Language in the Mind: An Introduction to Guillaume's Theory 9780773560413

Citation preview

LANGUAGE IN THE MIND

This page intentionally left blank

Language in the Mind An Introduction to Guillaume’s Theory WALTER HIRTLE

McGill-Queen’s University Press Montreal & Kingston • London • Ithaca

© McGill-Queen’s University Press 2007 isbn 978-0-7735-3263-2 Legal deposit third quarter 2007 Bibliothèque nationale du Québec Printed in Canada on acid-free paper that is 100% ancient forest free (100% post-consumer recycled), processed chlorine free. This book has been published with the help of a grant from the Canadian Federation for the Humanities and Social Sciences, through the Aid to Scholarly Publications Programme, using funds provided by the Social Science sand Humanities Research Council of Canada. McGill-Queen’s University Press acknowledges the support of the Canada Council for the Arts for our publishing program. We also acknowledge the financial support of the Government of Canada through the Book Publishing Industry Development Program (bpidp) for our publishing activities. The original versions of chapters 2 and 4 were published by Mouton de Gruyter and John Benjamins respectively. The present versions appear here with their permission.

National Library of Canada Cataloguing in Publication Hirtle, W. H. (Walter Heal), 1927– Language in the mind: an introduction to Guillaume's theory / Walter Hirtle. Includes bibliographical references and index. isbn 978-0-7735-3263-2 1. English language – Grammar -- Theory, etc. 2. Guillaume, Gustave, 1883-1960. 3. Cognitive grammar. 4. Linguistics. i. Title. p121.h556 2007

425.01

Typeset in New Baskerville 10.5/13 by Infoscan Collette, Quebec City

c2007-902617-6

To Roch Valin mentor and friend

This page intentionally left blank

Contents

Preface ix 1 Introduction 3 2 Language and the Ability to Speak 18 3 Words, Words, Words 30 4 Meaning: Representing Experience 51 5 A System for Representing 66 6 The Method of Analysis in Psychosystematics 85 7 The Substantive: A System of Subsystems 106 8 The Substantive and the System of the Parts of Speech 120 9 Some and Any 140 10 The System of the Verb 155 11 Auxiliaries: How Do They Help? 174 12 The Proof of the Pudding 189 13 The Noun Phrase 201 14 Concord, Discord, and the Incidence of Verb to Subject 213 15 Thought and Language 225 16 Conclusion 237 Glossary 243

viii

Contents

Notes 247 Bibliography 263 Index 271

Preface

The work of Gustave Guillaume is little known outside of France, partly because of his obscure life (see below) and partly because he published relatively little during his lifetime (see Tollis, 191), but mainly because he adopted a point of view that was quite new to his contemporaries and that, even today, is still unfamiliar to many linguists. Seen in retrospect, part of this originality was due to the fact that from the outset he integrated into his approach three currents of linguistic research that were to mark the twentieth century. His work reflects both the descriptivist’s concern for precise and extensive observation of real usage and the theoretician’s preoccupation with finding the explanatory principles lying behind usage, the structures or universals permitting us to understand what is observed. Furthermore, his constant focus was on the mental reality of language, leading one reviewer to remark that Guillaume in the 1940s, like cognitivists today, was seeking “to explain the properties of language in terms of general cognitive mechanisms” (Epstein, 308). What enabled him to combine these three elements into a unified approach to language was the method of analysis he adopted – the comparative method – but applied to language in synchrony. Saussure’s Cours de linguistique générale had divided the study of language into separate disciplines: diachronic and synchronic. Guillaume attempted to unite them into a single science through the method of analysis so successfully applied in the great nineteenth-century studies to language in historical time. Only in 1929 with his Temps et verbe did he succeed in applying the comparative

x

Preface

method to language in cognitive time, the cognitive time of the speaker during the act of speech. The result was a view of language as a mental reality, something we acquire thanks to our linguistic forbears and exploit according to the needs and circumstances of the moment. For many scholars of his day, however, this view constituted a challenge, requiring them to accept that a great many things happen in the speaker’s mind during the present of speech, the present instant. Throughout his fifty-year career as a linguist, Guillaume worked at developing his initial insight into a general theory and typology of human language, each year adding some new hypothesis, reworking some former analysis, providing more evidence for a given theoretical stand. Although he worked incessantly, the abundant documents he left contain no text for introducing the interested reader to his approach. Furthermore, most of his examples are drawn from his mother tongue, French, and this makes it even less accessible to most speakers of English. This then is the purpose of the present volume: to introduce Guillaume’s theory of language to the English-speaking reader, illustrating and supporting it with examples drawn from contemporary English. Comments from scholars working in other theoretical frameworks are mentioned as a means of highlighting Guillaume’s approach, but are not intended to constitute a systematic comparison with any other theory. Any merits this introduction may have are due largely to discussions with students and colleagues at the Fonds Gustave Guillaume in Quebec City, particularly during the weekly research seminar instituted there over forty years ago by Roch Valin. I am especially grateful to Joseph Pattee, Dennis Philps, Patrick Duffley, John Hewson, Ronald Lowe, and to the late Bob Uhlenbeck, as well as unnamed readers, for comments on earlier versions of this volume. And like most of those who have made the effort to explore Guillaume’s thinking on language, I am most indebted to Roch Valin for his rigour, encouragement, and vision.

LANGUAGE IN THE MIND

This page intentionally left blank

1 Introduction A language involves a system where everything fits together and has a wonderfully rigorous design. Antoine Meillet (cited in Guillaume 1989, 57)

AN OBSCURE LIFE: 1883–1960

It will help to bring out the originality of Guillaume’s work if we take a moment to situate him in the linguistic tradition of France at the beginning of the twentieth century. Guillaume’s life was, in his own eyes, an obscure one (passée dans l’ombre 1). His father, a talented painter, died when Guillaume was still a child. Little is known of his mother, with whom he lived until her death in 1926. School reports for the years 1891 to 1896 from the École Milton in Paris are the only record of his schooling until his name appears on the register of the École pratique des Hautes Études in 1914. A twelve-part Méthode Guillaume, handwritten in Russian but with a printed cover, for preparing teachers of French in Russia, dates probably from around 1905 and witnesses to his knowledge of Russian. Giving lessons to Russian émigrés was the stimulus for his lifelong interest in the article in French because of the need to find a means of explaining its use. The turning point in his professional life came in 1909 when he was working in a Paris bank where Antoine Meillet had his account. Meillet, impressed by Guillaume’s good general background in science, philosophy, and the arts, and especially his lively interest in language, encouraged him to undertake a career as a linguist with the aim of extending linguistics so that the same method of analysis could be applied to both the historical and the contemporary, or descriptive, dimensions. This is what Guillaume undertook in his first properly linguistic writings, three short studies published

4

Language in the Mind

in 1911, 1912, and 1913. The reason why Guillaume subsequently tried to hide these works throws light on all his later career. According to Valin (1994, 126) it is because “in this first stage of his research Guillaume appears to have gone astray. Hypnotized by the frequent necessity of using certain grammatical forms, he tries to reduce this necessity to the rules of an unconscious logic preceding conscious logic (in this he is right), but at the same time a universal logic, a set of fundamental equations applying to all languages, with the particular logic of each language reflecting just one of the possibilities of treating the terms of one of these primary equations. This certainly involves a mistaken approach, the mistake consisting precisely in its overly mathematical aspect. The logic of language – it does exist – is, as G. Guillaume was to show later, a sui generis logic which is not the learned logic of mathematics, even if there are serious reasons to believe that the intuitive sources of the latter could be in the former.” Guillaume’s continuing search for a method to analyze this “sui generis logic” of a language is manifest in his first major publication, on the article in French, which appeared in 1919. Although this study did not give him what he was looking for, it did pose the problem in terms of language potential vs language as actualized in usage, thus setting the stage for his first real breakthrough, published in his 1929 volume, Temps et verbe. The originality of this work is to describe the main lines of the subconscious system of the verb in terms of successive moments in the mental process of constructing the verb’s configuration of time during the instant of speech. It remained for him to find the same operational insight for analyzing morphemes, the elements constituting a grammatical system. This came in the early 1940s, when he finally found the method for analyzing the meaning potential of each article, making it possible to explain the different senses they express, a method essentially the same as that employed in comparative and historical studies. Thanks to this method he was able to describe the article system and to obtain his first view of a mental mechanism he was to postulate as the basis for all language systems. Throughout his remaining years he exploited the avenues of research opened by this view of language and its systems, and particularly the relation between languages with radically different systems for structuring words.

Introduction

5

On the personal side, Guillaume was dogged by illness most of his life and, after the First World War, lived in economic straits, not being eligible for a university position since he did not have a doctorate. He managed to support himself by giving private lessons and doing proofreading for a publishing house. Only in 1938 did he get a part-time teaching position at the École pratique des Hautes Études, where he lectured until the week before his death. Each of his lectures, usually one a week, sometimes two, occasionally three, was written out as a publishable text. These manuscripts, along with his many research notes, essays, etc. (some sixty thousand pages of manuscript, around a third of which have been published so far, in the Leçons de linguistique and the Essais et mémoires collections2) bequeathed to Roch Valin, are now kept in the Fonds Gustave Guillaume at Laval University in Quebec City. GUILLAUME AND THE NINETEENTH CENTURY

Fanciful accounts of language origins in some ideal language lost any appeal they may have had when Sir William Jones read his paper before the Bengal Asiatic Society in 1786. By declaring the affinity between Sanskrit, Greek, and Latin to be “so strong, indeed, that no philologer could examine them all three without believing them to have sprung from some common source, which, perhaps, no longer exists,” he introduced the dimension of time into matters linguistic to explain observed correspondences that are “too close to be due to chance” (Bloomfield, 12). He thereby awakened in anyone curious about human language a perspective stretching back in history to the earliest texts and even beyond into the darkness of prehistory. The work of the scholars of the nineteenth century retracing developments through time and reconstructing states of language that no longer exist constitutes the first application to a human institution of scientific method, with all the observation of detail and rigour of thought required to reconstruct in the mind what cannot be observed. Not the least among the nineteenth-century scholars was Saussure, who viewed language in a traditional way as basically a set of symbolic units consisting of a sign 3 and what it signifies: the significate or meaning. Moreover, aware that development in time can be discerned only by comparing the state of a language at different

6

Language in the Mind

historical moments, Saussure proclaimed the need to complement, with a synchronic or descriptive perspective, the diachronic or historical perspective holding sway at the beginning of the twentieth century. He is perhaps best remembered for insisting that language is systematic and that to describe a language at a given moment of its existence involves describing both usage (la parole) and the system (la langue). The posthumous publication of his Cours de linguistique générale in 1916 sparked discussion throughout much of the twentieth century, particularly with regard to Saussure’s conception of la langue, language as system. One of Saussure’s students, Antoine Meillet, known for his vast knowledge of the Indo-European languages and his authoritative works on the comparative method, introduced Guillaume to comparative grammar and historical linguistics as well as to Saussure’s thought. He confronted Guillaume with the need to find a method of analysis common to both the historical and the descriptive approaches to language in order to constitute linguistics as a science commensurate with its object. As a consequence Guillaume never lost sight of the diachronic perspective. He realized that whatever may be proposed for the contemporary state of a language is necessarily the outcome of developments over past generations. He also took to heart the traditional view that language units are symbolic, binary, made up of a physical sign tied to a mental significate consisting of a lexical and/or a grammatical meaning. He was most impressed by Meillet’s insistence that language is somehow systematic. Thus, as mentioned above, his first publications in linguistics consisted of three short attempts to systematize verb tenses and the article on the basis of logic of a mathematical sort, a point of view he soon abandoned in favour of an approach based on the coherent relations between meanings, an approach that today could be called cognitive. Guillaume also inherited the current view of typology, which divided languages into three or four types according to the manifestations of their morphology: isolating, agglutinative, polysynthetic, and inflexional. In the light of the meaning-based approach he was to adopt, this manner of distinguishing language types by means of their visible semiology would constitute a challenge for him to explore the mental conditions producing such variations in the physical morphology. In fact, in each of these areas – the relation between language and time (diachrony vs synchrony),

Introduction

7

the relation between sign and significate in the units of language, the relation between usage and system, and the relation between language types – Guillaume, in search of an all-embracing method, was soon led far beyond the ideas received from tradition and so proposed some original views that were not readily accepted by his contemporaries. GUILLAUME’S INNOVATIONS

Guillaume undertook to explore the implications of the view, commonly accepted after Saussure, that language as a grammatical system exists in synchrony, in the present, by examining the article in French. He soon realized that the relations between the physical signs of the morphemes, the symbols themselves, are not systematic. “Looking at things from the point of view of the signs, it is usually impossible to establish that there is a system,” as he put it later (2004, 25). As a consequence, he focused on the significates, attempting to observe a systemic relation between the meaning expressed by the definite article and that expressed by the indefinite article. This turned out to be impossible because the sense each article expresses varies from one use to another. That is, because of polysemy in the give and take of actual usage, the relation between a given sign and what it expresses is not constant. Confronted with the problem of polysemy, the major obstacle for an approach dealing with meaning, Guillaume realized that, although variable, the different senses expressed by a given word are related, like variations on some common theme, and this led him to the conviction that each word existed in la langue as a potential meaning whose various actualizations can be observed in la parole. This notion of a meaning potential, which is at least implicit in the work of previous scholars, did not reveal the potential meaning of either article, but it did focus his attention on the need to find a method for discerning it. Moreover, viewing meaning in this potential-actual way gave him a general framework for explaining polysemy and permitted him to continue exploring the idea of a language system based on meaning. Having observed that the system of the articles is not to be found in usage, in the relations between the signs or between the senses expressed, Guillaume concluded that the system of the articles does not emerge into consciousness. This led him to his first major

8

Language in the Mind

contribution: that the system exists not in an abstract norm as Saussure suggested for la langue but as a potential in the mind of each speaker, albeit far below the level of consciousness. (To distinguish between these two conceptions of the systemic level of language, I will retain la langue to designate Saussure’s concept and use the English term tongue for Guillaume’s.) Thus he viewed the relation between a system of morphemes and usage as that between a potential and its actualizations in the moment of speaking. That is to say, the articles exist as systematically related meaning potentials in the preconscious mind ready to be actualized in order to express what the speaker has in mind. (Understanding usage in this way, as what has been actualized, led Guillaume to substitute the term discourse for Saussure’s parole, since la parole, speech, as a physical realization, has its own state of existence as a potential in tongue.) This view provided Guillaume with a framework, or better, a general postulate for examining language: tongue exists as a potential providing what is necessary to produce discourse as the need arises. For Guillaume, this brought into focus the first aim of linguistics: through the observation of discourse, and particularly the senses expressed by different morphemes, to discern and describe the preconscious mental systems in tongue. He later called this la psychosystématique du langage, the psychosystematics of language, the study of the mental systems in tongue. The postulate of tongue as the potential for discourse entailed another constant in his approach, prompted no doubt by his historical studies and his knowledge of scientific method. While it is taken for granted in diachrony that every observed form can be explained as the result of a prior development, Guillaume’s innovation was to apply the same means for finding an explanation to synchrony by seeking in the speaker’s mental system prior developments or processes that can produce the observed results in usage.4 As a consequence, all his work is centred on how the speaker produces the words and sentences that constitute discourse since it is only by describing causal factors, conditions bringing about what we observe, that we can explain it. He realized of course that grammatical systems are not the only factors conditioning what a speaker says. Besides morphemes there are all the lexemes of a person’s vocabulary, and these too exist as potentials in tongue, though not organized in systems the way grammatical

Introduction

9

morphemes are. Moreover the message a speaker has in mind and wishes to express is necessarily a conditioning factor, but it is outside language, part of the speaker’s experience. Reflecting on the way the message conditions what is expressed, Guillaume realized that meaning potentials are exploited in such a way as to configure what the speaker has in mind to talk about. That is, he saw that for each word, the potential lexical meaning and the potential grammatical meaning(s) are actualized in such a way as to depict linguistically what constitutes the speaker’s message, an insight that later led him to characterize tongue as un univers regardant, a viewing universe, a mental universe for viewing our ongoing experience of the universe around us. He therefore postulated that lexemes and morphemes are means of representing something in the speaker’s experience. Each word is thus a representational unit for depicting what a speaker has in mind. By insisting that the relation between our ongoing momentary experience and meaning expressed is one of representing – he often repeated “no expression without representation” – he was able to show that in every language act the extralinguistic (our experience) conditions the linguistic (meaning), thereby bringing the study of language down to the here and now of an individual act unrolling in the present. Thus he confronted system and usage within the reality of a person speaking in the present of consciousness and in so doing, undertook one of the twentieth century’s most intriguing adventures into the human mind. In Guillaume’s day, people were becoming aware of “deep time” (cf. Gould, 1–19) – or, geological time – which, not being perceivable – because it stretches far beyond the limits of what is perceivable through the earliest geological records – is only conceivable. Guillaume’s insights led him into the realm of what might be termed “deep mind” to explore the psychomechanisms of language, the mental operations far below the surface of consciousness that permit speakers to instantaneously produce fitting words, and subsequently sentences, from the language system constituting their mother tongue. To guide him to these depths, to find that narrow passage between positivism and idealism, Guillaume had only the scientific method as practiced by the great comparatists. And so he begins his first major work, his study on the article, as follows (1919/1975, 11): “The present work is an attempt to apply the comparative method to the formal [grammatical] part of

10

Language in the Mind

languages.” This method called for careful observation of both physical sign and meaning expressed, as well as a gift for reflection that permitted him to move with ease from the general postulate of system in language to particular uses, and from particular back to general. His constant concern was to discern the mental processes speakers undertake in order to translate, to “commute” as he put it, some experience they have in mind, unsayable as raw experience, into something said. Rare were the linguists who could appreciate what Guillaume was trying to do, particularly in view of the fact that, in this first study, he did not succeed. He did not manage to discern the potential meanings of the articles and describe their system – it was to take him another twenty years to do this – but this volume did prepare the way for his first success in depicting a preconscious system some ten years later. In his 1929 volume Temps et verbe he describes three systems of the French verb (aspect, mood, and tense) in terms of mental processes. Meillet, one of the few who understood the significance of this achievement, once described the volume as “the study most apt to make explicit what Saussure understood by la langue,”5 thus bringing out the real significance of Saussure’s dichotomy. On the other hand, Guillaume’s manner of conceiving a grammatical system entailed something that most linguists would consider quite non-Saussurian: it introduced the idea of process, and therefore the dimension of time, into synchronic linguistics. Since any process involves time, describing a system of tongue as a mental process implies that it requires time, the micro-time required for preconscious mental operations. While commonplace today, to invoke preconscious psychomechanisms in Guillaume’s day was viewed by many as outlandish, as a sort of metaphysical speculation that had little to do with the reality of language, especially since it implies that whenever a verb is needed in a sentence it must be constructed, or rather reconstructed, by the speaker. Undeterred by this lack of comprehension, Guillaume pursued his reflections and soon realized that what he had discerned for the verb applies to all words: every time a word is needed for a sentence it must be mentally reconstructed by the speaker. This is how Guillaume’s whole linguistic career came to be focused on a single problem, what he considered the central problem for linguistics: how, from the possibilities of tongue, a speaker in the moment of speech constructs words to be used in a sentence.

Introduction

11

One can perhaps understand why even Guillaume’s subsequent work found such limited acceptance among his contemporaries. Most linguists have no difficulty reducing their view of language in diachrony stretching over many centuries, and even over the millennia of human existence, to a view of language in synchrony, in the present, as Saussure proposes. But some effort is required to focus on the present as it exists for the speaker because the present reduces to the instant of consciousness wherein any speaker must operate, to the “Cartesian instant” as Guillaume calls it.6 Even more effort is required to follow Guillaume when he takes us behind the scenes, behind the words that emerge into consciousness, to discern the mental processes unrolling within the instant, bringing words onto the stage of conscious awareness. Most disconcerting of all, perhaps, is his viewing words as processes, the constructive mental processes that bring them into existence. In fact, their resultative, observable state, albeit real, is so ephemeral when we speak as to be a mirage for some observers, a “linguistic figment.”7 Disconcerting though it may be, viewing words (and of course sentences) as merely passing over the stage of conscious awareness is inevitable once we introduce the time dimension into the synchrony of language, that is, into the process of someone thinking/ speaking and, as a consequence, of someone listening/understanding. The durable result of a given act of speech is neither words nor a sentence, but a message. Having set out on this road to discern how speakers produce the words they use – a road where, as he once wryly commented, he was not bothered by the traffic – Guillaume soon realized that a word’s function is determined by the way it is constructed, by its mental morphology or part of speech. Here too his views met with limited acceptance since many linguists, concerned with a word’s role in a sentence, were content to study syntax without any prior examination of the words involved. In fact he was sometimes criticized for neglecting the study of syntax, whereas in his own eyes he was establishing the necessary basis – an understanding of the grammatical makeup of words – that would permit us to explain syntax and not merely “account for it” through a more or less arbitrary description or set of rules. Regarding the word, and particularly the verb and the article in French, from the systemic point of view led Guillaume to consider that traditional historical studies were incomplete. To trace the

12

Language in the Mind

development of, say, the future conjugation from Latin to contemporary French by examining each of the morphemes involved – the history of their signs and their significates – is insufficient since it neglects the changes in the system, i.e. the changing relationships between the significates, between the future tense and the other tenses of the indicative. But again this view was not adopted by many working in historical linguistics perhaps because they did not consider the grammatical system, the verb in tongue, a mental reality whose development could be traced over the centuries. In the latter part of his professional life – see the first two lectures in the 1941–42 B series (2005, 1–26) – Guillaume’s attention was directed more and more to languages whose words are constructed on bases quite different from the part-of-speech basis common to languages in the Indo-European family. He came to realize that the element common to all languages, the universal on which a language typology can be based, is the word, or rather the vocable (a more technical term he employed to avoid foisting our conception of a word onto languages whose minimal sayable element of speech is radically different). In his last years he laid the foundations, both diachronic and synchronic, for a language typology, his Théorie des aires glossogéniques, a theory comparing different language types on the basis of the mental spaces – the “areas” – involved in their wordforming systems. He planned to present and illustrate this theory in a five-volume study but was able to complete an introductory essay and the first volume (Guillaume 2003, 2007, respectively) of this ambitious project before death intervened. GUILLAUME TODAY

What does Guillaume have to contribute to linguistics today? Perhaps the most striking contribution is his global view of language with its two modes of existence for any speaker, as a systemic potential and as actualized discourse, with the necessary mental processes for passing from one to the other (see below, chapter 2). The potential, a mental construct permitting the operations of representing and expressing anything arising in one’s experience, is a permanent possession of the speaker, whereas discourse is the result of intermittent acts of constructing words and sentences. Guillaume was quite aware that this operative view of language held implications both for areas of mental activity quite distinct from

Introduction

13

those involved in language and for the necessary physical substratum of language. Intent as he was on establishing linguistics as a science adequate to its object, however, he was careful not to step beyond the bounds of language itself and so made little attempt to explore obvious links with related disciplines such as psychology, neuroscience, and philosophy. Focusing on the word (vocable) is another basic contribution because, although a language universal, the element of language the ordinary speaker “feels that he knows best” as Bolinger (1963, 113) puts it, the word receives little attention from contemporary scholars. Even those who accept it as part of the linguistic environment often consider that “we retrieve a word from the mental lexicon,” as Croft and Cruse (109) put it, not perceiving the need to construct it in the moment of speech. Guillaume’s view of it as the fundamental unit of language – the unit of representation and therefore to be reconstructed each time we want to use it – provides a new way of looking at the word and its place in constructing a sentence (see chapter 3). Thus his speaker-oriented approach focusing on the psychomechanisms involved in constructing words marks him off from other approaches and invites research to bring to light the corresponding neuromechanisms. Another general problem, that of polysemy, which confronts anyone concerned with analyzing meaning, some consider a matter of “schema-instance relations” (Taylor 2002, 437), i.e. as involving a general/particular relationship, others as “a matter of isolating different parts of the total meaning potential of a word in different circumstances” (Croft and Cruse, 109), i.e. as involving a whole/ part relationship. In chapter 4 it will be shown that Guillaume’s manner of analyzing the senses expressed by a morpheme as actualizations of a single underlying potential meaning provides a different explanation for polysemy. This potential/actual meaning approach has proved valuable both in ESL (cf. Hirtle 1980) and in teaching the deaf (cf. Sadek-Kahil 1996). The view that every grammatical morpheme is part of an operational system brings out aspects of their meaning potential not otherwise discernible, as will be seen in the analysis of the system of number in chapter 5. Recognizing the micro-time required for a system to operate, Guillaume was led to postulate a type of mental operation underlying all systems, including that of the word. This general postulate concerning the mental basis for human language

14

Language in the Mind

offers interesting prospects to be explored as a framework for language acquisition and formation as well as a link with other spheres of mental activity. Of interest to those working in the diachronic perspective is Guillaume’s method for analyzing grammatical systems, which is essentially the same as the comparative method in historical studies. As is brought out in chapter 6, the main difference between the two applications of the method is in the timespan separating the observed fact from the reconstructed state. Having both a common object and a common method of analyzing the data, historical and descriptive linguistics can now be seen to form a single science. This achievement constitutes an important development for the history of science. Thanks to it Guillaume viewed linguistics as a science on its own, distinct from neighbouring disciplines like psychology, but he was aware of the contribution it can make to them, and particularly to anthropology through his Theory of Glossogenetic Areas. This obviously applies to contemporary studies of human consciousness (e.g. Merlin Donald’s A Mind so Rare: The Evolution of Human Consciousness). Insofar as morphology is concerned, one extensive study states in its conclusion that “one of the key unresolved questions in morphology is ‘what is a word?’” (Spencer, 453). This is the question Guillaume posed at the beginning of his study of morphology because of his systemic approach. The realization that all words have a morphology enabled Guillaume to make numerous original observations concerning their grammatical makeup and to extend his analysis to morphemes not provided with a distinct physical sign. In order to illustrate how words are formed grammatically by a part of speech in English, the genetic morphology of the substantive and of the verb are described below (chapters 7 and 10, respectively). His original view of the part of speech – a crucial question for Guillaume as we have seen since it involves the definition of a word in Indo-European languages – is perhaps his primary contribution to contemporary linguistics, and particularly his manner of conceiving a part of speech as a set of processes grammaticizing the word’s lexical import rather than merely as a class of items “stored in the speaker’s mental lexicon,” as Taylor (2002, 74) puts it. In syntax, the great variety of syntactic groups observed in discourse has led a number of scholars to develop construction grammars (cf. Croft and Cruse, 257–90). For Guillaume, such groupings are made

Introduction

15

possible by the meaning potentials, both lexical and grammatical, provided by tongue and actualized in the words involved. Thus he concentrated on describing the system of the part of speech whereby a word receives its grammatical meaning, since this is what permits it to establish a syntactic relation with another word or phrase (see chapter 8). That is, since syntax is the consequence of the makeup of the words involved – “every language has the syntax of its morphology” as he used to put it (2004, 352) – Guillaume’s approach would add a new construct to those generally included in construction grammars (cf. Taylor 2002, 561–2), the one on which the others depend, namely the word. A word-based syntax inevitably leads to the question of establishing a relation between one word and another, and so Guillaume comes to consider syntax as essentially operative, as the set of processes making one word or phrase “incident” to another. This manner of conceiving syntax will be illustrated by describing the forming of an auxiliary + verb compound (chapter 11), a noun phrase (chapter 13), and the relating of predicate to subject (chapter 14). From an operational point of view, then, the study of syntax will focus on the constructing of syntactic units rather than on the constructions resulting from it. As in the diachronic perspective, in synchrony it is the process leading up to the observed result that permits us to explain it. To see how far a word-based syntax can be pushed calls for a more fine-grained discussion of usage involving real examples where the variation of the lexical import and its influence on the grammatical import of the word can be shown to give rise to the expressive effect of the sentence. Recognizing the difference between the lexical and the grammatical meanings of words gave Guillaume a means of distinguishing between what is particular, depending on the specific words used, and what is general, arising from the grammatical systems called on in a given construction. This distinction, which has not been exploited in this way in other approaches, provides a basis for explaining some of the finer nuances of meaning expressed in ordinary usage, as we shall see in chapters 12–14. Finally, Guillaume’s concern with the word-forming system of a language focuses on what is general – every language has a system for forming words – and on what is particular to each language – no two languages, not even closely related ones, form words in

16

Language in the Mind

exactly the same way. Here it is not possible to describe types of word that differ structurally from the word in English and compare them within the framework of his general theory of typology, a theory based on the idea that different mental “areas” or spaces are required for forming words of different types, a notion of mental spaces to be compared with that found in Fauconnier. We will however be able to evoke briefly (chapter 15) what is perhaps the most general question raised by the title of the present study, the relation between language and thought. AN ALL-EMBRACING THEORY

Guillaume’s constant concern was thus to establish linguistics as a science adequate to its object, capable of embracing in one view the whole phenomenon of human language with all its variety in geographical space, in historical and prehistorical time, and in every speaker’s instant of consciousness. To do this he focused on the word, attempting to discern how it is constructed by speakers of different languages at different moments of historical time. Since the purpose of this volume is to present the main tenets of Guillaume’s theory of language through examples drawn from contemporary English, it cannot do justice to this panoramic view of language. It should therefore be supplemented by reading in his published work (some twenty volumes so far), and particularly in the one volume available in English.8 No attempt is made to include analyses of other languages based on Guillaume’s method of analysis or the way some principle of Guillaume’s has been applied in disciplines such as stylistics, language didactics and reeducation, translation, developmental psychology, and, of course, the philosophy of language. The website of the Fonds Gustave Guillaume (cf. note 2 above) provides a bibliography of these applications, as well as the linguistic analyses published to date. The general overview of a theory that, as sciences go, is still relatively young inevitably results in a chiaroscuro where, it is hoped, the many areas remaining to be explored will be seen as a challenge for further research. Indeed, this also applies to those areas already explored, since, as in any other science based on observation, no particular explanation or theory should be considered to be proven beyond question. All that can be considered as definitively acquired is the attitude proper to science, a postulate

Introduction

17

of system or order that has been called, hearkening back to Thales, the “Ionian enchantment” (cf. Wilson, 3–7). This “prejudice of order,” as Guillaume called it, sustained him throughout his fifty years of observing and reflecting on language, convinced, in spite of the numerous problems encountered in usage, that there is coherence at a deeper level. This is why he began the first lecture of his 1952–53 series with the following declaration: 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. (1984, 3)

2 Language and the Ability to Speak1 The whole of science is nothing more than a refinement of everyday thinking. Albert Einstein (1981, 290)

LANGUE AND COMPETENCE: SAUSSURE AND CHOMSKY

A few years ago a well-known linguist argued that we must take into consideration “the obvious ability of people to speak” if we are to reach “an understanding of language that is grounded in reality” (Lamb), implying thereby that the individual’s capacity or potential for producing discourse is part of this reality. This way of viewing language, as a dyad, a twofold entity made up of observable sentences and the non-observable ability or potential for producing them, is reminiscent of the langue vs parole dichotomy proposed by Saussure in 1916 and, in a very different way, the competence vs performance dichotomy proposed by Chomsky some fifty years later. Since neither of these proposals has widespread acceptance as a basic parameter today, at least in the English-speaking world, one may wonder what would justify making an apparently similar proposal so recently. The question is of interest because it involves, as we shall see, the efforts of linguists throughout the century to find a solid basis for linguistics as a science. It is well known that Saussure had conceived a means of reconstructing the system of vowels in Proto-Indo-European to explain their historical development down to attested examples in the earliest texts of the different Indo-European languages (1879/1987). Imbued with this comparative method in historical linguistics – observing attested examples, trying to understand the data by

The Ability to Speak

19

conceiving some prior state of affairs, explaining the data by describing the process whereby each of the attested examples must have developed from the prior system2 – Saussure attempted to adopt the same scientific approach for language in synchrony by eliminating the historical dimension, the temporal parameter. To explain observable parole, he proposed langue as the necessary state of language structure, the system governing the usage of individual speakers. Saussure’s notion of langue has been interpreted in different ways, one of which, suggested by Saussure himself (1916/1955, 25, 31, 38, etc.), being that it is the language of the community without the particularities of individual usage, a sort of norm. As such, language-as-langue came to be considered an abstraction, a purely theoretical construct that has no observable role to play in actual speaking because the community does not speak, only individuals. In fact, Saussure did not attempt to justify its existence by working out, as he had done for the theoretical construct in his historical studies, either the system of langue or the processes whereby the facts of parole could be shown to be the consequences of that system. It is therefore not surprising that his view of langue was not widely accepted as part of the reality of language. In Guillaume’s day, few linguists of the English-speaking world, influenced by positivism, gave such abstractions serious consideration as a basis for language analysis. It was not until Chomsky proposed his competence/performance dichotomy that the idea resurfaced, but in a different form: language-as-competence is the prerogative of the ideal speaker. Thus competence, not to be found in any real speaker with his or her limitations and subject to all the accidents involved in producing discourse, is also an idealization. Although it may be a useful conception for some purposes, it is not linked, through the actual operations undertaken by an individual speaker, to perceivable discourse. That is, understood in this way, competence, like langue, has no existence outside the linguist’s imagination and so can provide no basis for describing a real speaker’s language ability, or potential (to use a more general term). Small wonder, then, if the very notion of language-as-apotential is treated with suspicion, and even outright derision, when such renowned linguists fail to provide a satisfactory account of it. It remains, however, that two such failures are not sufficient grounds for considering language-as-a-potential to be merely an

20

Language in the Mind

explanatory expedient with no real existence, something like the ether of nineteenth-century physics. After all, in our everyday thinking we assume that the notion of a potential is a necessary component of other situations: the moment we see someone riding a bicycle or playing the piano we know that they have acquired the ability to do so and that this ability exists whether that person is actually performing or not. Similarly, when we hear someone speaking, we have no choice but to assume that they have acquired the ability or potential to speak that particular language.3 In other words, it is “obvious” for Guillaume that language-as-a-potential, far from being just a figment of the linguist’s imagination, really does exist in speakers even though they are quite unaware of it. Equally obvious, this potential exists whether one is actually involved in speaking or not, it being, in fact, a permanent acquisition of any person, barring accidents. Not only does our ordinary experience suggest this, but research on the brain is providing external evidence that some such potential is a cognitive reality. And so, taking cognitive in the sense of neurological, it has been suggested that the ability to speak resides in the brain, in the necessary physical support of language. Guillaume never envisaged any such hypothesis, probably because it would entail deriving from a physical potential the actual mental content, or meaning, we get from a sentence. While recognizing that language (or any other mental construct for that matter) cannot exist without a neurological support, he considered this support to be outside language and so beyond the linguist’s field of competence. Hence he postulated that language-as-a-potential is something nonphysical, a strictly mental part of language. Although the ability to speak is a cognitive reality, in the psychical not the physical sense of the term, it cannot emerge into conscious awareness. That is to say, if everyday thinking is any indication, language-as-a-potential is a permanent, albeit hidden, feature of the linguistic landscape whose existence can be ignored in linguistics only if we turn our backs on part of reality. It would seem, therefore, that the linguist’s efforts can best be employed in attempting to reconstruct this part of reality by theoretical means, and not in ignoring it, or placing it outside language in the neurological substratum, or dismissing it to the limbo of a non-speaking community or a nonexistent, ideal speaker.

The Ability to Speak

21

LANGUAGE-AS-A-POTENTIAL: GUILLAUME

The real problem for linguistics, then, is not whether this potential exists in the preconscious mind but rather how to come to grips with it – how to analyze it, how to describe it, how to relate it back to observable data – and this problem is inescapable for anyone who wants to understand and explain human language. As in other sciences based on the data of observation, it is a matter of finding the way to approach an area of reality whose existence appears to us necessary even though we cannot know it through observing it directly but only through observing its results. The approach one adopts will be determined in large part by the way one first conceives of this ability, by the initial vague notion one forms of what a language in its most elemental form is. This often uncritical assumption we make concerning the nature of language is important because from the outset it conditions what we look for in the data. Conceiving of it as either the system of an ideal speaker or as a sort of norm for a community of speakers has proved inadequate. The problem for linguistics as a science, which like any other science attempts to embrace the whole of its object, is this: how can our language potential be conceived in terms of an ability acquired and exercised individually by speakers constituting a language community? Initially we can attribute three characteristics to language-as-apotential: it is organized, dynamic, and mental. Since Saussure it has been commonplace to consider language as somehow organized in a system, if only because we can usually use our mother tongue so readily and with such unerring ease even in novel situations. Just as Saussure attributed the coherent set of relationships to the vowels in his PIE reconstruction and not to the set of actual results in the various attested languages, we are led to seek the system in the potential part of language. This relieves the linguist of the impossible task of trying to prove at all costs that actual usage is systematic in itself, that syntax constitutes an independent system. Since an ability is for producing results, this systematic potential must also be dynamic by nature. That is, a system providing the potential for carrying out certain productive processes must be organized in such a way as to make these processes or operations possible. And finally, since these operations are in large part

22

Language in the Mind

cognitive, consisting of thought processes, the ability permitting them must be in the mind. All this leads Guillaume to conceive language-as-a-potential to be a coherent construct in the mind permitting the repeated carrying out of certain operations, i.e. to be a mental mechanism, or rather a set of psychomechanisms. Compared with the Saussurian and Chomskyian approaches, this involves a very different way of regarding our ability to speak since it implies that each of us has a set of dynamic systems enabling us to realize the mental and physical processes required to construct and utter the words and sentences we need in order to talk about whatever we have in mind and to understand the discourse of others. Viewing a person’s language potential as a set of mental systems in this way involves both an operational conception far removed from Saussure’s idea of la langue as a set of static oppositions and a real-speaker oriented conception poles apart from the competence of an ideal speaker. It is closer to an approach that considers “A dynamic view of conceptualization … essential to a principled understanding of grammar” (Langacker 1997, 249) because here the ordinary speaker’s mental activity is considered crucial. Essentially, Guillaume views our ability to speak as the set of linguistic conditions in the mind necessary to produce whatever we say, a view he developed from 1919 on. This organized set of psychomechanisms constituting the language potential of any speaker, as we have seen, he calls tongue (la langue) to indicate that it is just as real as discourse, the actual speech and texts a speaker produces. Thus for Guillaume our mother tongue is an “obvious ability,” an acquired capacity of the mind that we activate whenever we want to talk about something. USING THE ABILITY

The upshot of all this is that in order to grasp the reality of language – and this includes both what can be seen and what cannot be seen – we must always keep in mind its two modes of existence, the potential and the actual, ability and speech (or text). But this is not all. To grasp its total reality we must also keep in mind the language processes whereby the potential is actualized to produce a particular unit of discourse. This brings us to the main problem in obtaining as complete and realistic a view of language as possible: what are the successive phases involved in a person undertaking an act of

The Ability to Speak

23

language and producing a certain discourse? Quite obviously, when we speak we do not put all our mother tongue into a sentence but rather call on certain of its resources to say what we have in mind. That is to say, certain of the resources made available by tongue must be activated in order to produce the discourse appropriate to whatever particular situation the speaker has in mind to talk about at that moment. It is this calling on certain formative elements, both lexical and grammatical, as well as their signs, to form words that represent and express what we want to say that constitutes the actualizing processes involved in any act of language. Thus to use a verb involves calling to mind the particular lexical element required (for example, in the present sentence the notion ‘involve’) with the appropriate tense, mood, person, etc. made possible by the system of the verb in English. (See my 2007 study on the verb for a detailed description of how a verb is formed.) And since the grammatical components – morphemes as Guillaume calls them4 – are organized in systems, and these systems integrated into the more general systems called parts of speech, we are led to view the grammatical side of tongue as “a system of systems” in Guillaume’s terms,5 as an organized set of linguistic systems each permitting certain mental operations for producing words. To help bring out the implications of this, it is worthwhile citing a passage distinguishing cognitive grammar from generative grammar: “As conceived in the present framework, the grammar of a language is simply an inventory of linguistic units. A grammar is not a “generative” description, providing a formal enumeration of all and only the well-formed sentences of a language. Nor do I employ the process metaphor and speak of the grammar as a device that carries out a series of operations and gives well-formed sentences as its output” (Langacker 1987b, 63). For Guillaume, on the other hand, a grammatical system is the dynamic potential for carrying out a series of operations to produce words in order to construct a sentence expressing one’s momentary experience (and not to conform to what a linguist might consider “well-formed”). His grammar is, however, not subject to the criticisms levelled against generative grammar (cf. Langacker 1987b, 64f) since it is generative in the common, non-mathematical sense and, as we shall see in the next chapter, it is first and foremost a means of producing words, and only as a consequence a means of constructing a sentence. A word is not a ready-made item in an inventory

24

Language in the Mind

but rather a made-to-order product, reconstructed on each occasion for use in the sentence under construction. In this respect, Guillaume’s position is quite different from that of both generative grammar and cognitive grammar. Thus, to understand the Psychomechanics of Language, as Guillaume’s theory is generally designated, it is essential to shift one’s perspective from the habit of thinking about language merely as an end-product, a text or speech, to be described, or merely as a set of items to be used in a sentence. It would be more realistic to take into account the means of production, that is, to consider language as a dichotomy opposing ability and sentences actually said. But even this is insufficient. To grasp the whole reality of language it is necessary to view it as a triad, a threefold entity wherein the operational component, language-as-operations, has its place between language-as-a-potential and language-as-actualized. This can be summarized in the following manner: actualizing language = language ability + operations + actual output

But even this formula, which suggests that one component is simply added to another to make up the whole, is inadequate because the third component is the outcome of the second, which is itself permitted by the first. We can bring out these condition/ consequence relationships more clearly with a more specific terminology and by depicting the whole of language as essentially dynamic in the following way: language = systemic potential

representing and expressing a given experience

sentence produced

The effect of introducing the dynamic or operative element into our consideration of language is to bring out the condition/consequence relationship between potential and actual by making explicit the temporal dimension involved in any operation. This gives a view of language in synchrony quite different from that proposed by Saussure. As a result of ignoring the temporal dimension, he came to consider language as something immobile, static, reduced to a thing, an object. In reality language is anything but static: it is dynamic from beginning to end. This operativity is obvious in the actual pronouncing of the sounds, but it includes far more than this. There are the operations involved in the syntax, all the

The Ability to Speak

25

preconscious mental operations required to relate the meanings imported by different words to give rise to the meaning of the sentence. And prior to this syntactic operativity there are all the representational operations involved in actualizing the meaning potential of the lexeme and of each of the morphemes to produce the particular meaning and syntactic potential of a word permitting it to play the role required of it in the sentence being constructed. Both language-as-a-potential (tongue) and language-as-actual (discourse) are realities, but neither should be considered as an object in itself: for the linguist, the former is a system of potential operations and the latter the result of a series of operations, a sentence (or a number of sentences).6 Thus what the above equation expresses could be formulated differently: language = actualizable operations

actualizing operations

result of actualized operations

Although the time involved in representing and expressing one’s experience is extremely short, keeping this temporal dimension in mind enables linguists to focus on the reality of language. Here again a basic tenet of Guillaume’s is paralleled in contemporary cognitive work: “Conceptual structure emerges and develops through processing time; it resides in processing activity whose temporal dimension is crucial to its characterization” (Langacker 1997, 249; emphasis in the original). As we shall see in the next chapter, it is the processing time, the time involved in any activity or operation – what Guillaume calls “operative time” – that provides him with the parameter for analyzing language activity. Since he often speaks of subconscious mental “operations” or “processes,” it is well to point out at the outset that he uses these terms in their common acceptation (i.e. as what is necessarily implied when linguists speak of language as “dynamic”), on the understanding that the processing time involved, the microtime of any mental operation, is so short it escapes perception. It is important to understand why an awareness of the temporal dimension keeps one in touch with the reality of language. A sentence cannot be seen as an object in itself but only as the final phase of an act of speech, the outcome of a speaker putting a momentary complex of experiential impressions into language. As speakers we always “language” some content of consciousness (i.e. put some experience into language), and, in fact, we cannot talk about anything but what we have in mind. This languaging, this

26

Language in the Mind

translating of some extralinguistic experience into the meaningunits provided by the words of one’s language, is part of the causal chain – a part often overlooked by linguists – giving rise to any coherent discourse. Since speaking necessarily involves talking about something one has in mind, nonsense strings – sets of words that have no correlate in the experiential awareness of the person stringing the words together – should not be confused with sentences, which are the outcome of real language processes actuated by speakers in order to represent and express some content of experience they want to talk about. It will perhaps bring out this fundamental point more adequately if we consider for a moment what is, for many linguists, the most remarkable characteristic of language: the capacity it gives us to represent and express whatever comes into our minds. This extraordinary ability – to give a linguistic form to something outside language by translating a passing experience into word meanings and saying it – is far more than a mere object or thing. My mother tongue is constantly meeting different requirements, adapting to new experiences, permitting me to represent and express novel messages – and, of course, to understand novel messages (like Hairsations on a shopfront, or scentsations in an ad). No system of post-factum rules, no commonly accepted norm for speaking, no set of sentences or inventory of linguistic units however numerous – none of these static ways of viewing language-as-a-potential could possibly account for the give and take of actual usage. This is the reality that a description of language in synchrony must deal with. Every time we want to use a morpheme we must actualize it again; every time we use a lexeme to represent and express something, we must form it again as a word along the lines provided in the system in tongue. What is called for, as Guillaume (cf. 1984, 79–99) so often stated, is to abandon the idealized, non-temporal view of synchrony left us by Saussure and reintroduce the parameter of time as the basis of a method of analysis, as the measure of the real operations that commute a speaker’s extralinguistic experience into something said. THE CHALLENGE: THINKING OF LANGUAGE AS A PROCESS

If this way of envisaging language does bring us closer to reality, one may well wonder why it is not adopted by all linguists. One reason is that it goes against the grain of a certain type of empiricism

The Ability to Speak

27

that would reduce linguistics to what can be induced from measurable data, as O’Kelly (57–61) makes abundantly clear. The experience of other sciences suggests another reason: the very fact of introducing the dimension of time into an area of reality hitherto seen as static is sometimes accepted with difficulty.7 The heritage left by Saussure of viewing language in synchrony as static combined with the lack of any method for analyzing the “psychological reality” of language operations helps to explain this difficulty in linguistics. But the reasons appear to go even deeper. For one thing, we acquire our mother tongue so naturally that we simply take it for granted and are normally quite unaware of the language potential we possess. Moreover, because almost all of the languaging processes are preconscious and so escape any direct observation, we see only their results, the sentences making up discourse, especially as written. Thus it is not surprising that, in the naïve view of the ordinary speaker, language is something static, like an object. And so when we learn the word language as youngsters its concept is formed with this limited scope, and this suffices for the needs of the vast majority of speakers. For scholars who focus on the study of language as such, however, adopting this common term as part of their learned vocabulary should lead to an examination of its popular sense so that they can be fully aware of the concept of language they are adopting in using it. Linguists can and should extend its meaning to cover all the reality of the object of their study, and this requires a conscious effort. This is what Saussure attempted to do, with only partial success, as we have seen, because he failed to bring in the operative component. Although some linguists have not attempted to do so, if one can judge by those who characterize language as “a set of sentences,” the fact that many do characterize language as “dynamic” shows that this static view is being superseded, even if not all linguists use this term in the same sense, as we shall see below when examining the type of process or operation it implies for Guillaume. In short, if we are prepared to do violence to our ordinary-speaker, uncritical assumption as to the nature of language by keeping clearly in mind a language-as-a-potential component and introducing a language-as-operations component in this way, we can begin to grasp the reality of language as a whole. Among the other theoretical approaches that view language as dynamic, many focus on syntax because here the operational is most clearly manifested in a language like English. Guillaume goes

28

Language in the Mind

far beyond this, as we shall see. He applied the notion of operativity not only to the constructing of the sentence but to the constructing of the word, focusing on the operational in morphology.8 He even proposed that the meaning of a morpheme or grammatical word involves a movement or process. This introducing of a dynamic view throughout the meaning side of language will be shown repeatedly below. It all resulted in a theory centred around the word, but before going on to that discussion a brief comment on terminology will be useful. Experience shows that, even after accepting the point of view that language is inherently operational, one is constantly confronted with the danger of lapsing back into static ways of thinking and speaking about it, simply because the ordinary-speaker “static” sense of the word language is constantly with us.9 This is why the terminology adopted to designate the three components, or better, phases of language is so important. The term discourse, quite widely used for both speech and text, is appropriate to designate languageas-actualized. There is no term in common usage to designate the operational phase of representing one’s experience of the moment and expressing the representation of it through speaking or writing. The term wordage, in the sense of “the use of words,” is far too rare. Less recondite, the term languaging in the sense of “putting one’s experience into language” is the best found so far to designate language-as-operations. Most important and most difficult in this “struggle for words,” as Einstein (1981, 327) in a similar situation put it, is to find a term for language-as-a-potential. It is hardly satisfactory to use the French term langue, which at best in English evokes the static view of Saussure. To use the term language itself, as in Gardiner (passim), is misleading because, besides using the name of the whole for one of the components, it is of no help in calling to mind the new view proposed by Guillaume. Even expressions like language as system or potential language, which point to the hidden part of language, would simply tend to make one think of resultative language, discourse, from another point of view. The advantage of tongue, the term adopted here, is that it designates language-as-a-potential as existing in its own right distinct from language-as-actualized, and so obliges us to enlarge our concept of language beyond the common-usage, resultative sense. Although some people have trouble accepting tongue in the sense of “the power of communication or expression through speech” (Webster’s,

The Ability to Speak

29

s.v.) as in expressions like the mother tongue and tongues of men and of angels, it is the most appropriate substantive available in English. Any initial hesitation in using it in this sense is perhaps a reflection of the fact that the reality it denotes is something hitherto unfamiliar and so easily overlooked. The following diagram will serve to suggest this view of language as a three-phase act carried out every time someone speaks10: tongue

languaging

discourse

By representing the reality of the language phenomenon in this way, we see these successive phases as links in a chain of causality.11 This provides synchronic linguistics with a solid basis for explanation, one shared with the other sciences based on observation because, as Aristotle remarked, “Here, as elsewhere, we shall understand things best if we consider them as they emerge from their origins” (cited in Waldron 1985, vi). In any case, since this condition/consequence approach works well in other sciences, and in comparative grammar, it should be exploited to the full elsewhere in the science of language as a means of reconciling the legitimate claims of both structuralism and psychologism in linguistics12 without sacrificing either the reality of the individual act of language or the reality of the hidden system. In short, it appears that the best way of reaching “an understanding of language that is grounded in reality” is to consider it “a commutation … that translates one’s momentary thought into speech” (cf. Guillaume 1995, 305).

3 Words, Words, Words Even the most detailed works of grammar have lost sight of the fact that the word has its architecture, that it is a system. This is nevertheless an important grammatical fact, the most important of all, and yet it is passed over in silence. Guillaume (1990, 60)

SENTENCES AS RESULTS

The mere observation of people speaking with such ease of whatever they happen to have in mind is a cause for wonder, and so for scientific investigation. It has led us to our basic postulate that for each speaker language is a potential (tongue) permitting innumerable actualizations (discourse) and that, as a potential, language is somehow systematic. This entails enlarging our initial conception of what a language is to include not only obvious, observable, actualized language but also what this necessarily implies, the nonobservable system and the partly observable languaging processes1: tongue

languaging

discourse

Adopting as our starting point this view of language as a threefold reality appears inescapable because it arises from that habit of our everyday thinking whereby we attribute to anyone carrying out an activity the ability to do so. Furthermore, this would seem to be a sound basis for analyzing language in order to understand and explain it since, it appears, science is a refinement of common habits of thought, such as seeking causes. We therefore adopt this postulate and then try to discern just what is involved in tongue and in the languaging processes that result in discourse. It is also part of our everyday thinking to consider language in terms of words and sentences, both words and sentences being observable features in the discourse of all languages.2 The fact that

Words, Words, Words

31

any extended speech or text can be divided into sentences shows why the sentence is commonly considered the unit of discourse: a sentence can be integrated as a component into a bigger whole. Furthermore, we sometimes observe that a sentence is interrupted, but then we get the feeling of something missing. The very fact of getting an impression of incompletion in such cases indicates that for the ordinary speaker the sentence is somehow recognized as a form constituting a whole, a unit. This is reflected in the traditional description of the sentence as “the expression of a complete thought,” a tradition stemming from the classical grammarian Dionysius Thrax, for whom “A sentence is a combination of words displaying a self-sufficient meaning.”3 Such descriptions evoke in a superficial way what a sentence does but do not go very far in telling us what a sentence is. That is to say, the sentence certainly does have the function of saying something about what the speaker has in mind and expressing it as a whole in some way, but this observation merely poses the problem for the linguist: how to reach a knowledge of the way a sentence is put together, of its makeup permitting it to function in this manner. Superficial though it may be, this traditional description is nevertheless invaluable as a starting point in common experience because it leads us to ask what preconditions must be met for a string of words, or in some cases a single word, to be able to express a complete thought and so to be a sentence. This working back from an observed result in an attempt to get an understanding of how something is put together is a common procedure both in everyday thinking – one is reminded of the natural curiosity of a youngster on observing the functioning of a watch – and in science, as shown by the well-known closed watch example given by Einstein and Enfeld (31) to illustrate the work of the scientist. As Guillaume remarks (1992, 97): “the linguist must learn to turn observed results back into genetic processes, in other words, reconstitute the mental process leading up to the result, the process whose conclusion is the observed result.” WHICH COMES FIRST: THE SENTENCE OR THE WORD?

