In the thirty years since this translation first appeared, Saussurean studies have not only flourished, but flourished as never before. The bibliography of Saussureana has swollen to an extent that makes it impossible to offer even a readable digest of it here. But the question may reasonably be asked whether anything new has come to light which affects the way we should read the Cours de linguistique générale today.
Little progress has been made with the problem of how Saussure’s editors conceived their task. Did they think of themselves as mere academic secretaries, supplying a tidied-up version of the oral teaching received by those who attended Saussure’s lectures? Or was their editorial ambition to produce the book that Saussure evidently had in mind, but never managed to write? These are two very different aims. Saussure himself sometimes referred to sections of the lectures as ‘chapters’, thus validating the notion that they could be treated as parts of a forthcoming book. But his terminology and his exposition were not always ideally suited to the printed page. Bally and Sechehaye, for their part, were handicapped by working at one remove from the original Saussurean material, for neither of them had attended the lectures in person.
The editors say in their Preface that their main aim was to present ‘an organic whole’ (un tout organique). This was certainly more than Saussure himself had achieved. They also describe their text as an attempt at ‘synthesis and reconstitution’. But is there not ab initio a conflict between synthesis and reconstitution? It might not have been a serious conflict had the three courses of lectures been merely viva voce variants of a single text. But that was far from being the case, as can be seen by anyone who reads separately and then compares the three sets of notes which Saussure’s first editors had amalgamated into one.
These are now available as (i) Premier Cours de linguistique générale (1907), d’après les cahiers d’Albert Riedlinger, edited and translated by Eisuke Komatsu and George Wolf, 1996; (ii) Deuxième Cours de linguistique générale (1908–1909), d’après les cahiers d’Albert Riedlinger et Charles Patois, edited and translated by Eisuke Komatsu and George Wolf, 1997, and (iii) Troisième Cours de linguistique générale, d’après les cahiers d’Émile Constantin, edited and translated by Eisuke Komatsu and Roy Harris, 1993.
Komatsu points out that although the editors said their text was based mainly on the Third Course, that was not in fact the order of presentation they adopted. The Third Course began with Geographical Linguistics, whereas in the Cours this topic was postponed until Part IV, when Diachronic Linguistics and other topics had already been dealt with. Thus the Cours as it stands cannot be regarded as a faithful record of Saussure’s oral teaching in the classroom.
What is one to make of the editorial reorganization of topics in the Cours? Komatsu sees the influence of Husserl. ‘The structure of the published text of the Cours seems to be in some respects significantly Husserlian.’ It explains, according to Komatsu, why Phonology is inserted immediately after the Introduction. Geographical Linguistics remains a loose end. Geographical Linguistics, as practised by Saussure’s generation in, for example, the publication of the French linguistic atlas (mentioned in Cours pp. 276–7), is manifestly a branch of Synchronic Linguistics. But it is not mentioned in Part Two of the Cours, ostensibly devoted to synchronic studies.
Students who attended the First Course were told straightaway that they should not expect a definition of linguistics. They must first understand what linguistics is not (ce qu′elle n′est pas). This is an initial example of the pedagogic technique that Saussure himself called négativité. But perhaps it masks Saussure’s own uncertainty at that time about how best to define the subject he was professing. Students who attended the Second Course, on the other hand, were spared Saussure’s finger-wagging warnings about confusing linguistics with neighbouring types of study. Instead the audience was presented in the first lecture with viable positive definitions both of la langue and of la parole. It is worth repeating them here.
La langue: ‘un ensemble de conventions nécessaires adoptées par le corps social pour permettre l’usage de la faculté du langage chez les individus.’ La parole: ‘l’acte de l’individu réalisant sa faculté au moyen de la convention sociale qui est la langue.’ Langue and parole are thus both defined by reference to the mental faculty involved (le langage), and not by reference to the community or to the community’s speech practices (which differ from one country to another).
