OSWALD DUCROT SLOVENIAN LECTUR ES Introduction into Argumentative Semantics Ed. Igor Ž. Žagar PEDAGOŠKI INŠTITUT / 2009 DIGITALNA KNJIŽNICA / DISSERTATIONES / 6 S L O V E N I A N L E C T U R E S OSWALD DUCROT SLOVENIAN LECTUR ES Introduction into Argumentative Semantics Ed. Igor Ž. Žagar PEDAGOŠKI INŠTITUT / 2009 DIGITALNA KNJIŽNICA / DISSERTATIONES / 6 Oswald Ducrot, Slovenian Lectures, Introduction into Argumentative Semantics Znanstvena monografi ja/Scientifi c Monograph Zbirka/Series: Digitalna knjižnica/Digital Library Uredniški odbor/Editorial Board: dr. Igor Ž. Žagar, dr. Jonatan Vinkler, dr. Janja Žmavc, dr. Alenka Gril Podzbirka/Subseries: Dissertationes (znanstvene monografi je/Scientifi c Monographs), 6 Urednik podzbirke in izdaje/Editor-in-Chief: dr. Igor Ž. Žagar Recenzentki/Reviewers: dr. Mojca Schlamberger Brezar, dr. Jana S. Rošker Prevajalec/Translator: dr. Sebastian McEvoy Oblikovanje, prelom in digitalizacija/Graphic Design, Typesetting and Digitalization: dr. Jonatan Vinkler Izdajatelj/Publisher: Pedagoški inštitut Ljubljana 2009 Zanj/For the Publisher: dr. Mojca Štraus Naklada izdaje na CD-ju/Circulation: 50 izvodov Izdaja je primarno dostopna na:/Th e web edition is accessible at: http://www.pei.si/Sifranti/StaticPage.aspx?id=70 Imetnik stvarnih in moralnih avtorskih pravic na tem delu je avtor Oswald Ducrot. To delo je na razpolago pod pogoji slo-venske licence Creative Commons 2.5 (priznanje avtorstva, nekomercialno, brez predelav). V skladu s to licenco sme vsak uporabnik ob priznanju avtorstva delo razmnoževati, distribuirati, javno priobčevati in dajati v najem, vendar samo v neko-mercialne namene. Dela ni dovoljeno predelovati. Rightsholder on this work is author Oswald Ducrot. Th is work is licensed under a license Creative Commons 2.5 Slovenia (Atribution-NonCommercial-NoDerivs). Th is license allows you to copy, distribute, transmit and use the work for non-commercial purposes. You may not alter, transform, or bulid upon this work. Terms available here: http://creativecom-mons.org/licenses/by-nc-nd/2.5/si/deed.en_CA. Aktivnosti v okviru projekta Perspektive evalvacije in razvoja sistema vzgoje in izobraževanja omogoča sofi nanciranje Evropskega socialnega sklada Evropske unije in Ministrstva za šolstvo in šport. CIP - Kataložni zapis o publikaciji Narodna in univerzitetna knjižnica, Ljubljana 81:1(0.034.2) DUCROT, Oswald Slovenian lectures [Elektronski vir] : introduction into argumentative semantics / Oswald Ducrot ; prevajalec Sebastian McEvoy. - El. knjiga. - Ljubljana : Pedagoški inštitut, 2009. - (Digitalna knjižnica. Dissertationes ; 6) Način dostopa (URL): http://www.pei.si/Sifranti/StaticPage.aspx?id= 70 ISBN 978-961-270-014-0 247891968  Oswald Ducrot was born on November 27, 1930 in Paris. Aft er he graduated in philosophy from the École Normale Supérieure, he was teaching philoso- phy in several high schools in France. Before he became the research director at the École des Hautes Études en Sciences Sociales in Paris (1973), he worked at the Centre National de Recherches Scientifi ques (VNRS) and col- laborated with André Martinet. Oswald Ducrot was lecturing at several universities in Europe, North Africa, North and South America, and he held courses, among others, at the universities of Campi- nas, Stuttgart, Montreal, Geneva, Berkeley nad Freie Uni- versität in Berlin.  Table of Contents In Place of Introduction  Lecture I  Lecture II  Lecture III  Lecture IV  Lecture V  Major Publications of Oswald Ducrot  Books  Collections  Participation in Joint Publications  Articles  Reviews  Prefaces  Index  9 In place of introduction In December 1991 Oswald Ducrot gave a series of fi ve lectures on the nature of his theory of argumentation in the language-system to postgraduate students of Discourse Studies at the ISH, Institute for the Study of Human-ities in Ljubljana, Slovenia. Since, on the one hand, there is growing interest for Prof. Ducrot’s theory, while on the other hand very few of his works were translated into English (the modern lingua fr anca, whether we like it or not), we decided to republish these lectures in digital form under the title Slovenian lectures - Introduction into Argumentative Semantics (the fi rst, bilingual edition appeared in 1996, and is no longer available). Th e lectures were translated by dr. Sebastian McEvoy. Th e Slovenian lectures have been conceived as an introduction to the theory of argumentation in the language-system, and in Prof. Ducrot’s opinion don’t need a special introduction. Let me therefore just point out that in the last ten years prof. Ducrot, in collaboration with Marion Carel, has been developing a new version of the theory of argumentation in the language-systeme, a theory of semantic blocks. A lot of people collaborated at this publication; my special thanks, as the editor of the volume, go to dr. Sebastian McEvoy, Danielle Charon-net, Peter Altshul, Simona Suhadolnik, Marjeta Doupona-Horvat, Zoja Skušek, and especially to Jonatan Vink ler who made this digital publication possible. Of course, Prof. Ducrot’s lectures, as well as this publication would not have been possible without the fi nancial support from the Ministry  Slovenian Lectures of Higher Education, Science and Technology of the Republic of Slovenia and the Educational Research Institute. Igor Ž. Žagar  Lecture I December 9 To begin, I would like to say that I am extremely happy to be here in Ljubljana. Th is is my fi rst visit to Slovenia, and I already hope it will not be the last. If I am lucky enough to come here a second time, I hope I will be able to say two, perhaps even three words in Slovenian to you, but I am not making a promise on that point. I would also like to thank all those who have collaborated with me on this seminar: to begin with, those who invited me and who have organised this meeting, particularly Igor Žagar. I would also like to thank both the Slovene and French institutions which have made this seminar possible. In the hand-outs which outline the seminar, you are told that I am going to present one of the most interesting (I am quoting, of course) linguistic theories of our times. I am afraid that Igor may have been slightly optimistic in writing that, but, in any case, what I am going to speak to you about are things which, personally, I fi nd interesting and which, I can even say, have fascinated me for around twenty years now – perhaps even more – that is to say, ever since I became involved in linguistics and especially that part of linguistics called semantics. In the fi ve lectures of this seminar, I am going to try to give you an overall view of the work I have been doing and which, in fact, I am continuing to do even as I present it to you in its most recent form: in the fi rst lecture, the one I am going to give today, I will develop a certain number of general topics; in the second, I will speak about a particular theory, the theory of polyphony, which is the basis of all my work, and then, in the last three lec-  Slovenian Lectures tures, I will speak about the notion of argument, around which all my work is now centered. To give you a general idea of my work, I will begin with a common- place hypothesis which sociologists very oft en make and I believe justifi - ably so. Especially in recent years, sociologists have been saying constantly that all social activity produces a representation or an image of itself in and through its very exercise. Th at is to say, once people get together to do something, they also produce a representation of that group and of what that group does. Th at is true of the diff erent professions, which all con- struct images of themselves. It is true also of every social class. Th ere are sociologists who insist upon the fact that one of the characteristics of the low-er classes is that they reproduce the image which in fact the ruling class has constructed of them. For example, French peasants develop an image of peasantry; but that image of peasantry developed by peasants is the image the ruling classes of the nation, for example the town-milieux, have constructed. One of the principal functions and one of the principal uses of the social sciences, in my mind, is to try to make that image which social groups construct of themselves explicit, and, when necessary, to criticize that image. Th is work which is carried out in the social sciences is absolutely necessary, it seems to me, because the representation which social groups give of themselves seems so obvious to them that, in general, they do not feel the need to make it explicit, to think about it. What is true for social activity in general is also true, I think, for linguistic activity, which is simply one social activity among others. When you use a language, you develop a certain image of language in general. Where is that image of language, which a language itself imposes upon us, to be found? Well, I think, in the lexicon of a language, in its vocabulary, which has a certain number of terms to speak about linguistic activity. For example, almost all languages have words like mean, express, say, promise, allow, etc. All these words, taken together, constitute a sort of description of what linguistic activity is about. I think that the linguist as a researcher who is concerned with that social phenomenon which language is, must manage to make that – so to speak – spontaneous representation a language gives of itself explicit, clear, refl ective. Moreover, we linguists, if possible, must question that self- representation which language constructs about itself and which is, so to speak, crystallized in the lexicon of a language. Now, you realise immediately that this work of the linguist’s is particularly diffi cult. It is particularly diffi cult, because the linguist’s situation is an extremely peculiar one. Indeed, to speak about language, the linguist can-Lecture I  not but use language itself. When, as a linguist, I speak about a language, I use that vocabulary, words like mean, express, say, which language has built up to represent itself. So that the linguist introduces the image that language has built up of itself into his discourse on language, an image that the linguist would like to describe and make explicit. Th e result is that the lin- guist is constantly running the risk of falling into the snares of language. He is constantly running the risk of taking the very thing he would like to criticize, or at least discuss, for granted. In as much as I deal essentially with semantics, from the outset of my research, I have had to use the vocabulary which concerns that aspect of language. In particular, to speak about language, I have had to use words like mean or meaning, – and all linguists have to use those words or their equivalents to speak about language. Now, I think that what I want to say, what I have always wanted to say, and hope to make you want to say by the end of these lectures, is that, ultimately, words do not mean anything, that discourse never means anything. Th at slogan is a slightly paradoxical and dangerous one. A few words of explanation may make it more acceptable. At fi rst, for a semantician to say “words do not mean anything” seems self-destructive. What am I doing here if indeed words do not mean anything? Why have I come to speak to you about semantics? Th at is just about what the Russo-American linguist Roman Jakobson would say laughingly to those who, as I have just done, claimed that words did not mean anything. Jakobson adopted the same line of argument as is oft en used against the sceptics. Th e sceptics, as you know, say “nothing is true”. Th e usual objection is: “Well, if nothing is true, then the statement ‘nothing is true’ is not true either”. Jakobson had the same type of argument about meaning: “When you say that words mean nothing, well, you make a sentence which cannot mean anything either, so that logically, the statement that language does not mean anything is one which has absolutely no meaning itself ”. Th erefore, the formula words do not mean anything, Jakobson went on saying, is self-destructive. As I have absolutely no intention of committing suicide, I must uphold my slogan words do not mean anything without being exposed to the fi re of Jakobson’s objection. I will go about doing that in the following way: I will say that in the formula words do not mean anything, the word mean must be taken as having its meaning in ordinary language. If by mean you understand what is usually understood in ordinary language, then language indeed does not mean anything. But there might be a conception of meaning which diff ers from the conception recorded in the vocabulary and which does not force one to say that words do not mean anything.  Slovenian Lectures * * * First, I will quickly try to describe that conception of meaning through which language represents itself and which, according to me, if accepted, should indeed lead semanticians to commit suicide. Th at conception of meaning inherent in the word mean and inherent in the usual representation of the standard use of language, I will call the informative or descriptive conception of meaning. More precisely, what I will try to do is to develop a conception of meaning which is not informative or descriptive. So, what does that informative or descriptive conception of meaning consist in? It consists in saying that the fi rst function of speech or discourse is to convey an image of reality, to provide information on whatever happens to be the case. Let us try to consider that a little more closely. Why can the words we use when we speak give information on reality? I think that to answer that question, you must have recourse to a distinction which, to my mind, is an essential one: the distinction between what I call sentences and utterances. What I mean by a sentence is a linguistic entity: a sentence is an element in a given language-system itself. A language-system makes it possible to construct sentences by combining words in a certain way. When we speak, we use sentences but under the form of what I call utterances. To take a very commonplace example, let us suppose that having already said “It’s warm”, I repeat “It’s warm”. What you have is two diff erent utterances, each of which has been produced at a particular moment of time (at an interval of a few seconds); but those two diff erent utterances, “It’s warm” and “It’s warm” are utterances of the same English sentence It’s warm, the structure of which is unique. So, a language-system provides a certain set of sentences and then the speaker uses those sentences in the form of utterances. Th us, that unique English sentence It’s warm is used millions and millions of times. Now, having made that distinction, I will return to the main point, which, I remind you, is to describe the informative or descriptive conception of meaning. Th at conception consists in thinking that the fundamen- tal value of a sentence consists in its truth-conditions. To describe the English sentence It’s warm is to say under which conditions it is true and under which it is false; it is to say how the world must be for that sentence to be true. Similarly, to describe the English sentence Peter is intelligent is to say what Peter must be for you to claim truthfully that he is intelligent. So, to describe a sentence under that conception of meaning is to give the truth-conditions for the use of that sentence. Given that conception, that utter-Lecture I  ances of a sentence should have informative or descriptive value is understandable. Indeed, you will then say that when you utter a sentence, you are pointing out that the reality you are speaking about is such as to make the sentence true. When I utter “It’s warm”, I am telling you that in the world around us, the conditions making the sentence It’s warm true are fulfi lled. Th e utterances one produces in discourse provide information, they describe the world, because those utterances consist in affi rming that the conditions that make the sentence uttered true are fulfi lled. So, the informative conception of the semantic value of utterances is connected to what I would call a truth-conditional or logical or again, – I think that would be more precise, – pseudo-logical conception of the value of sentences. Sentences are described in terms of truth and falsehood (that is the reason why I call that conception of the semantic value of sentences a “ logical” one) and then, given that, you explain that the utterances of those sentences should convey pieces of information about the world. To my mind, it is that conception of meaning which is at the root of our use of the word mean in ordinary conversation and, according to me, it is in that sense that one must maintain that words do not mean anything. When I say that words do not mean anything, I mean that words do not give data, do not provide information about the world – or, at least, that they provide information only in an extremely indirect way. * * * Most linguists, I think, have questioned that truth-conditional or pseudo-logical conception of the semantic value of sentences, which is related to a descriptive conception of the semantic value of utterances, however commonplace that conception may be. I am therefore going to speak to you for a while about the history of linguistics to try to show that for quite a few centuries now, the majority of linguists have been questioning that truth-conditional and informative conception of meaning. But I shall try to show you that the doubts they have raised are not radical enough, not suffi cient-ly decisive: I shall try to introduce a hopefully more radical form of criticism. I have said that most linguists had already questioned that conception, and I shall now give you a few examples. As a fi rst example, I shall remind you of things which were said in France by the Port-Royal grammarians in the seventeenth century and which, with slight terminological dif-ferences, were adopted and systematized by the Swiss linguist Charles Bally, at the beginning of this century. Th en I shall try to show that their criti- cism is not radical enough.  Slovenian Lectures Th e starting point for both the Port-Royal grammarians and Bally is one of Descartes’ theories, a theory understood in a very simple way: I do not know if it is really Descartes’, but at least it is the theory which most people regard as being his. Th at conception consists in saying that there are two fundamental faculties in thought: on the one hand, there is the understanding (or intelligence) and on the other, the will. Th e understand- ing is a passive faculty: it consists simply in perceiving a certain number of ideas which are representations of the world, and then the will adopts attitudes towards those ideas. Let us suppose for example that the understanding conceives of the idea that Peter will come tomorrow. Th e will can adopt a certain number of attitudes towards that idea: it can affi rm that idea by saying “Yes, it’s true, Peter will come tomorrow”; it can also deny the truth of that idea by saying “Peter will not come tomorrow”; it can also question it by asking “Will Peter come tomorrow?”. So, according to Descartes, there are two faculties: one is passive, the understanding; the other active, the will. One could also say that there is an objective aspect and a subjective aspect in all thought. So that Descartes’ theory belongs to the great Western tradition that distinguishes object and subject. By the way, even if this distinction seems obvious to us, I believe that it is obvious only within our cultural framework: for us, it seems to go without saying, because modern civilisation is based on it; but the Arabic grammars of the Middle Ages, in the thirteenth century for example, were not based in the same way on the distinction between object and subject. Closing that parenthesis, I will simply remind you that, for Descartes, thought is made up of the understanding which passively conceives ideas and of the will which adopts attitudes relatively to them. For the Port-Royal grammarians and for Bally, language is a representation of thought and each sentence is a small image of a thought. Given that in thought, there is a cooperation of two faculties, one passive, the other active, the understanding and the will, there must be a mark both of passiv-ity and of activity in the very structure of a sentence. Th at leads the Port- Royal grammarians and Bally on to say that in every grammatical sentence, two aspects must be distinguished: the fi rst is the modus (I keep the Latin term), which represents the attitude of the will; the second is the dictum, which represents the idea as conceived by the understanding. If for example you are describing a sentence like Peter will come – I am taking a simple example – you will say that the dictum is the association of a subject ( Peter), of a verb ( come) and a tense (the future); and besides that something in the sentence expresses an attitude of the will, and that is the grammatical Lecture I  mood (in this case, the indicative). Th e indicative indicates that the speaker adheres to the idea, to the dictum, according to which there is Peter’s coming in the future. If the sentence were May Peter come! , if the sentence had that quasi-subjunctive form, you would have the same dictum again but a diff erent modus, which would be the subjunctive: while the indicative indicates the subjective belief of the human being who adheres to the dictum, the subjunctive here would indicate the will – or rather, to avoid ambiguity, let us say the desire of the human being who wishes Peter to come in the future. So, all the grammatical moods express psychological attitudes, – beliefs, desires, for example – relative to dicta. (If the word will has seemed inadequate to describe the subjunctive, it is because it is preferable to reserve it to indicate the diff erent attitudes expressed by the modus, whatever they may be, in a general way.) Well, I do not intend to examine this conception in detail: the point is simply to show that, on the one hand, it already questions the truth-conditional conception I spoke of earlier on but, on the other hand, that it does so in a way which is not radical enough. It does question the truth-conditional conception in so much as, according to the Port-Royal grammarians, a whole part of the meaning of a sentence, namely, the modus, is distinct from the part, namely, the dictum, which represents reality or, according to Bally, there is something else in our sentences than a representation of reality. Th e representation of reality is given by the dictum alone. To that extent, the purely truth-conditional or informative conception of meaning is already criticized. But to my mind, that criticism is not radical enough in as much as the modus, which for example indicates belief or desire is also, in a certain sense, a description of reality: simply, it is not the same reality as that described by the dictum. Th e dictum describes the outer world, the modus the inner world, but fundamentally, in both cases, there is a representation of a certain reality: an outer reality in one case, an inner reality in the other. Th at has led a certain number of philosophers of language to produce a representation of meaning which is fundamentally similar to Bally’s and the Port-Royal grammarians’ but which diff ers on one important point: the subjective element is not psychological in nature. Yes, as my second example, I am going to take speech-act philosophy as developed by English and American philosophers, especially Austin in England and Searle in the United States. It is essentially about Searle that I am going to speak. Austin, whose position is far less clear-cut and possibly far closer to my own, I am leaving aside.  Slovenian Lectures So, I am going to speak about Searle’s systematization of Austin’s conceptions. Searle tells us that two things must be distinguished in the meaning of any utterance. One, which he calls the propositional content, is a description of reality, which can be either true or false and which therefore must be described in truth-conditional terms, in terms of true and false. Th en there is another part, which he calls an illocutionary force, which is applied to that propositional content. Th e illocutionary force is a certain type of speech-act. To go back to my examples: should he describe an utterance like “Peter will come”, Searle would say there is a propositional content which is exactly the same as Bally’s or the Port-Royal grammarians’ dictum: that is, the idea of Peter’s coming in the future, a content which can be either true or false. For the illocutionary force, Searle would not specify a belief but a type of act: the act here is the act of affi rmation. Were you to take the second utterance “May Pe- ter come!”, you would have the same propositional content but a diff erent illocutionary force: you would have another type of speech-act, the act of ordering for example. In saying “May Peter come!”, you perform the act of ordering, which is associated with the illocutionary force of the utterance, in connection with a certain propositional content, that is Peter’s future coming. You see the diff erence between that conception and Bally’s or the Port-Royal grammarians’. In Searle’s conception, the active, subjective aspect of meaning is no longer a description of the speaker’s attitude. It is made up of the act which the speaker performs or, in my view, makes himself out to be performing as he speaks. When I affi rm that Peter will come, I am not in- forming you on my beliefs: I assume a certain responsibility regarding Peter’s future coming. Once I have said “Peter will come”, I must be ready to give you reasons justifying my affi rmation, I must be ready to admit that I made a mis- take should it appear that, in fact, Peter is not coming. Th e force thus consists in an obligation which the speaker places himself under. Th e same goes for an order: when I say “May Peter come!” to you, I am not telling you that I desire his coming to be the case but I am performing an order, that is to say that I am making you out to be placed under the obligation of making Peter come, or else, possibly, making Peter out to be placed under the obligation of coming. Th at is, I make out a certain person to be placed under a certain obligation. Th e fact of placing, of making someone out to be placed under a certain obligation is something altogether diff erent from the representation of a psychological desire: I have perhaps absolutely no desire for Peter to come, but even so I can give the order “May Peter come!” In speech-act philosophy then, there is a criticism of the descriptive, representational, informative conception of meaning, which is beginning Lecture I  to be more radical than Bally’s since the subjective aspect of meaning, that is the illocutionary force, is not at all a representation of reality, not even a representation of a psychological reality. It is used to constitute the act which the speaker performs as he speaks. So, in the conception of the philosophers of language and in particular, of speech-act philosophers, the subjective aspect completely loses its informative aspect or character. But, aft er the little pause we are going to have now, I am going to try to show that this conception of speech-acts nevertheless gives too important a role to description. Indeed, for Searle, there is a whole part of meaning, that is the propositional content, which is a pure description of the world and which can be described in truth-conditional terms. Well, I am going to try to show you that in fact, that propositional content is purely and simply an illusion. I am going to try to show that there is nothing – I stress: nothing – in meaning which is a description of reality. Th at is what I am going to do in a mo- ment, if I have the courage to do so. * * * Let me begin by giving you the gist of what I was saying a moment ago. I said that language like all social institutions and like all social activity develops an image of itself. Th at image is to be found in that part of the vo- cabulary which is about linguistic activity, for example words like mean or meaning. It seems to me that one of the major tasks for a linguist is to criticize that image which language gives of itself and particularly, to criticize the conception of meaning which language develops in, so to say, a spontaneous way. Th at spontaneous conception of meaning consists in thinking that the fundamental function of our utterances is to provide information on reality and, consequently, that the semantic value of sentences consists in truth-conditions: to describe a sentence semantically is to say under what conditions it is true or false. I tried to show that this informative or truth-conditional or pseudo-logical conception of meaning had been regularly questioned, especially, in a Cartesian perspective, by the Port-Royal grammarians or by Bally and also by English or American speech-act philosophers; but, to my mind, their criticism is not radical enough. I am not going to go into speech-act philosophy any further. Let me just remind you, however, that in Searles’s formulation, it consists in saying that the meaning of every utterance is made up of two parts: a propositional content and an illocutionary force applied to that content (the illocutionary force is the type of act which in speaking, the speaker makes himself out to be performing: an order, an affi rmation, a promise, and so on). What I reproach speech-act philosophy for is that  Slovenian Lectures it maintains an objective part, an informative part within meaning: a part called the propositional content. To try to show that the maintaining of an objective propositional content is ultimately incompatible with linguistic reality, I am going to take a number of examples, I am going to try to show that those forms of analysis which I have just mentioned, ultimately, are not very interesting; sometimes, they are even impossible. Let us take a fi rst example like “Th at fi lm’s interesting”. (It is an abso- lutely commonplace example.) Were I to analyse that utterance in Searle’s way, I would say there is a propositional content, the fi lm’s being interesting, and then, an illocutionary force, an affi rmation. One should ask: what can that propositional content, the interest in a fi lm, possibly be? According to Searle, the propositional content is something which has truth-conditions, which can be true or false. But in your opinion, can the truth-conditions of the idea the interest of a fi lm be defi ned? Can that notion the interest of a fi lm be given a defi nition like the following: it is true that a fi lm is interesting, if such and such a set of conditions obtain? Personally, I do not see how one could do that, I do not see how the notion the interest of a fi lm can be described in terms of true and false. For my part, I do not know an-yone who has managed to do so. Should you see a possibility of doing so, tell me: I would be very grateful. You can see why the philosophers of language have constructed that notion of a propositional content. Th e utterances that they use as examples are always, like the logicians’: “Peter will come”, “Two and two make four”, “Th e snow is white”. Th ere, you can speak of truth-conditions and even then, I am not so sure. But if you think about everyday utterances, like “Th at fi lm’s interesting”, what can their truth-con- ditions be? For my part, I can see only one solution for a semantic description of that utterance, – and I will develop that kind of solution in the following lectures, – which is to look for the conclusion in view of which one can be brought to produce such an utterance. What is the purpose of uttering “Th at fi lm’s interesting”? To answer that question, you must, according to me, ask yourself how you could go on. For example, you could say: “Th at fi lm’s interesting. Go and see it!” In other words, the utterance “Th at fi lm’s interesting” is not a possible argument for a conclusion like “Don’t go and see it!” It is diffi cult to see how it would be possible to say that without coming up with an extremely complex context. What you could do is say “Th at fi lm’s interesting but...” (the word but makes its fi rst appearance here, a word we will speak a lot about) ... “but don’t go and see it!” Other examples of the same type are “Th at meeting was nice” or “Th at evening was nice”. In Searlian terms, I should say that there is a proposition-Lecture I  al content the niceness of the party: but what on earth can the truth-condition of “Th at evening was nice” be? I think a description of language by means of the notions of true and false is completely impossible for sentences of that kind, and in what follows I will try to show that it is impossible for any kind of sentence. Now, I will take the example of a non-assertive sentence. Imagine the utterance “Be reasonable!” Someone does something which according to you, should not be done, and to prevent him from doing so, you say to him “Come on, be reasonable!”