7b) the adverbial clause introduced by while provides a background proposition

which, combined with the proposition expressed by the associated clause, will yield contextual

implications. In order to confirm this dichotomy between CAC and PAC with respect to the

degree of their syntactic integration with the associated clause, many pieces of evidence have

been adduced in the literature, including coordination of likes, scope phenomena, VP ellipsis

and strict/sloppy identity.

3.1.1 Co-ordination of likes

Co-ordination of likes has been taken as hard evidence for the distinction between CAC and

PAC, given the typical assumption that only adverbial clauses with parallel interpretations can

be coordinated. Haegeman (2002) provides the following sentences in (8a) and (8b) to account

for the difference in interpretation between CAC and PAC, bearing in mind that CAC is

assumed to be event-related, whilst PAC to be discourse-related (Haegeman 2002; 2003):

(adapted from Haegeman 2002: 63):

(8) a. The party is also in danger of alienating older people above the poverty line, Mr

Cable argues. ‘Both these groups will swing to the Conservatives if the Tories are

smart enough and if we have nothing much to offer them.’

b. But if Sir Richard has been tainted by the affair, and if Mr Sixsmith’s role may

not have been as entirely well intentioned as he claims, the individual most

damaged by the row remains Stephen Byers.

In (8a), both conditional clauses have the same event-related interpretation, whereas adverbial

conditionals in (8b) have the same discourse-related interpretation, which is seen as peripheral

to the main event in the associated clause.

3.1.2 Scope phenomena

Haegeman (2002) argues that scopal properties can distinguish between CAC and PAC. The

main argument maintained is that CAC can be interpreted within the scope of operators in the

associated clause. By contrast, PAC is outside the scope of the operators in the associated

clause. This scopal difference can be illustrated in a number of different ways, such as negation

and focus.

3.1.2.1 Negation

Similarly, matrix clause negation has scope over CACs rather than PACs which cannot fall

within the scope of a negative operator in the associated clause. Consider the contrast in (9):


7

(9) a. He doesn’t drink while he is driving.

b. He never drinks while he is driving.

c. My husband doesn't smoke cigarettes, while he does occasionally smoke a cigar.

In (9a) and (9b), the negation can be said to range over the whole complex event: 'he does not

drink-drive', while in (9c) the two propositions (of the matrix clause and of the PAC) are

interpreted in parallel, hence only one of them is negated (Haegeman 2002: 66).

3.1.2.2 Focus

A focus operator in the matrix clause may range over CAC (as in 10a) but cannot do so over

PAC as in (10b). (Haegeman 2002: 68):

(10) a. It was after I left that I realised he was my former teacher.

b. *It was while my mother was a housewife that my father used to work in a

brickyard.

In (10a), clefting, as one manifestation of focus, holds grammatical in the sentence with CAC

but proves ungrammatical in sentence (10b) with PAC being located outside the scope of the

focus operator of the matrix clause.2

3.1.2.3 VP ellipsis and strict/sloppy identity

Haegeman (2002) indicates that CAC and PAC are also different with regard to VP ellipsis.

CAC may be affected by VP-ellipsis. When CAC contains a pronoun, VP-ellipsis may lead to

the so-called sloppy identity reading. Consider sentence (11) and its two readings in (12)

(adapted from Haegeman 2002: 69):

(11) Johni will leave the meeting before hisi paper is discussed and so will Bill.

(12) a. Bill will also leave the meeting before Johni’s paper is discussed.

b.Billj will also leave the meeting before Bill'sj paper is discussed.

The conjunct so will Bill can be interpreted as in (12a) or as in (12b), which signals the sloppy

identity reading. On the other hand, the PAC is not affected by VP-ellipsis and VP-ellipsis does

not lead to sloppy readings. The sentence in (13) has only one reading in (14):

(13) While hisi wife is unemployed, Johni has a high-powered job in the city and so

does Jamesj.

(14) Jamesj also has a high powered job in the city.

Commenting on the sentence in (13), Haegeman (2002: 70) states that ‘there is no way in which

one will assign a sloppy identity interpretation to the pronoun his assuming that James's wife is

also unemployed’. For her, the blocking of VP-ellipsis in the PAC is qualified as evidence that

PACs are attached outside the VP of the associated clause; hence, VP-ellipsis cannot affect the

adverbial clause.

2 Additionally, it was noted that interrogative operators in the matrix clause range over the CAC but cannot do so

over PAC (Haegeman 2002, 2003).


8

3.1.3 Conclusion

The whole discussion in this section shows clearly that a distinction between CACs and PACs

must be made. Unlike PACs, CACs are much more syntactically integrated with the associated

clause. This integration renders CACs local and positioned within the scope of (the operators

in) the associated clause. Accordingly, CACs are assumed to merge with the matrix clause at

an earlier point in the derivation than PACs. The former is adjoined to VP/vP, while the latter

is merged with a CP.

3.2 Internal syntax: root phenomena and main clause phenomena

It is well known that there exists a range of syntactic phenomena whose application is limited

to root clauses as well as embedded clauses with root properties (Haegeman 2004: 158). For

instance, Emonds (1970), Maki et al (1999), and Heycock (2006), among others, indicated that

English topicalization is one of these phenomena, termed as 'root phenomena' (Emonds 1970,

2000) or 'main clause phenomena' (MCP) (Hooper and Thompson 1973). Haegeman (2002,

2003, 2004, 2009, and 2010) assumed that MCP is not available in CAC. She builds her

argument on the fact that while argument fronting is available in root clauses and PACs, it is

prohibited in CACs. Consider the contrast between the sentences in (15) with CACs and

sentences in (16) with PACs (Haegeman 2002: 74):

(15) a. *Before this book, Mary read, John had already read it.

b. *Before my book, Mary bought, John had already bought yours.

c. *If some of these precautions you take, you will pass the exam.

(16) a. His face not many admired, while his character still fewer felt they could

praise.

b. [He had brought a number of records.] Although some of them I really

enjoyed, others were note very inspiring.

In (15), this book, my book, and some of these precautions are all topicalized in the given CACs,

hence the ungrammaticality of all sentences. However, topicalization does not incur sentence

ungrammaticality when it occurs in PACs as clearly shown in sentences (16) where his

character and some of them are topicalized, respectively. This contrast is also attested in some

other languages, including Japanese (Heycock 2006) and Bulgarian (Krapova 2002).

Following this line of reasoning, it has been advanced that CAC and PAC also differ in

terms of their internal syntax. Such difference in the internal syntax of CAC and PAC can be

illustrated by speaker-oriented epistemic modals and illocutionary Force (in addition to

argument fronting) (Haegeman 2003).

3.2.1 The speaker-oriented epistemic modals

CACs are different from PACs in that the latter may contain expressions of epistemic modality,

which is in principle speaker-related since it expresses the speaker's evaluation regarding the

likelihood of event (as in 17a). However, such expressions cannot be contained within CACs

(as in 17b). See the contrast: (Haegeman 2002: 71)

(17) a. *Mary accepted the invitation without hesitation after John may have accepted it.

b. The ferry will be fairly cheap, while/whereas the plane may/ will probably be too

expensive.


9

3.2.2 Illocutionary force

Following Declerck and Reed (2001), Haegeman (2002 and 2003) stated that PACs have

illocutionary force, whereas CACs do not have independent illocutionary potential, and they

are hence integrated in the speech act conveyed by the associated clause. One piece of evidence

for the availability of illocutionary force in the PACs and its absence in CACs comes mainly

from the observation that PACs may and CACs may not have their own question tags associated

with them (Haegeman 2003: 164). See the contrast in (18):

(18) a. Mary went back to college after/before her children had finished school,

didn’t she?

b. *Mary went back to college after/before her children had finished school,

hadn’t they?

In (18aa) the tag didn’t she is related to the matrix clause; a tag hadn’t they, which would have

to be related to the CAC, after/before her children had finished school, is not possible, hence

the ungrammaticality of sentence (18b). PAP such as, a contrastive while clause, may have its

own tag. Consider sentence (19) (Haegeman 2002: 73).

(19) a. Bill took a degree at Oxford, didn’t he, while his daughter is studying at UCL.

b. Bill took a degree at Oxford, while his daughter is studying at UCL, isn’t she?

Furthermore, it has been attested that PACs not CACs may also contain adjuncts pertaining to

the speech act, such as frankly, as illustrated by the following example:

(20) ‘[A referendum on a united Ireland ]…will be a ‘good thing, because frankly

they need to be taken down a peg and come down to earth and be a little bit more

sober in their approach to things.

Additionally, Verstraete (2002: 146) points out that some PACs may also have imperative force

markers, while CACs may not. Consider the contrast in (21).

(21) a. The fees should bring in more money, because remember we are expecting a

drop in the department funding.

b. *Mary should accept the invitation without hesitation while remember we

tell her the secret.

In conclusion, CACs and PACs are different with regard to their internal syntax. The

former does not have independent illocutionary potential and is integrated in the speech act

conveyed by the associated clause, while the latter does not. In order to account for the

asymmetry between PACs and CACs in terms of argument fronting, epistemic modality, and

illocutionary force, two main approaches have been proposed in the literature: the truncation

approach and the operator movement approach. The main assumptions of these two

assumptions altogether with their theoretical motivation are highlighted in the following

section.

3.2.3 A syntactic account

As stated above, there are two approaches advanced in the literature attempting to account for

the asymmetries between arguments vs. adjunct fronting. These two approaches are the

truncation approach and the operator movement approach.


10

3.2.4 The truncation approach

In this proposal, the reason why no fronting is permitted in CAC is that fronting, e.g.,

topicalization, is related to assertive illocutionary force which is encoded by the functional head

Force in the left periphery (Haegeman 2002). In CACs there is no any assertive illocutionary

force. Such clauses are thus structurally deficient in the sense that their left periphery is reduced.

They lack the functional projection ‘Force’ that encodes assertive illocutionary force. Due to

this deficiency, Topic Phrase and Focus Phrase being dependent on Force Phrase are not

projected either in CACs, and argument fronting is, consequently, ungrammatical (Haegeman

2003: 188). In other words, a constituent affected by a root transformation such as topicalization

and focalization is moved to a particular domain within the peripheral part of a clause. A clause

without such projections (e.g., CAC) cannot offer a landing site for a preposed constituent, and

hence, blocks the relevant transformation (Haegeman 2003, 2006, Munaro 2005, Bocci 2007,

Julien 2008, Nasu 2014).

This approach crucially assumes that there is a distinction between the head which

encodes illocutionary force (i.e., ForceP) and the head which serves simply to subordinate a

clause (i.e., to make it available for categorial selection independently of its force). (Haegeman

2003: 335). The ForceP (in the sense of Rizzi 1997) is split into two different projections: Sub

(a place where the subordinator is positioned) and Force (encoding illocutionary force of the

clause). In CACs, only Sub is available and Force and other projections depending on Force

(i.e., Topic Phrase and Focus Phrase) are truncated, which prevents argument fronting. On the

other hand, in PACs all projections in the left periphery (i.e, Sub, ForceP, TopP, and FocusP)

are available for any fronting, resulting in no restrictions in argument and/or adjunct fronting.

In such clauses, truncation is prohibited because PACs serve as a root clause having its own

assertive illocutionary force. Consider Table 3 (adapted from Haegeman 2003: 335).

Table 3 left periphery of clauses

Clause Type Projections available

Central adverbial Sub>Fin

Peripheral adverbial Sub>Force>Top>Focus>Top>Fin

Root clauses Force>Top>Focus>Top>Fin

On the basis of Table 3, CACs, PACs, and root clauses differ with regard to the projections in

their left periphery. This analysis makes available a straightforward account of the differences

between the temporal adverbial clauses introduced by gablla, gablma ‘before’, and baʕdma

‘after’ are the temporal adverbial clauses introduced by yoom and lamma ‘when’, in that the

former has a truncated left periphery exactly the same as CACs, whereas the latter do not, and

so both adjunct and argument fronting are permitted, as will be explained later in section 5.

3.2.5 The operator movement approach

Under this approach, a subordinate clause resisting a root transformation (such as CAC)

involves movement of an empty operator (Op) to its CP domain. As a result, this Op blocks any

argument fronting because the relevant transformation is ruled out as a minimality violation

(Haegeman 2007, 2010, and Haegeman & Ürögdi 2010). Following this approach, a central

adverbial clause is derived by movement of Op to a clause-initial position. As a result, a

topicalized argument landing in the peripheral position intervenes between the base positions

and the surface positions of the moving operator, giving rise to a minimality violation

(Haegeman 2010). Consider the following sentence and its schematic representation.


11

(22) a. *John left when the office Sheila left.

b. *John left [CP wheni the officej [IP Sheila left tj ti]]

Ø

Here, the topicalized DP the office blocks Op movement. Haegeman (2010) argues that

truncation of CACs’ is a by-product of Op movement. On the other hand, PACs are not derived

by Op movement; hence no restrictions are placed on the argument fronting. The main

advantage of this approach over the truncation one is that the latter cannot account for adjunct

fronting in CACs, if any, as in the example below (adapted from Haegeman 2009:5).

(23) When last year she started to write this column, I thought she would be fine.

The argument is that since the left periphery of CACs is truncated, how the adjunct fronting

can be accounted for, given the fact that in English and in some other languages, adjuncts can

be fronted to the left periphery of CACs. However, based on data from NA, it appears that

truncation approach is better than operator movement approach in accounting for the

discrepancies between the temporal adverbial clauses introduced by gablla, gablma ‘before’,

and baʕdma ‘after’ and the temporal adverbial clauses introduced by yoom and lamma ‘when’.

That is because both adjuncts and arguments are not allowed to show at the CP-domain of the

former, the issue which we are out to argue for in the next section.

4 CAC-PAC dichotomy in NA

Having shown the two prominent approaches most adopted in the related literature in

accounting for the inner structure of adverbial clauses, let us explore NA data. In subsection

5.1 the external syntax of NA temporal adverbial clauses is addressed, assuming that the

division between PAC and CAC is present in NA. In subsection 5.2 the argument is extended

to the internal structure of temporal adverbial clauses in NA, proposing that the truncation

approach is the one, which is valid in accounting for NA data.

4.1 External syntax of temporal adverbial clauses in NA

Indeed, the discrepancy found between gablla, gablma ‘before’, and baʕdma ‘after’ on one

hand, and yoom and lamma ‘when’ on the other provides us with a tangible clue that temporal

adverbial clauses in NA are not the same. Applying some diagnostic tests attested in the

literature for the dichotomy between CAC and PAC with respect to the degree of their syntactic

integration with the associated clause, it turns out that the temporal adverbial clauses introduced

by gablla, gablma ‘before’, and baʕdma ‘after’ are central, whereas the temporal adverbial

clauses introduced by yoom and lamma ‘when’ are peripheral. Prior to applying some of these

tests to corroborating this finding, it should be ascertained that the temporal adverbial clauses

introduced by gablla, gablma ‘before’, and baʕdma ‘after’ are amenable to the generalization

that CAC functions to structure the event which is expressed in the associated clauses, while

the temporal adverbial clauses introduced by yoom and lamma ‘when’ function as a structuring-

discourse device. Consider the following sentence:

(24) Fahd ʔaxδ al-ʕladʒ [gablma yidʒi

Fahd take.3SM.PAST DEF-medicine before come.3SM.PRES

ad-doctoor]

DEF-doctor



12

The temporal adverbial clause introduced by gablma ‘before’ structures the event expressed in

the matrix clause in that it indicates/entails that the event of the associated clause occurs prior

to the event of the temporal adverbial clause. This temporal entailment is actually built in the

meaning of the temporal subordinator gablma ‘before’. In other words, the whole event of the

entire sentence in (24) is divided into two temporal sub-events which are chronological in order.

Using the temporal subordinator gablma ‘before’, the speaker is able to structure these two sub-

events chronologically. The same logic is applied to gablma ‘before’ and baʕdma ‘after’ which

exhibit the same structuring role except for the fact that the latter (i.e., baʕdma ‘after’) entails

the reverse order between the sub-events of the entire sentence containing the adverbial clause

introduced by baʕdma ‘after’. See the following sentence:

(25) [baʕdma kala (V) Fahd (S) wadʒbah bi-l-mațʕam]

After eat.3SM.PAST Fahd meal in-DEF-restaurant

raḥ li-l-mistaʃfa

go.3SM.PAST to-DEF-hospital

‘After Fahd had eaten a meal in the restaurant, he went to the hospital.’

In sentence (25), the speaker explicitly indicates that Fahd had eaten a meal before showing up

to the hospital. The event of the adverbial temporal clause occurs before the event of the main

clause. The role these adverbial temporal subordinators play in structuring the sub-events of the

entire clause can be diagrammed as follows (>>>>>= preceding; <<<<<= following):

(26) [gablma; gablla ‘before’]: Event of matrix clause >>>> Event of temporal

clauses

[baʕdma ‘after’]: Event of matrix clause <<<< Event of temporal clauses

On the other hand, this structuring role is not exhibited when yoom and lamma ‘when’

are used. Although these two subordinators are followed by some events, nothing in the

sentence can entail whether the event of the matrix clause precedes or follows that of the

adverbial temporal clause. In addition, the event of the adverbial temporal clause introduced by

yoom ‘when’ or lamma ‘when’ is in principle intended to structure the discourse. The speaker

using such adverbial temporal clauses attempts to provide some background related to the event

introduced in the matrix clause. This background is discourse in nature since it might be the

context, setting, or even the reason for the event of the matrix clause. Consider the following

sentence:

(27) [yoom yʃtiri Fahd as-sayarah min al-ma ʕra ḍ]

When buy.3SM.PRES Fahd DEF-car from DEF-car show

kant dʒadi:dah

be.S.PAST new


Using the temporal subordinator, yoom ‘when’, the speaker introduces some discourse-related

information about the new car. For example, the car was bought from the car show. Unlike the

temporal clauses introduced by gablla, gablma ‘before’, and baʕdma ‘after’, the speaker does

not want to determine the chronological relation between the sub-events of the entire clause.

Rather, he/she introduces some discourse-related information about the car in that it was new

when it was bought from the car show. The same reasoning holds of the adverbial temporal

clauses introduced by lamma ‘when’. Pertinently, in order to negate the temporal clauses

introduced by gablla, gablma ‘before’, and baʕdma ‘after’, there is no way possible to negate


13

the adverbial clause per se but rather the only way permissible is to negate the event in the

matrix clause. Consider the following sentences:

(28) a. *Fahd ʔaxδ al-ʕladʒ [gablla


ma yidʒi ad-doctoor]

NEG come.PRES DEF-doctor

Intended: ‘Fahd had taken the medicine before the doctor did not come.’

b. Fahd ma-ʔaxδ al-ʕladʒ [gablla

Fahd Neg-take.3SM.PAST DEF-medicine before

ayidʒi ad-doctoor]

come.PRES DEF-doctor

‘Fahd had not taken the medicine before the doctor came.’

As is clear from the ungrammaticality of sentence (28a), it is not possible to negate the

adverbial temporal introduced by gablla ‘before’ since it should be contained within the scope

of the matrix clause. Negating the adverbial temporal introduced by gablla ‘before’ renders the

sentence anomaly with sharp ungrammaticality. The only way possible is to negate the content

of the adverbial temporal clauses introduced by gablma ‘before’ is to negate the matrix clause

whose scope is over the embedded adverbial clause, as in sentence (28b). The same conclusion

can be drawn to gablma ‘before’ and baʕdma ‘after’ alike. Consider the following sentences:

(29) a. *Fahd ʔaxδ al-ʕladʒ [gablma ma

Fahd take.3SM.PAST DEF-medicine before NEG

ayidʒi ad-doctoor]


Intended: ‘Fahd took the medicine before the doctor did not come.’

b. Fahd ma ʔaxδ al-ʕladʒ [gablma

Fahd Neg take.3SM.PAST DEF.medicine before

ayidʒi ad-doctoor]


‘Fahd had not taken the medicine before the doctor came.’

c. *Fahd ʔaxδ al-ʕladʒ [baʕdma

Fahd take.3SM.PAST DEF-medicine after

ma ayidʒi ad-doctoor]

NEG come.PRES DEF-doctor

Intended: ‘Fahd took the medicine after the doctor did not come.’

d. Fahd ma ʔaxδ al-ʕladʒ [baʕdma

Fahd Neg take.3SM.PAST DEF-medicine after

ayidʒi ad-doctoor]


‘Fahd had not taken the medicine after the doctor came.’

However, the adverbial temporal clauses introduced by yoom and lamma ‘when’ can be negated

on their own without the requirement of negating the matrix clause. Even if the matrix clauses


14

were negated, the negation does not have scope over the adverbial temporal clauses introduced

by such subordinators. Consider the following sentences:

(30) a. [yoom ma ʔʃtara Fahd as-sayarah

when NEG buy.3SM.PAST Fahd DEF-car

min al-maʕraḍ] radʒaʕit ʔal-floos

from DEF-car show return.1S.PAST DEF-money

‘When Fahd did not buy the car from the car show, I returned the money.’

b. [yoom ʔʃtara Fahd as-sayarah

when buy.3SM.PAST Fahd DEF-car

min al-maʕraḍ] ma radʒaʕit ʔal-floos

from DEF-car show NEG return.1S.PAST DEF-money

‘When Fahd bought the car from the car show, I did not return the money.’

As is evident in both (30a) and (30b), the adverbial clause fronted by yoom ‘when’ is not

associated with the matrix clause. In (30a), the adverbial clause is negated, whilst the matrix is

not. In (30b), the picture is to the reverse; the adverbial clause is declarative but the matrix is

negated. Furthermore, in (30b) the adverbial clause does not fall within the scope of the

negation as is clear from the translation, i.e., Fahd did actually buy the car from the car show

although the speaker did not return the money. If the negation has a scope over the adverbial

clause, it follows that Fahd did not buy the car from the car show, contrary to fact. The same

observation is held if the adverbial clause is posited to the right of the matrix clause as in (31a)

or the adverbial subordinator is replaced by lamma ‘when’ as in (31b,c).

(31) a. radʒaʕit ʔal-floos [yoom ma ʔʃtara Fahd

return.1S.PAST DEF-money when NEG buy.3SM.PAST Fahd

as-sayarah min al-maʕraḍ]

DEF- car from DEF-car show


b. [lamma ma ʔʃtara Fahd as-sayarah

when NEG buy.3SM.PAST Fahd DEF-car

min al-maʕraḍ] radʒaʕit ʔal-floos

from DEF-car show return.1S.PAST DEF-money


c. [lamma ʔʃtara Fahd as-sayarah

when buy.3SM.PAST Fahd DEF- car

min al-maʕra ḍ] ma radʒaʕit ʔal-floos

from DEF-car show NEG return.1S.PAST DEF-money

‘When Fahd bought the car from the car show, I did not return the money.’

