The Essence of Language: Wittgenstein’s Builders and Bühler’s Bricks Kevin Mulligan A BSTRACT . - What is essential to language? Two thinkers active in Vienna in the 1930’s, Karl Bühler and Ludwig Wittgenstein, gave apparently incompatible answers to this question. I compare what Wittgenstein says about language and reference at the beginning of his Philosophical Investigations with some aspects of the descriptive analysis of language worked out by Bühler between 1907 and 1934, a systematic development of the philosophies of mind and language of such heirs of Brentano as Martinak, Marty, Meinong, Landgrebe and Husserl. R ÉSUMÉ . - Y a-t-il quelque chose qui est essentiel au langage? Deux penseurs actifs à Vienne dans les années trente, Karl Bühler et Ludwig Wittgenstein, donnent à cette question des réponses qui sont apparemment incompatibles. Je compare ce que Wittgenstein dit du langage et de la référence au début de ses Investigations Philosophiques avec quelques aspects de l’analyse descriptive élaborée par Bühler entre 1907 et 1934, un développement systématique des philosophies de l’esprit et du langage des héritiers de Brentano tels que Martinak, Marty, Meinong, Landgrebe et Husserl. 1. The Essence of Language Wittgenstein mentions at § 65 of the Investigations “the great question that lies behind all these considerations”. For he has not so far directly faced up to the question: what is essential to language? His direct answer to this question is that what we call language is a range of phenomena sharing no one thing in common but akin to one another in many different ways (cf. § 92). He appears, he says, to be skipping the question of the general form of propositions. At first glance, this answer contrasts sharply with an earlier Austrian account of language, the empirico-teleological view, due to Franz Brentano, Anton Marty, Eduard Martinak and the South German psychologist, linguist and philosopher Karl Bühler, Schlick’s colleague at the University of Vienna. Bühler thinks human languages are definitely characterised by his four “axioms” - see § 2 below - “… the semantic relations constitute the object ‘language’” (ST 58; cf. especially 141); “language is what fulfills the four axioms” (ST x). They display “the essential similarity of structure” of all human languages (iv), “the structure of human language in the singular” (ST 141). That is why Bühler is happy to speak of “the essence of language” (III, v; elsewhere he notes that his axioms are perhaps only necessary conditions for something to count as a natural language). The empirico-teleological view of language contrasts with yet another K O D I K A S / C O D E Ars Semeiotica Volume 28 (2005) No. 1-2 Gunter Narr Verlag Tübingen Kevin Mulligan 72 Austrian tradition, inaugurated by Bolzano and developed by Husserl in his Logical Investigations, which takes the specification of the formal structure, grammatical and semantic, of propositions to be prior to any account of language in terms of intentions and use. In what follows, I shall argue that all the examples brought forward by Wittgenstein in § 1-21 to cast doubt on the claim that language has an essence are explicitly provided for in Bühler’s analysis. If this is true then it provides one apparent reason for thinking that Wittgenstein fails to sustain his negative thesis. In § 5 I mention some other instances where, in spite of deep similarities between their analyses, Bühler and Wittgenstein arrive at apparently different conclusions. This leads us to the question whether Wittgenstein and his Austrian predecessors understood the very idea of describing language in the same way. In my brief sketch of Bühler’s account of language I provide only what is necessary in order to understand the very limited comparison between Bühler and Wittgenstein to which this paper is devoted. 2. The empirico-teleological account of language and Bühler’s analysis Fortunately, the initial assumptions of Bühler and Wittgenstein are sufficiently similar to make such a critical comparison worthwhile. The empirico-teleological view consists of a claim about the way language develops and is acquired which yields a number of familiar claims about the way it works. Language, it is claimed, develops by complex extensions of the uses of signs which are to all intents and purposes tools. These extensions are due to the intentions of individual speakers. But the complex processes of adaptation and selection governing meaning change and the production of spontaneous linguistic order are not the sort of thing of which an individual could have a bird’s eye view. From this genetic claim Marty and Bühler draw the conclusion that many words are family resemblance terms and that language contains misleading or fictitious “pictures”. From the claim that words are tools they infer the falsity of that variety of Platonism about meanings which makes the latter out to be language-independent. The nominalist and naturalist strands in their view of language lead to an analysis that stands in much the same relation to Husserl’s views as Wittgenstein’s views stand to those of Frege. Bühler indeed provides us with an account of language as a biological and social phenomenon: language is based on blind training (Dressur), instinct and two types of behaviour common to human and non-human animals - expression and steering or signaling. Actions, from the most primitive extensions of drives to reading and ordering, are inseparable from dynamic perception (GEK 317). Thus all interpretation is based on and often in accordance with behaviour, in particular expression and steering: “the function of representation” developed “from something more primitive” (GEK 299). Language develops by adding new levers or joints. Language is a “tool for orientation in community life” (ST 48). But, as a good Austrian, Bühler thinks that biologism and sociologism, and indeed any purely causal theory of mind and language, fail to take into account the pervasiveness of internal relations (Strukturgesetze). Four axioms or platitudes, we said, together with a variety of further specifications, are supposed to give the essence of language. These axioms are: A. There are exactly three basic linguistic functions, expression, representation and steering or appeal. The Essence of Language: Wittgenstein’s Builders and Bühler’s Bricks 73 B. There are exactly three basic (material) types of linguistic entity: symptoms (criteria, indicators), symbols and signals. C. Two levels of formalisation of both (A) and (B) must be distinguished: Sprechhandlung and Sprachwerk, Sprechakt and Sprachgebilde. The former are subjektsbezogen, or bound to subjects, the latter subjektentbunden, free or independent of subjects, but not of linguistic conventions. A linguistic act is a type or species of which linguistic