Kant and the Platypus

Essays on Language and Cognition

Umberto Eco

• Publisher: Mariner Books
• Year: 2000
• ISBN:
  9780547563787
• Topic: Literatur
• Subtopic: Theorie der Literaturkritik

How do we know a cat is a cat... and why do we call it a cat? An "intriguing and often fascinating" look at words, perceptions, and the relationship between them ( Newark Star-Ledger ). In Kant and the Platypus, the renowned semiotician, philosopher, and bestselling author of The Name of the Rose and Foucault's Pendulum explores the question of how much of our perception of things is based on cognitive ability, and how much on linguistic resources. In six remarkable essays, Umberto Eco explores in depth questions of reality, perception, and experience. Basing his ideas on common sense, Eco shares a vast wealth of literary and historical knowledge, touching on issues that affect us every day. At once philosophical and amusing, Kant and the Platypus is a tour of the world of our senses, told by a master of knowing what is real and what is not. "An erudite, detailed inquirity into the philosophy of mind... Here, Eco is continental philosopher, semiotician, and cognitive scientist rolled all into one." — Library Journal (starred review)

Chapter Three

Cognitive Types and Nuclear Content

3.1 FROM KANT TO COGNITIVISM

If Kant had considered schemata early enough, Peirce said, they would have overgrown his whole work (WR 5: 258–59). In the previous chapter, I suggested that it was precisely the problem of schematism that obliged Kant to undertake a change of direction in the third Critique. But we might say something more: if we were to reconsider the problem of Kantian schematism, much of the semantics of this century, from the truth-functional to the structural variety, would find itself in difficulty. And this is what has happened in the area usually referred to as “cognitive studies.”

In point of fact, a hint of various forms of Kantian schematism (connected to a constructivist idea of knowledge) is present in the contemporary cognitive sciences, even though their practitioners are sometimes unaware of this connection.¹ Nevertheless today, when we come across notions such as schema, prototype, model, and stereotype, they are certainly not comparable to the Kantian notion (they do not imply transcendentalism, for example), nor can these terms be understood as synonyms.

Moreover these cognitive “schemata” are usually intended to account for phenomena such as perception and the recognition of objects or situations, while we have seen that Kantian schematism, created to explain how judgments such as All bodies have weight are possible, fell short of the mark precisely when it had to explain how we manage to have empirical concepts. Cognitivism drew attention back to empirical concepts and recommenced wondering about the same things Locke (and Husserl too, basically) had wondered about: What happens when we talk about dogs, cats, apples, and chairs?²

But to say that cognitivism asks questions about cats and chairs does not mean to say that the conclusions it comes to (which are many and discordant) are satisfactory. The ghost of schematism haunts much contemporary research, but the mystery of this secret art has not yet been revealed.

Nor have I any pretensions to revealing it in these pages, also because, as we shall see, I would rather not poke my nose into the black box of our mind or brain processes. I shall ask myself only a few questions regarding the relations between a possible neo-schematism and semiotic notions of meaning, dictionary and encyclopedia, and interpretation.³

Given the erratic character I should like these reflections to have, I shall not always try to identify positions, theories, research, or the schools of thought within contemporary cognitivism. Instead I shall recount, as will be seen, lots of “stories” (mental experiments in narrative form) that exemplify some of the problems.

Most of my stories are about something fairly similar to what Kant held to be empirical concepts: I mean to say that I intend to deal with the way in which we speak (i) of objects or situations of which we have or might have direct experience (such as dog, chair, walking, eating out, climbing a mountain); (ii) of objects and situations of which we have no experience but could have (such as armadillo, or performing an appendectomy); (iii) of objects and situations of which someone has certainly had experience but we can no longer have, and regarding which the Community nevertheless transmits us sufficient instructions to speak of as if we had had experience of them (such as dinosaur and Australopithecine).

Dealing with such elementary phenomena from a semiotic standpoint poses first and foremost a preliminary question: whether there is any sense in talking of perceptual semiosis, that is, of a semiotic aspect of perception.

3.2 PERCEPTION AND SEMIOSIS

The problem of perception as a semiosic process has already made its appearance in 2. Of course, those unfamiliar with the Peircean standpoint will find this concept difficult (or quasi “imperialist”), because if we accept that even perception is a semiosic phenomenon, discriminating between perception and signification gets a little tricky.⁴ We have seen that Husserl also thought that perceiving something as red and naming something as red ought to be the same process, but this process might have diverse phases. Between perceiving a cat as a cat and naming it cat, or indicating it as an ostensive sign for all cats, is there not a jump, a gap (at least that shift from terminus a quo to terminus ad quern)?