The first step in examining any complex entity like a sentence is to analyze it, to break it down into its component parts, since the way anything functions is conditioned by its makeup. Applied to

32

Language in the Mind

sentences, this procedure results in what is evident for the ordinary speaker: the elements of the sentence are words. Commonplace experience tells us that no sentence can exist without words, or at least a word, and that all sentences can be broken down into words. At the risk of belabouring the obvious, I insist on this point because some linguists focusing on how the sentence is constructed have tended to neglect it and as a consequence have failed to see the importance of the word in every analysis of language. Thus we are led to the view that, just as the sentence is the building block, the unit, of discourse, the word is the unit of the sentence because a word is a whole that can be integrated into a sentence. But there is a difference: a word has a particular grammatical function when used in a given sentence and this raises the problem of understanding how a word can fulfill this role. Again we must adopt the procedure of analyzing the word, of breaking it down into its component parts to see how it is put together. Only when we get a view of the formative elements of different words can we hope to explain how words function and, as a consequence, explain how a sentence fulfills its function. In short, as a necessity of analysis we must obtain some understanding of the nature of words in order to understand the nature of a sentence. A theory of the word is a prior condition for a theory of the sentence because morphology (in the sense of the grammatical components of a word’s meaning) conditions syntax. The many attempts to analyze the sentence without a prior analysis of the word have resulted in syntactic analyses, descriptions of the relations between words often expressed in the form of rules, but they provide no explanation for the observed relationships. For example, one can formulate rules to describe usage in cases like He forced me to do it and He made me do it, but to understand why forced takes to with the following infinitive, whereas made does not, requires an analysis of the words involved – the lexical meanings of force and make, the meaning of to and the grammatical meaning of the infinitive (cf. Duffley 1992a) – because a relationship is largely determined by the makeup of the entities involved. “It follows from this that, with a good method, in a sound linguistics, any study of the constructional mechanism of the sentence will be subordinated to a prior consideration of the structure of the word” (Guillaume 1971, 30). This is why Guillaume’s fifty years of reflection on language were devoted largely to working out a general theory of the word, and

Words, Words, Words

33

particularly of the different types of word found in languages that have established parts of speech as subsystems within the general system of the word. This focus on the word is, as we have seen, distinctive of his approach, which has sometimes been criticized for neglecting the study of the sentence by those who do not see that morphology conditions syntax. His word-based grammar follows from the potential/actual viewpoint outlined in the preceding chapter since the potential of tongue consists essentially in the system for constructing words. And this system, Guillaume found, is by no means the same for all languages. On the contrary, he saw that language types are to be distinguished on the basis of the overall system for constructing words, the most general system of any language. In any psychomechanical analysis, therefore, we first focus attention on the words involved. Only in this way can one come to understand how they work in the sentence and so get to the point where the syntactic processes involved in constituting the sentence can be described. The ordinary speaker is at least minimally aware of both words and sentences as forms. Since speaking as we know it would not be possible without them, it is part of the linguist’s task to understand and explain them. Ultimately, finding a satisfactory answer to these two fundamental questions – what is a word? what is a sentence? – will carry us a long way toward understanding what language is. THE FUNDAMENTAL UNITS OF LANGUAGE

Turning to the word, then, we might point out that our grammatical tradition provides no description of the word as it does in the case of the sentence. A reason for this may be that native speakers of English make practically no mistakes in forming words, only one word in a million according to one study.4 The fact that word fragments are simply not part of the ordinary Anglophone’s experience of language is perhaps the reason why so little reflection on the word as a unit, on what it takes to have a complete word, has arisen. Words always surface into consciousness well formed, so as ordinary speakers, and even as linguists, we tend to take them for granted. This is not, of course, to deny the importance of many contributions in the fields of lexical semantics, grammatical semantics, or word composition. The point is rather that there has been comparatively little concern with the system of the word, the word recognized as a unit by the ordinary speaker. One acute observer

34

Language in the Mind

of language puts the question this way: “Why is it that the element of language which the naïve speaker feels that he knows best is the one about which linguists say the least?” (Bolinger 1963, 113). Clearly the word poses a problem for linguists, or should pose one at any rate. If a sentence is a combination of words expressing a self-sufficient meaning, a complete thought (whatever that may be), what is a word and what does it do? In his Word Grammar, Hudson (4) alludes to the problems involved in defining the word in terms of recognizing its boundaries for the needs of linguistic analysis. But the question goes beyond delimiting it on the level of the sign to the very nature of the word as a mental unit, as Saussure (1916/1955, 154; my translation) had suggested: “the word, in spite of the difficulty one has defining it, is a unit which imposes itself on the mind, something central in the mechanism of tongue.” And much more recently, George Miller (5) raised basically the same question from the point of view of the word’s function: “Why are all languages wordy? Why are words a universal design feature of languages?” Like sentences, words are found in all languages and so their role must also be quite general. Even more: words are found when there is no sentence, only sentence fragments, since there can be no expressing by means of language, no act of speaking or writing in any language without words. (Hence the irony bordering on insolence in Hamlet’s response to Polonius, cited as the title of this chapter.) Words are as general as language itself and are, in fact, “the fundamental units of language” (Miller, 261), not in the sense of the smallest elements arrived at by linguistic analysis but in the sense Sapir (33), in the light of his experience with young speakers of Nootka, describes the word to be “a psychological reality,” “the existent unit of living speech.” Guillaume early recognized the primacy of the word, perhaps as a result of reading Saussure, but he did not leave the problem there. In fact his constant concern over the next forty or so years was to understand what is built into a word making it a recognizable unit and permitting it to fulfill its particular role in the sentence. The originality of his whole approach to language is based on his answer to this question: what is a word? First of all, therefore, we will attempt to bring out in a general way his view of the makeup and the functions of the word. Among those who have attempted to define the word, it has long been recognized that a word has a physical existence making it a sayable unit. This has been expressed in various ways: “the smallest

Words, Words, Words

35

speech-unit,” “a minimum free form” (both cited in Ullmann, 51), “a combination of vocal sounds, or one such sound … a vocable” (OED), etc. That is, from the point of view of their physical side, the words of a language are the smallest units that can be pronounced in ordinary speech. Although there are borderline cases where we may hesitate – cases requiring further study – the pronounceable character of words helps us distinguish a word from a suffix or prefix or root and its minimalist character distinguishes it from a phrase or sentence, except in cases of one-word phrases or sentences. The physical side of a word is fully actualized in ordinary speech but is only minimally actualized in that inner dialogue that accompanies much of our thinking, a situation that is reflected in the common remark: “There was so much noise I couldn’t hear myself think.” Even in the case of inner dialogue, however, the physical side fulfills its function of making the word perceivable, though in this case it is only made mentally perceivable and just to the speaker. The very fact that we can distinguish these two degrees of actualization indicates that the physical side of the word exists for the speaker not just as a set of sounds but also as a potential, as a set of phonemes drawn from the phonological system of the language. From this point of view, then, the word as a potential can be described as a sayable element of discourse, or, most succinctly, a vocable. It is also commonly accepted that a word is not just a set of sounds but has a mental side as well. To bring out this binary nature of a word, it has been described in the OED as “an ultimate minimal element of speech having a meaning as such,” and by Gray (cited in Ullmann, 51) as “the smallest thought-unit vocally expressible.” That is, the sounds constitute a sign that calls to mind a meaning; hence the word is a physico-mental unit. When we hear a word spoken, it is the physical part, the sign, we hear, and this sign, being permanently linked to a certain meaning, calls its meaning to mind so that we immediately become aware of the notion or idea the speaker is expressing. Likewise, when as speakers we have some meaning to express, we actualize the sign associated with it. At the very heart of the word’s makeup, then, we find the well-known signmeaning liaison, which Guillaume (1989, 13f) describes as follows: We have here a symphysis, a remarkable mental welding thanks to which a fragment of speech automatically calls to itself a fragment of thought, and conversely the fragment of thought calls up the fragment of speech.

36

Language in the Mind

The liaison is reversible, and this reversibility makes it possible for discourse to be understood by the person we speak to. This second person, to whom we speak, hears the fragments of speech and simultaneously evokes the corresponding parts of what is thinkable in tongue. The liaison is a reality in tongue.5

This is a two-way relationship that turns out to be more complex than it may appear at first sight. One complication arises from the fact that this relationship is not symmetrical. For the listener the sign is a means of calling to mind its meaning; however the meaning, once evoked, makes the listener think not of its sign but rather of what the speaker has in mind, what the speaker is talking about, the message. The sign is at the service of the word’s meaning, but not vice versa since we use a word for the meaning it can express and not for the way it sounds (unless we happen to be a poet). That is to say, the use of a word can be explained in terms of the meaning it expresses. In this respect, there appears to be much in common between psychomechanics and cognitive grammar: “The most fundamental issue in linguistic theory is the nature of meaning and how to deal with it” (Langacker 1987b, 5). Meaning is primordial in the word because the notion or idea a word expresses is linked to the message, to what the speaker intends to communicate. It is this link between the speaker’s intended message and the meaning of a word that justifies the use of a given word in a given act of language, a fact that throws light on another of the word’s functions. To make this clear and to bring out an important principle of psychomechanics, we must now examine more fully this relationship between word meaning and what the speaker has in mind to talk about. FROM MESSAGE TO MEANING

Few of us have ever stopped to reflect on how many different things we can talk about. We probably feel that the range of possible subjects of discourse available to us is quite unlimited because we can talk about anything that comes into our mind. As Linda Waugh (30) points out, citing Jakobson: “It is not the world outside which we talk about but an inner world, the ‘world of human experience, the world as it is perceived within us, the universe of discourse.’” And

Words, Words, Words

37

indeed, that seems to be the only limit: something must occur to us, must come into our consciousness, before we can talk about it. Conversely, in order to speak we must have something in mind to talk about. This intended message may be a passing thought or feeling, an outcome of reasoning, something arising from immediate perception or memory or imagination, from what someone has just said or from some other source, or any combination of these – in brief, we can talk about anything provided we are conscious of it, provided it is part of our experiential awareness. This particular momentary experience is not part of our language and so it exists whether we talk about it or not, but unless we represent it linguistically or in some other way, this experience remains strictly private, incommunicable, unsayable.6 This is why it is important to have some idea of what is involved in the process of representing something linguistically, of making the unsayable sayable. In order to talk about something, in order to make it sayable, therefore, we must call on the resources of our language, which have been instituted precisely in order to depict the momentary content of our stream of consciousness in such a way that we can grasp it more readily ourselves and make it available to others. That is, our language matches the complex of impressions making up our message with the lexical notions of the most appropriate words available, meanings that are shared by other speakers of the language. In short, as speakers we represent our extralinguistic experience thanks to word meanings and we express the resulting linguistic representations thanks to signs. Tongue provides the resources for this: the meanings representing experience, the means for forming them into grammatical units (words) and the accompanying signs to make the units sayable. The outcome of actually saying them in sentence units is discourse. This then is the primary function of words: representing the intended message, translating one’s unsayable experience of the moment into linguistically sayable units of meaning. Without someone representing their own experience in this way, no actual language can exist. This fact explains why words are “a universal design feature” not only of all languages (i.e. as potentials, as instruments ready for representing) but also, as actualized representations, of every act of language a speaker undertakes. Furthermore, this representing of our raw experience of the universe in a manmade form involves situating it in the conceptual framework of one’s

38

Language in the Mind

language. That is, each language formats the experiential data in its own way, as manifested most clearly by the more or less appreciable differences of vocabulary. Thus in learning their mother tongue, infants set about acquiring an inner universe of notions – a universe, because it is constituted for representing everything that may arise in our experience of the universe around us. This confrontation between the idea-universe within tongue and our experience of the universe outside the mind gives us a degree of independence from the inroads of the world around us and helps explain why, for Guillaume and many other scholars, language is the characteristic distinguishing human beings from other animals. “Man is man through language alone,” according to Humboldt (cited in Waldron 1985, 184). A contemporary philosopher speaks of being “constantly forced to the conception of man as a language animal, one who is constituted by language” (Taylor 1985, 246). Guillaume (1984, 157) develops this idea as follows: … accounting for the existence in thinking man of an expanding ideauniverse, 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 discourselanguage, and their discourse, with no mediation of representations, proceeds directly from experience.

We shall return to Guillaume’s thoughts on the humanizing function of language in the final chapter. REPRESENTATION AS A HALLMARK OF THE GUILLAUMIAN APPROACH

The point I am trying to make here is simply this: what we have in mind to talk about, our extralinguistic experience or knowledge, is to be distinguished from its linguistic representation as expressed in a sentence, just as one’s grandmother is quite distinct from some painting or photograph representing her. There is of course a similitude between what is represented and its representation, with the consequence that this distinction is not made by some linguists. Stern, for example, identified meaning with our experience when

Words, Words, Words

39

he remarked (45): “The meaning of a word – in actual speech – is identical with those elements of the user’s (speaker’s or hearer’s) subjective apprehension of the referent.” As a consequence, he adopted the position that “when the word camera is used of different cameras, the meaning changes in correlation to the change of referent” (40). Identifying linguistic representation with our everchanging experience in this way would make it impossible not only to consider a word as a meaning-sign unit, as a permanent potential in tongue available for use whenever required, but also to explain communication as we know it. This is why it is crucial to keep in mind the distinction between the extralinguistic intended message and its linguistic representation, a distinction that is not a “false dichotomy” but a reality of fundamental importance. This point is so important that I want to develop it in the larger context presented by John Taylor (1996, 26ff). To contrast Chomskyian and Langackerian approaches, he recalls the former’s postulate of a genetically determined language faculty distinct from conceptual and pragmatic faculties, with the result that linguistics is primarily concerned with “linguistic knowledge proper, i.e. linguistic knowledge ‘purified’ of conceptual and social knowledge” (27). Thus language processes operate independently of the “central thought processes” and as a result the linguistic meaning produced by the language faculty has nothing in common with conceptual and pragmatic knowledge other than their mental origin and link with language. In contrast with this modular approach, Taylor (28) maintains that cognitive grammar “rejects a strict compartmentalization of linguistic and non-linguistic knowledge. Language is assumed to be essentially symbolic in nature. Meaning is equated with conceptualization, and language serves to encode a speaker’s conceptualizations.” As a consequence, the linguistic meaning expressed by a sentence is not “ontologically distinct” from “a person’s conceptualization of a state of affairs.” In this he echoes Langacker’s characterizing such distinctions as “false dichotomies” (1987b, 154). The Guillaumian approach is crucially different from Chomsky’s since it considers that language is a product of “central thought processes,” not of genetically determined factors. Like the cognitivist and Jakobsonian approaches, it takes for granted that language is essentially symbolic since the purpose of any act of language is to call to mind the “state of affairs” the speaker has in mind.

40

Language in the Mind

However, Guillaume insisted on distinguishing between the particular experience or state of affairs and meaning, its representation by means of words. As he never tired of repeating: “no expression without representation.” That is, language does more than simply “encode” our experience (in the narrow sense of ‘providing a sign for’), a process that can also be accomplished by non-linguistic means, as when we utter a sigh of relief, raise a thumb to get a ride, etc. Language first translates a speaker’s particular mental experience into an abstract mental representation of that experience thanks to the words the speaker has learned to construct – words consisting of a sign and a double meaning input, lexical and grammatical. This is why people who speak different languages (and even those speaking the same language) represent the same situation or experience differently. Even granted this dichotomy between experience and representation in psychomechanics, and even though a given meaning can be rightly called a concept, we would not equate representation and conceptualization, meaning and “non-linguistic knowledge.” Not making the distinction would imply that human thought can operate only within the framework provided by language, whereas Guillaume maintained that it was thanks to our faculty of thought that language came into existence and developed. Indeed, those exploring new concepts have sometimes, as Einstein (1981, 327) put it, “struggled for words” to express them. We shall explore the double relationship between thought and language in chapter 15, the point here being simply that our capacity for thought cannot be limited to what the state of our language at a given moment permits. This is why, in order to make representation possible, there is no need to postulate the existence of a “language faculty” as in a Chomskyian approach. The purpose of the symbolic role of language is not to provide a makeshift means of expression (as when we try to communicate with someone whose language we do not speak) but rather to provide permanent means of expression readily available regardless of the momentary state of affairs or our experience thereof. For this reason, not sentences but words – more precisely the formative elements and processes for constructing words – are primordial. Being permanent resources, their meaning, both lexical and grammatical, is general, not tied to any momentary experience. This need to represent experience in a general way calls on the “central thought processes” to institute

Words, Words, Words

41

these resources by abstracting a durable mental schema or image from particular experiences. In short, Guillaume is led to postulate that it is not a specifically linguistic faculty or a series of particular mental abilities but the fundamental human faculty to generalize and to particularize that enables us to institute language and use it. To meet the requirements of expression over the ages, this innate faculty has enabled the speakers of English to constitute the means of forming diverse words as permanently available resources, as potential elements for constructing sentences. But if words express this abstract, schematic meaning, arrived at through conceptualizing experience, does this not put psychomechanics back in the cognitivist camp? In the most general sense of the term psychomechanics is “cognitivist” since it considers language both a product and an expression of human mentation, but it cannot be identified with the particular approach to language called cognitive grammar. This can be shown by considering Taylor’s rhetorical question (1996, 28): “What … would the linguisticsemantic meanings of words like leave and university be, if not the concepts designated by these words in actual utterances?” From a Guillaumian point of view, what a word “designates” in a sentence is something (an activity, an institution, etc.) in the speaker’s momentary experience, something outside language that has been represented thanks to the word’s permanent meaning potential or “concept.” What a word signifies in an actual sentence, on the other hand, is only one of the realizations of its meaning as a conceptual potential in tongue, a potential that guarantees the word’s existence for a speaker because in itself it is unchanging, though actualizable in various ways when used. That is, the actualized meaning of a word, observable in a particular sentence, must be distinguished from its potential meaning, which can never be observed since it cannot emerge from the preconscious mind (a lexeme or a morpheme can never be called on to express all its possibilities at once). Thus the distinction between potential meaning and actualized meanings (or senses), and more generally between tongue and discourse, sets psychomechanics off from other cognitivist approaches that do not call for a process of representation between a speaker’s experience and the meaning of words expressing it. Furthermore, this distinction provides a general linguistic basis for meeting the crucial problem of polysemy, a basis quite distinct from the psychological notion of prototype. In psychomechanics the

42

Language in the Mind

relationship between potential and actual, which is not simply that between a general meaning and specific instantiations, or between “basic” and “peripheral” meanings in Jakobsonian terms (Waugh, 72), provides for a unitary meaning as the necessary basis for human communication and, derivable from that meaning, various senses to meet the endless variety of our experience.7 Distinguishing potential and actual leads to viewing syntax in a quite different light as well. On the basis of the various ways words function in English, psychomechanics, like other approaches, concludes that there are nouns, verbs, prepositions, and the like. But here the similarity ends. Each of these word types is seen, not as an element manipulated by syntactic rules but as a word-forming process in tongue, a grammaticizing process enabling a word to play a certain syntactic role. That is, each part of speech is seen as a grammatical potential for shaping the lexeme to produce a word capable of carrying out the function required of it in the projected sentence. The syntactic function of a word is thus made possible by the word’s grammatical structure, its part of speech. That is, underlying each use is the most general system of a language, the system for forming words, with its part-of-speech subsystems (in those languages where the part of speech is built into the word). For Guillaume grammatical structure is thus primarily morphological, or better morphogenetic, i.e. a word-forming process. This is why he considers that in English, which has so few inflections, grammatical structure is largely covert, whereas for those, like Langacker (1987a, 26), who view it as primarily syntactic, “grammatical structure is almost entirely overt.” A final point of comparison here concerns the role of background or non-linguistic knowledge for the listener in understanding a sentence. For Taylor (1996, 77), since “expressions (and parts of expressions) can only have meanings in the context of background knowledge” it is “an intrinsic aspect of meaning” to the point that, in its absence, a verb like open “ceases to mean anything at all.” In psychomechanics, on the other hand, a listener or reader, quite ignorant of the context or situation of a sentence, may well understand its meaning, what is being said, but have no idea of what is being talked about, i.e. what the speaker or writer was referring to. Reference is the process of applying the linguistic meaning to the extralinguistic world of experience. This point, like the others mentioned above, will be developed more fully in what follows.

Words, Words, Words

43

REPRESENTING FOR USE IN A SENTENCE

A word fulfills its representational function in two ways. Besides its lexical component, whereby the word provides, as we have just seen, a conceptual or ideational representation of something in the speaker’s momentary intended message {I am now considering a word within the very brief moment when it is being formed – called to mind – to be integrated into a sentence being constructed), the meaning of a word has a grammatical component often expressed by grammatical inflections. Thanks to this grammatical meaning, a word can fulfill the particular function foreseen for it in the sentence under construction. That is to say, in a partof-speech language like English, if the purposed sentence requires a word that can assume, say, the specific role of predicating, then the word’s lexical component is formed by means of mood, tense, person, etc. in order to result in a finite verb, a part of speech whose role is precisely to predicate something of the subject. On the other hand, if a word is to serve as representing what is talked about in the purposed sentence, its lexical meaning must be categorized by means of gender, number, and case to result in a substantive8 or a pronoun, parts of speech that can assume the role of subject. Or again, if the word is required to represent a relationship between a noun phrase and some other part of the sentence, its lexeme must be formed to produce a preposition, the part of speech capable of playing this role. These examples help bring out the fact that a word’s function in a given sentence is made possible by the part of speech. That is, each part of speech is a potential making possible certain syntactic roles, and the system of the parts of speech provides for all syntactic relationships. This explains why no word can appear in discourse unless its morphological categorizing has been carried through to its most general form in a part of speech, though of course we may in some uses have trouble determining just which part of speech.9 The fact that all words in discourse belong to a part of speech suggests how general and abstract a system each part of speech must be. For this type of language, these are, in fact, the most general systems, systematically related to constitute a system of systems. This all embracing wordforming system provides the structure within which every word we use finds its form, thereby permitting it to assume its grammatical role in the sentence.10

44

Language in the Mind

Thus the preconscious act of representation involves two phases: ideogenesis and morphogenesis. ideogenesis is the set of processes involved in calling on the lexical resources of tongue to represent something in the intended message. This includes the operation of calling forth from the speaker’s idea-universe the particular notion or lexeme required, what distinguishes a given word from every other word in the language. It also includes the process of actualizing that notion, for example determining, in the case of a lexeme called to be a substantive, whether it is represented as ‘mass’ or ‘count,’ or, in the case of one called to be a verb, whether it is ‘stative’ or ‘dynamic,’ results that can be observed in discourse. An attempt to analyze the ideogenesis of the verb do is given below (chapter 11). morphogenesis 11 is the set of processes involved in calling on the system of the parts of speech to provide the grammatical meaning or morphemes required for any word to carry out the grammatical function foreseen for it in the sentence being constructed. These are categorizing, i.e. generalizing, operations that bring a particular lexeme actualized by ideogenesis to its most general grammatical form, the part of speech. As we will see in chapters 7 and 10, grammaticizing a lexeme as a substantive involves forming it through gender, number, and case, whereas to grammaticize a lexeme as a verb it must be formed by voice, person, aspect, mood and tense. To emphasize the fact that the two processes are integrated into the operation of forming a word and are concerned with meaning rather than with the sign, Guillaume speaks of them as “a synergy of two ideations: notional ideation and structural ideation or ideation of structure” (1997, 71). The outcome of ideogenesis and morphogenesis, of notional ideation and structural ideation, is a meaning-construct associated with its sign, an organized set of impressions that is sayable. When, during a given act of language, we construct, or better, reconstruct a word (we use the same formative elements and constructional processes every time we want that word), we translate something unsayable into something that is sayable. But at this point, at this precise instant in the process of sentence constructing, the word is sayable only, not yet said. That is, from the point of view of the different formative elements entering into its makeup, a word is an outcome, an actual unit resulting from activating the resources of tongue; but from the point of view of the sentence in which it is about to play a role, the word is a unit of potentiality, something to be said and

Words, Words, Words

45

integrated into a sentence, or as Guillaume pointed out (1990, 12): “It thus arises as a unit of potentiality with regard to what follows, i.e. the sentence, and a unit of actuality in relation to what precedes, i.e. the formative elements.” THE WORD AS A PROCESS

Regarded operationally in this way, the actual existence of a word is short-lived, ephemeral. All its formative elements and the processes required to combine them exist permanently in tongue, but this gives the word existence only as a potential. In discourse, in the sentence once completed, the word’s meaning has amalgamated with that of other words to form “a complete thought” so that the word as a meaning-sign unit no longer exists. Thus “words being processes, processes creating a linguistic result” (Guillaume 1992, 47), their actual existence is something transitional, arising only in the act of language as the output of an act of representation and the input for an act of expression. Indeed, one might well wonder if words really do exist, were they not a normal constituent of our everyday thinking and experience, especially in writing. Analyzing the word in this way throws light on the operational dimension of language in synchrony with its temporal parameter, a temporal parameter reduced to the space of an instant.12 But it is not easy to keep our attention focused on the operational in language: 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. (Guillaume 1984, 133)

The process of lexigenesis, which involves both the genesis of the significate (lexical and grammatical) and the genesis of the

46

Language in the Mind

sign, has a permanent existence in tongue for any speaker, but it exists as a potential, as a set of operations to be triggered by an intended message. Its actual existence, like that of any other process, is its coming into being. It must be analyzed in time, in the brief instant between its beginning and its end, because as a possibility lexigenesis is not yet a real process and once its result is obtained it is no longer a process. Inherent in this view of meaning is that a word, as observed in a sentence, is the result of a constructional process. Considering the word as a result only, one could maintain with Wierzbicka (1988, 8) that “There is no such thing as ‘grammatical meaning’ or ‘lexical meaning’” because the two components form a single mental entity. From this resultative standpoint, furthermore, “the notion of the meaning of a word in isolation is in any case a fiction” because “Meaning is conveyed by the utterance as a whole” (9), a view also proposed by Austin and Wittgenstein.13 Adopting this position is perhaps understandable for those interested in language only as a means of communication, but a concern for language as a phenomenon in itself, the concern of linguistics, calls for analyzing sentences and words in order to understand what enables it to function as a means of communication. And this leads to the view, first expressed by medieval grammarians, that words both signify a lexical meaning and “consignify” a grammatical meaning (cf. Michael, 46f). FROM WORD TO SENTENCE

If making representation possible is the primary function of the system of the word, then making expression possible is its secondary function. This is why words are constituted as binary entities made up of meaning and sign. Expression is secondary because, as Guillaume frequently stressed, we can express only what we have already represented. By this he meant that anything we want to express by linguistic means must first be represented by linguistic means. Not that our means of expression are limited to the linguistic: a cry of pain certainly expresses a meaning, but neither the experiential content expressed nor the sounds expressing it have been instituted in our language as a permanently available means of representation and expression. Between such obviously nonlinguistic means of expression as a cry of pain and the ordinary

Words, Words, Words

47

word there are various means of expression and in certain cases it may be difficult to discern the limit between what is instituted in tongue and what is not, a problem that calls for careful study of the means of expressiveness found in language usage (cf. Guillaume 1984, 84–8). The act of expressing what has been represented involves actualizing the signs whether orally (be it in rapid speech, carefully articulated speech, whispered speech, etc.) or scripturally or just mentally (in inner dialogue). It also involves actualizing the relationships between the different units of meaning provided by the words to integrate them into bigger meaning-units (phrases and clauses) and culminate in a meaning-whole, the “complete thought” expressed by a sentence. The need to establish syntactic relationships in most sentences arises from the inability of a single word to represent the whole of an intended message; except where a single word suffices (interjections, some imperatives, etc.) the speaker’s total experience must be broken down and analyzed into sayable parts and then reassembled syntactically to represent the original experience as a whole, as a self-sufficient entity. Actualizing the meaning relationships permitted by the part of speech of each word is the operational dimension of syntax, which therefore concerns the mental side of language, meaning. In order to bring the act of constructing a unit of discourse, a sentence, to a close, the various meanings must be amalgamated into one, what Guillaume calls the “summatory condition” for reaching a sentence, but there is no counterpart on the physical side: we do not blend all the signs into one, a sort of “complete sound,” because a sign is merely a means for calling to mind its own meaning and once this function has been carried out, the sign can be dispensed with. Thus an act of expression, the saying (or writing) of a sentence, is aimed at bringing out a complex of impressions, the sentence meaning. FROM MEANING BACK TO MESSAGE

Even when the sentence has delivered its meaning the act of language is not yet over. There remains one final operation without which language would not be able to confront extralinguistic reality and fulfill its communicative function. When we undertake an act of language, there is an intended message we wish to communicate to others (and often to ourselves, as in inner dialogue, when

48

Language in the Mind

we want to get a clearer mental grasp of what we have in mind). That is, the aim is to communicate a certain content of experience, not a sentence meaning. Once the sentence meaning has been obtained by the listener, therefore, it remains to translate it back into a content of experience as closely resembling the original as possible. The process of going from a linguistic representation, the sentence meaning, to an extralinguistic content of experience, also a mental reality, is called referring or reference. It is just the opposite of the process of representation discussed above, which went from experience to linguistic representation. Through reference, listeners attempt, in the light of the meaning expressed, to reconstruct the immediate referent, the content of experience that prompted the speaker to utter the sentence in the first place. Besides the sentence meaning, they make use of what they know about the speaker and the speaking situation, and in fact of any other knowledge in order to obtain the message the speaker wishes to communicate. Thus, although the immediate referent is something in the mind, an experience, it does not, in Guillaume’s eyes “belong to linguistics” (Jakobson, 320) since it is not part of language. This distinction between the sentence meaning and the resulting message – between “linguistic and extralinguistic knowledge” in Langacker’s terms (1987b, 154)– comes out clearly when we are unable to refer meaning to extralinguistic reality, as when renowned Egyptologist Alan Gardiner remarks (11): “Is it not something of a puzzle that especially in letters and in ancient documents of different kinds the meaning of the component individual sentences should often be perfectly clear, but that the reader should nevertheless be left in almost complete darkness as to what the document is really about?” In the case of inner dialogue, on the other hand, there is no need to reconstruct the intended message because we are already at least minimally aware of it. However, referring the sentence meaning back to the intended message does permit us, especially when writing, to compare what we have actually said with what we are talking about and if need be to rectify or improve the means of representing and expressing it. THE WORD IN LINGUISTIC ANALYSIS

As a lexical-grammatical construct, then, the meaning import of a word is a result of ideogenesis and morphogenesis. A remark by

Words, Words, Words

49

Lyons (1977, 19) makes this more explicit: commenting on the sentence The words FOUND and FIND are different forms of the same word, he points out that “The term ‘word’ is clearly being used in two different senses here.” Found and find are different by virtue of their grammatical meaning, their tense, whereas they have in common their lexical meaning. Speaking of them as the “same word” hints at their existence in tongue as a single lexeme ready to be actualized and to receive the grammatical form it will need in order to play its role in the sentence under construction. The two senses of the term “word” observed by Lyons thus give us a glimpse of the modes of existence of a word, as a potential in tongue, and as actualized for use in discourse. This glimpse is valuable because it helps clarify Guillaume’s surprising remark that there are no words in tongue: he is referring to words actualized as lexical-grammatical constructs in a sentence. The makings of a word do exist in tongue, particularly its lexical matter, which identifies any word as different from all other words in the language. Thanks to ideogenesis, the lexeme appropriate to the intended message is singled out from the whole universe of ideas and is then, thanks to morphogenesis, categorized, formed as a part of speech with the syntactic possibilities it will need in the sentence. To illustrate the way these two phases of a word’s ideation – notional and structural – work, the question of “conversion” will be discussed in chapter 8, once the system of the parts of speech has been presented. We are now in a position to appreciate Guillaume’s comment cited above, which remains as true today14 as when it was written in 1943: “Even the most detailed works of grammar have lost sight of the fact that the word has its architecture, that it is a system. This is nevertheless an important grammatical fact, the most important of all and yet it is passed over in silence.” In order to bring this fact into focus, I have proposed that all languages are “wordy” because words are the units of representation, units whose reconstruction is provided for in tongue to give unsayable experience a sayable representation in language. Furthermore, since this representation gets the grammatical form predetermining its syntactic relationships from the system of the parts of speech, the system of the word is central to linguistic analysis. The potential sign linked to the resulting meaning construct of each word involved can then be said (actualized) and this results in the actualization of the sentence as the expression of a “self-sufficient

50

Language in the Mind

meaning,” “a complete thought,” which is then referred back to the universe of experience. This general framework for analyzing the unit of tongue, the word, and the unit of discourse, the sentence, through the operativity of language in the instant of speech is sufficient to suggest why both are universals of language15 and why the word is a prior condition for the sentence. By considering the system of the word a fact, “the most important of all,” Guillaume ensures that his theory is based on a principle applicable to all human languages and to all acts of speech and writing. Furthermore, because the system of the word, of the vocable, is a necessary prior condition for that part of language which is observable, he ensures the scientific pertinence of his theory. This then is what constitutes the originality of Guillaume. He was, to my knowledge, the only linguist to focus his whole career on developing a theory of the word, revealing thereby a new dimension of language for linguists to explore, a new area for analysis: lexigenesis. This is why psychomechanics is quite different from an approach accepting the information about nouns and verbs provided in traditional grammars as “without question, substantially correct” (Chomsky 1965, 64) and an adequate basis for a theory of syntax. It would also differ from a “non-constructive conception of a grammar” (cf. Langacker 1987b, 65) if this excludes the repeated reconstruction of a word, but not from a grammar considered as “the successive combination of symbolic structures to form progressively larger symbolic expressions” (loc. cit., 97) if this involves the combination of meanings to form a word, the combination of word meanings to form a phrase, etc. In short, as a result of Guillaume’s theory, the challenge of thinking of language as dynamic is posed not only for phrases and sentences but also for its fundamental unit each time we speak. Guillaume’s approach provides a point of view for exploring words of different types. Although not all languages structure words on the same basis, it will not be possible to examine words with different structures in the present context.16 Rather, we will focus on English and try to describe several different kinds of word within its part-of-speech structure. But first, we shall explore the particular case of certain substantives and how they carry out their primary function of representation through grammatical number.

4 Meaning: Representing Experience1 … Trying to learn to use words, and every attempt Is a wholly new start, and a different kind of failure Because one has only learnt to get the better of words For the thing one no longer has to say, or the way in which One is no longer disposed to say it. And so each venture Is a new beginning, a raid on the inarticulate … T.S. Eliot (Four Quartets)

INTRODUCTION: USAGE VS GRAMMAR RULES

If we came across the expression one heads without any context, we would probably consider it either a verb with subject or a noun phrase with an error against the rule for using -s to indicate ‘more than one.’ Given the context where it actually occurred – Henry is certain to toss two heads or one heads or no heads… – and the fact that it describes someone flipping a coin in a book on chance,2 one might accept it. When one considers the change of meaning expressed if that -s is suppressed, there seems to be no alternative to one heads to express what the writer had in mind. These considerations are of little concern for the teacher of English language, because this use is too infrequent to be of practical importance. However, for linguists, or anyone else who wants to understand how the English language works, such facts provide an invaluable insight into one of the representational mechanisms of the substantive: the system of number. Like anyone else attempting to approach reality from a scientific point of view, linguists start with the assumption that there really is something orderly, coherent, systematic lying behind the observed facts because otherwise

52

Language in the Mind

they would be content merely to catalogue and classify uses but would not make any attempt to understand and explain the data – the observed facts – even such unusual uses as the above. A system is not, in itself, directly observable in discourse, so we can only infer it by abduction (the word is from Pierce3), that is, by working out the relationship between observed uses and then reconstructing it in imagination. Grammatical rules, often used in teaching, may afford an adequate description of the most frequent uses, those most easily observed, but cannot provide a view of all the manifestations of the system in discourse, nor can they provide a description of the underlying system. This is why it is important to seek out less frequent uses, like the one above, which are often more revealing when it comes to imagining the configuration of the system lying behind all uses. -S EXPRESSING BOTH ‘PLURAL’ AND ‘SINGULAR’: A HYPOTHESIS

In the case of grammatical number, this involves comparing and contrasting all uses of both -ø and -s morphemes. Here we can deal only with a few such uses, and will start by comparing one heads with an ordinary ‘plural’ use as in ten apples in order to discern what they have in common and what distinguishes them from the point of view of meaning expressed. There is, quite obviously, a distinction of sense because if we accept, with all grammars, to characterize the latter, prototypical use of -s as expressing a ‘plural,’ ‘more than one’ sense, the former use can only be seen as expressing the sense of ‘singular,’ ‘one.’ On the other hand, the two uses must have some element of meaning in common if the initial assumption of an underlying system is valid, because otherwise we would have to postulate two -s morphemes for number, each with its own quite distinct meaning, and this would be anything but coherent or systematic, especially in a language like English, where the number of grammatical suffixes is reduced to a minimum. The next task then is to discern what the above uses of -s, expressing ‘singular’ and ‘plural’ respectively, have in common meaningwise. In his lesson of 7 June 1945, Guillaume (1991, 206) points out the key relationship between ordinary singulars and plurals: “the affinity of the continuate with the singular and of the discontinuate with the plural are facts of prime importance which a linguist must

Meaning

53

always keep in mind.” What is of interest here is that the ordinary ‘plural’ sense of ten apples is in effect a ‘discontinuate’ because it depicts more than one entity, each represented as occupying its own place in space. With the remarkable singular one heads we call to mind one side of a coin to the exclusion of the other side. That is, the -s has the effect of marking a distinction, an imagined discontinuity, between one side of a coin and what necessarily accompanies it in our experience, the other side. Moreover, naming the other side in the example would have given one tails. This then appears to provide the element of meaning common to both the ordinary ‘plural’ use and the curious ‘singular’ use with -s: both represent a discontinuity. TESTING: EVERY MEANS, A FAIR AMENDS, ETC.

These considerations thus give us an explanatory hypothesis, but in order to provide some plausibility for it we must examine other such uses to see if it can provide an explanation for them as well. With heads and tails in the above use there is a sort of either-or situation where one possibility is necessarily opposed to and excludes the other. In an ‘exclusion,’ ‘opposition,’ ‘either/or’ type of situation, we experience an impression of division or binarity calling for a representation of ‘discontinuity’; this would explain why, in such cases, the ‘singular’ substantive is actualized with the -s morpheme. What gives this explanation some weight is, as we have seen, that with the -ø morpheme here, as in one head, the noun phrase would express a quite different sense. The following example with the indefinite article provides a more familiar use: A continuous belt is a means of power transmission. The same problem then arises here: why does the substantive means require the -s morpheme if it expresses something as ‘singular’? Again the lexical sense expressed here – ‘something to attain an end’ – would appear to provide the solution because it necessarily implies a binary situation: an opposition between the measures taken and the end pursued, and this involves separate places in time, the means expiring when the end is attained. Thus for this sense of mean, which the substantive cannot express with -ø morpheme,

54

Language in the Mind

the implied opposition with the end appears to involve an impression of ‘discontinuity’ calling for the -s morpheme. A similar impression would appear to underlie the -s in the following examples (Jespersen 1954, 167; see Wickens, 160, for other examples): I am friends with him. I was great pals with a man called Hicksey. Although there is no quantifier here, the substantive does designate a single entity in the speaker’s intended message, and so it can hardly be considered a plural, as Jespersen contends. Rather, the notion expressed by friends or pals refers to one person in relation to another. That is to say, evoking the relationship by means of the substantive denoting one friend or pal necessarily implies another friend or pal as the second term of the relationship. Representing one as linked to another brings in an impression of ‘discontinuity’ when evoking the singular entity and so calls for the -s morpheme. Similarly in: You have to take sides. (Conversation) it would not make sense to interpret the substantive as expressing ‘plural’ since the situation manifestly involved choosing either one side or the other. More plausibly, it can be understood to express a ‘singular discontinuate,’ a complex of impressions arising from representing by means of sides one entity in clear opposition with another, necessarily implied entity, the other side. A similar explanation can be put forward for the following examples (Wickens, 162–3): Australia is the antipodes … of England. Master Godfrey, what do you want with me? You’re my elders and betters, you know. Wickens (149–58) provides a convincing argument concerning compensation terms, as in: Meetings like these would make a fair amends. What a thanks I owe.

Meaning

55

Wickens points out that there is “a logically inescapable relationship” involved here whereby the second of the two entities implied in a compensation situation “necessarily presupposes the existence of the first.” In these two examples, the very fact of expressing the need for amendment and gratitude implies an existing lack or deficit and moral debt. Again it appears to be the impression of ‘discontinuity’ arising from representing one entity as a distinct member of a pair that calls for the -s morpheme. It is not always this easy to discern whether or not compensation terms are used to designate a single entity. Thus in uses like … for the wages of sin is death … to get one’s just deserts … to sue for damages a speaker might have in mind several compensations or just one (implying whatever calls for compensation). Whether used to express ‘singular’ or ‘plural,’ however, these expressions bring out an impression of ‘discontinuity.’ MORE EVIDENCE: ONE GAMES, A DOUBLES, A VERY STYLISH SUMMER PANTS

So far, then, the hypothesis that -s can express ‘singular discontinuate’ has proved valid to explain its use with substantives denoting a single entity implying a counterpart. This use could be illustrated more fully, but it will be more helpful to examine another, more frequent type of ‘singular’ usage, one which does not involve an opposition with some implied entity or position. The following is a typical example of this: … the most [records] ever set at any one Games, including Olympics… (Wickens, 192) What distinguishes this example from those just discussed is that the substantive designates a single composite entity, one made up of a series of elements. Thus the notion of one Games involves games or events occurring at a given place in space and moment in time. Granted the composite nature of such an entity, one might expect

56

Language in the Mind

it to be represented as an ordinary ‘plural,’ and indeed we often find these Games to denote the same reality. How then can one explain the representation of the substantive as a ‘singular’ with a or one as determiners? Again there appear to be different impressions involved: these Games brings out the different sporting events constituting the series, whereas a Games brings out more the series as a whole, the institution, the framework, the organization that makes all of the events a single manifestation. Where the latter impression, that of a single series or manifestation, is dominant in the mind of the speaker, the ‘singular’ representation is called on to express a single composite entity. Where the former impression of a number of diverse events dominates, the ‘plural’ representation is called on to express them as distinct entities forming a group. The contrast between these two ways of representing linguistically what is the same entity in extra-mental reality is clearly brought out by the following live TV commentary: We are only 24 hours from the end of these games, and what a games they have been! The commentator seems to have had in mind first the events yet to be held (these games), and then a sort of survey of the whole manifestation (a games). Whether games is represented as ‘plural’ or ‘singular,’ it is the impression of ‘discontinuate’ that calls for the -s morpheme. Another example of the same sort of use brings out even more clearly the effect of representing a composite entity: … an incredible little crossroads of six hardtop routes to nowhere in particular. (Wickens, 195) Had the sentence read six little crossroads there would have been no suggestion of a single entity but rather of six separate roads crossing a main highway at separate points. The ‘singular’ representation of a crossroads, however, depicts a number of roads meeting at a single point to form an intersection. Again one gets the suggestion of what are in effect distinct entities consolidated into one, thanks to the ‘singular’ form imposed on them. In the following remarkable example about a tennis tournament, the ‘plural’ and ‘singular’ uses of -s are contrasted:

Meaning

57

… will span three days – two singles on Saturday, a singles and a doubles on Sunday, and two singles and one doubles on Monday. (Ibid., 185) In three days and two singles the -s morpheme expresses ‘plural,’ ‘more than one’ with its normal ‘discontinuate’ element, so we understand ‘three consecutive days’ and ‘two separate matches.’ In a singles, a doubles, and one doubles, however, we understand either ‘a single match between two individuals’ or ‘a single match between two couples.’ Again, representing a composite entity calls for the -s to express the impression of ‘discontinuity,’ but the morpheme is restricted to its ‘singular’ sense to bring out the fact that the speaker has only one such entity in view. The use of these words without the -s (a single, a double) expresses a very different sense in a very different sport. Wickens (180–5, 279–85) provides a wealth of examples illustrating this use of -s to represent entities composed of two elements, mostly instruments and garments. The following from trade catalogues are typical: A general utility combination pliers. Has two jaw positions and shear-type wire cutter. A very stylish summer pants in the latest peg top effect. There are conflicting impressions involved here: each of these substantives designates a single entity made up of two functionally opposed elements. As a consequence, depending on whether the impression of one instrument or that of two grasping edges is dominant one could say a pliers or a pair of pliers, and where one impression arises to the exclusion of the other we find these pliers and even a plier. -S EXPRESSING ‘ALL’

A certain familiarity with the variety of ‘singular’ usage found with such terms is sufficient to convince one that the sign -s is not a simple reflex of ‘plural’ as rule-bound grammars usually suggest. Rather, the uses we have discussed so far show it to be an instrument for scanning certain impressions arising in the speaker’s experience and representing them grammatically either as ‘singular

58

Language in the Mind

discontinuate’ or as ‘plural discontinuate.’ There are, of course, other uses of -s expressing these two senses, such as the hiccups, the munchies, spirits, drippings, dibs, starters, and nearly five hundred other examples discussed in Wickens. All of these bear testimony to the remarkable sensitivity with which speakers of English make use of -s in everyday speech. Rather than dwell on these uses, however, it will give us a better idea of the representational capacity of -s if we examine a common use expressing a quite different sense: Dogs are vigilant. Cars contribute to pollution. Here the substantive denotes neither a single entity nor a certain number of entities but rather all entities that can be denoted by dogs or cars. That is to say, the -s here expresses a sense that goes far beyond the ‘plural,’ ‘more than one’ sense usually attributed to it since it takes in the whole range of the substantive: it expresses a ‘generic,’ ‘all’ sense. Furthermore there is an impression of ‘discontinuate’ here, as can be seen by comparison with A dog is vigilant. A car contributes to pollution. where the whole range of the substantive is evoked as well, thanks to the use of the article, which expresses a maximum of extensity, not of number, as we shall see in chapter 13. With -s but no article as in the previous two examples, it is as though the ‘generic’ sense were obtained by an accumulation of all possible individuals, to produce a sense that is the contrary of the ‘singular’ sense. Since it is quite familiar, no other examples are required to illustrate this ‘generic’ use. THE MEANING POTENTIAL OF -S

This rapid survey of the essential data shows that the three different senses – ‘one,’ ‘more than one,’ and ‘all’ – of -s vary in representing anything from the smallest number possible to the greatest possible. The morpheme can express all possible quantities of the entities

Meaning

59

represented by the substantive’s concept provided they are seen as somehow ‘discontinuate.’ It thus appears that underlying this polysemy, forming part of the systemic meaning of -s in tongue, is the meaning common to all its uses in discourse, what we can call ‘discontinuate quantity.’ The magnitude of the particular quantity actually represented and expressed has to be observed in any given use because it varies between the minimum and the maximum for the concept. That is, in the system in tongue, -s always offers the same underlying or potential meaning range of discontinuate quantity, whereas the actual sense expressed in discourse can be anywhere within this range depending on what the speaker has in mind to talk about. The result of our analysis of the diverse uses of -s is that its meaning in the underlying system is neither one or other of the quantities expressed, nor a collection of all these quantities, but rather the possibility of moving through the whole range of quantity seen as discontinuate, combined with the possibility of holding up this operation at the point providing a representation of the particular quantity required for the substantive. It is something like the range of possibilities provided by a given computer program: by moving the mouse and clicking at a given point, the particular possibility required is brought into view. Also part of the meaning potential of the morpheme, then, is the possibility of movement through the field of ‘discontinuate quantity’ and that of intercepting the movement. The meaning potential of -s can thus be depicted schematically as an operation carrying the mind’s focus from ‘singular’ or minimum quantity (as in most of the examples discussed above), through ‘plural’ or intermediate quantities (as in the most common use of -s), to ‘generic’ or maximum quantity (as in the dog and car examples). The mental process involved in this expanding type of movement is depicted in figure 4.1, where (m) symbolizes minimum, (I) intermediate, and (M) maximum. The oblique line depicts the gradually increasing scope made available as the operation signified by -s proceeds. Conceiving the potential meaning of -s in this way, as the expanding m M movement through the field of discontinuate quantity interceptible at any critical point, beginning, middle or end, we can understand how a single morpheme can represent various impressions and so express different senses. In this way we can explain the polysemy of -s.

Language in the Mind

60

m

I

M

Figure 4.1

Conceiving of a morpheme as an inherently dynamic mental entity in this way is a distinguishing characteristic of Guillaumian analysis. Both the movement through the field and intercepting it are mental operations made possible by underlying neural mechanisms but not to be identified with them. That is, such “operations of thought which form the basis and the framework for the structure of tongue” (Guillaume 1984, 49) are mental events dependent on but not reducible to neurological activity. It will help to make this point clear if we digress for a moment to compare Guillaume’s position with that of cognitive grammar. Langacker points out that “Meaning is equated with conceptualization” (1991b, 2). Elsewhere he specifies that “conceptualization can be analyzed at either of two levels; the phenomenological level (i.e. that of mental experience), and the level of cognitive events (i.e. neurological activity) … the structures at this [phenomenological] level must eventually be explicated with reference to neurological events” (1991a, 149). For Guillaume, on the other hand, analysis of meaning on the mental level, the task of linguists, includes observing the mental representation expressed by a word (its meaning) and discerning mental events leading up to it. Analysis on the cognitive (= neurological) level, however, would be outside the field of linguistics and would fall to the neurosciences. To avoid any confusion on this point, I will speak of mental rather than cognitive activity, so in what follows it will be a matter of mental processes or operations giving rise to the meaning a morpheme or word expresses at the phenomenological level. -Ø MORPHEME AND ITS DIFFERENT SENSES

As the counterpart of -s, -ø morpheme makes up the rest of the system of number in the substantive. A typical use is:

Meaning

61

They need a new car. Here too it is important not to limit observation to common examples expressing ‘singular,’ ‘one’, but to observe -ø substantives expressing other senses, to consider all the data. One such use is illustrated by the following example: They need water. This common use of what is often called the ‘mass’ or ‘noncount’ sense of a substantive also expresses quantity – a vague, undetermined amount – but what is most striking is the ‘continuate’ view of this quantity. This ‘continuate’ view is observable with the same substantive in the next example, though the quantity expressed is by no means the same: Water is made of hydrogen and oxygen. The ‘generic’ sense here indicates that the quantity represented is maximum, that is, great enough to include everything that can be designated by the substantive. These three examples show that -ø too can express any quantity from the greatest to the smallest, provided it is ‘continuate.’ Like -s, -ø can also express the quantity of substantives where there is a conflict of impressions, provided the impression of ‘continuate’ is dominant. For example, in There were two bear in there fishin’. both the verb and the numeral make it clear that the speaker has in mind more than one entity, giving rise to an impression of ‘discontinuate.’ One would therefore have expected bears, and indeed this would have been quite appropriate, but it would not have expressed the speaker’s experience with quite the same nuance. The use of -ø morpheme to represent a plurality of wild animals in this way gives the impression of a speaker with a certain expertise.4 The fact that this usage is characteristic of hunters, game wardens, naturalists, and other such cognoscenti suggests that uppermost in such speakers’ minds is not the notion of the individual animals but that of the species, with all its habits and

Language in the Mind

62

typical behaviour. This helps us understand why, for cognoscenti, an impression of something ‘continuate’ is often dominant in such situations. The fact that the particular nuance of meaning expressed by two bear is quite subtle indicates what a sensitive instrument of representation the morpheme is. Something similar arises with certain substantives naming members of a group – three resident faculty, four staff, five crew – where the notion of the group provides the ‘continuate’ impression. But this sort of use is not limited to speakers with special knowledge. In uses like two aspirin, the speaker’s main impression appears to be a type of medication, as opposed to two aspirins, where the idea of distinct tablets comes to the fore. Similarly for two dozen, two beer, these kind of cigarettes, and numerous other uses where the impression of something ‘continuate’ dominates that of ‘multiple’ in the speaker’s intended message and calls for the -ø morpheme. Thus it can be seen that -ø, like -s, is polysemous since it can express different quantities, and this observed polysemy leads to the hypothesis that, in the system of number, -ø also signifies the possibility of an operation. However it differs from -s in signifying movement not through the field of the discontinuate but through the opposite field, that of continuate quantity, so that regardless of its quantitative sense – ‘generic,’ ‘undetermined,’ or ‘singular’ – the substantive represents its referent as a spatial continuum. Furthermore, as we shall see in the next chapter, the operation signified by -ø morpheme is, by virtue of the type of movement involved, the contrary of that signified by -s morpheme: it is a movement of contraction starting with the total field of the substantive’s notion and ending with its minimal extent, M m, as in figure 4.2.

-ø M

I Figure 4.2

m

Meaning

63

REPRESENTING: MAKING THE UNSAYABLE SAYABLE

The point of these examples, which are typical of many more (cf. Hirtle 1982), is to illustrate the representational function of the two morphemes, the manner in which each one channels the most diverse elements of experience into a general form depending on the way the speaker perceives their spatial disposition. This applies to all representing of experience, whether by morpheme or lexeme: the resulting representation is always a generalization that can never depict the original in the intended message with complete faithfulness. A representation is not a facsimile. Only at this price can the unsayable particularity of any individual experience be made sayable by words, the instruments of representation available in language. But this has the advantage of permitting us to take our distance from the experiential world, to form our mental universe in such a way as to organize and anticipate our experience. On this depends the development of the mind’s potential. Our brief discussion of examples also serves to illustrate the method for analyzing particular morphemes: observing the meaning expressed by real examples in light of the potential/actual postulate, imagining a hypothesis permitting us to understand the data and testing the hypothesis by further observing usage as carefully and as completely as possible. This is, of course, the wellknown scientific method, with all its exigencies and dangers, which, because it is applied to the mental side of language, cannot resort to experiment but depends on the consensus of competent observers to establish the data, the meaning expressed. Paraphrasing an example, especially where the nuance is rather subtle, can never provide a completely adequate description of the sense expressed but does serve to indicate whether different observers are understanding the example in the same way. We shall discuss this method of analysis in more detail in chapter 6. This very inadequacy of paraphrasing highlights the important role of morphemes in grasping impressions, sometimes quite fleeting, arising in the intended message, impressions that cannot be as well represented by other means. The ability of the mind to intercept its stream of consciousness in this way is determined by the lexemes and morphemes instituted in tongue. They serve as

64

Language in the Mind

“viewing ideas” in Guillaume’s terms (cf. 1984, 124) to scan what is going through one’s mind and reflect it, “refract it” as he sometimes said, into the various representations making up the meaning of words. To make his view of the dynamic role of lexemes and morphemes more explicit, it can be contrasted with Jakobson’s view that “the speaker selects words” from his “lexical storehouse,” a sort of “filing cabinet of prefabricated representations” (117). Terms like “storehouse” or “inventory” (cf. Langacker 1987b, 47, 63) used to characterize lexicon and grammar imply elements with no inherent operativity. The essentially operational nature of lexemes and morphemes even when we are not speaking led Guillaume to speak of tongue as “an expanding idea-universe” (1984, 157), unconsciously categorizing whatever arises in our ongoing experience and ready to represent it linguistically if we decide to talk about it. In a later chapter we will come back to this doctrine of preconscious viewing ideas scanning our stream of consciousness. We touch on it here to recall the translating of our unsayable experience into representations attached to signs that will make it sayable – what Guillaume (1984, 136) calls the first mutation in any act of language: “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.” This view of a morpheme as a permanent means of focusing on and representing something in the intended message, imposes on the linguist the task of describing its meaning potential. However, in order to explain the polysemy observed in different uses it must be described dynamically, as was done above for -ø and -s. In fact, “a sign in tongue is the symbol of a movement, and of the whole of this movement” (Guillaume 1987, 53) because any potential involves an operation, the process of actualizing its potentiality. Without this versatility, language would be unable to come to grips with our ever-changing experience and with the mind’s inventiveness because, as Guillaume (1991, 46) remarks, “The mind moves freely within itself and is not a prisoner of words.” As a consequence, every attempt to articulate what we have in mind “is a wholly new start … each venture is a new beginning, a raid on the inarticulate,” as T.S. Eliot puts it. And this raid, this act of language,

Meaning

65

can be carried out only because our mother tongue provides the potential meanings and psychomechanisms for representing and expressing what otherwise would remain inarticulate. There remains the important task of examining the link between the two morphemes to explain why they alternate in the substantive. In the next chapter, we shall see how Guillaume’s approach permits him to propose the existence of a system, an entity that is not observable in discourse.

