LANGUAGE IN INDIA

THE ROLE OF VISION IN LANGUAGE LEARNING
in Children with Moderate to Severe Disabilities

Martha Louise Low, Ph.D.

The Role of Vision in Language Learning: Relationships between Visual Acuity, Looking Behavior, and Fast-Mapping of Novel Words onto Novel Objects in Children with Moderate to Severe Disabilities. A Ph.D. dissertation, University of Minnesota.

CLICK HERE FOR PRINTER-FRIENDLY VERSION.

CLICK HERE GO TO HOME PAGE.

CONTENTS

CLICK HERE FOR PRINTER-FRIENDLY VERSION.

CLICK HERE TO GO TO HOME PAGE

*** *** ***

A NOTE

At the time that this dissertation was completed in 1999, a few observations seemed worthy of further study, and have implications for effective language intervention for children with disabilities.

The first, a study related observation found on pages 59-62 (see Results, section on Differences in Inaccurate Following by Ocular Coordination Groups), is the finding that children with and without disabilities who have binocular coordination problems (e.g., "cross-eyed" or "wall-eyed") did not match the researchers visual target ("looking behavior" in the study) and did not fast map novel words to novel objects. Informally, these children seemed to tolerate greater amounts of "labelling ambiguity," without requests for clarification. At the time of this study, binocular coordination had not been studied or assessed in relation to children with language learning problems.

The second, an observation unrelated to the study, has to do with participants with autism. Although these children often had perfect visual acuity (no refractive errors), they usually did not make good use of novel visual intormation (e.g., a new toy). Instead, it seemed that children with autism were much more receptive to tactile information (e.g., touching objects). Sometimes the same object was ignored (or treated with irritation) when presented visually. This is significant because many interventions for children with autism use visual stimuli. Although these are often successful interventions, tactile methods may be useful for earlier stages of language learning or for children who are severely affected. At the time of this study, few interventions used tactile stimuli, such as three dimentional objects for language intervention.

Additionally, it was observed that children with autism were difficult to direct using temporal visual cues (e.g., a "vanishing" symbol such as a gesture). The study noted on page 74 (see Results, section on Intervention Conditions, Graph of Mean Looking Scores for Eye Gaze, Head Turn, and Show Gesture) that children with autism often did not respond to "slight cues" such as shift in researcher eye gaze, or head turn. However, when a dramatic arm gesture (show gesture) was used, which "cut a larger arc" in the child's visual field, then the child responded. This also has implications for teacher-pupil interactions during language interventions.

CONTENTS PAGE

CERTIFICATE

CONTENTS PAGE

ACKNOWLEDGEMENTS

I would like to gratefully acknowledge the many people and agencies that made this project possible. Those who facilitated my endeavor were with me at every step in the process of this research project. The first stage involved those on my committee, who helped me to conceive and refine the project. Special mention belongs to Dr. Hupp, my adviser, for the endless stream of advice she was asked to produce on my behalf. I appreciated her keen ability to direct this work in an efficient manner. Also, my committee members, who graciously gave me time for presentations and consultations, included Dr. Bart, Dr. Bauer, Dr. Rynders, Dr. Samuels, Dr. Weiss, Dr. Wilderson, and Dr. Windsor.

Funds for the completion of this project were granted by the Graduate School Fellowship Office at the University of Minnesota in the form of a Doctoral Dissertation Special Grant. Their welcome support of this project has enabled the timely completion of this project. Funds were used for translation costs, transportation, materials, copies, and postage.

Dr. Knowlton facilitated the acquisition of necessary equipment and materials for the project (photorefraction camera and film). She also provided technical support for the vision screening photograph interpretation and woodshop expertise for the novel toys. Special Education Programs provided the use of a camcorder to videotape the children in this study. Other welcome donations of materials and services were made by My Yen Vo (Vietnamese translation), Louisa Watson (Spanish proofreading), Chin Ching-Yu (1/2 inch videotapes), Shelly Kothe (familiar children's toys), the Low's (use of the family car), and the public school districts (copies of consent forms). Many thanks to Sandi Cassavant who coded the data for Interrater Observer Reliability, and for her friendship throughout the graduate school experience.

The participants were recruited from seven public school districts and a church children's program: Minneapolis Public Schools, St. Paul Public Schools, Roseville Public Schools, Richfield Public Schools, Bloomington Public Schools, Independent School District #196, Independent School District #191, and Bloomington Assemblies of God Church. There were countless administrators and teachers at these locations who helped me with the research approval process, and the identification of potential participants. A huge thank-you to every parent or guardian who signed and returned a consent form for his/her child's participation in the study.

Last, but not least, I gratefully acknowledge my loving husband, Kevin Low, and my son, Derek Low for laboring with me, and for gracing my life with their presence, love, and encouragement.

CONTENTS PAGE

DEDICATION

This work is dedicated to Jesus Christ of the Scriptures whom I thank so much for everything.

CONTENTS PAGE

ABSTRACT

The purpose of this study is to examine a particular strategy of language learning in young children with moderate to severe disabilities. At around 18 months, typically developing children begin to use the direction of a speaker's eyes to guess the meaning of some words (e.g., objects). The use of this strategy coincides with increased vocabulary growth, and may be a result of strategy use. Children with moderate to severe disabilities may also use this strategy, but it is not known if problems with visual acuity (e.g., near-sighted, far-sighted), which are more prevalent in moderate to severe populations, may hinder their ability to reference a speaker in this way. Although language delays are known to be related to developmental delays and/or cognitive deficits, it is not known whether these delays or deficits indirectly affect language development through inefficient interpretation of degraded visual information (e.g., poor visual acuity).

This study videotaped children with and without disabilities as they participated in a novel word learning exercise. After verbalizations containing the novel word, participants' responses were coded for eye gaze targets (speaker and direction of eye gaze). In one condition, the researcher's gaze was directed toward the child's novel toy. In the other, the researcher's gaze was directed toward her own novel toy. After this, the child was asked to identify which of two objects was the novel toy. Additional intervention conditions were administered to children that were not able to accurately map the novel word to the novel object. These conditions accompanied the novel word with shifting eye gaze, head turns, and gestures. In addition, visual acuity was measured by the use of a photo-refraction camera. Photographs from this type of camera are able to reveal the presence and degree of many mild and severe visual conditions (e.g., near-sighted, far-sighted, misalignments [strabismus], amblyopia, cataracts).

Data were examined for associations between subject characteristics and (1) visual acuity, (2) visual targets, and (3) language mapping ability. Other analyses were made for associations between (4) visual acuity and visual targets, (5) visual targets and language mapping, and (6) visual acuity and language mapping.

CONTENTS PAGE

LIST OF TABLES

Table 1 - Pearson correlations of the progressed age measures.

Table 2 - Spearman rank correlations of age progressions and looking variables.

Table 3 - Incidence of problems with acuity in the cognitive groups.

Table 4 - Incidence of ocular coodination problems in the cognitive groups.

Table 5 - Incidence of visual acuity problems for the diagnosis groups.

Table 6 - Incidence of ocular coordination problems for the diagnosis groups.

Table 7 - Spearman rank correlations of cognitive indexes and refractive error.

Table 8 - Frequencies of inaccurate mapping for cognitive groups.

CONTENTS PAGE

LIST OF FIGURES

Figure 1 - Graph of the chronological ages of participants

Figure 2 - Graph of the cognitive age progressions of participants.

Figure 3 - Graph of show gesture find-it scores for the diagnosis groups.

Figure 4 - Graph of total find-it scores for the diagnosis groups.

Figure 5 - Graph of looking variables for the entire study sample.

Figure 6 - Graph of total sequential scores for cognitive group.

Figure 7 - Graph of total sequential looking for diagnostic groups..

Figure 8 - Scatterplot of refractive errors and cognitive indexes.

Figure 9 - Graph of inaccurate follows for occular coordination groups.

Figure 10 - Graph of inaccurate mapping for ocular coordination groups.

CONTENTS PAGE

CHAPTER 1
INTRODUCTION

Children with moderate to severe disabilities are expected to have problems with language learning. Delayed language also is expected in cases of mental retardation (e.g., Down syndrome), with abnormal language development documented in cases of autism. Language research has given strong validation to social-interactionist theories of language development. In these paradigms, vision is an assumed, and necessary, skill in language development. The role of vision has been documented in various ways. For example, child vocabulary size has been correlated with mother-child joint attention characteristics (Goldfield, 1985-86-, Tomasello & Todd, 1983) and child joint attention skill in general (Morales, Mundy, & Rojas, 1998; Mundy & Gomes, 1998). Visual regard of a communication partner is part of the defining features of infant communication (Bates, Benigni, Bretherton, Camaioni, & Volterra, 1979). Effective language interventions usually emphasize joint attention activities or visual strategies for language learning (Matson, Benavidez, Compton, Paclawskyj, & Baglio, 1996).

There are a number of broad considerations to make when proposing a relationship between visual and language abilities. Children with refractive errors and normal cognitive abilities have no known difficulty with language learning. Children who are blind and have normal cognitive abilities are somewhat delayed in the development of speech but, eventually "catch up" to their sighted peers (Ferrell & Raver, 1991 - Hatton, Bailey, Burchinal, & Ferrell, 1997). However, children who have visual impairments and have cognitive disabilities are even more delayed in their language development. The cognitive deficits of this latter group of children may hinder them from making efficient use of their remaining vision. The concern of this research study is whether or not children with cognitive disability also are hindered in their language learning by the presence of refractive error (e.g., nearsighted or farsighted). Cognitive disabilities may hinder the efficient use of relatively good vision when refractive errors are present.

With regard to language learning, children with moderate to severe disabilities may experience cognitive overload in a number of ways. (1) Language is a complex, integrated process which usually involves many coordinated systems in the acts of comprehension and production (e.g., gesture, speech, and eye gaze) (Capirci et al., 1996; Caselli, 1983 -, Franco & Butterworth, 1996; Iverson et al., 1994; Zinober & Martlew, 1985). (2) Visual abilities are the most efficient means of determining the meaning of referents (e.g., "moon" or "yellow"). Tactile exploration of referents, for example, takes more time to process than visual exploration. (3) Refractive errors such as hyperopia and myopia may possibly add to the cognitive burden of interpreting the meaning of words. Refractive error could obscure certain important aspects of referents or social cues (e.g., direction of eye gaze). Refractive errors, particularly hyperopia, are much more common in children with moderate to severe disabilities than in normal populations.

A number of researchers have noted that the rate and quality of word learning at 12 months and 18 months is vastly different, even though children of both ages are similarly motivated to communicate (Nelson, 1973, as cited in Baron-Cohen, Baldwin, & Crowson, 1997; Baron-Cohen et al., 1997; Bakeman & Adamson, 1984). Children learn to speak at around 12 months of age, and combine two words at about 18 months of age. Word combinations herald a vocabulary spurt, where word learning occurs at a much faster rate. Differences in joint attention skill have been hypothesized to account for these differences in verbal production. At younger ages, children may use passive acquisition strategies or association with perceptual characteristics (e.g., visual appearance, functional use, association of a heard word and an event) (Baldwin, 1993; Baron-Cohen et al., 1997). During the middle of the second year, infants may switch to more active strategies of word learning (e.g., following speaker's direction of gaze to a novel referent).

