Ξ  

 Sobre a Deficiência Visual

Blindness and Children - Executive functions: memory, attention, and cognitive strategies

David H. Warren

Blind kid in Vietnam - foto de Tiet Ho
Blind kid in Vietnam - foto de Tiet Ho

 

Infancy
Memory span
Encoding of tactual information
Strategies of tactual information processing
Integration of information from different sensory modalities
Verbal and phonological issues in encoding and memory
Imagery


We consider now a set of interrelated cognitive processes that includes memory, attention, and cognitive strategies. These "executive functions" have not received much direct research attention in the visual impairment literature, particularly that with infants and children. However, many studies, some of which we have already mentioned in previous sections, bear at least indirectly on these issues.

Infancy

Several lines of innovative research with sighted infants have revealed that the neonate's attention is "captured" by certain perceptual events, and that the neonate has little if any volitional choice about which stimuli are actually attended. Initially involuntary, attention gradually comes under a degree of voluntary control. For example, early eye fixations are entirely determined by stimulus features but become largely volitional during the first year.

Among the visual events that are particularly attention-commanding are moving stimuli, facelike stimuli, and areas within the visual field that contain a moderate degree of complexity. Moving stimuli are especially effective in eliciting visual attention. Complexity is an intriguing dimension of visual attention: the evidence supports the notion that as the infant's visual information-processing capacities develop, the infant prefers to look at progressively more complex stimulus arrays. The other sensory modalities have been far less completely studied, and the existence of analogs to these visual-developmental principles in other modalities is for the most part a hypothetical question at present.

One of the most interesting lines of recent research in early perceptual and cognitive development has been in the area of memory. The use of ingenious research methods has revealed that by at least four months, infants begin to recognize and remember perceptual stimuli that they have encountered before. The implications of this development are clear: it is only when recognition and memory begin that conceptual representation and an understanding of the physical world can begin. Caution is in order, however: to say that the first evidence of recognition and memory occurs around four months of age is not to say that memory is mature at that point or that the capacity to represent aspects of the physical world conceptually springs forth fully formed.

Without research on perceptual development and executive functions such as attention and memory, our knowledge about the human infant's cognitive understanding of the physical world would be incomplete.

Impressive strides have been made in research on these issues with sighted infants. There is as yet little such work with visually impaired infants. Some inferences about memory can be made, and we reviewed the evidence for these in connection with our discussion of the development of object permanence in Chapter 3.

We turn now to consideration of these issues in preschool and schoolage children.

Memory span

The digit span subscale of the WISC can be taken as a measure of simple memory. Tillman and Osborne (1969) evaluated WISC verbal scale scores for groups of blind and sighted children, ages 7 to 11, for whom overall WISC scores were equated. Analysis revealed a significant interaction of scale (the six verbal scales) and group (blind and sighted). This was produced by superior performance of sighted children on the similarities scale, offset by superior performance of blind children on the digit span scale. This pattern of superiority of the blind children's memory, relative to their performance on the other scales, did not change with age. Print readers were excluded from the sample of blind children, and thus the results can be taken as applying to children with severe visual loss.

Smits and Mommers (1976), studying children in the Netherlands, reported a similar finding with children ranging from 7 to 13 years of age.

The pattern was exactly the same, with digit span performance relatively stronger for the children with visual impairments than performance on other scales. When the group was divided into blind and partially sighted subgroups, the overall WISC verbal IQ^was higher for the blind subgroup.

This difference appeared in each of the six scales and was apparently no stronger for the digit span than for the other scales.

From this evidence, then, there is clearly nothing wrong with the simple memory capabilities of children with visual impairments, and indeed this may be an area of relatively high function.

Encoding of tactual information

How is tactual information encoded in memory by children with visual impairments? Although it seems evident that the nature of coding must be tactual, the issue is not that simple, since it is possible that aspects of verbal or visual encoding may also be involved.

Davidson, Barnes, and Mullen (1974) varied the memory demand in a task involving matching of three-dimensional shape stimuli. The child explored the standard stimulus, then felt each of the comparison stimuli in succession and chose the one that matched the standard. Memory demand was varied by including either three or five items in the comparison set: since the incorrect members of the set were similar to the standard, exploration of them constituted tactual interference. Increasing the size of the comparison set increased the error frequency significantly.

The results thus support the hypothesis that features of tactually experienced stimuli are encoded in a specifically tactual form.

A useful paradigm for studying the nature of encoding involves inserting a delay between the experience of the standard and the choice of the comparison, during which various activities are interposed. The logic is that different kinds of intervening activity should interfere selectively with memory, depending on the nature of the encoding. For example, tactual intervening activity should interfere if the standard stimulus is tactually encoded, but not if it is encoded in some other form.

Following this logic, Millar (1974) used three-dimensional nonsense shapes designed to be easily discriminated but not easily labeled. The subjects were 9- and 10-year-olds who were blind from very early in life.

The procedure was to present a standard stimulus for a 2-sec inspection, then after a delay to present a comparison stimulus for the child's same—different judgment. The delay was either unfilled (no activity) or involved rehearsal (finger tracing of the shape of the standard on the flat floor of the apparatus), verbal distractor (counting backward by threes), or tactual distractor (a tactual manipulation activity). If memory of the standard is tactual, then rehearsal should facilitate performance, whereas the tactual distractor should interfere with it. The verbal distractor should presumably be neutral in its effect if the standard stimulus is encoded tactually.

Errors were were infrequent in all conditions, but response latencies varied with the nature of the delay activity. Responses were slow in the tactual distractor condition, a result that supports the tactual encoding hypothesis. However, rehearsal did not facilitate performance; furthermore, the verbal distractor did interfere with performance. These results do not conform to the hypothesis of tactual encoding.

