Ferdinand de Saussure (1910)

Third Course of Lectures on General Linguistics

Source: Saussure's Third Course of Lectures on General Linghuistics (1910-1911) publ. Pergamon Press, 1993. Reproduced here are the first few and last few pages of what are notes taken by a student of Saussure's lectures.

[28 October 1910]
Introductory chapter: Brief survey of the history of linguistics

The course will deal with linguistics proper, not with languages and language. This science has gone through phases with shortcomings. Three phases may be distinguished, or three successive approaches adopted by those who took a language as an object of study. Later on came a linguistics proper, aware of its object.

The first of these phases is that of grammar, invented by the Greeks and carried on unchanged by the French. It never had any philosophical view of a language as such. That's more the concern of logic. All traditional grammar is normative grammar, that is, dominated by a preoccupation with laying down rules, and distinguishing between a certain allegedly 'correct' language and another, allegedly 'incorrect'; which straight away precludes any broader view of the language phenomenon as a whole.

Later and only at the beginning of the 19th century, if we are talking of major movements (and leaving out the precursors, the 'philological' school at Alexandria), came 2) the great philological movement of classical philology, carrying on down to our own day. In 1777, Friedrich Wolf, as a student, wished to be enrolled as a philologist. Philology introduced a new principle: the method of critical examination of texts. The language was just one of the many objects coming within the sphere of philology, and consequently subjected to this criticism. Henceforth, language studies were no longer directed merely towards correcting grammar. The critical principle demanded an examination, for instance, of the contribution of different periods, thus to some extent embarking on historical linguistics. Ritschl's revision of the text of Plautus may be considered the work of a linguist. In general, the philological movement opened up countless sources relevant to linguistic issues, treating them in quite a different spirit from traditional grammar; for instance, the study of inscriptions and their language. But not yet in the spirit of linguistics.

A third phase in which this spirit of linguistics is still not evident: this is the sensational phase of discovering that languages could be compared with one another; that a bond or relationship existed between languages often separated geographically by great distances; that, as well as languages, there were also great language families, in particular the one which came to be called the Indo-European family.

Surprisingly, there was never a more flawed or absurd idea of what a language is than during the thirty years that followed this discovery by Bopp (1816). In fact, from then on scholars engaged in a kind of game of comparing different Indo-European languages with one another, and eventually they could not fail to wonder what exactly these connections showed, and how they should be interpreted in concrete terms. Until nearly 1870, they played this game without any concern for the conditions affecting the life of a language.

This very prolific phase, which produced many publications, differs from its predecessors by focussing attention on a great number of languages and the relations between them, but, just like its predecessors, has no linguistic perspective, or at least none which is correct, acceptable and reasonable. It is purely comparative. You cannot altogether condemn the more or less hostile attitude of the philological tradition towards the comparativists, because the latter did not in fact bring any renewal bearing on the principles themselves, none which in practice immediately opened up any new horizons, and with which they can clearly be credited. When was it recognised that comparison is, in short, only a method to employ when we have no more direct way of ascertaining the facts, and when did comparative grammar give way to a linguistics which included comparative grammar and gave it a new direction?

It was mainly the study of the Romance languages which led the IndoEuropeanists themselves to a more balanced view and afforded a glimpse of what the study of linguistics was to be in general. Doubtless the growth of Romance studies, inaugurated by Diehls, was a development of Bopp's rules for the IndoEuropean languages. In the Romance sphere, other conditions quickly became apparent; in the first place, the actual presence of the prototype of each form; thanks to Latin, which we know, Romance scholars have this prototype in front of them from the start, whereas for the Indo-European languages we have to reconstruct hypothetically the prototype of each form. Second, with the Romance languages it is perfectly possible, at least in certain periods, to follow the language from century to century through documents, and so inspect closely what was happening. These two circumstances reduce the area of conjecture and made Romance linguistics look quite different from Indo-European linguistics. It must also be said that Germanic studies to some extent played the same role as well. There the prototype does not exist, but in the case of Germanic there are long historical periods that can be followed.

