What drove the apes to abstraction?

The evolution of language raises two kinds of questions. Why did language evolve: what were the selective advantages of language? And how did language evolve: what morphological, cognitive and social changes had to occur to make possible the transition from indexical communication to symbolic communication?

Indexical communication is typical of animals and involves signals that cue their referents. For example, the roar of the stag deer is correlated with its vigour because the roar is energetically costly to produce. In contrast, symbolic communication, which is largely limited to humans, is defined by arbitrary connections between referents and signals. Nothing about "chair" as a word predicts the chair itself, although the two become conventionally associated through usage.

The advantages of symbolic communication are substantial. Consider that in human language, signals come to represent objects and events divorced from their contingent context. This is most striking for spatial and temporal correlations, such as "I saw a lion yesterday" or "John is in the tree at the end of the road", which stand in contrast to simply being able to yell, "Lion!" This required a transition from a signalling system in which communication was possible only about the "here and now", as in the "Lion!" case, to more abstract concepts, such as where the lion was yesterday. Thus, there are two issues: the evolution of a signalling system in which signal and referent are arbitrarily related and the evolution of a signalling system in which it is possible to signal about abstract concepts.

This second issue is really the subject of Stanley Greenspan's and Stuart Shanker's book, The First Idea . In particular, Greenspan and Shanker, a psychiatrist and a philosopher respectively, are interested in how and when the capacity for symbolic thought first arose. There is a huge body of work on these questions (Terrence Deacon's excellent book The Symbolic Species comes to mind). To a large extent the capacity for symbolic thought has been studied in the context of symbolic communication - typically vocal communication. Consequently, emphasis has been on the morphological changes to the vocal apparatus that had to occur to make speech possible. Increasingly, evolution of language research also emphasises the neural changes - changes to brain anatomy - that had to occur to make speech production, processing and perception, as well as symbolic representation, possible.

The other major line of inquiry, which has been more speculative, is the role that gene change has played in the evolution of language. This debate was in many senses sparked by the work of Noam Chomsky, who argued that certain features (principles) of language are universal, and that this universality suggests strong genetic underpinnings. Although Chomsky's universality notion has been questioned, at least a few features of language are universal in humans, and many of the cognitive capacities underlying the language faculty are influenced by changes at the genetic level.

Nonetheless, it need not be the case that gene change alone accounts for the evolution of cognitive capacities. For example, as Deacon points out in The Symbolic Species , processing long sets of indexical associations (fire for hot, hot for danger, and so on) into symbols (fire for danger) would have initially been burdensome for a poorly developed prefrontal cortex. Such processing would require longer attention spans - a computationally costly endeavour from the neural perspective. Thus behavioural plasticity (variation in capacity for long attention) sets the stage for exploring genetic space that might otherwise have been inaccessible. This is called the Baldwin effect.

A related question, which remains poorly addressed, is whether non-genetic inheritance mechanisms, such as social learning, have played a role in the evolution of genes affecting the capacity for language. Similarly, little attention has been paid to the role that emotional processes, at either the neural or behavioural level, might have played in the evolution of language.

Greenspan and Shanker tackle these issues. Their research, which has focused not on the evolution of language but on its acquisition during the course of child development, has strongly influenced their perspective. They note the importance of interation between an infant and its caregiver to a child's ability to think and communicate symbolically, and argue that it has played an important role in language evolution. Basically, they hypothesise that through emotionally mediated signalling interactions, the caregiver is able to shape an infant's thought processes towards greater abstraction and away from the here and now. Refinement of this skill over thousands of years, with each generation passing it down to the next through some kind of sociocultural inheritance, was, according to the authors, the critical factor allowing humans to acquire language.

Presumably this hypothesis rests on the assumption that the increasing intergenerational emphasis on emotionally mediated signalling changed the selective environment to favour other kinds of traits also conducive to the evolution of language. In this way, Greenspan and Shanker invert the phrase "ontogeny recapitulates phylogeny" to "phylogeny recapitulates ontogeny" - phylogenetic changes mirror increasingly complicated developmental requirements.

Here is an interesting and timely idea that gels with work in other areas of evolutionary biology on niche construction - that is, how individuals through modification of their environments are able to change the selection pressures to which they are subject. Unfortunately, none of this is clearly fleshed out in The First Idea . Instead, the book spends much time exploring the implications of emotionally mediated signalling (called "co-regulated affective signalling" by the authors) on a very large number of, at best, loosely related phenomena. These include individual development (from birth to death), the formation of group identity and taboos, the rise of Sumerian city-states and global interdependency.

What could have been an interesting treatment of factors often neglected in debates about the evolution of language is instead reduced to a theory of everything.

Jessica Flack is a postdoctoral fellow at the Santa Fe Institute, Santa Fe, New Mexico, US.