Since the tasks did not produce differential error rates, a second experiment was conducted with five- to seven-year-old children, using basically the same procedures. More errors occurred in the verbal distractor and movement distractor conditions than in the unfilled delay or rehearsal conditions.

The hypothesis that encoding is specifically tactual predicts that the tactual distractor should interfere most with performance, and specifically should interfere more than the verbal distractor. That the verbal and movement distractors both had interfering effects suggests that the interference was not a modality-specific effect that interferes with encoding, but rather was a matter of distraction of the child's attention.

Pursuing the attention versus modality issue, Kool and Rana (1980) hypothesized that a verbal distractor would interfere by distracting attention, whereas a tactual distractor would interfere specifically with the retention of tactual information. They used conditions like those of Millar (1974) with congenitally blind children ages 9 to 11, and ages 13 to 16 years in a second experiment. There was a decay of tactual memory with increasing delay in the unfilled delay condition. The verbal distractor interfered with performance, thus corroborating Millar's (1974) results.

It was assumed that the older children would have more tactual experience and would therefore be more inclined to encode tactually.

With these children, both verbal and tactual distractor conditions were effective. The effect of the tactual distractor was greater than that of the verbal distractor at every delay interval. This result supports the notion that the stimuli were tactually encoded. The significant effect of the verbal distractor suggests that there is also a general effect on attention, thus supporting Millar's (1974) conclusion.

These three studies support the hypothesis that tactual stimuli are indeed encoded in a tactual manner, but that the retention of tactually encoded information is affected by attentional factors that are not specific to sensory modality.

It is known that phonologically similar items can interfere with one another in memory and cause lower recall. Millar (1975a) asked whether a similar process might occur with tactual features encoded in memory.

Such interference would be evidence for the encoding of tactual features.

The blind children ranged in age from 4 to 10 years, and all had lost vision by the age of 18 months. The procedure required the recall of the position of an item in a series, with series length ranging from two to six items. The objects were presented sequentially, then one test object was given to the child with the request to replace it in its correct position in the series. Three types of series were used. One contained phonologically similar (but tactually distinct, e.g., rat, bag, man) items, another contained items that were tactually similar (but phonologically distinct, e.g., ruler, knife, comb), and the third was heterogeneous, containing items that were tactually and phonologically distinct (e.g., ball, watch, chair).

As expected, recall performance was worse for the phonologically similar series than for the heterogenous series. This finding shows the well-known phonological interference effect. The question of interest was whether a similar effect would be found for tactually similar series.

Overall recall of the tactually similar items was indeed worse than for items in the heterogeneous series, thus indicating that tactual features must have been stored in memory. Interestingly, the tactual interference effect was strongest for the smaller series and decreased as series length increased. No variation with age or other individual differences variables was reported.

Summary. This literature on the nature of encoding of tactual experience is small but interesting. Studies that show adverse effects of tactual distractors during a delay before recall suggest that tactually experienced information is encoded in a specifically tactual form. However, the issue is not that simple, since performance also deteriorates, though typically to a lesser degree, as a result of verbal distractors. This may be largely an effect of attention, but the relationship between initial encoding and retention variables is not completely clear. Finally, it is surprising that not more has been done to explore the possibilities that children with early visual experience encode tactual information in a different way that is somehow affected by that experience.

Strategies of tactual information processing

Performance in many perceptual and cognitive tasks is found to vary significantly as a function of the information-processing strategy that the subject adopts. There is a small but exemplary literature on this regarding children with visual impairments.

Using a very basic paradigm, Simpkins (1979) examined children's ability to recognize geometric shapes tactually. The child first felt a standard stimulus, then subsequently chose a match from a set of four sequentially presented alternatives. The children were four to seven years of age and varied in the amount of visual function.

There was little variation due to gender or visual status, but the older children performed better than the younger ones. Simpkins reported that in touching the stimuli, the younger children tended to attend to a peculiar topological property of a form (e.g., a hole in it) whereas the older children tended to hold the shape in one hand and trace its contour with the other hand. This shift in exploration strategy parallels that found with sighted children (Gliner, 1966), and it is not surprising that strategy shifts are related to performance in a similar way. Berla (1974) used irregular geometric stimuli varying in complexity from three to five sides. The child felt a shape in one orientation, then the shape was quickly rotated by 90, 180, or 270 degrees without the child touching it.

The child's task was to return the shape to its original orientation. The accuracy of performance improved with grade level from grades two through eight. Increasing complexity did not decrease accuracy but did increase the time required for performance. Berla noted that the graderelated performance differences seemed to be connected to the strategy of choosing a distinctive feature of the shape to concentrate on: the older children appeared to attend more to the distinctive features (e.g., sharp angles) of the shapes. Berla suggested that a consistent informationprocessing strategy was the basis for their better performance.

Berla's analysis of shape discrimination in terms of distinctive features is reminiscent of Gibson's (1969) formulation: attention to peculiar distinguishing features, or areas of high information content, improves the efficiency and effectiveness of shape perception. In a similar vein, Solntseva (1966) suggested that the difficulties that the blind child experiences in the formation of tactual images of the external environment are caused by problems in the ability to differentiate distinctive features of tactual experience. Tactual qualities such as texture and hardness are relatively attention commanding (Klatzky, Lederman, & Reed, 1987) and easy to discriminate, but the discrimination of shape requires a more systematic approach for the detection of critical features.

Davidson (1972), studying adolescents, used haptic judgments of curvature as a vehicle for studying the relationship of tactual scanning strategies and task success. The task was to judge whether an edge was convex, concave, or straight. Hand movement patterns were videotaped.

The most frequently used strategy was the "grip," in which all four fingers are spread out along the curve, followed by the "top sweep," which involved running the forefinger along the length of the curve. It is interesting that the blind subjects used the grip strategy much more frequently than a comparable group of blindfolded sighted subjects, and that the judgments of the blind group were more accurate. (When sighted subjects were instructed to use the grip strategy, their performance improved.)

