Purchase this article with an account.
C. Nilson, K. Joneleit, R. Smith, D. D. Hoffman; Color and part. Journal of Vision 2001;1(3):91. doi: 10.1167/1.3.91.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
We often learn the names of new objects and parts by ostensive definition — someone points and names the object or part. However, ostensive definition is ambiguous — the use of a word in some visual scene provides evidence consistent with many hypotheses, only one of which will usually be correct. Since word learning is fairly rapid and error-free, there probably are underlying principles or biases that we all use when learning the names for new objects and parts. Recent research indicates that learning new names for parts is guided by the minima-part bias, which uses the intrinsic geometry of shapes to define psychologically natural parts. It is proposed that perceptually salient color may also contribute to the learning of names for new parts. Experiment 1 shows that salient color can override the minima-part bias when minima boundaries are weak. Experiment 2 shows that when minima boundaries are weak and color is perceptually salient, subjects generalize the name for a part to a new color part that differs from the original by a translation or uniform scaling. Experiment 3 investigates the effects of order of stimulus presentation. Experiment 4 investigates the strength of color salience across the color spectrum.
This PDF is available to Subscribers Only