THE FIRST chapter discussed why universal psychometrics is needed, what it should address and why this is a good opportunity. In this chapter we discuss how all this can be achieved. We need to analyse how behaviours are to be classified and how behavioural features can be defined and measured. For this purpose, we need to determine the nature of these behavioural features and what the role of measurement is. In fact, we will see that there is a choice between two main kinds of measurement, pragmatic and representational, with very different explanatory power, applicability and accuracy of measurement. The methodology to analyse the space of behavioural features also depends on this choice, as also do the theoretical and experimental tools.

TAXONOMIES, MEASUREMENTS AND DEFINITIONS

The scala naturae, the great chain of being, was the predominant theory for the classification of organisms from Antiquity to the Middle Ages. Despite its limitations, Aristotle's History of Animals was the embryo of all subsequent taxonomies of living things, including Linnaeus's Systema Naturae, the father of modern biological taxonomy, and had a strong influence in early evolutionists, such as Lamarck, Blumenbach or even Darwin. Blumenbach viewed scala naturae as a basis “to order the natural bodies according to their greatest and most manifold affinity, to bring together the similar ones, and to remove from each other the dissimilar ones” (Blumenbach, 1790, pp. 8–9 cited in Mayr, 1982, p. 201). The principle of similarity could appear in terms of a self-evident affinity, such as the similarities between parrots and pigeons, or through analogy, such as the similarities between seals and penguins. In the end, the use of different concepts of similarity would lead to different taxonomies. Actually, many taxonomies use several features at a time, and taxonomists may have heated debates about which feature is most relevant. It is only very recently that species can also be classified through phylogenetic analysis using DNA information.