sign in sign up
<<
Home , Cognitive semiotics, Semiotics

Elmar Holenstein

Semiotics as a

The explanatory turn in the human sciences as well as in "artifi- cial intelligence" have led to a little discussed revival of the classical Lockean sub-discipline of dealing with mental representations - aka "ideas" - under the heading of "cognitive science". Intelligent be- haviour is best explained by retuming to such obviously semiotic catego- ries as , gmbol, , program, etc. and by investigations of the format of the mental representations. Pictorial and other "sub-linguisric" representations might be more appropriate than linguistic ones.

CORRESPONDENCE: Elmar Holenstein. A-501,3-51-1 Nokendal, Kanazawa-ku Yoko- hama-shl, 236-0057, . EMAIL [email protected]

The intfa-semiotic cognitive tum 's declaration that is a branch of was understood by traditional linguists as an attack on the autonomous status of their science, a status which older endorsed by incorporating (as a special system) into semiotics. Semiotics seemed to assure that genuinely linguistic relationships (grammatical and semantic relationships) were not rashly reduced to psychological or biological relationships (such as associations or adaptations). If it is looked at more closely in conjunction with the general development of the sciences in the 20th Century, the between Chomsky (1972: 28), who regards linguistics as a subfield of psychol- ogy, and (1916: 33), for whom it is primarily a branch of "semiology", comes down to the fact that Saussure assigns linguistics in a descriptive perspective to universal semiotics, whereas Chomsky assigns it in an explanatory perspective to a special semiotics. Chomsky founds linguistics in the sub-discipline of semiotics dealing with mental representations {alias "ideas") which was prominent in classical times and which has since had a revival under the title "cognitive science". But although inquir)' into the cognitive underpinnings of the use of is a matter of course from a classical point of view and in the of the development of the cognitive sciences since the end of the 1950s, in view of the

Cognitive Semiotics, Issue 3 (Fall 2008), pp. 6-19 SEMIOnCS AS ACOGNmVE SCIENCE | 7

prevailing self-image of semiotics in the 20th Century it does seem stränge. This self-image ranged within the Framework of the anti-mentalist attitudes in the human sciences between 1913 and 1959, the Start and end points of behaviour- ism: John B. Watson's Manifesto and Noam Chomsky's lethal criticism of the movement. The fact that there is no action without envisaging a goal, no complex action without a plan and thus no without a cognidve science had been ignored.

The philosophical explosiveness of cognitive science The best way to introduce cognitive science and its semiotic nature is based on a common distinction in . A distinction is made here between linguistic philosoply and . Linguistic philosophy is described in terms of (or research Stratege), the philosophy of language in terms of Contents. Philosophy of language addresses the structures and Functions of language regarded as philosophically relevant - with a more or less continuous transition to linguistics. It is a branch of philosophy like others, along with philosophy of nature, the object of which is nature, philosophy of history, the object of which is history, philosophy of art, the object of which is art, and so on. For linguistic philosophy, by contrast, language is primarily the Instrument and not the object of research. Its object is rather the centuries-old problems of philosophy. Linguistic philosophy attempts to solve these problems by analysing the language used to formulate them or the language in which the problems of philosophy become apparent; the language, as some of the advocates of this kind of philosophy believe, which causes these problems in the first place. Cognitive science and the science of can be analogously differen- tiated. Just as according to linguistic philosophy we must draw on the pertinent linguistic categories to solve the traditional problems of philosophy, according to cognitive science we must draw on cognitive categories, especially "representa- tion" and "computation", in order to analyse intelligent Performances. Thus, cognitive science, inasmuch as it is defined in such methodological terms — with a bold ontological commitment - stands in contrast to the science of Cognition, which is defined in substantive terms and has the various cognitive phenomena (perception, thought, memory and so on) as its subject matter. "Cognitive science" thus stands for a research strategy that assumes that processes that are most aptly described and explained with cognitive categories can be discemed both in human beings and animals, and - a revolutionary view 8 E. HOLfNSTBN

