Six Lectures on Sound and Meaning

8/2/2019 Six Lectures on Sound and Meaning

1/72


2/72

r

~ - - - - - -


3/72

II,

I:

IIIl..

SIX LECTURESon Sound and MeaningROMAN JAKOBSONS. H. Cross Professor Emeritus of Slavic Languagesand Literatures and ofGeneral Linguistics,Harvard University, and Institute Professor,Massachusetts Institute of Technology

with a Preface byCLAUDE LEVI-STRAUSSTranslated from the French byJOHN MEPHAM

The MIT PressCambridge, Massachusetts,and London, England


4/72

- - . "

First MIT Press edition, 1978

ptl;geIX

1

4 5 \ J ~ rI6989

109

ForewordContents

v-Lecture VLecture VI

Preface by Claude Levi-Strauss xiLecture I.Lecture IILecture III

~ , L e c t u r e IVAll rights reserved. No part of this book maybe reproduced in any form or by any means,electronic or mechanical, includingphotocopying, recording, or by anyinformation storage and retrieval system,without permission in writing from thepublisher.Printed and bound in Great Britain.

English translation published in Great Britainby The Harvester Press Limited, 1978.Copyright The Harvester Press Limited,1978. Translated from the French by JohnMepham.Original edition published in France byLes Editions de Minuit, 1976.

Library of Congress catalog card number:78-8728ISBN 0-262-10019-3

Vll

..


5/72


6/72

,,

,

....

PrefaceA BOOK BY Roman Jakobson has no need of a preface,and I would not have presumed myselfworthy of thegreat honour of writing one were it not for the factthat Jakobson himself wished me to contribute heremy witness as a member of his audience, and also, Iwould like to add, as his disciple. For these lectures,now one-third of a century old, which the author hasat last decided to publish, having often before formedthe intention of doing so, the project having beenpostponed each time to make way for more urgenttasks, were the first which I heard as Professor at theNew York Ecole libre des hautes etudes, during thatyear of 1942-3 when we began to attend each other'slectures.Rereading them today I rediscover that intellectualstimulation which I felt thirty-four years ago. At thattime I knew almost nothing about linguistics andJakobson's name was not familiar to me. It wasAlexandre Koyre who enlightened me as to his roleand who put us in touch with each other. Still keenlyaware of the difficulty which, as a result of my inexperience, I had met with three or four years earlierin trying to find an adequate notation to record thelanguages of central Brazil, I promised myself toacquire from Jakobson the rudiments which I lacked.In fact, however, what I received from his teachingwas something quite different and, I hardly need add,something far more important: the revelation ofstructural linguistics, as a result ofwhich I would later

Xl


7/72..

be able to crystallise into a body of coherent ideasvisions inspired by the contemplation of wild flowerssomewhere near the border of Luxemburg at thebeginning of May 1940, and the ambiguous feelings,a mixture of enthusiasm and exasperation, whichsome time later at Montpellier - where, for the lasttime in my life I performed for a while the job ofteacher of philosophy - had been aroused in me byreading Marcel Granet'sLes Categories matrimonialeset relations de proximite dans la Chine ancienne, as aresult on the one hand of the attempt to be foundthere to draw together apparently arbitrary factsinto a system, and on the other hand of the improbably complicated results at which this effortarrived.What I was to learn from structural linguistics was,on the contrary, that instead of losing one's wayamong the multitude of different terms the importantthing is to consider the simpler and more intelligiblerelations by which they are interconnected. Listeningto Jakobson I discovered that nineteenth century, andeven early twentieth century, ethnology had beencontent, like the linguistics of the neogrammarians,to substitute 'strictly causal questions for questionsconcerning means and ends' (p. 35). They were content, without having even properly described aphenomenon, to go back to its origins (p. 6). Thetwo disciplines, therefore, found themselves confronted by 'a stunning multitude of variations',whereas explanation ought always aim at the discovery of 'the invariants behind all this variety' (p.9). What Jakobson said about phonetics was applicable mutatis mutandis equally well to ethnology:

..Xll

It is true that the phonic substance of languagehas been studied thoroughly, and that suchstudies, especially over the last fifty years, haveproduced an abundance of illuminating results.But for the most part the phenomena under consideration have been investigated in abstractionfrom their function. In these circumstances it hasbeen impossible to classify, or even to understand, these phenomena.As for kinship systems, which were the subject ofmy lectures from that year 1942-3, it was to thecredit of those such as van Wouden (whose work Iwas not yet familiar with at that time) and Granetthat they had gone beyond this stage, but they hadstill not risen above focusing their attention on theterms to look rather at the relations between them.

In the absence of this approach they were unable torationally comprehend the phenomena, and weretherefore condemned to the endless task of searchingfor things behind things in the vain hope of reachingsomething more manageable than the empirical datawith which their analyses had to cope. What Jakobson writes here, about the phonic individuality ofphonemes, can be said about any terms whatsoever,real or imaginary: 'The important thing . . . is not atall each phoneme's individual phonic quality considered in isolation and existing in its own right.What matters is their reciprocal opposition within a. . . system' (p. 76).These innovatory ideas, towards which I was nodoubt drawn by my own thought, but as yet withneither the boldness nor the conceptual tools neces-

...XU!


8/72II...

sary to organise them properly, were all the moreconvincing in that Jakobson's exposition of themwasperformed with that incomparable art which madehim the most dazzling teacher and lecturer that I hadever been lucky enough to hear: the present text fullycaptures the elegance and the logical force of hisexposition. It is not the least of the value of thesepages that they testify, for all those who have neverhad the opportunity to hear Jakobson, to what hiscourses and his lectures were like, and what they arestill like now in his eightieth year.

In these lectures the discussion, presented with anoratorical talent which was as great in whateverlanguage Jakobson chose to express himself (eventhough we must assume that it is even greater whenit is displayed in his native language) is developedwith equal limpidity and rigour. Jakobson neverdevelops his abstract and sometimes difficult arguments at length without illustrating them with ex-amples drawn from a wide range of languages, andoften also from poetry and the modern plastic arts.His systematic reference to the great thinkers - Stoics,Scholastics, Renaissance rhetoricians, Indian grammarians, and many others - manifests his constantconcern to place these new ideas in perspective, and toimpress on the mind of his audience a sense of thecontinuity of history and of thought.

In Jakobson the order of exposition follows, stepby step, the order of discovery. His exposition therebyderives a dramatic power which holds his audience insuspense. With a wealth of theatrical effect, at oneminute off on a tangent and the next sweeping rapidlythrough a short-cut, the exposition strides swiftly to.XIV

t

its conclusions, whichare sometimesquite unexpectedand yet which always carry conviction.Taking their place besides those of his works whichwere always intended for publication, these six lectures represent a sample of his oral style which haslost none of its flavour for having been captured inprint. The first lecture gives an account of the state oflinguistics at the end of the nineteenth century. Itargues against the opinions of the neogrammariansfor whom sound and meaning belonged to completely distinct orders. I t gives proper credit to theresults of phonetic research but, by means of a distinction between motor phonetics and acousticphonetics, it shows that it is impossible to divorcesound frommeaning, linguisticmeans from their ends.If sound and meaning are inseparable what thenis the mechanism of their union? In the second lectureJakobson shows that the idea of the phoneme enablesus to resolve this apparent mystery; he defines thisidea, gives an account of its origins and discusses theinterpretations of it which were initially suggested.Continuing along the same track the third lectureintroduces the theory of phonology, based on theprimacy of relations and of the system. I t refuses to

get involved in a debate about the nature of thephoneme, an unnecessary and sterile problem, andvia an actual analysis it demonstrates the specificityof this linguistic entity in comparison with the morpheme, the word and the sentence. The only linguisticentity without conceptual content, the phoneme,which does not itself have a meaning, is a tool whichserves to discriminate between meanings.But this immediately raises two problems, and

xv

-r


9/72

,

I can now, many years later, recognise more clearlythan ever those themes in these lectures which most

The sixth lecture summarises and recapitulates theargument of the whole course. But Jakobson's endings are never merely repetitive. They take thelistener beyond that point at which, he believes, hewill be allowed to rest. In this particular case Jakobson takes him beyond the S a u s s u r i ~ n principle of thearbitrariness of the linguistic sign. jJhe sign does, ofcourse, seem arbitrary when looked at from the pointof view of resemblance, i.e. when we compare thesignifiers of one and the same signified in differentlanguages; but, as Benveniste has shown, it no longerseems arbitrary for each language considered in itself,when looked at from the point of view of contiguity,taking this as a necessary relation between signifierand s i g n i f i e d ~ In the former case the relation is internal, whereas in the latter case it is external. This isthe reason why the speaking subject seeks to compensate for the absence of the former by a recourseto the latter, by conferring on language a phoneticsymbolism. So the union between sound and meaningis once again achieved, this time at a level which, asJakobson shows, has an organic basis, one which wasignored by the traditional phoneticians not so muchbecause they reduced linguistic activity to its physiological substratum - a reduction which was criticisedin the first lecture - but, we can now see, because theywere content with a too superficial understanding ofthis aspect of language.