There has been little discussion of the distinction that Saussure drew very explicitly on May 5 1911. This is the distinction between une langue (‘a language’) and la langue (‘the language’). La langue, students were told, is a generalization designating what is common to every particular language. It is not clear whether this generalization is intended to apply retrospectively to Saussure’s previous uses of la langue. John Joseph (Joseph 2012: 580) remarks that the generalization ‘seems to belong more to the later conceptual world of Chomsky than to Saussure’; and indeed it is difficult to see what need there is for it in Saussure’s overall scheme of things. It presupposes, unrealistically, that the linguist is equipped with a knowledge of all languages (past and present). On p.121 of the Cours the linguistic system is described as unchangeable (immutable). The students’ notes, however, reveal a less absolute proposition: the linguistic system is never altered in its entirety (en entier, dans son entier).
The masterstroke of Saussure’s theorizing is to present his own failure to interlink the various facets of linguistic theory as being something more positive, that is, not as a failure but as a recognition of what he calls the first principle of linguistics: the arbitrariness of the linguistic sign.
* * *
Apart from the three publications listed above, the most significant development in Saussurean linguistics has been the discovery of Saussure’s ‘lost’ notes in a drawer in the orangery of the Saussure family home in Geneva. These have now been edited by Simon Bouquet and Rudolf Engler under the title Écrits de linguistique générale (Gallimard 2002).
The importance of the ‘lost’ notes is that they appear to show Saussure’s mature linguistic thinking moving away from the position commonly labelled structuralism. Since Saussure is widely regarded as the founder of structuralism in linguistics, this would be a radical shift.
But the matter is complicated by the fact that Saussure himself never uses the term structuralisme. Matthews (2007: 385) defines structuralism as ‘any school or theory in which language is conceived as a self-contained, self-regulating system, whose elements are defined by their relationship to other elements’. If we can take it that ‘defined’ here means ‘solely defined’, structuralism thus conceived is plainly holistic.
David Allison (Allison 1995: 770–2) picks out the characteristic that ‘each basic unit of language is itself the product of differences between other elements within the system’. (It would be more satisfactory if this read: ‘product of differences and similarities within the system’. For if there are only differences, it is difficult to see that the compilation of a dictionary would be feasible, or, if attempted, useful. No one wants a dictionary which says merely that each word means something different from every other word, but never gives a positive specification of its meaning.)
According to John Lyons, ‘we are not doing violence to Saussure’s thought if we say that a language is a structure, implying by this use of the term that it is independent of the physical substance, or medium, in which it is realized’ (Lyons 1981: 220). But this sits uncomfortably with the insistence in the Cours that speech alone (not writing) is the linguist’s concern. Furthermore, ‘not doing violence’ to someone’s thought is hardly the same thing as stating it accurately.
In Chapter VI of the Cours (p.45) we are told that l’objet linguistique n’est pas défini par la combinaison du mot écrit et du mot parlé; ce dernier constitue à lui seul cet objet. Critics have questioned whether Saussure has any business to adopt a stance which expels the written sign from linguistics ex cathedra in this way. For it appears to promote one medium at the expense of another. Nothing in the nature of the linguistic sign justifies this promotion. Culler is among those who have criticized Saussure on this score:
Sound is simply a way of manifesting the signifiers of a language, which are themselves defined in oppositional and combinatory terms without any reference to phonic material. So he ought not to assert, as he does, the priority of the spoken language. (Culler 1976 :111)
Yet this priority is asserted in all three courses. In the First Course, we are told that speech takes a ‘natural’ priority over writing: ‘tout naturellement (underlined in the original) nous accordons au signe écrit la prééminence sur le signe parlé’. In the Second Course, we are told, according to Riedlinger’s notes, that it is important not to forget ‘that the spoken language <alone> is the object of linguistics’. This is confirmed by Patois: ‘seul le mot parlé est l’objet véritable de la linguistique’ (underlined in the original). In the Third Course we are assured that linguistics
will take as its subject matter every kind of variety of human language: it will not select one period or another for its literary brilliance or for the renown of the people in question. It will pay attention to any tongue, whether obscure or famous, and likewise to any period, giving no preference, for example, to what is called a ‘classical period’, but according equal interest to so-called decadent or archaic periods. Similarly, for any given period, it will refrain from selecting the most educated language, but will concern itself at the same time with popular forms more or less in contrast with the so-called educated or literary language, as well as the forms of the so-called educated or literary language. Thus linguistics deals with language of every period and in all the guises it assumes. (Komatsu and Harris 1993: 3a-4a).