- even in machines. This is the explosive aspect of cognirive sciencc for philosophy. The gulf between animal and human beings that in the past loomed wide from the perspective of the faculty of language is thus riarrowed, at least to a certain extent. The continuity between animal and human being in the deeper layers of Cognition is stronger than in the higher, specifically linguistic levels. The connection between human being and machine is of even greater moment. It is of ontological and methodological significance. A cognitive now no longer makes a claim only to be theoretically coherent and empirically adequate, but also to be technicalJy reati^ble. A Computer that is programmed according to such a theory should exhibit the same Output as a human being or an animal if given the same input. A theory that is susceptible to formal operationalization is thus in principle also susceptible to technical operationalization. It has become operationalizable in a very tangible sense. There is now a relendess test for formalisms in the human sciences, which are often, and not without reason, suspected of being sterile bluffs. The machine that is bullt according to them either works, or it does not work. That takes care of the bluff. As far as the fruidessness of formalizations is con- cemed, the Situation is less clear. It is becoming increasingly apparent that even though certain cognitive adequately and precisely describe the intelligent Operations of machines, the limitations of the products of that have been realized up to now are also a sign of the fact that machines think, calcuJate or, to put it in more neutral terms, operate differendy from human beings. This insight has led to a reappraisal of the non-linguistic, non-rational as well as of the sub-cognitive layers of human epistemic Operations, of mental images and models, and increasingly of intuitions and feelings. Artificial intelligence originally adopted the traditional intellectualist claim that inttiitions and perceptions in general can be reduced to unconscious inferences or computations of the same type as we use for our explicit, rational reflections. This claim has been shaken. Cognitive science is comparable to intellectualistic psychology of the late 19th Century as far as its claims are concerned. It soon turned out that this also applies to its fate. Like intellectualist psychology, cognitive psychology is remarkably successful in a narrow field. What it explains so successfuUy is regarded by its critics as merely the tip of the "iceberg" of psychological processes. It had been rashly assumed that the remainder, which is actually more extensive, could be explained according to the same model. Just as in the SEMI0T1CS AS ACOGNmVE SCIENCE | 9 early decades of the 20th Century Gestalt psychology and phenomenology emerged in a move away from an overly intellectualistic psychology, the last decades of the Century brought a reaction against the excessively rationalist thought of early cognitive science. It cannot be excluded that the catchword "cognitive science" will again disappear as a slogan because of exaggerated Claims to universal pertinence just as "cybemetics" did in its time - despite enduring, important successes in many fields. Just as the category of goal- directedness (control by means of negative feedback) is not sufficient, so too are categories such as "representation" and "computation" insufficient to explain all psychological phenomena. The remarks made in this section can be summarized as foUpws: Cognitive science can be understood as an umbrella discipline embracing the study of the natural intelligence of biological organisms and the arrificial intelligence of man- made machines — with special attention to the fact that the two sub-disciplines (cognitive psychology and "artificial intelligence") function as models for each other. On the one hand, Computer models are undeniably of benefit to psychology. In an explicit sense, one has only understood something when one can also construct it, most convincingly in the form of a fiilly functioning machine. Similarly, in psychology, one cannot dismiss something because of ontological scruples if it has been successfully realized in machines. Something that is reality in machines, for example Operations that can only be adequately analysed with cognitive categories, must be recognized at least as a non- contradictory, ontologically unproblematic possibility for human beings. If the moves made by a chess robot can only be adequately understood under the presupposition that the visible Operations on the board are derived from invisible Operations of a cognitive kind taking place in the "interior" of the robot, then such an explanation can no longer be dismissed for ontological reasons in the case of a human chess player. On the other hand, the narrow limits of the intelligent Performances of machines, which have so bitterly disappointed the original lofty hopes, have shown the "knowledge engineers" that after each partial success they are well advised to orient themselves on how the human or the human brain solves the projected tasks. 10 E. HOLENSTHN