*

..XVll

*

B.XVI

these form the subject-matter of the fourth lecture.In the first place, it follows from the definition of thephoneme as a discriminative value that phonemesperform their function not by virtue of their phonicindividuality but by virtue of their reciprocal oppositions within a system. However, no logical relationcan be discovered between phonemes standing inopposition to each other: the presence of one of themdoes not necessarily evoke the other. In the secondplace, if the relations of opposition betweenphonemesconstitute the primary values which enable meaningsto be differentiated, how can it be that these relationsare much more numerous than are the phonemeswhich derive from them? Jakobson shows that thesetwo paradoxes both originate in an incorrect conception according to which phonemes are indivisibleelements. In fact as soon as they are analysed intodifferential elements we reach new kinds of relations,which on the one hand have the character of logicaloppositions and which, on the other, are in all languages fewer in number than the phonemes generatedby the different combinations of these oppositions.The fifth lecture illustrates these theoretical ideasby giving a description and an analysis of the consonantal system of French. This also affords the opportunity to deepen the idea of combinatory variation,and to resolve in a positive way the problem of thephoneme's operation on the two axes, of simultaneityand succession. This demonstration results in partfrom an original treatment of the idea of moraewhich, I recall, was to delight Boas shortly before hisdeath, during a dinner at his house to which bothJakobson and I had been invited.

..... .,


10/72L.

influenced me. However much ideas such as those ofthe phoneme and of the prohibition of incest might iseem incongruous, the conception which I was toform of the latter was inspired by the role assignedby linguists to the former. Just like the phoneme,which though it has no meaning of its own serves asa means bywhich meanings can be formed, the incestprohibition seemed to me to be the link connectingtwo domains hitherto held to be divorced from eachother. To the articulation of sound with meaningthere would thus correspond, on another level, thatof nature with culture. And just as the form of thephoneme is the universal means, in all languages,whereby linguistic communication is established, sothe incest prohibition, which, if we limit ourselves toits negative expression, is also found universally, alsoconstitutes an empty form which is nevertheless indispensable if the articulation of biological groupsinto a network of exchanges whereby they canestablish communication is to be both possible andnecessary. Finally, the meaning of marriage rules,which is incomprehensible when they are investigatedin isolation, can only emerge by seeing them as mutualoppositions, in the same way that the true nature ofthe phoneme does not lie in its phonic individualitybut in the oppositive and negative relations in whichphonemes stand to one another.'Saussure's great merit' , says Jakobson, 'was tohave understood clearly that [. . . ] something extrinsicis unconsciously brought into play' (p. 10). I t cannotbedoubted that these lectures also make an importantcontribution to the human sciences by emphasisingthe role played in the production of language (but

XVlll

also that of all symbolic systems) by the unconsciousactivity of the mind. For it is only on condition thatwe recognise that language, like any other socialinstitution, presupposes mental functions whichoperate at the unconscious level, that we can hope toreach, beyond the continuity of the phenomena, thediscontinuity of those 'principles by which languageis organised' (p. 11), and of which the subject whospeaks or thinks is not normally consciously aware.With the discovery of these principles, and especiallyof their discontinuity, linguistics and the other humansciences should find the way open to them to makerapid progress.This point is important because doubt has sometimes been expressed as to whether phonologicaltheory since its inception, and in particular inTrubetzkoy, really has implied a shift to underlyingunconscious structures. Yet one only has to look atthe critique of Scerba given here by Jakobson to seethat i t agrees on all points with that of Trubetzkoy,which is not at all surprising when one recalls justhow closely related they were to each other in theirthought: 'Scerba and someotherdisciples ofBaudouinde Courtenay', writes Jakobson, '. . . appealed to thelinguistic intuition of the speaking subject' (p. 38),because they did not understand that 'the elements oflanguage usually remain beneath the threshold ofour conscious deliberation. As the philosophers say,linguistic activity takes place without self-knowledge'(p. 39). And Trubetzkoy: 'The phoneme is an ideabelonging to linguistics and not to psychology. In thedefinition of the phoneme we must reject any reference to "linguistic intuition'" (Principes de phono-

XlX


11/72IlL.

logie, p. 42). The dissociation of the phoneme into't,distinctive features, adumbrated by Trubetzkoy but::achieved for the first time by Jakobson in 1938, should,make it possible once and for all, 'quite objectively "1and unambiguously', to reject any reliance on 'the '.subjective intuition of speakers' (p. 85). The dis- ,criminative value of these features constitutes theprimary fact, and our more or less conscious awareness of these elements is never any more than asecondary phenomenon (p. 38).There is only one aspect of these lectures on which iJakobson would probably no longer agree with the 'positionwhich he heldmore than thirty years ago. In j1942-3 he believed, quite rightly at that time, that he lcould say that ' language is the only system which iscomposed of elements which are signifiers and yet atthe same time signify nothing' (p. 66). Since then,there has been a revolution in biology with the dis-'covery of the genetic code, a revolution of which the!theoretical consequences cannot fail to have adramatic impact on the human sciences. Jakobsonunderstood this immediately: he was one of the firstto recognise and to elucidate 'the extraordinary:degree of similarity between the genetic information !system and that of verbal information' (Essais delinguistiquegenerale, II , 'Rapports internes et externesdu langage', Paris, Editions de Minuit, 1973, p. 51).After having listed 'all the characteristics which areisomorphous between the genetic code . . . and thearchitectonic pattern which underlies the verbalcodes of all human languages' (Essais, p. 54), he goesone step further and raises the question of whether'the isomorphism of these two different, genetic and

xx

verbal, codes can be explained by a simple convergence stemming from a similarity ofneeds, or whetherthe foundations of the manifest linguistic structures,intimately based on molecular communication, arenot directly patterned on the structural principles ofthe latter' (Essais, p. 55).This is a vast problem, and onewhich collaborationbetween biologists and linguists will perhaps make itpossible to resolve one day. But are we not already atthe present time in a position to state and to resolvea problem located at the other end of the hierarchyof linguistic operations, a problem of the same kindthough of infinitely more modest significance? Wehave in mind the problem of the relations betweenlinguistic analysis and the analysis of myths. Thisproblem involves the other side of language, thatwhich is oriented towards the world and societyrather than towards the organism, and here we findthe same problem of the relation between languageand another system (closer to language in this case,of course, since it necessarily makes use of language),a system which, in a different way from language, iscomposed of elements which are combined togetherto form meanings without in themselves, consideredin isolation, signifying anything.In the third lecture Jakobson shows that, contraryto the view of Saussure, phonemes differ from otherlinguistic units - words and grammatical categories in that they have a set of characteristics which are notfound altogether in any other unit. Grammat ica lcategories do, of course, have in common withphonemes that they are oppositive and relativeentities, but unlike the latter they are never negative;

XXI

r-'


12/72.....

in other words their value is not purely discriminative:each grammatical category taken in itself bears asemantic value which is discernible by the speakingsubject (p. 64). Now, the question can be raised as towhether all the characters of the phoneme do not .reappear in those entities which we have called 'my- 'themes': these are the elements from which mythicdiscourse is constructed, and they also are entitieswhich are at one and the same time oppositive, relative and negative; they are, to use the formula applied.by Jakobson to phonemes, 'purely differential and ,1contentless signs' (p. 66). For we must always dis- ,itinguish the meaning or meanings which a word hasin the language from the mytheme which this wordcan denote in whole or in part. In everyday languagethe sun is a heavenly body which appears in the day- ..time; but the mytheme 'sun' does not, taken in andfor itself, have any meaning. Depending on the particular myths under consideration it can range overa whole variety of different ideal contents. In factnobody, coming across 'sun' in a myth, would be ableto say in advance just what its specific content, natureor functions were in that myth. Its meaning couldonly be identified from the relations of correlationand opposition in which it stands to other my themeswithin this myth. The meaning does not, properlyspeaking, belong to any individual mytheme: it is aconsequence of their combination.We are aware of the risks that we run in seeking toindicate the formal correspondences between linguistic entities and those which we believe to be broughtto light by the analysis of myths. The latter do, ofcourse, belong to language, but within language they

XXll

constitute a separate order because of the principlesby which they are governed. In any case itwould be aserious mistake to believe that we take the mythemeto be of the same order as the word or the sentence,for these latter are entities of which the meaning ormeanings can be specified, albeit only ideally (foreven the meaning of a word can vary with context),and these meanings can be listed in a dictionary. Theelementary units of mythic discourse do, of course,consist of words and sentences, but these, in thisparticular usage of them, and without wishing topush the analogy too far, are rather of the order ofthe phoneme, in that they are units which, while theyhave no meanings of their own, do make it possibleto generate meanings in a system in which they standin opposition to each other, and this precisely as aresult of these oppositions.The relations between myth and language can bedefined by saying that statements in the discourse ofmyth reproduce the structure of language but onlybecause there is a shift of gear which disengages itfrom its normal operation: the basic elements ofmyth function like those of language, but they arefrom the start more complex in nature. As a result ofthis complexity mythic discourse becomes, in amanner of speaking, detached from the normal usageof language in such a way that there is only occasionally any precise correspondence in the results generated by the combination of these elements of differentorders. In contrast to a linguistic utterance whichcommands, questions or informs, and which can beunderstood by every member of the culture or subculture as long as they know the utterance's context,