Reasoning along these lines, I assume that the adverbial clauses introduced by gablla, gablma

‘before’, and baʕdma ‘after’ are much integrated with the associated matrix clause than the

adverbial clauses introduced by yoom and lamma ‘when’. An additional argument for the

difference between adverbial temporal clauses in NA can be adduced in reference to the

observation made by Haegeman (2002, and elsewhere) that the expressions of epistemic

modality cannot be contained in CACs. If we apply this observation to temporal adverbial

clauses in NA, it turns out that the adverbial clauses introduced by gablla, gablma ‘before’, and


15

baʕdma ‘after’ cannot contain epistemic expressions, whilst the adverbial clauses introduced

by yoom and lamma ‘when’ allow such expressions. Witness the contrast in the following

sentences (ESP= epistemic):

(32) a. *Fahd ʔaxδ al-ʕladʒ [gablla/galma yimkin

Fahd take.3SM.PAST DEF-medicine before ESP

yidʒi ad-doctoor]


Intended meaning: ‘Fahd took the medicine before the doctor might come.’

b. *Fahd ʔaxδ al-ʕladʒ [baʕdma yimkin

Fahd take.3SM.PAST DEF-medicine after ESP

yidʒi ad-doctoor]


Intended meaning: ‘Fahd took the medicine after the doctor might come’

c. [yoom/lamma yimikin ʃaaf-u ʔʃ -ʃurtˤah as-sayarah

when EPS see.3PM.PAST DEF-police DEF-car

bil-al-maʕraḍ] ʔʃtareena sayarah θanjjeh

in-DEF-car show buy.1P.PAST car second

‘When police might have seen the car in the car show, we bought another car.’

The ungrammaticality of sentences (32a) and (32b) is facia prima evidence for the fact that the

adverbial clauses introduced by gablla, gablma ‘before’, and baʕdma ‘after’ are quite different

from the adverbial clauses introduced by yoom and lamma ‘when’. Additionally, this

discrepancy signals the differences in their syntactic structure in that the former does not

possess a projection dedicated to epistemic modality while the latter does. What is important to

emphasise here is the assumption that the former adverbial clauses are not speaker-oriented

given the fact that epistemic modality is only permitted in clauses with speaker’s orientation

(as explained in the previous section).

Additionally, there is no way possible to coordinate adverbial clauses introduced with

either gablla, gablma ‘before’, and baʕdma ‘after’ with those introduced by yoom and lamma

‘when’. The only way possible as far as coordination of adverbial clauses is combine an

adverbial clause introduced by gablla ‘before’ with one introduced by gablma ‘before’ or

baʕdma ‘after’ and vice versa. The same thing is extended to adverbial clauses introduced by

yoom and lamma ‘when’ which can be combined with one another. These observations are

compelling evidence for the assumption that the adverbial clauses introduced by gablla, gablma

‘before’, and baʕdma ‘after’ are CACs, whilst the adverbial clauses introduced by yoom and

lamma ‘when’ are PACs. In this connection, following the related literature (e.g., Haegeman

2002; 2003), I propose that the former type of adverbial temporal clause functions as an adjunct

to the vP phrase, whereas the latter type enters the derivation as an adjunct to the TP or CP.

This assumption is supported by the fact the adverbial clauses introduced by gablla, gablma

‘before’, and baʕdma ‘after’ are much integrated with the associate clause more than the

adverbial clauses introduced by introduced by yoom and lamma ‘when’.

Having argued for the existence of the dichotomy between CAC and PAC in NA, let us

now examine why the truncation approach has priority over the operator movement approach

in accounting for the derivation of temporal adverbial clauses in NA.


16

4.2 Internal syntax of temporal adverbial clauses in NA

In this section, we argue for the contention that the derivation of the temporal clauses in NA

fares better within the truncation approach. As clearly explained above, within the operator

movement approach, arguments are blocked from fronting to the left periphery of adverbial

clauses not because there is no dedicated positions for them but rather because their movement

incurs locality violations. However, such a disruption (or locality violation) is not triggered in

case of adjunct movement. This being so, we argue that the operator movement approach is not

borne out in accounting for the ban in displacement to the left periphery in NA. That is because

simply both arguments and adjuncts are disallowed to be fronted. If we adopt the operator

movement approach, we are left with no options to account for the ban on adjunct fronting. One

possibility is to propose that adjunct fronting also triggers intervention effects. However, this

assumption is not supported by either theoretical or empirical considerations in cross-linguistic

syntax. Much literature attests that adjuncts movement is subject to different principles as

compared to argument movement, the same argument that has been taken as a conceptual and

empirical ground for the operator movement approach (cf. Danckaert 2012, De Cat 2012,

Yoshimoto 2012, and Authier & Haegeman 2015).

The ban on argument and adjunct fronting in adverbial clauses in NA is best analysed

with reference to the assumption that no dedicated projections for these items are available in

the left periphery of CAC. In line with this assumption, no Force Phrase is projected in NA

CACs. This assumption can be in part supported by the fact that no speaker’s oriented

expressions are permitted in such clauses. The lack of Force Phrase impinges on the existence

of Topic Phrase and Focus Phrase. Since there is no movement whatsoever to the left periphery

of the CAC in NA, it follows directly that these two projections which are discourse-directed

are not available, the assumption which accounts for the ban on argument and adjunct fronting

in NA CAC. Accordingly, CACs derivation in NA is schematically presented as follows:

(33)

Applying this derivation to all examples where the adverbial clauses in NA introduced by

gablla, gablma ‘before’, and baʕdma ‘after’, we end up with an account for the ban on argument

and adjunct fronting. For instance, the derivation of the temporal adverbial clause in the

ungrammatical sentence below is schematically represented in (35):

(34) *[gablma al-ʕamliah yswe-ha Fahd bi-l-mistaʃfa]

Before DEF-operation do.3SM.PRES-it Fahd in-DEF-hospital

ʔaxδ al-ʕladʒ


‘Before Fahd had the operation in the hospital, he had taken the medicine.’

Subordinator

Fin

T

DP V

VP

TP

FinP

SUB


17

(35)

It is impossible to move the direct object to the left periphery of NA CAC since there is no any

dedicated position in the left periphery in which the displaced DP can be accommodated in.

Additionally, the adjunct movement is also non-permitted for the same reason. Consider the

following ungrammatical sentence with its derivation.

(36) *[gablma bi-l-mistaʃfa yswe-ha Fahd al-ʕamliah]

Before in-DEF- hospital do.3SM.PRES-it Fahd DEF-operation

ʔaxδ al-ʕladʒ


‘Before Fahd had the operation in the hospital, he had taken the medicine.’

(37)

The same reasoning extends to other adverbial clauses introduced by gablla ‘before’ and

baʕdma ‘after’. Under this approach, there is no need to assume that subordinators are moved

gablma

Fin

yswe-ha

T

Fahd

DP

v

V

al?amliah

DP

VP

bi ilmista?fa

PP

VP

v'

vP

TP

FinP

SUB

gablma

Fin

yswe-ha

T

Fahd

DP

v

V

al?amliah

DP

VP

bi ilmista?fa

PP

VP

v'

vP

TP

FinP

SUB


18

from downwards but instead they are externally merged in Sub position dedicated to them. This

stipulation on their first merge of Sub goes hand in hand with the minimalist spirit on movement

which is only triggered when needed. However, even if CACs are derived in NA via operator

movement, the ban on adjunct and agreement fronting is only ascribed to the deficiency of the

left periphery of such clauses.

What is worth mentioning at this point is the observation that in NA CAC, the subject

is also unpermitted to move to Spec TP:

(38)

As indicated above, the only word order allowed in adverbial clauses introduced by gablla,

gablma ‘before’, and baʕdma ‘after’ is VS. The possibility that the verb moves to Focusº or

even Topicº should be dispensed with, given first that there is no any discourse-related reading

obtained for the verb and second following our line of investigating Focus Phrase or Topic

Phrase are not projected. The apparent conclusion is that the subject remains in situ in NA CAC.

This can be accounted for in assuming that the Spec of TP in NA CAC is also unavailable. Non-

availability of Spec of TP might be related to the lack of EPP in such clauses. In addition to the

fact that the thematic subject must show up to the right of the verb, no expletive element is

allowed to occur in such clauses. A different view that we can make is that the truncation of the

CP has a negative consequence on the EPP in that it is not triggered; however, I leave this issue

open pending further research.

On the other hand, the adverbial clauses introduced by yoom and lamma ‘when’ have a

full-fledged CP so both argument and adjuncts can be fronted. In addition, since the Force

Phrase is present in these clauses, speaker’s oriented expressions are valid. Furthermore, CP

being non-truncated and can be split when needed, no consequences on EPP are made, hence

both VSO and SVO are permitted in neutral-discourse occasions. In relation to this, the

sentence below where the direct object is fronted has the schematically representation in (40):

(39) [yoom as-sayrah (O) yʃtiri-ha (V) Fahd (S) min al-maʕraḍ]

When DEF-car buy.3SM.PRES-it Fahd from DEF-car show

kant dʒadi:dah

be.S.PAST new


Subordinator

Fin

T

Subject

v VP

v'

vP

T'

TP

FinP

SUB


19

(40)

In addition, in these clauses the subject can appear to the left of the verb, occupying Spec of

TP.

5 Conclusion

This research tackled the derivation of temporal adverbial clauses in NA with special reference

to argument and adjunct fronting. Firstly, it concluded that NA exhibits the dichotomy between

CAC and PAC like other languages such as English and Italian. In particular, it argued for the

notion that the CAC is much integrated within the syntax of the associated matrix clauses, the

issue that has certain consequences in their interpretation. On the other hand, NA PAC, as it

has been confirmed in cross-linguistic syntax, is less integrated with a unique interpretation and

scope. Both negation and epistemic modality were taken as diagnostics of bearing out this

assumption. The adverbial clauses introduced by gablla, gablma ‘before’, and baʕdma ‘after’

are CAC whereas the adverbial clauses, introduced by yoom and lamma ‘when’ are PAC. As

for the internal syntax of adverbial clauses in NA, it has been advanced that CAC has a truncated

left periphery in that no Force Phrase, Topic Phrase, nor Focus Phrase are projected. The left

periphery of such clauses contains only Sub position where the subordinators are situated

alongside Fin Phrase. The poverty of the discourse/speaker-related projections in NA CACs

precludes the use of epistemic modality. In a related vein, the deficiency of the CP of such

clauses makes it impossible for arguments and adjuncts alike to be fronted since there are no

dedicated projection for them. Thus, the current research does away with the operator

movement approach and adopts instead the truncation approach to account for the reason why

arguments and adjuncts resist fronting. Unlike CAC, NA PAC has a fully-articulated CP where

both argument and adjuncts can be displaced to the left periphery without imposing any

restriction, a result which on its own corroborates our main finding that operator movement

approach is not valid at least with regard to NA.

6 References

Authier, J., & Haegeman, L. (2015). ‘French adverbial clauses, rescue by ellipsis and the

truncation versus intervention debate.’ Probus, 27(1), 33-71.

yoom

assayrah

Top

Fin

y?tiriha

Fahd

DP

v

V DP

VP

min al-ma?ra?

PP

VP

v'

vP

TP

FinP

Top'

TopP

SUB


20

Benmamoun, E., & Choueiri, L. (2013). The syntax of Arabic from a generative perspective.

The Oxford Handbook of Arabic Linguistics, 115.

Bocci, G. (2007). ‘Criterial positions and left periphery in Italian: Evidence for the syntactic

encoding of contrastive focus.’ Nanzan Linguistics: Special Issue 3(1): 35-70.

Chomsky, N. (1993) A Minimalist Program for Linguistic Theory. In Keyser, K. H. a. S. J. (ed.)

The View from Building 20: Essays in Linguistics in Honor of Sylvain Bromberger.

Cambridge, MA: MIT Press. 1-52.

Chomsky, N. (1995) The minimalist program. Cambridge, Mass: MIT Press.

Chomsky, N. (2000) Minimalist Inquiries: The Framework. In Lasnik, H., R. Martin, D.

Michaels & J. Uriagereka (eds.) Step by step. Essays on minimalist syntax in honour of

Howard Lasnik. Cambridge. Mass: MIT Press.

Chomsky, N. (2001) Derivation by phase. In Kenstowicz, M. (ed.) Ken Hale. A Life in

Language. Cambridge, Mass: MIT Press. 1-52. .

Cinque, G. (2004). ‘Issues in adverbial syntax.’ Lingua, 114(6), 683-710.

Danckaert, L. (2011) On the left periphery of Latin embedded clauses. Ghent University.

Danckaert, L. (2012). Latin embedded clauses: the left periphery (Vol. 184). John Benjamins

Publishing.

De Cat, C. (2012). Towards an interface definition of root phenomena. Main Clause

Phenomena, 135-158.

Declerck, R. & Reed, S. (2001). Conditionals: a Comprehensive Empirical Analysis. Berlin:

Mouton de Gruyter

Demirdache, H., Uribe-Etxebarria, M., (2004). The syntax of time adverbs. In: Gue´ron, J.,

Lecarme, J. (Eds.), The Syntax of Time. MIT Press, Cambridge, MA, pp. 143–180.

Emonds, J. E. (1970). Root and structure-preserving transformations (Doctoral dissertation,

Massachusetts Institute of Technology).

Emonds, J. (2000). Unspecified categories as the key to root constructions. In Peripheries (pp.

75-120). Springer Netherlands.

Endo, Y., & Haegeman, L. (2014). Adverbial concord: mergind adverbial clauses. In Formal

approaches to Japanese Linguistics 7 (Vol. 73, pp. 25-44). Cambridge University Press.

Frey, W. (2012). On two types of adverbial clauses allowing mot-phenomena. Main Clause

Phenomena: New Horizons, 190, 405.

Haegeman, L. (2002). Anchoring to Speaker, Adverbial Clauses and the Structure of CP. In

Simon Mauck & Jenny Mittelstaedt, eds., Georgetown University Working Papers in

Theoretical Linguistics 2, pp. 117-180.

Haegeman, L. (2003). The syntax of adverbial clauses and its consequences for topicalisation.

Antwerp Papers in Linguistics. 107. Current Studies in Comparative Romance

Linguistics., p-61.

Haegeman, L. (2004). The syntax of adverbial clauses and its consequences for topicalisation.

Antwerp Papers in Linguistics. 107. Current Studies in Comparative Romance

Linguistics., p-61.

Haegeman, L. (2006). ‘Conditionals, factives and the left periphery’. Lingua, 116(10), 1651-

1669

Haegeman, L. (2007). ‘Operator movement and topicalization in adverbial clauses’. Folia

Linguistica., 18: 485-502.

Haegeman, L. 2009. Main clause phenomena and the derivation of adverbial clauses. In

Proceedings of the 18th international symposium on theoretical and applied linguistics

(ISTHAL), ed. Anastasios Tsangalidis (Vol. 1418).

Haegeman, L. (2010) 'The internal syntax of adverbial clauses', Lingua, 120(3), pp. 628-648.

Haegeman, L. and Ürögdi, B. 2010. Referential CPs and DPs: An operator movement account.

Theoretical Linguistics 36: 111-152.


21

Hengeveld, K. (1996) 'The internal structure of adverbial clauses', Functional Grammar Series,

pp. 119-148.

Heycock, C. (2006). Embedded root phenomena. The Blackwell companion to syntax, 174-209.

Hooper, J. B., & Thompson, S. A. (1973). ‘On the applicability of root transformations’.

Linguistic inquiry, 465-497.

Ingham, B. (1994). Najdi Arabic: Central Arabian (Vol. 1). John Benjamins Publishing.

Julien, M. (2008). ‘Embedded V2 in Norwegian and Swedish.’ Working Papers in

Scandinavian Syntax 80: 103-161.

Koeneman, O., & Zeijlstra, H. (2014). ‘The rich agreement hypothesis rehabilitated’. Linguistic

Inquiry.

Krapova, I. (2002). ‘On the Left Periphery of the Bulgarian sentence’. Working Papers in

Linguistics, 12, 107-128.

Maki, H., Kaiser, L., & Ochi, M. (1999). ‘Embedded topicalization in English and Japanese’.

Lingua, 109(1), 1-14.

Munaro, N., (2005). Computational puzzles of conditional clause preposing. In: di Sciullo, A.-

M. (Ed.),UG and External Systems. Language, Brain and Computation. John

Benjamins, Amsterdam/Philadelphia, pp. 73–94.

Nasu, N. (2014). ‘Topicalization in adverbial clauses: toward parameterization of embedded

root effects.’ Journal of foreign studies, 64(1), 61-84.

Nilsen, O., (2004). Domains for adverbs. In: Alexiadou, A. (Ed.), Adverbs across Frameworks,

Lingua 114, pp. 809–847.

Rizzi, L. (1997) 'The fine structure of the left periphery', in Elements of grammar. Springer, pp.

281-337.

Verstraete, J. (2002). Interpersonal Grammar and Clause Combining in English. Ph.D. diss,

University of Leuven.

Yoshimoto, K. (2012). ‘The left periphery of CP phases in Japanese.’ Acta Linguistica

Hungarica, 59(3), 339-384.


22

Ahmed Alshammari

School of English Literature, Language and Linguistics


Newcastle upon Tyne

Tyne and Wear, NE1 7RU

United Kingdom

[email protected]


Newcastle and Northumbria Working Papers in Linguistics 22 (2016) Alshamari

23

IN FAVOR OF CONTRASTIVE TOPIC PHRASE IN NORTH HAIL ARABIC LEFT

PERIPHERY: EVIDENCE FROM THE DISCOURSE PARTICLE TARA

Murdhy Alshamari

(Newcastle University) 1

Abstract This research investigates one discourse-related particle, namely tara in North Hail Arabic. It argues

that tara is a C-related particle heading the Contrastive Topic Phrase in the sense of Frascarelli and

Hinterhölzl (2007). tara enters the derivation endowed with an interpretable/unvalued [TOP] feature in

the sense of Pesetsky and Torrego (2007) in addition to a set of uninterpretable/unvalued φ-features.

These features are valued by the element that expresses the contrastive topic of the clause, which, as a

result of maintaining an Agree relation with tara, it bears a contrastive stress. Evidence for this

assumption comes from the fact that where the entity expressing contrastive topic is a DP, a clitic with

the φ-features of this DP is spelled out on tara. The significant point the study advocates is that such

valuation occurs while the entity expressing contrastive topic is in situ, through a probe-goal

configuration (Chomsky 2000, 2001, 2008). As such, the relevant entity is not triggered to move to the

Spec position of Contrastive Topic Phrase.

1 Introduction

The issue of C-particles in general and their role in determining the inner structure has recently

been one of the main concerns of the current syntactic investigation (Rizzi 1997, Cinque 1999,

2002, 2006, Belletti 2004, and Cinque and Rizzi 2008; Biberauer and Sheehan 2011;

Struckmeier 2014; Hack 2014; Biberauer et al. 2014 ). Hence, research in this specific domain,

which has largely been enriched from works on natural languages, has made available in-depth

insights into the structure of the left periphery and the interaction between C-domain and T/V-

domains. The present paper is seen as a continuation of this debate on the syntax of discourse

domain by exploring one C-particle used in North Hail Arabic (henceforth, NHA) within the

recent assumptions of the minimalist framework (Chomsky 2000, 2001, 2008).2 To this end, I

first introduce the relevant theoretical assumptions, namely Rizzi’s (1997) fine structure of the

left periphery and the topics typology of Frascarelli & Hinterhölzl (2007). Next, I introduce

some basic facts on tara’s syntactic behavior. Then, I analyze these facts within Chomsky’s

(2000, 2001, 2008) probe-goal mechanism of Agree. The major assumption is that tara is a

particle heading the Contrastive Topic Phrase and agrees with the element expressing a

contrastive topic while both are in situ.

2 Theoretical background

In this section, I introduce the basic theoretical assumptions my analysis for tara is based on.

These are Rizzi’s (1997) fine structure of the left periphery and Frascarelli and Hinterhölzl’s

(2007) topics typology.

2.1 Rizzi’s (1997) fine structure of the left periphery

Since the seminal work of Rizzi (1997) on the fine structure of the left periphery, many works

have addressed the inner structure of CP-domain in natural languages. One main advancement

1 I am grateful to my supervisors, Anders Holmberg and Geoffrey Poole, for their continuous support, valuable

comments and feedback. I would like to thank Ian Roberts, željko Bošković and Theresa Biberauer for the valuable

and helpful discussions we had at the ‘Rethinking Verb Second Conference’ and the ‘Rethinking Comparative

Syntax Conference’ at the University of Cambridge 2016. Special thanks are extended to the anonymous reviewer

at the NNWP for raising significant comments. 2 NHA is spoken in Saudi Arabia. See Alshamari and Jarrah (2016) and Alshamari (2015a,b,c) for works on this

dialect.

Alshamari Contrastive topic phrase in North Hail Arabic left periphery

24

of such studies is the appreciation of discourse role in sentence derivation. For Rizzi (1997),

what was already known as a CP is a fertile domain with several inner layers; each of them is

dedicated for a specific function. These projections include the following: Force Phrase, which

is the highest projection, Topic Phrase, Focus Phrase, and Finiteness Phrase which is, in turn,

the lowest projection in this hierarchy. See Figure 1.3

Figure 1: CP’s articulated inner structure

The main difference raised by Rizzi (1997) between the higher TopicP and the lower

one is that the former is not recursive, while the latter is recursive, giving multiple realizations

of Topics. On the other hand, although predicting that the new articulated CP includes several

Topic Phrases, Rizzi (1997) failed to consider the fact that topics encode different interpretive

properties, the matter I discuss in the next subsection.

2.2 Topics typology: Frascarelli and Hinterhölzl (2007)

Frascarelli & Hinterhölzl (2007) argue convincingly that there are at least three types of topics

which must be distinguished, namely Aboutness Topic, Contrastive Topic and Familiar Topic.

They propose that each type of Topic Phrase is associated with a distinct interpretive meaning.

Evidence for these three topics is adduced from both syntax and specific intonational contours.

The current study argues that these topics, with emphasis on the contrastive topic, can also be

detected through a set of discourse particles, merged in the left periphery. Let us first inroduce

definition for these three topics. Frascarelli and Hinterhölzl (2007) define the three topics as

follows:

i. Shifting Topic (or Aboutness-shift Topic, in the sense of Frascarelli 2008) (henceforth, S-

topic): The constituent referring to an entity which a sentence is about, and which is newly

introduced into the ongoing conversation (i.e., a new topic distinct from the preceding topic

which the ongoing conversation was about). Similarly, in situations where the ongoing

discussion is about one topic, but, for some reason, the conversation digresses from this topic

to a different one, the speaker may return to the original topic by means of an S-Topic. To

appreciate this point, consider the followiing dialogue:

(1) A: ʔafdˤal waqt li-l-qiraʔah ʔi-sˤubuħ

best time for-Def-reading Def-morning

The best time for reading is morning.’

3 * is recursive.


25

B: bus mumkin ma juku:n ʔal-waqt ʔal-ʔafdˤal ʔiða kint sahran

but migh Neg be.it Def-time Def-best if be.PST stayed.up

xusˤu:sˤan ʔiða kint laʕib mubarat qadam raħ tku:n

especially if be.PST played game football FUT be.FUT

taʕban min bukra

tired from tomorrow

‘But it might not be the best time if you stayed up the night, especially if you had already

played a football game. You will be tired the following day.’

