Bridges Between Psychology and Linguistics

A Swarthmore Festschrift for Lila Gleitman

Donna Jo Napoli, Judy Anne Kegl, Judy Kegl

• Publisher: Psychology Press
• Year: 2013
• ISBN:
  9781134760695
• Edition: 1
• Topic: Crescita personale
• Subtopic: Tecniche di scrittura e presentazione

Written as a tribute to Lila Gleitman, an influential pioneer in first language acquisition and reading studies, this significant book clearly establishes the relationships between psychology and linguistics. It begins with a thorough examination of issues in developmental psychology, continues with questions on perception and cognition, studies the realm of psycholinguistics, and concludes with an exploration of theoretical linguistics.

1

The Decline of Visually Guided Reaching During Infancy*

Emily W. Bushnell
Tufts University

The emergence of visually guided reaching and grasping has long been recognized as an important occurrence during infancy. The origins of this perceptual-motor skill and the precise nature of its precursors are matters currently under investigation and debate by developmental psychologists (for recent discussions, see Bower, 1974; Bushnell, 1981; Lockman & Ashmead, 1983; McDonnell, 1979; von Hofsten, 1982). In this commentary, these popular issues are avoided, and a later phase in the development of eye–hand coordination is focused on instead. The plan is to document that following its emergence and a period of intense exercise, visually guided reaching declines in practice and is replaced by reaching of a more visually elicited style in most circumstances. The nature of this decline is then discussed, and in the end, it is suggested that the decline may be of particular significance for infants’ cognitive development.

There is general agreement in the literature that visually guided reaching is first exhibited sometime between 3 and 5 months of age (cf., Bower, 1974; Bruner, 1969; Halverson, 1931; McDonnell, 1979; Piaget, 1952; White, Castle, & Held, 1964). At younger ages, infants have been observed to exhibit behaviors that indicate eye–hand coordination (Bower, Broughton, & Moore, 1970; Bruner, 1970; DiFranco, Muir, & Dodwell, 1978; McDonnell, 1979; Rader & Stern, 1982; von Hofsten, 1982; White et al., 1964), but these are not considered to be visually guided reaches. The behaviors have been variously identified as intentional reaches (Bower, et al., 1970), swiping (White, et al., 1964), “preadaptations” for reaching (Bruner, 1973), “prefunctional” reaches (von Hofsten, 1982), reaching-like arm movements (Rader & Stern, 1982), and “prereaching” (Trevarthen, 1975). In what follows, the last term is used for brevity’s sake, and it should be noted that this early eye–hand coordination is a matter for some debate. The very existence of prereaching has even been questioned (Ruff & Halton, 1978), but von Hofsten’s (1982) precise observations of newborns’ arm movements seem to authenticate the behavior. However, the continuity of prereaching and later forms of eye–hand behavior is uncertain. Von Hofsten (1984) recently reported that the amount of prereaching decreases abruptly at 7 weeks of age, at which time the form of the behavior changes, too. This break at 7 weeks is then followed by a period of practically no eye–hand behavior, after which reaching activity that can be described as visually guided begins to appear. Also at issue are whether prereaching is sensitive to spatial dimensions such as three-dimensionality (see Bower, 1972; Bower, Dunkeld, & Wishart, 1979; DiFranco, Muir, & Dodwell, 1978; Dodwell, Muir, & DiFranco, 1979) and whether it indexes the intention to touch or grasp the target object (see Bower et al., 1970; von Hofsten, 1982). Amidst the controversies about it, though, there seems to be consensus that prereaching differs from the reaching that emerges at about 4 months in at least three respects. Because these distinctions also figure in the analysis of the development of reaching during the second half of the first year, they are discussed next at some length.

The most obvious distinction between prereaching and later reaching is that prereaching is much less accurate. Bower (1974) reported that the arm extensions of very young infants result in contact with the target object only 40% of the time, whereas those of infants over 20 weeks old result in contact at least 80% of the time. Other researchers have found prereaching to be even less accurate than Bower noted; von Hofsten (1982) reported that his neonatal subjects touched the object on only 22 out of 232 reaches or 9% of the time, Ruff and Halton (1978) reported a hit rate of 7%, and Dodwell, Muir, & DiFranco (1976) reported one of just 3%.