The historical perspective that the Indo-Europeanists lacked, because they viewed everything on the same level, was indispensable for the Romance scholars. And the historical perspective revealed how the facts were connected. Thus it came about that the influence of Romance studies was very salutary. One of the great defects, from a scholarly point of view, which is common to philology and the comparative phase is a servile attachment to the letter, to the written language, or a failure to draw a clear distinction between what might pertain to the real spoken language and what to its graphic sign. Hence, it comes about that the literary point of view is more or less confused with the linguistic point of view, and furthermore, more concretely, the written word is confused with the spoken word; two superimposed systems of signs which have nothing to do with each other, the written and the spoken, are conflated. The linguistics which gradually developed in this way is a science for which we can take the definition given by Hatzfeld, Darmstetter and Thomas's Dictionary: 'the scientific study of languages', which is satisfactory, but it is this word scientific that distinguishes it from all earlier studies.

What does it take: 1) as its subject matter 2) as its object or task?

1) a scientific study 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.

Necessarily, it should be pointed out, in order to have documentation for all periods, as far as possible, linguistics will constantly have to deal with the written language, and will often have to rely on the insights of philology in order to take its bearings among these written texts; but it will always distinguish between the written text and what lies underneath; treating the former as being only the envelope or external mode of presentation of its true object, which is solely the spoken language.

2) The business, task or object of the scientific study of languages will if possible be 1) to trace the history of all known languages. Naturally this is possible only to a very limited extent and for very few languages.

In attempting to trace the history of a language, one will very soon find oneself obliged to trace the history of a language family. Before Latin, there is a period which Greek and Slavic share in common. So this involves the history of language families, as and when relevant.

But in the second place 2), and this is very different, it will be necessary to derive from this history of all the languages themselves laws of the greatest generality. Linguistics will have to recognise laws operating universally in language, and in a strictly rational manner, separating general phenomena from those restricted to one branch of languages or another. There are more special tasks to add; concerning the relations between linguistics and various sciences. Some are related by reason of the information and data they borrow, while others, on the contrary, supply it and assist its work. It often happens that the respective domains of two sciences are not obvious on first inspection; in the very first place, what ought to be mentioned here are the relations between linguistics and psychology - which are often difficult to demarcate.

It is one of the aims of linguistics to define itself, to recognise what belongs within its domain. In those cases where it relies upon psychology, it will do so indirectly, remaining independent.

Once linguistics is conceived in this way, i.e. as concerned with language in all its manifestations, an object of the broadest possible scope, we can immediately, so to speak, understand what perhaps was not always clear: the utility of linguistics, or its claim to be included among those studies relevant to what is called 'general culture'.

As long as the activity of linguists was limited to comparing one language with another, this general utility cannot have been apparent to most of the general public, and indeed the study was so specialised that there was no real reason to suppose it of possible interest to a wider audience. It is only since linguistics has become more aware of its object of study, i.e. perceives the whole extent of it, that it is evident that this science can make a contribution to a range of studies that will be of interest to almost anyone. It is by no means useless, for instance, to those who have to deal with texts. It is useful to the historian, among others, to be able to see the commonest forms of different phenomena, whether phonetic, morphological or other, and how language lives, carries on and changes over time. More generally, it is evident that language plays such a considerable role in human societies, and is a factor of such importance both for the individual human being and human society, that we cannot suppose that the study of such a substantial part of human nature should remain simply and solely the business of a few specialists; everyone, it would seem, is called upon to form as correct an idea as possible of what this particular aspect of human behaviour amounts to in general. All the more so inasmuch as really rational, acceptable ideas about it, the conception that linguistics has eventually reached, by no means coincides with what at first sight seems to be the case. There is no sphere in which more fantastic and absurd ideas have arisen than in the study of languages. Language is an object which gives rise to all kinds of mirage. Most interesting of all, from a psychological point of view, are the errors language produces. Everyone, left to his own devices, forms an idea about what goes on in language which is very far from the truth.