actions are tokens. The result of such an action, a Sprachwerk, is a product (as Twardowski had pointed out). A linguistic structure or pattern or paradigm is a type of which products are tokens. D. The two basic sorts of Gebilde (structure) are words and fields, which are mutually dependent, that is, internally related. There are verbal and non-verbal fields. Verbal fields are sentential, non-verbal fields are deictic, or behavioural or physical (ST 2-6). By “expression” or “utterance” Bühler means the function of indicating mental and vital states, by “appeal” or “steering” that of influencing and guiding the behaviour of an interlocutor or other type of creature. Expression and steering occur in any linguistic interaction but they may also cooperate in non-linguistic interactions, as when an apprentice hands something to his master or foreman whilst both are engaged in a common task and share a visual field (Bühler, 1927, 40; cf. 90). Strangely enough, Bühler, like Wittgenstein, seems to have been almost completely blind to the variety and structure of the linguistic episodes first described by Reinach and now called speech acts (promising, ordering). What does he understand by “representation”? Bühler assumes that the primary function of names and sentences is to represent [darstellen] things and states of affairs. They stand in an ideal or ideational (ideell) relation to things or processes and to the states of affairs containing things and processes. This is the relation of assignation or coordination (Zuordnung), a mathematical term used in similar ways by Meinong, Martinak and Carnap as well as Wittgenstein. He also distinguishes within the class of names between two fundamental word kinds, names of things and names of processes and activities. As he points out, the case systems and prepositional systems of Indo-European languages, have emerged to make possible the description of activities by subjects involving objects of different kinds. Another kind of sentence, impersonal sentences, describes yet a different kind of episode. Since Hobbes, Locke and Hume, Bühler (1909, 105) points out, it has been usual to analyse the processes in the speaker and his interlocutor according to the model of associations of ideas. “The old view was based essentially on two assumptions that were internally connected. It was believed that the functions of language could all be traced back to the naming function of words: every word is a name for something, its Bedeutung, a view most clearly formulated by Hobbes. And it was thought that the sentence contains essentially of an aggregate [Inbegriff] of names. And in accordance with this first assumption the processes of language learning were made out to be a learning to name objects. Both claims are false; the function of naming is only one of several functions of words and the fact that language learning is not based only on acquisition of the naming function is being shown more and more by systematic observation of children. Matters are essentially more complicated than they seemed to the first simple theory; just how complicated they are cannot be somehow deductively inferred but must be grasped on the basis of systematic observation of concrete cases of linguistic comprehension” (Bühler, 1909, 107). Kevin Mulligan 74 A wider and even more fundamental distinction between word kinds is that between words that name without the help of perception and those, such as indexical signals, that require perceptual help. That all language signs must be symbols of the same kind is: “an axiom that is too narrow for the theory of language; for some of them, such as the deictic words, prove to be signals. But the same work must not be demanded of a signal as of a (pure) symbol because there is a sematological difference between the two” (ST 107). Like Wittgenstein, Bühler criticises unnamed logicians who confuse indexicals and proper names. They are doubtless thinking of Russell and Carnap respectively. The relation of coordination is to be distinguished from those of meaning and association. Coordination is a relation between words and things “within the province of the conventions of the linguistic community for which the lexicon was made, in which ‘one’ uses the word”. Association is a psycho-physical connexion within the province of the speech dispositions of an individual between the acoustic image of a word and the image of the corresponding thing. Meaning is a mental episode but is characterised in terms of coordination. “The object named by a name is intentionally aimed at and also more or less intentionally reached in concrete speech-experiences; this is the case every time a member of a linguistic community himself uses the name meaningfully and correctly as a sender, or correctly understands it as the receiver of a verbal message in which it is used” (ST 164). Although all three relations “belong together” they are not equivalent and “terrible confusion will inevitably be the result” of failing to distinguish them (164). Meaning [meinen] is not a relation of association (58-9). Bühler likes to use such Husserlian expressions as “meaning in one pulse” “with one blow” (mit einem Schlage ST 365, 292). The external criterion for the fact that a speaker has produced one sentence, meant something in one blow, is a certain type of intonation. Both acts of meaning and linguistic actions are to be distinguished from types of linguistic products, which are ideal entities with a history - for example the “genidentity” of the German word “Vater”, which is one word for the linguist (62). Lexical and formal products - such as “the infinitive” - constrain acts of meaning and understanding. Axiom (D) distinguished verbal and non verbal fields. Since there are two basic types of non-verbal fields, we have three fundamental types of context or Umfeld (surrounding field) for symbols. fields or contexts consisting of other words the synsemantic field fields consisting of behaviour the sympractical field fields consisting of physical objects the symphysical field Examples of sympractical (Bühler also uses “empractical”, cf. ST 159) naming or pointing using isolated linguistic signs are the typical utterances of the customer in a café who says “black” to the waiter, or of the tram passenger who says “one way” to the conductor (155), as well as the interjections used to address others such as “Hey! ” or “Hello! ” (300). Demonstratives are used sympractically since they combine with gestures to function as signals (31). A sympractical use of a sign belongs to a surrounding field which is a “praxis”, it is built into a practise (158-159). Names can be physically attached to what they name, as with brand names, place-names, book titles, names inscribed on memorials. The place-names on signposts naming the places The Essence