. If I were to describe that “Be reasonable!” in Searle’s way, I would say that there is a propositional content You are reasonable ( You will be reasonable rather, since we have the future here) and then the illocutionary force of an order. Th at is, in telling you to “Be reasonable”, I am ordering you to be something and that something is to be reasonable. Yet, one feels that such an analysis is almost absurd, on the one hand, because there are no truth-conditions for being reasonable and on the other, because when I use that sentence Be reasonable! , I am not ordering you to be reasonable. Indeed, it is obvious that one must be reasonable, everyone knows that, I do not need to give that order: in the very notion of a reasonable action, there is the idea of something that must be done. If I say “Be reasonable” in the context I have imagined, the order which I give you is not the order to be reasonable but the order not to do what you are doing. It is the order “Stop doing what you are doing!” When I use the word reasonable, it is to give you an argument in favour of that order. If you wanted a paraphrase for “Be reasonable!”, it would be “Don’t do whatever you’re doing, because it’s reasonable not to do it”. So, I think to describe words like reasonable or unreasonable, you have to bring in the idea that those words count as arguments for some conclusion or another. Another example, this time a personal experience. Th e other day in the street, I saw a child who wanted to stroke a dog. His parents, to prevent him from doing so, said “Leave it alone, it’s dirty!”. Th at is a common enough way of speaking: no doubt, my own parents oft en said such things to me. Let us try to describe “It’s dirty!”. If I were to describe it in Searle’s way, I would say that there is a propositional content the dog is dirty and then, the illocutionary force of an affi rmation. Th e parents affi rm to the child that the dog is dirty. Th at is, I think, what the parents do have in mind. Th e par- ents think that they fi rst describe the dog, in saying that it is dirty, then that they ask the child to draw the conclusion that “it must not be touched”. But now try to see things from the point of view of children. What can “Th at dog’s dirty” mean for a child? What does the word dirty mean for a  Slovenian Lectures child? For a child the word dirty characterizes something to be stood away from, to move away from. He has learned that word dirty because he has been told “Th at dog’s dirty, don’t touch it!”, “Th at food is dirty, don’t touch it!”, “Th at shirt’s dirty, change it!”, “Your face is dirty, go and wash it!”, and so on. So, as far as the child is concerned, the word dirty is nothing but an argument to stand away from the thing characterized as dirty. When someone says a dog is dirty, he is not describing the dog. Besides, the parents themselves, I think, would be put in a real spot if the child was philo-sophically-minded enough to ask them “Well, what exactly do you mean by dirty?” If the child asked that, the parents would have to say something like “Something which is dirty is something you mustn’t touch”, which means that strictly speaking, their utterance is a tautology: “Th e dog is something you mustn’t touch, don’t touch it”. So, I think that you cannot give an informational value to the word dirty: in its meaning, there is rather the indication of a certain number of conclusions that you can draw from the ascription of dirtiness to an object. I take a last example. Suppose someone suggests walking back home to my hotel. And I answer simply “It’s far away”. You can easily imagine the dialogue: “Would you like us to walk back to your hotel?” Answer: “It’s far away”. You all agree that answer “It’s far away” counts as a refusal of the suggestion to walk back to the hotel. If I had wanted to accept the suggestion – suppose it were a suggestion from someone whom I really wanted to go for a walk with – I would not have said “It’s far away”, I would have said “Oh yes, it’s nearby”. Notice the sentences It’s far and It’s nearby are true or false under exactly the same conditions: there are no truth-conditions which make the expression far accurate and the expression near(by) false. Simply, It’s far away works as an argument, at least in that situation, to refuse a certain suggestion. Up until now I have been content with describing far in the particular context I have imagined but if I were looking for a general description of far as opposed to near, I would say that the word is used to represent the distance in question as an obstacle. When I say “It’s far away”, I represent the distance which separates me from a certain place as an obstacle to my going there. In my example, it is a question of going to that place by foot, so my utterance “It’s far away” indicates an obstacle to the suggestion which was made to me to go to that place by foot. If I had been advised to go to a certain place by car, and I had answered “It’s far away”, I would perhaps have been saying either that I did not want to go to that place or else that one should not go to that place by car, but rather for example by train or by plane. However, there again, the function of “It’s far away” would have been Lecture I  to represent the distance as an obstacle. Let us consider now the suggestion I go by plane, and my answering “It’s far away”: well, in that situation, that answer is rather diffi cult to understand. You would have to imagine a socie- ty where you could move about by satellite, or something: under those circumstances, the answer might have some meaning. What I want to say is that, in the very meaning of far, there is the indication of conclusions which that word suggests about the distance. By describing a place as far, I am saying “that distance is an argument for not going to the place in question or at least, for not going there by the means of transport suggested”. It is certainly not the only interpretation of far: I mean, the word far can work for other conclusions. What I have said is that there is one line of argument in which far represents the distance as an obstacle. But especially in our modern world, the purpose of describing a distance as great by means of an adjective like far can also be to indicate that it would be interesting to overcome it. Let us imagine your suggesting two places where to spend our holidays: you propose a choice between spending the holidays in Egypt or spending them in India, and I answer “India’s farther”. Perhaps I am using the second argumentative principle: given that it is farther, it seems more interesting to me (in the modern world, it is one of the ideas attached to the idea behind the word far). In that case, the purpose of my answering “India’s far” can be to accept your project; but it is still, by means of a certain argumentative principle associated with the very meaning of the word far. In other words, the purpose of describing a place as far can be to describe a distance as an obstacle but it can also be to describe the distance as a reason to be interested in a place; and there are a certain number of other possibilities too. But, in any case, those diff erent argumentative possibilities should be introduced in the very description of the word. * * * Such are the kind of diffi culties that I fi nd in speech-act theory and which have led me to suppress the idea of a propositional content and of truth-conditions as far as possible. I will not quite manage that, because the theory is not complete yet (however, I have a few years ahead of me hopefully), but I am going to try to go as far as possible in that direction. To show you how I work a little, I am going to point out an objection that was made to me a few months ago. I was giving a lecture in which I was developing this type of idea and someone made the following objection to me: you say that the meaning of words is essentially argumentative, well, prove there is argumentativeness  Slovenian Lectures in an utterance like “Th ere are seats in this room”. (Suppose we are speak- ing about a room where a meeting is to be held.) Th en, the person who was making the objection, asked me: “Try to describe the utterance ‘Th ere are seats in this room’ in a way which is not informative or truth-conditional”. My position was very diffi cult because that type of utterance does seem to favour the conceptions of speech-act philosophy: nevertheless, I think I can answer such objections. To do that, I suggest we carry out an experiment. Take the two utterances “Th ere are seats in the room” and on the oth- er hand “Th ey’re uncomfortable”. Suppose now that to connect those two utterances “Th ere are seats in the room” and “Th ey’re uncomfortable” we have to choose between two types of expressions: and moreover, and fur-thermore, and besides or, second possibility , but. Let me ask you which of those words you would put to connect “Th ere are seats...” and “Th ey’re un- comfortable”? I think it is almost necessary here, except if one imagines a very complicated context, to put a but in. Let us go on with this experiment. Let us now take as a second utterance “Th ey’re comfortable”, the fi rst utterance remaining the same. I think that the most likely connection is of the and moreover type: “Th ere are seats... and moreover they’re comforta- ble”. Th ose facts, of course, do not allow me to say that there is no propositional or truth-conditional content in “Th ere are seats”. I have not proved that, and I must say that I would be hard put to prove it. But what my example does prove is that in the utterance “Th ere are seats” there is some- thing else than an informative, descriptive or truth-conditional content. And that, at this stage, is not so bad. To prove that point, let me remind you of a general hypothesis that I make on the connective but and equivalent connectives in other languages (I believe all languages have some equivalent of but). My hypothesis is that but always, I stress always, connects two utterances that have the opposite orientations. Let us call R the orientation of what precedes but and non-R the orientation of what follows but. (Of course, I have not justifi ed that description of but. However, I do ask you to accept it for the time being – how shall I say this – to understand my argument.) But connects two counter-oriented utterances. If one accepts that hypothesis, one must say that “Th ere are seats” and “Th ey’re uncomfortable” are two counter-ori- ented utterances. In “Th ere are seats”, there is a move towards the possibili- ty of sitting down and in “Th ey’re uncomfortable”, there is a move towards the impossibility of sitting down. I consider besides that and moreover always connects co-oriented utterances. Th ere again, I reach the same conclu- Lecture I  sion: I will say that “Th ere are seats” moves towards the possibility of sitting down and so too “Th ey’re comfortable”. * * * In what follows I will be speaking in a far more detailed way about but but I would like to fi nish this lecture by a very quick analysis of a small example which will show you how, in my opinion, that description of but (according to which but counter-orients the utterances it relates) can be applied. My example is a short text from Proust’s Swan in Love 1. Th e narrator shows how in the small aristocratic world he moved around in, Swan’s image came to be defi ned as from the moment when Odette’s in-fi delity to him was known (or imagined): In the past, people used to think: “he’s not handsome in a standard way if you like (a), but he has great style (b). What with that quiff , that eyeglass, that smile of his! (c)...” When Odette started not loving him any more, people thought: “he’s not positively ugly if you like (a’), but he’s ridiculous (b’). What with that eyeglass, that quiff , that smile of his! (c’)” With the abbreviations I have proposed, the two successive segments of discourse can be schematized in the following way, (c) and (c’) being super-fi cially quasi-identical: I = a but b. What with c! II = a’ but b’. What with c’! Th e counter- orientation which, according to me, is always implied by but is clearly confi rmed here: a moves towards an unfavourable appraisal of Swan whereas b, on the contrary, moves towards a favourable appraisal, and, conversely, a’ is favourable whereas b’ is unfavourable. A second feature of but is also confi rmed. It is the point that the overall movement of an X but Y string is the same as that of Y. Th us, the fi rst part of the fi rst discourse segment ( a but b, where b is the favourable appraisal “He has great style”), taken as a whole, amounts to praising Swan, at the time when the success of his relationship with Odette gave him the image of a happy man. Similarly, the fi rst part of the second segment of discourse ( a’ but b’, where b’ is the disparaging judgment He’s ridiculous) amounts, on the contrary, to criticizing Swan. Th us, what betokens Swan’s downfall is merely the switch in position of the favourable and the unfavourable judg-ments from one side of but to the other. 1 Paris: Gallimard, Pléiade: fi rst edition, pp. 319-320; second edition, pp. 314-315.  Slovenian Lectures A third remark will concern c and c’ given to support, respectively, a but b and a’ but b’. Formally, nothing indicates that they have diff erent values: they are more or less homonymous. However, every reader immediately feels that the eyeglass, the quiff and the smile, which are admired in the fi rst segment of discourse, are given as being ridiculous in the second, so that one must utter c and c’ with utterly opposed intonations. Th at fact illustrates the idea, which is to be discussed at length in the following lectures, that, in argumentative strings of discourse, the argument and the conclusion have no semantic value independently fr om one another. Here, the arguments draw their meaning from the conclusions (to interpret c and c’, one must know the conclusions they move towards). Th at is for me the es- sential diff erence between discursive argumentation and logical reasoning: in reasoning, the argument must have a meaning which is complete in itself, irrespectively of the conclusion; in argumentation, on the contrary, the argument can be understood only as from the conclusion (I would say as much of the conclusion in its relationship with the argument). It is besides, I think, what Proust is out to show in that and many other texts: it is not the case, for him, that our attitudes are determined by our knowledge of things but the reverse. Th at makes one think of Spinoza: “We do not like things because they are good; they are good because we like them”. A fourth and last remark on the notion of orientation. I have said that a (“He is not handsome in a standard way”) and a’ (“He is not positively ugly”) moved, respectively, towards criticism and praise, and that it was that feature which enabled them to be connected, through but, to praise (b) and to criticism (b’). But that does not mean that in themselves, they constitute criticism or praise. To say of someone that he is not positively ugly is a strange kind of praise indeed. You have there a characteristic feature of the notion of argumentative orientation, to which, eventually, the whole of my lectures will be devoted, for it is that feature that constitutes the way of perceiving language which I am going to try to get you to accept and which is far more important, for me, than the whole technical apparatus that I am going to set up: if “the theory of argumentation in the language-system” has meaning, if it says anything, that meaning consists only in getting one to feel, getting one to perceive utterances as having a certain orientation, and that irrespectively of the information they give and the inferences that one can draw as fr om that information. Th at Swan at the time of his worldly down- fall is not “positively ugly” undoubtedly implies that ugliness prevails over beauty in him. But it so happens that, in the segment, the utterance moves towards a favourable appreciation – a movement which is then countered Lecture I  by the but. To give a certain orientation to an utterance does not necessarily mean to make a conclusion acceptable. Th at brief analysis of a passage from Proust was designed, on the one hand, to illustrate the general defi nition which I give for but, a defi nition which I will use a lot in the following lectures and, on the other hand, to illustrate the idea that words change their value according to their argumentative orientation. Th e quiff , the smile of the second segment are not like the quiff , the smile of the fi rst. I believe that even the eyeglass is not exactly the same eyeglass: even if it is materially the same, it is not seen in at all the same way. So that there again we have grounds for criticizing the truth-conditional or informative conception of meaning, according to which a part at least of meaning could be described in purely descriptive or representational terms. Without being afraid of giving a more radical twist to my ideas, I am going to try in the lectures that follow to develop that theme, that is that the notions of true and false, like the notion of information, are notions which do not enable one to describe semantic reality, even if they are notions which belong to the way the word meaning and its equivalents are used in ordinary language. I think that indeed in ordinary language, that is to say in the representation that language gives of itself, the word meaning is viewed as referring to a certain description of reality; but I think that language, here, gives a false image of itself and that the linguist should not fall into that snare. We have four lectures to try and not to. To avoid doing so, I am going to develop two theories: on the one hand, the theory of polyphony, which we will be concerned with next time and on the other, topoi and argumentation theory, which we will be concerned with until the end of this series of lectures.  Lecture II December 10 In the course of the previous lecture, I set out the general objective of all my work. Th e objective is to try to exclude the notion of informativity from semantic description as much as possible and even, if possible, to do away with it altogether. Indeed, I have the impression that nowhere in utterance-meaning nor anywhere in linguistic meaning is there a description of reality, so that for me the notions of true and false do not seem adequate ones to describe linguistic facts. In saying that, I am, it seems to me, turning away from the usual notion of meaning as conveyed in dictionaries, that is to say I am trying to construct a notion of meaning which is diff erent from the notion of meaning built into language itself and linguistically imposed upon us. What I am trying to do, if you like, is to see language diff erently from the way language itself sees itself. Th at is my general objective: what must now be attempted is an explanation of the way I am going to achieve that objective. Two theories in particular can, I think, make one progress towards that objective: the theory of polyphony, which I am going to speak about today, and the theory of argumentation, which I will speak about in the coming three lectures. * * * Well, I am going to start by developing the notion of polyphony, a notion which my research is centered around. Traditional linguistics, it seems to me, rests on a notion which it considers as clear and as free from any ambiguity: the notion of speaker. I am going to try to show that the speaker, usually taken by everyone for a clear notion, is in fact an extremely confused one and one which covers a number of wholly diff erent ideas. When  Slovenian Lectures the Port-Royal grammarians speak of modus and when Searle speaks of illocutionary force, they always refer back to the idea of the speaker: the modus is the speaker’s attitude towards the dictum; Searle’s illocutionary force, which can be an affi rmation, promise, order, and so on, is the type of act that the speaker accomplishes in producing an utterance. I am going to try to show that in fact, that notion of the speaker, which is at the root of the modus- dictum or illocutionary force- propositional content distinction, is for several reasons ambiguous. (I think I am going to need some more chalk. I am a great chalk-eater, really!) It seems to me that when one speaks of the speaker, one understands three diff erent things which must be distinguished. Th ose three ideas which are confused in the notion of the speaker are the notions of what I call the producer ( le producteur empirique), the locutor ( le locuteur) and the enunciator ( l’énonciateur)1. What do I mean by the producer of an utterance? Th e producer of an ut- terance is the one whose activity results in the production of an utterance. Th e producer is the one who carries out the phonetic activity, also the intellectual activity necessary for the production of the utterance. Whenev-er there is an utterance, there is obviously a social actor who must carry out a certain activity for that utterance to be produced. I would like you to realize immediately how uncertain that notion of a producer is, and it is that uncertainty which is going to lead me to distinguish the notions of locutor and enunciator, which, to me, seem clearer. Given any particular utterance, it is extremely diffi cult, it seems to me, to say who exactly is its produc- er. Who is the psychological or social actor behind a given utterance? At fi rst, it seems a very simple question to answer but when one starts thinking about it, one realizes that it is, in fact, a very diffi cult one. Suppose my son is a pupil: the school organizes, let us say, a walk in the country, and for my son to be able to go on that outing, I must give my per-mission. Th e school staff therefore give my son a form for me to sign. So my son brings me the printed form, or at least a typed form, saying something like “I allow my son to take part in the school outing”. At the bottom of the form, there is the word “signature”, and what I must do is put my personal signature under the word “signature”. Well, who is the producer of that form saying “I allow my son to take part in the outing”? Is it I, I who have signed? No, certainly not. I have done very little. I have merely put my name at the bottom of the form. Is it the school secretary who has typed 1 Translator’s note. Th e terminology is no doubt unusual in English but it has seemed preferable to stick as closely as possible to the French here, as the theory of polyphony radically departs from current theories on the speaker. Lecture II  out the form? If you like, but the secretary has typed the form because a member of the school staff has dictated the text to her. And as for the person who has dictated the text, he or she has not invented the text either: to draft the text, he or she has followed the general directives set out by the education authorities. So, it is very diffi cult to say who has actually produced that utterance. Is it the secretary? a member of the school staff ? myself ? Th e identifi cation of the producer here is no simple matter. I take another example to show how uncertain a notion the producer is. Sociologists have shown in what seems to me to be an altogether conclusive way that most of what we say consists merely in repeating things we have heard before. Personally, I fi nd it extraordinarily striking to see that three quarters, perhaps even more, of what we say is a mere repetition of what we have heard others say. For example, when I go on the underground in Paris, I enjoy listening to people’s conversations, and I notice that those conversations merely reproduce newspaper headlines. Quite naturally, I say to myself: “Th is one’s repeating what I read in Le Monde, that one’s repeating things I saw written in Le Figaro” and so on. So, who really produces the utterances of people on the underground? In a certain sense, they are; but in another sense, it is Le Figaro, Le Monde, and very oft en newspapers far worse than either Le Monde or Le Figaro. Th e last example on this point. I read in a newspaper recently the results of an inquiry on the theme “What do the French know?” Th e au- thor of that inquiry was shocked by the results of his work because they made out the French to be incredibly ignorant people: especially, it came out that three quarters of French people did not even know that the Earth revolves around the Sun. Th e journalist was shocked by this state of aff airs. Well, what astonishes me in this business is not that three quarters of the French people do not know that but that a quarter of them should! Because, if you stop and think, I think that there is no-one here, in this room, myself included, who really knows that the Earth revolves around the Sun. To know that would be to have some reasons to believe it. If someone asked me to prove that the Earth revolves around the Sun, even to give arguments in favour of that conclusion, for my part, I would not really know what to say. Besides, one would have to know exactly what “Th e Earth revolves around the Sun” means. Th at is not clear at all. When we say “Th e Earth re- volves around the Sun”, we are merely repeating something we have learnt at school but we have no inkling as to what it means. We cannot therefore say that intellectually, we really are the producers of that utterance. We are merely parrots.  Slovenian Lectures All that goes to show that the producer is not at all a clear notion. Now, it so happens that, as a linguist, I am fortunate enough not to have to worry about the producer. Why? As a linguist, I am interested in what is within utterances, in the meaning which utterances convey. I am not interested, as a linguist, in the causes that utterances stem from. So, as a linguist, I do not have to worry about the producer: as a linguist, I do not have to know if the utterances of people on the underground come from Le Monde or Le Figaro or some other newspaper. What I try to do is to describe what is within utterances. So, the problem of the producer, which is a very real, a very important one, is not my problem. It is a problem for psychologists, it is a problem for sociologists but I do not think that the linguist must deal with it. Unfortunately, the linguist must, on the other hand, deal with two other notions, which are also very diffi cult ones. Let me begin with the notion of the locutor. What do I mean by the locutor? What I mean is the person who, according to the very meaning of an utterance, is the person responsible for that utterance. I would like to stress that point: he is the person who is designated, in the utterance itself, as being the person responsible for that utterance. Th at person, who is responsi- ble for the utterance, it so happens, is in most languages, I think I can even say in all languages, designated by a particular morpheme, the fi rst person, that is to say in English: I, and also, of course, all the other grammatical forms that refer back to the fi rst person, such as the possessive adjective: my. Th e same reasons that justify my not dealing with the producer require my dealing with the locutor. I must deal with that person, because he or she is denounced, indicated in the very meaning of an utterance. I would like to show that the producer and the locutor, as I have just defi ned them, are, ultimately, two extremely diff erent notions. I shall attempt to do so with a number of examples. First, I am going to take a rather colloquial French expression and which, so I have been told, can be translated more or less exactly in Slovenian. Say we have two characters, Mr A and Mr B. Suppose that Mr A has been seriously indiscrete towards B: Mr A has looked through Mr B’s papers without having the least right to do so. Mr B catches Mr A looking through his papers and B, to reproach A for his indiscretion, says to him: “De quoi je me mêle?” or “De quoi je m’occupe?”, which is a very current French expression [more or less equivalent in meaning but not in structure to “Mind your business!” or “What business is it of yours!”]. Th at clearly means: “What are you, A, interfering with?”, and suggests: “You, A, are interfering in things which do not concern you!” If one looks for the produc-Lecture II  er of the utterance B addresses to A, the producer (I have said that one does not always exactly know who the producer is but here the situation is clear) is B. It is B who carries out the phonetic activity and makes the grammatical choices involved in the production of that utterance. But the je here does not refer back to B at all. Th e sentence does not mean “What is B in- terfering with?” Th e sentence means “What is A interfering with?” So that, to analyse such an example, I cannot but say that the je does not refer back to the producer at all but to someone, the locutor, who within the utterance, is represented as being the imaginary author of the utterance. In other words, B tries to imagine a question which A should have asked himself before committing his indiscretion. If I were to paraphrase B’s utterance, I would do it in the following way: “Before going through my things, you, A, should have asked yourself ‘de quoi je me mêle?’” Mr B thus makes Mr A speak. Another possible paraphrase but which, as far as the theoretical status of the locutor is concerned, leads to the same conclusion, would be: “Aft er committing your indiscretion, you should have an uneasy conscience and say ‘de quoi je me mêle en regardant les papiers de mon ami B?’” In any case, we have an example here in which the locutor, understood as being the person whom the I pronoun designates as being responsible for the utterance, cannot be identifi ed with the producer at all. I am now going to take another, perhaps more universal example than the fi rst (which perhaps can be found only in a few languages). It is extremely easy to use the I pronoun to designate persons, objects, which are incapable of the physical action which speech requires. I remember while travelling in Germany, seeing a sign posted up on the door of a shop bearing the words “Ich muss draussen bleiben”, that is to say “I must stay outside” and what stood for the signature under the utterance was the picture of a dog. Th e whole thing meant that dogs were not to go into the shop. Who is the producer of that sign? Perhaps the shop-keeper, perhaps the mayor of the German town. I do not really know who. But I do know who or what the locutor, designated by Ich, is, and that is all that matters to me, as a linguist: the locutor is the dog. A well-disciplined German dog is pictured, before going into the shop, as saying to himself: “Careful now, I must stay outside”. A creature is thus pictured as speaking who cannot be the producer of the utterance. Th at way of speaking is an extremely frequent one. Th ere is a sign oft en on dustbins such as “Don’t think twice about using me!” Th ere, it is the dustbin which speaks to passers-by and says: “Th row your cigarettes and throw your pieces of dirty paper inside this dustbin!” Th e producer, which, in this example, are the local authorities, makes the dustbin speak.  