A second distinction between prereaching and the reaching that emerges at about 4 months is that prereaching has been characterized as ballistic, or visually elicited, whereas the later reaching is visually guided. That is, in prereaching, the arm movement is preprogrammed or fully aimed at the target object before or at the moment the movement is launched. An arm movement with a trajectory that consists of a single movement element would by definition be ballistic, but arm movements with trajectories that consist of several elements at different angles or otherwise distinguished might also be ballistic. In either case, the critical property is that the course of the movement is not purposefully altered during its performance or upon its completion. In contrast, the reaches of infants beginning at about 4 months involve a continuous feedback process; the arm’s approach to the target object is monitored and may be adjusted throughout the reach.¹

The evidence that prereaching arm movements are ballistic is not as strong as that regarding their other properties. The clearest proof would derive from comparing prereaches toward stationary objects to those toward objects whose locations were changed after the initiation of the arm movement; however, this comparison has not yet been made. The evidence that does exist consists first of all of subjective descriptions of very young infants’ arm movements as “swipes” (White et al., 1964), “flings” (Bower et al., 1970), and “explosive” (Bruner, 1973). These observations that prereaches are very rapidly executed and therefore unlikely to incorporate constant vigilance and adjustments are substantiated by analyses of their spatiotemporal qualities. Von Hofsten (1982) found that the forward extensions of neonates came closest to the target object on the average just 1.04 seconds after their start. Under similar experimental conditions, the approach time for the (guided) reaches of infants 18 and 21 weeks old averaged about 1.8 seconds (von Hofsten, 1979). The inaccuracy of prereaching in a sense supports that it is ballistic, too. None of the researchers cited earlier for reporting that very young infants’ arm movements usually terminate without contacting the object has noted that such misses are followed by corrective actions. The results of two studies in which infants reached while wearing prism glasses also support the notion that prereaching is ballistic; these and the considerable evidence that the reaching that follows prereaching is guided are presented after the third distinction between the two is introduced.

The third distinction between prereaching and the reaching first exhibited at about 4 months has to do with the basis on which each sort of reach is accomplished. For eye–hand coordination, there are actually three variables that may be involved—the visually localized target (the seen target), the visually localized hand (the seen hand), and the proprioceptively or kinesthetically localized hand (the felt hand). Under normal circumstances, reaching can be successful (i.e., be related to the position of the target) if either the seen target and the felt hand or the seen target and the seen hand map onto one another or “match.” The available evidence (to be reviewed later) indicates that prereaching makes use of the former matching mechanism, whereas the reaching emergent at about 4 months utilizes the latter. That is, the eye–hand coordination of newborns and very young infants seems to be based on some sort of prewired visuoproprioceptive or visuomotor spatial coordination. The eye–hand coordination of somewhat older babies, though, is accomplished otherwise, by their monitoring and progressively reducing the “gap” between the seen target and the seen hand. Thus, for this later sort of reaching, the seen hand must be attended to and localized, but for prereaching, the seen hand is irrelevant.

The idea that eye–hand coordination changes early in life from using the felt hand to using the seen hand is strongly supported by the results of several studies in which the perception of the seen hand was interfered with. Lasky (1977) observed the reaching of infants under normal circumstances and in a mirror condition where they could see the target object but could not see their hands as they reached for it. Infants 3½ and 4½ months old reached equally well in the two conditions, indicating that they do not utilize the sight of the hand during reaching. In contrast, infants 5½ and 6½ months old reached more “clumsily” in the mirror condition than otherwise, which suggests that they do rely on the sight of the hand when reaching. The older infants also often aborted reaches and became upset in the mirror condition, presumably because they were “looking for” their hands, which did not appear.