Cognitive or semiotic categories physical or energetic categories The analysis of the intelligent Operations of human beings and machines with cognitive categories is radically different from the physical analysis of the processes in which they are instantiated. This becomes particularly clear in a conceptual comparison of cognitive categories with the physical categories to which a traditional physicalist adheres. The categories that he uses are (in addition to spatial and temporal categories) tbrust, pressure, attraction, repulsion, Impulse, tension, load and the like. Common to them all is the fact that they presuppose the category of power or energ^. It is revealing that in behaviourism, which attempted to explain human behaviour exclusively with mechanistic categories, a further physical category, ränforcement, was added. It is common to the categories of cognitive science, by contrast, that they are semiotic catego- ries. They are genuine sign categories (representation, signal, symhol, ) or categories that relate to the rule-govemed manipulation of signs {code, program and computatioti). It is decisive for the use of these categories that the processes described by them are meaningful, not only in the general sense of "func- tional", but also in the specific cognitive sense of "understandable" or "intelligent". In this sense, instead of "cognitive science", it is just as correct to speak of a "semiotic" or even (more provocatively) of a "hermeneutic science".

physical categories cognitive! semiotic categories thrust representation pressure signal attraction repulsion model impulse Information tension computation load code reinforcement program

In disciplines such as , psychology and , but also in , a physical (energetics related) analysis and a semiotic (cognitive or hermeneutic) analysis of one and the same process must be differentiated. Every semiotician is familiar with such a two-layered analysis, one that is at once physical and semiotic, from the sign processes that he studies, whether SEMIOnCS AS ACOGNmVE SCIENCE | 11 they be oral or written communicarions. They are always at the same time physical processes (signs transported by physical "sign vehicles" as it could be put drawing on Charles Morris's terminology). From psychiatry and especially from , too, we are familiar with processes that can be viewed physically and treated widi medication and which at the same time can be understood psychologically (cognitively) and treated psychotherapeuücally with verbal (semiotic) counselling. Psychoanalysis explicitly deals with processes that on the one hand seem to be subject to the laws of energetics (there is talk of drive abreaction and discharge of tension) and which on the other hand display a hermeneutic aspect that is just as pronounced. Drive behaviour is accompa- nied by images, it can be meaningfuUy interpreted and influenced by way of this interpretation. Hence, Paul Ricoeur (1965) righdy distinguishes between a psychoanalytic discourse in terms of energetics and one in hermeneutic terms.

The cognitive or semiotic tum of American psychologists

The cognitive renaissance in the human sciences is closely linked to how they developed in the United States. At the end of the 1950s, a number of American psychologists (cf. Miller, Galanter & Pribram 1960) found it simply too silly to continue to explain human behaviour in pure behaviouristic terms with trivial stimulus-response chains and with physical categories that are just as trivial after Computer science had long since started to work with much more complex structures and philosophically more sophisticated categories — the cognitive-semiotic-hermeneutic categories listed above - which, moreover, had earlier been regarded as specific to their own human science. Something that is possible in machines cannot be a priori denied to human beings. Today, or for someone who grew up in a different scientific and cultural tradition, it may sound perverse that one of the psychologists who helped to introduce the "cognitive tum", Jerome Bruner, writes in his autobiography about "the liberating effect of the Computer on the psychologist's image of what is humanly possible" (1983: 104).' But that was indeed the first effect that the artificial intelligence of the Computer had on the natural intelligence of the human being.

Similarly S. M. Kosslyn (1983: 22): "The first precise way to conceptualize how the mind works came from an unexpected comer, technology. [...1 What made the Computer so important for psychology was that it was a well-understood example of how a machine - a physical device — could process information". 12 E. HOLENSTBN

The restriction of the human sciences to what human heings observahly do proved to be totally inadequate for a science that adheres to the explanatorv' Claims ttiat are a matter of fact for every natural science, as Chomsky (1959) pointed out in his criticism of behaviourism, which was feit to be so sensational at the time. Behaviour must be explained. The decisive point is what human beings are capabk of. According to the explanation that is introspecdvely most obvious and up to now most fruitful, what human beings are capable of depends on what th^ know and how knowkdge is represented in them. The disposirions, with which many philosophers were satisfied in the 1940s and 50s - despite the fact that a straightforward recourse to dispositions tri^ers an association with Moliere's vis dormitiva, the somnolent faculty - must be underpinned in an explanatory perspective with representations and computaüons.

The cognitive tum of semiotics in the 20th Century

In the development of semiotics^ in the 20th Century, three stages can be discemed in the shift of attention to the subject, the user of signs: first, around 1930 a turn to the observable action of the users of signs, then around 1960 a tum to their cognitive competence and to their representations, and shortly thereafter, around 1970, a tum to the medium of their competence, to the format of their representations, to the question as to whether certain intelligent Operations of the human being can perhaps be better explained with pictorially stmctured than with propositionally structured mental representations, with mental images, or with more abstract mental models that are neither proposi- tionally nor pictorially stmctured.

