For example, the Ground should not even be an icon, if the icon is likeness, because it cannot have a relation of likeness with anything if not with itself. Here Peirce is swinging between two notions: in one sense, as we have seen, the Ground is an idea, a skeleton plan, but if it is such, it is already an Immediate Object, a full realization of Thirdness; in another sense it is a Likeness that does not resemble anything. All it says to me is that the sensation I feel is in some way emanated by the Dynamical Object.
In this case we must liberate (even if this means going against Peirce, who, by changing the term every time, confuses our ideas) the concept of likeness from the concept of comparison. Comparison occurs in relations of similitude, when on the basis of a given proportion we say—of a graph, for example—that it expresses certain relations that we must suppose in the object. Similitude (already riddled with laws) explains the workings of hypoicons, such as diagrams, drawings, pictures, musical scores, and algebraic formulae. But the icon is not explained by saying that it is a similitude, nor even by saying that it is a likeness. The icon is a phenomenon that founds all possible judgments of likeness, but it cannot be founded on likeness itself.
Therefore it would be misleading to think of the icon as a mental «image» that reproduces the qualities of the object, because in that case it would be easy to abstract a general image from many particular images, just as one abstracts (however this happens) a general idea of bird or tree from many birds or many trees. I do not wish to say that mental images must not be admitted or that in certain moments Peirce did not think of the icon in terms of a mental image. I am saying that in order to conceive the concept of primary iconism, the one that establishes itself in the moment of the Ground, we have to abandon even the notion of mental image.26
Let us try to eliminate the mental facts and make a mental experiment instead. I have just got up and, still half asleep, I put the coffee percolator on the gas. Probably I have put the gas on too high, or I haven’t put the percolator in the right place, but the fact remains that the handle has got too hot, and when I pick up the percolator to pour the coffee, I burn myself. Standard expletives emitted (and deleted), I protect my fingers and pour the coffee. End of story. But the next morning, I make the same mistake. If I were to verbalize the second experience, I would say that I put the same percolator on the gas and that I had the same painful sensation. But the two types of recognition are different. Establishing that the percolator is the same is the effect of a complex system of inferences (full Thirdness): I could have (as I do have) two percolators of the same type, one older and one newer, and establishing which of the two I picked up implies a series of recognitions and conjectures regarding some morphological characteristics of the object, and even the memory of where I put it the day before.
But «feeling» that what I feel today is the same (with some negligible variations in thermic intensity) as what I felt yesterday is quite another kettle of fish. I am pretty sure I have the same impression of burning, or, rather, I feel a painful thermic sensation that in some way I recognize as similar to that of the previous day.
I don’t think many inferences are required to activate this recognition. The handiest solution would be that the previous experience has left a «trace» in my neural circuits. But there is the risk of already considering this trace as a schema, a prototype of the feeling, a rule for recognizing similar feelings. Let us accept, if we wish, an idea current in neoconnectionist circles, whereby it is not necessary for the neural network to construct a prototype of the category, and there is no distinction between rule and data (i.e., the memory of the stimulus and the memory of the rule would have the same configuration, the same neural pattern). Even more modestly, we can assume that, the moment I feel the sensation of pain, a point in my nervous apparatus is activated that is the same one activated the day before and that this point, in activating itself, in some way makes me feel, along with the sensation of heat, a feeling of «again.» I am not even sure if one must presuppose a memory, if not in the sense that if on one occasion we have suffered an injury in one part of the body, the body has retained a «memory» of the injury and, when a new injury occurs, it reacts in a different way than if it were hurt in a part still undamaged. It is as if the first time I noticed a sensation of «heat1» and the second time a sensation of «heat2»
Gibson (1966: 278), while maintaining that the idea that feelings leave a trace is fairly reasonable and, all things considered, handy, and that the input present must in some way reactivate the trace left by the previous experience, nevertheless observes that an alternative explanation would be that the judgment of likeness between stimuli reflects a concordance between the perceptual system and the invariants of the informative stimulus. No trace, no preliminary «schema,» simply something that we cannot but call conformity.
It is not that we have plunged back into a theory of knowledge (or at least of its sensory anteroom) as adaequatio. We are dealing with a matter of simple correspondence between stimulus and response. Therefore, we do not have to tackle all the paradoxes in a theory of correspondence at higher cognitive levels: if, on perceiving a dog, we find it corresponds to our schema of dog, we must ask ourselves on what basis the judgment of correspondence is founded, and in seeking the model of correspondence, we enter the spiral of the Third Man. On the contrary, identity, the statistical correspondence between stimulus and response, tells us that the response is exactly that caused by the stimulus.
What does correspondence mean in this case? Let us suppose that somebody managed to record the process that comes about in our nervous system every time we receive the same stimulus, and that the record always has the configuration x. We should say therefore that x corresponds adequately to the stimulus and is the icon of it. Let us say, then, that the icon evinces a likeness with the stimulus.
This correspondence we have decided to call likeness has nothing to do (yet) with an «image» that corresponds point by point with the characteristics of the object or the field of stimulus. As Maturana (1970: 10) reminds us, two states of activity in a given nerve cell can be considered the same (i.e., as equivalents) if «they belong to the same class» and are defined by the same pattern of activity, without their having to possess the nature of a map with one-to-one correspondences. For example, let us accept as good Fechner’s Law, whereby the intensity of a sensation is proportional to the logarithm of the excitatory influence. If this were the case, and if the proportion were constant, the intensity of excitation would be the icon of the excitatory influence (in the formula S = K log R, the equals sign would express the relation of iconic similarity).
I think that for Peirce primary iconism lies in the correspondence whereby the stimulus is adequately «represented» by that sensation and not by another. This correspondence is not to be explained but only recognized. That is why it is the icon that becomes a parameter of similarity and not vice versa. If from that moment we mean to talk of other and more complex relations of similarity, or of calculated relations of similitude, it is on the basis of the model of that primary likeness, the icon, that we establish what it means, evidently in a sense that is less immediate, rapid, beyond dispute, to be similar to27
In 6.11 we shall see that a relation of this kind, nonmediated, beyond dispute (provided there is no interference from elements capable of «fooling» the senses), occurs in mirror images. But here I prefer to avoid recourse to an image of any nature, precisely to free the notion of iconism of its historic ties with visual images.
2.8.2 The Lower Threshold of Primary Iconism
If it is possible to define primary iconism in nonmental terms, it is because within Peirce’s thinking there is an encounter between two different but mutually dependent perspectives: the metaphysical-cosmological standpoint and the cognitive standpoint. Unless they are read from a semiotical point of view, Peirce’s metaphysics and cosmology are certainly incomprehensible; but the same thing ought to be said of his semiotics with regard to his cosmology. Categories such as Firstness, Secondness, and Thirdness, and the concept of interpretation itself, not only define modi significandi—that is, the ways in which the world can be known—they are also modi essendi, ways in which the world behaves, the procedures through which the world, in the course of evolution, interprets itself.28
From the cognitive standpoint the icon, seen in its nature as pure quality, state of awareness, absolutely unrelated, is a Likeness, because it corresponds to what called it
