Repository | Book | Chapter


(2022) Phenomenological philosophy of language, Genève-Lausanne, sdvig press.

Phenomenological structuralism and cognitive semiotics

Elmar Holenstein, Roberto Benatti

Publication details

Full citation:

Holenstein, E. , Benatti, R. (2022). Phenomenological structuralism and cognitive semiotics, in Phenomenological philosophy of language, Genève-Lausanne, sdvig press.

File clean-up November 23, 2021, 11:43 am sdvig press ( )

1 Benatti - Professor Holenstein, let us begin this interview1 with a general consideration of your research work. Your scientific activities in semiotics, phenomenology and structural linguistics show an extraordinary coherence, a continuous red line.

2Holenstein - The impression of »extraordinary coherence« results perhaps because of the contemporary background of my research work. All philosophers known to me within the German speaking countries who are more or less my age and are occupied with problems of language are dealing with either logical or pragmatic aspects of language. Right from the beginning, my main interests were the cognitive aspects of language and the interelationships between innate structures, experience, language and knowledge. Until the late eighties, I was relatively isolated in this line of research among German philosophers. What I did was regarded as empirical science and not as philosophy. It was not realized that there is no empirical linguistics without philosophical assumptions. But the epistemological preconditions and the ontological commitment (Quine's term) of the structural linguistics of the Prague Circle and the cognitive science of today are a challenge to the philosophers of the 20th century. For traditional phenomenologists, it is the conception of a »non-reductive naturalism«, and for philosophers like Apel and Habermas who maintain that the »philosophy of consciousness« is an outdated paradigm to be replaced by a »linguistic turn« in philosophy, it is the return (in particular and just in linguistics) to conscious phenomena. The present situation in Germany differs from the situation in the seventies. Nowadays cognitive problems are trend-setting among the younger philosophers in Germany as well as on a worldwide level.

3Benatti - One of the most interesting aspects of your writings is the re-discovery of the phenomenological memory of structuralism. (Holenstein 1973; 1974a; 1975a; 1975e; 1975f; 1976a; 1976g; 1979f; 1984b; 1984c) You indicate that phenomenology forms the historical and theoretical condition of the possibility of structuralism in connection with the replacement of naturalism. You combine typical structuralistic categories in an excellent formulation: »A phenomenon is no longer explained on its own by leading it back to a preceding event, of which it is a mechanical effect, but through the function it has within a system, or rather through the meaning it has to a subject.»(Holenstein 1985b, 12) What is Husserl’s contribution to the development of relationships between these categories?

4Holenstein - When speaking of »naturalism«, we must distinguish between two variants. According to »eliminative naturalism«, everything that is explained in biology and in the humanities with non-physical categories can be explained more easily and in a more general manner with physical categories. With the evidence of the autonomy of the laws of logic and psychology, Husserl made a revolutionary contribution to the overcoming of this naturalism. »Non-eliminative or ecumenic naturalism« recognizes the autonomy of logic and psychology. The axioms of logic and psychology cannot be derived as theorems from the axioms of physics which explain physical systems behaving in conformity with logical and psychological laws. Yet there is good reason for the assumption that nothing changes on the logical or psychological level of such a system without something changing on its physical level. The ecumenical naturalist asks: Which physical conditions allow a computer to behave in conformity with logical laws and a living creature in conformity with psychological laws which cannot be derived from physical laws? Husserl contributed only indirectly to this conception of a non-eliminative naturalism by trying to elucidate how a creature - to which the psychological laws apply - can behave in conformity with the laws of logic, which cannot be derived from the laws of psychology.

5Benatti - You were the first to point to the connection between the Prague School and phenomenology. You even spoke about Prague structuralism as a trend in the phenomenological movement. (Holenstein 1976, 56-75) I would like you to say something about this correspondence.

6Holenstein - Jakobson himself was the first to point to the connection between the Prague School and phenomenology, again and again. Another scholar pointing to it before I did was Jan Broekman (Broekman 1971). I was perhaps the first to follow up all evidence to be found in literature for this connection, and to try to allocate a systematic place within the phenomenological movement to Prague structuralism. The different trends within the phenomenological movement can be allocated to the different parts in Husserl's main work, his Logical Investigations. The third and the fourth logical investigations, about the theory of wholes and parts and about Universal Grammar, came to lead Jakobson’s way. These are the parts which were hardly taken to by philosophers belonging to the phenomenological movement, even though Husserl was always of the opinion that they were fundamental.

