The competition aspects of Calvin's model is extensible to population dynamics even at a very early stage in the evolution of language.
Imagine a time when there is no language as such, only utterances or even some type of linguistic noise that could constitute very early proto-language. This proto-language may employ previously used actions that are disarmed through a kind of quoting mechanism, such as we have described in previous chapters. These might be disarmed linguistic vocal responses. These linguistic vocal responses may constitute proto-vocables in which a kind of proto-attribution occurs. This attribution may be described as the kind of competition metaphors present. However these are not quite metaphors, despite the fact that proto-attribution represents an adaptation of previous uses of linguistic vocables. A proto-attribution also involves direct visual stimulus that differentiates the linguistic vocables from their original context of use. Hence the competition for neural space. As in the case of the metaphor this proto-attribution may require a vocal transaction to resolve the competition. A system of schematization could have promoted linguistic novelties early on in proto-attribution. Imagine that the category fruit does not exist, but that the shared experience of the category oranges does. Imagine further that an utterance exists that is characteristic of the presence of oranges, but that there is none characteristic of the presence of kiwi fruits. Imagine further that a kiwi fruit is presented by one member of a late hominid species to another member of the species. However there is no specific utterance that exists to describe classifiable features of a kiwi in particular or fruit in general. The strategy may be to use the utterance for the experience of orange and modify it somehow. If this is an attribution at all, it is certainly a wrongful attribution, however a gesture that mimics eating presents a competition between the wrongful proto-attribution of orange and the visual cue. The other member may infer something about eating, perhaps sweetness and some vague aspect of shape. There may be some back-and-forth gesturing to establish a common experience of the kiwi fruit and, in the process, modify the proto-attribution of orange to include - and exclude - perceptual features that will specifically acknowledge similarities and differences between oranges and kiwi fruits.
We suggest that competition between perceptual structures of groups of individuals bring about conventionalization. The transaction is of the sort that we might fancifully suppose would have been resolved by the overlapping synaptic structures emerging from exchanged reports of previous experiences, reports that would focus particular features. Hence the competition for neuron space may partly be a competition engaged in an arena involving two or more brains rather than one. We imagine that it is this sort of interaction that generates what we commonly understand as conventional use. On this view, the competition between vocal stimuli and visual cues provides a mix-and-match resolution problem that promotes a kind of generalization; or perhaps, proto-generalization since the vocal and visual cues may not yet constitute anything that we should want to label linguistic.