From the WAB archives: A selection from the Bergen Wittgenstein archives' Working Papers and audio-visual materials

• Home
• About
• Login
• Search
• Current
• Archives

Table of contents

• References

In the Yellow Book, that is to say the notes for the Cambridge lectures of the years 1932 to 1935, in the middle of a discussion about what makes the difference between hypotheses and grammatical rules, Wittgenstein suddenly asks us to consider all the things about which we would have to change our minds in order to be able to accept as an hypothesis that there is a hippopotamus in this room. He thus chooses a hypothesis which is at the limit of absurdity in order to compel us to consider that which we would otherwise escape. Wittgenstein has used the same trick many times. For instance, when, in the Philosophical Investigations, he considers the proposition "a rose has no teeth", saying that we would like to take it as obviously true, more sure than the one stating that a goose has none, but that this is not, in fact, so clear if we look at the matter more closely. Whereas we know where we have to look in order to check whether gooses have teeth or not, we do not know at all, at first sight, where we could find the teeth of a rose. There could be, for example, a drawing or a photography of a hippotamus somewhere in this room. Or there could exist a species of hippotamus not bigger than a mouse, or a fly, or even smaller, that has remained unnoticed up to now. What grants that an hippopotamus is a very big animal? His discussion about the teeth of the rose leads Wittgenstein to the very strange proposition that it is more accurate to say I know what you think than to say I know what I think. He says that we have here a whole cloud of philosophy condensed in a drop of grammar. From Wittgenstein's point of view, this is a question of grammar. The same holds for the proposition: "there is an hippotamus in this room". What we have to do with such a proposition depends upon what we mean by a hippotamus. The hippotamus can very well be a fat, more or less unpleasant, awkward, person, if we have in mind that an hippotamus is that kind of being. All this shows very clearly that Wittgenstein's concept of grammar is very far away from that of the linguists.

I cannot pretend to cover such a broad topic as "Wittgenstein and contemporary linguists"; I should have better entitled this paper: "Some remarks about Wittgenstein and contemporary linguists". I have choosen this topic because very few people seem to have been really interested in it. "Wittgenstein and language" that is almost a lieu commun. But "Wittgenstein and linguistics" or "Wittgenstein and linguists" that does not seem to have been much studied. We have linguists who have tried to use some of Wittgenstein's ideas about language. We have other linguists who have rejected vigorously any of those ideas as a kind of provocation. But who has tried to understand why at the same time this interest and this repulsion?

First of all, let me draw your attention upon the fact that Wittgenstein does not seem to have been interested in the works of the linguists. He does not seem to have been acquainted with any of those. Anyway, his main concern is with the philosophers and the logicians, not with the linguists.

For instance, Wittgenstein has discussed extensively names, and names have also worried Frege and Russell, as well as any of those we consider to be "philosophers of language". The question of names has been linked to that of the logical constants. What does Wittgenstein say about names? He does not try to tell us what is a name; he does not try to show that, as a linguistic expression, a name is a specific kind of such expression, having a specific linguistic function. He criticizes our way of using such expressions as "names of ..."; "names of numbers", "names of colors"; "names of directions (in space)" ... In the Brown Book he shows that we use the different expressions we call names in very different ways; he denies that there is something common to all the different uses of such expressions; he says that those uses must be studied case by case and that if some of them appear as governed by rules, we should not take as granted that there is a super-rule that we could call the rule of names of which the different rules of names would be special cases. In particular, he criticizes the idea that such a rule of names would he that any name refer to an object; he says that if we do so, we have to consider numbers, colors, directions, and so on, as objects, and then what we call objects appear to be so different things that we are compelled to say that we have different kinds of objects. We are then compelled, he tells, to say in particular that we have concrete objects on the one hand and abstract objects on the other. This is how philosophical problems arise, because as soon as we differentiate between abstract and concrete objects, we have to discuss what these two types of objects have in common, as well as what makes the difference; we would so be led to discuss the specific modality of existence of each kind of objects. Shall we, for instance, conceive this distinction in the line of that introduced by Plato between the tangible and the intelligible, between the numbers as numbers and the numbers used by merchants and speculators, between the numbers which can be handled only by thought and those which have a visible and tangible body. As you know, Wittgenstein attacks, in the Blue Book, those who say that a number and a numerical sign are two different kinds of objects (the number one and the sign "1") for instance. So Wittgenstein sees behind the problem of names a problem that has worried philosophers since philosophy exists, that of ontology. And as you know also, from Wittgenstein's point of view, philosophical problems are only pretended problems arising from our uncritical use of language.