A number of skill developments may aid in the language acquisition. Joint attention increases in complexity and length over the first couple years of childhood. At around nine months, infants engage in dyadic emotional play and participate passively in joint attention routines initiated by their mothers (Bakeman & Adamson, 1984). Triadic infant exchanges (mother, infant, and object/event) increase from an average of 7 seconds at 6 months of age to 34 seconds at 18 months of age.

Baldwin (1993) demonstrated that typically developing infants reference their communication partners in response to verbalization in similar ways throughout the second year. However, the infants did not use the information to map novel words to their referents. Only the 18-19 month old children were able to use the speaker's direction of gaze to successfully label novel objects in both conditions.

Baron-Cohen et al. (1997) used the same procedures to test word mapping strategies of children with mental retardation and children with autism (MA = 2 yrs.). Children with mental retardation performed almost as well as normally developing children, but children with autism exhibited poor use of speaker's direction of gaze strategy.

It could be that refractive errors hinder the use of visual information such as speaker's direction of gaze, because the eyes of the speaker are obscured (e.g., not in focus for the listener with refractive error). Furthermore, refractive error may obscure the referents being referred to by language. In either case, language learning could be further delayed, due to the lack of visual information. Cognitive disabilities may further compound this problem due to the inefficient use of visual information and the complex nature of language processing itself.

CONTENTS PAGE

CHAPTER 2
REVIEW OF THE LITERATURE

This paper is an exploration of possible relationships between visual acuity and the acquisition of symbolic language. Children with moderate to severe disabilities are usually delayed in the acquisition of verbal or gestural language. Language interventions for these children focus on extension of social and cognitive abilities, but may inadvertently presume upon visual abilities in the process. Vision is often relied upon for comprehension, imitation, and production of diectic gestures. The meanings of words are also first decoded by visual determination of an environmental reference (e.g., fast mapping). Although the absence of vision (blindness) does not prohibit the development of language when cognitive abilities are normal, mild deficits in vision may present significant barriers to language learning in cases of developmental delay or mental retardation. When children have both cognitive deficits and significant visual impairments, progress toward all developmental outcomes is slower than that for visual impairments alone (Ferrell & Raver, 1991; Hatton, Bailey, Burchinal, & Ferrell, 1997). Would it be possible that refractive errors in children with cognitive delays could also present significant barriers to language learning? It seems logical to suggest that if referents and language partners are not clearly seen, this may unduly tax the limited cognitive abilities of children with developmental disabilities.

Language takes many gestural and verbal forms. This paper will be limited to those forms that are the most communicative and age appropriate for preschool children. Diectic gestures, symbolic gestures made with the hands, and words are used by normally developing children throughout their lives. Because communicative forms should be chosen with regard to current and future functioning, these forms are logical targets of interventions (Kaiser, 1993; Siegel-Causey & Wetherby, 1993) and will serve as the primary focus of this paper. Production of gestural and speech forms requires motor and/or hearing abilities, and may be beyond the capabilities of some children with moderate to severe disabilities. The alternative modes of communication taught in these cases will not be discussed here.

First, the normal development of presymbolic and symbolic skills will be traced and compared with specific examples of moderate to severe disabilities. The language skills of Down syndrome will serve as an example of cognitive impairments. Autism was chosen as an example of social and language impairment. The characteristics of these populations are representative of many preschoolers with moderate to severe disabilities. Early Childhood Special Education (ECSE) best practices favor social theories of language learning. This has practical significance for intervention, as the qualities of social settings are more easily manipulated than neural nets or linguistic acquisition devices. As such, "social comments" will be made on the process of language learning where applicable to the products of early language.

Next, the characteristics of visual refractive errors will be described. Normal development of visual abilities will be compared with selected moderate to severe disability populations. Many neurological and physiological attributes and abilities are needed in order to see. This paper will consider refractive errors (acuity), rather than color vision, functional vision (tracking, fixating, motor control of the eyes and head), or neurological impairments which may prevent visual processing. Of these, acuity is the most accessible to effective intervention. Three commonly occurring types of refractive errors will be considered: myopia, hyperopia, and amblyopia. Associations with intelligence and academic achievement will also be reviewed.

Finally, the role of vision in the language development process will be discussed. As mentioned before, even mild visual problems may create significant barriers to language acquisition when present with concurrent cognitive deficits. Current language intervention practices in ECSE will be addressed in terms of their sensitivity to refractive errors in populations with moderate to severe disabilities.

Normal Development of Verbal and Gestural Language

In all cultures the normally developing infant begins to speak single words at about 12 months of age, and to combine two words at about 18 months (Capirci, Iverson, Pizzuto, & Volterra, 1996; Slobin, 1972). Using verbal measures alone, this presents a rather abrupt picture of development. When incorporating information on the development of gestures, a much smoother picture emerges. This seems to be true of Western cultures, which are the settings for the research cited in this paper. In general, once forms are acquired, they are utilized throughout life, although the frequency of the various forms changes (Evans & Rubin, 1979; Pechmann & Deutsch, 1982). Reported below is research on production and comprehension of presymbolic and symbolic verbal and gestural development. Most studies agree on the criteria for verbal and deictic gestural forms. However, definitions for symbolic gestures vary widely (Haslett & Samter, 1997). An attempt has been made to use research that contributes specifically to symbolic gestures made with the hands, rather than facial expressions or whole body movements. However, because not much research exists in this area (Haslett & Samter, 1997), some studies reviewed here include differing measures. Comments on social theory are made which relate to the acquisition of presymbolic and symbolic products (e.g., use of the eyes for joint attention).

Definitions

The definition of a symbol is found in Bates, Benigni, Bretherton, Camaioni, and Volterra (1979): The comprehension or use, inside and outside of communicative situations, of a relationship between a sign and its referent, such that the sign is treated as belonging to and/or substitutable for its referent in a variety of contexts; at the same time the user is aware that the sign is separable from its referent, that is, not the same thing" (p.43). Words and some gestures qualify as symbols.

The definition of gestural forms is found in Erting and Volterra (1990, as cited in Capirci, Iverson, Pizzuto, & Volterra, 1996). Erting and Volterra have dichotomized gestures into two categories: 1) Diectic gestures (DG), which include gestures of giving, showing, requesting, and pointing. These gestures cannot be interpreted without inspection of the environment. 2) Representational gestures (RG), include hand movements, whole body movements, and facial expressions which represent an object. For instance, a child may use a motor action similar to that used with a familiar toy to represent that toy. RGs come to be associated with relatively stable meanings across different contexts of production. Because RGs represent the referent, even in their absence, they qualify as true symbols according to Bates et al. (1979). On the other hand, DGs "point to" a present referent in the environment and do not qualify as symbols.

Selected Features of Language Development

Diectic Gestures. Diectic Gestures appear at around 9 months of age when the infant begins to use presymbolic gestural forms for giving, showing, requesting, and pointing (Bates et al., 1979; Haslett & Samter, 1997). Communicative intent is usually defined by visual checking and persistence in the behavior until the presumed goal is met (Bates, 1979). Deictic forms arise from instrumental gestures which may not acknowledge their communicative effect upon another social agent. Instrumental behaviors are abbreviated, or shortened for production efficiency, and conventionalized into "agreed upon" forms for social efficiency. Requesting and giving appear to develop first, and later, showing and pointing (Bates, 1979; Zinober & Martlew, 1985).

Social awareness appears to motivate the production of diectic gestures (Sugarman, 1984). Diectic gestures coincide with the development of coordinated joint attention (Rogoff, 1990; Bakeman & Adamson, 1984) and are divided into two categories: (1) declarative gestures, which focus on sharing information, and (2) imperative gestures, which focus on requests. When the point is used for declarative purposes, young infants look at the communication partner immediately after pointing (Franco & Butterworth, 1996). By 14 months of age, the declarative point is preceded by looking at the communication partner and more frequent looking during the act of pointing. The point for requesting is not preceded by social reference of the communication partner.

More positive affect is associated with declarative joint attention behaviors than with requesting behaviors (Kasari, Sigman, Mundy, & Yirmiya, 1990; Mundy, Kasari, & Sigman, 1992). The pleasure that children have in communication may fuel the development of cognitive and social abilities involved in more precise forms of communication.

Declarative pointing is a precursor of the development of speech (Bates, 1979; Franco & Wishart, 1995; Haslett & Samter, 1997; Sugarman, 1983). Taken together, it could be hypothesized that declarative pointing is prerequisite or facilitative of speech development. However, some cultures of the world are not responsive to infants' prelinguistic communication and do not acknowledge the child as a communicative partner until after the child starts talking (Schieffelin & Ochs, 1983). The speech development of these children is not temporally different than children from Western cultures.

Words appear at around 12 months of age, but the range of variation is quite wide (Bates, 1979; Thal & Bates, 1988; Bates et al., 1994). Words are first bound to specific contexts, and later become decontextualized (Rescorla, 1980). The majority of the first 50 words in Western studies are nouns and, secondly, verbs and then, social/personal words (e.g., "mommy," "hi") (Bates et al., 1994; Iverson, Capirci, & Caselli, 1994; Rescorla, 1981). Children with predominantly noun filled vocabularies (referential style) learn faster than those who have more personal-social words (expressive style) (Bates et al., 1994; Nelson, 1973, as cited in Haslett & Samter, 1997). Various researchers have tried to argue that words are learned or encoded by their perceptual characteristics (e.g., Clark) rather than their functional properties (e.g., Nelson), such as motor movements that are performed on objects (Haslett & Samter, 1997). However, in naturally occurring language learning contexts, these two are confounded. Few first words have solely one or the other property (Rescorla, 1980). Gopnik (1988; Haslett & Samter, 1997) infered from his study that "social" words appear first (e.g., "dada" and names of people), and then the "cognitive" words appear after that (e.g., "gone" and naming objects).

Deictic words (e.g., here, there, you, they, this, that) are late appearing and infants often use diectic gestures to substitute for or clarify diectic words (Clark & Sengul, 1978; Capirci et al., 1996). Several studies have found that diectic words were still less frequent than diectic gestures at 20 months (Capirci et al., 1996: Iverson et al., 1994). Clark and Sengul (1978) revealed a trend toward partial verbal mastery at 1 year to full mastery by 5 years. Diectic words have many verbal forms but, in the gestural mode, one form (point) can simply be used for many referents, thereby reducing cognitive requirements and improving social interactions.

The social environment may influence child output. The proportion of word types in children's vocabularies may be related to maternal input. Caregiver style that facilitated 12 month old child focus was associated with more nouns at 17 months (Tomasello & Todd, 1983). Directive styles were associated with more personal-social words. Goldfield (1985-86) claimed that it is the context that the dyad jointly constructs which determines vocabulary content: the referential style child is associated with interactions focused on objects, the expressive style child is associated with interactions focused on performance and participation. In a cross-cultural study, Tardif, Shatz, and Naigles (1997) demonstrated that caregivers of cultures whose speech contained more nouns than verbs had children who produced accordingly, and vice versa. However, Caselli et al. (1995) argued that some languages (e.g., Korean and Chinese) lack the complex morphology that might be expected to slow the acquisition of verbs. They claim that oriental children are learning verbs sooner, which makes them appear to have larger verb categories at the same age. In either case, both studies show that child language is related to characteristics of maternal cultural.

