These mental pictures of the world are most vividly triggered by perception and action. While an ape is attending to a scene, its structure is most vividly present to its mind, partially structured top-down by the animal's selective attention to particular objects and particular properties of theirs, in preference to other objects and properties, which are not attended to. To a limited extent, animals are capable of bringing scenes to mind in the absence of direct perception. Animals can plan and remember, and can make limited inferences over mentally represented scenes. Though made familiar by perception, and instantiated in the brain by sensory-motor `functional clusters' (Gallese & Lakoff, 2005) the mental representations of immediately prelinguistic creatures can be brought to mind during acts of planning and remembering. To this extent, they are cognitive, as well as perceptual, representations. Some animals show the faint glimmerings of episodic memory, a well-developed faculty in humans, by which non-current scenes and events can be brought to mind.
Many animals are capable of entertaining somewhat abstract relations and properties, such as SAME/DIFFERENT and OPPOSITE. Evidence of these can be elicited in experimental situations. It seems likely that the mental lives of animals in the wild seldom involve such abstract predicates; even more abstractly, properties-of-properties and relations-between-relations, are likely only to be available after some language training, or training with language-like objects, by humans.
The semantic predicate-argument structure attributed to animals is emphatically not to be confused with any linguistic structure, such as the Subject-Predicate structure of sentences. It is just unfortunate that the term predicate is ambiguous, applying to a structural constituent of a sentence, and to items in the vocabulary of a logical language only indirectly related to language structure.
The pre-linguistic mental representations so far discussed are all activated in individual heads and find no outlet (as yet in the story) in public communication of any kind. Pre-linguistic animals can, in two separate and very limited ways, communicate to others about various states of the world. The largely innate alarm calls of vervet monkeys and many other species are one way. The limited ability to make and interpret deictic pointing gestures is the other way. Pointing is almost totally absent in the wild, and brought out mainly in conditions of captivity and interaction with cooperative humans. In these very limited capacities of animals we can see the faint beginnings of communicative reference.
Apes have rich mental lives, but keep their pictures of the world to themselves, like all other animals besides humans. Only humans tell each other in detail about events and scenes in the world. And this is something of an evolutionary puzzle, because giving information away would seem prima facie to be against the individual interests of the information-giver. Evolutionary theory has come up with a number of solutions to the riddle of why humans should, in general, act relatively altruistically, with this (apparent) altruism manifested saliently in language. No one theory on its own, such as Kin Selection, Reciprocal Altruism or Sexual Selection, can adequately explain the unique human characteristic of freely giving information in such structurally complex ways as we do every day with language. A complex egalitarian social structure with shifting alliances, some degree of monogamy and equality betwen the sexes, and extended parental care of the young provides the most hospitable environment in which group-wide communicative codes can come to be adopted. Some degree of understanding of other minds is also required; some very limited evidence (Hare et al. 2001) has only recently become available showing that in some circumstances, a chimpanzee can know what another animal knows. But this is not enough, and I have adopted Tomasello et al's (2005) concept of shared intentionality as a key ingredient of humans' striking willingness to play complex language games with each other. This involves a degree of trust, a social attitude which is particularly well developed in humans. Our ape cousins have not evolved to exhibit shared intentionality or the appropriate degree of trust paving the evolutionary way for language.
We have reached a watershed. As soon as the breakthrough was made for animals to communicate their thoughts relatively freely to others, a cascade of other innovations were selected, designed to make the transfer of information more effective, and allowing humans to enrich their mental representations by thinking about the representations themselves, in what Quine (1960) has called `semantic ascent'. This cascade of consequences will be the topic of Volume 2. In it we will see examples of the evolutionary force of self-organization at work. Given creatures with a disposition to transmit messages to each other, there are pressures for these messages to become structured in particular ways, so as to be most easily learned and used by the creatures concerned. In this way, as will be seen in Volume 2, phonological structure and morphosyntactic structure evolved. In a co-evolutionary arms race, human capacities for efficient processing of these structures co-evolved with increasing complexity of the structures themselves.
I hope that when you have read the present volume, Volume 2 will be available or on the way, and that you will have an appetite to read it. For the present, thank you for joining me this far.