7 Benatti - How was your approach received?

8Holenstein - The first philosophers to quickly follow my indications to these parts of Logical Investigations and their evaluation by Jakobson, to apprehend the meaning for philosophical psychology and ontology, and to make it the theme of many ingenious writings, were three - at that time - very young British philosophers: Kevin Mulligan (today in Geneva), Barry Smith (today in Liechtenstein) and Peter Simons (today in Salzburg). Characteristically, amongst the elder phenomenologists, it was the Czech philosopher Jan Patočka who immediately agreed with my interpretation. He wrote a review of my book about Jakobson (Patočka 1976) with which Jakobson himself was very pleased. Jakobson was an extremely grateful man. It was my impression that he was a »typical Russian«. He who did something good for him or his work would be returned this kindness seven times over. When Patočka, as speaker of the »Charta 77«, died after an interrogation by the authorities in Prague, Jakobson wrote in an obituary: »There have been three Czech philosophers of world significance and exceptional moral power and purity: Comenius, Masaryk and Patočka (Jakobson 1977, 26-28). This was Jakobson!

9Benatti - Where Jakobson is concerned, it was you who brought his philosophia latente to light (Holenstein 1984b; 1984c; 1987d; 1990b). In your words, while Jakobson founded his fundamental and general principles in Hegel and in the Russian ideological tradition, he obtained the very specific formal laws of the foundation from Husserl. How is the relationship between Husserl and Jakobson characterized in terms of phonological laws?

10Holenstein - According to Jakobson, the sounds of a language form »a structural whole«. In Husserl's writings, Jakobson found a definition of wholes heuristically productive for phonology. One-sided and reciprocal implications arc constitutive for a whole. Distinctive characteristics facing each other in a relation of opposition (e.g. acute and grave) reciprocally imply each other. Something acute can only be perceived in contrast to something grave. Other traits are implied one-sidedly. It is thus a precondition for the presence of a nasalized vowel in a phonological system that a non-nasalized vowel is present. In Husserl’s terminology: One vowel is founded in another one.

11Benatti - Another example: You wrote that Jakobson's structuralism is a »Husserlianism« (Holenstein 1975a, 55-58) in the question of subject reference ...

12Holenstein - Here a brief biographical excursion may be quite informative. Jakobson was invited as visiting professor to Louvain in 1972. Since he was a long-time admirer of Husserl, he visited the Husserl- Archives where I was preparing the first volume of Logical Investigations for publication. I showed Jakobson the proofs of my dissertation on the Phänomenologie der Assoziation, which was just being printed. To my astonishment, he expressed immediate interest in a chapter which was of no interest to anyone else. It was a historiographic chapter about the concept of apperception (Holenstein 1972, 132-166). This concept attracted him. Why? For a historical and for a epistemological reason. It reminded him of a paper about the problems of apperception which he had written as a student at Moscow University in 1915 or -16 in a psychology seminar. Later, the concept of apperception became fundamental to the phenomenological conception of phonology. It is not the fact that sounds are perceived (apprehended), but rather how they are apperceived (comprehended) that is decisive for linguistic understanding. Since the term »apperception« has become obsolete, Jakobson replaced it in his American years by the genuinely phenomenological term »set« (Einstellung). When Paul Ricoeur labels structuralism as a »Kantianism without a transcendental subject«, this may be correct for some more formalistic branches within French structuralism, but it is surely not appropriate for Prague structuralism. This was the reason why I called Jakobson's structuralism a »Husserlianism«.