...XXlll


13/72

a myth never presents a specific meaning to those whohear it. A myth sets up a grid, solely definable in 'terms of the rules by which it is constructed. For the .members of the culture to which the myth belongs .this grid confers a meaning not on the myth itselfbuton everything else: i.e., on the picture they have of }'"he world, on the society and its history about w h i c h ~ ithe group members might be more or less accurately;'informed, and on the ways in which these things areproblematic for them. Normally these diverse factsfail to hang together and more often than not theyclash with one another. The matrix of intelligibilityprovided by the myth allows them to be connected"up into a coherent whole. I t is worth noting in pass- 'ing that this role which we are attributing to mYth i

leads on directly to that which a Baudelaire mighthave attributed to music.Do we not also find here - albeit at the other end

of the scale - a phenomenon similar to that 'soundsymbolism' to which Jakobson devotes much of hissixth lecture? Even i f i t derives from 'the neuropsychological laws of synaesthesia' (p. 113), or evenprecisely because of these laws, this symbolism isitself also not necessarily the same for everyone.Poetry has at its disposal many means for overcomingthe divergence, deplored by Mallarme, between thesound and the meaning of the French words jour'day' and nuit 'night'. But i f I might be allowed tocontribute here my own personal testimony, I confessthat I have never discerned this divergence as such:it only makes me conceive of these two periods indifferent ways. For me the day is something whichhas duration, the night something which is produced

.XXIV

or which comes about, as in the expression 'the nightfalls'. The former denotes a state, the latter an event.Instead of perceiving a contradiction between thesignifieds and the phonic particularities of theirrespective signifiers, I unconsciously take the signifieds to be different in nature from each other. Jourhas a durative aspect, congruent with vocalic gravity,nuit a perfective aspect, congruent with vocalic acuteness: which is, in its own way, a little mythology.We encounter at the two poles of language thisemptiness of which Jakobson speaks, and which callsout for some content to fill it. However, from onepole to the other the relations which are respectively /'present and absent are reversed. At the lowest level Vof language the relation of contiguity is present,whereas that of resemblance is lacking. In contrast,at that level which could be called hyperstatic (because there are evidenced there properties of a neworder) where myth bends language to its own ends, itis the relation of resemblance which is present unlike words, the myths of different peoples resembleeach other - whereas the relation of contiguity disappears because, as we have seen, there is no neces-sary relation between myths as signifiers and the concrete signifieds to which they can come to be applied.Yet in the one case as in the other these relationscan be complemented in a way that is neither inevitable nor determined in advance. At the lowest level,where language is under the direct sway of neuropsychologicallawswhich represent patternsofcerebralactivity between which homologies exist, sound symbolism can be expressed. At the highest level, in thatregion where myth, having transcended language,

xxv


14/72

Ir

,I'iiI

latches on to external r e a l i t ~ , we can ~ e the a p p e ~ r .. ,iance this time of a semantIc symbolism. Yet while:they are at opposite ends of the scale over which llinguistic functions are ranged, these two symbolisms, "the one phonetic and the other semantic, present a,clear symmetry. They each reflect mental necessities'>,of the samekind, oriented in the one case towards the'body and in the other towards society and the w o r 1 d ~ ,Jakobson might not find these potential extensionsof his theoretical thought acceptable, but in any case;they are a measure of the breadth of the domain'which he has opened up for research, and of t h e ~fertility of the principles which, thanks to him, can 'e,henceforth guide this research. Although they date (from many years ago these lectures are more than a I!mere illustration of the state of a science at somemoment in its history. Today as yesterday they b r i n g ~ 1to life a great adventure of the mind, of which the 'products have not ceased to appear in Jakobson'sown work, which is still striding ahead, and amongall those, whether linguists or specialists in other 'disciplines, to whom he has shown the way and whom,he continues to inspire.

CLAUDE LEVI-STRAUSS

XXVI

II AM SURE you are familiar with Edgar Allan Poe'sfamous poem The Raven, and with its melancholy refrain, 'Nevermore'. This is the only word uttered bythe ominous visitor, and the poet emphasises that'what it utters is its only stock and store'. Thisvocable, which amounts to nomore than a few sounds,is none the less rich in semantic content. I t announcesnegation, negation for the future, negation for ever.This prophetic refrain is made up of seven sounds seven, because Poe insists on including the final rwhich is, he says, 'the most producible consonant'. I tis able to project us into the future, or even intoeternity. Yet while it is rich in what it discloses, it iseven richer in what it secretes, in its wealth of virtualconnotations, of those particular connotations whichare indicated by the context of its utterance or by theoverall narrative situation. Abstracted from its particular context it carries an indefinite range of implications. ' I betook myself to linking/ fancy untofancy', the poet tells us, 'thinking what this ominousbird of yore -/ What this grim, ungainly, ghastly,gaunt, and ominous bird of yore/ Meant in croaking"Nevermore"./ This 1 sat engaged in guessing . . .This and more 1 sat divining. . . . ' Given the contextof the dialogue the refrain conveys a series of differentmeanings: you will never forget her, you will neverregain peace of mind, you will never again embraceher, 1will never leave you! Moreover this same word

1


15/72

can function as a name, the symbolic name whichpoet bestows upon his nocturnal visitor. "Yet this expression's value is not entirely accoun' 'for in terms of its purely semantic value, narrodefined, i.e., its general meaning plus its continge: .contextual meanings. Poe himself tells us that it ~ the potential onomatopoeic quality of the sounds 'the word nevermore which suggested to him its ass(i'ciation with the croaking of a raven, and which w/even the inspiration for the whole poem. Alalthough the poet has no wish to weaken the sness, the monotony, of the refrain, and while he Jpeatedly introduces it in the same way ('Quoth tlfraven, "Nevermore" ') it is nevertheless certain tb'variation of its phonic qualities, such as modulatiof tone, stress and cadence, the detailed articulatiof the sounds and of the groups of sounds, that suet,.variations allow the emotive value of the word to "quantitatively and qualitatively varied in all kindsways.The utterance of Poe's refrain involves only a vesmall number of articulatory motions - or, to lookat this from the point of view of the acoustic rathe:than the motor aspect of speech, only a small numbe:of vibratory motions are necessary for the word to "heard. In short, only minimal phonic means are r,quired in order to express and communicate a wealtb)of conceptual, emotive and aesthetic content. H e r ~we are directly confronted with the mystery of theidea embodied in phonic matter, the mystery of the!word, of the linguistic symbol, of the Logos, a mys-\tery which requires elucidation. ,Of course, we have known for a long time that aj

2

word, like any verbal sign, is a unity of two components. The sign has two sides: the sound, or thell1aterial side on the one hand, and meaning, or theintelligible side on the other. Every word, and moregenerally every verbal sign, is a combination of soundand meaning, or to put it another way, a combinationof signifier and signified, a combination which hasbeen represented diagrammatically as follows:

But while the fact that there is such a combination isperfectly clear, its structure has remained very littleunderstood. A sequence of sounds can function asthe vehicle for the meaning, but how exactly do thesounds perform this function? What exactly is therelation between sound and meaning within a word,or within language generally? In the end this comesdown to the problem of identifying the ultimatephonic elements, or the smallest units bearing signifying value, or to put this metaphorically, it is a matterof identifying the quanta of language. In spite of itsfundamental importance for the science of languageit is only recently that this set of problems has at lastbeen submitted to thorough and systematic investigation.

3

......


16/72

"

HI

III,

I t would certainly be wrong to ignore the b r i l l i a n ~insights concerning the role of sounds in languag,which can be found scattered through the work of tbthinkers of Antiquity and of the Middle Ages, foexample those of Thomas Aquinas, who was amon "the most profound of philosophers of language: anit would equally be wrong to ignore the subtle observations of the ancient oriental, and above all Hindu;lgrammarians. But it is only in the last two centuries'that our science has devoted itself really energeticall "to the detailed study of linguistic sounds.This interest in linguistic sounds derived at first:from essentially practical objectives, such as singin,technique or teaching thedeaf and dumb to speak: orelse phonation was studied by physicians as a com-;\plex problem in human physiology. But during the';nineteenth century, as linguistics gained ground, itwas this science which gradually took over researchinto the sounds of language, research which came to 1be called phonetics. In the second half of the nine- ,jteenth century linguistics became dominated by the!most naive form of sensualist empiricism, f o c u s i n g ' ~directly and exclusively on sensations. As one would jexpect the intelligible aspect of language, its signify- 'ing aspect, the world of meanings, was lost sight of,was obscured by its sensuous, perceptible aspect, bythe substantial, material aspect of sound. Semantics, :1or the study of meaning, remained undeveloped,while phonetics made rapid progress and even cameto occupy the central place in the scientific study oflanguage. The neogrammarian school of thought,which was the most orthodox and characteristiccurrent of thought in linguistics at the time, andwhich