C: laʔabt mubarat qabul ʔams w li-l-ħi:n taʕban ʔal-muʃkilah

Played.I game before yesterday and till-Def-now tired.I Def-problem

in-i kil ʕisbu:ʕ ʔalʕab θalaθ mubarayat

that-I every week play.I three games

‘I played a game the day before yesterday, and I am still tired. The problem is that I play

three games a week.’

A: binisbah li-ʔafdˤal waqt ʔil-qiraʔah ʔadˤin kil waħid l-uh

As for- best time for-Def-reading think.I everyone to-him

barnamadʒ-uh ʔal-muafadˤal

schedule-his Def-favorite

‘As for the best time for reading, I think everyone has his own favorite time.’

As clear in (1), the conversation was about the best time for reading, being the morning

time. As the conversation proceeded, the speakers diverted away from the main topic of the

conversation. As seen in Speaker A’s last utterance, the speaker shifted the conversation to the

main topic again, by means of the particle binisbali, the point at which the Arabic DP ‘the time

of reading’ expresses an S-Topic.

ii. Contrastive Topic (henceforth, C-topic): The entity which signals other topics with

which this topic creates oppositional pairs; this topic needs to be opposed to some other

topics. This type of topic is phonologically distinct with contrastive stress.4 Consider the

following dialogue:

(2) A: ʃlaon maharat Firas bi-l-malʕab

How skills Firas in-Def-field

‘How skillful is firas on the (football) field.’

B: MUNAWALAT-UH mumtazah

Passes-his great

‘HIS PASSES are great.’

(3) A: ʃ-rayak ib-di:n Salim

Q-opinion-you in-religious commitments Salim

‘How do you see Salim’s commitments to religious principles?’

B: SˤALAT-UH zi:nah

Prayer-his good

‘HIS PRAYER is good.’

4 Krifka (2007) argues that C-Topics are conceptually licensed through the fact that some utterances do not deliver

all the information that is expected by the speaker but rather deliver a limited or incomplete information that

diverges from the expectations of the speaker.


26

Speaker A in both dialogues asks about a particular entity related to some person. Upon

this, Speaker B creates a conversational common ground from which he selects one entity,

something that he considers the most relevant aspect of Speaker A’s inqiury. In (2B), for

instance, passing is selected out of a closed set of skills, all of which are familiar to both

speakers. Speaker B, then, could have selected dribbling, heading or even running, and Speaker

A would accept each because they are all members of a closed set of alternatives that are

implicitly understood, i.e, football skills.

iii. Familiar Topic (henceforth, F-topic): The entity which is textually given and

d(iscourse)-linked with the S-topic established in the conversation. Of note here is that

F-topics are typically realized as pronouns, given that they are to accessible through

the dialogue (cf. Givon 1983, Chafe 1987, and Pesetsky 1987). Unlike S-topics and C-

topics, F-Topics are recursive with a possibility thus to have many realizations of F-

topic per clause. Consider the following dialogues:

(4) A: Firas jiħib Barcelona w muʕdʒab ib-Messi w Nymar

Firas love.he Barcelona and like.he in-Messi and Nymar

‘Firas loves Barcelona and he likes Messi and Nymar the most.’

B: ʃae dʒami:l mumkin jiktisib min maharathum xusˤu:sˤan

Thing good might aquire.he from skills.thier especially

w in-uh sˤiʁi:r ‘

and that-he young

‘Such a good thing. He might aquire some of their skills, as he is so young.’

C: jtabiʕ-hum kil yaum ib-adʒizt-uh hata w hu bi-l- ɡatˤar

Watch.3SG.M-them every day in-gadgets even and he on-metro

‘He watches them every day with his gadgets; even while on metro.’

The conversation is about two entities, Firas and the two players: Messi and Neymar,

both expressing the F-Topics of the sentence. As we can see, both entities are mentioned in

the tail of the conversation and are accessible via an (non-contrastive) pronoun, for neither

was selected out of a set of members (a C-Topic), nor was returned to due to conversation

divergence (S-Topic).

Examining Italian and German left periphery, Frascarelli and Hinterhölzl (2007) point

out that these different types of topics are hierarchically ordered in a fixed fashion, as seen

in Figure 3 (adapted from Frascarelli and Hinterhölzl 2007: 89):


27

Figure 2: Frascarelli and Hinterhölzl’s (2007) topics typology

Further investigation on topic typology reveals certain restrictions on their occurrence

in some contexts. For instance, it has been argued that an S-topic can only be allowed in root

clauses, while C-topic and F-topic can occur in root and embedded clauses (Bianchi and

Frascarelli 2010).

Having introduced the main theoretical ssumptions the current research is based on, let

us now intrdouce the basic descriptive facts on the C-partcile tara in NHA.

3 Descriptive facts about the C-related particle tara.

Recent proposals view discourse particles as functional heads that are merged in fixed positions

and have the effect that they change the interpretation of the proposition expressed by the clause

(Biberauer et al 2014; Coniglio 2008; Zimmermann 2004). As for tara, I argue that this particle

is merged in the C-domain of the associated clause, colouring it with a topic-comment

interpretation. For this assumption, let’s look into conceptual and empirical evidence.

In the first place, tara occurs clause-initially, taking wide scope over the whole

clause,5 as in (5) below.

(5) tara Ali ʃAF as-sayarah

Prt Ali see.PST.3SG.M Def-car

‘Ali SAW the car.’

In (5), tara appears clause-initially and takes wide scope over the whole clause, having some

discourse interpretive effects over the proposition expressed by the clause. The contribution of

tara to the clause it is merged with is determined by the interpretation of the element that tara

marks. In this respect in (5), what tara (marks/ singles out/ stresses/ selects) is the verb ʃaf

‘saw’, which, as can be clearly seen, is the entity that bears the contrastive stress. With tara as

part of th numeration of (5), (5) is predicated about the action expressed. In syntax, this is

achieved by tara’s discharging its contrastive stress to the verb. Hence, the verb is the C-Topic

of the sentence. The interpretation of (5), then, can be conceived of as Ali saw the car (he didn’t

imagine it), where both the acts of seeing and imagining are entities that are available in the

5 The unmarked clause following tara can also be VSO. In both cases, the interpretation of the clause depends on

what element of the clause bears the contrastive stress tara assigns. In (5) it is the verb, where the topicalized

element of the clause is the action expressed by the verb.


28

conversational common ground (Bianchi & Frascarelli 2010; Büring 2003) of the given

discourse, and are accessible to the coversation interlocutors, but the former is selected over the

latter.

Moreover, in cases where the element that expresses C-Topic is a DP, it bears a

contrastive stress, and a clitic with the φ-features of this element is spelled out on tara.6.

Consider the following sentences:

(6) a. tara-ah ʔAL-BINT ʃaf-at ʔal-walad

Prt-3SG.F Def-girl see.PST.3SG-F Def-boy

‘THE GIRL saw the boy.’

b. tara-ah ʃaf-at ʔAL-BINT ʔal-walad

Prt-3SG.F see.PST.3SG-F Def-girl Def-boy

‘THE GIRL saw the boy.’

c. tara-h ʔAL-WALAD ʃaf ʔal-bint

Prt-3SG.M Def-boy see.PST.3SG.M Def-girl

‘THE BOY saw the girl.’

d. tara-h ʃaf ʔAL-WALAD ʔal-bint

Prt-3SG.M see.PST.3SG.M Def-boy Def-girl

‘THE BOY saw the girl.’

In (6), a clitic agreeing with the subject is spelled out on tara, regardless of the position

of the subject in relation to the verb (compare (6a,c) with (6b,d)). A possible context in which

(6a,b), for instance, occur is where the speaker emphasises that the girl not, say, her father,

mother, or sister , who are all available in the conversational common ground of the given

discourse, saw the boy. Hence, creating a set of alternatives (Frascarelli and Hinterhölzl 2007).

In addition to the contrastive stress that must accompany the element that tara marks, this

element agrees with the clitic appearing on tara. For instance, the subject in (6a,b), ʔalbint ‘the

girl’ agrees with the clitic ah, which is specified as [3SG.F], for which an initial observation is

that the clitic h spelled out on tara in (6c,d), is specified as [3SG.M] and agrees with the subject

ʔalwalad ‘the boy’. Under these charactersitics of tara with respect to the associated clause and

clause arguments, I propose that the clitic appearing on tara is a spell out of the unvalued φ-

features of tara, as a result of an Agree relation being established between tara and the (DP)

element expressing a C-Topic. What supports this thesis can be captured by (7) below, where

the clitic spelled out on tara agrees with the direct object (in such cases, the direct object must

be positioned to the left of the subject):

(7) a. tara-h ʃaf-at-uh ʔAL-WALAD ʔal-bint

Prt-3SG.M see.PST.3SG-F-3SG.M Def-boy Def-girl

‘THE BOY, the girl saw him.’

b. tara-ah ʃaf-ah ʔAL-BINT ʔal-walad

Prt-3SG.F see.PST.3SG.M-3SG.F Def-girl Def-boy

‘THE GIRL, the boy saw her.’

6 It must be pointed out that in all cases where tara is merged, in addition to the fact that tara is endowed with

contrastive stress and has to discharge it to one and only one element, be it a DP, V, PP, Adv, in case this entity is

a DP, tara is suffixed by a clitic that agrees with this DP.


29

In (7a), the direct objec ʔalwalad ‘the boy’ is the C-Topic; it is the element that is being

selected out of a set of alternative entities all of which are availabe in the given discourse, on

the grounds that the speaker asserts that who the girl saw was the boy, rather than the lady, the

man, the child, each of which is an entity established in the relevant discourse and is a part of

the common ground of the sentence constituting a closed set. The same analysis extends to (7b),

where the the direct object is ʔalbint ‘the girl’.

The last synatctic property of tara here is that, in addition to (5) above where tara

appears without any clitic spelled out on it, in which the C-Topic is the verb of the clause, tara

marks adjuncts (as in (8) below), with the interpretation that what is being C-Topicalized is the

temporal adverb taw ‘just now’, whose referent is being selected over any otherwise proposed

time referents at which Ali performed the act of seeing the car.

(8) tara TAW ʃaf Ali as-sayarah

Prt just now see.PST.3SG.M Ali Def-car

‘Ali JUST saw the car (not long before).’

The discussion thus far gives rise to the status of tara. With its morpho-syntactic

properties on the one hand and its semantic and discourse contribution to the interprtation of

the associated clause on the other hand, I argue that this C-particle is a head. Two pieces of

evidence support this contention. Firstly, as we have seen, tara hosts clitics, a property well-

known of heads (Chomsky 1995; Aoun et al. 2010). Another piece of evidence for this

contention comes from the fact that the verb is barred from appearing to the left of tara.

Consider the following ungrammatical sentence, followed by its schematic representation:

(9) a. *ʃaf tara Ali as-sayarah

see.PST.3SG.M Prt Ali Def-car

Intended meaning: ‘SEEING THE CAR, Ali did.’

b. [CP ʃaf tara [TP Ali T <ʃaf> [vp <ʃaf> [VP <ʃaf> assayarah].

The ungrammaticality of (9) is straightforwardly accounted for if tara is taken as a head,

causing an intervention effect against verb movement to a position to the left of tara. The verb

(or actually Tº in this case) being blocked from moving to the left of tara is captured by the

assumption that tara blocks T movement to any position to the left of it by virtue of the

Relativized Minimality Principle (cf. Rizzi 1990; see also Rizzi 2004, Friedmann et al 2009

and Boeckx 2009). T being a head cannot move across tara where this blockage is attributed to

the fact that the C-Topic head position is filled by tara, and, if T needs to move across it, T has

to first land in C-Top head position.

The interim conclusion is that tara is a C-Topic particle, heading the C-TopicP. All the

discussion below is dedicated to defend this hypothesis. One relevent point here is the account

of the clitic spelled out on tara when the entity expressing C-Topic is either the subject DP or

the direct object DP. I argue that this clitic is a reflex of an Agree relation being established

between tara and the element expressing C-Topic. Let us here make recourse to the recent

approach to Agree, the Probe-Goal approach of Chomsky (2000, 2001 and 2008), where

agreement is held when a probe, carrying some semantically uninterpretable and lexically

unvalued formal features, searches in its visible c-command domain for a matching active goal

carrying a semantically matching interpretable and lexically valued feature. Chomsky

(2001:122) set forth specific conditions on the Agree relation between the searching probe and

the matching goal. These conditions are mentioned below:


30

(10) The probe α agrees with the goal ß provided that:

a- α has uninterpretable φ-features.

b- ß has matchiinterpretable φ features.

c- ß is active by virtue of having an unvalued Case feature.

d- α c-commands ß.

e- There is no potential goal ϒ intervening between α and ß.

Let us now show clearly how this valuation occurs.

4 Analysis

In this section, I argue that tara is a C-Topic particle, heading C-TopicP, positioned above F-

TopicP and below S-TopicP, following Frascarelli and Hinterhölzl (2007). I propose that the

element which tara agrees with functions as the C-Topic, expresing the entity selected from an

exhaustive set of entities which are all familiar and available in the common ground. In order

to determine the existence and the position of the C-Topic Phrase in the left periphery of NHA,

it is imperative to highlight the main properties of tara (morphological, phonological and

syntactic), taking them as evidence of tara being a C-Topic particle, heading C-TopicP. First

and foremost, we need to provide evidence that tara is not a focus particle, despite the fact that

the element with which the clitic appearing on tara agrees has contrastive stress.

4.1 tara is not a Focus particle

An initial observation in support of the assumption that tara is not a focus particle is that, in the

first place, it cannot bear a clitic agreeing with an indefinite entity.7 tara only bears a clitic

agreeing with an entity that is specific, discourse-given and familiar to all the conversation

interlocutors, as seen above. Consider the following sentence:

(11) *tara walad ʃaf as-sayarah

Prt boy see.PST.3SG.M Def-car

Intended meaning: ‘A BOY saw the car.’

Another piece of evidence in favour of the claim that tara is not a Focus particle comes

from cliticization. In clauses containing focalized elements, tara does not bear a clitic with the

same φ-features of the focalized element. Hence, tara does not agree with it (12b). Rather, it is

cliticized with that of a non-focalized element (12a):

(12) a.wiʃi tara-h Ali ʃaf as-sayarahi

what prt-3SG.M Ali see.PST.3SG.M Def-car

‘What did ALI see?’

b. *wiʃi tara-ah Ali ʃaf as-sayarahi

what prt-3SG.F Ali see.PST.3SG.M Def-car

Intended meaning: ‘WHAT did Ali see?’

Sentence (12a) is grammatical because the clitic on tara agrees with the sucject DP Ali (which

is not being focalized), where this agreement is captured by the realization of the clitic h being

spelled out on tara, specified as [3SG.M] and agrees with Ali. In (12b), then, it follows that the

7 See Elghamry (2004), Danon (2008), Fassi Fehri (2012), and Jarrah & Zibin (2016) on definiteness in Arabic.


31

ungrammaticality is attributed to the fact that the clitic on tara agrees with the focalized object,

assayarah ‘the car’ where the clitic is ah, being specified as [3SG.F]. The same logic holds true

of clef-constructions, as in (13) below, where tara agrees with the DP Ali, resulting in the

ungrammaticality of the sentence:

(13) *Alii illi tara-hi ʃaf as-sayarah

Ali COMP prt-3SG.M see.PST.3SG.M Def-car

‘It’s ALI who saw the car.’

All these facts militate aginst the assumption that tara is a focus particle. This leaves us

with the other concept of discourse partitioning, Topic.

4.2 tara is a Topic particle

Having argued that tara cannot be counted as a focus particle and seems to behave as a topic

particle, let’s see the category of topic under which tara falls.

4.2.1 tara is not an S-Topic particle

Being a Topic particle, and occurring clause-initially after all, one might suggest that tara

serves as an S-Topic particle. However, when using tara, the speaker selects an entity out of a

set of entities, all of which are familiar to the speakers. In this way, the element marked by tara

exhibits a membership set, and it is singled out of a group of entities to which it is opposed

(Frascarelli and Hinterhölzl 2007; Frascarelli 2008; Bianchi & Frascarelli 2010; Büring 2003).

Consider the following dialougue:

(14)

Speaker A: ʔaxu:-i qal ʔinn-uh waħid min

brother-my say.PST.3SG.M that-3SG one of

ʔahal-i ʃaf l-bint

Familiy-my see.PST.3SG.M DEF-girl

’My brother said that one of my family members had seen the girl.’

Speaker B: tara-h OMAR ʃaf l-bint.

Prt-3SGM OMAR see.PST.3SG.M Def-girl

‘OMAR saw the girl.’

Speaker A creates a set of members belonging to one closed group, his family. Having

created this closed set of entities, all of which are familiar to Speaker A and Speaker B, the

latter selects one member (from the closd set, i.e., his family members) who has seen the

woman, and the entity expressing the selected member, Omar, is marked by tara (i.e., having

a clitic spelled out on on tara). Moreover, a common characteristic that S-Topic and C-Topic

share is that they do not allow multiple realizations; they are non-recursive (Frascarelli and

Hinterhölzl 2007 and Bianchi & Frascarelli 2010). Reasoning along these lines, tara can be

preceded by a constituent, which might suggest that this element is an S-Topic, hence, tara is

not an S-Topic particle. Evidence corroborating this hypothesis comes from clauses containig

S-Topic particles (like binisbali ‘as for’) that co-occurs with tara. In such clauses, the element

preceding tara is introduced by the S-Topic particle, rather than by tara. Consider the following

sentences:

(15) a. binisba-li- Ali, tara-ah ʔAL-MUSABAGAH ʔintahat.

Prt- Ali Prt-3SG.F DEF-COMPETITION.3SG.F ended.it

‘As for Ali, the COMPETITION has ended.’


32

b. *binisba-li- Ali, tara-h ʔAL-MUSABAGAH ʔintahat.

Prt- Ali Prt-3SG.M DEF-COMPETETION.3SG.F ended.it

Intended meaning: ‘As for Ali, THE COMPETITION has ended.’

c. *binisba-li- ALI, tara-h ʔal-musabagah ʔintahat.

Prt- ALI Prt-3SG.M DEF-competetion.3SG.F ended.it

Intended meaning: ‘As for ALI, the competition has ended.’

It is evident from (15a,b) that tara agrees with the DP following it; tara marks the DP

ʔal musabagah ’the competition’ carrying a clitic with the same φ-features as the DP ʔal

musabagah. Moreover, the contrastive stress tara is endowed with, which tara needs to

discharge to one and only one element, is assigned to the the DP ʔal musabagah rather than Ali,

which here expresses an S-Topic, hence, the ungrammaticality of (15c). Further evidence in

favour of the assumption that tara is not an S-Topic particle comes from the fact that, unlike

C-Topics, S-Topics are root phenomena in the sense that S-Topics cannot occur in embedded

contexts (Bianchi and Frascarelli 2010). In this regard, binisba-li doesn’t occur in embedded

clauses (16a), while tara can do (16b), a fact militating against the assumption that tara is an

S-Topic, as illustrated in the following sentence:

(16) a.*Firas sˤaraħ ʔin

Firas declaired.he COMP

binisba-li- Ali, ʔal- musabagah ʔintahat.

Prt- Ali Def- competetion ended.it

Intended meaning: ‘Firas declared that, as for Ali, the competition is over

b. Firas sˤaraħ ʔin

Firas declaired.he Comp

tara-ah ʔAL- MUSABAGAH ʔintahat.

Prt-3SG.F DEF-COMPETITION ended.it

‘Firas declared that THE COMPETITION has ended.’

Armed with these facts, I claim that tara is not an S-Topic. let’s now see if tara displays

any properties related to F-Topic, as will be dealt with in the next subsection.

4.2.2 tara is not an F-Topic particle

The entity marked by tara, as shown above, must bear a contrastive stress, where this entity is

interpreted as being selected out of a set of discourse-given entities with which all the

conversation interlocutors are familiar. Arguing that tara is an F-Topic, we must scrutinize the

main properties that distinguish F-Topic from S-Topic and C-Topic. As a first approximation,

though contextually given and available in the common ground of discourse (Frascarelli and

Hinterhölzl 2007), an F-Topic is inherently non-contrastive, for it is not being selected and

contrasted to a set of entities within the conversational common ground.

Related to this fact is that F-Topics are typically realized as non-contrastive pronouns,

given the ability to access them throughout the dialogue (cf. Givon 1983, Chafe 1987, and

Pesetsky 1987). Consider the following dialogue:

(17)

A1: binisba-li- Firas ʔal-mubarat intahat w Barcelona fazau

As for- Firas Def-match ended and Barcelona won

bi-l-butʔu:lah ʔal-aham


33

in-the-championship the most important

‘As for Firas, the game is over. Barcelona have won the most important

championship.’

B: bus ma fazau bi-ʔal-dauliyah

But Neg won in-Def-international

‘But they didn’t win the international championship.’

A2 : (*tara-h) hu miktifi bi-l-maħaliyah

Prt-him he satisfied.he in-local

‘He is happy with the local championship.’

A3: (*tara-hum) hum miktifin bi-l-maħaliyah

Prt-them they satisfied.they in-local

‘They are happy with the local championship.’

A4: *tara-h HU miktifi bi-l-maħaliyah

Prt HE satisfied with-Def-local

‘HE is satisfied with the local championship.’

A5: *tara-hum HUM miktifin bi-l-maħaliyah

Prt THEY satisfied with-Def-local

‘THEY are happy with the local championship.’

In (17 A1), the DP Firas serves as an S-Topic, being marked by binisbali while the familiar

topic is the DP Barcelona. As the conversation proceeds, Firas falls into the domain of F-Topic,

being pre-established by the S-Topic. It is now available and familiar, as can be captured by its

referent being expressed via the pronoun hu in (15 A2). This also extends to the other F-Topic,

the DP Barcelona, now being also accessible and expressed via the pronoun hum (15 A3).

Given this reasoning, their incompatibility with tara (15 A2, A3) and their non-contractiveness

property (15 A4, A 5) appears plausible. Under this analysis, it can be postualted that the DP

marked by tara is not an F-Topic, and that tara, as a result, is not an F-Topic particle.

4.2.3 Tara is a C-Topic particle

Having argued that tara cannot be categorised as S-Topic or F-Topic, and that tara displays the

properties of a C-Topic particle, let’s explore evidence in favour of this argument. For this, I

analyse the derivation underlying the role tara has in the sentence interpretation and the C-

Topicalization reading of the element tara marks. In this regard, consider the following

sentences:

(18) a. tara-h ALI ʃaf as-sayarah

Prt-3SG.M ALI see.PST.3SG.M Def-car

‘ALI saw the car.’

b. tara-hin ʔAL-BANAT ʃaf-an as-sayarah

Prt-3PL.F DEF-GIRLS see.PST.3PL.F Def-car

‘THE GIRLS saw the car.’

As discussed above, tara marks one element in the clause. This element can be identified

by the contrastive stress tara assigns it, and, in cases this entity is a DP, a clitic with the same


34

φ-features is spelled out on tara. For instance, in ((18a) and (18b)), the clitic on tara carries the

same φ-features of the subject Ali being specified as [3SG.M] and ʔalbanat ‘the girls’ specified

as [3PL.F], respectively. Put another way, tara agrees with the element it marks. It follows,

then, that when tara marks an element, it agrees with it in order to identify it as the C-Topic of

the sentence, while the latter is in situ. This now raises the question as to how this clitic is

spelled out on tara.