If we now turn our eyes towards the linguists, things are completely different. As a matter of fact, we can see that at least in some grammars, we also speak of names, of different kinds of names; common names and proper names, names of objects and names of persons, and so on. But, from a consistent linguistic point of view, this has to be justified, whatever grammarians of the past have done. It is not obvious and self-evident that the category of names is a linguistic category. If, as a linguist, we want to use such a category, we have to show that at least in some languages, or even in one language, we have a class of expressions which have all one common linguistic function, a function that differentiates the expressions of this class from other expressions. But it might well be that in a given language, the category of names does not exist. There has been, among linguists, a long discussion about whether there exist languages in which there are no verbs, the famous case being that of the Basque language (in which there is instead an ergative function which can be attached to almost any kind of words). We could have a similar discussion about the existence of names in some languages. But it might even be that we should conclude that the category of names does not exist as a linguistic category in any language. Finally, in the case that such a category exists as a linguistic category, it might well be that what linguists then call names did not always coincide with what we are used to call names (think, for instance, of the distinction between the masculine and the feminine).

So, linguists also have to discuss what we call names as well as about what they call names. But when they discuss that point they do not do it as Wittgenstein does. What Wittgenstein has in mind is our usual way of speaking of names, our usual way of distinguishing between different kinds of names given that this differentiation is also used by philosophers in an uncritical way, as if it was obvious and self-evident that it is grounded. He does not discuss the way linguists speak of names. That, in many cases, the linguists have also followed uncritically the usual way of speaking of names, is a point on which we cannot but agree. The question whether it is possible or not, from a linguistic point of view, that is to say from the point of view of the linguistic function of words and expressions, to speak of names and of different kinds of names, remain. And this has to do with what linguists require.

The linguists require at least two things. First, they require that it is possible, within any language, to distinguish between classes of expressions or units on the base of their linguistic function. Second, they require that the classes of expressions, their functioning, as well as their relations, can be defined in such a way that we get a basis of description of any language whatsoever. This second possibility has to do with that of general linguistics. These are the minimal requirements for any linguistics. They presuppose that we can speak of languages, i.e. that we can distinguish what belongs to a given language from what does not. And we know that this can raise many problems which are not solved by speaking of dialects instead of languages. To be able to speak in terms of dialects also requires that we can distinguish between what belongs to a given dialect and what does not. In that case too, we have to use notions such as those of grammaticality or correctness or the like. Without the possibility of using such notions, no linguistics is possible. But there is another specificity of languages. The case of languages is not alike that of an animal for instance. If we have to do with animals, and if that animal is not a dog then it is another animal, a cat or whatsoever. Despite the fact that the definition of zoological taxonomies might raise problems, the situation is not the same with languages. What does not belong to a given language, no matter whatever the criterion might be to decide that, does not necessarily belong to another language. Moreover, that which does not belong to a given language, without belonging to another, can still belong to a language and have a meaning. Furthermore, anything that has meaning does not necessarily belong to language. So linguists need to be able to speak of the language in general; they need to be able to put within language things that do not belong to any language but to language in general. In other words, they need to be able to distinguish what belongs to language in general from what does not. We all know that, for instance, gestures may have more or less the same meaning as linguistic utterances (pointing with one finger to an object, nodding one's head, and so on).

Wittgenstein has placed himself in a position in which he might appear as having get around all the difficult problems that these fundamental requirements of any possible linguistics arise. For instance, it happens that, in some cases, he puts on the same level gestures and linguistic utterances. It also happens that within an argument he refers to words or utterances belonging to different languages. But, precisely, in those cases, he does not deal with those words and utterances as belonging to different languages. On the other hand, his notion of language games raises problems which have some similarities with those which the linguists are confronted. What makes the difference between one language game and another game? What kind of relations exists between the different language games? What makes the difference between what is a language game and what is not a language game? Can we find within Wittgenstein's writing a definite concept of language?

