[Gustave Guillaume] Foundations for a Science of L(BookZZ.org)

FOUNDATIONS FOR A SCIENCE OF LANGUAGE

AMSTERDAM STUDIES IN THE THEORY AND

HISTORY OF LINGUISTIC SCIENCE

General Editor E.F. KONRAD KOERNER

(University of Ottawa)

Series IV - CURRENT ISSUES IN LINGUISTIC THEORY

Advisory Editorial Board

Henning Andersen (Copenhagen); Raimo Anttila (Los Angeles) Thomas V.Gamkrelidze (Tbilisi); Hans-Heinrich Lieb (Berlin) J.Peter Maher (Chicago); Ernst Pulgram (Ann Arbor, Mich.) E.Wyn Roberts (Vancouver, B.C.); Danny Steinberg (Tokyo)

Volume 31

Gustave Guillaume

Foundations f or a Science of Language

GUSTAVE GUILLAUME


Excerpts from the manuscripts translated by

WALTER HIRTLE & JOHN HEWSON Fonds Gustave Guillaume Université Laval, Québec

JOHN BENJAMINS PUBLISHING COMPANY Amsterdam/Philadelphia

1984

This is a translation of Principes de linguistique théorique (Québec) Paris 1973), a collection of passages selected from the manuscripts by

Roch Valin.

Library of Congress Cataloging in Publication Data

Guillaume, Gustave, 1883-1960. Foundations for a science of language.

(Amsterdam studies in the theory and history of linguistic science. Series IV, Current issues in linguistic theory, ISSN 0304-0763; v. 31) Translation of: Principes de linguistique théorique. Bibliography: p. 167 Includes index. 1. Linguistics. 2. Language and languages. I. Title. II. Series. P121.G85413 1984 410 84-9373 ISBN 90-272-3523-6

© Copyright 1984 - John Benjamins B.V. No part of this book may be reproduced in any form, by print, photoprint, microfilm, or any other means, without written permission from the publisher.

PREFACE

This volume presents, for the first time in English, a represen-tative view of Gustave Guillaume's thought. The texts, drawn mainly from Guillaume's manuscript notes for lectures at the Ecole Pratique des Hautes Etudes in Paris, were selected as far as possible f or their accessibility, as requiring no prior knowledge of his work. Further-more, the fact that the passages chosen deal with questions of general interest to the linguist and the man of science gives them added sig-nificance even in translation. The result is a panorama of the far-ranging and often provocative thought of one of the twentieth century's most original linguists, a fitting commemoration to mark the centenary of Guillaume1s birth.

The grouping of the passages, as well as headings and sub-head-ings, are those of the editor of the French volume. Any notes and marginal comments in the original manuscript have been integrated into the text so that all footnotes in the present volume have been added by the editor or the translators. Inevitably, some of the verve of Guillaume's style has been lost in translation. However, if the result is a text where the ideas of the original are faithful-ly portrayed, the aim of this book will have been achieved. To help the reader keep them in the right perspective, a brief outline of the author's career and guidelines for approaching these texts are given in the Introduction. A list of technical terms and a brief bibliog-raphy are also included at the end of the volume.

ACKNOWLEDGEMENTS

Others have contributed so much to this volume that it is diffi-cult to do them justice in a brief note. First mention goes to the team of students and scholars who compiled and published the original volume, Principes de linguistique théorique and to the Presses de l'Université Laval for permission to publish this translation. The long and difficult job of translating would have been much longer and more difficult without the expert help of Linda Mickelson-Cloutier. Both Professor Bob Jones at the University College of Wales in Aber-ystwyth and Professor Roch Valin, Head of the Fonds Gustave Guillaume in Quebec City have read the whole manuscript and made many pertinent suggestions concerning style and interpretation. Finally, it is only through the devoted and careful work of Line Simoneau, the secretary at the Fonds Gustave Guillaume, that we have been able to publish this translation at a fitting moment to mark the centenary of Guil-laume's birth. To each of these persons goes our gratitude and a portion of any merit this volume may find in the eyes of the reader.

John Hewson Walter Hirtle

TABLE OF CONTENTS

PREFACE V ACKNOWLEDGEMENTS VI INTRODUCTION

Gustave Guillaume 1883-1960 XI

Notes for the Reader XIX

PRELIMINARY COMMENTS 3

I - THE PROBLEMATICS OF A SCIENCE OF LANGUAGE 1. Linguistics and the Other Sciences 13

The Object of Linguistics 13 Logic and Linguistics 14

2. The Method of Analysis 17 Observation of the Entities of Tongue 17 The Mentally Seeable and the Sayable 18 Intuition 20 Working Hypothesis and Theory 22 Proof in Science 24

3. Gustave Guillaume and Linguistic Tradition 27 General Grammar 27 A Misuse of Logic 28 An Error of Historical Linguistics 30 Tongue and its History 32 A Shortcoming in the Saussurean Analysis 33 Comparative Linguistics and Psycho-systematics 38

4. Theory and the Science of Language 43 Theory: Understanding in the Superlative 43 Tongue is in itself a Theory 45 The Presupposition of Order in the Construction

of Science 46

VIII TABLE OF CONTENTS

II - FROM PROBLEMATICS TO SYSTEMATICS IN THE STUDY OF LANGUAGE The Simplicity Postulate 49 Psycho-systematics: Definition and Method 50 Psycho-systematics and Psycho-mechanics 51 Particularizing and Generalizing in the, Construction

of Language 52 The Law of Non-recurrence 54 System and Diachrony of Systems 56 The Continuing Causation of Language: Proposals

and Transforms 62 Problems of Representation and States of Tongue 64 Language and Silent Thought 65

III - SIGN AND SIGNIFICATE The Physical and the Mental: a Fundamental Duality. . . . 69 Signifying and the Mental/Physical Relation 70 Interiorizing and Exteriorizing in Language 71 The Suitability of the Sign to the Signifícate 72 Expressive Sufficiency: the Law of Psycho-semiology . . . 73 Reciprocal Accommodation between the Mental and the

Physical 74 Negative Morphology (Article Zero) 75

IV - THE ACT OF LANGUAGE The Nature of the Act of Language 79 The Act of Language: its Operational Chronology 82 Established in Language or Improvised?

Expression vs. Expressiveness 84 The Place of the Sentence in Language 88 Units of Potentiality and Units of Actuality:

the Word and the Sentence 89 The Precedence of Tongue: its Anticipatory Nature . . . . 90 From what is Thinkable to what is Thought:

Tongue and Discourse 91 Representation and Expression: Universality of

the Relation 94 Tongue as Conditioner 97

V - LANGUAGE AND SYSTEM 1. The Systemic Nature of Language 103

A Law Inherent in any System 103 A Difficulty Regarding the Study of Language

Systems 103 Tongue as a System of Systems 104

TABLE OF CONTENTS IX

The Established and Non-established Aspects of Language 104

System and Non-system in Tongue 106 The Historical Filiation and Renewal of Systems 107

2. The Systemic Nature of the Word 109 Discovering the System: an Appropriate

Analytical Technique 109 The Genesis of the Word: Matter and Form 1ll Operations of Word Construction: Discrimination

and Categorization 113 The Mechanism of Word Construction in the

Indo-European Languages 114 Morphology and the Genesis of the Word 116 A Structural Device: the Radical Binary Tensor 118 The Role of Incidence in Determining

Word Categories 119 Internal Incidence and External Incidence 121 System of the Word and System of the Article 123 Potentiality and Reality in Language 128

VI - THOUGHT AND LANGUAGE Language and Operationalism 133 The Operational Substratum of every

Language System 134 Language and what is Sayable: from the

Unsayable to the Sayable 135 Structural States of Language and the History

of Human Understanding 138 The Humanizing Function of Language:

Linguistics and Anthropology 139 Language as a Reducer of Mental Turbulence 143 The Specificity of Human Lucidity 144 The Reciprocal Causation of Thought and Language 145 The Common Mind and the Scholarly Mind 146

EPILOGUE Language and Scientific Curiosity 151 Language, Mathematics and Linguistics 153 Language, Man and the Universe 155 Language and Common Thought 161

BIBLIOGRAPHY 167 INDEX 171

INTRODUCTION

GUSTAVE GUILLAUME 1883 - 1960

One of the most striking characteristics of Gustave Guillaume as a linguist was his single-mindedness, his unwavering purpose. In the earliest of his writings (1911) he was already concerned with "the mind directing the act of language", and his lecture of January 28, 1960, written barely a week before his death, depicts the act of language as "a commutation within thinking man of his momentary thought into speech" and language itself as "a mechanism commuting what has been thought into something said". Between these two texts stretches nearly half a century of probing, of close observation and ever deeper reflection in an unflagging effort to fathom the relation between thought and speech, between mind and language.

THE PROBLEM OF POLYSEMY

Even though it is an oversimplification, one might characterize the main object of Guillaume's research during the first decade of his career as a linguist by means of a single question: where do the various contextual senses of a morpheme come from? As a young man, Guillaume had taught French to Russian emigres, and like many second language teachers, was no doubt confronted with the problem of reducing the polysemy of grammatical forms to a basic meaning. In this attempt to get to a deeper level in language he soon saw that the relation between the "underlying" meaning and the diverse con-textual senses must be one of potential to actual, an insight that led to his first major work. This was a detailed study of the arti-cle in French (Le problème de l ' a r t i c l e et sa solution dans la langue française), published in 1919 and dedicated to Antoine Meillet, at whose instigation Guillaume had undertaken his studies in linguistics some ten years before. In later years, the author himself looked back on this volume with a certain condescension, remarking that it contributed an idea "overwhelming in its banality: underlying an actuality there is a potentiality." With hindsight, this idea may well appear so obvious as to be banal. Yet the fact remains that

XII INTRODUCTION

some linguist had to adopt it as a starting point for reflecting on language, and, by the same token, as a basis for the science of lin-guistics. Guillaume was to spend the rest of his life exploring language in the light of this potential/actual relationship.

Guillaume's study on the article is also valuable for what it can tell us of the author's conception of scientific method and his lifelong concern with making linguistics a mature, theoretical sci-ence. Here we have perhaps the best illustration of what Guillaume had in mind when, in later years, he so often insisted on the impor-tance of observation in linguistics. Not only must observation be as meticulous and as extensive as possible, but it must, in order to be complete, focus on both the physical and the mental, both the form (the "sign" in Guillaume's terms) and the contextual meaning. In this respect, as a far-ranging collection of attested examples of the ar-ticle, each classified according to its contextual nuance of meaning, Le 'probléme de l' article stands unsurpassed. A similar respect for observed data characterizes all his work.

Guillaume never mentioned the importance of observation in lin-guistics without stressing that it must be accompanied by reflection. Convinced as he was that, on the mental side, the reality of language is largely subconscious and so extends far beyond what direct obser-vation can reveal, he realized that this hidden part can be reached only by analysis, by reflecting on observed data. He often spoke of observation and reflection in terms of their results — seeing and understanding — or in terms of their objects — the perceivable and the conceivable. And he frequently pointed out that the linguist must commute between the perceivable and the conceivable in his effort to understand what he sees and to see what he understands. Perhaps the most remarkable illustration of this view of linguistics as a two-way science resides in the fact that Guillaume spent over twenty years reflecting on what he had observed of the article before he conceived the theory which permitted him to understand. One reason it took him so long was that, in order to theorize the article, to reconstruct its system, he needed the basic parameter for analysis, and this came from his observing and reflecting on the verb.

OPERATION AND TIME

The second decade of Guillaume's career was largely concerned with the problem posed by the answer to his first question: if the various contextual senses of a morpheme arise as actualized meanings from a single potential meaning, where does the potential meaning itself come from? In his attempt to get to a deeper level of language he sought, as always, to determine the conditions presupposed by that facet of language he was trying to understand. It was not until 1927

INTRODUCTION XIII

that Guillaume had the insight that provided the cornerstone for all his later theorizing: he saw that something is potential before it is ac tual.

A use of the French subjunctive which opposed it to the indica-tive had led him to reflect on how we think the possible and the real. He saw that to represent some event as possible by means of the sub-junctive mood, the speaker must somehow give it precedence, must somehow think it prior to thinking it as real by means of the indi-cative mood. And the only time within which he could postulate that the subjunctive precedes the indicative was the "thinking time" re-quired by the mental process of representing a verb. This insight led him to imagine the system of mood as essentially a single, sub-conscious operation of thought which can be held up at successive moments, each such interception giving rise to the formal (grammat-ical) meaning that characterizes one of the moods of the French verb. Within two years, Guillaume had published his second major study, Temps et Verbe.

This second landmark is certainly significant for the systems of aspect, mood and tense proposed, but its basic importance lies in the principle of analysis underlying them: operative time. For Guillaume, a grammatical system is essentially a mechanism in the mind which produces the several meaningful constituents of the sys-tem. That is, each of the morphemes belonging to the system arises at a different moment in this operation of thought, as a potential meaning determined by its relative position in the operation involv-ed. As a consequence, each morpheme must be defined in terms of its position, early or late, in the micro-stretch of time required for this mental process to unroll. It is this "thinking" or operative time of a system ("It takes time to think, as it takes time to walk," as Guillaume used to say) which provides a necessary parameter for any language system, permitting the linguist to determine the "no-tional chronology" of the morphemes involved and to see their re-spective potential meanings as the consequences of their position — initial, medial or final — within the system.

This is why, from 1929 on, Guillaume's scientific work might be characterized by the motto "think operatively". For him, "everything in language is process," and to understand any language entity, be it the most particular use of a morpheme or the most general of systems, one must first have a view of the process that produced it. The habit of thinking operatively was so deeply ingrained that, for example, he tended to replace the term morphology by morphogeny and morphogenesis with their suggestions of generative processes. For Guillaume, even when we are not using our language, it resides in the depths of the mind as an organized set of possible processes, as a programme ready to be activated by the thinker/speaker.

Thanks to this persistent attempt to focus on the operational rather than on the more apparent resultative, Guillaume was able to

XIV INTRODUCTION

obtain a unified view of language. When, in 1916, Saussure's Cours de linguistique generale proposed the langue/"parole dichotomy, it left linguistics polarized. Guillaume, on the other hand, linked potential language to actual language by outlining the subconscious morphogenetic processes that produce a word, showing that a word, this "miniature bit of art" (to recall Sapir's striking description) must be assembled by the speaker at the moment of need before it can be used in a sentence. In describing how a speaker transits from language as a system of potentialities to an actual sentence when-ever he carries out an act of language, Guillaume was able to avoid the predicament of those who attempt to theorize the syntax of a sentence before theorizing the words that make syntax possible, or those who would analyze the functions of words without first analyz-ing the words themselves. As a result of trying to see everything in terms of operative time, Guillaume was able to get a holistic view of language, to see the potential and the actual as parts of a single phenomenon. This view constituted a major step toward making linguistics a theoretical science, even though, for the time being, it was limited in the main to the perspective of synchrony.

THE NOTION OF SYSTEM

Guillaume had proposed that the potential meaning of a morpheme resulted from its position in the system. However, this solution to his former problem itself posed a problem: where does the system containing the morpheme come from? That is, if one accepts, say, the system of the French verb or the system of the French article as rea-lities, as subconscious mental components of French, then one must seek out the conditions making these systems possible. Here, then, was an invitation to probe even deeper into language and for the next decade or so Guillaume set himself to this task.

In the years following Temps et Verbe, Guillaume explored the passing from potential to actual, from system to sentence, concen-trating mainly on further analysis of particular grammatical systems within the word, for the most part those concerned with representing time within the verb (chronogenes i s as he called it), but not to the exclusion of the substantive and its formal representation of space. The articles published during this period (collected in Langage et science du langage), as well as his study of the verb systems of Latin and Classical Greek (L 'Architectonique du temps dans les lan-gues classiques') , show not only a mind capable of probing beneath the observable facts of usage to find the hidden system, of "dis-cerning the invisible" as one of his early reviewers put it, but also a mind struggling to reach an even deeper level where all the individual grammatical systems of a language can be seen as partic-ular cases of one general system of representation. He was trying

INTRODUCTION XV

to demonstrate that grammatical systems are content systems and that as far as grammar is concerned, a language is a "system of systems". This led him to seek the most general system, that of the parts of speech, which provides the underlying structural mechanism of the word. Guillaume, however, was not able to discern this all-encom-passing system with any clarity until he finally managed to discern the form of the mechanism of that most transparent of all words, the article.

In 1938, Guillaume obtained a position as Charge de conference at the Ecole Pratique des Hautes Etudes in Paris. Each of the lec-tures he delivered there over the next 22 years was carefully written out and preserved in yearly series. During certain academic years, Guillaume gave more than one lecture each week, a practice resulting in several yearlong series, here designated Series A, B or C. Some 800 in all, these lectures provide a week by week record of the prog-ress of his thought and as such constitute a priceless scientific legacy. It was in 1941 that Guillaume finally saw the mental opera-tion underlying the system of the article: it consists of a double movement, the first of which carries from the greatest extent possi-ble for any notion (the universal) to the least extent possible (the singular); the second, following on the first in operative "thinking" time, carries inversely from the least to the greatest extent. To each movement of the system corresponds a sign: the first, a con-tractive movement, is expressed by the indefinite article, the second, an expansive movement, by the definite article. This psycho-mechanism he depicted by means of the following well known diagram:

This constituted an important breakthrough for Guillaume because it finally gave him an operational view of the system of the article, on the basis of which he was able to account for the numerous uses observed years before. It was a matter of determining where the movement was intercepted to produce the quantitative effect observed in discourse. However, the significance of this mechanism as a scientific landmark stretches far beyond the theory of the article itself.

XVI INTRODUCTION

Reflecting on the form of the movements involved here, Guillaume soon perceived that this mechanism, a device which he called the bi-nary tensor, could provide a representation of any variable relation-ship based on quantity. He showed how it provided a basis for the system of grammatical number within the substantive and he evoked it to depict the relation within the verb between time as an infinite stretch and time as a finite stretch (the present). On the level of the word, the binary tensor device permitted him to understand how a particularizing lexical meaning combines with generalizing grammatical meaning to receive a part of speech and thereby constitute a word. Here, then, he found the basis for his theory of the word and, by the same token, his theory of the parts of speech.

During the ensuing years, further reflection on this similarity of various systems in regard to their form — a mechanism consisting of a contractive, particularizing movement followed by an expansive, generalizing movement — provided Guillaume with a solution to his problem: the grammatical systems of a language like French, whether particular sub-systems like that of number in the substantive, or more general ones like that of the substantive itself, derive their form from the most general system of the language, that of the word. In this way he came to view the grammar of a language as a system of systems wherein the form of the general system of the word is reit-erated in each of the subsidiary categories of words (the parts of speech). Confronted here with the most general possible form of a language, Guillaume postulated that this all-embracing language mech-anism reflects one of the basic capacities of human thought: the ability to generalize and to particularize. In fact here Guillaume realized that he had reached the boundary between language and thought and that if he pursued his reflections any further in this direction, he would be leaving the domain of linguistics. And yet the solution he had worked out raised a further problem for the linguist: if each subsidiary grammatical system derives its form from the system of the word, where does the system of the word come from?

LANGUAGE TYPOLOGY

It is typical of Guillaume that he did not pause to work out in detail, with the years of observation and reflection it would require, the whole theory of the major systems of French — the parts of speech. Instead, satisfied that he had discerned the basis of the system, in particular when he described the mental relationship between noun and verb, he turned his attention to determining the operational condi-tions giving rise 'to the word in part-of-speech languages such as French and Latin. He soon realized that the word has the same gener-al structure in every Indo-European language although the architecture of the sub-systems may be different in each language. That is to say,

INTRODUCTION XVII

once he had glimpsed the mental infrastructure of the Indo-European languages, which gives rise to words necessarily provided with a part of speech, his constant practice of trying to understand by turning from a result to its prior generative operation led him to seek what had brought about this type of word structure. However, as mentioned above, Guillaume had already reached the confines of language in his analysis of the act of language, and so, unable to probe any deeper in synchrony, he was led to adopt a different point of view. That is, without denying that the part-of-speech structure is the starting point for the speaker of an Indo-European language, the potential from which he undertakes an act of language in synchrony, he began to regard this structure as a result from the dia-chronic point of view. This brought him to focus his attention more and more on the processes leading to this result, on the means where-by man over the ages was able to construct this mental edifice per-mitting him to represent and express his thought with such ease. So during the last decade of his life Guillaume left particular grammat-ical systems aside and turned to reflecting on their glossogeny — the progressive edification of the major language structures at-tested in today1s languages — and on the consequences which this new dimension of investigation held for the science of language.

Being careful not to foist onto other languages the type of word he saw in the Indo-European languages — to this end he often used the more general term "vocable" as being less tainted with Indo-European connotations — Guillaume attempted to discern the constructive mecha-nism, the sort of thought processes presupposed by the make-up of vocables observed in Chinese, in Arabic, in Basque, etc. The result is certainly one of the most original theories of language ever con-ceived: the Theory of Glossogenic Areas. This theory constitutes an attempt to explain how man, over the millenia, has been able to form the successive mental structures discernible beneath the dif-ferent types of vocable found today, and how each such general struc-ture governs the manner in which the architecture of vocables is realized in particular languages and how the vocables are actually produced as constituents of individual sentences. In short, this theory of glossogeny attempts to distinguish the different language types in terms of one constructional parameter and so to embrace the whole of language, both in diachrony and in synchrony, as a single phenomenon.

Guillaume devoted his last years almost exclusively to the Theory of Glossogenic Areas and to the conditions permitting linguistics to outgrow the phase of merely describing usage and become an explanatory science capable of throwing light on the nature of language. Unfor-tunately, he died before he could do much more than lay the founda-tions of this grandiose scientific edifice. What he actually achieved is more like the plans of a general theory of language — a landmark perceived on the horizon ahead — than a fully developed theory. Cer-tainly, his writings do include carefully worked out solutions to a

XVIII INTRODUCTION

large number of particular problems, as well as innumerable insights opening avenues for further observation and reflection. But since any scientific theory, however well founded is something to be de-veloped, rectified and ultimately outstripped and left behind, per-haps his most important legacy will prove to be the method he used from 1929 on to explore the hidden psychomechanisms of language processes: analysis in terms of operative time. Because it has already produced and is producing significant results, this method constitutes a challenge for any linguist, or at least any linguist who, interested in the mental nature of language, holds, with Guil-laume, that language is "a mechanism for commuting what has been thought into something said."

Walter Hirtle

NOTES FOR THE READER

The purpose of this section is to aid, as far as possible, the interested reader with Guillaume's terminology and ideas. It is acknowledged by those who attended his lectures that Guillaume was one of the most insightful and stimulating theoreticians that twen-tieth century linguistics has produced, but one must be careful not to be misled by expectations: he had a very distinctive view not only of linguistic theory, but also of what a language is and how it operates.

One must be careful, therefore, not to read Guillaume with the idée fixe that linguistic theory is a form of technology with a spe-cific kind of argumentation, the equivalent of a Euclidean theorem. For Guillaume, theory is concerned with the understanding of that which is not directly observable (the underlying operational system) through the medium of the "inevitabilities" (that which must be) in what is directly observable. He stresses the importance of observa-tion, likewise the importance of reflection. It is a view of science that is neither mechanist nor reductionist, and in that respect quite different from the traditional nineteenth century (or seventeenth century) views of science held by others. To understand and appre-ciate Guillaume, it is a difference that has to be taken into account.

There is a particular danger in this regard for a translation: a language relates to its own cultural matrix, and the latter remains foreign even in a translation. An English speaking reader familiar with the psychology of Piaget, with the anthropology of Lévi-Strauss, with the phenomenology of Merleau-Ponty, will quickly feel at home in these texts from Guillaume. But a reader reared on the psycholo-gy of Skinner, on the Cartesian mechanism of Chomsky, on the positiv-ism of Wittgenstein, may well suffer from culture shock, and not be able (or even willing) to make the necessary adjustments. The danger needs to be kept in mind at all times because in a book of selections a statement or a claim may be separated from its supporting evidence, thus exacerbating any differences of opinion on the proper function-ing of scientific method.

XX INTRODUCTION

LANGUE/LANGAGE

There are also some very difficult problems in the translation. A fundamental problem, for example, is that the French terms langue and langage need to be distinguished, but the English term language subsumes them both. In translating Saussure the traditional solution has been to translate langage as language and to use langue as a technical loan word. In the final analysis, this has not worked, since langue and parole have commonly been explained as meaning l a n -guage and speech, which returns us to square one.

Langage is the whole phenomenon of language, which includes human language, language systems, use of language, written language, foul language, any kind of language. It is open-ended, boundless, and consequently unknowable, as indicated by Saussure: "...le tout du langage est inconnaissable." Une langue, on the other hand, is "The tongue of one nation as distinct from others", to borrow a def-inition from Dr. Johnson.

Consequently, after much soul searching, the translators decided to use the terms tongue/language to translate langue /langage, fully aware that this leads to certain infelicities, and to an unexpected extension of meaning for the English word tongue, though possibly not significantly more than that given by Saussure to French langue. It may be noted, for example, that we commonly speak of "the mother tongue" (never "the mother language"), and that it is not unknown for linguists to speak of "the tongues of men."

This solution to a perennial problem, while it still presents difficulties, is made more digestible by the fact that Guillaume (as in Part One, Chapter Three below) replaces the Saussurean formula

langage - langue + parole

by a slightly different (and improved) terminology:

langage = langue + discours

Since discours, unlike parole, is easy to translate, the Guillaumean formula becomes in English:

language = tongue + discourse

Discourse in this sense does not require a paragraph; any connected, articulated language is discourse in the Guillaumean sense.

INTRODUCTION XXI

THE TONGUE/DISCOURSE DISTINCTION

A fundamental key to understanding Guillaume is a clear grasp of the tongue/discourse distinction. Tongue is the system, and dis-course is what is produced by the use or exploitation of the system. Guillaume, in fact, in an interestingly original way makes a process model out of what in Saussure is a purely static dichotomy: he links the two with the act of language, which has its own underlying time, as in Figure One:

FIGURE 1

Because there has been, and still is, so much confusion over these fundamental linguistic concepts, a concrete analogy is helpful in making clear what Guillaume intended by these terms. We may use a child's construction kit (such as a Meccano or Erector set, or Leggo blocks) as an analogy for tongue, and the constructions that the child makes (trucks, houses, helicopters, bridges) as an analogy for discourse, as in Figure 2 (which may be compared directly with Figure 1).

FIGURE 2

This analogy makes clear that a tongue is not a set of sentences, but a set of paradigmatically related parts that can be fitted together syntagmatically in significant ways to form an infinite variety of sentences. Similar views (though without the explicit operativity) are to be found in Hjelmslev and Jakobson (the latter, for example, distinguishes the paradigmatic from the syntagmatic axes of language). The analogy also illustrates what Saussure meant when he said that the sentence belongs to parole, not langue (CLG 172). It is a fun-damentally different viewpoint from that of linguists who define a language as a set of sentences.

XXII INTRODUCTION

SEMIOLOGY/PSYCHOSYSTEMATICS/PSYCHOMECHANICS

Guillaume does not use the term sémiologie in the Saussurean sense of "a science of semiotics", but more or less equivalent to Hjelmslev's term expression (For Hjelmslev content relates to the Saussurean signifié and is therefore meaning; expression relates to the Saussurean signifiant and is therefore the morphosyntactic means of conveying content). Because Guillaume uses the term expression in his own technical sense (see, for example, Part Four), Hjelmslev's more familiar terms could not be used here; consequently we have used the term semiology to translate sémiologie; the term means the total-ity of signs (Saussurean signifiants) that a language utilizes.

For Guillaume the grammatical systems of a given tongue are con-tent systems, systems of meaningful contrasts that the semiology, with its irregularities, reflects only imperfectly. This view (which parallels those of Hjelmslev and Jakobson) is discussed comprehen-sively in Part Three: it has been characterized as a "grammatical semantics." The study and analysis of these content systems falls within the domain of psychosystematics, since they are necessarily mental entities, and Guillaume uses the prefix psycho- to identify mental elements.

Psychomechanics is the term used for the study and analysis of the mechanical processes of the act of language: choice of lexeme, grammaticalization of lexeme, operation of systems, and so on. Psycho-semiology as may be expected, deals with morphological shape at a potential or mental level.

SYNAPSIS

Guillaume uses the term synapsis for motivated syncretism, that is for two (or more) separate but in some way similar elements that share a common morphology (e.g. the dative and ablative plurals of the Latin noun declensions, or such English verb forms as talked which serve as both preterit and past participle).

SAISIE = PREHENSION

It is a truism that 1 + 1 does not, strictly speaking, equal 2, since 2 is another distinct unit, a new saisie which replaces the two original units (e.g. 1 + 1) by amalgamating them into a new whole. A saisie is therefore a mental grasp of a new integral whole; we have translated saisie by the word prehension, again with the regret that, stylistically, we seem to have produced the Leaden Echo rather than the Golden Echo.

INTRODUCTION XXIII

The term saisie is an important one. The existence of deriva-tions is clear evidence that there are stages of processing, and that this processing will be carried as far as is necessary for the pur-poses of expression. At whatever point this processing is cut, a new prehension is achieved. There are three universal prehensions (common to all languages): that of the radical (basic lexical no-tion) , that of the word or vocable (no matter how different the words of different languages may be), and that of the sentence.

SUPPORT/IMPORT/INCIDENCE

Guillaume's syntax is essentially that of a dependency grammar, but of a quite original kind. Unlike Tesnière, who bases dependency on the verb and its valencies, Guillaume sees dependency as a matter of incidence, wherein a lexical import, when it finds a support else-where, thereby establishes a syntactic relationship. The lexical im-port of old in old book, for example, finds a support in book and thereby establishes a relation of external incidence in the first degree. Similarly very in very old book has an import which finds a support in old, thereby establishing a relation of external inci-dence in the second degree, the normal role of predication for an adverb. The germ of this idea, though not developed in such terms, is to be found in Jespersen's primary, secondary, tertiary. (In Guillaume's terms Jespersen's secondary has external incidence in the first degree, and his tertiary has external incidence in the second degree).

Jespersen's primary, which is the keystone of any such system (on which all the rest is dependent) is the substantive. For Guil-laume the incidence of the substantive is no longer external but internal (see the section in Part Five), a feature which gives it syntactic independence and allows it to be the support of other (secondary) elements.

The questions raised by these brief comments require the prep-aration of a book-length study of this kind of dependency syntax, a task which is presently in progress (see Hewson forthcoming). Al-though these questions are of very great interest to many linguists, discussion of them must consequently be postponed.

CONCLUSION

There is, of course, no way that a few brief comments on termi-nology can provide an adequate introduction to the work of a linguist as original and as provocative as Guillaume. These notes will have

XXIV INTRODUCTION

served their purpose, therefore, if they succeed in providing a pre-liminary map that will enable the reader to find his basic bearings and avoid the most obvious pitfalls.

John Hewson Quebec City, December 1983.


Reproduction of a page from Guillaume's manuscript.

PRELIMINARY COMMENTS

Science is founded on the insight that the world of appearances tells of hidden things, things which appearances reflect but do not resemble.

One such insight is that what seems to be disorder in language hides an underlying order — a wonderful order. The word is not mine — it comes from the great Meillet, who wrote that "a language involves a system where everything fits together and has a wonderfully rigorous design." This insight has been the guide and continues to be the guide of the studies pursued here.

The object of my research is to discover the system constituting tongue, and the plan of its architecture. I have been lecturing here since 1938 and those who have followed my teaching from the beginning are aware of the slow discovery of this architecture and of the methods of observation and analysis it has called into play. They remember how we groped our way on the days when nothing worked — and there were such days — and yet how, on the good days, our analysis advanced rigorously step by step. In this regard, I must thank my students from past years for their patience in following me when my efforts amounted to mere fumbling in the thick mists of research; the successful discoveries of the good days have rewarded such patience.

Along with the ups and downs of research, major breakthroughs did eventually come; it has been a long time since those early days when we perceived only vaguely, but with real intuitive conviction, that tongue is a s y s t e m , without realizing what this system was like on the inside. And it would be quite a story, though not a very glorious one, to recount our efforts, some fruitful, some unfruitful, which led us to see what the inside is really like.

1. The word "tongue" is used here as a translation of the French langue , denoting language as a potential and opposed to discours ("discourse") denoting language in use. For clarification of this point, and for other technical notes on the vocabulary used, see Notes for the Reader, p. XIX-XXIV.

4 FOUNDATIONS FOR A SCIENCE OF LANGUAGE

One of the earliest insights was that tongue is a system of s y s t e m s . This idea made our study somewhat easier because, setting aside for the moment the notion of an overall integrating system, we determined to gain knowledge of certain contained subsystems that are arrayed in a semiology1 which provides visible evidence of their existence .

Strangely enough, but for profound reasons that I did not immediately discern, I focused my attention on the system of the article first. When I recall those early investigations, moreover, the word system does not seem appropriate. I was not so much concerned with the system as with the problem of the article.

First of all I assumed — and it made good sense (even if I say so myself) — that, since the article had appeared independently and quite late in many languages, its appearance provided the solution for an unsolved problem, a problem latent in the mind: the patent solution to a latent problem. This was not really very much of a step, but it was useful, invaluable even. And what did I see underlying the patent solution revealed by the language? The problem explicitly posed. When, 36 years later, I look back over the book I then wrote — it is still not really out of date — I see mainly an explicit statement of the problem of the article.

The problem of the article, quite overwelming in its banality, is essentially this: underlying an actuality there is a potentiality, consequently, if we follow up the necessary sequence of commonplaces involved here, an element of tongue such as the substantive exists as a potentiality before existing as an actuality. Thus the speaker conveys it from potentiality into actuality during the instant when he is thinking/speaking. This necessarily involves the speaker in a transitional operation, and the problem is how to carry it out. Since it is necessary, this operation is always successful, but it took a long time for the right way of carrying it out to become clarified. For a long time the means of carrying out this transitional operation was left to linguistic categories that are also burdened with other functions, including those of number, gender and case, with number playing the principal role. Then, at a given point, these categories which were responsible for various functions unburdened themselves of the function involving transition of the potential noun to the actual noun. The consequence of this unburdening was the creation of the category of the article.

1. Guillaume used the term "semiology" to designate the physical means of expression, the sounds or writing, as opposed to what is expressed: meaning or content. It follows that by "system" Guillaume here means the system of abstract grammatical meanings, the content system underlying grammatical items.


My book on the article, Le problème de l'article et sa solution dans la langue française, does not succeed in giving a clear view of this unburdening process. This view did not come to me until much later when, from the point of a systematics rather than a problematics of language, I was led to look for,the respective positions of the two categories of number and the article, and to demonstrate that the latter has the same form of mechanism as the former.

Only at that moment did I realize what the system of the article is. I find it important to distinguish between these two moments of discovery. The former belongs more to a problematics than to a psycho-systematics of language. It runs as follows: the article exists, appears, independently in numerous vernaculars, without being borrowed. It is therefore the solution to a problem. What problem? As I said earlier, the substantive exists in the mind of the speaker as something potential before arising there as something actual.

Furthermore, this particular case of problematics brings with it the whole problem of the relation between tongue and discourse. Tongue exists in each one of us on a permanent basis prior to any act of expression. I speak and express myself by working from tongue. My speech, my discourse, is by nature momentary; on the other hand, tongue belongs to the non-momentary, the permanent within me. The tongue/ discourse transition is, therefore, a general problem and the problem of the article, which I have elucidated, is just one particular case which may be used to illustrate the general fact.