13 Benatti - Were there other reasons as well?

14Holenstein - Another reason was that for Jakobson, as well as for Husserl (and differently from Kant), the unconscious and intersubjectivity displaced the Ego as the unique carrier of the constitution. I can imagine that Jakobson would have agreed with Lacan's slogan »Decentering of the Ego«, but hardly or only with great restrictions with Foucault's slogan »Death of the Subject«. Only sick people - aphasiacs and schizophrenics - can lose the use of personal pronouns (compare Jakobson’s Hölderlin- analysis). How can one dream of a linguistics »without a subject« when one is dealing with two subjects in the use of language (a sender and an addressee), and when all languages dispose of personal pronouns and other deictic expressions? The human self-understanding implicit in the use of language over thousands of years is resistant to short-lived intellectual trends.

15Benatti - Patočka wrote that you did not only manage to show the starting point of Jakobson's structuralism in the first, third and fourth parts of Logical Investigations, but that you also showed which role the eidetic abstraction played in Jakobson's research work. Furthermore, it seems as though Patočka attached great importance to your »semiotic transcendentalism«. What seemed to be most important to him - as he said - was »an apprehension of the transcendental phenomenology [...] not exclusively using psychological terms but principally progressing semiotically« (Patočka 1976, 129) Under which circumstances do you see transcendental semiotics today?

16Holenstein - Today one has to speak up for the opposite position. A transcendental semiotics exclusively using semiotic and not psychological terms as well is sentenced to fail. Transcendental semiotics depends upon a supplementation of transcendental aesthetics. The relationship between transcendental aesthetics and logic remained an outer, mechanical relationship in Kant. The relationship between transcendental aesthetics and semiotics is interdependent.

17Benatti - You mentioned earlier that Jakobson had the opportunity to read your book Phänomenologie der Assoziation. How did he react?

18Holenstein - After I sent him my book, Jakobson replied with a truly encouraging letter (May 8, 1972): »Your book is particularly relevant to me because for decades I have felt the urgent need for a competent revision and rehabilitation of such concepts as association with its subclasses, and apperception.«

19 Benatti - With the comprehensive analysis of the association term, you made an important contribution to a basic transcendental-phenomenological concept of »passive consciousness«, to use Husserl's words. Later on, you supported the reintegration of formal and structural research in contemporary phenomenology. Would it be possible to take your association-theoretical conception of the two main axes of language (paradigmatic and syntagmatic) as an example of this? (Holenstein 1972a; 1974b)

20Holenstein - It is typical of mental and verbal human representations that logical as well as associative relations which are possible between them guide their production and their understanding. A reason for Jakobson's immediate interest in my dissertation in 1972 was its elaboration upon the importance of associative connections. In those days, they were completely undervalued as much in phenomenology as in Chomsky's linguistics. Today the trend is the other way around. Associative relations are overemphasized under the heading »connectionism«, and logical connections are undervalued. But the human language and the human mind are hybrid systems for which no uniform explanation exists.

21Benatti - Your position is clear to me. Furthermore, I want to remind you that the connectionists take care to connect themselves with phenomenology. They like to refer to confrontations between two types of philosophizing: logical and physical ontology, and the metaphor of computer and that of the brain. If one thinks about the background term (Hintergrund), one finds similarities with Husserl. All background processes are intentional, but not representational, and are filled with a finality. It can be said, in the phenomenological sense, that a horizon- intentionality of the network exists. You had already investigated the background term in comparison with similar problems in 1985. How do you see these aspects of a connectionistic apprehension of phenomenological tradition today?

22Holenstein - As I have already mentioned: I am of the opinion that connectionism is not a »universal key« to solving all problems. It is one method among many others, all of which must be adopted for the establishment of a »cognitive map« of the human mind. Connectionism has intensified for many people the sense of the fact that not only a numerically exact apprehension is essential for elucidating a phenomenon. A qualitative description of how something functions is just as important.

23Benatti - Let us now speak about your research on universals (Holenstein 1976a, 125-133; 1976d; 1976e; 1978b; 1979c; 1979d; 1980a, 53-70; 1980c; 1981c; 1985a; 1990a). You were of the opinion that the problem of universals can be regarded as an example of the manner in which Jakobson made use of Husserl. In this respect, you wrote: »With regard to the explanation, the contrast of the more recent research in universals is particularly striking in comparison with the classical teaching of universals. Universals were almost exclusively logically reasoned in the traditional philosophical theories. They are primarily explained biologically and psychologically in the more recent research.« (Holenstein 1985a, 2) And you continue: »The more recent linguistic research in universals, in stark contrast to this traditional conception, comes in on the level of expression categories, on the level of Jakobson's phonology [...] and on the level of the word order and the morphology of Greenberg.« (Holenstein 1985a, 7)

24Holenstein - Speaking of »language universals« has the effect of a red flag with many philosophers. Today, it is overlooked that completely different views are represented under this title than has been the case in the past. You cited two important differences.