Three stages in the tum of attention to the sign user around 1930 attention to observable behaviour around 1960 attention to cognitive competence around 1970 attention to the medium of cognitive competence

There are many people who see the significance of the familiar "", which was initiated by philosophy between the two world wars under the influence of behaviourism, in the dismissal of such problematical phenomena

2 This development applies primarily to semiotics in the United States. In Europe the picturc is different. The development here is characteri;;ed by a "decentralization of the ego", a differentiation of the subject, a turn from the ego to the unconscious and intersubjective media of the semiotic process. See Holenstein 1976: 47-51 and 61-69. SEMIOTICS AS ACOGNmVE SCIENCE | 13 as ideas. "Idea" seems to be a term related to a heterogeneous of fiizzy constructs that are both empirically and conceptually dubious; "language", by contrast, is deemed to be a tenn with which something intersubjectively observable and logically analysable is meant. In the course of the past decades, linguists and logicians have become more reserved. "Language", too, increas- ingly seems to be the tide for a completely hybrid entity. Another insight deserves more positive appreciation. It, too, is due to Computers. This is the insight that not only linguistic structures can be defined functionally, without recourse to subjective experience, but also structures which older philosophy attributed to mental images and other pre-linguisüc phenomena. The code to which brain processes are subject need not be a linguistic code. It can just as well be pictorial. Empirical data suggest that to solve certain tasks a code that is properly called pictorial is indeed decisive, as was always claimed with reference to the testimony of introspection. To function as a language, something does not have to be of an acoustic nature or of the nature of articulatory movements in the pharynx and mouth of a human being, and by the same token "something does not have to be a picture to function as one" (Kosslyn 1983: 22). Neuronal processes can also do this. If it is part of the self-image of science to do without subjective conscious- ness to the extent possible by defining intersubjectively verifiable physical structures in functional terms so that they can be used to explain something that in the past seemed to be explicable only by assuming the existence of consciousness, then the behaviourist turn to acoustic sounds is no more than one alternative. Another altemative, one which admits of numerous semiotic interpretations, is recourse to brain processes. There are empirical reasons that can be brought to bear against the assump- tion of the existence of mental images. Conceptual reasons cannot be adduced, although this was taught by philosophers for half a Century. Technology, however, has taught us that a physical device can process information just as well according to a pictorial code as according to a linguistic code.

Hermeneutics as a branch of semiotics Between the two world wars, attention shifted beyond the purely structural (syntactic and semantic) analysis of semiotic processes to their relationship to the human subject. But under the pressure of contemporary behaviourism, this resulted in a pragmatic rather than a cognitive tum. Social behaviour, which leads to or results from linguistic utterances, became the centre of attention; 14 L HOLENSTEIN

and for the fundamentalists among the behaviourists the focus was only on motor behaviour. Cognition was hardly mentioned, and at the beginning arid understanding were also ignored. But that above all is what is done with signs: one means something with them and one understands them. From this point of view, the centrepiece of the third sub-discipline of semiotics, pragmatics, is . And in view of the intuitive components of the mental models on which meaning and understanding are based, this hermeneu- tics is not only "grammatical", but also "divinatory" (see Schleiermacher 1838). The fact that today hermeneutic scholars, under the influence of the "lin- guistic turn" in philosophy, steer clear of the divinatory component of Schleiermacher's conception of hermeneutics as the devil steers clear of the Word from which "divination" is derived has a - demonstrably - sterilizing effect. Divination as a method is only problematical if the fact that it requires a Supplement is disregarded, if the fact that "grammatical" and "divinatory" analysis are complementary components of understanding is overlooked. Just as according to Kant without Intuition are "empty", a frail and fragile framework, and Intuition without concepts "blind", a groping in the dark, Schleiermacher feared that a hermeneut (i.e. an Interpreter) who restricts himself to grammatical analysis would be a "pedant" instead of a "virtuoso", and that one who makes do with divination would be a "nebulist" instead of an "artist". "Grammatical" hermeneutics needs "divinatory" hermeneutics for heuristic reasons and for Stimulation, "divinatory" hermeneutics needs "grammatical" hermeneutics for testing and substantiation. For wings, air is needed, and solid ground under the feet. The hermeneutic aspect of pragmatics becomes manifest in Charles Morris (1938) in the fact that he continues to use the term "Interpreter", the traditional designation for the user of signs. Moreover, it is not only in contemporar)' reflections on the use of signs that hermeneutics presents itself as a sub- discipline of semiotics. As early as the 18th Century, A. G. Baumgarten (1739: § 349) presented hermeneutics as a sub-discipline of the Säentia signorum or Characteristica (his "universal art of signs") together with heuristics (which deals with the of signs) and mantics (the specific subject matter of which is prognostic signs). According to Baumgarten, Hermeneuiica universalis (the SEMI0T1CS AS ACOGNmVE SCIENCE | 15