Students from past years know about my studies dealing with the problem of the transition from tongue to discourse, and so I can simply appeal to their recollection of the subject. A large part of my research was concentrated in the field of this transitional phenomenon whereby discourse is produced from tongue. Some years later, probably after reading what I had written, fellow linguists recognized the importance of this problem and saw, in acts of discourse as a whole, an actualization of what tongue provides. The word actualization was proposed by me, and used consistently, well before it was taken up by Bally. And nowhere has the actualization of tongue, necessary for the production of discourse, been more fully expounded and demonstrated than in my book on the article.

Thus, for a long time the range of my studies was that of a problematics. Everything that takes place in discourse stems from problems posed by tongue, by the virtuality of tongue. Then, without my being completely aware of it (scientific intuition — itself mysterious — was my guide), the time came to set aside problematics for systematics and so to look at things from a different angle. It was no longer a question of examining the tongue/discourse transition, but of seeing what was already instituted in tongue; from the examination of the momentary workings of discourse which stem from non-momentary, permanent tongue, already constructed, I moved to the study of the workings involved, in the mind, in the construction of


tongue itself. All my work of recent years has been done with this objective. I have replaced the study of the acts of expression, the constructors of discourse, with the study of the acts of representation, the constructors of tongue. This change in my research point of view took place about 1928.

Temps et Verbe, which dates from 19301 and marks the beginning of my work on acts of representation, deals with an act of representation of very great interest, the representation of time. This system of representation can be seen, semiologically, in the conjugation of the verb, which, as I showed, quite simply indicates the successive moments of this remarkable act of representation.

Already to be found in Temps et Verbe is the idea that time is constructed in terms of space on n dimensions: it has depth, represented by the succession of moods, and also width and height, represented by the system of tenses. A demonstration of this construction of time on the model of space is given in the monograph. But only later in my teaching did I introduce the principle that time is not representable by itself, and that it bases its representation, where it has one, on spatial characteristics. Thus the representation of time, called chronogenesis and chronothesis in my work — the latter being a cross-cut or interception of the former — is, in fact, a spatialization of time.

This is an operation that the human mind either does or does not carry out within itself. Where it does not carry it out, time, not representable by itself, has no representation. And I repeat: has no representation. That does not mean that it has no existence in human thought, but that it exists in thought only as our experience does. It has been put this way: the human mind is so made that it has the experience of time, but has no representation of it (it must, therefore, invent this representation, which will be a spatialization, representability being a property of space, and of space alone).

Later on, the reasons for representability belonging to space were studied here; this part of my teaching is unpublished. My book L'architectonique du temps dans les langues classiques, published in Copenhagen, presented a study of the spatialization of time in the classical languages, Latin and ancient Greek. This study is far superior to Temps et Verbe, although in the main the two works are in agreement.. The difference, essentially, is that in Temps et Verbe it was recognized that time is constructed like space on n dimensions, while in L'architectonique du temps it is established that time, not being representable by itself, borrows its means of representation from space, and that it is clothed in a spatial representation, in the absence of which we would know time only as experience (which would not be really knowing it).

1. In actual fact, Temps et Verbe was published in 1929.


At this point in the general progress of my research I began to realize that tongue comes into being through converting experience, which the human mind tends to get away from, into a representation, within which the human mind establishes itself. A moment comes when the experience of space is already promoted to the level of representation while the experience of time has not received this promotion. Time in such languages is not represented, which amounts to saying that it is not spatialized. These languages have no chrono-genesis, no conjugation.

I devoted considerable time to the study of the spatialization of time, and I was able to reveal its mechanism which, from a universal point of view, is not s i n g l e , but multiple. There is a trimorphic spatialization of time (past, present, future), found in ancient Greek, Latin, and the Romance languages, and a dimorphic spatialization, characteristic of the Germanic languages (past/extensive present).

In my own work I have described trimorphic spatialization, somewhat neglecting the description of dimorphic spatialization. This is because I have only recently seen clearly what the difference stems from. This year I shall have the opportunity to talk about this and to demonstrate that the representation of time differs only because the representation of space differs. The principle could be expressed this way: corresponding to a certain representation (of our experience) of space, say "A", there is a representation of time "A'", and corresponding to another representation of space, "B", there is a representation of time "B'". The demonstration of this will constitute a new contribution to the psycho-systematics of language this year.

The Slavic languages have spatialized time in an original way: they have the same dimorphic representation as the Germanic languages, with this difference: the means of spatializing the time within which the verb is lodged are borrowed from the spatialization of the time lodged within the verb.1 A diagram will help to clarify these ideas:

The spatialization of the time inside the verb helps in the spatialization of the time outside. Like the Germanic languages, the Slavic languages have only two time-spheres, the past and the extensive present, but they have developed a future, without recourse to an auxiliary (there being no future inflection), by using means borrowed

1. That is, the time between the beginning and the end of the event evoked by the verb.


from the representation of aspect, which is the representation of time inside the verb, of time involved in the event.

I have on occasion regretted spending so much of my career as a linguist on the study of the system for representing time, and neglecting the even more significant study of the general conditions governing linguistic systems. I began the latter study by discovering that the different systems of the category of number, of the article, and of the binary representation of the universe (space-universe/time-universe) — this being the basis of the parts of speech-had the same underlying mechanism.

And this led me to the idea, now sufficiently verified to put forward, that tongue is a peripheral system for apprehending the thinkable, a system which integrates within itself, by circumscribing them, more confined systems, with different limits, all of them reiterating the same general form. At the moment, my studies are based on the idea of a peripheral system consisting internally of repetitions of itself as far as the general form is concerned.

My idea of tongue is therefore that of a system of systems, with this much being clear: the general, containing system and the less general, contained systems do not differ in their general form; their difference is one of substance or of limits.

Last year, I outlined the overall psycho-systematics of tongue in terms of this view, and this year's studies will be pursued along the same lines. These studies have led me to see more clearly than before the inter-dependence of the integrating system of tongue with the systems it integrates. I am at this moment finishing a book which will describe quite plainly how the different systems linked and integrated in the overall system of tongue are delineated and individuated.1

One of the main ideas that led to my writing this book is that tongue as a whole lies at the origin of its constitutive parts. First, there was the whole, which was chaotic and unorganized; then there was the creative phase, which involved separating, discriminating, organizing, and progressively revealing its organization. This gives linguistics the character of a science of discrimination, a science which operates through contrasts.

1. This book, found in a nearly finished version, appears in the catalogue of unpublished papers under the title: Essai de mécanique intuitionnelle. Espace et temps en pensée commune et dans les structures de langue. A first transcription of this work has been produced, in spite of difficulties in reading the handwriting. It is one of the most curious of Gustave Guillaume's writings. The analyses and insights in it open unexpected perspectives on the two cognitive categories of space and time. Philosophers and scientists will find that it provides new food for thought on the problem discussed.


One principle that I shall make great use of this year in the research to be presented here is that we can think only by making contrasts. In particular, it will be shown that the contrast of space and time is the ultimate contrast, created to make thought possible when the mind, working upstream to its very sources, encounters there the non-contrast which is not thinkable. It is by this ultimate contrast that the representation of an empty universe eludes non-contrast.

Let me finish by assuring my listeners that there is nothing hazy in my work, no hidden basis of deep subsensory impressions which might suit a mind with a metaphysical bent, a mind which would like to lose itself in the fog of metaphysics. Rather, it is quite simply a question of discovering, beneath the reality of appearances, that reality — the true reality — of which the appearances are only the outer wrapping.

These studies have an austere beauty which should eventually attract more people than they have hitherto done. Interest in them has not been widespread, not because they have been criticized, but because they have not been criticized, for the simple reason that one cannot honestly criticize them without first of all following the rigorous step-by-step demonstration. I regret to say that people are not willing to make the effort. If a day comes when many minds, equal to the task and gifted in analysis, devote themselves to such a task of rectifying work already done, linguistics will see opened before it a future the magnificence of which, I make bold to say, has yet been only partially glimpsed by a few. With theoreticians such as myself, the difficulty is that the ordered sequence of prerequisite conditions of language structure must be retraced with me, followed step-by-step, if a proper understanding of the work I have done is to be obtained. And considerable retracing must be carried out before things become clear at all.

There are two possibilities. Either my new students — this they will best know themselves — are curious only about the directly observable aspect of language; in this case, I shall bore them by discussing things that they are not curious about and in which consequently they are not interested (this will put me in an uncomfortable position). Or else, on the other hand, they are truly curious and want to find out what lies beneath the surface of language appearances, and what mechanism — that of human understanding — is reflected in these appearances; if this is the case, the study that we are about to begin together will hold their attention permanently. Before anyone grants me a share of his time, therefore, let him ask himself in all conscience about the true nature of his curiosity. The most curious will be tomorrow1 s friends.

(Inaugural lecture for 1952-1953)

PART I

THE PROBLEMATICS OF A SCIENCE OF LANGUAGE

CHAPTER 1

LINGUISTICS AND THE OTHER SCIENCES

THE OBJECT OF LINGUISTICS

The most abstract mathematics provides the means for a finer conception of things. The most advanced linguistics can only reveal ways of conceiving that man already possesses and that he will not use any the better for knowing about them. Thus linguistics, the scientific study of grammar, gives man no new powers. It merely allows him to understand better the state and the nature of the mental potential he possesses at a given time, without this potential being in the least augmented by such knowledge. In short, being a linguist serves no practical purpose. Of all the sciences, linguistics is the least pragmatic. However, it leads us furthest into a knowledge of the means whereby thought manages to get a clear apprehension of its own processes. But I repeat, to know about these means, to discern them, in no way increases our power of thinking or of expressing our thought. The means that our language offers us in this regard do not become more readily available. The effort undertaken to discern them has, ultimately, no practical result, since the mind will acquire thereby no new skills; the only result will be an understanding of the mechanisms by means of which our thought is able to apprehend, to grasp what is going on within itself.

It should be pointed out that these mechanisms, which lie deep beneath the surface, are inaccessible to observation except in tongue, the mind's sole monument to its own activity. Granted, the whole of literature bears witness to man's power of thought, but literary works reflect the exercising of that power, not the power itself. Only in tongue do we find revealed the power of thought as such, this ability the mind has of grasping its own internal movements. No other monument to this power exists. Nor can it exist. And this is why linguistics holds such a unique position in the hierarchy of sciences: it is a science whose object is absolutely unique, with no counterpart in the universe.

(Lecture of January 13, 1944, Series A)


LOGIC AND LINGUISTICS

I must answer a question from a student, a most amiable tormentor, who wants me to explain the difference between coherence, a well-known feature in my teaching of language, and logic, a feature hardly ever mentioned except restrictively, with the idea of logic reduced to that of coherence. I have on occasion spoken of constructive logic in particular cases. Constructive logic is simply access to coherence.

The same student, again inquiring about logic and coherence, mentioned out of class that he had observed a growing tendency in my teaching to avoid the ending -logy and replace it with the ending - g e n y . Like everyone else, I once spoke of the morphology and ontology of language. Now I prefer to say morphogeny and ontogeny. In my view -logy carries with it an implicit reference to logic and -geny an implicit reference to coherence alone.

I must now define the difference, as I see it, between logic and coherence. To avoid any misunderstanding, let me begin by saying that the linguist need only be concerned with the coherence of language, leaving its logic entirely in the hands of the philosopher. To the best of his ability, the linguist follows the continuous coherent creation of language, focusing all his attention on the path it takes. His work is not concerned with the direct route — the straight line — of logic. The linguist is a very small man alongside that very great man, the philosopher. The linguist's objective is to find the reasons for things as they appear in the on-going, step-by-step process of their creation. The object of the philosopher is to relate this on-going, step-by-step process to its conditions of greater rectitude, of unsurpassable straightness.

Logic is the imaginary route things follow, one that does not have to take into account any accidents along the way, any problems and difficulties that things, because they are things and not just ideas, necessarily bring with them. As I recall, Leibnitz put it like this: "Things impede each other, ideas do not". But the fact that things impede each other does not prevent seeing order in them. Accounting for this order in the light of all the hindrances constitutes coherence. The fact that ideas do not hinder each other allows for an orderly, imaginary treatment which takes no account of the accidents en route. Everything is worked out from beginning to end with the utmost economy. To avoid some of the verbalism often found in discussions on the subject, one might say that logic is the very idea of the imaginary straight line. No accidents on the way between Paris and Rome, no detours, no turn-offs, no by-passes around obstacles: the straight line trip with no detours. Logic is simplicity itself, as fashioned by the imagination. I cannot think what language would be like if it were constructed along such imaginary lines. And I cannot hope to know what it would be like, since language of

PROBLEMATICS OF A SCIENCE OF LANGUAGE 15

this sort just does not exist. What I do know is that observable language does not follow straight, logical lines. It follows the road of coherence, where due account is taken of accidents along the way, accidents of thought, speech and writing. Coherence proceeds one step at a time, making allowance for the lie of the land and the detours needed in order to keep going. The zigzag road to the summit of a mountain is coherent: it is not logical, although it does have its own logic. (It is possible to speak of the logic of coherence, but that is a property of coherence, not of logic.) The road of logic has no zigzags; it is perfectly straight and therefore is imaginary, things being what they are. In fact, the zigzag road does just what the imaginary road would do if it really existed. Thus, one can always follow some section of the winding road of developing language as it makes its way around the obstacles and then look at it in the light of the imaginary road of straight line reasoning. In this way coherence can be reduced to logic, for in the end the result is identical, just as if logic, and not coherence, had been the guide. (The theory must emphasize this 'as if': identity of ends, difference in means.) Things impede each other but ideas do not. If you deal with things you must follow the paths of coherence; if you deal with ideas and not with things, you have before you the paths of logic. And so logic imposes a straight path. Coherence imposes orderly progress, but not along this straight path where one could say that the impediments caused by things are reduced to zero.

I hasten to apologize to my students for these reflections which have led me, as a linguist, to trespass upon the realm of the philosopher. Any success I have achieved as a linguist has been due precisely to the fact that I have always concentrated on the real ordering of coherence, and never taken logical ordering into consideration.

(Lecture of December 6, 1956)

CHAPTER 2

THE METHOD OF ANALYSIS

OBSERVATION OF THE ENTITIES OF TONGUE

The method that I recommend in linguistics, and in a general way in all fields of thought, is keen observation of concrete reality made even keener by deep reflection. I believe that the combination, in appropriate proportions, of these two mental faculties — observation and reflection — can lead to an ever-growing understanding of the universe. Although we remain in inevitable and permanent contact with this universe, we do have the inestimable ability to abstract ourselves from it, and this better enables us to see it within ourselves and to relate it to the exigencies of our reason, exigencies which ultimately determine the limits of the human mind.

The universe we come into contact with in linguistics is an inner one: the universe of the thinkable, made up of our own representations. The fact that it is within us adds greatly to the difficulties of observation: it is indeed difficult for us to grasp exactly what goes on deep in our minds. The difficulty stems largely from the fact that we always arrive too late to make observations. If we wish to observe an act of language, we can only do so once this act has been completed, which means that we have not observed the act itself, the process, but only its result. This is not enough.

Observing a fact of discourse, an act of language such as the construction of a sentence, is by no means easy. Observing a fact of tongue is even more difficult. A person can perform an act of language whenever he cares to, as the situation demands, and while he may not be aware of all that he has accomplished, at least he knows the general aim of his mental action. A fact of tongue is something else entirely because of its different position in the mind: it represents a process which has already taken place in the mind at some undetermined and completely forgotten time in the past. Indeed, who could say at what moment and by what operations of thought he learned the distinction between noun and verb, article and pronoun, between the different articles or all the various tense forms? All these past acquisitions are inherited, not created, by a


speaker whose mind retains no record of how or when it acquired them. It is in total ignorance of such facts. This being the situation, the facts of tongue would remain forever mysterious, unfathomable, were it not for one thing: in tongue itself, in the make-up and the arrangement of its constituent entities, we find visibly delineated those operations of thought which, at some unascertainable point in the past, brought these various entities into existence.

Thus, one of the tasks of the linguist is to examine these entities of tongue which, in the evolved vernaculars we are familiar with, have all taken the form of words. He must examine them very closely, with the help of deep reflection, to see how they are constituted.

The great lesson of historical grammar, which generations of scholars found so fascinating, is that the entities of tongue have neither a constant nature nor a constant form: they are subject to variation in time. Knowledge of this variation is, one must admit, of the highest interest. In order to observe it thoroughly, to discern all its elements and its ramifications, one has to learn to observe an entity of tongue at a certain moment of its existence, independently of what it might have been beforehand and of what it will subsequently become. It is this observation of entities of tongue at a single moment of their existence which constitutes descriptive grammar.

Exact observation of an entity of tongue is not easy, even if consideration of it is limited to a single moment of its existence. This kind of work requires an apprenticeship, which is exactly what we are all engaged in here.


THE MENTALLY SEEABLE AND THE SAYABLE1

One of the features of psychosystematics is that one cannot progress without the help of representative diagrams. The advice of Leibnitz was to think schematically. One might ask why this should be so. The answer is convenience. A diagram can bring out a system of relationships better than words can. And the moment one accepts, along with Saussure, that tongue is a system, the use of diagrams is advisable.

1. The French, visibilité mentale, evokes the characteristic of being mentally visible or seeable: "seeable-ness", if one can coin a term. Similarly, d i c i b i l i t é denotes the quality of being sayable, "sayableness".


But, valid as this may be, it is not the only reason. If we dig a little deeper we discover — and this discovery is important — that the economy of language consists in making sayable — in translating into something that can be said — certain mechanisms which, deep in our minds, are already seeable. In this translation lies the economy of language. The structure of tongue is, deep within us, something mentally seeable which language translates, keeping to what is necessary and sufficient, into something mentally say-able, then into something orally or scripturally sayable, and then into the actual saying in speech or writing. The steps are as follows:

The speaker pays no heed to what is mentally seeable; only what is mentally sayable interests him, and this is, in fact, quite enough for his needs. The speaker therefore is unaware that, prior to what is sayable which he makes use of, there is initially the mentally seeable, which the sayable translates. And the translation can actually be reversed — a retranslation into what is mentally seeable. The diagrams used in psychosystematics are this retranslation. They are no mere artifices of analysis. I could go so far as to say that they portray a profound reality. The diagram representing the category of the article in French1,

is exactly that: the retranslation of what is mentally sayable in human language back into what is seeable mentally, the source from which it derives.

1. The system of the article, often referred to by Guillaume in his teaching, is treated more explicitly below in the passage, "System of the Word and System of the Article".


Thus we have a series of operations. With the mind's eye I apprehend a structural mechanism within myself: what is basically seeable. In order to make use of it in language, I resolve it and change it, first into what is mentally sayable, then into what is orally or scripturally sayable, and finally into the actual saying, speaking or writing. These mutations are the work of the speaker, who 'follows the path of greatest utility and economy.

The work of the scholar, of the analyzing linguist — rara avis — comprises the series of counter-mutations. The structuralist must know how to change what is physically sayable, either orally or scripturally, into what is mentally sayable, and the mentally sayable into the incipient (I need a word to refer to the very first instant of something) mentally seeable.

The study which I am attempting in these lectures encompasses the whole structure of language. It is a translation, back into what is seeable, of the sayable produced in language through translation from what is at the basis, the seeable. Here we are dealing with the very important opposition between the active, interiorizing forces which lead back to the "seeableness" of the underlying mental acts, and the active, exteriorizing forces which bring about the translation of this interiorizing "seeableness" into an exteriorizing "sayableness". The scientific task of structural linguistics is to translate what is sayable, the only reality the speaker is aware of, back into what is seeable, which is unknown to him, though he possesses it deep within himself. That is what we are doing here. Hence our diagrams and the impossibility of progressing in our work without them.

(Lecture of January 17, 1957)

INTUITION

On intuition: "The truth comes to us not only from reason, but also from the heart; the heart gives knowledge of first principles and reason, not being involved in them, attempts to oppose them but to no avail... And it is the knowledge from heart and instinct that reason needs for support, as a basis for all its discursive activity".

The modern scholar does not claim to understand what intuition consists of and in what conditions it may operate. He usually defines it in a negative way. Mathematical truths, he says, are neither consequences of experimental facts nor results of logical construction or deductions. Therefore they presuppose some mode of perception which is not to be confused either with sense experience or reason. At times, adds the scholar, we are aware of practising this mode of perception (in the work of discovery) and we can see


that it is quite unlike demonstrative knowledge. If we attempt to isolate the mode of perception we have used, we manage to discern certain of its characteristics. However, we are forced to admit that it remains a mystery; in affirming its reality the mathematician is putting a question to the philosopher rather than helping him answer one.

Intuition: the inevitable credo. The credo which requires no rational justification. Its historical variance in relation toman's movement away from primitiveness. The inevitable credo of human suc-cessivities. The credo of the very primitive man. The credo of the modern scholar. The uncertainties of the credo. The intuition of uncertainty. The poet might say: "I have no Adriadne and I am without her thread. And my obscure hunger — obscure because of too much brightness striking it — for the light of day."

These credos are projected in man's works: art, poetry, religions and finally — it would be better to say firstly — language. It is human language, better than anything else, that embodies these inevitable credos which, in my opinion, constitute human intuition.

I am certainly not claiming I can elucidate these mysteries. Yet, as a linguist, as a structuralist, as one operating in the psycho-systematics of language, as an observer of psychosemiology, I have at my disposal the best existing document for the study of intuition and of this very subtle mechanics, known to us by instinct, which I call intuitional mechanics. It is this intuitional mechanics which is the operator of the structure of languages, while the structure of languages, in turn, is a faithful reflection of its activity.

All the workings of this intuitional mechanics are subconscious. Subconsciousness and intuition are much the same: the effectiveness of the operations of this intuitional mechanics as attested in the structures of language, where the results of these operations can be seen, proves conclusively that within us there is an order of activities over which we have no conscious control. And the purpose of these subconscious activities is to increase not our knowledge but our lucidity, that lucidity without which the acquisition of knowledge would be impossible.

I cannot help noticing in the functioning of intuitional mechanics something closely related to man's instinct for preservation, although in the present case what has to be preserved and jealously safeguarded against any threat is the human capacity to think.

I see intuition, with its own operations — which fortunately are mirrored in the structures of language — as a series of undertakings whose safeguarding purpose is often jeopardized by the very course taken by thought itself. Thought often goes off (one could almost say rushes off) to confront that which would utterly defeat it if the operation of safeguarding were not continually successful.


There is something intriguirigly evocative in what I have just said. Here we see the human mind instinctively committed to its own preservation, in other words, to safeguarding its lucidity. It is interesting to note, just in passing, that men of great knowledge can be distinguished from men of great lucidity. Sometimes they are one and the same, which of course is splendid, but even better and more beautiful is the fact that a man of little knowledge can nevertheless be one of great lucidity. Napoleon, who was a thinker, often reflected on this encounter between knowledge and lucidity. And time and again he would give greater weight to lucidity than to knowledge — of any kind. This lucidity which he sought, which he prized and believed in above all else, he called, in appropriate Napoleonic style, Ze juste sentiment des choses — the right feeling for things.

The human mind runs the risk of losing its lucidity when it encounters the infinite. At that extreme point, it has to operate an anastasis within itself, a resurrection of its expiring lucidity. The linguist, extraordinarily privileged in this regard, is able to follow this anastasis as an observable event, etched in the structure of language.

(Lecture of February 14, 1957)

WORKING HYPOTHESIS AND THEORY

Anyone can learn to cook but real chefs are born, not made. Similarly, personal experience leads me to think that one can learn how to interpret historical fact, whereas real theorists are born. And for better or for worse, I was born a theorist. What attracts me to theory is that it substitutes understanding for the type of seeing that takes in only observable facts, an understanding which leads to a seeing of a higher order, the order of the understanding itself. (For me theory substitutes for a seeing which is of the order of direct observation, a superior seeing which is of the order of understanding.) In my view understanding is theorizing to the extent that one understands, little though it may be. A good theory is the ultimate in understanding. Where tongue is concerned, one must distinguish carefully between the theories that can be formulated about it, and theory that tongue itself is.

A theory — any theory — must necessarily confront the facts. And this confrontation with fact is the critical moment for a theory. Yet merely fitting the facts is not enough to justify a theory. It is possible to construct a theory from the facts in order to explain them. And sometimes different theories are constructed in an effort to explain the very same facts; theories may actually be so different as to be quite incompatible, each excluding the others. Constructing


a theory from the facts in order to explain the facts is the equivalent of putting the fact (the example) in the position of protagonist one moment and in the position of antagonist the next, making the fact do two things. This is how theories founded on what are called working hypotheses are constructed. (Meillet was opposed to this cult of working hypotheses.) The following diagram illustrates my point:

Working hypotheses multiply theories and theorists. The circuit runs from observed facts back to observed facts, through an explanatory hypothesis. The observed fact is both protagonist and antagonist.

Theories of this kind are of little value in linguistics. Too many of them can be constructed for the same facts. That is why, being a theorist by natural inclination (whether this is for better or for worse is a matter of opinion) I have always refrained from constructing theories of this sort. To suit me, a theory, the ultimate in understanding, must satisfy the following formal conditions: it must confront the facts from the position of antagonist, of course, but it must itself be based not on fact but on some absolute and inevitable exigency. And it must proceed from absolute exigency to absolute exigency until it encounters the facts. The protagonist of the theory, then, is a certain absolute exigency, taken into consideration at the outset, and the antagonist is the fact, confronted when the theory has, in the words of the Apostle, "run its course". It is wise and only fair to let it run its course. This c u r s u s of the theory is not possible everywhere. I consider it possible in linguistics and, more generally, in everything that is a construct of thought. Such is the case with tongue and with language as a whole, both of which are constructed in thought before being constructed with signs.

The schema of a good linguistic theory is linear. This is a basic requirement.


The inevitable ► antagonistic (in the position of facts protagonist and as close as possible to (at the end elementary intuition). of the cursus)

The validity of a theory of this sort is doubly founded: it is rooted in the inevitable at the outset (this means much, although it involves little) and then there is its confrontation with the facts. Besides, between the two, there is the course of its step-by-step progress. In other words, we have the double test of a proper start and of a finish which corroborates the appropriateness of the start.

It should be pointed out that good method lets the theory work its own way through to the facts (and also to the proof based on fact) by its own momentum, by its own means, motu proprio. The route it works out for itself will be determined by its starting point and by each step along the way. A theory calls up a sort of mental creation of the facts it is approaching. Error, derailment, catastrophe would result if, during the normal step-by-step approach, the theory suddenly found, thrown across its path just when it was going to encounter them, the very facts it was interested in, presented in a manner alien to it and in total disregard of the way it was proceeding. A theory must be allowed to carry out the explanation it has begun. Care and tact are required to avoid suddenly substituting any explanation which would be foreign to the course the theory is pursuing as a consequence of its starting point and so would disrupt the explanation already begun. Meillet, who had a gift for finding the right word, called this avoir le tact de la docilité provisoire, being tactful enough to wait and see. A theory has to be given the chance to confront the facts successfully on its own.

(Lecture of February 7, 1957)

PROOF IN SCIENCE

At the beginning of the last century, scholars readily accepted the belief, or should I say the instinctive feeling, that tongue was based on distinctions of a metaphysical kind. This feeling they had about the underlying nature of tongue was certainly not false. Unfortunately, the philosopher-linguists of the time, unable to combine careful observation of facts with just the right proportion of deep, abstract reflection, could not clearly see what the metaphysical framework of language consists of. Failing to discover its true substance, they became lost in the clouds, often with only airy nothings in their grasp. Those with better minds, preferring solid ground,


soon tired of this and by their just criticisms threw into disrepute views which were actually based on a correct assumption and were wrong only in that they could not establish its correctness irrefutably. "Science lives on proofs, not on truths." This dictum of Meillet's is profoundly appropriate here; but it should not be allowed to dissimulate the fact that the human mind, especially in the case of those where it has reached its greatest acuity of perception, is made in such a way that it glimpses the truth long before it is able to produce proof. Science does indeed live on proofs and not on unproven truths; but, quite apart from the science that is capable of proving, there exists a prior state of science, a sort of prescience, characterised by its ability to grasp what is true from a distance, from so far away that the proof at that moment is not yet accessible. Some may disapprove of my daring to put forward such proposals, but I have been prompted to do so by requests for convincing proofs in cases in which I am not personally interested, cases where the proof is, by the nature of things, slow to appear, at least in any complete fashion. To evaluate these cases properly, one could take into account the faculty that certain minds possess of perceiving the truth not merely before the proof can be brought forward, but even under conditions that seriously inhibit its production. I am all the more reassured in expressing these opinions by the fact that up to now I have never needed to evoke on my own behalf the existence of perceived truth which could not be proved. I have in fact always succeeded in producing at least partial but valid proof of the theoretical views I have advanced. The fact nevertheless remains that I have often dimly perceived the truth long before being in a position to prove it. I have also at times suddenly come across the proof of a truth that up to then I had never even suspected. In actual fact, that is what has happened to me most often. I see the proof first and the truth afterwards. That is the natural bent of my mind.

(Lecture of February 3, 1944, Series B)

CHAPTER 3

GUSTAVE GUILLAUME AND LINGUISTIC TRADITION

GENERAL GRAMMAR

The old general grammar which was in favour in the 18th century — and which, even after Ferdinand de Saussure made descriptive grammar respectable, is sometimes ill-advisedly hearkened back to — had become bogged down in a basic error: that of seeing in the structural stage of highly evolved idioms evidence that certain distinctions are necessary and inherent to human thought. These inherent distinctions, it was claimed (without any real proof), were indispensable. But a more extensive study of human language in both space and time, one that took in all the languages of the world that have left any traces — there are several thousand of them — revealed that those distinctions formerly considered universal and necessary (such as the noun/ verb distinction) are not found in all languages. In fact in the form we know them, these distinctions were found in only a relatively small number of highly evolved idioms. In my lectures here, where I have been dealing with the different stages of development of the word — its psycho-systematics — we have seen that the word as found in French, and in European languages in general, is a relative latecomer in the evolution of language. There are numerous idioms in which the word is quite different from what we generally understand by this term, to the point that it is quite wrong to consider, as is sometimes done, the Chinese "word" and the French "word" as being equivalent forms. Syntactic comparisons based on the postulate of a formal equivalence of the Chinese word and the French word simply do not hold water. The only valid, acceptable syntactic comparison is between idioms whose words are at exactly the same stage of development or at least subject to the same basic constructive principles. This is a principle that is all too often overlooked or vexatiously disregarded.

But just because the facts of general grammar disappear in the light of observation, one should not conclude that they do not exist. Their disappearance simply indicates that such facts must be sought where they can truly be found. This point has been hitherto neglected,


although in recent years some small but appreciable progress not unrelated to my own teaching has been made in different sectors following this line of study, formerly misunderstood or even completely disregarded.

(Lecture of November 21, 1947, Series C)

A MISUSE OF LOGIC

The question I now wish to deal with briefly concerns the analysis that logicians have made of the part of speech called "verb", taken in its predicative function. They teach that in the mind of the speaker a verb with this function consists of an affirmation of the subject's existence plus an attribute. In je marche one is supposed to understand analytically je suis marchant, and presumably this is the situation for all verbs. My students — at least those who are familiar with my way of thinking — will have guessed that I find this explanation absolutely gratuitous and, what is more, inaccurate; being an ostensibly logical analysis, it is quite removed from the realities of language itself. If, in fact, the speaker does say il marche and not il est marchant, it is because he is logically thinking il marche and in no way feels that he would have a better insight into the verb, or a better logical analysis of it, if he were to use the paraphrase il est marchant.

The question raised by this small, somewhat specious problem of logical analysis at least serves to call our attention to the possible semantic content of the verbal category. In a language like French, the category of the verb is made up of the incidence1 of the whole of its semantic content to the ordinal person.2 Now, as some other languages have done,3 it would be possible to take a different point of view: the verb could be reduced simply to the relationship between the subject and the general notion of verb, leaving aside the verb's particular, lexical nature. One can see that in a language of this kind there would be very few verbs: there would be no more than the number of ways that the subject could be related to the general notion of verb. As a result, the verb category would be reduced to a small number of extremely abstract verbs (of which our auxiliaries can give us some idea), and it would be an almost entirely formal

1. That is, the fact of being incident upon, hence of being predicated about.

2. That is, the grammatical person of the subject: first, second or third.

3. Guillaume discusses one such language, Basque, in his Leçons de linguistique 1948-1949, Série B, pp. 79ff.


category including only a small part of the lexical meaning, the most general part, all the rest being left to the field of the noun.

This would be the case of French if, instead of forming the verb as we do without resorting to an allegedly analytical and logical periphrastic form, we normally said je suis marchant instead of je marche. The category of the verb would thus ipso facto be reduced to the single verb être and all the other verbs would have emigrated, under an adjectival form, to the field of the noun.

This is something that the French language has never considered. French has constantly tended to distribute lexical meaning equally between the two categories (nominal and verbal). Therefore when the logicians propose that we should analyse je marche by means of the paraphrase je suis marchant, they are merely displacing the lexical meaning of the verb and attributing to it a linguistic position that French has been careful not to adopt, a position that cannot be reconciled with the systematics — I almost said the logic — of French. The logicians, as it turns out, seem to consider quite unwittingly that the true place of lexical meaning is in the field of the noun, and that the logical analysis of a verb consists (goodness knows why) in putting it back there. Of course it is true that from a certain point of view it would be logical for a language to keep the lexical meaning strictly on the nominal side and reduce the verb to the status of support words, needed for the introduction of lexical meaning into the verb field at the moment of discourse. But it is also true that it is just as logical, and perhaps even more so, to localize lexical meaning equally in the two fields, nominal and verbal. One of the advantages of this system is that each category then has lexical meaning more in keeping with its nature. What is more, it would not require of discourse the rather uneconomical operation of continually transferring noun elements like the participle into the verb field, where these elements would have to find a small number of specially constructed support verbs to serve as the needed extra-nominal, verbal support.

It seems that these remarks inevitably point to the following conclusion: analyzing the verb as a verb affirming the subject's existence completed by an extra-verbal form playing the role of attribute is an exercise of no intellective value whatsoever. Furthermore, in the case of French it involves the serious error of ignoring the true systematization of the language. One of the systemic aims of French is unquestionably to have lexical meaning equally represented in each of the two opposed fields of tongue, noun and verb. This equal representation of lexical meaning in the two fields of tongue is in accordance with the logic of French which — like the logic of Latin, its ancestor — is a logic of symmetry. There exists no logic transcending the diverse interior organization of languages, but as many logics as languages. "Logic", as far as linguistic systematization is concerned, is an almost meaningless word. On the other hand, the word "symmetry", which I have just used, means a great deal


more, as I have had the opportunity to demonstrate on numerous occasions. I shall have just such another opportunity in a short while when I begin a detailed study — something I have not yet done here at the Ecole Pratique des Hautes Etudes — of the French sytem of vert forms, whose great symmetry, inherited from Latin, is so far not found outside the Romance languages.


AN ERROR OF HISTORICAL LINGUISTICS

It cannot be considered too harsh a judgement to say that historical linguistics has been far too preoccupied with studying how items are handed on from the past, while almost totally ignoring the study of the relationships established between these items and maintained for purposes of systematization.

This notable lack of interest in the systematic relationships established between items inherited historically is the reason why historical linguistics, which for a long time was presumed to be capable of explaining everything pertaining to language, has in reality explained nothing. All it has done (and this is by no means negligible) is relate events that have taken place within language. Traditional historical linguistics is, above all else, narrative. As soon as one wishes to describe a language system at a particular point in time, one moves away from historical narration and, like it or not, ends up examining the facts from another viewpoint: that of a systematic relationship established in the instant, in the "static instant", between handed-on items which, though accidental and irrational in themselves, are made rational by integration into the system of the language.

