Asymmetries in infants' attention to the presence or absence of features.

Three experiments were performed to test whether infants show a bias for detecting the presence of a feature in a stimulus rather than its absence. In the 1st experiment, 24 16-week-old infants were given 3 paired-comparison problems, each of which included a 25-s familiarization phase followed by 2 test trials. Infants were familiarized to 1 member of a set of capital alphabetical letters (E-F; Q-O; B-R). Then they were given a paired-comparison recognition test under 1 of 2 conditions. In the feature-present condition, the familiar letter (e.g., F) was paired with a novel letter containing the addition of a distinguishing element (e.g., E). In the feature-absent condition, infants were presented with a familiar letter (e.g., E) paired with a novel letter in which 1 element was removed (e.g., F). Infants showed a novelty preference to the letter in which the distinguishing feature was present, but there was no preference for novelty in the feature-absent condition. The 2nd experiment showed that infants' fixation to the letter containing the presence of the feature was not due to a simple preference for the letter with the greater number of elements. Finally, to test whether infants' failure to discriminate the absence was due to insufficient encoding time, 36 infants were tested in a 3rd experiment in which familiarization time was varied. After 20 s of familiarization, no evidence of discrimination was observed in either the feature-present or feature-absent condition. After 30 s, however, infants could discriminate the novel letter in the feature-present condition but not in the feature-absent condition. The significance of these results is discussed in terms of theoretical explanations for the development of the feature-presence bias.