The relationship of strategies and performance under variations in task difficulty is also of interest. Davidson and Whitson (1974) varied task difficulty by changing the number of items in the comparison set.

That is, a standard curve was presented and felt, and then the subject had to find the standard when it was part of a comparison set of one, three, or five curves. (In the case of the single curve comparison, a simple samedifferent judgment was required.) The congenitally blind subjects averaged 19 years of age.

When search strategy was unrestricted, errors increased regularly with the number of comparison alternatives, showing a basic effect of task difficulty. The question of interest, though, is whether strategies are differentially effective for various difficulty levels. The "grip" strategy was most frequently used regardless of difficulty level, but there was a tendency for the "top sweep" strategy to increase and the "grip" to decrease at the highest difficulty level.

Instructed strategies

In a second part of the same experiment, subjects were instructed to use a single strategy. There was a tendency for better performance when strategies were used in which more of the curved stimulus was simultaneously apprehended (e.g., the "grip," in which the subject's four fingers are spread along a substantial portion of the edge).

Berla and Butterfield (1977) examined the effectiveness of training procedures in improving tactual search performance. The subjects were braille readers in kindergarten through fifth grade, age range 6 to 17 years. The test stimuli were outline tracings of various states and countries.

The child felt the stimulus for 30 sec, then attempted to find the same stimulus in a set of four shapes. Based on a pre-test, children were divided into a training and a control group. The training group received three training sessions in which the child's attention was drawn to distinctive features of the shapes (e.g., "parts that stick out"). Following training, a post-test was given, and the trained group, which was matched to the control group on the basis of pre-test scores, performed very much better than the control group, with 84% of the training group showing an improvement.

In a second experiment, the test materials were changed to involve searching for a shape in a complex array of shapes. Training was similar to that in the first experiment but involved shapes embedded in more complex arrays. On the post-test, the trained group again performed significantly better and faster than the untrained control group. Thus training improved performance, apparently not only by drawing the child's attention to distinctive features but also by encouraging a more systematic search process as well.

Berla and Murr (1974) also instructed subjects in the use of specific search strategies while searching for features on a tactual map. The subjects were drawn from grades 4 through 12 and ranged in age from 11 to 19 years. All of the children were braille readers. Following a pre-test requiring the location of tactual symbols without strategy instructions, three groups were instructed to use either a vertical, a one-handed horizontal, or a two-handed horizontal scanning strategy. The children practiced the strategy for 4 min. Children in a fourth condition were free to scan as they wished. The task then involved finding as many target symbols on the map as possible. There was a modest general increase with grade level in the number of symbols located. Of more importance, the scanning strategies produced different rates of success: each of the three instructed strategies produced significant improvement compared to the pre-test, with the vertical strategy producing the greatest improvement.

The uninstructed control group did not improve over the pre-test performance. The benefit of the instructed strategies in general seemed to stem from the more systematic coverage of the map that resulted.

Berla (1981) further explored the effectiveness of training scanning strategies as a function of age. The children, all braille readers, were divided into group that averaged 11, 15, and 19 years. Each age group was divided into a control and a training group. Early in the test procedure, children in the training group were briefly instructed in the use of a systematic vertical scanning strategy. The task required the child to feel the parts of a nine-item "puzzle" and remember their locations, and then to recreate the puzzle on a blank board using the nine individual elements.

There were no obvious effects of the training on vertical location errors. For horizontal errors, however, the effect of training varied with age group. Specifically, the youngest group benefited from training, while the performance of the oldest group suffered from training. The performance of the middle group was not affected. Berla reasoned that the youngest children benefited from training because they had not yet established habitual search patterns, whereas the instructed search strategy may have interfered with search patterns that the older children had already established.

Summary. Davidson (1976) argued that the better search strategies facilitate the representation of the stimulus in memory, with attention as the mediating process: as attention is more organized, so is search, the result being more effective encoding of tactually perceived information. Whatever the exact mechanism, it is clear from the work of Berla and Davidson that more systematic strategies lead to better performance, and that furthermore, strategies can benefit from training.

Integration of information from different sensory modalities

There is extensive literature on sighted children regarding issues of intermodality relations, and particularly on what happens when information about events is received simultaneously from two or more sensory modalities. The literature on the visually impaired population is more limited, but several interesting studies illustrate important points, particularly about the perception of spatial structure.

One issue is the relative effectiveness of the perception of spatial and temporal structure. O'Connor and Hermelin (1972a) addressed this question using auditory stimuli that were distributed both temporally and spatially. A sequence of three spoken digits was heard from an array of three spatially separated speakers. The sequences were designed so that when asked for the "middle" digit, the child would have to choose between the digit that had occurred at the spatially middle speaker and the digit that had occurred in the temporal middle of the sequence. The children were 13-year-olds who had been blind since birth.

The results were very clear: the overwhelming choice was of the temporally middle digit, rather than of the digit that sounded from the spatially middle speaker. This pattern was strikingly different from that found with sighted children who saw the sequence of three digits at different spatial locations rather than hearing them. In this condition, the sighted subjects overwhelmingly reported the digit from the middle spatial location, rather than the digit in the middle of the temporal sequence.

Children with hearing impairments responded in much the same manner as did the sighted children in this condition. O'Connor and Hermelin argued that blind children do not naturally encode spatially distributed auditory information in terms of its spatial distribution (and that hearing-impaired children, correspondingly, do not naturally encode spatially distributed visual information in terms of its temporal distribution).

Battacchi, Franza, and Pani (1981) similarly evaluated children's ability to process the spatial structure of auditory events. They used a semicircular array of six loudspeakers that were separated by at least 25 degrees and therefore highly spatially discriminable from one another. A sequence of six names was heard, one from each speaker, at a rate of one per second. In the congruent condition, the sequence started at one end and proceeded regularly to the other end of the set of speakers, while in the incongruent condition the order of the names did not correspond to the spatial sequence of the speakers. After the presentation, the child was asked to say the names that had been heard at two of the speakers, chosen at random.