Because the explanations of historical linguistics never go much beyond a consideration of accidental items, they do not, for the most part, get us very far. This shortcoming can be illustrated by means of an example: the little word known in grammar as the article. Historical linguistics tells us here that the article goes back to a Latin demonstrative which, after the weakening of its demonstrative sense, became a regular introductory sign before the noun in modern French. Of the meaning of this sign, the role it plays, or the need it fulfills in our thought processes, nothing or almost nothing is said, and what is said is not very significant. The article is said to be a determiner of the noun. (Merely in saying this, historical linguistics is leaving its own axis for that of systematic grammar.) The traditional terms "definite article" and "indefinite article" stem from this notion. A noun preceded by the article le is considered a definite noun, a noun preceded by the article un, an indefinite noun. Nothing could be more vague. In fact one might go so


far as to say that nothing could be more meaningless. This is all just playing with words that are accepted without question but that do not relate to anything real or tell us anything about the underlying thought processes to which the article owes its existence, its raison d'être.

For lack of any such explanation — and anything that falls short of this is insufficient — linguists have been led to try to account for the regular usage of the article by explaining it away as habit. People must have somehow developed the habit of putting the article in front of the noun, and the habit was passed on and spread. In other words, one habit — since it was also a habit not to put the article in front of the noun — leads to a new, opposite habit. Yet something of the original habit seems to remain: this would explain the presence of quite a few nouns without an article in discourse. But explanations of this nature are hardly worthy of attention. Their inadequacy is exasperating.

Some grammarians have deliberately left aside the problem of the real nature of the article, dwelling mainly on the secondary fact that, having a variable form, the article is a semiological indicator of gender for nouns whose gender has otherwise no apparent mark: la chaise, le fauteuil. The inadequacy of these explanations is painful. And one is surprised to see them incessantly paraded, not only in the grammar books intended for teaching children, but also in works having a more lofty aim. That such explanations, which really explain nothing, are persistently repeated on all sides, is due to the grammarians1 obstinate refusal to look at anything other than the axis of items handed down historically, while completely ignoring the axis of systematic relationships. The habit of using the article results from a historical development: and being merely a result, there was no necessity to retain it. Speakers could have refrained from adopting the habit. If the habit of preceding the noun with an article was in fact retained, this was precisely because it seemed to satisfy the requirements of a relationship which had found a place, and an important place, in the system of the language. The real problem, therefore, is to find what it is in the system of the language that justifies the continuing role assigned to the article of introducing the noun.

To state the problem of the article this way indicates a radically new kind of curiosity which is, for the most part, alien to language historians, who have remained hypnotised by the observation of the individual facts of historical development.

(Lecture of February 13, 1948, Series C)


TONGUE AND ITS HISTORY

...This leads me, incidently, to make my position known on the controversy that began in 1916, with the appearance of Saussure's Cours de linguistique générale, over the possibility and even the necessity of eliminating diachrony — the historical approach — from descriptive grammar.

For Saussure, la tangue is, at each instant of its existence, a synchrony of relationships, a steady-state system whose reason for existing is found within itself, in its own organizational laws. There is a good deal of truth in this point of view, and consequently the Saussurean doctrine has received widespread acclaim and enthusiastic approval. It was strongly and clearly felt that the speaker mentally possesses his language system as a steady state, and that he possesses it exclusively in its present state, irrespective of any past states that have been left behind, abandoned or transformed. All of which is quite true.

But the widespread approval which greeted this idea in the beginning soon gave way to a certain reserve among knowledgeable linguists. When it was put into practice, difficulties arose in making a good descriptive grammar based entirely on synchrony, without any appeal, however slight, to diachrony, to history. There are certain cases — indeed major ones — where resorting to diachrony can be avoided: for example, the question of the representation of time in a number of languages. But there are other cases where it is impossible to avoid resorting to diachrony, if one wishes to account for the facts with any exactness.

This leads one to ask why certain questions in grammar require diachronic consideration, while others do not, and may even benefit from its exclusion. In languages, some things are better explained by considering only their present state without referring back to former states that have since disappeared.

The answer to this question is not easy. Careful examination suggests that it is necessary to distinguish between problems that have reached a solution and subsequently arise again,1 independent of the way this solution was reached, and problems which arise again before their solution has been reached, while it is still being worked out.

The first set of problems, having left behind the working out of their solution and so being independent of their historical past, calls for, an explanation which likewise will not need to be historical, while the second set, still dependent upon a historic past

1. For Guillaume, no problem of representation arrives at a definitive solution. See below, "Problems of Representation and States of Tongue".


that they have not yet left behind or transcended, will have to be explained by reasons drawn at least partly from history.

In other words, there are problems sufficiently removed from their past solutions for them to be considered intrinsically, separately, without having to refer to solutions that were given them in the past and that made them more explicit. Not all problems, however, reach this level of autonomy.

A certain number of problems have never really transcended their old solutions so that, when re-posed, they are not independent of the past. Therefore, when considering the facts of language corresponding to these problems one must take this incomplete liberation from the historical past, from diachrony, into account. These problems intrinsically remain more or less dependent on it. In order to be true, that is, to conform with reality, the explanation for these cases will have to remain historical to the degree required.

Thus the fact remains that the linguist who proposes to describe a language will need to do so without resorting to historical explanation whenever the facts he is examining are the solutions to problems that have transcended their former solutions and so become independent of them. And on the other hand, he will need to resort to some degree to historical explanation whenever the facts under consideration are the solutions to problems that are not yet free of their historical solutions, because the solutions are as yet incomplete.

I thought it would be useful to raise this question in order to ease the minds of those — and they exist — who, immersed in Saussurean doctrine, find themselves incapable of explaining the facts of language in a purely descriptive manner, without sometimes bringing in a little historical linguistics.

(Lecture of December 11, 1941, Series A)

A SHORTCOMING IN THE SAUSSUREAN ANALYSIS

As others have already pointed out, one rather special merit of Ferdinand de Saussure's Cours de linguistique générale was its opportunism. The work became almost immediately famous when it first appeared thirty years ago (the Cours dates from 1916). At the time, the prevailing doctrine in linguistics was, so to speak, essentially historical: history and the probing of the past permitted by the comparative method were expected to explain everything. This was of course an illusion, as could well have been deduced at the outset through careful reasoning, but when the Cours de linguistique générale appeared, this illusion was still the unshaken conviction of


most linguists and philologists if not absolutely all of them. Plainly the scientific situation was not good. No one could or would see the shortcoming inherent in the point of view everyone had adopted. And to aggravate the situation, the attention of linguists was turned almost exclusively towards parole,1 to the detriment of the study of langue, which is something quite different. It fell to Saussure to clarify the situation.

Among the errors of the time, I leave aside that of not separating, that is, of not considering as distinct and heterogeneous, the operations of thought which bring about the momentary construction of discourse and those which bring about the construction of tongue. This error had serious consequences; it persisted even after Saussure's book produced the effect that could be expected of it, achieving a unique and universal success. The work was highly esteemed, in theory a success. But while Saussure's general concepts were admitted and their appropriateness recognized and admired, practically speaking, little change came about in linguistics. The few attempts that were made to explain an état de langue from a purely static point of view on a "synchronic axis", to use Saussurean terminology, were, for various reasons, not very successful.

This description of the scientific climate prevailing when Saus-sure's work appeared shows how necessary it was to attempt to rectify the situation. Now if any such attempt was to be successful it had to come from a master respected throughout the scholarly world — someone already illustrious. Had it come from a little-known man, one who with all his scholarship had remained obscure, even a work of higher quality would not have had any immediate impact because it would have passed unnoticed. It is not enough for important things to be said: they must also be said by important men. Important men would be wise to seek out important things to say.

At any rate, in this particular case the author's reputation was of great service to the cause. And yet the cause would have been less well served had Saussure not been careful to put forward arguments that did not clash too strongly with the intellectual trends of the times, even if his ideas were basically quite revolutionary. It is this moderation in his attack, and the constant care on each page of the book not to increase opposition to the new ideas that he was advancing, which has been called Saussure's opportunism. There were certainly things that the master would have said, but the time was not ripe for them and he would not have found a favourable audience.

It may be worth our while to examine the exact nature of Saussure' s opportunism, of which I was aware quite early on, and which

1. To avoid confusion with Guillaume's similar but by no means identical use of such terms, Saussure's well-known terminology is left in the original French.


has also been pointed out by others. Saussure distinguishes langage, langue and parole, and he presents the following equation, which is fundamental for him:

langage = langue + parole

This equation must be interpreted according to a relationship which sees langage as the whole or totality, of a successivity, the suc-cessivity involved in moving from langue to parole: from langue, which is permanently present in us in a state of potentiality, to parole, which is present in us momentarily in a state of actuality.

This interpretation, which is my own, is not found in Saussure's book, but even though it does not appear explicitly, it is implied throughout. When the book is viewed as a whole, the idea is actually there, but it is an implicit one. This manoeuvre of keeping implicit a more penetrating explanation than the one presented is a clear-cut instance of the opportunism to which the work owes, if not a greater success, at least less resistance to its new doctrines.

As Saussure pointed out, langage is certainly a whole with two components, langue and parole. Before looking at this Saussurean equation with a more critical eye, we might represent it as follows:

This representation clarifies the fundamental question under discussion and helps to reveal the shortcomings of Saussure's formula:

langage = langue + parole

One factor which the Saussurean formula fails to consider, but which should be very carefully taken into account in any linguistic question, is the factor of time. Langage, as an integral whole, includes a successivity, the passage from langue — permanently present in the speaker (and consequently in no way momentary) — to parole, present in the speaker momentarily, for periods of time that may be close together or far apart.


Most opportunely, Saussure's work makes no mention of this suc-cessivity: the notion complicates the problem. Had Saussure taken it into account, it would have led him to a more complex yet truer view of things, but one that would not have found favour with historical linguists who tended somewhat simplistically to consider that the only explanation is diachronic, explaining historical consequences in terms of their antecedents. The mental pattern which historical linguistics adheres to is the following, very simple one:

antecedent (ascertained, attested)

regular (observed) development

consequent (generated by the development; result also ascertained, attested)

Then all one has to do is keep applying the same mental pattern, either working back upstream in time towards the origins or, inversely, working downstream towards the subsequent historical facts. The only real difficulty in the method is determining the stretch of time between antecedent and consequent to allow for a regular, differentiating development. Historians in fact generally realized that it was important, for rigor of method, to keep this stretch short. A recognized regular development, the kind of regularity which is abusively called a "law", is valid as such (as regular) only between fairly strict limits. In historical studies it often happens (and this is a source of many difficulties) that the limits in question are badly chosen. It might be added here that the choice of these limits depends solely on the accuracy of the historian's observation, and on nothing else.

But now, after this critical look at the mental pattern of historical linguistics, let us return to the Saussurean equation for the relationship between the three terms langage, langue , parole. If we include the factor of successivity between langue and parole, this relation becomes:


The speaker finds langue available to him, ready for use, and he uses it in order to speak. He makes the transition from tangua to parole. At this point the theory encounters an obstacle. During the moment of expression, the speaker certainly passes from langue to parole, that is, from langue to actual, momentary parole, which can be heard and has a physical existence. But although Saussure does not mention it, this transition from langue to parole is really only the passage from virtual speech, indissolubly linked to the mental side of tongue, to actual speech, realized and physical. Virtual speech, linked as it is to tongue by being an integral part of it, is speech which is silent, not physical, because the entities of tongue are mental. It is not difficult to become aware of the reality of this non-physical speech. Each notion at the level of tongue carries with it the idea of the sound or sounds that signify it, but only the idea of them, not the real sound or sounds.

It follows from this — and the partial science within linguistics called phonology has this as a foundation — it follows from this that ideal speech, which is a part of tongue, is something quite different from actual speech, which is a materialization of ideal-speech. And here we run into a relationship which is already familiar to us, namely that multiplicity is associated with the actual and unity with the potential. Opposed to ideal-speech, which as a conditioning factor is one, there is the immense diversity of actualized speech, which varies according to the speaker and even his particular situation.

Taking all these points into account, the Saussurian figure now becomes more complicated but more accurate:

1. Guillaume has replaced Saussure's parole ("speech") with discours ("discourse"). Parole ("speech") is now seen at both levels of language.


This analytical figure shows that the successivity involved in language goes from tongue to discourse, and that in tongue as in discourse there is a link and congruency between an entity of speech and an entity of thought. There are, however, differences between them that need to be brought out and emphasized. In tongue the link between mental content and speech is a link on the level of ideas where speech, which is in itself physical, does not actually get beyond the mind. On the level of tongue, speech, caught while still in a non-physical state, is a mental representation of its physical self.

On the level of discourse the story is different. The link between mental content and speech is mental of course, as it was in tongue, but in such a way that the physical nature of speech emerges actualized, materialized. Considered in itself, speech has thus left behind its original mental condition. On the level of discourse, speech takes on a body, a perceptible reality. It exists physically and is no longer merely a mental representation of itself. It does not cancel out what existed before but rather realizes it, conferring on its original mental representation a tangible materiality, an indispensable condition if language is ever to be exteriorized.

And so we may perceive and understand that speech is one thing at the deep level of tongue and something else at the surface level of discourse. Non-physical at the level of tongue, it becomes physical at the level of discourse, provided that discourse does not remain internal. This transformation is the one we have to take note of during the functioning of language, when we study speech. (Seen in this new light, speech corresponds to Saussure's signifiant.)


COMPARATIVE LINGUISTICS AND PSYCHO-SYSTEMATICS

...I shall try to show how one makes the transition from the semiological observation involved in traditional comparative linguistics to the psycho-systematic observation of a new kind of comparative linguistics whose very conception blazes a trail for the future — a trail which will be boldly taken by the researchers of tomorrow. (Any science has its capacity to observe. One day I think it will be justly recognized that I have considerably increased the capacity to observe in linguistics.)

I am going to start not with phonemes, but with a phase of language in which tongue is already composed of wholly constructed words. This phase goes beyond words like those found in the second


area of language,1 most of which (though not all) are words whose construction is not totally accomplished on the level of tongue but rather is completed during the passage — the transitus — from tongue to discourse. From the very earliest date, the Indo-European languages have had all the features of languages with words completely constructed in tongue, a fact which greatly facilitates their study. In these languages the word adds morphology to a radical, R, by placing one right after the other. This may be diagrammed as follows:

Radical + morphology ► part of speech

Through the generalizing effect of the added morphology, the word is borne along right to the part of speech, the final universalization, which specifies, characterizes, its total content.

When presented in these terms, the composition of the word in the Indo-European languages is psycho-systematic; but at the beginning, the linguist started by observing it semiologically. How, almost half a century ago, I saw the link between semiological observation (as ordinarily found in comparative linguistics) and psycho-systematic observation, which I was soon to consider devoting a lifetime to, I shall now explain.

Let us consider the well-known correspondence between the following words:

Avestic baranri (I carry) Sanskrit bhàrâmi (I carry) Armenian berem (I carry) English bear (to carry) Albanian birni (you carry) German gebaren (carry before bringing into the world) Russian beru (I take) Greek phero (I carry) Latin fero (I carry Gothic baira (I carry) Irish ber im (I carry)

This correspondence brings out a fact worth noting: all the words cited conserve the axial consonant r, whose role is to separate the radical from the morphological suffix (representing a structural

1. A term which designates, in Guillaume's theory of g lossogenic areas, those languages whose typology is attested in the ancient and modern Semitic vernaculars. Cf. his Leçons de linguistique 1956-1957.


"treatment" of the radical). The axial consonant separates the notional ideation linked with the basic phonemic group *bher from a transnotional ideation.1 This transnotional ideation stems, in variable proportions depending on the language, both from what has been inherited and what has been reconstructed in the semiological, physical component of the word and also in its meaning, its non-physical component.

This remark opens the way to a special study of what, in transnotional ideation, mental psycho-systematic construction and reconstruction is. And it also helps to maintain the link between this mental psychosystematic construction and those signs arranged in tongue to express it — that is to say, signs which, because of their physical nature, have the role of exteriorizing the non-physical interiority of a morphological system.

This work may be undertaken and carried out successfully in any Indo-European language, and, at the level of semiology, it regularly leads to bringing out the prime importance of the axial consonant in the structural semiology of the verb, and even outside the verb. The axial consonant's persistence in the above word correspondences is due to its important role.

Therefore, as a beginner in comparative grammar and grammar of any kind I paid a great deal of attention to the axial consonant. Some discoveries of real interest resulted. I was, for example, able to show that from a semiological point of view, the whole conjugation of the French verb is based on how the axial consonant is handled: as a general rule it is maintained and so intervenes everywhere between radical theme and audible inflection. The axial consonant is never allowed to drop before the audible inflection. And the axial consonant, systematically needed, is restored whenever necessary, if it happens to be lacking. Here are some examples. In the infinitive liRe, it is represented by r, which does not appear in the non-future forms of the conjugation of the verb. Before audible inflections, the axial consonant is, through systemic phonology, restored in the form of an s, nous liSons, ils liSent, ils lisaient. In renDre, the axial consonant d is kept throughout the conjugation. The regularity is perfect. In p r e n d r e , the axial consonant d appears in the infinitive, but not in the conjugated forms. We do not conjugate nous prendons9 although I have heard a child produce this construction through his own ability to say mentally — I know he had not heard it. I have also heard others, conserving the axial consonant, say mouru, il est

1. Guillaume here distinguishes between the forming of particularizing, lexical meaning connected with the radical, notional ideation, and the subsequent forming of generalizing, grammatical meaning evoked by the morphology, transnotional ideation. See below, "Operations Constructing the Word: Discrimination and Categorization" for a development of this distinction.


mouru. These childish mistakes reveal the efforts of the child to rediscover the constructive processes of tongue from what he hears: his task is one of intuition. The child knows the language when he knows its constructional mechanism and how to use it, and when, in order to use it, he has realized that it is connected with an aphys-ical mental mechanism, that is, tongue itself. The signs are only-needed to exteriorize its interiority.

It is important to note here that apart from the signs needed to exteriorize the aphysical mentalism, language admits nothing physical into it. Everything in language except for the signs is aphysical and qualitative, whence the reducing of the relationship universe/man, whose limiting terms evoke physical magnitudes, to its aphysical counterpart, the relationship universal/singular, fundamental to the structure of tongue...1

The discovery of the axial consonant's role and of the consequences stemming from its continual restoration gave me access to a semiology sufficiently well organized for me to discern, without too great an effort, the psycho-systematics whose inner workings it exteriorizes, providing something which is mentally sayable, but having no physical existence, with a means of exteriorization, making it physically sayable. And once I had started out on this road of mine — "You're not bothered by other people on it," a late colleague used to say to me — all I had to do was let myself be carried along by what the facts taught me. Not just what they made me see: what they made me understand.

Psycho-semiology tends to become a successful reflection of psycho-systematics. In the most evolved languages the result is remarkable. Simply by paying close attention to psycho-semiology, one soon perceives the psycho-systematics beneath it. At that point psycho-systematics becomes of overriding interest because it is so much more rewarding than describing perceptible appearances.

...Languages have two clockworks: the psycho-systematic clockwork, exclusively concerned with the construct of thought called tongue, and the psycho-semiological clockwork, whose rudiments are taught by comparative linguistics. No more than the rudiments, since one can see the psycho-semiological clockwork as it really is only if one clearly sees the psycho-systematic clockwork beneath it — for the role assigned to the psycho-semiology is merely to be a physical reflection of what lies deeper. And the opposite is true: one can see the psycho-systematic clockwork clearly only if one sees above

1. For a development of this idea, see below: "A Structural Device: the Radical Binary Tensor."

2. See above "The Mentally Seeable and the Sayable" for a development of these notions.


it, the psycho-semiological clockwork. What is learned from the one helps us to understand the other.

(Lecture of March 21, 1957)

CHAPTER 4

THEORY AND THE SCIENCE OF LANGUAGE

THEORY: UNDERSTANDING IN THE SUPERLATIVE

...I have just used the word theory. To theorize, in my teaching, means to understand to the highest degree. Theory is the superlative of understanding. This is why it is necessary to complete one's understanding of things by means of theory. Engineers particularly are not satisfied with the understanding they have of the facts they deal with until they have established a theory about them, one generally based on mathematics. Among linguists, theory and theories are the object of considerable mistrust. Many see in them an inadmissable inclination to plunge into the abstract, into fictitious abstractons, leaving behind the cherished horizons of the concrete which seem so scientifically reassuring. They have lost sight of the fact that in itself the concrete may be observed, but not understood. To understand even in the smallest way involves abstract reasoning. One could say that there is no seeing totally lacking in understanding, and consequently it could be said that seeing, in its full scope — an eminently powerful seeing — is that which grows, often quite suddenly, out of a powerful understanding. A linguist should be constantly concerned with seeking out the insight that is rooted in a profound understanding.

Without explicitly demonstrating the point, Saussure's Cours de linguistique générale teaches quite rightly that tongue is a system. Now a system can only be seen to exist by the person who understands it. For the person who does not understand it, because he sees only the apparent constituent parts revealed to him by the seeing of direct observation (it is possible to come across systems without seeing them), the system does not exist. A linguist obsessed by direct observation, resolved to see everything by simply observing and carefully avoiding the least substitution of the seeing of understanding for the seeing of observation, would have the impression (which actually appeals to some because of its vastness sui generis), that tongue is an immense disorder, an immense non-system, in which the mind


is irretrievably lost, and where it is foolish even to seek direction, since by nature disorder defies understanding.

This seeing of understan ding eventually attained in linguistics supposes a repeated movement to and fro, from seeing to understanding, understanding to seeing, seeing to understanding. The movement is like an oscillation and could be represented as follows :

Seeing1 Understanding1




and so forth, indefinitely. From Seeing1 to Seeing4 there is no difference in the materiality of the things seen, but there is a difference in the understanding arising from seeing them, and in the resulting seeing of understanding.

This point can be illustrated by the legend of Newton's apple. Everyone had seen apples and understood more or less why apples fall. With Newton, this very limited understanding suddenly became a far-reaching understanding, and the resulting seeing of understanding was the mechanism of universal gravitation. Yet it remains a seeing of understanding capable of stimulating a new and more powerful understanding from which will result a seeing that surpasses even Newton's. Thus science progresses. This is the only way it can progress. Experience, which is a seeing, has no virtue except insofar as it can give rise to a new understanding. Scientific activity in extenso is an oscillation from seeing to an understanding which transcends the seeing, and from this understanding to a seeing of its own order.

It could be said without any exaggeration that traditional linguistics has had too great a tendency to stay within the bounds of the seeing of observation, making little or no effort to go beyond simple observation to a seeing of far-reaching and powerful understanding. Because it leads linguists to see rather than understand, this tendency in linguistics, a tendency which has grown, has resulted in research directed more towards consequences than towards conditions. Thus, acts of expression have been persistently (I almost said obstinately) studied, while the prior acts of representation have been almost totally neglected.

(Lecture of November 29, 1956)


TONGUE IS IN ITSELF A THEORY

This second Thursday lecture — with a smaller audience than the first I am sorry to say (because, of the two, it is far more suitable for awakening interest in linguistics) — is my research lecture. In it I retrace the various paths my thought has followed, sometimes hesitantly, wondering which direction to take, in the quest to discover the truth of things, hidden beneath appearances in languages.

And this truth is in itself an order. The results achieved to date in the various areas of tongue prove it. I say "order" because tongue is a theory in itself, and consequently lends itself to theorizing. People are perhaps not sufficiently aware of what a theory really is. A theory is nothing other than a knowledge of the relationship of subordination that exists between a great number of particular facts and a small number (even as few as one) of general, overruling facts.

Now this is precisely what a language is: all the particular, contingent, fortuitous facts, born of chance, to which it apparently owes its existence, are dependent on a small number of general overruling facts, though one would hardly suspect it at first sight. While much less visible than the particular facts, these general facts are nonetheless the essential structural ones, and so it is most important to become acquainted with them first. But since they are hardly visible in themselves, they are the ones that are usually overlooked the longest. Studies in historical linguistics and comparative grammar would group and organize all their particular facts very differently were there a clear knowledge of those few overruling facts which control the others.

If the expression were not a trifle strange, one could say that it was by "theorizing" itself, through a kind of natural theorization, that the mind provided itself with the system of tongue. The theories that the mind has thus built naturally from itself are not recorded as such in discourse although they may be reflected there. So they are not to be sought in discourse: they are recorded at a deeper level of thought, in tongue. The overruling facts, on which depend the multitude of particular facts that go to make up a language, are few in number, and one notices that from age to age they are always the same, provided one is careful to remain in the same area of language.

(Lecture of December 16, 1943, Series B)


THE PRESUPPOSITION OF ORDER IN THE CONSTRUCTION OF SCIENCE

...It follows that this type of system is well illustrated by the forms of the Germanic languages, but the design that these forms reflect is situated at the deepest levels of thought, in the unconscious. Speakers of a Germanic language possess it, live it, use it, it is almost organic in them, but one could not say that they know it. They are quite unaware of it. They do not need to know it consciously in order to use it. The task of the linguist and the grammarian (our task here) is to discover this design, this systematic schema, allowing ourselves to be led by careful observation of the facts — of the world as it appears to the mind — and also, it must be said, by the presupposition that there is order among the facts, an order to be discovered.

This tenacious and intuitive presupposition, whose mental origins are quite obscure, is the great stimulus behind scientific research. Where it is not found, scientific research slumbers. But when the mind becomes concerned with the way the universe is constructed, the need to provide itself with representations of the universe that satisfy certain aesthetic conditions prepares it to undertake scientific research.

In studying a particular system of tenses, as we are doing here, one is led on by the idea that there reigns within it a certain order that is pleasant to contemplate. And experience shows that this assumption is not wrong.

As soon as the mind catches a glimpse of this order, it becomes concerned with discovering the principle behind it. Once the principle has been discovered, or simply glimpsed or even suspected, the mind uses it to try to piece together the edifice it is interested in, with the hope of grasping, both analytically and synthetically, the coherence it believes in.

Without this belief, deeply rooted in the human mind, that the universe is ultimately coherent, there would never be any active or prolonged scientific research. The point of view that the world is as one finds it leads nowhere. Science begins with the analytical discovery of a universe which has secrets it does not disclose at first contact and which do not yield to simple direct observation.


PART II

FROM PROBLEMATICS TO SYSTEMATICS IN THE STUDY OF LANGUAGE

THE SIMPLICITY POSTULATE

A truly linguistic curiosity, a curiosity of the highest order, which is stifled when restricted to a consideration of historical data only, is concerned with discerning the nature of the operations of thought which form the basis and the framework for the structure of tongue. Clearly, the question is a major one in structural linguistics.

It could be assumed that the basic operations on which the structure of tongue is built are extremely diverse. If this hypothesis is true, tongue, though still a system, would have to be an extremely elaborate, highly complex one. But then if the system's natural development does not tend in the direction of greater simplicity, we would have to explain how even the simplest man's mind manages to keep such a complicated system ready and available for use.

A second hypothesis, which appears more reasonable a priori,, is that the structure of tongue is founded on certain basic operations that are neither exceedingly numerous and diverse nor highly complex, but on the contrary, are few in number, basically little varied, and, in the last analysis, surprisingly uniform. This would explain why the construct called tongue is easily grasped and why even the simplest of men can learn a language.

The question of the nature of the operations on which tongue is founded is, as I mentioned earlier, a major problem in structural linguistics. It has held my attention for many years now, and after a great deal of observation along with much careful reflection, I have come to believe that the operations on which tongue is founded are essentially simple — and few in number. Applied repeatedly to the results they have already produced, they are by nature the very operations from which human thought derives its potential. These primordial operations from which human thought derives its potential are also those which form the basis of the structure of tongue. In the light of what was outlined earlier, this amounts to saying that


the relationships already existing between items from a past state of tongue and those being worked out to further the construction of tongue, are relationships determined by certain primordial thought operations aimed at augmenting the potential. To express it a little more concretely, they are the very operations that provide the potential for thought to operate.

In order to discover these processes on which tongue is founded — at a certain point in my scientific career, I found it necessary to state the problem in these terms — to discover the foundation processes of tongue, a question to be asked (and which I did ask myself) was: which basic operations are indispensable for the mind to operate? Indispensable to the point where getting rid of them, if it were possible, would leave the mind powerless. This question led me to begin research along different paths in the hope of finding a sound answer. It soon occurred to me that the foundation processes of the structure of tongue are the operations involved in the mechanism of extension: extension developed either in its inherent direction, that is, becoming more and more extensive and finally moving to the universal, or against its inherent direction, that is, becoming narrower and narrower and finally moving to the singular. Having arrived at this idea, whose striking elegance made it quite irresistible (the scholar is guided in his research by an exponent of elegance which he attributes to his own nascent ideas), having arrived at this idea, I determined to check its accuracy by methodically examining the major facts of tongue.

The outcome of this examination was that the facts, far from invalidating the idea (the idea that the extremely simple, foundation operations of tongue belong to the mechanism of extension and involve a tending toward or away from extension), actually provided nothing but confirming evidence. Not only historically early facts, such as those involved in the systematics of the word, but also historically late facts, such as the appearance of the article in evolved languages, offer the most convincing evidence.


PSYCHO-SYSTEMATICS: DEFINITION AND METHOD

The distinction always made here between tongue and discourse permits a much better view of linguistic facts than the view generally found in traditional grammar. Tongue is something that has been established in the mind, providing the starting point and the means for speaking. The act of speaking, as it gets under way, and even while it is being prepared, encounters tongue already pre-constructed in the mind. In its totality, tongue is a vast construct, built according to a general law: the coherence of the parts within

FROM PROBLEMATICS TO SYSTEMATICS 51

the whole. And, as experience tells us, this extensive, coherently-built construct (a system, thanks to its coherence), can be divided into several partial, interiorly coherent constructs, which are integrated systems within the overall integrating system. Like any system, these partial, integrated systems integrate their own constituent parts; they have their own unity, which makes each one a distinctly analyzable whole. Thus, in French the system of the noun is a whole, and likewise the system of the verb.

Experience tells us that tongue is a system of systems. The task of the linguist interested in tongue, the overall, constructed system, is a very special one. It is to reconstruct the great systems of tongue as they exist in the depths of the mind, before being called into use to provide one or the other of the forms they contain.

This reconstruction of the systems that together make up tongue is a special new branch of the science of language which I call psycho-systematics; it has an appropriate technique, in the process of being perfected, called positional linguistics. Essentially, this technique consists of regarding each linguistic phenomenon from the basic point of view of its longitudinal development, and analyzing it exactly the way the mind itself does: by cuts which intersect the longitudinal development.

Whatever the question treated, the technique remains the same. The linguistic phenomenon under consideration is represented by a vector which symbolizes its longitudinal development. The analytic grasp of this development is achieved by making interceptive, or, perhaps better, suspensive cross-cuts of the vector which represents the phenomenon as a whole.

(Lecture of January 9, 1948, Series C)

PSYCHO-SYSTEMATICS AND PSYCHO-MECHANICS

One question which has rather fruitlessly preoccupied philosophers is that of the close tie between language and thought. Some have identified thought and language, making one inseparable from the other. But the true situation is quite different. To get it in perspective, one must be aware of the distinction between thought proper, and the power of the mind to apprehend, or grasp, its own activity. These two are distinct in themselves and must not be confused.'

Language is absolutely independent of thought itself, although it tends to become identified with the mind's power to apprehend its own internal activity, whatever it may be. Thought is free, entirely free, unlimited in its free and active development, but the means it has to apprehend itself are systematized, organized and numerically


limited; language provides a faithful image of these means in its underlying structure. In tongue — considered in and for itself — the careful observer discovers the very mechanisms thought uses to prehend itself. These mechanisms belong to a systematics, the study of which constitutes a new branch of linguistics which I call the psycho-systematics of language.

Psycho-systematics does not study the relations between language and thought, but rather the mental mechanisms defined and constructed to enable the mind to apprehend its own activity. These mechanisms are faithfully reproduced in tongue — understandably so because a first requirement for an act of expression is for the mind to have the power to apprehend its own activity. Unless the mind can apprehend thought, no expression is possible.

All things taken into account, tongue is seen to be the set of means the mind has systematized and established in itself to ensure that it will always be able to make clear, rapid (immediate, if possible) apprehension of what is developing in itself, whatever the development or the matter involved. The study of the formal, psycho-systematic part of tongue does not lead to knowledge of thought and how it functions, as has been wrongly supposed, but to something quite different: knowledge of the means the mind has invented throughout the ages to permit almost immediate apprehension of what is taking place in itself.

As we shall see later, the means that the mind possesses for apprehending its own activity — whatever the activity — are of a mechanical nature. What we are confronted with here are psycho-mechanisms, whose constructive principle is their suitability for ready apprehension and also, once this has been made systematic, a higher degree of organization ensuring this suitability.


PARTICULARIZING AND GENERALIZING IN THE CONSTRUCTION OF LANGUAGE

My studies have concentrated on the following sequence of topics: linguistic time — its expression and its representation — the theory of the parts of speech, the general theory of the word, the theory of the noun, the theory of the verb, and auxiliary verbs. Throughout these studies, the guiding thought, either explicit or implicit, has been the distinction between the tendency to potentialize which gives rise to tongue, and the tendency to actualize which gives rise to discourse. This distinction has constantly grown in importance for me.

I am now convinced that the distinction between these two tendencies dominates the whole science of language. While basically


heterogeneous, the two nevertheless have something in common: one of them — aiming at the actual — takes over the result of what the other — aiming at the potential — has managed to produce. But this link is merely pragmatic, for utility's sake; it in no way prevents the potentializing tendency from making use of operations of thought essentially quite different from those the actualizing tendency brings into play.

The operations of thought brought into play by the potentializing tendency are few, yet essential: they are the source of the human mind's potential. The most important of these operations is the double movement of thought toward the singular and toward the universal, that is, in more general terms, a narrowing movement and a widening movement. The evidence indicates that the structures of tongue are all based on this double movement between the limits provided by the singular (narrow) and the universal (wide). Indeed, this double movement is found everywhere beneath appearances, which barely conceal it in most cases. Thus, it is found to be fundamental and yet quite discernible in the category of number and in the category of the article. It can also be discerned, even if it is concealed to a certain extent by the closely related operation of spatializing time, throughout the verbo-temporal system, where the mind constantly goes from time as a wide stretch to time as a narrow stretch — the narrow stretch of time made up of the present — and also from the narrow stretch of the present to time as a wide stretch, in the past and future time-spheres. In other words, there is alternation between time as a finitude and time as an infinitude. The same double movement is at the root of the entire system of the word; the theory of the parts of speech, related to the theory of the word, also bears its mark. In fact, underlying the distinction between noun and verb (which we have been studying in the second Thursday lecture) is the basic distinction between the space-universe and the time-universe, and so it too can be seen to originate from a success-ivity where mind passes from initial infinitude to finitude, and from finitude to a final infinitude; the initial infinitude is spatial, the final infinitude temporal.

The separation of time and space is not, as one might be tempted to think, a distinction in general grammar but in particular grammar. Actually, there are numerous languages where time and space are not separated. The existence of the article is also a fact of particular grammar, since there are languages which have no article. Likewise, the representing of time remains a fact of particular grammar because it differs from one language to another. And so, little by little, everything one might imagine as belonging to general grammar fails to qualify as such and comes down to particular grammar. What, then is left in general grammar after so much has been eliminated? Probably not a great deal, but what does remain is of prime importance. The two most important facts remaining are the following:


1. that language is made up of a potentializing tendency whose outcome is tongue, and an actualizing tendency whose outcome is discourse, and

2. that to construct tongue the potentializing tendency brings into play operations of its own order, potentializing operations, that is, those primordial operations that are the source of the human mind's potential.