Under the syntactic assumptions so far outlined and following Biberauer et al. (2014) in

that discourse particles affecr the interpretaion of the clause in which they merge, I argue that

tara has uninterpretable unvalued φ-features and an interpretable but unvalued [TOP] feature in

the sense of Pesetsky and Torrego (2007).8 Having an unvalued [TOP] feature, tara operates as

a probe for the assumption that unvalued, interpretable features must be valued before LF, due

to the principle of full interpretation. Thus, in (18a), for instance, once merged, tara begins

searching for a goal with a matching valued [TOP] feature within its visible c-command domain.

The DP Ali has an uninterpretable but valued [TOP] feature as well as interpretable/valued φ-

features, making it an active goal located within the visible c-command domain of tara. As a

result, a probe-goal relationship is established between tara and Ali. Following this, the

unvalued [TOP] feature of tara is valued by the matching valued counterpart on the DP Ali. At

the same time, the uninterpretable valued [TOP] feature on the DP Ali is deleted while being in

situ (in Spec-TP).9 Hence, the topicalized element is not triggered to move. Similarly, the

interpretable/valued φ-features of the DP Ali value the uninterpretable/unvalued φ-features of

tara, resulting in the spell out of the φ-features of tara, i.e. the realization of the clitic on tara.

The same analysis extends to sentence (18b) where a probe-goal configuration is etablished

between tara and the subject ʔalbanat ‘the girls’, resulting in spelling out the φ-features of tara,

the clitic ah, which is specified as [3PLF].

The derivation (18a) is schematically represented in (19) below (the dotted arrow

indicates tara’s probing the subject).

8 For Pesetsky and Torrego (2007), (un)interpretability and (un)valuedness must be conceived of as independent

notions. 9 This also occurs while the subject is in its thematic post-verbal position, Spec-vP as in (6b,d).


35

(19)

The question we should ask now is what about the cases where tara appears bare

(without a clitic appearing on it), as in (8) above (repeated below in (20))?

(20) a. tara ʃAF Ali as-sayarah

Prt see.PST.3SG.M Ali Def-car

‘Ali SAW the car (he didn’t just imagine it).’

b. tara TAW ʃaf Ali as-sayarah

Prt just now see.PST.3SG.M Ali Def-car

‘Ali JUST saw the car (not long before).’

As we have seen, when tara agrees with a DP, the latter values the φ-features of the

former resulting in the occurrence of the clitic being spelled out on tara. However, in (20), no

clitic is spelled out on tara, a fact which follows if we assume that tara’s φ-features are not

valued. This immediately raises the question as to what accounts for the convergent derivation

of the sentence, given that uninterpretable features (whether unvalued in the sense of Chomsky

2000, 2001 or valued in the sense Pesetsky & Torrego 2007), must, eventually, be deleted

during the course of the deivation, or the derivation would otherwise crash. Remedy to this

question lies in the assumption that tara in such cases agrees with the element that is not

endowed with interpretable φ-features. In (20a), for instance, tara agrees with the verb ʃaf

‘saw’, the C-Topic, as explained above. Lack of clitic on tara when agreeing with the verb,

then, follows from the fact that verbs have uninterpretable φ-features, which, by the time tara

is merged, would have already been valued and, more crucially, deleted, a fact which

consequently resluts in the φ-features of tara being unvalued and undeleted. Hence, unlike the

cases with DP’s, this results in lack of clitic spellout on tara. The same analogue is extended to

cases where the C-Topic is an adjunct, taw ‘just now’ (20b). Under such cases, I propose that,

following Epstein et al. (2010), uninterprtable φ-features of discourse related elements, rather

than lexical items, like tara are ignored at LF (and PF), since the interface levels recognize only


36

elements that contribute to the interpetation of the sentence, hence, those with interpretable

features; uninterprtable features need not be deleted. What is crucial for the convergence of the

given sentence containing tara is the valuation of the interpretable unvalued [TOP] feature of

tara.

With this conclusion, there are still several issues to be explored as far as tara is

concerned. Here, I single out two issues, namely the cases when tara agrees with the direct

object (where a clitic agreeing with the object is spelled out on tara) and the obligatory re-

merge of the C-Topic when tara is not part of the numeration. As for the first issue, the

interesting matter here is that the direct object must be in a position c-commading the subject.

When tara agrees with the direct object, the latter must move to a position which is closer to

tara, unlike the cases with the subject which can appear postverbally (in its thematic position)

while agreeing with tara. Regarding the second issue, it is evident that when tara is not part of

the sentence, the C-Topic must be discloated to the left peripehry even if the C-Topic is the

subject. It seems that there is a co-relation of some kind between the use of a probe-goal

configuration and the existence of an overt [TOP] head. However, I leave these issues open

pending further research.

5 Conclusion:

This research tackles the discourse-related particle tara in North Hail Arabic. The main

assumption of the current paper is that tara is a head instantiating the Contrastive Topic Phrase

in the sense of Frascarelli and Hinterhölzl (2007). The study argues that tara enters the

derivation endowed with an interpretable unvalued [TOP] feature in addition to a set of

uninterpretable φ-features which are valued by the element that serves as a Contrastive Topic.

This valution is conducted through a probe-goal configuration (Chomsky 2000, 2001, 2008).

Additionally, the study shows that the clitic appearing on tara when tara agrees with the subject

or the direct object DP’s is in fact a by-product (epiphenomonal) of the Agree relation

established between tara and the C-Topic. It is a spell out of the φ-features of tara as a result

of maintaining agreement relation between tara and the C-Topic, when the C-Topic has φ-

features, a DP. However, in case the C-Topic does not have φ-features, no clitic is reliazed on

tara, in which case the uninterpretable φ-features of tara are ignored at LF (and PF).

References

Al-Shamari, M.R. (2015). ‘Pragmatic analysis of the particle ʁadɪ in Najdi Arabic’.

International Journal of Linguistics, 7(2), pp.81-93.

Alshamari, M.R. (2015). ‘A Relevance-Theoretical Account of Three Discourse Markers in

North Hail Arabic’. Studies in Literature and Language, 11(1).

Alshamari, M. (2015). ‘Documentation of Discourse-related Particles in North Hail Arabic’.

English Linguistics Research, 4(4), p.p44.

Alshamari, M. R., & Jarrah, M. (2016). ‘A Minimalist-Based Approach to Phrasal Verb

Movement in North Hail Arabic’. International Journal of English Linguistics, 6(1), 24.

Belletti, A. (2004). Aspects of the low IP area. The structure of CP and IP. The cartography of

syntactic structures, 2, 16-51.

Bianchi, V., & Frascarelli, M. (2010). Is topic a root phenomenon? Iberia: An International

Journal of Theoretical Linguistics, 2(1).

Biberauer, T., & Sheehan, M. (2011). Introduction: particles through a modern syntactic lens.

The Linguistic Review, 28(4), 387-410.

Biberauer, T., Haegeman, L. and Kemenade, A., (2014). Putting our heads together: towards a

syntax of particles. Studia Linguistica, 68(1), pp.1-15.

Boeckx, C. (2009). Understanding minimalist syntax: Lessons from locality in long-distance

dependencies. John Wiley & Sons.


37

Carstens, V. (2010). Implications of grammatical gender for the theory of uninterpretable

features. Exploring crash-proof grammars, 3, 31.

Chafe, W. (1987). Cognitive constrains on information flow. In: Tomlin, R. (Ed.), Coherence

and Grounding in Discourse. Benjamins, Amsterdam, pp. 21--51.

Cinque, G. (1999). Adverbs and Functional Heads: A Cross-Linguistic Perspective. Oxford

Uiversity Press, Oxford.

Cinque, G. ed. (2002). Functional structure in DP and IP: The cartography of syntactic

structures (Vol. 1). Oxford University Press.

Cinque, G. (2006). Restructuring and functional heads: the cartography of syntactic structures

(Vol. 4). Oxford University Press.

Cinque, G. & Rizzi, L., 2008. ‘The cartography of syntactic structures’. Studies in linguistics,

2, pp.42-58.

Chomsky, N. 2000. Minimalist Inquiries: The Framework. In Lasnik, H., R.Martin, D. Michaels

and J. Uriagereka (eds.) Step by step. essays on minimalist syntax in honor of Howard

Lasnik. Cambridge. Mass: MIT Press.

Chomsky, N. (2001). Derivation by phase. In Kenstowicz, M. (ed.) Ken Hale. A Life in

Language. Cambridge, Mass: MIT Press. 1-52.

Chomsky, N. (2008). On phases. Current Studies in Linguistics Series, 45, p.133.

Coniglio, M. (2008). Modal particles in Italian. Working Papers in Linguistics, 18, pp.91-129.

Danon, G. (2008). Definiteness spreading in the Hebrew construct state. Lingua, 118(7),

pp.872-906.

Elghamry, K. (2004). Definiteness and number ambiguity in the superlative construction in

Arabic. Lingua, 114(7), 897-910.

Epstein, S.D., Kitahara, H. & Seely, T.D. (2010). Uninterpretable features. Exploring crash-

proof grammars, 3, p.125.

Fehri, A.F. (2012). Key features and parameters in Arabic grammar (Vol. 182). John

Benjamins Publishing.

Frascarelli, M., & Hinterhölzl, R. (2007). Types of topics in German and Italian. On information

structure, meaning and form, 87-116.

Friedmann, N., Belletti, A. & Rizzi, L. (2009). ‘Relativized relatives: Types of intervention in

the acquisition of A-bar dependencies’. Lingua, 119(1), pp.67-88.

Givón, T. (1983). Topic continuity in discourse: An introduction. In Topic Continuity in

Discourse: A quantitative cross-language study, T. Givón (ed.), 5–41. Amsterdam: John

Benjamins.

Hack, F. M. (2014). The Particle Po in the Varieties of Dolomitic Ladin–Grammaticalisation

from a Temporal Adverb into an Interrogative Marker. Studia Linguistica, 68(1), 49-76.

Jarrah, M., & Zibin, A. (2016). ‘On Definiteness and Information Trigger in Arabic’. Advances

in Language and Literary Studies, 7(2), 55-67.

Krifka, M. (2007). Basic Notions of Information Structure. in Fery, C., G. Fanselow and M.

Krifka (eds.). The Notions of Information Structure, Interdisciplinary Studies on

Information Structure 6: 13–55.

Danon, G. (2008). Definiteness spreading in the Hebrew construct state. Lingua, 118(7), 872-

906.

Fehri, A. F. (2012). Key features and parameters in Arabic grammar (Vol. 182). John

Benjamins Publishing.

Pesetsky, D. (1987). Wh-in-situ: movement and unselective binding. In: Reuland, E., ter

Meulen, A. (Eds.), The Representation of (In)definiteness. MIT Press, Cambridge, MA,

pp. 98--129.

Pesetsky, D., & Torrego, E. (2007). The syntax of valuation and the interpretability of features.

Phrasal and clausal architecture: Syntactic derivation and interpretation, 262-294.


38

Rizzi, L. (1990). Relativized minimality. The MIT Press.

Rizzi, L. (1997). The fine structure of the left periphery. In Elements of grammar (pp. 281-337).

Springer Netherlands.

Rizzi, L. (2004). Locality and left periphery. Structures and beyond. The cartography of

syntactic structures, 3, 223-251.

Struckmeier, V. (2014). Ja doch wohl C? Modal Particles in German as C‐related elements.

Studia Linguistica, 68(1), 16-48.

Zimmermann, M. 2004. ‘Discourse particles in the left periphery’. ZAS Papers in Linguistics,

35, 543-66.


39

Murdhy Alshamari



Newcastle upon Tyne


United Kingdom

[email protected]

[email protected]



Malessa & Lim An exploratory study of first exposure to Urdu

40

AN EXPLORATORY STUDY OF FIRST EXPOSURE TO URDU – DISCOVERING

WHAT MATTERS

EVA MALESSA & REBECCA LIM

(Newcastle University) 1

Abstract Little research has been conducted on how adults identify sound strings as words and map meaning to

these strings when they encounter naturalistic input for the first time. FIRST EXPOSURE studies

examine how adults process novel stimuli in an unfamiliar language. This paper reports on a replication

of Gullberg et al.’s (2012) work on Dutch speakers’ first exposure to Mandarin Chinese. In our small-

scale study, 38 English-speaking participants were exposed to Urdu in a four-minute weather report and

then tested on word recognition and sound-to-meaning mapping. Controlled variables included item

frequency, loanwords and visual highlighting. The current study found that the loanwords traffic and

degree facilitated word recognition. Results on item frequency and visual highlighting were not deemed

sufficiently conclusive to suggest facilitation effects. D-prime (d’) calculations for the tasks revealed a

very low discriminability of target and filler items. The current study is thus unable to support Gullberg

et al.’s (2012) findings that adults are able to extract language-specific information about a novel

language after minimal exposure. Future studies are needed to establish whether sensitivity to a novel

language is language-specific or not, and to further study the effect of first exposure with regard to

different novel languages.

1 Introduction to FIRST EXPOSURE

Making sense of the speech stream is a challenge for language learners confronted with a new

and unknown language for the first time. The first crucial steps include SEGMENTATION

(identification of relevant sound strings as ‘words’) and MEANING-MAPPING (assigning

relevant meaning onto those ‘words’). These steps constitute the complex ‘learner’s problem

of analysis’ (Klein 1986: 59). Shoemaker and Rast (2013: 167) point out that while the

limitations of second language (L2) speech processing at intermediate and advanced

proficiency level have been extensively studied, very little research has focused on 'how

learners initially break into the sound stream’.

Remarkably little research has, in fact, been conducted in second language acquisition

(SLA) on the topic of first exposure, investigating how adult L2 learners master segmentation

and meaning-mapping challenges when exposed to naturalistic input2 in form of complex and

coherent speech for the first time. This scarcity of relevant research has been noted by the SLA

research community (Carroll & Windsor 2015: 58, Han & Liu 2013: 145, Carroll 2004: 236).

This paper reports on a small-scale replication of Gullberg et al.’s study (2012), investigating

how adult learners process naturalistic yet controlled input after minimal exposure to a novel

language. The following section presents an overview of relevant studies. Section 3 illustrates

this study’s methodology. Results are presented in section 4. Finally, section 5 concludes with

WHAT really matters in first exposure.

1 We would like to acknowledge Professor Martha Young-Scholten, who was a great source of inspiration, for her

patient guidance throughout the process of this study. 2 In this paper, the term ‘input’ is used interchangeably with the term ‘stimuli’, following standard practice in SLA

literature. Carroll (2001: 8) points out that ‘most writers, however, do not bother to define what they mean by the

term’. Flege (2009: 175) makes a similar point, further defining L2 speech input as ‘all L2 vocal utterances the

learner has heard and comprehended, including his own’. In this study, the terms ‘input’ and ‘stimuli’ are

employed ‘in a general sense to refer to the language a learner hears’ (Piske & Young-Scholten 2009: 263).

Following Rast (2010), the terms ‘learning’ and ‘acquisition’ are also used interchangeably.

Newcastle and Northumbria Working Papers in Linguistics 22 (2016) Malessa & Lim

41

2 Background literature: From hearing noise to decoding

The initial stages of SLA have recently started to attract researchers’ attention, in particular

with regard to first language (L1) effect and L2 input processing and constraints (Carroll &

Windsor 2015: 58). Emerging findings indicate that learners go beyond their L1 perceptual

filters and point to sensitivity to phonotactic and other information providing cues for speech

stream segmentation (Carroll 2013; Shoemaker & Rast 2013). Carroll (2004: 229) stresses that

‘hearing words is no primitive operation’ as it results from ‘some rather complex phonetic and

phonological parsing and word-recognition operations’. Her focus is thus on how the learner

makes the transition from hearing incomprehensible noise to hearing some sequence of

syllables (ibid: 236).

Most first exposure studies (Carroll 2014, 2012, Rast 2010, Ristin-Kaufmann &

Gullberg 2014, Shoemaker & Rast 2013) have focused on the adult L2 learners’ ability to

segment prosodic words from continuous speech. They have shown that learners are able to

segment L2 sound forms after very little input, regardless of the similarity between L1 and L2

(Carroll & Windsor 2015: 59). Whether or not segmentation is easy and/or rapid remains a

debated issue. Carroll and Meisel (2015: 11) point to earlier research showing that

‘segmentation even at the initial stage is rapid’, while Han and Liu (2013) claim otherwise.

The role of INPUT remains a highly debated topic in SLA, in particular with regard to

adult learners’ capabilities to process natural speech (see Carroll 2014). Gullberg et al. (2012)

emphasize that while ‘input’ as a central issue in SLA research has received a lot of attention,

surprisingly little is known about HOW input matters. However, input undoubtedly has an effect

on SLA acquisition (Piske & Young-Scholten 2009: 1). The studies by Gullberg et al. (2010,

2012) indicate that ‘at the earliest stages of L2 acquisition […] the adult learning mechanism

can deal efficiently with very little input and very complex input’ (Gullberg et al. 2010: 18).

These studies differ from existing studies as they employ ‘natural input’ , visual cues, the effect

of item frequency, ‘implicit/incidental learning outside of a classroom context’ and extremely

limited exposure (Gullberg et al. 2012 : 244-245).

Ristin-Kaufmann and Gullberg (2014: 18) point out that while the contribution of

artificial and statistical language studies to the understanding of L2 learning has been

enormous, they are not comparable to ‘naturalistic’ language exposure studies. In the study

conducted by Gullberg et al. (2012), native Dutch participants were exposed to seven minutes

of controlled naturalistic input in Chinese Mandarin without explicit instruction and then tested

on their segmentation and meaning-mapping abilities. Their results suggested that adults draw

on frequency, gestural cues and phonotactic sensitivity in decoding a novel language. Gullberg

et al. (2012: 259) proposed that ‘the adult learning mechanism appears to be a great deal more

powerful than typically assumed in the L2 acquisition literature’.

2.1 Background of the present study, a replication of Gullberg et al.’s (2012) study

The following research questions, based on Gullberg et al. (2012), were adopted for this study:

1. Can adults EXTRACT information about word forms and meanings after minimal

exposure to naturalistic yet controlled input in a previously unknown language?

2. What are the EFFECTS of item frequency, visual highlighting (icons and gestures) and

lexical transparency (loanwords)?

Variables of this study included frequency of item appearance and visual highlighting. Lexical

transparency was included as a variable to investigate the lexical effects on segmentation as

suggested in existing literature (Carroll & Windsor 2015: 59). Shoemaker and Rast (2013: 169)

highlight that ‘transparent words’ show a higher detectability as they ‘are mapped onto existing

mental representations’.


42

It was hypothesized that participants would be able to segment sound strings after a

very short exposure in the pre-testing treatment. Learners were also predicted to be able to

extract word-related information from the complex and coherent speech they were exposed to

in the weather report. Finally, frequency and visual highlighting were expected to facilitate

detectability of target words and the presence of loanwords was anticipated to enhance

participants’ ability to segment and map meaning to form.

Our study investigated whether or not Gullberg et al.’s (2012) findings for Dutch

participants were possibly language-specific. This formed part of the motivation for this study,

in which English-speaking participants were exposed to Urdu. Urdu, an Indo-Aryan language,

was selected due to its typological distance to English and the fact that participants had had no

contact to Urdu prior to treatment. Urdu is a syllable-timed language, differing from stress-

timed English. In addition, its phonemic inventory consists of fewer vowels. There are,

however, more consonants in Urdu compared to English. Urdu and English also differ with

regard to orthographical and typographical characteristics. Urdu has, additionally, a complex

morphology. In terms of phonotactics, complex onset and coda clusters are, nevertheless,

allowed in both languages. (Schmidt 1999).

The following section presents the methods and materials employed in this study.

3 Methods and materials

3.1 Participants

Participants in this study were linguistics students from Newcastle University. The study

involved one experimental group consisting of 38 participants (7 males and 31 females). A

background questionnaire3 revealed that the majority of thirty English L1 speakers was

complemented by eight English L2 speakers. All but six English L1 participants reported that

that they were bilingual to some degree. Fourteen English L1 speakers reported knowledge of

one L2 and ten indicated having knowledge of two or more L2s. The non-native English

speakers were multilingual to various degrees (see Appendix 2 for details). All participants

reported no prior knowledge of Urdu.

3.2 Treatment

The current study incorporated a pre-testing treatment, in which the participants were exposed

to fully controlled, continuous novel stimuli in Urdu (see Appendix 3). During the treatment,

participants received naturalistic input in the form of a Pakistan audio-visual weather report.

The four-minute video, featuring a female native Urdu speaker, was made available online for

participants’ access. The video contained six visual weather charts. The script of the continuous

and coherent weather text consisted of 120 clauses and 24 target words, which the participants

were later tested on (see Appendix 5). The target word mosam ‘weather’ in the first clause of

the weather report is highlighted in Example 1 below.

(1) peer k din Pakistan main mosam k hal k sath, Main hu Qudsia

monday Pakistan in weather report with I am Qudsia

‘This is Qudsia with your Monday weather report in Pakistan.’

Item frequency and the speaker’s gestures were controlled. Target words were, however, not

controlled for sentence position. The participants were instructed to watch the four-minute

video at least twice at their own convenience within one week. In addition, participants were

3 The questionnaire required participants to self-indicate their knowledge and proficiency level of other languages.

It is therefore possible that participants either over- or underestimated their language knowledge/proficiency level.


43

told to pay attention to loanwords in the treatment video. All participants reported on their total

number of viewings in a background questionnaire during the testing (see Appendix 2).

3.3 Testing

The testing consisted of two tasks, completed in one session. Task 1 was a word recognition

task (see Appendix 4). Task 1 consisted of 42 items, including 20 Urdu target words, 12 Arabic

filler words and 10 Urdu filler words (see Appendix 5). The word order was randomized to

ensure they did not appear in the same order as the treatment video. Participants were instructed

to listen to audio clips recorded prior to testing by the same native Urdu speaker who presented

the weather report. After each clip participants had to indicate whether they had heard the word

in the weather report during the treatment by ticking the corresponding Yes or No box, see

Example 2. There was a time-limit of two seconds between each item to ensure instantaneous

response. The audio clips were played only once.

(2) Item 7 traffic

Yes No

Task 2 was a sound-to-meaning mapping task with six pre-recorded audio files containing four

Urdu target words, as well as one Urdu and one Arabic filler word (see Appendix 4 and 5).

For each item, a visual highlight (in form of an icon) used in the treatment video was shown

on the screen, followed by an audio clip containing a word. The icon order was randomized.

Participants were instructed to indicate whether the utterance in the audio clip correctly

represented the visual highlight shown on the screen by indicating Yes or No in the

corresponding boxes, see Example 3 below. Participants were given only four seconds between

each set of stimuli.

(3) Item 1 barf

Yes No

The following section presents this study’s results.