4

was dominant in the last quarter of the nineteenthcentury and up to the First World War, rigorouslyexcluded from linguistics all problems of teleology.They searched for the origin of linguistic phenomenabut obstinately refused to recognise that they aregoal-directed. They studied language but neverstopped to ask how it functions to satisfy culturalneeds. One of the most distinguished of the neogrammarians, when asked about the content of theLithuanianmanuscript which he had beenassiduouslystudying, could only reply with embarrassment, 'Asfor the content, I didn't notice it'. At this time theyinvestigated forms in isolation from their functions.And most important, and most typical of the schoolin question, was the way in which they regardedlinguistic sounds; in conformity with the spirit of thetimetheirview wasa strictlyempiricistand naturalisticone. The fact that linguistic sounds are signifiers wasdeliberately put aside, for these linguists were not atall concerned with the linguistic function of sounds,but only with sounds as such, with their 'flesh andblood' aspect, without regard for the role they play inlanguage.Linguistic sounds, considered as external, physicalphenomena have two aspects, the motor and theacoustic. What is the immediate goal of thephonatoryact? Is it the acoustic phenomenon or is it the motorphenomenon itself? Obviously it is the acousticphenomenon which the speaker aims at producing,and it is only the acoustic phenomenon which isdirectly accessible to the listener. When I speak it is inorder to be heard. Of the two aspects of sound it is,therefore, the acoustic aspect which has intersubjec-

5


17/72

I

tive, social significance, whereas the motor ph ,menon, in other words the workings of the VI 'apparatus, is merely a physiological prerequisite):,the acoustic phenomenon. Yet phonetics in the n.grammarian period concerned itself in the first p" ,with the articulation of sound andnotwith its acou:aspect. In other words it was not strictly speaking .sound itself but its production which was the focusattention, and it was this which formed the basisthe description and classification of sounds. Thisspectivemay seem odd or even perverse to us, but it"not surprising in the context of neogrammarian dtrine. According to this doctrine, and to all otb "which were influential in that period, the genetic pespective was the only one considered acceptable. Tbchose to investigate not the object itself but the coditions of its coming into being. Instead of describ'the phenomenon one was to go back to its ori, .Thus the study of linguistic sounds was replacedhistorical phonetics, i.e., by a search for their prottypes in earlier forms of each given language, w .-,so-called static phonetics was more or less entirelgiven over to the observation of the vocal apparatand its functioning. This discipline was incorporatinto linguistics in spite of the obviously h e t e r o g e n e o u ~character of the two domains. Linguists tried to pick!up a bit of physiology with results that are well illus...'trated by the following typical example: Edward W.Scripture, a famous phoneticianwho also had training,as a physician, ironically quotes the currentdescriptionof a particular laryngal articulation which would, had;this description been accurate, have inevitably resulted in the fatal strangulation of the speaker! But

even disregarding mistakes like this we can ask whatresults would t h study of linguistic sounds in theirJl1otor aspect arnve at.At first, even though linguists attempted to discusssounds in a strictly naturalistic manner and to scrupulously leave aside the problem of the functions theyperform in language, they did in fact unconsciouslyemploy properly linguistic criteria in their classifications of sounds, and especially in their demarcationof sounds in the speech chain. This illicit importationwas facilitated by the fact that linguists, and psychologists too, were as yet quite unfamiliar with the role ofthe unconscious, and in particular with its greatimportance in all linguistic operations. But as theobservation of phonatory acts was improved and asthe employmentof special instruments came to replacereliance on purely subjective experience, the linguisticcorrelate of the physiological phenomena was increasingly lost sight of.It was towards the end of the century that instrumental phonetics (or as it was usually but less accurately called 'experimental phonetics') began to makerapidprogress. Withthe helpofincreasinglynumerousand improved instruments a remarkable precisionwasachieved in the study of all the factors involved inbuccal articulation and in the measurement of expiration. A new era in the physiological investigation oflinguisticsoundswas openedup byX-rayphotography.X-rays, used in conjunction with sound film, revealedthe functioning of thevocal apparatus in all its details;the whole of sound production, the entire phonatoryact, was uncovered and could be actually seen as ithappened. When this method became practically and

I 6 c 7IIII


18/72

;1,I

I,I';(

1I.

Ii! l."

technically available to phoneticians a large nwnof the previous phonetic instruments became red"dant.I t was radiography above all which brought to Ii .. Ithe crucial role of the posterior parts of the v "apparatus, parts which are most hidden and whic::were until then most inaccessible to the availabmethods of experimental phonetics. Before the arri'of radiography there was, for example, very l itaccurate knowledge of the functioning in the proof the phonatory act of the hyoid bone, of the ep:"""glottis, of the pharynx, or even of the soft palate. Tb,importance of these parts, and especially of tbpharynx, was suspected, but nothing about them w.known in detail. Remember that the pharynx is atcrossroads fromwhich leads off, at the top, the p a s s a ~to the mouth cavity and the passage to the nascavity, and below, the passage to the larynx. Each 0these upper two passages is opened or closed by tb .velum whereas the lower passage, to the larynx, . 'opened or closed by the epiglottis. I t was only a fe'dozen years ago that one could read on the subject 0_;the pharynx, in the text-book of Ludwig Sutterlin, ai.well-known linguist and phonetician: 'The pharynx:'"seems to be very important in sound production, in;lthatit can be narrowed andwidened, but at thepresenttime nothingmore definite is known with certainty onthe subject' (Die Lehre von der Lautbildung, Leipzig,t,1908).As a result especially of recent work by Czech and 'Finnish phoneticians using radiography we do now.have a more adequate understanding of the function- .ing of the pharynx in phonation, and we can now ,

8

affinn that the phonetic role of this organ is no lessitnportant than, for example, that of the lips, whichare in some ways analogous to it. I t can be seen fromthese more recent observations that so long as thephysiological investigation of sounds had no grasp ofthe functioning of the pharynx and of contiguousparts, it was only possible to arrive at a fragmentaryand unsatisfactory description. A physiological classification of sounds which scrupulously takes intoaccount the varying degrees of opening of the mouthbutwhich fails to consider thevarying degrees ofopening of the pharynx can lead us into error. If phoneticians concentrated on the functioning of the lips andnot on that of the pharynx this was not because theformer had been shown to be the more important. I fthe physiology of sound production were to refuseto draw on other disciplines it would have no way ofestablishing the relative importance of the variousorgans involved. If phoneticians, in classifying linguistic sounds, took the labial factor but not thepharyngal factor into account, this was solely becausethe former was more accessible to observation thanthe latter. As it broadened the field of inquiry and as itbecame an increasingly precise discipline, the autonomous investigation of phonation decomposed thesounds which it analysed into a disconcerting multitude of detail without, however, being able to answerthe fundamental question, namely that of the valuewhich is assigned by language to each of these in-numerable details. In its analysis of the various soundsof a language, or of several languages, motor phoneticsuncovers for us a stunningmultitude of variations, butit has no criterion for distinguishing the functions and

9

....


19/72

i"

the degrees of relative significanceofall these observlvariations, and thus has no way of discovering tb"invariants among all this variety.Now the identification of individual sounds b ' phonetic observation isan artificialway ofproceedin ."To the extent that phonetics is concerned exclusivelwith the act of phonation, that is wi th the prodution of sounds by the various organst it is not in ,position to accomplish this, as Ferdinand de SaussUfl'had already made clear. In his Cours de linguistiqgenera/e, given between 1906 and 1911 and edited afte:his death (1913) by his pupils Charles BallyandAlbeSechehayet and published in 1916, the great linguissaid with foresight: 'Even i f we could record onall the movements of the mouth and larynx in pro..ducing a chain of sounds it would still be impossible t ....'discover the subdivisions in this sequence of a r t i c u ~ ilatory movements; we would not know where oneilsound began and where another ended. Withoutlacoustic perception how could we assertt for example,that inial there are three units and no t two or four?'Saussure imagined that hearing the speech chainwould enable us to directly perceive whether a sound ':had changed or had remained the same. But subsequent investigations have shown that it is not the 'acoustic phenomenon in itsel f which enables us to ,subdivide the speech chain into distinct elements;only the linguistic value of the phenomenon can do ;this. Saussure's great merit was to have understoodclearly that in the study of the phonatory act, whenweraise the question ofphonetic units and that ofdemar-

Published in English as Course in General Linguistics, trans.Wade Baskin, London, 1960.10

eating the sounds in the speech chain, somethingextrinsic is unconsciously brought into play. Twentyyears after his death the film that Saussure would haveliked to have seen was in fact made. The Germanphonetician Paul Menzerath made an X-ray soundfilIn of the workings of the vocal apparatus, and thisfilIn completely confirmed Saussure's predictions.Drawing on this film and on the latest results ofexperimental phonetics Menzerath and his Portugueseassociate Armando Lacerda demonstrated that theactofspeechis a continuous, uninterruptedmovement(Koartikulation, Steuerung undLautabgrenzung, 1933).Whereas traditional doctrine had distinguished between positional sounds, which are held steady, andtransitionalsoundswhich lack this stability and whichoccur in the transition from one position to another,these two phoneticians showed that all sounds are infact transitional. As for the speech chain, they arrivedat an even more paradoxical conclusion. From astrictly articulatory point ofview there is no successionof sounds. Instead of following one another the soundsoverlap; a sound which is acoustically perceived ascoming after another one can be articulated simultaneously with the latter or even in part before it.However interesting and important the study of linguistic sounds in their purely motor aspect may beeverything indicates to us that such a study is no morethan an auxiliary tool for linguistics, and thatwemustlook elsewhere for the principles by which the phonicmatter of language is organised.Even though they focused on the motor aspect oflanguage, phoneticians were nevertheless unable toignore the quite obvious, indeed tautological, fact