of Language: Wittgenstein’s Builders and Bühler’s Bricks 75 these point to are “attachments at a distance” (159). Yet another variation on the simplest case is provided by the relation between the names of owners or makers on their property or artefacts and the relevant people. Real attachment is common to all symphysical uses of names. The different types of “suppositio” make use of the different types of field in which signs can occur (312). Thus mention of a word involves a symphysical field - the word mentioned is a physical part of the singular term; it involves a sympractical field - the quotation marks function indexically; and of course, mention of a word will often occur in the context of a synsemantic field. What sort of semantic relations do symphysical and sympractical signs stand in? Interjections such as “Hello! ” occur neither in symbolic fields like names nor can they be counted among the deictic words “without reservation” … “It would also probably not be wrong to rank them in the one-class system of human and animal calls, thus separating them more thoroughly from genuine words” which belong to the two-class system consisting of words and sentences (300). Of other sympractical signs Bühler writes: “On unbiased examination it seems to be quite unimportant whether such words are deictic particles or have a naming function. The passenger in the tram car can, if he pleases, make it clear what he wants by pointing to one of the two books of tickets in the conductor’s hand instead of saying ‘transfer’. Otherwise, the particle ‘geradeaus’ [straight on, hence simple ticket], which might be interpreted as an ‘adverb’ (or perhaps not), will be on a par with the verb ‘umsteigen’ [transfer, hence correspondence ticket]. By the same token, it seems that the accusative ‘einen schwarzen’ [black] is on a par with the nominative; sometimes it is enough to nod the head or say ‘yes’ when the other person begins of his own accord and with an inquisitive demeanour to take the appropriate action, or one says ‘the other one today’ if that is called for. Naming words remain what they are even in such usage, they name something” (156). What is essential to all empractical uses of language emerges, according to Bühler, if we reflect on what and why it is sometimes necessary to introduce language into non-verbal transactions: “In Vienna the passenger used to be spared saying ‘transfer’ for there was only one type of ticket. Whoever saw the familiar transaction of buying a ticket taking place between silent partners without a hitch, knew too the borderline case in terms of which most so called ‘elliptical utterances’ have to be understood: islands of language emerge from within the sea of silent but unequivocal communication at the places where a differentiation, a diacrisis, a decision between several possibilities has to be made, and easily can be made by interspersing a word. They emerge and are welcome just as names and arrows are welcome on signs at the crossings of paths that one takes” (ST 155-156). Closely related to the symphysical uses just considered is the case of baptism. “If I have a name such as Charles or Maria solemnly conferred upon a child at baptism, this is a convention that those participating in the ceremony and later others who are informed by the participants observe. In smaller circles this first name suffices as an individual sign … Regarded in terms of the regulations on coordination are these and other proper names on a par with ‘class names’? With Mill, I say decidedly not. For this coordination at baptism is by no manner or means equivalent to a definition, but rather, seen from a distance, equivalent to making the chalk mark on a house. The fact that the individual sign, the proper name, is not branded onto the forehead of the newly baptised child makes no difference to our question. The people involved make a note of it on their own and are able (with the passing of the years better and better) to discern the person named as an individual distinct from others. This individual already exists and has been brought to be baptized; it has no need of a definition. Indeed, baptism is not a Kevin Mulligan 76 definition but (here it is tempting to continue: a sacrament) - it is rather an assignment analogous to attaching it; it is a deictic name conferral. Proper names are distributed deictically; it is not exactly the symphysical surrounding field but something analogous that becomes relevant here” (235-236; cf. 226). Thus if Bühler is right, the context principles of Frege and Husserl and of much contemporary linguistics are too narrow: words do require a context in order to signify, but the context need not be a sentential context. A space or field and its attendant internal relations are necessary, but physical and behavioural fields must be allowed in addition to sentential fields. Hitherto we have mentioned cases where linguistic signs form unities with non-linguistic elements which are not themselves signs. There are also non-linguistic signs. The distinction between synsemantic and symphysical fields has wholly non-linguistic analogues. Colourcontrast can be read off from a symphysical field of coloured patches. “But the case of the ‘context’ of the pictorial values in a painting as a whole is substantially different.” One and the same grey pigment on the palette can be used to represent a shadow, a light-reflection and the colour of an object. “The context of the pictorial values in a painting is analogous to the context of linguistic signs; in both there is a synsemantic surrounding field” (165). Nonlinguistic symbols require fields just as linguistic symbols do. Examples of non-linguistic fields are the music-paper used by musicians, maps, pictorial fields, the stage which is the field of the actor and graphical representations of curves (179-185). Examples of similar fields inside language are the system of action, agent and patient expressed by word-order or cases in Indo-European languages (195, § 15), the system of events, places and times expressed by impersonal sentences (§ 25.4), the system of internal relations expressed by verbs and internal accusatives (§ 15.4), the system of nominal sentences and conjugations (195). Bühler’s descriptions of extra-linguistic representational tools are introduced in order to throw light on the structures of representational uses of language. As he points out, it is curious that this sort of comparison is so infrequent (§ 12). The alphabet, the written decimal system of numerals and the spoken numerals, for example, are intermediaries, “medial implements” which Bühler calls files or organizers, “they are called intermediaries because they are placed between things, and organizers because what they do is comparable to what material ordering implements such as the