In the study, the authors investigated the effect of the words in the categorical perception effect that manifests itself in deterioration of the memorisation of individual characteristics of the objects when they belong to a single category. In the experiment, the subjects memorised the look of their new objects (images of butterflies) that had been previously associated with artificial names. The authors varied the degree of attention of the subjects to objects’ names. In one group, with the pronunciation of names, subjects were instructed that the butterflies were called by two names (artificial words: "tulnitsa" and "daryanka"), and they had to pronounce the names. In the other group, with an indication of the location, in addition to the name the subjects received a clue of where a butterfly would be and had to read aloud only the word for the location. According to the authors’ hypothesis, when accompanying actions with objects words do not always create the categorical perception effect, but rather only when a person's attention is directed to the name. As it turned out, the names led to the deterioration of memorisation of individual objects’ properties, but only when they were spoken aloud by the subjects; when the subjects did not pronounce the names, but instead the indication of the location was spoken aloud, the words did not create any categorical perception effect. In contrast to previous experiments (Lupyan, 2008; Kotov et al., 2012), the authors proved that in the perception of objects, the word alone ss not enough to create a categorical perception effect. In the case where a person's attention is directed to a verbal indication of the location of the object, the name of the object, which remains outside the attention focus, does not lead to the perception of the objects as a group.