11


20/72

I:i:

,,I .I,,L

that sound as such is an acoustic phenomenon.they believed that the investigationofthe productionA .sound, rather than of the sound itself, gave one 'motor equivalent of the acoustic phenomenon,equivalent which is more accessible, more i n s t r u c t ~ iand open to more profitablemethods of analysis.viewwasput forward, for example,byPierre R o u s s e l ~ 'They assumed that there is a one-to-one correspdence between the two aspects and that the classifi,tion of motor phenomena has an exact equivalentthe classification of acoustic phenomena. Thus 0need onlyconstruct the former, since the latter follo .automatically from it. Now this argument, whichbeen put forward timeand again rightupto thepresday, and which has many implications for the scienl"of linguistics, is utterly refuted, contradicted byfacts. Arguments against this position were put fo:'ward long ago, even before the very first hand-booon phonetics.We can mention, in the first place, a French boodating from 1630, which was called Aglossostom,graphie ou description d'une bouche sans langue qu'eLparle et fait naturellement toutes ses autres fonction[Aglossostomography, or the description of a tongul.."less mouth which speaks and naturally perfoall its other functions]. In 1718 Jussien published 1.the Memoires de l'Academie royale des sciencestreatise called 'Sur la fille sans langue' [On the .with no tongue]. Each of these works contained .detailed description of people who, though they halonly rudimentary tongues, were capable of an im,peccable pronunciation of all the sounds which .phonetics nowadays are called the 'l inguals', an

12

which are defined as sounds the emission of whichnecessarily involves the tongue. These interesting factshave since then been confirmed many times. Forexample, at the beginning of this century the physicianHermann Gutzmann, who was one of the best knownof researchers in the field of errors of pronunciation,was forced to admit thatwhile in Frenchthe very sameword (langue) is used to designate a part of themouth(the tongue) and language itself, in fact as far as thelatter is concernedthe former isdispensable, for almostall the sounds which we emit can be produced i fnecessary in quite a different way without the acousticphenomena being altered at all (Des Kindes Spracheund Sprachfehler, Leipzig, 1894). I f one of the phonatory organs is missing then another one can functionin its place, without the hearer being aware of this.Gutzmann, however, stated that there are exceptionsto this. Thus the sibilants - the fricatives z, s, and thecorresponding affricates - require the involvement ofthe teeth. Subsequent research, however, has shownconclusively that these apparent exceptions are not infact so at all. Godfrey E. Arnold, director of theVienna clinic for language disorders, has shown(Archiv fiir gesamte Phonetik, III, 1939) that evenwith the loss of the incisors the ability to pronouncethe sibilants correctly remains intact as long as thesubject's hearing is normal. In cases where dentalabnormality gives rise to errors of pronunciation onealways finds that the subject's hearing is impaired, andit is this that prevents the functional compensationfor the anatomical abnormality.Christoph Hellwag, who was one of the famouspioneers ofmotor phonetics and the discoverer of the

13


21/72

I:

II

i'

~ ,

I,

vocalic triangle, reported and discussed this impodiscovery in his treatise De formatione loq:(Tiibingen, 1781). At the beginning of his book'suddenly poses the following problem: if we owe 'faculty of speech to our art iculatory organs, how .it possible for the serpent, which lacks these organs,ltalk with Eve? Hellwag's strange question can beplaced by another, which though basically equival, ,to it is more empirical. Phonetics seeks to deducesounds ofour language from the various kinds of ~ tact between the tongue and the palate, the teeth,lips, etc. But if these various points of articula-' ..'were in themselves so essential and crucial then bwouldit bepossible for a parrotto faithfully reprod' ",so many of the sounds of our language in spite of '_fact that its vocal apparatus is so dissimilar to ounAll these considerations lead us to a conclusion whilis both simple and yet ignored in the vast majorityworks on phonetics. We cannot classify, nor even gia precise description, of the various articulations,less we constantly hold in mind the question: whatthe acoustic function of such and such a motor pformance?Thus, in investigating the consonants phoneticia ,have carefully noted the points where they are artie'lated, and in classifying the consonants they haranged them all on a line according to the positionthese points in the mouth: first the labials, then tndentals, next the palatals and finally the velar con"sonants ofwhich the point of articulation is behind tbhard palate. For a long time the fact that velar con'sonants change into labial consonants and vice versremained incomprehensible, or was explained by

14

lIlystical formula - extreme opposites come together.But ifwe give up treating the points of articulation asif they were independent variables and ask insteadwhat their purpose is, then we realise immediately thatthey functionmainly by forming two different types ofresonator. Both the labial consonants, articulated atthe lips, and the velar consonants, articulated at thesoftpalate, involve the formation of a long, continuousresonator; on the otherhand, in articulating the dentaland the palatal consonants the tongue divides theresonator, the mouth cavity, into two short compartments. Now, as we know from general acoustics, thepitch of the resonance increases as the resonatingcavity decreases. Thus, what the velar and the labialconsonants have in common is the length of theresonator and the consequentgravity of the resonance.Facts such as the change from lact- to lapt- and fromdirect to drept in Roumanian, which were for a longtime considered mysterious, are thereby explained.Moreover, the possibility of achieving the same acoustic effect by different articulatory means (and, inparticular, functional compensations for anatomicalabnormalities in the phonatory organs) permits us andencourages us to look for the common characteristicsof these dissimilar articulations which determine theidentity of their acoustical effects and which discloseto us the essence of the articulations, their pertinentaspects.Now, acoustic phonetics is not a recent invention.Since the middle of the nineteenth century physicistshave beeninterested in languagesounds, and especiallyin the acoustics of the vowels. But in contrast to thestudy of articulation, acoustic phonetics had no in-

15

III


22/72

"i

II,

'j:

~ I ; : '

fluence at all on tradit ional linguistics, andin partie'lar it made no mark on the work of the neogramarians. This is explained in part, as we have alread'indicated, by the genetic orientation of linguistics, an,'"in part bythe uncertain and over-speculative characof the first attempts at linguistic acoustics. But dudthelasttwenty or thirtyyears the acousticinvestigatio 'of linguistic sounds has made rapid, or one might ev,say miraculous, progress. Many things have permiUthis development, in particular the perfecting 01methods of empirical description in modern psy.'chology and phenomenology, illustrated notably .the fundamental works of Wolfgang Kohler CAkus,tische Untersuchungen', Zeitschrift fur Psychologie1910-15) and Carl Stumpf (Die Sprachlaute, B e r l i n ~1926). We can now draw upon research in acoustit 'whichwas performed in connectionwith developmenin telephonic communication, radio and soundin Europe and, above all, in America; and we can usc,the new precision instrumentswhich this research gave!rise to, in particular sound spectrographs, oscillo-,:graphs, etc. Thanks to the telephone, the gramophone"and above all radio, we have become accustomed to ,;hearing speechin the absence of a speaker.The phona"1tory act is overshadowed by its phonic products, andit is increasingly toward the latter that people tumtheir attention. tiiWhereas the study of articulation, even though it: 'was called 'experimental phonetics', remained at least'for the most part merely observational, modern '"acoustics by contrast employs a wide range of experi..mental methods.Sound is filtered, someof its elements,can be deliberately removed, it can be decomposed:

16

and recomposed. In the eighteenth century attemptswere made by the forerunners of modern phonetics tobuild a speaking machine by copying our vocalapparatus; but nowadays linguistic sounds can beinritated by producing their various acoustic components using special instruments. We have nowsucceeded in artificially reproducing i f not a homunculus then at least the phonic substance of his speech.For the first time we can hear human sounds not pro-duced by human beings. And this is not the furthestthat experimentation in acoustics can go in thisdirection. Sound film promises to takeus muchfurtherstill. The physical aspect of sounds, i.e., the complexvibratory motions produced in the air by the organs ofspeech, are now reproduced in an optical form on thereel of sound film. As anyone who has had the opportunity to examine closely a reel of sound film knows,each linguistic sound imprinted there has its ownspecific optical character. These are so distinctive thatfilm workers can learn to read the dialogue of a filmsimply from the reel of film itself. When the film isprojected these visual images of the sounds oncemoreturn into acoustic phenomena. This process opens upmany possibilities for phonetic experimentation. Aknowledge of the visual representation of each soundmakes it possible to directly draw the speech and thento transform it, via film, into an audible phenomenon.It is thereby made possible to hear speech which hasnever been uttered by anyone. And there is no need tobe limited to slavishly imitating sounds we alreadyknow. In drawing the sounds one can progressivelyalter and distor t their visual equivalents so as toachieve previously unfamiliar acoustic effects.