Slovenian Lectures You are aware besides that this distinction between the producer and the locutor is very close to the distinction made in the study of narrative between the author, who is on the side of the producer, and the narrator, who is on the side of the locutor. When in one of Proust’s novels, you have sentences in which there is the I pronoun, for example “For a long time, I went to bed early”, the I who went to bed early is not Proust the author, the producer of that utterance, but the character, the person who is telling the story as being his personal story and who, of course, is not necessarily Proust himself. Let me take another example to bring out this distinction between the locutor and the producer. You can certainly think of types of discourse that have no locutor, that is to say types of discourse in which there is no I and in which it would be impossible to bring in an I. You know that Benveniste distinguished two ways of speaking: on the one hand, what he called discours, that one fi nds, for example, in conversation, in poetry, in debates, etc. and, on the other, histoire, in which events, wholly disconnected from the person relating them, are recorded. Benveniste said, and he put this point perhaps too forcefully, that in historical narrative, there can be no marks of the fi rst person, there can be no I. When you write the history of Rome, you cannot introduce an I into that history. Th at is perhaps putting the point too forcefully, because there are probably certain passages of a historical narrative in which the author can introduce an I (if he makes some personal commentary about that history of Rome) but on the whole, one can say that historical narrative does not have a fi rst person. When a historian, the producer of the narrative, wants to write the history of a certain period, he writes as if history itself were telling the story, as if there were no particular person responsible for it. In that case, the historian is the producer but there is, properly speaking, no locutor. You know that there are also other ways of speaking, in which the locutor does not appear. Take the case of proverbs. In a proverb, you cannot bring in an I. A proverb is always an impersonal utterance: it is the utterance of no-one in particular. Now, proverbial speech is not an extraordinary phenomenon. Perhaps in modern societies, so-called intellectual societies, one has a tendency to forget about proverbs, but in old rural societies (I remember the society I was raised in as a child), peasants liked and perhaps still like to bring in proverbs into their talk, and in societies like the Arabic society, a very great part of what people say is made up of proverbs, which give force to what is said. Now, a proverb is essentially something that has no locutor, which does not mean that the utterance itself has no producer. Lecture II  So, I think that one must distinguish those two functions: the function of the one who is represented in a piece of discourse as being responsible for it, the locutor; and the function of the producer, that is to say the one who actually produces that piece of discourse. I move on now to the third notion, which seems to me to be encapsulated in the general idea of the speaker (it is the one I am going to speak about most here): the notion of the enunciator. I would like to point out that the term enunciator is a term that I have chosen in a wholly arbitrary way: it must be understood here as having no other meaning than the one I am about to defi ne. I have perhaps chosen an inadequate, a misleading term but the choice is behind me now and I cannot go back on it. What do I mean by enunciator? For me, all utterances represent one or several points of view: by enunciators, I mean the sources of those diff erent points of view which are represented within an utterance. In an utterance, one represents the state of aff airs one is speaking about as seen from the point of view of one or several persons: the enunciators are those persons from whose point of view the state of aff airs is viewed. I am going to try to show that the enunciator, as the source of a point of view, is wholly diff erent from the locutor. To do so, I am going to give you a certain number of examples in which there is no coincidence between the locutor, designated by marks of the fi rst person and, on the other hand, the enunciator, the source of the dif-ferent points of view represented in an utterance. First, I am going to take a very simple and quite untechnical example and then, I shall take examples which will require getting into linguistic technique properly speaking. My fi rst example concerns what are called echoic utterances. Suppose Mr A and Mr B are arguing. A says to B: “B, you’re a fool!” B answers using the very words A has used: “So, I’m a fool, am I! Well, just you wait!” It is not necessary for B to add anything aft er “just you wait!” but he can do so. Who is the locutor of that utterance of B’s “I’m a fool”? In other words, who is the person designated by the I pronoun? It is undeniably B. Indeed B is merely taking up the idea, if that can be called an idea, which was put forward by A and according to which B was a fool. So, the locutor, designated by I is B. Now, whose point of view is represented here? Unless B is incredibly lucid about himself, that point of view is not his: he does not at all claim to be a fool. Th e point of view which is represented is A’s: B in his utterance is expressing the point of view of a person diff erent from himself, a point of view of which he undoubtably does not approve at all and even, one can suppose, which he vehemently rejects. So, what we have here is a dissocia-  Slovenian Lectures tion between the locutor from the enunciator. Th e locutor is undeniably B but the point of view which is represented is foreign to B. Now, I am going to take more technical examples to try to show you that this dissociation of the locutor and the enunciator is an extremely frequent one. For me, it is a very general phenomenon that in discourse and even in a single utterance, one represents a certain number of points of view and very oft en points of view that one does not hold oneself accountable for and even which one rejects. To show this, I am fi rst going to consider the phenomenon of negation, which I am going to analyse in some detail. I begin by giving the general idea, the development of which will be presented aft er a very small pause. My analysis of negation is strongly in-spired by a famous paper which Freud wrote on negation, an article all the aspects of which, especially the psychological implications, I am not concerned with, but which, according to me, contains an idea which is linguistically very true. Freud’s idea is the following. A negative utterance – I symbolise a negative utterance by non-X – a negative utterance is a compromise proposed by the Ego, in its attempt to symbolically reconcile on the one hand, the Super-Ego and on the other, the Id, the deep urges of the libido (the Es in German). A negative utterance is thus a compromise which the conscious personality establishes between two unconscious authorities, the libido and moral censorship. For Freud, when someone says non-X, he is saying two things at the same time: on the one hand, he is saying X and on the other, he rejects X: it is the libido which says X, the Super-Ego which rejects it. In other words, to say non-X is at the same time to make a hid-den representation of the libido appear and at the same time, to symbolically satisfy morality by the rejection of that idea X, which is being denied. Negation is thus a kind of strategy, invented by the Ego, to fulfi l two con-tradictory requirements at the same time. An example which Freud gives is the one of patients who, as they describe the dreams they have had the previous night, say “Well, in that dream, I did not kill my father”. According to Freud, when a patient says that, one can be sure that in that dream, the patient has in fact killed his father. Th at is what the patient really wants to say to his analyst. But he says it in a negative form, imparting that murder of the father with a negative form which prevents censorship from repressing it. Before we stop for a pause, I would like to point out that one fi nds something similar in certain pictures or sculptures of the Middle Ages depicting the cardinal sins. Th ose representations are sometimes astonishing: the precision, the crudity and sometimes the voluptuousness with which the sins are depicted is quite extraordinary, and are such as to be unrecon-Lecture II  cilable with our understanding of the moral environment of mediaeval society; but if the painter or the sculptor of the Middle Ages can represent those images, it is because he represents them as images of sin, as images of actions which are to be condemned and, because of that moral condemnation, he manages to have those representations accepted. It seems to me that the title Sins which a sculptor or painter of the Middle Ages gives to his work works very much like linguistic negation: in saying Sins, he carries out a kind of negation on those images. He gives an extremely crude representation of pleasure, for example, and what allows him to represent pleasure in that extraordinarily vivid way is that, at the same time, he has negat-ed it by calling it Sin. In what follows, as unfortunately I am only a linguist, I am not going to develop the psychological aspect of Freud’s theory of negation, a thing which is quite beyond my scope. Besides, there is probably something slightly excessive in that theory of Freud’s; but it is Freud’s general idea which seems to me a very useful one and which for my part, I constantly use in the study of negation. My idea is the following: if I have to describe a non-X utterance, I say that the non-X utterance represents two points of view, in other words two enunciators, E1 and E2. So there are, if you like, two persons who express themselves through a non-X utterance, and this I will try to show by using strictly linguistic arguments: the fi rst person, enunciator E1, has a point of view which corresponds to the positive part of the utterance, that is to say X; and enunciator E2 disagrees with E1; but the two points of view are simultaneously present. Aft er the pause, as from a certain number of examples and as from a certain number of linguistic facts, I am going to try to show you that it is almost necessary to accept that representation of negative utterances: a negative utterance, properly speaking, has no unity... * * * Well, I was in the process of introducing the theory of polyphony, and more especially the notion of enunciator, using the example of negation. I remind you that for me a non-X utterance represents two enunciators, one whose point of view is represented by the positive part, X, and the other, E2, who disagrees with E1. Let us now apply that idea to a very simple sentence: “John won’t come”. We have an E1, whose point of view is that John will come and an E2, who disagrees with E1. How is one to justify that analysis? To start with, I shall be using what might be called syntactic kinds of proofs, by showing that this form of analysis enables one to understand the phenomenon of anaphora, that is to say of pronominal back reference,  Slovenian Lectures a phenomenon which it would oft en be very diffi cult to understand with- out a polyphonic analysis. Let us suppose that someone, having said “John won’t come”, adds ... “and I regret it, because it would have been nice”. In the second utterance we have two anaphoric pronouns, the it in “I regret it” and the it in “it would have been nice”. You can see immediately that the two occurrences of it do not refer back to the same thing at all. Th e it in “I regret it” refers back to the idea that “John won’t come”: what is regretted is that “John won’t come”. I will say therefore that it refers back to enunciator E2’s point of view. On the other hand, the second it refers back to enunciator E1’s point of view, that is to say, it refers back to John’s future coming: what would have been nice would have been John’s coming. If, within the “John will come” utterance, one has not distinguished the positive and the negative parts, it is extremely diffi cult, it seems to me, to explain the phe- nomenon of anaphora. Th at is true for explicit anaphora, like the one I have taken as an example, but it is true also for anaphora that one could call implicit. On that point, I am going to analyse a small text I found in a newspaper concerning the civil war in the Philippines. Th e journalist said that the Philippine rebels were completely independent and were not set on or supported by foreign powers but in the following utterance, the journalist said: “No state (I am thinking of China) is supporting the rebels”. Th at is exactly what he wrote, I assure you. At fi rst sight, that utterance is an extremely surpris-ing one. If no state is supporting the rebels, how can the journalist think of China in particular? For that parenthesis, “I am thinking of China”, to be understandable, it must (I can see no other solution) be analysed polyphonically. In “No state is supporting the rebels”, there are two points of view: the E1 point of view, according to which “A state is supporting the rebels”, and an E2 point of view, which disagrees with E1’s. Th e “I am thinking of China” does not concern the whole of the negative utterance but only the positive component which represents E1’s point of view or opinion, according to which a particular state is supporting the rebels: that is the opinion which the journalist is alluding to when he puts in “I am thinking of China”. If one does not have a polyphonic analysis, I fail to see how one can understand that kind of anaphora. Now, I will use examples which, properly speaking, are no longer syntactic but rather, let us say, pragmatic. Let us imagine a family quarrel between Mr A and Mrs A. Mrs A asks her husband, Mr A, to make himself useful about the house and do a certain number of things, for example clear the table, do the washing-up, then take the rubbish down. Mr A, annoyed Lecture II  by these requests, answers his wife in a somewhat polemical way: “I can’t do everything”. Much to my shame, I must confess that I must have given answers of that kind a few times – even if they are really shameful. Let us try to explain Mr A’s utterance “I cannot do everything”. Let us fi rst suppose that one does not have polyphonic analysis at one’s disposal. One will say that Mr A provides an argument “I cannot do everything” for the conclusion “I cannot take the rubbish down”. It is the only possible analysis, without the theory of polyphony. Th at analysis does not seem very interesting to me be- cause the force of Mr A’s utterance is really too weak. Th e fact that he can- not do everything, which is undeniable, in no way prevents him from nevertheless taking the rubbish down: surely, he does not need to be all-power-ful to take the rubbish down. One obtains, it seems to me, a far more interesting analysis if the theory of polyphony is adopted. In “I cannot do everything”, Mr A represents two enunciators: one enunciator (E1), according to whom Mr A can do everything and another enunciator (E2), who disagrees with E1. Who is that enunciator E1, whom Mr A puts forward, according to whom Mr A could do everything? Th e answer is, I think, quite easy: E1 is Mrs A. Mr A makes out Mrs A as having contended, to justify her request, that her husband can do everything. As that is obviously absurd, Mr A can side with enunciator E2, who disagrees with that absurd idea attributed to Mrs A. So, in that utterance, Mr A is something of a director, making up a little play, in which a woman tells him in an absurd way that he can do everything, and he steals the show, playing the part of a reasonable man – instead of putting forward a poor argument, as the non- polyphonic interpretation necessarily portrays him as doing (here, on the contrary, he argues in a perfect way by rejecting his wife’s abusive claim). One understands the strategy used in that discourse far better, I think, if the negative utterance is analysed polyphonically. I take another example of a discursive strategy which polyphony can bring out easily. Let us suppose that a mother leaves her fl at to go and do some shopping, leaving her children in the fl at. When the mother comes back home, aft er having done her shopping, she notices that a vase has been broken. Of course, she accuses the children of having broken it, and all the children say that they have not broken it: “No, no, we haven’t done anything, we haven’t been playing football around the house, we’re absolutely innocent”. Th e mother (I can imagine my own mother making an objection of this sort very well) tells her children “Well, I didn’t break it”. I think that a polyphonic analysis of that utterance of the mother’s “Well, I didn’t break it” enables one to understand the strategy used by the mother to accuse her  Slovenian Lectures children. She represents a fi rst enunciator E1, according to whom it is the mother who has broken the vase, and, undeniably, enunciator E1 is the children. She makes the children out as being forced to say “You have broken the vase” to defend themselves; and the mother has no diffi culty then in re- butting that justifi cation of the children’s, as she was not there at the material time. So, in saying “Well, I didn’t break it”, the mother makes the children speak, and she makes them do so in an absurd way. Th e mother, in the case in point, does exactly what Mr A, in the previous example, did when he spoke to his wife: they both put claims in their interlocutors’ mouths which makes their interlocutors seem ridiculous. In France and, I believe, in many Christian countries, there is a formula which enables one to replace the I in my . My mother could have said, not “Well, I didn’t break it”, but “Well, the Holy-Ghost [le Saint-Esprit] didn’t break it”, the Holy Ghost’s function being to do impossible things. My mother would thus have imagined in her discourse that, to defend myself against the accusation, I would have to ascribe the responsibility to the Holy Ghost. Th e remarkable thing is that ne- gation is not even useful here. My mother could simply have said, ironically, “In that case, the Holy Ghost [le Saint-Esprit] must have broken it”, without negation, but with irony in her tone of voice; and the irony would have played exactly the same role as negation. A last example to show the usefulness of polyphonic analysis. I will show how, through negation, one manages to construct an image of the other in one’s own discourse. I go back to my example “John won’t come”. “John won’t come” said by A. One can very well imagine that B, A’s addressee, answers something like “But I never said John would come!” One could even have a shorter version: “But I never said that!” Th e interesting thing is that in his utterance “But I never said that!”, B takes up the polyphony introduced by A. B interprets A’s utterance as representing E1 and E2, with E1 thinking that John is coming. B gives that analysis of A’s utterance and, moreover, he makes out enunciator E1, whom A represents in his utterance, as being identical with himself. B makes A out as having identifi ed B with enunciator E1. Th at is to say, B makes A out as having made him speak when A represented enunciator E1, according to whom John would be coming. Or in other words: B imagines that, in A’s speech, there is a representation of the addressee, that is to say B, with which B disagrees. And one could easily imagine the dialogue going on. A could defend himself saying “But I wasn’t thinking of you, I was thinking of somebody else, of C, who had claimed that John would come”. Suppose that A answered “I wasn’t thinking of you, I was thinking of somebody else, of C, who had claimed Lecture II  that John would come”. Th en, A, in that third utterance, gives a polyphonic interpretation to the fi rst utterance again, simply he proposes another identifi cation for enunciator E1. I think that there are a lot of discursive practices that rest on the ambiguity there is in the identifi cation of enunciators. * * * Now, I would like to apply this analysis of negation to the study of a short, and this time real, literary text, which will give me the opportunity, among other things, of conjuring up not only two but, – wait and see, – three enunciators. In certain analyses of mine, – but I will not present them, because it would take too long, – there are four, fi ve, six, even seven enunciators: the number of enunciators is absolutely unlimited. Th e example I am going to analyse seems interesting to me, at least interesting for the following lectures, because, in the course of the analysis, I will bring in certain notions about argumentation which were mentioned in the last lecture and which I will have to go into in greater detail in what follows. Th e text in question is a short text from Pascal [which to quote I am going to “modern-ise”, because it is written in a slightly old form of French, which is relatively diffi cult to understand but, – trust me, – I have not falsifi ed the facts]. Th e title of the text – very important for what follows – is Sound Opinions of the People. So, Pascal is going to indicate a certain number of opinions of the people which seem sound to him. And under that title, one fi nds the following text: “To be elegant is not so vain”. Pascal then goes on to explain why being elegant is not so vain. Being elegant is not so vain, because it amounts to showing that a great number of people work for you: it amounts to showing that one has a groom, a tailor, a perfumer, etc. Th e conclusion is: being elegant amounts to showing one’s power, it amounts to showing the importance one has within society and such is the reason why, according to Pascal, being elegant is not vain. It is the whole of the beginning which I would like to analyse, that is to say the utterance “To be elegant is not so vain”. I will claim that there are at least three enunciators in that utterance: E1, E2 and E3. Two of those enunciators are easy for us to identify, given the polyphonic analysis of negation that I have put forward previously. Th e utterance is a negative one. So, if one accepts my polyphonic analysis, there must be at least two enunciators: according to E2 “To be elegant is vain”; and then, there is E3 who disagrees with E2. Th at is something that springs directly from my polyphonic anal-  Slovenian Lectures ysis. Pascal makes someone speak, who criticizes elegance, who disagrees with E2. When I say makes someone speak, I am using a metaphor of course: enunciators do not actually speak. Enunciators do not have a mouth to speak with, they are merely points of view. One should rather say that Pascal makes enunciator E2, according to whom elegance is a vanity, think. Now, I claim that there is another enunciator whom I will call E1, an enunciator according to whom one must be elegant. Th at means that one would have a sort of dialogue with three turns: “One must be elegant”; “No, elegance is vain”; “No, elegance is not vain”. Why do I bring in that E1 enunciator which is, apparently, quite an arbitrary thing to do? I will say that enunciator E1 is contained in the word elegant itself. It is not easy to defi ne being elegant. If you wanted to describe that predicate in descriptive terms, you would be at pains to do so. I will say that being elegant is being well-dressed and the word well is crucial in my analysis, because in the very notion of elegance there is the notion of a way of dressing which is good, a point which, precisely, my paraphrase brings out. One could say the same thing in another way by saying that in the very notion of elegance, as a word of the language-system, there is the idea that elegance is a quality, and that idea is part of the very notion of elegance. So, it seems to me that in the word itself, as an item of the lexicon, there is a sort of justifi cation of elegance, a justifi cation which is like a fragment of discourse written into the word elegant: the word elegance in itself comprises a justifi cation of elegance. About an example studied yesterday (the example of parents telling their children not to touch a dog because the dog is dirty), I said that the word dirty in itself contained a criticism of dirtiness and that one could not understand the word dirty without introducing a sort of discourse according to which dirtiness must be kept away from. Similarly, but inversely, elegance is a way of dressing which is good. So, in Pascal’s text, there is enunciator E1 who picks up that element contained in the lexicon. I will say, if you like, that he is a lexical enunciator. Whereas the enunciators introduced by not are enunciators whom one could call syntactical, E1 is an enunciator incorporated in the lexicon of the language-system itself through whom viewing elegance as a quality is imposed. How is one now to justify that analysis? I will put forward two justifi cations. Th e fi rst justifi cation is that the analysis is wholly in conformi-ty with Pascal’s political theories, and especially those which are expound-ed immediately aft er that remark in Pascal’s text itself. For Pascal (let me remind you of his political theory), there are three possible attitudes towards the organisation of society and generally towards power. Th ere is fi rstly the Lecture II  attitude of the people. Th e people believe that the organisation of society is fair. Th e people, in Pascal’s view, believe in the justice of society: they believe that the king is a good, intelligent, handsome man endowed with all imaginable qualities; they believe that the laws are fair, that the members of the aristocracy are those who have the highest value in society, etc. Believ-ing that the organisation of society is fair, the people comply with the organisation of society because of the fairness which they credit it with: the people obey the laws because they believe the laws are fair. Th at is the fi rst possible attitude. Second attitude: the attitude of people Pascal calls the semi-wise. Th e semi-wise are people who have noticed that the organisa- tion of society is unfair, more exactly that it has nothing to do with fairness and is simply based on force: the people who dominate others are merely the strongest, and there is no legitimacy whatsoever to their domination over others. More generally, no institution has any legitimacy. So, the semi-wise have observed that the organisation of society has no justice in it, and from this observation, they draw the conclusion that since the organisation of society has no justice in it, since it is based merely on force, one must not submit to the organisation of society, one must avoid obeying it as much as possible: the semi-wise are revolutionaries; or at least, potential revolutionaries. Th en, there is a third possible attitude, which according to Pas- cal is the good one. It is the attitude of the wise, therefore of Pascal himself, which consists in this: the wise, like the semi-wise, have observed that the institution of society is wholly foreign to justice and is based on mere force. So, as far as the analysis of social reality is concerned, the wise wholly agree with the semi-wise and wholly disagree with the people. But the wise do not draw from that unjust character of the organisation of society the same conclusions as the semi-wise do: the wise, on the contrary, conclude that one must respect the organisation of society. Not respect it intellectually but respect it in fact: one must submit to, comply with the organisation of society. Th eir argument is the following. Man, by na ture, is unfair: man has neither a knowledge of justice nor, when he does have some idea of what justice might be, does he have any tendency in him to follow that idea of justice. For that reason, it is an excel lent thing that the organisation of justice should be unfair: if the organisation of justice were fair, men, being as they are, would constantly be fi ghting against it. An unfair society is appropriate to the unfair. For example, Pascal in a well-known text says the following: it is a very good thing that the king should be chosen at random, simply on the ground that he is the precedent king’s eldest son, because that greatly simplifi es the succession of kings. If one had to elect the fairest, the  Slovenian Lectures most reasonable, the most intelligent king, seeing that men are fundamentally unfair, they would never reach an agreement as to whom to elect. It is therefore wholly desirable to have a completely absurd society, consisting in naming the precedent king’s eldest son king: one thus knows who is going to reign, immediately, without strife, without civil war. Having observed the injustice of society, the wise conclude therefore that one must submit to society, because an unfair society is perfectly in accordance with what human nature is fundamentally: it is especially desirable that society should be governed by force, because we have a natural tendency towards submitting to force and a society which is ruled by force has a greater chance of being peaceful whereas a society claiming to be ruled by justice would be a society constantly dominated by anti-establishment activity. Anti-establishment in his opinions, Pascal was not at all so in practice: he was as conservative in fact as he was a revolutionary in his ideas. I can now label my three enunciators E1, E2 and E3 as the people, the semi-wise and the wise, for I think that enunciator E1, according to whom one must be elegant, is the people. Th e people consider elegance as a value, the eff ect of which is that it respects people who are indeed elegant, for it sees a proof of their superiority in their elegance. Now the semi-wise come in: they have noticed that elegance proves rigorously nothing – one can be elegant and perfectly stupid – and they conclude that one must dress in any old way, that one must make fun of elegant people, etc. Finally, we have the wise, that I identify with enunciator E3, and so with Pascal. Th e wise agree with the semi-wise on one point, in so much as elegance has a certain vanity in it, but from that vanity of elegance, the wise do not draw the same conclusion as the semi-wise do: the wise draw the conclusion that one must try to be elegant, at least to the degree that your social situation imposes upon you, in order to make the position you hold in society visibly clear. For there to be peace and order in society, one must be able to guess what someone’s social situation is on seeing him: one way of showing a superior social situation is to dress in a luxurious way, because that shows wealth. So, it seems to me that my analysis in terms of three enunciators enables one to relate Pascal’s text with his political theories in general. More precisely now, that polyphonic description makes one understand the title of the fragment, which could seem quite astonishing, if one had not carried out that polyphonic analysis. Th e general title is Sound Opin- ions of the People. Suppose that you do not have enunciator E1: well then, one no longer sees those sound opinions of the people appear. Th e idea that elegance is vain is an opinion of the semi-wise, it is not an opinion of the Lecture II  people. Besides, the attitude of the wise, which gives a value to elegance, but a wholly, how shall I say, indirect value, in so much as it is a manifestation of force – well, of course, that is not an idea of the people either. E1 is the only enunciator who can represent the opinion of the people. My analysis of Pascal’s sentence therefore shows how useful polyphonic analysis for the understanding of texts can be. It also enables one to introduce a notion which is to play an important role in what I will be saying next: the notion of lexical enunciator, since I put a point of view, a judgment within the word elegant itself. I do not think one can understand even the meaning of the word elegant without representing elegance as a quality to oneself. Th ere is no other way of understanding elegance than to introduce the notion of quality in it, or more precisely of goodness. Th us the function of lexical enunci- ator E1 corresponds to the idea that as their intrinsic meaning, words of the language-system contain a certain number of discourse fr agments, which, using a term taken from Aristotle, I will later call topoi. To develop that point will be the purpose of the next lecture. * * * I will conclude now by talking about the bearing this theory of polyphony has on what I said last time. Last time, you remember I said that I would try to chase the notions of true and false from linguistic semantics. To a certain degree, I have progressed in that direction with my theory of polyphony. But to a certain degree too, I am very far from having reached the point I had assigned myself. Why have I progressed in that direction? If one accepts the theory of polyphony, one must say that every utterance is a sort of small play, a sort of mini-dialogue. Now, there is not much sense in judging an utterance in terms of true and false. A dialogue is neither true nor false: a play is neither true nor false; it represents a certain number of opinions or positions, and one cannot ascribe the qualities true or false to the whole of a play. So, by bringing in polyphony, I have contributed to show that the notion of truth or falsehood cannot be applied to utterances taken as a whole. Th at is a progress towards the objective I have assigned to myself. But at the same time, I am still very far from that objective – for the following reason. If it is true that, according to polyphonic analysis, utterances themselves consist in dialogues that cannot be judged in terms of true and false, the fact remains that the diff erent enunciators that I have introduced within the semantic structure of utterances do represent points of view that one could perhaps describe in terms of truth and falsehood. If you want, my  Slovenian Lectures theory of polyphony, if it does chase truth-conditional considerations out from the global analysis of utterances does not chase them out from the detailed analysis of utterances. It may still be the case that the points of view of my diff erent enunciators can be described with true and false conditions. In other words, one could at this stage make the same criticism to me as I made to Searle. Concerning Searle, I maintained that by distinguishing the illocutionary force and the propositional content, by saying that the meaning of an utterance was the ascription of an illocutionary force to a propositional content, he prevented judging the utterance as a whole in terms of truth and falsehood: since the utterance is an act, it is, on that count, neither true nor false. Th at, I put to Searle’s credit but I also addressed Searle a reproach. I said: in the analysis that he gives of meaning, he maintains an area, that is the propositional content, which is defi ned in terms of truth and falsehood. Of course, he refuses to apply those notions of truth and falsehood to the whole of meaning but he does maintain a sort of little cubby hole, a little compartment where truth and falsehood are still relevant. One could make the same reproach to me: one could say “your theory of polyphony shows that an utterance, taken as a whole, cannot be given a truth-conditional description but the diff erent elements that you put within the meaning of utterances, well perhaps they could be described in terms of truth and falsehood”. It is that possibility which I will be discussing in the next three lectures when presenting my theory of argumentation. I will try to show that as elementary as they may be, the points of view of the enunciators cannot be described in truth-conditional terms.  