5 A System for Representing As an altogether general thesis, a system always has a before and an after. Guillaume (1971, 18)

MEANING AND POLYSEMY

Unless we happen to be particularly gifted artists, we cannot adequately express our thoughts and feelings about something, say a sunset or a sensation of hunger, let alone an idea in an essay, if we do not first represent them by means of the units of meaning made available to us by the words we have learned. As we have seen, all the formative elements of words are instituted in our tongue for just this purpose: to permit us to represent and express whatever we have in mind to speak (or write) about, though different kinds of words play this role in different ways. The representation provided by words is thus another mental entity, quite distinct from our unique (and so strictly private) experience of the moment. Furthermore, a linguistic representation is not always a satisfactory reflection of a particular experience but it does have the signal advantage of being sayable since it consists of meaning units linked to physically expressible signs, units which can be integrated into sentences. This permits others who speak the language to call to mind a like meaning when they hear the signs, knowing that this meaning is a representation of something outside language, a more or less faithful reflection of the speaker’s private experience. As a consequence they convert the meaning they get from some utterance back into a message, calling on any knowledge they have of the speaker, the speaking situation, etc., in order to obtain a message resembling as closely as possible what the speaker had in mind at the outset. For example, a simple sentence like It is ready

A System for Representing

67

will give rise to quite different messages if spoken by someone preparing lunch, by a mechanic in a garage, etc. In this way we can see how language links us with the reality of the speaker’s experience and thence, through that experience, with the reality of the universe around us. We really do talk about the world around us insofar as our experience of it, and the resources of our language, permit. Although much of what has just been brought out may seem obvious, it has been recalled here to ensure that the act of language as a whole is kept in mind. Indeed, in linguistics discussions, two crucial points are easily forgotten: 1) that the speaker’s extralinguistic experience – a complex of diverse impressions constituting what the speaker has in mind to talk about, here called the intended message – is distinct from the linguistic meaning representing it in a sentence; and 2) that the meaning the listener gets from the utterance is not the same as the message. For Guillaume, neglecting the two primordial operations of representing extralinguistic experience and referring sentence meaning back to the extralinguistic would result in an incomplete and so unrealistic view of language masking the relation between language and thought. To make this important point it is worth contrasting it with the view of other linguists. Jakobson (320), for example, considers “contextual meaning” an “aspect” of the referent “given by the whole context, by the universe of discourse,” whereas “the general meaning belongs to semantics.” For Langacker (1991a, 2) “meaning is equated with conceptualization … interpreted quite broadly: it encompasses … sensory, kinesthetic, and emotive experience…” The term “meaning” as used by Guillaume appears to be closer to what Langacker calls “semantic structure”: “The term conceptual structure will be applied indiscriminately to any such entity [thoughts, concepts, perceptions, images, and mental experience in general], whether linguistic or nonlinguistic. A semantic structure is then defined as a conceptual structure that functions as the semantic pole of a linguistic expression” (1987b, 98). To avoid assimilating the meaning expressed by some morpheme or word with the speaker’s experience, implying thereby that it gets its meaning from the referent, from what is outside language, Guillaume insisted on representation – the mutation of the unsayable into the sayable – as the beginning of the act of language and the translation of the sentence meaning into the message as its end.

68

Language in the Mind

Chapter 4 illustrated how the speaker represents impressions arising in the intended message by means of two number morphemes in the morphogeny of the substantive. We saw that each morpheme can express different senses, different quantities, thus posing the problem of polysemy. Reflecting on the relations between their different senses in the light of Guillaume’s view of language, we were led to depict their potential meanings as dynamic since this is the only way proposed so far to explain how an unvarying meaning-sign unit in tongue (a necessary condition for communication as we know it) can give rise to the varying senses expressed by a sign in discourse. Since this problem is quite general, arising with words as well as morphemes, it is worth mentioning other approaches to it in order to bring out the originality of Guillaume’s approach. One scholar poses the problem as follows: “a great deal that is problematic in linguistic semantics turns on the question of the autonomy of the word versus its context-dependence: is it possible to assign a portion of meaning to individual words or is ‘word meaning’, on the contrary, an illusion, a function really of the word’s occurrence in a particular context?” (Waldron 1967, 58). According to Cruse (80–1), linguists are divided on this issue: “Linguists who have worked in lexical semantics can be broadly divided into two categories: on the one hand, there are those who believe that a word form is associated with a number (perhaps finite, perhaps not) of discrete senses; and on the other, there are those who believe that the discreteness of lexical senses is illusory.” Perkins (26) is more explicit: Analyses of word-meaning can usually be divided into those which assign a meaning to a word in isolation from a specific context of use, and those which regard the meaning of a word as being largely, if not entirely, dependent upon a specific context of use. The ultimate expression of the latter ‘polysemantic’ approach is probably that of the later Wittgenstein, who argued that “every difference in a word’s use is a consequence of and evidence for a difference in its meaning” (Wertheimer 1972, 49). One recent example of the former ‘monosemantic’ approach is that of Bolinger (1977, x), whose stated purpose is to “reaffirm the old principle that the natural condition of a language is to preserve one form for one meaning, and one meaning for one form.” Both these approaches have their advocates, and it should be stated at the outset that neither is necessarily

A System for Representing

69

right or wrong; each can be judged only according to whether the phenomena it is used to interpret are thereby illuminated.

Since both approaches are based on observation, each has its merits, but neither can handle all the facts as long as the linguist is concerned with discourse only. Those who, like Jakobson, adopt a monosemantic approach consider language to be a code and consider the speaker and hearer as “coder and decoder” (Waugh, 15). A good illustration of this is given by Reid, who, to guarantee the unity of the morpheme, argues that -ø in the substantive has the “meaning ONE” (80). But he then observes that in ‘mass noun’ usage -ø is “employed faute de mieux to communicate messages for which it is not ideally suited,” an analysis suggesting that the relation between these two important uses of the morpheme is unmotivated, arbitrary. Our concern with the potential meaning in chapter 4, on the other hand, led us to seek the relation between the ‘singular’ and ‘mass noun’ senses – both represent space as ‘continuate’ – as a basis for considering them manifestations of the same morpheme in tongue. On the other hand, adopting a polysemantic approach like that contrasting prototypical and peripheral senses, with homonyms as the “the endpoint along the cline of relatedness” (Langacker 1991a, 268), provides no semantic unity for the word. Indeed, as Taylor (1989, 142) puts it: “The major advantage of this approach is that it frees the linguist from the obligation to search for a semantic component unifying all the different uses of a word.” In that case, if there is no such component, then the only basis for the unity of a word, or a morpheme, would appear to be the physical sign, but then meaning would no longer be central in language. Proposing a morpheme that is monosemantic in tongue and polysemantic in discourse solves the problem. If the potential meaning of -s is described in terms of a movement, more precisely the possibility of a movement, through the field of discontinuate quantity, that of -ø in terms of a possible movement through the field of continuate quantity, we can explain usage as observed so far, including some fairly subtle distinctions of meaning, such as the nuance between two bear vs two bears. If confirmed by further examination of data, this constitutes a major accomplishment because for the first time the remarkable variety of usage of these two morphemes can be traced back to two principles, two meaning

70

Language in the Mind

potentials. And the key to this was conceiving their meaning, not as something static like an item in an inventory, but as something dynamic, as a potential permitting a particular mental operation. FROM MORPHEME TO SYSTEM

Our analysis cannot stop here, however, because there are still unanswered questions. Why do both morphemes express the same range of quantities? What is the reason for the continuate/discontinuate opposition? In short, what is the relation between these two mutually exclusive morphemes? Raising the question of the relation between them carries our reflection to a second level of abstraction because the terms of the relationship – the meaning potential of the two morphemes in tongue – are themselves abstract entities at one remove from meaning directly observable in discourse. In other words, this raises the question of their system because a system is an entity of relationship, an immaterial, unsayable unit of tongue consisting of the coherent links between potential morphemes but with no sign of its own. This constituted an obstacle for those linguists in Guillaume’s day who could not admit the reality of anything in language unless it had a direct manifestation, a sign, observable in discourse.1 For them, therefore, a mental system was not really a fact of language. For Guillaume (1984, 103), however, such entities are real: Systems are entities of tongue which are just as real as the forms that represent them … 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.

Anyone who recognizes that a mental construct does have a reality of its own will, like Guillaume, seek any indirect manifestations of a linguistic system they can find in order to describe the system as the hidden part of language. This study of systems, what

A System for Representing

71

Guillaume calls the psychosystematics of language, has its technique of analysis permitting one to discern the outlines of a system. This technique, to be described in the next chapter, will enable us to explore the system of number in the substantive. Every use of a substantive in English involves one of the number morphemes expressing either a ‘continuate’ or a ‘discontinuate’ representation. This opposition is even more general than the -ø ‘singular’ vs -s ‘plural’ opposition, usually considered to be the basis of the system of number. We have seen, for example, that the opposition between two bear and two bears is not one of ‘singular’ vs ‘plural’ since both phrases express the same quantity. Similarly for one head and one heads. Furthermore, in Coffee is a stimulant, one can hardly call coffee a ‘singular’ substantive, whereas in A coffee, please it is considered a ‘singular’ because it expresses a limited, pluralizable quantity. In both examples, however coffee expresses an undivided quantity, i.e. a continuate view. Thus ‘continuate’ vs ‘discontinuate’, rather than ‘singular’ vs ‘plural’, appears to be the basic opposition in the system of number. Guillaume often pointed out (cf. 1991, 203–12) that this opposition involves a dependency relationship because representing something as discontinuate in space calls for depicting a series of entities and each entity is itself seen as continuate in space. That is, an experiential entity must be representable as a discrete continuate before it can be seen as a numerable set of discontinuates. This fact brings to light a crucial relationship: a ‘continuate’ representation is a necessary prior condition for producing a ‘discontinuate’ representation, but not vice versa. This is important because the system, being instituted to permit a process of representation to be carried out, is operational by nature. We have therefore to conceive of its fundamental relationship as one between two operations, and this allows us to determine the dynamics of the system, the order in which the two morphemes arise in the system. That is, because a representation of continuate space is a prerequisite for representing discontinuate space, we conclude that -ø morpheme precedes -s morpheme in the process of representation permitted by the system. We can depict their systemic positions by means of the following diagram: -ø

-s

Language in the Mind

72

Not only is the basic relationship of the system a dynamic one, but the terms of this relationship, the meanings of the two morphemes, are themselves operations, as we have seen. The fact that an ordinary ‘more than one’ use of -s involves a multiplication2 of a ‘one’ view of the notion suggests that the movement inherent in the meaning of the -s morpheme starts with a minimum space (m), a ‘singular,’ and proceeds as far as is required in the direction of the maximum quantity of space (M) producing at some intermediate point (I) a ‘plural,’ and finally, proceeding as far as possible, produces a ‘generic.’ This gives the expanding form of movement, m M, for -s, which expresses various discontinuate views of quantity as described in the previous chapter. The fact that this operation depends on the prior -ø operation having provided a view of continuate space at its minimum led us to propose for -ø morpheme the opposite form of movement, M m. We saw that at its very beginning (M) this operation permits a -ø substantive to express a ‘mass, generic’ sense, and at some intermediate point (I) an ‘indefinite quantity, mass’ sense and finally (m) an ordinary ‘singular’ sense. That is to say, the second, -s operation, which is the mirror image of the first, -ø operation, starts at the point where the first ends. Together they constitute the system of number in English.

-ø M

I

-s m

I

M

Figure 5.1

Figure 5.1 depicts better than words the set of relationships involved. This diagram brings out the unity of the system: the possibility of a process representing the quantity involved in any actualization of a substantive’s lexical notion. To be able to represent any quantity that a given lexeme may be called on to express, the system establishes the relationship between the greatest possible quantity and the smallest possible quantity since this includes all intermediate quantities. Furthermore, like any other relationship this one has two versions,3 depending on which term is considered

A System for Representing

73

first, and so the system is a binary unit in tongue. To ensure that both versions of a relationship can be represented, a grammatical morpheme always forms part of a system, which, having at least two components, has to be dynamic. The diagram also brings out this dynamic nature of the system. To represent a given quantity, it must be felt to be opposed to all other possible quantities. To think some notion as, say, ‘discontinuate plural,’ it must be implicitly opposed to ‘discontinuate one’ and to ‘discontinuate all’ as well as to any ‘continuate’ representation. “We think by contrast, unable to do otherwise,” as Guillaume often said. Similarly, the resulting representation, say, ‘continuate singular,’ is implicitly contrasted with all the other positions (possible points of interception) in the continuate phase as well as with any in the discontinuate phase. By holding up the underlying movement at some point in either phase, the speaker implicitly opposes it to the other phase. Activating a system and intercepting one of its operations in this way constitutes the actualizing of the meaning potential of the morpheme, the result being the actual representation of a certain quantity as opposed to all other possible quantities. This is why Guillaume postulates that the potential meaning of all morphemes and grammatical words involves a movement (cf. 1991, 68). The system of number is thus the set of mental conditions permitting a two-phase movement during which the speaker can call to mind the form of movement (contracting or expanding) and the position within it best representing what is being focused on in the intended message. This is a real mental process and so it requires time, real time, though it is so short that there is as yet no means of measuring it. The system resides in the preconscious mind of Anglophones and has no sign of its own to make us aware of its existence because we never need to represent and express all its meaning possibilities at one time, in any case an impossibility. Thus it is only through the indirect, circumstantial evidence provided by the use of its two morphemes that we can infer the existence and discern the makeup of the system. On the other hand, as in any other science based on observation, the greater the variety of data we explain by means of the proposed system the greater the likelihood of it corresponding to the hidden reality of language we are trying to describe. We will therefore examine more uses in the light of the system.4

74

Language in the Mind

Before going on to this, however, a comment concerning the expression “a mental operation” may be in order to avoid any misunderstanding. The notion of language as dynamic, widespread among linguists, implies movement or activity or process. The advantage of the term “operation” is not only the related terms it makes available (“operate,” “operative,” “operational,” and the more abstract “operativity”), but the fact that in ordinary usage it suggests a movement with a particular objective in view and so implies an activity or process with a beginning, a middle, and an end. This is why we can speak of a mechanism making an operation possible, indefinitely repeatable. As Guillaume (1984, 51–2) puts it: “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.” Guillaume thus saw the linguist’s task to be one of describing mental mechanisms insofar as they are instituted in language. Moreover he took for granted that the mind invents and institutes them in the different systems of tongue. He thus focuses on mental operators making possible operations of the mind for grasping (= representing) what is going on in the mind itself, and yet he does not consider that this reveals the nature of thought: “The study of the formal, psychosystematic 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” (ibid.). INDIRECT EVIDENCE FOR THE SYSTEM

One of the most intriguing characteristics of English substantives is the fact that their lexemes represent an experiential entity in either a unitary or a continuum mode. Conceived in a continuum mode with no internal divisions, and no external limits, the lexeme gives rise to what is called a “mass” or better (since more abstract) “continuum” noun.5 Conceived in a unitary mode with its own

A System for Representing

75

spatial limits to give a ‘singular’ or a ‘plural,’ the lexeme results in a “count” or “unit” noun. We have seen how the -ø morpheme can provide a form for both ‘continuum’ notions – by representing them as undetermined in quantity, that is, with no limits, internal or external, evoked6 – and for ‘unit’ notions – by representing them as determined in quantity, ‘singular.’ It is intriguing to compare uses of the same notion conceived in these two ways. For example, in There is wine on the table the undetermined quantity of the ‘continuum’ noun wine contributing to the meaning of the sentence could, depending on the speaker’s intended message, refer to a bottle of wine (as when setting the table), a few drops (as when cleaning up), or a much larger quantity (in a shop).7 On the other hand, to hear in a restaurant I have a wine you will like we would think ‘a vintage or type of wine,’ because here, as the indefinite article makes quite clear, the notion is understood in a ‘unit’ sense with the substantive wine representing a determined quantity, ‘one.’ It is this ‘minimum continuate’ sense that is multiplied to give a ‘discontinuate,’ ‘plural’ sense when the notion of wine is formed with -s morpheme, as in We have two new wines today. The lexeme ‘coffee’ can also be actualized as either a ‘continuum’ or a ‘unit’ notion and so used in similar ways, but as a ‘unit’ notion it can also have the sense ‘a cup of coffee’ as in I’ll get a coffee / two coffees. There are of course many words whose notions lend themselves readily to either a continuum or unit mode of representation, as is often pointed out (cf. Hewson 1972). Thus to be in a hospital is not the same as to be in hospital. In the former case the substantive calls to mind the physical limits of the institution, but in the latter case no such ‘unit’ notion comes to mind but rather the idea of the treatments or services constituting the function of the hospital. Similarly for in school, in prison, and a number of other such notions actualized not to represent a particular institution as a unit but rather the function provided by any such institution. Or again, substantives like bus, car, bicycle, etc. usually call to mind a vehicle, but in to travel by bus, etc. it is the means of transport not a particular unit that comes to mind. Similarly, we can compare looking for work, where the sense is ‘a means of earning a living,’ with a work of art, where it is a unit, the result of the process, that is expressed. All such contrasts show that the lexical matter can be actualized in different ways and that the system of number can provide the appropriate grammatical form for the lexeme.

76

Language in the Mind

Furthermore, in ‘plural’ uses of these lexemes – hospitals, cars, works of art – it is the ‘unit’ sense that is expressed, as the position in the system would lead one to expect. In a few cases, it is very difficult to describe the expressive effect of ‘plural’ usage precisely because no corresponding ‘singular’ usage has been found. For example, in European waters or the waters of the lake, one gets an impression of different areas or depths, but we do not find the ‘singular’ a water in a corresponding sense. Similarly for the sands of the desert and The Snows of Kilimanjaro. This brings out another trait of the system of representation: it is there to depict the speaker’s experience and as long as the particular complex of impressions calling for a given representation has not arisen, that particular position in the system of number remains a mere possibility for a given notion. A word like cattle, which is found neither in the ‘plural’ *cattles nor in the ‘singular’ *a cattle makes this quite clear. The fact that it is commonly used to represent a fairly large number of animals suggests that it is the impression not of a species (as in the case of two bear) but of a herd that dominates here because two or three animals do not constitute a herd. Even in the case of two cattle, for example, the implication is that the animals belong to a larger group. Other positions in the system remain possibilities for cattle if ever the set of impressions it expresses is felt to be appropriate.8 Similarly, the substantive riches is found commonly in ‘plural’ or ‘generic’ uses, perhaps because the sense of ‘possessions, means’ involves an impression of multiplication. But again, the very fact that certain lexemes cannot be used in one way or another shows that each position in the system is a potential means of intercepting and forming a given lexical matter. Because some lexemes are seldom if ever thought in one mode, striking innovations can arise in usage. We usually think the lexeme of behaviour in a continuum way as one’s ‘manner of conduct or comportment’ but when in psychology we read of different behaviours the -s morpheme indicates that we are to think it in a unit way as ‘different types of behaviour.’ Likewise, we normally find continuum usage like not enough evidence and all the information, but occasionally we find unit uses like every evidence we now have (Webster’s, s.v.) and in technical usage, an information. In a lot of house for your money, ‘house,’ normally thought as a unit, must be actualized as a ‘continuum’ notion such as ‘whatever goes to make up the quality of a house.’ A good example of innovation arose when

A System for Representing

77

a rather tall person using a public telephone remarked: There’s not enough phone booth here for me. On the spur of the moment the speaker actualized the notion to express not the unit with its physical limits but the space contained in it, using other words (not enough) to characterize its limits. The fact that ordinary speakers can make words fit their own momentary message in this way shows clearly that words exist in the mind, not as fixed units ready to be inserted into a sentence, but as representational potentials – as “viewing ideas” and grammatical systems with all the constructional processes these involve – ready to be reconstructed in the way required by the intended message. This has been depicted by a striking metaphor: “Thought and expression are but the two sides of the same prism” (Sayce, 35). Tongue constitutes the prism refracting unsayable experience, or thought, into sayable representations of language. A MECHANISM FOR REPRESENTING

The last example discussed above is interesting both for its novelty and for what it suggests of the system of number: it brings out in a graphic way the distinction between an intermediate and a final interception of the -ø operation. Representing the space involved in a lexeme to the exclusion of its limits is what produces any ‘continuum’ noun with its undetermined quantity, and this is the way the notion ‘(phone) booth’ is formed by -ø morpheme: the morpheme’s movement is intercepted before it reaches its minimal, ‘singular’ representation. This results in a curious impression of vague, unconfined, extendible space. On the other hand, representing the space involved as contained is what results in a ‘unit’ notion with its ‘minimum quantity,’ ‘one’ sense. That is, when the -ø movement is intercepted only at its final instant the notion is formed with the limit called for by a ‘unit’ view, as in a phone booth. Figure 5.2 is intended to suggest that the -ø morpheme imposes a spatial limit on the notion, conceptualized as a self-contained unit, to give rise to a ‘singular’ substantive at the end of its contractive movement. The expansive -s movement, starting from the same ‘singular’ position, can be diagrammed as in figure 5.3. Intercepted at an intermediate point, the -s movement usually gives rise to a ‘plural’ sense and at its final instant to a ‘generic’ sense. Intercepted at its first instant, before it has moved away from

78

Language in the Mind

Figure 5.2

Figure 5.3

the ‘singular’ position it gives rise to uses like one doubles and this crossroads, expressing a ‘unit’ view of what is constituted by a number of components. Even uses like one heads and a means depict units drawn from larger entities. The suggestion of something ‘discontinuate’ in each of these uses indicates that the focus is not on the space contained in the unit but rather on the space containing it. Considering the transition from the -ø phase of the operation to its -s phase in this way – as a maximal approach to the limit defining the space within vs a minimal withdrawal from that limit, contrasting with the space outside – is a result of trying to imagine what is going on in the hidden workings of language, an example of the type of reflecting that goes on in any theoretical work as a result of problems posed by the data. In this case the proposal is at two removes from the observed facts of discourse and so is subject only to indirect or circumstantial evidence. One such bit of evidence is the very notion of limit: a limit is an abstract entity, an imposition of the mind, because in physical reality a limit like the border between two countries occupies no space.9 As a consequence, there is no position possible “in” the limit; at best one can occupy a position immediately adjacent to it, on one side or the other. But these two positions, these two points in space, are separated only by a limit, that is, by no space. Further by way of circumstantial evidence, if representing ‘plural’ involves representing ‘more than one’ as the grammar books point out, then it is necessarily an iteration of a minimal unit quantity, made representable by the -ø operation. This involves the possibility of mentally lining units up one after another, of viewing them

A System for Representing

79

as distinct parcels of space, not of focusing on the amount of space contained in the units. That is to say, this embracing of several parcels of space in one view can be obtained only by situating them in a single stretch of space. Passing like this from an inner view of space to an outer view would, in fact, appear to be the only way to get from a representation of space as ‘continuate’ to a representation of space as ‘discontinuate.’ This proposal thus helps confirm the coherence of the system. Finally, this manner of understanding the passage from the first to the second phase of the operation of representing number brings out the continuity of the operation involved. The -ø phase of the operation expires at the point where it reaches the limit of the space contained and the -s phase begins at the point where that limit is transcended. Since the same ‘minimal space’ position provides the end point of the first movement and the starting point of the second, there is no break between the two and the unity of the system of representation is ensured. This theory of number in English is based on data from English, real examples that often require considerable effort to observe. The manner of viewing the data, however, was drawn from Guillaume’s principles, which postulate, as we have seen, that any use of -ø or -s is a result of a representational operation made possible by the system of number in tongue. In each of its uses the morpheme represents a position occupied in the movement constituting the morpheme’s potential meaning. Although French does not exploit its system the way English does, Guillaume had already proposed a general theory of number in the substantive where the distinction between continuate and discontinuate is the basis of the two-phase operation permitting the mind to represent space, as we have seen: “The affinity between continuate and singular, and between discontinuate and plural are primordial facts that a linguist must always keep in mind” (1991, 206). He calls the two movements “kinetic singular” and “kinetic plural,” designating them by the point they move toward giving rise to the most common result in discourse “static singular” and “static plural.” He also points out (203) that “the interceptions producing” the various resulting representations, expressed in English by -ø and -s, “arise throughout the longitudinal progression of the two movements.” On the basis of the figure depicting the system, he even points out in the same lesson of 7 June 1945 the possibility existing in

80

Language in the Mind

tongue of “not one, but two static singulars belonging to different movements” (205). This was illustrated above in our discussion of -s ‘singular’ vs -ø ‘singular’ (a crossroads vs a crossroad, and other examples). The fact that this was observed in the English data some thirty years later is a remarkable comment on Guillaume’s power of insight because he had no language in view where these possibilities were exploited in discourse and his statement could be verified.10 He realized that the same position must be the end point of the first movement and the starting point of the second movement if the system is a single unit in tongue, based on a single operation. This last point, to the effect that the continuity of the basic operation from one phase to the next depends on one instant common to both phases, has a far more general application than just the system of number. It appears to be a condition for any multiphase operation and so should be discernible in other grammatical systems, since, as we have postulated, all systems are operational entities. Of course other systems would not operate in the field of number with ‘continuate’ and ‘discontinuate’ phases bringing into relationship maximum and minimum quantities. But regardless of the particular field, the type of operation just outlined, the mechanism, has been discerned in other systems, that of the article and that of some and any, for example, and even at the most general level, the system of the word. We will see this in later chapters. A HYPOTHESIS THAT EXPLAINS

Those observations that preceded and led to discerning the system – the different quantitative senses expressed by each morpheme – are of course explained by it. The next step is to use the system as a hypothesis to explain other facts of usage. The validity of an explanation can be judged only by the extent to which it permits us to understand why the speaker used a particular form, the point being that if a system of morphemes is really a coherent set of potential meanings, then all uses must be actualizations of that potentiality. This aspect of proof is demanding because it often requires us to live with a problematic use for a long time before discerning its relation with the underlying system, as we shall now see. The explanatory capacity of the hypothesis for -ø can be illustrated by more details of usage expressing the internal plural of animal

A System for Representing

81

names. It was pointed out above that uses like the following11 are common for cognoscenti, speakers knowledgeable with regard to wildlife: The quail were calling. Bear vary greatly in size. They shot several elk. … over 200 hundred gazelle… The fact that such speakers are cognizant of the species led to the hypothesis that their dominant impression is that of a continuum, the species, which is instantiated by a number of individual animals or specimens. This was borne out by comparison with the use of the -s substantive in similar contexts: the slight difference in expressive effect corresponds to the view of the ordinary speaker, whose dominant impression is that of individual animals. In one study (Allan, 102) it has even been pointed out that a comment like the following would sound pretentious on the part of the average tourist: We saw three elephant in the game park. It would suggest a speaker assuming a certain expertise in order to impress others. It is also pointed out that the -ø ‘plural,’ unless “facetious or precious” would not be acceptable in: We saw three elephant at the zoo. Besides suggesting a casual visitor as a speaker, the sentence depicts animals in an environment inhibiting their normal, speciesanimated behaviour. On the other hand, the example We have reared three white rhino at the London zoo. coming from, say, a zoologist, would be quite acceptable, perhaps because it depicts growing to maturity, a species-governed activity. Such facts suggest that with -s ‘plural’ an ‘individual animals’ impression dominates that of ‘species-relationship.’ This is borne out by another, quite uncommon use: even the substantive deer, usually with -ø ending for ‘plural,’ has been found with -s to designate stuffed animals in a museum. It remains however to clarify in

82

Language in the Mind

the light of our hypothesis why undesirable animals not generally hunted, like skunks, bats, moles, rats, and the like are not, according to authors consulted, designated by the -ø substantive. Similarly for insects. Perhaps detailed observation of the usage of cognoscenti will provide a more complex picture and suggest an explanation. This hypothesis also helped us to explore the problem of why the internal plural is not found with domestic animals. In the same study Allan depicts the situation with the following uses: The farmer shot some duck. The farmer fed some ducks. The first sentence, suggesting a hunting scenario, depicts the birds as wild, whereas the second, evoking a barnyard scene, depicts them as domestic and requires the external plural with -s. This requirement can be understood if we remember that an animal is domesticated to the extent that its species-animated instincts have been curbed. That is, in our ordinary experience of domesticated animals, the impression of the species never dominates. The fact that cattle and swine express ‘plural’ appears at first sight to counter this explanation. However details of usage, such as their incapability of expressing a ‘singular’12 or taking the -s inflection, suggest that they present a different case. What appears to give them a ‘continuate’ basis is not a ‘species’ notion but a ‘belongingto-a-herd’ notion in their lexeme, as in the case of head (fifty head of …). (The same may be true of sheep, or it may be an “irregular” sign inherited from Old English.) In the case of humans, we find examples like three resident faculty, 1000 staff, the skipper and five crew where the substantive names a collectivity, the ‘internal plural’ sense providing a representation of its members. This is to be contrasted with a faculty/faculties, a staff/staffs, a crew/crews, where the ‘singular’ and ‘external plural’ senses represent the group or groups, not the members. People is perhaps the best known substantive of this type: fifty people, a people, the peoples of Europe. In the title of a book, How a People Die both the ‘collectivity’ and the ‘member’ senses are expressed, but by different means.13 Uses like many of these western Carrier and about 2000 Eskimo are common, whereas substantives like Canadian or European are not found with -ø ‘plural.’ It would seem that a collectivity based on a geographical or political entity does not provide the conditions for a ‘continuate’ view, but a native ethnic group does. Here too

A System for Representing

83

it remains to explain some uses; for example, should Chinese in a ‘plural’ sense be considered a -ø or an “irregular” -s substantive? The manner in which the hypothesis for -s confronts new data can be illustrated by means of examples and comments taken from Wickens (99–118) for substantives like pliers. Here it is pointed out that the use of -s with substantives designating an individual binary tool has usually been attributed to the fact that the object is made of two parts. However, the fact that we find expressions like a pair of spaghetti tongs, designating an implement “made of a single, continuous band of metal,” suggests rather that it is a functional opposition between the two extensions which is dominant in the speaker’s mind: “In these various devices, the two terms – the two more or less identical and complementary legs, limbs, branches or blades moving toward each other in opposite directions – cannot be perceived as occupying the same space; they must be perceived as occupying separate parcels of space” (102). Since an impression of functional binarity involving opposed movements necessarily implies spatial discontinuity, it calls for the -s. This explanation based on the ‘continuate’/’discontinuate’ opposition provides a means for confronting other uses. Although no prior studies had brought out -ø usage, many examples like the following are cited from primary sources14: We originated the tongue and groove plier. A cutting plier is a lever with the fulcrum point at the center of the joint. Such uses, found in “encyclopedic and historical studies on tools and especially in trade literature, such as product reviews and catalogues” (104), express a more abstract sense, a type or model of the tool. As a consequence, it is not the functional opposition of an individual tool that is dominant, as in practical situations involving the use of the tool, but rather its “nature or inherent properties,” and so no ‘discontinuate’ impression arises to require -s. Another example of this usage is: A plier is not a hammer or a pry tool or a wrench. Here the noun phrase expresses a ‘generic’ sense, ‘any plier.’ In the following, the substantive is used in the sense of a “functional capacity,” not of an individual tool:

84

Language in the Mind

Pipe plier, wire cutter, and flat nosed plier combined. This example suggests a ‘continuate’ use analogous to that of in hospital, by train, and the like, i.e. a ‘continuum’ actualization of the lexeme. There are even cases of ‘singular’ -s usage (Wickens, 282): … and then the forward bends are made to be looped back parallel around the jaws of a round-nosed pliers. These few examples suggest how the system of number provides the form for different manners of actualizing the lexeme ‘plier.’ This applies to other tool names listed in grammars along with pliers: clippers, forceps, nippers, nutcrackers, pincers, pinchers, scissors, shears, snuffers, tongs, tweezers. Wickens provides a similar illustration and explanation of usage for trousers and some twenty other names for bifurcate garments. The following example will suffice to suggest that here, too, the proposed system can provide the required quantifying form for the sense the speaker wants to express by means of the substantive: There is a rumor going about That skirts are “in,” and pants are “out,” That Fashion will no longer brook The old, androgynous denim-look, That lovely legs will steal the scene, But not in trouser, slack, or jean. This concludes the presentation of the system of number, but we have yet to situate it in the wider context of the substantive as a part of speech. First however we will take stock of what has been outlined in the last two chapters from the point of view of the method of analysis employed to bring out what Guillaume considered basic to all research in psychosystematics.

6 The Method of Analysis in Psychosystematics The golden rule guiding our work here is to convert the observed result back into a process, a genetic process. Guillaume (1984, 133)

THE COMPARATIVE METHOD

In all his scientific work, Guillaume applied the method he had learned from Meillet when studying the comparative grammar of the Indo-European languages. In his eyes, this involved applying the scientific method to language, a method, already well worked out in the physical sciences, presupposing some order or system lying behind and giving rise to observable phenomena. He held the work of the nineteenth-century comparatists in high esteem, considering it the first successful attempt to adapt and successfully apply this method with all its rigour to a human artefact. However since the historical approach deals with only the diachronic dimension of language, he realized that, in order to have a complete science, the synchronic dimension must also be dealt with. Furthermore, while the comparative method had been applied with success to the physical side of language, the sign, it had not been successfully applied to the mental side – meaning – and this too must be dealt with. This project of including all aspects of language thus involved a challenge: to find a way of making use of the same method to analyze language as both a physical and a mental reality and this in synchrony, in the present of speech, in the instant in which the speaker speaks. Only in this way could linguistics become a single discipline, a science capable of treating the whole of its

86

Language in the Mind

object. Without this, a general theory of human language would not be possible. Before going on to describe how Guillaume met this challenge it will be useful to recall briefly the comparative method to bring out more clearly how he adapted it in psychomechanics. I will summarize the presentation from “The Comparative Method in Historical Linguistics and in Psychomechanics” (Valin 1996), replacing the example from the Romance languages by one from the Germanic languages: English foot, Dutch voet, German Fusz, Danish fod, and Swedish fot (cf. Bloomfield, 9, 299), words denoting the same reality. Since the link between sign and significate in language is contingent, arbitrary, it seems “absolutely inconceivable” that the correspondence of the signs and the concordance of the significates could be a matter of chance. Furthermore, once Sir William Jones (cited above in the introduction) observed the relationship between Latin, Greek, and Sanskrit, scholars were no longer led “to derive contemporary languages one from the other” (Bloomfield, 9). Only one alternative remained: these words developed historically from a common ancestor. As a consequence, all five contemporary forms are assumed to have developed from one more ancient stem, the Proto-Germanic *fo¯t-, with a phonetic form reconstructed in such a way that it could give rise to each of them. *Fo¯t- here is an explanatory fact with regard to the observed facts, not because it is observable – like most explanatory facts in historical linguistics, it is not attested – but because it provides the most likely phonetic conditions required for the five attested forms to come into existence. Furthermore, it existed long enough before the facts to be explained for the phonetic changes to have taken place. Figure 6.1 depicts in a simplified way this inferred phonetic development in historical time by means of arrows. The important relationship here, insofar as explanation is concerned, is that one reconstructed form is seen as a conditioner making possible various observed forms, which are seen therefore as conditioned facts. To paraphrase Valin (1996, 39), substituting our example from Germanic: the definitive explanation for the footvoet-Fusz-fod-fot concordance is not simply the fact that *fo¯t- is the oldest form, but the all-important fact that its existence conditions the later existence of foot, voet, Fusz, fod, and fot, each of these forms representing a possible phonetic development of *fo¯t-.

Method of Analysis

87 foot voet Fusz fod fot

BEFORE single form (often reconstructed)

AFTER various forms (attested)

EXPLANATORY FACT conditioner not necessarily observable

FACTS TO BE EXPLAINED conditioned must be observable

Figure 6.1

This is what gives explanations in diachronic linguistics their validity in the eyes of science: “In the mental argumentations which create science, the explanatory fact is seen as containing potentially and virtually the variety and the number, small or great, of facts to be explained” (loc. cit.). The conditioning relationship is thus what gives the comparative method its heuristic power, and it was to this method that Guillaume turned in order to give explanations in synchronic linguistics the same scientific validity. MEETING THE CHALLENGE

Meeting the challenge called for two adaptations of the comparative method: changing the time frame and applying it to meaning. Reducing the time frame from a diachronic stretch of centuries and even millennia to the synchronic span of an instant posed a problem because science explains by means of causal factors, which must somehow exist prior to the facts to be explained. Guillaume early realized that the part of language that resides permanently in the preconscious mind – tongue – is the prior conditioning factor that gives rise to words and sentences in the present of speech while the speaker carries out an act of language. This view may have raised a few eyebrows when first proposed, but today, with all the research going on in the neurosciences, we are aware that even an ordinary act of perception involves many neural operations, all of which are carried out “nearly instantly,” “in mere milliseconds.”1 There is then nothing surprising in Guillaume

88

Language in the Mind

proposing that the many mental operations involved in actualizing the potential to reconstruct a word for use in the sentence are all carried out during the present of speech. As Valin (1996, 42) puts it: “the crucial moment under discussion here is actually the extremely narrow lapse of time when the conversion of potential language to actualized language takes place.” That is, like any other potential, physical or mental, a potential meaning or a psychosystem exists before the actualized meaning comes into existence, long enough before for the actualizing operation to take place. In this way, whether the explaining fact – be it the reconstituted form in diachrony or the system of potential meanings in synchrony – is separated from the facts to be explained by millennia or by milliseconds, it pre-exists and so is in the position of a conditioner conditioning the data and permitting the analyst to understand and explain an observed use. In this respect, then, the method of analysis Guillaume proposes for synchrony is essentially the same as that found in diachronic studies. On the other hand, Guillaume realized that the nature of the object under study determines how it must be observed and analyzed and so he was never tempted to identify the way the scientific method is implemented in the experimental sciences with the way it is implemented in linguistics. In the former, it is often possible to determine all the factors conditioning a given phenomenon and so predict its occurrence; in linguistics, however, this is not possible, if only because of the speaker’s inherent individuality and freedom, so that one can never predict, in the scientific sense, what a person will say. Linguists must be content with isolating the conditioning factors arising in tongue – systems and the potential meanings of lexemes and morphemes – since as linguists they are not competent to analyze the social, psychological, physiological, and other factors conditioning the speaker. This sets Guillaume off from some contemporary linguists: he did not consider linguistics to be part of another discipline, such as cognitive psychology, because he could not assume that the method of analysis he had worked out for language was applicable to fields of inquiry outside language. On the other hand, he maintained that it is applicable to whatever is part of language. That is to say, as a linguist his aim was to analyze language, the whole of language, and nothing but language. And this led him to focus on its raison d’être, meaning.

Method of Analysis

89

Finding how to apply the comparative method of analysis to the mental side of language in synchrony called for lengthy reflection on the part of Guillaume. Only after he had discerned the operational relations between the moods in the French verb (see below, chapter 10) did he realize how systems, hidden entities of tongue, were constituted. He understood that the meaning components of a grammatical system, the morphemes, are temporally related in the mind, sequenced in time by the need to think one before another – what he called their “notional chronology.” We saw this in the system of number, where it appeared necessary to access the position of ‘singular,’ ‘one,’ in order to access the position of ‘plural,’ ‘more than one.’ Because of this notional chronology we concluded that in the system of number the field of the continuate necessarily precedes that of the discontinuate. Recognizing a sequence between ideas in this way showed Guillaume how to apply the comparative method to meaning: examine the actual significates for indications suggesting a temporal order, a notional chronology, between the notions expressed by different items belonging to the same system. By thus introducing a temporal dimension into grammatical systems, he was led to conceive of them as operational by nature. His analysis of different systems in effect demonstrated that even in its systematic framework, its structure, language is dynamic. In this sense, Guillaume considered himself a structuralist. This insight into the dynamics of language permitted him to analyze a number of systems, but it took further years of reflecting on the articles in French before he made the next important step: the realization that the different actual significates of a single morpheme are also temporally linked. He was thus able to extend the method to the individual morpheme, observing its polysemy to seek indications of a sequencing of senses. We saw this in examining the uses of -s: taking for granted that all three quantitative senses are expressed by the same morpheme, it seemed necessary for the ‘one’ sense to arise before the ‘more than one’ sense, and both before the ‘all’ sense in the field of the ‘discontinuate.’ This notional chronology led us to assume that the potential significate of -s is dynamic, permitting a movement whose three necessary moments (beginning, middle, and end) give rise to three different results in discourse. Examining the different senses of -ø for

Language in the Mind

90

indications of sequencing in time led to the conclusion that they too arise from an underlying dynamic potential meaning. In this way Guillaume’s positional technique (positioning in time) has permitted us to see that the actualized significates (the diverse facts to be explained) are all conditioned by the potential significate (the one explanatory fact).

BEFORE single meaning (always reconstructed)

‘all’ ‘more than one’ ‘one’

discontinuate quantities

M -s I m

field of ‘discontinuate quantity’

potential significate actualized significates

AFTER various senses (attested)

EXPLANATORY FACT conditioner not observable

FACTS TO BE EXPLAINED conditioned must be observable

Figure 6.2

Figure 6.2, for -s morpheme, where the arrows depict the microtime required to actualize the potential, summarizes this description. This is how Guillaume adapted the comparative method to the mental side of language in synchrony. In Valin’s words (1996, 42): “Just as the comparativist, guided by the unity of meaning attached to the correspondences, must imagine the single phonetic form that can explain the diversity of attested pronunciations, so the Guillaumian, guided by the unity of the sign to which the various observed expressive effects are attached, must imagine the single potential significate which can explain all such effects.” Temporally based analyses of systems and morphemes convinced him that microtime is the dimension along which the analyst can gain access to the hidden potential of any morpheme or system. A phenomenon observed in discourse he assumed to be the actualization of one of the positions in the hidden phenomenon, the mental operation. Adopting a dynamic conception of morphemes

Method of Analysis

91

and systems in this way – a conception of structure that, to my knowledge, is not found in any other theory – gave Guillaume a basis for observing any problem in discourse: “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” (1984, 134). In thus postulating that intercepting a preconscious operation gives rise to the observed use in discourse Guillaume provides a method of analysis applicable to whatever arises from the resources of tongue because it uses the same parameter for all: operative time, the time involved in the representational operation inherent in any entity of tongue. That is, be it a lexeme, a morpheme or a system, its different manifestations in discourse can be distinguished by the moment in the underlying operation that gave rise to each one. Being general, this method provides a starting point for exploring any problem because it tells us what to look for in the data, but by the same token it imposes a requirement that is never easy to meet: to describe the mental operation as a potential in tongue, as capable of producing all the observed senses. Until this is done, there is no understanding an observed use, no explanation. Although Guillaume did not succeed in providing a satisfactory analysis of lexemes, he did meet the challenge in the case of a number of morphemes and systems in French, thus showing how effective the method can be. On the basis of what we have seen in the two previous chapters, we can now discuss what is involved in applying it. Aware that linguists, whether working in diachrony or in synchrony, are limited in their field of competence, Guillaume considered their task one of working out the prior language factors conditioning what is observed in discourse, but not the extralinguistic factors. Being thus limited to language, his method permits us to understand and explain the use of a given form or type of expression as the actualization of a linguistic potentiality. Since linguists cannot obtain a clear view of the necessary and sufficient conditions of an observed phenomenon permitting them to predict it, proof through experimentation based on predictable results is not possible. To provide proof for linguistic hypotheses in synchrony like those proposed above for -s and -ø, as in diachrony, the only alternative is to accumulate more and more data, observations

92

Language in the Mind

of usage, that the hypotheses explain satisfactorily. It is important, therefore, to understand what is involved in observing the actualized significates of a morpheme. PRECONDITIONS FOR OBSERVING

When Guillaume uses the expression positional linguistics (1984, 51) to designate the technique involved in applying this method, it reminds us that the guiding principle when observing is to discern indications of the position in the underlying genetic process giving rise to the sense of the morpheme in the observed use. That is, Guillaume, like many others, approached the observation of usage with certain preconceived ideas as to what he was looking for. Not that he necessarily had a particular hypothesis for a given problem before examining the data, but rather he approached any problem with certain postulates concerning the nature of language that appeared to him incontrovertible, inevitable. “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” (1984, 23). Thus a theory, like that proposed above for number, has the “double test of a proper start and of a finish which corroborates the appropriateness of the start” (1984, 24). In an attempt to make these exigencies explicit, Guillaume starts with one that is obvious: to have language one must have human beings, speakers in touch with the universe (including other humans) through their experience. Obvious though this may be, it does lead him to observe that whatever one speaks about is derived from the experiential relationship between universe and speaker. Furthermore it led him to view our experience as the starting point from which all grammatical constructs are derived by generalization (for an example see the discussion of auxiliaries below). To hypothesize some sort of preconceptual framework, universal logic, or autonomous formal system as a basis for constructing a grammar did not appear inevitable to him. Rather, like most scholars, he assumed the ability to generalize and particularize to be native to human thought. This capacity to generalize from experience enabled the mind to construct the systems of tongue and provided him with the postulate on which to reconstruct them.

Method of Analysis

93

Working from these assumptions, Guillaume considered that our ongoing experience, our conscious awareness, is something mental and is a product of the mind, thus attributing to the mind its own existence and functioning distinct from, but not independent of, that of the brain. This is not an unexamined a priori for Guillaume but rather a principle confirmed time and again by the analysis of grammatical systems and by the nature of the word itself. As a consequence, characterizing mind, thought, concept, etc. as “convenient reifications,” or considering that “a thought is the occurrence of a complex neurological, ultimately electrochemical event,” or that “Mental experience… is what the brain does” (Langacker 1987b, 100), he would view as presuppositions not supported by evidence from language. He of course recognized that without the neuromechanisms permitting the processes of the brain the human mind cannot function, but instead of reducing the mental to the physical in this way he postulated the existence of psychomechanisms permitting the mental processes that produce the meanings observable in discourse. In the final analysis, questions such as the nature of thought or of the mind he considered to be so general that they exceed the competence of linguists. Like anyone else seeking a scientific explanation, Guillaume further assumes (cf. 1984, 46) that there is, hidden beneath the observable in language, a “design,” a “systematic schema”; it is “a certain order that is pleasant to contemplate” because it allows us to understand data. That is, discerning this hidden order is a prerequisite for explanation since “in itself the concrete [a particular use] may be observed, but not understood” (1984, 43). Moreover, as in other sciences based on observation, this order provides conditions for bringing the data into existence: it is a system permitting mental operations and so requiring time, operative time. Thus he conceives of the hidden system of language as an orderly set of psychosystems making possible the preconscious operations that form our lexemes to commute a personal experience into the particular use to be observed. The psychosystem is, as we have seen, a coherent relating of potential meanings, of morphemes, whereas the different signs linked to the morphemes must be expressive enough to reflect the meaning system but are not in themselves systematic. That is, there may be “irregular” signs but not “irregular” meanings2: “in mental systematization it is the law of greatest possible coherence that prevails … In semiology the need for coherence is secondary: coherence

94

Language in the Mind

can always be sacrificed without harm as long as expressive sufficiency benefits” (Guillaume 1984, 74). The greatest possible coherence for the significates, sufficient expressiveness for the signs – these are postulates that Guillaume explicitly adopts for psychosystematics. Since the semiological forms are not necessarily systematic, coherently related, this allows for “irregular” forms like the strong verbs in English. Guillaume’s position is therefore not that of Jakobson for whom “formal differences are always meaningful” (Waugh, 43). Guillaume thus observes any sentence from the point of view of both the speaker’s momentary experience and the language potential with its universe of viewing ideas and set of psychosystems. Thanks to this potential, tongue, the sentence expresses the speaker’s representation of that experience, the intended message, in the sentence. A linguist-observer also brings to bear all that is known from previous studies since any valid observation, regardless of its theoretical setting (if any) constitutes a fact to be explained. For example, that -ø signifies ‘one’ or ‘singular’ and -s ‘more than one’ or ‘plural,’ and that both have to do with number, as all grammars of English observe – these were, when we undertook our study, already known facts guiding further observation of usage. These were moreover viewed as facts to be explained within the framework of the theory, i.e. as actualizations of an underlying systemic potential. As a consequence we were not tempted to consider that this is all -ø and -s could express, a point of view often found in grammars, where other uses (e.g. ‘plural’ -ø and ‘singular’ -s) are, if mentioned at all, considered as marginal or as exceptions. In their consideration of the system of number, many studies do not even include the use of -ø with a lexeme in a ‘continuum’ sense, presumably because -ø here expresses something that is not numerable, as Jespersen pointed out. Our approach, on the other hand, based on the assumption that we were dealing with a system of representation (and not just a means of communication), confronted us with the problem of how a single morpheme like -ø could represent both what is numerable and what is not numerable, and this ultimately led us to propose a more general system than that of ‘singular’ vs ‘plural,’ one based on ‘continuate’ and ‘discontinuate’ representations of quantity. Without the assumption that a morpheme is a dynamic component of a dynamic system in tongue, the polysemy of items tends to be

Method of Analysis

95

ignored as a reality of discourse, or at least its importance for the observer is not recognized. With this assumption in mind, however, the observer actively seeks the unusual, less frequent use (like the -s expressing ‘singular’) since what it reveals of the underlying system may be different from what the more common, prototypical use (-s expressing ‘plural’) reveals. Granted that meaning and how to deal with it is, as Langacker (1987b, 5) puts it, “the most fundamental issue in linguistic theory,” this dynamic view of meaning constitutes an important contribution because it provides an orderly way for dealing with polysemy. WHAT TO OBSERVE

Observers must then approach usage – facts to be observed – with a knowledge of what has already been observed and an awareness of the theoretical postulates conditioning their observing. Another condition for adequately observing language is, obviously, to have as clear an idea as possible of what one is trying to observe. For Guillaume it must be a piece of normal discourse, something said or written to express a particular experience. That is to say, notwithstanding its different theoretical framework, psychomechanics is, like cognitive grammar, “a usage-based theory” (Langacker 1987b, 46) in which the observation of the different uses of a word is the prerequisite for, and a consequence of, analysis. This means that an example invented to illustrate questionable or even unusual usage like ‘singular’ -s does not provide valid data because the listener or reader, if unable to interpret it, cannot relate it back to a message, the experience that prompted it. This constraint in turn imposes on the observer the painstaking task of finding sufficient examples of a new or unusual use,3 in context, to permit understanding the use and getting back to the speaker’s intended message before attempting to analyze it in the light of the theory. In psychosystematics, the facts to be observed in any example are the grammatical sense expressed by an inflection or grammatical word (and even a syntactic position) and what it contributes to the meaning of the phrase and ultimately the sentence. A meaning, however, is a mental entity and so cannot be given a physical mode of existence and made measurable. It is therefore observable only through introspection, though its presence is indicated by some semiological means. Thus anyone who speaks the language

96

Language in the Mind

understands it and becomes aware of what the sentence means, but they are not necessarily in a position to isolate the sense of a particular element in the sentence. However, granted the ordinary speaker’s ability to understand, a morpheme’s sense can be observed by anyone who takes the trouble to isolate it through commutation in minimal pairs and varying the context in numerous examples.4 The outcome of observing meaning in this way is crucial: a consensus of competent observers is required to ensure the validity of the data constituting the empirical basis of the theory. The problem involved here – that of adapting one’s means of observation to the mental reality to be observed – calls for comment. In most cases, distinguishing the import of the morpheme causes little difficulty because the meaning expressed by the phrase or sentence makes it clear to the reader or listener how the speaker has thought the word. For example, all observers would probably agree that in ordinary uses like There are forks on the table the -s has its ‘more than one’ sense, whereas in Dogs are vigilant it has its ‘generic’ sense. Furthermore, even in unusual uses like a crossroads or two bear, observers would agree that the noun phrase denotes a single unit in the first case and more than one animal in the second, and this suffices to ensure that the data is valid, even though there may be no consensus on the level of analysis in accounting for the presence of -s or -ø. The importance of a consensus concerning the data can be brought out by looking at a different explanation which was put forward during a discussion of these two uses. Someone advanced the counter argument that in such cases the -s of crossroads has a ‘more than one’ sense and the -ø of bear a ‘one’ sense but that the quantifiers one and two tell the listener to interpret the substantives respectively as ‘singular’ and as ‘plural.’ This amounts to arguing that the substantives get their meaning insofar as grammatical number is concerned from another element in the context. The objection to this proposal is that it is not based on the observed facts: the ‘one’ sense attributed to the -ø morpheme of bear is not part of the meaning of the noun phrase, not part of the data. Certainly the quantifier enables a listener to interpret the noun phrase already constructed and spoken, but the linguist who wants to explain what it expresses must discern the conditions leading up to the sentence meaning and so must adopt the point of view

Method of Analysis

97

of the speaker who constructs the phrase. And from this operative point of view one wonders why any speaker would represent as plural something experienced as singular, or vice versa. Rather, as we shall see when we examine the syntax of auxiliaries and of articles, the speaker represents each element in the phrase (or sentence) so that it will contribute to the meaning to be expressed, a position quite distinct from those who, adopting the point of view of the listener, assume that the import of some item may be foreign to the meaning of the phrase or sentence. I got the impression that the person who raised the above counterargument was prompted by the assumption that a morpheme is monosemous5 – that -s always means ‘more than one’ and -ø ‘one’ – an assumption probably based on the realization that communication as we know it would be impossible if a morpheme did not have a constant, unchanging meaning. This is why Guillaume considered the potential meaning in tongue to be constant and unchanging for a speaker, whereas he expected to find in discourse all the variation permitted by the potential. To suppose monosemy of a morpheme or grammatical word in discourse would be a serious error. It can be avoided if observers keep in mind the inherently operational nature of a potential meaning because this will lead them to look for evidence of polysemy as an indication of the underlying meaning in tongue. This dynamic view of the underlying meaning does not, on the other hand, warrant the assumption that all morphemes are necessarily polysemous in discourse. Only observation of the usage of each one can determine whether or not its meaning potential has been exploited to give rise to different actual senses. The observer’s aim is, then, to ascertain in what sense the speaker has thought the morpheme, i.e. what actualized meaning it contributes to the “complete thought” of the sentence. In order to do so, the observer, like the ordinary listener, is dependent on context and situation to provide guidelines. This is what happens, for example, when someone begins a sentence with a substantive like “Horses…” or “Wine…” and then pauses. As listeners we must have more context before we can work out how the speaker has actually thought the substantive (cf. Horses require exercise ‘generic’ vs Horses were left to roam ‘plural’; Wine is made from grapes ‘generic’ vs Wine will be served with the meal ‘continuum,’ ‘indeterminate

98

Language in the Mind

quantity’). On the other hand, it would be a mistake to consider that the speaker must await the rest of the sentence in order to think the substantive, i.e. in order to call to mind the lexeme formed by one of the number morphemes expressing a particular sense. It is the intended message arising before the act of language, what the speaker wants to talk about, that determines the lexeme and its number. The same holds for the other words that go to make up the context, because each word is called on to represent something in or arising from the speaker’s momentary experience (and to fulfill a role in the sentence). Thus the observer tries to get back to the relation between morpheme and intended message, tries to discern how the potential meanings in the system of number focus on the experiential entity in order to represent it. By concentrating on the speaker in this way, psychosystematics attempts to trace the causal chain from its first link, representation, through to the resulting discourse and reference, since discerning prior conditions is what permits us to understand and explain what we observe. To get back to the crucial point: attempting to observe the sense in which the speaker actualizes each morpheme (and lexeme) in this way inevitably raises the question of the validity of resulting observations. How can one be sure that this is what the speaker thought? After all, meaning is a mental construct that can never be “physified,” never occur in a physical state like its sign, and so any observation of it is necessarily introspective. As pointed out above, the only basis for considering that any such observations are valid is that they are based on the consensus of competent observers, a condition that applies to all scientific data. In our case, this comes down to the ordinary circumstances of communication because a competent observer is one who understands English, who feels the difference between two bear and two bears, between one heads and one head, between a single and a singles, who gets the message that many Micmac and these people represent more than one individual, whereas this people and a headquarters and an outstanding opening ceremonies represent single entities of a particular sort. Such facts, which the ordinary speaker takes for granted, are data for the observer, facts to be brought into full awareness and explained along with the data provided by the physically perceivable signs. These then constitute the mental reality the

Method of Analysis

99

psychosystematics of language tries to observe and then to organize by means of the positional technique in order to explain them as observed facts. HANDLING THE DATA

Confronted with diverse senses expressed by a given morpheme, the observer must try to discern some order or coherence in the data. This is no longer a matter of observing but of reflecting on how observed facts relate to the underlying potential postulated for them in tongue. Thus in the case of -ø it seemed fairly obvious that there is an order between its different numerical senses as illustrated by the following ordinary uses of the substantive salmon: (1) Salmon migrate. (‘generic’) (2) They caught some salmon yesterday. (‘plural,’ ‘more than one’) (3) Someone gave us a salmon. (‘singular’, ‘one’) All of these uses can thus be considered as arising from the same mental operation, that of bringing the experiential entity in the intended message (the species, the fish making up yesterday’s catch, a single fish) into relation with the whole range, from the greatest possible quantity to the least possible quantity, of what can be designated by the substantive. And this is what any numerical use does: in expressing, say, ‘more than one’ it implicitly excludes the others, ‘all’ and ‘one.’ Since the proposed operation covers all possible numerable experiences of an entity, it provides a meaning potential capable of giving rise to all such uses. As is so often the case, however, other uses were found that complicated the picture: (4) We had some salmon for supper. (‘continuum,’ ‘indeterminate quantity’) (5) Salmon is a delicacy. (‘continuum,’ ‘generic’) The problem of how to relate this type of use to the proposed potential was not solved until ambiguous examples like the following were examined:

100

Language in the Mind

(6) I bought some salmon. (‘continuum,’ ‘indeterminate quantity’ or ‘plural,’ ‘more than one’) While the ordinary listener would understand the sentence in the light of the person speaking (e.g. someone who has just done the week’s shopping vs a buyer for a fish market), the very fact that ambiguity is possible indicates that salmon can be understood in two different ways, as either the flesh or as individual fish. Since ambiguity like this is not very common, occurring mainly with names of edible fish, we concluded that it is the lexeme ‘salmon’ that makes it possible, that the lexeme is itself a potential in tongue offering these two possible actualizations. Although the form of operation involved in actualizing the lexeme has not yet been discerned, it seems clear that the ambiguity does not involve any difference in the morpheme itself. That is, regardless of whether the lexeme has been actualized in a ‘continuum’ or a ‘unit’ sense, the operation permitted by the potential meaning of -ø intercepted at its beginning gives a sense of maximum quantity, and intercepted at some midpoint in its course gives the sense of an intermediate quantity. Three points of method are illustrated here. First, it is important to distinguish between the grammatical and lexical imports of a word. Only when this was done were we able to see that -ø is not limited to representing ‘singular’ but is capable of representing quantity whether numerable or not. That is, this morpheme is quite general, capable of quantifying whatever lexical matter is provided to form a substantive. Second, in order to handle new data, here ‘continuum’ (‘mass’) uses, we had to reflect on, try to imagine, the conditions of representation in tongue which could give rise to both formerly observed usage (ordinary ‘singular’ substantives) and the non-numerable uses, yet to be integrated into the system. This led to the proposal that the field of -ø in tongue is that of continuate space, a point of view that ultimately allowed us to explain the subtle distinction between two bear and two bears. Third, what permitted us to generalize our initial view of the meaning of -ø as ‘singular’ (as it is presented in most grammars) to see it as just one case of ‘continuate quantity’ in this way was the careful observation of more and more uses and relating each new use back to the hypothesized potential meaning to check if it can give rise to each use as one of its actualized senses. And this second phase of observation aims at finding, not more cases of common uses,

Method of Analysis

101

but examples of less common uses that may have been overlooked, because they may either confirm a hypothesis or reveal something requiring it to be modified. This manner of observing differs from that in generative grammar, where concern with the formal conditions of grammaticality leads to focusing on common uses and making little attempt to find the rare use. As Guillaume puts it (1984, 23), in psychosystematics the need for the proposed meaning potential to “confront the facts from the position of antagonist” is crucial because it is the facts, the different senses expressed by a given sign, that determine in the final analysis the validity of an explanation. As a consequence, every effort is made, as in cognitive grammar, to omit no pertinent data, regardless of its frequency in usage. On the other hand, in psychosystematics, as we have seen, there is no attempt to distinguish between prototypical and peripheral senses but rather to seek the relation between senses in notional chronology, regardless of their prominence, in order to throw light on the hidden potential. As a consequence, there is no temptation to derive one observable use from another, which would amount to repeating the error of early comparatists of ”deriving contemporary languages one from the other,” instead of seeking a prior condition for explaining all uses. The method we have just described in some detail for relating the various uses of -ø and arriving at a hypothesis concerning its potential meaning is applicable to other morphemes and so there is no need to recall its implementation for the -s inflection. More important is to reflect on the two resulting hypotheses in order to see how the two morphemes are related so as to constitute a coherent whole, a system. RELATING THE MORPHEMES

We have just seen how the postulate of order behind the apparent “chaotic diversity” (cf. Einstein 1956, 98) of observed usage led to discerning the relation between the different senses expressed by -ø as different magnitudes of something represented as a ‘continuate,’ and for those expressed by -s, as different magnitudes of something represented as a ‘discontinuate.’ The same postulate led to reflecting on the relation between the two potential meanings signified by -ø and -s, and this brought to light an order in everyday

Language in the Mind

102

thinking, the notional chronology implied by these two ways of representing a spatial entity: it is necessary to be able to think ‘one’ as a prerequisite to thinking ‘more than one.’ That is, to represent any entity as a discontinuate, as ‘plural,’ it must be representable as a minimal continuate, as ‘singular.’ Thus there appeared to be a relation of order between the two morphemes, -ø preceding -s, an order in the extremely brief operative time required to represent either ‘continuate’ or ‘discontinuate’ in one of its possible magnitudes. The outcome of these reflections was the system as proposed in the previous chapter (see figure 6.3).