Symbolic gestures

Symbolic gestures, like words, also appear at around 12 months (Acredolo & Goodwyn, 1988). Like words, symbolic gestures also undergo a process of decontextualization (Acredolo & Goodwyn, 1988; Caselli, 1983). These gestures can be spontaneously produced from an existing motor repertoire (e.g., pantomime of sniffing a flower to represent "flower") (Acredolo & Goodwyn, 1988). Symbolic gestures seem to incorporate functional elements rather than perceptual elements (Acredolo & Goodwyn, 1988). Word classes corresponding to the meanings of first gestures are roughly equivalent to first words; nouns predominate, verbs are second (Iverson et al., 1994). Parent interviews indicated that the gestural symbols are acquired for words that are not yet a part of child verbal vocabulary and that, once words are acquired for those items or actions, the gestures are no longer used (Acredolo & Goodwyn, 1988). Iverson et al. (1994) reported that gestural and verbal vocabularies had only a 10% overlap at 16 months of age. The frequency of gesturing was found to be at its peak around the time that infants acquired 10 words, and 80% of gestures were made by the acquisition of 25 words (Acredolo & Goodwyn, 1988). At that point gesture production became greatly reduced, although it was present throughout the 50 word period (Capirci et al., 1996). Combined findings indicate that as words increase in number, gestures decrease (Acredolo & Goodwyn, 1988; Capirci et al., 1996; Iverson et al., 1994).

The frequency of gesturing can be expected to vary according to the culture. As stated before, some cultures emphasize verbal forms of communication (e.g., Kaluli) (Sugarman, 1983; Schieffelin & Ochs, 1983). Research in cultures that gesture more (e.g., Italian) have shown differences with cultures that gesture less (e.g., American). Children from Italy showed a greater prevalence of gestural communication than that found in the American samples (Capirci et al., 1996, commenting on Goldin-Meadow & Morford, 1985). Factors within cultures that produce variability in onset may be related to caregiver availability. Both gesture onset and speech onset have been shown to be correlated with the number of hours per week spent in day-care; the higher numbers of hours were associated with higher ages of onset in both modalities (Goodwyn & Acredolo, 1993). Maternal availability would also explain why first-born children gestured more in Acredolo and Goodwyn's study (1988). Some symbolic gestures may be learned from adult input or from caregiver routines (Acredolo & Goodwyn, 1988; Capirci et al., 1996; Haslett & Samter, 1997; Zinober & Martlew, 1985). In general, the data seem to indicate that infants' communication partners determine child input, which may also be predetermined to some degree by partners' culture.

Timing of words versus gestures

A commonly debated issue is whether or not verbal and gestural symbols share a common cognitive base. Studies are inconclusive as to whether symbolic gestures precede first words. If onset is simultaneous for both modalities, this indicates support for maturation of some cognitive mechanism, which would employ either modality in the task of symbolic representation (Goodwyn & Acredolo, 1993; McNeill, 1985). Proponents of a "sign advantage" claim that gestures are easier to employ in tasks of reference because motor schemes are well established by the time infants develop referential abilities (Piaget, 1962, as cited in Thal & Bates, 1988). Werner and Kaplan (1963, as cited in Capirci et al., 1994) also theorized that "distancing" occurs through initial use of a motor action to represent a referent. Greater "distancing" is then made possible through use of the verbal mode. Goodwyn and Acredolo (1993) reported that a small, but statistically significant, gestural advantage exists (about 3 weeks, range .6-3.6 months). Even without a proven "sign advantage" the replacement of gestural vocabulary with verbal vocabulary suggests that infants find some words too taxing to produce until further cognitive development has taken place (Acredolo & Goodwyn, 1988). Furthermore, gesture advantage may be related to maternal sensitivity, a variable which attempts to measure caregiver responsiveness to infant signals for aid or comfort. The infants in Goodwyn and Acredolo's (1993) study who demonstrated a gesture advantage of at least one month differed from those who did not demonstrate an advantage, based on higher maternal education, a variable noted for correlation with maternal sensitivity (Bakeman & Adamson, 1986; Seifer, Schiller, Sameroff, Resnick, & Riordan, 1996).

Two-word Combinations. At around 18 months the typically developing child combines two words in a single utterance, often after a "critical mass" of 50 words have been acquired (Bates et al., 1994; Capirici et al., 1996; Slobin, 1986). Deaf infants also combine two signs (referential symbols) at the same time that hearing infants combine two words (Caselli, 1983; Goodwyn & Acredolo, 1993). A few studies have attempted to document the way in which two word combinations are associated with symbolic gestures. Authors of these studies often emphasize the individual variation exhibited by their subjects when attempting to make generalizations from the data (Capirci et al., 1996; Iverson, et al., 1994; Zinober & Martlew, 1985).

The most salient finding is that gesture+word forms are used throughout the second year, before and after the acquisition of two-word combinations (Capirci et al., 1996; Goldin-Meadow & Morford, 1985; Iverson et al., 1994; Zinober & Martlew, 1985). Goldin-Meadow and Morford (1985) also indicated that gesture+gesture forms also appear before the two word stage, although much less commonly. Two diectic gestures (give + point) make up the majority of these forms. These researchers report that gesture+gesture forms disappear once word+word combinations appeared. Specific types of gesture+word combinations were measured by Capirci et al. (1996) and Zinober and Martlew (1985). Data suggest a developmental trend where combinations of the two modes are first redundant, and, later, supplementary in their message content. Supplementary gesture+word combinations may appear just prior to two-word combinations (Capirci et al., 1996).

Before the end of the first year, diectic gestures are a primary means of intentional communication (Bates, 1979). Words and gestures appear separately at first, and then become increasingly coordinated during the second year (Zinober & Martlew, 1985). At the beginning of the first year, diectic or symbolic gestures and words are used equally often for communication tasks and finally, gestures take a more supportive role to verbal communication (Iverson et al., 1994; Zinober & Martlew, 1985). Especially illustrative of this is that the number of symbolic gestures decreases as words replace them (Acredolo & Goodwyn, 1988; Iverson et al., 1994). The symbolic gestures that continue to be produced seem to fulfill more supportive roles in communication (e.g., adjectives for number or size) (Iverson et al., 1994).

Children may show a preference for gestural or verbal modality before two-word combinations develop, regardless of the number of vocabulary items in either mode. These preferences were demonstrated at 16 months in Capirci et al.'s (1996) study but, by 20 months, the verbal mode was strongly preferred over the gestural mode. Iverson et al. (1994) measured increases in diectic gestures (68-80%) and diectic words (5-13%) at the same ages (16 and 20 months). Pointing accounted for the majority of diectic gestures at both age points.

From prelinguistic communication to two word combinations, the communication of normally developing children fulfills socially oriented purposes of behavior regulation, social interaction, and joint attention (Wetherby, Cain, Yonclas, & Walker, 1988). In normal children, these functions seem to emerge simultaneously. Context also influences the functions of child productions and confounds comparisons of study measures (Wetherby & Rodriguez, 1992; Zinober & Martlew, 1985).

Vocabulary prediction

Attempts have been made to predict vocabulary size toward the end of the second year by earlier verbal measures taken during the beginning of the second year. Predictions are improved by adding gestural measures or verbal comprehension measures (Thal & Bates, 1988). Frequency of gesture+word combinations have been correlated with later vocabulary size (Capirci et al., 1996; Goodwyn & Acredolo, 1993). Both age at the acquisition of 10 words and the propensity to imitate gestures at 17 months were related to later verbal vocabulary size (Acredolo & Goodwyn, 1988). Ability of 1.6 yr. olds to follow the gaze and point of an experimenter is positively correlated with receptive language development (Mundy & Gomes, 1998). Receptive vocabulary at 17 months is correlated with the tendency of infants to follow a point and gaze of a tester at 14 months (Mundy & Gomes, 1998).

Comprehension

Differences exist between the sizes of verbal comprehension and production vocabularies: children understand more words than they can produce (Benedict, 1979; Rescorla, 1981). Two stages have been described in the vocabulary development of 1-2 yr olds (Goldin-Meadow, Seligman, & Gelman, 1976). First, the children said fewer nouns than they understood. They said no verbs, although they understood many. In the next phase, they said virtually all the nouns they understood plus their first verbs. Children in the one-word stage can comprehend and respond to two-word directions, even when they describe actions unfamiliar to the child (e.g., "kiss book") (Sachs & Truswell, 1978). Children at 18 months of age accept novel gestures or words as communication (Namy & Waxman, 1988). Older infants at 26 months of age could understand novel words, but not novel gestures, unless additional practice was given.

Studies have been conducted to assess whether or not children use extra-linguistic cues to improve comprehension. Although gestures improve verbal comprehension in adults, gestures don't seem to help children under two years of age in the same way. Maternal gesturing seems to increase child attention, but not understanding (Schnur & Shatz, 1984). Children in Morford and Goldin -Meadow's study (1992) performed better with familiar diectic gestures (e.g., "point" & "give"), rather than "shake" and "throw" gestures, which were new to them. Morford and Goldin-Meadow demonstrated that children's comprehension was improved by gestures, particularly the redundant type, where words and gestures carry the same meaning. Supplementary gestures also produced higher levels of performance than the word alone condition.

The inconsistency of findings on whether or not gestures improve comprehension of words may be a problem with the chosen measures. The studies above used acts of compliance to indicate comprehension in children under one year old. It has been demonstrated that children aged 1.4 - 1.6 were more likely to act if a gesture accompanied a directive or was used alone (Allen & Shatz , 1983). In this study, children tended to respond in the mode that the directive/question was given in. If a gesture accompanied speech, the child responded with some kind of action; if no gesture accompanied the request, words were more often used. The tendency to act when gestures are used may be instrumental in the long term task of word comprehension.

Other studies have attempted to reveal whether the child relies on one modality more than another in order to decipher messages (Allen, 1991; Allen & Schatz, 1983; Thompson & Massaro, 1986). When presented with conflicting verbal and gestural cues subjects tend to respond in the mode in which the directive was given (Allen & Shatz, 1983). Thompson and Massaro (1986) systematically varied the ambiguity of simultaneous gestural and verbal directives. Preschool age subjects (2.5 - 3.11 yrs.) were unable to use the clear source of information to decode messages. The ability to integrate sensory information improved with age.

IMPAIRED DEVELOPMENT OF VERBAL AND GESTURAL LANGUAGE

Children with mild disabilities and resulting language delays may not be identified until three years of age (Rice & Schuele, 1995). This puts children's development at a disadvantage, because much time is lost when interventions might have been taking place. Children with moderate to severe disabilities, on the other hand, are usually identified closer to birth (Morrison & Polloway, 1995). Language delays or impairments are often expected, in spite of the best efforts of prescribed intervention services. The notable exception is autism, a severe disability which typically has an onset before 36 months of age. Individuals with moderate to severe disabilities have a wide range of different disabling conditions. Each child shows a great deal of variation in his/her manner of communication (e.g., comprehension mode, production mode, contexts of participation). A comprehensive picture of symbolic communication development in the moderate to severe range of functioning is nearly impossible. Instead, specific conditions, such as Down syndrome and autism, are provided as examples of possible effects of moderate to severe disabilities on language development.