Lecture III December 11 Let me begin by reminding you of where we have got to so far in this series of lectures. As I told you the fi rst time, my general objective is to construct a conception of linguistic meaning which relies neither on the notion of truth or falsehood nor on the notion of information, or at least which relies on those notions as little as possible. For me, the meaning of an utterance – at least deep down – is not the information which that utterance provides about the outer world; it is not the description which that utterance gives of reality. Last time, to eliminate that usual conception of meaning, according to which meaning and information, or truth-conditions, are one and the same thing, I introduced the theory of polyphony. Th is theory leads one to think that each one of our utterances represents a multiplicity of points of view, some of which can diff er from the locutor’s. For example, in a negative utterance such as »Peter will not come«, we have one point of view according to which Peter will come and another point of view, which disagrees with the fi rst. You remember that I called the origin of those points of view enunciators. Th at led me on to say that in the utterance “Peter will not come”, there are two enunciators, that I labelled E1 and E2: E1 has a point of view according to which Peter will come and E2 disagrees with E1. If one accepts that, one sees the notion of information or of description disappear. Indeed, if an utterance is a sort of mini-dialogue between two enunciators, it becomes extremely diffi cult to assess it in terms of truth and of falsehood: a dialogue in itself is neither true nor false, and one fails to see how one could apply the notion of truth to the utterance if, as I have suggested it be, it is described polyphonically, as a dialogue, a confrontation of diff erent enunciators. Th at was where we had got  Slovenian Lectures to last time. Now, I am still quite far from having completed the task I have set myself (as a matter of fact, fortunately so, as I have another three lectures to fi nd material for!). I have only shown that an utterance could not be described in terms of truth and of falsehood but I have not shown that the diff erent enunciators’ points of view could not be described in terms of truth and falsehood. I have forced out those notions of truth and falsehood from the global meaning of an utterance but they might remain valid as far as the enunciator’s points of view are concerned. Th e task I must now accomplish consists in forcing the notion of truth and of information out from not only the global meaning but also the enunciator’s points of view. I will have to show you a way of describing those points of view which does not bring in the notion of truth or of information. Th at is what I am going to do, or try to do, with the theory of argumentation, about which I am going to speak in the last three lectures. Today, I will give you a general picture of this theory, two aspects of which I will develop in the two remaining lectures. I have built up this theory of argumentation with several collaborators: the one I have most worked with is Jean-Claude Anscombre, with whom I have written several books and articles, and we have called our theory the theory of argumentation in the language-system, in short the TAL. I will be introducing the general idea of the TAL now. To do so, I will have to impose a few defi nitions upon you fi rst, which I apologize for. Th en, I will try to illustrate those defi nitions with as many examples as possible. But I cannot avoid starting with a few austere and repelling defi nitions. * * * Th e general thesis of the TAL is that the argumentative function of a discourse segment is at least partly determined by its linguistic structure, whether the segment in question is an utterance or a segment of an utterance. When I speak of linguistic structure, I mean the sentence, which in a previous lecture I have defi ned as being a linguistic entity. Th e argumenta- tive function of a discourse segment is at least partly determined by its linguistic structure, and irrespective of the information which that segment conveys about the outer world. So, the argumentative function or value (I have not yet said what that was but I shall do so in a moment) is determined by the language-system itself. Th at is why we have called our theory the the- ory of argumentation IN the language-system. [In Saussure’s terms], for us, the argumentative value of parole has its origin in langue: the language-system determines what speech is argumentatively used for. Now, I must de-Lecture III  fi ne the notion of argumentative function. Th e argumentative function of discourse segments consists in their representing enunciators whose points of view have an argumentative orientation. What I mean by argumentative orientation must now be defi ned. Th e point of view of an enunciator E’s having argumentative orientation means that it is represented as being able to justify a certain conclusion, or to make that conclusion acceptable. Th e argumentative function of a segment consists in representing enunciators whose points of view are argumentatively oriented. My general thesis then is that the representation of argumentatively oriented points of view is determined by the linguistic structure of discourse segments, – irrespective of the information which those segments provide. I will now explain my defi - nitions: to do that, I am going to take a certain number of examples. I think the clearest case, the most obvious discourse segment having an argumentative function is a segment which is given as an argument for a certain conclusion. It is the case where in a piece of discourse, one fi nds a string of the following type: A, so C. A could be “Th e weather’s beauti- ful”, for example, and C, “Let’s go for a walk!” One could also have said: C, since or because A: “Let’s go for a walk” (C), since “Th e weather’s beauti- ful” (A). When in discourse, you have a string of that kind, it is clear that A, given as an argument aimed at getting C accepted, does have an argumentative function, that is to say that A represents an enunciator whose point of view is represented as leading on to the conclusion. In my examples, A represents an enunciator, who ascribes the quality beautiful to the weather and who considers that quality the weather has as a good reason to go for a walk. I have deliberately taken a very simple example, in which it is diffi - cult to see any other enunciators than the one ascribing the quality beautiful to the weather and representing that satisfactory quality of the weather as a good reason for going for a walk. Now, an argumentative function can be attributed to discourse segments which do not state the targeted conclusion, that is, in which the conclusion is completely implicit. I go back to a previous example. Someone proposes to walk somewhere with me, for example to walk back to my hotel, and I simply answer: “It’s far away”. Even if I have not made the conclusion explicit, the only possible way to understand my answer “It’s far away” is to understand it as being oriented towards the conclusion: “I don’t want to walk there, the distance is too great for the means of going there which you propose to be an acceptable one”. So, even if it is not followed by a So something, my “It’s far away” is intrinsically oriented towards the conclusion “Let’s not walk there”. Take another example in which the conclusion is not  Slovenian Lectures explicit. Let us suppose that someone proposes to go and see a fi lm, and that you answer: “I’ve already seen that fi lm”. Or again, someone suggests seeing some monument, and you say: “Oh, I’ve already visited that monument”. Or again, someone suggests going for a walk, and you answer: “I’ve already gone on that walk”. I think that in those three cases, everyone will understand the answer as being oriented towards a refusal of the suggestion made to you. “I’ve already seen that fi lm” means “No, it’s not worth our seeing it together”, “I’ve already gone on that walk”, “so I don’t want to go again”. Notice, by the way, how strange that is: it is rather strange, because the fact that one has already gone on a particular walk does not prevent one from going again. Personally, I take enormous pleasure in walking around Paris and all the walks I go for are walks I have gone on a hundred times. However, it so happens that the expression “I’ve already been on that walk” is immediately understood as a reason for not going on that walk again. I would rather like to know why. I hope I can give you an idea of why that is so by the end of the last lecture but I wanted to make you feel that there is a problem here. So, sometimes, we have A segments, which are argumentatively orient-ed, even if they are not connected to a conclusion. I now take another example of a segment which, though the conclusion is not apparent, is nevertheless argumentatively oriented. Let us suppose that you have string of the following type: A but B. Someone suggests going for a walk (I am taking the same example), and you answer: “Th e weather’s beautiful but I’m tired”. Everyone, I think, will understand that your answer “Th e weather’s beau- tiful but I’m tired” is a way of refusing the suggestion of going for a walk. You cannot say: “All right, let’s go for that walk, the weather’s beautiful but I’m tired”. You will have to say: “No, let’s not go for that walk, it’s true the weather’s beautiful but I’m tired”. To describe that situation, I say that when the two segments, A and B, are connected by but, they are each oriented towards opposite conclusions. In my example, A is oriented towards the conclusion Let’s (“Th e weather’s beautiful” is viewed as a reason for going for a walk) and segment B (“I’m tired”) is viewed as a reason for not going for a walk, that is to say it leads towards the opposite conclusion Let’s not. It seems to me that if one wants to give a general description of a conjunction like but, one must say (personally, I can see no other solution) that the function of this conjunction is to represent the two segments it connects as being oriented towards opposite conclusions. Th e very fact that you put a but between “Th e weather’s beautiful” and “I’m tired” leads you to read into the “Th e weather’s beautiful” segment a point of view favorable to the walk and into “I’m tired” another point of view, oriented against the walk. Lecture III  To show that this argumentative value given to but is specifi c and peculiar, I am going to try to imagine other possible ways of understanding the connective. A certain number of my American colleagues say that one need only bring in a notion which for ten or so years has had a lot of success in the United States: the notion of implicature. Th at notion is claimed to make argumentation theory unnecessary. To begin, let me say a word on the notion of implicature. [...] I think I used the notion [or its French equivalent] at a time when it was not yet fashionable. Well then, what is an implicature? An implicature is information necessary for a given utterance to be in conformity with the general rules governing discourse behaviour. We all know that one cannot say any thing at any time: there are a certain number of norms or maxims which govern or command discourse. I am not going to say, all of a sudden, “Th e weather’s beautiful” (although I have just used that sentence a good ten or so times as a linguistic example). I am not going to interrupt my lecture to say: “Th e weather’s beautiful”. Indeed, there is a rule according to which, when one says something, that thing must be, as one oft en says, relevant, that is to say must have some bearing on the utterance-situation. So I cannot say to you “Th e weather’s beautiful” with- out implying that given the utterance-situation, I have reasons to say so to you. One of the possible reasons for saying “Th e weather’s beautiful” could be the intention of suggesting that we go for a walk (but one can imagine many others, so long as they make the utterances conform with the norm of discourse according to which our utterances must be relevant to the situation). When you have to interpret my hypothetical “Th e weather’s beau- tiful”, you will always imagine some reason of that kind, for example that I wanted to suggest going for a walk to you, intention which thus becomes an implicature from my utterance. For the American linguists I am alluding to, what I call argumentativity is simply implicature: in at least a certain number of circumstances, the utterance “Th e weather’s beautiful” implies that one would like to go for a walk because otherwise it would not be relevant to utter it. According to the objection made against me, the but connective serves, as they say, to defeat the implicature. I told you “Th e weather’s beautiful”. You have concluded that I wanted to go for a walk, because otherwise my utterance would not be relevant and would not conform to discourse norms. In adding “but I’m tired”, I am simply trying to defeat that predictable implicature. I would like to show you now that this description which aims at de-priving the notion of argumentation of its usefulness is wholly insuffi cient in a great number of cases. In the example I took, it is just about suffi cient  Slovenian Lectures but there are a good many cases where it is not at all operative and where the but absolutely does not behave as an implicature-defeater. On that point, I will simply take two examples. First example: “Peter doesn’t know all wines but he does know some”, which is a wholly natural use of but. Let us try to analyse that stretch of discourse with the notion of implicature. Th e classi-cal view is that “He doesn’t know all wines” implies “He knows some”. Indeed, it would be absurd and contrary to the rules of conversation to say that someone does not know all wines if thereby one did not imply that he does not know some. When I say “I don’t know all French wines”, you immediately conclude that I know some. If I did not know any, I would be misleading you in saying that I did not know all of them. But then you do see that in my example but absolutely does not defeat the implicature: quite the contrary, it confi rms it. Th e implicature was “Peter knows some wines”, then one has a but and one adds “He knows some”. So the description of but as an implicature-defeater is visibly inadequate, at least in that particular case, even it is just about operative in other cases, like the fi rst I envisaged. Personally, I will not use the notion of implicature to describe but (which does not mean that I believe the notion of implicature is useless: I simply think that it does not help to describe but). I think that the function of but is to bring out the argumentative potential of discourse segments. In my personal description, I will say therefore that the utterance “Peter does not know all wines” is oriented towards a certain conclusion, like Peter’s ignorance as to wines (if you like, in saying “Peter does not know all wines”, I am criticizing, or at least devaluating whatever knowledge Peter may have of wines). By contrast, “He knows some” is oriented towards the opposite conclusion: it is a way of bringing out the knowledge of the person I am speaking about. So, for me, in putting in but between “Peter doesn’t know all wines” and “He knows some”, one is making the argumentative function of “Peter doesn’t know all wines” clear or, in other words, one is making it clear that this discourse segment represents an enunciator who justifi es an unfavourable appreciation of Peter’s knowledge as to wines. I take a second and last example. Perhaps I am overstressing this but for me, it is extremely important to bring out what is original about argumentation. Take sentences like “Peter would have liked to come but he didn’t” or “Peter would have liked to come but he wasn’t able to”. I have used the conditional mood deliberately. Th e use of that mood clearly implies that Peter did not come. If I thought that Peter had come, I would have said “Peter had a great desire to come” but I would not have used the conditional. So, I have an implicature “Peter did not come” (the Americans, in that Lecture III  case, would speak of a conventional implicature). But that implicature is not all defeated by what follows but, that is “He was not able to”. On the contrary, the implicature “Peter did not come” is confi rmed, or at least is explained by what follows but. So, here again, one cannot say that but functions as an implicature-defeater: once again, the notion of implicature does not seem relevant to describe the function of but. For my part, I would describe the function of but in that stretch of discourse by saying that “Peter would have liked to come” represents an enunciator indicating the desire Peter had of coming, an enunciator who gives an argument for Peter’s coming. Of course, in saying “Peter would have liked to come”, in a sense, I am saying that Peter did not come but at the same time, I am indicating the desire Peter had of coming and that desire, like any desire, is an argument for its own fulfi lment. So, I will say that the segment “Peter would have liked to come” is argumentative: it has an argumentative function, in the sense that it represents an enunciator whose point of view justifi es the idea that Peter was to come; aft er that, but introduces a segment that is oriented towards the opposite conclusion, Peter’s not coming. It seems to me therefore that but for the same reason as so can bring out the argumentative function of discourse segments, that is to say the argumentative orientation of the points of view which a stretch of discourse represents. I do not think it can be described otherwise. Now, let me give you the example of another word, about which we will have a lot to say later: the word even. Let us suppose that I am telling you about a meeting I was at, and that I say to you “Th ere was Peter, and even John”. What does even do, which links the idea that there was Peter and the idea that there was John? For my part, I describe the function of even with the help of the notion of argumentation and I cannot describe it otherwise. I say that when one links, when one connects two discourse segments with even, one represents the two segments as being oriented towards a same conclusion, that I arbitrarily call R, the second segment being a more forceful argument than the fi rst relatively to that conclusion. If you want to understand my stretch of discourse “Th ere was Peter and even John”, you must ask yourself “Why did he say even?” To answer that question, the only solution (at least, I cannot see any other) seems to be to ask yourself what I have tried to show, that is to say, what the conclusion that I have tried to justify is. To understand my even, you must fi nd a conclusion which is justifi ed by “Th ere was Peter” and which is also justifi ed and, if I may say so, even more justifi ed by “Th ere was John”. Th at conclusion is not specifi ed in the dis- course. It could be “We had fun: there was Peter, who is very amusing and  Slovenian Lectures even, there was John, who is even more amusing”.It could mean “We got very bored: there was Peter, who is very boring, and there was even John who is worse than boring”. It could mean “Th e meeting was a great social event: there was Peter, who is the Prime Minister, and even, there was John, who is the President of the Republic”. It could mean “Th e meeting was a great failure: there was Peter, who has really no social status and there was John, who has even less of a social status”. You can imagine what you like. But once you see an even between two segments, you must say to yourself that both segments are oriented towards a common conclusion. Or, in my terminology, each of those segments represents an enunciator whose point of view is oriented towards a certain conclusion. Aft er having tried to explain what I mean by argumentative orientation, I would like to make what I have just been saying clearer, and then aft er a little pause, I will try to justify the general thesis that I formulated at the beginning of this lecture (for the moment, it is not at all justifi ed). I would like you on the one hand to notice the expression represented, which I used in my defi nition of argumentative points of view. Let me read my defi nition over to you again: to say that an enunciator’s point of view is argumentatively oriented is to say that it is represented as being able to justify a certain conclusion. I must insist upon the reasons that have led me to say represented as, and not simply justify a certain conclusion. Imagine a short stretch of discourse, in which one is speaking about someone whom I will once again call Peter: “Peter is stupid, so he’ll be successful”. It does not seem at all stupid to me to say things of that sort. Stupidity, in a great number of cases, is wholly an asset for at least a certain type of success. Let us suppose that I say to you “Peter is stupid, so he’ll be successful”. Given the presence of so, one must (if the preceding analyses are correct) accept that stupidity is oriented towards a conclusion of the success type. But that does not at all force the linguist, as such, to think that stupidity, in general, is a cause for success. It only means that in a particular discourse, stupidity is represented as a cause for success, as a factor of success. Th e important thing is what is rep- resented in a piece of discourse as justifying a conclusion, not what eff ectively does justify it. Second and last clarifi cation. To say that a segment A has an argumentative function is quite diff erent from claiming (this point is for me absolutely essential) that A is an argument in the logical sense of the word, that is to say a premise for a conclusion in a piece of reasoning. Remember my example “Peter would have liked to come but he couldn’t”. I said that the fi rst segment was argumentatively oriented towards Peter’s coming. But it Lecture III  is quite obvious that one is not going to use it, in a piece of reasoning, to prove that Peter did come. My notion of argumentativity is linguistic: it does not correspond to logical inference. * * * In the fi rst part of this lecture, I tried to explain to you what I meant by the argumentative function of a discourse-segment, that is to say, by the argumentative orientation of a point of view. Now, I must try to prove my thesis, according to which the argumentative function of a discourse-segment is at least partly determined (I would like to be able to say totally but I am not confi dent enough to say that) by its linguistic structure, that is to say, by the sentence which that segment is an occurrence of, and that irrespective of the information which that segment provides. To show this, I am going to give examples, which will all be of just about the same nature: I am going to be comparing two sentences which provide the same information, or more precisely, the utterances of which provide the same information but which nevertheless have wholly diff erent argumentative functions. So, I am going to indicate a certain number of pairs of sentences, in which there is no change as to the informative value but a very important change as to the argumentative value. To begin, I am going to compare “It’s eight” and “It’s only eight”. Let us suppose that someone starts a stretch of discourse saying “It’s eight”. One can very well imagine his going on with a conclusion of the following type: “It’s early, don’t hurry, take your time”. But one can also imagine that aft er having said “It’s eight”, the same locutor should go on with: “It’s late, hurry, there’s not a minute to waste”. So, two types of follow-ups are possible after “It’s eight”. But notice now what happens aft er “It’s only eight”. I think it is impossible to follow up with: “It’s late, hurry”. Th e string “It’s only eight, it’s late, hurry!” makes no sense. What would be possible on the other hand would be “It’s only eight BUT it’s nevertheless late, you must hurry”. What there would be no diffi culty in understanding is “It’s only eight, so it’s ear- ly. Don’t hurry!” whereas it would be absurd to replace so by but. Th e argu- mentative function of the utterance of those two sentences is thus extremely diff erent: “It’s eight” can be oriented either towards earliness or towards lateness; “It’s only eight” can be oriented only towards earliness and not at all towards lateness. Now, I think you will admit without diffi culty that the information provided by those two utterances is exactly the same. I am giving exactly the same piece of information when I say “It’s eight” as when I say “It’s only  Slovenian Lectures eight”. You cannot answer: “It’s eight, that’s true but it’s false that it’s only eight”. If you fi nd one true, you fi nd the other true, if you fi nd one false, you fi nd the other false. So, we have here an example of utterances which provide the same information but which do not have, which cannot have the same follow-ups at all. In other words, the point of view about the time is oriented diff erently in “It’s eight” and in “It’s only eight”. Th at does not at all depend upon the information given: it depends simply on the linguistic structure, that is to say, the sentence of which the segment is an occurrence of. It is the word only which forces you to choose a follow-up of the “It’s early” type, and makes a follow-up of the “It’s late” type impossible. To show you how diffi cult the problem is, compare “Peter will be ar- riving at eight” and “Peter will be arriving only at eight”. I am not going to resolve the problem but I am simply pointing it out to you. “Peter will be arriving at eight” has two possible follow-ups to it: “Peter will be arriving at eight, that’s really early”; “Peter will be arriving at eight, that’s really too late” – Peter’s arrival can be represented in whichever way, as late or as an event that is to happen early. Now, let us look at “Peter will be coming ONLY at eight”. Th at means “He will not be arriving before eight, he will not be arriving any earlier than eight”. Th e utterance “Peter will be arriv- ing only at eight” is necessarily oriented towards lateness; it cannot be oriented towards earliness. You can see the linguistic problem involved here (it would take several hours to try to resolve it): “It’s only eight”, as we have seen, is oriented towards earliness but “He’ll be arriving only at eight”, on the contrary, is oriented towards lateness! So, only, from the argumentative point of view, has two diff erent eff ects, depending on whether it modifi es the time it is (“It’s only eight”) or the time at which an event happens! Th at is the type of problem one has to deal with when one is in argumentative linguistics, and it is by no means an easy one. I have, for my part, tried to fi nd reasons for which only has those opposite eff ects, but all I wanted to point out was the kind of problem one encounters when one goes into what I call argumentative linguistics. We know that only has two opposite argumentative eff ects. Well then, why? I take another example, but which will still be of the same nature. Th e purpose is still to show that two utterances with the same informative value do not have and cannot have the same argumentative value. Compare “It’s almost eight” on the one hand and “It’s not eight yet” on the other, and look for their possible follow-ups (the argumentative description of an utterance requires looking for its follow-ups, because these bring out the argumentative orientation of the points of view which the utterance repre-Lecture III  sents). “It’s almost eight”, I think, can be oriented only towards conclusions of the “It’s late” type: “It’s almost eight, hurry! It’s almost eight, walk fast-er if you want to be on time!” I could not say: “It’s almost eight, you’ve got plenty of time”. With almost eight, the time is viewed as late. On the contrary, when I say “It’s not eight yet”, it is the opposite type of conclusions which become possible: “It’s not eight yet, you’ve got plenty of time! It’s not yet eight, there’s no need to hurry!” However, you will agree with me that the information given by almost eight and not eight yet is exactly the same: in both cases, it is, I don’t know, say, fi ve to eight, ten to eight. So, once again, we have two utterances which give the same information but which have a wholly diff erent argumentative function and that argumentative function is necessarily linked to the words themselves, the words those utterances are made of, that is to say, to the language-system, to the linguistic structure, quite irrespective