Sighted children perform this task better in the congruent condition than they do on the incongruent condition: apparently the spatial structure of the speaker array facilitates their processing of the auditory information.

In contrast, neither partially sighted nor blind children (ages 8 to 10) performed better in the congruent than in the incongruent condition: the regularity of the spatial sequence in the congruent condition did not facilitate their processing of the auditory information. In fact, the performance of the blind children was not above chance. The performance of the partially sighted children, however, was better than that of the blind. (A group of blind young adults did show a performance advantage in the congruent condition, suggesting that this ability develops, albeit slowly, with age.)

We should stress that in these experiments, the task is to process auditory spatial information. On the face of it, there is no reason to expect that impairment of vision should interfere with this ability. However, the empirical evidence is clear.

In another approach to the concept of "middleness," O'Connor and Hermelin (1972b) assessed the encoding strategies of seven- to nine-yearold blind children in a three-term series problem. Two types of sentences were constructed, each expressing a relationship among three items. In one type, the sequential order of presentation corresponded to their logical relationship, and in the other the sequential and logical orders did not correspond. The child was asked two types of question, one dealing with the logically middle member of the triad and the other dealing with one of the two logically extreme members. There was a tendency to report the sequentially middle item when it was incorrect. O'Connor and Hermelin suggested that the blind children did not have a readily available spatial code for use when it was appropriate, and instead tended to rely on a temporal code even when it was inappropriate.

Axelrod (1968) used still another method to approach the same issue: an oddity problem, in which the child is required to identify a characteristic that distinguishes one member of a triad from the others. Children who had lost vision earlier than 18 months had greater difficulty than children who had lost vision later than two years in learning such problems when the key characteristic was that the item occupied the temporal or spatial "middle" of a triad. Axelrod also evaluated the formation of intermodality learning sets, which have to do with the ability to transfer a solution from a problem learned in one sensory modality to a similar problem presented in another modality. When the problem was initially learned tactually or auditorially and then presented in the other modality, children with early visual loss were again worse than those with later visual loss.

With respect to the question of information-processing strategies applied to spatial and other tasks, Millar (1981a, 1982) argued that there is nothing inherently different in the information-processing capabilities of blind children, but rather that preferred strategies develop as a result of the typical ways that children gain their primary information. Thus, with respect to spatial perception, "If blindness leads subjects to neglect external cues, they will learn less, and know less about directional connections between external cues. This, in turn, strengthens the preference for strategies derived from the remaining modalities" (1982, p. 119).

This is not to say that the strategies actually chosen for spatial (and presumably other) tasks are necessarily the optimal ones: a visualization strategy may not be optimal for a given task, but the child with residual vision may use it nonetheless because of the effectiveness of visualization in many prior experiences. Similarly, the blind child may have external spatial-referential strategies available but tend not to use them because the primary source of spatial information (touch) tends to elicit internally referenced strategies. Similar conclusions were reached by Robin and Pecheux (1976), working with tasks requiring reproduction of two- and three-dimensional spatial models.

Summary. It is clear that the processing of even nonvisual information about spatial structure is hampered by impaired vision. These results again underscore the important role of vision as a vehicle for the organization of spatial structure, regardless of modality. However, the issue is complicated by the issue of information-processing strategies: strategies tend to be selected based on the particular sensory modality through which information is received. Although this association of strategy and sensory modality may be natural, Millar suggests that it is not inviolable.

The implication is that training studies designed to help children select appropriate strategies of information processing may prove useful.

Verbal and phonological issues in encoding and memory

We turn now from spatial issues to those related to the encoding of verbal and phonological information. Much of this research uses braille characters as stimuli: these are especially interesting as research stimuli because they have both tactual and verbal-phonological properties. Our intent here is not to review how braille characters are learned or how braille reading is acquired, but rather to examine the nature of encoding and memory of verbal and phonological information, particularly as it is obtained via touch. The issue, in short, is the nature of encoding of information in memory.

Tactual versus phonological encoding

As we noted earlier, Millar (1975a) demonstrated that tactual information is stored in memory in a specifically tactual form, since interpolated activity of a tactual nature during a delay interfered specifically with the recall of tactually experienced information. Millar (1975b) examined the corresponding question with braille stimuli. That is, would braille stimuli, with both tactual and phonological properties, be stored phonologically, tactually, or perhaps in both forms?

Three sets of stimuli were used, one consisting of items that were tactually dissimilar but phonologically similar, another of items that were tactually similar but phonologically dissimilar, and the third of items that were dissimilar both tactually and phonologically. Set size ranged from two to six items. The blind children ranged in age from 4 to 12 years and had lost vision within the first 18 months of life. They were screened by pre-test to ensure their ability to discriminate the letters tactually, and in the case of the older children, to identify the letters. The child felt each letter of a sequence in succession, then was given one of the letters and asked to indicate where it had occurred in the series.

Evidence for both phonological and tactual interference was found for all ages, indicating that both the phonological and the tactual properties of the stimuli were encoded. However, the younger children tended to show stronger evidence of tactual than phonological encoding. It was also clear that different processes were involved for tactual and phonological information, since there were different relationships of tactual and phonological interference effects in relation to overall memory demand.

Additionally, there was a tendency for children with higher IQ_generally to perform better than those with lower IQ.

Overall, Millar's (1975b) results for braille stimuli corroborated her (1975a) results for purely tactual stimuli in confirming that tactual encoding does occur. However, the addition of phonological properties added a specifically phonological form of encoding as well.