There is no rationale here for according priority to the spoken word. Nor is any offered elsewhere in the students’ notes.
So it seems that the priority of the spoken language occupies a curious place in Saussure’s vision of linguistic studies. It is not a theoretical principle but an independent article of faith that stands outside the rationale of the linguistic sign itself. However, one can see why it occupies the place that it does. Without it, writing would be on a par with speech, and graphology with phonology. That would be the end of any attempt to construct a unified theory of the sign. For there is nothing in common between the restrictions that these two media in their primitive form impose on the signs belonging to their domain. Writing requires marks on a surface: speech does not. (I say ‘in their primitive form’ since more technically sophisticated forms go beyond this; e.g. if we regard a sound spectograph as a visible form of speech. But no one resorts to sound spectography in everyday communication, and the linguistics of Saussure’s day had no sound spectography in any case.)
* * *
To judge by the ‘lost’ notes, the direction in which Saussure seems to be moving – albeit tentatively – is towards what would nowadays be called integrationism. (See Integrational Linguistics: a First Reader (1998), edited by Harris and Wolf, and Rethinking Linguistics (2003), edited by Davis and Taylor. Cf. the entry Integrational Linguistics in Matthews 2007: 197.)
Integrationism proposes a different perspective altogether from any form of structuralism.
For integrationists the individual is the focus of attention. What passes for ‘the language system’ (langue) is simply a social extrapolation from communication between individuals. The individual is the language-maker, and language is made – every day and every hour – by integrating one person’s communicative activities with another’s. This integration is temporary and context-bound. It does not establish any quasi-permanent set of forms and meanings. The purpose it serves is ephemeral.
A more radical way of putting this is to say that, for the integrationist, languages (in the sense in which English, French, German, etc. are commonly held to be languages) do not exist at the level of first-order linguistic experience. They are merely second-order social constructs. This proposal commonly meets with scepticism or outright rejection from orthodox linguists, whether they are Saussureans or not. The domain of linguistic orthodoxy treats the world in which we live as already divided into separate linguistic communities, to which such language-names as English, French and German can readily be attached, just as continents may be divided into countries, each with its own name (England, France, Germany, etc.). In the linguistic case, each such name supposedly designates the language spoken and written mainly by those who live in a particular area of the world. In short, we are dealing with factors that Saussure would doubtless have subsumed under Geographical Linguistics.
The great lacuna here is the linguistique de la parole that Saussure promised his students but never delivered. We can only conjecture how Saussure would have tackled the range of problems that the study of speech acts presents. But there are already clues in the text published by Saussure’s editors in 1916, even though they are few and far between, and even though none of them is elaborated in any definitive way.
First of all, there is the notion of the act of parole as one involving both the will and the intelligence. La parole est ... un acte individuel de volonté et d’intelligence (Cours p.30). The word intelligence does not appear in any of the students’ notes and seems to be an interpolation of the editors. Why did they think it necessary? Presumably to rule out the possibility that some acts of speech might be automatic or spontaneous functions of the human vocal apparatus. The intervention of intelligence in any human activity implies a rational selection of means chosen to accomplish a given end, as opposed to a course of action dictated by physiological or other factors. Nevertheless, it is notoriously difficult in any complex programme of human activities to distinguish with assurance between what is rationally selected and what is not. This is the first major problem that would have faced Saussure’s linguistique de la parole.
Clearly related to this is the controversy over whether to regard la phrase as a unit of langue or a unit of parole. For it raises the issue of the extent to which the speaker is in control of the syntagmatic structures of speech, as opposed to the extent to which they are dictated to the speaker by the system. According to Culler (1976: 84), Saussure did not realize ‘that it is possible to construct a finite system of rules which will generate structural descriptions for an infinite number of sentences’. Culler (1976: 82) says that Saussure’s conception of syntax ‘seems exceptionally weak’ because of his ‘failure to include sentences within the linguistic system’. Culler hastens to put himself safely on the side of the angels, who are all singing from the same generativist hymn sheet. Generativists privilege something called ‘rule-governed creativity’. What this amounts to in practice is restricting the speaker’s freedom to constructions allowed by the ‘rules of grammar’. Saussure, it is true, advances no theory of grammar as such. Nor, an integrationist would add, does he need one. Linguistic rules are the inventions of theorists, not laws of Nature or patterns devised by language-users.