Bower (1976) observed infants reaching for objects while they were wearing prism glasses that laterally displaced the seen object and the seen hand. Under such circumstances, seen target–seen hand reaching will still be successful, but seen target–felt hand reaching will result in a “miss.” Bower reported that infants younger than 24 weeks of age routinely missed the object and, following the miss, pulled their hands back and reached erroneously again. Infants older than 24 weeks started their reaches off in the inappropriate direction, but corrected the trajectory as soon as their hands came into view. These “mid-reach” corrective actions reveal not only that the seen hand was attended to by the older infants, but also that their reaches were guided rather than ballistic. Conversely, the younger infants’ failure to remedy their misses either within or between reaches supports the claim that prereaching is ballistic. McDonnell (1975) also observed infants who were wearing prism glasses. He reported that infants ranging in age from 4 to 10 months all corrected the trajectories of their reaches, thereby evidencing both use of the seen hand and guidance. However, the youngest infants’ trajectories were the most deviant and seemed to reflect an abrupt switch to visual control after a substantial mis-aimed component. Thus, McDonnell’s results, like Bower’s, indicate that reliance on visual feedback of the hand during reaching increases between 4 and 6 months.

Finally, the observations of von Hofsten (1979) add to the evidence that, starting at about 4 months, reaching is guided rather than ballistic. He found that 4-and 5-month-old infants’ approaches to targets were relatively slow and devious (i.e., made up of zigzag and roundabout movements), consisted of several elements each of about the same duration, and often contained elements directed away from the target. Though such features do not necessarily mean that the reaches were guided, ballistic approaches would ordinarily be more rapid and direct. The infants also generally fixated the target throughout the reach, as is characteristic of guided reaching. Whether the apparent guidance involved the seen hand or the felt hand is not indicated by von Hofsten’s data; though the target rather than the hand was fixated, the hand was undoubtedly within the visual field during most of each approach.

The case has been presented that the eye–hand behavior evident during the earliest months of life is inaccurate, ballistic, and based on a mapping of vision and proprioception, whereas the reaching that emerges at about 4 months is more accurate, guided, and based on matching the seen target and the seen hand. It should be noted at this juncture that the second and third distinctions that have been discussed are logically independent; a reach based on matching the seen target and the felt hand, for instance, could be either ballistically aimed or involve continuous monitoring and adjustment (see Hay, 1979). Empirically, however, these two distinctions are often blurred, as techniques for studying one of them often provide information concerning the other, too (as when prism glasses are employed, e.g.). Furthermore, it seems that in practice a ballistic reach could rarely involve the seen hand, because the latter would often be unavailable when the reach was initiated. In what follows, where the term “prereaching” is no longer appropriate, the term “visually elicited” will be used to refer to reaches that are both ballistic and based on a visuoproprioceptive mapping, and the term “visually guided” will be used to refer to reaches that are not ballistic and that are based on reducing the gap between the seen target and the seen hand.

It should also be acknowledged at this point that the particular age at which prereaching is succeeded by the later form is identified somewhat differently by the several investigators whose work has been cited. Thus, Lasky (1977) and Bower (1976) located the transition between 4½ and 6 months, whereas von Hofsten’s several studies (1979, 1982, 1984) considered together indicate that it occurs between 2 and 4 months. McDonnell (1975) found that reaching was sensitive to visual feedback of the hand by 4 months, but also noted that it became more so between 4 and 6 months. These differences most likely derive from differences in the subject selection procedures, stimulus conditions, and constraints on posture and vision employed in the various studies. Although the effects of such variables are interesting, what is important for the purposes of this commentary is that by all accounts, guidance and the seen hand become more and more integral aspects of reaching over the first half-year of life.