It is well known that the grouping of items within a serial string of verbal material facilitates memory of that material. Presumably such a grouping effect should also occur with tactual material that has phonological correlates. Indeed, Millar (1978) used strings of braille letters and found that grouping facilitates recall. This result further corroborates the evidence of phonological influence on the tactual encoding of verbal material. However, would similar facilitatory effects of grouping occur with tactual material without verbal association? The answer was a clear no: when the stimulus strings were nonsense shapes without phonological correlates, grouping actually interfered with recall. The results supported the hypothesis that tactual encoding is significantly different when the stimuli have verbal associations than when they do not. Overall performance improved with increasing age over the 7- to 11-year range, but the difference between associative and nonassociative stimuli did not change with age. Mental age (as well as digit span) was similarly related to overall performance but also did not interact with stimulus type effects.

These findings constitute further evidence for the existence of different memory processes for verbal and tactual information, and particularly for the interaction of these processes when verbal-phonological information is involved.

At another level of phonological-tactual interaction, pronounceability may facilitate braille letter recognition. Such a study was reported by Pick, Thomas, and Pick (1966). The subjects were braille readers ranging in age from 9 to 21. They varied in age at visual loss, amount of residual vision, and braille reading experience. The stimuli were letter groups containing from three to six characters. In one condition the stimuli were pronounceable, whereas in another condition, the letters of each group were rearranged to render it unpronounceable. The child's task was to scan the letters tactually and name each letter as quickly as possible. It was hypothesized that for the pronounceable stimuli, the sound sequence would facilitate discrimination of the letters. Indeed, there was a dramatic speed difference in favor of the letters occurring in pronounceable groups, and fewer errors occurred for letters in these groups, again showing the facilitative role of phonological context. The results did not differ as a function of either age or braille reading experience; variation with age at visual loss or amount of residual vision was not reported.

Summary. In sum, the evidence about tactual and phonological encoding supports the view that when stimuli have both tactual and phonological properties, these are both encoded. The two kinds of encoding follow somewhat different processes. However, these processes operate interdependently, particularly in the encoding of braille.

The role of touch in semantic coding

In the case of print reading, if letter sounds are encoded one-by-one, as novice readers may be inclined to do, reading suffers. This is particularly so if the sound of the word is not congruent with the sequential sounds of the letters. (For example, it is difficult to arrive at the sound of the word eat by combining the sequence of sounds of the letters e-a-t.) Does such an effect also occur for braille? In fact, the effect might be stronger because of the sequential nature of encountering braille characters.

Pring (1982) asked whether a phonological code is generated as the braille letters are contacted tactually, or whether touch simply serves as a channel, with phonological encoding occurring at some higher level. If individual letters are phonologically encoded, a phonological-lexical conflict would be generated in an incongruent condition where the sounds of individual letters do not match their role in the sound of the word (e.g., in the pair steak-leak) and reading should be slower. In contrast, in the congruent case (e.g., stake-leak) no conflict is present and reading should be faster.

Pring studied children who were congenitally blind, were rated as good braille readers, and ranged in age from 11 to 13. Relatively few errors were made in reading the word-pair lists, and the primary analysis was of response latency. The time required to read the target word was longer in the incongruent than in the congruent condition. This result supports the notion that phonological encoding occurs at the tactual level, and thus that pronunciation is constructed by assigning phonological properties to individual letters and generating the sound of the word (and thus eventually its meaning) by combining the individual sounds.

However, if the children used only this form of tactual-phonological encoding, then errors of pronunciation would be expected for words with irregular spelling. That such errors were not generally found supports the notion of direct access via touch to the known meaning of the word, rather than a constructive process from individual letters. In short, the results yield evidence for both processes.

In a second experiment with the same children, Pring tested the hypothesis that this process is mediated specifically by grapheme-phoneme (letter-sound) correspondence. This was done by having the children read lists of words that were either regular (words whose pronunciation corresponds to their spelling, such as wood and dance) or irregular (those whose pronunciation does not correspond to their spelling, such as pint and talk). Errors were again few, and latency data showed that regular words were pronounced more quickly than irregular words. This suggests that extra processing is required for irregular words, or that an alternative processing route is used for them.

Overall, these results show interesting parallels to the processes involved in reading print visually, but they do not yet yield a clear picture of how the processes work.

Attention to features of braille stimuli

In earlier sections we addressed the question of the relative effectiveness of different information-processing strategies in mediating performance in tactual perception, and the issue arises here as well for the perception of braille stimuli. Millar (1984) investigated children's attention to various features of braille characters, exploring in particular the relative attention to phonological and tactual properties of the characters in relation to the children's level of skill in braille reading. The children were congenitally blind and ranged in age from 7 to 12 years. They were divided into three reading groups based on reading rate: these groups overlapped substantially in age and were moderately differentiated by IQ_ score.

The test was a multidimensional oddity problem. Three stimuli were presented on each trial, and the child had to choose which of the three stimuli was the "odd one." The child was instructed that a stimulus could be "odd" by differing in meaning, sound, shape, or number of dots from the others. From the child's choice it is possible to discern which of the dimensions governed the child's choices.

Differences were found between the reading groups. Faster readers based their choices more on semantic features and less on shape features.

This choice pattern was even more highly related to mental age than to reading level. Each child was additionally classified as a normal or a retarded reader, based on whether or not his or her reading proficiency score was within a year of chronological age. The normal readers did not show predominance of any of the dimensions, whereas the retarded readers tended to focus on phonological features. This indicates that their reading strategy was to construct the sound of the word by combining the sounds of the individual letters, which also characterizes poorer sighted readers.

In a second experiment, children were instructed to use specific features for judgment. Faster readers were better able to respond to different features in accordance with the instructions, whereas slower readers were less able to escape their own spontaneous strategies.

Thus, Millar's work shows that there are relationships between attentional propensities and reading capability. The direction of causality, of course, is elusive.