25Benatti - I would like you to depict two other sides of your universals research work. At times you confront the thesis of language universals with Wittgenstein's thesis of »family resemblances« (Holenstein 1976, 125-133; 1980a; 1985a)

26Holenstein - It is to Wittgenstein's merit that with the catchword »family resemblances«, he pointed to something important: For the affiliation with a »natural kind«, it is not necessary (»essential«) that each member possess all characteristics typical of the kind. No individual feature is necessary for this. But Wittgenstein’s concrete statements, and in particular his analysis of the word »game«, are superficial and wrong. They do not reach the level of the investigations of folk taxonomies in ethnolinguistics, which show how things are arranged in word fields in natural languages. Nor do they reach the level on which the Russian psychologist Lev Vygotsky showed (also using the expression »family name«) how children classify things linguistically. I described these alternatives in detail and evaluated them in the final chapter of my book Sprachliche Universalien (Holenstein 1985, 169-208)

27 Benatti - Can you give an example?

28Holenstein - The manner in which we structure the world cognitively and linguistically varies greatly. But many of Wittgenstein's followers overlook the fact that there are universal constraints in variability. Not all variations that come to mind exist in natural languages. And not all of them have the same stability. Not everything that is logically possible is natural.

29Benatti - With regard to cognitive sciences, permit me to introduce our readers to another quotation from your writings: »At present, the greatest challenge of the phenomenological - or rather - mentalistic philosophy is that all mental performances can also be explained physically.« (Holenstein 1980, 152; cf. also 1981b; 1982d; 1985a; 1988c) Do you see in the phenomenological categories the possibility of an integrated investigation of mind and brain?

30Holenstein - Most suitable for this are the categories Husserl specified under the title »formal ontology«, the already mentioned categories of one-sided and reciprocal foundations. But using such categories alone, one is moving on a very abstract level.

31Benatti - In his book The Mind's New Science (Gardner 1985), a »classic work« of the cognitive sciences, H. Gardner very much neglected Husserl, yet different researchers pointed to surprising anticipations found in Husserl's noetic and noematic analyses.

32Holenstein - It is to Gardner’s merit that he - as an American scholar, in his book on the historical background of cognitive science - did not completely disregard the European forerunners of this movement. Thus he refers to the genetic epistemology of Piaget, the structural anthropology of Lévi-Strauss and also to Jakobson and the Prague School. Nobody is omniscient, and particularly not in historical connections.

33Benatti - That is clear, but Dreyfus already spoke about Husserl as the father of cognitive psychology and artificial intelligence (Dreyfus 1979; 1982; PEtitot 1989). Do you think that this perspective is exaggerated?

34Holenstein - Dreyfus has the merit of having written very early a critique of the project of »artificial intelligence« from a phenomenological point of view. Thereby he exposed a weak side of this too ambitious project and has become very influential because of this. He has done more for the understanding of phenomenological viewpoints than the great number of phenomenological colleagues charging him with this and that misinterpretation of Husserl. That is why I do not like to criticize Dreyfus. Dreyfus tends to simplify and personalize philosophical positions. Husserl is »the bad guy« with regard to »artificial intelligence« and Heidegger and Merleau-Ponty are »the good guys«. This has to be attributed in part to the fact that, in Husserl's case, Dreyfus mainly refers to the first volume of Ideen (1913) - an ideological text. He thereby did not take into consideration that in Husserl’s earlier and later works much more circumspect analyses can be found by means of which the project of »artificial intelligence« can at the same time be justified and restricted in an original manner (cf. Holenstein 1988b).