The operations of thought that are the source of the mind's potential are also those it uses to construct tongue, since constructing tongue is the whole point of the potentializing tendency. Among the potentializing operations from which tongue derives its structure, prime importance must be given to the alternating suc-cessivity in the mind of the generalizing movement (carrying the mind toward the universal, as opposed to the singular) and the particularizing movement (carrying it toward the singular, as opposed to the universal). More generally, it might be stated in canonical form (as a mathematician might say): toward the wider, as opposed to the narrower, or toward the narrower, as opposed to the wider. To make it even more general, we could invoke the principle (belonging to the highest sphere of general grammar) that, when constructing tongue, the mind carries out its constructive action between limits (provided by the mind itself, according to the problem to be solved) and gives itself freedom to move in both directions between these limits.


THE LAW OF NON-RECURRENCE

At this point another principle should be brought in: for a relationship (A/B) to become established in tongue through a mental act, it must meet the requirement of being seen as a whole. To meet this condition of wholeness, there must be a movement in the two directions in an "additive" way which does not bring one back to the starting point. This is the very important law of non-recurrence. One should avoid a representation such as this:

A ► B = 1/2 relationship


As well as one like this:

This sort of representation brings us right back to where we started — to the starting position — and the relationship is cancelled out

It would be better to adopt an additive successivity:

With no return, no turning back, one stretch, AB, is added to the other, BA.

Notice that this way of representing things is temporal, not spatial. For example, consider the journey from Paris to Versailles and back: in space, the return trip brings me back to Paris, my starting position. The position on returning is the same as the one we started from. But in time, since the journey took n minutes, the position on returning is not the same as the starting position. The two are different. The position is new and non-renewable. In space, returning is possible; in time, returning is impossible.

This temporal way of representing things is what makes the structure of languages. The necessary limits of extension being the singular and the universal, any movement begins at one of these two limits and proceeds consecutively in each of the two directions without returning, like this:

(Lecture of January 4, 1952, Series B)


SYSTEM AND DIACHRONY OF SYSTEMS

I have just used the word "system" several times. The word does not find equal favour with all linguists; in fact, it divides them into two camps. It enjoys a favorable reputation with an increasing number who admit that linguistics can some day become a theoretical science, through keen observation of facts, constantly made keener and more penetrating by deep reflection. On the other hand, it is received coldly and with considerable reserve by those fellow linguists who fear that linguistics will fall back into the errors of the old general grammar. This fear is not at all justified today; it only leads them to deny the science of language the right to do more than merely describe and classify, as it has been doing since the discovery of Sanskrit late in the 18th century. After that very important date in linguistic history, general grammar was superseded by historical linguistics and comparative grammar and the experimental method replaced the deductive method.

I should add here that this was a most fortunate change. By making linguistics a geographical and historical science based on comparison, linguists chose the only road that could lead to success. The only mistake they made after that was to become excessively positivist, to be overly suspicious of abstract reflection, and consequently to misjudge how much the timely intervention of reflection contributes in power and insight to the observation of concrete reality.

As practised for over a century now, traditional historical and comparative linguistics has been exclusively concentrated on the "facts", having lost confidence in reasoning, which was considered to have led it astray. But the rather curious and unexpected consequence of this mistrust is that it has blinded linguists to many facts, including the most important ones, as we shall see in our studies here. The facts they have focused on have always been the most visible ones, those falling within the scope of direct observation and requiring little effort of thought. However there are others that are not less operative or less real — in fact they are even more operative and more real — but, not being immediately visible, they cannot be reached through direct observation. Traditional historical and comparative linguistics, which for some time now has been oblivious to these more or less hidden facts, is just beginning to discover them, and my own work has brought some of them to light.

Gone now is the period of excessive positivism that made linguistics a science which, had it gone much further in that direction, could have become nothing more than a catalogue of facts whose overall mechanism and various laws of assemblage would never have been fathomed. Many linguists are now undertaking research to uncover the most hidden facts of language history, and remarkably enough

FROM PROBLEMATICS TO SYSTEMATICS 5 7

all of these are, in the first instance, general facts of thought, inherent to human thought. Although languages as a whole may at first sight appear to be based on more particular and more concrete facts, in reality the particular facts always depend on and are strictly subject to the general ones, which are therefore dominant or sovereign facts. Yet these are the very ones linguists have been oblivious to.

No matter what scientific field a theory belongs to, it is never anything but a relation ship established between a general fact (or the smallest possible number of general facts) and the particular facts subordinated to it. It is an affirmation of the sovereignty of one fact over others. Once it is glimpsed, this relationship can be discerned more and more clearly. Indeed, after more reflection and closer scrutiny, the relationship may well appear to be different from what it was first thought to be. The theory can then be revised.

Theories are unstable, since the task of discovering the subordination of the particular to the general never ends, nor can it end. A general fact, which at first is seen to have a host of related facts subject to it, can suddenly be perceived as itself a particular fact and subject to some more general facts, unnoticed up to then. The instability of theories should not be considered a liability. It is simply a clear mark of their ability to progress, benefitting continually from new facts furnished either by observation or abstract reflection. Whether mathematical in form or not, reflection can furnish facts which, though all too often neglected in linguistics, are just as valuable as facts furnished by observation.

Highest in the hierarchy are those sciences which have succeeded in combining careful observation of concrete facts and abstract speculation in just the right proportions. Linguistics may well have greater difficulty defining its theories than certain other sciences of observation, because between particular, visible, subordinated facts of the physical universe and general, secret, subordinating facts there is a stable relation, whether this has been discovered or not, whereas in language this relation is perpetually being renewed.

Hence it is by no means easy for a linguistic theory to become completely general. It is liable to represent no more than a passing stage of the relation between the particular, contingent facts of a language and the general, sovereign facts subordinating the others, because although the relation as such is permanently established, it is constantly being diversified. This constant changing of the real and objective relationship in tongue itself (which is, in fact, none other than the system of tongue) between particular governed facts and the general governing facts, makes it difficult enough to develop a theory; but if to this one adds the problem that


the relationship may well have been inexactly observed, then one can begin to appreciate the difficulties involved in developing a general theory of language.

When a physicist confronts the universe, an observer reasonably capable of accurate observation confronts a stable system, universal and unvarying. The only mishap that might befall the physicist is to misunderstand the system he is observing.

In linguistics it is not at all the same: the intrinsic varying of the system makes accurate observation more difficult. The difficulty becomes quite appreciable when a long stretch of time is being considered, — and a theoretician aiming at the general cannot refrain from doing so — because the object to be observed is no longer one system but a systematic succession of systems, one following the other, one replacing the other in time. The system of the universe studied by the physicist is a constant, objective entity, whether his knowledge of it is clear or not.

Since the system of tongue which the linguist tries to observe is objectively a changing and inconstant entity, it follows that linguistic theory should be historical in form. The theoretician in linguistics cannot help being a historian. (As compared to the historian of the visible (the contingent, to some extent), he is above all the historian of the invisible (the necessary).) The very nature of his object makes the theoretician in effect a historian of the most general hidden facts. The historian, who is not a theoretician, perceives only the most visible facts, those within the scope of simple, direct observation, that is, only the particular and accidental ones.

One of the tasks of linguistics today is to identify the different systems in each language — the component parts of each system of systems — and to establish a filiation, which in the most favorable circumstances will be historical, between the systems identified. When, however, the historical filiation cannot be determined for lack of documentary evidence, it will be an obligatory, necessary filiation based on extra-social, simply human factors.

Something that should always be kept in mind when studying the systems of tongue in any language is the fact that while the inner state of a language is of course derived from what has been transmitted, inherited, from.the past, the mind is always organizing this inheritance from moment to moment according to its own necessities — call them laws if you like. And these laws become more explicit through dealing with this inheritance.

From instant to instant, tongue is both an inheritance from the past and a human organization that transcends the inheritance received. A moment's thought reveals that this is none other than the well-known distinction between diachrony and synchrony introduced by Saussure in the Cours de linguistique générale. I use


different words here in order to deepen our understanding of things. Diachronic linguistics grasps things longitudinally, in time, which makes them change, perturbs them, disorganizes them, and would destroy them if some contrary, organizing force did not intervene. Synchronic linguistics grasps them by cutting across, not longitudinally along their disorganizing movement, but latitudinally with the opposed organizing or systematizing movement which then determines the relationship between things by making it subject to the most deep-seated laws of human thought.

For Saussure, the two dominating images are those of time, which flows, and the instant, which stops and immobilizes. The following is a schema of his view:

This view is penetrating but rather summary. The systematizing Saussure attributes to each immobilized instant in the longitudinal progression of time is not in fact instantaneous: it took, takes and (since it changes) will take time, just as does the reverse process of disorganization, which prompts the systematization.

In language, systematic organization works on the disorganization that a language inherits from instant to instant. Actually, two opposing forces are involved here, the one descending and disorganizing, the other ascending and organizing. And the two meet. An organized system results when the descending force of disorganization is stemmed by the ascending organizational force. A suitable image to represent the interaction between disorganizing diachrony and organizing diachrony would be as follows:


Organization — the system — appears at the top of the upward movement, where there is relative (astatic) equilibrium between the two impulses.

This figure and the complex interplay it brings out provide considerable food for thought. Of the two impulses involved, the disorganizational descending one and the organizational ascending one, the first constantly seems to have the advantage. Indeed, acquired states of tongue are constantly being abandoned. But in the last analysis this advantage granted to the disorganizing impulse is illusory because tongue does not abandon an acquired state unless it be to respond once again, and more fully, to the ascending organizing power. The later in the history of the language this organizing power meets the descending, disorganizing power, the more operative it is. It stands to reason that the organizational power should have little to do when the disorganizational power has done little itself. And history substantiates this observation. Great linguistic systematizations and moments of great perturbation coincide. This also explains why the systematization of evolved languages surpasses in certain analytical qualities the systematization of less advanced, more ancient or more conservative languages.

It is unfortunate that a man of genius like Saussure was not tempted to examine the idea that each synchronic system derives from a prior system which, while different, was just as rigorous. Failure to do so leaves his view of the interaction of forces involved in the formation of language incomplete. Had he paused over this idea, which is substantiated by data from observation, he would have realized that there is not just one diachrony, one concrete history of linguistic facts considered in isolation without the system of which they are integral parts; there are two. The second is a diachrony — a history — of the systems: a diachrony of synchronies. For, and here I must diverge from Saussure's thought somewhat, we are never really confronted with a system. The instant a system becomes established, it has virtually begun to remake itself. However, in most cases the remaking is so barely incipient that the system can be fixed, pinpointed in time, and described as if it were a stable entity.

As I see it, this diachrony of synchronic states should constitute the general framework for the history of language. The view historical grammars give of it is incomplete, in fact for the most part practically non-existent, since they fail to mention or even discern the history of tongue as an entity, an admittedly abstract but nonetheless real entity constituted by the system existing at any instant.

At any given instant, the particular facts of a language form a system. These facts change. They can change without the system varying. Numerous phonetic and semantic changes can occur within a system without modifying its mechanism in the least: languages


have been known to change vocabulary without changing their systems. One might speak French with English words, as is sometimes done in certain snobbish circles, and the change would have little or no effect. The most important and interesting changes are those that trigger or accompany a systematic change. In such cases it is important, while keeping in mind the history of the particular concrete facts, to discern the abstract history of the system itself, that is, not the history of its material components but of the relation established between them. In other words: the history not of the things linked, but of their linkage or coherence.

This history has no definite or even appreciable place in traditional historical linguistics, which consequently could be criticized, not for being too historical (this was Saussure's view), but for not being historical enough.

To be complete, a historical grammar should include numerous chapters on the history of grammatical systems instead of relating little else but the individual histories of the forms composing these systems. Moreover, the history of the forms composing the systems and the history of the systems themselves as abstract entities (showing how the existing forms cohere) throw light on one another. And the true language historian, as I said before, would be the one who could see both the concrete changes affecting the forms, and the abstract, less visible changes affecting the system, composed as it is of forms assembled according to the permanent laws of human thought.

I wish I could represent the history of language by means of a diagram showing isolated forms and systems of forms, and clearly portraying their development. This is not easy to do. Conventions would have to be laid down to cover all the kinds of changes that can affect the constituant forms of any system. Because of the difficulties involved in constructing representative diagrams, I have to be satisfied with describing in the ordinary way, using words.

On the one hand, a system can remain stable, with forms that change, just affecting the system externally. This often happens. As long as this situation goes on, the system has no history, only its component forms do. On the other hand, a system can vary without the forms changing outwardly. They may simply change positions within the system. Almost every time a form changes position within a system, there is a systematic change of some importance. Thus, when Spanish inherited verb forms ending in -ra from Latin, they came to occupy a very different systemic position from the one they had held before. What is important in this case is the change of position, and to grasp it clearly it is necessary to discern how the transformation of the system came about that is, the systematic diachrony must be traced. In simpler terms, the history of that abstract, not directly observable entity known as the system must


be traced. It is also possible, though rare, for one form in the system to take the position of another which, in turn, takes exactly the position of the one that replaced it. In this case, the abstract system has not changed. The change, limited to the level of concrete expression, is actually nil.

I do not have time today to outline all the possible cases and those which can result from interference between different cases. The main idea to be remembered is that in tongue there can occur changes with no effect on the system, indifferent to the system, and changes that do alter the system. These two types of changes must neither be confused, nor, in a good historical grammar, treated in the same place. Long experience with the most difficult of linguistic questions has shown me that the way things change in language cannot be represented accurately if, beneath the apparent changes that initially attract attention, the underlying changes in the state of the system of tongue are not discerned. And so, to understand how the verb changed between Latin and French, it is necessary to outline the verb system of Latin clearly and then contrast it with the verb system of French. Then the transformation of the system can be seen, and, with the help of the more visible history of the component forms, its history can be traced.

(Lecture of November 11, 1943, Series A)

THE CONTINUING CAUSATION OF LANGUAGE: PROPOSALS AND TRANSFORMS

The causation of language is continuous, a constantly operating causative mechanism of systematization. At the outset it gives rise to an interiorly systematized construct, which I shall call Constructed Causatum1. The causative mechanism continues to operate within this construct, eventually modifying it and so giving rise to Constructed Causatum2. The Continuing operation of the causative systematizing mechanism within Constructed Causatum2 brings about a development of it in turn and gives rise to Constructed Causatum3. Within it the causative mechanism keeps on operating and eventually Constructed Causatum4 appears. It will be treated the same way as the constructed causata produced before it. Thus causation continues from constructed causatum to constructed causatum indefinitely — indefinitely here meaning as long as there remains a possibility of producing in this way a constructed causatum which is different from the constructed causatum before it.

The possibility of perpetually renewed differentiation does not increase, but decreases: the differentiating slows down. There is an asymptotic approach to a limit where the possibility of differentiating would be followed by the impossibility of differentiating.

FROM PROBLEMATICS TO 'SYSTEMATICS 63

In other words, the differentiating of the constructed causata slows down to the point where it borders on, without actually reaching, the limit of extinction. During the approach to this limit, the possibility of differentiation is assured, but the probability of differentiation becomes remote.

The differentiation through successive constructed causata now operant in French is slow: it is a movement of a form which entails slowing down the differentiating considerably. The possibility is still there but the probability of differentiation is becoming remote. This possible/probable relationship can be seen in the morphology of Modern French. In diagram form, glossogeny thus seems to be something like the following, where C stands for the initial causative mechanism, Cc1, Cc2 and Cc3 the successive constructed causata, and the Greek letter A the differentiating treatment:

In Cc2, there is a differentiating treatment A of Cc1, in Cc3 a differentiating treatment A of Cc2 and so on... (no need to go on.)

The filiation of successive constructed causata thus turns out to be, from moment to moment, the change of a previous proposal into a subsequent t r a n s f o r m . This produces a glossogenetic succession, whose general progression is as follows, where p stands for what it proposed and ρ' for the resulting transform:

[ ρ ] ► [ ρ' ]

and applied to French:

[ Ρ ] [ Ρ' ] Old French Modern French

(proposal of the (transform of the transform ρ') proposal ρ)

1. Glossogeny: the historical development of the word (the vocable) as a form of thought.


Thus, a linguist observing French and the development of language in general can adopt two possible points of view: a non-retrospective one where only the transform is observed, and a retrospective one with observation moving from the transform back to the proposal. The first gives a view of the present state of the French language, and the second a comparative view of this state and the former state found in Old French, that is, a comparative view of the transform ρ' and the proposal ρ.

Old French is the proposal and present-day French its transform. Likewise, present-day French can, through anticipation, be considered as the proposal for a transform yet to arise but not yet arisen as such. The shape of the phenomenon, in diagram form, is:

The cornerstone of the Cours de linguistique générale is Saussur e ^ distinction between synchrony and diachrony. The distinction made by means of these two Greek nouns is exactly the same as the distinction between a study of language restricted to the transform, and a broader study of language that includes comparison with the proposal. To limit study to the transform is to choose synchrony. To extend the study to the proposal is to choose diachrony.

Old French is the proposal of Modern French, which, in turn, is the transform of this proposal.


PROBLEMS OF REPRESENTATION AND STATES OF TONGUE

In linguistics, it is always the acquired solution that confronts the mind with the problem to be solved. A linguistic solution, that is, any state of tongue, has the double effect of solving problems of representation confronting the mind, and of presenting the "solved" problems to the mind again in the form of the solutions just worked out. Problems of representation are never definitively solved: in linguistics the solution is not just a solution, but also a restatement of the problem already solved.


This explains why linguistic development is limitless. A solution, however elegant and however successful, raises the problem anew, and calls for a fresh attempt to solve it. The process goes on like this ad infinition.

(Lecture of December 5, 1947, Series C)

LANGUAGE AND SILENT THOUGHT

...Following this line of thought, a point I would like to bring to your attention in finishing is that we find registered in tongue not only the needs of thought about to undertake an act of expression, but also what might be called the needs of silent thought, when the mind is occupied, outside the actual act of language, in scanning itself, searching for the best means of apprehending what is taking place there. It is wrong to link the construction of tongue too closely to what happens during the act of language. The truth, I believe, (and this has become a principle of my teaching) is that tongue is of course created in us through use, but partly also outside of use, during that deep, ceaseless musing that goes on endlessly in the minds of thinking men, (which actually includes all men, or at least the great majority). In my opinion the most profound part of tongue stems more from the deep musing of the human mind than from the direct exercise of the act of language which, in a multitude of cases, calls into play things already discovered by the mind outside the act of language, at the level where thought recollects itself. The metaphysical part of languages, so visible in our study of the great linguistic systems, is largely the expression of the ceaseless working of thought that extends far beyond the relatively short moments when the power of speaking with which we have been endowed is actually being put into practice.


PART III

SIGN AND SIGNIFICATE

THE PHYSICAL AND THE MENTAL: A FUNDAMENTAL DUALITY

We can explain to the extent that we have understood. We can understand to the extent that we have observed. In other words, we can explain only to the extent that we have observed. To be complete, observation in the science of language must take in both what is physically (and so immediately) visible as well as what is mentally (and so non-physically) visible beneath the physically visible part of language.

Of all the relations and correlations in language and in the science which observes it, the most important, and most neglected (the least taken into account), is that between the physical and the mental.

If I may put it this way, language "physifies" what is mental. In language, the mental calls on the physical to make it perceptible, through sight or hearing. It appeals to sensory means whose limited role is to provide a "physified" representation of what is mental. This representation will never be a completely faithful image of the mental original it tries to render. Throughout its long structural and architectural history, human language, in its continuing causation, has been a never-ending search for the optimal rendering.

I have just said optimal. This term should be taken to mean the best that can be achieved considering the state of development of the language at a given age in a given civilization. An optimum beyond which it would be impossible to find something better is inconceivable. It fades away into the distance, until it reaches its limit at a moment analysts have perhaps not paid enough attention to, when a better rendering, while still possible, is not at all probable — in fact is improbable to the nth degree.



SIGNIFYING AND THE MENTAL /PHYSICAL RELATION

My observation of the synapses in declensions is based on the principle that the cases involve fleeting impressions linked to signs recognized as suitable to represent them. These impressions have strong accretive powers enabling them to group together mentally, to cluster under signs considered apt to signify their mental combination. I tend to regard the formation of tongue not so much as a sort of linear, etymological derivation, but rather as a continuing accretion of fleeting impressions linked to formative elements which agglutinate to provide the physical counterpart of the mental accretion. On the mental side, I see the words of a language as a set of more or less fleeting impressions with accretive powers causing them to fuse together, and on the physical side as a set of corresponding formative elements to which the impressions are attached.

The line followed here for the construction of the vocable is not the one followed by the science of etymology. Etymology does not depict the historical genesis of words as a line where groupings take place, where there are successive accretions of fleeting mental impressions attached to formative elements, their physical counterparts. It depicts it rather as a line along which words seem to arise one from the other, simply by derivation. This leads to difficulties when it comes to explaining the mental filiation of senses, and the physical filiation of the phonetic changes. Both of these filiations remain conjectural more often than is mentioned.

In traditional comparative grammar little more than the physical side of vocables has been observed. No attention is paid to the relation — one of prime importance in the structure and the architecture of human language — between the physical and the mental. Change in this unobserved relation has not been seen, for lack of sufficient insight to discern it.

A little reflection on the subject is enough to suggest that it would be possible in a comparative grammar to observe the changed physical appearances of language and, at the same time, the changed relation between the physical appearances and the mental aspect underlying them. This is the comparative grammar of the future, for those linguists bold enough to blaze new trails.

Strangely enough, however, present efforts are all tending away from this sort of comparative grammar, or so it seems. Although signifying is the necessary attribute of language, every effort is made to consider only what is physical in language; yet this physical part is insignificant if it does not overlie something mental, the prerequisite for signifying. For the study of language, the only suitable point of view is one that sees the dyadic relation throughout language, and is able to trace its distribution among partial physical/mental dyads. (In French the most important of these, the most general, are the parts of speech.) The structural

SIGN AND SIGNIFICATE 71

and architectural history of language is, for any given moment, an account of the distribution of the dyadic relation — sine qua non of the existence of any language — among partial dyads, each of which forms a system within the overall system, tongue.

Tongue — any tongue — thus seems to be a system that meets a requirement of oneness and contains a certain number of partial systems. One characteristic of the overall containing system is that it is not contained in any wider system, nor does it have an exterior.

Tongue also appears to be made up of the sub-systems contained in each partial system. In French, there is the system of the noun and the system of the verb; within the noun there are the substantive and adjective sub-systems, and within the verb the sub-systems of the impersonal constructions (para-nominal) and the personal constructions. For one's view of any given tongue to be complete, one must work one's way from the systematic whole inward to the systems and sub-systems that make up the whole and understand their mental filiation within it as revealed in the corresponding physical filiations. In every observed language the task of the linguist is to work his way in to the systems and sub-systems without ever losing sight of the common thread of movements and forms of movement1 running through them all.

{Lecture of January 22, 1959)

INTERIORIZING AND EXTERIORIZING IN LANGUAGE

Up to now, linguists have been overinclined to observe the physical means of exteriorizing language, and insufficiently inclined to observe the non-physical, mental means of interiorizing — two orders of means which progress equally in opposite directions. This has been a mistake.

The symphysis of the two means, never disturbed in the normal speaker but variously disturbed in the pathological speaker, has never received the explicit attention of linguists. It is a mechanism which, as a mechanism, has been ignored. No linguist has said to himself: corresponding to the language I see being exteriorized under physical signs, there is always a counterpart, a language interiorized by mental means.

1. For an example of one form of movement, see below, "A Structural Device: The Radical Binary Tensor".


In other words, the error of traditional linguistics has been its failure to see (with the mind's eye, the eye of understanding) that any interiorizing step calls for an equal exteriorizing step, and that what we know of the means of exteriorizing language thereby helps us to see what the mental, interiorizing means have been. Thanks to this symphysis, a true science of language can be built.

The constant aim of these lectures is to look for and discover the means of interiorization in language that are always found beneath its correlative means of exteriorization. To one who knows how to question them, insisting until they yield an answer, the means of exteriorization are like a mirror. Let me make this clear: in French, when I encounter a sign (the means of exteriorizing) such as that of the article, I must ask myself what interiorization this exteriorizing agent stands for.

The interiorizing process is exclusively mental; the exteriorizing process, on the other hand, is a mutation from mental to physical. The same terms could be used to describe any linguistic fact which is observable and therefore provided with means of exteriorization.


THE SUITABILITY OF THE SIGN1 TO THE SIGNIFICATE

The relation between sign and signifícate: a vast subject. It is a relation of suitability. A law: the suitability will never be excessive. The two types of suitability: a) material

b) formal. There would be material suitability if the sign by itself — as a result of its materiality — conveyed a significate. No need for linguistic convention. The constructing of language has not developed in this direction.

The suitability of sign to signifícate is not based on the former's materiality but on conditions of adaptation of a very different nature.

The problem in psycho-semiology: an idea cannot invent a suitable sign for itself, but can look in the already existing semiology for a sign that might be transferred to it; because the sign has not been made specifically for the idea, it becomes suitable to it only

1. In order to avoid confusion, here and elsewhere we have translated as if Guillaume had always used the terminology adopted during his latter years, i.e., =


by losing its former suitability. This is the way things develop. This is why the linguistic sign is arbitrary. Inventing it

entails a loss of suitability: the new suitability is based on this loss.


EXPRESSIVE SUFFICIENCY: THE LAW OF PSYCHO-SEMIOLOGY

In the semiological system freedom prevails. Anything is good enough, anything is suitable if it manages to signify sufficiently. The only law in constructing semiology is bare sufficiency. There is the greatest possible freedom in choosing the means, as long as they are sufficient. This explains the diversity, for example, of the verb conjugations and the large number of verbs whose conjugations are apparently irregular. This diversity and these apparent anomalies are consequences of the freedom prevailing in semiology — freedom because success in "signifying enough" is all that is required of semiological means, whatever they may be.

In the field of semiology, strange and unforeseen combinations for evoking meaning must be expected. Semiology always presents us with a great wealth of combinative inventions. A bird's eye view of semiology, seen in historical perspective, reveals a continual search for more and more extensively applicable means of signifying. Those with a narrower range are replaced by more extensive ones that can be used over a wider linguistic field.

This substitution does not prevent less extensive semiological elements surviving along with the more extensive ones. The survival of elements whose extensity has been deemed insufficient (and which therefore have lost their propagative power in tongue) is represented by what are called strong forms. And weak forms are those which still retain a propagative power in tongue.

On the mental side things are very different because we are no longer confronted with signs which theoretically can be as diverse or heterogeneous as desired so long as they signify sufficiently, but rather with the signifícate which, as much as possible in the formal part of tongue, must have a fundamental unity. It must have acquired systematic coherence.

Systematic unity is acquired in the mental field long before the corresponding unity is achieved in the semiological field. Thus for French, in the category of the verb, the verbo-temporal system for representing time has acquired a unity, on the mental side, which can be considered perfect. Mentally there are not several systems in the field of the verb, each claiming a certain part of the verb


category, as is found in less evolved languages, but one single system covering the whole of the verb category. Nothing is left outside under another system. In the semiological field, however, nowhere near the same unity has been achieved, even though the tendency toward increasing systematic coherence is strong. And so we find diverse semiologies operating suppletively, one beside the other, in the field of the verb. Hence the multiplicity of conjugations, and the numerous irregularities above and beyond this multiplicity.

What separates semiological systematization from mental systema-tization can be understood by considering that in mental systematization it is the law of greatest possible coherence that prevails, whereas in semiological systematization it is the law of the best possible expressive sufficiency. In semiology the need for coherence is secondary: coherence can always be sacrificed without harm as long as expressive sufficiency benefits.

It should be added that we find coherence developing even in semiology. But coherence is not a direct outcome of semiology and its requirements, which are of a different sort. Rather, it arises because semiology, in order to be operative, must pattern itself after the mental side, reproducing it sufficiently. It thereby tends to reflect up to a certain point the coherence it finds established there.


RECIPROCAL ACCOMMODATION BETWEEN THE MENTAL AND THE PHYSICAL

A linguist enjoys tracing the details of the small, systematically constructed facts of tongue because in a striking way they illustrate the general law of congruence constantly maintained between the physical and mental parts of language structure. To be sure, the whole organization or lay-out of the French verb system is based on physical, phonic variations that can be directly explained from within the field of phonetics. But on closer inspection, we find that these phonetic variations have always been accompanied by corresponding mental phenomena whose essence, whose fundamental nature, is to reflect the more or less contingent accommodation between the mental and the physical.

This accommodation is more or less reciprocal: in numerous cases, accident permits the phonetic construction to adapt itself almost directly to the mental construction. So it is in the relationship between the imperfect and the conditional. In other cases, however, when things remain more dependent on the historical basis — when things are continued rather than remade — a very fine, subtle accommodation of the mental construction to the physical (phonic)


construction can be noted. Marchèrent is a case in point. Without giving up anything essential, the mental formation of this form has managed to adjust to phonetic variations foreign to it, so that finally, through channels that happened to be open to it, the deep-lying symmetry of the future and the aorist1 was successfully marked, even in this case where it seemed it could not be made visible. Only in the obvious cases of the double person2 is the symmetry of the French aorist and future left unmarked. This fact can of course be explained quite well phonetically; but here again, through analysis we can discern the process of accommodation, where the mental part adapts to the physical. In French and in any other language the double person is something on its own, with its own particular conditions of existence. The norms governing the simple person cannot be applied to it.

(Lecture of April 16, 1942, Series A)

NEGATIVE MORPHOLOGY (ARTICLE ZERO)

The day the history of article zero in French is written, it will be seen to be primarily an article of negative morphology, a critical one, if I may say so. It is an article which spots the expressive deficiencies of the other articles in particular cases and replaces them, not so much in order to say what it has to say (which is nothing, since it is an absence) as to avoid saying what the other articles would say inappropriately in that instance.

I have always had the impression that not enough attention has been paid to the existence of a negative morphology — and syntax for that matter — based not so much on the mind's attraction to a form as on its aversion to the other forms. It has even been suggested to me by some fairly extensive studies that there may well be languages where the morphology is more negative than positive, the expressive sense of a form consisting not so much in what it signifies as in what it intends not to signify. Such languages would be the most difficult to analyze from the point of view of morphological nuance.

I have been getting quite a persistent impression of mainly negative morphology during my study of the verb in the Semitic languages. In these languages it is very rarely a matter of positively indicating, say, a tense. They have nothing that could really be called tenses, that is, which would correspond to tense in our languages. Instead, in these languages it is always a matter of not

1. Guillaume considered the French past definite an aoristic form. 2. That is, with nous and vous, which involve more than one person.


indicating a temporal situation that would clash with the particular idea one intends to express. It is by means of avoiding the unsuitable at every turn that the suitable finally comes to the fore : suitability as a result of avoiding the unsuitable.

Article zero is of interest not only for what I have shown from the systematic point of view, which in itself would be ample to make it worthy of the closest attention as an entity in tongue. But it is also of interest when it occurs in a modern language whose nuances we are sensitive to, because it offers a golden opportunity to follow the twisting and turning path of this negative, essentially critical morphology to whose existence I have just been drawing attention. Negative morphology is found in every language, with, it seems, certain languages particularly favoring its use.


PART IV

THE ACT OF LANGUAGE

THE NATURE OF THE ACT OF LANGUAGE

The act of language does not begin just at the moment when words are uttered to express thought. It starts with an earlier, underlying operation, when thought, seeking expression, resorts to tongue, a permanent resource of the mind. The fact that tongue is permanent frees us from having to imagine makeshift means of expression every time we need to express something. It is important to remember this as being the principal function of any language. But is is extremely difficult, impossible even, for us to understand fully the secret of the mechanism whereby thought resorts to tongue, and finds an immediate response. (The rapidity of this response cannot help but astonish anyone who takes the trouble to reflect on it.) If we could but see what was taking place in our minds in the instants immediately preceding an act of expression — in what could be called the preparatory phase or phase of tension, to be followed by the instant of release and the phase of impulsion, to use Bergson's terms for very general analysis — if our mind could discern its internal activities at these great depths, we would be in possession of extremely important linguistic information which is now denied us. And a multitude of unsolved or highly controversial problems would then be as easy to explain as it is for us to express thought by means of speech.

Access — at least direct access — to the mental operations preceding the release of the act of language is denied us, and thus the most essential part of the act eludes direct investigation. Only the very last instants of the act of language can be observed. The first instants, when contact is established between thought seeking expression and tongue, the mind's permanent resource, are instants that defy direct observation. Any knowledge of them must be acquired through analytical interpretation of later instants, an interpretation which must take into account both what occurs in discourse and also what has become permanent in tongue in the form of lexemes, morphemes and systems.


Lexemes and morphemes are elements of tongue that can be observed by traditional linguistics: they are, in a sense, corporal, represent ed in tongue by signs whose role is to carry them into discourse when the need arises. The act of language — a rapid and efficient act in our more developed languages where many of its cumbersome elements have been dispensed with — consists largely in conveying from tongue to discourse the lexemes and morphemes required for thought to be expressed. For lexemes and morphemes to be thus conveyed they need a sign in tongue, that is, a representation of sound linked to what they mean, to the mental entity that constitutes their signifícate.

The general function of morphemes is, through diffusion, to subsume whole series of lexemes. The series thus covered vary in extent. A fact ingnored by traditional linguistics is that morphemes are an integral part of systems. Before it can choose the appropriate morpheme, the thinking mind seeking expression must "see" the morpheme within its system, the system which lends the morpheme its meaning according to its position therein. This is why, to use a single verb form, the mind must rapidly evoke the system of verb conjugation as a whole, an operation so rapid and deep-seated that we are totally unaware of it. As the system of the conjugation is rapidly evoked, all the forms it contains and integrates are offered as equal possibilities — all the forms, that is, which were assembled by the mind in a previous, defining act that brought the system into existence. From among these forms presented all together within the system, one beside the other, as it were, and also one beneath the other if the system is tridimensional, the mind, preparing to express something, chooses the one form that at the moment is judged to express most aptly and faithfully what it intends to express, leaving the others aside.

The system presents all its forms together in such a way as to allow a total, synoptic view of them at a single glance. Like any other operation of language, that of choosing one form from among those presented requires time — concrete, experiential time which, while very short, is nevertheless real. And this time can be greatly shortened by the quality of the system the forms belong to. As a rule, the speed with which the mind chooses the form most suitable for discourse depends on the stage of development of the system and on the positions that the various forms occupy within it — positions that determine the essential meaning of each form. A system whose positions are elegantly arranged, following each other in one constructive movement makes it easier to discover the form required for the act of expressing discourse. The process becomes faster and surer.

1. Guillaume uses "lexeme" in the sense of a unit of lexical meaning, and "morpheme" in the sense of a unit of grammatical meaning.

THE ACT OF LANGUAGE 81

If a system were badly constructed or insufficiently developed, or did not, as a whole, constitute a single act that integrated the component forms, the search for the form best suited to the momentary aim of discourse would take more time and, in the end, would not be as successful. If a system for integrating the forms in tongue were totally lacking, it would be not only difficult but in fact impossible to find the most suitable form, for it is within the system and only within the system that the forms have a meaning of their own, predestining them for whichever particular contextual sense is required in discourse. It would be a mistake — one that some very good linguists have often come close to making and may even have made — to postulate that forms have no meaning in themselves but acquire their meaning only when used in discourse. Such reasoning would imply that that which has no meaning in itself could, when applied in a context, mean something. This does not make sense: how could the mind possibly apply the meaning of something which has no meaning to be applied? Meaning can be applied only if some basic meaning is available. Zero meaning cannot be "applied" to anything.