No more than I have attempted to discuss what Wittgenstein says about names or language games, will I try to discuss whether the requirements of the linguists are grounded or not. Let me just draw your attention upon the fact that those requirements define an ideal. How this ideal can be reached, as well as to which extent it can be reached, is an other point. I have recalled all this only to stress the fact that Wittgenstein's point of view and that of the linguists appear as radically different. And that raises a question: does what Wittgenstein has written or said about language have any relevance for linguists and linguistics? This will be my question today.

Saying that Wittgenstein has made of language a crucial concern for philosophy implies ipso facto that his conceptions deserve to be considered by linguists has, I think, induced much confusion. More precisely, I will try to show that the attempts to apply – I insist on that word "to apply" – some of Wittgenstein's ideas to linguistics, have generated confusion because those ideas were not relevant from the point of view of linguistics. These attempted applications have induced confusion on both sides; they have led to betray Wittgenstein's conceptions and to betray linguistics. But I will also try to show that if we drop the idea of applying Wittgenstein's ideas to linguistics, then we can see that some of his ideas are relevant for linguistics, but in a very specific way. In order to avoid the confusions just mentioned, I think that this relevance must be very precisely defined and circumscribed.

I have stressed the difficulties linguists are confronted with. I have more or less presupposed that there exist such a thing as the linguistic point of view. Linguists need to believe that such a point of view exists and that it can be precisely and without ambiguity defined, but that does not imply that this belief is grounded. As for the contemporary linguists, we can see that very few of them have attempted to clear up that point whereas it is obvious that among all those who claim to be linguists, there is no agreement about what is linguistics, what it should be or can be. The question which is raised today, bears precisely upon the possibility to reach such an agreement. Today, in France at least, this question is even raised at an institutional level because linguistics has been put among what is now called the "sciences of language" and because, among those pretended sciences, the specificity of linguistics appears as diluted. That among those sciences of language we find that part of psychology (cognitive or not) that deals with language, that part of neurophysiology or neuropathology that deals with the use of language, its acquisition, its disturbances, that part of sociology which, under the name of sociolinguistics, deals with language, and so on – all this cannot but rewake the old devils that, since the very beginning of linguistics, linguists have perceived as threatening their knowledge and identity. When I discuss with my linguists friends today, many of them appear to be more or less in a state of anxiety or even of anguish, seeing threats all around them. Others seem to be ready to burn all the saints they have revered and to pass with weapons and luggage in the camp of the threateners. Some of them, eventually, call for Wittgenstein for support and legitimization of what they are doing.

On the other hand, if we consider the past history of linguistics, we can ask ourselves whether linguists have not always felt more or less insecure about what they are doing. The case of Baudouin de Courtenay gives a very exemplar instance of what I am talking about. As you now, Baudouin has with his pupil and colleague Kruszewski contributed in a very decisive way to the elaboration of the concept of phoneme and thus to the foundation of phonology. Baudouin is the one who claimed, in his inaugural lecture in Saint-Petersburg in 1870, that linguistics should be considered as an independent science, not to be confused with psychology or physiology. Baudouin is also the one who has agreed for a while with the opinion of Kruszewski according to which the only firm base which the science of language can have has to be found within language itself and not outside it (i.e., among other possibilities, within the brain or the society ...). This conception of what should be linguistics has led Baudouin and Kruszewski to elaborate a whole set of concepts or notions which have been later retaken and re-elaborated by Saussure and other pioneers in linguistics. But Kruszewski died prematurely and it seems that his death has been quite dramatic. At least, this death has been very dramatic for Baudouin. What has happened is that, after Kruszewski's death, Baudouin has rejected the whole of what he had done in collaboration with Kruszewski. Baudouin had followed the teaching of one of the founders of the Leipzig school, Alfred Lieskens. Whereas, at the same time of his collaboration with Kruszewski, he had taken some distance with that psychophysics, it is on that base that he has rebuilt his conception of language and of linguistics during his last years. The psychophysics of the Leipzig school imposed to him theoretical as well as methodological orientations: a strong separation between the physical aspects of language, which had to be studied by physical and physiological methods, and its psychological aspects which had to be investigated by means of introspection, both being linked together according to conceptions inspired by Fechner. The psychological aspect of the phoneme, in order to be able to be studied by means of introspection, had to be conceived as belonging to the conscious linguistic representation of the individual. This led Baudouin to deny that there exist such things as languages; he was driven to argue that the reality of language is purely individual and that the only reality of language is that of idiolects. To speak of the properties of the Russian language, for instance, became, from his point of view, a convenient but misleading way of speaking. When we look more closely at Baudouin's last texts, we can see that all his anxiety became focused on the modality of existence of linguistic entities, especially that of the phoneme. He seemed to have also feared in a very acute way what, when we are speaking and listening, remains out of our conscious control. On the contrary, what had so much impressed Kruszewski when he met Baudouin was the stress the later was putting on the unconscious feature of the forces operating within language. Baudouin has ended his life in a state of deep depression, saying that all the disappointments he had to face during so many years made him pessimistic and have deprived him of any desire to live. He said that he considered himself as a superfluous and worthless being.