35Benatti - Let us continue to speak about the relationship between phenomenology and cognition sciences. Husserl saw the task of semiotics to be the following: »The great question of the logic of signs - semiotics - is how it is possible that a blind mechanism of sensual signs can replace and save logical thinking.« (Husserl 1970, 394) It is remarkable that Husserl's observations, which can be regarded as anticipating present cognitive research, are principally devoted to semiotics. Nevertheless, this conception gives rise to some problems. How can semiotics be understood as cognitive science, according to Husserl?

36Holenstein - The central concepts of cognitive science are all semiotic concepts (representation, symbol, information, code, program etc.) or imply semiotic concepts (computation). When cognitive scientists define themselves by proclaiming that they deal with Physical Symbol Systems, it is only a historical accident that one does not speak of semiotic science but of cognitive science. From a phenomenological point of view, it is a lucky accident because cognition is not exclusively logically and symbolically structured.

37Benatti - By formulating the thesis that cognitive science is also a semiotic science, you also touch aspects of the history of semiotics. I cite: »In the tradition of the Stoics and of Locke, semiotics can be designated in its first part as a cognitive science, just as it is possible from the other side to determine the cognitive science of the last decades as a semiotic science in consideration of its central categories and problems.« (Holenstein 1988b, 26; cf. 1988c) How can this comparison be explained?

38Holenstein - That we, in our cognitive performances, rely on »internal signs« (mental images and models, plans, linguistically structured considerations) is a common and well-founded introspective experience. The question of whether or not and to what extent the unconscious physical processes in our brains can be interpreted as semiotic processes is controversial.

39Benatti - Today this seems to be one of the central problems of a »non-eliminative« semiotics. With this concept, I would like to mention that in phenomenological-oriented semiotics, the attempt is being made to imagine something different from the traditional dualism between the cognitive and the physical sphere. Nevertheless, I see - as you so fittingly said - an Entsinnlichung of semiotics in the explanation of physical processes as relevant to semiotics if these are not understood with a phenomenological conceptuality.

40Holenstein - The word Sinn in German has several meanings. It is used in psychology for the »senses« (sense-organs), in logic and in linguistic theory for the »senses« or meaning of expressions and in philosophy of mind for the »inner sense« or consciousness. Along this line, it can be said that there developed a kind of Entsinnlichung in semiotics in the 20th century in three respects. We can see (1) a desensualization, (2) a deidealization and (3) a despiritualization of semiotics. (1) It is no longer a precondition that something be perceived sensually or experienced phenomenally in order for it to function as a sign. It is sufficient that it is a physical incident which can be causally effective. (2) An immaterial mind is no longer a precondition for the comprehension of the meaning of a sign. It is sufficient that the user of the sign is capable of making use of physical structures functioning as signs and following rules of a logical and teleological kind. (3) For something to be classified as a sign, it is no longer a prerequisite that an »idea« connected with it be mentally observed (in a subjective, introspective experience) and understood. It is sufficient that a certain use is made of it: a teleological use for the control of behavior, or a logical use for arriving at conclusions. A phenomenologically trained philosopher, in view of this development, spontaneously raises the question of whether or not something has been eliminated that can be »phenomenally experienced«, and without which the specifically human performances cannot be explained. Is it not the case that something is too quickly eliminated here only because it does not fit into one's philosophy for dogmatic reasons?

41Benatti - For some time now, much has been said about phenomenological semiotics. Your work is understood as a prerequisite for the construction of a phenomenological semiotics in German speaking countries. For example, W. Nöth summarizes entirely heterogeneous directions like your work and that of M. Heidegger or J. Kristeva under the designation of phenomenological semiotics. He arrives at the following conclusion: »In spite of the different approaches and the great number of relationships between semiotics and phenomenology, it should be too early to be speaking already about a phenomenological direction of semiotics.« (Nöth 1985, 5-6) Other authors also point to phenomenological semiotics as one of the current directions (Bentel & Bystrina 1978). Nevertheless, it seems as though they just did not understand the main historical and theoretical characteristics, and basically mixed several approaches which are not really compatible from the viewpoint of a structuralistic and semiotic interpretation of Husserl's phenomenology. How do you judge the attempt to design a phenomenological semiotics?