At the level of the system, it is the contextual meaning of the form that is at zero grade; only in discourse will this contextual meaning be determined. On the other hand, the systemic meaning, which exists prior to any contextual meaning, is at full grade in the system where each form represents a different moment of the system's mental construction, or, to be more precise, in the single, homogeneous act of definition that the system represents. Thus, in the temporal system, each component form, whether mood or tense, represents part of a homogeneous act — the architectural construction of the time image. Each denotes a given moment whose distance from the other moments within this constructional act is determined by the sequence of the forms in the system.

It cannot be too strongly emphasized that, in tongue, the systemic meaning of a form is totally acquired, whereas the total contextual meaning is yet to be acquired. The role of tongue is to permit a fairly wide variety of contextual meanings in discourse by presenting forms in systems. Each systemic meaning makes possible a certain range of contextual meanings in discourse.

Up to now, morphology has focused on the contextual meaning of forms, and to all intents and purposes has neglected their systemic meaning: the systemic point of view has not found favour, both its existence and its importance being ignored by most linguists. The real reason why morphology has thus taken no notice of the fact that before forms take on a contextual meaning in discourse they have a permanent 'systemic meaning in tongue, is that linguistics has remained oblivious to systems. In the hierarchy of form— the hierarchy which starts with the singular and develops toward the universal — linguistics has stopped at the level where morphemes subsume series of lexemes; it has not reached the higher level where systems subsume and integrate morphemes in series of various lengths.


Thus halted in its development at a point where half its possibilities are left unexploited, linguistics has been a rather disappointing field — and criticized as such by some fine minds — because after scoring a brilliant series of successes in pushing back in time and discovering origins, it has nonetheless taught us very little about the true nature of tongue and the essential laws of its formation in the mind.

I am absolutely convinced that linguistics has only to set out resolutely along the path that is open to it in the study of systems in order to cease being an incomplete science which does not tackle questions related to the mind — questions that exacting but enlightened critics will never stop expecting it to elucidate and solve, deeming that they are within its scope and that it fails in its calling by declining to study them.

Observation reveals that all these questions of a higher order which linguists have avoided studying are related to the study of systems. It thus becomes evident that the urgent task in the science of language is now to orient its research along the lines followed by our approach. Before examining the contextual sense of a form, one should first make a point of reconstructing the system of which the form is an integral part and from which it derives its essential meaning — a meaning which, prior to any contextual meaning in discourse, exists in the mind even though we cannot become aware of it directly, since we do not have direct access to these deep-seated operations.

(Lecture of April 27, 1944, Series A)

THE ACT OF LANGUAGE: ITS OPERATIONAL CHRONOLOGY

The act of language, taken as a whole, involves an internal chronology which reflects the main lines of the general system. This chronology can be divided into two phases:

A. an initial phase of potentiality, from the formative elements of the word to the constructed word;

B. a final phase of actuality, from the word to the sentence, that is, from the unit of the potential to the unit of the actual (in French, this would be from the words to the thought expressed).

Whereas the mental operations of the second phase, B, may be consciously observed by the speaker, the more deep-seated operations of phase A cannot. Conscious observation comes into play too late


to discern them, at a moment when for the mind they are already accomplished, completed, left behind. That is to say, compared to words, which are already constructed in the mind of the speaker, the sentence, in the psycho-mechanism of language, is for him something yet to be constructed. In terms of time, words are finished constructs representing a constructing that is past and left behind; sentences are possible constructs yet to be assembled. Mutatis mutandis, we have here the distinction between the already accomplished and the yet-to-be accomplished, that is, between past and future.

What I have just been discussing brings out the complexity of linguistic facts. This complexity, which I have been commenting on by presenting one of its more interesting sides, can be summarized in the following, relatively simple diagram, if we restrict ourselves to the points just dealt with:


One of the errors in grammar as a science has been the failure to distinguish clearly between the two kinds of mental operations: those which reach completion on the potential level with the construction of a unit of potentiality — a word, in our developed languages — and those more easily discernible operations which reach completion later, on the level of the actual, in the sentence, a short-lived, limited piece of discourse whose construction and limitation are governed by conditions in tongue. In other words, these conditions belong to the predetermined mental operations, established permanently on the level of potentiality.


ESTABLISHED IN LANGUAGE OR IMPROVISED? EXPRESSION VS. EXPRESSIVENESS

The established system which language — the act of language — resorts to is tongue. In cases where, for any reason, this established system is lacking, the act of language — insofar as the lack is felt — resorts to means of its own on the spur of the moment, purely expressive means which do not establish notional differences under different signs. To establish notional differences under different signs would in fact amount to constructing something permament: tongue

As long as thought is expressed only through improvised language without having recourse to the established system, there is no need for notional differences to be fixed by distinct signs. Thus, language in its entirety involves recourse to two kinds of resources:

-those which come into play late in the act of language, pertaining to the improvised; -those which come into play early in the act of language, pertaining to the established. The act of language tends to rely less and less on its own means

of improvisation. The more that tongue can furnish already prefabricated, the less there is to be improvised during the act of language. The only thing that remains to be improvised is usage, and not even all of that — only what is singular or ephemeral in it; what is general, systematic in usage is already governed by the established system.

At the very beginning, when there was no established tongue, the act of language had to accomplish everything, to express everything by its own means, which were quite different from those of established language. Expressiveness was then its sole, quite inadequate resource This situation can be compared roughly with that of a man who has to communicate with another man without knowing the other's language, his tongue, and who finds he must resort on the spur of the moment to


improvised language, precarious though it be. But the further language, in its general history, moved away from its original state, the more the established system took over part of the task of the language act and,' in so doing, reduced the load of improvisation.

Improvisation still has its part to play, although it is no longer directly concerned with the content of the act of language. In our languages, the resources of tongue are now available for this purpose. Improvisation is now only required for the various means of expressiveness involved in bringing out this content, and this only insofar as tongue, with all its established resources, is inadequate.

An essential characteristic of discourse in our developed languages is that little is demanded of improvisation in comparison with what is demanded of established language, that is, tongue. Those demands that are met by tongue make up what we call e x p r e s s i o n . Those demands which are not met by tongue, and which the language act must therefore meet, through suppletion, using its own improvised means, make up expressiveness. As I have often pointed out, the act of language taken as a whole can be represented by the following formula:

expression + expressiveness = 1.

It should be recalled that expression involves resorting to what is established, and expressiveness resorting to what is improvised, the resources of the language act itself being of the improvised or non-established type, and those of tongue being of the non-improvised or established type.

One's way of addressing someone through speech generally involves improvisation. There are a thousand ways of saying the simplest things with the same words. One's way of addressing someone, which makes up the expressiveness of the act of language, is added on to the expressed content, which is rendered by the established means of tongue.

In certain languages, expressiveness in the act of language can well bring about considerable variation in the meaning that particularly susceptible vocables can have in discourse. One need.only think of the variety of sense and nuance in some words of several southern languages, particularly Spanish. Expressiveness, generously added to expression, becomes all important. A particularly striking example of this is the Spanish word quiero, whose variety of senses defies limitation.

The equation:

expression + expressiveness = 1


is very important for a clear understanding of what happens in syntax. Syntax gains from what expressiveness loses, and expression becomes all important. With a minimum of expressiveness and well-developed syntax, one could say: There will be a gala performance at the Opera this e v e n i n g , and with more expressiveness and less-developed syntax (insofar as the verb is concerned): Grand gala performance at the Opera this e v e n i n g . Generally, it can be postulated that a gain in expressiveness in our languages brings about a loss in expression and, therefore, a loss in syntax. Verbless sentences, of which there are countless examples, of countless varieties, are sentences where expressiveness replaces the syntactic development normally introduced by the verb.

Expressiveness, carried far enough, can make a sentence of a single word. For example: S i l e n c e ! C u r t a i n ! But if the expressiveness is withdrawn, these are no longer sentences. Exclamations such as S i l e n c e ! or C u r t a i n ! are felt to be sentences because considerable expressive momentum has been added to the words used. Without it, the word would not be felt as a sentence. The equation

expression + expressiveness = 1

allows for any variation, from a minimum of expressiveness, where expression accounts for nearly everything, to a maximum of expressiveness, where expression has a diminished role; what expressiveness gains, expression loses, and vice versa.

The interjection is an extreme case of this variation in language. It represents expressiveness which has somehow absorbed expression or done away with it. An interjection is a sentence whose dynamic force is not the verb, but the expressive movement, carried to its limit.

In the psychomechanism of language, the expressive movement is a variable which must be carefully taken into account in any serious analysis. In general, the expressive movement supplements expression: that is, whatever the circumstances, expressiveness compensates for what expression has failed to provide, given the intended message. And, in order to satisfy this intention the expressive movement, in its development, relieves expression of elements that would encumber it. This decrease in expression is compensated for by an increase in expressiveness.

This inverse relationship between expression and expressiveness is a fundamental fact of general grammar. In the beginning, when language was a matter of improvising, expressiveness carried everything and expression nothing. Now, in the language of the educated who speak with a certain care, expressiveness counts for little and expression for almost everything. Popular or everyday language


functions to the detriment of expression and to the benefit of expressiveness. ..

...In the general history of language, expressiveness has been primary, expression secondary. At some time, therefore, every language has been characterized by a certain loss of expressiveness. This loss was always offset by a compensatory gain in expression, since expressiveness in itself depends on improvisation and expression on what has been established in tongue. Indeed, it would be just as appropriate to say that any given language, at a given time, is characterized by the development of richer resources of expression, which in turn reduces what is required of the means of expressiveness.

Before the act of language occurs, the means of expression already exist, pre-registered in tongue, the established part of language. They are therefore available for immediate use. On the other hand, the means of expressiveness are developed later during the occurrence of the act of language itself, of which they form an integral and inseparable part.

There is good reason to state as a general principle that in our languages expression "absorbs" expressiveness. Through frequent use, the resources of expressiveness in the language become codified and established as resources of expression. To the extent that a means of expressiveness becomes established, it becomes a resource of tongue, a means of expression.

It is often difficult to measure exactly to what extent any of these resources have become established in tongue, and to what extent they still depend on improvisation. Full integration into tongue has certainly been achieved by those notions which have developed corresponding signs. Integration is less complete in those cases where what has been incorporated into tongue is the way of using signs. There are nevertheless some rigidly established ways of using signs which are conveyed by the signs themselves and are as much a part of tongue as are the signs.

Here we have the syntax of expression, and added to it, the syntax of expressiveness — a less well established syntax in spite of its having been established as a counter to the first. In the sentence Vetev walks in, the syntax of expression has taken precedence. In walks Vetev is based more on the syntax of expressiveness. What is called the grammatical order of words is completely established and belongs to the syntax of expression; this order is no longer followed if the speaker resorts to the rival syntax of expressiveness. Therefore, to be able to maintain that there is a rigorous grammatical order of words in a language like French, one must gloss over the syntax of expressiveness.

It is not easy to determine the degree to which the syntax of expressiveness has been established in tongue in cases where it is used to make up for a lack in the fully established syntax. The problem is more doctrinal than theoretical.


Tongue, as a whole, is made up of everything that has been established. But the fact must be taken into account that there are different degrees within what is established, and that the more established something is, the deeper its level in tongue, while less established items are at a more superficial level.

What is most fully established is always the least expressive. Thus the search for expressiveness constantly leads speakers to break away, more or less, from the deep-seated, commonplace, established resources of tongue. If used frequently enough, the new resources become, in turn, established as resources of tongue, thereby augmenting its resources and making what is already established more complex.

(Lecture of May 7, 1948, Series C)

THE PLACE OF THE SENTENCE IN LANGUAGE

A basic principle behind my teaching, often mentioned in my lectures, is that expression starts with what has already been represented. This is a constant which characterizes not the ontogeny of language, but its praxeogeny. The present constructional state of our modern languages makes it easy for us to verify the accuracy of the principle. It is simply a matter of examining what goes on in our own minds, in our mental world. If we have something to say (and in linguistics it all comes down to saying something, the constant goal of language) we look within, seeking among the representational possibilities permanently established there, to find the words to express what we want to say, and the words come to us, more easily in some cases than in others. "What is well conceived", wrote a poet and thinker, "can be stated clearly, and the words to say it come easily." Boileau had no doubt — and neither do I — that we call on words — units of potentiality in tongue belonging to what is represented — to construct sentences — units of actuality in discourse belonging to what is expressed. By examining what takes place in the mind, we see that this is so.

Scholars could simply have accepted this view of things. At our present stage of language development it would have been quite justifiable for them to judge linguistic reality by what they know of it firsthand. However, they have not done so. Like Molière's physician, they have changed all that, rejecting all the observable mental physiology and imposing the idea that people think in sentences without getting involved in words. And they have insisted a great deal on this point — though they have not yet managed to teach the language without using the words in it. The sentence, so it has been taught, is the only reality in language. And yet it remains true that in


order to construct sentences expressing what I want to say, I must seek and find appropriate, preconstructed words. That is what speaking is all about.

Strangely, declaring the sentence to be the sole reality of language — and thus taking things at their final stage, at the consequence and not at the condition — is equivalent to returning to that long-lost period, at one remove from primitiveness, when language had to operate within the first glossogenic area (the only one then available) in order to represent and to express. At that stage, there was a certain interference between the act of expression and the act of representation.


UNITS OF POTENTIALITY AND UNITS OF ACTUALITY: THE WORD AND THE SENTENCE

A long-standing principle in my teaching has been that expression starts with what has been represented. Representation is in tongue: tongue consists of acts of representation, each of which is represented by a unit of potentiality called a WORD. What is expressed belongs to discourse: discourse consists of acts of expression, each of which when finished results in a unit of actuality called a SENTENCE.

In the languages we are more familiar with, the sentence arises from a grouping — not an agglutination — of units of potentiality. These units, which consist of lexical substance with grammatical form, result from an instantaneous lexical prehension that precedes the sentence prehension. This lexical prehension gives rise to and results in the word.

Lexical prehension, the first formal prehension, is both grouping and agglutinating. The radical formative elements, singled out in the radical prehension , are therein grouped and agglutinated: they are given an initial form from which will arise the units of potentiality of language. Sentence prehension, the second formal prehension, which gives rise to the sentence, the unit of actuality where units of potentiality are assembled, is grouping only (a general law of language). Sentence prehension comes into play after the lexical prehension and consists essentially in grasping, under a form which is itself not destined for permanency, those established forms that were imposed on the lexical substance of the word in the earlier lexical prehension.

It is lexical prehension which gives rise to the words man, is, and mortal, which are units of potentiality in tongue. It is sentence prehension, an assembling of units of potentiality (words) that gives rise to the sentence Man is mortal.


The sum of all the lexical prehensions of notional substance carried out by the human mind constitutes tongue. The number of these lexical prehensions is finite: it could be found in a dictionary which left out none of the words of the language. For each of us, this number is the sum of the words we possess. The sum of all sentence prehensions, which are prehensions of units of potentiality, constitutes discourse. The number of sentence prehensions is not finite; therefore, while it is possible to compile a dictionary of the words contained in tongue, a dictionary of the sentences made possible through the choice and use of these words cannot be written.

Tongue is something established; discourse is not established. In discourse, there is great freedom, which is contrary to establishment; in tongue is non-freedom, which is consistent with establishment. A sentence that becomes established in the mind, and so is no longer felt to be ephemeral, becomes a word in tongue. In French some short sentences, through becoming established, are now felt as words: on-dit, qu'en dira-t-on, m'as-tu vu, etc.1 In ancient and archaic tongues like Sanskrit, long sentences can be found which, through becoming established and losing their ephem-erality, have been reduced to the state of nouns, involving special conditions of juncture, of joining formative elements...


THE PRECEDENCE OF TONGUE: ITS ANTICIPATORY NATURE

It is not a natural thing for the mind to turn inwards in order to ascertain what the language it has acquired as an inheritance — the mother tongue — really is. The mind's natural bent is to use language to express some thought either to communicate it to others or simply to make it clearer. So the mind must be forced, so to speak, to go against its natural tendency in order to focus on the mental operations behind the construction of tongue, which should never be confused with discourse.

Discourse is the answer to a projected actualization: undertaking an operation of discourse obviously involves a desire to act, to have an effect on someone. Without some such aim there is no discourse; thought remains "silent". This "silence" does not mean non-possession of tongue, but simply that tongue is left unused for the moment in the depths of the mind, where it can always be found wholly made, wholly constructed and ready on the spur of the moment to

1. Examples of the same phenomenon is English might be: a what-do-you-call-it, a whodunit.


provide the services that we expect because we possess it. In linguistics, one must never forget that tongue exists in its entirety in the mind even when there is no speaking or intention to speak.

By definition tongue exists prior to the act of language. The act of language takes place between tongue — which is a construct in the mind, an inheritance received after birth from those with whom one lives — and discourse — which can be derived at will from the means permanently placed at one's disposal by tongue.

In an abstract but very real way, tongue should be pictured as seated at a deep level in the mind, at the level of potentiality, while discourse, on the other hand, is not seated deep enough in our minds to be other than momentary — it has no permanence. Quite the contrary, it is a construction carried out as the need arises, whenever we have something to say. Thus discourse belongs to the level of actuality.

Analytically speaking, tongue and discourse are not constructed in thought at the same time. Tongue is something past, something accomplished; discourse is what we will actually accomplish with the means provided by the tongue we have acquired, whose acquisition accompanied the very formation of our minds.

This difference of level between tongue and discourse in the mind, (which having managed to create tongue within itself, uses it to construct discourse when momentarily useful) is something that one should never lose sight of when treating any linguistic question. It helps one to understand what the mind expects of tongue, and also what it expects of discourse. Potentiality and ease of expression are what the mind expects of tongue; of discourse all it expects is skilful exploitation of the potential readily furnished by tongue.

For tongue to provide the mind with the potential required for ease of expression, it must systematically anticipate all the needs of expression that could arise, however different they might be. It is therefore justified to consider tongue as an anticipatory system, established and stored in the depths of the mind.


FROM WHAT IS THINKABLE TO WHAT IS THOUGHT: TONGUE AND DISCOURSE

Items inherited from the past constitute the matter of tongue, and tongue itself, in relation to this inheritance, constitutes their form. Tongue is the form of the relationship, of the association which these inherited elements have forged among themselves. Tongue


exists only as a form. If it were not a sum of relationships established between items from the past, it would not exist. Inherited elements seen without any orderly and systematic relations encompassing them would not make up a tongue. They could not be used to express thought because the expression of thought requires tongue to be, in anticipation, a general representation of what is thinkable: in other words, to be one way of representing a general view of the universe.

A simple but important distinction to be made here is between what is thinkable and what is thought. Expression involves what is thought; one expresses what one has thought. Expression takes place in discourse where what has been thought is found being expressed. Representation, on the other hand, involves only the thinkable. Through representation, what is thinkable is divided, subdivided, internally organized — in short, systematized — and the result of all these different systematizing operations is tongue. Problems of representation are dealt with in tongue, while problems of expression, being of a different type, are left to discourse.

This difference of type is basically due to the fact that the matter of discourse (matter for expression) is what is thought, while the matter of tongue (matter for representation) is whatever is thinkable. Discourse is a form adopted by what is thought, in order for it to be expressed, whereas tongue is a form adopted by what is thinkable, in order for it to be represented.

The tongue which dwells within a person, offering him the potential for readiness of expression, is an all-inclusive representation of the thinkable subjected to a certain internal systematization. It offers the speaker a total and immediate image, without any particularizing momentariness, of everything that can be thought but which, being thinkable and only thinkable, is not yet thought.

Discourse is quite different. It can exist only if the speaker — who is the thinker actively thinking — has momentarily constructed some thought with the help of the thinkable, which is permamently available to him in his mind where it is represented as an internally systematized whole by tongue. The operation of constructing what is thought from what is thinkable is both a prerequisite of, and involved in, discourse.

There are many fundamental differences between what is thought and what is thinkable, between the thought intended for expression and the thinkable intended only for representation. The thinkable is the totality of potential thought, an integral of potentiality. The thinkable must provide in advance for any thought that a speaker might actually want to produce. In practice tongue, which establishes the thinkable in systematic form, must be able to allow for the expression of any thought. Tongue is the only integrator of the potential; this property is not found elsewhere. Discourse, involving only what has been thought, is constructed out of the thinkable but


never includes more than a fraction of what the thinkable allows for. Vectorially — vectorial representation is often convenient in linguistics — the thinkable could lead to any thought of any nature, whereas whatever has been thought, although derived from the thinkable, has left the thinkable behind and made use of only a part of it, the part that was suitable.

Thus, in passing from the thinkable of representation (tongue) to the thought of expression (discourse) there is a transition from the integral to the differential, from the whole to the part, from the whole potential — immediately available to the mind — to the part actualized, produced in the fleeting moment of need and necessarily momentary. Everyone is aware of this relation between tongue, the whole potential, and discourse, the parts actualized. Everyone immediately senses that what is thought is not as broad as what is thinkable, and that everything that has been, is being and ever will be thought, will never equal what could be thought. The thinkable definitively integrates, as a potential, all that is thought, even though thought continues to be multiplied.

Further reflection on the relation between what is thinkable and what is thought brings out the fact that while, in the mind, the thinkable has an integral, tongue, what is thought does not have any such integral. No one could hold and integrate in his mind all that is thought, a never-ending story, for people have thought and people will go on thinking.

Yet no matter how long or how deeply one thinks, actual thought can never get outside the potentially thinkable. Thoughts being expressed, though different from one another, will arise successively without ever ceasing to belong, even with all their differences, to the thinkable which, at the level of representation, integrates them potentially. As a potential, tongue is an integral of actual differences, which are unlimited.

The contents of tongue, immediately available to the mind, are the sum total of what is thinkable, presented not in a state of expression but in a state of systematized representation. And the content of any discourse is part of the thinkable, differentiated and taken from the rest for use in producing what is actually thought.

In the human mind, what is thought is differentiated whereas what is thinkable is integrated. These two opposed characteristics apply to discourse and tongue respectively. In discourse, thoughts arise which are momentary, singular, different from one another, and which are expected to be different rather than identical. In tongue, nothing arises', everything is, all immediately available. The thinkable is thus made immediately available — the thinkable internally organized, systematized.

In the general theory I am developing, the distinction between the terms expression and representation is most important. Tongue


in itself expresses nothing; it represents, it is representation. Expression takes place only in discourse, which expresses something already represented, using the means provided by representation. In linguistics and in grammar a great deal of confusion has resulted from the failure to distinguish clearly between expression, the role of discourse, and representation, the role of tongue.

(Lecture of March 12, 1948, Series C)

REPRESENTATION AND EXPRESSION: UNIVERSALITY OF THE RELATION

The subject of this lecture, representation and expression in the grammatical system of the French language, is very important in itself, but it also brings out a general, human fact as well as a social fact.

The human fact is the representation/expression relationship itself. It is a fact of language at all times, in all ages, in all societies and social groups. Which means that everywhere, always, spoken language has to confront the problem of balancing representation and expression. It has to construct tongue with acts of representation and discourse with acts of expression.

To have language spoken by man, the relationship between representation and expression is an absolutely necessary one, a human one. But the realization of this relationship, allotting some proportion to each of its necessary terms (representation/expression), is a social fact.

The social aspect of language has been so emphasized and the human aspect so neglected that it might be useful to pause here over the difference between them. A language like French has reached a certain equilibrium — the equilibrium of French. — between representation and expression. It has its own way of solving the equation

representation + expression = 1

where 1 is the symbol of the whole, that is all spoken language. Clearly, French, a highly evolved language, has solved this problem quite elegantly. Its own elegant solution is the social factor here. But the problem itself and the need to pose it, if not to solve it, is the human factor.

The human principle underlying language is that expression is possible only if something has first been represented. The necessity of representing something before expressing it is universal in space


and time. It is a sub-social fact. The social fact involves a certain reciprocal relativity, variable from age to age, between representation and expression. It would be a social fact not to have any representation, to have, in the place of representation, a state of spoken language which would replace the missing representation by experience. In other words, it would be a social fact to have to render what one wants to signify on the basis of a relation of this sort:

representation = 0 + expression = 1

experience = x 1

Since representation makes up tongue, in a case with no representation such as this, one would have to speak without any established tongue, proceeding directly from one's experience of things. This would be the situation of the animal. There would be language, but not spoken human language. Transition from animality to humanity takes place when the basis for expression turns into representation and so leaves mere experience behind.

Man carries the germ of representation in his mind. Moreover, all the great developments of thought reflect progress in representation. Basic arithmetic is representation born of man's experience of number. Algebra is a higher representation that springs from numerical representation. Progress is registered on a curve moving away from experience; representation can, with both profit and loss, be substituted for experience as such. Thus understood, progress is human. But as soon as a certain reciprocal relativity is established between representation and expression, we are dealing with a social factor. In other words, the human factor is:

representation (= x) + = 1

expression (= y)

and the social factor

representation (= a) + = 1

expression (= a')


It must be stressed that, for the problem considered here, the social factor can be explained only in the light of the human factor. Before any attempt is made to determine the proportions allotted within a given language to representation and expression, it is important to realize that language when spoken is the sum of the two. In a recent lecture, looking at language from a social point of view, I mentioned that tongue is what is established in the mind and discourse what is non-established. Discourse exists only because everything has not been established. If the established part covered the whole of language, I would be mute: everything would already have been said; there would be nothing left to say. The relation

is impossible, imaginary. However, like many such imaginary relationships it helps reasoning by providing a limit.

Another impossibility — radical, this time, rather than final — is the relation

Speaking with no established tongue, with none of the representations of tongue, would be a state of torment, animal torment, which we are spared by human representation, socially developed.

We all realize that the questions raised here are important to many different fields, including pathology. Studies of aphasia, if well conducted, reveal that certain aphasias at least appear to be basically a question of faulty representation; often it seems that a part of representation has slipped away or faded out. There seem to be gaps, systematic gaps in representation, causing systematic aphasias. In one case of aphasia the person could not substantivize, although he could still adjectivize. This created a pathological need to render thought (which was unimpaired essentially) by means of adjectives and verbs without the possibility of giving them a substantive as support.

(Lecture of November 23, 1951, Series B)


TONGUE AS CONDITIONER

Any grammatical paradigm, a conjugation for instance, involves a limited number of formal cases of representation. The representation corresponding to each form, to each formal case, is one. Its oneness is absolute; however, the oneness of this single representation has been established through using an internal mechanism that permits, on the basis of the single representation, thousands of different effects, different sometimes to the point of being oppo-sites. The example of the imperfect indicative in French will help make this clearer.

The imperfect represents the idea of an on-going event already accomplished to a certain extent. The analytical formula corresponding to this representation is:

positive, accomplished portion + = 1

prospective, yet-to-be accomplished portion

In this formula for the imperfect, the accomplished/prospective relation is not quantified, which means that in discourse the speaker has free play within the variation permitted. This free play allows the quantum of the accomplished portion to be increased or decreased at will. An imperfect can thus have a substantial accomplished portion, one which has lasted for some time, before the prospective portion, as in Pierre marchait . An imperfect can also confront the prospective portion with an accomplished portion which is not extended, one which has not yet lasted any length of time, as in Le lendemain Pierre arrivait2. Here, arrivait introduces an idea of suddenness, of something that interrupts the normal course of action and foreshadows the unexpected. There would have been no foreshadowing if we had said arr iva 3 , which is aprospective, that is, not internally prospective. Indeed, arriva, a past definite, is a form which represents accomplishment in extenso, without any already accomplished portion. A narrative in the past definite evokes a series of occurrences. Wherever an occurrence is represented as having some accomplished portion, be it ever so small, the imperfect is used. Thus, the imperfect has an invariable condition of representation, which can be written as follows :

1. = Peter was walking. 2. = approximately: The next day, there was Peter arriving. 3. = arrived.


accomplished portion + prospective portion = 1 (large or small at will)

Even if the accomplished portion is reduced practically to nothingness it still exists in the representation, though its existence is of so little consequence that the imperfect would be almost equivalent to the past definite, and the two would be grammatically, if not stylistically, interchangeable. It is thus possible to say both Le lendemain P i e r r e arrivait and Le lendemain Pierre arriva with only the slightest shift in nuance.

In representation, then, there is an absolute condition:

accomplished portion + prospective portion (positive)

and in expression, the accomplished portion may be quantified freely as one wishes. In tongue, the imperfect imposes its absolute condition for representation. In discourse, respecting this condition allows complete latitude within the set limits; the imperfect is free to do anything except transgress the condition whose image it reflects, this condition establishing the presence of an accomplished portion preceding the prospective portion, but not the quantity of the accomplished portion opposing the portion in prospect. This quantity is not established in tongue, and we know that non-established elements are part of discourse. This amounts to saying that, successively, we find:

-the imperfect in tongue, involving an accomplished portion of any length + a prospective portion;

-the imperfect in discourse, involving an accomplished portion of a certain length + a prospective portion.

There are therefore two ways of studying the imperfect. It can be studied in discourse alone, in the sentence as expressed, in which case the imperfect has not just one shade of meaning, but a thousand. Any attempt to reduce all thousand of these to any one of them considered basic would be wasted effort; one would inevitably get lost in making distinctions. (In fact, the better one discerns the differences, which may well be extreme, the more completely one can get lost.) The alternative is to study the imperfect first in tongue, to discover the unchanging systematic condition that it represents there, namely :


accomplished portion + prospective portion (no quantification)

and then later to study it in discourse, where all possible quantifications of the accomplished portion can be seen to occur quite freely, with all the various expressive effects they give rise to. The form1 is provided in tongue. The quantitative variations possible within the limits of this form are determined only in discourse. They are not established. They are free, as is discourse itself. Their form alone is established in tongue — a form which the imperfect cannot and does not transgress as it takes on any of the different quanta of accomplished portions permitted by the requirement that the imperfect must have such an accomplished portion.

I have used a simple fact to demonstrate what distinguishes something in tongue (formal established representation) from something in discourse (the free functioning of what is established). The established element never exceeds its limits; it leaves complete freedom for discourse to operate within the set limits.


1. Guillaume uses form, here and elsewhere, in the sense of the grammatical or "formal" meaning, as opposed to matter, the lexical or "material" meaning. For this reason some of his analyses have been described as "grammatical semantics".

PART V

LANGUAGE AND SYSTEM

CHAPTER 1

THE SYSTEMIC NATURE OF LANGUAGE

A LAW INHERENT IN ANY SYSTEM

Any system is necessarily both one, by virtue of its governing principle, and several internally, by virtue of the positions it contains. No system can contain just one element. By nature and by definition a system is binary: it must contain at least two elements to be a system at all.


A DIFFICULTY REGARDING THE STUDY OF LANGUAGE SYSTEMS

Systems are entities of tongue which are just as real as the forms that represent them. If, in spite of their importance, systems have not been studied up to now, it is due solely to the fact that they are abstract realities. Systems do not relate directly to the perceptible material reality of linguistic entities: they involve the more abstract reality of systemic links between a number of entities in tongue which together allow a certain number of complicated relationships to be expressed. These relationships are subject to the twofold law of coherence and reciprocal relativity which is always respected in the construction of language systems.

It is easy to see why the study of abstract entities, mere entities of relationship — a linguistic system is the subconscious organization the mind imposes on its own representations — should be less readily undertaken than the study of concrete entities, which have a different type of existence. I might add that studying systems in no way keeps us from considering facts; it simply leads us to extend the meaning of linguistic fact so that the systemic dependence of two or more forms is considered just as much a fact as is the existence of the forms themselves.

(Lecture of February 10, 1944, Series A)


TONGUE AS A SYSTEM OF SYSTEMS

Any tongue taken as a whole is a vast, rigorously coherent system made up of several systems linked to one another by relationships of systemic dependence and together forming a whole. Since this is true for any tongue, and since psychosystematic linguistics specifically aims to study the abstract entities of tongue, that is, the purely relational entities called systems, one of our problems is to identify the different component systems that make up the vast containing system, tongue. For the overall system is actually nothing more than the reciprocally relative systematization of the subsystems which have become differentiated within it.

Tongue is a systematic whole englobing the entire range of what is thinkable and containing systems each of which covers one specific part of the thinkable. These sub-systems have a natural tendency to become individualized and to stand as wholes, integrated within the vaster overall whole, tongue. Their individualization as wholes within the whole of which they are parts is not always clear-cut; the theoretical linguist must be careful to remember that although, as was just pointed out, the different systems of a tongue tend to become individualized within it, they also tend to avoid any gap between each other, and this makes the passage from one to another almost imperceptible. A tongue is an antinomical construction always striving to achieve opposing aims. If a sentence, for example, is to have meaning, the words must be distinguishable, and yet their distinctiveness must also be erased for a brief moment. In its internal structure every tongue, as an integrating system, tends to distribute itself over a number of integrated systems each one of which forms a distinct whole within the containing whole. At the same time, an opposing drive tends to keep the integrated systems as close together as possible. These two antinomical tendencies — the one toward the separation of systems and their identification as wholes, the other toward the maintenance of a close and almost continuous link between them — are forever in search of a satisfactory balance. This balance is, in general, elegantly achieved in highly evolved languages, especially those of advanced civilizations.

(Lecture of March 2, 1944, Series B)

THE ESTABLISHED AND NON-ESTABLISHED ASPECTS OF LANGUAGE

Insofar as its technique is concerned, the psychosystematics of language is based on the principle — this is both a principle of structure and a principle of analysis — that the mind can become aware of itself only by intercepting its own activity.

LANGUAGE AND SYSTEM 105

This principle applies to any formal structure in tongue, that is, to all the forms which through alternation within a closed system are "transportable" to various ideas. For example, let us consider the forms of the verb conjugation: they constitute a closed system and the speaker, instantly grasping the system as a whole, chooses from among the finite number of forms contained therein the one that best suits his aim in discourse, that is, what he intends to express. He picks from among a finite and never very great number of forms; and if, for example, the forms are taken one after the other, a moment soon comes when it is necessary to take some form over again. It could be said that the alternation takes place in a closed circuit.

Among linguists the question whether or not a tongue is a system has been argued time and again. The discussion is over and done with the moment we agree that a system exists wherever forms alternate in a closed circuit, and is absent — a non-system — wherever forms in the established part of language (tongue) alternate in an open circuit. One example is the system of number in languages with two numbers, singular and plural. Another is the system of gender in French, which involves two genders, masculine and feminine. These are systems because the forms, being only two, alternate in a closed circuit. The cases of the Latin declension form a system, since the alternation operates in a closed circuit.

But on the other hand, wherever notions can be substituted for others in an open circuit, no system exists. This is true for the particular ideas that form the content of the parts of speech. They can be substituted for one another without expressly closing the field of substitution. Under the substantive form I can list a very large number of notions and, if need be, freely increase this number without meeting any limits in i n t e l l e c t u . The case of the parts of speech themselves is quite different. Their number is finite, in fact quite small and they alternate in a closed circuit. They form a system. The endlessly recurring controversies on this subject seem to me to be settled by this explanation which clarifies both system and non-system, both the free and the non-free in tongue, the established part of language.

When applying the technique of mental interceptions in our last lecture, we divided the thought ➙ language relation, the object of the science of language, into two parts: the established part which we called tongue, and the non-established part, which we called discourse. It should be said right away that systems exist only in the established part. Let no one argue that the sentence belongs to discourse and yet is a system; the system of the sentence is in fact an integral part of tongue. It is not the system but the use of the system along with the free, momentary choice of the ideas the sentence is to express that makes the sentence. This free choice is not systematically conditioned, and although the system of the sentence is made up of a finite number of systemic conditions (with no increase possible) , the use of the system makes a far larger number of consequences


and combinations possible. Only the use of the system, within its set limits, belongs to discourse. The line is not easy to draw: the system governs usage, and usage becomes established.