DOWN SYNDROME

Language development

Down syndrome is a genetic condition that causes varying degrees of mental retardation. Among children with Down syndrome, verbal and gestural language development is much slower and shows a great deal of individual variation (Franco & Wishart, 1995; Rondal, 1988). Although many cognitive and developmental skills are delayed, language is delayed even more so. Pointing and first words may appear at 20-24 months. Two word combinations could appear at 3-6 years. The two-word combinations of children with Down syndrome are similar to that of normal children at the same stage of language development (Rondal, 1988). Comprehension vocabulary contains more items than receptive language vocabulary. However, compared to receptive language matched peers, children with Down syndrome produce more gestures (Caselli et al., 1998; Mundy, Sigman, Kasari, & Yirmiya, 1988). This means that children with Down syndrome gesture more than expected for their level of language development. As Zinober and Martlew (1985) suggested, perhaps this is so because of the extended delay in verbal production. However, the frequency of nonverbal requests is less than would be expected. Nonverbal requests are correlated with later expressive language scores (Mundy, Kasari, Sigman, & Ruskin, 1995; Mundy et al., 1988).

Social aspects of language development

Children with Down syndrome demonstrated social awareness in the execution of their diectic gestures (Franco & Wishart, 1995). They exhibited the skills of normally developing children in declarative gestures, in that they looked at the partner before pointing. The subjects' mothers were the recipients of more communicative gestures than were unfamiliar peers. Unfamiliar peers, however, received more social referencing looks than did the mothers, which may indicate that the subjects were aware of the greater need to monitor their attention.

Children with Down syndrome do not differ significantly from normal subjects on measures of coordinated joint attention (Kasari, Freeman, Mundy, & Sigman, 1995). However, during joint attention episodes, they spend more time looking at a communicative partner's face than do nonretarded children (Kasari, Mundy, Yirmiya, & Sigman, 1990). Studies on maternal input have not revealed any differences between mothers of normal and impaired children on the proportion of nouns or diectic words (Cardoso-Martins & Mervis, 1990).

Language intervention

Interventions for children with Down syndrome have focused on prelinguistic requests and comments (Warren, Yoder, Gazdag, Kim, & Jones, 1993). Nursery rhymes and parent administered milieu techniques have been used to teach vocabulary (Glenn & Cunningham, 1984). Successful interventions have also used redundant word + gesture combinations (Reich, 1978) and two word combinations modeled during play (Jeffree, Wheldall, & Mittler, 1973). Salmon, Rowan, and Mitchell (1998) contrasted the more didactic milieu techniques with responsive (interactionist) techniques in an alternating treatments design. Milieu produced higher rates of intentional communication, and interactionist techniques yielded a more balanced distribution of comments and requests and discourse functions (e.g., initiations and responses).

AUTISM

Language development

Autism is a condition that causes disordered language and social development. This means that several developmental domains are not commensurate with a child's chronological or mental age, or that development proceeds at an erratic rate (P. Pulic, psychologist, personal communication, January 1998; S. Patterson, psychologist, personal communication, January 1998). Because of diagnostic changes made over the past 50 years, children who receive an autism label do so for a wide range of behaviors (APA, 1994; Cohen, Volkmar, & Paul, 1986). Onset of the disorder may occur after verbal language has begun to develop. They may exhibit echolalic (communicatively nonfunctional) speech or repetitive behaviors. Children with autism may or may not be mentally retarded. Those with lower IQs are likely to be mute, while those with higher IQs are likely to have some functional language (Waterhouse et al., 1996).

Deictic gestures of request are more frequent than those for declarative purposes (e.g., show, point) (Mundy, 1995; Mundy & Sigman, 1989; Wetherby & Prutting, 1984). If pointing does develop, it may be used for requesting, rather than declaring (Goodhart & Baron-Cohen, 1993; Franco, & Wishart, 1995; Mundy, Sigman, Ungerer, & Sherman, 1986). If a declarative pointing gesture develops, functional speech may appear soon afterwards, in accordance with the normal sequence of development (Sugarman, 1983).

Spontaneous use of symbolic gestures is inconsistently noted by researchers and interventionists. For instance, social gestures of greeting are less likely to be offered spontaneously, and less likely to be returned if given the opportunity (Hobson & Lee, 1998). Communication interventions often focus on teaching these gestures as an alternative to speech (see below).

If verbal skills do develop, they may do so in unconventional ways. Words or phrases that are produced do not appear to be functional communication. Whole chucks of language heard from a caregiver, such as a phrase or sentence, may be parroted over and over without regard to their uncommunicative effect (echolalia). It is often difficult to ascertain the idiosyncratic meaning of these utterances. Some repetitions are thought to occur when communication is not understood (Light, Roberts, Dimarco, & Greiner, 1998; S. Merzer, psychologist, personal communication, January 1998). Mitigation of the echolalic response has also been associated with the ability to learn speech (Bebko, 1990). However, some mute autistics have acquired language after long periods of time (Windsor, Doyle, & Siegel, 1994).

Comprehension of verbal or gestural language in children with autism has been used to develop screening measures for autism. Gestural joint attention (a point) could correctly identify 70-80% of the autistic and the language age matched kids at both the initial and follow-up assessments (Mundy, Sigman, & Kasari, 1990). Integration of sensory information may be a characteristic weakness of autism (Pierce, Glad, & Schreibman, 1997). A number of researchers have suggested that cognitive processing problems are minimized by using visual formats (Bebko, 1990). Interventions typically make use of visual presentation to support comprehension (Dijkxhoorn, Berckelaer-Onnes, & Ploeg, 1996; Light et al., 1998; Peterson, Bondy, Vincent, & Finnegan, 1995).

Social aspects of language development

Children with autism are noted for a lack of eye contact (Hobson & Lee, 1998; Willemsen-Swinkels, Buitelaar, Weijnen, & van Engeland, 1998), which is a precursor to normal communication development (Bates, 1979; Bakeman & Adamson, 1984; Sugarman, 1983). Preschool age autistic subjects returned fewer gazes than did developmentally delayed or normal subjects (Willemsen-Swinkels et al., 1998).

Autistic children's joint attention behaviors are worse than those of their developmentally delayed peers matched for language, chronological age, and mental age and language (Mundy, Sigman, Kasari, 1990; Roeyers, Van Oost, & Bothuyne, 1998). They do not display higher levels of affect during joint attention when compared with requesting episodes (Kasari, Sigman, Mundy, & Yirmiya, 1990; Mundy, Kasari, Sigman, 1992), contrary to normal development. In fact, autistic subjects were lacking in affect, regardless of context.

If children with autism do produce diectic gestures, the function of these gestures usually is to request, rather than declare/refer (Goodhart & Baron-Cohen, 1993; Mundy, 1995; Mundy, Sigman, Ungerer, & Sherman, 1986; Mundy, Sigman, & Kasari, 1994; Wetherby & Prutting, 1984). If declarative gestures are used, preschool autistic subjects look at the communication partner after making the gesture (Willemsen-Swinkels et al., 1998), contrary to Down syndrome and normal development. Just as in normal development, where language skills are associated with mental age, autistic deficits in the form and function of prelinguistic communication are also related to level of developmental attainment (Mundy, Sigman, & Kasari, 1994; Willemsen-Swinkels et al., 1998).

Verbal input of mothers to their autistic children follows child focus to the same degree as mothers of normal children (Watson, 1998). However, in a study by McArthur and Adamson (1996), unfamiliar adult partners gave fewer conventional bids for attention to autistic subjects than to language and age matched language delayed subjects. Instead, child attention was redirected using literal bids (e.g., interposing an object in line of gaze). The adults were experienced in language disorders, but did not know child diagnoses.

Joint attention behaviors have also been associated with predictive outcomes on language measures (Mundy, Sigman, Kasari, 1990; Willemsen-Swinkels et al., 1998). Words may also be incorrectly learned because the child is not able to follow the speaker's line of gaze, and uses his/her own line of sight to map word meanings (Baron-Cohen, Baldwin, & Crowson, 1997). The fact that children with autism experience social deficits that seem to hinder language acquisition, and that improvements in social functioning bring improvements in language functioning, is highly suggestive of the necessity of social contributions to language development (Rogers-Warren & Warren, 1984; Sugarman, 1983).

Language intervention

Interventions which involve a communication partner have attempted to stimulate normal development (e.g., facilitate speech or joint attention) or to teach the use of alternative modes of communication (e.g., sign or picture symbols). Although interventions may produce more functional language, deficits in social interactions usually persist throughout the life-span. Below are interventions for facilitating normal communication development.

Facilitating joint attention for children with autism is often difficult, due to autistic behaviors and limited attention spans. Most interventions imply facilitation of joint focus, but vary in the way in which this is accomplished. Behavioristic methods have been successful in teaching language skills (Lovaas, 1997; Harris & Weiss, 1998; Matson, Benavidez, Compton, Paclawskyj, & Baglio, 1996). These methods provide stable (even rigid) contexts where the focus of attention is interventionist supplied. Once speech is taught using behaviorist methods, it is often not generalized to other contexts outside of the intervention setting (Matson et al., 1996).

Another program that specifically targets joint focus uses child focus to build joint attention skills ("Floor Time," Greenspan & Wieder, 1998). Turn-taking skills and the affective side of interactions are given a strong emphasis in this program. Few research studies exist which support child responsive interventions (Hewill, 1998), but there is a growing number of case studies that report improvements in social or language skills (Greenspan & Wieder, 1998; Rollins, Wambacq, Dowell, Mathews, & Reese, 1998).

Milieu techniques are better supported by the research literature. Natural language teaching paradigms lie somewhere between the ultra-structured and ultra-unstructured approaches. Sessions emphasize child focus by utilizing an object that the child is interested in. The therapist plays with the object and verbally models the target language while waiting for the child to respond. Natural conversation is preserved, and child response is optional. The technique has demonstrated improvements over serial trial techniques (Koegel, O'Dell, & Koegel, 1987) and has been effective in reducing disruptive behavior during language intervention (Koegel, Koegel, & Surratt, 1992). Incidental teaching and time delay techniques also improved generalization (Matson et al., 1996).

Few intervention studies have addressed receptive language, a common deficit in children with autism (Matson et al., 1996). However, many interventions use visual means or object attainment to support comprehension. Multi-modal presentation of speech and sign has demonstrated gains over speech alone or sign alone methods (Matson et al., 1996). However, both sign and speech use transient symbolic mediums which employ a temporal sequence. This may present difficulties for autistic children's attention deficits. Peterson et al. (1995) demonstrated the superiority of using a static visual system (pictures) for interventions with autistic children. The picture only conditions had greater gains than speech alone or a combination of visual and speech, and had lower rates of disruptive. Perhaps the positive results of the picture only condition are due to the support they lend to joint attention deficits. Autistic children's difficulties in comprehension and/or lack of integration of sensory information may also be due to joint attention deficits (Pierce, Glad, & Schreibman, 1997). There has been a long-term established emphasis on visual strategies for supporting communication intervention for children with autism (Bebko, 1990; Dijkxhoorn, Berckelaer-Onnes, & Ploeg, 1996; Hodgeson, 1995; Light et al., 1998; Peterson, Bondy, Vincent, & Finnegan, 1995).