42Holenstein - Phenomenology has become a philosophical movement with heterogeneous approaches and concerns. That is why it is by no means astonishing that today heterogeneous movements in semiotics are labelled »phenomenological semiotics«. As a student, I gained insight into phenomenological philosophy through reading Heidegger. But soon it became clear to me that I had to go beyond Heidegger's positions (1) if I did not want to give up my intuitions concerning my outlook on the world, which were always close to scientific experience, and (2) if I wanted to follow the state of research in the humanities and the natural sciences with my philosophy. The most obvious way to accomplish these two objectives was to return from Heidegger to Husserl. Besides that, on the way back to Husserl, there lay the human-scientific currents of the first half of the twentieth century, which were congenial to Husserl's phenomenology and partly developed from it, Gestalt Psychology and the structural linguistics of Saussure and the Prague Circle. When you are searching for a model with such an orientation beyond Heidegger, in a return to Husserl and the mentioned scientific currents, you may keep to Merleau-Ponty. He took this path in his times. This path should be followed again today, if one wants to construct something like a phenomenological semiotics. Concerning Husserl, it would be most wise to start with his earliest semiotic considerations, which are found in Philosophy of Arithmetics and Logical Investigations. The semiotic analyses carried out by Husserl, while for many years endeavouring to revise the sixth logical investigation (Husserl 1975), should be helpful in doing so. At present, these analyses are being prepared for publication by Felix Belussi for a volume of Husserliana. One will be able to find the connections between the present attempts in cognitive science and these early writings by Husserl rather quickly. The present attempts try to clarify the ontological status, the structure and the function of mental representations (i.e. signs). One of my candidates for a doctoral degree in Bochum, Rolf Schlegel, is working on such a project under the working title »Inner Signs«. »Inner Signs« is a more suitable title than »mental signs«. It is an ontologically neutral title, and indicates that he who is trying to understand human performances must take into account not only mentally realized but also physically (neurally) realized signs. Today, it is no longer a matter of designing an exclusively phenomenological semiotics. The urgent task is a non-reductionistic multi-dimensional semiotics, a semiotics with a nonreduced phenomenological dimension.

43 Benatti - Could you please explain in more detail what you mean by a »non-reduced phenomenological dimension«?

44Holenstein - This depends on the conversation partner: With a hermeneutician, I would emphasize that not only sense relationships but also formal relationships (similarity, nearness, patterns of Gestalt) can be phenomenally apprehended and are fundamental to the constitution of a »text«. With a linguistic-analytic philosopher, I would emphasize that human linguistic usage is not understandable without intuition. We only know intuitively what is meant by equality and inequality. With a representative of a »materialistic« philosophy of mind, I would emphasize that the experience of consciousness cannot be reduced to its causal role. We are capable of a non-reducible experience of a phenomenal quality of consciousness.

45Benatti - The role of phenomenology in Ch. S. Peirce is continually being re-assessed from different directions, and this is due last but not least to Jakobson, who integrated the Peircean semiotics in structuralistic methodology. It seems to me that an investigation of the possibilities of a semiotics with a non-reduced phenomenological dimension should also take into consideration Peirce's writings on phaneroscopy. Do you believe that a comparison between Peirce and Husserl could be productive today? Some regard Peirce, where the concept of »unlimited semiosis« is concerned, as a forerunner of the »drift of the signifiers«. Within this framework, J. Derrida says - in his book, De la grammatologie - that the differences between the two are fundamental (Derrida 1967). To name just two: The concepts of the »sign« and »the metaphysics of presence«. He also says that Peirce was closer to the inventor of word phenomenology, Lambert, when he intended to reduce the theory of things to the theory of signs...