4 Results

This study’s data was analysed by obtaining the percentage of correct identification of TARGET

words (used in the treatment) and FILLER words (not used in the treatment video). Task 1 and

2 contained 24 target and 24 filler items in total. Each item, correctly identified as either target

item (heard in the treatment) or filler item (not heard in the treatment) was awarded a score of

1. An incorrect response was awarded a score of 0. The data of three participants (*36, *37,

*38) were excluded as these participants were directly involved in the set-up of the treatment

and testing, rendering their data inappropriate for the study (see Appendix 2). This reduced the

number of subjects in the data analysis to 35.

4.1 Overall performance

In Task 1, participants were able to correctly identify 55.24% of all 42 items as either target

words or filler words. On average, they were able to correctly map the meaning of the word to

icons for 60.95% of all six items in Task 2.


44

Table 1: Percentage of overall performance

Overall performance (Target + Filler)

Task 1 55.24%

Task 2 60.95%

Task 1 + 2 53.99%

The overall performance in both tasks was just above chance (53.99%), which was lower than

anticipated. One possible explanation for the low performance might have been the lack of

control in ensuring that all participants viewed the treatment video at least twice. This lack of

control might have resulted in participants not watching the video at all, or only watching it

once. Since the number of viewings was not a controlled variable, this study relied on the

collaboration of the participants and their honesty in reporting their number of viewing times.

4.2 Comparison of the performance on target and filler items

In Task 1, the two loanwords traffic and degree were included as target items. Due to the

participants’ bias resulting from the viewing instructions, it came as no surprise that

participants were highly successful in recognizing both L7 traffic and L22 degree, with 85.71%

and 88.57% success rates respectively (see Appendix 7). Performance for these target words

was therefore distinctively higher compared to the performance of target words (see Table 2).

Table 2: Percentage of performance of target words

Overall performance (Target only)

Task 1 43.13%

Task 2 38.57%

Task 1 + 2 42.40%

The overall performance on target words (42.40 %) was very low, compared to the overall

performance on filler words (67.14%), as shown in Table 3 below. In Task 1, both Arabic and

Urdu filler words were identified as new input in 67.92% of cases. A more detailed breakdown

revealed greater detectability of Arabic than Urdu filler words. On average, 68.79% of Arabic

fillers and 65.19% of Urdu filler items were successfully recognized.

Table 3: Percentage of performance on filler words

Overall performance (Filler only)

Task 1 (Arabic) 69.52%

Task 1 (Urdu) 66.00%

Task 1 (Overall) 67.92%

Task 2 (Arabic) 60.00%

Task 2 (Urdu) 57.14%

Task 2 (Overall) 58.57%

Task 1 + 2 (Arabic) 68.79%

Task 1 + 2 (Urdu) 65.19%

Task 1 + 2 (Overall) 67.14%

Based on the results of performance, i.e. the correct identification as either previously heard

items or new items, participants seemed to perform better at correctly identifying filler words

in both tasks as non-target words (i.e. new items) in contrast to identifying previously heard

items as target items. In order to obtain a more distinct indication of performance, response


45

bias (β) and D-prime (d’) were measured. The following section presents the results for the β

and d’ calculations and discusses some problematic issues of the current study.

4.3 Response bias and sensitivity

In order to separate detection, i.e. correct identification, from RESPONSE BIAS, ‘the general

tendency to respond yes or no’ (Stanislaw & Todorov 1999: 139), values for the factors

response bias (β) and sensitivity (d’), in the obtained performance results in Task 1 and 2, were

calculated, based on the Signal Detection Theory (SDT)4.

The obtained value for sensitivity (d’)5 in Task 1 (see Table 4 below) indicated low

discriminability, signifying participants’ minimal sensitivity and hence, inability to distinguish

target items from filler items. For Task 1, the quantification of the response bias6 with β

indicated a bias towards responding with no (< 1), i.e. they have not heard the word before.

This bias towards responding no possibly accounts for the unexpected higher detectability for

filler items (67.14%) in comparison to target items (42.40%), as mentioned in section 4.2.

Table 4: Calculation of response bias and sensitivity for all participants

Task 1 Task 2

d’ 0.245 - 0.056

β 1.087749 0.984402

For Task 2, the quantification of the response bias with β showed a bias towards responding

with yes (> 1). As two thirds of the tested items in Task 2 were target items, this bias may have

resulted in the higher overall performance in Task 2 (60.95%) compared to the overall

performance in Task 1 (55.24%). In addition, the negative value of d’ (-0.056) suggests

response confusion, i.e. ‘responding yes when intending to respond no, and vice versa’

(Stanislaw & Todorov 1999: 139-140).

Both the bilingual, and in particular the multilingual native-English speakers, were

affected by this response confusion in Task 2, while the monolingual speakers were only

slightly affected and the non-native speakers remained unaffected (see Appendix 11). Based

on the calculations for d’, no similar observations were made for Task 1 (see Appendix 10).

Interestingly, the response bias values for the monolingual English-native speakers and

bi-/multilingual native speakers differed slightly. While the monolingual speakers tended to

respond no, according to their response bias value (1.07), the bi-/multilingual speakers

favoured yes (0.98), see Appendix 11 for details.

4.4 Item frequency and visual highlighting

For the purpose of this small scale-study, an item was classified frequent7 if it occurred five

or more times in the weather report. Only 18.75% of the target words were classified as frequent

in this study, with the majority of target items appearing infrequently. The most frequent word

T19 ilaqay ‘areas’ had an unexpectedly low detectability (22.86%) compared to an average of

40.48 % for target items in Task 1, as shown in Table 5 below.

4 See Table 1 and Table 2 in Stanislaw & Todorov (1999: 145). 5 A value of 0 for d’indicates an inability to distinguish signals from noise. Larger d’ values (> 0) indicate a

correspondingly greater ability to distinguish signals from noise. 6 A β value of 1 indicates no bias. A value less than 1 signifies a bias towards yes, and a value more than 1 signifies

a bias towards no. 7 In Gullberg et al.’s (2012: 245) study frequent words occurred eight times or more.


46

Table 5: Percentage performance on frequent and infrequent target words

Overall Performance (Target Items)

Task Frequent Items Infrequent Items Overall

1 40.48% 38.18% 55.24%

2 45.71% 31.43% 60.95%

Q7 traffic and Q22 degree, both loanwords as well as frequent target items in Task 1,

constituted the most successfully detected items in the study, with over 85% recognition

success (refer to Appendix 6 for details). This might have affected the outcome of the overall

performance for frequent items. The recognition success of frequent items (40.48%) in Task 1

is thus seen as comparable to that of infrequent items (38.18%). Contrary to prior expectations,

this study did not find any significant correlation between item frequency and the successful

recognition of target items, and differs, therefore, from Gullberg et al.’s (2012: 254) finding

that ‘if a word appeared frequently in the input, it was more likely to be recognized than if it

had appeared infrequently’.

In Task 1, target items denoting directions, T5 shumal ‘north’, T8 janub ‘south’ and

T26 mashraq ‘east’, were accompanied by a gesture in the treatment video. The average

performance for these three target items was extremely low, under 40.00%. It is important to

consider that the lack of rehearsal by the speaker prior to recording the treatment video might

have resulted in spatial and temporal inaccuracy of the gestures. As a result, the gestures in the

treatment video might not have been sufficiently obvious to participants, providing no ‘reliable

deictic link between gesture, speech and weather charts’ (Gullberg et al. 2012: 246).

In Task 2, only one outlier (5F) correctly identified all six items in the task (see

Appendix 9). The meanings of the aural items were only 45.23 % of the time successfully

mapped by all participants to the corresponding icons. It seems very likely that the role of visual

highlighting in form of icons might have been overestimated in this study due to the small

number of items in Task 2 (6 items in total: 4 target items and 2 filler items) compared to Task

1 (42 items in total: 10 target items and 12 filler items).

Our initial results seem to suggest no effect of visual highlighting on word recognition

in this study. This would be in line with the observations by Gullberg et al. (2012: 254). Taking

into account the overall results and problematic issues such as the frequency of target words,

number and distribution of target and filler words as well as participants’ sensitivity (d’) and

response bias (β), this study’s findings could, however, not convincingly show whether item

frequency and visual highlighting facilitate detectability.

5 Discussion – What really matters

D-prime measurements revealed the participants’ miniscule ability to distinguish target from

filler items. This is also reflected in the overall performance results, which were minimally

above chance (see Table 1). Further this study found that participants were compromised by a

response confusion in Task 2. Based on these shortcomings, this study’s results were

interpreted as non-significant regarding the original hypothesis that adults are able to extract

language-specific information after minimal exposure to naturalistic yet controlled input in a

novel language. Future studies are needed to shed more light on the (in)effectiveness of first

exposure.

This study found, nevertheless, that, according to prior expectations, LEXICAL

TRANSPARENCY displayed by the two loanwords traffic and degree MATTERS, as it enhances

segmentation and meaning-mapping in a novel language. Previously, Shoemaker and Rast

(2013) have demonstrated that loanwords facilitate segmentation and meaning-mapping in

speech streams and have also provided evidence of such lexical effects on segmentation (see

Carroll & Windsor 2015: 59).


47

In order to boost scientific rigor of the current study, monitoring the treatment input in

terms of length, frequency and setting as well as employing a control group is seen to be

essential. Additionally, controlling participants’ language backgrounds to ensure

methodological rigor in future studies is considered crucial. According to Rast (2010: 80) not

only the learners’ native languages, but also other languages play an important role in the

acquisition of an additional L2. Carroll and Windsor (2015: 55) claim that ‘familiarity with

other L2s can offer learners phonetic variants as possible targets for segments recognized as

functionally equivalent’. Similarly, Carroll (2004: 235) proposes that bilinguals use a

segmentation strategy that is ‘based on exposure to the prosodic properties of their dominant

language, and transfer it to their weaker one’. It seems clear then that knowledge of both L1

and L2(s) matters in first exposure studies.

In addition, aptitude, especially phonetic coding ability (Skehan 1989), appears to be

an interesting variable that should be considered in future first exposure studies. Shoemaker

and Rast (2013: 180) suggest that generalizations from studies involving experienced language

learners to ‘less experienced learners should be made with caution’. It is obvious that this also

concerns generalizations from less experienced/monolingual learners to multilingual learners.

Generalizing from one particular type of input, in this case a specific text type of a

‘pseudo-weather report’ (Carroll 2014: vii-viii) is certainly likewise problematic. The lack of

variability in text type and structure needs to be taken into account (Carroll 2013: 137).

Variables including the number of speakers used in the input, speech rates and environment

(noise), which are typical of normal speech, need to be considered as well (Carroll 2014: ii). In

a follow-up of the present study, input containing two or more speakers engaged in a two-way

conversation could be employed to increase the authenticity of the input and improve

variability. Rast (2010: 80) emphasizes further ‘the relative importance of interacting variables,

such as frequency and transparency, on different types of language activity’. Carroll (2012:

40), on the other hand, stresses that ‘repeated exposure to the same input is not necessary to

memorize a sound token’. Is less then really more in the initial stages of learning? (cf.

Denhovska et al. 2016). There remains a lot to be discovered about the implicit learning

mechanisms of adults, and first-exposure studies are just the tip of the iceberg revealing what

matters in SLA.


48

Appendix 1: Questionnaire

QUESTIONNAIRE

School of English Literature, Language and Linguistics,

Percy Building,

Newcastle upon Tyne,

NE1 7RY,UK

First name + initial of family name: ________________________________________

First language(s): ___________________________________________________

Other languages:

Beginner level ___________________________________________________

Intermediate level ___________________________________________________

Advanced level ___________________________________________________

Have you ever had any contact with Urdu before watching the video (e.g. Urdu-speaking friends, shopping at a

Pakistani grocery store)?

___________________________________________________________________________

How many times did you watch the video?

once twice three times or more

When was the last time you watched the video? _____________________________


49

Appendix 2: Participants’ background information based on the questionnaire

Participants are referred to according to a number followed by the letter M (male) or F (female), e.g. 1M.

BEG INT ADV

No.Gende

rFirst Lg

Urdu

Contact

Times watched

videoLast watched

Number of

Languages

1 M EL NIL N 2 6-Dec 1

2 F EL NIL N 1 30-Nov 1




6 F EL NIL N 2 6-Dec 1

10 M EL FRENCH N 1 30-Nov 2

24 F EL FRENCH N 1 30-Nov 2






20 F EL FRENCH N 1 3-Dec 2




14 F EL SPANISH N 1 30-Nov 2


29 F EL SPANISHY

Friends2 30-Nov 2

7 M ELSPANISH

GERMANN 2 7-Dec 3

8 M ELITALIAN

FRENCHN 2 2-Dec 3

9 M ELFRENCH

SPANISHN 1 6-Dec 3

21 F ELFRENCH

SPANISHN 2 30-Nov 3

22 F EL FRENCH GERMAN N 1 30-Nov 3

23 F EL GERMAN FRENCHY

Travel1 30-Nov 3

30 F ELGERMAN

FRENCH

Y

Classmate2 30-Nov 3

16 F ELFRENCH

GERMANN 1 30-Nov 3

17 F EL

DUTCH

DANISH

ITALIAN

FRENCH SPANISH N 3 30-Nov 6

18 F EL

ITALIAN

GREEK

BSL

FRENCH N 2 6-Dec 5

31 M CANTONESE MANDARIN EL N 2 30-Nov 3

32 M CANTONESE JAPANESE KOREANEL

MANDARINN 1 30-Nov 5

33 F POLISH ITALIAN PORTUGESEGERMAN

ELN 3 3-Dec 5

34 FFINNISH

SWEDISHMANDARIN FRENCH

GERMAN

ELN 3 6-Dec 6

35 F MALAY EL N 1 30-Nov 2

*36 F MANDARIN EL CANTONESE N 2 30-Nov 3

*37 F RUSSIAN MANDARIN EL

Y

Knows

Arabic

2 30-Nov 3

*38 F MANDARIN HOKKIEN EL N 1 30-Nov 3

Other Lg

NATIVE'

NATIVE' + L2

NATIVE' + L2+

NON-NATIVE'


50

Appendix 3: The Pakistan weather report

Target words are marked in bold. Target words also highlighted by a gesture are marked in green.

Text associated with weather slide 1

1. Subho bakhair

morning good

‘Good morning,’

2. peer k din Pakistan main mosam k hal k sath, Main hu Qudsia

monday__ Pakistan__in____weather report with I am Qudsia

‘this is Qudsia with your Monday weather report in Pakistan.’

3. Guzishta roz k mosam k muqably main

yesterday ’s weather in comparison with

‘In comparison to yesterday’s weather,’

4. Aj ka derja herarat berqarar nhe hai

today ’s temperature stable not is

‘today's temperature is not stable,’

5. Subo bohet kum

morning very low

‘very low in the early morning,’

6. Jab k bhar ker challis degree centigrade tak pohnch gaya

whweras rise forty degree centigrade upto reached

‘and could climb to as high as 40 degrees Celsius’

7. Bad dopehr

late Afternoon

‘by the late afternoon.’

8. Lakin mulk k kuch ilaqay

but country of some areas

‘But some areas of the country’

9. Abi bi dhop main kuch lamhat se lutf andoz ho saktay hain

still sunshine in moments enjoy can

‘can still enjoy a moment of sunshine.’

Text associated with weather slide 2 (snowy)

10. Aayiay shumal ka ahwal jantay hain

let us north situation know

‘Let’s look at the north.’

11. Shumal main

north in

‘In the north,’

12. Abi barf giray ge

still snow fall will

‘it will be snowy.’

13. Darja-hararat kam ho ga

temperature low be will

‘The temperature will be low.’


51

14. Ye munfi pandra degree centigrade tak jae ga

this minus fifteen degree centigrade to go will

‘It will be minus 15 degree Celsius.’

15. Aur mazid munfi taees degree centigrade tak girnay ka andesha hai

and more minus twenty-three degree centigrade to go down chance is

‘and it might go down to minus 23 degree Celsius’

16. Subo k ibtadaee auqat main

morning of early time in

‘during the early hours of the morning.’

17. Apko bohet thand mehsooos ho ge

to you very cold feel be will

‘You will feel very cold.’

18. Sham k waqt

evening of time

‘In the evening,’

19. Mazid barf bari ho sakti hai

more snowfall be can is/will

‘there will be more snowy,’

20. Ku k shumali Peshawar num hawao ki lupait main hai

because northern Peshawar moist air of Surrounded by is

‘because the moist air will gather north of Peshawar,’

21. Khas toer se wo ilaqay

especially those areas

‘especially in the areas’

22. paharu k karib derja herarat nukta e injimat per hai

mountains of near temperature below zero on are

‘near the mountains with below-zero temperature.’

23. Sarku pe pani jam jany se phislan ho ge

roads water frozen from slip be will

‘The roads will be very frosty and slippery.’

24. Bahir jatay huway ahtiat kijeay

out going while careful be

‘Be careful when you are out.’

Text associated with weather slide 3 (rainy)

25. Ab janub ka ahwal janiay

now south of situation know

‘Let’s look further down at the south.’

26. Janub main

south in

‘In the south,’

27. Barish k asaar hain

rain of chances are


52

‘It will be rainy.’

28. Aur kuch had tak garaj chamak k sth yalabari k imkanat b hain

and some of Thunder with hail fall of chances there are

‘and also has a bit of hail and thunder’

29. Darja - hararat kam ho ga

temperature low be will

‘Temperature will be lower.’

30. Sham main

evening in

‘In the evening,’

31. Ye taqriban das degree centigrade tak pohnchay ga

it approximately ten degree centigrade to reach will

‘it could be 10 degree Celsius.’

32. Ap thand mehsus kerain gay

you cold feel will

‘You will feel cold,’

33. Apna chataa yaad se rakhain

your umbrella Remember take

‘Remember to take your umbrella’

34. Aur apni barsaatian b

and your raincoat

‘and your raincoat.’


out going while be careful

‘Be careful when you are out,’

36. Jeb ap Gari chala rhy hun

when you (vehicle) driving are

‘when you are driving.’

37. Khas toer se Karachi k beshter ilaqu main

especially Karachi of most areas in

‘Especially in most areas of Karachi,’

38. Ziada barish ho ge

very rainy be will

‘it will be very rainy’

39. Aur taiz andhi

and very windy

‘and very windy.’

40. Trafic mutasir ho ga

traffic affected be will

‘The traffic will be affected,’

41. Ku k beshter ilaqu main mustaqil barish ho gi

because most areas over continuously rain be will

‘because it will rain continuously over most areas.’


53

Text associated with weather slide 4 (sunny, clear)

42. Ain mashraq ki terf dekhien

let’s east at look

‘Let’s look at the east.’

43. Mashraq main

east in

‘In the east,’

44. Pura din barpoor dohp mutwaky hai

all day a lot of sunshine expect is

‘we can expect a lot of sunshine all day.’

45. Zyada ter ilaqon main matla saaf rhy ga

most areas in the sky clear be will

‘Most areas, the sky will be very clear.’

46. Derja herarat kafi blund rhy ga

temperature quite high be will

‘The temperature will be quite high,’

47. jo ibtadaee subha k waqt apnay urooj pe ho ga

which early morning of time its highest on be will

‘and much higher in the early morning.’

48. Jabk Lahore k zyada ilaqon main

whereas Lahore of most areas in

‘But most areas of Lahore,’

49. Dohp dair tk rhy gi

sunshine last longer will

‘the sunshine will last longer.’

50. derja herarat main tezi k sath bohet garmi rhy gi

temperature in high of with very hotter be will

‘It will be very hotter with high temperature’

51. Dopehr gey tak

afternoon late by

‘by the late afternoon,’

52. Mashraq se aany wali garm lehron ki weja se

east from coming heat wave due to

‘due to the heatwave coming from the east.’

53. Derja herarat bery ga

Temperatures climb up will

‘Temperatures will climb up to’

54. 30 se 40 degree celsius tak

30 to 40 degree Celsius upto

‘as high as 30 to 40 degree Celsius.’



‘Please be careful when you are out,’

56. Aik dasti chata rakna soodmand hai

a handy umbrella posession useful is


54

‘a handy umbrella could be useful.’

57. Jeb ap bahir gari chala rhy hun

when you out go ing are

‘When you are driving out,’

58. Khas ker Pur hujom sarkon per

especially crowded roads on

especially on the crowded roads,’

59. Pur sakon rehye aur ehtiyat beratye

calm be and careful be

‘remember to be calm and careful,’

60. Garm surat-e-hal ki wja se

hotter stituation, due to

‘due to the hotter stituation,’

61. Aur blund derja herarat

and higher temperature,

‘and the higher temperature,’

62. Is se traffic masail ho sakty hain

it cause traffic problems be can

‘it can cause the traffic problems.’

Text associated with weather slide 5 (partly sunny)

63. Ain eb darul hukomat Islamabad me mosam per nazr dalty hain

let’s now capital Islamabad in weather at look

‘Now let’s look at the weather in the capital Islamabad’

64. Log eb b dohp k lamhat se lutaf andoz ho sakty hain

people still sunshine moment enjoy can be

‘People can still enjoy a moment of sunshine’

65. Ku k juzwi toer se dohp rehay ga

because partly sunny be will

‘because it will be partly sunny’

66. Acha mosam shehr pe cha jay ga

fine weather city over spread will

‘Fine weather will spread over the city.’

67. Mosam boat garm rhy ga

weather very warm be will

‘The weather will be very warm.’

68. Gerchy chamakti dohp thori dair he rhy gi

though bright sunshine shortly stay will

The bright sunshine will stay shortly, though.

69. Jumarat ko asmaan badlon se daka rhy ga

Thursday on sky clouds by coverd be will

‘The sky will be covered by clouds on Thursday.’

70. Baadal door shamaal me jama hon gy

clouds far south in gather will

The clouds will gather in the far south.


55

71. Aur darul hukomat aur karibi ilaqon ki terf berien gy

and capital(city) and near areas upto move will

‘and move up to the capital city as well as cities in nearby areas.’

72. Lakin derja herarat zyada tabdil nhe ho ga

but temperature too much change not will

‘But the temperature will not change too much.’

73. Derja herarat blund ya am ho sakta hai

temperature high or low could be

‘The temperature could go up or go down’

74. 24 se 28 degree Celsius key kareeb

24-28 degree Celsius around

‘to around 24-28 degree celsius.’

75. Mujmuee toer pe mutla juzwi aber alood rehay ga

collectively sky partly cloudy remain will

‘In general, it will be partly cloudy,’

76. Shumal se aany wali hewaon ki wja se

north from coming breeze due to

‘due to a breeze coming from the north’

77. Aur us k bad shandar dohp ho gi

and afterwards brilliant sunshine be will

‘and afterwards there will be brilliant sunshine.’

Text associated with weather slide 6 (windy and rainy)

78. Ain maghrib ki terf dekhien

let’s west at( towards) look

‘Let’s look at the west.’

79. Subha main

morning in

‘In the morning,’

80. Badil batadreej bardh rehay gy

clouds gradually grow ing will

‘clouds will gradually accumulate,’

81. Aur ye inn ilaqun k darmiyan rhy gy

and these areas to center be will

‘and it will center mainly around these areas.’

82. Boat andhi chaly gi

very windy be will

‘it will be very windy.’