Jespersen distinguishes between ‘formulas’ and ‘free expressions’ (Jespersen 1924: 18ff.). How do you do? is a formula and I gave the boy a lump of sugar a free expression. In the formula, he says, ‘everything is fixed’, whereas in the free expression everything has to be ‘created in each case anew by the speaker, who inserts the words that fit the particular situation’.
This must not be confused with the Saussurean distinction between langue and parole. Every utterance is an instance of parole, regardless of whether it is a formula or not. Jespersen never explains how the hearer recognizes whether the speaker has uttered one or the other. (It’s raining might be either.)
Once it is realized that faits de langage do not come neatly sorted into two categories (faits de langue and faits de parole), cracks begin to appear in the whole Saussurean edifice of synchronic linguistics. There is no longer any question of defining one category by reference to the other, or claiming that they are connected by a relationship of mutual presupposition (Cours, p.37: se supposent l’un l’autre). Mutual presupposition is a euphemism for circularity. The simplistic equation ‘langage = langue + parole’ emerges as one of the most misleading dicta in the whole of the Saussurean corpus. For if there is no clear bipartition of faits de langage in the first place, then langue can no longer be langage minus parole, any more than parole can be langage minus langue. What previously looked like a well-mapped territory for the linguist to explore immediately becomes an uncharted no-man’s-land. In the no-man’s-land, it is indeed possible to discern landmarks. The trouble is that there is no consensus about these landmarks and how they are related to one another. What one person takes to be a linguistic fact is not necessarily recognized as such by that person’s interlocutors or correspondents.
This is a point already made centuries ago by John Locke, although there is no evidence that Saussure had ever read him. Locke was the first to draw attention to the question-begging assumption that when a speaker uses a word the hearer will automatically attach the same meaning to it as the speaker does. What is the basis of this psychological premise? It lies at the heart of the langue/parole distinction.
How far did Saussure’s doubts about this distinction go? If we were limited to the evidence of the Cours, the answer would have to be ‘Not very far’. But the ‘lost’ notes published in 2002 throw a rather different light on the matter. Take, for instance, the masterly critique of the orthodox lexicographical concept of a ‘word’, of which Bally and Sechehaye were seemingly unaware (Bouquet and Engler 2002: 77). In a monolingual dictionary, the lemma gives the orthographic form of the word, while the gloss supplies its meaning. But the gloss is just another set of words. Nothing has been explained. The problem of the ‘word’ has simply been carried over from one example to another. However far this deferral is carried, no positive or permanent identification of the unit emerges:
vouloir épuiser les idées contenues dans un mot est une entreprise parfaitement chimérique, à moins peut-être de se borner à des noms d’objets matériels et d’objets tout à fait rares, par exemple l’aluminium, l’eucalyptus, etc. Déjà si l’on prend le fer et le chêne, on n’arrivera pas au bout de la somme de significations (ou d’emplois, qui est la même chose) que nous donnons à ces mots, et rien que la comparaison de fer avec deux ou trois mots comme acier, plomb, or ou métal, rien que la comparaison de chêne avec deux ou trois mots comme saule, vigne, bois ou arbre représente un infini travail (...) une vie humaine pourrait sans exagération s’y passer.
What this implies is that words do not exist as the dictionary portrays them, that is, as fossilized exhibits like dead butterflies in the entomologist’s glass case. The life of words is to be sought in the daily exchanges of actual language-users. Thus far from langue taking priority over parole, it will be parole that takes priority over langue.
This conclusion does not square at all with Saussurean structuralism as commonly construed. At one stroke it destroys the autonomy of the language system. For the system cannot be reconstituted from one day to the next. It is not like the state of the larder or the garden. The lexicographer’s self-imposed task is futile; not because life is too short but because the task rests on a misconception of linguistic meaning. Meaning arises from the act of communication itself and the ongoing activities thereby integrated.