I could comment upon this by just saying that to commit oneself to the study of language is not free from any threat upon one's psychological equilibrium. But the story I have just told raises, I think, a more fundamental question as similar events seem to have repeatedly occurred all along the history of linguistics. Think on what happened to Saussure when he started to find anagrams in any piece of language and not only in Latin verses. Saussure seems to have experienced, in those circumstances, something which was close to madness. If later he seems to have been more quiet, he has dropped his investigations on anagrams. But, first, that does not mean that his experience with anagrams has not had a decisive influence upon his later conceptions, those expressed in his famous lectures on general linguistics. Second, if Saussure has adopted in his Cours a position in the line of that taken at the beginning by Baudouin and Kruszewski, in his investigation of the Germanic legends which he carried on at the same time, he was somehow contradicting what he was saying in the Cours. This leads me to consider the strange kind of desire which might pushe people to become linguists, a kind of desire which might be somewhat masochistic since it might be that, in the last resort, linguists can never be sure that the knowledge to which they devote all their efforts is not purely illusionary.

One may say that the case of the linguist is not so specific as it may appear. We have the cases of leading scientists who, like J.F. Mayer or Georg Cantor, have experienced such a deep insecurity about the grounding of what they had discovered that this has left them without defenses in front of serious mental disorders. Nevertheless there seem to be something specific to the study of language, at least with the attempts to build a science of language as did Saussure, Baudouin and Kruszewski. If we agree upon the fact that language has something to do with our existence as subjects, if not with our existence as living beings, then studying language concerns our existence as subjects. Language is not any kind of object of scientific investigation, especially if one considers that the proper ground for this investigation has to be found within language and not outside it. That may explain why those who have devoted themselves to the study of language have been so strongly and recurrently inclined to escape this fate by looking outside of language for bases able to account for it. Considering language as a form of behavior, eventually as a rule governed form of behavior, or as a mean of communication of thoughts (while taking as granted that thought exists independently of language), or as a code for communicating information (also existing independently of language), stand among the available means which have been recurrently used for this end. This is what Baudouin, going back to a psychological account of language, did at the end of his life. As long as he has not been alone facing the intrinsic evanescence of linguistic entities (the languages, the phonemes ...) Baudouin has been able to stick to his position, even if we can find in his early writings some confusions or indications which can explain the later drift of his thought. It seems that after Kruszewski's death he has not been able anymore to face alone uncertainty and doubts about the existence of linguistic entities. His comments about Kruszewski's works, in the necrological note he wrote about his former friend, sound like the bitter echoes of an unhappy love-affair. He was much unfair in his evaluation of Kruszewski's contribution to linguistics because, without that friend, he was pushed to withdraw as quick as possible to positions that he felt more secure for himself, even if those positions revealed themselves afterwards, as the end of his story tells, to be nothing but pis-allers.