Thus it is equally legitimate in that respect for linguistics today to Claim to be able to put many ideas right, to throw light on areas where the general run of scholars would be very liable to go wrong and make very serious mistakes.

I have left on one side the question of languages and language in order to discuss the object of linguistics and its possible utility.


[4 November 1910]
Main sections of the course:

1) Languages 2) The language 3) The language faculty and its use by the individual.

Without for the moment distinguishing terminologically between languages and language, where do we find the linguistic phenomenon in its concrete, complete, integral form? That is: where do we find the object we have to confront? With all its characteristics as yet contained within it and unanalysed? This is a difficulty which does not arise in many other disciplines - not having your subject matter there in front of you. It would be a mistake to believe that this integral, complete object can be grasped by picking out whatever is most general. The operation of generalisation presupposes that we have already investigated the object under scrutiny in such a way as to be able to pronounce upon what its general features are. What is general in language will not be what we are looking for; that is, the object immediately given. But nor must we focus on what is only part of it.

Thus, it is clear that the vocal apparatus has an importance which may monopolise our attention, and when we have studied this articulatory aspect of languages we shall soon realise that there is a corresponding acoustic aspect. But even that does not go beyond purely material considerations. It does not take us as far as the word, the combination of the idea and the articulatory product; but if we take the combination of the idea and the vocal sign, we must ask if this is to be studied in the individual or in a society, a corporate body: we still seem to be left with something which is incomplete. Proceeding thus, we see that in catching hold of the language by one end at random we are far from being able to grasp the whole phenomenon. It may seem, after approaching our study from several angles simultaneously, that there is no homogeneous entity which is the language, but only a conglomerate of composite items (articulation of a sound, idea connected to it) which must be studied piecemeal and cannot be studied as an integral object.

The solution we can adopt is this:

In every individual there is a faculty which can be called the faculty of articulated language. This faculty is available to us in the first instance in the form of organs, and then by the operations we can perform with those organs. But it is only a faculty, and it would be a material impossibility to utilise it in the absence of something else - a language - which is given to the individual from outside: it is necessary that the individual should be provided with this facility - with what we call a language - by the combined effort of his fellows, here we see, incidentally, perhaps the most accurate way of drawing a distinction between language and languages. A language is necessarily social: language is not especially so. The latter can be defined at the level of the individual. It is an abstract thing and requires the human being for its realisation. This faculty which exists in individuals might perhaps be compared to others: man has the faculty of song, for example; perhaps no one would invent a tune unless the community gave a lead. A language presupposes that all the individual users possess the organs. By distinguishing between the language and the faculty of language, we distinguish 1) what is social from what is individual, 2) what is essential from what is more or less accidental. As a matter of fact, we shall see later on that it is the combination of the idea with a vocal sign which suffices to constitute the whole language. Sound production - that is what falls within the domain of the faculty of the individual and is the individual's responsibility. But it is comparable to the performance of a musical masterpiece on an instrument; many are capable of playing the piece of music, but it is entirely independent of these various performances.

The acoustic image linked to an idea - that is what is essential to the language. It is in the phonetic execution that all the accidental things occur; for inaccurate repetition of what was given is at the root of that immense class of facts, phonetic changes, which are a host of accidents.

3) By distinguishing thus between the language and the faculty of language, we see that the language is what we may call a 'product': it is a 'social product'; we have set it apart from the operation of the vocal apparatus, which is a permanent action. You can conjure up a very precise idea of this product - and thus set the language, so to speak, materially in front of you - by focussing on what is potentially in the brains of a set of individuals (belonging to one and the same community) even when they are asleep; we can say that in each of these heads is the whole product that we call the language. We can say that the object to be studied is the hoard deposited in the brain of each one of us; doubtless this hoard, in any individual case, will never turn Out to be absolutely complete. We can say that language always works through a language', without that, it does not exist. The language, in turn, is quite independent of the individual; it cannot be a creation of the individual-, it is essentially social; it presupposes the collectivity. Finally, its only essential feature is the combination of sound and acoustic image with an idea. (The acoustic image is the impression that remains with us the latent impression in the brain (D.)). There is no need to conceive it (the language) as necessarily spoken all the time.

