Perceptual learning: An analysis based on selective attention measurements

Abstract

An experiment was carried out to assess the relative merits of the differentiation and association theories of perceptual learning. Two Hebrew and two Roman characters were used to form stimuli describable by three orthogonal dimensions: form, system, and name. Two groups of subjects, who differed in their ability to read Hebrew, did a series of card sorting tasks. The results indicated that the subjects who could read Hebrew were able to use the name dimension in a classification task, whereas the non-Hebrew readers were not. This finding shows that long-term associative learning can facilitate perceptual classification.To explain the role of learning in perception, J. J. Gibson and E. J. Gibson (1955) argued that there are essentially two theories of perceptual learning. According to the association theory, perception becomes enhanced by memory associations acquired during learning. As more memories are accumulated, a constant sensory input gives rise to more diverse experiences. In contrast, the differentiation theory states that, as a result of learning, perception corresponds more closely to the stimulation in the environment. The perceiver makes finer discriminations because he/she has learned to respond to variables of stimulation that were not previously responded to in a given sensory input.It is useful to consider the evidence that has been used to support the differentiation theory.

Familiarity effects in visual comparison tasks and their implications for studying human intelligence.

“…(For more information on the distinction between a selective attention task and a biconditional classification, see Garner, 1978;Posner, 1964. ) In Experiment 1, the stimuli from the Tetewsky and Garner (1986) experiment were used to design a letter-matching task similar to the one used by Posner and Mitchell (1967). Subjects were required to classify the stimuli as being the same or different on the basis of four distinct rules of sameness: physical identity, form, system, and name.…”

“…These stimulus dimensions are orthogonal in that whenever a given pair of stimuli can be classified as being the same on one of the dimensions, that pair also differ on the other two dimensions. (For a diagram showing the dimensional structure of the stimuli, see Tetewsky and Garner, 1986. ) The stimuli were displayed on a Macintosh Plus microcomputer.…”

“…The method that was used to study familiarity effects was motivated by an experiment carried out by Tetewsky and Garner (1986). Although this experimental paradigm is related to the way Jackson (1980) operationalized familiarity effects, it has the added advantage of using alphabetic stimuli that have been learned over a number of years, as opposed to artificial stimuli that have been learned over several training sessions.…”

“…In the Tetewsky and Garner (1986) experiment, two Hebrew and two Roman characters were used to form a set of stimuli that could be described by three orthogonal dimensions: form, system, and name. By designing the stimulus set in this way, Tetewsky and Garner were able to analyze familiarity effects by using three different classification tasks.…”