Refractive Errors in the Normal Population

Although language development is normally complete by the age of five years (Haslett & Samter, 1997), this is not so for children with moderate to severe disabilities. Therefore, the refractive development of the eye will be reviewed from birth to maturity, in order to provide information on possible visual states that may exist for children who are delayed in their language development. Research demonstrates that refractive errors change over the life span of an individual. However, these studies are complicated by the various methods of assessing refractive errors. Three common types of refractive errors will be presented: (1) Myopia, (2) Hyperopia, and (3) Amblyopia. Astigmatism is not addressed because it often co-occurs with other refractive disorders. Correlations of refractive errors with IQ and achievement scores are discussed along with theories that explain these associations. Interventions for mild vision problems are briefly outlined.

ASSESSMENT

Screening methods for vision problems include the familiar Snellen Chart (e.g., E-chart). The procedure used for this quick assessment and its variants requires that the chart be placed at a distance of 20 feet from the assessed. This procedure will certainly screen for near-sighted (myopic) people, but far-sighted (hyperopic) people will generally pass this test (Getman, 1985; Kirschen, 1954; Manley & Schuldt, 1970). Myopia causes the visual image to be focused in front of the retina, causing difficulty discerning a given target (Skrtic, 1995). This effect is ameliorated by viewing targets at closer distances. Hyperopia causes the visual image to be focused in back of the retina, also causing difficulty discerning the image. This effect is ameliorated by viewing targets at greater distances. Testing of each eye independently can sometimes reveal differences in acuity. Also, eye chart assessments do not reveal any information about how a subject sees at close distances, which is where most school work is done.

A more thorough assessment is done by the optometrist or ophthamologist. These professionals define refractive error with a unit of measure called a diopter (D). Plus numbers (+) indicate hyperopia, while minus (-) numbers indicate myopia. Zero (0) refractive error indicates normal vision. It takes a small measurement to define myopia, usually -.5 D. There is less agreement about defining hyperopia, which ranges from +1.25 D to +3.00 D. Because of the lack of agreement, two researchers could conceivably report different percentages of myopic, emmetropic (normal), and hyperopic participants when reviewing the data from a single sample (Hirsch, 1964).

During assessment the eyes (actually the irises) are usually dilated with a drug. This is referred to as the cycloplegic method. Cycloplegia has been criticized because it can overestimate refractive errors. Non-cyclopeged methods involve relaxing the irises by other means, such as placing the assessed in a dark room. (Proponents of cycloplegia say that "dark refraction methods" under-estimate the amounts of refractive error.) There is generally one (1) diopter (D) of difference between cyclopleged and non-cyclopleged techniques (Howland, 1988). However, in infant populations these methods are not well correlated (Maino & Gerhard, 1984, cited in Kohl & Samek, 1988).

In addition to these disparities of measurement methods, the boundaries that determine the differences between hyperopia, emmetropia (normal vision), and myopia are placed inconsistently by various researchers.

SELECTED FEATURES OF VISUAL DEVELOPMENT

Myopia, Emmetropia, and Hyperopia

1. Birth: The erratic gaze and blinking reflex of newborns make the assessment of their vision difficult (Howland, 1991). Howland's (1991) summary of studies on refractive errors in infants report that infants are generally myopic (near-sighted) when using dark refraction procedures and hyperopic if cycloplegic methods are used. In addition to the unreliable assessment of infants, different genetic endowment may dictate the varied responses to cycloplegia. Newborns also lack the focusing power of the irises, which is called accommodation (Getman, 1985; Howland, 1988). Accommodation is a function of the autonomic nervous system (Howland, 1991; Getman, 1985), and quickly develops to adult levels by six months of life.

2. First Year: Refractive assessments are more reliable as infants age, because gains in head control facilitate refractive techniques (Howland, 1991). The distribution of refractive errors becomes more restricted toward the close of the first year. This may be because of accommodation gains, and/or it may be because of the refractive qualities of the lens, cornea, and axial length of the eye (Howland, 1991).

3. First Five Years: Cyclopleged findings indicate that mean refractive errors tend toward hyperopia over the first five years (Howland, 1991). As children age, they become increasingly hyperopic, and this trend decelerates as they approach five years of age. Noncyclopleged results show no increasing trend toward hyperopia, but do indicate that children are slightly hyperopic (+.75 D). Other cyclopleged results have shown a mean of +1.27 D in children 1, 2, and 3 years old (Dobson, Fulton, Manning, Salem, & Petersen, 1981), and low hyperopia for 0-2 year olds (Kohl & Samek, 1988). The distribution of refractive errors continues to narrow in range over the first five years (Howland, 1991).

4. 5 Years-Adolescence: Most children enter kindergarten with some measurable amount of hyperopia (Hirsch, 1964; Howland, 1991). As children approach puberty, the eyes begin a trend toward myopia (Hirsch, 1964). Children who enter kindergarten with +.5 D to +1.25 D of hyperopia are likely to be emmetropic at adolescence. Children who are emmetropic during kindergarten tend toward myopia as they become adolescents. Children who enter kindergarten with measurable amounts of myopia tend to become more myopic as they age. Those with greater amounts (+3.0 D) of hyperopia will tend to remain hyperopic. (Those that are likely already myopic tend to remain so.) Twelve percent (12%) of students entering grammar school wear corrective lenses, 28% of those students wear corrective lenses upon graduation (Hirsch, 1951-52, as cited in Kolb, 1962).

Amblyopia

Amblyopia is a condition where both eyes do not focus in the same way (Ameder, Peck, & Howland, 1990). Incidence varies greatly depending on the study sample. Estimates range from 1-5% of the school age population (Ameder, Peck, Howland, 1990; Ottar, Scott & Holgado, 1995). There are two subtypes of amblyopia. (1) Anisometropia amblyopia, refers to the inability of the two eyes to accommodate equally on a given target (usually one eye is out-of-focus). (2) Strabismus, refers to a lack of binocular alignment or coordination. One eye may be permanently or intermittently turned inward (esotropia), outward (exotropia), upward (hypertropia), or downward (hypotropia). Amblyopic conditions may also cause the cognitive suppression of visual information from the eye that performs badly. Additionally, gross and fine motor problems are often associated with amblyopia (Slavik, 1982).

Some theories propose that amblyopia begins with a lack of accomodation during early development (Ameder et al., 1990). This can result in a lack of neurological "feedback loops," which are necessary for the reinforcement of binocular focus and alignment. Animal studies have shown that errors in accommodation can cause both kinds of amblyopia (Ameder et al., 1990; Howland, 1988). These errors are usually induced by prolonged covering of one eye. However, this progression has only been inferred inconsistently in human populations (Ameder et al., 1990; Ottar, Scott & Holgado, 1995). It may be just as likely that strabismus causes anesometropia/amblyopia (Ameder et al., 1990; Howland, 1991). Amblyopia is usually detected by the age of five years (Ameder et al., 1990).

TREATMENTS

Prescriptive treatments

Myopic people frequently are given prescriptions for refractive errors as small as -.5 D. For hyperopia, glasses are rarely prescribed, unless the errors are at least +2.0 or +3.0 D (M. Knowlton, Ph.D., class on vision screening, University of Minnesota, October 1998). This is especially true of children, whose eyes are still changing in their refractive characteristics. Because of the adverse effects of hyperopia on reading, Rosner and Rosner (1997) recommended prescription of glasses for +1.25 D or more of hyperopia (see below for a more complete discussion of refractive error and reading scores). The prescription of glasses for anesometropia is regarded by some as a possible prevention of amblyopia (Almeder et al., 1990). Some clinicians discourage the use of corrective lenses for strabismus because eye alignment problems may become correspondingly worse (Carter, 1964).

Amblyopia is often treated by "occlusion," which prescribes covering the better eye, in order to strengthen the worse eye (Mohindra, Jacobson, Zwaan, & Held, 1983). Occlusion is effective in increasing the acuity of the worse eye, but the occluded eye temporarily decreases in acuity. Monoccular occlusion also adversely affects alignment (Herman, Tauber, & Roffwarg, 1974). Strabismus is also effectively treated with surgery (Norcia, Hamer, Jampolsky, & Orel-Bixler, 1995).

Behavioral treatments

Vision Therapy, a highly controversial treatment, has been used to address problems with myopia (Blount, Baer, & Collins, 1984), hyperopia (Leung, Yap, & Lagrow, 1992), amblyopia (Flax, 1993). Getman (1985) explained that eye exercises have been known to help, but former explanations were implausible. Eye exercises do not increase muscle strength, because the eyes are already 100 times stronger than they need to be. Also, because accomodation is controlled by the autonomic nervous system, it is not subject to fatigue. Getman (1985) suggested that changes in acuity have taken place because the mind has been trained to perceive visual targets differently. The exercises taught the subject to use available visual information more efficiently. Behavioral techniques for acuity do not generalize well. In addition to training exercises, vision therapists sometimes prescribe lenses that over or under correct refractive errors in an attempt to bring long-term correction to vision problems.

COROLLARY FINDINGS

Intelligence scores

Students with myopia have been associated with higher intelligence test scores than students with normal or hyperopic vision, even after controlling for age, gender, father's occupation, and/or nonverbal performance IQ (Hirsch & Nadell, 1958; Manley, & Schuldt, 1970; Williams, Sanderson, & Share, 1988). Amblyopia has been associated with lower intelligence scores (Stewart-Brown, Haslum, & Butler, 1985). Some studies reveal differences in favor of myopia on verbal IQ subtests which employ reading, and no differences on nonverbal/nonreading IQ subtests (Williams, Sanderson, & Share, 1988; Young, 1963). Although the evidence strongly supports a relationship between intelligence and acuity status, these studies are possibly confounded by prerequisite reading skills needed to take IQ tests (see next section).

Achievement scores

Varying acuities have been associated with teacher ranked ability groups: high, medium, and low reading ability; and academic vs. vocational educational programs (Grosvenor, 1970). Myopia is found mostly in high and medium ability groups, and hyperopia is found mostly in medium and low ability groups. Myopia is associated with higher ability levels and academic programs. Schwartz (1938) reported that 43% of 1,000 poor readers were hyperopic. Hyperopia has links with lower achievement scores, (Rosner, & Rosner, 1997), particularly reading (Hirsch & Nadell, 1958; Williams, Sanderson, & Share, 1988), even after equating achievement scores for IQ (Stewart-Brown, Haslum, & Butler, 1985.) Because of the adverse effects of hyperopia on reading, Rosner and Rosner (1997) recommend prescription of glasses for +1.25 D or more of hyperopia. However, studies are divided on whether or not corrective lenses cause significant improvements in subjects' reading scores (Stewart-Brown, Haslum, & Butler, 1985; Grisham & Simons, 1986).