Can we detach the phenomenon of semiosis from the idea of sign? There is no doubt that when we say that smoke is a sign of fire, the smoke we notice is not yet a sign; even if we accept the Stoic standpoint, smoke becomes a sign of fire not in the moment in which it is perceived but when we decide that it stands for something else. In order to pass on to that moment, we must leave the immediacy of perception and translate our experience into propositional terms so that the observation of smoke becomes the antecedent of a semiosic inference: (i) there is smoke, (ii) if there is smoke, (iii) then there is fire. The passage from (ii) to (iii) is a matter of inference expressed propositionally; while (i) is a matter of perception.

On the contrary, we speak of perceptual semiosis not when something stands for something else but when from something, by an inferential process, we come to pronounce a perceptual judgment on that same something and not on anything else.⁵

Let us suppose that someone with almost no knowledge of English but nonetheless accustomed to seeing English titles, names, or phrases on record sleeves, postcards, or various tinned goods receives a fax that, as often happens, has superimposed or distorted lines and illegible letters. Let us suppose (by transcribing the illegible letters as X) that he attempts to read Xappy neX Xear. Even without understanding the meaning of the words, he remembers seeing expressions such as happy, new, and year and presumes that these were the words the fax was intended to transmit. He will therefore have made inferences solely on the basis of the graphic form of the terms, from what was there on the sheet of paper (the expression plane) and not from what the words stood for (and so he would have to consult a dictionary).

Therefore any phenomenon, for it to be understood as a sign of something else and from a certain point of view, must first of all be perceived. The fact that the perception may be successful precisely because we are guided by the notion that the phenomenon is hypothetically understood as a sign (otherwise we would pay no attention to certain stimuli) does not eliminate the problem of how we perceive it.⁶

When the phenomenological tradition speaks of “perceptual meaning” it refers to something that legitimately precedes the constitution of meaning as the content of an expression; and yet (see A Theory of Semiotics 3.3) if I descry an indistinct animal form in the darkness, the success of the perception (the judgment That is a dog) is governed by a cognitive schema, something that I already know about the dog and that can legitimately be considered as a part of the content I usually assign to the word dog. In such a case I have made an inference: I have surmised that the indistinct form I descried in the darkness was a token of the type dog.

In the example of the fax, the letters -ear stand, in the inferential process, for the y they make it possible to hypothesize. The subject of our example possesses the (purely graphical) knowledge of at least one English word that could end with those letters and therefore guesses that -ear is an (incomplete) token of the lexical type denominated year. If on the other hand he has a good knowledge of English, he will also have the right to assume that the missing letter could be chosen from among b,d,f,g, h, n, p, r, t, and w (with each of which one can form an English word that makes sense), without being able to assume c, i, o, q, or u. But if he extends the inference to the whole syntagm Xappy neX Xear, he notices that one solution is more probable than the others, because he assumes that the whole string (incomplete in three places) is none other than Happy New Year (a stock phrase and a highly codified expression of good wishes).

We could then say that even in such an elementary process the token stands for its type. But what happens in the perception of unknown objects (such as the platypus)? The process is certainly more adventurous, that to stand for is contracted through processes of trial and error, but the relation of mutual referral from type to token is fixed once a perceptual judgment has been established.⁷

If (as is reiterated in Eco 1984) the basic characteristic of semiosis is inference, while the equivalence established by a code (a = b) is only a scleroticized form of semiosis, fully found only in ciphers (i.e., in the equivalences between one expression and the other, as in Morse code—see Eco 1984: 172–73), then the perceptual inference may be considered a process of primary semiosis.⁸

Naturally it might be decided that the question is wholly nominalistic. If it were established that semiosis occurred only when institutionalized sign functions appeared, then any talk of semiosis in the case of perception would be purely metaphorical—and in such a case we would have to say that so-called primary semiosis is only a precondition of semiosis. If this makes it possible to do away with pointless discussion, I have no problems in speaking of perceptual presemiosis.⁹ But things would not change that much, because, as we shall see in the following story, the relation between this primary phase and the successive development of full-fledged semiosis presents no evident fractures; rather, it constitutes a sequence of phases in which the preceding one determines the following one.

3.3 MONTEZUMA AND THE HORSES

The first Aztecs to hasten to the coast witnessed the landing of the conquistadors.¹⁰ Although only a very few traces of their first reactions remain and the best information we have depends on Spanish reports and indigenous chronicles written after the event, we know for sure that various things must have completely amazed them: the ships; the Spaniards’ awesome and majestic beards; the protective coverings that lent those fully armored “aliens” with their unnaturally white skins such a frightening air; the muskets and the cannons; and finally, apart from the ferocious dogs, those unheard-of monsters, the horses, in terrifying symbiosis with their riders.