83. Ghaliban barish ho sakti hai ksi bi wakt

probably rain be can any time

‘It will probably rain at any time.’

84. Derja herarat zyada blund ya kam nhe ho ga

temperature too high or low not be would

‘The temperature wouldn’t be too high or too low.’

85. Log thori thand mehsos krien gy

people bit cold feel will


56

‘People will feel a bit cold.’

86. Sara din (matla) zyada ter aber alood he rhy ga

throughout day mostly cloudy be will

‘It will be mostly cloudy throughout the day.’

87. Khas ker sham main

especially evening in

‘Especially, in the evening,’

88. Wakfy wakfy se garaj chamak ho gi

time to time from thunder be will

‘thunder will appear from time to time,’

89. Aur isi doraan andhi be chaly gi

and at the same time windy be will

‘and at the same time it will be windy,’

90. ahistagi se barti Sard hewa k sath

slowly moving cold air with

‘with the cold air moving slowly’



‘Please be careful when you are out,’

92. Aik dasti chata rakna soodmand hai

a handy umbrella posession useful is

‘a handy umbrella could be useful,’

93. Aur apki barsati b

and your raincoat also

‘as well as your raincoat.’

94. Jeb ap bahir gari chala rhy hun

when you out go ing are

‘When you are driving out,’

95. Khas ker Pur hujom sarkon per

especially crowded roads on

‘especially on the crowded roads,’

96. yad rakhye pur sakon rehna hai aur ehtyat beratni hai

remember to calm be and careful be

‘remember to be calm and be careful,’

97. Khrab mosam ki wja se

bad weather due to

‘due to the bad weather,’

98. Aur waqt ba waqt chalti andhi

and time to time blowing wind

‘and the wind blowing from time to time,’

99. Is se trafic masail ho sakty hain

it cause traffic problems be can

‘it can cause the traffic problems.’

Text associated with weather slide 7 (weather warnings)


57

100. Akhar mai aik nazar sahili ilaqun pe

lastly in a look coastal areas upon

‘Lastly, let’s have a look at the coastal areas.’

101. Door janub main

far south in

‘Here in the far south,’

102. Derja hararat maloom nai kia ja sakta

temperature known not do go can

‘the temperature during the day couldn’t be expectable.’

103. Loag mosami italat se juray rehain

people weather report to connected be

‘People need to stay tuned for the weather report,’

104. Aur trafic k halaat baray

and traffic of situation about

‘and the traffic report,’

105. Khasusan subo sweray bheer k waqt

especially morning early rush of time

‘especially in rush hours like early in the morning’

106. Ya pahir shamko

Or then evening

‘or in the evening,’

107. Jab har aik gar se kam k liay sarak pe hota hai

when everyone home from work for roads on be would

‘when everyone are on the roads from home to work.’

108. Samandaer se uthnay walay ganay badil

ocean from rising massive clouds

‘due to massive clouds from the ocean,’

109. Moosla dhahar barish ka baies bntay dikhaeee de rehay hain

cats and dogs rain of cause becoming are

‘it will rain like cats and dogs.’

110. Lihaza tofani barish se shadeed sailaab mutwaky hai

so heavy rain from severe flood expected is

‘So, severe flooding from heavy rain is expected,’

111. Braye mehrbani ehtyat kijye

please careful is

‘please be careful,’

112. Ghair zaruri safar/drive se ijtinab kijeay

non-urgent drive from avoid do

‘avoid non-urgent driving out of the house’

113. Jin ilaqun main sailab ki paishangoi ki gai hai

to areas in flood of warning issued is

‘to areas which are issued a flooding warning.’

114. Kuch sarkain band rehnien gi

few roads block be will

‘There will be a few road closures,’


58

115. Khasusan bheer k auqat main traffic musael paida ho saktay

especially rush of time in traffic problems create be can

‘especially in a rush hour that can causes the traffic problems.’

116. Humaray andazay main ganay badil mutwaqo hain

our pridiction in masssive clouds expected are

‘We predict that massive clouds will appear.’

117. Aur tez andhi ay gi

and srong wind come will

‘and there will be a strong wind.’

118. Natijatan garaj chamak k imkanat hain

as a result thunder light of chances are

‘This can cause thunder and lightning.’

119. Sahili ilaqu k janubi shehr kai roaz tak mutasar rehain gy

coastal areas of south cities several days to affected be will

‘Southern city in coastal areas will be affected for several days.’

120. Aj k mosam mai bus itna he... mazid jannay k liay humaray sath rehiay

today's weather in only this much more information for us with let's be

‘That's all for today's weather. Stay tuned.’


59

Appendix 4: Participants’ answer sheet

Part 1:

Listen to the following 42 audio files.

Decide whether you had heard these words in the weather report: Yes? No?

Don’t think, just tick! (You have only 2s for each item).

1. Yes No

2. Yes No

3. Yes No

4. Yes No

5. Yes No

6. Yes No

7. Yes No

8. Yes No

9. Yes No

10. Yes No

11. Yes No

12. Yes No

13. Yes No

14. Yes No

15. Yes No

16. Yes No

17. Yes No

18. Yes No

19. Yes No

20. Yes No

21. Yes No

22. Yes No

23. Yes No

24. Yes No

25. Yes No

26. Yes No

27. Yes No

28. Yes No

29. Yes No

30. Yes No

31. Yes No

32. Yes No

33. Yes No

34. Yes No

35. Yes No

36. Yes No

37. Yes No

38. Yes No

39. Yes No

40. Yes No

41. Yes No

42. Yes No

Part 2:

Look at the icon shown on the screen AND listen to the audio file.

Decide whether the audio file represents the icon: Yes? No?

Don’t think, just tick! (You have only 4s for each item).

1. Yes No

2. Yes No

3. Yes No

4. Yes No


60

5. Yes No

6. Yes No

Appendix 5: List of tested items (filler and target items)

Part 1

1 Juma FW

2 Nujoom FW

3 Barid FW

4 Ehtiat TW

5 Shumal TW

6 Kursi FW

7 Traffic TW

8 Janub TW

9 Chataa TW

10 Hafta FW

11 Thalj FW

12 Kharif FW

13 Mosam TW

14 Itwaar FW

15 Lahore TW

16 Bahaar FW

17 Mosam garma FW

18 Arabah FW

19 Ilaqay TW

20 Budh FW

21 Khazaan FW

22 Degree TW

23 Phool FW

24 Fannan FW

25 Kutta FW

26 Mashraq TW

27 Sayara FW

28 Dopeher TW

29 Karachi TW

30 Thand TW

31 Rathath FW

32 Badal TW

33 Andhi TW

34 Chala TW

35 Sarku TW

36 Mard FW

37 Derja herarat TW

38 Parinda FW

39 Lail FW

40 Sham TW

41 Barsaatian TW

42 Daftar FW

TW – target

word

FW – filler word

Part 2

1 Barf TW

2 Fajr FW

3 Dohp TW

4 Garja chamak TW

5 Barish TW

6 Janwar FW


61

Appendix 6: Frequency of target words

Note: The accuracy in classification of frequent and infrequent items might vary to some degree, due to multiple

acceptable orthographic variations in Urdu, e.g. ilaquay and its variations ilaqu, ilaqon, ilaqun.

Target words Frequent Infrequent Target words Frequent Infrequent

Ehtyat, ahtiat 6x Thand 3x

Shumal(i) 4x Badal

(badil/baadal)

2x

Traffic (trafic) 5x Andhi 5x

Janub(i) 4x Chala 3x

Chataa 1x Sarku 1x

Mosam(i) 9x Derja herarat 9x

Lahore 1x Sham 3x

Ilaqay/ilaqu/

ilaqon/ ilaqun

11x Barasaatian

(barsati)

2x

Degree 5x Barf 2x

Mashraq 3x Dhop (doph) 7x

Dopehr 1x Garaj chamak 1x

Karachi 1x Barish 6x


62

Appendix 7: Raw scores for target items (Task 1)

Legend:

T4 – target item

L7 – loanword

P15 – place name

No. Gender T4 T5 L7 T8 T9 T13 P15 T19 L22 T26 T28 P29 T30 T32 T33 T34 T35 T37 T40 T41TOTAL

CORRECT% CORRECT

1 M 1 1 1 0 1 1 1 1 1 0 0 1 0 1 0 1 0 0 1 0 12 60.00%

2 F 0 1 1 0 0 1 1 1 1 1 0 1 1 0 1 1 1 1 0 1 14 70.00%

3 F 0 0 1 0 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 3 15.00%

4 F 0 1 1 0 0 0 1 1 0 1 0 0 0 1 0 1 1 0 0 0 8 40.00%

5 F 0 1 0 0 1 0 1 0 1 1 0 1 1 1 1 1 1 1 1 1 14 70.00%

6 F 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 5.00%

7 M 0 0 1 0 0 0 1 0 1 0 0 1 0 0 1 0 0 0 0 0 5 25.00%

8 F 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 10.00%

9 F 1 1 1 0 0 0 0 0 1 0 0 1 0 0 1 0 0 0 1 0 7 35.00%

10 F 0 0 0 1 1 0 1 0 1 0 0 0 0 1 0 1 1 1 1 0 9 45.00%

11 F 1 0 1 1 1 1 1 0 1 0 1 1 1 0 1 0 0 0 1 1 13 65.00%

12 F 0 1 1 1 1 0 1 0 1 1 1 1 0 1 1 0 0 0 0 1 12 60.00%

13 F 0 0 1 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 2 10.00%

14 F 0 0 1 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 2 10.00%

15 F 0 1 0 0 1 0 0 0 0 1 0 1 0 0 0 1 1 0 1 1 8 40.00%

16 F 1 1 1 0 1 0 1 0 1 0 1 1 0 0 1 1 0 1 1 0 12 60.00%

17 F 1 0 0 1 0 1 0 0 1 0 1 0 1 0 1 0 0 0 1 0 8 40.00%

18 F 1 1 1 0 0 0 0 0 1 0 0 0 0 0 0 1 1 1 1 0 8 40.00%

19 F 0 0 1 0 0 1 0 0 1 0 0 1 1 1 0 0 1 0 0 0 7 35.00%

20 F 0 1 1 1 1 1 0 0 1 0 1 1 0 1 1 1 0 0 1 0 12 60.00%

21 M 0 1 1 0 1 1 1 0 1 1 1 1 0 1 1 0 1 1 1 0 14 70.00%

22 F 1 1 1 0 0 0 1 1 1 0 1 0 1 0 1 0 1 0 0 0 10 50.00%

23 F 0 0 1 0 0 1 0 0 1 0 0 0 1 0 0 0 0 0 0 0 4 20.00%

24 F 0 0 1 0 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 3 15.00%

25 M 0 1 1 1 1 1 1 1 1 1 1 1 0 1 0 1 0 1 1 0 15 75.00%

26 M 1 1 1 0 1 1 1 0 1 0 0 1 1 1 1 1 1 0 1 0 14 70.00%

27 F 0 0 1 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 2 10.00%

28 F 1 0 1 1 1 1 1 1 1 0 1 0 1 1 1 0 1 0 1 0 14 70.00%

29 F 0 0 1 1 0 0 1 0 1 0 0 1 1 0 0 0 0 0 0 0 6 30.00%

30 F 0 1 1 1 0 0 1 0 1 1 0 0 0 1 0 1 0 1 0 1 10 50.00%

31 M 0 0 0 1 0 0 0 1 1 0 0 1 1 0 1 0 0 0 0 0 6 30.00%

32 F 0 0 1 0 0 0 0 0 1 0 0 1 0 0 0 0 0 0 0 0 3 15.00%

33 M 0 1 1 1 1 1 1 1 1 0 0 0 0 0 0 1 0 1 0 1 11 55.00%

34 F 0 1 1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 0 0 8 40.00%

35 F 1 1 1 0 1 0 0 0 1 0 0 0 0 1 0 1 1 1 1 0 10 50.00%

11 18 30 12 16 12 17 8 31 8 10 18 11 13 14 15 13 10 15 7 289 41.29% 35

31.43% 51.43% 85.71% 34.29% 45.71% 34.29% 48.57% 22.86% 88.57% 22.86% 28.57% 51.43% 31.43% 37.14% 40.00% 42.86% 37.14% 28.57% 42.86% 20.00%

TARGET RESPONSES (TASK 1 - WORD RECOGNITION)

NATIVE

MONO

NATIVE BI

NATIVE MULTI

NON-NATIVE

6

14

10

5


63

Appendix 8: Raw scores for filler items (Task 1)

Legend:

U1 – Urdu filler item

A2 – Arabic filler item

No. Gender U1 A2 A3 A6 U10 A11 A12 U14 A16 U17 A18 U20 U21 U23 A24 U25 A27 U31 U36 Q38 A39 A42TOTAL

CORRECT% CORRECT

1 M 1 1 0 1 1 1 0 1 0 1 0 0 1 1 0 0 0 0 1 0 0 0 10 45.45%

2 F 1 0 0 0 0 0 1 1 1 1 0 1 0 1 0 0 1 1 1 1 1 1 13 59.09%

3 F 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 22 100.00%

4 F 1 1 0 1 1 1 1 0 1 1 1 1 1 0 0 0 1 1 1 1 0 0 15 68.18%

5 F 1 0 0 1 1 0 0 1 0 0 0 1 0 0 1 0 0 1 1 1 0 1 10 45.45%

6 F 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 22 100.00%

7 M 1 1 0 1 1 1 1 0 0 1 1 0 0 0 1 1 1 0 1 1 1 1 15 68.18%

8 F 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 21 95.45%

9 F 1 1 0 1 1 1 1 0 1 1 1 0 0 0 1 1 1 1 1 1 1 1 17 77.27%

10 F 0 1 0 0 0 1 0 1 0 1 1 1 0 0 1 0 0 1 0 1 1 0 10 45.45%

11 F 0 0 1 1 1 0 0 1 1 1 0 0 1 0 0 1 1 1 0 1 1 1 13 59.09%

12 F 0 1 0 1 0 1 1 0 0 0 1 0 1 1 1 0 1 1 0 0 1 1 12 54.55%

13 F 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 22 100.00%

14 F 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 22 100.00%

15 F 1 0 0 0 1 0 0 0 1 0 1 0 0 1 0 0 1 0 1 0 1 0 8 36.36%

16 F 1 1 1 1 1 1 1 0 1 0 0 0 1 0 1 1 1 0 1 0 1 0 14 63.64%

17 F 0 1 1 1 0 1 1 0 0 1 1 0 1 0 1 1 1 1 0 1 1 0 14 63.64%

18 F 1 1 1 1 1 1 1 0 0 0 0 1 1 1 1 1 1 0 1 0 1 1 16 72.73%

19 F 1 1 0 0 1 1 0 0 1 0 0 1 0 1 1 1 0 1 0 1 0 1 12 54.55%

20 F 1 0 0 0 1 1 0 1 0 1 0 0 0 1 1 0 1 1 0 0 1 1 11 50.00%

21 M 0 0 1 1 0 1 1 1 0 1 1 1 0 0 1 1 1 1 0 1 1 0 14 63.64%

22 F 1 1 0 1 1 0 0 0 0 1 1 0 1 0 0 1 1 1 0 1 0 1 12 54.55%

23 F 1 1 1 1 1 0 1 1 1 1 1 1 1 0 1 1 1 0 1 1 0 1 18 81.82%

24 F 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 0 1 1 1 20 90.91%

25 M 0 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 0 8 36.36%

26 M 1 1 0 0 1 1 1 0 0 1 0 0 1 0 0 1 0 0 0 1 0 0 9 40.91%

27 F 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 21 95.45%

28 F 0 0 1 1 1 1 1 1 0 0 1 0 0 0 0 1 1 0 0 1 1 0 11 50.00%

29 F 1 0 1 1 1 1 1 0 1 1 1 1 1 1 1 0 1 1 1 1 1 1 19 86.36%

30 F 1 1 0 1 1 1 1 1 0 0 1 1 1 0 1 1 0 0 1 1 1 1 16 72.73%

31 M 0 1 1 1 0 1 0 1 0 1 1 1 1 0 1 1 1 1 1 0 1 1 16 72.73%

32 F 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 21 95.45%

33 M 1 0 0 1 0 1 0 0 1 1 0 0 1 0 0 1 0 1 0 1 1 1 11 50.00%

34 F 1 0 0 1 0 1 1 0 0 1 1 1 1 1 1 0 0 0 0 1 1 0 12 54.55%

35 F 1 1 1 0 1 1 0 0 1 1 1 1 1 1 1 1 0 1 1 0 0 1 16 72.73%

27 25 19 28 27 28 23 19 19 26 24 21 24 15 25 23 25 25 22 27 27 24 523 67.92% 35 URDU

77.14% 71.43% 54.29% 80.00% 77.14% 80.00% 65.71% 54.29% 54.29% 74.29% 68.57% 60.00% 68.57% 42.86% 71.43% 65.71% 71.43% 71.43% 62.86% 77.14% 77.14% 68.57% ARABIC

6NATIVE

MONO

14NATIVE

BI

10NATIVE

MULTI

5NON-

NATIVE

FILLER RESPONSES (TASK 1 - WORD RECOGNITION)


64

Appendix 9: Raw scores on target and filler items (Task 2)

Legend:

Q1, 3-5 – Target items

Q2 – Arabic filler item

Q6 – Urdu filler item

No. Gender Q1 Q2 Q3 Q4 Q5 Q6TOTAL

CORRECT% CORRECT

1 M 0 0 0 1 1 0 2 33.33%

2 F 0 1 0 0 0 1 2 33.33%

3 F 0 1 0 0 0 0 1 16.67%

4 F 1 1 0 1 1 1 5 83.33%

5 F 1 1 1 1 1 1 6 100.00%

6 F 1 0 0 0 0 1 2 33.33%

7 M 0 0 1 0 0 0 1 16.67%

8 F 0 0 1 0 0 0 1 16.67%

9 F 0 0 1 0 0 0 1 16.67%

10 F 0 1 0 1 1 1 4 66.67%

11 F 0 1 1 1 1 1 5 83.33%

12 F 0 0 1 0 1 0 2 33.33%

13 F 1 1 1 1 0 0 4 66.67%

14 F 0 1 0 1 1 1 4 66.67%

15 F 0 1 0 0 0 1 2 33.33%

16 F 0 1 0 0 0 1 2 33.33%

17 F 0 1 1 0 1 0 3 50.00%

18 F 1 1 1 0 0 1 4 66.67%

19 F 0 1 1 1 0 1 4 66.67%

20 F 0 0 0 0 1 0 1 16.67%

21 M 0 0 1 1 1 0 3 50.00%

22 M 0 0 0 1 1 0 2 33.33%

23 M 1 0 1 0 0 0 2 33.33%

24 F 0 0 1 0 1 1 3 50.00%

25 F 0 1 0 1 1 0 3 50.00%

26 F 0 1 0 0 0 1 2 33.33%

27 F 1 1 0 1 0 1 4 66.67%

28 F 0 0 0 0 0 1 1 16.67%

29 F 0 1 0 0 0 0 1 16.67%

30 F 0 0 1 0 1 0 2 33.33%

31 M 1 1 0 1 1 1 5 83.33%

32 M 0 1 0 0 0 1 2 33.33%

33 F 0 0 1 0 0 1 2 33.33%

34 F 0 1 0 1 1 1 4 66.67%

35 F 0 1 1 0 0 1 3 50.00%

8 21 16 14 16 20 95 45.24% Arabic filler

7 22.86% 60.00% 45.71% 40.00% 45.71% 57.14% Urdu filler

NATIVE MONO

NATIVE BI

NATIVE MULTI

NON-NATIVE


65

Appendix 10: Values for sensitivity (d’) and response bias (β) for Task 1

All participants(35)

miss rate d' 0.245

signal 289 411

hit rate signal % 41% 59%

filler 247 523

filler % 32% 68% β 1.087749

Native Monolingual participants (6)

miss rate d' 0.348

signal 52 68


filler 40 92

filler % 30% 70% β 1.126236

Native Bilingual participants (14)

miss rate d' 0.146

signal 107 173


filler 101 207

filler % 33% 67% β 1.055852

Native Multi-lingual participants (10)

miss rate d' 0.347

signal 92 108


filler 72 148

filler % 33% 67% β 1.099731


66

Non-native participants (5)

miss rate d' 0.193

signal 38 62


filler 34 76

filler % 31% 69% β 1.080643


67

Appendix 11: Values for sensitivity (d’) and response bias (β) for Task 2

All participants 35)

miss rate d' -0.056

signal 53 87


filler 28 42

filler % 40% 60% β 0.984402

Native Monolingual participants (6)

miss rate d' 0.220

signal 10 14


filler 4 8

filler % 33% 67% β 1.073177

Native Bilingual participants (14)

miss rate d' -0.047

signal 21 35


filler 11 17

filler % 39% 61% β 0.986289

Native Multi-lingual participants (10)

miss rate d' -0.572

signal 15 25


filler 12 8

filler % 60% 40% β 0.9815


68

Non-native participants (5)

miss rate d' 0.896

signal 7 13


filler 1 9

filler % 10% 90% β 2.110556

References

Carroll, S. E. & Meisel, J. M. (2015). Input, learner populations, and the human language-

making capacity. In C. Hamann & E. Ruigendijk (eds.), Language Acquisition and

Development. Proceedings of GALA 2013. Newcastle upon Tyne: Cambridge Scholars

Publishing, 1-16.

Carroll, S. E. & Windsor, J. W. (2015). Segmental targets versus lexical interference.

Production of second-language targets on first exposure and the result of minimal

training. In Peukert, H. (ed.), Transfer effects in multilingual language development.

Amsterdam: John Benjamins Publishing Company, 53-86.

Carroll, S. E. (2014). Processing ‘words’ in early stage SLA: A comparison of first exposure

and early stage learners. In Z-H. Han & Rast, R. (eds.), Input processing in second

language acquisition. Cambridge: Cambridge University Press, 107-138.

Carroll, S. E. (2013). Introduction to the special issue: Aspects of word learning on first

exposure to a second language. Second Language Research 29(2): 131-144.

Carroll, S. E. (2012). First exposure learners make use of top-down lexical knowledge when

learning words. In K. Braunmüller, Gabriel, C. & Hänel-Faulhaber, B. (eds.),

Multilingual individuals and multilingual societies. Amsterdam: John Benjamins, 23-

45.

Carroll, S. E. (2004). Segmentation: Learning how to ‘hear’ words in the L2 speech stream.

Transactions of the Philological Society 102: 227-254.

Carroll, S. E. (2001). Input and evidence: the raw material of second language acquisition.

Amsterdam: John Benjamins.

Denhovska, N., Serratrice, L. & Payne, J. (2016). Acquisition of second language grammar

under incidental learning conditions: the role of frequency and working memory.

Language Learning 66 (1), 159-190.

Flege, J. E. (2009). 'Give input a chance!. In T. Piske & Young-Scholten, M. (eds.), Input

matters in SLA. Bristol: Multilingual Matters, 175-190.

Gullberg, M., Roberts, L. & Dimroth, C. (2012). What word-level knowledge can adult learners

acquire after minimal exposure to a new language? International Review of Applied

Linguistics 50: 239-276.