letterfile, which keeps our letters in order, catalogues and the like do” (193). They are “indirect, mediated coordinations” (192). Files play the same role as what Marty had called inner linguistic forms, but unlike these are always replaceable by external counterparts. They are indirect, mediated coordinations, in contrast to the direct coordinations considered above. Bühler gives the following simple example. Consider two ways of assigning letters to the corners of a polygon, arbitrarily or by following the order of the alphabet from A through to F. The second solution, unlike the first, introduces an element of mapping and this projection of the corners of the polygon on to the associational series of the letters of the alphabet provides a number of advantages. Thus “even when the object has been removed from the sphere of actual perception it is still possible to say much about it that can be immediately verificd with the associational series alone” (ST 192-193). The blind orders of the alphabet or the numerals are learnt by association. But once the convention that, for example, certain signs should have the import of one, tens, hundreds and so on from right to left, has been learnt, “structural insights become possible and can be made use of when dealing with the numerals, insights that could not be drawn from a blind associa- The Essence of Language: Wittgenstein’s Builders and Bühler’s Bricks 77 tional chain as such” (ST 194). Bühler attaches great importance to the ways inner linguistic forms of the sort just mentioned function. “It would”, he thinks, be easy enough to show “that within spoken language as a system of signs many associational chains and networks can be encountered which, seen from a psychological perspective, are on the same level as the alphabet chain, and which do us similar service in the extensive and comprehensive problem of the ordering of our knowledge of things and imparting this knowledge to others …” (ST 193). Symbolic fields, whether linguistic or not, are characterised by abstractive relevance. Their properties are due to conventions and tradition and are not physical properties. To fail to see this point is to be guilty of a stoffliche Entgleisung, of materially going off the rails: “A blank piece of paper in front of me is no field. By the same token, the raw succession in the sound stream of human speech is not yet a field; rather, something additional is needed in the succession of sounds, something that corresponds to the net of geographical coordination lines or the stave of five parallel lines on music-paper, in order to gain a field or fields from the temporal order” (ST 181). Failure to take into account the conventionally fixed, sematological internal relations leads either to the error that consists in overlooking all but the physical properties of signs or to a “magical” view of language which attributes mysterious properties to it. The two errors are combined in accounts of language that reduce the phenomena of abstractive relevance and internal relations to mysterious causal relations (cf. ST 46-47, Wittgenstein’s 1930-1931 manuscript TS 213, quoted by Hilmy 1987, 108). Yet another error would be the assumption that the internal relations necessary to language require Platonism. Internal relations emerge from and depend on contingent facts. For Bühler as for Husserl all sentences about objects in time are in one way or another deictic (ST 373). But the fact that language is a system consisting of two types of field, symbolic and deictic fields, with the former dependent on the latter, is, on Bühler’s view, itself rooted in the role of deixis in learning. He (ST 385) calls deixis in the ordinary use of language “object deixis” and its role in learning “learning deixis” (Lerndeixis, acquisitional deixis). Every speaker “has gathered the meaning [Bedeutung] of all naming words from things and states of affairs pointed out directly or indirectly and then retained it by practice” (383). The mechanism is the same when we are confronted with new symbols in logic or science: “‘Look at this: we use this sign written on the blackboard, on the page of the book before your eyes for this or that.’ That is how one goes about granting meaning to all symbols, and without these deictic clues it would in fact be impossible to give intersubjective currency to any symbolic system” (Bühler, 1934, 383). Granting or giving meaning, then, is not for Bühler, as it was for Husserl, something brought about by acts of meaning. It involves correlations between marks or sounds and perceptible, public objects. Since “acquisitional deixis continues to have effect always in our understanding of all sentences” (ST 385), whatever non-linguistic items are necessary to the acquisition of language, exemplars, colour patterns or colour tables, belong to language. Bühler’s “two-field theory claims that the several modes of perceptual pointing and presenting are just as much a part of the essence of natural language as are abstraction and the conceptual grasp of the world, and that they are equally close to the essence of language” (ST v). Kevin Mulligan 78 This is a claim that holds of both the way in which language is used and of the way it is learnt. Every linguistic act or action is embedded in an act-history which in turn is embedded in the speaker’s history which includes his learning history, the history of his acquisition of linguistic skills (Bühler, 1933, 51-52). Bühler’s account of the role of perception in making coordination possible is a revised version of Husserl’s realist and verificationist account of meaning and intentionality. But Bühler pushes such an account in the direction of nominalism and naturalism. In particular, there are three main twists that Bühler gives to earlier Austrian accounts of language. First, consider Husserl’s claim that singular reference depends on perception and that predication depends on the possibility of perceptual justification. Bühler clearly agrees with Husserl about the first claim although it is not clear whether he accepts the second part of Husserl’s (realistic) verificationism. On Husserl’s view the senses of expressions depend on their relations to perception. But on Bühler’s view expressions do not have senses. They are used meaningfully to the extent that they are internally related to perception and so to behaviour. Secondly, Bühler places linguistic rule-following firmly in the context of perception, behaviour and instinct: “all concrete speech is in vital union [im Lebensverbande] with the rest of a person’s meaningful behaviour; it is among actions and is itself an action” (ST 52). Natural perception and natural language are, for Bühler as for Scheler, the two central components of the one and only human form of life, the natural attitude of common sense. Finally, Bühler