17

,


23/72

:1:1il

ii:,Ii.:HI

I,'I

iI:II:

"1,1:'1ill:1Iii

J;,1.i ii

Acoustic phonetics, which is developing and .creasing in richness very rapidly, already enables us tl'solve many of the mysteries of sound, mysteries whic:motor phonetics could not even begin to solve. Hoever, even though it has infinitely greater organisepower, acoustic phonetics, no more than motophonetics, cannot provide an autonomous basis fothe systematisation and the classification of the phonil"phenomena of language. Basically it is faced withjus'the same obstacles as is motor phonetics. At fir, "acoustics attributed to the different sounds onlylimited number of characteristic features. This did nomean that these part icular features were the mos'essential ones. The limits weredue above all to the fac'lthat the analytical capacities of the new d i s c i p l i n ~were as yet ra ther restricted. But i f we consult "thoroughly modern work in the field of a c o u s t i ~phonetics, such as for example the fine monograph by.;Antti Sovijarvi on the Finnish vowels and nasals, D i e ~gehaltenen, geflilsterten und gesungenen Vokale unNasale derfinnischen Sprache (Helsinki, 1938), we findourselves once again confronted with a stunning;multitude of details concerning the features of eachsound, the sound being decomposed into an innumer-:ablevariety offractions. Motor and acoustic phonetics'have proved equally incapable of offering any guid-'ance in this chaos, of identifying the pertinent characteristics, the constitutive and inalienable features ofeach sound. Acoustics can provide us, in impressivedetail, with the micrographic image of each sound, butit cannot interpret this image; it is not in a position tomake use of its own results. It is as i f they were the

18

hieroglyphics of an unknown language. When, as isalways the case, two sounds show both similarities anddissimilarities, acoustics, having no intrinsic criteriafor distinguishing what is significant from what is not,has no way of knowing whether it is the similarity orthe dissimilarity which is crucial in any given case. I tcannot tell whether it is a case of two variants of onesound or of two different sounds.This crucial difficulty is faced not only by experimental acoustics but by any method ofphonetic transcriptionof auditory phenomena, to the extent that thetranscription is based solely on purely auditory perception. Such transcriptions, being obliged to note allnuances of pronunciation, even the most subtle,scarcely perceptible and fortuitous among them, are asAntoine Meillet pointed out, difficult to read and difficult to print. This is not a purely technical difficulty. I tis once again the vexing problem of identity withinvariety; without a solution to this disturbing problemthere can be no system, no classification. The phonicsubstance of language becomes as dust. When facedwith a similar problem in relation to motor phoneticswe had to make reference to an extrinsic criterion andto ask about the immediate aim of articulations, ormore precisely about their acoustic aim. Now we mustask what is the immediate aim of sounds, consideredas acoustic phenomena? In raising this question westraight away go beyond the level of the signifier, beyond the domain of sound as such, and we enter thedomain of the signified, the domain of meaning. Wehave said that we speak in order to be heard; we mustadd thatweseek to be heard in order to be understood.

19

r- ,


24/72

IIIII,"I

(II:;1

"

I'!

iII,

!IIi! ~ j i

i,I,i'II

,"

iii,IIiiIi'r):',II, i

The road goes from the phonatory act to sound, 'narrow sense, and from sound to meaning! Atpoint we leave the territory ofphonetics, the disci: "which studies soundssolely in theirmotor andaco iaspects, and we enter a new territory, that ofphlogy, which studies the sounds of language inlinguistic aspect.One hundred years ago the Romantic Rus;writer Vladimir Odoevskij told the story of aman .received from a malevolent magician the gift of t.. ~ able to see everything and to hear everything: 'Evl'thing in nature became fragmented before him, 'nothing formed into a whole in hismind', and forunfortunate man the sounds of speech became tr,formed into a torrent of innumerable articulat"motions and of mechanical vibrations, aimless'without meaning. The victory of naive empiriccould not have been foretold and represented urimore forceful way. In the laboratories of the scient"of this tendency the phonic resources of languwere split up into a multitude of microscopic fi .which theyproceeded to measurewith great care w.deliberately neglecting their goal and raison d'etre. "was in conformity with this approach that metriststhat time taught that one can only study verse if 0forgets both the language it is written in andtheme,ing which it conveys. The study of the sounds of la ,guage completely lost touch with the truly linguisproblem, that of their value as verbal signs. The 40heartening picture of the chaotic multitude of f:inevitably suggested the antithetical principle, thatunity and organisation. 'Phonology', said the mast,

20

of French linguistics, Antoine Meillet, 'frees us from akind of nightmarewhich hadweighed upon us.' In thenext lecture we shall try to state more exactly whatphonology is and how it succeeds in reconnecting theproblem of sound with that ofmeaning.

21


25/72

I I

23

IT IS SAID that every word, and more generally everylinguistic sign, is an entity with two sides. Every lin-guistic sign is a unity of sound and meaning, or inother words, of signifier [significant, Latin signans],and signified [signifie, Latin signatum]. Remember thediagram that is used to represent this:

I t is rightly said that the two components are inti-mately related, that they call for each other, as is indi-cated by the arrows inthe diagram. Take, for example,the French word which is written pain 'bread'. Thisgraphic form, which is the way in which this word isreproduced in writing, is a form stemming from tradi-tion or history, and i t no longer corresponds to theway in which the word is actually pronounced; insome dictionaries i t is complemented by a more orless detailed phonetic transcription. What is thepresent day phonic form of the word? I t is pe (theconsonant p plus the nasal vowel), and this is thesignifier of the word. The dictionary goes on to tell uswhat the word means: 'Food made of dough, withadded yeast, and baked in an oven'. This is the signi-fied of the wordpain 'bread'.

l'

!'I;iji

I.

i

[,

'J:

Ii,IIi i

I ' 'II:


26/72

Ii i

IIIIii,Ii

:1,

'I.:

~ ' iII

.!

I,;"I

lilli,;1,I,'I1, 'fl"

'1 i,

If someone says pE, this signifier evokes in uscorresponding signified, i.e., the idea of food madedough, with added yeast, and baked in an oven. .the other hand, ifwe think of this kind of food, andour thought happens to be captured in the Frenlanguage, then the motor and acoustic representatiowhich springs to mind is the phonic representation p ,This intimate relation between sounds and m e a - ~ - - 'is perfectly manifest and clear but, as we have alrea,pointed out, it is only recently that the structurethis relation has been studied systematically, and f"study is as yet far from being complete. We know tha,the chain of sounds acts as the supportof themeanin, 'but we need to know how the sounds perform f "function. Wemade use above of a metaphor: we Sai,that this comes down to the problem of discoverin,the quanta of language, i.e., of identifying the smalles'phonic elements bearing signifying value.

I t is true that the phonic substance of language ha"been studied thoroughly, and that such studie.especially over the last fifty years, have produced :abundance of illuminating results. But for the mo ,part the phenomena under consideration have bee!'investigated in abstraction from their function. - 'these circumstances it has been impossible to c l a s s i f ~ 'or even to understand, these phenomena. In the sam,way, itwould be impossible to understand and classifmachines or other instruments so long as a t t e n t i o ~was focused exclusively on the materials with which:,they are made, or on their external form, with no consideration of what they are used for. In order toable to interpret and classify the diverse actions of ourphonatory organs it is essential that we take intol

24

account the acoustic phemonena that these actionsaim at producing, for we speak in order to be heard;and in order to be able to interpret, classify and definethe diverse sounds of our language we must take intoaccount themeaning which they carry,for it is in orderto be understood that we seek to be heard.

If we look at some phonic phenomenon - for example, stress - from a purely phonetic, i.e., purelysensualist, point of view, the direct observation of themotor and acoustic facts and the instrumental analysisof them would show us that the observable characteristics of this phenomenon are essentially the samein various different languages. Auditory intensity andits physiological determinants have been studied andthis has thrown light on the role played by the lengthof the vocal cords. To make a sound louder we putmore force into the flow of air; this mechanical forceincreases the length of the vocal cords, their vibrationsincrease in amplitude, and as a result the sound becomes louder. In comparing stress in different languages it has been noted that it can differ in degreeand that it can be related in differentways to pitch andduration, but basically the operation of stress is initself identical in the different languages. In contrastto this, the use which language makes of it, the linguistic functions of stress, vary from one language toanother.To illustrate this we can compare a simple sentencein two Slavic languages, precisely because althoughthey have an extensive common heritage and althoughthey are very similar in verymany respects, the Slaviclanguages differ from one another completely in theiruse of stress.

25


27/72

IIJj~ j'I'


28/72

il l

I ~ .,

IIii''j:iii11:,

iI,I, ;,'I:. 'I

iiil,', 'iiIi 'I IIII')1I iif '

!.::!ItI,!IiII,11 :I;1"I,j"l' "'l' i'

keeping in mind the multiplicity of the lin,functions of sounds, it is their differentiating fun,which we must consider first of all.If we look at French words such as de (de) ,and dais (de) 'canopy', we can see that the differ,between two sounds - closede and open e - funchere to distinguish the two words.And i f we 100'"the phonic repetory of Russian we can findamong the stressed vowels, two analogous sounds.,more closed e and a more open e: mel' (= m'el') ,..,ground' and mel (= m'el) 'chalk'. In Russi(closed) only ever appears before the p a l a t a ~ 'consonants; e (open) appears in all other positio:Remember that the palatalised consonants are 'nounced by pressing thetop of the tongue against'palate, i.e., with a flattened buccal resonator,that they are therefore acute (high-pitch). Thereflin Russian e (closed) ande (open)cannot appearinsame position and so they arenot able to differenbetween words. You can see then that there is,;;fundamental difference between the paire-e in F r e ~and the analogous pair in Russian. This pairpossein French, a differentiating value which i t lacks "Russian. 'Sounds which have differentiating value, thsounds which are able to distinguish words, havebeegiven a specific name in linguistics. They are caUphonemes. Thus in Russian closed e and open eonly two variants of one and the same phoneme; tbare called combinatory variants, because they depeQsolely on the combination of sounds: before palalised consonants the vowel e is closed and in otbcombinations it is open.