of the information given. I take a third example (taking examples is all I am going to be doing until the end of this lecture). I am going to speak to you about the two expressions little and a little. In English, in most Romance languages [like French] and also in some other Germanic languages like German (perhaps also in some other languages that I do not know), those two expressions are, – how shall I say? – “built” in the same way: the idea of a little is obtained by putting an indefi nite article before the word which designates the idea little. In other languages, things are more complicated and the diff erence between the two notions is marked in a more subtle way. We are going to compare the semantic eff ects of little and a little. Let us compare “Peter has worked a little” and “Peter has worked little”. Within an argumentative framework of analysis, to compare these expressions means looking for their possible follow-ups. Aft er having said “Peter has worked little”, it seems wholly reasonable to me to go on and say “He can’t be tired”. On the contrary, aft er “Peter has worked a little”, I would go on in the opposite way: “He must be more or less tired” or even quite simply “He must be tired”. So the expressions go in completely diff erent directions. However, I think one can say (although there is much debate on this point) that the information given by “Peter has worked a little” and “Peter has worked little” is nearly identical. To show that the information given by those two segments “Peter has worked a little” and “Peter has worked little” are roughly identical, I would like to make you notice that one can very well say “Peter has worked LITTLE but still, he has done so A LITTLE”. I point out that fact to show, I remind you, that in both cases, there is the same information. Indeed, when  Slovenian Lectures one connects two discourse-segments with but, those two segments must be compatible with one another. I cannot say “Peter has come but he has not come” or “Peter is fair-haired but he is brown-haired” (at least in the ordinary use of language, that is the use which depends on the idea, or the illusion, of informative value). Consequently, if one can say “He has worked LITTLE but still, he has done so A LITTLE” without any diffi culty, it is because little and a little are compatible. To complete my demonstration, I must point out that if those two quantities are compatible, they must be identical. If there were a diff erence between little and a little, the two indications would be incompatible. For example, if little were less than a little, there would be a contradiction in saying “little but a little”, – a contradiction which no-one is aware of ! So there again is a case where two expressions with wholly diff erent and even opposite argumentative values cannot be distinguished from the informational point of view (for argumentative purposes, I am making myself out as accepting the use of that notion of information!). I am now going to take examples from another area (if I have time after that, I shall come back on little and a little) so as to show that what we have here is a wholly general phenomenon in language. First, a commonplace remark, one that has been made for a long time: how is the diff erence between “Th e bottle is half full” and “Th e bottle is half empty” to be described? From the informational point of view, the diff erence is rather hard to establish. If I make a little drawing to depict the bottle which is half empty, it is going to be the same as the drawing depicting the bottle which is half full, is it not? But from the argumentative point of view, the two expressions have fundamentally diff erent values. Th e utterance “Th e bottle is half empty” can be connected with conclusions of the type: “It must be fi lled up” or “Another one must be bought”, in other words, with conclusions which are relevant to the empty or half-empty state of the bottle. On the contrary, if I start by saying “Th e bottle is half full”, the expected con- clusions are of the type: “It’s not worth buying another straight now”, “We can still wait a little”, and so on. Th e conclusions we have here are relevant to the fi lled-up state. Th at is brought out in a spectacular way if you put an interjection like “unfortunately!” before or aft er either sentence. In saying “Th e bottle is half empty. Unfortunately!”, you are complaining about the empty state of the bottle. It is the remark to be expected from a heavy-drinker on his noticing that half his wine is already gone. But in saying “Th e bottle is half full. Unfortunately!”, what you are complaining about is that there should be something left in the bottle. Th at, our heavy-drinker is not Lecture III  going to say. Who will? Most probably his wife, annoyed at the idea that her husband still has half a bottle of wine to drink. In any case, the fi lled-up state of the bottle will be what the argumentative follow-up aft er “Th e bottle is half-full” will be relevant to and aft er “Th e bottle is half-empty”, it will be relevant to its empty state. Th at is so, although the information is exactly the same. So, facts, information, in such cases, do not determine argumentative value at all (even if the existence of such facts is brought into linguistics – which, as I told you at the beginning, is not something I do – they are irrelevant in the examples I have just been taking for argumentative analysis). Now, I am going to take examples exclusively related to the lexicon itself. I remind you of an example that I took, I think, during the fi rst lecture. We had compared far and nearby: “It’s far”, “It’s nearby”. At fi rst, you have the impression that the information given is not the same. But that is an impression which disappears if you look at these words more carefully. You can say either “It’s far” or “It’s nearby”, without a change in the situation and about the same distance which you have no uncertainty about. I am going to take the example I gave a couple of days ago. Someone suggests I walk with him to a certain place. I know what the distance in question is exactly. Let us even suppose that I can say how much it measures: I know it measures, say, a mile. Whatever the situation, I can very well say: “It’s far”. But I can also very well say: “It’s near”. Th e information has therefore no rel- evance to our choice of the words. What then, is the diff erence between the two answers: “It’s far” and “It’s nearby”? Th e diff erence is that the answer “It’s far” will be understood as a refusal of the suggestion which has been made to me whereas the other answer “It’s nearby” will, on the contrary, be understood as an approval, an acceptance. If someone suggests “Let’s walk to your hotel!”, I can answer either “Oh, no. It’s far” or “All right, it’s nearby” but the distance remains the same: simply, in one case, I refuse to walk there and in the other, I accept. Th e diff erence then is not informational: it is purely argumentative. Yet another example. Take these two words: thrift y, avaricious. Is there any informational diff erence between my describing someone as thrift y and my describing him as avaricious? I defy you to say: “It’s true he’s thrift y if such and such conditions obtain”; “It’s true he’s avaricious if such and such other conditions obtain”. It is not at all on the grounds of the information provided that you can distinguish the thrift y from the avaricious, it seems to me. Th e diff erence is in the attitude you adopt towards the person you are speaking about. You can say to someone “He’ll be a good husband,  Slovenian Lectures he’s thrift y”, because the language-system makes out thrift iness to be a good quality (even if it can be considered as a rather dismal one) but it will be far more diffi cult to have the same follow-up with avaricious. You can hardly say “Oh, I rather like him, he’s avaricious” (even if you appreciate the fact that the person in question spends little). You cannot like avarice: what I mean is that you cannot say that you like someone and at the same time call avarice the quality you say you like in him. Th at does not mean that the information is not the same in the two cases. But it is the language-system which imposes one argumentative orientation rather than the other. * * * It is facts like those (of which there are many, many more, I think) which have brought us to say that argumentation is written into the language-system itself, into the most linguistic aspect of the structure of our utterances. As I have fi ve or ten minutes left , I would like to go back now to my example of little and a little and consider an objection which is sometimes made to me. I have told you that in saying “He’s worked little” or “He’s worked a little”, the same information was being given. Some people think that this is not true and that there is an informational diff erence, a diff erence in the facts being described, between “He’s worked little” and “He’s worked a little”. If, for example, one measures work in terms of the time spent working, those people tell me that “He’s worked little” indicates less time spent working than “He’s worked a little”. Let us suppose for example that “He’s worked little” means one hour of work: then “He’s worked a little” will mean, say, two hours. So – this is the objection that is made to me and which I am going to try to answer in a moment – “He’s worked a little” means more work than “He’s worked little”. To justify the idea that the quantity designated by “He’s worked a little” is superior to the quantity designated by “He’s worked little”, someone once carried out the following experiment, which seems conclusive, but which, as I will try to show, in fact, is not. Th e experiment is this. People were asked to imagine the following situation. Take two children, Peter and John: Peter has worked two hours and John has worked one hour. Th e people, the experimental subjects, were asked to describe the two children, one as having worked little, the other as having worked a little: you must say which has worked a little and which has worked little. Th e problem having been set in that way, all those questioned gave the same answer (I presume that answer is also yours, and it would also be mine): if I have to use the expressions worked a little and worked little to describe the two children, I will say about Lecture III  Peter that “he’s worked a little” and about John that “he’s worked little”. Th e members of the audience who had raised the objection therefore concluded: “Well you see, there is a quantitative diff erence between a little and littl e, a little does designate a greater quantity than little”. Now, I am going to try to answer those objections, which rely on facts of that sort. To do so, I shall use my defi nition of little and of a little. I will say that to work little is an argument, say, for the insuffi ciency of the amount of work and that to work a little is an argument for the opposite conclusion, that is to say, for the idea that the amount of work is suffi cient. In a class-room situation (and that was the situation for the experiment) , He has worked little is an argument for the insuffi ciency of the amount of work: for example, for some form of punishment. On the contrary, He has worked a little is indirectly an argument for some sort of reward. So, quite disregarding any fact or information, and simply from my argumentative description of little and a little, you can predict that He has worked little can be used as an argument for punishment and He has worked a little can be used as an argument for reward. So when the question for the experiment is asked (here are two children, Peter has worked two hours and John has worked one hour, about which of the two do you say that he has worked little?), given that to say He has worked little is one reason among others to punish him and to say He has worked a little is a reason to reward him, it is pretty clear that the experimental subjects are going to choose worked little for John, who has worked for an hour, and worked a little for Peter, who has worked for two hours. It is merely a sense of fairness which makes you attribute the little, which is a source of punishment, to John and the a little, which is a source of reward, to Peter. But that stems from cultural considerations related to our notions about justice, and not at all from the meaning of the expressions worked little and worked a little themselves. So, the informational diff erence which you read into those two expressions are really extremely indirect diff erences, which can appear in very particular situations, such as the one which was imagined here, where someone, the subject of the experiment, had to choose between the two expressions. But that by no means changes the fact that in the linguistic structure of little and of a little, there is no informational diff erence to distinguish them by, no information-related diff erence: there is only an argumentation-related diff erence. * * * What am I going to do in the next lectures? Well, today I tried to introduce and justify the general thesis of the theory of argumentation in the  Slovenian Lectures language-system. In the two remaining lectures, I am going to try to show you how argumentation can be understood as being within the language-system. Today, I have shown you that argumentation had to be understood as being within the language-system but the thing now is to discover how this is so. Th at is the problem which I will try to deal with in the follow- ing sessions by bringing in the notion which I am working on the most for the moment: the notion of topos. To do that, tomorrow, we will be studying argumentation under its rhetorical aspect, and, aft er tomorrow, we will be studying it under its linguistic aspect.  Lecture IV December 12 Yesterday, I tried to show that the argumentative function of discourse-segments was written into the language-system, that is to say that the linguistic structure of the sentences occurring in those discourse-segments directly determined their argumentative functions, and that, irrespective of the information those segments conveyed. What now remains to be done is to see how that argumentative function is written into the language-system or which semantic elements of a sentence determine the argumentative function of discourse-segments. To do that, in this and the next lectures, I am going to explain a theory which Jean-Claude Anscombre and I have been developing for a certain number of years: topoi theory. Our use of the Aristotelian term topos de-forms its meaning a little perhaps but I think we are more or less faithful to the idea Aristotle put behind it. Last time, I told you that this argumentative function of discourse-segments could be discovered chiefl y when those discourse-segments were linked to one another in discourse. For example, when a string uses a conjunction like so: “Th e weather’s beautiful, so let’s go for a walk!” Or else, when there is a but, or an even. Th at leads me to try to see what the argumentative function is to be ascribed to, to study argumentative discourse-strings more closely. Th at is what we are going to be doing today. In the lecture tomorrow, I shall try to show how the results obtained up until now by analysing argumentative discourse-elements are as it were prefi gured within the language-system itself. But today we are going to stay at the discourse level, and especially study argumentative strings of the following type: A, so C (A  C), in which a segment A is given as an argu-  Slovenian Lectures ment for conclusion C. Th e fi rst idea that I want to bring out is that when you have an argumentative string of the type argument  conclusion, there is always a reference to a third term which is distinct from both the argument and the conclusion, and which allows you to bridge the gap from one to the other. S.E. Toulmin called that third term, the warrant. When I say A, so C, I am supposing that there is something, a warrant, which allows me to bridge the gap from A to B. Let us take an example which I have already used several times again: “It’s warm, let’s go for a walk!” When you say that, you are supposing that there is a principle which allows you to bridge the gap between the warmth, stated in A, and the suggestion to go for a walk, the conclusion reached in C. Th at principle, which ensures the validity or the legitimacy of the move from A to C is what, using the Aristotelian term, I will call a topos. Let us go back to our example: “It’s warm, let’s go for a walk!” When you say that, you are in certain way presupposing that Warmth makes a walk pleasant. Unless there is this principle, the A  C string would not achieve its conclusive move. What are the characteristics of this topos (or warrant) which is to be found behind argumentative discourse-segments? My claim is that the topos, as encountered in that type of string, has three characteristics: fi rst, it is general; second, it is represented as a shared belief, that is a belief which a certain group of people already accept; and third, it is scalar. * * * Firstly, the topos is general. What I mean is that the topos is represented as being also valid in situations other than the one which the current discourse is about. Th e topos which allows the move from “It’s warm” to “Let’s go for a walk” is not the Th e warmth today will make the walk pleasant to- day but something far more general: Warmth IN GENERAL makes a walk pleasant IN GENERAL, a principle which also goes for an infi nite number of situations other than the one which a particular piece of discourse may be about. I will formulate this general character of the topos in the following way: I will say that the topos relates two properties: a fi rst property P, connected with the argument, A, and a second property Q, connected with the conclusion, C. In my example, property P is the general property of warmth and property Q, connected with the conclusion C, could be called pleasantness of a walk. In saying “It’s warm, let’s go for a walk”, you are supposing that in general those two properties, warmth and pleasantness of a walk, are connected with one another in a certain way which will be defi ned more precisely later. Lecture IV  I would like you to notice that I have characterized the topos as being general: I have not said that it was universal. Th at is a crucial point for me. To say that it is universal would be tantamount to saying that it allows no exception whatsoever. Now, that is not at all what we are supposing in the utterance “It’s warm, let’s go for a walk”. When we say that, we do admit that there might be exceptions but that does not prevent the topos from being valid, which is the point this highly famous formula attributed to Aristotle makes: “exceptions make it possible to uphold the rule in unforeseen cases”. Th at is to say, in cases which the rule does not foresee: in such cas-es, the notion of exception makes it possible to uphold the validity of the rule nevertheless. Let us suppose that we are having a walk one day and on that day, it is warm but that the walk is very unpleasant: I can say it is an exception to the rule but that exception does not make the rule void. For Aristotle, the possibility of there being exceptions to general rules is of prime importance. Besides, it is a point connected with his general principles. I would like to say a few words about this. You know that, for Aristotle, the world is divided in two zones: a zone which is above the moon and where the stars are in motion and another zone which is below the moon, the sub-lunary world, where we, both you and I, have the ill-fortune of residing. What is the essential diff erence between those two worlds? Well, it is that in the world of stars, there are universal rules, that allow of no exceptions whatsoever whereas in our world, admittedly, there are rules also but those rules always have exceptions. In most languages, there is a word to mark that one is faced with a case which, relatively to the rule, is an exception, the rule being nevertheless upheld as valid: it is the word yet or nevertheless. Th us I could say to you: “It was warm, yet it was an unpleasant walk”. In using a yet to join the segments “It was warm” and “It was an unpleasant walk”, I am upholding that there is a rule, connecting warmth and the pleasantness of a walk but that, unfortunately, we were faced with an exception to that rule because of extraordinary factors. So, when I say that the topos is general, I do not at all mean that it is universal; it is essential, on the contrary, for it to have exceptions. How is the general character of the topos to be proved? We only need to consider the refutations of an argument: because oft en those refutations take into account the generality of the topos. Let us suppose that I am still making that suggestion for a walk and still on the grounds of the warmth argument. You can object: “It was also warm yesterday and yet it was an unpleasant walk”. Th at is say, you are pointing out that there are exceptions to the rule which I have used and in saying that, you are suggesting that  Slovenian Lectures perhaps I have no right to use the rule for the particular case which occa-sioned my discourse. In pointing out that there are exceptions, you recognize that the rule which I have used is a general rule but at the same time, you are telling me that I do not have the right to use that rule in my particular situation. You do not deny the generality of the rule at all, you are simply showing that there are exceptions to it and you are suggesting that we may be in one of those exceptional cases (which are foreseen by the rule itself, since the rule foresees the possibility of exceptions). Or again, I say “Peter is wealthy, he must be happy” on the grounds of a very commonplace topos according to which wealth is a factor of happiness. You will answer “But I know a certain number of people who are both rich and very unhappy”. In saying that, you are pointing out that there are exceptions to my rule, you are insisting upon those exceptions: it is therefore not certain, according to you, that I have the right to apply it in the particular situation which we are speaking about. Concerning that point, it is interesting to note that the purpose of a great number of proverbs in many of our societies is to point out that there are exceptions to rules. Th at does not amount to denying those rules. All it means is that they must be used with caution, because there are cases in which they do not apply. Th at is what a lot of English proverbs do. Th ey have the same form in general. Here are a few examples. I took the following example of argumentation just now: “Peter is wealthy, he must be happy”. To refute it you can use a proverb, which works as an anti- topos, and it is this: Money can’t buy happiness. Th at means that there are exceptions to the rule according to which when you are rich, you are happy. Th at does not prevent your admitting that in wealth, there is an intimation of happiness; but that intimation can sometimes be deceptive. Or again, you have this proverb which, I think, is just about a universal one, at least in Western society: All that glitters is not gold. Th at does not mean that everything that glitters is valueless. It means that the rule we normally use, and which from a glittering appearance concludes to real quality, has a certain number of exceptions, and that care must be taken in our use of it. Take yet another proverb like His bark is worse than his bite. Th at does not mean that when a dog barks, you need not worry and can be sure it will not bite you, – which would be contrary to experience – and proverbs do not run against experience. All that means is that there are exceptions to the rule according to which a dog that barks is about to bite, so that you need not worry too much if you do see a dog barking, because the situation is perhaps one in which the dog barks but does not have the intention of biting. Lecture IV  Th at is what I wanted to say concerning the general character of a topos. It is a connection between two properties, P and Q, related respectively to argument A and to conclusion C. * * * Second characteristic: the topos is represented as a shared belief, a belief that has been accepted beforehand by a community which at least the locutor and usually also the allocutor, or addressee, belong to. Th at is the property which gives the argument its constraining force. If it is reasonable to move conclusively from the idea of warmth to the idea of a walk, it is because the move is based on a rule which the locutor has not invented: it is represented as accepted by a certain community. Not perhaps by all men and women but at least by a small community of reasonable men and women whom the locutor and, he hopes, also the allocutor belong to. Fail-ing that, the so would be completely impossible. A topos then is presented as one of a certain community’s shared beliefs. Th at explains a certain form of irony, of which, in France, there are many examples in Voltaire, and which consists in using topoi which obviously no-one accepts. You reason (really, I should not use that word reason, as I have distinguished reasoning and argumentation), you argue on the grounds of topoi which no-one accepts, or only very few persons do. If someone says “He’s rich, so he’s unhappy”, he thinks that he is being interesting, amusing, because he has made out a certain topos to be obvious which on the whole, society, in fact, does not accept. In Voltaire, you could fi nd sentences like “Mr X’s standard of morality was high: that is why he harmed all those he knew”. Here, you posit the connection between the fact of having a high standard of morality and the fact of doing harm as being a topos; and you cannot do that otherwise than ironically, since you are making out something to be obvious which generally is not considered as such, and even which is contrary to what is obviously the case (which amounts to being paradoxical while giving the impression of speaking like everyone). * * * I move on now to the third characteristic, the one which we will have to speak most about, because it is the most problematical one and, for a linguist, I think, the most important: scalarity. Th ere is not much to object about the generality and the apparent sharedness of a topos I believe. But it is about scalarity that everyone makes objections to me and it is on that  Slovenian Lectures point that I am going to defend myself. When I say that the topos is scalar, I am saying two things. First, properties P and Q themselves are scalar. Th at is to say, that they are properties which you can have more or less of. In the example I took, it is quite obviously so. Th ere the topos, which was the basis of the argument, was the one according to which warmth makes a walk pleasant. Now, obviously, there are diff erent degrees of warmth and there are also degrees of pleasantness. I will formulate that scalarity of a topos by saying that predicates P and Q, which a topos connects, must be considered as scales. Th ere are diff erent degrees of intensity in the possession of characteristic P and in the possession of characteristic Q. Th at does not at all mean (I would like to avoid a misunderstanding on this point) that the arguments and the conclusions are scalar. Th e properties within the topos are scalar but not the propositions used in discourse as arguments or conclusions. I take an example. Consider the following argument: “It’s less than ten degrees, take a coat with you”. Well, there is no doubt that neither A nor C are scalar: it cannot be more or less ten degrees, and you cannot more or less take a coat. So, the indications contained in A and in C are not scalar ones. But that does not prevent the topos, which is the warrant for that string, from being describable in scalar terms. Th e topos here is that when it is cold, you must dress warm: it relates one property P, which is the cold, and another property Q, which is, say, garment warmth. Th e indications contained in dis- course-segments A and C, “It’s less than ten degrees” and “Take a coat with you” represent degrees within those general properties P and Q. Minus ten (-10°C) is a degree of cold: there are lesser degrees of cold and greater degrees of cold. A coat is a type of warm garment, and there are still warmer ones (say, the outfi ts skiers put on) and also less warm ones, for example a jacket. So, when I am speaking about the scalarity of predicates, I am not speaking about the scalarity of A and of C but, I stress, I am speaking about the scalarity of properties P and Q connected to A and C. Such is the fi rst idea contained in my contention that a topos is scalar: the two propositions P and Q are scalar. Now, there is a second idea, an even more diffi cult one to accept, and I will certainly have diffi culties in getting you to accept it. Th e idea is that the relationship which a topos establishes between P and Q is itself scalar. We have seen that P and Q are scales: a topos indicates (I hope to have arguments to justify this thesis) that there is a scalar relationship between the degrees of property P and the degrees of property Q. Th at is to say, that go- ing along the scale of property P in a certain direction also means going Lecture IV  along the scale of property Q in a certain direction: a move up or down one scale means a move up or down the other. I go back to my example “It’s less than ten degrees, take a coat with you!” To say that the topos used here is scalar means that the degree of cold implies a degree in garment warmth. Th at can be formulated (a way of doing so which I keep for later but I am antici-pating here) by saying: the cold-er it is, the warm-er you must dress. * * * Well then, how is one to go about proving that scalarity? I am going to give a certain number of examples, the fi rst of which are of a rhetorical order, that is to say concern discourse-strategies and especially refutative discourse-strategies. Th en, I will take examples which are more directly lin- guistic. I will begin with the rhetorical examples. Let us suppose that I suggest going for a walk, using my sempiternal example: “It’s warm, let’s go for a walk!” Having no desire to go for a walk with me, you are going to refuse my suggestion politely. Th ere are many ways for you to do that: you can say that you are tired, you can say that, as a matter of fact, it is not all that warm. You can also say (and I fi nd this solution interesting because it makes you feel the scalarity of a topos): “Tomorrow, it’ll be even warmer, let’s put off that walk till then” – and the next day, you just forget about it. Let us think about that type of argument a little. If you fi nd that argument a rather clever one, it is because you have recognized the scalarity of the topos I have used and you turn it against me. I used a topos according to which the more you went up along the scale of warmth, the more you went up along the scale of pleasantness. And what do you do? You turn that scalarity against me, saying that tomorrow there will more warmth and therefore more pleasantness: consequently, putting off that walk is a reasonable thing to do. Th e ar- gument is a clever one because, so to speak, I am beaten at my own game: you use a principle against me which I was the fi rst to use in my interest. What am I going to be able to answer? No doubt many things, but there is one thing I cannot do, which is to reject your argument as irrelevant, because your argument rests on something upon which mine also does. Here is another type of counter- argument which brings out the scalarity of a topos: it is what I call exaggeration or, refutation by exaggeration.. I will stick to the same example and once again look for a way for you to answer my argument. If you use the strategy of refutation by exaggeration, you can say: “Well, in the Sahara desert, I suppose you’d spend your time going for a walk, wouldn’t you?” If you can resort to that strategy, it is be-  Slovenian Lectures cause I have established a scalar relationship between the two properties of warmth-heat [chaleur] and