"universal art of interpretation") deals "de cognoscendis signorutn signatis" (with the knowledge of what the signs signify).^ To my knowledge, the first person to make the overdue cognitive revision of Morris's tripartition of semiotics into , and pragmatics was a chemist at the Swiss Federal Institute of Technology, Hans Primas, who, when a textbook on quantum mechanical chemistry, feit obliged to reflect on the philosophy of science. According to Primas (1981: 19ff.) every science can be regarded as a semiotic system. As such a scientific theory has to fulfil three criteria corresponding to the three branches of semiotics. In syntactic respect, a good theory is logically coherent, in semantic respect it is empirically verified and in pragmatic respect it is understandable. For Primas, intuition is a centrepiece of understanding: "A good theory is consistent, confirmed, and intui table".

Three requirements to be met hy a sdence (1) syntactic logical consistency (2) semantic empirical confirmation (3) pragmatic/hermeneutic intuitive understandability

It may be surprising that it is a quantum theorist who makes a case for intuition. The discussion of the counterintuitive theoretical constructs of quantum physics was a major factor leading to the disregard for intuition in the philosophy of science. In the physical sciences that investigate objects that escape intuition, however, the medium of knowledge (i.e. the signifiers) comes to replace the object of knowledge (i.e. the signified data) as the field of intuitive reflections and intuitive claims (among them aesthetic claims). Consider the role played by intuitive models, including mental models, for creative research- ers, and the role played by the concrete visual shape of mathematical formulae (especially their elegance).

Note that in Baumgarteti's wording, the object of hetmeneutics is not the understanding of the sign or the text, but rather of the thing designated. It is to the credit of later hermeneudcs that it has pointed out that ai] understanding of things is dependent on the expressions that are created for them, on the signs that are used for them. Today, in the attempt to build machines with linguisdc competence, the reverse dependency has come into the focus of attendon, the dependency of the understanding of language on knowledge of things (expert knowledge, knowledge of the world). The dependency is redprocal. It also means that linguis- tics, semiotics and hemieneutics are only relatively autonomous sciences, and require addi- Üonal sciences defined in material terms as their complements. 16 E HOLENSTEIN

The field of application of Intuition old: signata object of knowledge new: signantia medium of knowledge

The cognitive or semiotic tum in the philosophy of science

A development can be discemed in the philosophy of science of the last hundred years from the view that science is an inductive processing of objective data to the insight that all experience is theory bound and finally to the rediscovery of the dependency of all theory on Intuition and Imagination. The theory dependency of experience is commonplace today. "Seeing is a 'theory- laden' undertaking", is a key claim of one of the advocates of this position (Hanson 1958: 19). The fact that all theory is also intuition-dependent is gradually gaining wider recognition under the influence of cognitive science. "When scientists hold a theory, they hold a parricular mode of imager)' as well", reads a new key (Miller 1984: 312).

Thne stages in the history of the philosophy of science (1) From the inductive processing of objective data (2) via the theory dependency of all experience (3) to the image dependency of all theory.