Let us come down to details; let us consider the language as a social product. Among social products, it is natural to ask whether there is any other which offers a parallel.

The American linguist Whitney who, about 1870, became very influential through his book The principles and the life of language, caused astonishment by comparing languages to social Institutions, saying that they fell in general into the great class of social institutions. In this, he was on the right track-, his ideas are in agreement with mine. 'It is, in the end, fortuitous,' he said, 'that men made use of the larynx, lips and tongue in order to speak. They discovered it was more convenient; but if they had used visual signs, or hand signals, the language would remain in essence exactly the same: nothing would have changed.' This was right, for he attributed no great importance to execution. Which comes down to what I was saying: the only change would be the replacement of the acoustic images I mentioned by visual images. Whitney wanted to eradicate the idea that in the case of a language we are dealing with a natural faculty; in fact, social institutions stand opposed to natural institutions.

Nevertheless, you cannot find any social institution that can be set on a par with a language and is comparable to it. There are very many differences. The very special place that a language occupies among institutions is undeniable, but there is much more to be said-, a comparison would tend rather to bring out the differences. In a general way, institutions such as legal institutions, or for instance a set ,of rituals, or a ceremony established once and for all, have many characteristics which make them like languages, and the changes they undergo over time a.-e very reminiscent of linguistic changes. But there are enormous differences.

1) No other institution involves all the individuals all the time; no other is open to all in such a way that each person participates in it and naturally influences it.

2) Most institutions can be improved, corrected at certain times, reformed by an act of will, whereas on the contrary we see that such an initiative is impossible where languages are concerned, that even academies cannot change by decree the course taken by the institution we call the language, etc.

Before proceeding further, another idea must be introduced: that of semiological facts in societies. Let us go back to the language considered as a product of society at work: it is a set of signs fixed by agreement between the members of that society; these signs evoke ideas, but in that respect it's rather like rituals, for instance.

Nearly all institutions, it might be said, are based on signs, but these signs do not directly evoke things. In all societies we find this phenomenon: that for various purposes systems of signs are established that directly evoke the ideas one wishes; it is obvious that a language is one such system, and that it is the most important of them all; but it is not the only one, and consequently we cannot leave the others out of account. A language must thus be classed among semiological institutions; for example, ships' signals (visual signs), army bugle calls, the sign language of the deaf-and-dumb, etc. Writing is likewise a vast system of signs. Any psychology of sign systems will be part of social psychology - that is to say, will be exclusively social; it will involve the same psychology as is applicable in the case of languages. The laws governing changes in these systems of signs will often be significantly similar to laws of linguistic change. This can easily be seen in the case of writing - although the signs are visual signs - which undergoes alterations comparable to phonetic phenomena.

Having identified the language as a social product to be studied in linguistics, one must add that language in humanity as a whole is manifested in an infinite diversity of languages: a language is the product of a society, but different societies do not have the same language. Where does this diversity come from? Sometimes it is a relative diversity, sometimes an absolute diversity, but we have finally located the concrete object in this product that can be supposed to be lodged in the brain of each of us. But this product varies, depending On where you are in the world, what is given is not only the language but languages. And the linguist has no other choice than to study initially the diversity of languages. He must first study languages, as many languages as possible, and widen his horizons as far as he can. So this is how we shall proceed. From the study and observation of these languages, the linguist will be able to abstract general features, retaining everything that seems essential and universal, and setting aside what is particular and accidental. He will thus end up with a set of abstractions, which will be the language. That is what is summarised in the second section: the language. Under 'the language' I shall summarise what can be observed in the different languages.

3) However, there is still the individual to be examined, since it is clear that what creates general phenomena is the collaboration of all the individuals involved. Consequently we have to take a look at how language operates in the individual. This individual implementation of the social product is not a part of the object I have defined. This third chapter reveals, so to speak, what lies underneath - the individual mechanism, which cannot ultimately fail to have repercussions in one way or another on the general product, but which must not be confused, for purposes of study, with that general product, from which it is quite separate.