Pring (1984) used a word-recognition task to explore a similar issue at the semantic level. The question at issue was the degree to which semantic or tactual information would govern children's ability to recognize words. Word pairs were constructed to contain semantically related (e.g., bread-butter) or unrelated {nurse-butter) members. Nonword combinations were also included. The child's task was to determine, as quickly as possible, whether or not the stimulus was an English word. Words in semantically related pairs were correctly recognized faster than those in unrelated pairs. This semantic facilitation effect is evidence that the children attend to the semantic context while processing information about an individual word.

However, when the braille stimuli were tactually degraded by physically reducing the height of the braille dots, the semantic facilitation effect did not occur. Apparently reducing the legibility of the braille dots redirected the child's attention from the semantic to the perceptual characteristics of the stimuli. The children were congenitally blind, were of normal to high intelligence, were relatively experienced braille readers, and averaged 10:6 years of age.

Summary. The implication of the Millar and Pring studies is that there are indeed individual differences in children's attentional propensities, and furthermore that these are related to reading level. Pring's results in particular support the view that the child's attention is limited and that its allocation is flexible, depending on the balance of cognitive and perceptual task demands.

The relationship between verbal and pictorial information

When sighted children look at pictures, their recognition and memory can be facilitated by accompanying verbal information: this is an example of verbal mediation. The question arises whether a similar phenomenon occurs when the verbal information is experienced via braille. Pring (1987) examined the role of verbal mediation in the recognition of pictures by congenitally blind children ranging in age from 7 to 16, all of whom were braille readers. The picture stimuli were raised-line drawings of objects with which familiarity could be expected (e.g., shoe, hand, or sofa). In a matching task, each picture was paired with a word and printed in braille; the child's task was to report whether the word went with the picture ("same") or not ("different"). Following this, a recognition test was performed in which a series of pictures and words was presented.

Some of the words and the pictures had been experienced in the matching task and some had not. The child's task was to judge whether each item had or had not been encountered in the matching task.

Performance was in general very good. For recognition of pictures, performance was best for pictures that had been encountered together with the matching word, and specifically better than for those with a mismatched word. Pring suggested that the results may reflect not verbal mediation in the true sense, but rather a dual encoding of the stimuli such that the picture and the verbal information are encoded in parallel.

Whatever the exact mechanism, association of verbal material and pictures clearly occurred. In fact, this association also operated in a negative way in the results. Having encountered a word in the matching phase increased the likelihood that an erroneous "yes" recognition response would be given to the picture corresponding to that word. That is, false positive picture recognition responses were made as a result of previous exposure to the word.

Although it is not strictly an example of verbal mediation, the question also occurs whether there might be a reciprocal effect, such that picture information aids in the recall of verbal information. Pring and Rusted (1985) found positive evidence. A short prose passage containing specific facts corresponded to each of six raised-line pictures of animals or plants that were explored tactually. Each of the passages contained information represented in the picture, as well as other information not depicted. The child heard the prose text twice, once with and once without the picture available. When the picture was available, the child was encouraged to explore the picture and identify features as they were mentioned in the text. Immediately after each text, the child was asked to describe the subject of the text (and picture). After completing the trials, an interpolated task of braille letter naming was used for 15 min, after which a delayed recall task was given.

Immediate recall for pictured information was, not surprisingly, better than delayed recall. When the text was accompanied by the picture, immediate recall of the depicted information was better than when there was no picture. Thus, there was a positive effect of depicted information, indicating that an effective association was being made between the verbal and the pictorial information.

When recall was delayed, the results were more complex. Two groups were tested, one of congenitally blind 13- to 15-year-olds, and another that had lost visual function after age two. Two-thirds of the latter group had some residual pattern vision, although it was not sufficient to discriminate the pictures. On delayed recall, the group that lost vision later showed the same pattern of facilitation by pictured information as in the immediate recall task. The pattern of the congenitally blind children showed that pictured information was recalled better in the illustrated condition, but that nonpictured information was better recalled in the unillustrated condition. Pring and Rusted suggested that this was a result of attention: the availability of a picture draws the child's attention to the textual information pictured, at the expense of information not pictured.

Further, when a picture is not available, the child devotes less attention to information that is picturable. This is a provocative pattern of results, since it suggests that strategies of attention and information encoding differ as a function of early vision.

Summary. Several major summary points emerge from this body of research. First, it is clear that when stimuli have both tactual and phonological properties, as in the case of braille characters, separate processes of tactual and phonological encoding occur. However, these forms of encoding can have reciprocal effects on one another, and thus they are not completely independent.

Second, the evidence shows that the child does not operate with a limitless reservoir of attention, but instead allocates attention variously to tactual, phonological, and semantic features of letters and words as the demands of the task vary. Third, in the case of tactually perceived pictorial stimuli, there are clear effects of related verbal information, and the reciprocal influence also occurs.

This body of evidence has been primarily directed to demonstrating the operation of basic processes of information encoding and to elucidating the variables that affect their operation. This is certainly a valid and valuable pursuit. However, the literature has generally not addressed issues of individual differences, aside from some interesting evidence of strategy variations in relation to reading.

Imagery

For decades there has been interest in the nature of the mental images that blind adults and children have, and in how their imagery may vary as a function of such variables as partial vision or an early period of visual experience.

For example, Fernald (1913) reported a study of imagery in two university students, one blind from birth and the other partially sighted.

Reportedly, the latter used visual imagery abundantly whereas the totally blind student never used visual imagery. Schlaegel (1953) reported interesting variations between children with differing amounts of vision in the imagery characteristics that words evoke. Test words and phrases were presented orally to the child, who was asked to report the sensory experience evoked by the "first mental image." The predominant image reported by visually impaired children, as by sighted children, was visual.