28

In Czech also closed e and open e are unable todifferentiate between the meanings of words. Hereagain they are but two variants of one and the samephoneme, but the distribution of the two variants isquite different from in Russian. In a style that wemight call neutral, Czechusesan open e, whereas in anaffected style - but more particularly in vulgar style,in gutter language - a closede can be heard. Whereasin Russian the two vowels are combinatory variants,which vary with the phonic contextof the phoneme inquestion, in Czech they function as stylistic variants:the vocative pepiku! ('Joe!' and simply 'fellow!')becomespepiku! in speechwhich is more freeand easy.While open e and closed e are both pronounced inRussian and in Czech - in the former varyingwith theneighbouring sounds, in the latter varying with thestyle of speech - it is nevertheless difficult for bothRussians and Czechs to use the openeand the closede of French correctly as different phonemes - or evento notice this difference without effort in pairs ofwords like Ie dais and Ie de, or Ie lait (Ie) 'milk' and IeIe (Ie) 'width '. This is explained by the fact that inthese two Slavic languages the difference betweenthese two vowels cannotmark the distinction betweenthe meanings of words.On the other hand Czech, and also Hungarian, contain, besides the dental consonants, a neighbouringseries of prepalatal consonants, which are articulatedfurther back than the dentals, i.e., a series articulatedat the front part of the hardpalate. For example, usingconventional Czech orthography,sit 'to sow', and sit''the net'. These are, therefore, two different phonemes,the one dental, the other prepalatal. There is an

29

r


29/72

q(,i';1:'I"I:1 1

:1'I',',I', '.1 1 1. ;~ Ii;iIjl'~ iI11I,"Ii,!I''iiIii'

p

riliI!':! :"li l

Ii:II iI'i

IY'

8' ,I,:"I ~J:'JI',II

'IJ,r":[1.

analogous pair among the voiced phonemes: in,ventional orthography dej! 'give!' (imperative):dej'the action'. Now the same prepalatal consoare found in popular French pronunciation, Dample before the semi-vowel of the Frenchword;i'pity', as spoken by a working-class Parisian. B,contrast to the Czech and Hungarian l a n g u a g e s ~ iocclusives of this French parlance do not oppalprepalatal phoneme to a dental phoneme. H e r e ~dental and prepalatal occlusives are only combina,lvariants of a single phoneme. The prepalatal voccurs before the prepalatal semi-vowel, and',dental variant occurs in all other positions. Whileprepalatal consonant has a place in the phorepetory of popular French it does not perfodifferentiating function. Similarly, the two varian'k - one a velar (back) consonant, articulated at'soft palate (the velum), and the other palatal, artilated in a more forward position, on the hard pala-'\both occur in the pronunciati9!J, of French. The fr,.,variant of this consonant is used in various Fre "parlances before front vowels, especially before 1,There is often a very clear difference betweeninitial sounds of the words cas 'case' and qui 'whbut they are only two combinatory variants, anddifference has no independent value in French.contrast in Polish, and also in Roumanian, thesetwo quite distinct phonemes. For example, in Ro'manian the palatal variant of the occlusive in chi'cries' orchiar 'same' (givenherein conventionalorth'graphy) are opposed to the velar occlusive of cu 'wi-or car 'cart'.Irish uses the presence or absence of the voice n

30

only to distinguish d from t, g from k, etc., but alsotWO different lateral phonemes, a voiced I and an un'Voiced I: la and lao Now these two sounds are alsoused in the pronunciation of French, but a Frenchspeaker ~ h ~ a no ~ n o ~ l e d g e ~ P ? o n e ~ i c s ~ o u l d notnotice thIS SInce thIS parr, which In Insh IS used todistinguish between words with different meanings,cannot perform this function in French. In this language they are combinatory variants: the unvoiced I,which is pronounced without vibrating the vocalcords, occurs at the end of words after an unvoicedconsonant, as for example in peuple ' the people'; inallother positions 1is voiced, as for example inpeupler'to people'.The English language distinguishes between twodifferent phonemes: a labiodental, written v, and abilabial, written w. In Slovak, the labiodental v andthe bilabialware two combinative variants of a singlephoneme which occurs as the labiodental v beforevowels and as the bilabial w in all other positions.The two liquids r and I have such clearly distinctfunctions in our languages (cf. ray-lay, fur-full) thatit seems strange to us that in some other languagesthey are simply two combinatory variants of a singlephoneme.Thus inKorean this phoneme is representedby I at the beginning and by r at the end of a syllable(the Indo-European language was probably similar inthis respect originally). It is natural that a Koreanwhois trying to learn English will at first pronounce roundwith an initial I, sellwith an r at the end, and will reverse the order of the two liquids in rule which willthen be confusedwith lure. Again, in French there is adistinction between three vocalic phonemes in words

31


30/72

ii',I,':\'I''I"I:.;/i!:

i.,"I:ii!,I': I,'I' '

II:I, !

III! .

, Ir ,;1'II:I:,:;' I

'IiIii'

'I ii,,:I,, '!1,.;1.

such as si, SU, and sou, whereas in Cherkess thesemerely threecombinatoryvariants of a singlephonem(narrow vowel) and the choice of variant depends o'the nature of the preceding consonant.These few examples, although elementary, shoul"be enough to make clear the fundamental differen ".between the strictlyphonetic point of view, which a'only at drawing up an inventory of the sounds oflanguage considered simply as motor and acousti,phenomena, and thephonological point of view whic:requires that we examine the linguistic value of tbsounds and that we list the phonemes, Le., the syste: 'of sounds considered as elements which serve to dis4tinguish the meanings of words. If we compare the:two inventories it will be seen that they are q u i ~ ,different and that of the two the collection of p h nemes is at once very much more restricted, morlclear-cut and more discrete in the mathematical senof the term. It reveals to us a coherent andcoordinat,system. If we compare any two particular languag,we will see that from an acoustic and motor point 0. ,view their sounds could be identical while the way iwhich they are grouped into phonemes is d i f f e r e n t ~For example, in the Far East there are found several;,neighbouring languages which all use thedental soundl,r; yet in some of these languages, for example'Tungusic, rand 1are separate phonemes; in o t h e r s , ~ ,for example inKorean, rand 1are the twocombinatoryj'variants of a single liquid phoneme; a third p o s s i b i l i t ~ 'is represented by old Gilyak inwhich rand t were tw-r,combinatory variants representing a single dentalphoneme. In an intervocalic position the occlusion ..Jthe closure of the breathpassagewhich is necessary foJ.:

32

the articulation of a t - was not complete and in theseconditions the dental phonemewas pronounced in theform of an r. On theother hand an essentially identicalphoneme can be represented in different languages bysounds among which there is a significant variationfrom an acoustic and motor point of view. For example, in the majority of Far Eastern languages thereis but a single liquid phoneme, but whereas in Chinesethis phoneme takes the form of an 1, in Japanese ittakes the form of an r, and in Korean, as we havealready said, it is represented by two combinatoryvariants: these purely external differences in no wayalter the fact that in allof these languages there is onlyone liquid phoneme.In linguistics the idea of the phoneme, of the dis

tinctive sound, or rather the idea of that in the soundwhich is distinctive, is not a recent one. In the historyoflinguistics the credit for having initiateda discussionof this problem goes primarily to Baudouin deCourtenay. This great Polish linguist introduced theidea of the phoneme in 1870, when he was twenty-fiveyears old, in his inaugural lecture at the University of8t. Petersburg. From the very beginning he had considered, besides the purely phonatory and auditorystudy oflinguistic sounds, 'their role in themechanismof language, and the significance attributed to themby the linguistic intuition of speakers'. The youngBaudouin had understood that this latter aspect doesnot always coincide with the classification of thephonic data on the basis of their physical and physiological properties: in short, on his view of the matter,What is important in linguistic sounds, for both thelinguist and the speaker of the language, is primarily33


