There is a`theory of mind' theory of autism. Meltzoff has recently developed a procedure that gives a nonverbal assessment of`theory of mind'. A group of children with autism and a matched control group of normally developing infants were given three of Meltzoff's tasks and three conventional, gestural imitation tasks. The children with autism showed the expected deficits on gestural imitation, but were significantly better than the normally developing infants on the Meltzoff tasks. The implications of these results for a number of theoretical issues are discussed.

It has been proposed that speech is specified by the eye, the ear, and even the skin. Kuhl and Meltzoff (1984) showed that 4-month-olds could lip-read to an extent. Given the age of the infants, it was not clear whether this was a learned skill or a by-product of the primary auditory process. This paper presents evidence that neonate infants (less than 33 h) show virtually identical patterns of intermodal interaction as do 4-month-olds. Since they are neonates, it is unlikely that learning was involved. The results indicate that human speech is specified by both eye and ear at an age when built-in structural sensitivities provide the most plausible explanation.

While there are many theories of the development of speech perception, there are few data on speech perception in human newborns. This paper examines the manner in which newborns responded to a set of stimuli that define one surface of the adult vowel space. Experiment 1 used a preferential listeningahabituation paradigm to discover how newborns divide that vowel space. Results indicated that there were zones of high preference flanked by zones of low preference. The zones of high preference approximately corresponded to areas where adults readily identify vowels. Experiment 2 presented newborns with pairs of vowels from the zones found in Experiment 1. One member of each pair was the most preferred vowel from a zone, and the other member was the least preferred vowel from the adjacent zone of low preference. The pattern of preference was preserved in Experiment 2. However, a comparison of Experiments 1 and 2 indicated that habituation had occurred in Experiment 1. Experiment 3 tested the hypothesis that the habituation seen in Experiment 1 was due to processes of categorization, by using a familiarization preference paradigm. The results supported the hypothesis that newborns categorized the vowel space in an adult-like manner, with vowels perceived as relatively good or poor exemplars of a vowel category.

No abstract

The vowels /i/, /y/, /u/, and /unrounded u/ define one surface of the vowel space, whether defined in articulatory or acoustic terms. Adults can give goodness ratings to vowels from the space surrounding prototypic vowels, the prototypic vowels typically obtaining the highest ratings. Does this reflect unlearned structural sensitivities or some averaging process during life? Is it true in newborns? Attractiveness or ‘‘goodness’’ was measured by how long newborns listened to first presentation of a set of vowel stimuli. The longer the listen, the better the newborn perceived that vowel to be. The edges of the surface defined by /i/, /y/, /u/, and /unrounded u/ was broken into 40 steps. Twenty newborns (7–33 h) were presented with 21 stimuli defining two edges of the vowel surface, five infants begining at each corner of the vowel surface. The corner, prototype vowels are preferred over intermediates; even /y/ and /unrounded u/ which do not occur in Texas. However, /i/ and /u/, which do occur in Texas, do elicit more listening than the other two corner vowels. Unlearned, structural sensitivities have been affected by experience in infants as young as 7 h of age.