SYSTEM AND NON-SYSTEM IN TONGUE

Tongue being always the totality of what has become established, one of its important characteristics is that both freedom and law, non-system and system, have been incorporated into it, each confronting the other without detriment to either. There is a sort of systematic division of labour, two tendencies: one leading to the development of an open circuit of ideation, the other to closed circuits of ideation.

Tongue, the established part of language, is systematic in toto but its systematization includes both free established elements (non-systematic, found in open circuits) and non-free established elements (systematic, found in closed circuits). The free, non-systematic part operates within the framework of the non-free, systematic part. The systematization of the non-free, systematic part has already been achieved, while the systematization of the free, non-systematic part is yet to be achieved. The morphology of any tongue is the establishment of closed circuits of ideation within it. If open circuits are established everywhere, the language is without morphology.

To understand what I mean more clearly, suppose that French contained nothing but particular ideas that could be freely substituted for one another. It would be a language without any kind of morphology. But this is not in reality the case. In tongue, as a consequence of its structure, closed circuits of ideation have been set up with interchangeable notions that are all transportable. These circuits are the systems of tongue, represented in ordinary grammars by paradigms .

It might be better to say that tongue establishes a free ideation whose conditioning factors remain unacquired, systematically undetermined, and then subjects this free ideation to a non-free, systematically conditioned ideation. Non-free, systematic ideation operates within a closed circuit, that is, within a circuit of forms internally transportable to the same linguistic field. (The speaker must choose among several forms, finite in number: e.g. 3 moods, 2 aspects, 15 verb forms: it is necessary to choose among the prehensive forms wher confronted with matter.)

Free, non-systematic ideation, which operates within an open circuit, is the free activity of the mind, with negative systematization,


relating not to the law of the activity itself (the creation of ideas — which knows no laws) but to the law of a secondary mental activity: the activity the mind carries on within itself in order to prehend its own activity. On the one hand, then, there is the free, non-systematic activity of producing ideas, and on the other, the systematic activity of shaping, prehending what has been produced. In the structure of tongue two things must thus be distinguished, both situated in the mind: the activity of the mind in operation and the activity of the mind in prehending its own operation. Systematiza-tion is not found in the mind's operational activity, in its free activity, but in the prehension of this activity by the mind. Because — let us not forget this basic truth — the mind can perceive its own existence only insofar as it can prehend itself and so distinguish the different moments of its own internal activity. This prehension coincides with representation: it is what representation is.


THE HISTORICAL FILIATION AND RENEWAL OF SYSTEMS

There is an aspect of the historical development of the spatio-temporal system in the Indo-European languages which we have yet to make clear. The establishing of a system is based on primordial principles inherent in the human mind. Once established, the system perpetuates itself by producing consequences stemming from its nature. Over a given period of time, these consequences develop to a certain extent — to the greatest possible extent if the development takes place in a period of linguistic calm.

A system given free rein to follow this pattern of development through may well encounter historical circumstances that counter its development and bring about its destruction. It will then become something past, cancelled out, through loss of its mental significance, as if it had never existed. When reconstructing, the mind cannot build upon this past which has been eliminated; since no other basis is available, it is necessary to start over again, with the constructional principles permanently inscribed deep in the subconscious. Thanks to previous constructions, the mind has a fairly clear view of these principles and at the same time some experience of the constructional combinations that can come from skillfully exploiting them.

To understand clearly the history of the systems adopted successively by the Indo-European languages, it is essential to remember that the collapse of an established system entails a return to a situation where no account is taken of what was built up before. Making a fresh start, the mind goes back to the primordial principles


of thought that have already been put to work many a time in previous systematizations.

This mental cancellation of what was already constructed — in some cases it can be almost complete — complicates the history of systems. It gives rise to new beginnings in system construction, so often undertaken before and so often carried out successfully. Under these conditions it is easy to see why an examination of systems does not give a very accurate idea of their historical filiation. Strictly speaking, the filiation is broken every time the reconstruction of a collapsed system involves returning to the primordial principles used at the very beginning. Depending on the language, the construction built on the basis of these primordial principles, which must be resorted to every time the formerly constructed system fails to perpetuate itself, represents a development — sometimes considerable, sometimes less so — of these principles. Perhaps the degree of development is what gives us the best indication of their relative historical age.

(Lecture of January 27, 1944, Series B)

CHAPTER 2

THE SYSTEMEC NATURE OF THE WORD

DISCOVERING THE SYSTEM: AN APPROPRIATE ANALYTICAL TECHNIQUE

In any language the word constitutes a system. Discovering this system, that is, discovering the constructional mechanism behind the word, is the task of the psychosystematics of language, with its special technique called positional linguistics. A study of the constructional mechanism of the word in Indo-European languages will give an idea of the technique, developed by and for this new branch of the science of language.

When we apply the positional technique, a linguistic phenomenon is always represented vectorially, so to speak. That is, its dynamism is represented as a vector movement which can then be examined inter-nally by means of successive cross-cuts. These cuts mark positions occupied along the vector movement, whence the term positional lin-guistics, coined to denote this technique which is regularly used in psychosystematics.

The positional method has enabled us to discover the system of the word not just in Indo-European languages but in the languages of the world in general. Wherever it is applied, the method produces notable results and has the merit of being uniform throughout. Know-ing from previous reflection that the primordial mental operations in the case of the word in evolved languages are universalization and singularization continually alternating with one another, I have found that in the languages we are most familiar with the word is based on an initial particularization which is then matched with a concluding generalization. The following vector lines illustrate this movement:

This diagram represents the activity of the mind, which itself prehends this activity by the only means available: by making cuts across the vector lines. A cut intercepting the. first, particularizing


operation, gives the WORD BASE. The successive cuts, intercepting the second, universalizing operation give the vector forms added to the WORD BASE. A last cut at the end delivers the part of speech. Schematically, to show clearly what positional linguistics essentially involves, we have the following:

Particularization Universalization

First cut Successive trans- Final cut (result: material material cuts (result: part part of the word) (result: vector of speech)

forms)

No linguistic problem exists which cannot be elucidated, usually in a most striking way, the moment it is conceived in a form allowing this special analytical technique, positional linguistics, to be applied. The greatest difficulty in applying the positional technique — an unvarying technique, remember — is to discover the primordial vectorial form of the phenomenon under consideration. In the case of the word, for instance, it was first necessary to see that in evolved languages the word depends on a double process of particularization and generalization to bring about its construction. Concluding or rounding out the word is always left to.the generalizing operation of thought which, as the second and last operation, endows the word with its general form: the part of speech. Once the primordial vectorial form of the phenomenon has been found, the rest offers no serious difficulties: it is quite simple to discover what the various vector movements have produced in the entity of tongue under consideration.


1. For a discussion of the term vector forms see below "The Mechanism of Word Construction in the Indo-European Languages" and "Morphology and the Genesis of the Word".


THE GENESIS OF THE WORD: MATTER AND FORM

...A fact that has not yet been recognized: the relation between form and matter is one of order, of sequence, and nothing more. Lexical matter is the product of the mind's first movement away from one of its extreme limits, singular or universal; and grammatical form is the product of the mind's second movement, back to its starting point.

It follows that form will involve universalization in any language which generates its vocables by starting from the universal, withdrawing far enough to obtain the degree of particularization in meaning required and then returning to the universal.

This is the case for the evolved tongues we know best. In our minds, having gained a certain power of abstraction, we start out from the universal in the direction of the singular, stop en route at the point corresponding to the intended degree of particularizaon, and then return to the universal. It is this return movement that generates the form. The first movement, away from the universal toward the singular, generates matter. In a diagram this may be shown as follows:

The superimposing of the two directed movements can be sensed in any lexeme in English2 and so is directly verifiable. Let us consider, for example, a lexeme with a very general sense, like thing. Insofar as its matter is concerned, it represents a minimal movement away from the universal followed by a return (corresponding to the form) to the starting point. In a diagram:

1. In the following three diagrams all that is involved, curiously, is the movements; the systematic order in which they arise is not shown in operative time. The preceding passage, "Discovering the System: an Appropriate Analytical Technique", indicates that we are to understand the second operation (= form) as being triggered at the point where the first is intercepted.

2. Guillaume actually uses examples from French.


For a less general word, man for example, the diagram would be:

It is the movement that generates matter which provides the word in our languages with its initial meaningful content; the movement that generates form gives the word in finem its distinguishing mark as a part of speech.

The part of speech is a universalization which integrates and finalizes the word's development, and represents in every case — since the movement of universalization has reached its limit — a confrontation of the universal view with itself. The final limit of the universalizing movement may not be surpassed, and so has no sequel with which it may be contrasted. The only remaining possibility is to oppose the universal view to itself without stripping it of its identity as a universal. What results is a curious dichotomy of the universal view which takes the mental form of two antinomic universes: the space universe and the time universe. For new students I shall go back over the movements of thought responsible for this dichotomy.

Words in our languages, as a consequence of the form-generating movement, which relates them once more to the universal from which the process began, finish either in the time universe or the space universe. When they finish in the space universe they are nouns; when they finish in the time universe they are verbs. Whenever they finish in the time universe, they take in finem the marks of representational categories pertaining to time, that is, mood and tense, and the ordinal person (the person which can be declined and change rank). The categories of mood, tense, and ordinal person are the


visible determinants of the verb. As soon as these determinants have appeared, the word is a verb. Whenever words finish in the space universe, they assume the mark of the categories of spatial representation, that is, the non-ordinal (non-locutory) third person, number, gender and case (already indicating the function or functions the word can fill directly in the sentence, without recourse to a preposition) .

(Lecture of November 25, 1943, Series A)

OPERATIONS OF WORD CONSTRUCTION: DISCRIMINATION AND CATEGORIZATION

Languages have shown a universal tendency, from some stage of their evolution, to mark the distinction between noun and verb and to make it more and more formal and categorical. The linguistic distinction between noun and verb is closely bound up with the noo-logical distinction between time and space. Although the question is a profound one for general linguistics, a brief outline will at least bring out its main features.

Every lexeme is the outcome of two mental operations which tend toward opposite ends: a discriminating operation and a categorizing operation. The first, discriminating operation tends to abstract the particular from the universal and so enables the lexeme to acquire its individual identity. Individualizing the lexeme is a function of the abstracting operation which withdraws the particular from the universal. Should the mind have difficulty in carrying out this first operation — a hypothesis of this sort should not be neglected in linguistics at the most general level — the lexeme would have trouble acquiring its individualization. If such a difficulty ever existed when languages first originated, it was always overcome quite early, and the lexicons of our languages bear witness to the power acquired by the human mind to abstract from the universality of the thinkable the notions it wants to be able to consider individually. Abstracting the particular from the universal, discriminating through abstraction a notion drawn from the mass of thinkable notions in order to consider it distinctly is therefore an operation that the human mind now has learned to carry out with power and ease.

In our evolved languages, the discriminating operation is automatically followed by the reverse operation which is, in a way, its mirror-image. The particular is now returned to the universal from which it was abstracted in the first place. This second operation is not one of discrimination, aiming to individualize the lexeme, nor one of anti-discrimination, aiming to de-individualize it and thus undo what has been done. It is of a different type: a categorizing operation which, while letting the lexeme keep all its


acquired individuality, aims to produce in the mind, in the understanding, a categorization that will be as general —as non-particular-as possible. This categorization is expressed by the part of speech.

To arrive at this categorization, which should be as general, as non-particular as possible, the universal must be opposed not to the particular but to itself, the universal. The universal can be opposed to itself through two different, mutually exclusive (antinomical) modes of categorization. These two antinomical modes of categorizing the universe are space and time. (Although the deepest of thinkers have often dwelt on the problem, the exact mental operation involved in the genesis of these two modes of categorization remains mysterious.) Accordingly the verb can be defined as a word whose categorization reaches completion in time; the noun, as one whose categorization reaches completion outside of time and, consequently, in space (space here being everything that is not time).

Every lexeme in an evolved tongue is thus a notion abstracted from the universal by a discriminating operation and returned to the universal by a categorizing operation which does not undo (this is an important point) what the discrimination has achieved. Thus as the mental origin of the word, we have a full universe containing the particular notion to be abstracted and, as the mental term of the word, an empty universe from which everything particular has been withdrawn. The full, original universe is internally specified by the particularity of its contents. The final universe is one which, being void of notion, can be specified only in relation to itself. It is this specification, undertaken under purely formal conditions (since the universe is thought of as being void of specifiable matter) which ends up as the space-time contrast: abstracting the universal from the universal. The separation of noun and verb is based essentially on this contrast.

(Lecture of February 19, 1942, Series A)

THE MECHANISM OF WORD CONSTRUCTION IN THE INDO-EUROPEAN LANGUAGES

...These laws apply in the following way: the operations of particularizing and generalizing interact with each other in the first, material phase of word construction, without the generalizing ever being pushed to the limit, developed as far as the universal and becoming a whole.

Indeed, beneath the particularization in the initial, material part of the word a universalization develops which is carried as far as possible but never as far as the overlapping particularization. For if ever the two did become equal, the particularization would disappear: the WORD BASE would no longer exist.


The formula corresponding to the construction of the initial, material part of the word, and one which should always be kept in mind for the question under discussion here, is:

Particularization = 1

Underlying universalization = 1 - q

In other words, the universalization under the particularization must always and everywhere satisfy the condition:

universalization < particularization.

Thus when the initial, material part of the word reaches its term, the particularization is completed, whole, whereas the universalization involved is incomplete, partial. To be completed it must continue developing and, in order to continue, it needs supports, which it finds in the vector operations leading to a part of speech.

In the system of the word the vector operations leading to the part of speech correspond to the portion of universalization that had to be postponed so that the particularization carried out would not be absorbed, would not be destroyed by this opposing action.

We can represent all this quite concisely in a diagram:

Once the word is constructed, a condition of completude for both the particularizing and the universalizing movements has been satisfied, namely:


This double condition of completude, whereby a complete particularizing must have a complete universalizing as its counterpart in the overall system of the word, is a condition universally respected in the word of the Indo-European languages, and everywhere satisfied in these languages.


MORPHOLOGY AND THE GENESIS OF THE WORD

Between the two operations — the initial one of discrimination, producing the WORD BASE, and the final one of categorization, producing the part of speech which concludes the word — there are mediating operations whose essential purpose is to carry to its term the final universalization sought.

Thus, the structure of the word — the system of the word in our evolved languages (for the word is a system there) — is as in the following diagram:

The mediating operations take care of the transition between the initial discrimination, which brings a word its notional substance, its matter, and the final categorization, through which the word receives its general form. Without them, the word could not be constructed. Reasons for this will be given later, but for the time being we shall merely point out that their existence may be observed in all languages with words.

These mediating operations between the WORD BASE and the PART OF SPEECH which they lead up to, all consist of grammatical indications constituting the formal, morphological part of the word. It is worth dwelling on this for a moment to clarify some important points of theory.

The mediating operations that convey the word from word base to part of speech are certainly formal operations, as all grammarians would agree. But they do not constitute the form of the word; they


only lead the word up to the form, which is the part of speech. It follows that where forms are concerned a distinction should be made between the final, conclusive forms — the parts of speech — and the mediating, vector forms which provide a support for thought as it advances toward the final form.

Grammatical indications of gender, number, case and person for the noun and indications of mood, tense and person for the verb represent in tongue, where they are incorporated into the word, mediating vector forms entrusted with conveying the word to its absolute formal conclusion.

Although they prepare this conclusion, vector forms are not themselves the conclusion; as we know, the conclusion is a part of speech. We can say that the part of speech is a conclusive form. The distinction between pre-conclusive vector forms and the conclusive form of the word is important because it enables the formation of the word to be schematized in a convenient new way, as follows:

Besides their own meaning, the pre-conclusive forms signify in the word, through their position, a movement away from matter toward form. In noting this, one is implicitly recognizing that they are found in the word prior to the pure form, which is the final form. In other words, formal though they are, the pre-conclusive vector forms incorporated into the word retain something material1, something pre-formal in themselves.

Only the part of speech is absolutely formal. Having nothing material, it concludes the word by halting its development. This absolutely formal nature of the part of speech has semiological consequences which should be pointed out.

Not being absolutely free of notional content, the pre-conclusive vector forms may have a material representation in tongue, a representative sign, a mark. In a language such as French, gender, number, person, mood and tense are rendered by material signs. It is quite different for the conclusive form, the part of speech, which has no representative sign of its own, in French or in any other language. To establish itself in the mind as a last step (when the word reaches its finalized state) the part of speech does have determinants which precede it and which, one can correctly say, trigger its occurrence.

1. That is, they have some meaning of their own.


But — the fact must be stressed — the part of speech is not represented materially. It is produced not explicitly, with its own sign, but implicitly, at the end of a series of vector determinants which together provoke its occurrence but which are not themselves the part of speech, either separately or together.


A STRUCTURAL DEVICE: THE RADICAL BINARY TENSOR

...To complete the analysis all we need is to know what mechanism gives the human mind its power. Pending better results through analysis, some knowledge of this mechanism can be gained a priori, without interpreting or assuming too much, through ordinary reflection.

We can begin with the obviously well-founded idea that the mind obtains its power from its ability to particularize and generalize. Deprived of this double ability, which makes up an internally binary whole, the human mind would be without force, inoperative.

If these two operations which give the mind its power — particularization and generalization — are stripped by abstraction to their bare mechanism, they amount to two movements of thought, one from wide to narrow (particularizing), the other from narrow to wide (generalizing) . Reduced to its quantitative basis, particularization is a movement from more to less, and generalization a movement from less to more.

The mechanism giving the mind its power is the adding together of two tensions with no recurrence, that is, no moving backward: tension I, closing, progressing from wide to narrow, and tension II, opening ad infinition from narrow to wide, as in the following figure:


This mechanism, basic to the power of the mind, has, in this study, been appropriately named the RADICAL BINARY TENSOR1. (Extract from an unfinished, unpublished manuscript entitled Essai- de mécanique intuitionnelle. Approximate date of composition: 1951.)

THE ROLE OF INCIDENCE IN DETERMINING WORD CATEGORIES

A word with a material meaning, a word which is a lexeme, contains indications as to both its fundamental meaning and its intended use — the role, defined within certain limits, it is slated to play in the sentence; within these limits the word is delimited and its category determined. Limiting the role the word is going to play limits the word itself. To clarify this very important point let us examine a verb such as to walk in the infinitive. It has as its meaning an action or process, but notice that this is only a small part of what it contains. Its containing or annexing power extends in fact to all it is intended for. (It involves providing for only certain applications, and by the same token involves excluding by anticipation other applications, foreign to those actually provided for.) A verb in the infinitive like {to) walk begins by providing for its categorization in time; this involves belonging to a mood, which is an indication of its aptitude to take the mark of time. This is the infinitive mood, the case of {to) walk. Materially, the idea of the noun walk (as in take a walk) and the idea of the verb walk are the same, except that the latter provides for a temporal conjugation whereas the noun walk, by its very nature, excludes this, although it has exactly the same notional comprehension. From a certain point of view, therefore, the verb walk is far from being the equivalent of the noun walk. The two words do not have the same possibilities where the faculty for taking the mark of time is concerned. Such a possibility is non-existent for the noun walk; on the other hand, it is highly developed for the verb walk. But providing for an effect or, if you like, a use to be attached to a word, involves not only the possibility for the word to be finally categorized in time or in space but also other modes of application, among these being the mode I call incidence, whether internal or external.

Incidence is a property which grammarians, regrettably, have never taken into account; yet it is the one which has the most direct influence in determining the category of a word. The substantive and the adjective, together making up the part of speech called noun, obviously differ in their mechanisms of incidence. The adjective is

1. This device is often easily recognizable in Indo-European languages, particularly in the article and in number as well as in the general form of the word. See Langage et Science du langage by Gustave Guillaume.


never incident to itself; rather it is incident to a support that it does not concretely provide for ahead of time. For example, the adjective deep can be said of many different supports that discourse makes it momentarily incident to, since everything that discourse produces is momentary, just as everything belonging to tongue — because it belongs to tongue — is non-momentary. Everything in tongue is permanent. Everything in discourse is a momentary use of the permanent resources of tongue. Deep, or for that matter any adjective, permanently provides for incidence to something other than itself. To a large extent, this is what makes an adjective an adjective. On the other hand, the substantive depth permanently provides for a type of incidence which will not go outside its own signification or, to use a very precise scholastic term, its connotation.

The infinitive (to) walk is a form of the verb which makes permanent provision for the verb to be incident to what it signifies, an incidence thus identical to that of the substantive. This identity of incidence allows us to consider the infinitive as the nominal form of the verb. When I say A walk can be tiring, walk is not one lexeme said of another lexeme represented by a noun or pronoun, but one said of itself. It thus provides both the import of meaning and the support for the imported meaning. If a word can find a support for the meaning it imports without having to go outside this meaning, that is if a word does not transport its meaning to a support chosen outside itself, this is enough to give it the appearance of a substantive. When I say Only the true can be loved, the word true imports a meaning and this imported meaning finds a support for itself without the mind's having to go beyond it. As a consequence the word true becomes a substantive in use. When, in French, I say Un saint trouve sa jo ie dans la pauvreté, a word which is an adjective (sa in t ) also becomes a substantive because it has been given a support without going outside its own meaning. But this time, the support found in the field of what the word saint connotes, or signifies, is found not by an extension of the idea to all it encompasses, but rather by an anti-extension, a singularization which provides it with a setting in the narrow image of the individual1.

Thus an adjective seems to be a word for which provision has been made in tongue for incidence to some support outside of what it means. However, when used in discourse, it may eventually find a support for itself without going outside what it signifies, in which case the adjective automatically becomes a substantive.

A walk can be tiring and To walk can be tiring are two sentences, each with a word that finds a support within the idea it expresses, or imports. But there is a difference. On the one hand, the substantive

1. This use of an adjective as a substantive representing one individual is common in French. In English an adjective is more likely to become a noun representing a collectivity, e.g. the rich, the deaf, the needy.


walk rules out a priori any final temporal categorization which would give it the capacity of taking the mark of time; on the other hand, although the infinitive walk is like the substantive walk in being incident to a support found within its own meaning, it is characterized by the fact that it proposes or provides for a final categorization in time, and is thus ready to take the mark of time.

Thus the infinitive and the substantive are identical if considered only from the point of view of incidence, which is internal to the meaning in both cases, but different insofar as the final categorization of the word is concerned: in the substantive walk it is effected outside of time, in the infinitive walk, in time. This difference is so important that it keeps grammarians from putting the infinitive in the category of the noun. The infinitive is a position adopted by the mind within the category of the verb, but barely within it, where it begins, just beyond the category of the noun.


INTERNAL INCIDENCE AND EXTERNAL INCIDENCE

It is through its internal incidence that the substantive differs from the adjective, whose incidence is external to the meaning imported. For example, when we use the adjective beautiful, we apply its meaning not to something included in the idea of beautiful, but to something outside it which beautiful by itself cannot designate. It follows that the adjective beautiful can be said of many different things, with no tangible limits (beautiful work, a beautiful book, a beautiful landscape, a beautiful woman). The ease with which the meaning of beautiful can be transported to supports as varied as one wishes makes it perfectly clear that the incidence of beautiful is external to its own meaning. Transporting meaning in this way to many different supports is impossible in the case of a substantive, whose capacity of application, or aptitude for incidence, does not go beyond the meaning imported. As a substantive, man can be said only of beings belonging to the collectivity subsumed by this word. The incidence does not go outside this collectivity.

And yet the adjective beautiful only needs to be made incident to what it means and be conceived within the limits of this incidence to appear as a substantive. In more general terms, with the adjective the possible support extends beyond the import. With the substantive, this extension is prohibited. In the example The beautiful and the true are two sides of the same coin, "beautiful" is a reapplication of the import' beautiful to itself. The import has become a support: here beautiful is no longer felt to be incident to many different


beings through external incidence; it is felt to be incident, through internal incidence, to its own meaning. In every word, of whatever kind, two consecutive, related mental operations can be distinguished:

1) the importing of meaning 2) the transporting of the imported meaning to a support with,

as far as this second operation is concerned, two types: a- the type where the support is found within the imported

meaning; b- the type where the support is not found within the import

ed meaning. It is the grammatical person, ultimately, that forms the support

for the meaningful import of the word. A word contains the notion of logical person only insofar as the imported meaning contains reference to a support. It is the reference to a support which gives the word logical person. In other words, the logical person is present in the substantive, and can be considered absent in the adjective. With the substantive there is import of meaning, and the meaning imported holds the incidence, or application to a support, strictly within the limits of the import. Analytically, then, we are dealing with a term which provides for a relation between an import and a support, and the presence of the support entails the presence of person.

Just like the substantive, the adjective imports meaning, but here the search for a support remains unconditioned. The adjective can be said unconditionally of many different supports. From this it can be safely concluded that the adjective does not itself contain any indication of its support, nor, consequently, any indication of logical person.

This reasoning, confirmed by the way we use the two types of words, indicates that the substantive is a word satisfying the condition:

meaning import

logical person positivized

and the adjective a word satisfying the condition:

meaning import

logical person negativized.

In the adjective a support is not a prerequisite; in the substantive it is.

(Lecture of June 4, 1948, Series C)


SYSTEM OF THE WORD AND SYSTEM OF THE ARTICLE

Historically speaking, the article-noun is a recent part of speech. Understandably so: it brings to an end the endogeny of the parts of speech, which is an historical process (long-lasting, sometimes slow or sometimes more rapid) before being a systemic one (instantaneous in the speaker's mind). Here we find the distinction between historical time and operative time. Operative time involves the projection in homogeneous, surface dimensions of a process previously undertaken and carried out in heterogeneous, historical dimensions. The historical process is finally deployed along another axis in the form of a result, one which reproduces the operative system by making it instantaneous. Operative time in structural psycho-systematic linguistics is the time involved in this deployment. The deploying itself is somewhat complex; within it are processes which have their own deployment.

The complexity of this linguistic phenomenon, whereby the processes involved in reaching a solution are retained in the solution itself, stands in the way of a broad, synthetic description of the phenomenon, and necessitates a presentation which breaks it up into successive moments, each containing other moments, presented in profile. Here we have a whole projective geometry functioning deep in the mind and based on an intuitional mechanics which is reflected in the structures of tongue because they reproduce it within themselves. We would know nothing of this mechanics if we could not see it in the structures of language. The proper aim of language is what is useful, what is strictly necessary, and so language does not portray this intuitional geometry in itself as something basically see-able, but only the translation of what is seeable into something sayable. The speaker need only know how to use the mechanism of the verb, of the moods, and of the tenses; he is spared the trouble of showing how it works. The economy of language — and it really is an economy — calls for numerous reflections which may seem quite novel. Let me mention a few.

Language presupposes that mental activities are mentally perceived and prehended; but language need only produce something effectively sayable from this mental perception, need only make it expressible. To explain the efficacy of this making something sayable out of something thus seen, the linguist must be able to translate it back into something seeable. The linguist's task, wherein lie both his value and his pathway to knowledge, is to translate back — to discover how to translate back — into visible entities (made visible through explanatory diagrams) that which language reveals only indirectly, without the intervention of analysis, through making it effectively sayable.

From his writings, Leibnitz appears to have felt this difference between what is perceivable or primary in the mind and what is expressible or secondary, and brought out in human language. That is why


he so wisely advised thinking in diagrams. "Things impede each other, ideas do not." Diagrams are still things, but less so than the signs language uses to exteriorize its interiority. To think in diagrams is to keep things, to a large extent, from impeding each other. But to arrive at the exact diagram needed requires sustained reflection carried out with rigour and finesse. There is a real risk of constructing false diagrams which fortunately is greatly reduced by the fact that in order to construct a diagram one must start with extremely simple, elementary views with highly plausible exigencies.

The value of the psycho-systematics of language lies both in the successes so far achieved, and in the plausibility — to the point of being indisputable — of its postulates, such as the idea that man dwells in the universe and consequently, following the chain of plausibilities, that he belongs to the universe in which he dwells, and that to win his autonomy in it he must belong less to the universe and make the universe belong proportionately more to him. This is done by opposing the universe where he, as man, dwells to a universe that dwells within him, for which thinking man has become the setting. This universe is tongue, consisting of viewing ideas transformable into viewed ideas. The article is an operator serving to transform the virtuality of a viewing idea into the reality of an idea as actually viewed. The article in human language is a small word that "realizes" things or makes them real. Its non-use is therefore one cause of non-realization.

Proof of this (science lives not on truths, but on proofs) can easily be produced in French, a language with a high degree — an extraordinary degree — of internal systematization, making its study of great value in structural linguistics. Compared to highly systematized languages, less systematized languages tell us little about what language systematics is capable of; they tell us more about how much non-systematization languages will accept. Language systematization is at its most observable in the great modern languages of Indo-European origin, a fact which makes them of special interest to structural linguistics. Following this line of thought, remember that it is not the less systematized that explains the more systematized, but the more systematized that explains the less systematized. . .

Just a moment ago, when I was saying that the article realizes, I had an easy, immediate proof in mind: pas and point evoke very small things; they are words, impression-collectors, giving an impression of smallness. For example, in Vous n'avez avancé que d'un pas2, pas means very little, a very little whose reality, in spite of smallness, is affirmed by the article. Un pas is a very short,

1. These French nouns mean, literally, step and point or dot; both are used in negations.

2. = You have advanced only one step.


real distance between two very close, consecutive points. Pushed to the limit, the reality becomes precarious: if I withdraw the realizing article which is the guarantee of its conservation, this precarious reality disappears. It is in this way that the substantive-noun PAS becomes a negation through systematic mental mutation, as in Vous n'avez pas avancé1, or, with supplementary recall of the small, real quantity un pas, a recall emphasizing its absence, Vous n'avancez pas d'un pas2.

What I have just said about pas also applies invariably to its mental counterpart or replica, the word point, as well as to other less frequently used words which involve the idea of smallness as a limit of decreasing size. The words pas and point in themselves do not represent fixed smallness but decreasing size; when the article is applied to them, the decrease is suspended and made into some fixed smallness. E.g.: Il n'y faut ajouter qu'un point. Il n'y faut point ajouter*.

Another word needing the article to evoke a reality is personne; this is because of its qualitative, not its quantitative, deficiency. E.g.: Une personne est venue qui vous a demandé, Personne n'est venu, Personne ne vous a demandé4. The psycho-systematic process involved can be reconstituted. Une personne, given the weak designating power of the word, is anybody as realized by means of the article. Once the article is withdrawn, it is anybody in unreality, that is to say nobody, personne. Here we see the subtle play of quantities of movement, which is not totally unlike the play of quantities of movement observable in physics.

The article owes its realizing power to the fact that, although deprived of the substance5 of a noun, it has retained the movement of a noun. The noun is a compound of movement, quantity and form. This combination can be made clear by means of a diagram6 which includes all of its constituents:

1. = You have not advanced. 2. = You are not advancing a step. 3. In the first sentence, point expresses something small, in the

second, a negation. 4. = A person came and asked for you. Nobody came. Nobody asked

for you. 5. Throughout this discussion, Guillaume uses "substance" to desig

nate the meaning underlying the phonic or written sign, hence "matter" or "matter-substance" is the lexical meaning, "formal substance" the grammatical meaning.

6. Here and elsewhere, U stands for "universal", S for "singular".


The result — the nominal form — both proceeds and results from the form-substance which leads to it and which, in leading to it, determines it step by step through the grammatical notions it evokes. The nominal form results from this movement which proceeds toward the definition of a form and which consists of a series of grammatical notions acting in sequence.

On the other hand, these same grammatical notions belonging to the formal substance are each individually complementary to the nominal material substance, being its attributes: they declare its number, gender, etc. They can be complementary to the nominal material substance only if this material substance exists. If I withdraw it, their role as attributes of the nominal material substance will cease to exist and the formal substance will retain only its serial role as determinant of the nominal form.

A graph taking into account the two roles of the formal substance would show the formal substance as two different vector lines, one below the other: one vector line would show the accession to the noun category, and under it, another line would show the role of complementing the nominal material substance assigned to the notions of number, gender and function. For the formal substance, this would give the following diagram:

(1) sequential determination of the noun category

Form-Substance

(2) complementarity of the material substance


Thus, if I withdraw from the noun the matter-substance engendered from U1 to S1 in Tension I, the question of complementing this material substance no longer arises, and line (2) of the form-substance vanishes. Line (1) is thus freed from its relation to line (2), based on their containing identical grammatical indications. All that remains is line (1), which is then transferred and takes up a position as a category determiner outside the noun system, here reduced to its two tensions. This is how it becomes an article-noun. The internal composition of the article, with the constructive operations involved taken precisely into account, is therefore as follows:

Tension I Tension II U1 1S2 U2

(no mat ter -subs tance) (no form-substance complementary to the matter-substance — line (2) — but formal substance retained — line (1) — as a determiner of the nominal category)

In this way the article is constructed. Nothing is missing, and although it is deprived of material substance, we can see why it has the same formal appearance as the noun.

The formal substance of the noun is, in the substantive-noun, the sign of two things: a) the complementarity of the matter-substance, declared singular or plural, masculine or feminine; b) the directing of the word to the noun category (through which it becomes a noun). In the article-noun the formal substance of the noun, which is retained, is now the sign of only one of these: the second, the movement involved in the formal substance of the word as it accedes to the noun category. One has to have analysed this mechanism to really understand — the linguistics taught here is a seeing of comprehension — that the number and gender already indicated in the noun are indicated again in the article1. This is how agreement in number and gender is established between the two — not agreement between an adjective and a noun, but between a substantive supplied with material substance and a substantive (the article) deprived of it.

A number of grammarians — including Grevisse in Le Ban Usage — concerned with correctly classifying words into categories have made the error of regarding the article as a kind of adjective governed by the rules whereby the adjective agrees with the substantive. This is a mistake. The article is not adjectival with regard to the noun; it does not indicate any of its qualities. In fact, the agreement between article and noun is the reverse of that between adjective and noun

1. In French, of course.


because it is not the article which is said of the noun, but the noun which is said of the article. The feeling everyone has of this can be seen in the question what, asked after the article, when one has not understood the noun. The question what sometimes arises when the noun is an unusual one which the listener wishes to have repeated to make sure he has heard it correctly. For example, after I have said A chronothesis is the result of this, I have known people to reply: A what? One places the question what after the article when confronted with a noun one cannot accept.

Thus in the substantive/article relationship it is the substantive that should be described as functioning as an adjective with regard to the article. The substance of the article itself is one or the other of the two tensions; as such, a tension has neither number nor gender. It thus follows that the article declares a number and a gender that are not related to its own contents and that call for contents which it does not have — the contents belonging to the noun announced by the article.

This operative mechanism is noteworthy for its elegant economy. The number and gender of the article (the article itself containing no substance which can have a number and a gender) are, in discourse, in the sentence, an appeal for substance, a substance that the substantive-noun will provide.

(Lecture of March 7, 1957)

POTENTIALITY AND REALITY IN LANGUAGE

As soon as human language exists (the question of its origin is not under consideration here), it discovers two vocations; one, to become tongue, the other, to become discourse. It becomes tongue when, tending in the direction of the potential, it becomes institutionalized, thus freeing itself from the condition of being transitory; it becomes discourse when, tending in the direction of the actual, it occurs as non-institutionalized language, subject to the condition of being transitory. When not freed from this transitory condition, language, which is then discourse, is seen as an intermittent reality, alternatively present and absent, and in this case the basis of its representation is discontinuity. Freed from the transitory condition and hence made tongue, language is seen as a constantly present potentiality, never absent, and in this case the basis of its representation is continuity. In structural linguistics one of the most important conditions is transitoriness: discourse results when tran-sitoriness is allowed, tongue when it is forbidden.