insists, against Husserl and Scheler, that the representative function of language is a tool-like function. But this claim does not lead him to say that language has a goal external to it, as does the activity of cooking. Marty’s account of language assimilates it to cooking: a speaker, he thinks, aims to modify the cognitive states of an interlocutor. To the extent that language is action, Bühler insists, this action must be conceived of as praxis in Aristotle’s sense. It is only to the extent that language use involves bringing about products and their types that the language of goals becomes appropriate (ST 52-53). The “semantic devices or relations” - expression, steering and representation - help constitute the life of human communities (KP 39). A great deal turns on the distinction between the proper functions of language and the goals of language use (cf. KP 123). If a type of linguistic use has no external goal, then its normative dimension is that of a categorical norm; if it has an external goal then its normative dimension is that of a hypothetical norm. This distinction emerges already in the crucial role of games in the acquisition of language and in other types of learning. If language is, to begin with, the result of an interaction between training and games (KP 208f.), the first such games are “Handlungsspiele” followed by “Werkspiele” (ST 53, GEK 220f, 467f) 3. Wittgenstein’s examples and distinctions As we shall now see, the opening of the Investigations illustrates the analysis of language just presented. Many readers of the Investigations are struck by the masterly way in which § 1 adumbrates the themes of Part I. But some are disturbed by the apparent disconnectedness of the examples and distinctions that multiply in quick succession. Someone who has followed my The Essence of Language: Wittgenstein’s Builders and Bühler’s Bricks 79 sketch of Bühler’s analysis of language has a good chance of experiencing a Gestalt switch - what Bühler called an “Aha-Erlebnis” - as he goes through the opening paragraphs of the Investigations. He may come to see a house where before he had seen only bricks, or a wood where before he had seen only trees. The Investigations begins with a quotation from Augustine in which a number of concepts such as “calling”, “indication”, “pointing”, “meaning to point out” and “understanding” are employed. Wittgenstein introduces three distinctions and a use of language. The first distinction is between learning and using language. The second distinction is that between a certain “picture” of language and a certain “idea”. The third is between kinds of words. The use of language is the first of many concrete descriptions of the way language works. According to the picture, words and objects stand in the naming relation and sentences are combinations of words. This picture is the root of the following idea: all words stand in the relation of coordination [Zuordnung] to, of standing for, meanings [Bedeutungen]. Later Wittgenstein goes on to introduce the related “semantic” terms “bedeuten, bezeichnen, benennen” and “sich beziehen auf”. In the detailed descriptions that follow of how words are and can be used we are presented with cases for which the picture is appropriate and it is suggested that the idea of language is an extreme oversimplification. The relation of coordination is later distinguished from two other relations. There is the associative connexion (§ 6), which is sometimes also presented by Wittgenstein as involving memory (§ 53; cf. 256, 508, 271). A third relation is mentioned at § 19: the relation of meaning [meinen] a thing with or without the help of a sign (§ 20). The counterpart of this is understanding a word “in one blow”. What is grasped in such cases is not the temporally extended “use” of a word (§ 138; cf. 139, 191, 197). The final distinction introduced in § 1 is between names of things and names of processes and activities and yet other types of words. Later, in § 8, Wittgenstein introduces indexical expressions which are yet later distinguished from names (PI § 38), and in § 15 from proper names. Functions and signs in context Consider the use of language described in § 1. Here a shopper takes a slip marked “five red apples” to the shopkeeper who opens the drawer marked “apples”, looks up “red” in a table and finds a colour sample opposite it and counts up to “five”, finally producing five red apples from a drawer. This use of language illustrates the three categories of organisers, sympractical and symphysical uses of signs. The signs “five, red apples” combine with a series of actions which form a field for them. The word “apples” attached to the drawer is an example of a word belonging to a symphysical field. Both the table containing “red” and colour samples as well as the series of cardinal numbers are organisers. One plausible aspect of organisation of the first sections emerges immediately. Having introduced examples of organisers, sympractical and symphysical signs, Wittgenstein goes on to consider these types of use singly and in different combinations. A second use of language (§ 2) is a sympractical use of language only. The builder, A, calls out “block”, “pillar”, “slab”, or “beam” and the assistant, B, “brings the stone which he has learnt to bring at such and such a call” (cf. Innis, 1988, 80; BB 77F; Waismann, 1965, 198f). Kevin Mulligan 80 A further use of language at 8 adds two organisers and two deictic expressions to the language of the builders. The deictic expressions are “there” and “this”. The first organiser is the series of letters of the alphabet, the second a number of colour samples. In a variant of the builder’s language-game, Wittgenstein mentions a symphysical use of language in which proper names are attached to objects, “the object signified is marked with the sign. Suppose that the tools A uses in building bear certain marks. When A shows his assistant such a mark, he brings the tool that has that mark on it” (PI § 15; that these marks are proper names is indicated at § 41; cf. Waismann, 1965, 198f). Wittgenstein says that: “The word ‘to signify’ [bezeichnen, designate], is perhaps used in the most straightforward way when the object signified is marked with the sign. […] It is in this and more or less similar ways that a name signifies and is given to a thing. It will often prove useful in philosophy to say to ourselves: naming something is like attaching a label to a thing” (PI § 15, cf. 26). Whereas Wittgenstein suggests that naming may consist in attaching a label to a thing, Bühler claims that such an attachment is a criterion for the obtaining of the naming relation: “When the optical name image is materially attached or connected to the perceptible thing named, that is, when there is a connection that under such circumstances