46Holenstein - For Jakobson, it was not the divergences but the convergences between Peirce and Husserl which were fundamental. Differences between different authors are most telling if one does not undervalue the commonalities, in so far as they exist. For Husserl, the concept of sign is central but not as comprehensive as it is for Peirce. Husserl was not a pan-semiotician. This also applies to Lambert. Lambert does not reduce the theory of things to a theory of signs. Lambert only showed that it is possible to investigate an ontological complexity in a system of signs if the signs which are put together are not arbitrarily selected. Thus it is possible to clear up a complexity which cannot be intuitively overlooked »without looking back on the matter itself«, as Lambert said. In his Philosophy of Arithmetics, Husserl followed Lambert in this point. There Husserl showed the importance of signs for the constitution of mathematical concepts and the execution of mathematical operations. But Husserl himself probably did not read Lambert - surely not as thoroughly as Peirce did. Both Husserl and Peirce became important to semantics; Husserl with the concept of intentionality, the structure of which he clarified far beyond Brentano; and Peirce with the concept of the interpretant, the function of which he showed in a purely relational analysis of sign structure. Husserl is a phenomenological structuralist, Peirce a so-to-speak pure structuralist. Both were equally attractive to Jakobson, the former analyzing how relationships are subjectively experienced, and the latter analyzing relationships functionally and in their systematic connection.

47Benatti - As we all know, Jakobson supplied some bases for the application of R. Thom's theory of catastrophes to languagey2 especially for the processing of the theory of linguistic morphogenesis, and therefore it is no coincidence that some researchers of catastrophe theory connect up with phenomenological structuralism. In terms of this connection, I am of the opinion that your writings have important points of contact with a structuralistic resistance against an understanding of semiotics which underestimates the dependency between form and content (e.g. in considering categorical perception, the theory of prototypes and the teleological derivation of structures). Indeed, researchers like J. Petitot (Petitot 1989, 192) or K. Pomian (Pomian 1981, 733) regard your interpretation of phenomenological structuralism as a precondition when trying to bring Jakobson's transfer of eidetic phenomenology in phonology back into the game. How do you regard this development of a phenomenological method in a discipline uniting theoretical biology and mathematics, as in Thom's works?

48Holenstein - In 1974 and 1975, Jakobson spent the summer months in Switzerland, in Einsiedeln near Zurich. He wrote an article there about »structuralism« for the Enciclopedia Einaudi. Unfortunately, he never completed this article. In this context he said that for him there were »three great structuralists«: Trubetzkoy, Lévi-Strauss and Thom. Saussure and especially Hjelmslev did not belong among them. According to Jakobson, they were both formalists and not structuralists. For a structuralist, »form« and »content« are not independent of each other. The interdependency of »form« and »content« is nowhere else as clear as it is in natural languages, in phonology and in semantics. At that time, Jakobson asked me what I thought of Thom. I said that I did not understand him and that, as far as I believe I understand something, it speaks to me intuitively. Jakobson answered that Lévi-Strauss said something similar to him, and that the same applies to himself. So I can only say: I assume that what Jakobson spoke intuitively about was exactly this connection of linguistics, biology and mathematics about which you asked. Biology like linguistics is a science in which it is impossible to avoid functional, i.e. teleological categories, as well as information-theoretical, i.e. semiotic categories. Think about the genetic code! Biology, furthermore, has to do with natural dynamic systems, as is the case with linguistics. Furthermore, structural constraints play an eminent role in the development of these natural dynamic systems. About mathematics, Jakobson used to say that only non-mathematicians believe that mathematics deals exclusively with numbers. Mathematics is more than just numerical mathematics, and especially more than just statistics about which many humanists think when they hear about mathematical linguistics. Topology was an attraction to Jakobson because it deals with the formalized analysis of relationships and qualities. That was precisely a concern of structuralism. Jean Petitot belongs to those few philosophers who recognized that phonology is an Eldorado for philosophical analyses, and in particular for those of an epistemological kind. I suppose that today, the strongest approach to a semiotics with an unreduced phenomenological dimension can be found in Petitot.

49 Benatti - I know that lately, you have been working specifically on the difference between the concepts of representation and idea. Could you describe the focal points of your recent investigations?