[8 November 1910]
Part One: Languages

This heading contrasts with that of my second chapter: the language. There is no point in giving a more detailed specification and the meaning of these two contrasting headings is sufficiently self-evident. Just as, although comparisons with the natural sciences must not be abused, it would likewise be immediately evident what was meant in a work on natural history by contrasting 'the plant' with 'plants' (c.f. also .'insects, versus 'the insect').

These divisions would correspond reasonably well even in content to what we shall get in linguistics if we distinguish between 'the language' and 'languages'. Some botanists and naturalists devote their entire careers to one approach or the other. There are botanists who classify plants without concerning themselves with the circulation of the sap, etc., that is to say, without concerning themselves with 'the plant'.

Considerations relevant to the language (and equally to some extent to languages as well) will lead us to consider languages from an external point of view, without making any internal analysis; but the distinction is not hard and fast, for the detailed study of the history of a language or of a group of languages is perfectly well accommodated under the heading 'languages', and that presupposes internal analysis. To some extent one could also say that in my second part 'the language' could be expanded to read 'the life of the language', that this second part would contain things of importance for the characterisation of the language, and that these things are all part of a life, a biology. But there are other things that would not be included: among others, the whole logical side of the language, involving invariables unaffected by time or geographical boundaries. Languages constitute the concrete object that the linguist encounters on the earth's surface; 'the language' is the heading one can provide for whatever generalisations the linguist may be able to extract from all his observations across time and space.


[30 June 1911]

Reversing the order of the two series I have considered, we can say that the mind establishes just two orders of relations between words.

1) Outside speech, the association that is made in the memory between words having something in common creates different groups, series, families, within which very diverse relations obtain but belonging to a single category: these are associative relations.

2) Within speech, words are subject to a kind of relation that is independent of the first and based on their linkage: these are syntagmatic relations, of which I have spoken.

Here of course there is a problem, because the second order of relations appears to appeal to facts of speech and not linguistic facts. But the language itself includes such relations, even if only in compound words (German Hauptmann), or even in a word like Dummheit, or expressions like s'il vous plait ['if you please'] where a syntagmatic relation holds.

When we speak of the structure of a word, we are referring to the second kind of relation: these are units arranged end to end as exponents of certain relations. If we speak of something like a flexional paradigm (dominus, domini, domino) we are referring to a group based on associative relations. These are not units arranged end to end and related in a certain way in virtue of that fact.

Magn-animus: the relation involving animus is syntagmatic. Idea expressed by juxtaposition of the two parts in sequence. Nowhere, either in magn or in animus do you find something meaning 'possessing a great soul'.

If you take animus in relation to anima and animal, it is a different order of relations. There is an associative family:


Neither order of relations is reducible to the other: both are operative.

If we compare them to the parts of a building: columns will stand in a. certain relation to a frieze they support. These two components are related in a wax which is comparable to the syntagmatic relation. It is an arrangement of two co-present units. If I see a Doric column, I might link it by association with a series of objects that are not present, associative relations (Ionic column, Corinthian column).

The sum total of word relations that the mind associates with any word that is present gives a virtual series, a series formed by the memory (a mnemonic series), as opposed to a chain, a syntagma formed by two units present together. This is an actual series, as opposed to a virtual series, and gives rise to other relations.

The conclusion I should like to draw from this is as follows: in whichever order of relations a words functions (it is required to function in both), a word is always, first and foremost, a member of a system, interconnected with other words, sometimes in one order of relations, sometimes in another.

This will have to be taken into account in considering what constitutes value. First, it was necessary to consider words as terms in a system.

As soon as we substitute term for word, this implies consideration of its relations with others (appeal to the idea of interconnections with other words).