pleasantness. Th ere again, I am beaten at my own game: you are saying “So, when the degree of warmth-heat is extreme (the Sahara desert is considered as one of the hottest places in the world), one should also have an extreme degree of pleasantness”. As that is obviously false, my argument comes out looking unreasonable. And if you study real arguments and real controversies, you will oft en see that type of refutation being used. I recall an argument which I jotted down at a time when, in France, the reduction of the school-syllabi was being discussed. A number of so-called market-orientated teaching specialists contended that children would learn far better if they were given fewer things to learn. Somewhere, I had found the following argument put forward by someone who precisely was against the cut: “For those people, schools will be perfect when nothing at all is taught there any more”. Th at is a typical use of the exaggeration strategy. You go right up the scales of a topos till you reach a result which is generally considered as unacceptable, and thereby you think that you have successfully refuted your addressee. You could object that sometimes people base their arguments on principles which are not scalar. Let us suppose, for example, that someone has been murdered, even here say, at four thirty, and that he has been stabbed to death (a very important detail for my demonstration). Th e culprit is being looked for and the police suspect a certain French linguist who is present-ly in Ljubljana: that linguist had reasons to resent his victim, who had been very unpleasant about the theory of argumentation in general and about scalarity in particular; moreover, the wound could very well have been made with the dagger which that linguist usually has in his luggage. At that moment of the inquiry, a new piece of information reaches the police: the information that at four thirty, the time of the crime, the French linguist was at his hotel and obviously could not have stabbed someone here. In vir-tue of the following argument, he is found not guilty: “It cannot be him, as he was at his hotel at four thirty”. Such an example does seem to show that the principles which arguments rest upon are not necessarily scalar. In that case, the argument rests on a principle according to which When a person is not in a place, he cannot do anything there, and there seems to be nothing scalar about that principle at all. Up to me then to show you now that there is something scalar in the argument in question. We are going to stick to the same situation. Well, the police have just received the information that at four thirty, the linguist was at his hotel. Th en, all of a sudden, some more information reaches them ac- Lecture IV  cording to which in fact, the linguist was not at his hotel but much further from the place of the crime, for example that he was visiting a castle situated out of town, in the country. Now, having said “At four thirty, he was at his hotel”, a policeman may very well say to correct what he has just said: “In fact, he was even visiting the castle”. I think that the policeman would really tend to use an even to correct the fi rst piece of information. Now, remember my description of even. I say that even relates two arguments mov-ing towards a common conclusion, the second argument being represented as more forceful than the fi rst. So, “He was at the castle” is a more forceful argument than “He was at the hotel” for the conclusion aimed at (“He’s not guilty”). Why more forceful? If it is a more forceful argument, it is because the topos which the policeman was using was not When a person is not in a place, he cannot do anything there but rather Th e further a person is fr om a place, the lesser he can do something there, so that the linguist being in the castle at the material time, he was even less likely to have committed the murder than if he had been at the hotel. A last word about that example. It is an interesting one to distinguish those two notions which I have already spoken about: the notions of argumentation and reasoning. Indeed, in his reasoning process, the policeman (I have no reason to take him for a fool) was surely using the non-scalar principle When a person is not in a place, he cannot do anything there: that was how he reasoned in his head. But what I am concerned about is what the policeman said, not his reasoning process. In so much as the policeman uses an even (“He was at his hotel, and in fact, he was even at the castle”), what he says implies a scalar vision of the relationship between distance and the possibility of action. I must distinguish on the one hand, reasoning, which is not necessarily scalar, and on the other, argumentation, relevant to what people say, which, in my view, is always scalar. If one accepts that description of even which I have given, a description according to which even joins two arguments, the second of which is represented as more forceful than the fi rst for their common conclusion, then one must say that, in what he says, I stress, in what he says, the policeman does use a scalar principle, even if it is not at all the principle which commands his reasoning. Th at example is worth noticing only if one clearly distinguish- es reasoning, by which from certain facts certain other facts are concluded (and that is not my business, it is something for logicians to worry about) and on the other hand, the arguments through which speech conveys them. My thesis is that argumentation, as formulated in speech, is based on scalar principles. Everything that I have said supposes a clear-cut distinction  Slovenian Lectures between reasoning and argumentation, and as a linguist, what I am interested in is what goes on in speech, not what goes on in people’s heads, not reasoning. From the logical point of view, the policeman does not need to rely on a scalar principle but once he opens his mouth, he injects scalarity into things, which in themselves have none. Scalarity is a constraint which speech imposes upon us. Let me take a last example, which can be classifi ed as a more linguistic one, and which is going to involve the connective even again. You can very well imagine the following two pieces of discourse: (1) “It’s twelve, or even thirteen degrees, let’s go for a walk”; (2) “It’s twelve, or even eleven degrees, let’s go for a walk”. We know besides that in our community, we have two diff erent topoi, which I call T1 and T2: one, T1, according to which warmth makes a walk pleasant; the other, T2, according to which, it is cold which makes a walk pleasant. At diff erent times of our lives, we use now one, now the other. Now, I ask you: which is the topos used in (1) and which in (2)? I have the impression (I hope you have the same) that in (1), it is T1, the one which views warmth as a cause of pleasant-ness, and in (2), it is T2, the one which on the contrary views cold, or simply coolness, as a cause of pleasant-ness. I therefore hold it as a fact that string (1) uses T1 and string (2), T2. Now, as a linguist, I have the following question to ask myself: Why do we feel that (1) uses T1 and (2), T2? To answer, the scalarity of a topos seems very useful to me. Perhaps there are other ways of explaining the fact that I have posited, a fact which seems undeniable to me, but in any case, that fact can be explained in a satisfactory way if you have recourse to the scalarity of a topos. I remind you that according to me two arguments joined by even move towards a same conclusion and that the second is more forceful than the fi rst. In string (1), “thirteen degrees” is therefore a more forceful argument than “twelve degrees” for the common conclusion “Let’s go for a walk”. On the contrary, in string (2), “twelve degrees” must be weaker than “eleven degrees” for the same conclusion. All that stems from the description of even which I have put forward. Obviously, if you do not accept that description, the whole of my demonstration fails. Linguistic demonstrations are always indirect: a thesis can be demonstrated only through a number of hypotheses which are taken for granted. Th e hypothesis I ask you to grant me for the rest of my demonstration is my argumentative description of even, according to which the second segment is a more forceful argument than the fi rst for the same conclusion. I ask you to grant something else, which I could try to justify but it would take up too much time. I ask you to grant Lecture IV  that when even is used for a rectifi cation, that is to say to correct what has been said previously, the two arguments use the same topos. Not only do we have two arguments for a same conclusion but, when the locutor uses even to correct what he has just said (to say “I’ve made a mistake; what is true is more probably this other thing”), in that case, the two segments joined by even use the same topos. As from those two hypotheses, and if you moreover accept the idea that a topos is scalar, you will understand the facts that I have presented. Let us begin with utterance (1). If my analysis is correct, the locutor uses a scalar topos, which relates the degrees of warmth and the degrees of pleasantness. From that, you understand why that topos must necessarily be topos T1. Given that thirteen is a superior degree of warmth to twelve, a temperature of thirteen degrees, according to T1, will entail a degree of pleasantness superior to the one produced by a temperature of twelve degrees. So “thirteen degrees” will be a better argument than “twelve degrees” for the conclusion “Let’s go for a walk”. I repeat my demonstration. In string (1), “thirteen degrees” must be a more forceful argument for pleasantness than “twelve”, given that it is used through even to correct “twelve”. On which scale is “thirteen” a superior argument to “twelve”? Th e warmth scale. So, the topos must be topos T1: Warmth makes a walk pleasant. When you go up the scale of warmth, you also go up the scale of pleasantness. Let us now take string (2). Th ere again, there is a link between two scales. Th e second scale is the pleasantness scale. What can the fi rst be? Given that there is an even between twelve and eleven, we know that eleven is superior to twelve in that fi rst scale. So the fi rst scale must be the cold scale, and the topos must be topos T2: Cold makes a walk pleasant. Th e principle behind my argument is the following: in each case, the second scale is pleasantness. Th e problem is what the fi rst is. In case (1), thirteen is superior to twelve, so the scale must be the warmth-scale; in the second, eleven is superior to twelve, so it must be the cold-scale. I think that if you did not accept scalarity, you would not understand why string (1) alludes to topos T1 and string (2), to topos T2. Th at is the argument (the linguistic argument properly speaking) which I wanted to develop and which brings in the connective even. Th at notion of the scalarity of a topos justifi es the introduction of a new idea: the distinction between topos and topical form, a distinction which is entirely related to the notion of scalarity. Take a topos relating property P and property Q in a scalar way. Th at topos says that when you move along scale P in one direction, you also move along scale Q in one direction: for example, when you go up P, you go up Q. But you notice imme-  Slovenian Lectures diately that saying: Th e more you go up P, the more you go up Q amounts to the same thing as saying: Th e more you go down P, the more you go down Q. If the more you go up the warmth-scale, the more you go up the pleasantness scale, it must be the case that the more you go down the warmth-scale, the more you go down the pleasantness scale. So that a same topos, which relates warmth (P) and pleasantness (Q) in a scalar way, can have two forms, which I symbolise as on the one hand, +P, +Q and on the other, -P, -Q. I will say that those are the two topical forms, TF’ and TF’’, of the same topos T. Th e same relationship between warmth and pleasant-ness can be considered under two forms but it is viewed positively in one case and negatively in the other. I would like to say that the notion of topical form can explain a certain number of things. Here again, I am going to put myself at the rhetorical level, that is to say the level of discourse, and I am going to try to show that the duality of topical forms for a same topos can explain certain interesting rhetorical phe-nomena, for example a discourse-strategy which I call refutation of an argument by the refutation of the converse [ réciproque] argument. Let me explain what I mean by that immediately. Take an A, therefore C argument: I call the not-A, therefore not-C, the converse argument of that fi rst argument. Th e converse argument of “It’s a beautiful day, the walk will be a nice one” is the argument “It’s not a nice day, the walk will be an unpleasant one”. Th e converse argument of an argument is therefore the argument which moves from the negation of the argument to the negation of the conclusion. For a logician, those two movements of thought A, therefore C and not-A, therefore not-C are completely diff erent, and it would be a very serious fallacy (well-nigh a mortal sin for a logician!) to confuse the two. Now, a fact which I fi nd very interesting is that in speech, it is extremely frequent to confuse A, therefore C and not-A, therefore not-C. I am going to give an example of that, even several. When a logician comes across such a phenomenon, he says: “Th e reason behind those monstrosities is that people have not studied logic enough, do not think enough, are mentally undisciplined”. In fact, I believe that confusion between an argument and the converse argument has its roots in the conditions of discourse itself, and especially in the existence of two topical forms for a single topos. Th at is what I am going to try to show. What does the strategy, which consists in refuting an argument by refuting the converse argument, consist in? Someone says: A, therefore C and to refute the argument, you say: Not-A, therefore not-C, which is the converse of what he has just said. Th e strategy goes by unno- ticed in everyday conversation. I am going to make up an example, perhaps Lecture IV  not a very convincing one, but which is clear, and then I will take real examples. Someone says “Peter works, so he’s going to succeed”: A, therefore C. And as an objection to that argument, he is reminded of the following: “Many people do not work, and nevertheless succeed”. He is given the example of such and such, who spends his life resting, but who, having stocks and shares, thrives wonderfully. Let us try to see the underlying mechanism. What have you shown in saying “Many people do not work and nevertheless succeed”? You have shown that there is no connection between X does not work and X does not succeed, between not working and not succeeding. And that fact of having shown the lack of a connection between not working and not succeeding is easily taken as a refutation of your opponent’s argument, which moved from work to success. Now, I am going to give you examples which I have really observed, and then I shall try to explain them with my theory. You know that in France at the moment, there are great debates on the status of immigrant workers. Th ere are discussions especially about whether they should have the right to vote, at least in local elections. Th ose who are in favour of giving immigrant workers the right to vote oft en rely on the following argument: “Th ey pay taxes, so they have a right to vote”. Indeed, immigrant workers do pay taxes in as much as they get wages, and the argument, which seems a very reasonable one, consists in concluding that they must be allowed to vote. At a certain time, the supporters of that point of view would stick up posters with an immigrant going to the tax-collector and bringing the money for his taxes: the tax-collector would take the money with a big smile. In a corner of the poster, you could see the same immigrant at the poll-station; but there he would be pushed out in a hostile and indignant way. Th e contrast, which was supposed to highlight the absurdity of the situation, was designed to suggest that when you paid taxes, you had the right to vote. In a debate on the right to vote for immigrants, which was reported in the papers, someone said the following to object to the argument: “So, in your mind, those who do not pay taxes should not vote”. Th e fi rst argument was: “Th ey pay taxes, so they have the right to vote”, and the other picks that up and says: “So, in your mind, those who pay no taxes do not have the right to vote”. Th e refutation was an extremely clever one, because in the history of the French democracy, one of the great dates was 1848 (I think it is 1848) when universal suff rage was introduced, when the right to vote was granted to all citizens, irrespective of whether or not they paid taxes, whereas before only those people who paid taxes had the right to vote. So, the position of the fi rst speaker was made out to seem con-  Slovenian Lectures trary to the laws of the French democracy, contrary at least to one of the great acquisitions of the French democracy. Let us try to bring out the mechanism underlying that debate. Th e fi rst speaker says: “Th ey pay taxes, so they have the right to vote” (T, V). Now, in the French democratic tradition, withdraw-ing the right to vote from people who do not pay taxes is unacceptable. Th at explains the idea that to conclude that people have the right to vote from the fact that they pay taxes is a ridiculous one. Th e second speaker has taken the converse argument: he has shown that the converse argument of the fi rst argument is unacceptable, and concludes that the fi rst argument itself is unacceptable. I wonder how one could rebut that second speaker’s statement, at least in a way which, in the present state of society, would be eff ective. Let us now try to explain that way of arguing with the idea of topical form. Let us suppose that someone argues A  C. If what I have said is true, the argument relies on a topos, which connects two properties, P and Q; A is related to P and C, to Q. To argue A  C, he relies on the idea +P, +Q. So, as A attributes a certain degree in property P to something, you have to accept C, which attributes a certain degree in property Q to something else (or the same thing). Now, let us suppose that you put forward the converse argument: non-A, so non-C. On what principle do you rely? You rely on the converse of the topical form which the fi rst argument relied on. You rely on something like -P, -Q. So the person who said: “When you pay taxes, you have the right to vote” (A  C) relied on the topical form: Th e more you pay taxes, the more you have the right to vote (+P, +Q). His opponent identifi es that argument A  C and the argument non-A  non-C (You cannot pay your taxes, so you cannot vote), which relies on the topical form -P, -Q ( Th e less you pay taxes, the less you have the right to vote). What makes that polemically very useful assimilation linguistically legitimate (I am not saying logically legitimate, because for a logician, the assimilation is a culpa-ble one) is the fact that +P, +Q and -P, -Q are two forms of the same topos: the topos relating P and Q in a scalar way. It is the scalarity of a topos which implies that it must have two forms: Th e more you go up P, the more you go up Q; Th e more you go down P, the more you go down Q. Given that the arguments A  C and not-A  not-C rely on those two forms, they tend to be more or less confused: consequently, when you have made out the argument non-A  non-C to seem unacceptable, you give the impression that you have successfully destroyed the validity of argument A  C. I will now sum up everything I have said so far. A same topos in so much as it establishes a scalar relationship between two properties can appear under two forms: the more you go up one, the more you up the other; the more Lecture IV  you go down one, the more you go down the other. Th ose two forms cannot be separated: you cannot refuse the one but accept the other. Now, that duality of topical forms, which stems necessarily from the scalarity of a topos, explains a lot of things I think: it explains a number of linguistic issues, which we will speak about later, and it also explains certain rhetorical phenome-na, such as the one I have just analysed. * * * Now, with the help of topoi theory which I have introduced, I would like to answer a question which Žagar asked me. Žagar said to me: “when you speak about polyphony, you always take examples in which there are only two enunciators. I would very much like to fi nd examples in which there are several enunciators”. So, I am going to try to give you one. I will perhaps not have time to justify my analysis completely but at least I will give you the result I have reached. Let us suppose you have to describe a string made up with but, for example the string (you must be getting to know it by now): “It’s warm, but I’m tired”, a string used to answer, and reject, a suggestion for a walk. Someone has suggested going for a walk, because the weather was warm and you answer: “It’s warm, true, but I’m tired”. For me, there are at least four enunciators in that string (there would certainly be even more, if the two segments were more complex, for the segments are rather simple ones). Two enunciators are related to the fi rst segment and two other ones, to the second segment: E1 and E2, are related to “It’s warm”; E3 and E4, to “I’m tired”. E1 describes the weather: he describes it by saying “It’s warm”. But “It’s warm” is represented as an argument in favour of a walk. So, when he says “It’s warm”, he is alluding to a particular type of warmth, that type of warmth which favours walks. So, I will say that, about the situation you and he are speaking, that is about the weather, E1 calls up a topical form like Th e warmer it is, the more pleasant whatever is. According to him, that situation justifi es the use of the topical form Th e warmer it is, the more pleasant whatever is about that situation. Th en, another enunciator, E2, comes in, who from E1’s point of view concludes to the walk. Aft er having characterized the weather as having that type of warmth it takes to make a walk pleasant, there is the conclusion that a walk would be a good idea to take advantage of the weather. E3, on the contrary, whose voice can be heard in the segment “I’m tired” uses a topical form like Th e less one’s physical state is good (being tired is not a very good physical state), the less going for a walk is pleasant. So, in giving “I’m tired” as an argument for not going for a walk, the physical state is be-  Slovenian Lectures ing represented as a property making a walk unpleasant. Finally, there is an E4 enunciator who concludes from E3’s point of view to no walk. E4 is to be identifi ed (at least, if the utterance is not an ironical one) with the locutor himself, which explains why the utterance, taken as a whole, serves as a refusal to go for a walk. In saying “It’s warm, but I’m tired”, you put forward four enunciators: the fi rst characterizes the weather as making a walk pleasant (E1); the second (E2) concludes for a walk; then, the physical state of the locutor is characterized as making the walk unpleasant, following the topical form Th e less the physical state is good, the less going for a walk is pleasant (E3); and lastly, E4 (identifi able with the locutor) moves on to conclude to no walk. Why have I distinguished those four enunciators? those four points of view? I think I can justify this distinction in the following way. Th e locutor who says “It’s warm but I’m tired” completely agrees that it is warm, and even that it is warm in a way that makes a walk pleasant: that is what I mean in saying that the locutor accepts E1’s point of view. What he does not accept is E2’s point of view: he refuses the conclusion that he should go for a walk. To do so, he imposes E3’s point of view, and he also imposes E4’s conclusion. Th us, what justifi es distinguishing E1 and E2 is that the locutor does not have the same attitude towards the two: in one case, he agrees; in the other, he disagrees. To see the diff erence between E1 and E2, you have to accept (this is my main thesis) the idea that there is a diff erence between summoning a certain topical form (which is what E1 does) and using that topical form to draw the corresponding conclusion, say “So, we must go for a walk”. Th e locutor recognises that the weather is favourable to going for a walk but he absolutely refuses to draw the conclusion. So much for the distinction between E1 and E2. Now that I have distinguished summoning a topos and using that topos for a particular conclusion, I must also distinguish E3 and E4, who also stand for the summoning of a topos and the using of a topos. Having distinguished E1 and E2 (and according to me you cannot but do so), you must distinguish E3 and E4 merely for the sake of coherence. Moreover, but there is no time to go into this, it so happens that in certain situations, the locutor can have diff erent attitudes regarding the two points of view (he can simply accept E3 but identify himself with E4). With that example, I hope to have shown that there are other cases of polyphony than the negative utterances which I spoke of two lectures ago: there are much more complicated ones. I hope that I have succeeded in also suggesting that there is a relationship between the theory of polyphony and Lecture IV  topoi theory. Without explaining the point, I told you that the enunciators are argumentative entities and that their points of view are argumentatively oriented. My analysis of but is an example. E1’s point of view consists simply in summoning a topos under a certain form about a given situation, and similarly, E3’s point of view. As for E2 and E4, they draw the conclusions, and it so happens that their conclusions disagree. Th at is what I wanted to say. I am afraid I have been rather too long but next time, to allow for a discussion, I will try to stop before the time of doom.  Lecture V December 13 Let me start by reminding you where we have got to so far in this series of lectures. For two lectures, I have been trying to introduce the theory of argumentation in the language-system. Th at theory must enable the points of view of the diff erent enunciators in discourse to be characterized as argumentative and not as informative, truth-conditional or logical. In my account of the theory of argumentation in the language-system, I started justifying the general thesis according to which the argumentative function of utterances was largely determined by their linguistic structure, that is to say by the sentences those utterances are occurrences of. Th e problem which remains to be solved is how that linguistic structure can determine the argumentative function of utterances. I tried to show that it determined it, but how does it do so? Last time, to resolve that problem, I studied argumentative strings, and especially relatively simple strings of the A, therefore C (A  C) type: an argument, therefore a conclusion. I contended the following: behind those strings, there is a third term, a warrant, which authorizes the move from A to C. I call that third term a topos. I tried to characterize topoi by saying (1) that they were represented as the shared beliefs of a certain community, (2) that they were general, that is to say, that they indicated a link between two general properties, P and Q, connected with respectively A ( argument) and C ( conclusion). Finally, I tried to show (3) that those topoi were scalar. What I meant by that was on the one hand, that properties P and Q were scalar properties and on the other, that the relationship between them was a scalar one too. Th at internal scalarity of a topos enables one to say that a topos which posits the presence of P as being favorable for the presence of Q can be represented under  Slovenian Lectures two topical forms, which I call converse. One of those topical forms says, that when you go up or down in P, you also go up or down in Q. So, my topos T, relating P and Q, has two topical forms, TF’ and TF’’: TF’ = +P, +Q; TF’’= -P, -Q. Th at was where we had got to in the last lecture. Now what there is left for me to show is how those topoi, which command argumentative strings, are written into the language-system itself. Th e argumentative strings themselves belong to speech, or discourse: they are relationships between two discourse-segments, one of which is represented as justifying the other. To show that those topoi are written into the language-system, I am going to study a certain number of examples. To begin with, I am going to study words which belong to the lexicon, to the vocabulary, and then I shall study words which belong more to grammar. * * * Let us consider the following four adjectives: courageous, timorous, prudent, rash. We all feel that those four adjectives belong to a single category, and that they describe the same kind of conduct, but viewed in thoroughly diff erent ways. In the four cases, the question is a man’s possible attitudes to danger (I shall indeed be taking the word courageous only in the sense of active physical courage, the kind of courage consisting in confronting danger; I shall not be speaking about courage in the moral sense, the courage there is in upholding a paradoxical idea, nor of passive physical courage, the courage there is in not screaming when a dentist pulls a tooth out). Many people who have thought about language have noticed that there was something those four adjectives had in common. Th ose who work within Greimas’ semiotic perspec- tive say that those four adjectives are the four angles of a square – the Greimas square being a sort of adaptation of Aristotle’s logical square. I am not going to go into a criticism of those conceptions: I prefer to give you my own way of describing those four adjectives straightaway. I am going to accept the idea that, in the language-system itself, we have two topoi, T1 and T2, which I shall call contrary [ contraires]. I am not saying converse, because I used the word converse to characterize the two topical forms of one and the same topos. So, we have two topoi, which are contrary to one another: topos T1 ascribes value to the fact of confronting danger, to the fact of taking risks, which I express by saying that it relates the notion of risk and the notion of goodness; topos T2, on the contrary, relates the notion of risk and the notion of evil; in one case, the fact of taking risks is viewed as something good, in the other, as something evil. Each one of us, I think, is aware of those topoi: at times, depending on our discursive in-Lecture V  tentions, we represent a risk as worth taking and we have consideration for the person who takes it and at others, on the contrary, in our discourse, we represent the fact of taking risks as a bad thing. If one accepts the existence of those two diff erent topoi, one can already see how those four adjectives might be classifi ed: two of them implement topos T1 and the other two, topos T2. Which? I hope that you have the same feeling as I do about this. Courageous implements topos T1: when one says that someone is courageous, one is praising him, and one is praising him, because he dares to take risks, or has dared to do so sometimes in his life; what you have in the adjective courageous is a positive valorization of risk-taking. In the case of the adjective timorous, I would say that the topos used is still topos T1, the topos which values risk-taking positively. When I say that someone is timorous, I am blaming him. I am blaming him, because he does not dare take a risk, or has not dared to: which, according to me, implies that risk-taking is good, at least in certain circumstances. Courageous and timorous are therefore based on the same topos T1, but courageous is used to praise those who dare take risks and timorous is used to criticize those who do not manage to do so. Let us now take the two remaining adjectives: prudent and rash. As you expect, I am going to say that they both implement the same topos, this time topos T2, a topos which depreciates risk-taking. When I say that someone is prudent, except if I do so ironically, I ascribe a certain quality to that person, and I praise him because he can keep away from risks: in that way, I consider risk-taking as an evil. So, here we have topos T2. In the case of rash, the topos used is the same again. But this time, when I describe someone as being rash, I am criticizing him, I am blaming him for taking risks in an unacceptable and unjustifi ed way. So, I am blaming him for not implementing topos T2, just as I am congratulating the prudent person for implementing that topos. But, in either case, I am referring to topos T2. Now, aft er that fi rst classifi cation of the four adjectives, I must distinguish courageous and timorous on the one hand and prudent and rash on the other. I have two groups but I must make subdivisions within each of those two groups. To obtain those subgroups, I am going to bring in the topical forms. Each topos can, as we have seen, be implemented as two converse topical forms. Let us write those topical forms down: as far as topos T1 is concerned, I have two topical forms: TF ’ and TF ’’; and similarly, as far as T2 is con-1 1 cerned, we have TF ’ and TF ’’. What are those topical forms? TF ’ will be 2 2 1  Slovenian Lectures something like “Th e more one takes risks (+R), the worthier one is (+V)”. Topical form TF ’’ will be the converse of the fi rst topical form, that is “Th e 1 less one takes risks (-R), the less one is doing what one should (-V)”. Having distinguished those two forms under which topos T1, which values risk, can appear, I can, or at least hope I can, distinguish courageous and timorous, which both refer to that topos. I will say that courageous implements topical form TF ’ “Th e more one takes risks, the worthier one is”: in saying 1 that someone “is courageous”, I am stressing both the magnitude of the risks he is taking and the worthiness he thereby evinces. As for timorous, the effect is exactly the opposite: what we have is topical form TF ’’. When I say 1 that someone is timorous, I am applying topos T1 under the form “Th e less one takes risks, the less worthy one is”, which accounts for the unfavourable conclusion about the person, who has been categorized in that way. Th e same thing can be done with the two adjectives involving topos T2, which depreciate risk-taking: TF ’ (“Th e greater the risk, the greater the 2 evil”) and on the other hand, TF ’’ (“Th e lesser the risk, the lesser the evil”). 