FIGURE 6.6 Illustration of a 2 - 1 = 1 or 2 trial. Two Mickey Mouse® dolls (Wynn, 1992) or eggplants (not shown; Hauser et al., 1996) were placed on a stage. A screen was then raised to hide the objects. A hand entered the scene and removed one object. The screen was then lowered to reveal one (expected) or two (unexpected) objects. The time a monkey or human infant spent looking at the final outcome was measured. (Reprinted by permission from Nature 358, 749, © 1992 Macmillan Publishers Ltd.)

recorded, and stimuli are presented repeatedly until a substantial reduction in looking time occurs (habituation criterion). Infants are then shown new exemplars of the same number and a new number. Infants tend to look longer at the new number, demonstrating that they discriminate the novel from the familiar numerosity. In the first study of this sort, Starkey and Cooper (1980) tested infants between 16 and 30 weeks on their ability to discriminate two vs. three or four vs. six dots. They found 2 vs. 3 but not 4 vs. 6 discrimination (for a similar demonstration with slides of household objects, see also Strauss and Curtis, 1981). Antell and Keating (1983) tested newborn infants and also found evidence of 2 vs. 3 but not 4 vs. 6 discrimination. Van Loosbroek and Smitsman (1990) found 2 vs. 3 and 3 vs. 4, but not 4 vs. 5 discrimination in 5-month-old infants tested with dynamically changing displays rather than static images. Bijeljac-Babic et al. (1993), using the high-amplitude sucking procedure, found evidence that 4-day-old infants discriminate two vs. three consonant-vowel syllables. Collectively these studies suggest that infants' ability to make numerical discriminations may be limited to small set sizes.

In a recent study, Wynn et al. (2002) habituated infants to two or four randomly moving dots (see Figure 6.7) and then tested both groups of infants with two groups of four dots or four groups of two dots. Infants habituated to two dots looked longer at the four groups of two dots, and infants habituated to four dots looked longer at the two groups of four dots. This is a particularly powerful finding in that the test displays contained the exact same eight dots and were thus precisely controlled for surface area and perimeter. In another study from Wynn's (1996) research group, infants were habituated to two or three puppet jumps and then tested with both numbers of jumps. Infants' looking time was measured to the static puppet after the jumps were completed, and the timing of the jumps was carefully controlled so that total duration was not available as a cue. Infants looked longer after the novel number of puppet jumps, showing that they discriminated two from three puppet jumps even when total duration and tempo were controlled.

sample frames from sample frames from habituation movies testmovies

FIGURE 6.7 Sample frames of habituation and test movies used by Wynn et al. (2002) to test 5-month-old infants in a 2 vs. 4 discrimination. (From Wynn, K., Bloom, P., and Chiang, W., Cognition, 83, B55-B62, 2002.)

A new method makes use of infants' reaching behavior and shows that both the duration and number of reaches an infant makes into an opaque box are correlated with the number of objects that the infant observed being placed into the box (Van de Walle et al., 2000). Infants were shown one or two objects being placed into a box. The experimenter surreptitiously removed one of the objects from the box and then allowed the infant to reach in and withdraw one object. After the experimenter took the first object away from the infant, the infant was then allowed to continue searching the box. The time infants spent searching the empty box was then measured. Infants who observed two objects being placed into the box were more likely to reach for a second object than infants who observed one object.

6.4.2 Large Number Discrimination

Contrary to the view that infants can only discriminate small values, Xu and Spelke (2000) found that 6-month-old infants discriminate 8- vs. 16-element displays (see Figure 6.8). Infants were habituated to 8- or 16-element displays. The 8- and 16-element displays contained the same cumulative surface area of dots, and stimulus size was constant in habituation such that density covaried with number. In test, the continuous variables that had been held constant in habituation now covaried with number, whereas the variables that covaried with number in habituation were held constant in test. Specifically, element size was held constant in the 8- and 16-element test displays, whereas the stimulus size varied such that density was constant. The particular element size used in test was chosen so that the total surface area in test differed from that of habituation for both the 8- and 16-element displays by a constant amount. Infants looked longer at the novel numerosity. Supporting these findings, Brannon (2000b) replicated this experiment and Lipton and Spelke (2001, 2002), using the head-turn procedure, have also found large number discrimination in 6-and 9-month-old infants with tones rather than visual displays.

6.4.3 Cross-Modal Matching

Starkey et al. (1983, 1990) reported that when 6- to 8-month-old infants heard two or three drumbeats, they looked longer at a visual display that contained the matching number of elements than at visual displays of the other numerosity. This finding, however, was not replicated by other researchers (Mix et al., 1997; Moore et al.,


0 0

Post a comment