-ø M

I

-s m

I

M

Figure 6.3

This system calls for several comments because it is typical of the type of result obtained in psychosystematics. In the first place, conceiving a grammatical system in terms of an operation of representation involved in the word’s “psychogeny” (cf. Guillaume 1984, 134) – the construction of a word’s mental content, its binary significate, as opposed to its sign – makes Guillaume’s view of grammar distinctive, setting it off, for example, from approaches that view grammar as essentially a theory of syntax concerned with the relations between words. It is also to be distinguished from a “non-constructive conception of a grammar” proposed in cognitive grammar, which “does not consider the grammar a constructive device” (Langacker 1987b, 64f). It requires the speaker to activate the potential operation underlying the system of number whenever confronted with an experiential entity whose lexeme is to be formed as a substantive. This of course leaves the possibility open for a speaker to make an innovative use of a morpheme with a given lexeme, to construct the substantive in a new way to correspond with the intended message, as in the telephone booth example discussed previously. Such innovations are usually nonce uses, but

Method of Analysis

103

if someone characterizes a recurrent experience with such aptness that the new use is taken up by other speakers, especially children acquiring their mother tongue, it may lead to a long-term change in exploiting the lexeme. In this way, considering grammar as a mental system operating in synchrony provides a basis for initiating changes in diachrony. Another point to be brought out is the degree of abstractness of the system. There is no possibility of expressing a specific quantity other than ‘one’ by means of either morpheme (though of course other elements in the sentence often specify the quantity). The reason for this is that the particular magnitude expressed depends on the position in the operative time best corresponding to the referent in the intended message, the point where the morpheme’s movement is intercepted. Because in any movement there are three necessary moments – beginning, middle, and end – there are only three critical positions for a morpheme. As a consequence the intermediate position (I) for -s may correspond to specific quantities from two to millions, whatever quantity one may wish to imagine, provided it is neither the fewest possible (one) nor all possible entities, because that would be the M (maximum), i.e. end, position. Furthermore, because the same sign expresses the three different senses, the listener, like the speaker, must activate the system to reconstruct that actual meaning of the morpheme that is most appropriate to the particular context and situation. The above system also illustrates an important constructional principle of all language systems: “the mind has the potential, the capacity, of intercepting its own processes, of taking cross-sections of them” (Guillaume 1984, 134). It is the possibility of intercepting the movement permitted by a morpheme that gives rise to its observable polysemy, a capacity, it should be noted, Guillaume considers not as specifically linguistic but rather as a potentiality of the mind itself.6 This principle underlying systems led him to the view that a representation obtained by means of language is essentially spatial, that even time with its inherent mobility must be spatialized, held up, and represented by spatial traits, as we shall see in the chapter on the verb. Because of this spatial character of linguistic representations, the relationships involved in systems can be conveniently depicted by means of figures like the above. On the other hand, systems require the microtime of mental processes

104

Language in the Mind

to produce results, and as a consequence this operative time provides the parameter for analyzing them. The fact that the system of number (and other systems) is so abstract, making quantity representable whether it be numerable or non-numerable, led Guillaume to reflect on its binary nature, on why all possible magnitudes from the greatest to the least should be representable in two different versions. Because the operation underlying the system consists of two movements, one the reverse of the other from the point of view of quantity, he saw them as stretching or tending toward different ends (often calling them “tensions”) and so conceived of the system as a bringing into relationship of the extreme terms, the greatest and the smallest. To obtain a complete view of any relationship calls for seeing each of its terms in the light of the other. That is, a system must be binary (at least) to represent the relationship as a whole and so provide for all possible manifestations in discourse. This brings us to what Guillaume calls “the very important law of non-recurrence” (1984, 54), always implicit in psychosystematics. Since going backwards in real time is an impossibility, the movement signified by -s cannot be conceived as simply the reverse of that signified by -ø. Spatially it is a mirror-image of the -ø movement, but temporally it is different because it extends away from the -ø movement, occupying a different place in microtime. That is to say, if a system is dynamic its second phase cannot be conceived as returning to the point in time where the first phase started since this would involve recurrence in real time, in operative time. This succession in time is what distinguishes the system’s two phases and gives different senses to the morphemes even when they express equivalent quantities. Reflection on the system of number and on that of the article led Guillaume to conclude that the mechanism of bringing into relationship two terms was a mental operator. He called it the radical binary tensor because he found this twofold movement at the root of a number of systems. And he ultimately proposed that it is an operator underlying all language systems, identifying its source in a generally accepted characteristic of human mentation, our ability to generalize and to particularize. By tracing synchronic systems back to a native potentiality of the human mind itself, he introduced a systemic dimension into the diachronic study of grammatical systems.

Method of Analysis

105

THE METHOD IN PSYCHOSYSTEMATICS

The brief survey of usage in the previous chapter shows how the method can handle the data by reconstructing the system and providing an explanation for the observed expressive effects. In most cases, what is a vague impression arising from the initial observation of a use is made much more precise by discerning the contribution of the number morpheme to the meaning of the noun phrase. Furthermore, observing discourse from the viewpoint of the system brought to light many uses that had not caught the grammarian’s eye. Such results help to confirm the theory and support Guillaume’s method, which “is a continual back-and-forth between the concrete and the abstract. Abstract reflection is required to guide observation of the concrete and make it more penetrating” (1991, 192). In this way he attempts to discern the mental reality of a grammatical system. In chapter 3 we saw that Guillaume’s first contribution to linguistics is a genetic theory of the word, an overall view of how a word is formed so that it can play its part in the sentence. His second contribution, a corollary to the first, is this method for analyzing the data and obtaining a hypothetical reconstruction of the preconscious system, essentially the same method as that employed in comparative and historical work, by seeking temporal relations between observed facts and imagining the process giving rise to those facts. Besides uniting linguistics, hitherto divided between the diachronic and the synchronic, this contribution opens up new perspectives for research throughout the systematic part of language – in grammatical semantics and in syntax, and even in lexical semantics to the extent that items are systematized. To my knowledge, there is no other approach that applies this method to language in synchrony. To bring out the importance of this contribution we will now go on to sketch briefly the makeup of the substantive, the system of the parts of speech and the verb.

7 The Substantive: A System of Subsystems A Substantive or Noun is the name of any thing that exists, or of which we have any notion. Murray (1799, 32)

INTRODUCTION

Murray, like many other grammarians, defines the substantive in terms of its extralinguistic referent, what it names or denotes. That is, he tries to characterize in a very general way what its lexeme represents. This manner of conceiving the substantive has not found favour among linguists, who consider it a grammatical entity. As a consequence they tend to regard the substantive in terms of the different syntactic roles it can play in the sentence. Each of these two approaches reflects one of the substantive’s functions, but neither provides an overall view permitting us to characterize the substantive as such. Guillaume’s approach was, as always, to work back from what is observed in usage in an attempt to understand what permits the substantive to function in these two ways. This led him to examine its morphogenesis, the process of forming a lexeme into a substantive. In this chapter we will outline what has been discerned so far in the morphogenesis of the English substantive. The overall categorization, i.e. the resulting part of speech, will be treated in the next chapter. Because of the grammatical inflections in former stages of English and throughout the Indo-European languages it is traditional to accept gender and case along with number as formative elements of the substantive. Many grammars consider all three to be categories of the English substantive even today. While nobody questions the presence of number, the existence of gender and

The Substantive

107

case in contemporary English is sometimes questioned. Our first task, therefore, is to see if there is reason to consider them as phases of the morphogenesis of the substantive. The high visibility of number in English has permitted us to examine it in some detail. Because these other formative elements are far less visible they have not yet been explored to the same point. However, bringing out the little we do know of them in this chapter will not only indicate the direction of present research but will also help to situate the system of number and give us a first view of the overall system of the substantive. GENDER

In French, Guillaume distinguished between gender as expressed by adjectives and determiners and gender as expressed by the substantive. The former, as in la chaise (‘the chair,’ feminine) and le fauteuil (‘the armchair,’ masculine), he called “fictive” gender because it corresponds to no distinction on the level of meaning.1 The latter, sometimes expressed by the semiology of the substantive, as in berger/bergère (‘shepherd, shepherdess’), he called “real” gender because it reflects a distinction of meaning between ‘masculine’ and ‘feminine,’ two manners of actualizing the same lexeme. In all uses of substantives to designate a particular person the same distinction arises even though it may not be reflected by the semiology. Thus mon médecin (‘my doctor’) is thought as either ‘feminine’ or ‘masculine.’ The possibility of this distinction does not arise with lexemes designating things. This leads Guillaume to say (1999, 40): “The hallmark of the masculine and the feminine is to be, each in its own way, the singular of a whole composed of two which complete one another and make up the pair.” And so he concludes that real gender based on an ‘animate’ vs ‘inanimate’ opposition is part of the substantive’s morphogenesis in French, with a ‘masculine’ vs ‘feminine’ opposition within the ‘animate.’ To justify the existence of gender in the English substantive, grammars often cite pairs like waiter/waitress and prince/princess where -ø vs -ess makes the distinction between ‘masculine’ and ‘feminine’ gender clear. Pairs such as boy/girl and king/queen express the same distinction of meaning without any suffix. Far more common to designate persons, however, are words like student, friend, driver, etc. that do not provide the evidence of paired

108

Language in the Mind

signs but only that of obliging us to think either ‘masculine’ or ‘feminine’ when they designate a specific individual, as in I consulted my doctor. When such words are used non-specifically, as in He needs a doctor, where the speaker has no specific doctor in mind, we understand neither ‘masculine’ nor ‘feminine’ but rather a more general notion of gender that, if actualized in an individual, would be either one or the other. Thus it seems that when the speaker uses a substantive to designate a person, its lexeme is given the grammatical form of ‘masculine’ or ‘feminine’ or the non-committal epicene form allowing for either. This has sometimes been called personal gender since it is found mainly, but not exclusively, in substantives designating a person. A number of substantives designating higher animals also have distinct signs for male and female individuals: lion/lioness, dog/bitch, stallion/mare, etc. Although we usually speak of lion, dog, or horse without making any distinction of gender because an impression of the sex of the animal is not generally part of our intended message, for anyone breeding and raising such animals it would be only normal to think of an individual as either male or female and so to represent them as ‘masculine’ or ‘feminine.’ In order to avoid excluding cognoscenti usage, therefore, it is more appropriate to speak of “animate” rather than “personal” gender, with the understanding that for most speakers animate gender in the substantive is usually restricted to substantives designating persons. These observations of the use of substantives provide evidence of animate gender when the speaker’s experience of an entity in the intended message involves a distinction of sex. This is not, however, to suggest that gender is simply a representation of the sex of the individual. It is the viewing idea, the lexeme distinct from all other lexemes, that represents certain characteristics of the entity in the intended message. Grammatical categories configure this lexical input in order to constitute a word capable of playing a role in the sentence. Animate gender configures a lexeme in such a way as to provide for a further opposition within the lexeme – “otherness within sameness” as one linguist puts it (Morris, 201) – whereas inanimate gender forms it without any such possibility of opposing two versions of the lexeme. Number, as we have seen, also forms the lexeme but what distinguishes it from gender is its variability from one use of a substantive to another. There is however no evidence to show that the gender of a substantive like lawyer

The Substantive

109

or desk can vary between ‘animate’ and ‘inanimate’ in this way,2 and this would explain why, unlike number, it requires no specific sign: it is determined by the lexeme. That is to say, gender appears to be a generalization of something inherent in the lexeme, something with an effect on its spatial disposition. Understood in this way, gender calls for representing certain entities by means of a lexeme that can be grammatically formatted as either a ‘masculine’ or a ‘feminine.’ Guillaume recognizes in lexemes permitting this “a symbiosis of two different beings” (1999, 38) constituting a dual, one of whose constituents is actualized when representing an individual person. How about all the other cases where there is no impression of a sexual being in the intended message? Words like book or economy represent such a situation and are called inanimates or neuters. They express a meaning permitting no such binary categorization of the lexeme and so do not arise in pairs like actor/actress, nephew/niece, or even in paired expressions like male driver / female driver. That is to say, even when designating an individual entity, a substantive formed as ‘inanimate’ can call to mind neither ‘masculine’ nor ‘feminine.’ There is therefore reason to accept the idea that the ‘inanimate’ vs ‘animate’ opposition is an element of meaning, a grammatical form of the substantive, even though there is no sign to indicate it in the vast majority of cases in English. When it forms the lexeme as ‘animate,’ the substantive arises in discourse expressing ‘masculine’ or ‘feminine’ or ‘either.’ When it forms the lexeme as ‘inanimate,’ the substantive appears in discourse expressing ‘neither masculine nor feminine,’ that is, ‘neuter.’ Viewing gender in this way brings out its indirect link with our experience of sexual distinctions, a link often manifested by the euphemistic use of the term “gender” to designate the extralinguistic phenomenon rather than its grammatical representation. Being a grammatical form imposed on the lexeme, gender categorizes an impression inherent in the lexeme, itself a representation of the experiential entity. This view of gender also serves to bring into focus some questions requiring further observation and reflection. One such case arises with expressions like male language, female voice, female sensitivity. The message one gets from these expressions is that a given language, voice or sensitivity is characteristic of male or female persons, not that a language, a voice or sensitivity themselves have a sex. Here it seems that the fact the substantive is ‘neuter’ helps

110

Language in the Mind

the listener determine in what sense to take the adjective. On the other hand, the fact that the substantive is ‘animate’ in male/female driver appears to condition our understanding of the adjective so that the message we get is that the driver is a being with sex. So far so good, since this helps confirm the view of gender suggested above. However, when we characterize some animals in terms of their sex – female mosquito, male lobster – and even plants, do we actually think such substantives as ‘masculine’ or ‘feminine’? Certainly the impression of a binary opposition does not arise whenever the ordinary layman calls to mind their lexeme as it does with substantives designating persons. On the other hand, it may well be that informed laymen, or experts doing research on some animal, become aware of sex distinctions to the point where they think the lexeme with the gender opposition corresponding to their experience. If further observation proves this to be the case, it would be another example of cognoscenti usage. Words like printer and sorter raise another problem. As ‘animates’ they represent persons, either male or female, who carry out a certain function. As ‘inanimates’ they represent machines that carry out the same sort of function. Should they be considered different words, homonyms, because they denote entities of a different nature, a person and a machine? If so, printer ‘machine’ would be an ‘inanimate,’ whereas printer ‘person’ would be an ‘animate.’ On the other hand, homonyms are words with quite distinct meanings, and here they have at least one important part of their meaning in common. Should they then be considered the same word actualized in different ways? If so, it would suggest that gender is similar to number, that a particular gender is not fully predetermined for a notion in tongue but is rather the result of the way the lexeme is actualized for use in discourse. Conversely, the fact that most substantives always arise in usage with the same gender would indicate that their lexeme does not offer the possibility of being actualized differently. A solution to this problem may well depend on our acquiring a clearer view of how lexemes are organized in tongue, but in any case it shows how closely grammatical gender is tied to lexical makeup. As can be seen from this brief discussion of distinctions of meaning expressed, our knowledge of gender in the substantive is limited. As yet we can do little more than tentatively recognize its existence and assume that, like number, it constitutes a subsystem of the

The Substantive

111

substantive. How the system is organized is, for the moment, not known. It does however appear that in its functioning, in forming a notion as ‘inanimate’ or as ‘animate,’ as ‘masculine’ or ‘feminine’ or ‘either,’ the system is conditioned more by the particular nature of the lexeme than is the system of number. More light will be thrown on the system when we examine it in the light of the personal pronouns, where the conditioning influence of a particular lexical matter does not arise.3 CASE

According to a number of scholars, the -’s in expressions like the girl’s book is the last remnant of the inflections for case in the Old English substantive, and so it has often been considered in grammars to be a “genitive” or “possessive” case in contemporary English. This would make the -’s a grammatical inflection like the -s of number and lead to the assumption that there is a system of case in the substantive wherein the genitive girl’s is opposed to the “common” case girl. Such an interpretation raises a problem, however, because the -’s is not related to the substantive in the same way as other grammatical morphology. Thus in the president of the university’s car, the -’s is at the end of the noun phrase and so can hardly be considered an inflection of the substantive designating the possessor. As Quirk et al. (328) point out: “it is necessary to revise the idea … that the genitive is a noun inflection. The -s ending is not a case ending in the sense which applies to languages such as Latin, Russian, and German.” That is to say, the fact that -’s always arises at the end of the noun phrase indicates that, unlike the -s inflection for number, it is not the sign of a morpheme that is part of the substantive’s morphogenesis. Just what its status is, however, has not yet been worked out. We do occasionally find other morphemes at the end of noun phrases – for example, mother and fatherless – and this suggests that since Old English it may well have developed into an element of lexical morphology. Whatever it may be, -’s does not appear to be an element of grammatical morphology today. Thus there is no visible morphology of case within the substantive. Is this sufficient evidence to permit us to dismiss case as a subsystem of the substantive? We have seen that a morpheme like ‘continuate’ number needs no sign to signal its presence in the substantive.

112

Language in the Mind

With number, however, it is an either/or situation: the substantive must be formed as either ‘continuate’ or ‘discontinuate’ in number, so -ø morpheme has a meaning insofar as it is systematically opposed to -s. We have also seen that there is no overt sign for gender in most substantives because gender is a direct consequence of the lexeme, of which the root of the word is the sign. What provided evidence that gender is a phase in the morphogenesis of the substantive was a general opposition between substantives on the level of meaning – the opposition between lexemes that emerge into consciousness with the possibility of ‘masculine’ vs ‘feminine’ categorization and lexemes that do not offer this possibility. The lack of inflections would therefore hardly be a sufficient argument for dismissing case as part of the grammatical makeup of the substantive. The problem is complicated by the fact that case differs considerably from gender and number, as can be seen in languages where inflections clearly indicate that case is a phase of the substantive’s morphogenesis. We have seen that both gender and number are based on impressions arising from the intended message – either the nature of the entity or its momentary spatial manifestation in the speaker’s experience. Case on the other hand is concerned primarily with function in the sentence, with providing the means for the substantive to fulfill its syntactic role of subject (nominative), direct object (accusative), etc. In English, similar syntactic roles arise of course, but they are expressed mostly thanks to word order and prepositions. Furthermore, it is the noun phrase, not just the substantive, that is the unit carrying out these functions, so that to explore them more fully it is necessary to get beyond the system of the substantive to see how it relates to other words in the phrase, an area we will explore in chapter 13, once we have a first view of the makeup of the substantive. Insofar as the substantive itself is concerned, however, it remains that we do speak of the functional or syntactic unit as the “noun” phrase and consider the substantive its “head,” at least lexically. Moreover, when the substantive stands alone, as is often the case of a proper noun, it is capable of assuming any role the noun phrase can, and in this it resembles the personal pronoun, which is inflected for case. Observations like these are suggestive, but the strongest argument for proposing case as a morpheme of the substantive comes from our basic postulate: behind any actual function or behaviour lies

The Substantive

113

the potential to realize it since “It is the state acquired in tongue which predetermines uses” (Guillaume 1997, 66). That is, the substantive’s actually assuming different roles in the sentence is an observable reality. We can only assume that the potential permitting this reality is somehow built into the substantive’s grammatical makeup because otherwise the substantive would not be able to exercise its different functions in the sentence. Hence Guillaume proposes case as one of the phases of the substantive’s morphogenesis in Modern French. If we adopt the same point of view, this can be proposed for English as well but just how this subsystem is instituted in tongue is yet to be discerned clearly. Certainly Modern English is quite different from Old English, with the latter’s four cases forming a system in tongue. An important development of Guillaume’s ideas on case in such languages is proposed by Valin (1994, 383–94) when he makes the point that all the oblique cases function by carrying the meaning of the substantive over to the meaning of something else. That is to say, a substantive in the accusative, the genitive, the dative, etc. characterizes something else (the verb, another substantive, etc.) in the sentence by applying its meaning to that of the other element. In the nominative, on the other hand, the substantive usually serves as a support to receive meaning brought in – imported – by something else in the sentence, the predicate.4 Thus even where there are a number of different cases instituted in tongue, each with its own sentence functions, these appear to be of two types: characterizing something else and supporting a characterization. As in a number of Indo-European languages, the historical development of English has resulted in the generalization and amalgamation of individual cases to the point where there now appears to be only one case in tongue enabling the substantive to help fulfill the two types of sentence function. Guillaume named these two functions import and support in order to have sufficiently general terms to cover all the variations possible when one meaning is related to another in a sentence. These terms remind us of the whole purpose of meaning in discourse: to say something (import) about something (support), to characterize something. Guillaume thus extends this notion of the role of meaning to all “semantic words” when he says “every lexeme is an import of meaning which must find a support” (1990, 122). It even extends to the sentence at the moment when the predicate is applied to the subject as an

114

Language in the Mind

import of meaning to its support, thus constituting the sentence meaning, the “complete thought” a sentence expresses. We shall see that the operativity inherent in these two functions suggests a basis for analyzing all the syntactic relationships permitted by the parts of speech, but before exploring this field we must complete our brief survey of the substantive. Such considerations help bring out the idea that the substantive can, through the noun phrase, provide for two different sentence functions, that of import and that of support. Being the potential for the substantive’s functions in the sentence, the disposition of the system of case in tongue must provide for these two possibilities and it must provide for them in such a way that the noun phrase itself can assume the syntactic role assigned to it. These glimpses of case clearly indicate how it differs from gender and number in being turned toward use in the sentence under construction rather than toward representing the intended message, and yet it resembles them in providing the possibility of two actualizations for the substantive’s meaning – as either a support or an import. SUBSTANTIVE AND LEXICAL MATTER

The grammatical subsystems constituting the substantive’s morphogenesis give form to – categorize – the lexical matter resulting from its ideogenesis. We shall now examine briefly the place of this lexeme, lexical content, or dictionary meaning,5 which is combined with its grammatical content or meaning in the process of psychogenesis 6 to constitute the mental side of a word. Although Guillaume did not explore the field of lexical semantics, he did establish as a requirement for a word the import of both a lexical and a grammatical meaning, and so we can at least discern the relation between matter and form within this complex. It is traditional in grammars to distinguish proper nouns from common nouns, and indeed various syntactic facts – for example, on their own Shakespeare or Canada, like pronouns, function as a noun phrase – suggest that we have here a specific type of substantive. The interesting thing for us is that such proper nouns provide a clear and unambiguous reference to the extralinguistic reality they represent. It seems that their meaning consists in representing not just ‘a poet’ or ‘a country’ but a particular poet or a particular country, not just a member of a species or category, but an individual.

The Substantive

115

And because it provides a representation of enough particularities to identify an individual – its lexical matter is particularized to the point of being qualitatively singular, unrepeatable, and so is not found as such in the plural – a proper noun characterizes a specific individual without any qualifiers or determiners. The difference between this sort of syntactic behaviour and that of common nouns is striking. When one and the same substantive, say horse, can be used to designate different individuals, it is said to be a common noun. The term “common” is to be understood as applying not merely to the physical sign but primarily to the mental significate, to the lexeme, which is common to a whole series of individuals: the meaning is sufficiently general to represent and refer to any number of individuals without differentiating between them. It is in effect a concept, the linguistic characterization of a category, the representation of an abstraction (something based on, but not found as such in, what we perceive of external reality). Not that in common-noun usage the substantive can designate only the category; on the contrary it is more frequently used to represent and refer to an individual or a set of individuals as belonging to the category. In fact, as a common noun, a substantive like horse has the potential of designating anything from the category itself to a single individual, including some or all the individuals in the category as well as one or more sets (subcategories) within the category. For example in The horse is a mammal the substantive horse designates the category, i.e. the species, whereas in The percheron is a horse for heavy work it designates one type or breed. In Your horse needs exercise, on the other hand, the speaker has an individual animal in mind, in These horses need exercise a number of individuals, and in Horses are mammals all the individuals in the category. It is this remarkable representational/referential range made possible by their lexical matter, this designating potential inherent in common nouns, that opposes them to proper nouns.7 With proper nouns, the designating potential is reduced to a minimum, a single individual. Notwithstanding this difference, the lexical matter in both types of substantive plays, in one respect, a similar role: it particularizes a substantive in discourse, distinguishing it from every other substantive, even those that have an identical sign (homonyms).8 The particularizing effect of lexical matter is quite apparent in a proper noun, with its single-individual characterization. Although the lexical

116

Language in the Mind

matter of common nouns is more general, it too exercises the same particularizing effect on the word. No matter how general or how particular the concept a substantive may import into a sentence, it serves to distinguish that substantive from others since no other word can express exactly that.9 This singularity of a word’s lexical matter has a consequence on the physical side: because each lexeme is different, the tendency is to provide it with a distinctive sign. It yet remains to analyze satisfactorily the lexical matter of a substantive and its place in the “idea-universe,” i.e. its relation with the other lexemes that constitute a mental universe confronting the external universe of our experience. Until this is described in terms of a meaning potential, however, we will not be able to see clearly how a speaker gets from, say, the experience of something small, round, and red to its linguistic representation as expressed by apple in a given sentence. Concerning this initial phase of lexigenesis where lexeme confronts extralinguistic experience, therefore, we can only take for granted that an act of representation is accomplished since we know that ideogenesis always produces a particular lexical matter. We will now examine how this lexical import is formed in the second phase, morphogenesis. IDEOGENESIS + MORPHOGENESIS

We have seen that the subsystem of gender categorizes a lexeme on the basis of characteristics inherent in it, characteristics corresponding to something in the very nature of the experiential entity denoted by the lexeme. It is much more closely bound to the lexeme than either number or case, so closely in fact that for a given lexeme variation between ‘animate’ and ‘inanimate’ does not seem possible. This link suggests that gender is the first subsystem made operational in morphogenesis once the lexeme has been actualized. On the basis of characteristics in the particular lexeme, gender categorizes it, giving it a general form – ‘animate’ or ‘inanimate’ – that makes it comparable to any other lexeme. Number is far less lexeme-bound than gender judging by the fact that it is rare to find a substantive that cannot be actualized as either ‘continuate’ or ‘discontinuate’ according to the varying needs of discourse. This greater independence of grammatical number with regard to the lexical matter suggests that the system of number arises later than the system of gender in the process of

The Substantive

117

morphogenesis. On the other hand, the fact that particular lexemes do condition the working of the system of number – for example a predisposition for a ‘unit’ rather than ‘continuum’ representation, or vice versa – shows that number is not completely free from the influence of the substantive’s lexical matter and, through it, of the experiential entity. As in the case of gender, the system of number categorizes, that is, gives a general mental form to, impressions originating in the speaker’s experience. Case is different from both gender and number in this respect since no substantive in English is limited to either a support role or an import role because of its lexeme.10 This is understandable in view of the fact that case is a matter of sentence function and so is determined not by some characteristic of the lexeme representing something in the speaker’s experience but by the role foreseen for the substantive in the purposed sentence. This suggests that case arises later in morphogenesis than number, and in fact brings this categorizing process to its final point where it results in the most general category possible in a word, the part of speech, a form that concludes the process of psychogenesis, of meaning-formation, integrating all that has arisen within it. When morphogenesis is seen as a process of forming the lexical matter through successive categorizations moving from the particular lexeme to the part of speech, we can see why Guillaume considered it to be a generalizing process for languages where the part of speech is incorporated into the word. In fact it led him to view the word-constructing process as a two-phase operation consisting of a particularizing movement, ideogenesis, followed by a generalizing movement, morphogenesis. In the first phase, ideogenesis starts with the universe of ideas and ends when the particular lexeme required, distinct from every other lexeme, comes into focus. The second phase takes this result and forms it through successive categorizations until the ultimate generalization is reached in the part of speech. Figure 7.1 (cf. Guillaume 1984, 110, 126) suggests this view; the diagram depicts the constructing of every substantive in English as the speaker, calling it to consciousness, prepares to express it during the act of language: first discriminating a lexical matter and then forming it by successive generalizations that carry the word to its final form, the part of speech. Guillaume’s dynamic view of language is evident here, with its two-phase – lexical then grammatical – sequence, and within the

part of speech case

number

morphogenesis gender

universe of ideas

ideogenesis

lexeme

Language in the Mind

118

Figure 7.1

grammatical the sequence of subsystems, each categorizing all that has preceded it. That is, even a system like that of the substantive is dynamic in the sense that it provides the potential for an operation in which the different items – lexeme and morphemes – arise sequentially, one after the other in the microtime involved. Notwithstanding particular differences from one language to another, Guillaume saw in this binary operation the basis of the general word-constructing system in Indo-European languages. Furthermore, the similarity between this diagram and that proposed for the system of number brings out the fact that it is the same mental mechanism at work in both, the radical binary tensor. In fact, as we shall see later, Guillaume sought at the basis of all grammatical systems this binary process of relating general and particular, universal and singular, nonfinite and finite, etc. as a psychomechanism reflecting an operational potential inherent in the human mind. The above sketch outlining how Guillaume conceived the substantive leaves much unsaid and should not be considered an attempt to justify his views, which would be a far lengthier undertaking. It nevertheless brings out how important he considered a knowledge of the makeup of each type of word as a prior condition for examining syntax. Indeed, some linguists whose first concern is syntactic analysis have found his preoccupation with analyzing the word to be excessive. For them gender, number, and case are of secondary importance in dealing with the substantive’s role in the sentence. This attitude however neglects the fact that a knowledge of these subsystems is necessary to understand what a substantive is as a grammatical form, just as an understanding of the

The Substantive

119

substantive’s makeup is necessary to explain how it functions in the noun phrase. The fact that morphology conditions syntax (cf. chapter 3) means that a syntactic analysis that ignores morphology cannot give satisfactory results. Guillaume’s position in this matter sets him off from most other linguists. Cursory though it is, the above description of the substantive does show how Guillaume’s presuppositions of potentiality, system, and operativity provide guidelines for observation and a framework for further research to test the validity of what he proposes. Moreover, by bringing out the internal coherence of the substantive as a system of subsystems this outline permits us to proceed to more general considerations in the next chapter, where we shall examine how Guillaume viewed the substantive in its external coherence as a part of speech, in its relations with other parts of speech.

8 The Substantive and the System of the Parts of Speech Languages like Sanskrit … already weave into the unity of the word its relations to the sentence. Humboldt (128)

NOUN: SUBSTANTIVE AND ADJECTIVE

I have been using the term “substantive” rather than the more current term “noun.” The reason for this will lead us into a discussion of the parts of speech and to a first understanding of Guillaume’s conception of their operational system. From classical times on, the distinction between verb and noun has been that between words cum tempore, with time or tense, and words sine tempore, without time or tense. Words inflected for case but not for tense were nouns, but within this category certain words, substantives, had a fixed gender and could stand alone in the sentence whereas others, adjectives, varied in gender and could not stand alone (cf. Michael, 89). The advantage of keeping the term “noun” for the more general category is to remind us that substantive and adjective have something in common, but to grasp this clearly we must first see why they differ in these two ways, and particularly why a substantive noun can stand alone whereas an adjective noun cannot. HOW AN ADJECTIVE NOUN MODIFIES

We have already seen that the gender of a substantive noun is imposed by its lexical matter, and as a consequence it does not vary.

The Parts of Speech

121

On the other hand, in languages where the adjective noun is inflected for gender the fact that the gender of the adjective varies according to the gender of the substantive clearly shows that its lexeme does not determine its gender. In fact it indicates that the adjective is said of the substantive, as the etymology (< L. adjectus past participle of adicere, ‘added to’) suggests. Even today teaching grammars define adjectives in essentially the same way: “a word qualifying a noun or pronoun,” “a word used with a noun or other substantive as a modifier to describe or define it.” Like many everyday definitions, these leave so much implicit that one is tempted to disregard them. They do, however, depict something resulting from using an adjective, something in the reality of syntax, and so, in order to get a clearer view of what they are getting at, I propose to unravel what is implied here since taking the descriptions at their face value would not make sense. Thus in an ordinary noun phrase like a big boy the adjective characterizes not the substantive itself but its meaning, but even this fails to bring out how the adjective fulfills its role. To make this explicit, we must be more precise and point out that the relation between adjective and substantive is far more intimate, that it is the very way we think the particular lexeme ‘boy’ that is qualified or modified by the particular lexeme ‘big.’ That is, the adjective’s import of meaning, the notion it brings to the sentence, is carried over and added to the substantive’s import of meaning, laid on it as on a support, resulting in an amalgamation of the two into one sense, a fusion that could be suggested by writing ‘bigboy.’ This carrying over or transporting of an import of meaning to a support of meaning is an example of what Guillaume calls incidence, a process arising in any sentence. He introduced this term, which comes from the Latin incidens (= ‘falling upon’), to denote the process involved in bringing the meaning of one word or expression into relation with that of another. Other terms such as “modify” or “describe,” as we have just seen, characterize the result of making an adjective incident to a substantive and so leave the operational implicit. Further, the term “incidence” has the advantage of being sufficiently general to be applicable to all cases of syntax, to all relating of one meaning to another. That is, this term “refers to the movement, absolutely general in language, whereby, always and everywhere, there is an

122

Language in the Mind

import of meaning, and referring of the import to a support. The relation import/support is covered by the mechanism of incidence” (Guillaume 1971, 137). The very fact that the process is so general led Guillaume to distinguish different kinds of incidence. In the case of the adjective, its lexical import “falls upon” a support in another word, the lexical import of the substantive, so he characterizes the adjective as a word with external incidence. Since this is the function any adjective carries out in discourse, from the point of view of its grammatical potential in tongue the adjective can be defined as a word capable of making its own lexical import incident to that of a substantive. HOW A SUBSTANTIVE NOUN NAMES

How about the substantive? Again popular definitions like “the name of a person, place, or thing,” “a name of anything” leave much unsaid. The problem is to grasp the sense in which name is used here. If we ask someone their name, they might answer “Mary” or “James.” The point is that the above definitions are adequate in a superficial way for proper nouns since, as we have seen, a proper noun by itself, without a determiner, can name that to which it refers outside language through referential incidence. On the other hand, if you ask someone their name they would never reply “girl” or “boy.” Does this mean that a common noun does not name anything? It certainly names something but not in the way just seen for proper nouns. It might perhaps be maintained that a common noun names something, not in the sense of designating something outside of language, but rather in the sense that it calls to mind its lexeme. In this sense, however, all words name something – a sign always brings to mind its meaning – so this cannot be the distinguishing characteristic of the substantive. The popular description, which has after all stood the test of time, suggests that a substantive names things in its own way, and to see this we must examine a relationship inherent in the lexeme of all substantives, that between comprehension and extension. In the last chapter, our discussion of proper and common nouns brought out the fact that in tongue the lexeme of a particular substantive provides the potential for representing some entity or entities in the intended message but not others. Its lexeme consists

The Parts of Speech

123

not only of a lexical matter or concept – traditionally called its comprehension – for representing a certain complex of impressions arising from one’s live experience of the moment, but also of the concept’s range of applicability, called its extension in logic. “Extension” is an appropriate term because it calls to mind the range or field of possible application over which the designating capacity of a substantive’s lexeme can be extended or exercised. Furthermore it is important to keep in mind the notion of ‘extension’ because this is a permanent characteristic of any substantive’s lexical matter. However in borrowing the term from logicians we give it an operational sense. That is, a word’s extension must not be thought of as something static but rather as a potential for actualizing the particular extent or scope required by the intended message. As a consequence, in speaking of the extension of, say, horse we think of its unvarying range for representing, a potential that may be variously actualized in discourse so that whether we use horse to represent a single individual, a type of horse or the species itself, this referent is always characterized as having the nature of ‘horse,’ as belonging to the category and so as part of the word’s extension.1 Thus thanks to its comprehension a substantive always says what something is2 by representing it somewhere within the extension. To designate that portion of the extension thus actualized in any given use Guillaume introduces the term extensity: “Extensity is a variable of discourse; extension, imposed by the comprehension, is a constant of tongue” (1982, 155). We can therefore say that a substantive names by actualizing its comprehension to represent the nature of something in the intended message and then applying this representation to its extensity, that portion of its extension that must be actualized to coincide with what is in the message. In terms of incidence, one could say that a substantive names by making its actualized comprehension or lexeme – its import – incident to its extensity, that part of its own range corresponding to the message the speaker has in mind. Thus the reason the traditional definition of a substantive noun leaves so much unsaid is that the notion of ‘extensity’ is generally neglected in grammars. This is quite understandable in the case of proper nouns used as such because extension and extensity are always equivalent for them. That is to say, because its extension in tongue is minimal, limited to one individual, a proper noun always

124

Language in the Mind

has the same extensity in discourse: one individual. This is why it normally requires no adjectives or determiners. On the other hand, because its range of representability is not minimized in tongue, a common noun in discourse varies in extensity from maximum to minimum portions of the extension, from expressing a ‘universal’ or ‘species’ sense to expressing an ‘individual’ or ‘least possible’ sense. In a later chapter this realization of how much the extensity of a substantive can vary will help us understand the need for constructing a noun phrase: to characterize sufficiently that stretch of the lexeme’s extension that has been actualized for a particular use. The point here, however, is to grasp a relation within a substantive’s lexeme, that between its lexical matter and its range of applicability, between its comprehension and its extension, because this will help bring out the nature of the substantive as a part of speech. INCIDENCE: INTERNAL AND EXTERNAL

Observation of usage has thus led grammarians and logicians to see that in usage a substantive has a twofold import of meaning arising from its comprehension and its extension. That is, within the substantive itself there is a representation both of what the substantive says (all words import some notion) and of the nature of what the substantive applies its notion to (all meaning is about something), so that the application of what is said to what it is said about is made possible within the substantive itself. In more technical terms, for a substantive the import of meaning and the support for this import are both word-internal, represented inside the word. This is what led Guillaume to define the substantive as a word providing the conditions for internal incidence. “The substantive … has its incidence in the very field of what it signifies; that is to say, the support it is to characterize is, insofar as its nature is concerned, declared in advance by its import” (1971, 137). A lexeme formed as a substantive declares the nature of its own support and so can stand on its own. An example will help to make this more concrete. Calling to mind the lexeme of a particular substantive – say ‘dog’ – determines its range of possible reference, its extension: anything we perceive as canine by nature, to the exclusion of anything feline, bovine, chevaline, or whatever other conceivable nature. The

The Parts of Speech

125

intended message determines the actual extent of reference the notion is to be made incident to within the extension of ‘dog’ – whether it be the whole species, an individual, or an intermediate extensity that is being talked about. This is quite different from an adjective. Calling to mind a lexeme such as ‘big’ does not delimit a field of possible referents because the notion does not name the nature of what is being talked about, of its support. We can see here what opposes the substantive to the adjective: an adjective is made incident to “a field not delimited by its semantic import” and so requires an outside support for its import. The support intended for the adjective “is not a support whose nature is declared in advance” (Guillaume 1971, 137) by the adjective’s own import. As a consequence, the adjective cannot stand on its own. As we have just seen, it is for this reason that Guillaume calls the adjective a word providing the conditions for external incidence. That is to say, he analyzes parts of speech not in terms of the roles they play in the sentence but in terms of the conditions that permit them to play these roles, in terms of providing a word, a unit of potentiality, with what it requires to operate as the speaker intends it to operate in the sentence being constructed. Thanks to this notion of internal incidence we can perhaps understand in a less superficial way the popular description of a substantive. By bringing to the sentence both its notion-import and the nature of its extensity-support, the substantive categorizes what it represents, tells what it is (in the eyes of the speaker, of course), and in this sense “names” what it designates. This is equally true of the proper noun, which “names” not only by designating a unique entity but also by telling what it is (a country, a city, a person, etc.). Thus we can see how the substantive differs from the adjective, whose notional import seeks, through external incidence, a support in a substantive to “qualify, modify, describe,” etc. Furthermore, it should be kept in mind that, as a discriminant between the two, the type of incidence – external or internal – is not a static characteristic but rather a potential making possible an operation of incidence, an operation that may be triggered early or late. Thus, for example, Guillaume explains the position of the adjective in French – before or after the substantive – in terms of the instant when the incidence of the adjective is effected: before the process of forming the substantive is complete or after. Similarly, Hewson (1972, 58) has suggested that the diminished lexical

126

Language in the Mind

impact of the adjectives in an expression like Poor old John!, where John need be neither poor nor old, is a result of the early intervention of the adjective in the process of substantivation. If we view the incidence of a part of speech in this way, as a sort of syntactic program for the notional import, we can understand how a word like rich can, through external incidence, characterize an indefinite variety of entities – a man, a country, a food, etc. In a phrase like the rich, however, it has caused grammarians considerable difficulty (cf. Michael, 90) because it functions as a substantive. From the point of view of incidence, we can see that besides expressing the idea of ‘affluent,’ rich calls to mind that which is affluent, the set of people being talked about. That is, in this use the notion is understood as naming the nature of what is being talked about and so it delimits its own field of application. In technical terms, the lexeme, having been substantivized, characterizes the support of its own import and so has internal incidence. Much remains to be explored in this area of usage but it would carry us too far afield to follow up any of the particular questions that arise. Rather, it remains for us to determine what the adjective and the substantive have in common, what justifies grouping them both in the category of noun. THE NOUN: A “SPACE WORD”

This brings us back to the original opposition in the parts of speech, that between the “time word” and the “non-time word.” Based on the traditional distinction between verbs, which are inflected for tense, and nouns, which are not, this opposition was repeatedly explored by Guillaume in order to bring to light more and more of its implications. His first breakthrough into the domain of deep mind came with his Temps et verbe in 1929, an analysis of how the system of the verb provides a temporal form for its lexeme, showing in detail how, as grammarians had long proposed, tense situates an event in time. We shall examine this in some detail later. In his later research, he was equally concerned with the grammatical makeup of the noun, attempting to discern how gender, number, and case help structure a “non-time word,” a word sine tempore. At the very beginning of his teaching, in 1938,3 he points out that a noun forms its notion as something occupying space, a verb as something occupying time (cf. Guillaume 1992, 3).

The Parts of Speech

127

To my knowledge, this is the first time the noun vs verb opposition has been analyzed in terms of such general categories of experience. He goes on to relate each of the parts of speech to this basic opposition, thereby establishing it as the basis of the structure of words in French and the other Indo-European languages. Guillaume is careful to point out that, although space and time are part of everyone’s experience, not all languages have instituted the opposition between them in tongue as a formal basis for noun and verb, and the other parts of speech. As a consequence not all languages have parts of speech within the word. They do not all “like Sanskrit [and the other Indo-European languages] already weave into the unity of the word its relations to the sentence,” as Humboldt puts it. It is important to keep this point in mind – that the part of speech, i.e. the potential for a word’s syntactic relations, is “woven into” the word, is consignified by it, as medieval grammarians would say – in order to understand why Guillaume refused to consider that noun and verb are “universal grammatical categories,” a question discussed in Duffley 1992b. Langacker remarks (1991a, 60) that “If the schematic characterizations [of noun and verb] to be proposed are essentially correct, they are no doubt universal rather than language specific,” but considers that the question of universality “is really a nonissue,” presumably because he regards noun and verb as functional classes expressing elements common to human experience. Guillaume of course recognized the fact that all speakers, whatever their language, have the functional means of expressing their experience of time and space. To consider noun and verb simply as functional classes in discourse is however quite different from considering them as parts of speech,4 i.e. grammatical categories of representation instituted in tongue as basic parts of the word-forming system. Consequently he often pointed out that verb and noun are not found in all languages as parts of speech. Guillaume’s insight concerning the noun as a “space-word” arose largely as a result of his reflecting on the representation of grammatical number, which, with its ‘continuate’ vs ‘discontinuate’ opposition clearly provides a categorization of the substantive’s notion in terms of the manner its referent is perceived to occupy space. Gender as determined by the notion (i.e. “real” gender, not the “fictive” gender of French) also has its roots in a spatial impression

128

Language in the Mind

in Guillaume’s eyes since it appears to be based on the fact that certain lexemes involve a duality resulting in a binary opposition within their extension, whereas other lexemes do not. This distinction on the level of the lexeme’s spatial disposition provides the basis for the system of gender. It can be seen that by categorizing the notion resulting from ideogenesis, both gender and number help give it a spatial form. Case, on the other hand, provides what is necessary for the syntactic role of the substantive – more precisely, the noun phrase – in the sentence and so helps determine its place in the sentence being constructed. From this we can begin to see how the substantive’s lexeme is formed by each of these subsystems, particularly by the system of number, and is ultimately represented as a spatial entity in the speaker’s mental universe. How about the adjective? In languages where it is declined for number and gender, and even case, it is clear that its lexeme is formed along the same grammatical lines as the substantive’s, except that, insofar as gender is concerned, the notion of the adjective cannot predetermine a particular gender, as the substantive’s notion does; the adjective requires gender as a form, but leaves the particular gender to be determined by its support, the substantive. Moreover, the fact that the adjective agrees with the substantive in this way indicates that its lexeme must find its support in the place defined by the substantive to which it is incident. Thus it requires the possibility of being made incident to a lexeme whose internal incidence has defined its own place in space. In English, where there is no agreement of adjective with substantive, it is generally assumed that these subsystems are no longer part of the makeup of the adjective. Before accepting this view, however, certain facts of usage must be taken into account, such as the use mentioned above, the rich, which expresses ‘plural’ but which cannot be used as a singular: *a rich. Furthermore, it is well known that certain adjectives like asleep are restricted in the syntactic roles they can assume. Whether or not such details indicate that number and case do arise in the adjective’s morphogenesis remains to be seen. What is clear, however, is that the adjective’s lexeme is formed in such a way as to be applicable to that of a substantive. Where a lexeme is not provided with the grammatical form required for this, incidence to a substantive’s lexeme is not possible: for example, expensive clothes but not *expensively clothes. And this brings us to the adverb.

The Parts of Speech

129

THE ADVERB AND THE PREDICATIVE PARTS OF SPEECH

An adverb is “used to qualify a verb, an adjective, or another adverb,” according to the traditional description, which we can now understand to mean that the lexical import of the adverb is carried over and applied to that of another word. The adverb can of course assume other roles in the sentence, so that Quirk et al. (438) consider it “the most nebulous and puzzling of the traditional word classes,” but the point I want to make here is that it cannot “qualify” a substantive. This led Guillaume to view the adverb as a part of speech that, like the adjective, forms its lexeme for external incidence, but, even further removed from the substantive, external to a second degree. As a consequence, unlike the adjective, the adverb cannot be incident to a notion with internal incidence. For Guillaume, then, the notion imported by the adverb is applied to another import, itself in the process of being made incident to its support. As Guillaume sometimes put it (cf. 1982, 130), an adverb is incident to an operation of incidence in progress, by which he means incident to a support that is itself involved in a process of external incidence. Thus in a very bad cold, for example, very is made incident to bad as the adjective is being transported to its support, so that it is the complex adjectival notion ‘verybad’ that becomes incident to ‘cold.’ Although it has yet to be shown clearly how the morphogenesis of the adverb differs from that of the adjective, that it is different seems apparent in view of the difference of syntactic function,5 a conclusion borne out by other languages where the adverb is never declined as is the adjective.6 Considering the adverb in this fashion, as a part of speech forming a lexical matter in such a way that it becomes incident to another notion, itself seeking an external support, helps bring out the essentially operational nature of syntax. Syntax consists in relating the meaning of one word or phrase to that of another, applying an import to a support. “Every lexeme is an import of meaning which must find a support” as Guillaume puts it (1990, 122), because the whole purpose of a lexeme, of a word with full-fledged lexical meaning, is to say something about something. And it is the part of speech – substantive, verb, adjective, or adverb – that gives the lexical matter the potential for realizing its incidence to a support.