We must not begin with the word, the term, in order to construct the system. This would be to suppose that the terms have an absolute value given in advance, and that you have only to pile them up one on top of the other in order to reach the system. On the contrary, one must start from the system, the interconnected whole; this may be decomposed into particular terms, although these are not so easily distinguished as it seems. Starting from the whole of the system of values, in order to distinguish the various values, it is possible that we shall encounter words as recognisable series of terms. (Incidentally: associatively, I can summon up the word dominos just as easily as domino, domine, domin-?; syntagmatically, I have to choose either dominos or domini.)

Attach no importance to the word word. The word word as far as I am concerned has no specific meaning here. The word term is sufficient; furthermore, the word word does not mean the same in the two series.


Chapter V. Value of terms and meanings of words.
How the two coincide and differ.

Where there are terms, there are also values. The idea of value is tacitly implied in that of term. Always hard to keep these two ideas apart.

When you speak of value, you feel it here becomes synonymous with sense (meaning) and that points to another area of confusion (here the confusion will reside more in the things themselves).

The value is indeed an element of the sense, but what matters is to avoid taking the sense as anything other than a value.

It is perhaps one of the most subtle points there is in linguistics, to see how sense depends on but nevertheless remains distinct from value. On this the linguist's view and the simplistic view that sees the language as a nomenclature differ strikingly.

First let us take meaning as I have represented it and have myself set it out:

image going into concept The arrow indicates meaning as counterpart of the auditory image

In this view, the meaning is the counterpart of the auditory image and nothing else. The word appears, or is taken as, an isolated, self-contained whole; internally, it contains the auditory image having a concept as its counterpart.

The paradox - in Baconian terms the trap in the 'cave' - is this: the meaning, which appears to us to be the counterpart of the auditory image, is just as much the counterpart of terms coexisting in the language. We have just seen that the language represents a system in which all the terms appear as linked by relations.

series of cricles joined by double-headed arrows At first sight, no relation between the a) and the b) arrows. The value of a word will be the result only of the coexistence of the different terms. The value is the counterpart of the coexisting terms. How does that come to be confused with the counterpart of the auditory image?

Another diagram: series of slots:

line divided up into sections the relation inside one slot and between slots is very hard to distinguish.

The meaning as counterpart of the image and the meaning as counterpart of coexisting terms merge.

Before example, note that: Outside linguistics, value always seems to involve the same paradoxical truth. Tricky area. Very difficult in any domain to say what value consists of. So let us be very wary. There are two elements comprising value. Value is determined 1) by a dissimilar thing that can be exchanged, and that can be marked | [an up-arrow] and 2) by similar things that can be compared <- -> [left-right arrows].

up-arrow with to and from arrows either side These two elements are essential for value. For example, a 20-franc coin. Its value is a matter of a dissimilar thing that I can exchange (e.g. pounds of bread), 2) the comparison between the 20-franc coin and one-franc and two-franc coins, etc., or coins of similar value (guinea).

The value is at the same time the counterpart of the one and the counterpart of the other.

You can never find the meaning of a word by considering only the exchangeable item, but you have to compare the similar series of comparable words. You cannot take words in isolation. This is how the system to which the term belongs is one of the sources of value. It is the sum of comparable terms set against the idea exchanged.

The value of a word can never be determined except by the contribution of coexisting terms which delimit it: or, to insist on the paradox already mentioned: what is in the word is only ever determined by the contribution of what exists around it. (What is in the word is the value.) Around it syntagmatically or around it associatively.

You must approach the word from outside by starting from the system and coexisting terms.

A few examples.

The plural and whatever terms mark the plural.

The value of a German or Latin plural is not the value of a Sanskrit plural. But the meaning, if you like, is the same.

In Sanskrit, there is the dual.

Anyone who assigns the same value to the Sanskrit plural as to the Latin plural is mistaken because I cannot use the Sanskrit plural in all the cases where I use the Latin plural.

Why is that? The value depends on something outside.

If you take on the other hand a simple lexical fact, any word such as, I suppose, mouton - mutton, it doesn't have the same value as sheep in English. For if you speak of the animal on the hoof and not on the table, you say sheep.

It is the presence in the language of a second term that limits the value attributable to sheep.

mutton / sheep / mouton (Restrictive example.)

