camouflaged Amazonian insect known as the walking stick in English because its elongated torso looks so much like a twig. Its Quichua name is shanga. Entomologists call it, appropriately, a phasmid—as in phantom—placing it in the order Phasmida and the family Phasmidae. This name is fitting. What makes these creatures so distinctive is their lack of distinction: they disappear like a phantom into the background. How did they come to be so phantasmic? The evolution of such creatures reveals important things about some of the “phantomlike” logical properties of semiosis that can, in turn, help us understand some of the counterintuitive properties of life “itself”—properties that are amplified in the Amazon and Runa ways of living there. For this reason, I will return to this example throughout the book. Here I want to focus on it with an eye to understanding how the different semiotic modalities—the iconic, the indexical, the symbolic—have their own unique properties at the same time that they stand in a relation of nested continuity to each other.
How did walking sticks come to be so invisible, so phantomlike? That such a phasmid looks like a twig does not depend on anyone noticing this resemblance—our usual understanding of how likeness works. Rather, its likeness is the product of the fact that the ancestors of its potential predators did not notice its ancestors. These potential predators failed to notice the differences between these ancestors and actual twigs. Over evolutionary time those lineages of walking sticks that were least noticed survived. Thanks to all the proto–walking sticks that were noticed—and eaten—because they differed from their environments walking sticks came to be more like the world of twigs around them.19
How walking sticks came to be so invisible reveals important properties of iconicity. Iconicity, the most basic kind of sign process, is highly counterintuitive because it involves a process by which two things are not distinguished. We tend to think of icons as signs that point to the similarities among things we know to be different. We know, for example, that the iconic stick figure of the man on the bathroom door resembles but is not the same as the person who might walk through that door. But there is something deeper about iconicity that is missed when we focus on this sort of example. Semiosis does not begin with the recognition of any intrinsic similarity or difference. Rather, it begins with not noticing difference. It begins with indistinction. For this reason iconicity occupies a space at the very margins of semiosis (for there is nothing semiotic about never noticing anything at all). It marks the beginning and end of thought. With icons new interpretants—subsequent signs that would further specify something about their objects—are no longer produced (Deacon 1997: 76, 77); with icons thought is at rest. Understanding something, however provisional that understanding may be, involves an icon. It involves a thought that is like its object. It involves an image that is a likeness of that object. For this reason all semiosis ultimately relies on the transformation of more complex signs into icons (Peirce CP 2.278).
Signs, of course, provide information. They tell us something new. They tell us about a difference. That is their reason for being. Semiosis must then involve something other than likeness. It must also involve a semiotic logic that points to something else—a logic that is indexical. How do the semiotic logics of likeness and difference relate to each other? Again, following Deacon (1997), consider the following schematic explanation of how that woolly monkey that Hilario and Lucio were trying to frighten out of her hidden canopy perch might learn to interpret a crashing palm as a sign of danger.20 The thundering crash she heard would iconically call to mind past experiences of similar crashes. These past experiences of crashing sounds share with each other additional similarities, such as their co-occurrence with something dangerous—say, a branch breaking or a predator approaching. The monkey would in addition iconically link these past dangers to each other. That the sound made by a crashing tree might indicate danger is, then, the product of, on the one hand, iconic associations of loud noises with other loud noises, and, on the other, iconic associations of dangerous events with other dangerous events. That these two sets of iconic associations are repeatedly linked to each other encourages the current experience of a sudden loud noise to be seen as linked to them. But now this association is also something more than a likeness. It impels the monkey to “guess” that the crash must be linked to something other than itself, something different. Just as a wind vane, as an index, is interpreted as pointing to something other than itself, namely, the direction in which the wind is blowing, so this loud noise is interpreted as pointing to something more than just a noise; it points to something dangerous.
Indexicality, then, involves something more than iconicity. And yet it emerges as a result of a complex hierarchical set of associations among icons. The logical relationship between icons and indices is unidirectional. Indices are the products of a special layered relation among icons but not the other way around. Indexical reference, such as that involved in the monkey’s take on the crashing tree, is a higher-order product of a special relationship among three icons: crashes bring to mind other crashes; dangers associated with such crashes bring to mind other such associations; and these, in turn, are associated with the current crash. Because of this special configuration of icons the current crash now points to something not immediately present: a danger. In this way an index emerges from iconic associations. This special relationship among icons results in a form of reference with unique properties that derive from but are not shared with the iconic associational logics with which they are continuous. Indices provide information; they tell us something new about something not immediately present.
Symbols, of course, also provide information. How they do so is both continuous with and different from indices. Just as indices are the product of relations among icons and exhibit unique properties with respect to these more fundamental signs, symbols are the product of relations among indices and have their own unique properties. This relationship also goes only in one direction. Symbols are built from a complex layered interaction among indices, but indices do not require symbols.
A word, such as chorongo, one of the Ávila names for woolly monkey, is a symbol par excellence. Although it can serve an indexical function—pointing to something (or, more appropriately, someone)—it does so indirectly, by virtue of its relation to other words. That is, the relation that such a word has to an object is primarily the result of the conventional relation it has acquired to other words and not just a function of the correlation between sign and object (as with an index). Just as we can think of indexical reference as the product of a special configuration of iconic relations, we can think of symbolic reference as the product of a special configuration of indexical ones. What is the relationship of indices to symbols? Imagine learning Quichua. A word such as chorongo is relatively easy to learn. One can learn that it refers to what in English is called a woolly monkey quite quickly. As such, it isn’t really functioning symbolically. The pointing relationship between this “word” and the monkey is primarily indexical. The commands that dogs learn are very much like this. A dog can come to associate a “word” like sit with a behavior. As such, “sit” functions indexically. The dog can understand “sit” without understanding it symbolically. But there is a limit to how far we can go toward learning human language by memorizing words and what they point to; there are just too many individual sign-object relationships to keep track of. Furthermore, rote memorization of sign-object correlations misses the logic of language. Take a somewhat more complex word like causanguichu, which I discussed earlier in this chapter. Non-Quichua speakers can quickly learn that it is a greeting (uttered only in certain social contexts), but getting a sense of what and how it means requires us to understand how it relates to other words and even smaller units of language.
Words like chorongo, sit, or causanguichu do of course refer to things in the world, but in symbolic reference the indexical relation of word to object becomes subordinate to the indexical relation of word to word in a system of such words. When we learn a foreign language or when infants acquire language for the first time there is a shift away from using linguistic signs as indices to appreciating them in their broader symbolic contexts. Deacon (1997) describes one experimental setting where such a shift is particularly apparent. He discusses a long-term lab experiment in which chimps, already adept in their everyday lives at interpreting signs indexically, were trained to replace this interpretive strategy with a symbolic one.21
First,