As a classical philosophical discipline, philosophy of science deals mainly with Problems of the justification of knowledge and with normative questions. It studies how, in the ideal case, scientific knowledge is logically structured. As an object of research, the acquisition of knowledge is assigned to the history of science and empirical psychology. But the acquisition of scientific knowledge leads to explanations that rely on cognitive categories that are füll of philoso- phical Problems. It is thus understandable that the advocates of the cognitive sciences (linguistics, psychology, neurology and anthropolog)) that now cultivate contact with philosophy, from which they had supposedly definitiveh emancipated themselves in the course of the 19th and early 20th centur)' on the basis of their empirical procedures (which at the time meant "inductive"), are no longer concerned with formal, methodological problems as was the case in the middle of the last Century. It is not questions of the logical structure of scientific explanations to which they expect an answer from philosophy, as was the case earlier, for example the question whether functional explanations are in SEMIOUCS AS ACOGNmVE SCIENCE | 17 line with the deducrive-nomological model of scientific explanarion. What they reflect on today are rather substantive, ontological questions, for example the question as to whether mental structures intrinsically have to do with the functional Organization of physical structures and how any functional Organiza- tion can be instantiated equivalendy in physical systems that are totally different from each other (in punched cards, chips and neurons). Beyond the philosophy of science, there is a renewed demand for philosophy of niind - in particular, for a whose central categories are semiotic, categories such as "representation" and ""; a philosophy of mind that includes the question of the possibility of mental images, of non-linguistically structured knowledge among its central problems. In the past, semiotics was mainly practised as a descriptive science. The priority was the Classification of signs. What Chomsky convincingly demanded for linguistics, namely that it be established as an explanatory science, is also necessary for semiotics. Significandy, such a tum of focus to questions of explanation will remain, as pointed out, a largely intra-semiotic matter. It attributes the use of perceivable signs and large parts of communicative action to the possession of mental representations. It thus reverts to the pre- behaviourist semiotics of the Stoa and to .^

An historical appendix: John Locke redivivus

Locke's tripartition of the sciences in the last chapter of his Esse^ Conceming Human Understanding (IV.XXI.2-4) into physike, praktike and semeiotike is well known. As Locke's adoption of Helienic terminology shows, this is a traditional division scheme. It is derived from Stoic philosophy. Locke's most conspicuous deviarions from his ancient precursors are the new, alternative designations that he proposes. Here, only the term for the third discipline, semäotike, is of interest: "the Doctrine of Signs, the most usual whereof being Words, it is aptly enough termed also logike, Logick".^ The significance of semiotics for Locke is generally viewed in association with the Interpolation of a "Book of Words" into his "Treatise of Knowledge". He justified the addition of a linguistic investigation to the Essay with his

4 And from Locke to the young Husserl who was investigating the foundations of mathemat- ics. See Hoienstein (1978: 53ff, and 1988a). 5 The usual Stoic designation for this third discipline was indeed either "" or "dialectics" (cf. Diogenes Laertius: 1.17-18; 7.39^7; X.29-31). 18 E. HOLENSTEN

increasing insight into the part played by words and in thc acquisitdon and examinadon of our knowledge (III.IX.21; IV.VI.1-2). But Locke expücitly disdnguishes beiween two Idnds of sign: ideas and words (III.III.ll; IV.XXI.4). He states that words are the commonest signs ("the most usuai whereof being words"), but with respect to the acquisition of knowledge and the quality of knowledge provided by the two types of signs,'' the ünguisric signs, being the "Signs of our Idea/', only rank second in his view (IV.VI.l; IV.XXI.4). Their function is mainly ("which is the chief end of language" III.V.7) - a very traditional view — and (in private use) mnemonic (II.XXXII.7; IV.XXI.4). Ideas rank first. They serve the mind, to which the things are not immediately present, "as a Sign or Representation of the thing it considers" (IV.XXI.4). It can thus be conjectured that Locke introduced the new term semeiotike precisely because in the narrower sense the traditional term, "logic", only Covers linguistic signs. Furthermore, it is tennpüng to assign the Clements of the twosomes that are so prominent in the description of signs in the final chapter of the Yissty to the two types of sign, [1] to the ideas and [2] to the words ("distincdy weighed, and duly considered"): "the ways and means where by the knowledge [...] is [1] attained and [2] communicated"; "[1] the understanding of Things, or [2] conveying its knowledge to others" (IV.XXI.l & 4)."' In the Stoic-Lockean tradition, semiotics can be called a cognirive science in its first and fundamental part, just as the cognitive science of recent decades can be deemed to be a semiotic science in view of its central categories and Problems. There is a historical explanation for the fact that this was not discussed from the very beginning even though for computation there was no lack of paraphrases such as "symbol manipulation", and for Computer science no lack of slogans such as "physical symbol systems". At the time of the (re-)birth of cognitive science around 1960, there was no semiotics which included a pragmatics interpreted in cognitive or hermeneutic terms as its sub- discipline. Furthermore, in the wake of the rampant anti-mentalism of the preceding decades, the traditional Stoic-Lockean conception of semiotics had been pushed onto the sidelines. There is no need to point out that the traditional philosophical conceptions of need revision in many aspects. But only the blind can