THEORIES

Does myopia give a student a reading advantage, and consequently, an IQ advantage (Grosvenor, 1970; Hirsch & Nadell, 1958)? Differences in IQ scores may be due to ability to do near work, particularly for extended periods of time. Uncorrected myopia produced better reading scores than corrected myopia (Grisham & Simons, 1986). This may also lead to another advantage of knowledge previously acquired through reading (Grosvenor, 1970; Young, 1963). These advantages may account for higher scores on intelligence tests by children with myopia.

Does myopia cause a student to become a reader, or does reading cause myopia (Hirsch, 1964)? Genetic predispositions toward myopia could lead to the natural reinforcing of particular experiences and competencies. The myope may prefer reading, which may in turn may further degrade visual acuity, but also strengthen the ability to do near work (Grosvenor, 1970 &1971). Hyperopia may be a hindrance to efficient reading because of the effort required to accommodate (focus) on closer targets. Even moderate amounts of hyperopia can cause physical symptoms during reading, such as rapid fatigue, discomfort, and nausea (Grosvenor, 1971). The same hyperope might pass a screening test for near vision using single letters, because the time on task is not sufficient to bring about any physical symptoms.

REFRACTIVE ERRORS IN DISABILITY POPULATIONS

The research literature has few dissenters on the higher incidence of refractive errors in moderate to severe disability populations. Some specific syndromes are also known to be associated with visual impairments and/or refractive errors. In general, the incidence of refractive errors increases in association with more severe disabilities. Below are some specific disabilities and some studies conducted on populations with moderate-severe disabilities, including Down syndrome and autism.

Assessment

Disability populations have been difficult to assess because many are lacking receptive/expressive language or have behavior problems (Kirschen, 1954). Cooperation and understanding of tester directives are necessary for accurate assessment. Modified Snellen charts may use familiar outlines instead of an "E" in various orientations. Apple, House, Square, or Circle on the chart are matched nonverbally with replicas that the child can point to instead of verbalizing (LH Symbols: Measurement of Visual Acuity for Children, Hyvahyvarinen, 1991). Teller Acuity Cards, another nonverbal means of assessment, uses visual preference to determine acuity (Teller, Morse, Borton, & Regal, 1974). Photorefraction (Howland, 1985; MTI Photoscreener, Ottar, Scott & Holgado, 1995) is an easy, quick, cheap, and nonintrusive screening method, but it underestimates refractive errors compared with dark refraction by about 1 D (Howland, 1982). Some studies resort to inferring acuity by behavioral observation (Castane, Peris, & Sanchez, 1995). Cycloplegic and noncycloplegic assessments are not well correlated at any age in disability populations (Kohl & Samek, 1988).

DISABILITY POPULATIONS

General disabilities populations

Many studies conducted on moderate-severe populations have shown higher incidences of hyperopia (45-59%), myopia (22-24%), and amblyopia/strabismus (28%) (Courtney, 1971; Castane et al., 1995; Kirschen, 1954; Kolb, 1962; Manley, & Schuldt, 1970). Although some of these studies included instiutionalized emotionally disturbed subjects along with moderate to severe disabilities, these too had higher incidences of hyperopia than Hirsch's youngest school population (ages 5-6 yrs.), which would be expected to have higher incidence of hyperopia than normal older populations (Courtney, 1971). In fact, all of the above percentages for types of refractive errors are higher than in Hirsch's study, and higher than would be expected in the normal population (M. Knowlton, instructor, Univ. of MN, class on photorefraction, October 1998).

Specific disability populations.

Contraction of reubella during pregnancy has been associated with increased incidence of severe myopia, cateracts, and glaucoma (Cooper, 1969). Fifty percent (50%) of children with cerebral palsy, 25-33% of children with hydrocephally, and 80% of children with myelodysplasia are estimated to have vision problems (Jackson & Vessey, 1996). Down syndrome populations have a high frequency of myopia (23%), hyperopia (22%), amblyopia/strabismus (57%), and cataracts (11%) (Caputo, Wagner, Reynolds, Guo, & Goel, 1989; Courtney, 1971).

Studies conducted on subjects with autism do not seem to indicate problems with acuity (Janicki, Lubin, & Friedman, 1983). However, this may be an artifact of a lack of comprehensive screening, due to the fact that these children are difficult to assess. Children with autism exhibit behavior which is indicative of visual problems, such as lack of visual responsiveness, squinting, scrutiny of detail, use of peripheral vision, head tilting, reliance on near space, lack of fixation on targets in far space, and posture and motor coordination problems (Bourgeois, 1971; Kaplan, Carmody, Gaydos, 1996; Kohen-Raz, Volkmar, & Cohen, 1992; Schulman, 1994). Use of photo-refraction may help to give a clearer picture of acuity for children with autism. In some cases, Vision Therapy has been used to modify the visual behavior of affected children (Kaplan et al., 1996; Rose & Torgerson, 1994; Schulman, 1994).

Treatments

Treatment of refractive errors with corrective lenses is a logical way to improve visual functioning. Glasses are usually prescribed to be optimal at 20 feet, which may compromise classroom functioning (e.g., near work) (M. Knowlton, Ph.D., University of Minnesota, class on photorefraction, October 1998). However, because much of the optometrists/ophthamolgists procedures require extensive cooperation, it is possible that clients with low verbal and/or behavioral skills will not be accurately assessed. It may be that the higher functioning a subject is, the more likely s/he will be prescribed an appropriate set of corrective lenses. If the client cannot give feedback on lack of improvements that glasses have made, s/he must depend on teachers or parents to infer acuity from behavioral cues. Furthermore, errors of correction may result if the client does not wear the glasses properly, or if the frames become bent (S. Endris, itinerant vision teacher, December 10, 1998). A simple way to test the prescription of corrective lenses is to use photorefraction. Appropriate prescriptions will cause the eyes to appear emmetropic. Often refractive errors go uncorrected because the child refuses to wear glasses (Sally Endris, itinerant vision teacher, personal communication). Other times refractive errors are undetected because a cognitive deficit masks visual problems (Sally Endris, vision itinerant teacher, personal communication). For instance, a hyperopic student may just look like s/he is not able to pay attention, when they are uncomfortable focusing on targets which are too close.

Theories

Why is there a higher incidence of refractive errors in disabled populations? Are IQ or achievement related to refractive errors? It could be reasoned that the correlations of IQ/achievement and refractive error were attenuated on the lower end of ability because those students were not included in the studies. If the average and gifted populations were combined with the subaverage intellectual populations, it seems logical to think that stronger correlations would be found for hyperopia and lack of intelligence or achievement. Even among the moderate-severe populations, it has been stated that more severe visual problems were found in subjects with lower IQs (Courtney, 1971; Kirschen, 1954). Severe visual impairments can restrict perceptual information which is important for development, especially in populations with severe disabilities (Ferrell & Raver, 1991; Hatton, Bailey, Burchinal, & Ferrell, 1997). Perhaps this is also true of refractive errors in disability populations.

Frequently mentioned in these studies is the history of birth injuries as a possible explanation for refractive errors (Courtney, 1971; Kirschen, 1954; Manley & Schuldt, 1970). Trauma to the brain and/or nervous system would affect the eyes as a part of that system. Both the brain and the eyes are more highly developed than other body parts, and have less growth to accomplish after gestation (Courtney, 1971).

Additionally, refractive errors could be caused by a lack of physiological development. Many developmental disabilities are accompanied by a lack of growth, which may also affect the brain and/or the eyes (Courtney, 1971; Hirsch & Nadell, 1958). Retarded development of the eyes could explain the higher incidences of hyperopia, which are also found in young children. Also, reports of "exaggerated" responses to cycloplegia in disability populations is reminiscent of early development (Manley & Schuldt, 1970).

CONTRIBUTIONS OF VISION TO ACQUISITION TO ACQUISITION OF VERBAL AND GESTURAL LANGUAGE

Other than literature on development of blind children, little research can be cited that directly addresses the effects of vision on language development. In the following section, research on blind, delayed, and normal development are brought together to suggest possible relationships between vision and language development. Additional comments are made about social factors which facilitate language.

Selected Language Features

Joint attention

Implicitly or directly stated, facilitation of joint attention is a necessary component of language intervention. Joint attention measures have been associated with language measures (Mundy, Sigman, & Kasari, 1990). Some interventions may focus on the mere sharing of affect (Greenspan & Wieder, 1998), because positive affective experiences are associated with joint reference, rather than requests (Kasari, Sigman, Mundy, & Yirmiya, 1990). Requests are easily engineered by structured milieu techniques. But more unstructured responsive interaction interventions may be more likely to stimulate shared positive affect through encouragement of declarative functions (Salmon, Rowan, & Mitchell, 1998).

Even though other cultures demonstrate that prelinguistic diectic gestures are not necessary for language development, it may be advantageous for preschoolers with disabilities to be taught these skills. Because verbal language is often severely delayed in preschoolers with disabilities, it is beneficial for them to learn other methods of effective communication. Until symbolic communication develops, diectic gestures may nurture motivation for communication. However, these interventions are rarely mentioned (brief outline in Dijkxhoorn et al., 1996; one line in Rossetti, 1991; Warren et al., 1993).

Perhaps it is possible that the experience of affective sharing is "dampened" by the a mismatch between interaction distance and refractive error. Hyperopic students may not enjoy close encounters because it is too difficult to focus on the communication partners face. Myopic students may miss out on affective encounters because interaction distances are too great. Mismatches may produce less positive affect and lead to lower motivation for communication.

Diectic gestures

Visual requirements are inherently involved in the comprehension of gestures. They must be seen to be communicated. Congenitally blind children do not acquire diectic gestures without direct instruction (Rowland, 1984). Is it possible that refractive errors in developmentally delayed children could reduce the saliency of visual information contained in gestures? At a distance, myopes might not interpret a pointing gesture differently from a reach. If diectic gestures are not clearly seen and comprehended, they may not be imitated and produced. The presence of amblyopia/strabismus may also interfere with the production of gestures because of associated fine and gross motor problems (Slavik, 1982). Additionally, is it possible that reduced saliency of environmental targets could hinder the development of communication via diectic gestures? If the diectic referent is not clearly seen, the purpose of the gesture may be misunderstood.

Symbolic gestures

Symbolic gestures may be spontaneously produced or taught to the child (Acredolo & Goodwyn, 1988). If they are spontaneously produced, they are likely to be based on motor movements that are used with referents. Research with blind children reveals that they make motor gestures associated with favorite toys at around 5 months (Fraiberg, 1977). Although these are not symbolic, they have the potential of becoming so through reinforcement. If these motor movements could be identified in normal or delayed development, responsive intervention might maintain them through reinforcement until they eventually became referential symbols.

If gestures are taught directly to the child, the motor movements could be modeled or directly manipulated (Reich, 1978). Hand over hand produced motor movements may prove especially salient to the hyperopic child. In this way the communication interventionist might overcome the problems of being too close to be seen by the child, or too far away to be salient. Hand over hand may also aid the child with amblyopia/strabismus in production of gestures, which may require fine motor skills.