The visually impaired group was divided into three subgroups. The predominance of visual imagery varied regularly with the amount of residual vision: those with the least vision reported the fewest visual images, and those with the most visual capability reported the most. It may be that children with partial vision did indeed experience a greater frequency of visual images, but an equally plausible explanation is that there was a response artifact, such that children in this subgroup were more inclined to report visual images.

As a procedural matter, the difficulty of studying imagery should be noted. Two approaches are possible, and each involves its own assumptions.

On the one hand, the child may be asked, as in the work by Fernald and Schlaegel, to describe the nature of his or her images. This procedure is open to the question of whether habits of language use are artifactually biasing the outcome: are reported variations in imagery really that, or just differences in the use of particular words to report images? Generally, this approach cannot generate unequivocal results.

The second approach involves the functional aspects of imagery: tasks can be designed on which performance should differ in predictable ways depending on the images hypothesized to be involved. The work on mental rotation, discussed in an earlier section, serves to illustrate this point. For example, when Carpenter and Eisenberg (1978) found longer reaction times to make judgments about letters that were rotated from the upright, they reasonably concluded that imagery must have been the mediating mechanism, and specifically that cognitive rotation of a mental image had occurred. As sound as this reasoning may be, it is good to remember that images are being examined not directly, but indirectly by inference based on the nature of their mediation of behavior.

Imagery in spatial tasks

The imagery work may be broadly divided into that which involves the use of imagery in performing spatial tasks and that which involves imagery in other learning tasks. We have considered much of the work on spatial behavior in earlier sections, and a brief mention of the imagery aspects of that work should suffice here. Both Knotts and Miles (1929) and Duncan (1934) asked their subjects to report the nature of their imagery in solving maze problems: in both cases, subjects who reported using a verbal approach performed better than those who reported using visual images or kinesthetic-motor images. Worchel (1951) used tasks involving various tactually perceived shapes and solicited reports of the subjects of the nature of their imagery. The responses of the congenitally blind subjects tended to refer to the "feel" of the shapes, whereas the adventitiously blind subjects tended to refer to "mental pictures." Worchel interpreted this result as indicating that visual imagery results from early vision. Interestingly, the performance of the later blind subjects was better than that of the congenitally blind, thus suggesting that visual imagery can effectively mediate performance.

Imagery in verbal tasks

We turn now to the issue of imagery in learning tasks, which have involved primarily verbal material. It is known that words that evoke images are easier to learn, for example in a paired-associate task, than those which do not. (The typical paired-associate task involves the presentation of a list in which both words appear, then testing for the recall of one word with the other as a cue.) This paradigm has been used to assess the imagery of children with visual impairments. Kenmore (1965) studied third and sixth graders from schools for the blind. About half of the children were blind, while visual function in the remainder ranged from object perception to 2/200: age of visual loss was not reported. The speed of paired-associate learning was assessed in conditions involving verbally and tactually presented material of varying familiarity.

Overall, the sixth graders learned more quickly than the third graders, and children with higher IQ_ scores performed better than those with lower scores. No variation in results as a function of residual vision was reported. More substantively, Kenmore hypothesized that since the school experience of visually impaired children is highly verbally structured, it should lead to stronger verbal imagery in older children because of their longer experience in the environment. The older children were indeed better than the younger in learning verbally presented pairs.

Conversely, the older children were worse at learning tactually presented material. Kenmore suggested that this may be a result of the relative neglect of tactual learning strategies in schools for the blind, which should in turn lead to less tactual imagery. (The inferential dangers of imagery work are evident here: since imagery was not measured directly, its role as a mediating mechanism is uncertain even though the results are consonant with that formulation.)

Paivio and Okovita (1971) studied visual and auditory imagery using a paired-associate learning paradigm. The congenitally blind children, whose ages were 14 to 18 years, were all "above average" in IQ Lists of word pairs were constructed to be high in both visual and auditory imagery (e.g., ocean-clock), or high in visual but low in auditory imagery (e.g., green-palace). Performance was significantly better throughout for words with high auditory imagery, although the children did learn both the high- and low-auditory imagery lists.

In a second experiment, pair lists were created to contain words with high visual and low auditory (and tactual) imagery, or with high auditory and low visual (and tactual) imagery. Again, performance was better with the lists that contained words high in auditory imagery, although the differences decreased over the course of the experimental session. Both experiments clearly showed the ability of the children to benefit from auditory imagery, and a relative lack of ability to benefit from visual imagery. No variation in results was reported as a function of age or IQ_, but this is not surprising given the limited range of these variables in the sample.

On the other hand, Zimler and Keenan (1983) studied younger children, 7 to 12 years of age, who had lost sight within the first six months of life. Word-pair lists of four types were created: high visual and low auditory imagery in both (V-V), high auditory and low visual imagery in both (A-A), and mixed imagery (V-A, A-V). The children's performance did not differ as a function of list type. The lack of an advantage for the lists with high auditory imagery stands in contrast to the results of Paivio and Okovita (1971). Although the children were younger than those of Paivio and Okovita, it is not clear how this variable may have affected the results.

Again we can turn to the issue of cognitive strategies in pairedassociate learning, with a study by Martin and Herndon (1971). The words were not chosen for their modality-specific imagery; rather, the purpose of the study was to investigate the nature of verbal strategies in remembering word pairs. One member of each pair was a real word (and therefore presumably "imageable"), while the other was either a pronounceable nonword or a very low-frequency word (presumably less imageable). In a control condition, children were not instructed as to strategy, whereas in an "aided" condition they were instructed in the use of associative strategies such as recognizing superordinate relationships between the two members of a pair. Learning performance was significantly better in the aided condition.

The children's reports of their strategies were classified according to the type and complexity of cues used (after Martin, Boersma, & Cox, 1965). There was a significant correlation between performance and the level of associative strategy. This result, coupled with the overall superiority of the aided group, suggests that learning is better when associative strategies are used, and that such strategies can be effectively instructed.