130

Language in the Mind

The system of these parts of speech provides for all possible relations of incidence involving words whose notional ideation results in a lexeme, a lexical meaning that predicates something of what is being talked about. Either a word represents the nature of its own support and so can realize its incidence internally (the substantive) or it does not. In the latter case, the word must effect its incidence to a notion in another word: either to a substantive’s notion by external incidence of the first degree (adjective, finite verb) or, by external incidence of the second degree (adverb), to a notion itself seeking an outside support. By making the adverb incident to an incidence in progress, this analysis helps explain its remarkable syntactic versatility. Guillaume called these the predicative parts of speech precisely because their type of incidence allows for notionally characterizing something in the speaker’s intended message. The system of the predicative parts of speech is complete in the sense that it permits all the relations of incidence between notional imports required in constructing a sentence. On the other hand, the system has a second phase giving rise to parts of speech that form grammatical words focusing on the operations involved in constructing the sentence. THE TRANSPREDICATIVE PARTS OF SPEECH

In its second phase, the system gives rise to the nonpredicative, or better, transpredicative parts of speech, a term intended to suggest that grammatical or function words arise in a second phase of the system, at a point beyond that of the predicative parts of speech. Guillaume considered that these parts of speech form closed systems (he described in some detail the system of the articles in French) in the sense that one cannot make up or learn a new pronoun or auxiliary in English as one can a new substantive or adjective. He threw some light on the auxiliaries and the personal pronouns, but he did not succeed in analyzing their systems nor that of the other pronouns, or the conjunctions. Likewise for the most numerous of the transpredicatives, the prepositions: he considered their system to be made up of a series of interrelated subsystems based on the lexical import of each individual preposition. This section will therefore offer no more than an attempt to situate these different parts of speech with regard to one another,

The Parts of Speech

131

but in the next chapter, one such subsystem, that of some and any, will be described. The role of the transpredicative parts of speech is to represent the conditions imposed by the language for saying something about something. That is, these grammatical words serve not to characterize a support in terms of the speaker’s experience but rather to represent the relationships that arise in a sentence between an import and a support.7 “The predicative parts of speech are those whose incidence, internal, external 1 or external 2, is tied to the event the sentence relates. Those parts of speech whose incidence is tied to the mechanism of the event constituted by the sentence itself are not predicative” (Guillaume 1982, 130). Abstract though they may be, these “function” words, as they are often called, are constructed on the same pattern we saw above for “content” words: ideogenesis + morphogenesis. As we have seen, Guillaume also designated this bi-phase constructional program “notional ideation” and “structural ideation” to make it explicit that the result of forming a word lexically and grammatically is, in both cases, something mental. He thus rejects the view, quite frequent among linguists, that some words are meaningless.8 The first such part of speech, often derived historically from the adverb,9 is the preposition, whose function is to represent and establish the relation of incidence between a noun phrase and its intended support. A preposition is called for whenever a noun phrase cannot relate directly to what it is to characterize, when the noun phrase cannot, for some reason, make its import part of the sentence. In Guillaume’s terms (1971, 154f): “The preposition can therefore be quite accurately defined as the type of word which is already predestined in tongue to intervene in discourse between two words separated by an interval which is not covered by a mechanism of incidence in operation… It is the type of word that comes into play whenever the mechanism of incidence between two parts of the sentence is left in abeyance.” For example, it would not make sense to say *a book pictures, whereas a book with pictures or a book of pictures or a book without pictures does make sense because in each case the preposition establishes the relationship of incidence between the notion of pictures and its support. That is, as a part of speech the preposition has the capacity to relate an import of meaning that already has its own internal support to a support of meaning elsewhere in the

132

Language in the Mind

sentence. It is a word whose syntactic function is to establish incidence between the two but it characterizes neither. Rather, by means of its own specific import of meaning, each preposition characterizes the particular relationship it establishes between two parts of the sentence, so that, in effect, a preposition is incident to the relationship, the gap between them, as Guillaume puts it. Because it helps characterize not what is being spoken of (though by implication it may do so) but the means of speaking of it, the preposition is a transpredicative part of speech even though, in some cases, its lexical content is quite similar to that of an adverb. The role of a conjunction is to effect the incidence between two parts of a sentence and characterize the relationship thus established. Where it is a matter of making a clause with its verb phrase incident to a support, a subordinating conjunction is called for, as in I’ll go if you will. Without if, the incidence between the two clauses would not be realized. This function is often combined with that assumed by a pronoun, giving rise to the relative pronouns, but just how the two parts of speech work together in this way remains to be clarified. Coordinating conjunctions, on the other hand, can relate words, phrases, or clauses that are on an equal syntactic footing, as in a series. Since they can link two sentence parts with the same function, they are not limited, as are subordinating conjunctions, to establishing the incidence of a clause to another part of the sentence. On the other hand, since coordinating conjunctions do not establish the incidence of one element to the other as import to support, one of the items they connect is not said of the other, does not characterize it explicitly. Auxiliaries constitute another transpredicative part of speech, the counterpart of full verbs, from which they were originally derived by a process variously designated as semantic bleaching, lexical depletion, or, as Guillaume calls it, dematerialization. This is an historical process whereby some of a word’s lexical matter is abandoned, resulting in a more abstract lexeme. This process, carried to an extreme in do, be, and have (see below), gave rise to auxiliaries with the capacity of making an event imported by another verb, an infinitive or participle, incident to its subject while not characterizing lexically either the event or the subject. As we shall see in a later chapter, representing an event by means of an infinitive or participle leaves one of its grammatical parameters, its place in time, unspecified, and it is this which calls for the use of an auxiliary.

The Parts of Speech

133

The counterpart of the noun in the transpredicative parts of speech, the pronoun, brings to the sentence as a result of its notional ideation, not a lexeme representing what is being spoken about, but rather some characteristic of it – possession (possessive pronouns), position within the speaking relationship (personal pronouns) or outside it (demonstrative pronouns), etc. Certain personal pronouns (I, he, etc.), used only as subject, are situated in time by the finite verb. Through its structural ideation, the pronoun represents the spatial form of what is being spoken of. In some cases this suffices for the pronoun to constitute a noun phrase by itself, usually playing a suppletive role with regard to a lexical import expressed elsewhere. In other cases, however, the pronoun plays a completive role in a noun phrase by providing a formal support for the lexical import represented by the rest of the phrase. Thus Guillaume (cf. 1997, 204) would consider he, mine, this (in its so-called “pronominal” uses), and many others to be pronouns whose function is one of suppleting, whereas what are often called “determiners” he considers to be completive pronouns whose role is to give a formal or grammatical representation of the substantive’s support, thus making the substantive’s incidence fully explicit. That is, in uses like this book, some problems, Guillaume does not consider this and some to be adjectives since they do not operate as a predicative part of speech, bringing an import of meaning to the lexeme of the substantive. On the contrary, the substantive carries its meaning over to the formal support provided by the determiner, as we shall see later with some and any, and particularly the articles, the most abstract of these pronouns, dematerialized to the point where they represent no more than the substantive’s extensity. THE SYSTEM OF THE PARTS OF SPEECH

This rapid evoking of the parts of speech helps suggest why Guillaume considers them to form a system in several dimensions or operational possibilities, each reflecting an opposition in the mind. Although Guillaume never gave a definitive version of his view of the system of parts of speech in French, particularly the transpredicatives, it will be useful to summarize some of the relationships involved here and suggest (see figure 8.1) the sort of system he seemed to be working toward.

Language in the Mind

134

verb finite verb

substantive adjective noun

adverb adverb

conjunction preposition

pronoun

auxiliary

pronoun

article space

space

infinitive -ing participle

time

time

PARTS OF SPEECH

predicative

transpredicative

Figure 8.1

The fundamental opposition depicted in this diagram, that arising at the outset of the system and persisting throughout it, reflects the space-time nature of our experience, and constitutes the basis of the noun vs verb dichotomy discussed above. Operationally, this provides the speaker with the possibility of giving a grammatical form to a lexical import arising from notional ideation. Categorizing the lexeme during the process of structural ideation prepares it to be situated in either the universe of space or the universe of time. The next opposition is between words that can define their own positions in either universe, and those that cannot. It gives rise to the internal incidence of substantives as opposed to the external incidence of adjectives and of adverbs seeking incidence to the universe of space through the substantive. This also gives rise to situating verbs in time, as defined internally by their tense (though the finite verb must then be made incident to its external spatial support in the subject, as we shall see in chapter 10), and the external incidence of adverbs to a temporal entity in the verb. The third opposition is that between “semantic words” (cf. Guillaume 1990, 85) characterizing something in the intended message and grammatical words characterizing the means of saying, i.e. the relation between two meaningful units (e.g. prepositions), referring to the intended message (e.g. the demonstratives), providing a support of meaning (e.g. the articles), etc. This gives rise to the predicative parts of speech and the transpredicative parts of speech, the latter reiterating, under different conditions, the basic opposition between space and time, and that between internal and external incidence.

The Parts of Speech

135

This diagram illustrates how the tenets of psychomechanics offer guidelines for exploring the most general of systems in English. For seeking the potential permitting the syntactic behaviour of each part of speech, Guillaume proposes incidence as a common capacity and distinguishes each one according to the way it implements this capacity in discourse. The operativity of the system of constructing words can be seen in its two phases – predicative (representing the particular message) and transpredicative (representing the general mechanism of the sentence), the latter necessarily following the former in the system – and in the succession of positions in each phase, and finally in the subsystems implicated in a given part of speech. This double operation reflecting the underlying binary tensor mechanism opposes the representation of space and time in the particular lexemes of the predicative parts of speech to the representation of space and time in the more abstract content of the transpredicatives. Figure 8.1 is presented here as a framework for reflection, a tentative hypothesis. As such, it does not depict the various subsystems, which are to be imagined developing within each part of speech on another dimension. This view of the parts of speech will constitute an explanatory theory only to the extent the different morphemes discernible in discourse can be situated in their place in the subsystem within the part of speech. Even as a preliminary sketch, however, it affords a glimpse of this pluridimensional structure of thought wherein the mind can intercept its own operations time and again so that we can begin to understand what Guillaume had in mind when he defined tongue as “a system of systems.” For him the formal part of tongue is so highly systematized that when the mind takes up a position within a subsystem of a part of speech, it is situated with regard to all the rest. And so, as he says, “One example – it’s always the whole of tongue that is evoked” (1997, 64). The system of the parts of speech is thus concerned throughout with the syntactic incidence of an import to a support and ultimately with making the sentence meaning incident to the intended message through referential incidence. As such the system makes possible all the relations between words in a sentence and so imposes a form on all the words we use, except interjections. Interjections enter into no relations of syntactic incidence and so can be neither an import or a support of meaning for other words. Rather, an interjection expresses a “complete thought” (if one can

136

Language in the Mind

call the impressions it evokes a thought) all on its own, and through its own expressiveness applies its import of meaning directly to the intended message through referential incidence (cf. Guillaume 1984, 86). Thus it constitutes a part of speech peripheral to – or in a way, a non-part-of-speech opposed to – the system of the parts of speech outlined above. This contrast helps bring out the fact that the system of the parts of speech is not some sort of quasi-logical pipe dream of grammarians imposed on language. Guillaume arrived at his view as a result of reflecting on correspondences in the Indo-European languages (cf. 1984, 38ff), particularly the verb and the noun, where the distinction between lexical and grammatical import is clearly indicated in the semiology. That is to say, regardless of whether his view of the system for French or English proves valid in all respects or not – only further research and reflection will tell us that – the very existence of such a mental system cannot be called into question without revoking our basic postulate concerning the potential presupposed by any observed actualization. Since every word we observe in usage in English belongs to a part of speech, there must be some potential, some system for structuring words in this way, a system making possible the structural ideation of words, thus providing their formal significate. Only when this system has been clearly discerned and described will it constitute a theory of the word as a grammatical entity in English. As such, it will provide the basis for a theory of the sentence as a grammatical entity, for a theory of syntax. HOW THE SYSTEM WORKS: “CONVERSION”

The system of the parts of speech thus provides the general systemic sequencing of morphogenesis, the second phase of word formation. The content and the structure of all words is provided by this two-phase mechanism, whereby a lexeme is first distinguished from all the other lexemes and then related to all other words through categorization as a part of speech (see figure 8.2). This system offers a basis for analyzing the mental side not only of inflected words, where the sign reflects different meaning components, but also of words where the sign gives no indication of the complexity of the significate because, as Guillaume realized, even if a language like English has lost most of its grammatical

morphogenesis

137

part of speech

ideogenesis

lexeme

universe of ideas

The Parts of Speech

Figure 8.2

inflections we cannot conclude that most words have lost their grammatical import as well. In proposing that every word involves a systematic arrangement of lexical and grammatical significates, he is making explicit what is often left implicit or ignored by other linguists, but he is proposing it as a word-forming activity carried out each time a word is used. This, I believe, is an important innovation. To bring this out, we can look at a process of word-formation called “functional shift” or “conversion” (cf. Quirk et al., 1558ff), defined as “the derivational process whereby an item is adapted or converted to a new word class without the addition of an affix.” Examples are: • • • • •

“deverbal nouns”: a want, a doubt, a bore, a throw; “de-adjectival nouns”: a daily, a regular [customer], newly marrieds; “denominal verbs”: to bottle, to gut, to parrot, to canoe; “de-adjectival verbs”: to dirty, to yellow; “denominal adjectives”: a brick garage, a cotton dress.

This approach is basically static in presupposing that, when called on for use in a sentence, “an item” (a word) already exists in the lexicon or inventory as a member of a given word class. Hence the need for the word to be converted from one class to another. This is not “a historical process, but rather … a process now available for extending the lexical resources of the language” (loc. cit.). For Guillaume there is no such storehouse of words in tongue. What is stored in tongue is, as we have seen, the whole series of lexemes functioning as viewing ideas with regard to one’s experience. When called on by the process of ideogenesis to represent that experience, the lexeme has no grammatical category, and so

138

Language in the Mind

it must be categorized by morphogenesis in a predicative part of speech at the moment of speaking. That is to say, the lexical input is processed by the grammatical input in view of the role it is to play in the sentence being constructed. This does not involve “converting” but simply forming a substantive or a verb or an adjective or an adverb out of the unformed lexical matter provided. Every word is formed anew; none are reconditioned versions of another word. One can thus understand why Guillaume, having rejected the static notion of a lexicon, says that there are no substantives or verbs or adjectives in tongue – no words – but only the formative elements and the means for forming words. Thanks to the distinction between ideogenesis and morphogenesis in the predicative parts of speech, we can account for uses like those given above: a want is not a “deverbal,” but simply a lexeme (‘want’), which although usually formed as a verb is here formed as a substantive; a daily is not a “de-adjectival” but a substantive formed from a lexeme that is normally categorized as an adjective. And so on. It remains that no lexeme is used in all parts of speech, and this appears to arise from the nature of the particular lexeme. Although ‘daily’ is found as substantive, adjective, and adverb, one has trouble imagining it being used as a verb, presumably because it cannot be actualized as designating some action or state. Similarly for ‘want,’ which is not found as an adverb. On the other hand ‘forward’ can be formed in all four predicative parts of speech. Perhaps our understanding of the lexical semantics of these words will eventually permit us to explain these constraints. Of course the needs of expression in a given situation may prompt a speaker or writer to forge a new word, as in Shakespeare’s But me no buts and more recent inventions like Small is beautiful and Is better always good? Such uses, which may surprise us at first, are clear indications of a speaker making use of the part of speech system to obtain particular expressive effects in discourse from lexical resources available to all speakers. The formative capacity of the predicative parts of speech is not limited to treating lexemes in tongue. Though they belong to no “word class,” expressions like an up-in-the-air feeling or one of these everything-turns-out-well-in-the-end films are obviously processed on the spot to produce a syntactic unit. (Analyzing such cases however is more complex since each word and its relations with the others must be described to bring out how their various imports produce

The Parts of Speech

139

the lexical matter of the expression.) Such examples show that the part of speech system, far from being a static structure, endows a lexical input with a means of establishing incidence in order to carry out its function in the sentence. For this reason, Guillaume maintains that although sentences are not found in tongue, the system of the sentence is found there; that is, it is the possibilities of incidence built into words by the system of the parts of speech that give each sentence its particular architecture. To the extent that a sentence expresses a “complete thought” by means of more than one word, operations of incidence are involved because the different meaning imports must be assembled, integrated into a coherent whole. Speakers are able to analyze any experience into different lexemes thanks to the lexical resources available in tongue, a veritable universe of ideas confronting the external universe that gives rise to our experience. Thanks to the system of the parts of speech, speakers can form these lexemes into different words and then synthesize them through incidence into a whole representing the original message. Thus the original experience of the speaker can be reconstituted into a spatiotemporal complex thanks to the fundamental oppositions provided by the parts of speech. We have yet to examine what provides the temporal dimension of a sentence, the system of the verb, and its implementation as both full verb (predicative) and auxiliary (transpredicative). Before turning to this topic, however, it will be valuable to describe one of the subsystems in the pronoun to give a clearer idea of what is involved in analyzing the transpredicatives.

9 Some and Any Grammatical symbols of tongue cover movement. Their matter is movement. The principle just recalled is new and, from a practical point of view, very important. Guillaume (1991, 68)

INTRODUCTION

Some and any are often linked in grammars, where their distribution – some in positive contexts, any in negative contexts, and both in interrogative contexts – must be described for teaching purposes. In linguistic studies they are sometimes considered as members of a system of quantifiers, or a pair of suppletives in discourse. These various approaches are based on different observations of usage in discourse, but whatever their merits these approaches have not provided an adequate description of the import of these two grammatical words. Whether limited to distinguishing their respective distributions, or describing the expressive effects they contribute to (‘assertive’ or ‘real’ for some vs ‘nonassertive’ or ‘possible’ for any), or isolating one sense they express (‘universal’ for any, ‘unspecified quantity’ for some), none of the approaches attempts to describe the precondition permitting all these observed manifestations, namely the meaning potential of each word. In this chapter we will examine selected cases of usage in order to work back to the meaning in tongue for each pronoun and then show how they are systematically related. Since our aim is to illustrate one subsystem of the system of pronouns, there will be no attempt to give a detailed demonstration of how all their different uses are derived from their underlying potential. For a fuller description of usage and for references, one can refer to my 1988 article.1

Some and Any

141

ANY

Any has been aptly described as a “many-splendoured thing.” Not the least of its splendours can be attributed to the fact that it is a quantifier. This view, accepted by most scholars, raises a question that is not always easy to answer: if any is a quantifier, what quantity does it express? The problem here is not that it expresses no quantity, but rather that it expresses a wide variety of quantities, so wide in fact that it is difficult to reconcile this polysemy with the idea that any is a single word insofar as the expression of quantity is concerned. And in fact some scholars have proposed that there is more than one any in English. Since our first concern will be to show its unity, the “thing” underlying its “many splendours,” we will focus first on its quantitative import. We can best begin by examining some clear cases where the quantity expressed is not in doubt. For example, if I want to evoke the application of a rule, I might say: Any students registering late will be charged a late fee. Here the speaker seems to have in mind something like ‘all,’ i.e. ‘whatever students register late.’ Similarly in an example to do with publishing Any changes could be made in proof. the sense is something like ‘all changes that may be required.’ Thus these two examples show that any has the possibility of expressing an all-inclusive quantity, a totality. Any has also been called a universal quantifier in uses with ‘singular’ substantives like the following because it implies no matter which policeman: Any policeman will be able to tell you. However a recent grammar interprets this as expressing “a free choice as to which policeman can be selected, but no matter which policeman it is, that policeman will be able to tell you” (Huddleston and Pullum, 382). This same “free choice” sense of any is expressed

142

Language in the Mind

in the following sentence, which might be said to help someone find the university library: Once on the campus, ask any student you meet. Here the speaker obviously intends the person addressed to pick out one student at random from among those met and ask directions. Similarly, when doing a card trick one might say: Take any card. Again it seems obvious that the speaker intends the other person to choose a single card out of the deck. The quantitative meaning of any in these three cases can therefore be paraphrased as ‘one.’ Thus it seems clear that any can express quantities as different as ‘all’ and ‘one,’ but these are not its only senses. A third sense can be discerned in the following sentence: The closest scrutiny is owed to the Anglo-Saxon kennings and the Homeric epithets; if any words and phrases are formulaic, they will be. Here the expressive effect is something like ‘if there really are certain words and phrases that are formulaic…’ Quantity-wise, any here expresses the sense of ‘a certain number of,’ i.e. neither ‘one’ nor ‘all’ but some unspecified quantity between these two. The following use gives the same reading insofar as quantity is concerned: That any people retain doubts on these questions is symptomatic of two unfortunate conditions of modern existence. The writer has in mind something like ‘that there should still be a number of persons who…’ Again one understands an intermediate quantity, neither ‘one person’ nor ‘all people’ but somewhere in between the two. These examples will suffice to recall what others have already observed: any can express not just one quantity, or several quantities, but a whole range of quantities. Webster’s Dictionary, for example, has the following entry for any: “one, some or all indiscriminately of whatever quantity.” Assuming that any is the same word in each

Some and Any

143

of our examples, and I believe most speakers of English would hesitate to declare it otherwise, then underlying these different senses there must be some principle that ensures the notional unity of the word, some meaning potential that can give rise to the expression of quantities from ‘all’ to ‘one.’ We have here a problem of polysemy similar to that posed by -s morpheme, discussed above in chapter 4, and will resort to the same postulate of operativity to reconstitute the meaning potential that will explain the observed polysemy. We postulate that the meaning of any (and any other grammatical word or morpheme with a number of observable senses) is not something static (a particular quantity, for example) but something dynamic, some mental operation making it possible for the speaker to obtain a representation of the various senses observed. That is, any always has a potential meaning of ‘quantity,’ but the size of this quantity in any given use will depend on where in the field of quantity the representation is obtained. Inherent in its meaning therefore is a movement proceeding through the field of possible quantities from the greatest possible, through intermediate possibilities, to the least possible, or the reverse – from the least to the greatest (we will see below which is the form of movement for any). Only in this way can we explain the spectrum or gradient or cline of resultant senses in discourse: the mind intercepts this movement at the point in the field corresponding best with the quantity to be represented. The significant point here is that this dynamic conception of the potential meaning as a movement between limits permits us to explain the notional unity of any while accounting for its variability in usage. SOME

Rather than give more examples of any at this point it will be useful to turn to some to see if it too can express different quantities. The following example provides one clear expression of quantity: I read it in some newspaper. Quantity-wise, this clearly expresses the sense of ‘a newspaper’ – i.e. ‘one’ – with the added suggestion of ‘I don’t remember which,’ a nuance we will return to below. Likewise in

144

Language in the Mind

Some day, many years in the future, you will remember this. some clearly has the sense of ‘one’; in fact it could be replaced by one. This sense can be contrasted with that expressed in: You had better check because some flights have been cancelled. Here we understand ‘a number of flights but not all.’ Similarly, as a suppletive pronoun in the following sentence describing the patterns on snakes Some are annular, some are reticulated, some are dotted and some are akin to spirals. some depicts ‘a certain number’ belonging to each type. The following common use brings out the same expression of intermediate quantity, ‘more than one but not all’: I used some of the eggs for an omelette. These findings for some, clearly paralleling what was found for any, encouraged us to look for the third sense of ‘all,’ especially since Webster’s has as an entry for some: “being always at least one, but often a few and sometimes all of.” However no use was found where some has the reading of ‘all’; this is why, unlike any, it is never considered a universal quantifier. On the other hand we were not able to describe satisfactorily the sense it expresses when unstressed, perhaps its most common use: I went in and found some students working on a project. Some tells us little about quantity here, but in Sahlin, one of the most detailed studies, it is described as “slightly quantifying” (15, 56) in this use. (This description can be appreciated if some is stressed here because then it would have the same sense as in , ‘a certain number but not all.’) Some scholars have considered some in this use the equivalent of an article, the plural of a. Any has a similar unstressed use, as can be seen in the following pair: Did you buy any/some apples?

Some and Any

145

In a later section we will discuss the difference between the two pronouns in such uses, the point at the moment being the problem they raise: to determine whether they function as quantifiers here, and, if so, how. A curious characteristic of this use is that the completive pronoun can be omitted without a major change in the meaning expressed by the sentence (cf. Did you buy apples?). To understand why they could be omitted in this way, we had first to discern another element of their potential meaning. This permitted us to solve the problem of their function here and eventually to see how they are related to the articles. PARTITIVES

One of the most common expressive effects of the two quantifiers is to “indicate a subset of an explicitly singled out set” (Sahlin, 38), as in: The American dream was compounded of many strains. Some were clearly of Christian origin. There was no sound or light in the entire house to indicate that any of the occupants were awake. In such cases the quantifiers express a part of the total number expressed by “many strains” and “ the occupants.” This expressive effect is found with ‘continuum’ notions as in , as well as with substantives expressing a single whole, as in : He snatched a basin of water and sprinkled some upon her face. Let us conceive Shakespeare writing Macbeth in an age of “exact history.” Hardly any of the play would be left. To include uses like the last two one could not speak in terms of “subset” and “set,” but the more general terms of “part” and “whole” are appropriate. This led us to explore the idea that some and any represent a quantity as part of a whole, that they are partitive quantifiers. In the examples just given, the partitive effect comes out clearly because both part and whole are expressed, but when some and any are used as completive pronouns as in , it was not so clear what expresses the whole to give the effect of ‘a certain portion

146

Language in the Mind

but not all.’ An example from Sahlin posed the problem even more clearly: Some women get a thrill out of housework. Here we understand the noun phrase to mean ‘a part or subset or number of, but not all, women.’ This appears to be due to the quantifier because omitting some gives a sentence with a very different meaning: Women get a thrill out of housework. Comparing these two sentences brought out another significant fact: regardless of whether the sentence is about a certain number of women , or all women , each of them suggests the notion of ‘women in general.’ In light of the theory of grammatical number outlined in chapters 4 and 5 we interpreted women in as a substantive whose lexical import had been intercepted at the end of the -s movement through ‘discontinuate’ space in the number system, giving it this ‘maximum’ or ‘generic’ reading. Example , however, is about a subset of women and so we understand the substantive’s import to be formed numerically by intercepting the -s movement somewhere before its end to give an ‘intermediate’ or ‘plural’ reading. Since the partitive sense arises with the introduction of some, it appears that the determiner depicts the quantity of the substantive’s import but represents this quantity as part of a greater quantity. That is, as a partitive quantifier, some represents the same quantity as women but implies a relationship with another quantity – ‘all women, women in general’ – by representing it as occupying a portion but not all of the greater quantity. Any in this type of noun phrase can be analyzed in the same way; it represents the quantity of its substantive as a subset or part of another quantity the speaker has in mind, ‘women in general’: Do any women get a thrill out of housework? The important point here is that the substantive’s lexical import, quantified by the system of number, brings to the sentence a certain scope – maximum, intermediate, or minimum, as we have seen. In and , this scope provides the quantity depicted by some and any as a part of the maximum whole the speaker has

Some and Any

147

in mind. This analysis also applies to examples where the speaker does not have in mind the maximum lexical scope. Thus in , flights gives a spatial representation of the number of departures that have been cancelled and some represents this quantity as part of the number the speaker has in mind – say all those from the local airport on that day. Similarly in An earlier difficulty was overcome by making it clear that individual libraries in any area might join or not, as they saw fit. Some library boards are wary of the plan. we understand that some library boards is a subset of the set the speaker has in mind, as suggested in the first sentence. Likewise in , out of the set of individuals who know about these questions, any people represents the possible existence of a certain number who retain doubts. A sentence like the following, with dominant stress on the quantifier, gives another example: Does he have any good qualities? Thanks to the partitive representation of any, the expressive effect here of ‘even a few’ suggests a contrast between what one would expect in an individual, a certain number of good qualities, and what one hopes to find in this particular person, ‘just a few.’ With ‘singular’ substantives the partitive effect is different because the substantive calls to mind not a set but a single entity. Thus in an example already seen I read it in some newspaper. some depicts the quantity involved in a single entity but depicts it as a part, as one among others, perhaps those newspapers the speaker has read recently. This is what gives the reading ‘one I don’t know which.’ Similarly for day in above: ‘one day, I don’t know which.’ On the other hand, any with a ‘singular’ substantive has quite a different expressive effect. Thus we saw in Any policeman will be able to tell you. there is the suggestion of a “free choice” among policemen (among those in the area where one asks for directions). Similarly

148

Language in the Mind

evokes one student as the future choice among those that happen to be met, and one card as the one that will be chosen among the set presented. One gets the impression that any depicts the single item or individual as one possibility among the set the speaker has in mind. This explains the common expressions in any event and in any case: in a scenario where the speaker envisages diverse possible developments any represents the outcome that eventually becomes a reality, whichever ‘event’ or ‘case’ it may be, as one among several possibilities. We will see below why, with ‘singular’ substantives, some gives rise to a ‘one I don’t know which’ expressive effect whereas any suggests a ‘free choice’ effect. These examples show that any and some are partitive quantifiers both when they represent an indefinite number of parts from the whole the speaker has in mind and when they situate a single entity expressed by a ‘singular’ substantive as one part of a whole. How about any when it represents the totality? Is it still functioning as partitive? We saw in that any depicts the totality of students registering late. Similarly for any changes in . These examples show that any as a partitive can represent all the parts in the whole, however many that may be. That is to say, we have to understand the term “partitive” not just as a single part or a number of parts of a whole but also in the sense of all the parts. Another interesting thing about these two examples is that the quantifier could be omitted without a noticeable change in the quantity expressed. This detail of syntax is quite understandable in light of our analysis. Quantity-wise, the whole is equivalent to all its parts, so since students and changes express the whole, omitting any in its ‘all-inclusive’ sense here would involve no change in quantity expressed, though the sentence may give the effect of being less explicit. This possibility of omitting the quantifier reminded us that in unstressed uses like and the quantifier can also be omitted, and so we were led to have a second look at the problem they presented. Having realized that any can be omitted when it expresses all the parts of the quantity the speaker has in mind, we examined unstressed uses for the quantity expressed by the substantive as in: Did you buy any/some apples? Assuming this is addressed to someone who has just done the week’s shopping, it is understood to evoke the quantity of apples normally bought in those circumstances. Since omitting any would

Some and Any

149

entail no change in the quantity expressed, we concluded that it represents the same quantity as that expressed by the substantive, which represents the total quantity the speaker has in mind. That is, this appears to be another case where any depicts the totality of the whole the speaker has in mind, but in a situation where the quantity is known, is taken for granted. As a consequence the expressive effect of ‘all, whatever the number’ as in and does not arise. This result of our analysis was quite unexpected because of any’s reputation of being a universal quantifier. In both its unstressed use and its ‘all inclusive’ use any represents the substantive’s quantity as equivalent to the total set, the whole, the speaker has in mind, and this is what permits expressing the same quantity without any. The difference between the unstressed use and the stressed use in and depends on whether the quantity is known or hypothetical, and this conditions the contexts in which each use is found, as we will see. How about some in this use, clearly indicated by both lack of stress and a weakened vowel? The fact that it is often found in the same contexts as unstressed any, as in , suggests that some can be analyzed in the same way. Similarly for the fact that it is often in complementary distribution with any in declarative contexts: Yes I bought some apples. No I didn’t buy any apples. Not only can both quantifiers be omitted here but they could be described as a “slightly quantifying” plural version of the singular article a. That is, some here seems “to do nothing more than express the size of the set of entities denoted by the nominal” (Milsark, 23). Furthermore, when it is stressed in many such uses the difference in quantifying the part can be observed. Thus in , unstressed some simply evokes a group of students working, but with stressed some the sentence would express something quite different: ‘a number of the students were working but not all.’ That is, stressed some would represent the quantity of the substantive as a part that is smaller than, not equivalent to, the whole set the speaker has in mind, and in this case it could not be omitted without a noticeable change in the quantity expressed. This analysis of the common, unstressed use thus showed that, as a completive pronoun, some, like any, can play its role as a partitive by representing the quantity of the substantive as not only a

150

Language in the Mind

single part or an indefinite number of parts but also all the parts of the whole the speaker has in mind. All of which goes to show that both quantifiers include in their potential meaning an element ‘partitive.’ Furthermore, it is better to speak of its actualization in terms of “minimum,” “intermediate,” and “maximum” part or portion, rather than “subset,” since this analysis also applies to substantives expressing ‘continuum’ notions, as in and in the following examples. The stressed quantifier in and represents the substantive’s quantity as a small (intermediate) portion of the whole the speaker has in mind, whereas the unstressed version in represents it as a maximum portion equivalent to the implied whole: Some butter is not salted. Is any butter not saturated? Did you buy any/some butter? Even granted the capacity of some to represent its substantive’s quantity as equivalent to the whole as in (28), it remains that some does not have give rise to an ‘all inclusive, however many’ expressive effect and in this differs from any. To understand why, we will now examine the third element in their meaning, that which differentiates the two pronouns. EXISTENTIAL IMPORT

Although we have seen that both some and any can express a quantity as a minimal, intermediate or maximum part or portion of the whole the speaker has in mind, they do not express it in the same way. Thus in the following, the two questions do not say the same thing: Would you like any/some more of this pudding? With any “we hint that we expect a negative answer” whereas with some “we seem to elicit an affirmative answer” (Poutsma 1916, 1204). Likewise in a conditional context If you find any/some mistakes in my translation, don’t be angry.

Some and Any

151

the clause with some suggests that “the speaker is afraid there will be some mistakes” whereas that with any suggests that “the speaker does not know whether there are any mistakes” (Kruisinga and Erades 1960, 577). These comments suggest that the determining factor here is how the speaker views the situation, and particularly the existence of the part represented by the quantifier. This is confirmed by and , where some is used when its existence is affirmed, any when it is negated. Various scholars have described the difference involved in such uses, suggesting that some characterizes its existence as real or actual, any as possible, virtual, or hypothetical. The ‘real’ in some’s meaning can be seen both in positive contexts that affirm, and in questions that presume, the existence of the part. The ‘hypothetical’ in any’s meaning can of course arise where existence is denied (negative sentences) or seen as merely possible (questions), but also when it does exist, provided it is viewed as highly unlikely, as in: It was little short of a miracle that any in the train escaped with their lives. Without this impression of extreme unlikelihood – if, for example the main clause were It was quite normal that… – any could not be used here. What distinguishes the two quantifiers is thus the opposition between representing the part as ‘hypothetical’ and representing it as ‘real.’ This explains not only their complementary distribution in negative and positive sentences but also the possibility of using either one in questions and conditional clauses. It also explains why some is not possible in positive expressions like in any case, which evoke one eventuality among several but nothing actual. Finally it helps us understand why some is not found as a ‘universal’ quantifier, as in: Any tragedy, he maintains, has six elements: plot … Any relates the substantive to a whole that is so general in scope that it takes in both existing and all possible tragedies, whereas some, as we have seen, represents its substantive as a reality. This completes our analysis of the potential meaning of the two quantifiers. We have seen that common to both is a representation

Language in the Mind

152

of quantity as part of a whole, and that the size of this part can vary from a minimum to a maximum. We have also seen that, although they may express identical quantities, they characterize their imports differently as to their existential status: any always depicts it as ‘hypothetical,’ whereas some depicts it as ‘real.’ It remains for us to examine the relation between these two meanings in order to reconstruct their system in tongue. THE SYSTEM

As for any other problem of polysemy, we postulate an operation inherent in, or underlying, the meaning of each partitive pronoun in order to explain the different-sized parts it can express. That is, their potential meaning includes the possibility of a movement through the field of quantity that can be intercepted at the point best corresponding to what is being represented. What remains to be clarified, however, is the form of this movement – from the least to the greatest quantity or vice versa. To work this out, we must first locate the two phases in the system: which comes first, any or some? When one considers the notions we have called “hypothetical” and “real”2 as applied to the existence of something, one perceives a temporal relation between them. The existence of whatever one may have in mind, be it a holiday, a meal, the solution to a problem, etc., must be seen as possible, as being prepared, as sought, i.e. as hypothetical, before it is seen as actual, as ongoing, as possessed, i.e. as real. In other words, the hypothetical is not yet real, whereas the real is no longer merely hypothetical. From this necessary relationship we infer that any precedes some in the binary system. The system is therefore basically a two-phase process between maximum and minimum such as depicted in the following diagram by means of its inherent operative time: any

some

This to and fro movement through the field of quantity can be imagined as the viewing idea provided by each quantifier scanning the whole the speaker has in mind in order to situate in it the extent of space covered by the lexical import of the substantive and so to actualize it as a “part.” Starting with the total extent of the whole, which contains all possible parts, the viewing idea of any, like an adjustable lens, can reduce the scope it covers until it

Some and Any

153

focuses on the desired extent. If any’s scope is reduced as much as possible (it must remain a positive quantity) for the lexeme without the part coming into focus as something ‘hypothetical,’ the second phase of the system is reached. Its opposite form of movement begins at this most reduced scope, with the viewing idea of some scanning the whole in the reverse direction by expanding its scope until the part corresponding to the substantive’s extent of space comes into focus as something ‘real.’ In more abstract terms, the mental system makes possible a twophase, interceptible operation that begins with maximum quantity (M) for any since this is the scope that includes all possibilities of representing partitive quantity. If any’s contractive movement is not intercepted before or at the point of minimum quantity (m), the second phase begins. The expansive movement of some proceeds back toward the point of maximum quantity seen as ‘real’ (but not back to the starting point in time since this is impossible, as we have already seen). Because the system offers all these points of interception as possibilities, wherever either movement is intercepted, at (M) or (m) or at some intermediate point (I), the particular quantity actualized, i.e. the extent of the part, is seen in relation to all the other possible quantities, ‘hypothetical’ or ‘real.’ Because the point of inversion of this operation at (m) is common to both phases it occasions no break in the process, and this ensures the unity of the system in tongue. The relationships involved in this description can be diagrammed as in figure 9.1.

M

any I

m

some I

M

Figure 9.1

The similarity between this system and that of number in the substantive (cf. chapter 5) is due to the fact that both are based on the binary tensor mechanism. The difference between them, however, should not be lost from sight. The representation of space as continuate or discontinuate by means of number is common to all

154

Language in the Mind

substantives because each lexeme must be formed for internal incidence, for being incident to its own extensity. The representation of space as part of a whole by means of a partitive pronoun is far less general since not all noun phrases involve a notion to be presented as a part. Other pronoun subsystems present the notional import of the noun phrase in different ways, such as locating it with regard to “here” (the demonstratives). The most general of these subsystems is the article, as we will see in chapter 13. The above sketch of how the system was arrived at calls for a return to the data to test its capacity to explain other uses. We will not pursue the discussion of the partitive pronouns any further, however, since the aim here is to illustrate the type of system to be found in the grammatical words formed by the transpredicative parts of speech and to give an example of how pronouns make explicit the spatial support of the noun phrase. We will see that the auxiliaries, formed by another transpredicative part of speech, make explicit the temporal support of the verb phrase, but before examining them the system of the verb, a predicative part of speech, must be described.

10 The System of the Verb [Aristotle’s] definition of the verb … that which … ‘consignifies time’ or ‘signifies time in addition’. Padley (13)

CHRONOGENESIS

So far we have traced the way Guillaume’s view of language can be applied to the noun phrase in English. Assuming that any actual use betokens a potency, we started with observed facts of number usage and inferred the potential meaning of each morpheme, then worked back from there to the subsystem of number containing the two potential meanings, and thence to the even more abstract level of the system of the substantive, to reach finally the system of systems constituting the formal side of tongue. We have seen that each of these levels of analysis is concerned with something that really exists in tongue, not because the particular description given here necessarily corresponds to its preconscious reality – indeed, several systems within the substantive and some outside it are as yet seen only in hazy outline – but rather because of what the postulate on which all Guillaume’s work is based obliges us to accept: that there must exist a mental reality corresponding to what we have called the potential meaning of -s, to the subsystem of number, to the system of the substantive, and to the system of the parts of speech. Let us recall that Guillaume postulates as inevitable for the use of any word what his study of the article revealed: “The problem of the article, quite overwhelming 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

156

Language in the Mind

conveys it from potentiality into actuality during the instant when he is thinking/speaking” (1984, 4). Not accepting this analytical procedure of working back from observed data to the successive levels of systematization would involve refusing both the habit of everyday thinking that tells us we have the ability to speak our mother tongue and its corollary that permits a scientist to postulate the existence of some order behind the disparate results of direct observation, behind “the chaotic diversity of our sense-experience” as Einstein (1956, 98) puts it. In short, without Guillaume’s or some other postulate sufficiently general to condition any use, the chaotic diversity of usage remains chaotic, unexplainable. It is now time to illustrate further the method of analysis outlined in chapter 6 by applying it to the other basic part of speech, the word conjugated with tense to express time. We have had little to say about the verb up to this point, and yet it is thanks to a certain use of the verb in French that Guillaume was first able to discern the workings of a system in tongue and in so doing to perceive the parameter for analyzing all systems. He had been observing and reflecting on the articles in French, un and le, for over ten years without seeing clearly that their system is essentially operational, making possible an operation of representation. Then, one day in 1927 while walking down Boulevard St-Michel in Paris, he was reflecting on the verb, turning over in his mind the sentence: Si vous le faites et qu’il s’ensuive un accident, on vous en tiendra rigueur.1 He was pondering the fact that a clause introduced by si (= ‘if’), which is used to present something viewed as possible, was coordinated with a clause introduced by que (= ‘that’), which is used to present something viewed as real. Suddenly he saw a relationship between the two moods involved here: he became aware that for the two coordinated clauses to be equivalent the subjunctive, used with que (= ‘that’), must represent its event as possible, whereas the indicative, used with si (= ‘if’), must represent its event as real. But then he realized that this presupposed a temporal relationship between the two moods: since something is always possible before it is real, he concluded that there must be a notional chronology, as he was later to call it, linking subjunctive and indicative, that the speaker must somehow think the subjunctive before the indicative in the system of mood. For the first time, he had discerned between two morphemes a relation – a sequence between their respective representations – that was

The System of the Verb

157

not static but dynamic, implying that the two arise at different points in the same operation. This insight soon led him to two important conclusions. He saw that a grammatical system of morphemes, existing as an ordered set of potential meanings in tongue, is instituted to make possible an operation of representation that actualizes the morpheme called for by the projected sentence. He also realized that this operation of actualizing requires time, a time that is, of course, far too short for us to be conscious of or even to measure.2 On this basis he went on to develop his theory of chronogenesis,3 integrating the three systems of aspect, mood, and tense into an operational program that can give rise to the various images of time expressed by the different forms of the French verb. Published in 1929 as Temps et verbe. Théorie des aspects, des modes et des temps, this first breakthrough into the grammatical workings of the preconscious mind gave him the substratum of all systems and, by the same token, the parameter of all his subsequent analyses, operative time. We have already made use of this parameter in analyzing the system of number in the substantive and the system of some and any. And we have seen how the different phases of the system of the parts of speech are related in operative time, the underlying microtime required for the mind to occupy the position in the system called for by the sentence being constructed. “It takes time to think, as it takes time to walk,” Guillaume used to say. We shall now make use of this same analytical measure to explore the verb in English. Although English and French have the same part of speech structure, we cannot simply apply the theory of chronogenesis of the French verb to the English verb since each language has developed a different architecture within the basic structure. The theory conceived by Guillaume does, however, suggest in a general way what to look for in examining the various uses of the verb in English: we must be particularly attentive to anything suggestive of order in operative time, before and after, with the purpose of discerning the operations involved in generating the time image that, according to grammars since Aristotle, the verb brings to the sentence. The originality of this analysis is to include mood as an essential part of the system representing time, Guillaume being, to my knowledge, the only linguist to have integrated grammatical mood as a necessary component of the system of the verb. Besides making

158

Language in the Mind

the following description different from the many other descriptions of the English verb, incorporating mood in this way leads to a more explicit understanding of tense itself. We will start with tense, the most obvious of the verb’s characteristics, and work back from observations of usage in discourse to the system of representation in tongue.4 TENSE AND MOOD

Tense is traditionally considered a category for situating in time. As one well-known study puts it, tense is the “grammaticalised expression of location in time” (Comrie 1985, 9), a definition that leaves implicit what is located but does bring out where it is located: in time. Furthermore, as grammarians point out, this stretch of time is represented as an endless horizon, capable of containing anything we can imagine in time, and divided into two parts by the present instant. “The present moment, ‘now’, is nothing but the ever-fleeting boundary between the past and the future” according to Jespersen (1951, 258). Quirk et al. (175) are even more explicit: “time can be thought of as a line (theoretically, of infinite length) on which is located, as a continuously moving point, the present moment. Anything ahead of the present moment is in the future, and anything behind it is in the past.” Thus the two time-spheres or better, time-stretches of the indicative in English, the past and the nonpast (a term designating both the stretch of time contemporary with the act of speech and anything beyond it, the future), can be described respectively as stretching from a point as far back as one wishes to imagine up to the present, and as stretching from the present on as far as one wishes to imagine. In a diagram: past

nonpast

The present, represented as the limit between the two stretches of containing time, cannot as such contain anything since, as we have seen, a limit occupies no space, has no interiority.5 Thus there are only two tenses in the indicative, one for locating before the present in the past, the other for locating beyond it in the nonpast. As a consequence sentences like I see them today or I saw them yesterday are quite ordinary, but not *I see them yesterday because the

The System of the Verb

159

nonpast tense can situate only in the stretch of time beyond the instant represented as the present.6 Such observations are so commonplace that without some idea of what is pertinent guiding one’s observation it is easy to overlook far less frequent uses like It was suggested I see them yesterday, where it is not incompatible for yesterday, by means of its lexeme, to express where see situates in time. The fact is that in this syntactic context one can use I see with today, tomorrow, or yesterday. This indicates that the verb in this use is not constrained to locating something with reference to the present but can situate it anywhere on the horizon of time. To explain this curiosity of usage, which is one of those revelatory facts opening a new vista to anyone curious about the conditions permitting it, we are led to propose that the horizon of time is represented as an infinite stretch, as above, but that it is not divided into two parts by the representation of the present instant. The following diagram depicts this representation:

That is to say, in this type of use the finite verb see brings to the sentence a representation of time as an infinite horizon but without any representation of the present instant dividing it into two time-stretches. This, however, raises another problem: how is it that see, or almost any other finite verb, can bring to the sentence two quite different representations in this way, one with the present represented, the other without it? There is, after all, a major difference in meaning between the two. For some grammarians, in fact, the category of tense necessarily involves reference to the present, a position implying that see here has no tense, is a word sine tempore, a nonverb. Others, however, have proposed that there are two kinds of tense, “absolute” tenses, which necessarily provide reference to the present, and “relative” tenses, not providing reference to the present. This seems to be what is going on here when we consider the two very different ways of using see. It remains to discern what it is in the system of the verb that permits these two different representations of tense. If we return to the observation of usage, another unusual fact is brought to light by varying the person of the subject. Thus, in It was suggested that I/you/he see them yesterday/today/tomorrow, the -s inflection of the third person singular does not appear. This indicates

160

Language in the Mind

that the verb is in the subjunctive mood here. Other observable facts of usage, such as the form of the verb to be (It was suggested that I/you/he be there.) and expressing the negative without do auxiliary (…that I/you/he not see…) confirm that this is the subjunctive (cf. my 1964 article for more examples). Furthermore, because the event is not situated in either the past or the nonpast by being referred to the present, it is represented as a possibility, not a reality as in the indicative mood, and this is what permits it to be negated without an auxiliary, This view of the subjunctive is developed more fully in my 2007 study of the verb (120–44). We can then conclude that in the subjunctive the tense of see is relative, whereas in the indicative it is absolute. The important point here is that what distinguishes one mood from another is the way it represents the infinite horizon of time: in the subjunctive the present instant is not represented, and so there is no reference to the present and no locating in a particular time-stretch; in the indicative it is represented, and so there is reference to the present and locating in a time-stretch. This difference in constructing an image of time helps us to understand the different uses of verbs in each mood: because the present usually represents the instant of live experience for the speaker, a verb in the indicative presents its event in terms of reality, whereas a verb in the subjunctive cannot refer to the moment of experiential reality, so it depicts its event only in terms of possibility. The details of usage cannot be explored in this introductory sketch, nor need the status of the so-called past subjunctive be discussed. Rather we must turn to the nonfinite forms making up the other mood of the English verb to see how the horizon of time is represented. The infinitive and participles can have adjectival or substantival uses, a fact that led Guillaume to speak of them as constituting the quasi-nominal mood. Their most striking characteristic as verbs is that they are nonfinite, that is, they cannot predicate their event of a subject without the intermediary of a finite auxiliary. Furthermore none of them is capable of situating in a particular timestretch, only in undivided time, and in this they resemble a verb in the subjunctive. This shows that the quasi-nominal mood does not provide a representation of the present, that the horizon of time is depicted as an endless, unbroken stretch. However, it differs from the representation of time in the subjunctive because it is

The System of the Verb

161

oriented in the opposite direction, away from the future toward the past, as in the following diagram:

This image of time is the most elementary view of the infinite horizon of time, reflecting our ordinary experience of things coming to pass. It can be illustrated by the auctioneer’s call Going, going, going, gone. This corresponds to our experience of things arriving from the future (cf. the common expression The time is coming when…), passing (cf. as time goes by), and falling away into the past (cf. Time bears away all things). It reflects the impression of events taking place, not through the subject’s bringing them about, as in the finite moods, but through the inevitable movement of time itself bringing them to existence and carrying them away. This past-oriented movement of time is the way the infinite stretch of all-containing universe time is represented in the quasi-nominal mood, a proposal that cannot be discussed in detail here because this would entail examining its effect on the three tenses of this mood as they are used in the various verb compounds of English. The aim here is to provide a hypothetical basis for explaining the difference between a mood whose tenses cannot take a subject and the moods whose tenses do take a subject. Furthermore, I want to suggest that these two views of time, pastoriented and future-oriented, are not just a linguist’s invention but are abstract representations of impressions found in the common experience of speakers.7 The examination of tense in both simple and compound forms in my 2007 study helps bring out the important effect this difference in representing universe time has on the person of the verb and its contribution to the event’s realization. To make more explicit the first element that is involved in the commonplace that tense locates something in time, we have applied the theory of mood as a system for representing universe time and described how the three moods represent the horizon of time as a place for situating events. Resorting to Guillaume’s insight concerning the nature of grammatical systems, we can discern the relation between the moods which constitutes the operational basis of the system. It is a temporal relation whereby the different ways of representing arise at different points in the operation depending on how complete a representation is required. This operation

Language in the Mind

162

of chronogenesis can be depicted as in the following diagram by means of a vertical vector with a chevron for each of the three possible interceptions. The resulting representation, distinct for each mood, Guillaume calls a chronothesis (cf. 1991, 8). They can be depicted successively on a horizontal axis, as in figure 10.1, which is intended to show both the sequence of moods arising in their system and the representation of universe time each mood provides when actualized. quasi-nominal mood

chronogenesis

subjunctive mood

past

nonpast

indicative mood

Figure 10.1

Since every verb has tense, the system must be activated every time a speaker needs a verb for the sentence under construction. The least grammaticized representation of time, that of the quasinominal mood, arises right at the beginning of the movement, at the first instant of operative time. It depicts time as past-oriented, descending, bringing events to their place of existence, and carrying them beyond it to where they no longer exist. In descending universe time there is no representation of the present and no possibility of situating a grammatical person that can be made incident to a subject. The most complete representation, that of the indicative, arises at the end of the operation, at the last instant of operative time. It depicts time as extending in the opposite direction, ascending toward the future. This future orientation is required in English for situating in time not only the subject’s position but also the present instant dividing the horizon into two time-stretches. At some midpoint in the operation of chronogenesis arises the subjunctive representation of universe time, futureoriented but with no present instant represented. This view of the system of mood gives another good illustration of the way Guillaume conceives of the dynamism of a psychosystem.

