In defense of categories.

Ritchie et al. (this issue) argue that the complex landscape of selectivity in the human occipitotemporal cortex is better understood in terms of behavioral relevance than in terms of category. At the descriptive level, their suggestion is obviously very powerful because it allows to describe almost any type of selectivity that is found. Yet, it does not seem to suffice to explain the selectivity and its properties. Categories still play a crucial role in clarifying the functional organization of object representations in the brain, including the tuning for dimensions and the relationship to behavioral relevance.