31/72

35

aJlllIlarian thought is the continual substitutionof~ c t 1 y causal questions for questions concerningJlleans and ends. Any attempt to define a linguistichenomenon in terms of its function would have been~ n d e m n e d in this period as an unacceptable heresy.The etymological phonetics, or in other words thefunctional phonetics, contemplated by the youngBaudouin, was replaced in this scientist's own laterworks, in conformity with the spirit of the times, bywhat he called 'psychophonetics'. The new disciplinein the process of formation was no longer centred onthe function of sounds, on the ends which they serve;in short it was no longer conceived in terms of theproblem of the relation between sound and meaning.And whereas etymological phonetics had been conceived by Baudoin as a bridge between phoneticsandgrammar, psychophonetics on his own account wouldattempt to build a bridge between phonetics andpsychology. Phonetics would study the productionand the audition of language sounds, and psychophonetics would have the taskof throwing lighton thepsychological determinants of phonation and audition.Yet if we ignore the phraseology and the terminology ofBaudouin's programme, and i fwe look insteadat the essence, at the actual content, of his works inthis field, we can observe that he did in fact treat language sounds not as a psychologist but as a linguist.Right from the start hehad grasped the importanceofdifferentiation, he had brought to light the distinctivekernel of sounds - in other words, the phoneme itself.His investigation of the phonic aspect of languagewasbased precisely on the concept of the phoneme. Yet

34

their role in the ordering of words. Baudouin /Courtenay proposed the creation of a new lin . '.,discipline, to be called 'etymological phonetics'.new discipline would have as its task, accordingtofounder, the analysis of the relations betweenmotor-acoustic properties of sounds and their lexi,'and grammatical values. ,Baudouin's creativity enabled him with a s t o n i s ~ foresight to raise and to undertake a preliminarycussion of the central problems of linguistics asknow it today, but the ideological weakness, orcertainty, of his time prevented this scientist of gem,from fully exploiting his own discoveries, and fr, "having any direct successors. We have quoted fr 'Baudouin's inaugural lecture which was given in18We can see from this date the uncommon indedence of thought at work in Baudouin'sjuvenilia. .was, for international linguistics, a period of debaland fermentation, a period which was favourable flthe instigationof bold, individual ideasand initiativiI t was only at the end of the 1870s that the neogra,marian school, centred on Leipzig, stabilised and b'came an identifiable and lasting force. This current 'thought soon came to exercise a dominant influen 'over linguistic thought on an international scale, aD! "it succeeded in maintaining this positionup to the t'of the First World War. While Baudouin never stricti'speaking fully identified himself with the neogramarian school, it must be admitted none the lessthalike almost all linguists in that period, he was influenced by this school, and clear indications of thi:influence can be found in his work. 'Perhaps the most characteristic

II j

.It

'i''I'IiiiI

;r'i

,d

. ~ I,

[:,i: I!';II '

lilll i i l '1'1"ii! :1"I," ,I ,I'! 1'III' III";j 1,,1,IIIIi

W!: '

: I ~ .JI:

, I. 'I


32/72

i "I Lj1 :.!),I"",t:,11:,1,Ii'l,P I,"Ii,II'.:1:1'1:::1]::iillij'li;111!1

i1II'.".!" '1"~ . ,"il,/:I! 'j:1:':;"f',:I,:"I '

while he was an original and subtle thinker in lintics, Baudouin de Courtenay's philosophical""'!'psychological views remainedwithin the framewor'ideas current in his time. And since it was requir,that time that any phenomenon be defined noterms of its function but solely in terms of its orl,Baudouinattempted to formulatea genetic concepof the phoneme, in conformity with the do . fideology. To legitimate the idea of the phoneme!felt himself obliged to answer certain troubquestions: where is the phoneme located? indomain of reality does it have its roots?He though'could deal with these problems by projecting.phoneme, in facta purely functional, purely lingu'"idea, into the realm of mental images. He t h o u g h ~had succeeded in providing proper foundationsthe phoneme in defining it as 'the psychic equivalof the sound'. Baudouin's 'psychologism' was onlcamouflage which served to legitimate his in:Q.ovastudies in the eyes of his contemporaries, and inowneyes too. But this camouflageprevented him fr,finding his way among his own great discoveriesfrom drawing out their implications.What people learned from Baudouin's workfortunately shared this ambiguous character. Forample, the distinguished Russian linguist LevScerba, one of Baudouin de Courtenay's best p u p ~ 'in his book on the Russianvowels (published in 191which represented an important point in the develment of the Baudouin School and of linguistics 'general, paid careful and detailed attention toconcept of the phoneme, and identified the phone:as the 'fundamental element' in linguistics. In defi .

36

it in this way Scerba placed greater emphasis than hadBaudouin on the functional aspect of the phoneme,but at the same time he tied this concept, even morefirmly than had his master, to the genetic andmechanistic dogmas of traditional psychology. I t is true thatfor Scerba the essential characteristic of the phonemeis its capacity to distinguish between words, but atthe same time this scientist insisted on psychologicalcriteria for identifying phonemes. For him the phoneme and the sound are not two aspects of one and thesame phenomenon, but two contiguous phenomena.Instead of taking the phoneme to be the functionalaspect of the sound and the sound to be the substratum of the phoneme, he distinguished between thesound and the phoneme as being an external, objectivephenomenon on the one hand, and a subjective,psychic phenomenon on the other. This conception ismistaken. To be convinced of this it is enough to refer. to our interior, non-externalised speech.We speak to ourselveswithout emitting andwithouthearing any sounds. Instead of pronouncing or hearing we imagine ourselves to be pronouncing or hearing.Thewords ofour interior speech arenot composedof emitted sounds but of their acoustic and motorimages. And if a Russian, in his interior speech, pronounces in imagination the wordsmel andmel', whichwe have already discussed above, the former wordwill include the acoustic andmotor image of an open 8,and the latterwordwill include the imageofa closed e.Therefore, the identity of the phoneme in contrast tothe variety of the sounds - for example, in Russianthe identity of the phoneme leI in relation to its twovariants, the sounds open 8 and closed e - cannot be

37


33/72

and what we are not conscious of is very shifting andindefinite as regards everything concerned with language and its elements. As a rule language is not forus an end in itself but only a means, and the elementsof language usually remain beneath the threshold ofour conscious deliberation. As the philosophers say,linguistic activity takes place without self-knowledge.And even if a speaker with no special training were tosucceed in isolating some of the functional elements oflanguage, in particular some of the phonemes or thegrammatical categories, he would still not be in aposition to discover the laws which relate them one toanother, i.e., the system of grammatical categories orthe system of phonemes. As we would expect, Scerba,having based his investigation of phonemes on whatspeakers were conscious of, found himself compelledto give up any attempt at classifying these entities.In spite of all these ambiguities the solid core of \Baudouin's doctrine, the idea of the differentiatingvalue of phonemes, did in the end gain admittance intolinguistics. Moreover, some other nineteenth-centurylinguists introduced similar ideas, inde endently ofthe school we have been discussing. W pld,men-tion in particular t he Swissdialectologis - telerwho, in a brilliant monograph on a S Germandialect of the canton of Glarus (published 876) notonly blazed a trail for scientific dialectology but alsoclearly indicated the necessity of not confusing twodistinct kinds of phonic differences: those which in agiven language areused to mark lexical or grammaticaldifferences, and others which lack this function. ButWinteler's book came out at the same time as did thefirst important works which were to spread the doc-

fl'.ll!:1:r,,1:1, .I' '

I

I' ,'Ii!

I,

!,,, I

I,"

!'! "' I.;i, ;Ii:

:'1,: ''i! I,I:i: ': 'I , , ' " i'I' I,i:I'I,:,',:,

,! "II,:l 'i '

l

interpreted as the identity of the psychic image in cotrast to the variety of actually emitted sounds. What "the meaning of the identity of a phoneme, for e x a m p ~ lthe identity of the phoneme lei in Russian? I t is tthe difference between a closed e and an open e is n,_'put to work in the semantic system of this languag,"that this difference does not serve to distinguish b 'tween words. What does serve to distinguish betWiwords is the intermediate character of an lei (whethopen or closed) in contrast to the high vowel Iii (m"'dear') and the low vowel lal (m'al 'crumpled'). II t is only by an analysis of the functioning of soun - ' in a given languagethat the systemofphonemes of th:language can be established. Scerba and some otbdisciples of Baudouin de Courtenay chose to rely ondifferent method, that of psychological i n t r o s p e c t i o ~ 'They appealed to the linguistic intuition of the speaing subject. On their view the phoneme is an acousti.motor image which the speaker is himself in a positio:to identify. I t is true that we are much more cosciously aware of those elements of a language whi,have an independent differentiating role than of thowhich lack this function. But the primary fact is pr lcisely this differentiating value of any particulelement; conscious awareness of it is a consequence ::this value. I t is therefore logicalto take as theanalyHcriterion this primary fact, i.e., the differentiafvalue of the elements to be analysed, rather than f 'secondary fact, Le., our more or less conscious awar,ness of these elements. This latter criterion WOWItake us out of the territory of linguistics and into thof psychology. But the greatest disadvantage of tcriterion is that the borderline between what we arl

38 E 39


34/72

I.I11

I.,

i ,',::I,,.,

I ,I

'_:;,

'

trines of the neogrammarian school, and since h i s ~ ~basic idea went so much against the stream it wentalmost entirely unnoticed. Two famous phoneticians,:;:;the Englishman Henry Sweet and his Danish disciple\;Otto Jespersen, distinguished in principle a m o n g ~phonic phenomena between those which are

Six Lectures on Sound and Meaning

Documents

Transcript of Six Lectures on Sound and Meaning