It is interesting that in the notes published by Bouquet and Engler in 2002 (p.72), Saussure himself uses the term intégration. He writes:
Le phénomène d’intégration ou de postméditation-réflexion est le phénomène double qui résume toute la vie active du langage.
What Saussure proposes here is something very close to the integrationist conception of the way linguistic communication works. Communication depends on signs. But the signs are not established in advance of the act. Their making is part of the act itself. The process is open-ended. Since the number of those signs and their reciprocal and relative aspects change from one moment to the next, communication likewise changes accordingly. That is what we experience in everyday conversation, whether with friends or with strangers.
In an integrationist perspective, there is no need for the postulation of invariant collective codes that ensure the identity of one individual’s thoughts with another’s. Just as synchronic and diachronic linguistics require two quite different standpoints and sets of assumptions, so a linguistique de la parole would require a third perspective. It will not do to treat the study of la parole as no more than an examination of the individual’s post facto execution of units and constructions supplied in advance by la langue.
Saussurean linguistics, as it was bequeathed to posterity at Saussure’s death, represents at best only two thirds of a comprehensive linguistics. The missing third is the problem that hangs over any modern attempt to judge Saussure as a linguistic theorist. What is lacking includes, at the very least, a theory of reference, a theory of deixis, a theory of proper names, a theory of interlocutory roles, and a theory of context. In the whole of Saussure’s teaching there is no detailed analysis of a sustained stretch of discourse. All Saussure’s examples – and there are few enough of these – are decontextualized snippets.
Dans chaque signe existant vient donc S’INTÉGRER, se postélaborer une valeur déterminée [...] qui n’est jamais déterminée que par l’ensemble des signes présent ou absent au même moment; et, comme le nombre et l’aspect réciproque et relatif de ces signes changent de moment en moment d’une manière infinie, le résultat de cette activité, pour chaque signe, et pour l’ensemble, change aussi de moment en moment dans une mesure non calculable. (Bouquet and Engler 2002: 88)
This overtly integrationist pronouncement contains the nucleus of what might have been, had Saussure lived, his mature linguistique de la parole.
It would be hasty to conclude that what Bally and Sechehaye published is an inadequate rendering of Saussure’s linguistic thinking, and therefore an English translation of their text is superfluous to academic requirements. For, as Culler puts it (Culler 1976: 17),
Saussure’s importance in linguistics and in other fields rests less on what he ‘really’ thought than on what is contained in the Cours.
However far students of Saussure may stray from the Cours de linguistique générale, it is to the Cours that they must return in the end. The authorial figure created by Bally and Sechehaye is alone responsible for most of the insights and puzzles in the history of structuralism. That is why a translation of the posthumously published text, although originally intended for students lacking a fluent command of French, retains its interest for those familiar with that language. For reading the French side by side with the English – which is nowadays the best way to read it – draws attention to possible ambiguities of interpretation which might – and almost certainly would – otherwise remain unnoticed.
Saussure’s most enthusiastic supporters would hardly describe him as a lucid or rigorous thinker. On the contrary, he was given to taking up ideas which he found attractive but had not examined in the detail required. His 1916 editors commendably made the best of a bad job. But the best of a bad job is what it was and remains. The Cours and the Course need each other.
References
Allison, D., ‘Structuralism’. In R. Audi (ed.), The Cambridge Dictionary of Philosophy, Cambridge, Cambridge University Press, 1995, pp.770-72.
Cours: F. de Saussure, Cours de linguistique générale, ed. Ch. Bally et A. Sechehaye, 2éd. Paris, Payot, 1922.
Course: Harris, R., F. de Saussure, Course in General Linguistics, London, Duckworth, 1983.
Culler, J., Saussure, Glasgow, Fontana, 1976.
Jespersen, O., The Philosophy of Grammar, London, Allen & Unwin, 1924.
Joseph, J.E., Saussure, Oxford, Oxford University Press, 2012.
Locke, J., An Essay concerning Human Understanding [1706]. ed. A.C. Fraser, repr. New York, Dover, 1959.
Lyons, J., Language and Linguistics. An Introduction, Cambridge, Cambridge University Press, 1981.
Matthews, P.H., The Concise Oxford Dictionary of Linguistics, 2nd edn, Oxford, Oxford University Press, 2007.