2 (It goes without saying that those representations of the topical forms are short-hand schemata.) According to me, the two adjectives prudent and rash implement the two topical forms TF ’ and on the other hand, TF ’’. 2 2 Th e prudent person is the one who does not take risks and whom one congratulates on that count. So, in saying that someone is prudent, one is implementing topical form TF ’ about him. And similarly, as far as rash is con-2 cerned, it is still the same topos but this time under topical form TF ’’: “Th e 2 more one takes risks, the more one is being evil”. When I say that someone is rash, I am saying that he takes risks in an unjustifi ed way, which directs my discourse toward conclusions which are negative, unfavourable for the person I am speaking about. Th e form of analysis which I have suggested for courageous, timorous, prudent and rash can be applied to many other groups of adjectives. You can if you like amuse yourselves inventing or, to keep up the “scientifi c” style of saying things, establishing groups of four adjectives which could be described in that way with two contrary topoi and two converse topical forms for each of those topoi. I will simply point out another one, leaving you the task of dealing with it in detail yourselves, if you are interested. Within the group generous, avaricious, thrift y, spendthrift , the same relationships are to be found as those which I have indicated for the courageous-timorous-prudent-rash group. Finding other analogous groups of words is easy. Now, I would like to compare the solution I have proposed for the description of those adjectives with an apparently far simpler and more rea-Lecture V  sonable solution which is oft en given but which, for my part, I fi nd unsatisfactory, and I will say why. One could think of describing those adjectives using the notion of connotation, a notion which is oft en used in con-temporary linguistics. For my part, I do not use the notions of denotation and connotation: they do not come into my theoretical framework at all. When one uses those notions, one defi nes denotation as the representation of a fact (denotation is therefore factual in nature) and connotation as the representation of a psychological attitude regarding that fact. So, the denotational aspect of an utterance or of a word is the objective information it provides about reality and its connotational aspect is the information it gives on the speaker. To apply those notions to the description of adjectives I have studied is tempting. Let us see, for example, how one could describe the diff erence between prudent and timorous. One would say that prudent and timorous are adjectives that have the same denotation, that is to say designate the same facts, but which have diff erent connotations. So, to describe “Peter is prudent”, one would say that there is a denoted component, that I symbolise as “a” and which one could paraphrase as “Peter avoids risks” – that would be what the word denotes; and then a connoted component, that I label “b” which one could paraphrase as “I approve” – and implicitly – “Peter’s avoiding risks”. So, on the one hand, the locutor describes what is in eff ect the case – “Peter avoids risks” – and on the other, he indicates his attitude towards what is denoted. Th e denoted compo- nent would be exactly the same for “Peter is timorous”: it would still be “Peter avoids risks”. Th e two words would diff er only through their connoted component, that is “b”. For “Peter is timorous”, the connoted component would be something like “I disapprove of Peter avoiding risks”. Th e same thing could be done for courageous and rash. I shall say why such a solution is contrary to the very spirit of the theory which I have been introducing for now four lectures and then, why it seems to me that the solution does not really pull through. First, why do I not like that solution? I think you can guess why. Scientifi cally, it may not be very relevant to say what my likes and dislikes are but, even so, I would like to show you the reactions that I can have towards that theory, given all that I have said up to now. Th at theory which opposes denotation and con- notation is based it seems to me on a decision which governs a whole part of Western philosophy and linguistics, at least since the seventeenth century: the decision to oppose the objective and the subjective. Of course, denotation is on the side of the objective and connotation, on the side of the subjective. You remember that the opposition of the objective and the sub-  Slovenian Lectures jective, of which denotation and connotation are a particular case, is the opposition I spoke about in the fi rst lecture between what the grammar of Port-Royal called the dictum and the modus or again, the opposition between what speech-act theorists call on the one hand, the propositional content and on the other, the illocutionary force. In the three cases, the basis is ultimately the same concepts but used in a diff erent way. Now, precisely, I am trying not to use that facile opposition between the objective and the subjective. I try not to use it because I do not really know what the objective is – personally, I have never come across the objective in my life, whatever eff orts I may have made to do so. So, for me, that opposition, which is apparently a very clear one, because we are accustomed to it, is in fact a very confused one. What I am attempting to do, as I have been explaining from the beginning of this series of lectures is (how shall I say?) to amalgamate the objective and the subjective. Th e notion of argumentation which I am developing here mixes what could be called the objective and the subjective components of meaning together. You remember that I blamed speech-act philosophy, or the grammar of Port-Royal, with maintaining a separate slot for a sort of objective, factual, truth-functional capture of reality: the dictum or propositional content. So now, I cannot use the denotation- connotation opposition, which, ultimately, is of the same type. Th at is why I feel a certain repugnancy (the word may be a little too strong, but never mind) from those notions. Now, I must also show (this is the “serious” part of my criticism) that in the analysis of the adjectives I am interested in, that distinction between denotation and connotation produces relatively unsatisfactory results. Th e theory I am discussing at the moment posits that there is a common component, “a”, in prudent and timorous, which is the denoted component “Peter avoids taking risks”: from the denotational point of view, the same thing is alleged to be said in either case. But, does that idea correspond to the meaning we give to the words prudent and timorous? I think not. When I say that someone is prudent, what are the risks which, according to me, he avoids taking, and justifi ably does? Th e risks he avoids taking are unreason- able risks, bad risks. Th e prudent person avoids taking risks which should not be taken, and I cannot possibly approve his taking such risks. When on the contrary, I say that someone is timorous, I am of course saying that he avoids taking risks. But which risks? Th e risks which he avoids and which I blame him for avoiding are reasonable risks, or merely apparent ones. Th e timorous person keeps away from risks which in fact, he should take whereas the prudent person keeps away from risks which are unreasonable, which Lecture V  are unjustifi ed. Th ere is therefore abuse it seems to me in maintaining a common denoted component for prudent and timorous, because the risks in question are not at all the same in the two cases. One could say that in another way by saying that the connoted component comes out on to the denoted component and transforms it: from the fact that I approve Peter’s avoiding risks, those risks become unreasonable; from the fact that I disapprove, they become reasonable risks. Now you understand that reasonable and unreasonable are not denotational notions. An advocate of the theory I am criticizing could not say to me: “Well, I am going to improve my theory to answer your objection, and I will say that when one uses the word prudent, the denotation is ‘He avoids unreasonable risks’ and when one uses the word timorous, the denotation is ‘He avoids taking reasonable risks’”. An advocate of the connotation- denotation theory cannot give me that answer, because reasonableness and unreasonableness have nothing to do with denotation: for a risk to be reasonable or unreasonable is not a fact (nor is it a fact for the risk to be real or apparent). To try to give you a real example of that distinction between reasonable risk and unreasonable risk, and so between prudence and timorous-ness, here is a factual experience. Personally, when I drive my car, rightly or wrongly, I have the impression of being prudent, that is to say of avoiding risks which seem to me unreasonable. Now, when my children judge the way I drive, they think I am timorous: that is to say, that I avoid taking even reasonable risks, or only apparent ones. Probably, if I were to judge myself as I was thirty years ago, I would fi nd myself not prudent but timorous. Th at goes to show that there is nothing objective in that distinc- tion between prudent and timorous, because the notion of risk present in those two notions is completely subjective. Th at was what I wanted to say on the group of four adjectives and, generally, on the use of topoi and topical forms to describe lexical words. * * * Now, I am going to take an example from a completely diff erent area: grammar. My fi rst example will be the word almost. And there again, I am going to try and show that to describe that word, which is a constituent of sentences from the language-system, it can be very interesting to bring in argumentative notions. We are going to study structures of the almost X type, in which X is a numerical or more generally quantitative expression. For example, “It’s almost eight o’clock”, “it’s going to last almost ten minutes”, “It  Slovenian Lectures costs almost a thousand francs”, and so on. As I said, I am restricting my study to almost modifying expressions of a quantitative type. First, what is the usual description (that is to say, as you have guessed, the one which I am then going to reject)? Th e usual description consists in saying that almost X indicates a quantity inferior to the quantity indicated by X: almost X is a little less than X. Th e description seems altogether rea- sonable, at fi rst. If you ask what the price of a book is and I answer “It costs almost a hundred francs”, you immediately draw the conclusion that it must cost ninety or ninety-fi ve francs. If you ask me what time it is and I say “It’s almost six o’clock”, you draw the conclusion that it must be slightly before six o’clock, for example fi ve to six. Or again, if you ask how long my lecture is going to last, my answer “It’s going to last almost another ten minutes” suggests that it is going to last seven or eight minutes. So, to describe almost X as meaning a quantity a little inferior to X does seem reasonable. In doing that, one thinks one has characterized almost in a purely factual, truth-functional way, without bringing in any argumentative notions. Unfortunately, things, I fear, are far less simple than that: I have noticed quite a large number of examples in which almost X indicates a quantity which is superior to X; in which almost X means (if one can use the word mean to say that) more than X. I am going to give a few examples, and then I shall try to deal with them using the notion of topos. At a time when the dollar was going down enormously in France, I remember reading in the papers: “Th e dollar is almost down to fi ve francs” (there was an economic cri-sis in the U.S.A. at the time, and the dollar was going down every day). In that context, “Th e dollar is almost at fi ve francs” meant that it was worth a few centimes more than fi ve francs: so, it meant “more than fi ve francs”. I take another example. I have an appointment with a friend at 8 o’clock. My friend arrives at ten past eight, and I start getting angry with him, because I consider he is making me lose my time. He can very well answer: “Ah, don’t start getting angry with me. Aft er all, it’s almost eight!”, and his “it’s almost eight” in that case is ten past eight, that is to say a little bit more than eight. Last example, an example I have experienced recently. Imagine a car, in which there are three persons. At the front, there is Mr X, the driver and, next to him, Mrs X. Behind them, there is a passenger. Th e passenger, it so happens, is me. Mr X and Mrs X were taking that passenger to a perform-ance of a play, which the three persons all wanted to see. Negligently, Mr X had forgotten to take some petrol and also to map out the route. Whence this reproach from Mrs X to her husband: “We’re going to be late because of you”. Th ereupon Mr X answers: “Not at all, we’ll be there on time. Th e Lecture V  theatre is almost fi ve minutes away from here”. Upon my questioning him on his “almost fi ve”, he confi rmed having meant “between fi ve and ten”... Before the pause, I was questioning the usual description of almost according to which almost X is less than X and to do so, I gave you a certain number of examples in which almost X meant on the contrary “more than X”. With facts of that kind, the linguist is faced with an alternative. A fi rst solution consists in saying that there are two words behind almost, one meaning “less than” and the other, “more than”. So, one could claim that there was a phenomenon of ambiguity here, or, in a more moderate version of that solution, of polysemy resulting from a double usage of a same word. One could also say, second solution, that the word almost, properly speaking, means neither “less than” nor “more than” but something more general, that is “about”: saying “almost eight o’clock” would amount to saying “about eight o’clock”. Th at solution is a possible one but I think it is not really an accurate one, because we feel a marked diff erence between saying “I’ll be there at almost eight o’clock” and “I’ll be there at about eight o’clock”. Th e use of “almost” implies a certain intention which is foreign to “about”. Besides, if one is to interpret almost in a given context, one always knows if it is to be understood as “more” or “less”, which is not the case for “about”. Now, I am going to try to introduce an argumentative solution which gives almost a single description, and which nevertheless avoids confusing almost with about. I will posit that to understand almost X, one must know the argumentative orientation of the segment in which almost X is to be found: that is, one must know the conclusion (let us call it “R”) for which “almost X” is said, and if one does not know conclusion “R” for which one says “almost X”, then one cannot understand “almost X”. More precisely, I will posit the following rule: “almost X” is oriented towards the same conclusion as “X” but is argumentatively less forceful than “X”: it moves less forcefully towards that conclusion than “X” does. What there is left for me to do now is to show that the description I have just given allows one to predict in which cases “almost X” means “less than X” and in which cases, “more than X”. Let us suppose that I say “It’s almost eight o’clock” with the argumentative intention of getting you to notice that it is late (strictly speaking, I should not say “argumentative intention”, since I am not concerned with the locutor’s intentions: I should simply say “orienting what I say towards a conclusion like ‘It’s late’”). In that case, we shall say that a topos or one of its topical forms like “Th e more time goes by, the more one is late” is be- ing implemented. For example, the point is to blame you for arriving late:  Slovenian Lectures what I say, “It’s almost eight”, is based on this remark which stems from common sense that “Th e more time goes by, the more one is late”. Follow- ing the rule which I have given for almost, “eight o’clock” must therefore be a more forceful argument than “almost eight o’clock” for the conclusion “One’s late”; and if “eight o’clock” must be a more forceful argument than “It’s almost eight” in favour of the conclusion “One’s late”, which is reached through the topical form “Th e more time goes by, the more one is late”, then “almost eight” inevitably means something like “ten to eight”, that is to say is a less forceful argument than “eight o’clock” for the conclusion “One’s late”. So, with a conclusion like “One’s late”, with a topical form like “Th e more time passes, the more one is late”, “almost eight” must indicate a moment of time inferior to eight o’clock, for example “ten to eight”. Let us suppose now that “almost eight” is directed towards the opposite conclusion, that is to say is directed towards the conclusion “We’re not late”. It is the case in the example of my appointment with my friend: my friend says “It’s almost eight” to have me notice that aft er all, we are not late, and so that I should not be blaming him. In that case again, according to my rule, “almost eight o’clock” must be a less forceful argument than “eight o’clock” for the conclusion “One’s not late”. In that argument, the topical form is however the converse of the one used in the fi rst example: the topical form is “Th e less time has gone by, the less one is late”. It can be repre- sented as a mapping of the degrees of one scale into those of another: the fi rst scale represents the decreasing quantities of time that has gone by; the second, the diff erent degrees to which lateness can be increasingly denied. Th at is what the following table is meant to show: decreasing quantities increasingly forceful of time gone by denials of lateness It’s fi ve to being early It’s eight being on time It’s ten past being not all that late1 It’s a quarter past being not very late (N.B. What that table shows is not of course the topical form used as considered under its general form, because as such it does not mention the particular hour in the argument scale, that is, the left -hand scale. Here we have the topical form as applied in the particular situation which is being 1 Translator’s note. Neither little nor a little can be used here: the fi rst, for grammatical reasons; the second, for semantic ones (“I’m a little late” does not belong to the scale of denials but of avowals of lateness). Lecture V  envisaged, where the quantities of time are indicated by moments of time. Th e only thing that matters for my argument is that, in the scale, earlier moments, like fi ve to eight, should be above the later moments, like a quarter past eight, because they imply a lesser quantity of time gone by.) All I have to do now is to apply the general rule which I have suggested for almost to that topical form, and one immediately foresees the temporal value of “almost eight o’clock” in the example I am analysing. Being argumentatively less forceful than “eight o’clock”, and seeing the topical form used, that expression must indicate a quantity of time gone by situated below the quantity indicated by “eight” in the argument scale (i.e. a superior quantity of time gone by, since the left -hand scale goes from the greater to the lesser). Th at quantity which “almost eight” indicates is therefore the quantity implied by a moment aft er eight o’clock, which could be fi ve past eight, ten past eight, or half past eight, depending on the conception the locutor has of precision. But, in any case, given the argumentative orientation envisaged and the topical form which that orientation imposes, it must be a moment of time “superior” to eight o’clock, justifying one’s saying “It’s more than eight o’clock”. To conclude, an argumentative description, it seems to me, enables one to understand the opposite eff ects which almost X can have: sometimes, it is “more than X”, sometimes “less than X”. It is the argumentative orientation of what one says which determines whether “almost X” is “more than X” or “less than X”. Th e only unchanging thing is that “almost X” is a less forceful argument than “X”. But to be a less forceful argument than “X”, “almost X” indicates a quantity now superior, now inferior to “X”. From the theoretical point of view, that explanation seems to have two advantages. First advantage: it is better than the description saying that “almost X” is “about X”. It is better because the description with the notion “about” does not account for the relationship between the argumentative orientation and the moment of time designated by X. “About X” designates a moment of time close to X, and that, whatever the argumentative intentions may be (again, I should not be saying “intentions” but “ orientation”). Now, it is remarkable that “almost eight”, in a given case, should never indicate the global environment of “X” but, in some very precise cases, a superior quantity and in other, also very precise, cases, an inferior quantity. Th at is the reason why the argumentative description I have suggested seems to me to account for more facts that the one identifying “almost X” and “near to X” or “about X”. I will put the same point in a diff erent way. When I say “It’s around eight o’clock”, the only way for you to know if it is more than  Slovenian Lectures eight o’clock or less than eight is to look at a watch or to reason as from objective signs (for example, you have not heard eight o’clock strike yet). But when I say that it is “almost eight o’clock”, you always know whether it is more or less than eight, or rather you know if you know the argumentative orientation of what I am saying, that is to say if you know that the conclusion is based on lateness or earliness. Th at is the kind of fact, the kind of data which an argumentative description of almost can account for. I would even say that only an argumentative description can account for such a fact, fabricated as it is with the concept of argumentative orientation. A second advantage of my description of almost, and a far more general one. You remember that in the description of meaning, my purpose was to do away with the informational component. Having said that, there is no doubt that the informational component does obtain. When I use an utterance, I am giving you a certain amount of information in a certain way. If I say “It’s eight o’clock”, I am giving quite a precise piece of information, and if I say “It’s almost eight o’clock”, the piece of information I am giving you is undoubtedly vaguer but is still a piece of information. So, the informational component unquestionably does have a certain reality, which I cannot deny. What I have just done, in the analysis I have suggested for almost, amounts to deriving the informational value from the argumentative value. From the point of general linguistics, that is something which seems important to me. At least in the particular case of almost, one can derive the informational value from the argumentative value: that is to say, from the fact that the utterance is directed towards lateness or towards earliness. Examples of that kind mitigate the paradoxical or even scandalous quality which my desire to constitute non- descriptive semantics may have. What I would like to succeed in showing is that in a certain way, information can derive from argumentation. For almost, I really think I have achieved that result. Th ere are many cases for which I have not managed to do so but it is in that direction that I am working. You may remember that in a preceding lecture, without over-insist- ing on that point, I already showed you a case in which information was derived from argumentation. I had to admit that “Peter has worked little” usually indicated a quantity of work inferior to the quantity indicated when saying, in the same situation, “Peter has worked a little”. “Peter has worked little” suggests, for example, that he has worked for one hour whereas in the same situation, “Peter has worked a little” suggests that Peter has worked for two hours. So, in a certain number of cases, the use of those expressions, a little and little, can convey relatively precise information. You Lecture V  remember the experiment carried out by a psychologist which showed that, under certain circumstances, when there were two quantities to assess using little and a little, it was always the weaker quantity which summoned little and the greater, a little; but given the discourse-situation, I had deduced or derived that informational diff erence between little and a little from an argumentative diff erence which to me seemed deeper, deeper in this sense that argumentation seemed to explain information in a certain number of cases whereas information never seems to explain argumentation. I am not claiming that the results which I have reached are fi nal nor especially that they are general enough but as from little and a little, I did want to show the direction I am working in. Th e point is to introduce argumentation in linguistic meaning, and from it, to derive the informational components that one observes in the meaning of utterances produced in particular discourse-situations. * * * I take another example, the last example of this series of lectures: the example of negation. I would like to show how I bring in topoi and topical forms within the description of negation. First, I remind you of what I said when I introduced negation from a polyphonic point of view. Take the utterance “Peter hasn’t worked”. I told you that the utterance represented two enunciators, E1 and E2: E1 is a point of view from which Peter has worked and on the other hand, E2 disagrees with E1. Th at description that I gave was a provisional one which I hope I will be able to improve upon through the theory of argumentation. You can see why I am not completely satisfi ed with it. Th e thing is that at least apparently, it maintains an informational point of view at the level of enunciator E1: E1 gives the information that “Peter has worked”, and then E2 disagrees with that. So, in that polyphonic description of negation, as I have explained it, I have maintained an informational component within the linguistic description of negation. Well, I think I can now get rid of that informational component, at least in a certain number of cases. I must make a remark before going on. For what reason, in that example, do I want to avoid bringing in a propositional content (in Searle’s sense) or a piece of information like “Peter has worked”? Th e reason is that the ex- pression “Peter has worked”, properly speaking, does not convey information. What is one saying when one says that “He has worked”? One does not really know. Many people, granted, think when I stride back and forth in my offi ce muttering linguistic examples that I am not working; and they may be perfectly right to think so. Many of my friends in Paris think that, here in Ljublja-  Slovenian Lectures na, I am not working: “You’re not going to get us to believe that you’re going to Ljubljana to work. Lecturing isn’t work. Th ose who work are here in Par- is.” One could spend hours discussing what work means: there is no clear notion of work. According to me, the defi nition of the word work must bring in topical forms. Two topical forms at least constitute the meaning of this word. Work, fi rstly, is activity represented as tiring: work is tiring. Th at is what I express under the following topical form: “more work, more tiredness”. To show that someone has not worked, it is indeed usual to show that he is not tired. If someone is in top form, he will have great diffi culty in proving that he has been working. (When I go back to Paris, if I want to prove that I have been working in Ljubljana, I will have to look as if I have not been resting too much.) Th e word work contains another idea: “Work produces results”. Work is an activity which changes something in the world. If I have been content with swaying my body from side to side, even if that has made me very tired, you will not agree that I have been working. Incidentally, that is, I think, one of the reasons why in the Catholic world, up until the twentieth century, so called ‘intellectual’ work was allowed on Sundays whereas manual work was not: I remember that when my parents, walking in the country on Sundays, saw farmers ploughing the fi elds, they had the impression that those farmers were infringing upon a religious rule; but they would never have reproached me for doing my homework on a Sunday, and reading, thinking or perhaps even writing was a perfectly accepted thing to be doing on Sundays. Why? It is, I think, because intellectual work was not considered as productive but as being a mere contemplation of truth. He who devotes himself to intellectual work contents himself with bringing reality into his mind: he brings about no change in the world. It is also one of the reasons why priests were allowed to carry out intellectual work whereas they were absolutely forbidden manual work. A Catholic priest could be a researcher, because scientifi c research had nothing to do with a modifi cation of the world: it merely consisted in admir-ing the world, and, as such, was not really work. In the notion of work, I will therefore bring in not only the topical form “more work, more tiredness” but also “more work, more production”. (Th ere are also other topical forms, and I do not claim the description to be complete.) Let us go back to negation. In the description of E1’s point of view, I am now going to be able to replace the informational indication that “Peter has worked” with an indication of an argumentative