Gullberg, M., Roberts, L., Dimroth, C. Veroude, K. & Indefrey, P. (2010). Adult language

learning after minimal exposure to an unknown natural language. Language Learning

60: 5-24.

Han, Z-H. & Liu, Z. (2013). ‘Input processing of Chinese by ab initio learners’. Second

Language Research 29: 145-164.

Klein, W. (1986). Second language acquisition. Cambridge: Cambridge University Press.

Piske, T. & Young-Scholten, M. (eds.) (2009). Input matters in SLA. Bristol: Multilingual

Matters.


69

Rast, R. (2010). The role of linguistic input in the first hours of adult language learning.

Language Learning 60: 64-84.

Ristin-Kaufmann, N. & Gullberg, M. (2014). The effects of first exposure to an unknown

language at different ages. Bulletin Suisse de Linguistique Appliquée, 17-29.

Schmidt, R. L. (1999). Urdu: An Essential Grammar. London: Routledge.

Shoemaker, E. & Rast, R. (2013). Extracting words from the speech stream at first exposure.

Second Language Research 29: 165-183.

Skehan, P. (1989). Individual differences in second-language learning. London: Edward

Arnold.

Stanislaw, H. & Todorov, N. (1999). Calculation of signal detection theory measures. Behavior

Research Methods, Instruments & Computers 31(1): 137-149.


70

EVA MALESSA



Newcastle upon Tyne


United Kingdom

[email protected]

REBECCA LIM



Newcastle upon Tyne


United Kingdom

[email protected]



Newcastle and Northumbria Working Papers in Linguistics 22 (2016) Jarrah

71

Weightlessness Preservation in Jordanian Arabic Varieties: A Stratal OT Analysis

Marwan Jarrah


[email protected]; [email protected]

Abstract This research explores the morphological form of 2nd person possessive clitics in Jordanian Arabic

varieties within Stratal OT. It assumes that possessive clitics formation in these varieties undergoes two

levels of constraint interaction, namely the stem level and the word level. Additionally, it argues that

the surface differences between these varieties regarding formation of 2nd person possessive clitics is

reducible to the relevant ranking of one constraint that bans weightless consonants of the input to have

moraic weight in the output, at the word level where inputs can be constrained. This constraint is coined

as NONFINAL(c)-IO. The argument is that in Mainstream Jordanian Arabic, NONFINAL(c)-IO

outranks LINEARITY-IO, hence the possibility for metathesis as a repair strategy when the input of

level 2 includes a weightless consonant (yielding, as a result, the surface form –ik). On the other hand,

in the so-called Karaki Arabic, the order between NONFINAL(c)-IO and LINEARITY-IO is reversed

(LINEARITY-IO >> NONFINAL(c)-IO), the matter which excludes metathesis as a repair strategy

(yielding, as a result, the surface form –ki). The study shows that other repair strategies, including

deletion and insertion, are blocked in Karaki Arabic, given that NONFINAL(c)-IO is outranked by

*COMPLEXONS, *COMPLEXCOD, MAX-IO, and DEP-IO.

1. Introduction

Jordanian Arabic (JA) is a blanket term that covers several sub-varieties, including Bani Hassan

Arabic, Rural Jordanian Arabic, and Ma’ani Arabic, to name just a few (see, Suleiman 1985;

Zawaydeh 1998; De Jong and Zawaydeh 1999; Zuraiq and Zhang 2006; and Rakhieh 2009).

Between these sub-varieties there are a number of phonological differences that involve a range

of phonological constructions, i.e., syllables, words, etc. One of these differences lies in the

formation of the 2nd Feminine Singular (2FS) possessive clitic, which is the principal concern

of the current research. For the purposes of the present study, I divide JA varieties into two sub-

varieties. The first variety is what I call Mainstream Jordanian Arabic (MJA), and the second

variety is Karaki Arabic (KA), spoken in Al-Karak City.1

In MJA, the 2FS possessive clitic is in most cases –ik, whereas it is always –ki in KA.

It is the ultimate aim of the current research to investigate this discrepancy in the order of the

segments that form the 2FS possessive clitic between these varieties. Consider table 1 that

shows the paradigm of 2nd person possessive clitics used in MJA and KA in addition to their

corresponding forms in Modern Standard Arabic (MSA) to the noun daftar ‘a note book’

(possessive clitics are bold-faced).

1 Al-Karak city is 140 kms to the south of Amman, the capital city of Jordan.

Jarrah Weightlessness preservation in Jordanian Arabic varieties

72

Table 1: 2nd possessive clitics in MJA, KA and MSA (P: Person; N: number; G:

Gender)2

P N G MJA KA SA

2nd

S

M daf.ta.rak daf.ta.rak daf.ta.ru.ka

F daf.ta.rik daf.tar.ki daf.ta.ru.ki

P

M daf.tar.kum daf.tar.kum daf.ta.ru.kum

F daf.tar.kin daf.tar.kin daf.ta.ru.kunna

As clear from table 1, 2nd possessive clitics in Arabic varieties under discussion do not have an

invariant morphological form. Rather, they display overt morphological agreement with the

possessor. For instance, if the possessor is 2SM, the possessive clitic used is different from

what is used with a 2PM possessor. A closer look at table 1 reveals that MJA and KA have the

same paradigm of possessive clitics except for 2SF. The immediate question that arises is

whether this difference boils down to some phonological principle/condition. Using the current

generative phonological enterprise, namely Stratal Optimality Theory (Kiparsky 2000 and

2003, and subsequent work), the current research argues that this difference is in essence

attributed to some constraint that demands extra-metrical material, in the sense of Hayes (1995),

to remain weightless in the output, the state of affairs I label as Weightlessness Preservation.

All below discussion is dedicated to defend this hypothesis. Section 2 provides some

preliminary observations and analysis. Section 3 explores the cases where all varieties have the

same form of the 2SF, i.e. -ki, namely when the base ends with a vowel. Here I focus on the

ensuing vowel lengthening, assuming it is forced by stress assignment. Section 4 includes the

main analysis, postulating that the switch between the segments forming the 2FS possessive

clitic occurs in MJA grammar, demanded by one constraint I coin as NONFINAL(c)-IO

(weightless consonants of the input remain weightless in the output). I argue that

NONFINAL(c)-IO is a highly ranked constraint in MJA grammar, outranking LINEARITY-

IO. The switch in the segments forming the 2FS possessive clitic is taken as a repair strategy

against the violation of NONFINAL(c)-IO (NONFINAL(c)-IO >> LINEARITY-IO). On the

other hand, in KA, NONFINAL(c)-IO is outranked by LINEARITY-IO, hence no switch in the

segments forming the 2FS possessive clitic follows (LINEARITY-IO >> NONFINAL(c)-IO).

Section 5 concludes the discussion with some pointers to further research.

2. Preliminary observations and analysis

Before addressing the main core of my hypothesis (the existence of a constraint that demands

weightlessness preservation), let us first explore the segmental make-up of 2nd possessive clitics

in MSA. A closer look at table 1 reveals that both MSA and KA have the same order of the

segments that form the 2SF possessive clitic. In both varieties, -ki is used. On the other hand,

they are different with respect to two facts. Firstly, in MSA the possessive clitic used is

preceded by the vowel [u]. Secondly, there is a switch between the segments that form the 2SM

possessive clitic in KA, if we take MSA as a departure point. Consider table 2:

Table 2: 2nd person possessive clitics in KA and MSA

KA MSA

ak uka

ki uki

kum ukum

kin ukin

2 For some speakers of MJA and KA, [m] is dropped in the 2PM possessive clitic. In the following discussion, I

put this tendency aside in favour of the unmarked situation at no peril to the analysis.


73

Regarding the first fact (the possessive clitic in MSA is preceded by the vowel [u]). Much

evidence can be adduced for the assumption that this vowel [u] is not a genuine part of the

possessive clitic, but a morphological reflection of the overt Case MSA enjoys. It is widely

known that one of the main syntactic differences between MSA and other Arabic varieties,

including KA and MJA, is the lack of overt morphological Case on the part of the latter, where

Case is abstractly expressed. No overt Case markers are used in vernacular Arabic varieties, a

result broadly assumed by many works on Arabic (see, e.g. Holes 2004; Ryding 2005; and Fassi

Fehri 2012). Compelling evidence that [u] is a morphological realization of (nominative) Case

rather than an underlying part of the possessive clitic comes from the fact that it changes to [a]

or [i], depending on the structural position that the respective word occupies in the associated

sentence. Consider the following examples:

(1) a. marar-tu bi-baab-i-ki

Passed-1S with-door-GEN-your.2SF

‘I passed along your door.’

b. ʕaraf.tu baab-a-ki

knew1S door-ACC-your.2SF

‘I knew your door.’

In (1a), the vowel preceding the possessive clitic is [i] instead of Nominative [u]; the word

baabiki ‘your book’ is the object of the preposition bi ‘in’, hence the environment of Genitive

Case assignment. In (1b), the vowel changes to [a], since the word baabaki ‘your book’ is the

direct object whose Case must be Accusative in Arabic grammar (Holes 2004 and Wright and

Caspari 2011). Examples in (1) demonstrate clearly that the vowel preceding –ki in MSA is a

Case-sensitive form rather than a part of the underlying make-up of the possessive clitics.

Concerning the second fact pertaining to the switch between the segments forming 2SM

possessive clitic (the possessive clitic is VC in KA, while it is CV in MSA), one might entertain

the possibility that the underlying form of 2SM possessive clitic in MSA is different from that

of KA, hence the difference in the surface form. This possibility must be ruled out, though. As

indicated in table 2, the switch of the segments forming 2nd possessive clitics in KA is only

related to the 2SM clitic, as compared to MSA. There is no conceptual ground or actual

motivation to single out this clitic with a different underlying form, namely VC. Pertinently,

the 2SF possessive clitic in MJA is VC (-ik), while it is CV in KA and MSA (-ki). It is hardly

plausible to suggest that MSA and KA share the same underlying form for the 2SF possessive

clitic, while they are at odds as far as the 2SM possessive clitic is concerned. It will become

clear that the underlying form of possessive clitics in all Arabic varieties under discussion is

the same and any surface differences must come down to some phonological reasons.

3. Possessive clitics and vowel lengthening

Table 2 indicates that MJA maintains a different order of the segments that form the 2SF

possessive clitic. However, this difference falls when words ending with a vowel are taken into

account. Consider the following examples from MJA:

(2) /dawa/ ‘medicine’ → [da.'waa.ki] ‘your medicine’

/sama/ ‘sky’ → [sa.'maa.ki] ‘your sky’

/mawla/ ‘leader’ → [maw.'laa.ki] ‘your leader’

/ʔabu/ ‘father’→ [ʔa.'buu.ki] ‘your father’


74

In (2), the 2SF possessive clitic has the form CV rather than VC. With this being the case, the

following observation can be made, capturing the morphological form of the 2FS in MJA:

(3) In MJA, when the word ends with a vowel, the form of 2SF is –ik; otherwise the form

is -ki.

Though the observation in (3) might account for all cases in my MJA data, the question

remaining unanswered is why this must be the case. The observation in (3) is descriptive in

character without any explanatory value. Additionally, the observation in (3) says nothing about

KA and MSA, where the 2SF possessive clitic has an invariant form (i.e. CV), regardless of the

value of the final segment of the base (vocalic vs. consonantal).

Of note here is the fact that the final vowel of all examples in (2) must undergo vowel

lengthening in MJA and KA. One possibility is that the underlying form of the 2SF possessive

clitic in MJA is VCV, where the consonant and the final vowel have an already determined

value as –ki. The value of the initiating vowel is determined by context (i.e., phonological

environment) when suffixation of the possessive clitic to the base occurs. This possibility might,

at face value, account for vowel lengthening occurring to the final vowel of the base. However,

it must be eliminated since no vowel lengthening occurs when the base ends with a consonant

(Consider table 1). Even if we grant that the initiating vowel of the putative VCV might be any

phonological material (e.g. a consonant), no gemination to the final consonant is demanded,

implying the inescapable conclusion that vowel lengthening must be ascribed to a different

factor rather than the underlying make-up of the possessive clitic.

A promising possibility is that vowel lengthening under such cases is forced by stress

assignment (rules). According to Jordanian Arabic assignment rules (Btoosh 2006 and Abbas

2012), the penultimate syllable is typically the locus of the primary stress unless the final

syllable is superheavy. When the possessive clitic is attached to the end of the base, the

penultimate syllable is what precedes the possessive clitic. If we assume that in MJA and KA

grammars, the vowel in the syllable with CV shape gets lengthening when assigned primary

stress (CV → CVV/ V’), vowel lengthening forced by suffixation of the possessive clitic

follows. Several works argue convincingly that the syllable with primary stress tends to be

heavy (cf. Crosswhite 2001; De Jong 2004; and Fudge 2015). This tendency is formulated

within the OT-theoretical account by Prince and Smolensky (1993) as a markedness constraint

that requires heavy syllables to be stressed:

(4) WSP (Weight-To-Stress-Principle)

Heavy syllables are prominent both on the grid and foot structure.

If the base forms in all examples in (2) are used, primary stress falls on the first syllable. On

the other hand, when the possessive clitic is attached to the base, primary stress migrates to

lodge on the penultimate syllable, CV, which feeds the ensuing vowel lengthening. The

question now is as why no vowel lengthening is required when the base ends with a consonant.

Following WSP, it can be advanced that the existence of a coda (i.e., the consonant) suffices

for rendering the penultimate syllable heavy when the possessive clitic is attached to the base.

This argument fits KA, given that the final consonant of the base remains a coda to the same

syllable when suffixation of the possessive clitic applies (daf.tar → daf.tar.ki). On the other

hand, in MJA, the final consonant of the penultimate syllable becomes an onset to the following

syllable (daf.tar → daf.ta.rik), the matter which makes the stress fall on the anti-penultimate

syllable (daf.tar → 'daf.ta.rik). It seems that stress does not fall on a syllable whose code is

annexed to the following syllable. Anyway, I leave this issue open pending further research.


75

6 Main analysis

We are left with the main question which is as why there is a switch in the segments that form

the 2SF possessive clitic in MJA. One might raise the preliminary interrogation as what is the

evidence that the morphological switch happens in MJA not in KA or MSA? Put it a different

way, is –ki the underlying form, and –ik is the variant? Above, I treated MSA as the departure

point but revealing the actual underlying form is of paramount importance for any sound

generalization we need reach. Contingent on the available data, it can be advanced that our

departure point is right; KA and MSA display the unaffected cases regarding the morphological

form of the 2SF possessive clitic, and the switch thus occurs in MJA grammar. Support for this

contention comes from the full paradigm of possessive clitics in the three varieties under

discussion. Consider table 3.

Table 3: The full paradigm of the possessive clitics in MJA, KA, and MSA

P N G MJA KA MSA

1st

S

M -i -i -i

F -i -i -i

P

M -na -na -na

F -na -na -na

2nd

S

M -ak -ak -ka

F -ik -ki -ki

P

M -kum -kum -kum

F -kin -kin -kun.na

3rd

S

M -uh -uh -hu

F -ha -ha -ha

P

M -hum -hum -hum

F -hin -hin -hun.na

A closer look at table 3 implies that the morphological make-up of the possessive clitics can be

decomposed. [k] appears only with possessive clitics whose person value is fixed as [2nd],

whereas [h] with those with person valued as [3rd]. As such, [k] and [h] are the morphological

realizations of 2nd and 3rd person values, respectively. The question here is as what is about the

possessive clitics with 1st person value? There are two possibilities. The first one is that these

clitics do not have person feature, hence the appearance of such clitics without a segment

dedicated for person.3 However, this possibility blocks a unified account of possessive clitics

in Arabic varieties, given that possessive clitics with 2nd and 3rd person values show overt

realization of person feature. This gives rise to the second possibility that 1st person is marked

with zero morpheme. In other words, 1st person feature is present with no phonetic spell-out.

Let us assume the second possibility to be the case, the morphological paradigm of person in

Arabic varieties is summarized in table 4.

Table 4: the morphological paradigm of person in Arabic varieties

PERSON 1st 2nd 3rd

Lexically valued as ∅ k h

An important point here is the fact that in almost all cases the segment showing the person

feature initiates the possessive clitic, a part from the 2SF possessive clitic in MJA and 3SM in

3 We cannot assume that [n] is the morphological realization of 1st person because 1st singular cases lack it and [n]

co-occur with 2nd and 3rd person with feminine gender.


76

MJA and KA. Following this observation, -ki is the base whereas -ik is the form with switched

segments. Accordingly, our main question is now crystalized as why the segments that form

the 2SF possessive clitic undergo metathesis in MJA.

I argue that the answer for this question lies in extrametricality (Liberman and Prince

1977; McCarthy 1979; Hayes 1982; Hayes 1995; Hammond 1999; Kiparsky 2003; Watson

2007). Following Hayes (1995), the final consonant in the word daftar ‘a note book’ is extra-

metrical being a coda of a CVC syllable that appears word-finally. In order to form a word with

a possessive clitic in Arabic grammar, the possessive clitic must be suffixed to the base, which

is here the noun. On the basis of the MJA data and as far as the 2SF possessive clitic is

concerned, the assumption is that the switch in the possessive clitic segments follows from the

demand that the final consonant in the base must be weightless in the output even when the

possessive clitic is attached to it. Everything else is equal, what is common between the base

daftar ‘a note book’ and the output daf.ta.rik ‘your book’ is that [r] remains weightless in both

of them. On the other hand, while [r] in the base daf.tar is weightless in KA, it has weight in

the output daf.tar.ki; [r] is not an onset nor an extrametrical element because it loses its

peripherality once the possessive clitic is attached to the base. Empirical evidence for that [r]

contributes to the syllable weight in KA but not in MJA is that the primary stress falls on the

penultimate stress daf.'tar.ki, while it falls on the antepenultimate syllable in MJA 'daf.ta.rik.

In the OT-theoretical account, extrametricality is reformulated as NONFINALITY

Constraint, though NONFINALITY has been slightly different from extrametricality in the

sense that its focus is stress peaks (Prince and Smolensky 2008). NONFINAL, the OT successor

of extrametricality, is formulated as follows:

(5) NONFINAL: The prosodic head of the word does not fall on the word-final syllable

In her turn, Crowhurst (1996: 415) defines NONFINAL as:

(6) NONFINAL: The final syllable of a PrWr is not stressed4

With other formulations of NONFINAL, Hyde (2003) assumes that NONFINAL is a general

mechanism for achieving descriptive invisibility through focusing on the parsability of the final

segment (be it the final foot, syllable or mora)

(7) NONFINAL-SEG (μ, σ, F, PR): The final (mora, syllable, foot or prosodic word) is

not parsed into a higher prosodic structure

Accordingly, the NONFINALITY constraint has different permutations depending on the final

‘material’ intended. The following constraints can be thus noted (cf. Al-Jarrah 2011: 5):

(8) NONFINAL(c): The final consonant is not parsed into the next higher prosodic constituent

(i.e. the syllable)

NONFINAL(v): The final vowel is not parsed into the next higher prosodic constituent (i.e.

the syllable)

NONFINAL(σ): The final syllable is not parsed into the next higher prosodic constituent

(i.e. the foot)

NONFINAL(F): The final foot is not parsed into the next higher prosodic constituent (i.e.

the prosodic word)

4 PrWr stands for Phonological Word.


77

NONFINAL(PR): The final prosodic word is not parsed into the next higher constituent (i.e.

the lexical word)

The main concern of the current research is the first version, NONFINAL(c). Crucial here is

that weightlessness preservation (i.e. weightless material of the input remains weightless in the

output) cannot be pursed utilizing the full strict parallel OT theory (Prince and Smolensky 1993,

McCarthy and Prince 1993a, b). Firstly, weightlessness preservation cannot be reformulated as

a faithfulness constraint. That is because strict parallel OT has no mechanism to allow inputs

to be constrained. Assuming that weightless segments of the input remain weightless in the

output leaves us with the serious problem of constraining the input, and hence violating

Richness of the Base Principle. Secondly, if weightlessness preservation is formulated as a

markedness constraint, we must do away with the identity relation, which is nonetheless

required to keep weightless segments of the base/input as such in the output. I argue that the

puzzle of accommodating weightlessness preservation within OT can be resolved if we appeal

to the Stratal OT (Kiparsky 2000 and 2003, and subsequent work).

Stratal OT consists of several serially ordered mappings from input to output. Consider figure

(1) which shows how Stratal OT works (Kager 1999: 382).

Input

Stratum 1: GEN, EVAL

Stratum 2: GEN, EVAL

Output

Figure 1: Stratal OT Design

In Stratal OT, the output of each stratum serves as the input of the following one (cf. Rakhieh

2009: 41). Following Kiparsky (2003), there are three levels of constraint interaction, namely

stem level, work level and post-lexical level. Let us assume that the possessive clitics are

adjoined to the given noun in the second level (word level) whose input is originally the output

of the stem level. Additionally, since the input of level 2 is actually an output of level 1, it can

be advanced that this input (of level 2) is subject to constraints: the structure of the output of

Level 1 can stand in an identity relation with a candidate at Level 2.5 Said this, it can be now

suggested that weightless consonants of the input of level 2 must remain weightless in the

output of level 2. Weightlessness preservation can be translated into the following faithfulness

constraint which I name as NONFINAL(c)-IO:

(9) NONFINAL(c)-IO: weightless consonants of the input remain weightless in the output.

Let us now explore how NONFINAL(c)-IO can provide us with an account of the behaviour of

the 2nd possessive clitic in Arabic varieties under discussion. When the input of level 2 ends

5 Thanks for the NNWPL anonymous reviewer for pointing this to me.


78

with an extrametrical consonant and the possessive clitic is suffixed to the noun, peripherality

is no longer a possible option to keep the extra-metrical consonant of the base. One direct way

is through re-syllabification. The extra-metrical consonant of the base changes to an onset of

the following syllable rather than remaining a coda of a word-medial syllable whose coda must

be parsed. When the extra-metrical consonant of the input of level 2 becomes an onset, it has

no metrical value, the wanted result.

(10) /daftar + ki/ → [daf.ta.rki]

What appears however, problematic at this point is that the re-syllabification process leads to a

complex onset, the matter which is totally prohibited in MJA grammar. *COMPLEXONS is a

highly ranked constraint.

(11) *COMPLEXONS (Prince and Smolensky 1993):

A syllable must not have more than one onset segment.

If we assume that the two NONFINAL(c)-IO and *COMPLEXONS are highly ranked whose

violations incur fatal violations in MJA grammar, repair strategies must be invoked. One of

these strategies is to delete /k/.

(12) /daftar + ki/ → [daf.ta.ri]

When /k/ is deleted, the two NONFINAL(c)-IO and *COMPLEXONS are satisfied given that /k/

is weightless and is not part of a complex onset, respectively. However, this repair strategy

must be eliminated given that MAX-IO is highly–ranked constraint in MJA grammar as well,

thereby the impossibility to delete /k/.

(13) MAX-IO (McCarthy & Prince 1995):

Every segment of S1 has a correspondent in S2.