must be interpreted in terms of an effective surrounding field, this visible attachment becomes an indication (Indizium) of an ideal coordination” (ST 164). The existence of a criterial relationship between physical contiguity and naming implies that the two relations are distinct. And this is indeed the case: “Real attachment is not sufficient to make a word a name of its bearer. After all print is physically attached to the page but is not a symphysical sign. The relation between the paper in books to the black figures on it is quite different from that of the product to the product name printed on it, and in general the relationship of every bearer to the language sign it bears when it displays it as its name or the like. In such a case attachment becomes the physical, sensibly manifest criterion of coordination” (ST 162). At § 37, Wittgenstein tells us that the naming relation may consist of a number of different things. He mentions three possibilities. The first has already been mentioned: it may consist in what Bühler calls symphysical naming. It may also consist in sympractical naming or even in association (cf. GEK 230). Tools and functions The functions of words in these examples are as varied as the functions of tools (§ 11-14). They are bound up with the activities of sending someone shopping, taking a slip to the shopkeeper (§ 1), communication, building, passing stones (§ 2), pointing (§ 8), orders (§ 2, § 8) and Wittgenstein mentions languages that might be imagined to consist only of orders and reports in battle or questions and yes/ no answers (§ 19). Although Wittgenstein cautions against such simplification, we might say that in all these cases signs are used as signals to steer an interlocutor. More controversially, we might say that the users of these signs indicate or express something. But it is a feature of Wittgenstein’s descriptions that they contain no reference to psychological states that are so expressed. The Essence of Language: Wittgenstein’s Builders and Bühler’s Bricks 81 At § 3 and § 17, Wittgenstein mentions an analogy which will become more and more important, between words and chess-men, between the description of language and the description of games. Wittgenstein likes to compare words to chess figures. Their properties are not only physical properties (PI § 108). They and the fields they belong to are to be characterised in terms of conventional properties. The use of a word or of the King in chess can only be explained to someone who already knows its “Platz” or position (PI § 31). The contrasts that, according to Bühler, are essential to such fields, to their different Plätze or field-values, and the values of elements in them, are often mentioned by Wittgenstein who nevertheless stresses the danger of supposing the psychological reality of possible contrasts (PI § 20). Fields and words in fields are not mysterious mental or Platonic entities. But their properties are not only physical properties. Similarly, when we describe the rules governing the uses of pieces in chess we are not describing the physical properties of the pieces (PI § 108, 31). The physical movement of a piece in chess is a move in chess only in the circumstances we call ‘playing a game of chess’ (PI § 33). To confuse physical and conventional properties is to adopt an analogue of what Bühler calls the phonetic approach to phonology. Non-verbal parts of language Wittgenstein says it is most natural and causes least confusion to reckon the colour-samples among the instruments of language (PI § 16). For if one says: Pronounce the word “the”, the second “the” counts as part of the sentence. The role it plays resembles that of the coloursample. This type of suppositio, mention, like the use of a colour-sample, makes use of a symphysical field. The field in question consists of words. There are many different sorts of patterns which, in non-linguistic fields, nevertheless belong to language, although not to the word language (PI § 50, 56, 72, 73). Colour tables and other organizers are the expressions of rules (PI § 53). Wittgenstein’s brief remarks here have a long history in his own thought (just as Bühler’s descriptions of the role of perception in language have their origins in Husserl’s account of the way simple linguistic sense is completed by perceptual content in indexical sentences). In “The Thought” Frege formulated an account of indexicals which, while it contains no explicit account of their sense and content, does contain an account of the representational media involved and of their Bedeutung. On this account an indexical singular term consists not only of a demonstrative pronoun but also of the accompanying circumstances. In other words, a deictic expression is not wholly linguistic and it refers to its non-linguistic part. Wittgenstein returns again and again to this distinction of Frege’s which he likes to mark by distinguishing between “Symbols” - which are not purely verbal - and “signs”. Bühler and Wittgenstein often describe many different types of deictic expression. The most basic type, nominal deixis, where “this” for example is completed by a perception of an object, exhibits many subtypes. There is also “so” deixis as when the reference is to a pattern exemplified by a proffered piece of behaviour (“Play it thus”). And in addition to the egocentric deictic field, the “origin” of which is the speaker, there are topomnestic fields in which a familiar object takes over the role of “origin”. Finally, deictic expressions can be replaced by names - for which Bühler coins the term “prodemonstratives” - although of course, this does not suffice to eliminate the functions of pointing and perception. Here as elsewhere Bühler finds a system in what Wittgenstein apparently presents as part of the inexhaustible variety of language. Thus nominal deixis exemplifies the category of Kevin Mulligan 82 symphysical naming; “so” deixis steers an interlocutor in a way that goes beyond the steering peculiar to nominal deixis in that the interlocutor must track the piece of proffered behaviour; topomnestic orientation depends on egocentric orientation. It is because of this structure, according to Bühler at least, that the “field values” of one deictic system are so easily translateable into those of another system. Learning and use In a community whose language was just the builder’s language an important part of the training of children will “consist in the teacher’s pointing to the objects, directing the child’s attention to them, and at the same time uttering a word” (PI § 6). Wittgenstein calls the teacher’s activity ostensive teaching of words. It “forms an important part of the training” and “can be said to establish an association between the word and the thing”. The ostensive teaching of words differs from giving ostensive definitions of words in that the latter are given to children capable of asking what the name of an object is whereas the former is provided before the child can ask, of the object pointed to, what its name is (cf. GEK 230, 397). We may perhaps say that ostensive teaching and the associations it establishes are more primitive than ostensive definitions. 