The System of the Verb

163

Any such system is the potential for a mental operation between particular limits, a process whose actualization can be intercepted at the point required for what is to be represented. Indeed, he maintains that the only way the mind has of knowing itself is to take a cross-section of its own preconscious activity in this way. Hence, as we have already seen, the method of analysis in the psychosystematics of language is always the same: in the light of observed uses, to infer the operation made possible by a system and the point within this operation where each morpheme with its various uses arises. This is the linguist’s task: “to discover and reconstruct the operation leading up to the result” (Guillaume 1997, 94). In this way, system by system, we can gradually manage to describe more and more of our language potential, our mother tongue, which provides the “obvious ability to speak.” So far our examination of tense as, in Comrie’s terms, the “grammaticalised expression of location in time” has led us to see that the time in which tense locates is represented by the system of mood. It follows that the system of mood must produce its result before the system of tense can do its locating. It follows also that tense is not limited to the indicative, to absolute tenses, but is found as relative tenses in the subjunctive and the quasi-nominal moods. “Each chronothesis is a little system of tenses which has its date of definition in chronogenesis” (Guillaume 1991, 8). It remains for us to examine the element left implicit by Comrie’s definition, namely what is situated in time, because this too is part of the time image resulting from chronogenesis. That is to say, both the boundless stretch and what is situated in it are temporal entities, and both must be represented before tense can operate to give the part of speech verb. We call event the representation of any particular happening, be it a process or a state, that is located in time, using this term to designate the representation of what the speaker wishes to talk about by means of a verb. A theory of how time is represented grammatically within an event constitutes another important contribution made by Guillaume. TENSE AND ASPECT

It is a commonplace that any particular happening in our experience is always thought of as unfolding from its beginning to its end. Even in cases where we are not aware of either limit, as in The

164

Language in the Mind

St Lawrence flows past Quebec City, we understand the subject to be in the middle of the state’s duration, at some moment between its beginning and its end. This led Guillaume to oppose the duration of an event to the time in which it is situated. He described this opposition in various ways: time implicated vs time explicated; time inside vs time outside an event; time contained in an event vs time containing an event. The terminology finally adopted is that proposed by Valin (cf. 1994, 42) event time, the grammatical representation of the duration of an event, as opposed to universe time, the grammatical representation of the endless horizon on which any event can be situated.8 Having realized that these two representations of time are necessarily found within any verb and that universe time is represented by the system of mood, Guillaume had then to find which of the systems of the verb provided a representation of event time. Being familiar with Russian – it was his second language9 – he was well aware that verbs in Slavic languages represent event time in two ways, imperfective and perfective, and that these are considered aspects of the verb. He perceived a somewhat similar way of representing the event’s duration in Latin. He thus proposed that the system of aspect provides a representation of event time, a view quite consonant with that later proposed by Comrie (1976, 3): “aspects are different ways of viewing the internal temporal constituency of a situation.” Although his familiarity with Russian and Latin gave him a clear view of aspect in these languages, Guillaume made no attempt to impose the same imperfective/ perfective opposition on French. In fact here, as elsewhere, he was guided by the principle that “the mental side of tongue, of a construction in tongue whatever it may be, must be sought by appealing to its appearance, to its outside, to the semiology belonging to it, and should never be sought elsewhere” (1999, 2). Thus, by respecting the new means of expression invented by the Romance languages, he realized that they had all developed another basis for the system of aspect.10 Instead of opposing degrees of completude of an event as in Russian and Latin, aspect in French, Italian, etc. opposes the interior of the event, whatever its degree of completude, to the situation arising after an event, as a result of it. The former, expressed by the simple form of the verb in French, he named immanent aspect because it represents time contained within the event, its duration, or realization phase. The latter,

The System of the Verb

165

expressed by a compound form of the verb avoir 11 (= ‘have’) + past participle, he calls the transcendent aspect because it represents time beyond the event expressed by the past participle, its aftermath, or result phase. In English (and the other Germanic languages) a compound form parallel to that in the Romance languages has developed, have + past participle. Here too the auxiliary represents the result phase, depicting the person support12 as involved in whatever arises as an outcome of the past participle’s event. More precisely, the auxiliary, which cannot evoke a particular result, represents a place in time subsequent to the past participle event and linked to it in such a way that whatever particular state results from that event is seen to exist at the moment where the auxiliary is situated in time. For example in I have forgotten your name the verb does not focus on the forgetting (as would I forgot your name) but rather on the situation arising as a result of forgetting. The examination of many such uses (cf. Hirtle 1975) shows that this compound form always expresses the aftermath of the past participle event and so confirms that this is the transcendent aspect of English.13 This form can be diagrammed as follows: past participle

have

The immanent aspect of English, which represents the time within the event, is expressed by any form of the verb other than auxiliary have + past participle. This includes the simple form of the verb, as in French, but also do + infinitive and be + present participle (the progressive), because these forms also represent the inside of the event, each from its own point of view.14 A few characteristic uses of these compounds will be discussed later, but it will be useful at this point to examine briefly three characteristic uses of the simple form in the indicative in order to illustrate both the immanent aspect and the basic principle whereby a form exists as a potential in tongue to be actualized in diverse ways in discourse. This will show that the immanent aspect is not just an abstract form imagined by linguists but a means of representing the different ways we experience the duration of happenings. Grammarians have observed that a sentence with the simple form often tells us the subject begins, carries out and completes the event expressed by the verb in the past or the nonpast, as in:

Language in the Mind

166

I drew a circle. He shoots! I promise to do it. In each such case, the verb expresses the realization of the whole event by its person support, depicted moving through the event to the final instant of its duration. In the following diagram, the vector depicts the unfolding of the event from beginning B to end E, and the chevron shows the final position of the support: B

E

The simple form can also express an event whose realization is not carried out but is left wholly prospective: Either that alligator goes or I go. If it rains, we won’t go. Perhaps the most common use of this type is the imperative: Shut the door, please. Here the person support is situated at the beginning of the event, in a position to carry it out, but none of the event’s realization is represented as actually carried out so it remains prospective. This prospective realization can be depicted in a diagram by means of a dotted line: B

E

Finally, the simple form is often used to represent an event at some mid-point of its realization without representing either its beginning or its end: They need money. He weighs two hundred pounds. It is eleven o’clock. This type of event can be diagrammed most simply by means of a short arrow somewhere between the beginning and the end of

The System of the Verb

167

the event. This depicts the moment in the event’s duration where the person support is situated, which is the portion of event time represented by the verb, and leaves implicit the already realized portion of duration (if any) and its subsequent duration (if any), neither of these stretches being represented by the verb:

When we bring together these three manners of representing an event and consider the relationships involved, we get a view of the potential provided by the simple form in tongue: the duration of an event with its three necessary moments – beginning, middle, and end. In terms of aspect, it can be seen that the simple form provides a representation of the time within the event, with three possibilities of representing the person support: just at the beginning of this stretch of duration, situated somewhere in it, or proceeding through to its end. Being a potential offering different possibilities for representation, the stretch of duration is depicted by a dotted line in the following diagram: B

E

Since the person support must be either at the beginning, in the middle, or at the end of an event, there are no other possibilities for the inside of an event.15 That is to say, envisaging the simple form as making possible the situating of the support person anywhere within the event in this way provides a potential meaning in tongue capable of generating all uses in discourse. This is why many grammars propose ‘dynamic’ and ‘stative’ as the mains uses of the simple form, uses that correspond respectively to endpoint and midpoint interceptions. However the various ‘prospective’ uses, resulting from an initial interception, are not grouped together in grammars to show what they have in common. Indeed, this type of use came into focus only when the simple form was viewed as a morpheme with a dynamic meaning in tongue capable of being actualized with different senses in discourse. Again we can see why the principle that “grammatical symbols in tongue cover movement” is very important. A similar type of analysis applies to the transcendent aspect. Viewing the auxiliary, the finite component of the transcendent, dynamically, we can see characteristic moments in the aftermath where the support person can be situated. In a sentence like

168

Language in the Mind

I’ve already seen that film. we get the impression that the subject is at some place in time after seeing the film, but how long after is left undetermined. That is to say, this indefinite stretch, long or short, separating the seeing from the moment in the result phase that exists at the time of speaking is not represented by have, and so it is depicted by a dotted line in the following diagram: seen

have

Quite different is the impression given by an exclamation like Eureka! I’ve found it! where the moment of speaking is felt to be just after the finding, as though no time had elapsed between the end of the finding and the point in the aftermath where the subject is situated. Likewise in We’ve lived here for a long time. there is no gap between the long time and the point where the subject is situated. Since the point immediately following the past participle event is necessarily the first instant of the result phase, both these examples can be depicted as follows16: lived / found

have

Thus the support person can be seen either at the beginning of the event’s aftermath or somewhere in the middle of it. The fact that no examples are found where the subject is situated at the final instant of the aftermath is significant: there is no end to being after something done. When we bring together the two manners of representing an aftermath found in discourse and consider the relationships involved, we get a view of the never-ending result phase of an event. And in terms of the system of aspect, it can be seen that in tongue have auxiliary offers the potential of representing a point in the transcendence of a past participle event, either at its first instant or anywhere beyond that, to situate the person support. The potential for these

The System of the Verb

169

different ways of actualizing have is suggested in the following diagram by the dotted line and the two possible interceptions: past participle

have

Since the support person must be either just after an event or somewhere beyond that, there are no other possibilities for the aftermath of an event.17 That is to say, envisaging auxiliary have in this way provides a meaning potential capable of generating its different uses in discourse, as we will see below. This is why many grammars consider that the two main uses of the nonpast tense of the transcendent express ‘event-lasting-up-to-the-present’ and ‘indefinite-past,’ which correspond respectively to intercepting the result phase at its initial point and at some later point. Again, when confronted with the polysemy of a form, we can see the advantage of viewing meaning in tongue as inherently a movement. We have seen how the same analytic method as applied to the number morphemes above can be applied to both the simple form and have auxiliary to obtain a view of their potential in tongue. This method consisted in observing the representations of time a form brings to the sentence and discerning a temporal relationship between these expressed meanings to establish the operation of representation the morpheme makes possible in tongue. This way of viewing data – as the result of intercepting a mental operation of representation at different moments of its development – is invariable in psychosystematics since it always leads us back to the condition in tongue, the potential meaning of a morpheme giving rise to its observed polysemy in discourse. As in the case of number, the next step is to discern the relationship between the two morphemes that binds them together in a system. This step consists in arranging their potential meanings in an operative sequence to get a view of the act of representation giving rise to them. In the case of aspect this notional chronology is quite obvious: the immanence of an event must precede its transcendence. The system of aspect can thus be diagrammed as in figure 10.2, taking the simple form as the most common way of expressing the immanent aspect. Whenever we use a verb in English, event time must be represented because no event can be imagined without duration. Whether the immanence or the transcendence of an event is represented

Language in the Mind

170

Immanent Aspect

Transcendent Aspect

simple form past participle

have Figure 10.2

depends on the speaker’s intended message. It is this representation produced by the system of aspect that must then be located in time. TENSE

Reflection on the traditional definition of tense has led us to see that it involves a double representation of time: event time represented by aspect as the duration of any event or its aftermath (which, because it occupies a different place in time, is in effect a different event) to be located in time, and universe time represented by mood as an infinite horizon for locating the event. That is to say, both aspect and mood must first provide their representations if tense is to carry out its role of locating the event with its event time in universe time. This gives us a glimpse of the order in which the systems involved in chronogenesis are activated. As a dynamic system, then, tense situates an event, formed by an aspect, in time, formed by a mood, but since each mood provides a different representation of universe time, the conditions for situating the event vary. It is clear how tense situates an event in time in the indicative, the mood most grammarians have in mind when speaking of tense: everything is referred to the present instant. As Quirk et al. (175) point out: “Anything ahead of the present moment is in the future, and anything behind it is in the past.” A typical actualization of the tenses of the indicative18 can be diagrammed as follows: went past

goes nonpast

On the other hand, we have seen that relative tenses do not locate or situate an event in a particular time-stretch but leave its place in time undetermined, often to be related to something else in the sentence. To capture this quite different way of “locating” an event in time it is preferable to speak of tense in terms of making an

The System of the Verb

171

event incident to time. The term “incident,” it will be recalled, has its etymological sense of ‘falling on’ and so focuses more on the operation involved and less on the result of being in a determined place. This more general way of conceiving how tense operates includes both absolute and relative tenses and so allows for viewing as tenses forms arising in moods other than the indicative. In the quasi-nominal mood, there is no determining the event’s place in universe time by relating it to the present. As a consequence, the tenses in this mood reflect not different positions of the event in universe time but different phases of its actualization relative to where its spatial support is in time, wherever that may be. (These forms must include the representation of a spatial support because without that, an event would not be representable, as Duffley 1992a shows.) There are three such relations: an event can be seen approaching, passing through or receding from its undetermined place in time, and these three views correspond to the three tenses of this mood, the infinitive, the present participle, and the past participle respectively. Examples like an hour to go as opposed to an hour gone, or the well-known going, going, gone! give an idea of the sort of expressive effect each tense can give rise to.19 Figure 10.3 depicts this system of tenses with reference to any given point. go go ing gone Figure 10.3

These forms, included in the tense system of English because they can express an event in time, raise a question: how can they evoke something taking place if they have no subject to bring it about? It was to answer this question that past-oriented universe time was proposed as a hypothesis for this mood. A representation based on the impression of time as “an ever-flowing stream” bringing things into existence and carrying them away into the past provides the condition for an event to take place without evoking the intervention of its spatial support as subject. Thanks to this hypothesis we can extend the system of tenses to include the quasinominal forms and explain why this is a nonfinite mood.

172

Language in the Mind

In the subjunctive, as in the indicative, future-oriented universe time is proposed as a condition for viewing the event as an activity of (or at least dependent on) the subject. This calls for a beginningto-end view of event time. That is, event time is represented as a stretch open to the activity of the person support. More precisely, the person support is positioned to help bring about the event’s realization, and so the verb takes a subject. In the subjunctive, however, tense does not refer the event to the present and so can express only its possible accomplishment anywhere in time. The event is therefore depicted in figure 10.4 by means of a dotted vector. (he) go Figure 10.4

In figure 10.5, by combining the three moods and the forms they contain, we finally get a view of all the tenses of English.20 go go ing

chronogenesis

gone

(he) go

went

goes nonpast

past

Figure 10.5

This synoptic view of the tense system of English shows that reference to the present is not an essential. What is common to both absolute and relative tenses is that they bring together a double representation of time, event time and universe time. The fact that the constituents of tense, as conceived by many grammarians, are made explicit in this way enables us to explain the uses

The System of the Verb

173

of the different tenses, both relative and absolute. This was made possible by Guillaume’s discovery of the place of mood as an essential component of the system for representing time, a basic insight distinguishing his view of the verb from that of other linguists and grammarians. If tradition is right in characterizing a verb as a word cum tempore, this combining of time as what is contained in a particular event and time general enough to contain any event appears to be essential to the verb as a part of speech within the word. Before any firm conclusions can be drawn at this level of generalization, however, the place of the other two subsystems of the verb – voice and person – must be more carefully described, a task that would take us far beyond the scope of our aim here, which is only to illustrate the analytic method of psychosystematics in the field of the verb. A more complete description of the verb system with an attempt to include person and voice is provided in my Lessons on the English Verb (2007).

11 Auxiliaries: How Do They Help? Dematerialization can be complete or incomplete. The effect of incomplete dematerialization is to produce words which are no longer semantemes and are not yet quite morphemes. This is more particularly the case of the auxiliaries. Guillaume (1991, 145)

THE PROBLEM

It is traditional in grammars to consider have an auxiliary verb in the have + past participle construction but an ordinary verb in its other uses. A number of recent studies have questioned whether an auxiliary is really a verb in view of the fact that its syntactic performance is so different from that of a prototypical “full” verb. According to the tradition, auxiliaries are verbs because they have the same morphological inflections as verbs, but to justify how they differ in syntax nothing is said except that they are “helping” verbs, little more than a tautology. To make the point, it would seem necessary to show how auxiliaries help other verbs. Otherwise, the argument of tradition is simply that auxiliaries look like verbs, so they are verbs. The counterargument – they behave differently from verbs so they are not verbs – is hardly more convincing, because it would seem necessary to show why they behave the way they do. Both approaches are based on observed data – on grammatical inflections in the one case and on syntax in the other – but unless all the data is taken into account, there is no solid basis for demonstrating whether these words are verbs or not. This twofold problem, morphology and syntax, is particularly acute in English, a language known for its reduced grammatical inflections and extensive development of auxiliaries and compound constructions. Attempting to see if auxiliaries are verbs will

Auxiliaries

175

provide an occasion to apply the theory of chronogenesis and bring out the important principle already seen that a word brings to the sentence both a lexical and a grammatical import and that this makeup conditions how the word is used. Examining the other facet of the problem, syntactic behaviour, will provide an opportunity to introduce the widespread process of dematerialization. Finally, this discussion will permit us to explore the relations between three auxiliaries – do, be, and have – with a view to describing their system in the transpredicative parts of speech.1 VERBS OR NOT?

Since this question presupposes some notion of what a verb is, it will be addressed in light of the theory of chronogenesis presented in the preceding chapter. We saw that the hallmark of a verb is tense, a grammatical form making event time (represented by the system of aspect) incident to universe time (represented by the system of mood). Bringing together this double representation of time inherent in any verb appears to be what characterizes the part of speech itself, since it indicates final categorization of the lexical input, the event, in time and not in space. It is therefore important at the outset to observe whether auxiliaries incorporate tense the way full verbs do. As the auxiliary of the progressive, be is conjugated in both tenses of the indicative (They are/were studying), a clear indication that it incorporates a representation of universe time divided into two time-stretches by the present instant. It is also found in the subjunctive (I suggest you be studying when…), a fact showing that it can be formed with an image of undivided universe time oriented toward time ahead. In the quasi-nominal mood, be auxiliary is commonly found as an infinitive (I ought to be studying) and as a past participle (They have been studying), again showing that it can be conjugated in this mood with its descending, undivided view of universe time.2 This last example also shows that be is found in the transcendent aspect as well as the immanent aspect (the other examples given). From this it can be concluded that be auxiliary is formed by the system of aspect, by the system of mood, and subsequently by the system of tense just as any full verb is. As the auxiliary of the transcendent aspect, have is also found in both tenses of the indicative (They have/had studied it). It is common

176

Language in the Mind

in the quasi-nominal mood as an infinitive (I ought to have studied it) and as a present participle (Having studied it, I…). Again we see that have is formed by the system of mood to give it a representation of universe time.3 Furthermore, the fact that have auxiliary does not occur as a past participle indicates that its own event time is represented only on the inside, never from the point of view of its aftermath.4 This corroborates what was brought out in the last chapter, where we saw that the event time of have, which is the aftermath of the past participle event, cannot itself be represented at its final instant. Thus although have auxiliary occurs only in two moods and one aspect, we must conclude that it includes as formative elements representations of both event time (always immanent) and universe time (either undivided and descending or divided and ascending), and so its tenses are of the same nature as those of a full verb. Do auxiliary is not conjugated in the quasi-nominal mood5 and is rarely used in the subjunctive mood. It does however occur commonly in both the past and nonpast tenses of the indicative (They do/did study). Although its conjugation is thus practically restricted to the indicative mood, this clearly indicates that do auxiliary does situate a representation of event time in universe time and so suffices to show that its tenses are like those of any full verb. This brief survey thus indicates that the three auxiliaries are words cum tempore, understood in the sense of being formed by the process of chronogenesis. As verbs, then, we can see that they share other characteristics with full verbs, and particularly that of making the import of the verb phrase incident to a support. That is to say, auxiliaries carry out the primordial syntactic function of any verb in the finite moods, predicating something of the subject, a clear indication that auxiliaries can incorporate a representation of support person like any other verb. On the other hand, auxiliaries cannot fulfill this function on their own, but require the input (or at least the recall) of a full verb’s import. Understanding the reason for this difference will bring out how auxiliaries differ from full verbs and provide a basis for explaining other syntactic differences. HOW DO AUXILIARIES DIFFER?

Although various authors consider do a “dummy,” a word with no meaning, more cautious writers try to avoid the contradiction involved in that approach by speaking of the auxiliaries as lexically

Auxiliaries

177

depleted or bleached. Langacker, for example, considers that do auxiliary expresses whatever “content is inherent in the notion of a process … nothing beyond this bare minimum” (1991a, 239) but does not describe this inherent content. Guillaume, one of the first to explore this question (1938/1964), also considered that the lexical meaning of the full verb giving rise historically to the auxiliary was reduced to its most elemental state by a process of dematerialization. (The advantage of this term is that it focuses on the material significate, the lexical matter, and does not call into question the formal significate, the grammatical meaning, which, as we have just seen, is like that of any full verb.) It will be recalled that this is an historical process of leaving aside particularizing impressions with the result that the lexeme becomes more and more general. As Guillaume pointed out, when this process is not pushed to the limit it produces grammatical words, “words which are no longer semantemes and are not yet quite morphemes” (1991, 145). This is the in-between state of auxiliaries, a state that prompted one scholar to ask whether auxiliaries should be treated as part of the morphology or part of the syntax of a given language (cf. Heine, 25). When we examine the forming of a verb compound below we will see that this static manner of posing the question is inadequate, since it is the way they are formed as words, their morphology, that permits them to function syntactically as auxiliaries. Because of dematerialization, do, be, and have as auxiliaries can no longer express an event, which is always the representation of some particular happening, of something arising in the speaker’s experience, but they do express some condition so general that it may be found in any event. This condition constituting the material residue of an auxiliary verb is so general in fact that it is difficult to observe and describe directly, the way one can observe and paraphrase the meaning of full verbs like eat or drink. We must therefore proceed by way of analysis to discern the result of ideogenesis in an auxiliary. To get beyond the particularity of the great variety of events represented by individual full verbs various scholars have proposed types of event. Some, concentrating on the manner in which the happening occurs, have proposed event types such as accomplishments, achievements, activities, states, etc., the most general such grouping being the distinction between dynamic and stative events. Others, focusing more on the way a happening relates to time,

178

Language in the Mind

group events as inceptive, instantaneous, durative, iterative, terminative, etc. Since an auxiliary helps express events of all types, it would seem that its residual lexical element must be common to all these different types. The one thing that is found in any happening, and in fact constitutes a necessary condition for any process or state to be part of our experience is a certain duration. Be it long or short, complete or incomplete, punctual or ongoing, in the offing or already over, no occurrence, process, state, etc. can be experienced or even imagined without a stretch of time between its beginning and its end.6 Thus inherent in the very lexical makeup of any word representing an event is the impression of a stretch of duration. One can imagine, in a sort of thought experiment, what would happen if we took the lexical matter of any verb and withdrew all impressions particular to its event and to its type of event, leaving only what is most general, the “content … inherent in the notion of a process.” The result would be the representation of a stretch of duration emptied of any particular activity or state. This appears to be what happened in the historical development of auxiliaries, each of which was derived from a full verb, but one whose meaning was already far more general than that of most verbs. Thus have, with its meaning of ‘possession,’ developed a sense without impressions of ‘holding’ or ‘ownership’ and the like, leaving merely a representation of the stretch of duration in which the state of ‘possessing’ would be seen to exist if this dematerialization had not taken place. In like fashion, from the meaning of ‘carrying out,’ ‘undertaking,’ ‘realizing,’ do developed a sense, perhaps through its use as a causative in Old English, where only the time involved in any activity or state is represented. Finally, the ‘existence’ meaning of be appears to have been the source of the dematerialized sense found in the auxiliary. All three auxiliaries, then, developed in the same direction, retaining what is most general, though each has its own distinguishing trait, as we shall see. The point here is that they all bring to the verb compound a stretch of duration, a span of time occupied by no event.7 This gives us an overall view of the makeup of an auxiliary within the general structure of the word. Like any other word, its meaning consists of matter and grammatical form, but what distinguishes it from full verbs is that its lexical matter has been reduced to a minimum,8 whereas its grammatical form is left intact. Guillaume

Auxiliaries

179

liked to express this by means of the following notation (cf. 1938/ 1964, 78), where (M) stands for lexical matter, (F) for grammatical form, and (-q) for the quantity depleted: Auxiliary = M(-q) + F

It is obvious from this formula why the auxiliary cannot be used alone. Thanks to its aspect and mood it is capable of representing both event time and a stretch of universe time, but this event time – the grammatical representation of the durational element constituting the auxiliary’s remaining lexical matter – is empty of any particular event. The event time of a non-event, of nothing, would be literally “non-sensical” since it would express nothing, and so would be of no use in discourse. This is why, in the ordinary usage of auxiliaries, a second verb form is required, one bringing in the representation of a particular event. It requires a nonfinite form – infinitive, present participle or past participle – to provide a complement of lexical matter to make up for what is lacking in the auxiliary. That is to say, when the lexical import of the infinitive or participle is made incident to the stretch of event time represented by the auxiliary it constitutes an event situated in time and ready to be predicated of a subject. In this way, bringing together an infinitive or participle with an auxiliary to form a verb compound reconstitutes a lexical-grammatical whole on a par with any simple verb. This view of auxiliaries as highly dematerialized verbs may appear to be such an abstract hypothesis that it has little to do with the practical reality of discourse. To show that it does correspond to observable facts, I want to compare do as “full” verb, as suppletive, and as auxiliary, in an attempt to trace its dematerialization through its three characteristic uses. THE VERB DO

As most commonly used, do has been described as “a generalpurpose agentive transitive verb” (Quirk et al., 135) and is found in a great variety of contexts: to do the dishes, to do history, to do five years, to do well, etc., etc. It is often contrasted with make in ESL classes to avoid errors like: *I nearly did a mistake.

180

Language in the Mind

To distinguish between the two verbs in a way that will help students is not easy, especially when confronted with pairs like to do a dive vs to make a dive, to do the beds vs to make the beds. This difficulty illustrates clearly that do, though quite abstract even in this its most concrete sense, cannot cover the whole field of activities that a verb like French faire or Spanish hacer can express. Compared with this general-purpose use, do in its suppletive, propredication, or vicarious use has a greater range since it can substitute for many other verbs including make: I was so nervous about making a mistake that I nearly did so. We asked him to make supper and he did so. This fact is significant since it has long been established in semantic analysis that a greater extension, or range of application, is the consequence of a reduced comprehension, or import. That is, we are led to infer that the lexical input of do in this second use is less, that the element (whatever that may be) opposing it to make in the first use is not present, that the import of suppletive do is dematerialized when compared with the import of general-purpose do. In light of our postulate that all lexemes in tongue consist of a meaning potential, we can understand these two uses of do as arising from two different interceptions of its ideogenesis, the process of actualizing the lexical import of a word. That is, the partly materialized sense of suppletive do can be understood as arising from its ideogenesis before it reaches its end, whereas the fully materialized sense of general-purpose do results from intercepting the operation at its end. However until we can see clearly what element of meaning differentiates the two, this hypothesis cannot be considered to be confirmed. Turning now to the third use of do, we can compare its range of application as an auxiliary with its range in the suppletive use. It has been pointed out (cf. Quirk et al., 878–9) that in the latter use do can supplete for verbs expressing an activity, as in and , and in: I asked her to learn the poem and she did so. but it cannot substitute for verbs expressing a state: *I asked her to know the poem and she did so.

Auxiliaries

181

That is, do has a narrower range in its suppletive use than in its auxiliary use, since the auxiliary is used with verbs of any sort, dynamic or stative: I asked her to learn the poem and she did. I asked her to know the poem and she did.9 If then the auxiliary has a range of application greater than either the suppletive or the general-purpose verb, we can conclude that its lexical import is more dematerialized than either of these main verb uses, that its content is reduced to whatever “is inherent in the notion of a process [or a state] … nothing beyond this bare minimum.” This suggests that the auxiliary’s lexical import is obtained by intercepting do’s ideogenesis very early, at its very beginning, before anything more than the essential has been materialized. It is this type of comparative work that led to posing the question: what is common to all events, without which an event is inconceivable? In light of Guillaume’s analysis of tense in terms of event time and universe time it appeared that a certain duration was a necessary condition for representing a lexical import as an event and that the lexical import of do auxiliary is reduced to this bare minimum. The same can be said for be and have. In this way there emerged a view of auxiliaries as verbs whose lexical import is no longer a particular matter but an element of such abstractness that it cannot be used in a sentence without a material input. Because its degree of abstractness, of dematerialization, is unsurpassable for a verb, the auxiliary is situated at the final limit of the transpredicative parts of speech. Recognizing that auxiliaries differ from main verbs in this way helps us understand why they are used with the quasi-nominal form of a full verb: to make up for the lexical deficiency of the auxiliary. This is not the whole story however because it does not tell us how the “helping verb” helps, or why the mind instituted in tongue the form for a lexically deficient verb, a verb lacking the lexical matter required to function as a unit of discourse. To understand this, we must take a closer look at the infinitive and participles. HOW DO AUXILIARIES HELP?

We have seen that the tenses of the quasi-nominal mood can neither situate their event with regard to the present nor take a

182

Language in the Mind

subject. Arising at the beginning of chronogenesis, they are lacking the grammatical development found at the final instant of chronogenesis in the tenses of the indicative. The role of the auxiliary is to make up for this lack, to provide the quasi-nominal form’s event with the grammatical determinations required. To show how the auxiliary does this calls for a rather technical explanation (cf. Hirtle 1997a), which is worth summarizing here to recall that word meaning consists of matter and form and to illustrate once again that it is the way of realizing this basic structure that determines how a word fits in with other words in the sentence, how it plays its role as a “part” of speech. As we saw in the previous chapter, the tense of the quasi-nominal form situates its event in past-oriented time but leaves undetermined its relation to the present. In a verb compound, we observe that the quasi-nominal event (or, when used with auxiliary have, its result) is always made incident to the time-stretch, the temporal support, provided by the auxiliary. That is, the process of compounding appears to make the event of the infinitive or participle incident to the auxiliary’s place in time. In this way, the event’s relation to the present is determined, with the result that the event takes on the absolute tense of the auxiliary. This is how an auxiliary in the indicative mood serves to provide the quasi-nominal form with a determined support in time. In the subjunctive and quasinominal moods the auxiliary cannot provide this determination because tense in these moods permits no reference to the present. Thus we can say that auxiliaries help by completing the grammatical forming of the event. This is the counterpart of a point made above, where we saw that the infinitive or participle provides a lexical refill for the dematerialized auxiliary. Our analysis thus brings out the two strands, matter and form, which, woven together, constitute the meaning of a verb compound. The coherence of this double import can be expressed by recalling Guillaume’s formula for the auxiliary – with M(-q) for quantity of lexical matter dematerialized and F for grammatical form – and adding a formula for the quasi-nominal forms, with (-t) for undetermined tense and (t) for determined or absolute tense, and a formula for the resulting compound (see figure 11.1). These formulas bring out the contribution of each word to the resulting verb compound, which is such a closely knit construction that it is generally referred to as “the verb” of the clause or sentence, even though it is composed of two verbs, two constructs

Auxiliaries

183

Infinitive or Participle = M(q) + F(-t) Auxiliary (indicative) = M(-q) + F(t) Verb compound

= M

+ F

Figure 11.1

of chronogenesis. (This contrasts with adjective + substantive combinations, which are never referred to as “a noun.”) The reason for such a tightly integrated construction, neither component of which can function as a finite verb on its own, is that the lexical import receives its final grammatical determination in the auxiliary in a way analogous to that of ordinary words, where the lexical import resulting from ideogenesis is grammaticized through the word’s own morphogenesis. That is, a verb compound gives us, at the moment of syntactically linking the two words together, a partial replica of what happens during the constructing of any word to prepare it to play its part in the sentence. One effect of compounding in this way is that the person support of the quasi-nominal event is also situated in the stretch of time provided by the auxiliary. That is, identified with the person of the auxiliary, the event’s spatial support disposes of a future-oriented span and so is in a position to condition the actualizing of the event or its result phase. As a consequence, the compound can be made incident to its external support, predicating the event of the subject. This manner of viewing how verb compounds are formed is based on the concept of chronogenesis – the idea of providing a more complete grammatical form from mood to mood – and on the lexical depletion of the auxiliary. It raises the question of why so many verb compounds have developed in English and ultimately poses the problem, basic to the English verb, of the opposition between simple and compound forms. This problem can be profitably addressed only after an examination of each verb compound to bring out the operativity involved in making the import of a particular quasi-nominal form incident to the support offered by a particular auxiliary. Only then will we be able to discern what is common to all verb compounds. We will illustrate briefly the three compounds in the next chapter, but before leaving the auxiliaries it remains for us to describe how they relate to one another.

184

Language in the Mind THE SYSTEM OF DO, BE, AND HAVE

We have seen that a verb in the quasi-nominal mood represents an event with a spatial support but with no determined place in time and that an auxiliary in the indicative is a verb dematerialized to the point where all that is left of its event is an empty stretch of duration situated in either the past or the nonpast. The incidence of the event to this stretch of duration forms the compound verb. On comparing the three auxiliaries, however, we find that this stretch of duration is characterized differently in relation to the event that is to be made incident to it. For example, have’s stretch of duration always follows the event, represented as accomplished by the participle, as we saw in the last chapter. That is, it situates the spatial support after the event in a position to actualize some moment, the initial moment or a later one, of the event’s result phase, as depicted in figure 11.2. event

result phase

gone had Figure 11.2

As the auxiliary of the progressive, be’s stretch of duration coincides with some moment within the participle’s event. That is to say, it situates the spatial support at a moment somewhere between the beginning and the end of the event so that it is seen involved in realizing the next moment of an event represented by the participle as already begun but, unlike what we saw above, not yet over, as depicted in figure 11.3. This permits the expression of an ongoing event by means of the progressive form with all its nuances and effects in discourse (see the next chapter for details). going was Figure 11.3

Like be, do auxiliary provides a way of expressing the immanent aspect because its stretch of duration coincides with the interior of the event. On the other hand, do can provide a stretch not just for the next moment but for the whole event. Thus it situates the

Auxiliaries

185

spatial support at a moment where, as we saw in the previous chapter, it can realize the whole event, represented by the infinitive as prospective (see figure 11.4). This permits the do + infinitive construction to express the event’s realization as an alternative to its non-realization, thus giving rise to its well-known uses in negating, questioning, and emphasizing the event, as we will see below (cf. my 2007 study for more details). event go did Figure 11.4

In fulfilling their role of helping the event find its place in grammatical time, each auxiliary thus imports as its remaining lexical matter a stretch of duration involving a particular relation with the time involved in the nonfinite event. As a consequence, each auxiliary positions the spatial support differently with regard to the event. This can be shown by comparing the three following sentences: Did he eat it? Was he eating it? Had he eaten it? Where questions the realization of the event from beginning to end, questions the realization of a moment in the middle of the event, and questions the event through the realization of a moment in its aftermath. To be able to evoke its event in the past in this way each auxiliary must bring to the verb phrase a stretch of duration that can accommodate the event as formed by its quasi-nominal tense. Comparing them with regard to their remaining lexical import, we see that: 1 do offers a stretch of time for the infinitive event as prospective, putting the spatial support at the beginning in a position to realize the event; 2 be offers a stretch of time for the participle’s event as already begun but not yet over, putting the spatial support in a midposition to realize its next moment;

Language in the Mind

186

3 have offers a stretch of time for the aftermath of the participle’s event as accomplished, putting the spatial support in a position to realize the next moment of the aftermath. B

did

E

was had Figure 11.5

The relationships between the auxiliaries can be depicted as in figure 11.5, where the stretch of duration of each auxiliary is indicated by a dotted line to suggest that it offers different possibilities for situating the spatial support. This system of auxiliaries thus permits the speaker to represent stretches of duration situating the event in time from its beginning on, at some point in its middle and with regard to some point after its end. These are the three necessary moments in the realization of any event, a forward movement that carries the support from the point where the event comes into existence to the point where it goes out of existence leaving only its consequences in time. These auxiliaries thus bring to the sentence conditions for representing the event as a reality in time, not conditions for representing the event as a potentiality, a role left to the modal auxiliaries. A TRANSPREDICATIVE PART OF SPEECH

Like the partitive pronouns some and any examined above, the auxiliaries arise as transpredicative parts of speech and provide another example of how grammatical words are organized in tongue on the basis of their matter, what remains of their lexical meanings. One advantage of examining this particular system is, as we saw above, that each of these verbs still has full verb uses. This fact allows us to imagine, with a certain plausibility, the diachronic process of dematerialization required to institute in tongue for each verb a potential meaning sufficiently abstract to make possible, as one of its actual meanings, the residual meaning observed in its auxiliary use. Thus, in the case of have, one constant characteristic

Auxiliaries

187

of the notion of ‘possession’ is that it is the result of some process of acquiring, so it is hardly surprising that, when dematerialized, it situates the subject of the compound in a position after the event. Similarly, because the notion of ‘existence’ implies necessarily a position between the point where something begins in time and the point where it ends, be auxiliary situates its support in the middle of the event. And if do did develop from the causative use of Old English don, we can readily understand why the auxiliary came to focus on the point where the subject brings the event into existence, the point where the cause becomes effective. These suggested developments require, of course, careful historical observation for any available evidence to confirm them or to suggest some other line of development. The point here is that observation of usage today indicates that the meanings of these verbs have all been dematerialized to give similar, but not identical, results that are now instituted in tongue. This system of auxiliaries is thus no accident, but the result of that “ceaseless working of thought,” which Guillaume (1984, 65) considers to be the source of language systems: “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 worth bringing out the coherence of this system, typical of formal systems in the transpredicative parts of speech. Thus each of the three auxiliaries depends on the position expressed by the others to express its own position vis-à-vis the event: to have an impression of being after, the beginning and middle (and even the end) must have been left behind; to have an impression of being at the beginning, the middle and end must be seen as prospective; to have an impression of being in the middle, the beginning must be felt as left behind, the end as in the offing. That is, the system consists of a set of positions in a movement inherent in any event’s duration and a mechanism of interception enabling the mind to hold up the movement at the appropriate point. The same type of mechanism was described above in the system of mood. From the point of view of their system, auxiliaries are in striking contrast with full verbs, which do not appear to have the same sort of coherent arrangement. One need merely consult a dictionary of synonyms for, say, to eat to realize that the proliferation of full verbs corresponds to the unending variety of our experience with

188

Language in the Mind

regard to eating, and not to the unvarying conditions for representing duration in the abstract for any event whatsoever. As a consequence of this marked difference between predicative and transpredicative parts of speech, the meanings of words arising as transpredicatives are part of closed systems to which an individual speaker cannot add a new item, whereas the meanings of words arising as predicative parts of speech are part of open systems to which we can at any time add new items by increasing our vocabulary or by inventing a word in order to enable the universe of ideas in tongue to keep pace with our ever-expanding experience of the universe around us.10 The distinction between predicative and transpredicative raises a problem for the auxiliaries, and it will be well to conclude our discussion with a question that must be left open. In light of this distinction, how can we deal with the fact that do, be, and have can be used as full verbs and as auxiliaries? Should we consider that there are two formative elements in tongue, realized as two different parts of speech? After all we find cases like I do do them. I am being careful. I have had it, where they have both roles, transpredicative and predicative. To propose that there are two do s, two be s, two have s in tongue would imply that we were thinking of the word as a unit already existing in tongue ready to play its part in the constructing of a sentence. However if we think of tongue as the state of potentiality for constructing the word itself, as we saw above when discussing “conversion,” it can be seen to provide the formative elements and their operative possibilities ready to produce the unit of potentiality, be it full verb or auxiliary. In this case it would appear that the auxiliary arises as a result of ideogenesis delivering the dematerialized, or rather the “non-materialized,” version of the lexeme to the morphogenesis of the verb, whereas the full verb arises from formatting in morphogenesis the “materialized” version of the lexeme. An analysis of do in this latter sense was presented above (cf. also my 1997c article), but the question can be settled for all three auxiliaries only when we get a clearer view both of the relation between predicative and transpredicative parts of speech and of the process of ideogenesis itself. As in any other scientific theory, a proposed solution to any question, like the analysis of the auxiliaries presented here, serves to raise the next question and thus open the next avenue of research.

12 The Proof of the Pudding The method used by psychosystematics is that used by all the sciences based on observation that have managed to establish themselves as theoretical sciences: back and forth between the concrete and the abstract, not a vague, hazy abstraction, but a clear design, a mental construct of a well finished form. Guillaume (1997, 139)

INTRODUCTION: FROM THEORY TO HARD FACTS

The system of chronogenesis and the system of the three auxiliaries of realization provide hypotheses that are coherent both internally and in relation to one another, so they appear quite plausible. Coherence, both internal and external, is certainly a necessary condition for any valid hypothesis but is not sufficient when the aim is to have a theory that permits us to understand. For Guillaume a valid theory must have an incontrovertible starting point, as we have seen, but must also lead to observationally verifiable results: 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” … The schema of a good linguistic theory is linear. This is a basic requirement. [see figure 12.1](1984, 23–4)

In the case of language – this bears repeating – as in any other human activity, it is inevitable to assume there is a potential behind

190

Language in the Mind

the inevitable (in the position of protagonist and as close as possible to elementary intuition)

antagonistic facts (at the end of the cursus)

Figure 12.1

the actual. (And so we have seen in chapter 10 how, for example, the meaning potential of the simple form makes possible three different actualizations depending on where the event’s realization is intercepted.) From there, the theory must work its way through to what is observable because, as Guillaume liked to point out, citing a dictum of Meillet, “Science lives on proofs, not on truths.” Regardless of how well founded and coherent a theory may be, to gain admittance to the hall of science it must also successfully confront the facts. The facts for Guillaume are examples of real usage in their context, the linguistic expression of some speaker’s intended message in the normal use of language. As a consequence he is not concerned with problematic examples made up for testing grammaticality or usage prompted by artificial conditions such as psychological testing, but rather with the far more demanding task of explaining real usage, what people actually speak or write in ordinary discourse.1 This however means that the job of testing a theory can never be completed because of the remarkable diversity of usage and the unending output of discourse. In this chapter, then, we will examine verb compounds briefly, considering only their most ordinary uses, in order to illustrate the way theory confronts facts. (A more complete coverage of usage can be found in my 2007 study, 178–252.) This is just the opposite process from that illustrated above in chapter 4 on number in the substantive, where we saw how a theory can be arrived at starting from observed usage. Both processes, from particular facts to theory, the discovery process, and from general to particular, the explanatory process, are necessary to establish any theory. One of the main difficulties in the latter process is to distinguish what the particular grammatical form contributes to the “complete thought” expressed by the sentence. Fortunately the well-known practice of contrasting through commutation permits us to observe the distinct meanings expressed by two sentences, one with a verb

The Proof of the Pudding

191

compound and one with the simple form. We will begin with the have + participle compound. THE SIMPLE FORM VS HAVE + PAST PARTICIPLE

The discussion of aspect in chapter 10 brought out that the auxiliary have situates the subject after the event evoked by the past participle, either immediately after or at some later moment. This was diagrammed as follows: past participle

have

The earlier interception expresses an event lasting up to the present of speech, whereas the later one often gives an expressive effect characterized as ‘current relevance.’ The latter effect brings out the fact that the auxiliary situates the subject in the result phase of the event. A good example of this is: Newton has explained the movements of the moon. where, according to Jespersen (1954, iv, 66), the speaker considers that Newton’s explanation “is still known or thought to be correct.” This being a present result for Newton, one can understand why we would not say the same thing of Ptolemy. The simple past could also be used in the above sentence, but it would make one think of an occurrence in the life of Newton rather than its current influence. The auxiliary does not of course make a specific result explicit, so when it is found that in an example like It would seem that great changes have occurred in the style of myths… “the ‘element of result’ does not seem to be evident” (Allen 1966, 142–3); it is clear that the changes are not specified. On the other hand, the necessary result of changes occurring to myths is quite clear: the style of myths is now different. Thus the result phase is merely implied by the position expressed by have and as such may be characterized by no palpable effect – what someone has called “nil results” – as in:

Language in the Mind

192

I have read this book but it has left no impression on my mind whatsoever. On the other hand, granted an appropriate context (such as someone suggesting Let’s ask her), an example like She has lived in China. implies that, as a result of living there, the subject now has a certain knowledge of China, a use that has been called “the perfect of experience.” Turning now to the early interception of the result phase, at its first instant, the transcendent nonpast expresses the result phase of a recent event, one that extends up to the moment of speech, as in: We have always discussed these things on Thursdays. discussed

have nonpast

past Figure 12.2

The repeated discussions are understood to have started at some unspecified point in the past and to have continued up to now, as suggested in figure 12.2. The impression of ‘continuing up to the present’ arises because the subject is situated in time immediately after the last actualized instant of discussed. The example does not specify the result – whether the speaker envisages a change or not – but does indicate that the question of continuing or not is open at the moment of speaking. On the other hand, the simple past in the sentence would have represented a period that began and ended in the past. Similarly, the question Have you run all the way home? might be addressed to someone who arrives out of breath, whereas at a later moment the immanent past Did you run… would be used. Again in

The Proof of the Pudding

193

True to most predictions made last spring, the nation’s overheated housing market has distinctly cooled over the summer months. the sentence expresses the present result of an activity that began early in the summer and has extended up to the moment of writing. Such examples suggest why some grammarians consider that the have + participle compound expresses a “recent past.” However, describing the nonpast transcendent as either a recent or an “indefinite past” as some do is misleading, both because of the tense of the auxiliary and the nature of the transcendent aspect. Moreover, examples like the following are anything but past: You must come and see us when we have found a place to stay. Because of its nonpast tense, have represents the event’s aftermath in the nonpast time-sphere but here its location is further specified by when, which situates it and the event found at some distance beyond the present of speech. That is, both “finding a place” and its result phase are yet to be realized (figure 12.3 depicts this). Only by distinguishing tense and aspect in a verb’s morphogenesis can its meaning in uses like this be satisfactorily explained. found

have

nonpast

past Figure 12.3

In the past tense, the same relation between immanent simple and transcendent compound can be observed. Thus in When they got to the stadium the game started. the starting arises after the arrival at the stadium, whereas in When they got to the stadium the game had started.

194

Language in the Mind

the arrival coincides with the aftermath of the starting. And finally, in reported speech, They told me George had died two days before. the result phase of the dying is contemporary with the telling, whereas the simple past here (died) would indicate that the speaker had represented the immanence of the event at a point prior in time to the reporting. These commonplace examples will suffice to illustrate how the theoretical considerations of the two preceding chapters can explain the difference between the simple form and the have + participle compound. THE SIMPLE FORM VS BE + PRESENT PARTICIPLE

We saw in the last chapter that in the progressive form the participle situates its internal spatial support corresponding to the subject at some midpoint in the event’s realization. As a consequence, an event is represented as incomplete by means of the progressive and so in the nonpast is usually felt to be ‘in progress’ at the moment of speaking, as in the following example from a hockey commentary: He is going behind the net. Here we know where the subject is headed for but there is no indication that he actually gets there. On the other hand, in He goes behind the net. we are told that he gets there. These two verbs in the immanent aspect can be diagrammed as in figure 12.4. Since the progressive situates in time only one moment of the event, that realized during the present of speech, we get the impression of an unfinished action, but with the simple form one has the impression of something that happens so quickly that it is realized from beginning to end before the commentator has finished speaking. The frequent description of the progressive as expressing an event “going on at the moment of speaking” applies here, but, interestingly enough, it applies to the simple form as

The Proof of the Pudding

195

going is nonpast

past goes

nonpast

past Figure 12.4

well because, as is common in sports, the event’s duration is so short it is represented as complete within the present of speech. This contrast between representing activities as incomplete or as complete is quite common in the past. The progressive is often described as expressing a past activity going on when something else happened, as in: When I arrived they were having dinner. Here the arrival is represented as occurring during the meal. That is, the time of the arrival coincides with the stretch of duration in the past represented by were to situate the next moment of having dinner, as in figure 12.5. having were arrived past

nonpast Figure 12.5

The relationship between the two events would be quite changed with the simple form: When I arrived they had dinner. Here had represents the accomplishment of the whole event, beginning to end, and so the arrival is situated not during the meal but just before it, as in figure 12.6. This might suggest that they awaited my arrival to start dinner.

Language in the Mind

196

arrived

had past

nonpast Figure 12.6

Similarly in This morning she was finishing her essay. the sentence does not indicate whether the essay was actually finished, whereas in This morning she finished her essay. we are told that it was, that the activity of finishing was carried to its end during the morning. The grammatical terms for this contrast between complete and incomplete accomplishment of an activity are perfective and imperfective. Although they designate a difference of aspect in Slavic languages, these terms will be adopted here to distinguish how simple and progressive forms represent an event, a distinction that is not one of aspect, as we saw in the previous chapter.2 We will now see, however, that in the English verb these terms designate something more general than the partial or complete accomplishment of an event. When comparing simple and progressive in the nonpast, the contrast between complete and incomplete activities is far less common than the following contrast: He is smoking a pipe. He smokes a pipe. Teaching grammars often describe the meaning of the two forms in terms of an ‘ongoing event’ for the progressive and a ‘habitual event’ for the simple form. While the former expresses the accomplishment of the contemporary moment of the event, as in previous examples, the latter does not express a beginning to end view of the event’s accomplishment: there is no indication at what point in time the habit began or when it will end. All we are told is that the habit exists at the moment of speaking (see figure 12.7).

The Proof of the Pudding

197

smokes past

nonpast Figure 12.7

In this case it can be seen that the simple form does not express the event’s accomplishment from beginning to end and so is not a perfective in this sense. This is in fact a verb representing its event as a settled disposition of the subject, i.e. as a state, and not an activity. A moment’s reflection on this common distinction between activities and states shows that the ‘dynamic’ character of the former involves change and development, whereas the ‘stative’ character of the latter involves no change from one instant or phase to the next of its accomplishment; it is “like-parted” (Quirk et al., 198). This means that a state is a monophase event, consisting of one phase, one set of circumstances, the same throughout its duration, whereas an activity is a metaphase event, made up of diverse phases, differing sets of circumstances, as its accomplishment proceeds. The simple form can represent both the contemporary phase of a state and all the phases of an activity, because in each case it represents in time the total lexical import of the particular event. The simple form can thus be considered a perfective because it represents enough event time – an instant for a state, the total duration for an activity – for the lexeme as a whole to be situated in universe time. The progressive, on the other hand, cannot express a state because, as a monophase event, a state cannot be incomplete, cannot be represented as imperfective. The view that the simple form generally expresses the contemporary instant of a state regardless of how long the extralinguistic happening may exist can be illustrated by ordinary examples like: The Earth revolves around the sun. He looks tired. Other components of the sentence indicate that the first sentence expresses what has been characterized as an “eternal truth” whereas the second evokes only a momentary condition. Moreover the distinction between situations judged by a speaker to be

Language in the Mind

198

changeable, and so represented as metaphase, or as unchanging, and so represented as monophase, can be quite subtle: I’m hoping to finish tomorrow. I hope to finish tomorrow. An example of the progressive in the transcendent aspect will complete this survey of its most common uses. Although it may seem surprising or even inconsistent at first sight, metaphase events represented as imperfective by means of the progressive can also be formed by the transcendent aspect to give the result phase of an incomplete activity. Thus in She has been sleeping since 10 o’clock last night. the auxiliary of the progressive is formed as a past participle, been, and so it coincides in time not with the next-to-be-accomplished moment of sleeping but with its already accomplished portion. This portion of the event extends up to the present, and has situates the subject just after its last accomplished instant, in the present of speech (see figure 12.8). sleeping been has nonpast

past Figure 12.8

Because the event is represented as incomplete, the sentence suggests that it continues at the moment of speech. This and other uses show that the transcendent aspect is found with both perfective and imperfective events. THE SIMPLE FORM VS DO + INFINITIVE

The do + infinitive compound is found in interrogative, negative, and positive contexts, the first of which brings out its makeup most clearly. Ordinary questions such as Do you like pumpkin pie?

The Proof of the Pudding

199

are called yes/no questions because they evoke two options, affirming or negating the realization in time of the event ‘liking pumpkin pie.’ This is made possible by like, which, being an infinitive, represents its event as to-be-accomplished, prospective, and so leaves open the possibility that it will not be accomplished. This double possibility is expressed in the sentence because the incidence of the verb to its subject is suspended, not realized. That is, do auxiliary, like any other finite verb, is formed grammatically to be incident to – to predicate its import of – the support provided by the subject. Here the inversion of finite do and subject indicates that this operation of syntactic incidence is not carried out, so the event’s two alternatives are left open. Thus forming interrogative sentences with do is not to “avoid the inconvenience of the old verb-inversion” (Sweet II, 90), nor simply to have “a ‘bearer’ … of tense-coding” (Ard, 447). Rather, “grammatical factors require an auxiliary” (Langacker 1982, 296), and the analysis in preceding chapters indicates what these factors are. In a positive sentence such as He does like pumpkin pie. there is no inversion, and so the event’s realization is predicated of the subject but usually with an expressive effect of emphasis. This use often arises to counter a prior statement or suggestion evoking the contrary, and so it seems that the speaker, confronted with both options, uses the infinitive here to call them to mind but actualizes only the positive alternative, implying thereby a rejection of the negative one. This implication would not arise with the simple form here (unless suggested by stress or some lexical means). In a negative sentence such as He doesn’t like pumpkin pie. there is no inversion, and so the event’s non-realization is predicated of the subject. This is of course the effect of not (n’t), which exercises its negating effect on the realization alternative, “a denial … [being] always a response to an implied possibility,” as Reid (10) points out. This grammatical role of not can be brought out by contrasting its two uses in the following sentence: They didn’t not like them.

200

Language in the Mind

While not here serves as a lexical import to give the infinitive the sense of ‘dislike’ (Quirk et al., 798), n’t indicates that the event ‘not liking them’ is refused access to the stretch of past time where did situates the subject. The role of n’t here appears to be one of holding up the incidence of the infinitive to its auxiliary. Similarly in the previous example, the ‘liking pumpkin pie’ is not situated in the stretch of time represented by does. In this way, negation by means of do + infinitive declares that, in the stretch of time where the auxiliary situates it, the subject does not accomplish the event. CONCLUSION

This concludes our brief survey of the verb compounds. It is intended to show that the abstract system of the grammatical auxiliaries, combined with the system of the quasi-nominal forms, provides a basis for explaining observable facts of discourse. No attempt was made here to analyze the operations of syntax involved in forming a verb compound. This and a far more extensive confrontation of theory and usage are presented in my Lessons on the English Verb. It has been claimed that “an accurate account of the facts is beyond the scope of … any … theory hampered by its reliance on discrete categorization” (Sag, 85). If this claim is to be countered and the reconstructed systems are to be recognized as valid, the interaction between the theory of incidence and the theory of the verb with its subsystems must be examined in detail. In particular, the fact that the auxiliaries constitute a mental system, a system called on for the meanings its components can bring to the verb compound, has often been overlooked because of the abstractness of these imports, an abstractness that appears to be unsurpassable in the field of the verb, as seen in chapter 11. In this respect the system of the grammatical auxiliaries is comparable to the system of the articles, which will be examined in the next chapter.

