A possible mechanism that can explain the ability of developing categories is investigated. Even if categorization might well rely on a multiplicity of mechanisms, the one proposed here appears to play a key role, especially in the categories built in the first period of life. This mechanism, here named coincidence detection and coding, is motivated by the state-of-the-art in the neurobiology of learning, is coherent with evidence given by developmental psychology, and is supported by results of neurocomputational models. One of its main aspects is the special sensitivity of organisms in detecting that two or more elements, in the stream of stimuli from the environment, happen with remarkable synchrony. The other is the ability of the organism to code traces of repeated coincidences of the detected elements. One advantage is that subsequent experiences of just one (or a few) of the coded elements might elicit the activation of the internal representation of the other(s).