While discontinuous discourse is present for variable intervals, but only at those sporadic moments marked by the occurrence of speech (whether outer and corporeal, or inner and non-corporeal), tongue, by contrast, dwells continually within thinking man, furnishing him with the means — no time restrictions here — to express, if he sees fit to do so, what he has momentarily conceived. In fact, at no time is it possible for the speaker to abrogate its presence.

Thanks to its inherent discontinuity, thanks to the contrast of presence and absence which characterizes it, discourse is willed activity in thinking man, activity produced outside what escapes natural control, although at the beginning perhaps just barely outside. On the contrary, thanks to its inherent continuity, which excludes the presence/absence contrast, tongue involves involuntary activity for thinking man, activity retained within that part of the mind that escapes natural control, although perhaps only barely within.

Be it tongue, situated prior to the threshold of consciousness, or discourse, situated beyond this threshold in human understanding, language is based on knowing the conditions of representation which, through continued creating over thousands of years, tend toward a superior state of explicitness attained historically by the languages most commonly studied. Studying them reveals this knowledge to be deliberately founded, through a constructive psychomechanism, on two representational postulates: the fundamention of notional ideation, and the fundamentum of structural ideation. In these languages, all highly evolved even when they first appeared historically, the two ideations — the notional, which is primordial, and the structural, which is transprimordial — are superimposed on each other. Hence it is impossible, in ancient Greek as in French, to evoke the notion "man", for example, without at the same time evoking the structure of substantive. Notional ideation involves a discriminating process which, however general the ideas conceived may be, brings out their singularity and attributes to them a proper name as it were; the ideation of structure involves a categorizing process which, once the discriminating process is transcended, is concerned with the ideas conceived only insofar as they imply something universal.

The superimposing of the two ideations, notional and structural, is obligatory and immediate in the Indo-European languages. It is also obligatory, although not absolutely immediate, in the Semitic languages, where a separative time, as short or as close to nothingness as desired, arises between the notional ideation signified by the consonant root and the structural ideation signified by the morphological vowels and, in a complementary way, by affixes. The root in Semitic languages cannot be spoken but it can be written. The historical and systemic significance of this remark needs no comment. It is not always necessary for the two ideations to be superimposed, and languages exist where notional ideation answers for everything, as in the case of Chinese. These languages are built simply on the basis of the primordial postulate of representation. One might well


consider the superimposing of structural ideation on notional ideation in numerous related and unrelated languages as a sort of algebra of linguistic representation which, like the algebra of mathematicians, would enable us to understand much while seeing little (Henri Poincaré).

The postulates of representation on which language is built have been identified and in our last lecture were diagrammed with lines and points. They are the outcome of a subtle and unceasing discussion of a relationship that is inseparable from a more general view of things, the relationship between size and form. Characterized by the fact that they exist in the mind prior to any undertaking of reason, these postulates of a representation — representation in language — are both closer to intuition and simpler than are, or can be, those postulates, with a greater load of implicit logic, upon which geometries of reasoning are founded....

(Report on the- lecture series for the year 1953-1954. Yearbook of the Ecole pratique des Hautes Etudes, 1955, pp. 53-55).

PART VI

THOUGHT AND LANGUAGE

LANGUAGE AND OPERATIONALISM

My listeners cannot help but have noticed — and I hope that it has not wearied them unduly — the uniformity of the method I use to analyse the grammatical facts of French. In my research the overriding principle, absolutely unchanging and monotonous, is that tongue is made up of results, and accounting for them is a matter of discovering the creative thought processes giving rise to them. In other words, the golden rule guiding our work here is to convert the observed result back into a process, a genetic process. Thus behind the substantive, which is a visible result in tongue, we have distinguished the necessary, prior process of substantivation; and behind the adjective, visible too and also considered a result, the process of adjectivization.

The substantial insights that we have achieved into the very delicate question of the relationships of position and meaning between adjective and substantive are all results of systematically applying the method of converting the result back into a process. This method of ours, having proved itself successful in solving many other problems, including some most important ones, inspires more and more confidence as our studies progress.

Sometimes, however, we fall into old habits; we try to examine the facts, the results, directly, without going to the often considerable trouble of analytically converting the result back into a process. We always end up regretting the wasted time and effort.

Everything in tongue, in fact, is a process. And the results that we observe are, if I may say so, illusions of sorts. There is no substantive; in tongue there is simply a substantivation that is arrested early or late in its movement. There is no adjective; rather, there is an adjectivization that has run a certain part of its course when the mind prehends it. There is no word; rather, there is an extraordinarily complicated genesis of the word, a lexigenesis.


There is no tense; rather, there is a phenomenon of formation of the time image — chronogenesis — which must be retraced if one is even to begin to understand the systematics of the French moods and tenses.

(Lecture of March 18, 1943, Series B)

THE OPERATIONAL SUBSTRATUM OF EVERY LANGUAGE SYSTEM

With these explanations, which review and elaborate on what has already been outlined, I particularly intended to illustrate, using very specific facts, two important constructional principles that you now know, namely:

a) that the mind has the potential, the capacity, of intercepting its own processes, of taking cross-sections of them;

b) that entities of tongue maintain the systematic character they get from their psycho-systematic origin.

These two principles governing the psychogeny of language entities are obviously of great importance for the researcher. In accounting for a linguistic phenomenon, nothing could be more precious than to know a priori, that such a phenomenon might well be nothing more than a simple vertical cross-section of an underlying phenomenon with its own sequence. By starting with this analytical principle, I managed to establish chronogenesis and chronothesis as the universal sources of the architecture of time in language. Chronogenesis — and I refer my listeners to my book Temps et Verbe — is the genesis of the potential to construct time, and chronothesis is the exploitation of this potential as it is being acquired. Every time such potential is exploited, the process of acquiring it is intercepted at a characteristic point and a cross-section taken. Each cross-section thus obtained constitutes a mood.

A mood sums up a certain potential already acquired during chronogenesis but which is perhaps not yet final. The tenses of the mood express the exploitation of this potential — to the extent that it has been acquired. In a diagram:

THOUGHT AND LANGUAGE 135

It is no less valuable to know a priori that an entity of tongue arising on a given level must necessarily always reflect that level and is incapable of shedding the characteristics acquired from that level.


LANGUAGE AND WHAT IS SAYABLE: FROM THE UNSAYABLE TO THE SAYABLE

We know what conditions of size and form must be met to satisfy the existential definition of language. In order to have a complete view, however, we must now look at the conditions of use to be met to satisfy the functional definition of language, the only definition linguists consider, as a rule. In other words, after establishing its ontogeny, which develops in historical time, we must now recognize its praxeogeny, which is formally identical to itself at any historical moment, although its constructed state can change because the ontogenetic space within which praxeogeny must actually function varies in extent according to the historical moment, the extent to which the ontogeny has progressed by transforming its virtuality into reality. From the existential point of view, language satisfies the same structural conditions as all non-finite beings, especially conditions of time. From the functional point of view, however, the structural conditions met by language are exclusively its own; the conditions envisaged here concern human language alone, the only sort of language linguistics deals with.

What are these conditions? Let us examine them in the order in which functional possibility is acquired. To carry out its function fully, language must produce three successive mutations within itself, the first making the others possible:

136 FOUNDATIONS FOR.A SCIENCE OF LANGUAGE

1. Mutation of the unsayable into something sayable; 2. Mutation of the sayable into saying; 3. Mutation of the saying into a something said.

First mutation: the unsayable becomes something sayable. Prior to this mutation, there is no language. The unsayable does not belong to language; it confronts language, but is not brought into being by language. Human language presupposes that the unsayable ➙ sayable mutation has been realized. Let me say right away that I am here talking about acquiring the characteristic of being sayable mentally, to which will be added that of being sayable orally and scripturally, and the relation between the two. It is worth remembering the existence in human language of these three manners of being sayable: the mental, the oral and the scriptural. It should also be noted that two of the three, the mental and the scriptural, are silent, which makes them in one respect congruent. Only one of the three, that of being orally sayable, is not silent (physical existence more visible: the preponderance of phonetics).

The distinction between being orally and scripturally sayable is of little interest in languages where the phoneticizing tendency has been fully exploited; in this way of making things sayable, writing is merely a transcription of speech. In non-morphogenetic languages, character-languages, the two coexist side by side, with the scripturally sayable having the advantage. In Chinese, the spoken sound betokens the written figure. As a result, written Chinese is understood by hundreds of millions of persons, whereas spoken Chinese, because of the instability of speech, is understood only by restricted groups.

The "characterology" of Leibnitz (writing ideas directly — a preoccupation for Leibnitz) would have been a form of writing read by all human beings — a universalized Chinese — which, moreover, would have had all the qualities of the word-languages we are used to. Leibnitz as it happens does not seem to have realized the full extent of the difficulty involved. We can see it just by listing the number of written signs needed to write the simple group des hommes as it is thought in our languages. It would require:

for the article -a reverser of extensity: de -a sign for extensity: le -a sign for the plural: s -a sign for gender: masculine -a sign for function: (the single case of French) -a sign for the whole (the part of speech) with the meaning of the article des;


for the substantive

-a sign for the notion (highly figurative) -signs of agreement with the article: (plural, gender, function, i.e. the same signs)

-a sign with the sense of the part of speech substantive-noun.

And still more: all this writing would have to be the product of a single act of categorization. Impossible. Chinese, with the charac-terology it has created, makes use of only one unvarying sign in these circumstances. It is, therefore, a language very different in type from our languages. The Chinese type of language, whose visible mark is non-morphogeny, is the perfectum of tongues which have not been able to escape — or perhaps more precisely, which have been able not to escape — from the initial area of language typology. In other words, this type represents the highest perfection of conserved prim-itiveness; it is the mark of a mentation which has carried the initial phase of language to its highest development instead of substituting for this initial phase the initial plus medial phases. Chinese represents a kind of peak of civilization in conserved primitivism. Although quite obviously a great achievement, it is nonetheless one that makes further development impossible, as we shall see later on.

Saussure and others before and after him say that language is a system, meaning that its internal arrangement is not arbitrary, that it is codified. They do not say what the system is. But this we now know. It is a system of three ways of being sayable — the mental, the oral and the scriptural — with this particularity: the orally sayable has tended to become nothing more than a notation of the mentally sayable, and the scripturally sayable nothing more than a notation of the orally sayable. The order of these three notations has best been established in those Indo-European languages with widespread intellectual traditions. Consequently, everything is, as it were, known of the structure of our languages once we have fathomed what is mentally sayable in them. This is the domain of tongue.

The term mentally sayable deserves a moment's pause. To be sayable by means of signs — to be semiologically (orally or scripturally) sayable — is possible only with regard to something thought, and not thought in just any way whatever: freely, but in a proper fashion so that a sign can be given to it. This kind of thinking makes something mentally sayable. The attribution of a sign judged to be suitable makes something orally or scripturally sayable, gives it semiol-ogical "sayableness" — psycho-semiological "sayableness" (the choice of a sign having its reasons, its mental motivation).

Let us now return to the first functional condition met by human language, the mutation of the unsayable into the sayable. We are all acquainted with it: we encounter it when we have to say something subtle, something hard to grasp in our minds. But in the structural history of language, the term unsayable applies to something other


than what is difficult to express; it applies to what was before the sayable existed. What was before was human experience, the experience man had from his presence in the physical universe. This experience, because of its vastness, because of its incoherent diversity, because of its internal multiplicity, was not representable and hence not sayable. It belonged to the unsayable. Making it sayable meant resolving its incoherent diversity into series that lead, that converge into one and the same representation. It is easy to produce examples. Take the experience t r e e : this is a repeated, diverse, incoherent experience which, when seriated, leads to the representation t r e e , to making tree mentally sayable, and once a suitable sign is found, to making it orally sayable.


STRUCTURAL STATES OF LANGUAGE AND THE HISTORY OF HUMAN UNDERSTANDING

There is a history yet to be written — the history of human thought, of the condition behind both its potential and the safeguarding of its potential. The history of the potential of human thought is faithfully mirrored in the structures of languages; it is found inscribed in these structures with hidden strokes that we must learn to decipher.

I shall finish today by adding that the structural history of language, represented by the historical sequence of the structural states of language, is the only document available for studying the history of the potential of human thought; it is the sole record of this unwritten history. This gives linguistics a special importance, a sort of preeminence in a fully developed history of the sciences. For all sciences are built upon the pre-science of lucidity, that entirely potential pre-science which constitutes the very structure of language. Although this pre-science provides no positive knowledge of the universe, it does provide a way of viewing the universe with a degree of penetration that has varied considerably throughout the long history of human thought (represented in the history of language structures). To avoid any misunderstanding, let me add that language structures do not expressly show the potential for thought possessed by men of a given place and age of humanity; rather, they show the level of potential beneath which the minds of those men could not fall, however crude they might be Individually. What language structures reveal is the civilization of the mind common to some human collectivity at a given period, the minimal conditions, what could not be less in that civilization. Indo-European understanding involves a minimum level of potential above which it can rise as far as can be imagined and determined, but beneath which it cannot descend.


Chinese understanding involves a level of understanding above which it too can rise as high as can be imagined, but beneath which it cannot descend. This impossibility for the understanding of a human collectivity to descend beneath the level it has reached historically involves a deep-seated hierarchy whose distinctive principle is not to keep men from climbing individually to the heights of thought, but to keep a whole human collectivity from falling beneath a certain level, rock-bottom, beneath which no further descent is possible. To know where one's rock-bottom level is in the history of human understanding, the question to ask is not: what shall I not climb up to? for which no answer is written into our document, the structural history of languages, but just the opposite: what shall I not, can I not descend to? In the history of human understanding, a civilization of the mind is the answer to the subtle question: quo non descendam? To see roughly what I mean, imagine that circumstances put us back into a primitive state: what primitive person, what sort of primitive person or barbarian would I become again essentially? I hasten to add here that these last considerations are outside the scope of my teaching, which will be concerned solely with discovering, through as powerful a comprehension as possible, the psychomechanisms of the language structures of each successive stage in the structural history of language.


THE HUMANIZING FUNCTION OF LANGUAGE: LINGUISTICS AND ANTHROPOLOGY

Linguistics had aroused a certain expectation: that of providing better knowledge of what language in itself is. It was expected that this knowledge would be found by working back toward origins. Going back in time showed that to each state of language there corresponds an earlier state, and that the subsequent state can be accounted for on the basis of the prior one.

But continuously passing from the subsequent to the prior does not throw any light on the question of the nature of human language. It could even be said that the further back one works, from subsequent to prior, using the means of historical observation, the more obscure the problem of the nature of language becomes. Working back from French to one of the Indo-European languages — the ordinary means of comparative grammar make this quite possible — tells us nothing pertinent about the nature of language. From this point of view, a specialist in Sanskrit knows no more than a person who specializes in French. It could even be maintained without any paradox that in one way it is easier to begin probing the nature of language through the study of French than through the study of Sanskrit.


This fact is striking enough to justify the conclusion that the systems of Sanskrit and French are different, and that the difference involves progress in systematization which becomes more and more explicit. Thus we can see that from a certain point of view, the general direction of language history — and this could well be examined more closely by historical linguistics — is for a coherent construction to develop from implicitness to an explicitness of a higher order.

This conclusion is not a very important find. But it is a find, and unimportant though it may be, the linguist seeking a good place to start his study will nevertheless profit from realizing that when he approaches origins by using the methods of historical observation he is faced with systems that are less developed inwardly, less marked or explicit than they will be later on.

The conclusion is implied in the process of working back from subsequent to prior: little by little, this retracing becomes, for the observer, who is observing his own construct — observing language, observing his observing — ...1 This conclusion is enhanced by another small step forward in the matter, namely the idea that the problem of language is made more and more explicit by the solutions found though the ages.

The problem is always the same. It is presented right at the beginning, but in a form which does not make it explicit. It exists, but lacks the required elements that would present it explicitly. And the history of language is a succession of provisional solutions to the problem presented, each one a more explicit presentation. In the final analysis the history of language — and more especially the history of its structure (that is, of its interior systematization) — is the history of the successive explicitations of the problem of language. English, German, French and Spanish are four parallel solutions to the problem of language, a problem common to the whole of humanity. We must remember, however, that these four languages also make the problem explicit, presenting it and calling for a new solution.

Moreover, the more successful the solution reached, the less urgent the call for new solutions is. In other words, the better the problem is presented, the greater the delay in starting a mental discussion of it, deep in the subconscious, which will cause it to be presented differently in the form of a newly worked-out solution.

I must add here that to progress in the way just outlined is simply to progress like any apprentice. I have profited from listening carefully to young locksmith apprentices speaking to one another: one of their problems is to learn how to file things level so as to have a flat surface, one the file will not slide off. There is a

1. In the manuscript the sentence seems to be incomplete.


question of dimension involved here. The apprentice learns through experience. For a given problem he offers a faulty solution. Because of its faults and shortcomings the solution itself poses the problem again.

This observation could easily provide the basis for explaining the rate of progress in apprenticeship. The apprentice always finds a temporary solution which makes the problem explicit enough for the definitive solution to be found. If the apprentice were a philosopher, he would be able to see clearly in his own mind — intellectually — this manner of progressing. But he does not do this because he thinks instinctively, more with his fingers and hands and the stance of his body than with his mind. And yet, as he works from experience his mind is at work, although its work is in the middle1: things guiding the mind more than the mind guiding things.

To a large extent, in this middle activity of the mind lies the future of humanity: this guiding of things by decision of the mind in turn guides the mind. But human development is also due to activities of pure reasoning carried over as directly as possible into experiential reality. Certainly bridges have been constructed by building bridges, but from this trial-and-error science of bridge-building has come a theoretical science of bridge-building, that of the engineer. Theory transcends trial-and-error.

Linguist-engineers do not exist, nor does a science for constructing language. Language is constructed by methods of trial-and-error correction: a more successful solution is tried in place of an already successful solution and the solution substituted always presents the problem anew.

I was saying at the beginning that the language problem constantly confronting man shifts from the less explicit to the more explicit. The historical observation of language yields proof of this increasing explicitness. At the outset there is a problem of viewing, of outlook. Man dwells in the universe and sees it with his physical eyes. But he does not see it humanly, he does not see it with a human outlook, unless he looks at it again, within himself.

We see the external universe only through the medium of the universe-view we carry in our minds. This medium is part and parcel of the human outlook. A properly human view of the universe is the outcome of our ability to deal with the universe within us. Within me, making up part of my inner universe, is the image man. To see a man, to see him as such, humanly, is to subject this image man, integrated into my mental universe, to a treatment which will make it the equivalent of an image belonging to the external universe. I grasp the image within my mind and bring it out to look at it, and this

1. To be understood grammatically: in the middle voice; cf. G. Guillaume, Langage et Science du langage, pp. 127 ff.


movement coincides with a realizing treatment. I never see anything but mental inwardness realized mentally.

If instead of this view of what is realized mentally — a view excluding any other — I had a direct view of the real, I would not be a human being. To do away with a human being's view of reality through the compulsory medium of the image of reality that he carries within him would be to do away with the human being, to descend from the human to the animal. To replace an animal's direct view of reality by a view which is the result of treating an image of reality carried within, would be to promote the animal to the condition of man, in other words, to deprive it of its immediate view of the universe and substitute a mediate view through the channel of a prior mental representation. Possessing one state entails losing the other, And the comparison of the two states supposes prior reasoning powers enabling one to conceive of it but not to experience or see it.

We have moved from the field of linguistics to that of philosophy, with advantages to both linguist and philosopher. The dialogue between linguist and philosopher must be credited with a number of reciprocal insights for both disciplines. The starting point of this dialogue is the institutionalized language, and the main point to be made is that language has, through its own means which are not those of philosophy, internally recognized and systematized relations which belong to philosophy, the universal/singular relation for example. But it must not be forgotten that such philosophical relationships are discovered in language through means which are more anthropological than philosophical. Therefore, prior to the dialogue between philosopher and linguist begins the dialogue between linguist and anthropologist, from which I think the linguist has more to gain than from his dialogue with the philosopher. In any case, one dialogue calls for the other; the dialogue between philosopher and linguist leads to the dialogue between anthropologist and linguist.

Anthropologists are interested in the development of man from the earliest of primitive states. One of their tasks is to mark the stages or phases of this development, and so to examine the physical documents that bear witness to it: monuments left behind, food, clothing, housing conditions, materials known, tools and other artifacts, engraved drawings, and the art discernable in these things.

By examining these documents the anthropologist can recognize the successive ages of civilization of material objects and also what these successive ages reflect concerning civilization of the mind:

a) in the moral sphere (beliefs, religions, superstitions — because here too successive ages belonging to mental civilization can be discerned);

b) in the physiological sphere, where observation turns to man's physical structure according to the evidence provided by bones preserved and dated by various means;


and finally — what has hardly been touched up to now — c) in the linguistic sphere (golden age of anthropology).

It is rather surprising that although many documents regarding the development of man have been examined, the incomparable document provided by the linguistic state — by the succession of linguistic states — has not been examined from the anthropological point of view, and that nobody has had the idea of distributing human development into linguistic ages. This gap is a serious one; to fill it, the matchless document of human language will have to be examined.

Only the linguist can carry out this examination. He alone knows the linguistic facts. His science consists in seeing them, seeing not only the facts which appearances reveal, but also those which appearances cover up and, in a way, hide. A linguistics is now emerging which will teach the anthropologist about the ages of the deep-seated mental civilization of man; this civilization is not represented explicitly by the dominant ideas of thinking man at different times and places, but rather by the lucidity he has acquired.

(Lecture of May 23, 1957)

LANGUAGE AS A REDUCER OF MENTAL TURBULENCE

Human language exists only when raw experience is transmuted into representation...

In all thought there is cogitation. Cogitation involves turbulence. Representation involves abolishing turbulence. To think well is to abolish the turbulence of cogitation. Mental mechanisms can regulate turbulence.

Writing, more than speech, obliterates the turbulence of cogitation. With regard to spelling reform: difficult spelling, with all its arbitrary elements, retains an educational value as an extinguisher of mental turbulence. I am not at all sure that this important process is the aim our educators have in mind.


At the outset it can be seen that human language — and subsequently the potential of human thought — is created when the turbulence of mental cogitation has been extinguished to a sufficient degree. Indeed, reducing turbulence is instinctive.

A study of the origins of language must be careful to take writing and its social preponderance over speech into account, that is, the social stature of the one who knows writing, the KATIB, and the reflection of this status on the book, the KITAB, what is written, which has retained something of its prestige, in a new form: what is printed. The basic thing is the escape from turbulence.

(Lecture of April 4, 1957)

THE SPECIFICITY OF HUMAN LUCIDITY

Thought unaware of the extent of its own potential construction would be turbulent thought, left in the mental turbulence of its origins — a turbulence which human thought, by its very nature, has done away with to a high degree.

With the uprise of human language, turbulent thought has been replaced by thought where turbulence is no longer inevitable, by a type of mental activity where it is inevitable to keep a certain distance from the original mental turbulence, to which man is no longer permitted to return.

The fact that he cannot return is what makes man man. His mind is denied the possibilities afforded by the original mental turbulence, whereas these quite considerable possibilities remain a privilege of the animal. How many things the animal knows differently from the way man knows them or is able to know them! Man cannot even know what sort of knowledge the animal has of them. Animal lucidity is based on the physical development of the species, and cannot transcend it. Weighing on animal lucidity is the impossibility of transcending the inherent finality of this development, which follows the paths of the evolution of life. Weighing on man, on human lucidity, is the impossibility of not transcending the finality of this development. In other words, human lucidity must be based on itself rather than on the physical development of the species.

In the universe a race is going on between the minimized lucidity of the animals — which the development of the species brings them— and the deminimized lucidity of humans, who seek their lucidity outside this development, transcending it, in a field where it no longer rules absolutely. Since animal lucidity depends on and cannot transcend the point reached by the species' physical evolution, the animal belongs body and soul, if I may say so, to the forces of life


in the universe, with regard to which it has no autonomy. Since human lucidity depends on transcending the degree of the species' physical evolution, then in proportion to this transcendence man no longer belongs to the forces of life in the universe. He has won a certain autonomy over them, and even opposition to them. It is an increasing, though not total, autonomy.

In this kind of lecture I would prefer to avoid such questions concerning the specificity of human thought, but they cannot be avoided. The science of language inevitably leads us back to them, because linguistics is knowledge, not of the physical universe within which man dwells and of which he is a part, but of a mental universe — tongue — that dwells within him. It is the victorious confrontation of tongue with the universe he inhabits that gives man his relative autonomy within it.

It would not be a misuse of words to speak of the outer universe and the inner universe, tongue, and of the unending drama of their continued collision. We could speak at some length of the physical results of this collision, and of its non-physical results reflected in the history of human language. We could speak of two very different spheres of progress: the progress constituted by the physical results, and the progress constituted by the non-physical, essentially linguistic results. We could also speak at some length of the different goals of each type of progress, and of the conflict produced when they meet. Unwittingly, literature, philosophy, and science make unceasing though implicit reference to the different goals of two types of progress.

(Leeture of January 8, 1959)

THE RECIPROCAL CAUSATION OF THOUGHT AND LANGUAGE

Thinking man constructs his own thought, which would not exist1 if he did not do so. Without language, man would not know — and it is essential for him to know — how far he had advanced in the construction of his own thought. From age to age, era to era, moment to moment, language provides man, the constructor of his own thought, with a view of what has been constructed so far.

This progress of a constructed work — whose constructor expects it to provide the very potential for continuing that construction — is what Delacroix, the philosopher, was referring to in his brilliant insight: "thought makes language while being made by language". Delacroix did not add any comment; I would like to do it for him.

1. That is, as specifically human thought.


It means that in what it has constructed the mind finds a view of what it has to undertake so that thought, in constructing itself, will not stop at the point language indicates it has reached.

Since the mind is informed by language of what is done and of what remains to be done in order to continue, the construction of thought within the mind is a process that is sure to continue. Language gives the mind the power to safeguard and to increase the power it has acquired, which is that of whatever constructed state it has reached. At the root of this operation, giving it its special character, is human lucidity, the lucidity proper to the human species.


THE COMMON MIND AND THE SCHOLARLY MIND

...After careful reflection, we find that the major laws of representation, which have given us the geometrical, mechanical and mathematical sciences, operated in tongue prior to being used elsewhere, and that if they had not been previously brought into operation in tongue, where the notions we think with are found, they would not have come into operation anywhere else. This brings us to the principle, which I have already stated more than once, that tongue is the pre-science of science. The loftiest speculations of science are built on the systematized representations of tongue.

The best starting point for a history of human thought is in the constructed states of language, which provide background possibilities:


The postulates of this pre-science tend to constitute their own axiomatics. Establishing this set of axioms means moving out of a state of turbulence.

Extinguishing the turbulence of thought is the basis for developing a language. There is a hierarchy here, and there are revolts against the hierarchy. Highly evolved tongues have a geometrical, mechanical beauty and less turbulence, and also, better means of evading turbulence. Beneath the non turbulent structures of reason there lies the turbulence of ideas: to stream the best of turbulent cogitation into structured non-turbulence. This characterizes highly evolved languages. Evolution means the advance towards the most successful conditions of representation, a hidden progress towards a greater underlying potentiality.

This progress is accomplished in our minds as we move away from the primitive state, a movement that we know to be a matter of occupying more and more of the developmental space which language, by vocation, is called to occupy. In other words, it means approaching the time when we can make the equation: historical development = development by vocation. But at present we can only write: historical development < development by vocation, though equality is not so far away. And perhaps we may already be justified in writing: historical development ≤ development by vocation (that is, is less than but almost equal to). But we must not make such statements rashly.

(Lecture of May 16, 195 7)

EPILOGUE

EPILOGUE

LANGUAGE AND SCIENTIFIC CURIOSITY

...Something resembling a search for complex conciliation — another word for harmony — between elements not flatly contradictory is to be discovered in these inner activities of the mind which are reflected in the structure of language and nowhere else. Studying these activities with this idea in mind is obviously more the work of philosophers than linguists, although linguists cannot ignore it completely. It is up to philosophers to bring out the general implications of the insights into structural linguistics as presented here. Often our views are closely related to questions now being discussed, rather subtly and from another angle, in the philosophy (general theory) of the physical sciences and, by extension, of mathematics. As has often been noted, mathematics arises from language and, from the point of view adopted in this essay, is a special form of scientific curiosity.

The starting point for scientific curiosity may be at different levels along the linguistic ascent from seeing to understanding. The particular form it takes depends on how high its starting point is in the ascent. If it starts at the lower levels, scientific curiosity is more involved with the reality perceived by the senses (a reality which can be observed) than with the reality perceived by reason (a reality which must be proved). If it starts from higher levels of the ascent, scientific curiosity has more to do with the conceived reality of reason than with the perceived reality of the senses; when it springs from the highest reaches of the linguistic ascent from seeing to understanding, scientific curiosity is wholly involved with conceived reality and proportionately withdrawn from perceived reality, which is considered insufficiently probative. Mathematics is scientific curiosity that deals exclusively (or almost exclusively) with the realities — the lofty realities — of reason.

Mathematics tends to seek proof by reason, and to consider this the only fully satisfactory method. The same tendency is found in theoretical science. This is why the theorizing sciences — physics


in particular — are constantly calling on mathematics either to justify through reason what experiment has established as fact, or to anticipate, guide and orient experimental discovery with views arising from the higher, mechanical type of reasoning found in calculus .

What the linguist and the philosopher should keep in mind is the link between scientific curiosity and the constructed state of the language, for in a given spatio-temporal area this state governs what scientific curiosity will actually be like at that point — that is, if we disregard anything that is borrowed and more or less assimilated. When the linguistic ascent from seeing to understanding has not reached a very high level in the commonly spoken idiom, scientific curiosity is also at a low level: not very discerning or penetrating. When the ascent in the commonly spoken idiom is a strong one, scientific curiosity is proportionately powerful. All of which is quite understandable if we remember that the linguistic ascent from seeing to understanding (the extent of which is expressed in the structure of tongue) generates lucidity.

Consequently, the bases of scientific discoveries will historically depend on the constructed state of the tongue naturally spoken by the people of a particular place and era. The state of mathematics expresses fairly exactly the sort of scientific curiosity provoked by a certain typological and structural state of tongue. There are scientific questions that the human mind would never have raised if the unquestioned pre-science of representation, tongue, had not already included the latent basis for them. This remark can perhaps throw light on the delicate and controversial question of intuition in mathematics: its source would appear to be linguistic.

The emergence of a given scientific curiosity definitely seems to depend on the constructed state of tongue, which contains it as a deep-seated virtuality. But it should be noted right away that this statement cannot be verified, because it is impossible to observe and compare the sciences that large groups of people speaking different languages would have produced if left to themselves (without any outside influence), working from the sole potential of their own linguistic lucidity. Borrowing ideas through travel, reading, cultural relations, borrowing artifacts and the ideas they bring with them, and all the ways of imitating the foreigner obscure and relax the functional relationship between linguistic structure and the scientific curiosity latent in that structure, and constituting in and through the structure a source of constructive imagination. Judging, however, by the works of the known history of humanity, this relationship does not seem to be overly conjectural.

(From an unpublished, unfinished essay entitled Prolégomènes à la linguistique structurale, 1954)

EPILOGUE 153

LANGUAGE, MATHEMATICS AND LINGUISTICS

Problems of representation solved in tongue sometimes lead to intermittent problems of expression solved in discourse under conditions set by the problems of representation settled in tongue. Since the two types of problem are essentially different, they should, in any good method, be studied separately. We represent in order to represent; we express in order to express. Between expression and representation the one and only link is the fact that representation supplies expression with the means it can use.

Throughout the different branches of linguistics a stubbornly persistent error is found in the failure to recognize that language solves problems of two different orders: problems of representation, solved outside discourse in tongue, before use, and problems of expression, solved outside tongue in discourse, which is use of pre-established tongue. Tongue is not formed while one speaks but while one does not speak, in the silence of a mind in constant and unconscious search of an ever deeper knowledge of itself; tongue depends far less on the strictly social relation between man and man than on the extrasocial relation, rooted in the infinite, between man and the universe within which man exists and affirms his power and his relative and growing autonomy. As I have already pointed out and demonstrated, the acts of representation making up tongue are far more an analysis of the great man-universe confrontation than the smaller man-man confrontation.

The simple fact of separating, for the needs of analysis, acts of representation making up tongue, from acts of expression giving rise to discourse, completely renews the science of language if the consequences of this separation are exploited. Avenues are opened up for linguistics to become, in every sense of the word, a science. This scientific character is brought out very clearly in the study of acts of representation, the acts which give the human mind its power. This power is based on deep-seated mental operations which are necessary and not contingent; without them it would not exist in its present state.

These deep-seated operations make possible both the acts of representation of tongue, and the human power of thought, which depends on their being carried out. Taking them as a whole, I have given them the general name intuitional mechanics and have tried to bring out their subtle yet simple rigour in my teaching. Their rigour — the fact that in them a step just made governs the following step — recalls the equal rigour of mathematics and immediately suggests the possibility of applying mathematics to linguistics. In the light of experience and further reflection, however, this idea seems to be neither sound nor well-founded, since the mathematics involved in intuitional mechanics is not a conscious use of the discursive faculties of the human mind, whose workings are the basis of learned


mathematics, but rather an unconscious use of the primordial faculties of representation. The discursive faculties of the human mind are founded on these primordial faculties; they arise from them and constitute, on a conscious level, their result. Creators of the conscious, the faculties of representation exist before consciousness and hence are not under its control for it is not yet set in motion when the above faculties are in operation.

Thus as a conscious activity of the mind bringing the discursive faculties into play, mathematics seems to leave behind it in the unconscious the operations of the intuitional mechanics which are produced by the primordial faculties of representation. Therefore, if in a certain profound sense the intuitional mechanics explains mathematics, the converse is not at all true. The sui generis mathematics of intuitional mechanics is elementary mathematics, as profound as it is subtle; in the potential chronology of the human mind it comes before the reasonings of arithmeticians, algebraists, geometers, and analysts, and, far from being explained by their reasonings, would rather appear to explain their own deep-seated sources. What is involved here is a sort of pre-mathematics of representation which is over when discursive and measuring mathematics arises; the first step is for arithmetical number (1,2,3, etc.) to take over from linguistic number (plural, ..., trial, dual, singular). Arithmetical number is a number for computing, while representational number resides in the mind beneath the faculty of computation, which is discursive in essence.

Linguistics is on the wrong track and can lead one quite far astray when it attempts — as has recently been done — to apply mathematics in the form it has acquired as a discursive discipline to the science of language. To apply mathematics adequately to the science of language would require linguistics to be completely reworked as a sort of pre-mathematics showing how the faculties of representation, through being exercised, give rise to the discursive faculties of reasoning: the former in thinking man are forerunners of the latter. Before undertaking this rather difficult task it would be better to catch the operations of the intuitional mechanics in the act through straightforward means of observation and analysis: to observe their function and sequence in languages whose development has made these operations clear enough for observation of them to be useful.

(Prolégomènes, 1954)

EPILOGUE 155

LANGUAGE, MAN AND THE UNIVERSE

Entirely founded on the space-time dichotomy, tongue (that is, what is represented or sayable, along with its associated mental signs) constitutes a continually expanding idea-universe in thinking man. Among thinking beings, man alone can delineate for himself an idea-universe whose form is renewed from age to age. Its variation represents the progress of civilization at the most profound level of man. The civilization of an era is the idea-universe that thinking men of that era have managed to delineate and establish deep in their minds. This idea-universe is tongue, present in men in each instant of live consciousness without any possible lapse.