13 The Noun Phrase By its very makeup, a word brings its own associative possibilities to the sentence. And so the structure of the sentence appears to be conditioned, and worked out, by the structure of the word. Guillaume (1971, 30)

INTRODUCTION

In chapter 3 it was brought out that in order to speak we establish a first relation between our intended message and language by analyzing the content of our experience into, and representing it through, the sayable units of meaning provided by words. We have also seen how a nonfinite verb and an auxiliary combine to form a verb compound that functions as a single verb in discourse. The theory of syntax involved in this view of the verb compound is basically very simple: the meaning of one word is applied to and integrated with that of another word resulting in a new complex of meaning. Similar processes of incidence are repeated over and over again from word to word, from phrase to phrase, and so on until all the operations of syntax foreseen for the sentence have been carried out. That is to say, as elements of the sentence, as quanta of meaning, words coalesce through incidence to produce the sentence meaning, a sort of fusion in which the elements need no longer be distinguished. The last link in the chain of operations that constitutes the act of language involves re-establishing the relation between the meaning actually expressed and the speaker’s original message. As speakers or writers we may be concerned with the degree to which the meaning actually expressed corresponds to our intended message, and as listeners we are certainly concerned with getting to the message the speaker has in mind. On the basis of the meaning

202

Language in the Mind

expressed, listeners try to work out the speaker’s experience, often calling on other factors outside language itself – what they know about the speaker, the speech situation, and about the world in general – to help them find out what is being talked about and what is being said about it. The point is that the act of language ends as it began – by establishing a relationship between what belongs to language (its mental representations) and what is outside language (experience). At the beginning, the act of language involves the process of representing, i.e. finding the most adequate linguistic correlate of the experience constituting the content of the message, whereas at the end it involves the reverse process of reference, finding the experiential correlate of the meaning constituting the content of the sentence. Linguistic reference is always a matter of establishing a relationship between two mental entities, meaning expressed and experience, the content of the sentence and the content of the message. Whether the speaker’s experience corresponds to a reality outside the speaker’s mind is of crucial importance to the logician in establishing the truth value of a proposition, as it is to a courtroom judge or a psychiatrist. But this is of no concern to a linguist because the same means of representation and expression would be called for whether there is an extramental correlate of the speaker’s experience or whether there is not (as when we describe a dream). Hence the act of language finishes with linking the sentence to the referent as it exists in the speaker’s mind, because every sentence, even questions and commands, correspond to some experiential complex.1 And to carry out this referential function, there is a particularly well adapted construct in English, the noun phrase. THE NOUN PHRASE

We have seen that applying the predicate to the subject constitutes the culminating operation of syntactic incidence in a full-fledged sentence. That is, the meaning import of the verb and all that is incident to it is said of its external support, the meaning import of the noun phrase, which is a representation of what the speaker is talking about. Only when this final incidence has been effected can the resulting sentence meaning be referred to the speaker’s experience. That is, it is through the support provided by the subject

The Noun Phrase

203

that reference can be effected, and so it is by examining the noun phrase that we can gain some insight into this final operation of the act of language, referential incidence. Not that the noun phrase is a sine qua non for reference. Many sentences, from interjections to imperatives, have no noun phrase subject, and each such case must be examined to see how referential incidence is effected. Nor does a noun phrase necessarily trigger reference. When it functions as direct object or as object of a preposition, for example, it does not provide a support for the sentence meaning and so cannot relate it to the intended message.2 Our concern here will therefore be limited to the noun phrase functioning as subject of the sentence, where its syntactic function of support reflects most clearly the internal incidence characterizing the makeup of the substantive. When discussing how a substantive noun names, we saw that the part of speech substantive permits a notional import, its comprehension, to determine its own support, its extension. The lexeme to be formed as a substantive includes what is required to represent the nature of its referent and so implies its range of possible application, which extends to all entities recognized as having this nature. Not that a substantive in usage must always be made incident to the full range of its extension; on the contrary, from one use to another its actual field of application varies. Depending on what the speaker has in mind, a substantive may have as its support different extensities – the full range of its extension, or only a minimal portion of it, or some part of it in between these two extremes. This too we have seen when exploring the system of number, but the expression of extensity is not limited to grammatical number. Guillaume was the first to bring into focus from a grammatical point of view the substantive’s extensity, its variable field of application in different uses, as opposed to its extension, an invariable.3 It is important to have a clear idea of extensity at this point because it constitutes the internal support of the substantive’s import, and so of the lexical import of the noun phrase, which provides the support for the predicate of the sentence. That is, the effecting of syntactic incidence in both the subject and the predicate ends up applying the import of the sentence to that portion of the substantive’s extension actualized by the speaker since this is the formal representation of what is being talked about. This may be clarified

204

Language in the Mind

by the following example, which could be understood in different ways, depending on how the speaker has thought the extensity of the substantive: The beluga is not going to survive. With a single individual in mind, as in an aquarium, the speaker would think beluga with minimal extensity so the sentence expresses concern about one individual. With the polluted conditions of a given area in view, the speaker would think beluga with intermediate extensity so the sentence expresses concern about all the ones that frequent those waters. If the speaker has the whole species in mind, the predicate would be incident to beluga with its maximum extensity so the sentence expresses the probability of extinction. That is, the same word beluga is used in three different senses. There are, in fact, three different intended messages and three different sentence meanings here so even though the same set of words may be uttered in each situation, three different acts of language are involved, resulting in three different sentences, each appropriate to its own intended message. The point of an artificial example like this (artificial because it is unlikely that in actual speech a listener would find it ambiguous) is to bring out what we are not aware of unless it is pointed out: in every use of a substantive, its extensity is determined with regard to the intended message. Another point brought out by the example is that the extensity may vary independently of the number. For each of the three readings, the number expressed by the substantive is ‘singular.’ This shows that the operation involved in actualizing the extensity is distinct from that involved in representing by means of the system of number the space, continuate or discontinuate, occupied by the lexeme, but it is not always easy to discern the difference since both extensity and number are present in every use of a substantive. Fortunately, just as there are signs, -ø and -s, for the operation of assigning a number there are signs for the operation of assigning an extensity: the articles a and the. THE ARTICLES

Grammarians (cf. Quirk et al., 265) have pointed out that determiners (completive pronouns), including the articles, express the

The Noun Phrase

205

extent of reference, the size of the field in the speaker’s message to which the import of the noun phrase is applied. Examples of this are: A tiger can be dangerous. A tiger in captivity is a sorry sight. A tiger is sleeping in the cage. In each of these examples, the substantive tiger is ‘singular’ in number, and yet the noun phrase refers the sentence to quite different situations in the world of experience – taking in the whole species, a subset of it, and a single individual, respectively. The definite article can express a similar variation, as in the beluga example above or in: The tiger can be dangerous. The tiger in captivity is a sorry sight. The tiger is sleeping in the cage. This difference in the scope or extent of reference of the noun phrase is attributed to a variation in representation on the part of the article. At first sight it may appear redundant to have two means of expressing the same extensity. However on comparing the two articles, it can be seen that they express these senses in different ways, whence the usual way of distinguishing the articles as ‘indefinite’ and ‘definite.’ It is pointed out in grammars that in the ‘single individual’ sense, the definite article as in often expresses its substantive’s import as anaphoric, already known, whereas the indefinite article, as in , introduces its import of meaning, making it known. Insofar as the ‘generic’ or ‘universal’ sense is concerned, a tiger as in implying ‘whatever tiger you may think of,’ suggests that the speaker has an individual tiger in mind and that it is typical of any other tiger. On the other hand, the ‘universal’ sense expressed in depicts nothing individual but rather something quite general, inherent to the species. This difference between a and the at their widest scope has been illustrated by means of an example like The tiger is in danger of becoming extinct.

Language in the Mind

206

where the indefinite article would not be used because a species may become extinct, but not the individuals that make up the species. Finally, for an intermediate extensity, the article in expresses a ‘non-specific’ sense, bringing out all the individuals similar to the one the speaker has in mind but not specifying it. The article in expresses what has been called a ‘parti-generic’ sense, evoking a subcategory of tiger. For over twenty years after the publication of his detailed study on the articles in French (1919), Guillaume continued to reflect on the articles. He was confronted with a complex problem: working out the arrangement of their system, explaining their polysemy, and, most important, discerning what in the makeup of the substantive they represent. Once he had discovered the different moments occupied by the three moods in the system of chronogenesis in the French verb, he realized that the two articles occupied different moments in their own system. Uses such as with its ‘anaphoric’ effect suggested a position in the system: the later phase from which one can ‘refer back.’ That is, a precedes the because of the notional chronology involved here: something is necessarily unknown or ‘indefinite’ before it is known or ‘definite’; in a diagram: a

the

In his efforts to explain the polysemy of the articles, Guillaume reflected on number in the substantive, where the same problem arises. This finally led him to see that inherent in the potential meaning of each article is a dynamism, the possibility of a certain operation. He considered this a “new principle” (1991, 68) because up to then he had discerned the dynamism of the system but not that of each component in the system. This gave him the basis for explaining polysemy, since any potential operation can be carried out partly or fully, giving rise to different results, different actual meanings, and so, as we saw in chapter 5 in discussing the partitive pronouns, he was later to extend the principle to all morphemes and grammatical words manifesting polysemy. The following diagram depicts this possible movement in each phase of the article system: a

the

The Noun Phrase

207

To explain what permitted the articles to express a substantive’s extent of reference, Guillaume pointed out that they depicted the momentary actualization of the substantive’s extension, i.e. its extensity. And since this extensity could be equivalent to the full extension or to any part of it down to the least possible extent, he postulated that the first article begins with the maximum (‘universal’), since this necessarily contains the extensity of whatever can be designated by the substantive (a resembling any in this respect), and extends to the minimum (‘singular’). To represent the other leg of the relationship between the extremes of extensity, the second article’s movement begins with the minimum and extends to the maximum. In this way he first conceived of a system (cf. 1999, 131) in the form of what he was later to call the binary tensor (see figure 13.1).

a

the Figure 13.1

Thus Guillaume realized that the support of the substantive’s lexical import is, thanks to the article, expressed outside the substantive. That is, each article’s “matter” (the movement depicted in the above diagram is too abstract to be called a lexeme) serves as a viewing idea for scanning the intended message to find how much of the substantive’s extension is to be actualized. The portion of its extension actualized by the article, its extensity, can then serve as a support for the substantive’s comprehension to be applied to. This is why the substantive’s incidence to the article, to its own extensity in the article, remains an effecting of internal incidence. That is, unlike the adjective, which calls for a support found in another lexeme, the substantive always requires a support, a particular extensity, found within its own lexeme. Without its incidence to the article in the above examples, the substantive’s lexical import would be left with all possible extensities offered by its extension but would have no actual extensity and so would not be able to realize its internal incidence.

208

Language in the Mind

Thus the article is the sign of an operation that must be carried out every time a substantive is constructed for use in discourse because the extensity thus determined provides the support for the substantive’s import of comprehension, thereby permitting its internal incidence. Moreover, even when the noun phrase requires no article or other determiner, the incidence of the substantive’s import of comprehension to its support of extensity is carried out but there is no indication of it outside of the substantive itself. With the article representing the extensity, however, the substantive’s incidence to the article effects its internal incidence in a more visible manner, making explicit the actualization of the substantive’s potential for naming and giving us a means of viewing the internal syntax of the noun phrase. THE OPERATIONAL SYNTAX OF THE NOUN PHRASE

Like the sentence, the noun phrase is a construct of discourse, a set of words assembled through incidence into a single meaning complex. To understand the syntax of a noun phrase, therefore, we try to discern the relation between the words in terms of incidence. That is, since “every lexeme is an import of meaning which must find a support” (Guillaume 1990, 122), it is a matter of working out the way each adjective’s and adverb’s import of meaning finds its support of meaning in the substantive’s lexeme, which is then made incident to its own extensity. Here, then, we have something similar to the syntax of the verb compound outlined in my 2007 study, where it is shown that the nonfinite verb first represents the event, the auxiliary then represents its own place in time, and the event is finally made incident to it, thereby completing its formal determination. We will examine4 a typical noun phrase such as the subject in: An old tiger is sleeping in the cage. It is the lexical head, the substantive tiger, which is first called on to provide its import of comprehension formed by gender, number, and case but whose incidence to its own extensity is not yet effected. This import may then be complemented by calling on, for example, the import of one or more adjectives and making these notions incident to the import of the substantive until the

The Noun Phrase

209

meaning complex is judged adequate as a representation of the experiential entity in the intended message. Finally, the extensity of the substantive is represented by the article, and the total import of meaning is made incident to it. Thus for a simple noun phrase like the one above, the order in which the elements are represented, the order of construction in the mind, is the opposite of the order of pronouncing them in speech. This sequence of operations, actual and virtual, can be detailed as in figure 13.2. In the case of each word, the process of psychogenesis forms its mental import, which can then be made incident to a support. 1. The psychogenesis of tiger, bringing to mind its import of comprehension formed by gender, number, and case, but not yet made incident to its own extensity: psychogenesis of tiger

‘tiger’

2. The psychogenesis of old, formed as an adjective, i.e. with external incidence (depicted here by the dotted arrow as a prospective operation): ‘tiger’

psychogenesis of old

‘old’

3. Effecting the incidence of ‘old’ to ‘tiger’: ‘old’

‘tiger’

4. The psychogenesis of an to represent the lexeme’s minimal extensity and the substantive’s prospective incidence to it: ‘oldtiger’

psychogenesis of an

‘an’

5. Effecting the incidence of ‘oldtiger’ to ‘an’: ‘an’

‘oldtiger’

6. The resulting meaning complex with its minimal extensity, constituting the noun phrase: ‘anoldtiger.’

Figure 13.2

210

Language in the Mind

The important point brought out by the diagrams in figure 13.2 is that the noun phrase as a unit of discourse depends on the internal incidence of the substantive. That is, applying the substantive’s import (its own comprehension completed by that of the adjective) to a portion of its own extension is realized here by representing that portion outside the substantive and then effecting the incidence to it. In this way, the link between substantive and article is a grammatical one bringing to completion the morphogenesis of the substantive, and yet it remains an internal incidence because ‘tiger’ – unlike ‘old’ – is applied to a part of its own range of application, the part represented by the article. This summary description brings out the fact that the accumulation of meaning effected by the successive operations of incidence and resulting in the noun phrase is not paralleled by an accumulation of signs. Unless each word’s sign remains sufficiently distinct in discourse, it cannot play its semiological role of prompting the listener to undertake the appropriate operation of lexigenesis to call to mind the particular word’s meaning. On the other hand, in actual speech, in order to effect the liaisons involved in even a simple noun phrase like this (cf. an not a, and the manner of actualizing final [d] before initial [t]), the speaker must already have in mind the next word’s potential sign. This foresight is made possible because the noun phrase is constructed mentally as outlined here: by calling to mind the meaning imports (psychogenesis) in the order opposite to that of actualizing the signs (semiogenesis). The above description also brings out how the article represents the substantive’s extensity as the support of its lexical import. In this way the substantive’s internal incidence is made grammatically explicit whenever there is an article. When there is no article, as in Old soldiers never die. the substantive’s internal incidence remains implicit, effected within the substantive as the final operation in its morphogenesis. Thus we can understand why Guillaume considered the article a sign of the actualizing of the substantive, a means of signifying in discourse one of the extensities made possible by the substantive’s extension in tongue. Furthermore, this extensity is always equivalent to the extent of reference of the noun phrase, reflecting how wide or narrow a field of application it has in the speaker’s message.

The Noun Phrase

211

Other completive pronouns also provide a support for the substantive within the noun phrase since they all represent the space occupied by its designatum in the intended message. They do not, however, represent this space in terms of the lexeme’s extension in tongue, as its extensity. We have seen that any and some represent the extent of reference as a part of some whole. The demonstratives depict the space occupied by what is being talked about by situating it in the intended message. Much remains to be explored in the psychosystems of pronouns, both suppletive and completive, and Guillaume (1991, 79) provides the basis for this work of analysis when he remarks: It has been pointed out on several occasions in the preceding lessons that, in the depths of the mind where they reside in permanence, ready to serve when needed, when discourse calls on them for its own construction, the most important grammatical words of the language represent not a position, but a movement. That is to say, the meaning linked to them, that they convey, is not static in character but kinetic. This kinetic character of grammatical words is the way they are in tongue. The conversion of this kinetic character, of the signified movement, into a position occupied by the mind within the said movement is their state in discourse.

This is why a grammatical word like an article does not make explicit the particular extensity it represents when used in a noun phrase.5 Rather it makes explicit for the listener the fact that it signifies a movement, the position of this movement in the system and the form of this movement, and finally the fact that the movement has been intercepted at some point during its development. Thus, in the case of grammatical words, passing from tongue to discourse involves passing from the kinetic to the static. This explains why the kinetic character of grammatical words has not been recognized up to now. This character has in fact something special about it in that it never surfaces completely without restriction into discourse where movement, which the grammatical word symbolizes, invariably appears intercepted by a crosscut and finally immobilized at a certain instant of its development, at the point it has reached. (Ibid.)

Enough has perhaps been said to show that the relation between substantive and article is not exactly the same as that between substantive and determiners like some. And this suffices for present

212

Language in the Mind

purposes, since the aim is simply to give an idea of the diverse ways the operations of incidence can be effected so that the resulting noun phrase can provide a support for the predicate and a means for carrying out referential incidence. In the next chapter we will examine how the noun phrase in carrying out its function of subject helps determine the way the verb actualizes its external incidence. In this way we can get a first view of how the “complete thought” of the sentence is constituted.

14 Concord, Discord, and the Incidence of Verb to Subject Language … is a commutation, operated in thinking man, of his thought at a given moment into speech. Guillaume (1995, 305)

INTRODUCTION

The noun phrase was presented in the last chapter as a unit consisting of the imports of its component words integrated by means of the syntactic possibilities of each word. One of its main sentence functions is that of subject, i.e. support of the predicate’s import. The incidence of predicate to subject is operated by means of the verb through its external incidence, a process traditionally called agreement or concord. In English this commonplace syntactic operation offers some surprising possibilities that have been neglected by many approaches to syntax. For the great majority of grammarians and linguists the phenomenon of subject-verb concord in English can be summarized as follows: a verb in the nonpast of the indicative agrees in number with its subject in the third person, expressing this agreement by the presence or absence of the suffix -s. This description applies to the vast majority of sentences with a verb in the nonpast and so provides a convenient syntactic rule for teachers of English since it entails the agreement of the verb with the substantive, which is inflected for grammatical number. The cases where it does not apply, not very frequent, are generally considered marginal, or requiring other rules. In the last chapter we saw that the syntax of the noun phrase led to the integration of its various imports and the constitution of a

214

Language in the Mind

new unit of meaning that is the support for the predicate. That is, the subject of the verb is the noun phrase, which integrates the words contributing to it so that they are no longer considered as separate grammatical units. This view of the noun phrase is therefore hardly reconcilable with the rule making the verb agree with the substantive. To bring out that this rule is not adequate as a basis for analyzing the syntactic relationship involved here (though it may be useful for teaching purposes), we will examine a number of cases where it does not apply. The aim here will be to show that the verb agrees not with the grammatical number of the substantive but with the import of the noun phrase, and this both for the cases where the syntactic rule does not apply and those where it does. In other words, it will be argued that verb agreement is not an automatically applied rule but a meaningful operation of syntax since the role of the predicate is to say something about the subject. FACTS TO BE EXPLAINED

Among the cases that do not reflect the syntactic rule, the most common type is probably that of collectives as in: This company are superbly managed and their products will continue to be in great demand.1 To be noted here is that company, with no -s inflection and accompanied by the demonstrative this, is clearly singular. The verb are, however, is the form found with plural substantives. In this example, the speaker could have established singular agreement for the verb, but the meaning of the sentence would not have been quite the same. A similar case is found in the following example, from Quirk et al. (758): The audience were enormous. Collectives like company and audience all express the idea of a group and so call to mind both the impression of “one” (the group) and the impression of “many” (the members making up the group). Agreement will depend on which impression dominates at the moment of speaking, as the following pair from Reid illustrates: This afternoon our panel are three male singers.

Concord, Discord

215

Our panel of opera experts today is certainly a lively one. In the second type of use the substantive also evokes a group, but here the members of the group are explicitly evoked. The following two examples illustrate the two possibilities: A combination of audiotapes, slides, and a specially prepared study guide provide the core for much of the learning activity in the course. A combination of rising wages and excess profit margins has put prices up. In it is the impression of the whole, of the combination of the two elements working together, that dominates and calls for has, whereas in it is the impression of the different elements making a contribution that is dominant and calls for provide. As support-substantive we find not only collectives like bunch of, group of but also, according to Juul, expressions like set of, majority of, whole of, kind of, whole succession of, spate of – all of which let the speaker bring out by means of verb agreement either impression, that of the whole or that of the constituent parts. The following is another example of this type: An increasing percentage of Negroes as compared with Whites are gaining employment. The third type involves a phrase of the same sort, but instead of a substantive a quantifying pronoun provides the support and designates the whole. For example in Each of you own $100. the pronoun each is grammatically singular, but it evokes a set of entities. Similarly, in the following example, because of any, the support-pronoun one calls to mind a multiplicity of entities: Any one of these claims are enough to overwhelm citizens of other rising cities in the South. In these examples the speaker could have used owns and is to bring out an individual as typifying all. Similarly, in

216

Language in the Mind

Among girls only 1 in 5 were pursuing any form of further education. we understand a number of individuals depicted by 1, whereas with was it would suggest simply a proportion. Again it can be seen that the choice offered permits the speaker to express either of the impressions proposed by a complex intended message. In the fourth type of use, the noun phrase has a substantive with a quantifier, as in: The last few years has seen the advent of enclosed antique supermarkets. Thanks to the agreement in the singular here, there is the impression that years depicts one stretch of time, whereas with agreement in the plural (have) one would have the impression of successive years. Between the following examples, the same difference can be discerned: Another 4ft are being added to the 20ft-high west walls of Canterbury prison. … and a further 12,000 miles has done little to dampen my regard for the car’s many excellent qualities. Where gives the curious effect of multiplicity, of adding to the walls foot by foot, simply evokes the total quantity. In the following examples the effect of a single quantity comes out clearly: Three rooms is enough. Two drops deodorizes anything in your house. Granted the conversational situation in which this example arose, it would even have been out of place to use are. Finally, a curious case: … the continual danger both this nation and the world faces. One wonders if the verb reflects the speaker’s view of the world as constituting a whole containing the nation.

Concord, Discord

217

In a fifth type of use we find the same relation between the parts and the whole, but there is no expression of quantity. Thus in Extra teachers is the next major item. even though the substantive is plural, there is a clear impression that the noun phrase expresses the notion as a unit, perhaps in the sense of ‘the problem of…’ Similarly in High costs, the report also finds, inhibits reform by way of small classes. Juul (162) brings out the unitary impression by paraphrasing the subject as ‘the fact that there are high costs.’ Likewise, where the idea of singular dominates, it has been said that there is “notional concord” in: Good manners is a rarity these days. Of interest here is the fact that the impression of a single entity can be dominant even when the noun phrase is reduced to an adjective and a plural substantive. In the following example the noun phrase is reduced to a minimum, but even here we must distinguish between the grammatical import of the substantive and the support provided by the noun phrase for the predicate: Results is what I want. The next sort of use consists of sentences with a multiplicity of lexical imports calling for coordinated substantives in the subject: Bread and butter are nourishing. Here the two substances are said to be nourishing, whereas in Bread and butter is nourishing. the combination is evoked. There are variations in the relations involved here (cf. Quirk et al., 759ff.), but in all cases of singular

218

Language in the Mind

agreement one gets the impression that the different entities designated by the substantives constitute a single reality in the speaker’s mind. The following example, taken from an ad in a periodical, brings this out clearly: The Game Parks, the palm-fringed beaches and the immense business potential of East Africa is now less than half-a-day from London. The author very effectively evokes a single region through its most attractive features (though in a later number of the periodical the verb had been “corrected”). The same sort of grouping effect in the noun phrase is brought out by the following example from Reid: Galloping horses and thousands of cattle is not necessary to cinema; I call that photography. The last type of use to be mentioned is that where the subject consists of a clause introduced by the pronoun what. In cases where only one impression, whether of unicity or of multiplicity, is involved, no variation of agreement is possible. Thus in What appear to be large windows in the second story are glass heat collectors. ‘multiplicity’ is the only impression suggested by the what clause, so it would not be possible to replace are by is. On the other hand, in What interests and a little confuses me are the names he chose to call. the author could have used is instead of are, and here, as in other such cases where agreement is variable, one gets a double impression of unicity and multiplicity. It is a matter of emphasis, as though suggests ‘the names he called constitute a subject that interests and confuses me,’ whereas with is it would imply ‘the subject that interests and confuses me consists of the names he called.’ It may even happen that the speaker hesitates between the two impressions, as in the following example:

Concord, Discord

219

What they want us to tell you about is – uh – are these things. This is not intended to provide a complete list of the types of subject with which the “principle of grammatical concord” (Quirk et al., 757) is not respected, because the verb does not reflect the “grammatical marker” for number in the substantive. Examples like the last one above suggest that this phenomenon is more frequent in speech than in the “more inhibited medium of writing” (ibid., 758). In any case, the studies from which the above examples were selected give enough data to show that this kind of usage cannot be ignored as incorrect or abnormal or deviant. In fact, it poses the problem of the fundamental relation in the sentence, the operation of incidence between predicate and subject, by inviting a comparison with “normal” usage in order to find what they have in common. LOOKING FOR A FACT THAT EXPLAINS

It is obvious that a syntactic rule that does not account for the above uses must be considered inadequate as an explanation of the data. A more interesting, though inadequate, attempt at explaining the phenomenon is the “proximity principle” or “attraction,” which maintains that the verb is attracted by a substantive situated closer to the verb than the substantive-support. From this it follows that in, for example, the plural agreement of are is due to the attraction of Negroes and Whites. Likewise in it would be argued that we find are rather than is because of the attraction of names. Although there does seem to be something valid in this approach, as we shall see below, it is not an adequate explanation if only because there are cases – and among others – where it does not apply. In seeking an explanation for the last type, where the subject is a what clause, the following question was raised: should the clause itself be considered grammatically plural or grammatically singular? No grammarian has, to my knowledge, proposed this as a possibility, and one can understand why: there is no semiological indication, no inflection, in a clause that would support such a claim. Furthermore, to maintain that a noun clause varies in grammatical number would imply that it involves a morphogenesis like

220

Language in the Mind

that of a substantive or pronoun, and this would not be tenable since, as we have seen above, morphology precedes syntax in the constructing of a phrase or clause. There are no grounds for considering that a clause, or a phrase for that matter, is declined for grammatical number. As a last resort, confronted with the impossibility of finding a general rule for attributing every case of verb concord to some other element in the sentence, some scholars have proposed that verb agreement is weakening in English, that it no longer holds for all uses. In support of this, one could recall that there is no longer any evidence of agreement when the verb is in the past tense or the subjunctive mood (cf. my 2007 study of the verb). However it remains that, for the great majority of sentences with verbs in the nonpast tense, agreement must be accepted as a reality since no variation is possible. Without some description of the precondition making variation possible and what makes it impossible, this explanation is also inadequate. Other than these three, the only explanation that has been proposed is an appeal to the “notional principle,” “agreement of verb to subject according to the notion of number rather than with the actual presence of the grammatical marker for that notion,” as Quirk et al. (757) describe it. But here too it has been applied to only a few uses and does not include all the commonplace uses where the problem of agreement does not arise. That is to say, none of the explanations proposed is sufficiently general to account for all uses. This suggests that verb agreement has not been envisaged as a sign of a general phenomenon, the incidence of the predicate to its subject, including all the particular cases that appear as exceptional or abnormal if one adopts a more limited view of agreement. In the various types of usage examined above, agreement certainly seems to be linked to the notional content the speaker wants to express. In each case of variable agreement, there is a double impression – ‘unity’ and ‘multiplicity’ – in the intended message. That is, the notional principle of agreement, constructio ad sensum, as some grammarians put it, does hold, and there is certainly nothing surprising in this. After all, if we accept Guillaume’s view that an act of language commutes the speaker’s momentary thought into something said, what is more obvious than to propose that what one wishes to express, the intended message, helps determine the most important syntactic link in the sentence? Moreover, if this

Concord, Discord

221

argument is valid for those sentences where agreement is variable, why can it not be just as valid for all other sentences, those where agreement is not variable and poses no problem? This point of view will be adopted here: in every case verb agreement is motivated by the notional import of the noun phrase and not merely the grammatical import of the substantive. To see if this point of view helps us understand the agreement of the verb better than the other approaches, we will try to discern when and how the incidence of the predicate to its support takes place. To explain the phenomenon, to see what determines agreement, we cannot invoke what the sentence says because the message resulting from the sentence arises after the agreement has been effected, too late to explain something that took place during the construction of the sentence. We will look rather at what happens before agreement is effected, at how the intended message or content of experience – which, as we have seen, precedes and triggers the act of language – is represented. And in the constructing of the sentence, it appears necessary that the subject be constructed before the verb’s agreement with it can be effected. We have seen that much more is involved in constructing the noun phrase than simply the psychogenesis of the substantive with its lexeme, its grammatical number, etc. The various imports of adjectives and adverbs must be made incident to the substantive’s lexical import. The incidence of this totalized import to the substantive’s extensity (usually represented in a determiner) to effect its internal incidence makes referential incidence possible. All this must be carried out in order to complete the noun phrase. That is, to have a verb agree with its subject, speakers must already have represented (but not necessarily expressed) by means of a noun phrase that portion of the intended message they wish to talk about. The above discussion of examples has brought out that the support provided by the subject is the global meaning of the phrase including both what prompted the grammatical representation of number in the substantive and other experiential impressions involving quantity, if any. It is this resultant meaning, the synthesis of all these impressions as represented in the noun phrase, that constitutes the support of the predicate and motivates agreement. This manner of understanding agreement obviously permits us to explain ordinary agreement in sentences where the grammatical number of the substantive reflects as either ‘continuate’ or

222

Language in the Mind

‘discontinuate’ the only impression of quantity inherent in the global notion of the noun phrase. It also permits us to explain the agreement observed in the various uses discussed above, where contrary impressions of quantity are discernible in each case. Furthermore it suggests that once the construction of the verb is complete as a part of speech permitting its incidence to an external support in a subject, the speaker establishes this external incidence in view of the only impression of quantity in the noun phrase or, where there are conflicting impressions, with the dominant impression at the moment of establishing incidence. This gives a view of agreement as a single phenomenon whether or not it involves discrepancy between the verb and the substantive’s “grammatical marker.” It also permits us to understand why grammarians have proposed the principle of proximity: when the substantive in the noun phrase calls to mind a group with its impression of unity (cf. Combination in and percentage in ), the substantives intervening before the verb call to mind the elements making up the group and so bring out the impression of multiplicity also inherent in the global meaning of the phrase. Likewise in , it is the substantive names that is supposed to have attracted the verb; in fact, if our approach is valid, it is not the substantive itself but what it corresponds to in the global meaning of the what clause that calls for plural agreement. We can also understand spoken sentences like where evidently the speaker, confronted by the noun phrase with a double impression, changed his mind while expressing the incidence of the predicate to its support. Finally, we can understand a type of agreement not mentioned above, that where the same noun phrase provides a support for two verbs, one with singular agreement, the other with plural agreement: The company market Vogue and Butterick paper patterns and, in addition, through a subsidiary markets sports wear. The committee says that the proposals for the first cycle depend on the acceptance of the diploma as a valuable qualification, and say “no award can acquire high prestige overnight.” The Copenhagen school, and particularly Hjelmslev’s glossematics, is generally very poorly known in this country, and thus, no doubt, deserve a volume of their own.

Concord, Discord

223

In such cases, what changes as the constructing of the sentence proceeds is the dominating impression of quantity involved in what is being talked about. That such cases should arise is anything but surprising, especially in the case of written discourse with more lengthy sentences. There remain, of course, uses that have yet to be analyzed, such as the following from Quirk et al. (758): More than one member has protested against the proposal. Many a member has protested against the proposal. Only after an analysis of the quantifying expressions involved here, more than and many a, can such uses be examined in the light of the general hypothesis of verb agreement outlined above. Similarly, it has yet to be explained why a noun phrase with each and every, which call to mind all the members of a group, is rarely found with variable agreement. Only after it is shown how their “distributive effect” differs from the effect of collectives examined above can this problem of agreement be explored. Considering verb agreement as a process of making an import incident to a support has the advantage of viewing it as a particular case of a more general phenomenon, the incidence of the meaning of any word or group of words to that of another word or group of words. Agreement is not therefore simply the effect of applying a particular rule but another manifestation of the activity inherent in any act of language, that of saying something about something. That is, we thereby situate this grammatical question within the general situation of anyone undertaking an act of language, of constructing a sentence in order to commute their intended message into something said. That English should have managed with its meagre inflectional resources to render subtle experiential impressions in this way is a noteworthy achievement. We can best finish with two examples of this syntactic resource that illustrate the finesse and inventiveness with which it can be used. The first is the title of a book: How a People Die. Although the subject designates an ethnic group, one of the native peoples of Canada, thanks to plural agreement the verb attributes

224

Language in the Mind

dying not to the group but to the individuals in the group, one by one. The second example is the title of an article echoing a wellknown advertising slogan: Hey America, Coke are it! The capacity of a language that permits us to represent a popular soft drink as a team and express in this way the novel situation of marketing it in two different versions cannot be reduced to the application of a grammatical rule.2

15 Thought and Language … I pitied thee, Took pains to make thee speak, taught thee each hour One thing or other. When thou didst not, savage, Know thine own meaning, but wouldst gabble like A thing most brutish, I endowed thy purposes With words that made them known. The Tempest (I, II, 353–8)

CAN WE THINK WITHOUT LANGUAGE?

Like some of the subjects touched on in preceding chapters, the thought-language relationship has so many ramifications for a theory of language in the mind that I can do little more here than suggest several avenues for reflection. We can perhaps best approach such a fundamental topic through a question sometimes raised by students: can we think without language? By this they often have in mind language not just as the spoken word but also as that inner dialogue where words emerge into consciousness without their signs being physically realized. The fact that both a negative and a positive answer to the question can be justified indicates that there is a possible ambiguity in the question. A moment’s reflection suggests that it resides in the sense we give to the verb think and its substantive thought. To make our discussion fruitful, therefore, it will be necessary to call to mind different acceptations of these words. We can begin by considering the word thought as it is used in the traditional definition of a sentence: a “complete thought” clearly designates what is represented by words and assembled by incidence to form the sentence meaning. In this resultative sense, thought obviously cannot arise without language (or some language surrogate)

226

Language in the Mind

in the form of discourse. With this in mind, we can understand Shelley’s remarkable couplet from Prometheus Unbound: He gave man speech, and speech created thought, Which is the measure of the universe. (II, 4, 72–3)

Thinking in this sense may be understood as a matter of manipulating concepts called to mind by words or other symbols, or “as a discourse that the mind carries on with itself about any subject it is considering,” as Socrates put it (cited in Vendler, 99). It is this sense we find in common expressions like “I couldn’t hear myself think,” “thinking out loud,” and “to read my thoughts,” and in Oscar Wilde’s remark that we should consider language as “the parent, not the child, of thought.” We can depict the relationship involved in such language-dependent thought as follows: language

thought

Besides designating reflective processes of this sort, however, think can be used to evoke being aware of anything in our momentary experience, just as the substantive thought can be used to designate what we are aware of. We find this sense in “How can I know what I think till I see what I say?” and in William James insisting on the continuity of consciousness: “In talking of it hereafter, let us call it the stream of thought, of consciousness, or of subjective life” (cited in Bartlett, 648). Northrop Frye (1124) describes thought as “a process stumbling through emotional entanglements, sudden irrational convictions, involuntary gleams of insight, rationalized prejudices, and blocks of panic and inertia, finally to reach a completely incommunicable intuition.” In this acceptation, then, we can think without discourse, we can have something in mind without the conscious intervention of language. In fact, as we have seen, thought in this sense is a prior condition for language because without this experiential content framed as an intended message to prompt a speaker, no act of language would be undertaken since there would be nothing to represent, nothing to talk about. And so we can also speak of language as the child of thought, a relationship that can be depicted as follows: thought

language

Quite obviously the relation should not be considered a simple one since thought can be seen both as a precondition for the

Thought and Language

227

exercise of language and as a result of it. In fact, we are led to see this relationship as binary, its two versions involving a succession in time, as in the following schema: thought 1

language

thought 2

Although it might involve some oversimplification, in order to distinguish that portion of our stream of consciousness constituting the intended message from the sentence meaning and message we get as a result of discourse, one is tempted to speak of pre-language thought and post-language thought. And between the two there is the mechanism of commutation, the viewing ideas and systematized mental processes of language provided by tongue for translating the one into the other. Of course the latter, the thought already expressed by someone else (thought2), sometimes for me becomes the former (thought1), a content of consciousness I might want to talk about, as in a lively discussion. It remains, however, that every time we speak, this two-phase relation comes into play because speaking converts what is unsayable into something said. Thanks to its possible ambiguity, the question raised by students thus brings out the binary operation involved here and throws light on some of the problems raised in discussions concerning the thought-language relationship. The question is not raised just by students. For example, this is the way Einstein (1981, 327) poses the problem: “What is it that brings about such an intimate connection between language and thinking? Is there no thinking without the use of language, namely in concepts and conceptcombinations for which words need not necessarily come to mind? Has not every one of us struggled for words although the connection between ‘things’ was already clear?” It is not, of course, given to everyone to have the “things” in their pre-language thinking “already clear.” Probably most of us have to “see what we say” at least through inner dialogue to be able to focus clearly on what is in our stream of consciousness. Quite obviously, the thoughtlanguage-thought relationship is not realized in the same way by all speakers.1 Nor should the expression of what we have in mind be limited to the means provided by language, as the composer Arvo Pärt makes quite clear: “Talking about my music traps me in a vicious circle and it’s very seldom that I manage to escape it. If I’m writing a new piece then I mustn’t talk about it because if I do then I have no impulse to write it any more. Once it is written, then there is nothing left to say. That’s very apparent to me. It’s a

228

Language in the Mind

matter of thinking in music, and I hope my music finds a direct way to the listener without any further explanation.”2 Keeping the binary character of the relationship in mind helps us understand what is going on here. It also leads us to see that the converse question – can we speak without thinking? – would probably give rise to little disagreement. The reason for this will lead us to consider the word thought in another of its acceptations and bring out other aspects of this complex relationship. CAN WE SPEAK WITHOUT THINKING?

Provided one does not simply dismiss thinking as an illusion or an irrelevant epiphenomenon, to ask if we can speak without thinking would bring forth a negative answer. Whether we consider thought as that which language represents, or as that which language expresses, or both, we assume some thinking process is required to produce language. That is, regardless of the particular content of thought required for speaking, the processes of representing and expressing involved in languaging are themselves mental operations, processes of thought instituted for the very purpose of carrying out an act of language. And so language without thinking is quite inconceivable. Indeed, that such operations constitute the major part of the act of language appears to be an incontrovertible fact, a fact that provides the basis for the Psychomechanics of Language and its attempt to discern what kind of thinking processes are carried out every time we speak. For this reason, Guillaume insisted on the importance of psychosystematics, the study of particular systems in tongue, as a branch of linguistics because it is only through an adequate analysis of particular morphemes and their systems that the linguist can discern the mental operations that make them functional. Having analyzed the systems of number and of the article, he was struck by the fact that the operations involved in them have a similar form. Work in English has shown that the same type of binary operation, involving a double movement from greatest to least to greatest, is implemented in certain systems: that of number (genericsingular-generic), that of the articles (maximum-minimummaximum), that of any and some (whole-part-whole). Thus it appears that these systems are based on a similar underlying mechanism. We have also seen that this mechanism, the radical binary tensor, is

Thought and Language

229

fundamental for producing words in English, all of which are formed through the processes of ideogenesis and morphogenesis: ideogenesis, a particularizing movement discriminating or “discerning” the specific lexeme required from all those latent in tongue, followed by morphogenesis, a generalizing movement that grammaticalizes the lexeme, situating or “comprehending” it in the most general system of the language by giving it a part of speech (see figure 15.1). This operation of psychogenesis, implementing the binary tensor mechanism to discern a specific notional matter within the universe of ideas and then subsume or comprehend this matter in a general system of grammatical forms, is activated every time we produce a word in English, as we have seen.

Figure 15.1

Confronted with findings of this nature, Guillaume traced this mechanism back to its roots in the mind and identified it as a reflection of the very mental capacity that characterizes humans. “This mechanical substratum that the confronting of the category of the article and the category of number brings to light is nothing less than the mechanism of potency for thought: the mechanism giving thinking man his double aptitude to particularize and to generalize alternatively, an aptitude without which, in that thinking being, man, there would be no human thought” (2003, 69). Thus he speaks of “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” (1984, 118). For this reason he calls it an “intuitional” mechanism, suggesting thereby that it is inherent in the very makeup of the mind and comes into play whenever the mind confronts the universe of experience. All this brings us to another sense of the substantive thought: ‘the capacity to think,’ i.e. the faculty, a sense it does not share with

230

Language in the Mind

the verb. With this sense in mind, we can discern yet another aspect of the thought-language relationship. LANGUAGE AND THE FACULTY OF THOUGHT

As a faculty or capacity, thought is a sine qua non of language, as it is of all thinking. That is, it constitutes the potential lying behind and permitting not only pre-language thought and post-language thought, but also the mental operations involved in language ensuring the commutation from the one to the other. But Guillaume went even further when he proposed that this intuitional mechanism reflecting the faculty of thought is instrumental in instituting a new system in a language. The establishing of a system is based on primordial principles inherent in the human mind… 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. (1984, 107–8)

The idea that tongue is founded on principles inherent in the faculty of thought itself is what leads him to advance the hypothesis that all its subsystems have the same general form. Even more, he considered that all the Indo-European languages belong to the same type because they have a common structure. That is to say, although their various subsystems are architectured differently, each of these languages has the same general word-forming structure, namely the part of speech system. And in his theory of language typology, Guillaume argued that the same basic mental capacity provides a different structure for each type of language. Implemented differently according to the type of language under consideration, these same primordial principles provide an underlying structure of tongue, a mechanism for word formation, in every human language, a structure that, in languages that do not have the part of speech system, is quite unfamiliar to native speakers of English. At first sight, this hypothesis may well appear unduly speculative and perhaps even no more than wishful thinking. On the other

Thought and Language

231

hand, when we examine what led Guillaume to such a proposal we can see that it is not to be rejected without serious consideration. When he refers to the “obviously well-founded idea” that the mind derives its power from the binary capacity to generalize and to particularize he is simply expressing in operational terms a commonplace among those who have thought about human nature: that humans are rational animals, capable of abstract thought, of conceptualizing (which implies the capacity to generalize and particularize), etc.3 Furthermore, it is one of Guillaume’s major contributions to have analyzed certain systems and demonstrated that this human capacity, conceived of as a binary mechanism, is manifested in them. As a consequence, when his background in comparative linguistics with its emphasis on the physical sign permitted him to see that the universal-singular relationship is “fundamental to the structure of tongue” in the Indo-European languages (cf. 1984, 38–42), that all our words are formed on this basic pattern, and when he found the same relationship, implemented differently, in quite different languages (Chinese, Basque, Arabic, etc.), it is not surprising that he was led to the hypothesis that it is the mind that institutes systems in tongue and that these systems mirror in different ways the underlying mechanism of the mind. Guillaume’s daring hypothesis concerning the ultimate source of the form of language systems constitutes the last link in the causal chain we have followed up because it is a source outside language, inherent in the mind. As such it throws a new light on the relation between thought and language. It permits us to conceive of the faculty giving rise to language, not as a specific language-producing ability, but as the faculty of thought dealing with “the problem of language, a problem common to the whole of humanity” (1984, 140). At any given time in any given language community, our human faculty of thought attempts to institute and make use of the best means available for translating the unsayable into something said, for carrying out the activity of producing words and sentences in order to “express complete thoughts.” For this reason Guillaume’s theory is known as the Psychomechanics of Language – in order to keep its field distinct from that of any other human activity calling into play the mental operations made possible by the faculty of thought. Viewing the relation between language and thought in this way – the speaker’s mind imposing its own means of organizing and

232

Language in the Mind

systematizing on language – implies, as we have seen, that historical development is not completely haphazard. The speaker’s mind takes advantage of anything among the accidents arising in diachrony that will permit it to implement its own systemic view. This presupposes that the mind is active even when not engaged in an act of language: “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” (1984, 65). This “deep musing of the human mind” goes on far below the level of consciousness, “at the level where thought recollects itself” to systematize the means of grasping, of representing, the stream of conscious experience in a constant effort to make it ever more readily languageable. This “ceaseless working of thought” provides solutions to problems of representation. “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” (1984, 153). However no solution can ever be definitive since language can never provide a representation that is completely adequate to all that is involved in a given intended message. “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 infinitum” (1984, 65). This brings into focus once again the relation between language and human experience but from a point of view that will permit us to discern a new facet of it. LANGUAGE AND EXPERIENCE

We have already seen that the purpose of a sentence is to express a representation of some particular experience, some content of awareness arising from perception, memory, imagination, etc. However, even when not engaged in an act of language, the preconscious mind is ceaselessly active, scanning its experiential content through the viewing apparatus provided by tongue. That is, Guillaume considered an ordinary human experience, such as seeing someone or something, not as a product of physical perceiving alone but as a percept brought into focus by the notion in tongue, the potential meaning, best corresponding to it. A human

Thought and Language

233

experience is thus situated by what he calls the “viewing” ideas of tongue; the raw material of a percept is reframed by our inner universe of ideas so that we know what we see. Speaking of “the language problem constantly confronting man” as “a problem of viewing,” he remarks (1984, 141): 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 universeview 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 [for example] 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.

Contrasting this “mediate” view of the universe with the animal’s “direct” or “immediate” view, as he calls it (1984, 142), where no “prior mental representation” intervenes, Guillaume points out that the discussion has “moved from the field of linguistics to that of philosophy” and so carries it no further. It is nevertheless valuable to be able to grasp this conception of tongue as intervening in the ordinary process of vision to provide a frame or format for what is perceived.4 In Guillaume’s eyes, thought without this unconscious intervention of language would be turbulent. 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. (1984, 144)

If the very fact of possessing a human language involves framing raw experience in terms of the concepts, the viewing ideas, of tongue, one can understand why Guillaume saw language as a

234

Language in the Mind

means of reducing the mental turbulence involved in our constant relation with the universe. For Guillaume, language categorizing our experience in this way provides that specific lucidity, that specific capacity to see, which characterizes homo sapiens sapiens. Attributing to the mind, as he does, the inherent tendency to organize and systematize thus helps bring out the primary function of language as the mind’s first and most important instrument for confronting the universe. 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. (1984, 145)

Language is thus our means of quelling the turbulence of our experience, endowing our thoughts “with words that make them known” as Shakespeare puts it. It provides a sort of viewing apparatus, a universe of viewing ideas that constantly formats our experience, giving it meaning, and this without triggering an act of language leading to a sentence. (We do not have to talk about everything we see in order to understand what it is.) That is, Guillaume saw tongue itself as a sort of common man’s theory of our experience of the universe, providing ahead of time a place for whatever may arise in our stream of consciousness. And so the linguist does not have to make up a theory of English, but to describe the theory, the system of tongue and its viewing ideas, that native speakers have interiorized and use both to understand their experience and to speak about it. As such, even without the conscious intervention of language, tongue provides the basis for human thinking, and for him this includes even the scientist’s theorizing: “tongue is the prescience of science. The loftiest speculations of science are built on the systematized representations of tongue” (1984, 146). Not that Guillaume considers human thought to be language bound, limited to the forms and lexemes of one’s tongue. On the contrary, he

Thought and Language

235

emphasized the freedom of thought with regard to language but recognized that, wherever our thinking may lead, language provides a means of grasping the content of thought. “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” (1984, 51). CONCLUSION

This remarkable way of understanding the relation between thought and language led Guillaume to see the constructed state of language as the work of the mind’s inherent systematizing power and at the same time as an indication of what is to be undertaken to continue this construction and increase the mind’s ability to cope with experience. It is in this way that he understands the “brilliant insight” of a philosopher, Delacroix: “thought makes language while being made by language” (1984, 145–6), i.e. the development of thought, the findings of “silent thought” resulting from its “deep musings” as it contemplates the stream of consciousness and the adequacy of language, permits the development of tongue. With greater adequation of tongue to its object, the turbulence inherent in our experience of the universe is further quelled, thereby permitting thought to contemplate its contents in view of the next step to be undertaken in the endless development of language. This give-and-take relationship between thought and language has been evoked by others, like Wordsworth when he remarks “language and the human mind act and react on each other,” but it is rare to find a linguist trying to focus on it to the point of considering it the object of linguistics. Certainly Guillaume is the only linguist to spend the greater part of his career exploring this general relationship and trying to understand details of usage in its light. And he is, to my knowledge, the only linguist whose observation and reflection evoke such a vast panorama of human language, one extending far beyond the view of language as a means of communication: 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

236

Language in the Mind

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 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. (1984, 160–1)

Guillaume had a rare gift for seeing the general in the particular and the particular in the general. From the very beginning, this incited him to seek the system manifested in a particular use and the particular use as a possibility of the system. Highly developed by dint of constant observation and reflection, this remarkable capacity to keep universal and singular constantly before his mind, combined with a realism leading him to see everything developing in time, permitted him to construct an all-embracing theory of language considered by some to be unparalleled for its scope and originality. And for those linguists who, on reading his work, have caught a glimpse of the “wonderful” order of language in the mind, it has revealed a new dimension for the development of linguistics as a science.

16 Conclusion The most beautiful thing we can experience is the mysterious. It is the source of all true art and science. Albert Einstein (cited in Bartlett, 763)

At the end of this introduction, one naturally wonders where this approach to language with its many original insights can lead. Guillaume’s work opens avenues of reflection for the psychologist, the philosopher, the anthropologist, and of course for any linguist who is prepared to consider language in the mind. As with any original thinker, however, to approach his work one must have the docility involved in that “willing suspension of disbelief” required to understand what he has to say. This may well call for jettisoning inapplicable prejudices like those arising from positivism, materialism, and determinism, and retaining the only prejudice essential to the adventure of science, that there is an order hidden behind the directly observable facts. Like any other scientist Guillaume has the right to be wrong, but since his speculations are never divorced from observable data, his explanations are generally plausible and often convincing. In any case, his theories and hypotheses deserve the treatment accorded to any scientific work: confrontation with the facts and replacement by what explains better. For a non-linguist, one of the most striking features of Guillaume’s theory is perhaps the way it poses the problem of consciousness. He insists on the preconscious activity of the mind in forming systems and in exploiting them. Although his position is clearly mentalist, his scrupulous care in respecting the limits of linguistics led Guillaume to consider the relation between the mind and the brain “an impenetrable mystery” for the linguist. Of interest in this regard is the following comment reported by one of his former students (Toussaint, 104): “I also heard him say, towards the end of his life,

238

Language in the Mind

that one day PM [psychomechanics] would show neurologists and neurosurgeons the way forward.” When I first read this, it reminded me of what Wilder Penfield once said after addressing a group of linguists: that he was the one who should be asking them questions, and they should be telling him what to look for in his research on the brain. Both these comments, it seems to me, can be understood if, in the words of a contemporary neuroscientist, “the neural circuits are etched, so to speak, with the impressions of a newly learned fact” (Damasio, 116). Be it a newly learned fact or a system in tongue, a mental entity’s relation with the neural circuits has yet to be clearly described. In the meantime to avoid undue reductionism, elementary scientific prudence would seem to indicate that, much like contemporary physicists with their four forces, one can only accept this dualism as an unsolved problem. For the linguist it is quite obvious that, in proposing to describe how language exists in the mind before it is uttered as observable discourse, Guillaume’s theory opens new vistas. This involves turning the linguist from observing what is most visible to what is hidden. Worth remarking is that Guillaume operates this reorienting of the linguist’s regard in the manner of the scientist: he effects a scientific revolution by building on the solid achievements of his predecessors, by going beyond them, rather than by destroying them as in the case of many a political revolution. Instead of turning everything upside down he turns things around, inviting linguists, both historical and descriptive, to explore the mental origins of what they observe, because for him everything in language bears the stamp of the human mind. Not that Guillaume was the first to presume the mental origins of language, but he was the first to provide the means for exploring these origins. By adapting a well established method, he enabled the linguist sufficiently observant of usage, and convinced of the systematic potential giving rise to it, to probe that area of language forever hidden from our direct awareness. His method of analysis ensures both the scientific probity of his approach and its focus on the reality of language. Thanks to this method, whereby any manifestation can be seen as arising at a particular position in the operative time underlying the phenomenon under observation, the particular explanations he puts forward can be appreciated by working back up the chain of causation from the data to the hypothesis proposed. The hypothesis can be tested by working down the chain to new data, in an

Conclusion

239

effort to extend and refine the hypothesis, or even to replace it with a better one. As long as other scholars can verify data and results, the margin of error, the risk of going astray, is reduced to a level acceptable in scientific discourse. This “positional” linguistic technique, applied to problems and to languages that Guillaume had never focused on, has produced results supporting his claim that it is applicable to any linguistic problem involved in the act of language. This of course opens a vast field for research both in the languages we are most familiar with and in those outside the IndoEuropean family. The great advantage of using this method when examining a language other than one’s mother tongue is that it permits an analysis independent of personal prejudices. Analyzing an unfamiliar language in terms of the categories learned in Latin the way many early missionaries did, or in terms of the unconscious categories of English the way some modern linguists tend to do, is a matter of unwittingly prejudging one language on the basis of another, unrelated language. Guillaume also prejudges, as indeed any observer must, in order to have some idea of what he is looking for. But he knowingly prejudges languages, all languages, on the basis of what he considers a universal, an incontrovertible fact of human speech: the need to think in order to speak. Representing each of these thought processes as requiring time and so as interceptible at different points provides a yardstick that, because it is common to all languages, can, in a very general way, tell linguists what sort of thing to look for in the data. Not that the method can be applied mechanically as a sort of “discovery procedure” – it calls for a detailed knowledge and sensitive appreciation of language usage as well as all the linguist’s intuitive insight to work out the form and quantity of movement implied by what is observed – but it does help free one’s insight from the trammels of the mother tongue. This is why it can be argued that discovering operative time as the parameter of analysis, and illustrating the positional technique it makes possible, is Guillaume’s greatest contribution to linguistics. Be that as it may, this method of analysis is distinctive in attempting to keep the whole of language reality in view by obliging the linguist to have in mind both its visible and invisible parts, as well as the operations linking them. In this respect, one might argue that Guillaume is the most realistic of linguists, because he tries constantly to embrace language, the whole of language, and nothing

240

Language in the Mind

but language. This is not to deny that those who focus on directly observable usage, bringing to light innumerable individual facts to be explained, do an indispensable task. Nor does it suggest that those who reflect on general conditions, universals, of usage cannot provide valuable parameters, though here the margin of possible error is much wider if the linguist does not begin with what are necessary, incontrovertible facts of humans speaking. It does, however, follow from Guillaume’s concern for the total reality of language that real progress will be achieved only by keeping in mind at all times both the invisible and the visible, the general and the particular, the potential and the actual – as well as the processes that lead from one to the other. And for this reason, reading his Leçons de linguistique can at first be quite disconcerting because a discussion of some particular details of verb inflection will be “interrupted” by considerations on a general topic. Likewise, he frequently starts a lesson by reflections of a broad, systematic nature before going on to discuss particular manifestations of them. Even when he remains on the most general level, he always has in mind particular facts, the stepping stones that led him to the summits of the theory, and presumes that his reader has them in mind as well. Thus while Guillaume takes his stance in the reality of language, one foot in language-as-actual, the other in language-as-potential, his method of analysis obliges him to work from the observed actual back to the potential, just as in explaining he must proceed from the potential to the actual. Without this conscious effort to focus on the operational, on languaging, without this resolute attempt to situate everything in the microtime required to produce words and sentences, the linguist, like a cyclist who stops pedalling, sooner or later falls into one of the static views of language that can be observed in former centuries. The eighteenth century saw the parts of speech generalized abusively – with no supporting evidence – to provide a logical framework for all languages, whereas the nineteenth century, preoccupied with the observation of detail in texts, tended to view with suspicion speculation about anything not derived directly from observed data. Both these forms of reductionism can, of course, be observed today, be it as a logical formalism or as an empirical realism. Some approaches do attempt to grasp both, but as the horns of a dilemma and not as the beginning and end of a process, so that at best their explanations strike one as ingenious rather than convincing, and they tend to veer off in the direction of some adjacent field, such as the neurological

Conclusion

241

substratum or the pragmatics of the social situation involved when we speak. To embrace language, the whole of language, and nothing but language, linguists must keep their eye, their mind’s eye, fixed on what the speaker (and the hearer) does, on the mental and physical processes required to produce the observed sentence. All this suggests that the prospects of further exploration are not without pitfalls. If the theory of Guillaume is accepted as a set of inalterable dogmas it will be reduced to the status of an abstract logical structure affording shelter for selected data. The Psychomechanics of Language would be denatured into an ideal construct with no real link to the languaging processes constituting the language under study. On the other hand, if the theory is taken as simply a means of classifying observed facts, it will survive as just another empirical approach whose relation with a human speaker is only accidental. In either case language would be reduced to something static, quite distinct from its dynamic reality. If however the Psychomechanics of Language is understood as a working theory, one whose postulates and findings are constantly open to verification, then the method will remain central, languaging will not be lost from sight, and the relation between potential and actual will be ever-present, whether one is discussing the most general categories of tongue or some minute particularity of discourse. Whatever the fate of his work, it will be to Guillaume’s credit that he was among the first to attempt to trace observable items of speech step by step back to their mental origins, back to that specifically human endowment without which language as we know it remains inexplicable. In so doing he has given us a glimpse of the human mind’s wonderful faculty for organizing not only its experience coming from the outside world, but even its own operations, its own capacity to systematize, so that it can better theorize its experience. By giving a first glimpse of how our language-making ability arises from the mind’s innate potential for categorizing and discriminating, for drawing the general from the particular and the particular from the general, Guillaume highlights the anthropological dimension of language as a record of the development of homo sapiens sapiens on the basis of that initial endowment. To the extent that his theory is developed and continues to provide explanations of languages in their existential reality, it will throw more light on that mysterious potential, which, as Guillaume puts it, makes us “beings of exception in nature.”