6 On Locke's conception of the cognitive function of signs cf. Holenstein 1978: 43 f. 7 The distinction between mental and linguistic signs is, of course, also a traditional teaching ot ancient seniiotics. Cf. Diogenes Laertius: V1I.43. SEMIOTICS AS AC0GNIT1VE SCIENCE | 19 overlook the fact that mental representatdons exist and play a role in the use of linguistic signs and in communicative action in general. In the course of his "Treatise of Knowledge", Locke feit obliged to Supplement the "Book of Ideas" (Book II) with a "Book of Words" (Book III). Exploring linguistic usage (and also in the context of psychological theories of action), a need was feit in the second half of the 20th Century to write new books about mental represen- tations.

References

Baumgarten, A. G. (1739). Metaphysica. Halle: Hemmerde; Reprint, Hildesheim: Olms (1963). Bruner, J. (1983). In Search of Mind. Esst^s tu Autobiograp/y. New York: Harper & Row. Chomsky, N. (1959). Review of B. F. Skinner, Verbal Behavior (1957). l^nguage 35, 26- 58. Chomsky, N. (1972). \Mtiguage and Mind, Enlarged Edition. New York: Harcourt. Diogenes Laerüus (3rd Century CE [1931-38]). IJves and Opinions of Eminent Philosophers (Greek and English). Cambridge, MA: Harvard University Press. Hanson, N. R. (1958). Pattems of Discovery. Cambridge University Press. Holenstein, E. (1976). Roman jakohson's Approach to luinguage: Phenomenoio^al Structural- ism. Bloomington, IN: Indiana University Press. Holenstein, E. (1978). Semiotic PhUosophy? In R. W. BaUey et al. (Ed.), The Sign: Semiotics Around the World, (pp. 43-<57). Ann Arbor: Michigan Slavic Publications. Holenstein, E. (1988a). Eine Maschine im Geist: Husserlsche Begründung und Begrenzung künstlicher Intelligenz. Phänomenoh^sche Forschungen 21, 82—113. Holenstein, E. (1988b). Einfuhrung: Semiotica universalis. In R. Jakobson, Semiotik Ausgewählte Texte 1919-1982 (pp. 9-38). Frankfurt am Main: Suhrkamp. Kosslyn, S. M. (1983). Ghosts in the Mind's Machine: Creating and Using Images in the Brain. New York: Norton. Locke, J. (1690'-17065). An Essi^ Conceming Human Understanding (ed. by P. H. Nidditch). Oxford: Clarendon (1975). Miller, A. I. (1984). Image^ in Scientific Thought. Boston: Birkhäuser. Miller, G. A., Galanter, E. & Pribram, K. (1960). Plans and the Structure of Behavior. New York: Holt. Morris, C. (1938). Foundations of the Theory of Signs. In an the General Theory of Signs (pp. 13-71). The Hague: Mouton (1971). Primas, H. (1981). Chemistry, Quantum Mechanics and Reductionism. Berlin: Springer. Ricoeur, P. (1965). De l'inthpretation: Essai sur Freud. Paris: Seuil. Saussure, F. de (1916). Cours de linguistique generale. Paris: Payot. Schleiermacher, F. (1838). Hermeneutik und Kritik (ed. by F. Lücke). Berlin: Reimer; Reprint: Frankfurt: Suhrkamp (1977). Watson, J. B. (1913). Psychology as the Behaviorist Views It. P^choloffcal Review 20, 158-177.