So the | arrow is not enough. The <- -> arrows must always be taken into account.

Something similar in the example of decrepit.

How does it come about that an old man who is decrepit and a wall that is decrepit have a similar sense?

It is the influence of the neighbouring word. What happens to decrepit (an old man) comes from the coexistence of the neighbouring term decrepit (a wall).

Example of contagion.


[4 July 1911]

It is not possible even to determine what the value of the word sun is in itself without considering all the neighbouring words which will restrict its sense. There are languages in which I can say: Sit in the sun. In others, not the same meaning for the word sun (= star). The sense of a term depends on presence or absence of a neighbouring term.

The system leads to the term and the term to the value. Then you will see that the meaning is determined by what surrounds it.

I shall also refer back to the preceding chapters, but in the proper way, via the system, and not starting from the word in isolation.

To get to the notion of value, I have chosen to start from the system of words as opposed to the word in isolation. I could have chosen a different basis to start from.

Psychologically, what are our ideas, apart from our language ? They probably do not exist. Or in a form that may be described as amorphous. We should probably be unable according to philosophers and linguists to distinguish two ideas clearly without the help of a language (internal language naturally).

Consequently, in itself, the purely conceptual mass of our ideas, the mass separated from the language, is like a kind of shapeless nebula, in which it is impossible to distinguish anything initially. The same goes, then, for the language: the different ideas represent nothing pre-existing. There are no: a) ideas already established and quite distinct from one another, b) signs for these ideas. But there is nothing at all distinct in thought before the linguistic sign. This is the main thing. On the other hand, it is also worth asking if, beside this entirely indistinct realm of ideas, the realm of sound offers in advance quite distinct ideas (taken in itself apart from the idea).

There are no distinct units of sound either, delimited in advance.

The linguistic fact is situated in between the two:

This linguistic fact will engender values which for the first time will be determinate, but which nevertheless will remain values, in the sense that can be attached to that word. There is even something to add to the fact itself, and I come back to it now. Not only are these two domains between which the linguistic fact is situated amorphous, but the choice of connection between the two, the marriage (of the two) which will create value is perfectly arbitrary.

Otherwise the values would be to some extent absolute. If it were not arbitrary, this idea of value would have to be restricted, there would be an absolute element.

But since this contract is entirely arbitrary, the values will be entirely relative.

If we go back now to the diagram representing the signified and signifying elements together signified and signifier with up arrow

we see that it is doubtless justified but is only a secondary product of value. The signified element alone is nothing, it blurs into a shapeless mass. Likewise the signifying element.

But the signifying and signified elements contract a bond in virtue of the determinate values that are engendered by the combination of such and such acoustic signs with such and such cuts that can be made in the mass. What would have to be the case in order to have this relation between signified and signifying elements given in itself ? It would above all be necessary that the idea should be determinate in advance, and it is not. It would above all be necessary that the signified element should be something determined in advance, and it is not.

That is why this relation is only another expression of values in contrast (in the system). That is true on any linguistic level.

A few examples. If ideas were predetermined in the human mind before being linguistic values, one thing that would necessarily happen is that terms would correspond exactly as between one language and another.



cher ['dear']

lieb, teuer (also moral)

There is no exact correspondence.

juger, estimer
['judge, estimate']

urteilen, erachten
have a set of meanings only partly coinciding with French juger, estimer .

We see that in advance of the language there is nothing which is the notion 'cher' in itself. So we see that this representation: idea and auditory image although useful, is only a way of expressing the fact that there is in French a certain value cher delimited in French system by contrast with other terms.

It will be a certain combination of a certain quantity of concepts with a certain quantity of sounds.

So the schema signified and signifier is not the starting point in the language.

The value cher is determined on both sides. The contours of the idea itself is what we are given by the distribution of ideas in the words of a language. Once we have the contours, the schema can come into play. signified and signifing

This example was taken from vocabulary, but anything will do.