The horses must have been no less perceptually puzzling than a platypus. At first (maybe also because they did not distinguish the animals from the pennants and armor that covered them), the Aztecs thought that the invaders were riding deer (and in so thinking they behaved just like Marco Polo). Oriented therefore by a system of previous knowledge but trying to coordinate it with what they were seeing, they must have soon worked out a perceptual judgment. An animal has appeared before us that seems like a deer but isn’t. Likewise they must not have thought that each Spaniard was riding an animal of a different species, even though the horses brought by the men of Cortes had diverse coats. They must therefore have got a certain idea of that animal, which at first they called magçatl, which is the word they used not only for deer but for all quadrupeds in general. Later, since they began adopting and adapting the foreign names for the objects brought by the invaders, their Nahuatl language transformed the Spanish caballo into cauayo or kawayo.

At a certain point they decided to send messengers to Montezuma to tell him of the landing and of the terrifying marvels they were witnessing. We have posterior evidence of the first message they sent to their lord: one scribe gave the news in pictograms, and he explained that the invaders were riding deer (magaoa, the plural of maçatl) as high as the roofs of the houses.

I don’t know whether Montezuma, confronted with such incredible news (men dressed in iron with iron weapons, perhaps of divine origin, equipped with prodigious instruments that hurled stone balls capable of destroying all things), understood what those “deer” were. I imagine that the messengers (worried about the fact that in their neck of the woods, if the news was not to the hearer’s liking, there was a tendency to punish the bearer of it) screwed up their courage and integrated the report with more than just words, since it seems that Montezuma was wont to require his informers to provide him with all the possible expressions for one and the same thing. And so they must have used their bodies to hint at the movements of the maçatl, imitating its whinnying, trying to show how it had long hair along its neck, adding that it was most terrifying and ferocious, capable in the course of the fray of overwhelming anyone who tried to withstand it.

Montezuma received some descriptions, on the basis of which he tried to get some idea of that as yet unknown animal, and goodness knows how he imagined it. That depended both on the skill of the messengers and on his agility of wit. But he certainly understood that it was an animal, and a worrisome one too. In fact, still according to the chronicles, at first Montezuma did not ask other questions but withdrew into a distressing silence, with head bowed and wearing an absent, sorrowful air.

Finally the encounter between Montezuma and the Spaniards came to pass, and I would say that, no matter how confused the messengers’ description may have been, Montezuma must have easily identified those things called magaoa. Simply, faced with the direct experience of the maçatl, he must have adjusted the tentative idea he had conceived of them. Now, like his men, every time he saw a maçatl, he too would recognize it as such, and every time he heard talk of maçaoa, he would understand what his interlocutors were talking about.

Then, as he gradually got to know the Spaniards, he would learn many things about horses, he would begin to call them cauayo, he would learn where they came from, how they reproduced, what they ate, how they were reared and trained, what other uses they could be put to, and to his regret he would very soon understand how useful they could be in battle. But according to the chronicles, he must also have harbored a suspicion regarding the invaders’ divine origins, because he was told that his men had managed to kill two horses.

At a certain point the learning process whereby Montezuma was gradually increasing his knowledge of horses stopped, not because he could not learn any more but because he was killed. And therefore I will leave him (and the great number of those who were massacred along with him for having had the revelation of Horse-hood) in order to observe that in this story a great number of different semiotic phenomena come into play.

3.3.1 The Cognitive Type (CT)

At the close of their first perceptual process, the Aztecs elaborated what we shall call a Cognitive Type (CT) of the horse. If they had lived in a Kantian universe, we should say that this CT was the schema that allowed them to mediate between the concept and the manifold of the intuition. But for an Aztec where was the concept of horse, given that he did not have one before the Spaniards landed? Of course, after having seen some horses, the Aztecs must have constructed a morphological schema not that dissimilar to a 3-D model, and it is on this basis that the coherence of their perceptual acts must have been established. But by speaking of a CT, I do not mean just a sort of image, a series of morphological or motor characteristics (the animal trots, gallops, rears); they had perceived the characteristic neigh, and perhaps the smell, of horses. Apart from the appearance, the Aztecs must have immediately attributed a characteristic of “animality” to the horse, given that the term maçatl was immediately applied, as well as the capacity to inspire terror and the functional characteristic of being “rideable,” since it was usually seen with human beings on its back. In short, let’s say that the CT of the horse was of a multimedial nature right from the start.

3.3.1.1 The Recognition of Tokens

On the basis of the CT thus elaborated, the Aztecs must have been immediately able to recognize as horses other exemplars that they had never seen before (and this apart from variations in color, size, and vantage point). It is precisely the phenomenon of recognition that induces us to talk of type, in fact, as a parameter for the comparison of tokens. This type has nothing to do with an Aristotelian-Scholastic “essence,” and we have no interest in knowing what the Aztecs grasped of the horse (perhaps wholly superficial features of a kind that left them unable to discriminate between horses and mules or donkeys). But it is certain that by talking of type in this sense, we conjure up the ghost of a Lockian type of “general ideas,” and some might object that we have no need of these to explain...