50Holenstein - On one hand, I am endeavouring to establish a definition and clarification of the concept of representation. I am working on the following definition: A representation is something (1) that co-varies systematically with something else - in prototypical cases: isomorphically (this is the structural component of the definition); (2) that serves to control behavior - in prototypical cases: to draw conclusions (this is the functional component of the definition); and (3) that is acquired, obtained and modified in prototypical cases, depending on this function (this is the genetic component of the definition). Furthermore, as a phenomenologist, I am interested in how conscious ideas differ from purely physically realized and purely functionally defined representations. For philosophers with a traditional phenomenological education, it is important to realize that today one naturally speaks of representations in theoretical psychology if it can be shown that a physical structure in a computer or in a biological system plays a causal role, which was formerly assumed to need conscious ideas. Perhaps one must get away from assuming that conscious representations of human beings make possible behavioristically defined performances, because a physical system with only physically realized representations is basically not capable of doing so. A being with conscious representations is perhaps only - in »real time« - superior to a merely physical competitor. It may perform tasks within a period of time and under circumstances which a robot, within the same time limit and under the same conditions, could not accomplish. One thing is certain: Animals and human beings with conscious ideas have developed within a period of time and under conditions which are insufficient for an automatic development of robots. For a structuralist a la Jakobson, the meaning of a linguistic expression is not independent of its phonic and syntactic structure. Expressions in different languages can therefore be equivalent with regard to an individual function, but never with regard to all functions they can take on, because their connotations are dependent on the phonic and syntactic structure. And I ask myself whether something analogous also applies to internal representations - to consciously realized and to merely physically realized representations. Merely physically realized representations are basically indeterminate for linguistic expressions, as was convincingly shown by Quine. Conscious representations are different; they arc obviously determinate. When speaking of a rabbit, I know that what I mean by the word is certainly not an undetached part of a rabbit, as a reader could interpret my word »rabbit« - according to Quine's analysis - to mean. This is a difference, but unfortunately not simple behavioral proof.

51 Benatti - I understand your work in this way: there exists a tendency in semiotics which allows for the understanding of the cognitive sphere and does not simply reduce it logically and symbolically, while the phenomenological structuralism of Jakobson and of the Prague Linguistic Circle forms its epistemological and methodological - or rather categorical - basis. Moreover, as a phenomenologist, you try to demonstrate the negative influences of a behavioristic and physical reduction of the phenomenological dimension of the mind, which have dogmatically restricted theoretical research in this field. In conclusion, could your point of view be summarized in this way?

52Holenstein - Two questions must be kept apart: (1) Is it in principle possible to analyze human cognition outside a symbolic and logical paradigm? (2) Do we have good empirical reasons to assume that we humans actually make use of logical structures in our cognitive performances, although this would not be absolutely necessary? In physics, parsimony proved to be a fertile principle. Another principle proved to be fertile for biological systems: redundance. Biological systems make use of structures in quantitative as well as in qualitative respects, and they are not absolutely necessary, but they are advantageous. This is one of the most important methodological insights that I gained from Jakobson's structuralism.

53Benatti - Thank you for the interview. I am convinced that the approaches you have presented will be very stimulating for the readers of Scripta Semiotica.


  • 1 This interview was originally conducted in German, in 1991, through an exchange of correspondence. The English translation was edited by Amy Williams.
  • 2 Among other writings, Jakobson's »dynamical theory of morphogenesis« is explained in (Jakobson 1963). For Thom see »Topologie et linguistique« (Thom 1970)




Günter Bentele, Ivan Bystřina

Stuttgart, Kohlhammer



Jan Broekman

Freiburg im Breisgau, Alber

De la grammatologie


Jacques Derrida

Paris, Éditions de minuit

What computers can't do


Hubert L Dreyfus

New York, Harper & Row

Passive Genesis


Elmar Holenstein

Tijdschrift voor filosofie 33/1

Einleitung des Herausgebers


Elmar Holenstein

in: Logische Untersuchungen. Erster Band, Den Haag : Nijhoff

Jakobson und Husserl: Ein Beitrag zur Genealogie des Strukturalismus


Elmar Holenstein

in: History of linguistic thought and contemporary linguistics, Berlin : de Gruyter

Le tournant morphologique de l'objectivité et de la cognition


Jean Petitot

in: The nature of culture, Bochum : Brockmeyer

Topologie et linguistique


René Thom

in: Essays on topology and related topics, Berlin : Springer

This text is available for download in the following format(s)