Eye–hand behavior during the second half-year of life has also been of interest to developmental psychologists. Observations of older infants have generally focused on manual sensitivity to various object properties (e.g., Lockman, Ashmead, & Bushnell, 1984; Pieraut-le Bonniec, 1985), on handedness (e.g., Ramsey, 1980), or on responses in special situations such as those involving multiple objects or those demanding detour reaching (e.g., Bruner, 1970; Lockman, 1984). However, some evidence concerning the status of visually guided reaching after its initial appearance has accumulated. McDonnell and Abraham (1979) studied perceptual adaptation in infants who repeatedly reached for objects while wearing laterally displacing prism glasses. They reasoned that adaptation should be minimal when reaching is visually guided, because no intercue discrepancy exists to motivate it; with prism glasses, closing on the seen target with the seen hand still results in a successful reach, because these are not displaced relative to one another. McDonnell and Abraham generally observed aftereffects indicating that adaptation had occurred, but they noted that the magnitude of adaptation decreased between 5 and 7 months of age. This finding implies that the seen hand is relied on increasingly over this age period, as discussed earlier. Interestingly, though, McDonnell and Abraham found that, after dropping off at 7 months of age, the magnitude of adaptation then increased again by 9 months. McDonnell (1979) interpreted this pattern of results as evidence that visually guided reaching “peaks” at around 7 months of age and is then gradually replaced by a more ballistic style of reaching. He cited Halverson’s (1931) descriptions of the target approaches of infants of different ages as further support for this return to visually elicited reaching.

McDonnell’s suggestion that reaching becomes less visually guided toward the ¾ mark of the first year is supported by the secondary results of several recent investigations. Bushnell (1982) studied 8-, 9½-, and 11-month-olds as they reached for an object that appeared to be located under a little shelf. The shelf was actually a half-silvered mirror that reflected the image of an object hidden above the shelf. The reaching situation was therefore similar to that studied by Lasky (1977); although the hand did not disappear entirely as it passed under the shelf, its appearance became somewhat transparent and “ghostly.” Although the purpose of the study was to observe the infants’ reactions when the object actually located under the shelf differed from the object seen, a number of main effects of age were found in addition to the condition effects of major interest. The 8-month-olds looked at the mirror image for shorter durations than the 9½-and 11-month-olds did, they touched the object under the shelf for shorter durations, and they looked and touched simultaneously for shorter durations. Basically, the younger infants were rather reluctant to interact with the apparatus. Bushnell speculated that the strange appearance of their hands was noticed by the 8-month-olds and disrupted their reaching, as the hand’s absence interfered with the reaching of 5½-and 6½-month-olds in Lasky’s study. The 9½-and 11-month-olds, though, did not seem to be as concerned with the appearance of their hands.

In another task in the same study, Bushnell (1982) compared the behavior of the infants as they reached for an ordinary, palpable object and for an optically created, nonpalpable image of the same object. Both the 9½-month-olds and the 11-month-olds accompanied greater percentages of their touching with looking on the unusual, virtual-object trials (86% and 89%, respectively) than they did on the control trials (69% and 73%, respectively). The 8-month-olds, however, did not exhibit this difference. They showed a ceiling effect, accompanying the same high percentage (84%) of their touching with looking on both sorts of trials. Evidently, 8-month-olds visually monitor their prehensile activities intently in most situations, even routine ones, but older infants do not; they monitor their prehensile activities intently only in special circumstances.

Further evidence that the role of visual guidance during reaching becomes less important toward the end of the first year was reported by Lockman, Ashmead, and Bushnell (1984). They observed the hand positioning and looking behavior of 5-and 9-month-old infants as they reached for horizontally and vertically oriented dowels. Infants of both ages usually looked at the dowel throughout the reach, at least until they had touched the target. Som...

Table of contents

• Cover
• Halftitle
• Title
• Copyright
• Contents
• Introduction
• 1 The Decline of Visually Guided Reaching During Infancy
• 2 Programs for Movement Sequences
• 3 Similarity and the Structure of Categories
• 4 Perception of a Unified World: The Role of Discontinuities
• 5 Infant Vocalizations and Changes in Experience
• 6 Mechanisms for Listener-Adaptation in Language Production: Limiting the Role of the “Model of the Listener”
• 7 Linguistics and Dyslexia in Language Acquisition
• 8 Spatial Language and Spatial Cognition
• 9 Farewell to ‘Thee’
• 10 Linearity as a Scope Principle for Chinese: The Evidence from First Language Acquisition
• 11 On Interpreting Partitives
• 12 On the Relevance of Traditional Phonological Analysis to the Abstract Patterns Found in ASL and Other Signed Languages
• 13 Phonology as an Intelligent System
• 14 Linguistic Theory and the Naturalist Approach to Semantics
• Author Index
• Subject Index