type. I shall say that, concerning Peter’s activity, E1 summons one of the topical forms associated with the word work: for instance, but not necessarily, “more work, more tiredness” Lecture V  or again “more work, more production”. I shall say that E1’s point of view consists in characterizing Peter’s activity as a tiring or a productive one. In that way, the informational component vanishes from E1’s positive point of view. Now, I shall also be able to describe E2’s point of view, which confl icts with E1’s, in a more precise way. I had simply said that E2 disagreed with E1. What does “E2 disagrees with E1” mean? I can answer, I hope, using my theory of topical forms. I would say that, for the same question as E1, E2 summons a certain topical form: the converse topical form of the same topos. In my example, discussing Peter’s activity, E2 summons the following topical form: “less work, less tiredness”. More generally, I will say that semantically, a negative utterance must be understood as representing [here as previously, the word must be taken in the theatrical sense of the word] two enunciators: E1 and E2. About a given situation, E1 summons a topical form associated with the predicate which occurs in the utterance (in the example, I took “more work, more tiredness”) and E2 brings in the converse topical form, about the same situation or the same object. So, I can now describe the relationship between E1 and E2 more precisely: the relationship is that which obtains between the topical form of a topos and its converse. I would add that the locutor rejects E1’s point of view: that is, he refuses to apply the topical form summoned by E1 to the object being discussed, here the topical form “more work, more tiredness” to Peter, and he identifi es with E2’s point of view, which amounts to saying that he applies the topical form “less work, less tiredness”. Th ere is no time for me to bring out all the advantages such a solution may have but I will point out one, without being able to really develop it. “Intui-tively”, one does feel that negation is not the only morpheme to have a negative value: there are many others, for instance the morpheme little, which I have repeatedly spoken to you about. Instead of saying “Peter has not worked”, I could say “Peter has worked little”, and, between you and me, that would not make an enormous diff erence. What is the diff erence however? When you say “Peter has worked little”, in a certain way, you accept that “Peter has worked” but you declare that the work is insuffi cient, or does not really deserve to be called work. Nevertheless, even syntactically speaking, there is much in common between full negation with not and weaker negative forms such as the morpheme little. An advantage of the solution I have suggested for not is that I could give little a very similar description: thereby, I would account for the similarity which we feel obtains between the two utterances “Peter has not worked” and “Peter has worked little”. To analyse “Peter has worked little”, I will postulate the same two enunciators E1 and E2 as for the analysis of “Peter has not worked”. If one considers  Slovenian Lectures only the enunciators, those utterances can be described in the same way: one enunciator applies the topical form with more, and another, the converse topical form with less. I shall account for the diff erence by considering the attitude of the locutor. For “Peter has worked little”, I shall say that the locutor rejects E1’s point of view but admits or recognises that he does not regard it as completely unjustifi ed. In the case of not, on the contrary, his attitude towards E1 was absolute rejection. Th ereby, I acknowledge the fact that in saying “Pe- ter has worked little”, to a point, one does admit that Peter has worked, which, within my theoretical framework, means a recognition that, in another discourse, a positive topical form associated with the word work could be applied to him. But, in what he is now saying, the locutor identifi es with enunciator E2 and, again in what he is now saying, he is using a negative topical form, for example “less work, less tiredness”. By bringing in topical forms, I can therefore establish a parallel between diff erent forms, that are felt as being related to one another, of an attitude which I call negativity. Th e diff erence between those forms is merely in the attitudes of the locutor regarding the diff erent enunciators. With that last example, which I have barely outlined the analysis of, what I have time to tell you about the introduction of argumentation within linguistic analysis is complete. * * * If you are not at the end of your patience, I would now like to say a few more words to conclude. You will have noticed that in these lectures, I have oft en spoken badly about the notion of truth. In the fi rst lecture already, I told you that I refused to work with that notion. But that does not prevent my coming across it all the time. Consequently, I would like to speak of the relationship there may be between a linguist and the notion of truth. Th ree types of encounters can occur between them, I think. First type of encounter: the notion of truth can be an object for the linguist. In all languages I think, there are words like true and false: I have never heard of a language that does not have equivalents for the words true and false found in Indo-European languages. Th e linguist, who describes the words of a language, must also describe the words true and false: that is part of our work in so much as we describe the lexicon of languages. So, that fi rst type of encounter between the linguist and truth is an inevitable, ines-capable one. But those words of the language-system, let me point out, do not at all have the value that those same words have in the language of logicians: the “true” and the “false” of the language-system are totally diff erent from those of logicians. In French, I can say “Th is bank-note is a false Lecture V  [i.e. counterfeit] dollar”. Try to work that out with the logical notion of falsehood! Admiring a work of art, I can say “Th is work is a true [or real] master-piece!”. Try to explain that expression “true” with the logical notion of truth. And when to say ill of someone, I say “He’s a true [i.e. real] fool”, what does the logical notion of truth or of falsehood have to do with it? In German, to say that someone has gone the wrong way, one says “He has taken the false way ( den falschen weg)”. What does the notion of falsehood used here have in common with the logical notion that has the same name? Let me sum up: we have spoken of truth and falsehood as notions within the language-system but those notions are quite diff erent from the homonymous notions in logic. Th e “true” in the language-system certainly has something to do with the good and the “false” with the bad: there is approval in the adjective “true” and disapproval in “false”. Of course, the analysis should be carried on further. I only wanted to show that to describe those words, one had to turn away from the logical notions. So much for the fi rst type of encounter. Second type of encounter: the linguist can use the ideas of truth and falsehood as a means of carrying out his enquiry. For example, the semantician can say: “To study language, I am going to ask under what conditions sentences are true in the logicians’ sense of the word and under what conditions those sentences are false, there again in the logicians’ sense of the word”. Logical linguistics, which uses the logical notions of “true” and “false”, does exist: it is called truth-conditional semantics. Th at form of research can be extraordinarily subtle and complex, because, of course, one is not going to be content with ascribing truth-conditions to each sentence of the language-system one by one. On the contrary, one is going to construct the calculus whereby the truth-conditions of a sentence can be obtained from the description of its components. Let us take for instance the case of operators, that is to say words which transform a simple sentence into a more complex sentence (thus one can postulate that little transforms Peter has eaten into Peter has eaten little). Th eir semantic description will consist in saying how they transform the truth-conditions of the simple sentence to produce those of the complex sentence. Such a conception of semantics can be considered as a “structural” one in that it describes each component of the language system, not in itself, but in relation with its combination with others. Everything I have said from the beginning of these lectures makes you see that this form of structuring seems impossible to me: the infl uence of a word on another through its phrastic combinations cannot be described in truth-conditional terms according to me, since I refuse  Slovenian Lectures to describe the semantic value of sentences in those terms. What the combinative semantics I am proposing tries to explain is the way a word determines the argumentative orientation of the sentences in which it occurs and thereby, the argumentative usage of the utterances of those sentences. Th ere is however a third possible, or even necessary, type of encounter between linguistic research and the notion of truth, and this one does put me in an awkward position: it consists in the fact that most linguists (including myself, when I am not careful) make themselves out as formulat-ing true propositions, or at least propositions susceptible of being true or false about some particular language or another or about language in general. Such a claim must seem an unreasonable one to make if one admits that, as I have repeated over and over again, sentences of the language-system, given their semantic structure, cannot run as candidates for truth or falsehood. Indeed, until now, it is within natural languages, which according to me are argumentative and not truth-functional, that linguistic thinking is carried out and its results formulated. In that situation, I can envisage only two possible positions to adopt. On the one hand, the linguist can try to construct an artifi cial truth-conditional metalanguage. Th at is what all the natural sciences do: even if they use material signs provided by natural languages abundantly, they try to ascribe new values to them making truth-conditional discourse possible: concepts with precise appli-cation conditions are made to correspond to words. Th at goes on even for the use of connectives like so or if which in mathematics, for example, have a very diff erent meaning from the one they have in ordinary language. Th ink also of the value given in mathematics to words like necessary and suffi cient, which have only a remote bearing on their ordinary use. I would certainly like to be able to do the same thing in linguistics. But I cannot, and do not wish to, close my eyes to the fact that I am far from having achieved this. Undoubtedly, in my lectures, I have tried to construct a few defi nitions but I know full well that I have done no more than characterize the directions for the use of the words I use. For example, I cannot formulate a defi nition of the following kind: “Discourse is argumentative if and only if such and such conditions obtain”. Perhaps I will be able to do so on another visit to Ljubljana but, between you and me, I would like to come back before. It is therefore the second possible position with which I must be content, at least provisionally. It consists in consciously giving linguistic discourse itself not a truth-functional but an argumentative and especially polemical end. Its end is not to say what language is but simply to question certain simplistic images that people entertain about it; and among those images, the Lecture V  truth-functional conception through which in our daily lives, we oft en interpret the commercial, religious, sentimental or political discourse which we are addressed. I have tried to make you feel that the claim made by a discourse to be describing things is a snare which we must perpetually be wary of, not because it can be mistaken or dishonest but because it does not fulfi l even the basic conditions allowing one to hope it is true. Were fi ve lectures necessary to reach that altogether banal conclusion, and even without being able to really show it is true? Th e only excuse I can put forward is the importance and the diffi culty of putting it into practice.  Major Publications of Oswald Ducrot B o o k s L1 – Le structuralisme en linguistique. Published as chap. 1 of the joint publication Qu’est-ce que le structuralisme? , Seuil, Paris, 1968, then as a separate volume in the series Points, n° 44, Seuil, Paris, 1973. Italian trans. (I. L. I., 1970), Japanese (Chikuma Shobo, 1972), Spanish (Losada, 1975), Brazilian (Cultrix, 1970), German (Suhrkamp, 1973). L2 – Dictionnaire encyclopédique des sciences du langage (in collaboration with T. Todorov), Seuil, Paris, 1972. Later published in the series Points, n° 110, Seuil, Paris, 1979. Italian trans. (Isedi, 1972), Portuguese (Dom Quixhote, 1973), Spanish (Siglo XXI, 1974), Japanese (Asahi Shuppensha, 1975), German (Athenaïon, 1975), Brazilian(Perspectiva, 1977), American ( John Hop- kins Univ. Press, 1979), English (Blackwell, 1981), Albanian (Rilindija, 1984), Serbo-Croatian (Prosveta, 1987), Greek (Diatton, 1992). L3 – Dire et ne pas dire, Hermann, Paris, 1972., 2nd, enlarged, edition, 1991. Brazilian trans. (Cultrix, 1977), Italian (Offi cina, 1979), Spanish (Ana- gramma, 1982). L4 – La prevue et le dire, Mame, Paris, 1973. Brazilian trans. (Global Universitaria, 1982). L4’ – A part of this book, revised and enlarged, was published under the title Les échelles argumentatives, Minuit, Paris, 1980.  Slovenian Lectures L5 – Les mots du discours (in collaboration), Minuit, Paris, 1980. L6 – L’argumentation dans la langue (in collaboration with J. C. Anscombre), Mardaga, Bruxelles, 1983. L7 – Le dire et le dit, Minuit, Paris, 1985. Spanish trans. (Paidos, 1987), Brazilian (Pontes, 1987), Slovenian (Stu-dia Humanitatis, 1988). L8 – El decir y lo dicho, collection of articles translated in Spanish, Hachette-Argentine, 1984. L9 – Logique, structure, énonciation, Minuit, Paris, 1989. L10 – Polifonia y argumentacion. Univ. del Valle, Cali, Colombie, 1990. C o l l e c t i o n s R1 – In collaboration with E. Coumet et J. Gattegno: Logique et linguistique ( Langages, n° 2, Larousse, Paris, 1966). R2 – In collaboration with M. C. Barbault: Linguistique et mathématiques ( Langue fr ançaise, n° 12, Larousse, Paris, 1971). R3 – In collaboration with M. C. Barbault: Enseignement du fr ançais et enseignement des mathématiques (Recherches pédagogiques, n° 56, I. N. R. D. P., 1973). R4 – Des grammaires générales au transformationalisme (selected, translated, and commented texts). Panorama des sciences humaines, N. R. F., 1973, chap. 5 (p. 475-531). R5 – Linguistic part of the l’ Introduction à l’histoire des sciences, published under the editorship of G. Canguilhem, Hachette, 1970. P a r t i c i p a t i o n i n J o i n t P u b l i c a t i o n s T1 – Participation in the French translation of Le langage, by L. Hjelmslev, Minuit, 1966. T2 – Participation, for the linguistic part, in the draft ing of the report, presented by J. Viet to the Conseil International des Sciences Sociales. Report published under the title Les sciences de l’homme en France, Mouton, La Haye, 1966. A r t i c l e s A1 – “Logique et linguistique”, introduction to the collection R1, pp. 3-38. Republished, with modifi cations, in L4. Major Publications of Oswald Ducrot  A2 – “Le roi de France est sage: implication logique et présupposition linguistique”, Etudes de linguistique appliquée, 1966, pp. 39-74. A3 – “Quelques illogismes du langage”, Langages, n° 3, 1966, pp.126-139. A4 – "La commutation en glossématique et en phonologie", Word, 23, 1967, pp. 101-121. A5 – “Chronique linguistique (à propos de Benveniste et de Priéto)”, L’homme, 1967, pp. 109-122. A6 – “Le problème de la négation dans diverses grammaires françaises” (in collaboration with S. Barnicaud, A. M. Comparé, A. Vidal), Langages, n° 7, 1967, pp. 58-73. A7 – "Langage et communication", Travail social, 1966-7, pp. 34-38. A8 – “La glossématique”, Supplément à la grande encyclopédie Larousse, 1968. A9 – “La notion de présupposition et la description sémantique des énoncés français”, L’homme, 1968, pp. 37-53. A10 – “Peu et un peu”, T. A. Informations, 1969, pp. 11-13. A11 – “Les actes de langage”, Sciences, n° 60, 1969, pp. 8-13. A12 – "La chaîne parlée; la syntagmatique", Guide alphabétique de la linguistique (under the editorship of A. Martinet), Denoêl, 1969, pp. 37-53. A13 – “Logique et langage”, Guide alphabétique de la linguistique (under the editorship of A. Martinet), Denoêl, 1969, pp. 232-239. A14 – “Présupposés et sous-entendus”, Langue fr ançaise, n° 4, 1969, pp. 30-44. Republished in L7. German translation in the collection by Petöfi , Präsuppositionen in Philosophie und Linguistik, Athenaïon, 1973, pp. 241-273. A15 – “Peu et un peu”, Cahiers de lexicologie, 1970, pp. 21-52. English translation in the collection by F. Kiefer et N. Ruwet, Generative grammar in Europe, 1973, pp. 178-202. A16 – “Les indéfi nis et l’énonciation”, Langages, n° 17, 1970, pp. 91-111. Republished in L4. A17 – "Réponse à A. Trognon, sur les diffi cultés de la théorie de la présup- position", L’homme, 1970, pp. 82-83. A18 – “Logique et langage”, L’éducation, n° 105, 1971, pp. 16-18. A19 – "Langue et pensée formelle", in/dans R2, pp. 3-12.  Slovenian Lectures A20 – "L'expression en français de la notion de condition suffi sante", in R2, p. 60-67. A21 – "A propos de la Seconde provinciale”, in R2, pp. 90-92. A22 – “Analyse logique d’un texte de Montesquieu” (in collaboration with J. Depresle), in R2, pp. 93-97. Th e articles A19, A20, A21, A22 were republished, with some modifi - cations, in L4. A23 – “D’un mauvais usage de la logique”, in the collection by J. Martinet De la théorie linguistique ŕ l’enseignement de la langue, P. U. F., 1972, pp. 137-151. A24 – “Enseignement du français et enseignement des mathématiques”, in R3, p. 5-22. A25 – „Analyses de langue: ou, la notion de cause, le rôle de la négation dans le langage ordinaire“ (in collaboration with M. C. Barbault), in R3, pp. 86-106. Th e articles A24 and A25 were republished, with some modifi cations, in L4. A26 – „La description sémantique en linguistique“, Journal de psychologie, 1973, 1-2, pp. 115-133. A27 – “Les présupposés, conditions d’emploi ou éléments de contenu?”, Actes du congrès de linguistique de Varsovie, 1968, in the collection J. Rey-Debove, Recherches sur les systèmes signifi ants, Mouton, La Haye, 1973, pp. 243-257. Polish translation: Pamietnik Literacki, 1975, pp. 269-282. A28 – “Humboldt et l’arbitraire linguistique”, Cahiers internationaux de symbolisme, 26, 1974, pp. 15-26. Republished in L5. A29 – “Je trouve que”, Semantikos, 1, 1976, pp. 69-88. Republished in L5. A30 – “Car, parce que, puisque”, (in collaboration with the group L-I), Revue romane, Copenhague, tome 10, fasc. 2, 1975, p.258-80. A31 – “Quelques implications linguistiques de la th éorie médiévale de la supposition”, in the collection by H. Parret, History of Linguistic Th ought and Contemporary Linguistics, de Gruyter, Berlin/New York, 1976, pp. 189-227. Republished in L9. A32 – “L’argumentation dans la langue” (in collaboration with J. C. Anscombre), Langages, 42, 1976, pp. 5-27. Translated in the Brazilian edition of L4. Major Publications of Oswald Ducrot  A33 – “Mais occupe toi d’Amélie” (in collaboration), Actes de la recherche en sciences sociales, 6, 1976, pp. 47-62. Republished in L5. A34 – “Illocutoire et performatif ”, Linguistique et sémiologie, Univ. Lyon II. Republished in the 2nd edition of L3. Translated in the Spanish edition of L3. English translation in Semioti-ca, supp. 1981, pp. 181-201. A35 – “Deux mais en français?“ (in collaboration with J. C. Anscombre), Lingua, 43, 1977, pp. 23-40. A36 – “L’énonciations dans la description linguistique”, Actes de la session de linguistique de Bourg-Saint-Maurice, Ed. de l’Univ. Paris III, 1977, pp. I.1.-I.47. A37 – “Structuralisme, énonciation et sémantique”, Poétique, 33, 1978, pp. 107-128. Republished in L7. Translated in the Italian and Brazilian edition of L3. Polish translation: Pamietnik Literacki, 1980, pp. 321-342. A38 – “Echelles implicatives, échelles argumentatives et lois de discours” (in collaboration with J. C. Anscombre), Semantikos, vol. 2, n° 2-3, 1978, pp. 43-65. A39 – "Lois logiques et lois argumentatives" (in collaboration with J. C. Anscombre), Le fr ançais moderne, vol. 46, n° 4, 1978, pp. 347-357, and vol. 47, n° 1, 1979, pp. 43-65. Texts A32, A38, and A39 were, modifi ed, republished in L6. A40 – "Atti linguistici", Enciclopedia Einaudi, published between 1977 and 1982. A41 – “Dicibile/indicibile”, ibid. A42 – “Enunciazione”, ibid. A43 – “Presupposizione/alluzione”, ibid. A44 – “Referente”, ibid. A45 – Texts A40-A44 were translated in the portuguese edition of Enciclopedia Einaudi, t.2, Impresa nacional, 1985, ( Linguagem - Enunciaçao). A46 – “Présupposés et sous-entendus: réexamen”, in the collection Stratégie discursives published by Presses Universitaires de l’Université de Lyon II, 1978, pp. 33-48. Republished in L7. Translated in the Spanish edition of L3.  Slovenian Lectures A47 – “Les lois de discours”, Langue fr ançaise, 42, 1989, pp. 21-33. Republished in L7. A48 – “L’imparfait en français”, Linguistische Berichte, Wiesbaden, avril 1979, pp. 1-23. Republished in the collection by F. J. Hausmann, Etudes de grammaire fr ançaise descriptive, pp. 25-44, Julius Groos Verlag, Heidelberg. A49 – „Quatre textes, un seul but“ (in collaboration), Langage et société, 5, 1978, pp. 53-73. A50 – “Décidément: la classifi cation dissimulée” (in collaboration), Journal of Pragmatics, vol. 3, n° 2, 1979, pp. 127-149. Republished in L5. A51 – "De magis à mais: une hypothèse sémantique” (in collaboration with C. Vogt), Revue de linguistique romane, 1979, n° 171-2, tome 43, pp. 317-341. A52 – "Analyses pragmatiques", Communications, 32, 1980, pp. 11-60. A slightly diff erent version of the same work was published in the collection by H. Parret, Le langage en contexte, John Benjamins, Amsterdam/Philadelphia, 1980, pp. 489-575. A53 – “Note sur les obligations conditionnelles (à propos d’un article de J. C. Gardies)”, Revue philosophique, n° 4, 1980, pp. 413-420. A54 – "L'argumentation par autorité", in the collection L’argumentation, Presses universitaires de Lyon, 1981, pp. 9-27. Republished in L7. Italian translation: Materiali fi losofi ci, n° 4-5, 1981, pp. 7-25. A55 – "Langage, métalangage et performatifs", Cahiers de linguistique fr ançaise, Univ. de Genève, n° 3, 1981, pp. 5-34. Republished in L7. A56 – "Interrogation et argumentation" (in collaboration with J. C. Anscombre), Langue fr ançaise, n° 52, 1981, pp. 5-21. Republished in L6. A57 – "La valeur argumentative de la phrase interrogative", Actes de colloque de pragmatique de Fribourg, 1981, Peter Lang, Berne, 1983, pp. 79-110. A58 – “La notion de sujet parlant”, Philosophie et langage, Université de Grenoble, 1982, pp. 65-93. A59 – „Note sur l‘argumentation et l‘acte d‘argumenter“, Cahiers de linguistique fr ançaise, Genève, n° 4, 1982, pp. 143-163. A60 – " Justement, inverseur argumentatif ” (in collaboration), Lexique, Presses de l’Univ. de Lille, 1982, pp. 151-164. A61 – “Puisque, essai de description polyphonique”, Mélanges C. Vikner, Revue romane, n° spécial 24, 1983, pp. 166-185. Major Publications of Oswald Ducrot  A62 – "Opérateurs argumentatifs et visée argumentative", Cahiers de linguistique fr ançaise, Univ. de Genève, n° 5, 1983, pp. 7-36. A63 – "Analyse énonciative des Animaux malades de la peste” (in collaboration with G. Declercq), Actes du colloque d’Albi, 1983, Université de Toulouse-Le Mirail, 1984, pp. 5-38. A64 – “Polyphonie”, Lalies, n° 4, 1984, pp. 3-30. A65 – " Enfi n, opérateur métalinguistique” (in collaboration), Journal of Pragmatics, n° 9, 1985, pp. 199-239. A66 – "Informativité et argumentativité" (in collaboration with J. C. Anscombre), in the collection in homage to Ch. Perelman, published by M. Meyer, De la Métaphysique à la rhétorique, Bruxelles, 1986, pp. 79-93. English translation (by J. C. Anscombre) in the English edition of that collection, 1990. A67 – “Sous un mot, une controverse: les emplois pragmatiques de toujo-urs” (in collaboration), Modèles linguistiques, 1985, fasc. 2, pp. 105-124. A68 – “Charles Bally et la pragmatique”, Cahiers Ferdinand de Saussure, 1986, vol. 40, pp. 13-37. Republished in L9. English translation: Diacritics, John Hopkins Univ. Press, vol. 21, n° 4, 1991, pp. 3-19. A69 – " Toujours (suite): le cas des conclusions assertives” (in collaboration), Modèles linguistiques, 1986, fasc. 2, pp. 115-122. A70 – “Pour soigner le minimalisme” (in collaboration with J. C. Anscombre), Journal of Pragmatics, 1986, vol. 10, pp. 435-440. A71 – “L’interprétation en sémantique linguistique: un point de départ imaginaire”, Confr ontations, Aubier, Cahier 17, 1987, pp. 119-133. Republished in the 3rd edition of L3. A72 – “Sémantique et vérité, un deuxième type de rencontre”, Recherches linguistiques, Univ. de Paris VIII, 16, 1987, pp. 53-63. A73 – “Argumentation et topoï argumentatifs”, Actes de la 8ème rencontre des professeurs de fr ançais de l’enseignement supérieur, Helsinki, 1987, pp. 27-57. Portuguese translation in the collection by E. Guimaraes, Historia e sen-tido na linguagem, Pontes, Campinas, 1989, pp. 13-38. Spanish translation in the collection by B. Lavandera, Lenguaje en contexto, vol. 1, n° 1-2, 1988, Buenos Aires, pp. 63-84.  Slovenian Lectures A74 – "Topoï et formes topiques", Bulletin d’études de linguistique fr ançaise, Tokyo, 1988, n° 22, pp. 1-14. A75 – "Topoï et sens", Actes du 9ème colloque d’Albi, 1988, Univ. de Toulouse-Le Mirail, 1989, pp. 1-22. A76 – “Albine, ou les surprises de l’innocence: explication de la première scène de Britannicus”, Actes du 11ème colloque d’Albi, 1990, Univ. de Toulouse-Le Mirail, 1990, pp. 1-29. A77 – “Argumentation et persuasion”, in W. de Mulder, F. Schuerewegen, L. Tasmowski (éds.), Enonciation et parti-pris: Actes du Colloque d’Anvers, février 1990, Rodopi, Amsterdam/Atlanta, 1992, pp. 143-158. A78 – “Opérateurs argumentatifs et analyse de textes”, Linguistic perspectives on Romance Languages: Selected Papers fr om the XXI. Linguistic Symposium on Romance Languages (Santa Barbara, 1991), in the series Current Issues in Linguistic Th eory, n° 103, John Benjamins, Amster- dam/Philadelphia, 1993, pp. 45-62. A79 – "A quoi sert le concept de modalité?" in N. Dittmar, A. Reich (éds.), Modality in Second Language Acquisition, de Gruyter, Berlin/New York, 1993. A80 – “Les modifi cateurs déréalisants”. Journal of Pragmatics, 1994. A81 – “Les topoï dans la théorie de l’ Argumentation dans la langue”, in Christian Plantin (éd.) Lieux communs, topoï, stéréotypes, Kimé, Paris, 1994, pp. 233-248. R e v i e w s C1 – “Y. Bar Hillel: Language and Information”. L’homme, 1966, pp. 136-138. C2 – “T. Todorov: Recherches sémantiques”. L’homme, 1966, pp. 120-121. C3 – “A. J. Greimas: Sémantique structurale”. L’homme, 1966, pp. 121-123. C4 – “J. Lewitt: Th e Grammaire des grammaires de Girault-Duvivier”. Lin- gua, 24, 1969, pp. 98-100. C5 – “O. Jespersen: Philosophie de la grammaire”. La quinzaine littéraire, n° 134, février 1972, pp. 18-19. C6 – "A. Marty: Psyche und Sprachstruktur". La linguistique, 1972, n° 2, pp. 153-158. Republished in L9. Turkish translation in Türk Dili Dergisi, 1991, n° 27, pp. 16-20. Major Publications of Oswald Ducrot  C7 – "H. J. Uldall: Outline of Glossematics". La linguistique, 1973, n° 2, pp. 149-152. C8 – "R. Martin: Inférence, antynomie, paraphrase". Revue romane, 1979, pp. 150-153. C9 – “H. Nolke: Les adverbes paradigmatisants: fonction et analyse”. Revue romane, n° 23, 1986, pp. 137-143. P r e f a c e s P1 – “De Saussure à la philosophie du langage”, introduction to the French translation of Speech Acts, de J. R. Searle, Hermann, Paris, 1972, pp. 7-34. P2 – Preface to O intervalo semântico, de C. A. Vogt, Editora Atica, Sao Paolo, 1977, pp. 1-20. P3 – “Note sur la présupposition et le sens littéral”, epilogue to Le mauvais outil, de P. Henry, Klincksieck, Paris, 1977, pp. 171-203.  Index Argumentation\discursive argu- A mentation 26 Anscombre 48, 63 Argumentation\theory of argu- Argumentation 26, 27, 29, 41, 46, 48, mentation in the language sys- 51, 52, 53, 60, 62, 67, 71, 72, 81, 86, tem 26, 48, 62, 81 92, 93, 96 Aristotle 63, 65, 82 Argumentation\argument 12, 20, Austin 17, 18 21, 22, 23, 26, 37, 39, 49, 53, 54, 61, 63, 64, 65, 67, 68, 69, 70, 71, 72, 73, B 74, 75, 76, 81, 90, 91 Bally 15, 16, 17, 18, 19 Argumentation\argumentative Benveniste 34 analysis 59 Argumentation\argumentative de- C scription 57, 61, 72, 91, 92 Argumentation\argumentative Connotation 85, 86, 87 function 48, 49, 52, 53, 54, 55, 57, Conversation 15, 31, 34, 52, 75 63, 81 Argumentation\argumentative ori- D entation 24, 25, 26, 27, 49, 53, 54, Denotation 85, 86, 87 55, 57, 60, 89, 91, 92, 98 Descartes 16 Argumentation\conclusion 20, 21, Descriptive 14, 15, 18, 24, 27, 42, 92 22, 23, 25, 26, 27, 49, 50, 52, 53, 54, Dictum 16, 17, 18, 30, 86 55, 58, 61, 64, 67, 68, 71, 72, 73, 74, Discourse 13, 14, 15, 25, 26, 34, 35, 78, 81, 89 36, 39, 40, 42, 45, 48, 49, 51, 52, 53,  Slovenian Lectures 54, 55, 58, 63, 64, 66, 68, 69, 72, 74, M 81, 82, 83, 84, 93, 96, 98, 99 Meaning 13, 14, 15, 17, 18, 19, 20, 22, F 23, 24, 26, 27, 29, 32, 35, 45, 46, 47, 48, 61, 86, 92, 93, 94, 98 Freud 36, 37 Modus 16, 17, 30, 86 G O Grammar 16, 82, 86, 87 Objective 16, 20, 85, 86, 87 Greimas 82 P I Pascal 41, 42, 43, 44, 45 Implicature 51, 52, 53 Polyphony 11, 27, 29, 37, 39, 40, 45, Informative 14, 15, 17, 19, 24, 27, 55, 46, 47, 77, 79 56, 58, 81 Polyphony\polyphonic 38, 39, 40, 41, 42, 44, 45, 47, 93 J Port Royal 15, 16, 17, 18, 19, 30, 86 Jakobson 13 Propositional content 18, 19, 20, 21, 23, 30, 46, 86, 93 L Proust 25, 26, 27, 34 Language 12, 13, 14, 16, 19, 21, 26, R 27, 29, 32, 42, 45, 48, 57, 58, 60, 62, 63, 82, 87, 96, 97, 98 Representation 12, 14, 16, 17, 18, 19, Lexicon 12, 42, 96 27, 36, 37, 40, 49, 85 Linguistics 11, 15, 29, 56, 59, 85, 92, Rhetorical 62, 69, 74, 77 97, 98 Linguistics\argumentative linguis- S tics 56 Searle 17, 18, 19, 20, 21, 30, 46, 93 Linguistics\linguistic activity 12, 19 Semantics 11, 13, 45, 92, 97, 98 Linguistics\linguistic structure 48, Sentence 14, 15, 16, 17, 19, 21, 22, 45, 49, 55, 56, 57, 61, 63, 81 48, 55, 63, 81, 87, 97, 98 Logic 97 Social sciences 12 Logic\logical reasoning 26 Speaker 14, 17, 18, 19, 20, 29, 30, 35, Logic\pseudological 15, 19 85 Logical 15, 54, 55, 72, 81, 82, 97 Speaker\enunciator 30, 35, 37, 39, 40, 41, 42, 44, 45, 46, 48, 49, 53, 54, 77, 78, 79, 81, 93, 95, 96 Index  Speaker\locutor 30, 32, 33, 34, 35, V 36, 47, 67, 73, 78, 95, 96 Speaker\producer 30, 31, 32, 33, Value 14, 15, 19, 22, 26, 27, 48, 51, 55, 34, 35 56, 58, 59, 82, 92, 96, 98 Speech 14, 33, 34, 40, 48, 71, 72, 74, Voltaire 67 82 Speech act 17, 18, 19, 20, 23, 24, 86 Spinoza 26 Subjective 16, 17, 18, 19, 85, 86, 87 T Topos 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 76, 77, 78, 81, 82, 83, 88, 89, 95 Topos\topical forms 73, 74, 76, 77, 78, 82, 83, 84, 87, 89, 90, 91, 93, 94, 95, 96 Topos\topoi 27, 45, 63, 67, 72, 77, 79, 81, 82, 84 Toulmin 64 Truth conditions 14, 19, 20, 21, 22, 23, 47, 97 Truth conditions\falsehood 15, 45, 46, 47, 48, 97, 98 Truth conditions\truth 15, 45, 46, 47, 48, 96, 97, 98 Truth conditions\truth conditional 15, 17, 18, 19, 24, 27, 46, 81, 97, 98 U Utterance 14, 15, 18, 19, 20, 24, 25, 26, 27, 29, 30, 31, 32, 33, 34, 35, 36, 37, 39, 40, 41, 45, 47, 48, 51, 55, 56, 57, 60, 79, 81, 85, 92, 93, 95, 96, 98 Utterance situation 51 Document Outline Oswald Ducrot, Slovenian Lectures, Digitalna knjižnica/Digital Library, Dissertationes 6 (Front cover) Oswald Ducrot, Slovenian Lectures, Digitalna knjižnica/Digital Library, Dissertationes 6 (Title page) Copyright page Table of Contents In Place of Introduction Lecture I Lecture II Lecture III Lecture IV Lecture V Major Publications of Oswald Ducrot Books Collections Participation in Joint Publications Articles Reviews Prefaces Index