Like man himself, who through the eyes of the body and the eyes of the mind is a spectator and observer, this idea-universe — tongue in thinking man — is a viewing universe made up of notions, each one the outcome of a ponderation of particularizing intention (that is, comprehension) and generalizing extension. Fixed in tongue — in what is represented but not yet expressed — before discourse comes into play, any vocable, any unit of potentiality is a stable, unchanging balance of the contractive movement of intension and the expansive movement of extension. This is the case for the word homme, which includes a fixed relation between intension (= comprehension) and extension, differing from the equally fixed relation between intention and extension on which the mental construction of the word animal, for example, is based.

The invariance of the relation between intension and extension in every vocable whatever its kind is a fact established in tongue. Once tongue is left behind, this relation is complicated by another sort of variation: the variation of extensity (or better still, of "tensity") which is involved in every vocable as an eventuality of discourse alone, but not of tongue, where it does not arise. The variation of extensity allows for understanding a word like homme in a narrow sense, intensively, or in a wider sense, more extensively. Furthermore, it permits this in two different ways: the variation of extensity permits a wide sense of the word at the very beginning of the contractive movement and also another wide sense toward the end of the expansive movement. The resulting nuances are those which differentiate the two examples I have already quoted a number of times because they illustrate the difference well: Un homme est un homme1

and L'homme est l'homme2. Once the extensity is determined, and this

1. = A man is a man. The article un involves a contractive movement which is here intercepted at its beginning to give the sense of a universal. (Cf.above, p. 19 the diagram of the category of the article).

2. = Man is man. The article le involves an expansive movement which is here intercepted at its end to give the sense of a universal as well, but with a different nuance.


is the point to be emphasized here, the word homme is no longer a viewing idea of tongue but a viewed idea of discourse. It has changed horizons in the mind.

Since the variation of extensity involves only discourse and not tongue, the ideas linked to signs that make up tongue (a word presupposes an idea which has annexed a sign to itself on a lasting basis) are all viewing ideas and together make up an idea-universe which is entirely viewing.

In their chapters on the noun, grammars used in teaching — the best and the oldest — usually bring out the correlation between comprehension and extension, pointing out, for example, that the difference between dog and animal is that dog in itself has more comprehension and less extension and animal has less comprehension and more extension. It is regrettable that these grammars have not brought out the variation of extensity, which is characterized by its complete indifference to the comprehension of the word: the variation of extensity makes no change in the comprehension; all it does is widen or narrow its field of application. The field of application is wide if we say L'homme est mortel1, narrow if we say UN homme entra or L'homme s'assit2, wide at the beginning of a movement going from wide to narrow, as in Un homme est un homme or in Un enfant est toujours l'ouvrage de sa mère3, and wide again, but differently and somehow more powerfully, when the reverse movement is caught far along in its course from narrow to wide, as in L'homme est l'homme or in L'enfant a droit à la protection de la société4.

The property of extensity is assigned to the noun quite late. If it is withdrawn from the noun by deflection and designated on its own, a grammatical word is constructed. Its express duty will be to designate, insofar as form and size are concerned, the extensity of a noun yet to be expressed. This word also will bear, in a complementary fashion as a result of deflection continued even further, the structural attributes of the noun foreseen but not yet expressed. This word is the article...

...The part of speech grammarians call the article is the outcome of the series of psycho-mechanical operations just described. Their constructional logic is such that the way they depend on and link up with one another could quite easily be represented in diagrams .

Clearly, one of the tasks in structural linguistics and its two auxiliaries, psycho-systematics and psycho-semiology, is to extend as far as possible our knowledge of the acts of representation making up

1. = Man is mortal, 2. = A man came in. The man sat down. 3. = A man is a man. A child is always the product of his mother. 4. = Man is man. The child has a right to be protected by society.

EPILOGUE 157

tongue; the article is one of the most interesting of them. Another, very different task in structural linguistics is that of accounting for the existence in thinking man of an expanding idea-universe, destined to grow in quantity and quality, an inner universe that he alone of all thinking beings is capable of building up within himself. This idea-universe that the human mind interiorizes and infolds is tongue. Animals have no tongue since nature has refused them the faculty of delineating an idea-universe within themselves; they have only discourse-language, and their discourse, with no mediation of representations, proceeds directly from experience.

We should see this difference, which makes man an exception among thinking beings, as the consequence of a measurement existing in man alone: a measurement of the relationship of independence between the thinking being and the universe within which he lives and which, insofar as he thinks, he is aware of being in. Although in animal thought this relation of independence cannot be measured, it does present itself in human thought as a measurable object whose finesse has been measured. Its measurability throughout the different ages of humanity, as a sort of substratum, conditions and determines the uniqueness and the power of human thought.

Constantly variable, this measurement of the autonomy of the human person with regard to the universe — whose forces the human person confronts with nothing but his own forces — is the great mental fact which enables thinking man to be what he is and is becoming within the universe, the place of his existence. It also enables him to delineate and possess his viewing idea-universe, tongue, an idea-universe for which he, thinking man, is the place of existence. The basic contrast involves an inverse relationship: that between a physical universe, man's place of existence, and man, the place of existence of an opposed mental (non-physical) universe.

The main lines of the psycho-mechanism of this incessant and unconscious contrast — unconscious because consciousness is born of it and not it of consciousness — can now be outlined. Between man and the universe the relationship is one of belonging. Its extreme, strictly theoretical forms appear to be:

a) man belonging entirely to the universe and, correlatively, the universe not at all belonging to man, whence man completely subjected to the forces of the universe, their toy;

b) the universe belonging entirely to man, and correlatively, man not at all belonging to the universe, whence the forces of the universe subjected to man's forces — to an infinite and absolute human understanding.

It is between these two extreme, purely theoretical, forms that the real form of the relation between man and universe is established. In reality, man's subjection to the universe was extreme at the


beginning, though not quite absolute. Down through the ages man has been freeing himself, because he knows how to measure the relationship. (This is his privilege.) Whatever irrevocable autonomy in relation to the universe has resulted from this for the people of a given spatio-temporal area can be identified with the most virtual and fundamental aspects of the civilization of that area.

The human mind is like a three-dimensional picture: on its nearest horizon is what is viewed in thought, and on the more distant horizon is what is used for viewing thought, though not yet actually viewed itself. Discourse, and the sentences that compose it, and the words in these sentences belong to the horizon of what is viewed. The words that make up tongue belong to the viewing horizon; their appointed role in tongue is to await the transitional treatment which from time to time will carry them from their permanent place on the viewing horizon to a momentary place on the viewed horizon.

For the substantive, this transitional treatment involves adding to the necessary and sufficient conditions of comprehension and extension met by the word in tongue a complementary condition of exten-sity that must be met by the word in discourse. Expressed in a formula, the mutation is as follows:

1. viewing idea of tongue = comprehension + extension (fixed relationship);

2. viewed idea of discourse = the same fixed relationship between comprehension and extension + extensity varying to suit the end sought in discourse, variable by its very nature.

Ideas that may be viewed as the occasion arises are far more numerous than viewing ideas which are the only ones in language with the right to a representation of their very own. From a single viewing idea such as homme , very different viewed ideas can be obtained by varying the extensity — as widely different as homme with the article un, and homme with the article le in: UN HOMME qui fait honneur à l'HOMME1. From a single viewing idea, the little word le, one can obtain ideas as widely different as le in L'homme s'endormit and le in L'homme est mortel2.

Thus we can see that in order to lighten the load of ideas carried permanently — whether there be discourse or not — the human mind has reduced tongue to viewing ideas only, and arrives at the viewed ideas whereby it depicts the real universe by applying an actualizing treatment to the viewing ideas contained in its (wholly viewing) idea-universe.

1. = A man who is an honour to man. 2. A similar example for the definite article in English would be

The horse is in the pasture and The horse is a quadruped.

EPILOGUE 159

The elegant economy of this linguistic mechanism and the hints it gives as to the nature of tongue and the nature of the human mind prompt one to examine all the particularities of its workings — a vast field of fruitful study for structuralists.

The transition from the viewing noun, not burdened with the unstable condition of extensity but only with the stable and inversely proportional conditions of comprehension and extension, is a transition from the viewing noun, the potential noun, to the viewed, actual noun. The transition is implicit in tongues where this ex-tensity, useful by what it adds, has no sign of its own and so is conferred on the noun by simply being assigned. It is made explicit in languages where the complementarily useful extensity, having its own sign, the article, is conferred on the noun by means of this part of speech.

In languages where it has become established, the article is the sign for the transition from potential noun to actual noun. I defined it, incompletely but accurately, in these terms in my first publication on language, a study of the article1.

Language is a social phenomenon. This is an undeniable fact. What better means than language could there be for men, brought together by their proximity in space, to escape from their individual solitude and to tighten the bond uniting them by extending it from the physical to the mental? Language enables men (in whom it has been lastingly established in the stable form of tongue) to communicate ideas and feelings of all kinds to one another. But in its human form of discourse based on a tongue, does language have just this one aim? To see in language nothing more than a means, the best means, of bringing man into relationship with man, is this not to belittle it and miss the essential? Is it really in this intermittent relationship that language comes into being, takes form? At first sight these questions may seem paradoxical, and yet they must be asked the moment we do not accept what is really visible as what is really true.

For people possessing language in common (with the same notional and structural ideation), it provides a useful bridge between those with things to say. What they have to say does not expressly concern their relationship within the society to which they belong, but rather a relationship of a very different sort: the relation of each and every one of them to the universe, the place of their existence. Only about this relationship can men converse, this relationship which is the substratum of all others, even the direct social relationship. They cannot escape from it. In the last analysis, there is no other possible subject of converse between them.

1. Cf. G. Guillaume, Le problème de l ' a r t i c l e et sa solution dans la langue f r a n ç a i s e , Paris, Hachette, 1919 (republished, 1975).


As perspicacious philosophers have seen, human language is based not on the small Man-Man confrontation, but on the great Universe-Man confrontation. This is its source. The structure of tongue bears undeniable evidence of this, evidence already known to the reader. Indeed, the further the language being observed has progressed in its ascent from seeing to understanding, the more visible the evidence is...

The commonplace view that tongue and language are social phenomena is one of those simplistic, insufficiently scrutinized views which have been quite detrimental to the progress of structural linguistics; it has focused the attention of researchers on the Man-Man relationship — which has contributed little to the structure of tongue — and drawn it away from the Universe-Man relationship — which has contributed everything or nearly everything to it. Besides, whatever the Man-Man relationship has contributed is integrated into the Universe-Man relationship, from which tongue, by definition a viewing idea-universe, cannot escape.

Proof of this integration is the fact that behind what is made explicit in the first person (speaker) and second person (spoken to), whose relationship derives directly from the Man-Man relationship, there is left implicit — and this grammarians fail to mention — a third person, spoken about, which, like the noun and the third person pronoun substituting for the noun, depends on the Universe-Man relationship, which includes the entire content of tongue. The first person is the one who speaks; in speaking he speaks of himself, thereby constituting not only the speaking first person but also, by synapsis, a third person spoken about; this third person can be a more or less extensive being belonging to the universe or even, at the limit of extension and extensity, the universe itself. The second person is the one spoken to, who is also being spoken about; hence besides being the second person spoken to it is, by synapsis, a third person spoken about. When explicitly signified by itself outside of any synapsis of rank, the third person is the one spoken about, which, being passive in the exchange, neither speaks nor is spoken to.

The persistence of an implicit third person beneath the explicit persons of a higher rank (first and second) is an outcome of the un-severable link between the small Man-Man confrontation and the great Universe-Man confrontation. The former emerges from the latter but does not break away from it.

Through the use men make of language as a means of exteriorizing and communicating their thoughts and feelings, tongue can be seen as a social phenomenon; but unless it is seen as an essentially human — and therefore extra-social — phenomenon in man silently thinking, not speaking (taken up not with his intermittent relation with his fellows, but with his never ceasing, continuous relation with the universe) any possibility of conceiving (and subsequently discerning) its structure is excluded. Indeed, the whole structure of tongue is

EPILOGUE 161

the outcome — and structural linguistics will have to make room for this idea — not of the meeting of man with man, but of the eternal confrontation of man and universe, and of the specifically human conditions of their confrontation which tongue mirrors, so to speak, in its structure.

(Prolégomènes, 1954)

LANGUAGE AND COMMON THOUGHT

Universal and singular... The architecture of language is universally based on this relationship. The concrete root of this relationship is the universe-man relation; its anti-concrete root, opposite to the concrete is the finite-infinite relation. In a diagram:

A common property of the universe and the infinite is the absence of exteriority: there is interiority and nothing else. The idea defies our understanding. The human mind, author of the operation, encounters the universal-singular relationship, but what sort of mind? That of the scholar certainly, late in the history of the human mind, but also, earlier on, the mind of the non-scholar, of the common man. With no need to reflect: straight off, inevitably, intuitively, unconsciously, with one of those operations of the intuitional mechanics which forge the tongues of men, which are the means of constructing language.

A sure basis for this intuitional mechanics, very close to the concrete (through observation, in the sphere of the quantitative) would be the man-universe relationship, or far removed from it (in the sphere of the qualitative), the singular-universal relationship. More explicitly, the universe (infinite) has an interiority which is total, and an exteriority which is nil; man (finite) has an interiority which is not total1 and an exteriority which is not nil. From this starting point, or from one of its likenesses (there have been three representations so far)2, the human mind begins its own circuit, which gives it its exercisable potential (its power) of thought. The human mind undertakes the construction both of the power of thought

1. That is, an interiority which does not contain everything, as opposed to that of the universe containing absolutely everything.

2. Universe-man, universal-singular, infinite-finite.


and of language — the two progress together — under conditions of u n t h i n k a b l e n e s s which have been averted and conditions of thinkable-ness which have been created. These conditions are:

-for the unthinkable: total interiority (all) and nil exteriority (nothing)

-for the thinkable: interiority not total and not nil and exteriority not nil and not total.

Fundamental for constructing the power of thought is the mind's operation of grasping thought. Without this prehension, thought does not exist for the mind1...

It is through language that the mind can gauge the circuit it runs, with varying speed, within itself —where it stands with regard to the power of prehending its own potential within itself.

Human language has an anthropogenetic aspect. It tells the human mind the point it has reached in its natural course within itself. In order for the mind to be operative, the idea of the infinite deprived of exteriority is overlaid with the idea of a finitude having exteriority. This is always done by means of the same mechanism: the infinite is conceived of as a potential for mental construction that encounters a center of inversion within itself. On both sides, right and left, of the center of inversion the infinite is a going beyond, an exteriority. The whole constructive mechanics of language is the movement of thought — the interiority of the mind's movement — encountering a centre of inversion within itself. In a diagram, it may be represented as something like this :

Let us take a closer look. A universalizing movement of the type U1 U2 (interiority alone with no exteriority) is unthinkable. Exteriority must be reinstated, by inserting interiority, so that there may be a contrast between an infinitely containing exteriority and an infinitely contained interiority. Once the mind sets out on its circuit, two images present themselves:

-a line (endless), the image of something infinitely containing,

-a point, the image of something infinitely contained.

1. A page of the manuscript is missing here.

EPILOGUE 163

In other words, to safeguard its potential the mind revokes the condition of unthinkability U1 U2, having found the necessary mechanism to do so; this revoking leads to U2. Here S is minimized interiority and are maximized exteriorities of S, that is, two exteriorities which are both exterior to the centre of inversion.

It should be noted that because of their extraordinary simplicity these operations became an essential condition for the common thought of man. These operations of common thought are very "scholarly", yet have nothing "scholarly" about them. I believe they are inevitable in human thought, in my human thinking faced with the unthinkable (U1 U 2 ) . From this unthinkable my thought must invent the thinkable

(U1 S U2)

using the only means available to it. This symbolization is frequently used in psycho-mechanics:

U1 S U2

Time will tell its real value. It is very convenient in linguistics. Elsewhere it might be:

which I would find satisfying enough because it would bring out the image — the intuitive image — of the lengthwise (longitudinal)

and of the crosswise (latitudinal)1.

For the longitudinal, the important thing to keep in mind is the varied use of the basic form

1. Lengthwise longitudinal here is synonymous with operative and crosswise latitudinal with resultative.


applied as a means of solving all the difficulties the mind has encountered in trying to safeguard its potential and in the related task of constituting language, its means of safeguarding this potential. In this way similarities in language are discovered (by the seeing of comprehension) by reapplying the basic form to many different cases. These similarities are delineated in language and the resulting configuration is the grammar of the language and something more. In this richly diverse diachronic delineation the universal-singular relationship is always found: there is never any place greater than the universal and never any place smaller than the singular, within (in the centre of) the universal. All of this reduces the mechanism to the elementary task of finding the right place.

...One remark here: at the root we find the universe-man relation. Man is the smallest place of all, the seat of a specific type of thought, human thought (which is not animal thought); universe is the seat of a type of thought which, if man is to.understand it, it must be the formal equivalent on a very large scale of what human thought is on its own very small scale. There must be a difference in size, not in form. Deep in the human mind, underneath all its speculations, the thought of the universe is assimilated to the thought of man. The subject is vast. Reduced to the case of language, it becomes a relation between universalization and singular-ization...

...In language the relation U-S-U, which is always the same, can be configured in many ways. The intuitional mechanics is the subtle varying of this invariance from within. The standard formula U-S-U is basically concrete with this form:

Next comes the man-man relation, discourse. Since discourse begins where tongue leaves off, for the whole of language we have:

EPILOGUE 165

This is the realistic way of viewing language. Why? Because that is just what language is, if the whole of it is viewed: it is a finalistic piece of work being constructed to safeguard the power of thought, and complernentarily, to allow man to communicate with man through the means offered by language. But using it as a means is secondary because the real object of the science of language is how this means is constructed. There is the constructing and the constructed stage; fortunately, the former (the constructing) can be found in the latter (the constructed state). The constructing goes ahead by itself at a good pace and then slows down thus allowing it to be refracted so that what has been constructed up to that point in the process may be deployed. This deployment of the constructive process at some point within itself — made possible by the process slowing down — is what gives rise to the constructed state of language. Thus, an image of the constructive process can fortunately be found in the constructed state. We must be grateful for this because if it were not so, there would be no men speaking human languages and no linguists.

Note carefully — I insist on this point — that throughout this discussion I have been not a philosopher, nor a logician, nor a psychologist, but a man, a man thinking in common with any other man (le tout venant humain1 as a learned Latinist once said to me). I have been a man using common thought — really common to all — to grapple with the problem, also quite common, of the place that thinking man is (a narrow space — greater spaces exist); of the place that thinking man has (ultimately, the universe — n o greater space exists, and it is impossible to go outside it); and, correlatively, of the constructing of an inner universe, tongue, within him which he uses to assert his autonomy in the outer universe; and lastly, of the use made of this inner universe, that is, of discourse with his fellow men (who are in the outer universe).

These results of observation and analysis are quite clear, but they would not be so if I had been a philosopher, logician, psychologist or linguist in the traditional sense. To discover these simple insights I had to be — to make myself — the anybody of common human thinking. The problems tackled here are unavoidable in human thought. To avoid them, one would have to descend beneath the level of human thought. Once homo toquens came to exist as such in historical time these problems became unavoidable. And because it is necessary for him to know about them, he must solve them and, in order to solve them, give himself an image of their solution. This image is language.

(Lecture of May 29, 1958)

1. Roughly: all comers, everyman.

SELECTED BIBLIOGRAPHY

PRINCIPAL PUBLICATIONS OF GUSTAVE GUILLAUME

PUBLISHED BY THE AUTHOR

1911 Études de grammaire logique comparée. Les passés de l'indicatif français , allemands et russes, Paris, Fischbacher, 84 p.

1912 Études de grammaire française logique. Le lieu du mode dans le temps, dans l'espace. Fascicule 1 : L'article, Paris, Fisch-bacher, 115 p.

1913 Études de grammaire française logique. Le lieu du mode dans le temps, dans l'espace. Fascicule 2 : Les temps, Paris, Fisch-bacher, 136 p.

1917/1975 Le problème de l'article et sa solution dans la langue française, Paris, Hachette, 318 p. Prix Volney (1917). [Reissue: Paris-Québec, A.-G. Nizet-Presses de 1'Université Laval, 1975, XVI-318 p.]

1929/1965 Temps et Verbe. Théorie des aspects, des modes et des temps, Paris, H. Champion, 1929, 134 p. Prix Volney (1931). [Reissued with L 'Architectonique du temps dans les langues classiques : Paris, H. Champion, 1965.]

1942-43/1945/1965 L'Architectonique du temps dans les langues classiques, Acta Linguistica, 3, 2-3, pp. 69-118. [Reissued in a separate volume: Copenhague, Munksgaard, 1945, 66 p. and in one volume with Temps et Verbe. ]

POSTHUMOUS PUBLICATIONS

1964 Langage et science du langage, Québec-Paris, Presses de 1'Université Laval-A.-G. Nizet, 287 p. [Posthumous collection of articles published between 1933 and 1958.]

168 BIBLIOGRAPHY

1971 Leçons de linguistique, 1948-1949, Série A. Structure sémio-logique et structure psychique de la langue française 1, published by R. VALIN, Québec-Paris, Presses de l'Université Laval-Klinck-sieck, 269 p.

1971 Leçons de linguistique, 1948-1949, Série B. Psycho-systématique du langage : Principes, méthodes et applications 1, published by R. VALIN, Québec-Paris, Presses de l'Université Laval-Klincksieck, 222 p.

1973 Leçons de linguistique, 1948-1949, Série C. Grammaire particulière du français et grammaire générale 4, published by R. VALIN, Québec-Paris, Presses de l'Université Laval-Klincksieck, 256 p.

1973 Principes de linguistique théorique, Collection of unpublished texts, editor - R. VALIN, Québec-Paris, Presses de l'Université Laval-Klincksieck, 279 p.

1974 Leçons de linguistique, 1949-1950, Série A. Structure sémio-logique et structure psychique de la langue française 2, published by R. VALIN, Québec-Paris, Presses de l'Université Laval-Klincksieck, 223 p.

1982 Leçons de linguistique, 1956-195?'. Systèmes linguistiques et successivité historique des systèmes II, published by Roch VALIN, Walter HIRTLE and André JOLY, Québec-Lille, Presses de l'Université Laval-Presses de l'Université de Lille, 311 p.

YET TO BE PUBLISHED

Various essays, among which the Essai de mécanique intuitionnelle. Espace et temps en pensée commune et dans les structures de langue and the Prolégomènes à la linguistique structurale. Discussion et continuation de la théorie saussurienne de la diachronie et de la synchronie.

Some thirty volumes in the Leçons de linguistique series. For details, see Bibliographie de la recherche en psycho-systématique du langage (H. CURAT and L. MENEY, Presses de l'Université Laval, 1983).

BIBLIOGRAPHY 169

FURTHER READINGS

IN ENGLISH FRASER, Thomas, and André JOLY. 1975. Studies in English, Grammar,

eds., Paris, Ed. Universitaires.

GUIRAUD, Pierre. 1970. "Gustave Guillaume and Generative Grammar" Language Sciences, 10, pp. 1-6.

HEWSON, John. 1972. Article and Noun in English, The Hague-Paris, Mouton, Janua Linguarum, Series Practica, 104.

1972. "The Essential Guillaume: a Critical Explication", Linguistics, 92, pp. 13-27.

1974. Review of Guillaume, Leçons de linguistique 1948-9, vols 1 and 2, Linguistics 131:91-102.

1976. "Langue and parole since Saussure", H i s tor iograch ia Linguistica, 3, 3, pp. 315-348.

1981. "The Guillaumian Tradition in Canadian Linguistics", Canadian Journal of Linguistics, 26, 1981, pp. 161-170.

Forthcoming. Fundamentals for a Word Based Syntax.

HIRTLE, Walter. 1967. The Simple and Progressive Forms. An Analytical Approach, Cahiers de Psychomécanique du Langage, Québec, Presses de 1'Université Laval.

1975. Time, Aspect and the Verb, Cahiers de Psychomécanique du Langage, Québec, Presses de l'Université Laval.

1977. "Already, Still and Yet", Archivum Linguisticum, 8, pp. 28-45.

1982. Number and Inner Space. A Study of Grammatical Number in English, Cahiers de Psychomécanique du Langage, Québec, Presses de l'Université Laval.

JOLY, André. 1967. Negation and the Comparative Particle in English, Cahiers de Psychomécanique du Langage, Québec, Presses de l'Université Laval.

170 BIBLIOGRAPHY

1975. "A Psychomechanical Approach to the Verb Do" , in S. de Vriendt, et al., Grammaire generative et psychoméconique du langage, Bruxelles-Paris, A.I.M.V.-Didier, pp. 123-128.

JONES, R.M. 1970. System in Child Language, Cardiff, University of Wales Press.

1976. "The Article in Welsh", Studia Celtica, 10-11, pp. 326-344.

POSNER, Rebecca. 1974. "Can Guillaume Teach us a Lesson?", Archivum Linguisticum, 5, pp. 83-98.

TEYSSIER, J. 1968. "Notes on the Syntax of the Adjective in Modern English", Lingua, 20, 3, pp. 229-249.

Anonymous. 1972. "Guillaume and the Guillaumeans", Times Literary Supplement, September 1, p. 1030.

IN FRENCH

VALIN, R. 1964. La méthode comparative en linguistique historique et en psychomécanique du langage, Cahiers de Psychomécanique du Langage, Québec, Presses de l'Université Laval.

1981. Perspectives psychomécaniques sur la syntaxe, Cahiers de Psychomécanique du Langage, Québec, Presses de l'Université Laval.

WILMET, Marc. 1978. Gustave Guillaume et son école linguistique, P a r i s - B r u x e l l e s , Nathan-Labor, Langues e t Cul ture , 12.

INDEX

accomplished (accompli): 97. act of expression: 5-6, 89. act of language (acte de

langage): 17, 65, 79, 82-87. act of representation: 6, 89. actual: 4, 128, 159. actuality: 4, 82, 88. actualization: 5. adjective: 119-122, 133. algebra: 95. animal: 95-96. aorist: 75. aphasia: 96. arbitrary: 73. architecture: 3, 70. arithmetic: 95. article: 4-5, 8, 19, 30, 124-128, 156, 159.

article zero: 75-76. axial consonant: 39-40.

Bergson: 79. binary tensor (tenseur binaire

radical): 118-119.

categories: 4. categorization: 114. categorizing operation: 114. causata: 63. causation: 62. change: 62. Chinese: 27, 129, 136-139. chronogenesis (chronogénèse): 6-7, 134.

chronothesis (chronothèse): 6, 134.

cogitation : 143-144. coherence: 14-15, 50, 73-74, 103. comparative grammar: 70. comparative linguistics: 38, 41. comprehension: 156-159. conditional: 74. congruence: 74. constructed causatum (causé cons

truit): 62. content: 105. contextual meaning: 81. continuity: 129.

Delacroix: 145. diachrony: 32-33, 58-64. diachrony of synchronies: 60. direct observation: 56. discontinuity: 129. discourse (discours): 5, 17, 37-39, 45, 52, 54, 80-81, 89-93, 96, 98-99, 105-106, 120, 128-129, 153, 158-159, 164-165.

discriminating operation: 114. double person (personne double):

75. dyadic relation: 70-71.

etymology: 70. expression: 65, 84-89, 92-96, 153 expressive sufficiency: 73-74. expressiveness: 84-87. extension: 155-160.

172 INDEX

extensions: 50. extensity (extensité): 155-160. exteriority: 162. exteriorization: 41, 72. exteriorizing: 40, 71-72. exteriorizing forces: 20. external incidence (incidence

externe): 121-122. external universe: 141.

filiation: 58, 63, 70-71. first glossogenic area (aire

■prime): 89. first person: 160. form: 111-112, 117. formal structure: 105. formal substance: 126. form-substance: 127. free: 106. French: 62-64, 70-75, 87, 94,

97, 105-106, 117, 129, 133-134, 139-140.

future: 75.

general facts: 57. general grammar: 27, 53, 56,

86. generalization: 109-110, 118. generalizing: 114. generalizing movement: 54. glossogeny (glossogénie ) : 63. grammar : 84. grammatical order of words: 87. Greek: 129. Grevisse: 127.

historical filiation: 107. historical linguistics: 30, 36,

56. history: 61. human: 95-96.

ideation: 106. idea-universe (univers idée): 155-158.

imperfect: 74, 97-98.

import (apport): 120-122. improvisation: 84-85. incidence (incidence): 28, 119-121.

Indo-European languages: 39-40, 107, 114, 116, 124, 129, 137-139.

inevitable: 23-24, 163. infinitive: 121. inner universe: 165. intension: 155. interception: 6. interceptive: 51. interiority: 162. interiorization: 72. interiorizing: 71-72. interiorizing forces: 20. interjection: 86. internal chronology: 82. internal incidence (incidence

interne): 121-122. intuition: 5, 20-21, 41. intuitional mechanics (mécanique

intuitionnelle): 21, 123, 153, 161.

langage (fr.): 35-36. language (langage): 145-147,159-

165. language, animal and human: 143. langue (fr.): 35-37. Latin: 61-62, 105. law: 36. law of non-recurrence: 54. Leibnitz: 14, 123, 136. level of actuality: 91. level of potentiality: 91. lexeme: 80, 111, 113-114, 119-

120. lexical meaning: 29. lexical prehension (saisie lexi

cale): 89-90. lexigenesis ( l ex igénèse ) : 133. linguists: 82. logic: 14-15, 28-29. logical person: 122.

INDEX 173

Man-Man relationship: 160, 164. material substance: 126. matter: 111-112, 117. matter substance: 127. meaning: 80, 133. mechanism: 5. Meillet: 23-25. mental: 69-74. mental content: 37-38. mental operations: 79, 82, 84. mental universe: 141, 145. mentally sayable: 19-20. mentally seeable: 18-19. mind: 52-53, 58, 64-65, 79-80,

82, 91, 93, 103-104, 106-107, 113, 118, 121-122, 130, 146, 154, 158-159, 162.

moods: 6, 134. morphemes: 80. morphogeny (morphogénie): 14. morphology: 14, 39-40, 75, 81, 106, 116.

Napoleon: 22. negative morphology: 75. Newton: 44. non-free: 106. non-systernatic: 106. notional ideation (idéation

notionnelle): 40, 129-130. noun: 113-114, 119, 121, 125, 127-128, 156, 159.

number: 4-5, 8.

observation: 57, 69, 82. ontogeny: 14, 88. operations: 49-50. operations of thought: 53-54. operations of tongue: 50. operative time (temps opératif): 123.

orally: 19. orally or scripturally sayable: 20.

order: 46. ordinal person: 28.

paradigms: 106. parole (fr.): 35-37. particular: 113-114. particular facts: 57. particular grammar: 53. particularization: 109-110, 115, 118.

particularizing: 114, 116. particularizing movement: 54. parts of speech: 105, 112, 114, 116-118.

pathology: 96. philosophers: 51, 151. phonetics: 74. phonology: 37. physical: 69-70, 72, 74. physical universe: 145. Poincaré: 130. positions: 80, 133. positivism: 56. positional linguistics (linguis

tique de position): 51, 109-110.

potential: 128. potentiality: 4, 82, 88. potential noun: 4, 159. praxeogeny (praxéogénie): 88. prehension (saisie): 89, 107. prehensive forms: 106. pre-science: 146-147, 152. problematics: 5. process: 133-134. proof: 25. proposal: 63-64. prospective: 97. psychogeny: 134. psycho-mechanics (psychornécanique):

51. psycho-mechanism of language: 83. psycho-mechanisms: 52. psycho-semiology (psycho-sémiolo-

logie): 21, 41, 72-73. psychosystematic linguistics: 104. psychosystematics (psycho-systéma

tique) : 5-8, 18-19, 21, 38, 40-41, 50-52, 104, 109, 123.

radical: 39.

174 INDEX

radical prehension (saisie radicale): 89.

reflection: 56-57. representation (représentation): 6-9, 17, 55, 64, 92-98, 107, 129-130, 143, 147, 153-154.

Sanskrit: 139-140. Saussure: 27, 32-38, 43, 58-61,

64, 137. sayable (dicible): 18, 123, 135-136, 138.

science: 46, 82, 84. scientific curiosity: 151-152. science of language: 69, 82, 105, 145.

scripturally sayable: 19. second person: 160. seeable (visible): 20, 123. seeing (voir): 151-152. seeing of observation (voir

d'observation): 43-44. seeing of understanding (voir

de compréhension): 43-44. semiologically: 6. semiology (sémiologie): 4, 40, 73-74.

Semitic: 75, 129. sentence: 83, 88-90, 105. sentence prehension (saisie

phrastique): 89-90. significant (signifiant): 38. signifícate (signifié): 72-73. signs: 23, 40, 72-73, 137-138. singular: 53, 111, 161. singularization: 109. space: 113-114. space-universe: 8, 112. Spanish: 61, 85. spatialization of time: 6-7. spatio-temporal system: 107. speaking: 50. speech: 37-38, 144. structural ideation (idéation

structurelle) : 129-130. structure: 49, 70, 106. structure of language: 151. structures of tongue: 53, 107.

subconscious: 21. substance: 128. substantive: 4-5, 119-122, 133, 158.

sub-systems: 71, 104. successivity: 35-36, 55. suitability: 72-73. support (support): 120-122. synapsis (synapse): 160. synchrony: 32, 58-59, 64. syntax: 86. syntax of expression: 87. syntax of expressiveness: 87. symphysis: 71-72. system: 3-5, 18, 31, 43, 46, 49,

51, 56, 58, 60-62, 71, 73-74, 80-82, 85, 103, 105-109, 137.

system of gender: 105. system of number: 105. system of systems: 4, 8, 51, 104. system of the word: 115-116. systematic: 106, 134. systematics (systématique): 5, 29.

systematization: 106. systemic meaning: 81. systems of tongue: 106.

tenses: 6, 134. theory: 22-24, 43-45, 58. thinkable: 8, 17, 92-93, 104, 163.

third person: 160. thought: 57, 59, 61, 65, 79, 91-93, 105, 117-118, 138, 143-147, 157-158, 161-165.

time: 35, 114. time image (image temps): 81, 134.

time universe: 8, 113. tongue (langue): 3-8, 13, 17,19, 22, 24, 37-39, 41, 43, 45, 49, 50, 52, 54, 56, 58, 60, 62, 64-65, 71, 73, 76, 79, 82, 84-85, 87-96, 98-99, 103, 106, 117, 120, 124, 128-129, 133, 135, 137, 145-146, 152-153, 155-160, 164-165.

INDEX 175

transform: 63-64. transitoriness: 128. transnotional ideation: 40. turbulence: 143-144, 147.

understanding: 43, 151-152. unit of actuality: 89. unit of potentiality: 89. universal: 53, 111-114, 161. universalization: 39, 109, 115. universalizing: 116. universal-singular : 164. universe: 124, 159, 161, 165. Universe-Man relationship: 160-

161, 164. universe-view: 141. unsayable (indicible): 135-136, 138.

unthinkable (impensable): 163.

vector: 51. vector forms: 110, 117. vector movement: 109. vector operations: 115. vectorial forms: 110. verb: 6, 28-29, 73-74, 113-114, 117, 119-121.

viewed idea (idee regardée): 156, 158.

viewing idea (idée regardante): 158.

vocable: 70, 85.

word: 27, 83, 88, 109-110, 112, 116-117, 119, 133.

word base: 110. writing: 136, 137, 144.

Documents

[Gustave Guillaume] Foundations for a Science of L(BookZZ.org)