Now, we can understand why Wittgenstein's ideas upon language have exerted such a mixture of fascination and repulsion upon linguists. To put it in a few words, I may say that Wittgenstein has put his finger straight in the middle of the wound with a precision, an accuracy and an insistence which has never been matched. Furthermore, we can say that he has not only put his finger straight in the middle of the wound; he has at the same time cut all the possible ways of withdrawal. More systematically than anyone else, Wittgenstein exposes the traps of all forms of psychologism. Those who have attempted to interpret what he means by language games in a sociological way have obviously betrayed his thoughts. And for him, as for the founders of linguistics, language cannot be accounted for from the outside. In other words, more than any one else, Wittgenstein compels linguists to be faithful to themselves and to face their fate; he compels them to clear up their point of view, what makes its specificity, what it requires. Today, given the situation of confusion in which linguistics is, that cannot but deserve consideration. But that does not imply that Wittgenstein gives to the linguists the means for grounding linguistics. On the contrary, his position may be understood as denying the possibility of any linguistics whatsoever.

Some linguists have well seen that. They belong among those who have shown their concern with the problem of the epistemological status of linguistics, and who have tried to give to it at least the beginning of an answer. There are rather few of them. But I know at least one who has attempted to grapple with that problem in a book recently issued: Jean-Claude Milner in his book Introduction a une science du langage (Paris, Le Seuil, 1990). In his attempt, Milner could not but encounter Wittgenstein. In fact, he refers to him only twice, but at crucial steps. Furthermore, he says something that I feel very important; that linguists cannot demonstrate that Wittgenstein is wrong in his basic conceptions of language, even if those imply that no linguistics, conceived as a science of language, is possible. As you know, the fact that you cannot demonstrate that a statement is wrong does not imply that this statement is right and, conversely, the fact that you cannot demonstrate that a statement is true does not imply that this statement is wrong. Mathematicians know that well. In that sense we might speak of a Wittgensteinian conjecture about language. Moreover, I insist upon that point, formulating statements that cannot be demonstrated as being right or wrong is, first of all, not easy at all, not trivial, and, furthermore, always highly meaningful.

But we have a problem. What does it mean to say that we cannot demonstrate that Wittgenstein is wrong (or right)? When Milner says that, it is of course in reference to a precise issue, not about Wittgenstein's sayings as a whole. It concerns what I have recalled at the beginning, i.e. that from Wittgenstein's point of view we just have words and expressions that we can use, that our uses of words and expressions must be studied case by case, that if those uses appears as governed by rules, those rules do not combine themselves together so as to make a whole or an unified system. Then, says Milner, if we follow Wittgenstein, we have to conclude that there is no such a thing as what we call a language, that we have just language games, and that what we call a language or the language does not exist or exists only as exists for Spinoza the general idea. Wittgenstein's point is that there are no rules common to all possible language games, even that speaking of all possible language games is meaningless. Having in mind, when we speak of a language, instead of what we call, for instance, Norwegian, or French, or English, but dialects or anything of that kind whatsoever, does not change anything in that respect. What is the case is that Wittgenstein rejects the minimal requirements under which any linguistics is possible. Therefore, says Milner, if we follow Wittgenstein, we have to declare that linguistics is illusionary.

We might be inclined to say that Wittgenstein's position is close to that one of Baudouin de Courtenay at the end of his life. But there are in fact fundamental differences between the two conceptions. First, even such a thing as Baudouin's idiolect does not exist for Wittgenstein; for him, there is no unified systematicity of that kind either. Anyone is, by turns, engaged in one or another language game, but there is no general, unified systematicity of all language games. Second, Wittgenstein does not establish his views upon language on psychological considerations as Baudouin did. There is a link between these two points because what linguists have always tried to save is the idea that any language makes a unified homogeneous system. That holds at all levels, that of the languages, of the dialects or even if, like Baudouin, they make use of the concept of idiolect, at that level too. Of course linguists agree that within language one must distinguish between at least a phonological, a syntactical and a semantical level. But the general theoretical assumption that there is a single type of organization common to all these levels has been dominant within linguistics. According to this general assumption, language as an object would be describable by progressive complexification (combinations of phonemic traits making phonemes, combinations of phonemes making words, combination of words making sentences...). Structuralism has given this assumption its greatest extension in linguistics, making of language a structure of structures. In that conception of language, this one is organized in levels of stratification, the law of organization of each level being the same, for instance that of a linear structure in which all the properties can be stated in terms of contiguity (syntagmatic contrasts) and of distinctivity (paradigmatic opposition). (See, for instance, Roman Jakobson, "Two aspects of language and two types of Aphasias", in Jakobson, R. and Halle, M., Fundamentals of Language. S'Gravenhague, Mouton, 1956, pp. 55-86 or Noam Chomsky, Aspects of the theory of syntax, Cambridge (Mass.) M.I.T. Press, 1965, n. 2, p. 131).