Another example. Idea of different tenses, which seems quite natural to us, is quite alien to certain languages. As in the Semitic system (Hebrew) there is no distinction, as between present, future and past; that is to say these ideas of tense are not predetermined, but exist only as values in one language or another.

Old German has no future, no proper form for the future. It expresses it by means of the present. But this is a manner, of speaking. Hence Old German present value is not the same as in French future.

Similarly if we take the difference between the perfective aspect of the verb and the imperfective aspect in the Slavic languages (difficulty in the study of these languages). In Slavic languages, constant distinction between aspects of the verb: action outside any question of time or in process of accomplishment. We find these distinctions difficult because the categories are unfamiliar. So not predetermined, but value.

This value will result from the opposition of terms in the language.

Hence what I have just said: The notion of value was deduced from the indeterminacy of concepts. The schema linking the signified to the signifying element is not a primary schema. Value cannot be determined by the linguist any more than in other domains: we take it with all its clarity and obscurity.

To sum up, the word does not exist without a signified as well as a signifying element. But the signified element is only a summary of the linguistic value, presupposing the mutual interaction of terms, in each language system.

Chapter VI

In a later chapter, if I have time: What I have said by focussing on the term value can be alternatively expressed by laying down the following principle: in the language (that is, a language state) there are only differences. Difference implies to our mind two positive terms between which the difference is established. But the paradox is that: In the language, there are only differences, without positive terms. That is the paradoxical truth. At least, there are only differences if you are speaking either of meanings, or of signified or signifying elements.

When you come to the terms themselves, resulting from relations between signifying and signified elements you can speak of oppositions.

Strictly speaking there are no signs but differences between signs.

Example in Czech: zhena, 'woman'; genitive plural, zhen.

It is clear that in the language one sign is as good as another. Here there is none.

(zhena, zhen functions as well as zhena, gen. pl. zhenu which existed previously.)

[This example shows that only the difference between signs is operative.

zhenu works because it is different from zhena.

zhen works because it is different from zhena.

There are only differences; no positive term at all.

Here I am speaking of a difference in the signifying element.

The mechanism of signifying elements is based on differences.

Likewise for signified elements, there are only differences that will be governed by differences of an acoustic nature. The idea of a future will exist more or less, depending on whether the differences established by signs of the language (between the future and the rest) are more or less marked.

Aller ['to go'] functions because it is different from allant ['going'] and allons ['(we) go'].

aller | allons | allant

English going = aller, allant

Unsegmented, given no acoustic difference between two ideas, the ideas themselves will not be differentiated, at any rate as much as in French.

So the whole language system can be envisaged as sound differences combined with differences between ideas.

There are no positive ideas given, and there are no determinate acoustic signs that are independent of ideas. Thanks to the fact that the differences are mutually dependent, we shall get something looking like positive terms through the matching of a certain difference of ideas with a certain difference in signs. We shall then be able to speak of the opposition of terms and so not claim that there are only differences (because of this positive element in the combination).

In the end, the principle it comes down to is the fundamental principle of the arbitrariness of the sign.

It is only through the differences between signs that it will be possible to give them a function, a value.

If the sign were not arbitrary, one would not be able to say that in the language there are only differences.

The link with the chapter entitled Absolute arbitrariness, relative arbitrariness is this: I have considered the word as a term placed in a system, that is to say as a value. Now the interconnection of terms in the system can be conceived as a limitation on arbitrariness, whether through syntagmatic interconnection or associative interconnection.

So: In couperet syntagma between root and suffix, as opposed to hache.

(Interconnection, syntagmatic link between the two elements.)

Hache ['axe'] is absolutely arbitrary, couperet ['chopper'] is relatively motivated (syntagmatic association with coupe ['chop']),


syntagmatic limitation absolutely arbitrary.

plu ['pleased']
plaire ['to please']

associative limitation

In this course only the external part is more or less complete.

In the internal part, evolutionary linguistics has been neglected in favour of synchronic linguistics and I have dealt only with a few general principles of linguistics.

These general principles provide the basis for a productive approach to the details of a static state or the law of static states.