Another possibility is to undo the consonant cluster resulted from the re-syllabification process

by an epenthetic vowel, inserted between the members of the consonant cluster:

(14) /daftar + ki/ → [daf.ta.ri.ki]

However, this output seems less-harmonic for violating the highly-ranked DEP-IO.

(15) DEP-IO (McCarthy & Prince 1995):

Every segment of S2 has a correspondent in S1 (S2 is “dependent on” S1).

A promising possibility is to undergo metathesis between the second part of the consonant

cluster and the vowel, resulting in the most-harmonic candidate.

(16) /daftar + ki/ → [daf.ta.rik]

Even if the candidate in (16) violates LINEARITY-IO, it shows up as the optimal output.

(17) LINEARITY-IO (Pater 1995):

S1 reflects the precedence structure of S2, and vice versa.


79

This option turns out to be economic, given no deletion nor addition is required. All of these

observations can be schematically shown in the following Level 2 constraints interaction

tableau:

Base: daftar

Input: /daftarki/ *COMPLEXONS

NONFINAL

(c)-IO

DEP-

IO

MAX-

IO

LINEARITY

a- daf.tar.ki *!

b- daf.ta.rki *!

c- ☞ daf.ta.rik *

d- daf.tar.i *!

e- daf.ta.ri.ki *!

Up to this point, the interim ranking of the relevant constraint in MJA is as follows:

(18) *COMPLEXONS, NONFINAL(c)-IO, DEP-IO, MAX-IO >> LINEARITY-IO

Back to the observation that in MJA, there is no segmental switching of the make-up of 2SF

possessive clitic when the input of level 2 ends with a vowel. Consider the following examples

from MJA in (2), reproduced below for convenience:

(19)

/dawa/ ‘medicine’ → [da.'waa.ki] ‘your medicine’

/sama/ ‘sky’ → [sa.'maa.ki] ‘your sky’

/mawla/ ‘leader’ → [maw.'laa.ki] ‘your leader’

/ʔabu/ ‘father’→ [ʔa.'buu.ki] ‘your father’

The answer should be simple. Under such cases, no extrametrical material is involved, given

that the last syllable is a vowel rather than a consonant. No violation for NONFINAL(c)-IO is

invoked, and, hence no syllabification of any type (repair strategies) is required. This reasoning

extends to MSA where the base ends with a vowel (a reflection of overt Case), yielding no

trigger to NONFINAL(c)-IO.

Let us now consider how this reasoning helps us account for cases in KA. I argue that

the difference between MJA and KA in terms of the order of the segments forming the 2SF

possessive clitic is the ranking of NONFINAL(c)-IO in relation to other constraints. In KA,

NONFINAL(c)-IO is a low-ranking constraint, which is outranked by LINEARITY-IO.

Therefore, no syllabification is required. /r/ is not forced to remain extrametrical and hence the

possibility for it to be a coda with a moraic weight contributing to the metrical structure of the

given word. Consider the following Level 2 constraints interaction in KA grammar.

Base: daftar

Input: /daftarki/

*COMPL

EXONS LINEARITY

DEP

-IO MAX-IO

NONFINAL(c)-

IO

a. ☞daf.tar.ki *

b. daf.ta.rki *!

c. daf.ta.rik *!

d. daf.tar.i *!

e. daf.ta.ri.ki *!

The question promptly arises here as why there is a switch targeting the 2SM possessive clitic

in KA. A closer look at table 1 above, which I reproduce below, indicates that the segmental

switch is also exhibited in KA regarding 2SM possessive clitic despite the assumption that

NONFINAL(c)-IO is outranked by LINEARITY-IO in KA grammar.


80

Table 1: 2nd possessive clitics in MJA, KA and MSA (P: Person; N: number; G: Gender)

P N G MJA KA SA

2nd

S

M daf.ta.rak daf.ta.rak daf.ta.ru.ka

F daf.ta.rik daf.ter.ki daf.ta.ru.ki

P

M daf.tar.kum daf.tar.kum daf.ta.ru.kum

F daf.tar.kin daf.tar.kin daf.ta.ru.kin

One way round this apparent problem is to assume that the underlying structure of the

2SM in Arabic varieties is C instead of CV. The vowel preceding –k in both MJA and KA is

an epenthetic vowel demanded by the interaction of the high-ranking constraints. Evidence for

this assumption comes from some varieties of Najdi Arabic (spoken in Arabian Peninsula)

(Ingham 1994 and Al-Essa 2009). In these varieties where *COMPLEXCOD is low-ranking, no

vowel preceding the 2SM is exhibited (cf. AlAmro 2015). Consider the following examples:

(20)

ʔu.ladk ‘your son’

raask ‘your head’

ʕabaatk ‘your dress’

Further evidence can be adduced from KA and MJA themselves. When the 2SM possessive

clitic is suffixed to nouns ending with a vowel, the only change is the vowel lengthening which,

as we have seen above, is ascribed to stress assignment. Unlike the cases with the 2SF (-ki),

only –k is used. Consider the following table:

Table 5: 2SF forms vs. 2SM forms

Base Meaning 3SF 3SM

/dawa/ ‘medicine’ da.'waa.ki da.'waak

/sama/ ‘sky’ sa.'maa.ki sa.'maak

/mawla/ ‘leader’ maw.'laa.ki maw.'laak

/ʔabu/ ‘father’ ʔa.'buu.ki ʔa.'buuk

/gafa/ ‘back’ ga.'faa.ki ga.'faak

Data in table 5 indicate that the underlying form of the 2SM in both MJA and KA is /k/ rather

than –ak. Following this line of analysis, it can be advanced that when the 2SM possessive clitic

is suffixed to the base that ends with a consonant (daftar + k), a complex coda is formed

(daftark). As a result, violation for the high-ranking constraint *COMPLEXCOD in both KA and

MJA grammars incurs. The optimal repair strategy followed in such cases is insert an epenthetic

vowel that breaks down the resulting cluster. What this means is that DEP-IO can be violated

for the sake of MAX-IO which bans deletion. As such, DEP-IO is relegated a step down in

MJA and KA constraints hierarchy. For the case in hand, consider the relevant constraints

interaction:


81

Base: daftar

Input: /daftark/ *COMPLEXCOD MAX-IO DEP-IO

a. ☞daf.ta.rak *

b. daf.tark *!

c. daf.ta.ra.ka **!

d. daf.tar *!

e. daf.tak *!

NONFINAL(c)-IO plays no role in determining the optimal candidate in such cases.

The relevant question here is as why weightless consonants are forced to parse in 2P

possessive clitics, both in masculine and feminine forms. For instance, the extrametrical

consonant in the base daftar is forced to be a coda in the output daftarkum or daftarkin,

contributing to the syllable weight where it shows up. This state of affairs is not problematic

for KA grammar, given that NONFINAL(c)-IO is low-ranking. As for MJA, I argue that the

violation of NONFINAL(c)-IO in MJA grammar is tolerated in favor of not incurring violation

of the higher-ranking constraints, including *COMPLEXONS, *COMPLEXCOD and MAX-IO.

Consider the following tableau that shows the Level 2 constraint interaction of the input

daftarkum in MJA grammar:

Input: /daftarkum/ *COMPL

EXONS

*COMPLEX

COD

MAX-

IO

DEP

-IO

NONFINAL(c)-

IO

a. daf.ta.rkum *!

b. daf.ta.ra.kum *!

c. ☞ daf.tar.kum *

d. daf.ta.rum *!

e. daf.ta.kum *!

f. daf.ta.rukm *!

g. daf.ta.ru.kam *!

In the suboptimal candidate, daf.ta.rkum, *COMPLEXONS is violated, requiring a repair

strategy if NONFINAL(c)- IO is not to be violated. An epenthetic vowel is inserted between

the parts of the consonant cluster so as to undo the resulting cluster. However, DEP-IO is

violated, yielding to the sub-optimal daf.ta.ra.kum. This indicates that DEP-IO outranks

NONFINAL(c)- IO, as we referred to above. Candidates in (d and e) violate the high-ranking

MAX-IO, whereas the candidate daf.ta.rukm incurs violation of *COMPLEXCOD. As a result,

the most harmonic candidate is the one violating NONFINAL(c)-IO. Accordingly, the

constrains ranking in (18) above must be modified as follows as far as MJA is concerned:

(21) *COMPLEXONS, *COMPLEXCOD, MAX-IO >> DEP-IO >> NONFINAL(c)-

IO >> LINEARITY-IO

In KA, the following hierarchy is maintained:

(22) *COMPLEXONS, *COMPLEXCOD, MAX-IO, LINEARITY-IO >> DEP-IO >>

NONFINAL(c)-IO

7 Conclusion

This research argues that the difference in the morphological forms of some 2 possessive clitics

in Jordanian Arabic varieties is not arbitrary, but rather follows from interaction of the relevant

constraints. Using Statral OT, this research advocates a unique constraint banning weightless


82

consonants of the input (of level 2) to be moraic in the output. This constraint is coined as

NONFINAL(c)-IO. I base my discussion that NONFINAL(c)-IO outranks LINEARITY-IO in

MJA grammar, resulting in the switch in the segments forming the 2SF possessive clitic.

Differently, NONFINAL(c)-IO is outranked by LINEARITY-IO in KA, hence the appearance

of the relevant possessive clitics safe without any metathesis. All other relevant issues of 2nd

person possessive clitics were addressed, resulting in a unified analysis for all of them.

Furthermore, this research brings further evidence in favor of necessity of distinguishing

between different levels, and hence superiority of stratal OT over Parallel OT. On the other

hand, this research opens up questions more than it solves, such as the morphological form of

other possessive clitics with (1st and 3rd) person value. Additionally, NONFINAL(c)-IO is to

be examined in other languages so as to validate its theoretical and empirical ground or even

decline it in favor of a more promising mechanism. At first inspection, it is plausible to assume

that weightlessness preservation follows naturally since faithfulness need not be restricted to

the segmental structure but extends to metrical structure of a given word.

References

Abbas, K.H.A. (2012) Topics in the phonology of Jordanian Arabic: An Optimality Theory

perspective. LAP Lambert Academic Pub.

Al-Essa, A. (2009) 'When Najd meets Hijaz: dialect contact in Jeddah', Arabic dialectology. In

honour of Clives Hole on the occasion of his sixtieh birthday. Leiden-Boston: Brill. 203-

222.

Al-Jarrah, R.S. (2011) 'Extrametricality Revisited', LiBRI. Linguistic and Literary Broad

Research and Innovation, 2(1), pp. 3-35.

AlAmro, M. (2015) 'Syllabification in Najdi Arabic: A Constraint-Based Analysis', Arab World

English Journal, 6(4).

Btoosh, M.A. (2006) 'Constraint interactions in Jordanian Arabic phonotactics: an optimality-

theoretic approach', Journal of Language and Linguistics, 5: 102-221.

Crosswhite, K. (2001) Vowel reduction in optimality theory. Psychology Press.

Crowhurst, M.J. (1996) 'An optimal alternative to conflation', Phonology, 13: 409-424.

De Jong, K. (2004) 'Stress, lexical focus, and segmental focus in English: patterns of variation

in vowel duration', Journal of Phonetics, 32: 493-516.

De Jong, K. and Zawaydeh, B.A. (1999) 'Stress, duration, and intonation in Arabic word-level

prosody', Journal of Phonetics, 27: 3-22.

Fassi Fehri, F. (2012) Key features and parameters in Arabic grammar. John Benjamins

Publishing.

Fudge, E. (2015) English word-stress. Routledge.

Hammond, M. (1999) The Phonology of English: A Prosodic Optimality-Theoretic Approach:

A Prosodic Optimality-Theoretic Approach. Oxford University Press, UK.

Hayes, B. (1982) 'Extrametricality and English stress', Linguistic inquiry, 13: 227-276.

Hayes, B. (1995) Metrical stress theory: Principles and case studies. University of Chicago

Press.

Holes, C. (2004) Modern Arabic: Structures, functions, and varieties. Georgetown University

Press.

Hyde, B. (2003) 'Nonfinality. Ms., Washington University, St. Louis, Missouri. ROA-633,

Rutgers Optimality Archive'.

Ingham, B. (1994) Najdi Arabic: Central Arabian. John Benjamins Publishing.

Kager, R. (1999) Optimality theory. MIT Press.

Kiparsky, P. (2000) ‘Opacity and cyclicity’. The linguistic review 17: 351-367.


83

Kiparsky, P. (2003) Syllables and Moras in Arabic. In Caroline Fery and Ruben Var De Vijver

(eds.) The Syllable in Optimality Theory. Cambridge Unversity Press: Cambridge.

Liberman, M. and Prince, A. (1977) 'On stress and linguistic rhythm', Linguistic inquiry, 8:

249-336.

McCarthy, J.J. (1979) 'On stress and syllabification', Linguistic inquiry, 10: 443-465.

McCarthy, J.J. and Prince, A. (1993a) Generalized alignment. Springer.

McCarthy, J.J. and Prince, A. (1993b). Prosodic Morphology I: Constraint Interaction and

Satisfaction. Ms., University of Massachusetts, Amherst and Rutgers University

Prince, A. and Smolensky, P. (2008) Optimality Theory: Constraint interaction in generative

grammar. John Wiley & Sons.

Rakhieh, B.A. (2009) The Phonology of Ma'ani Arabic: Stratal Or Parallel OT. Unpublished

PhD thesis. University of Essex.

Ryding, K.C. (2005) A reference grammar of modern standard Arabic. Cambridge university

press.

Suleiman, S.M. (1985) Jordanian Arabic between diglossia and bilingualism: Linguistic

analysis. John Benjamins Publishing.

Watson, J.C.E. (2007) 'Syllabification patterns in Arabic dialects: long segments and mora

sharing', Phonology, 24: 335-356.

Wright, W. and Caspari, C.P. (2011) A grammar of the Arabic language. Cosimo, Inc.

Zawaydeh, Bushrah A. (1998). ‘Gradient Uvularization Spread in Ammani-Jordanian Arabic’.

Perspectives on Arabic Linguistics XI: Papers from the Eleventh Annual Symposium on

Arabic Linguistics. E. Benmamoun, M. Eid & N. Haeri. Amsterdam / Philadelphia: John

Benjamins. 117–41.

Zuraiq, W., & Zhang, J. (2006). ‘Phonological assimilation in Urban Jordanian Arabic’. Kansas

Working Papers in Linguistics, 28, 33-64.


84

Marwan Jarrah


Percy Building


Newcastle upon Tyne

NE1 7RU

United Kingdom

[email protected]; [email protected]



85

Style Sheet

The style sheet is for the most part based on the Unified Style Sheet for Linguistics Journals,

as followed by the Journal of Linguistics. Please treat this style sheet as a checklist, and make

sure you have followed every point before you submit your paper. For any queries regarding

this style sheet, please email us at: [email protected]

1. Page limit

The page limit is 20 pages, excluding references and any appendices. Don't feel

obliged to use all the allotted pages, but definitely do not exceed the limit.

2. Paper size and margins

A4, with a 2.5cm margin on all four sides of all the pages.

3. Font

Times New Roman; size 12pt for text, 10pt for footnotes.

4. Title

12pt bold, left-aligned, italic capitals.

5. Author name

Leave one blank line under the title, and then put your name in 12pt capitals. On

the line below this, put your institution in brackets in 12pt ‘title case’ (each word

with an initial capital – as you would normally write a university’s name).

7. Abstract

Your paper should begin with an abstract.

Leave two blank lines beneath the institution name.

Insert the heading, Abstract (12-point bold, left-aligned).

Leave one blank line beneath this heading.

Start the abstract text. (See ‘Text’, below, for text style.)

Do not indent the first line of the abstract.

If your abstract has more than one paragraph, indent the 2nd (3rd, etc.) paragraph

1.25cm (0.49 inch).


86

Example:

THIS IS YOUR TITLE

ANNE AUTHOR


Abstract

The first line of the abstract…

8. Text

Text should be single-spaced.

Text, footnotes, and references should be fully justified.

Do not indent the first line of paragraphs under section headings but leave a blank

line between the heading and the first line of the paragraph.

Indent subsequent paragraphs by 1.25cm.

Use SMALL CAPITALS for emphasis in the main body of the text or footnotes.

Titles of books, journals and dissertations in the running text should be in italics.

Any language material in the running text should be in italics. The translation or a

gloss (where applicable) should immediately follow the example and should be

enclosed in single quotation marks, e.g. moja matka ‘my mother’.

Terms used in a semi-technical sense or terms whose validity is questioned are

enclosed in ‘single quotation marks’.

"Double quotation marks" are used for quotations within quotations only.

9. Section and subsection headings

All headings should be left-justified, 12pt, bold, not underlined and in sentence case

(the first letter of the first word capitalized) not title case (the first letter of every

word capitalized).

Number your section headings, starting with ‘1’. (Even if your first section heading

is Introduction, the number should be ‘1’.)

Insert a dot after each section number or subsection number.

Leave one blank line after section and sub-section headings. (Except if there is no

text between a section heading and a sub-section heading, in which case it is not

necessary to leave a blank line after the section heading.)

Example:

3. Results

The text of section 3 starts here, after one blank line. ...

3.1. Experiment 1

The text of section 3.1 starts here, after one blank

line. ...

Any headings in numbered examples should be in italics.

10. Examples

87

Examples should be in the same font and font size as the text of the paper.

The example number should not be indented.

The body of the example should begin 1.25cm (0.49 inch) from the left margin.

Where examples are lettered as well as numbered, the letter should be 0.75cm

from the left margin.

Example:

(1) a. Example example.

Gloss gloss

'Translation.'

Leave one blank line before and after examples.

Linguistic category labels appearing in the gloss should be in SMALL

CAPITALS (not capitals with a reduced font size – go to the Font tab and expand

it to find this option in Word 2007).

Translation is included in single quotation marks and sentence-final punctuation

is within the quotation marks.

Examples within the text should be referred to as (4a), (4a, b), (4a-e), (4)-(7).

Emphasis in numbered examples should be in bold.

Words in the examples should be aligned by using the Tab key, not by inserting

spaces.

11. Tables and figures

Tables and figures should be in their actual positions in the paper, not placed at

the end or on separate pages.

Number tables and figures separately, and give each table or figure an

informative title. The style for titles is as described under ‘Headings’.

Leave one blank line between the title of the table (or figure) and then insert the

table (or figure).

Example:

Table 2. This is the title of the table

Insert table here

12. Quotations

88

Quotations of under 25 words should be included in single quotation marks in the

running text. Longer quotations should be set out as a separate paragraph on a new

line, indented at the left margin throughout, without any quotation marks and with

no extra indent on the first line.

13. Footnotes

Use footnotes, not endnotes.

Footnotes should be single-spaced and should not be indented.

If your first footnote consists of acknowledgments, mark it with * (an asterisk).

14. Appendices

Any appendices should come after all material except the references. (The

references should be the last item in the paper, not the appendices).

15. References

i. Short references in the text:

Short references in the running text should follow the general format author-date,

e.g. (Chomsky 1995).

Include a single space between the colon and page number, e.g. (Chomsky 1995:

135)

Use a comma only between authors’ names, not between author’s name and year,

e.g. (Chomsky 1995, Rizzi 1997)

Where more than one work is listed, works are ordered chronologically, not

alphabetically, unless two works by different authors have the same year of

publication.

When citing more than one author, use the ampersand (&) immediately preceding

the surname of the second or last author, e.g. (Smith & Jones 1995).

ii. References at the end of the paper:

Leave one blank line between the last line of your article, and the heading,

References.

Leave one blank line between the heading, References, and the first reference.

References should be single-spaced in Times New Roman, 12pt.

Second and successive lines for each reference should be indented 1.25cm.

Do not leave any spaces between reference entries.

Justify both margins.

Be consistent with punctuation throughout your reference list and should follow

the style in the example below, namely dates should appear in brackets, book,

journal and thesis/manuscript titles should be italicized (paper/chapter titles

should not, and should not be in single quotation marks), and do not include any

abbreviation for ‘pages’.

Only include references to articles cited in your paper.

Please use only initials (not full names) for the first names of authors and editors.

The initials should always follow the surname of the author/editor.

89

Please do not use abbreviations in the case of journal titles or publisher names

e.g., Oxford University Press (not OUP).

US state names are given using the standard two-letter abbreviation, e.g. MA

(not Mass).

In the case of joint authors, please use the ampersand (&).

Use a hyphen between page numbers, e.g. 3-63.

If more than one work by an author is listed, please give the full reference

(including name) each time (not a horizontal line).

If more than one article is cited from a single edited volume, a short reference to

the volume appears in the article entries and the full details of the volume appear

in a separate entry (see example below).

Example:

References

Abraham, W. (1997). The interdependence of case, aspect, and referentiality in the history

of German: The case of the verbal genitive. In van Kemenade & Vincent (eds.),

29-61.

Cole, M. (2000). The syntax, morphology, and semantics of null subjects. Ph.D dissertation,

University of Manchester.

Holmberg, A. (2003). Null subjects and uninterpretable features: evidence from Finnish.

Durham Working Papers in Linguistics 9, 11-24.

Kemenade, van A. & Vincent, N. (eds.) (1997). Parameters of morphosyntactic change.

Cambridge: Cambridge University Press.

Lattewitz, K. (1996). Movement of verbal complements. Ms., University of Groningen.

Lieber, R. (1992). Deconstructing morphology. Chicago: University of Chicago Press.

Roberts, I. & Holmberg, A. (2005). On the role of parameters in Universal Grammar: A

reply to Newmeyer. In Broekhuis, H., Corver, N., Huybregts, R., Kleinhenz, U. &

Koster, J. (eds.), Organizing grammar: Linguistic studies in honor of Henk van

Riemsdijk, 538-553. Berlin: Mouton de Gruyter.

Sells, P. (1995). Korean and Japanese morphology from a lexical perspective. Linguistic

Inquiry 26, 277-325.

16. Correspondence details

Below your references, start with a new page and type your name, affiliation,

postal address and e-mail address. This text must be left-aligned and in italics.

Example:

Anne Author

Department of Linguistics

Address line 1

Address line 2

Country

[email protected]

17. Page numbers and headers

Do not number your pages.

Do not use headers.

90

18. Submitting your paper

Please submit two electronic copies (Microsoft Word) of your paper. On one copy, delete the

author name below the title and the correspondence details at the end of the paper. If you have

acknowledged your supervisor or anyone else who might cause reviewers to identify you,

please delete this. If you have referenced yourself (e.g. ‘In a previous work I showed that…’),

also delete this – replace it with ‘reference deleted for anonymity’ or similar.

The filename of your named paper should be in the following format: the first three words

of the title of your paper_ named (e.g. newhorizonsin_named.doc).

The filename of your anonymous paper should be in the following format: the first three

words of the title of your paper_ anon (e.g. newhorizonsin_anon.doc).

Address for submission:

[email protected]

Correspondence address:

Newcastle and Northumbria Working Papers in Linguistics

CRiLLS

Room 3.01 & 3.02, Level 3, Old Library Building


Newcastle upon Tyne NE1 7RU

United Kingdom


Documents

NEWCASTLE AND NORTHUMBRIA VOLUME 2 ISSN: 2041-1057