This page intentionally left blank

Glossary

chronogenesis: the mental operation of constructing the time image expressed by a verb. chronothesis: a representation of time resulting from intercepting the operation of chronogenesis at some point in its development. completive pronoun: a grammatical word used as a determiner in a noun phrase (cf. suppletive pronoun). dematerialization: the reduction of the lexical matter of a word resulting in a more abstract meaning. discourse: the actual speech and texts a speaker produces. event: the representation of any happening (state, process, action, achievement, activity, etc.) as a durational entity. event time: the grammatical representation of the duration of an event. expression: the operations involved in manifesting by means of signs what has been represented by means of the resources of tongue. extensity: the particular field of application to which a notion is applied in any given use. external incidence: the operation whereby the lexical import of one word is applied to a meaning support outside itself. ideogenesis: the operations involved in calling forth the particular notion or lexeme of a word; notional ideation. immanent aspect: the grammatical form for representing time contained within the event, the time involved in the event’s realization phase. imperfective: a representation provided by the progressive form whereby only part of the event’s lexical matter is situated in time. import: the meaning a word, phrase, or clause brings to the sentence. incidence: the operation whereby an import of meaning is applied to a support of meaning.

244

Glossary

intended message: the experiential content of consciousness, or that part of it, that a person wants to talks about. internal incidence: the operation whereby the substantive’s import of meaning is applied to a support within its own extension, i.e. to its extensity. lexigenesis: the operations involved in reconstructing a word, both meaning and sign, at the moment of speaking. metaphase event: a ‘dynamic’ type of event whose every instant is susceptible to bring about a new phase of the event. monophase event: a ‘stative’ type of event whose every instant involves the same phase, which persists throughout the event’s duration. morphogenesis: the operations involved in calling on the system of the parts of speech to provide the grammatical meaning of a word; structural ideation. notional ideation: see ideogenesis. operative time: the time required for an operation to unroll; in particular the microtime required for the mind to realize one of the preconscious mental operations made possible by tongue. perfective: a representation provided by the simple form whereby the whole of the verb’s lexical matter is situated in time. positional linguistics: the technique involved in discerning the position in the underlying genetic process giving rise to the particular sense of a morpheme in a given use. predicative parts of speech: types of word (substantive, adjective, verb, adverb) instituted to represent something in the intended message. psychogenesis: the operations involved in forming the meaning of a word, both lexical and grammatical. psychosystematics of language: that part of linguistics concerned with the study of grammatical systems of tongue. quasi-nominal mood: the mood of the verb arising at the earliest interception of chronogenesis containing nonfinite tenses, often used as nouns (adjectives or substantives). radical binary tensor: the mental operator consisting of a process of generalization and a process of particularization for instituting a relationship between the universal and the singular, the greatest possible and the smallest possible, etc. representation: the operations involved in depicting something in the intended message by means of the lexical notions of tongue (ideogenesis) and the forms of the appropriate grammatical system (morphogenesis).

Glossary

245

sign: the perceivable mark that calls to mind the meaning. structural ideation: see morphogenesis. suppletive pronoun: demonstratives, possessives, etc. functioning as noun phrases (cf. completive pronouns). support: that which an import of meaning characterizes by being made incident to it. tongue: the viewing ideas (lexemes), the organized set of mental systems, and the signs constituting the language potential of a speaker. transcendent aspect: the grammatical form for representing time beyond the event, the time involved in the event’s result phase. transpredicative parts of speech: types of word (grammatical words) instituted to bring out something involved in the operativity of the sentence. universe time: the grammatical representation of time conceived as an endless stretch capable of containing any event.

This page intentionally left blank

Notes

CHAPTER ONE 1 Cf. Valin 1994, 928–96 for the biographical details. 2 For the list of volumes published so far, as well as a list of the unpublished manuscripts: www.fl.ulaval.ca/fgg/. 3 Terms introduced in small caps are listed in the glossary. The term sign is to be understood here not in the Saussurian sense but in the traditional, common language sense of the perceivable entity that calls to mind something else, namely the meaning or significate (= that which is signified by the sign). 4 One is reminded here of a similar situation confronting Copernicus and later Galileo, described by Shea and Artigas (18): This raises the question of what Ptolemy was trying to achieve. He was certainly not attempting to devise a unified cosmology. Rather he seems to have assumed that his job as an astronomer was “to save appearances,” as the phrase went, namely “to account for the way heavenly bodies appeared,” not to offer a physical explanation of their motion. If a planet showed an irregularity in speed, and another in size, the astronomer was at liberty to explain the first by an epicycle and the second by two epicycles or vice versa! The question of the reality of these constructions was never raised by Ptolemy. Copernicus was dissatisfied with this arbitrary way of doing astronomy and he proposed a radically different system by moving the Sun to the center and locating the Earth among the planets.

Guillaume, not satisfied with makeshift ways of accounting for usage, sought to explain in terms of mental reality the way morphemes are used and his theory was received by contemporaries in much the same way as that of Copernicus.

248

Notes to pages 10–27

5 Cited in Guillaume 1938/1964, introduction, 20. 6 See the remarkable discussion of Saussure’s theory of diachrony and synchrony and the Cartesian instant, the instant where “I think therefore I am,” in Guillaume 2004, 3–19. 7 This is how Malinowski described isolated words in his Coral Gardens and their Magic. 8 Foundations for a Science of Language, 1984. Citations from other writings of Guillaume have been translated from the French by myself. CHAPTER TWO 1 A first version of this chapter has appeared in Productivity and Creativity: Studies in General and Descriptive Linguistics in Honor of E.M. Uhlenbeck, Mark Janse, ed. (1998, 93–101). This version appears with the permission of the original publisher, Mouton de Gruyter (Berlin and New York). 2 See Valin (1996, 37–45) for a clear description and convenient example of this method. 3 One is reminded of the dictum of the scholastics, Ab actu ad posse valet illatio, to the effect that observing something as real is sufficient grounds for inferring the existence of its possibility. 4 Guillaume used the term “morpheme” to designate a grammatical component of word meaning, opposing it to “lexeme,” the word’s lexical component, and “phoneme” a component for forming the physical part of the word, the sign. 5 The late Professor Uhlenbeck has pointed out (private communication) that the conception of language as a dynamic system of systems was proposed by the Prague School in the 1920s. 6 As a result the spoken sentence normally has an ephemeral existence that does not extend beyond the moment when its message has been understood. 7 This difficulty is illustrated by Harris (11) when, criticizing Guillaume’s view, he argues that Saussure (1916/1955, 113) did introduce time when discussing langue and masse parlante. The point is that Saussure is referring to time over “several centuries,” in diachrony, whereas Guillaume is concerned with time in synchrony, the micro-time involved in someone speaking.

Notes to pages 28–30

249

8 Guillaume generally uses the term “morphology,” not in the sense of inflections but to designate the set of morphemes incorporated into a word, i.e. the grammatical components of a word’s meaning. 9 The most striking example of the way the ordinary-speaker meaning of a word can influence our thinking is provided by Saussure himself, who redefines the key word signe in the Cours de linguistique générale (1916/1955, 99) as a technical term yet uses it in the common-usage sense more than fifty times in the same volume. 10 For a more abstract, and so more exact, description of this view of language, see Guillaume’s lecture of 7 January 1960 (1995, 283–5). Although it was implicit as early as his study on the article (1919), only in this lecture, a few weeks before his death, does Guillaume adequately designate the second phase: langage puissanciel || effection || langage effectif: Literally: “potential language || effecting || effected language.” Up to this point he often used discourse, occasionally tongue, in an operational sense, a practice that can result in a certain ambiguity for the reader. 11 Again, one can appreciate the difference between Guillaume and Saussure, for whom la langue is language minus speech (“La langue est pour nous le langage moins la parole”: 1955, 112), a view Guillaume summarized by the formula: langage = langue + parole. It is not clear why Harris (11) finds this interpretation of Saussure “disconcerting.” 12 See Geeraerts (663–74) for an interesting comparison of these two types of approach from the point of view of lexical semantics. CHAPTER THREE 1 The processes involved in realizing the meaning of a word or a sentence do not lie within the field of the observable. Those processes that are directly observable, whether by ordinary perception or by technical means, are involved in the realization of the physical sign. 2 To be more exact, it should be pointed out that, to avoid imposing our Indo-European view of a sayable unit on character languages like Chinese, Guillaume was careful not to apply the term “word” to such languages. He used the term “vocable” as a generic to bring out the distinction between the character and the word. However, since our examples are drawn from English, it will not be necessary to take this distinction into account in the present context, and so the term “word” will continue to be used as a generic.

250

Notes to pages 31–44

3 Cited in Gardiner (99); see the discussion in Michael (39–40). 4 Even children, who can be heard saying things like “he comed,” are applying the processes for forming the sign before they have learned the irregularities of certain verbs. 5 Guillaume’s view that, thanks to an appeal to this liaison in tongue, the sign calls its meaning to the speaker’s mind differs from Jakobson’s view of the physical sign as “matter which transfers something in addition to itself” (Waugh 39). 6 Others have expressed this much more vividly. Thus, Northrop Frye (1124), the critic states: “it is clear that all verbal structures with meaning are verbal imitations of that elusive psychological and physiological process known as thought, a process stumbling through emotional entanglements, sudden irrational convictions, involuntary gleams of insight, rationalized prejudices, and blocks of panic and inertia, finally to reach a completely incommunicable intuition.” And Robert Frost: “A poem … Begins as a lump in the throat, a sense of wrong, a homesickness, a lovesickness … It finds the thought and the thought finds the words” (cited in Bartlett, 749). 7 It follows that homonyms, quite unrelated in meaning, are words with different meaning potentials in tongue. 8 In chapter 7, I will discuss the reason for using “substantive” rather than the more current term “noun.” 9 This applies to part-of-speech languages like English, but as many languages have not instituted these most general categories as grammatical systems within the word, Guillaume was careful to avoid the frequent mistake of assuming that all other languages are like his own in this respect. Indeed, his general theory of the word, the Theory of Glossogenetic Areas, is based on the idea that different language types correspond to different systems of word formation, each providing a different structure for the word. 10 Because of the omnipresence of parts of speech in English, it is often hard for Anglophones to become aware of their grammatical role. We can get a glimpse of it from the expression long time no see, where the built-in grammatical component of the words is reduced to a minimum and the syntactic relationships giving rise to the total meaning of the expression must be established largely without it. 11 To bring his operational perspective to mind, instead of speaking of morphology Guillaume (cf. 1984, 14) often used the terms morphogenesis and morphogeny since they make implicit reference to

Notes to pages 45–51

12

13

14

15

16

251

the coherence of processes leading up to the results often observable in the sign. A moment’s reflection on the nature of an instant – ”an infinitesimal space of time” – will dispel the apparent paradox involved in proposing not only an operation, but a succession of operations, in an instant. See Valin (1994, 123–78) for a discussion of this point and an outline of the various relationships between language and time. “It may justly be urged that, properly speaking, what alone has meaning is a sentence” (J. Austin, cited in Lehrer, 136). For a discussion of Wittgenstein’s dictum, “Don’t look for the meaning of a word, look for its use,” see Lyons 1968 (410–11). A glance at the table of contents and the index of recent books in linguistics provides a convenient check on the place given the word as a linguistic unit in contemporary studies. The fact that, except on the level of the sign, the word is often taken for granted makes Guillaume’s point. Again one should recall that in some languages a word’s grammatical meaning is not organized into parts of speech as in English. Furthermore, one should keep in mind the distinction between the word as treated here and the character in Chinese. Guillaume often maintains that, as a unit of tongue, the character does not include a grammatical component predetermining what syntactic relationships it can entertain with other vocables in the sentence (cf. inter alia 1987, 167, 173; 1982, 65–70). See Lowe 1996 for a psychomechanical description of the word in Inuktitut, which is formed on structural principles quite different from those underlying the word in English. CHAPTER FOUR

1 A first version of this chapter appeared as Hirtle 2000. The present version is printed with the permission of the original publisher, John Benjamin’s. 2 Cf. Wickens (189). Many examples given here will be drawn from this, the most complete study to date on -s as a number morpheme in English. Without an appreciation of the extraordinary variety of usage of -s, illustrated by Wickens’s mine of examples, one can hardly hope to reach an adequate view of the underlying system of number.

252

Notes to pages 52–75

3 “Abduction is the process of forming an explanatory hypothesis … Deduction proves that something must be; Induction shows that something actually is operative; Abduction merely suggests that something may be … [E]very single item of scientific theory which stands established today has been due to Abduction” (Pierce, 106). 4 The speaker was in fact a rancher in the foothills of the Canadian Rockies. CHAPTER FIVE 1 The following citation from a well-known volume that appeared around the middle of the twentieth century will serve as a reminder of a widespread attitude at that time: “For the purposes of this book, therefore, the linguist’s assumption that language is a set of sounds will be adopted” (Hill, 3). 2 Dennis Philps points out (private communication) that in many languages, including some creoles, the ‘plural’ is formed by reduplication of the ‘singular’ form, a semiology suggesting that the singular arises first here as well. 3 To have a complete view of, say, the father/son relationship it must be considered from both points of view. Likewise for the quantitative relation between, say, 2 and 3: 2 is smaller than 3, but 3 is bigger than 2. 4 For an account of how the system of number was discerned in the first place and a more detailed examination of usage, see Hirtle 1982. 5 Christophersen (25) pointed out the inadequacy of the term “mass” as applied to abstract notions, proposing rather “continuate.” To avoid confusion with the system of number, I adopt “continuum” and, like Christophersen, oppose it to “unit” to designate the two manners of actualizing a substantive’s lexeme. 6 Even in cases where there are internal limits present, as in two bear, these are subordinated to a ‘continuate’ impression arising from a view of the species, as we have seen. 7 This implies that we are, in effect, discussing three different sentences here because in Guillaume’s view, as we have already seen, a sentence exists as a reality of language only as the expression of someone’s intended message. Different intended messages therefore give rise to different sentences, each expressing a different “complete thought.” Thus as Hewson 1992 makes clear, it would be

Notes to pages 76–96

8 9

10 11 12 13

14

253

a lack of realism to consider that only one sentence with different possible meanings is involved here. We occasionally find a use like Cattle was the only means of livelihood. Here, of course, it expresses a ‘continuum’ notion. A limit like a property line may of course be marked, i.e. made perceivable, by some physical means, such as a hedge or fence, which occupies space. The property line itself, however, does not consist of a strip of land between two properties. He knew English only through grammars, and these, to my knowledge, make no mention of this remarkable trait of number. Examples in this section are taken from either Hirtle 1982 or Wickens. Applied metaphorically to a human, swine can of course express a ‘singular.’ Since a substantive cannot be both ‘singular’ (a) and ‘plural’ (die) at the same time, this use led us to see that verb agreement is not always a reliable criterion for determining the grammatical number of the substantive, as Reid (1991) and others suggest. This observation has important implications for the analysis of the noun phrase. The importance of primary sources should be emphasized here. When first asked if a plier was acceptable in English, I replied that it would suggest to me a speaker for whom English is not the mother tongue. I was quite unaware of usage in special contexts where the conceptualization of the tool is different from that found in ordinary practical contexts. CHAPTER SIX

1 This is how Davis (184) describes the speed with which neural networks permit us to “experience an experience.” 2 The so-called irregular signs for the plural (e.g. mice, leaves, geese) tend to be replaced by the regular sign when the lexeme signified is felt to be distinct. Thus: mouses for the computer accessory, Leafs for the hockey team, and gooses for the tailor’s instrument or the poking gesture. 3 Guillaume’s earliest major study (1919/1975) was on the article. It shows the meticulous care he exercised in observing the nuances expressed in usage. 4 In psychosystematics, there is no attempt to “ask language users to describe what is going on in their minds when they produce and

254

Notes to pages 97–114

understand words and sentences” (Ungerer and Schmid, xi), because only the resulting meaning, the meaning expressed, emerges into consciousness where it can be observed introspectively; the processes of producing and understanding it unroll in the preconscious mind. 5 See the discussion above, chapter 5. 6 Thus this capacity is manifested in activities other than language, for example in ordinary physical activities like walking (the length of a step) or opening the mouth (drinking with a straw as opposed to being in the dentist’s chair), where our intention to act determines the extent (minimal, intermediate, or maximum) to which the process is actualized. CHAPTER SEVEN 1 With the consequence that one sometimes hears Francophones arguing about which fictive gender, le or la, to attribute to an unfamiliar substantive. 2 It is not uncommon in grammars to bring in the use of personal pronouns like he and she to suggest that gender in the substantive does change. What justifies the assumption on the part of such grammarians that the gender of the pronoun can be attributed to the substantive is not made clear. 3 Many of the above remarks are based on Gender in English: System and Uses, an unpublished doctoral thesis (Quebec City: Université Laval, 1991) by Lori Morris. This thesis is largely concerned with a detailed study of the use of gender in the personal pronouns. 4 Valin points out that, in languages where the substantive is declined for case, the oblique cases can all be governed by a preposition, whose function is to apply the substantive’s meaning to a meaning support elsewhere in the sentence. The nominative is the only syntactic case that cannot be governed by a preposition, evidence that the function of a substantive in the nominative is to be a support for meaning brought in by some other part of the sentence. 5 The various ways of designating the mental side of a word deserve comment since each brings out a different aspect of the same reality. The term “content,” often used by grammarians and linguists, implies a relationship with a container, usually understood as the physical sign; this, however, does not portray the real nature of the relationship involved here: a sign does not contain, but rather

Notes to pages 114–26

6 7

8

9

10

255

signifies, its meaning. The term “matter,” often used by Guillaume, implies a relationship with a form, and so suggests the relation between lexical and grammatical meanings, the latter providing a mental form or category for the former. To use the term “representation” calls to mind the relation between meaning and its extralinguistic correlate in the intended message. The most common term “meaning” and its technical equivalent “significate” characterize the mental side of the word simply insofar as it is signified by the sign, and so are the most general terms. The term is formed from “psychogeny” (cf. Guillaume 1984, 134). In Valin’s “L’incidence interne: une évidence?” (in 1994, 329–42), an article on which many remarks in the present chapter are based, it is pointed out that a concept is defined in the common noun, whereas in the proper noun there is no limit to the characterizing traits, and so no concept is defined. This applies to proper nouns as well. When two or more individuals have the same name, these constitute for anyone who knows them a set of homonyms. That is to say, there is no category or concept of persons named John, or of cities named London, but rather a series of differently represented entities with the same sign. This would suggest that, for a given speaker, there are no absolute synonyms, a point borne out by observation, since no two words seem to have quite the same possibilities of usage in discourse. This degree of independence is not found in all languages. As Valin points out (1994, 383–94), when functioning as subject, ‘inanimate’ substantives have the sign for the accusative case in languages like Classical Greek and Latin, declined for case. And in Spanish, ‘animate’ substantives require the preposition a when in direct object function. CHAPTER EIGHT

1 This holds even when horse does not have a ‘unit’ representation, as when designating the meat of the animal. 2 Except in metaphorical usage, where the substantive says what something or someone is like. It would take us too far afield to explore this important question here. Cf. Hirtle 1992 for discussion. 3 Unable to teach in a university because he had no doctorate, Guillaume finally (at the age of fifty-five) obtained a post at the

256

4

5

6

7 8

9

Notes to pages 127–52

École pratique des hautes études in Paris thanks to the posthumous influence of Antoine Meillet. It has become common to speak of “word classes” instead of “parts of speech,” but the change of concept involved is not often mentioned. While a “class” may group any number of items that fulfill the same observable function, a “part” here signifies an item formed to integrate with other parts to constitute a whole, the sentence. That is, “part” calls to mind the grammatical meaning forming a word and so conditioning its function. This term is, therefore, more oriented toward analysis because it evokes a conditioning factor. For a careful and insightful examination in light of psychomechanics of the many syntactic questions raised by the adverb in English, see Guimier. The fact that the adverb is not declined for gender, number, and case may well be due to the fact that its immediate support, the lexeme of an adjective, a verb, etc., is not in itself categorized as ‘animate,’ ‘plural,’ etc. It follows that the adverb’s lexeme need have none of these forms imposed on it to adapt it to its immediate support. Like everything else in the sentence, however, these grammatical relationships do reflect something in the speaker’s experience. The view implied by describing some words as “semantically empty” (Hudson, 91) is quite opposed to that of Guillaume, for whom the whole purpose of a word is to make a unit of meaning expressible. Similarly, his view of the morpheme is opposed to that implied in “most (if not all) grammatical morphemes are meaningful” (Langacker 1987b, 18), Guillaume maintained that all morphemes have a meaning import. And still very closely linked to the adverb in English. Cf. Look out the window! and Look out! CHAPTER NINE

1 There are significant differences between the analysis of usage presented here and that found in the 1988 article. 2 It should be recalled that paraphrases like these, which describe the meaning of a word or expression by means of other words, are at best suggestive but can never adequately signify the meaning of what is paraphrased.

Notes to pages 156–61

257

CHAPTER TEN 1 “If you do [indicative] it and [that] there result [subjunctive] an accident, you will be held responsible.” See Valin (1994, 9f) for an account of this discovery. 2 For a full discussion, see Valin (1994, 123–77). 3 This term, introduced by Guillaume, obviously does not designate the generating of time but rather the generating of an image or representation of time. 4 The description of chronogenesis in English given here is developed from Guillaume’s latest views, presented viva voce by Roch Valin. 5 In ordinary usage, the term “present” is not, of course, restricted to this sense of a point or limit with no duration. For example, when we speak of “the present of speech,” we are thinking of the stretch of time during which someone carries out an act of language. “Theoretically it [the present] is a point, which has no duration, any more than a point in theoretic geometry has dimension … But in practice ‘now’ means a time with an appreciable duration, the length of which varies greatly according to circumstances,” as Jespersen (1951, 258) points out. When speaking of the verb in English, however, it avoids confusion to restrict the term “present” to the ‘limit’ sense given here, using some such expression as “present of speech” to designate a stretch of time capable of containing something, and to use the term “nonpast” to designate the time-stretch containing the present of speech and the future. In a language like French, as Guillaume demonstrated, the grammatical present is represented as a short stretch of time capable of containing an event and not just as a limit. 6 The use of the nonpast tense as a “historical present” and the “science fiction” use of the past tense to refer to some later century show that the instant represented as present does not necessarily correspond to the speaker/writer’s experienced present. Such uses, along with the “irrealis” use of the past tense, do not conflict with the view of tense presented here but in fact help confirm it. They can, however, be profitably discussed only in a more detailed study of the verb presenting a full panorama of tense usage. 7 The following citations from a well-known physicist will indicate that these impressions are not just an invention of grammarians. In The Nature of Physical Theory (New York: Dover Publications, 1936: 32),

258

8

9 10

11 12

13

14

15

16

Notes to pages 164–8

P.W. Bridgman remarks: “We usually think of the future stretching before us, and ourselves as going to meet it.” This should not, however, be considered a universal since “The Greek, on the other hand, thought of himself as facing the past, with the future behind him, coming up over his shoulder as the landscape unfolds to one riding back to the engine.” Observations like this suggest that it is our language that determines how we organize our experience of time. By analogy with the usual use of the term universe to designate the container of any imaginable spatial entity, universe time designates time represented as the container of any imaginable temporal entity. Among the Guillaume manuscripts, there is a considerable correspondence in Russian. For a contrast between the Slavic/Latin and the Romance/ Germanic manners of instituting aspect as a subsystem of the verb, see Valin 1994 (37–52, a reprint of his 1965 article), a translation of which appears as an appendix in Hirtle 1975. Or, in some cases, être (= ‘be’). The expression “person support” denotes the representation of person within the verb, the spatial element that permits incidence to the subject. It remains, of course, that the uses of the transcendent aspect are by no means identical in English and French, a difference due in large part to the very different ways of representing the present in the two languages. This conception of aspect as a grammatical system in English is quite independent of distinctions represented by the verb’s lexeme (to be about to, to begin, to continue, etc.). Being based on the immanent/transcendent opposition, it also differs from approaches that consider either the progressive or will + infinitive expressing ‘future’ to be an aspect (cf. Quirk et al. 197ff; Hewson 1997a, 87). The presentation of the simple form is necessarily schematic here, leaving many questions unanswered. See my 2007 study for a more detailed discussion, and particularly for a discussion of the difference between the representation of a stative event by means of the simple form and the representation of a developing event by means of the progressive. There is, of course, a difference of expressive effect between these two sentences: with found we have the impression the event is over,

Notes to pages 169–78

17 18 19 20

259

whereas with lived we have the impression the event may continue. This difference arises from the different realizations of the lexeme, as dynamic or stative, but in each case the person support is situated immediately after the last accomplished instant of the past participle event, i.e. at the first instant of the result phase. The aftermath of an incomplete event is expressed by the progressive in the transcendent aspect. Goes here would be a representation of usage in a sports commentary such as He goes into the corner. For a detailed comparison of the infinitive and the participle/ gerund, see Duffley 2006. To provide a complete picture, the past subjunctive, if it is still a productive tense in English, would have to be included here. CHAPTER ELEVEN

1 The following discussion brings out only those points that help illustrate the general theory. 2 The fact that be as auxiliary of the progressive is quite rare as a present participle (cf. Jespersen 1954 IV, 205) is to be attributed to its verbal complement, itself a present participle. As an auxiliary of the passive, it is of course commonly found in this form (You are being called). 3 It is rarely found in the subjunctive, if at all, because of the unlikelihood of anyone wanting to evoke the result phase of an event as a mere possibility. 4 In French, the auxiliary of the transcendent aspect, avoir, does occur in the transcendent aspect, giving rise to the third aspect of French, the bi-transcendent aspect, as in Quand j’ai eu fini…, a form whose particular nuance is difficult to translate into English. 5 For the reasons why do auxiliary, unlike its full verb counterpart, has no infinitive or participles, and so is not found in the transcendent aspect, see my 1997b article. 6 A good illustration of this is the fact that even something that takes no time in reality is represented as having a minimal duration, as in It is exactly midnight. 7 It need hardly be pointed out that this is a far cry from simply characterizing auxiliaries by a feature (AUX), which, as Langacker (1987b, 853) points out amounts to “ad hoc labeling but not explicating the notion AUXILIARY.”

260

Notes to pages 178–203

8 But not eliminated completely as we have seen, since this would have resulted in a grammatical inflection, as in the case of the future endings of the French verb (cf. Guillaume 1938/1964). 9 To distinguish between suppletive and auxiliary uses here, one can test for the ability to take enclitic n’t. One could finish and with but she didn’t. This would not be acceptable in and : *but she didn’t so. 10 Like the fifteen-year-old on the radio: “They recently unemployed five hundred peope.” CHAPTER TWELVE 1 The point is that an example must be interpretable by other speakers, competent observers, if it is to serve as data. Where a sentence is not readily understood in terms of the intended message, sufficient indications of context and situation must be provided, whence the need for attested examples. Where it contains an immediately interpretable use, there is no need to provide such indications, and so in such cases made-up examples are acceptable. The examples given in this chapter illustrate commonplace usage because a theory should be able to explain both ordinary and extraordinary uses. 2 These terms are not used in the same sense as that described by Langacker (1991a, 85–91), for whom Paul knows the answer contains an imperfective, Tom is learning the answer a perfective. CHAPTER THIRTEEN 1 There is, of course, no attempt to discuss reference in detail here. Even from the point of view of the speaker, there is much more to be explored; for example, the very fact of languaging some experience may well bring it more clearly into focus and lead the speaker to see it differently. Some of the questions involved here are discussed in my 1994 article. 2 This is not to deny that throughout the constructing of a sentence, especially when writing, we can check a word with the intended message in order to see if it adequately represents what we have in view and, whenever needed, find a more adequate word. This sort of checking on the part of the speaker to rectify or refine

Notes to pages 203–33

261

representation is not the same as the referential incidence that arises once the sentence is achieved. 3 Invariable for any speaker, but not in diachrony, where we observe words “narrowing” or “widening” their extension over a number of generations. 4 The following syntactic analysis is based on Valin 1981. 5 The same is, of course, true of any polysemous morpheme or word. CHAPTER FOURTEEN 1 Most of the examples given here are taken from Reid, Juul, Pulgram, and Christensen 1955. See my 1984 article for references. 2 This last example heads an article (Newsweek, 22 July 1985: 40) with photo on the well-known and the newly introduced versions of the same soft drink. I presented it to some fifty scholars at the end of a paper on verb agreement in English. The reactions were significant: “Could it be a mistake?” “Is this really English?” and the like. The general note of disbelief suggested to me that some linguists, inhibited by a view of language as “rule-governed activity,” do not appreciate the resources made available by their mother tongue to cope with new situations. Fortunately this does not inhibit the exploiting of these resources by ordinary speakers, including linguists. CHAPTER FIFTEEN 1 See also, for example, Vendler’s “Wordless Thoughts.” 2 Cited in Gramophone, September 1996, 14. 3 See for example Plato in The Sophist (235c), who speaks of “the method of people who are able to chase a thing through both the particular and the general,” and Humboldt (110), for whom “all thought consists in separating and combining.” 4 Guillaume of course was not the first to propose the intervention of the mind in forming our perceptions. Of interest is that some recent research in the neurosciences calls on higher functions, including a “semantic” input, to interpret perception. Cf. Ramachandran and Blakeslee (112, 275n16 ); also Arguin et al.

This page intentionally left blank

Bibliography

Allan, Keith. 1976. “Collectivizing.” Archivum Linguisticum 7, 99–117. Allen, R.L. 1966. The Verb System of Present-Day American English. The Hague: Mouton. Ard, Josh. 1982. “Auxiliary Do: Support or Emphasis?” Linguistics 20, 445– 66. Arguin, Martin, Daniel Bub, and Gregory Dudek. 1996. “Shape Integration for Visual Object Recognition and Its Implication in CategorySpecific Visual Agnosia.” Visual Cognition 3 (3), 221–75. Bartlett, John. 1980. Familiar Quotations. 15th ed. Boston: Little, Brown. Bloomfield, Leonard. 1933. Language. New York: Holt, Rinehart & Winston. Bolinger, Dwight L. 1963. “The Uniqueness of the Word.” Lingua 12, 113– 36. – 1977. Meaning and Form. London and New York: Longman. Chomsky, Noam. 1965. Aspects of the Theory of Syntax. Cambridge, MA: MIT Press. Christensen, F. 1955. “Number Concord with ‘What’-Clauses.” American Speech 30, 30–7. Christophersen, Paul. 1939. The Articles. Copenhagen: Einar Munksgaard and London: Humphrey Milford. Croft, William and Allan F. Cruse. 2004. Cognitive Linguistics. Cambridge: Cambridge University Press. Comrie, Bernard. 1976. Aspect. Cambridge: Cambridge University Press. – 1985. Tense. Cambridge: Cambridge University Press. Cruse, D.A. 1986. Lexical Semantics. Cambridge: Cambridge University Press. Damasio, Antonio R. 1999. “How the Brain Creates the Mind.” Scientific American, December 1999, 112–17.

264

Bibliography

Davis, Joel. 1994. Mother Tongue: How Humans Create Language. New York: A Birch Lane Press Book. Donald, Merlin. 2001. A Mind so Rare. The Evolution of Human Consciousness. New York: W.W. Norton & Company. Duffley, Patrick. 1992a. The English Infinitive. London and New York: Longman. – 1992b. “The Definition of the Verb in Langacker’s Cognitive Grammar: How to Deal with Grammatical Meaning?” The Eighteenth LACUS Forum 1991, Ruth M. Brend, ed. Lake Bluff: Linguistic Association of Canada and the United States, 229–34. – 2006. The English Gerund-Participle: A Comparison with the Infinitive. New York: Peter Lang. Einstein, Albert. 1956. Out of My Later Years. Secaucus, NJ: The Citadel Press. – 1981. Ideas and Opinions. New York: Dell. Einstein, Albert and Leopold Infeld. 1966. The Evolution of Physics. New York: Simon and Schuster. Epstein, Richard. 1991. Cognitive Linguistics, 2/3, 298–309. Fauconnier, Gilles. 1984. Espaces mentaux: aspects de la construction du sens dans les langues naturelles. Paris: Éditions de Minuit. Frye, Northrop. 1971. “Ethical Criticism: A Theory of Symbols.” Critical Theory Since Plato, Hazaro Adams, ed. New York: Harcourt. Gardiner, Alan. 1960. The Theory of Speech and Language, 2nd ed. Oxford: Clarendon Press. Geeraerts, Dick. 1988. “Cognitive Grammar and the History of Lexical Semantics.” Topics in Cognitive Linguistics, B. Rudzka-Ostyn, ed. Amsterdam and Philadelphia: John Benjamins. 645–77. Gould, Stephen Jay. 1987. Time’s Arrow Time’s Cycle: Myth and Metaphor in the Discovery of Geological Time. Cambridge, MA, and London: Harvard University Press. Guillaume, Gustave. 1919/1975. Le problème de l’article et sa solution dans la langue française. Paris: Nizet; Québec: Presses de l’Université Laval. – 1938/1964. “Théorie des auxiliaires et examen de faits connexes.” Langage et science du langage. Paris: Nizet; Québec: Presses de l’Université Laval. – 1971. Leçons de linguistique de –, vol. 2: 1948–1949, Série B.: Psychosystématique du langage. Principes, méthodes et applications (1). Québec: Presses de l’Université Laval; Paris: Klincksieck. – 1982. Leçons de linguistique de –, vol. 5: 1956–1957. Systémes linguistiques et successivité historique des systémes (II). Québec: Presses de l’Université Laval; Lille: Presses universitaires de Lille.

Bibliography

265

– 1984. Foundations for a Science of Language. Amsterdam and Philadelphia: John Benjamins. – 1987. Leçons de linguistique de –, vol. 8: 1947–1948, Série C. Grammaire particulière du français et grammaire générale (III). Québec: Presses de l’Université Laval; Lille: Presses universitaires de Lille. – 1989. Leçons de linguistique de –, vol 9: 1946–1947, Série C. Grammaire particulière du français et grammaire générale (II). Québec: Presses de l’Université Laval; Lille: Presses universitaires de Lille. – 1990. Leçons de linguistique de –, vol 10: 1943–1944, Série A. Esquisse d’une grammaire descriptive de la langue française (II). Québec: Presses de l’Université Laval; Lille: Presses universitaires de Lille. – 1991. Leçons de linguistique de –, vol. 11: 1944–1945, Série A. Esquisse d’une grammaire descriptive de la langue française (III). Série B. Sémantèmes, morphèmes et systèmes. Québec: Presses de l’Université Laval; Lille: Presses universitaires de Lille. – 1992. Leçons de linguistique de –, vol. 12: 1938–1939. Québec: Presses de l’Université Laval; Lille: Presses universitaires de Lille. – 1995. Leçons de linguistique de –, vol. 13: 1958–1959, 1959–1960, Québec: Presses de l’Université Laval; Paris: Klincksieck. – 1997. Leçons de linguistique de –, vol. 15: 1951–1952, Série A. Psychosystématique du langage. Principes, méthodes et applications (IV). Québec: Presses de l’Université Laval; Paris: Klincksieck. – 1999. Leçons de linguistique de –, vol. 16: 1942–1943, Série B. Esquisse d’une grammaire descriptive de la langue française (I). Québec: Presses de l’Université Laval; Paris: Klincksieck. – 2003. Essais et mémoires de –, vol. 1. Prolégomènes à la linguistique structurale I. Québec: Presses de l’Université Laval. – 2004. Essais et mémoires de –, vol. 2. Prolégomènes à la linguistique structurale II. Discussion et continuation psychomécanique de la théorie sausurienne de la diachronie et de la synchronie. Québec: Presses de l’Université Laval. – 2005. Leçons de linguistique de –, vol. 17: 1941–1942 B, Théorie du mot et typologie linguistique: limitation et construction du mot à travers les langues. Québec: Presses de l’Université Laval. – 2007. Essai de mécanique intuitionnelle: espace et temps en pensée et dans les structures de langues. Québec: Presses de l’Université Laval. Guimier, Claude. 1988. Syntaxe de l’adverbe anglais. Lille: Presses universitaires de Lille. Harris, Roy. 2003. Saussure and his Interpreters. 2nd ed. Edinburgh: Edinburgh University Press. Heine, Bernd. 1993. Auxiliaries: Cognitive Forces and Grammaticalization. New York and Oxford: Oxford University Press.

266

Bibliography

Hewson, John. 1972. Article and Noun in English. The Hague and Paris: Mouton. – 1992. “The ideal sentence as a linguistic datum.” Journal of Pragmatics 18, 579–89. – 1997a. The Cognitive System of the French Verb. Amsterdam and Philadelphia: John Benjamins. – and Vit Bubenik 1997b. Tense and Aspect in Indo-European Languages: Theory, Typology and Diachrony. Amsterdam and Philadelphia: John Benjamins. Hill, Archibald A. 1958. Introduction to Linguistic Structures: From Sound to Sentence in English. New York: Harcourt, Brace & World. Hirtle, Walter. 1964. “The English Present Subjunctive.” The Canadian Journal of Linguistics 9, 75–82. – 1975. Time, Aspect and the Verb. Québec: Presses de l’Université Laval. – 1980. “Meaningful Grammar Teaching.” Bulletin CILA 31, 9–20. – 1982. Number and Inner Space. Québec: Presses de l’Université Laval. – 1984. “L’accord en anglais et l’incidence du verbe au sujet.” Modèles linguistiques 6, 99–108. – 1988. “Some and Any: Exploring the System.” Linguistics 26, 443–77. – 1992. “La métaphore: une idée regardante?” ALFA 5, 137–50. – 1994. “Meaning and Referent: for a Linguistic Approach.” Word 45, 103–17. – 1997a. “The Support Role of Auxiliaries.” Linguistica Atlantica 18/19, 75–85. – 1997b. “DO auxiliary – A meaningful Support and Operator.” Lingua 100, 111–49. – 1997c. “Un petit mot: DO. Du verbe ‘plein’ au verbe ‘vide.’” Psychomécanique du langage, problèmes et perspectives. Actes du 7 e Colloque International de Psychomécanique du langage, P. de Carvalho and O. Soutet, eds. Paris: Champion. – 2000. “Number in the English Substantive.” Between Grammar and Lexicon, Ellen Contini-Morava and Yishai Tobin, eds. Amsterdam and Philadelphia: John Benjamins, 59–71. – 2007. Lessons on the English Verb: No Expression without Representation. Montreal: McGill-Queen’s University Press. Huddleston, Rodney and Geoffrey K. Pullum. 2002. The Cambridge Grammar of the English Language. Cambridge: Cambridge University Press. Hudson, Richard. 1984. Word Grammar. Oxford and New York: Basil Blackwell. Humboldt, Wilhelm von. 1836/1988. On Language. Translated by Peter Heath. Cambridge: Cambridge University Press.

Bibliography

267

Jakobson, Roman. 1990. On Language. (Linda R. Waugh and Monique Monville-Burston, eds.) Cambridge, MA, and London: Harvard University Press. Jespersen, Otto. 1951. The Philosophy of Grammar. London: George Allen & Unwin. – 1954. A Modern English Grammar on Historical Principles. London: Allen & Unwin; Copenhagen: Munksgaard. Juul, A. 1975. On Concord of Number in Modern English. Copenhagen: Nova. Kruisinga, E. and Erades, P.A. 1960. An English Grammar, vol. 1, part 2. Groningen: Noordhoff. Lamb, Sydney. 1996. “A Word from the Chair.” LACUS Newsletter 2,1. n.p. Langacker, Ronald W. 1982. “Remarks on English Aspect.” Tense and Aspect, P.J. Hopper, ed. Amsterdam: John Benjamins, 265–304. – 1987a. “Nouns and Verbs.” Language 63: 53–94. – 1987b. Foundations of Cognitive Grammar Vol. I. Stanford: Stanford University Press. – 1991a. Concept, Image, and Symbol: The Cognitive Basis of Grammar. Berlin and New York: Mouton de Gruyter. – 1991b. Foundations of Cognitive Grammar Vol. II. Stanford: Stanford University Press. – 1997. “A Dynamic Account of Grammatical Function.” Essays on Language Function and Language Type. Dedicated to T. Givon, Joan Bybee, John Haiman, and Sandra A. Thomson, eds. Amsterdam and New York: John Benjamins. 249–73. Lehrer, A. and K. 1970. eds. Theory of Meaning. Englewood Cliffs, NJ: Prentice-Hall. Lowe, Ronald. 1996. “The Internal Syntax of the Eskimo Word.” Tollis 1996: 74–99. Lyons, John. 1968. Introduction to Theoretical Linguistics. Cambridge: Cambridge University Press. – 1977. Semantics. Cambridge: Cambridge University Press. Michael, Ian. 1970. English Grammatical Categories and the Tradition to 1800. Cambridge: Cambridge University Press. Miller, George A. 1991. The Science of Words. New York: Scientific American Library. Milsark, G. 1977. “Toward an Explanation of Certain Peculiarities of the Existential Construction in English.” Linguistic Analysis 3, 1–39. Morris, Lori. 2000. “The Grammar of English Gender.” Between Grammar and Lexicon, Ellen Contini-Morava and Yishai Tobin. eds. Amsterdam and Philadelphia: John Benjamin’s, 185–203. Murray, Lindley. 1799. English Grammar. 5th ed. York: Longman and Rees.

268

Bibliography

O’Kelly, Dairine. 1996. “Facts of Discourse and Facts of Language.” Tollis 1996: 57–68. Padley, G.A. 1976. Grammatical Theory in Western Europe 1500–1700. Cambridge: Cambridge University Press. Perkins, M.R. 1983. Modal Expressions in English. London: Frances Pinter. Pierce, C.S. 1934. Collected Papers of Charles Sanders Pierce, Vol. V Pragmatism and Pragmaticism. Charles Hartshorne and Paul Weiss, eds. Cambridge, MA: Harvard University Press. Poutsma, H. 1916. A Grammar of Late Modern English, Part II Section I.B. Groningen: Noordhoff. Pulgram, E. 1972. “The Frailty of Grammatical Concord.” Linguistic and Literary Studies in Honor of Archibald A. Hill, II : Descriptive Linguistics, M.A. Jazayery et al., eds. The Hague: Mouton, 91–100. Quirk, Randolph, Sidney Greenbaum, Geoffrey Leech, and Jan Svartvik. 1985. A Comprehensive Grammar of the English Language. London and New York: Longman. Ramachandran, V.S. and Sandra Blakeslee. 1998. Phantoms in the Brain. New York: William Morrow. Reid, Wallis. 1991. Verb and Noun Number in English: A Functional Explanation. London: Longman. Sadek-Kahil, Denise. 1996. “Psychomechanics and the Pathology of Language.” Tollis 1996: 183–5. (For complete references in French, see Bibliographie, Orthophonie et rééducation du langage at www.fl.ulaval. ca/fgg/.) Sag, Ivan A. 1973. “On the State of Progress on Progressives and Statives.” New Ways of Analyzing Variation in English, Charles-James N. Bailey and Roger W. Shuy, eds. Washington: Georgetown University Press. Sahlin, Elisabeth. 1979. Some and Any in Spoken and Written English. Stockholm: Almqvist & Wiksell. Sapir, Edward. 1921. Language. New York: Harcourt Brace. Saussure, Ferdinand. 1879/1987. Mémoire sur le système primitif des voyelles dans les langues indo-européennes. Hildesheim: G. Olms. – 1916/1955. Cours de linguistique générale. Paris: Payot. Sayce, A.H. 1875. The Principles of Comparative Philology. London: Trübner & Co. Scheffer, Johannes. 1975. The Progressive in English. Amsterdam and Oxford: North-Holland. Shea, William R. and Mariano Artigas. 2003. Galileo in Rome: The Rise and Fall of a Troublesome Genius. Oxford: Oxford University Press. Spencer, Andrew. 1991. Morphological Theory: An Introduction to Word Structure in Generative Grammar. Oxford: Basil Blackwell.

Bibliography

269

Stern, Gustaf. 1931. Meaning and Change of Meaning. (Reprint, n.d.) Bloomington: Indiana University Press. Sweet, Henry. 1898/1958. A New English Grammar Logical and Historical. Oxford: Clarendon Press. Taylor, Charles. 1985. Human Agency and Language. Philosophical Papers 1. Cambridge: Cambridge University Press. Taylor, John R. 1989. Linguistic Categorization: Prototypes in Linguistic Theory, Oxford: Clarendon Press. – 1996. Possessives in English: An Exploration in Cognitive Grammar. Oxford: Clarendon Press. – 2002. Cognitive Grammar. Oxford: Oxford University Press. Tollis, Francis, ed. 1996. The Psychomechanics of Language and Guillaumism, LynX’ 5. Valencia, Spain: Departament de Teoria des Llenguages, 1996. Toussaint, Maurice. 1996. “From the Psychomechanics of Language to Analytical Neurolinguistics.” Tollis 1996: 103–19. Ullmann, Stephen. 1957. The Principles of Semantics. 2nd ed. Oxford: Blackwell. Ungerer, Friedrich and Hans-Jörg Schmid. 1996. An Introduction to Cognitive Linguistics. London and New York: Longman. Valin, Roch. 1981. Perspectives psychomécaniques sur la syntaxe. Québec: Presses de l’Université Laval. – 1994. L’envers des mots. Analyse psychomécanique du langage. Québec: Presses de l’Université Laval; Paris: Klincksieck. – 1996. “The Comparative Method in Historical Linguistics and in the Psychomechanics of Language.” Tollis 1996: 37–45. Vendler, Zeno. 1977. “Wordless Thoughts.” Language and Thought. Anthropological Issues, William C. McCormack and Stephen A. Wurm, eds. The Hague and Paris: Mouton. Waldron, R.A. 1967. Sense and Sense Development. London: André Deutsch. Waldron, T.P. 1985. Principles of Language and Mind. London: Routledge & Kegan Paul. Waugh, Linda R. 1976. Roman Jakobson’s Science of Language. Lisse: Peter de Ridder. Webster’s. 1969. Webster’s Third International Dictionary of the English Language. Springfield, MA: Merriam. Wickens, Mark A. 1992. Grammatical Number in English Nouns. Amsterdam and Philadelphia: John Benjamins. Wierzbicka, Anna. 1988. The Semantics of Grammar. Amsterdam and Philadelphia: John Benjamins. Wilson, Edward O. 1998. Consilience: The Unity of Knowledge. New York: Alfred A. Knopf.

This page intentionally left blank

Index

adjective, 120–2, 124–6, 128, 134. See also incidence, external adverb, 129, 134 agreement. See incidence, predicate to subject articles, 204–8, 134; system of, 207 aspect: immanent and transcendent, 164–9; system of, 170; in usage, 191–4 auxiliaries, 132, 134, 174; as dematerialized verbs, 176–9; do, 179– 81; as helping verbs, 181–3; system of, 184–6; as transpredicative part of speech, 186–8; as verbs, 175–6 binary tensor, 109, 118, 207, 229– 31, 134 Chomsky. See generative grammar chronogenesis. See verb chronothesis. See verb cognitive grammar, 23, 36, 39–42, 60, 67, 95 comparative method, 9–10, 85–92 completive pronoun. See pronoun conjunction, 132, 134

continuum noun, 74–7 conversion, 137–8 count noun. See unit noun dematerialization, 132, 174–5, 177–9, 181, 186 discourse, 8, 12, 24–9, 30–8, 41, 50 event. See verb event time. See verb expression. See representation Guillaume: life, 3–5; Saussure’s influence, 5–8, 10 generative grammar, 18–19, 23–4, 39–40, 101 ideogenesis, 44, 48–9, 116–18; and morphogenesis, 118, 131, 137–8, 229, notional ideation, 44, 130–4. See also meaning, lexical imperfective. See verb import, 48, 96–7, 207–10; grammatical, 15, 136–7, 175, 217, 221; lexical/notional 14–15, 125, 133, 152, 179–81. See support

272

Index

incidence, 121–2, 201–3, 207–12; external, 122, 124–6; internal, 124–6; of predicate to subject, 213–24; referential, 48, 66–7, 114, 124–5, 202–3, 205 intended message, 9, 25–6, 36–7, 47–8, 67 interjection, 135–6

noun. See substantive number in the substantive: -ø morpheme 52, 60–2, 70–3, 81–4; -s morpheme, 52–60, 70–3; system of, 51–2, 70–84, 116–17

Jacobsonian approach, 36, 42, 64, 67, 69

mass noun. See continuum noun meaning: actual vs potential, 7, 13, 41, 58–62, 89; grammatical, 51– 65; lexical vs grammatical, 8–9, 15, 44–6, 49; potential, 68–70, 80, 90. See also import; polysemy metaphase event. See verb monophase event. See verb mood, 158–63; indicative 158–62; quasi-nominal, 160–2; subjunctive, 159–62; system of, 162–3 morphogenesis, 44, 48–9, 106–7, 111–18, 136–8, 229; structural ideation, 44, 131, 133–4, 136. See also ideogenesis; meaning, grammatical

parts of speech, 14, 43–9, 126–7, 256–4; predicative, 129–39, 154, 188; system of, 133–6; transpredicative, 130–5, 154, 175, 181, 186–8 perfect. See aspect; simple form perfective. See verb person support. See verb polysemy, 7, 13, 41–2, 89, 94–5; in articles, 204–5; vs monosemy, 67–9, 97; in number morphemes, 52–62, 100; in simple form, 165– 7; in some and any, 141–50; in transcendent aspect 167–9 positional linguistics, 90–2, 99, 239 positional technique. See positional linguistics preposition, 131, 134 progressive. See simple form pronoun, 133–4; completive, 133, 145, 149, 204, 211; suppletive, 144, 211. See also articles; quantifiers psychogenesis. See word psychomechanisms, 9, 13, 22, 74, 77–80, 150–4 psychosystematics of language, 8, 71–4, 98–9; method of analysis in, 87–105

notional chronology, 89, 101–2, 156, 169, 206 notional ideation. See ideogenesis

quantifiers: any, 141–3; existential import, 150–2; partitives, 145– 50; some, 143–5; system of, 152–4

language: as operations, 10–13, 24–9; potential vs actual, 19–22; and thought, 225–36; as usually conceived, 27. See also discourse; tongue lexeme, 8–9, 25–6, 41–4, 63–4, 72–7, 108–11 lexigenesis, 45–6, 50, 116, 210

operative time, 10, 25, 91, 93, 102–4, 162, 238–9

Index radical binary tensor. See binary tensor reference. See incidence, referential representation: and expression, 9, 26, 46; and the message, 9, 25– 6, 36–42, 63–7; and the word, 13, 38–49, 66–7 rules, 4, 11 Saussure, 10, 248n7; langue and parole, 18–19 sentence, 30–3, 49–50, 139 sign, 5–7, 34–6, 47, 49, 64 simple form, 165–7; vs be + present participle, 194–8; vs do + infinitive, 198–200; vs have + past participle, 191–4 structural ideation. See morphogenesis substantive, 127–8, 134; case in, 111–14, 116–17; common nouns, 115–16; comprehension and extension in, 122–4; extensity of, 123, 203–8; gender in, 107–11, 116–17; ideogenesis in, 114–18; morphogenesis of, 107– 14, 116–18; proper nouns, 114– 15. See also number in the substantive suppletive pronoun. See pronoun support, 133, 154, 165–72, 181–7, 207–23; and import, 113–14, 121–6, 129–32 syntax, 11, 27, 15, 32, 43–5, 47; construction grammar, 14–15; of the noun phrase, 208–11; of subject and predicate, 213–24 system, 7–8, 17, 70–84, 88–92, 134; and usage, 8–10, 12; of wordforming, 15–16, 23, 33, 49, 230–1

273

tense, 158–64, 170–3; absolute and relative, 159–63, 170–1; system of, 172 theory of glossogenetic areas. See typology tongue: and la langue, 7–8, 10, 12, 20–2, 28; as potential language, 23–9, 46; as a viewing universe, 9, 38, 44, 63–4, 232–4 typology, 12, 16, 33, 230–1, 250n9 unit noun, 74–7 universals, 50 universe time. See verb unsayable. See intended message unsayable to sayable. See representation verb, 134, 155; chronogenesis, 155–8, 162, 170, 175, 182–3; chronothesis, 162, 163; event, 156, 160–73; imperfective vs perfective, 164, 196–8; metaphase vs monophase, 197–8, 244; person support, 165–9; present, 158, 257n5; universe time vs event time, 160–4, 170–6, 179, 181, 197. See also aspect; auxiliary; mood; simple form; tense viewing ideas, 64, 94, 137, 227, 234. See also tongue, viewing universe vocable, 12, 35, 249n2 word 30–50, 136–7; construction of, 10–16, 23–4, 44; functions of, 37, 43, 46; psychogenesis of, 114, 117, 210, 229; and syntax, 15. See also conversion; vocable