In Part II of this book, two windows on the evolution of language – the shell-bead window and fossil-skull window – are to be discussed as instances of the type of windows that has been dubbed ‘correlate windows’. So a natural question is: ‘What is it that makes a particular window a correlate window?’ In a nutshell: it is the nature of the inferential steps permitted by it. To see this, consider again the skeletal characterisation of the shell-bead window and the fossil-skull window provided in Section 1.2. In the case of both windows, the three inferential steps link the entities concerned in terms of correlation. What this involves is made concrete in Figure 3.1 in Section 3.1 with reference to the shell-bead window. In this figure, that is, the tick shells represented in block A are taken to be correlates of the beads represented in block C; these beads are taken to be correlates of the symbolic behavior represented in block E; and this behavior is taken to be a correlate of the fully syntactic behavior represented in block G.’

But what does it mean to say that two entities (or variables) X and Y are correlates? On a conventional definition, it means that the link between X and Y is such that when one of them changes, the other changes too. Correlations between two entities can be manifested in various ways. In the case of the shell-bead window, for instance, fully syntactic language has been considered ‘an essential requisite for’ symbolic behaviour, a point to be discussed at some length in Section 3.4. ‘Being an essential requisite for’ is but one of the ways in which a correlation can be manifested. Two entities X and Y can be correlated in various other ways, including the following: ‘X causes Y’ (or ‘X results from Y’), ‘X precedes Y in time or some other dimension’ (or ‘X follows Y’), ‘X is located in the proximity of Y’ and so on. The question of the precise nature of the correlation between tick shells and Middle Stone Age (MSA) beads, between MSA beads and symbolic behaviour, and between symbolic behaviour and fully syntactic language will be dealt with in some detail in Section 3.4.