Words

Studies on children who are legally blind, who are without mental impairments, indicate that they develop more slowly on language and social measures in the first years of life than partially sighted or normally developing children (Ferrell & Raver, 1991; Hatton, Bailey, Burchinal, & Ferrell, 1997). Language and developmental domain gains are directly proportional to the amount of useable vision. In cases of concurrent cognitive or developmental deficits, gains are more depressed. Is it possible that mild visual impairments (e.g., refractive errors) could also depress the developmental gains of preschoolers who have moderate to severe disabilities?

Phonetic production of verbal language may also utilize the visual field. In normal development, speech sounds which are more visible (e.g., "m," "b," & "f") make up 33% of a child's early verbal vocabulary (Vihman, in press, cited in Goodwyn & Acredolo, 1993). Children who are congenitally blind produce these sounds in significantly smaller amounts, and blind children make more phonetic errors of substitution (Mills, 1983). Refractive error could obscure these visual models of speech production. Concomitant disability and refractive error may create barriers to phonetic imitation.

Comprehension

Comprehension of early verbal language is normally very dependent upon the visual field. If a child cannot clearly see what is being referred to, comprehension, acquisition, and production may be inhibited. Refractive errors could "dampen" environmental stimuli and hinder word acquisition via the meaning mapping processes. Congenitally blind children eventually use other sensory abilities to form understandings about the meanings of words. However, words are slow to generalize to other contexts in blind children's language (Dunlea, 1989). The blind child's first words are more context bound than normal children's words and, thus, are lacking in symbolic power. Most of blind children's word extensions (generalizations to other contexts) are made on the basis of tactual similarities (Dunlea, 1989). The tactile field of perception is not as instantly available to an infant as the visual field. It takes greater activity to explore the fewer items that are available for tactile exploration. Because sight is the most efficient mode for processing environmental stimuli, it is beneficial for children with disabilities and refractive errors to make the best use of their vision. Interventionists who are sensitive to the effects of children's refractive errors might arrange instruction and the classroom environment to enhance visual function.

Because children with developmental disabilities are slow in their physical maturation, they may be slow to integrate information from various sensory channels, due to the presumed retarded neurological maturation. Interventionists might need to consider the assumptions made when communicating. Prompts and responses might need to be established in one mode (e.g., gestural or verbal) before presuming that cross-modal responses would take place, whether or not simultaneous input from another mode is given. This also implies that the interventionist would model the mode that the student is to respond in, which may facilitate acquisition. Additionally, use of redundant gesture+word combinations by the interventionist may also facilitate comprehension for students with cognitive disabilities, provided the refractive errors and distances are compatible.

LANGUAGE RELATED FEATURES

Reading

Reading has implications for the development of language. For this reason, speech-language pathologists are now considering the various types of language that are involved with academic work (Butler, 1995). Although not every child with a disability will learn to read, they are usually expected to engage in activities which involve books, worksheets, or computers at near distances. Additionally, the emphasis on mainstreaming will bring children with disabilities into regular education classrooms where close seat work is the norm. Books, worksheets, and computers become the referents to which language refers. If children with disabilities are to be fully integrated, and to get the most (language learning) out of their mainstream experiences, the referents should be discernable. Corrective lenses may aid significantly in the quality of these encounters. However, the 20 feet standard distance correction may not produce positive effects for near work (approximately 2 feet). If glasses are not used, the teacher could be sensitive to the child's visual needs by restructuring an activity. For instance, a child may not engage with books which are presented at close distances, but may do so when stories are presented on a felt board at a greater distance.

Social Aspects

Social development may also develop more slowly for children with refractive errors and concomitant disabilities, because pragmatic cues may be obscured by refractive errors (e.g., facial expressions or environmental cues). Children who are congenitally blind socialize one-fourth as much as their sighted peers, and are retarded on measures of social development (Dote-Kwan & Chen, 1995). Children with concomitant blindness and cognitive disability are delayed socially even more (Ferrell & Raver, 1991; Hatton et al., 1997). Is it possible that refractive errors may additionally limit the social development of children with moderate to severe disabilities, in a similar manner? If a child is unresponsive to pragmatic cues, communication partner input may change (McArthur & Adamson, 1996). While this may serve an adaptive function, it can also serve to limit the richness of the social/language model, reducing the likelihood of exposure and production of more mature communication (McArthur & Adamson, 1996).

Social factors interact in complex ways throughout development (Hewitt, 1998; Sameroff & Chandler, 1975). A child that gives unreadable communicative signals to potential partners may initiate a cycle that sustains that breach. Caregivers may be less motivated to communicate with the child, and so fewer communication opportunities are utilized. If intervention is attempted, restrictive methods designed to produce desirable behavior or eliminate undesirable ones may alienate the child. Instead of learning the joys of communication, they are actively avoiding it. Similarly, a child that is motivated by attention may initiate a cycle of chasing their communicative partners. Interactions with peers may be less manipulable than those with adult interventionists. Because of the important of social competence with peers, corrective lenses may be the most facile vision enhancing ploy. Corrective lenses are not usually recommended at young ages for hyperopia, but are more commonly recommended for myopia. Depending on the severity of refractive errors, this may not pose a significant barrier. Mildly hyperopic children will see the social environment well in group settings. Peer attitudes toward children with disabilities can be positively affected through interventions, but may take significant teacher time in monitoring to remain effective (Lerner, Lowenthal, & Egan, 1998).

INTERVENTIONS

Early Childhood Special Education (ESCE). Intervention in the ECSE classroom is greatly concerned with the development of language. The techniques used to remediate language deficits are designed from theories of social interaction (Lerner et al., 1998). Classrooms are designed and programmed to facilitate high levels of language and social interactions between peers. Classrooms that use best practices more frequently also demonstrate greater language gains among the children that they serve (Schwartz, Carta, & Grant, 1996). However, a restricted range of language intervention styles were used for students with moderate-severe disabilities (Schwartz et al., 1996). Most frequently, the students with moderate to severe disabilities were the focus of language intervention, instead of other objects, events, or people. Interaction usually took place in 1:1 settings, rather than in small or large groups. Taken with information on refractive error, this means that preschoolers with disabilities, who are likely to be hyperopic, are likely to be interacted with at close distances. That the interaction is probably at close range, may partially compensate for this mismatch by increasing the saliency of the interventionist (e.g., the interventionist will take up more of the field of view).

It should not be ignored that children are also capable of compensating for their lack of visual ability. Even children with severe disabilities may make good inferential use of remaining vision. Students with severe hyperopia may rely more on touch and auditory cues when an interventionist is interacting at close range with them. They may even produce the behaviors that the interventionist is looking for. However, the issue here is to optimize language interventions by making the best use of the learner's visual abilities, as the visual sensory mode is the most efficient for quickly processing visual information. If optimal use can be made of vision in the language intervention context, it seems that this would enhance interventions. Perhaps the removal of visual cognitive-processing barriers ought to liberate cognitive space for language learning.

Assessments for ECSE are primarily concerned with the developmental age of the child. Vision screening is employed in ways that do not involve distance acuity (D. Ornelas, school nurse, personal communication, December 1, 1998). Even if hyperopia was identified at preschool age, it is not likely that corrective lenses would be prescribed, as normal children's eyes are still developing toward emmetropia. This places greater burden upon the teacher in the ECSE classroom to use environmental accommodations for children with refractive errors.

Developmental disabilities

Down syndrome interventions have been sensitive to the connection between the disorder and the need for corrective lenses. Other children with developmental disabilities are not necessarily as likely to receive corrective lenses. Often, cognitive disabilities mask visual problems (S. Endris, itinerant vision teacher, personal communication, December 10, 1998). Whether or not glasses are prescribed, intervention can take information about refractive errors into account. Interventions for children with cognitive deficits often relies on making important information salient to the student. Optimizing intervention for students with hyperopia might emphasize communication at a distance. Participation in large motor activities contingent upon verbal or gestural communication might include being thrown a large ball, running to a distant target, or jumping into a pool. Other language targets, such as locatives, are easily woven into motor activities. If near space had to be used with a hyperopic child, utilizing visual information of a broad nature could be used (as opposed to detailed information). For instance, large crayons and bold outlines could be used for coloring worksheets in an activity where children are to verbalize what they are coloring and which colors they are using. Optimizing the intervention for a child with myopia might emphasize use of near space, which is very natural in 1:1 exchanges. Using interaction contexts at greater distances might involve use of a social interaction song. When the child who is "it" calls out the name of the next child, all the children can repeat the name and point (redundant word + gesture) to the next person. When the teacher works with a child with amblyopia, additional considerations should be taken to favor the better eye during interactions, especially if the teacher sits next to the student. Some children with amblyopia use one eye for near distances and the other for far distances (M. Knowlton, Ph.D., Univ. of MN, class on photorefraction, October 1998).

Autism

Autism intervention often relies heavily on visual mode to communicate important information. Although there is no current research which links autism and refractive errors, visual deficits should be ruled out in order to ensure optimal intervention effects. It may be that cognitive factors prevent children with autism from functional use of their vision (e.g., attention to stimuli, processing of stimuli, cognitive implications of stimuli). Any additional refractive errors can only exacerbate learning problems through the visual mode. Optimizing interventions for children with hyperopia might mean use of color coding, bold lines, or texture schedule icons. (Visual schedules are often used for autistic preschoolers to establish child expectations about classroom activities.) Social Stories, a popular book technique for establishing expectations of child behavior, might need to transferred to the felt board. Optimizing interventions for children with myopia might involve large wave gestures when greeting the children from a distance (word+gesture). If a video monitor is to be watched from some distance, the room could be darkened, to make the image more salient.

SUMMARY

The saliency of the visual information connected with gestural or verbal referents is somewhat dependent upon the refractive state of the child's eye. Language or social development is dependent upon the processing of relevant symbols and social or environmental cues. Children who have cognitive or social impairments and, thus, are prevented from acquiring presymbolic or symbolic communication may benefit from more efficient use of available visual information. Disability populations have a higher incidence of refractive errors than that found in normal populations. Language remediation, which is often done at close interaction distances, may be too close for hyperopic students. Those students with myopia may not be able to discern distant environmental features which are critical to language learning. With respect to the specific needs of children with moderate to severe disabilities, we do not know the degree to which refractive errors exacerbate their developmental delays. A variety of interventions exist and may be tailored to their needs. These may take the form of corrective lenses, environmental arrangement, and/or teacher engineered interactions. However, these interventions have not been studied sufficiently with developmentally disabled children who also have refractive errors. Clearly, much work remains to be done.

Based on this review of the literature, this study was designed to examine several relationships between language and visual status in young children with and without disabilities. Participants with disabilities of the chronological ages of 3-5 years were compared with a sample of children without disabilities who had the similar cognitive levels and chronological ages of 1½ - 2½ years. The study planned to measure participants' (1) ability to learn novel words (language mapping), (2) frequency of visual targets (e.g., looking at the speaker's face and/or direction of gaze), and (3) visual acuity. Furthermore, analysis was planned for participants' relationships between (4) acuity and looking, (5) looking and language mapping, and (6) acuity and language mapping. Research for each of these six topics are described below.

1. Would participants with disabilities differ from typical children of comparable cognitive ability in terms of their ability to learn novel words? Would participants with specific disabilities, such as Down syndrome or Autism, have greater difficulty with novel word learning? Are higher developmental levels associated with higher accuracy in novel word learning? When participants are accurate in learning novel words, are those with higher developmental levels more likely to integrate vocalizations in their nonverbal responses?