4. Curing the ellipsis plague The distance between Bühler and Husserl, on the one hand, and between Wittgenstein and, say, Frege, on the other hand, emerges most clearly in the discussions by Bühler and Wittgenstein of the phenomenon of ellipsis. The label unfortunately prejudges the issue. As Bühler and Wittgenstein stress, many uses of words are easily regarded as elliptic or abbreviated expressions of thoughts and their full sentential expression. But easy though it is to take this step, the temptation should, they think, be resisted. There is, says Bühler, a temptation succumbed to by the “hardened adherent of the general idea of ellipsis” (ST 157) who reads into every isolated occurrence of a word or expression an implicit or unexpressed linguistic environment. Bühler opposes a persistent theoretical temptation, present in many approaches to sympractical and symphysical uses of words, to his own descriptive analysis. “The fact that” naming words in a sympractical field “sometimes march step in step with a variety of linguistic and non-linguistic signs […] can easily tempt the theoretician to give all cases the same summary interpretation. But he should proceed with caution” (156). Bühler admits to having succumbed to the temptation himself: “… that is how I first worked until I realized how arbitrary and forced my completions often were. Sometimes one feels one is behaving like a silly school boy or (perhaps more correctly) like a pedantic schoolmaster when one starts to theorize with sentence completions although naive practice is completely unequivocal” (ST 157). How is the temptation to be dealt with? “[T]he flood of ellipsis will be checked before it swells up if it can be shown that the [following] presupposition is wrong: all meaningfully used words must be situated in a synsemantic field, they must be borne by a [linguistic] context. That alone is the effective and radical cure for the ellipsis plague that has been with us now for two millenia.” (ST 167-168) The Essence of Language: Wittgenstein’s Builders and Bühler’s Bricks 83 The form the temptation takes according to Bühler is that of an overgeneralisation of a genuine phenomenon: “Of course there are ellipses. There are uncompleted buildings (think of medieval cathedrals) and all sorts of other human works that have come to a standstill in the course of realization, amongst them uncompleted utterances” (ST 166). To correctly recognize the facts of the sympractical and symphysical use of sound signs makes it possible for the theoretician to thoroughly depopulate the denizens’ ghetto (Metökendorf) around the sentence’s palace. Those evacuated live in their own right and do not need to be measured against the “full sentence”, that is, against synsemantically integrated and “completed” speech (ST 366). Contact which is sparing in the use of sound “must not be summarily and generally characterized as poor, primitive, incomplete speech. That would be just as false as regarding traffic in goods with little or no money as the expression of a primitive and imperfect economic order. Rather, both admit of much refinement. There is an advanced culture of ‘elliptical’ speech in which the field values of the situation are used to fulfil and make precise the meaning of the phonetic islands” (ST 88). Where there is no linguistic context the theoretician must, Bühler warns, take care to avoid general hypothetical constructions. It may be that “the speaker reproduces a part of a sentence and spares himself and the hearer the rest; that the linguist recognizes from this or that moment of form a determination of the sign’s syntactic position [Platzbestimmtheit]. What are we to make of this? Hardly much more than that the language sign as it was uttered in this case could also have occupied a definite position within a linguistic context and that as a rule it does”. The hardened adherent of the general idea of ellipsis takes these real possibilities to be everywhere true of occurrences of isolated words. But “it would be nothing other than a thorough misjudgement of the psychological conditions if this interpretation were to be regarded as necessary and sufficient for all cases” (ST 156-157). The hardened adherent “might point out that a sentence can always be constructed around the case of empractical naming. The answer is that his point cannot be denied but does not prove anything. An interpreter with a gift for language can indeed provide a more or less appropriate text for every phase of a completely silent communicative exchange; the right arm of the passenger raised holding the money in the tram car ‘says’ to the conductor, ‘Please give me a ticket’. Well of course, and the gesture says it about as unequivocally as the raised front paw of a whining, begging dog says to its master at a meal ‘Please give me a bit, too’” (ST 156-157). Wittgenstein, too, thinks that the question of the type of field or context required by words comes into sharp focus with the problem of “ellipsis”. His discussion of “ellipsis” at PI § 19-20 contains a decisive consideration against the tendency to assume that the use of a single word, such as “Slab! ” in the builder’s game is really an elliptic expression of a propositional thought or of a longer sentence. “Slab! ” “is surely a shortened form of the sentence ‘Bring me a slab! ’ […] - But why should I not on the contrary have called the sentence ‘Bring me a slab! ’ a lengthening of the sentence “Slab! ”? - Because if you shout ‘Slab! ’ you really mean ‘Bring me a slab! ’ - But how do you do this: how do you mean that while you say ‘Slab! ’? Do you say the unshortened sentence to yourself? ” (PI 19). Kevin Mulligan 84 Bühler appeals to the tool analogy to show that it is wrong to assume that we think in sentences and merely express these elliptically. “When the taciturn café customer says ‘black’, he is reproducing a handy chunk from among the linguistic dispositions in his memory, and thus behaves more or less like a practical person who wants to hammer a nail and grasps the next best object that he can get his hands on. It does not have to be a proper hammer, it can be a climbing boot, a pair of pliers, or a brick. In the situation imagined in the café, a choice must be made between a few equally likely drinks, and to this end the naming word ‘black’ or even the isolated preposition ‘without’ is enough” (ST 157). Witggenstein points out that a “sentence is ‘elliptical’, not because it leaves out something that we think when we utter it, but because it is shortened - in comparison with a particular paradigm of our grammar” (PI 20). Such a grammatical paradigm is an example of what Bühler calls a type of linguistic product, a “Gebilde” or structure (cf. PI 108). 