It has been almost spontaneously admitted that language has a homogeneous, unitary, structure. But such a hypothesis, whatever has been its fecundity, is a very strong one. It has led linguists constantly to look for speculative justifications. Saying that language is a psychological reality or a social reality or that the basic organization of language is determined by that of the brain, are the justifications which have been most frequently ventured because the mind, or the society, or the brain, have been conceived as having also a homogeneous, unitary, structure. But it must be noticed that, first, such justifications have not had any concrete incidence upon linguistic descriptions and, second, that, if on the theoretical level, we have had this hypothesis of homogeneity, at that of empirical description, even within structuralism, language has been treated as heterogeneous and not homogeneous. When Wittgenstein says that language is a toolbox, that words are tools which like any tools can be used in so many different ways, that there is nothing in common to all these different ways of using wards as tools, he adopts a conception of language as heterogeneous and not as homogeneous. In that sense we might say that Wittgenstein can help linguists to break with the hypothesis of the structural homogeneity of language and, so, to put their theory in agreement with their practice. But the insistence of Wittgenstein upon the heterogeneity of language has been developed for philosophical ends, not in view of linguistic description and theory. In fact, Wittgenstein does not give any key for that description or theory. What he can bring to linguists is essentially critical and not constructive. He insists upon a problem which has always worried linguists without giving them the means for solving it. We nevertheless should not underestimate the importance of Wittgenstein's critic of the assumption of the structural homogeneity of language.

As already said, Milner mentions Wittgenstein once more in his book. And this concerns another crucial point, a critic of the use by Chomsky and the Cambridge School, as well as by the cognitivists, of the notion of "unknown rule". What Wittgenstein has written about what it is to follow a rule has been used to criticize Chomsky. Milner does not reject those critics. He just draws from them conclusions which are the opposites of those which have been made (See for instance, G.P. Baker and P.M.S. Hacker, Language, Sense and Nonsense, Oxford, Blackwell, 1984). He recognizes that the way Chomsky uses the notion of "unknown rule" is very problematic. But instead of saying, as it has been the case, that this should led us to reject any of the theories of language of the kind proposed by Chomsky, Dummett, Davidson or others, he proposes to drop the notion of a rule, at least within linguistic theory. Following up to a certain point Searle, Milner says that what we are used to name rules within linguistics (rules of grammar, rules of transformation, or the like), should be and can be reconsidered in terms of laws. Moreover, he says that it is because he uses the notion of "unknown rule" that Chomsky has been led to look in the direction of psychology and, beyond psychology, in that of biology. The problem for Chomsky was that of giving a ground for the notion of "unknown rule", even if this ground has no incidence upon the way he concretely accounts for such or such property of language. The concrete identification of those properties could not but remain based on purely linguistic considerations, biology or psychology being in fact of no help. On the other hand, Milner contests that what Wittgenstein says about rules can give a ground for using such a notion within linguistics. If we are faithful to Wittgenstein's conceptions, we cannot separate the notion of rule from that of a game. Then, if we agree upon the fact that dealing with language in terms of language games, as Wittgenstein does, is in contradiction with the minimal requirements for any possible linguistics, we are led to conclude that what Wittgenstein can bring to linguists is to push them to give up their attempts to deal of language in terms of rules (tacit or explicit). There too, we can see that this does not tell to linguists what to do, but at least what not to do. And that might be priceless.

Refbacks

• There are currently no refbacks.