Convergence between G and g in three monkey species (Sapajus spp, Ateles geoffroyi, and Macaca fascicularis).

Abstract

An open question in comparative psychology is whether the source of correlations among different measures of ability (the g factor) is shared between species, or is distinct. This is examined using data on the performance of three monkey species (tufted capuchins, black-handed spider monkeys, and longtailed macaques) on 16 cognitive ability measures. The differences between species pairs across measures are not generally strongly related to the degree to which the g factor loads on each subtest. Iteratively removing the subtests with the lowest coefficients of variance (CV), and recorrelating the species differences with subtest g-loadings was found to increase the association between the two. Across iterations, subtest pooled CV strongly and positively predicts the increase in the degree to which g-loadings are predictive of species difference in two comparisons, but is a weaker predictor in the comparison between tufted capuchins and black-handed spider monkeys. These associations were not related to phylogenetic distance but were very strongly related to species differences in the means of certain ecological factors. g-scores computed for each species on the basis of the three subtests with the highest CV values exhibited very-high magnitude (..9) associations with species-level G-scores. G is simply the species-level equivalent of g, resulting from covariation among cognitive ability measures at the level of species differences. Finally, as with previous studies, subtests that show the greatest species differences seem to draw on executive functioning, and attention, suggesting that these may be a phylogenetically conserved source of g across many animal taxa.