The results of studies using the paired-associate learning task suggest a facilitory role of auditory imagery in paired-associate learning, although the relationship to CA is uncertain. Furthermore, there are variations in performance with associative strategy.

Serial learning tasks have also been used to investigate the role of imagery in learning. In this paradigm, the subject's task is to learn items presented in a serial list. After each run-through of a list, the subject is asked to remember as many of the items as possible, either in order or not (free recall). Craig (1973) used this method with adolescents whose IQ scores were in the normal range. Age at visual loss ranged from birth (70%) to six years, and all were braille users. Lists of high- and lowimagery words were created (the imagery characteristics of the words were not further specified). More items were recalled from the high- than from the low-imagery lists. In serial learning tasks there is a general tendency to find higher recall of items both early and late in the list than in the middle; this effect was found for both high- and low-imagery lists.

Groups of hearing-impaired, visually impaired, and sighted subjects were tested. The subjects with both sight and hearing performed better than either the visually impaired or the hearing impaired (who experienced the lists visually rather than auditorially). Following the reasoning of Paivio and Okovita (1971), Craig concluded that sighted subjects have two codes (visual and auditory) potentially available for mediating the task, whereas the visually impaired and the hearing-impaired subjects do not perform as well because in each case one of the codes is unavailable.

Although there was apparently some range of visual function in the visually impaired group as well as variation in the age of visual loss, the possible relationship of these variables to performance was unfortunately not reported.

In the research noted above, Zimler and Keenan (1983) studied the free recall of serial word lists that differed in their common attributes.

Three attributes were used, "redness," "loudness," and "roundness." The rationale for this choice was that the visual attribute "redness" should facilitate recall by sighted children, the auditory attribute "loudness" should facilitate recall by blind children, and the attribute "roundness," which is accessible both visually and tactually, should facilitate the two groups equally.

Four words of one category were presented seriatim, then four of the next, and four of the third. The blind children were indeed better at recalling the "loud" words, but they were also better at recalling the "round" words than the sighted children, and contrary to expectations, they were equal to the sighted in recalling the "red" words. These results do not correspond to expectations based on a modality-specific coding hypothesis.

Summary. There is no doubt, based on this work, that imagery facilitates verbal learning. More specific questions arise about the specific form of imagery and how it exerts its effect. On the one hand, some results (e.g., Craig, as well as Paivio & Okovita) support the notion of modalityspecific imagery, and specifically that visual imagery is not facilitative of learning by children with visual impairments, while auditory imagery is.

However, other results (e.g., Zimler & Keenan) cast doubt on the modality-specific formulation. It is possible that this varies with individual differences characteristics such as visual experience, but for the most part the research has unfortunately not explored this issue. Kenmore's work raised an important issue in finding age-related shifts in imagery and in questioning whether these are experience related. There are obviously many unanswered questions in this area.

Developmental shifts in imagery

According to Bruner (1966), experience is encoded in a series of stages that proceeds developmentally from actions to images to symbols. Enactive representation refers to an action; ikonic representation refers to an image that is pictorial (and free of action); and symbolic refers to an arbitrary or more conceptual form of representation, as in the case of language labels. Hall (i98ia,b) used Bruner's framework for representation as a starting point. Based on a review of the literature on imagery in relation to the performance of various tasks by children with visual impairments, Hall suggested that these forms of representation may not be tapped in the same ways as in sighted children, and specifically that because of their experiential structure, children with visual impairments may not use ikonic representation as much but may rely more on symbolic and enactive modes of representation.

Hall designed a series of tasks to explore the use of representational modes in blind children who had lost vision within the first year. Three tasks were used: a concrete task, a verbal task with high-imagery words, and a verbal task with low-imagery words. Questions were designed to tap classification strategies, and specifically to show whether grouping would be done on the basis of perceptible (sensory), functional (referring to the function of an object), or nominal (name) attributes. It was expected that the children's classifications would be based primarily on perceptible attributes in the early years, with functional and nominal groupings more frequent with increasing age. In addition, the formation of equivalence groupings was expected to vary with the degree of concreteness and imagery level of the task.

In the concrete task, children tended to classify based primarily on perceptible attributes over the entire age range from 7 to 17. As expected, nominal and functional strategies increased slightly over age for both of the imagery tasks, although the use of perceptible strategies in these tasks remained high. Surprisingly, perceptible attributes did not diminish in use with age. Based on this result, Hall (1983) suggested that the use of concrete tasks in the educational setting may not promote cognitive growth and higher-level thinking skills. The relationship is evident between this possibility and Kenmore's (1965) suggestion of shifts in imagery tendencies as a result of school experience.

Summary. We should reiterate the difficulty of studying imagery, and particularly the danger of relying on subjects' reports of the nature of their imagery.

Nonetheless, some studies that obtain performance indicators along with self-reports (e.g., Worchel, 1951) tend to support the validity of selfreports.

Other studies use performance indicators such as the pairedassociate or serial learning task as a basis on which to infer the nature of imagery. Much of this work has been done with children who lost vision at birth or early in life and who have at most LP, and consequently information about visual experience variables is unfortunately lacking.

However, there is a body of evidence that suggests, though not unequivocally, that performance varies as a function of the imagery characteristics of the stimulus words; blind children's performance using words with visual imagery characteristics is not facilitated, whereas auditory imagery characteristics are facilitative. All in all, though, the literature on imagery is not very satisfying.

The End

 

ϟ  

Blindness and Children - David Warren

excerpt:
part II - chapter 6 in

Blindness and children - An individual differences approach
DAVID H. WARREN
Department of Psychology
University of California, Riverside Cambridge University Press, 2009

https://doi.org/10.1017/CBO9780511582288

Δ

22.Abr.2024
Maria José Alegre