5. Conclusion The examples of ways of using words given by Wittgenstein in the opening paragraphs of the Investigations neatly illustrate the taxonomy of language provided by Bühler. In order to determine the exact relation between Bühler’s account of the essence of language and Wittgenstein’s denial that there is any such thing it would of course be necessary to consider the relations between the empirico-teleological account as a whole and what Wittgenstein says elsewhere about language and mind. Thus at least the following questions would need to be considered. Wittgenstein’s opening paragraphs deal with relations between words and referents that are perceptually given. But what is the relation between such cases and cases where we talk about objects that are not so given? Wittgenstein merely implies at 15 that there is a similarity between the two cases. What is the relation between the accounts of Husserl, Bühler and Wittgenstein of the ways in which perception fixes not only the reference of indexicals and demonstratives but also the reference of those proper names whose sense is not simple? Just how much weight attaches to Wittgenstein’s claim that “the meaning (Bedeutung) of a name is sometimes explained by pointing to its bearer” (§ 43)? What is the relation between the accounts given by Bühler and Wittgenstein of the ways in which perception determines the use and meaning of predicates and their analyses of criteria and indicators, as distinct from symptoms (cf. Mulligan, 1990)? What is the relation between Wittgenstein’s views on this matter and verificationism, for example the realist verificationism of Husserl? Bühler and Wittgenstein suggest, as we have seen, that the internal relations between language and perception, particularly between language and exemplars, are brought into being via learning. But, as they both realised, any such account must deal with the problem of the acquisition of superordinates, the tree structures characteristic of most concepts. What is the relation between Wittgenstein’s account of the interrelations between “rule”, “same” and “agreement” (PI § 241-242, § 355, § 429, § 224-225) and the empirico-teleological account of “agreement in judgement”? Martinak (1901, 43) argues (and Bühler agrees) that “agreement with a norm” and the type of “sameness” this involves require judgements, The Essence of Language: Wittgenstein’s Builders and Bühler’s Bricks 85 and above all dispositions thereto, about coordinations between words and, for example, perceptually presented objects. Such judgements are definitions, ostensive and verbal. Were there no such judgements, he argues, mere associative connexions would not yield the stable connexions essential to language. What is the relation between such judgmental definitions, ostensive teaching or acquisitional deixis, and Wittgenstein’s agreements in definitions, in judgements, in opinions, in language and in form of life? A part of the answer to these questions is provided by Bühler’s attempt to understand rule awareness and rule following by bringing Husserl’s theory of categorial intuition down to earth. A question more fundamental than any of these concerns the conception of “essence” that Wittgenstein had in mind in denying that language has one. Perhaps Bühler and Wittgenstein use “essence” in different ways. Wittgenstein seems to have thought that the sense of “essence” in which language does not have one is related to a conception of analysis as something that leads to simples. But as Bühler points out, echoing Husserl, it is important to distinguish analysis as it is practised by the butcher (ST 58) from analysis that makes morphologically correct divisions (ST 153, Bühler, 1933, 36). Perhaps, after all, Wittgenstein did think he could describe the essence of language, that it is only the notion of a hidden essence that he wanted to reject (cf. PI § 92). Descriptive analysis in the Brentanian tradition also aims at making explicit something that is not hidden, not there to be discovered, in the sense in which one can make empirical discoveries, but is rather difficult to get into focus (cf. Mulligan, 1993). But what exactly is the relation between Wittgenstein’s understanding of morphological description and that common amongst his contemporaries in Austria-Hungary? References and abbreviations Bühler, K. 1909: “Zur Kritik der Denkexperimente”, Zeitschrift für Psychologie, p. 108-118. - 1927, Die Krise der Psychologie, Jena, Fischer. - 1930, Die geistige Entwicklung des Kindes, Jena, Fischer (I refer to Bühler, 1930, as GEK). - 1933, “Die Axiomatik der Sprachwissenschaften”, Kant-Studien, 38, p. 19-90. - 1934, Sprachtheorie. Die Darstellungsfunktion der Sprache, Jena, Fischer (reprint 1965, Stuttgart, Fischer; English Translation by D. Goodwin, Theory of Language. The Representational Function of Language, 1990, Amsterdam, Benjamins). I refer to Bühler, 1934, as ST, followed by the German pagination which is indicated in Goodwin’s translation. Eschbach, A. 1984: “Verstehen und Interpretation. Karl Bühlers synchytische Begriffe und Ludwig Wittgensteins Familienähnlichkeiten”, A. Eschbach (ed.), Bühler-Studien, vol. II, Suhrkamp, p. 175-206. Eschbach, A. (ed.), 1988: Karl Bühler’s Theory of Language, Amsterdam, Benjamins. Hilmy, S. 1987: The Later Wittgenstein. The Emergence of a New Philosophical Method, Blackwell. Innis, R. 1988: “The Thread of Subjectivity: Philosophical Remarks on Bühler’s Language Theory”, Eschbach (ed.), 1988, p. 77-106. Martinak, E. 1901: Psychologische Untersuchungen zur Bedeutungslehre, Leipzig, Barth. Mulligan, K. 1990: “Criteria and Indication”, Wittgenstein. Towards a Reevaluation, Proceedings of the Kirchberg Wittgenstein Centenary Celebration, 1989, Vienna, Hölder-Pichler-Tempsky, p. 94-105. - 1993, “Description’s Objects: Austrian Variations”, B. Garrett & K. Mulligan (eds.), Themes from Wittgenstein, Working Papers in Philosophy, 4, RSSS, Philosophy Program, Research School of Social Sciences, Australian National University, Canberra, p. 62-85. Ortner, H. 1983: “Wittgenstein und Bühler. Kritik der cartesianischen Sprachauffassung”, Mitteilungen aus dem Brenner-Archiv, p. 34-49. Toccafondi, F. 1995: I Linguaggi della Psiche.Teoria della Mente, della Percezione e del Comportamento da Würzburg a Vienna, Milan, Guerini. Kevin Mulligan 86 Waismann, Fr. 1965: The Principles of Linguistic Philosophy, London, Macmillan. Wittgenstein, L. 1968: Philosophical Investigations, Oxford, Blackwell (PI). - 1972, The Blue and the Brown Books, Oxford, Blackwell (BB).