2. Would participants with disabilities differ from their "cognitive comparisons" without disabilities in terms of their frequency in looking at visual targets? Would different visual targets (e.g., the speaker's face or the speaker's direction of gaze) be associated with different profiles of looking for participants with and without disabilities? Would participants with specific disabilities, such as Down syndrome or Autism, produce differences in the amount and types of looking behavior? Are higher developmental levels associated with higher frequencies of looking? Do specific interventions, such as shifting eye gaze, head turns, or use of hand gestures, facilitate looking frequency? Is the intervention effectiveness associated with participant characteristics, such as cognitive ability or documented disability?

3. Would participants with disabilities differ from their cognitive comparisons without disabilities in terms of their visual acuity? Would participants with specific disabilities, such as Down syndrome or Autism, have different visual acuity profiles when compared with participants without disabilities? Do specific interventions, such as shifting eye gaze, head turns, or use of hand gestures, facilitate language learning? Is the intervention effectiveness associated with participant characteristics, such as cognitive ability or documented disability?

4. Would participants with disabilities differ from their cognitive comparisons without disabilities in terms of the relationships between visual acuity and visual target frequency or type? Would participants with specific disabilities have different relationships when compared with participants without disabilities?

5. Would participants with disabilities differ from their cognitive comparisons without disabilities in terms of the relationships between visual targets and language mapping? Would participants with specific disabilities have different relationships when compared with participants without disabilities?

6. Would participants with disabilities differ from their cognitive comparisons without disabilities in terms of the relationships between visual acuity and language mapping?

The answers to these questions may yield information which is useful for language intervention design for children with disabilities. Language intervention for these children may assume certain language sub-skills (e.g., visual acuity or language learning strategies) which may be similar, more frequent, or deficient when compared to children without disabilities. If the specific information alluded to in the above questions is known, language intervention might become more accurate and effective for children with disabilities.

CONTENTS PAGE

CHAPTER 3
METHODS

This study examined the visual acuity and looking behavior of participants with and without disabilities in the context of a fast-mapping language exercise (novel word learning). A modified version of Baron-Cohen et al.'s (1997) warm-up procedures with familiar toys were used to support inferences of the children's demonstration of their knowledge of novel words. Children were asked to identify familiar toys in order to validate the use of a similar procedure to test whether or not fast-mapping had occurred. Baldwin's (1993) procedures for novel word learning were used to compare the effects of visual targets (e.g., speaker and speaker's direction of gaze). Under naturalistic-like conditions, the researcher uttered a novel word five times while looking at a novel object (held by the child or the researcher). Additional intervention trials which used eye gaze shifting, head turns, and hand gestures were administered to facilitate performance on looking targets and fast-mapping for participants who did not successfully map novel labels. To determine whether novel word learning had occurred, participants played a "find-it" game in which the novel objects were hid temporarily under a towel. Finally, visual acuity was assessed with a photo-refraction camera to determine the presence and degree of vision problems.

PARTICIPANTS

Participants were chosen from Early Childhood Special Education Programs in the Twin Cities Metro and Suburban areas. Seven school districts agreed to participate in the study, and teachers within these districts were asked to nominate children 3-5 years of age who were functioning developmentally at half or less of their chronological age. A control group was drawn from typical children from a church nursery program who were approximately 1½ - 2½ months of age.

After teachers nominated participants for the study, consent forms (see Appendix A) were sent home to parents explaining the study and inviting them to allow their children to participate. For participation in the study, parents would receive the results of their children vision screening, which could aid in referrals for optical or medical services. If parents agreed, the vision screening results would be placed in the child's educational file.

The number of forms that were sent home with parents (ECSE programs) or mailed (church children's program) to parents was 146, 106 in the public schools and 40 in the church children's program. Sixty-nine (69) forms were returned, 42 from public school student guardians and 27 from the church children's guardians. Sixty (60) students were actually participated in the study, 38 from the schools and 22 from the church. The typically developing children attended a suburban church program (n = 22). The seven school districts were from both urban (2 districts) and suburban (5 districts) areas. The total of public school children from urban schools was 22 and the total of public school children from suburban schools was 16. The total for urban study participants was 22 and the total for suburban participants was 38.

MATERIALS

Six familiar toys were presented in three pairs: (1) car-shoe, (2) baby-boat, (3) ball-phone. Objects within pairs were of similar size. Eight (8) novel toys were created to be at least 5 inches long and brightly colored to ensure that far-sighted participants could see them easily. The toys were hand made from hardware store parts (knobs, colored string, rings, etc.) to ensure that the subjects would have never seen the toys before. The novel toys had simple, one-step functions similar to Baldwin's toys. A list of the supporting equipment used in this study is as follows: card table (.91 meters X .91 meters), high chair, white tablecloth, white towel (for find-it game), large popcorn tin for familiar toys, camcorder, MTI photo-refraction camera (Ottar, Scott & Holgado, 1995), and large sheets of paper to block out light from windows (use of the photo-refraction camera requires a darkened room).

PROCEDURE

Data collection took place at the educational or church sites. In the schools, children were taken to another room within the school building when school was in session (e.g., occupational therapy room, empty classroom). In the church, children were taken to an empty nursery room when children's programs were in session.

Upon entering the testing room, participants were shown a brightly colored popcorn tin on a square folding table which was draped with a white table cloth. The children's interest in the tin was piqued through various means: (1) drumming on it, (2) asking them to pick it up, (3) exclaiming about the characteristics of the tin, etc. When the children were sufficiently interested, they were seated in a high chair in front of the table. The tin was then emptied of its contents, the familiar toys. The child was allowed to investigate the items, while the researcher commented on the items, readied the camcorder, and positioned herself across the card table from the child.

To determine the validity of the novel word testing procedures, a modified version of Baron-Cohen et al.'s (1997) "warm-up" activity was conducted with the participant. If the participant could correctly identify familiar toys in response to a verbal prompt, it was reasoned that they might be able to do the same for a novel toy, whose label had just been learned. The warm-up activity was conducted using the familiar toys (see materials). Three pairs of toys were presented serially to the participant. While the first pair was hidden under a towel, a song was sung ("Gonna Find It," see Appendix F). When the towel was removed, the researcher asked, "Where's the [familiar toy]?" If children reached for, picked up, touched, or pointed to the correct toy, the researcher cheered the child's response. If no response was made, the researcher modeled the correct response by holding up the correct toy and cheering the child (as if s/he had made the response). During videotape coding, only child-initiated correct responses were credited as correct. Sometimes the child added a verbal or vocal remark to his/her gestures or manual choices. These were also noted during the coding of the video tapes to answer questions about integration of vocal and gestural modes.

Phase One of the fast-mapping exercises were similar to Baldwin's (1993) procedures. Each child was presented with two novel toys, one of which became the child's and the other became the researcher's. In one condition, the researcher would utter a novel word in verbal context ("It's a toma.") while looking at the child's toy (follow-in condition). In the other condition, the researcher would utter the novel word while looking at the researcher's toy (discrepant condition). The novel word would be said when the child's attention was directed to the child's toy, and the researcher would continue to look at the toy of focus for four seconds.

After five utterances using the novel word, the researcher would take both toys and place one toy on the right side of the table and the other on the left. To determine if novel word learning had occurred, the researcher would ask the child to identify the novel object of focus (e.g., "Where's the toma?"). The researcher would look at the child's face only. Child replies were acknowledged in a neutral manner ("Oh - O.K."). After the child answered (or did not answer) s/he was asked again to identify the novel object ("Let's do that again."). If the child answered correctly both times, the researcher would play the find-it game (similar to the warm up activity) with the child (described below).

If the child answered incorrectly or ambiguously, the researcher would perform an eye gaze shifting intervention. This was done by first causing the child to look at the researcher's face (say the child's name and/or make gestures). Then the researcher would shift her eye gaze to the correct object and say, "Look at the toma!" Eye gaze shifting intervention was done twice. Then the child was again asked twice to identify the location of the novel toy.

Next in the protocol was the find-it game, which was similar to the warm-up activity. The two objects were placed under a towel in reverse orientation and the song "Gonna Find It" was sung. When the towel was removed, the child was asked to indicate where the novel object was ("Where's the toma?"). The game was repeated again after the researcher had switched object positions again. Toys were held beyond the reach of the child during the interventions and novel language prompts All participants received the follow-in and discrepant conditions. To aid in the management of children's attention, a song was sung to introduce the new toys for each condition ("New Toys Comin' Up," see Appendix F).

Phase Two tests were additional conditions that were administered to children who were not successful in mapping the novel words to the novel objects. These interventions consisted of a head turn condition and a show gesture condition. Like the eye gaze shifting cues in Phase One, they were intended to increase looking at target toys, and perhaps facilitate language mapping.

As at the start of other conditions, the song "New Toys Comin' Up" was sung to introduce another pair of novel toys. After the child was allowed to play with the toys, they were placed in the right-left orientation on the table and the participant was given two trials of eye gaze shifting intervention. Then the child was asked twice to identify the novel object. If the subject was correct, the language test was administered ("find it" game). If the child was incorrect, the child was given two trials of novel toy labeling intervention using a head turn or a hand gesture ("show"). The head turn intervention was included eye gaze shifting. The show gesture included a head turn and eye gaze, thus these language cues varied in salience from eye gaze (least salient), to head turn, to hand gesture (most salient). After the head turn or show gesture intervention, the child was again asked twice to identify the novel object. Then the find it game was played twice, reversing the object positions both times. A flow chart of language learning and testing procedures can be found in Appendix D.

After the language exercises, the child's vision was assessed with the photo-refraction camera to determine the presence and degree of refractive errors. Other types of vision problems, such as binocular coordination problems were also noted in the data. Photographing of the participant after language activities prevented after-images from the flash from interfering with the child's vision during the fast-mapping exercise.

During the warm-up activity the order of pair presentation of the familiar toys and the requested toy position (right or left) were counterbalanced for the disabilities and typical groups. For the experimental conditions (follow-in, discrepant, head turn, and show gesture) novel toys, novel labels, toy position were counterbalanced for the groups. Cue cards were developed to aid the researcher in the proper execution of counterbalancing (see Appendix E).

CODING AND INTERRATER RELIABILITY

Coding - Once the camcorder was recording and before the language tasks took place, "calibration" of the child's eye gaze was performed. This permitted the inference of child eye gaze targets. Participant eye gaze "calibration" was performed by directing the child through word or hand movements to look at specific places: (1) the researcher's face, (2) the researcher's play area on the table, (3) the right side of the table, (4) the left side of the table, and (5) the child's own play area on the table. The placement of the camcorder, the high chair, and the size of the table all created standard distances and angles which aided in the coding of the videotapes. Calibration helped to sort out any coding difficulties due to ocular coordination problems, where it was possible that only one eye would be focused on any coded targets.

Videotapes were coded for data during the warm-up activity, Phase One, and Phase Two. The warm-up activity data consisted of (1) correctness of responses to the six prompts to identify warm-up activity, (2) number of correct responses, (3) number of times vocalization occurred simultaneously with the correct responses, and (4) percentage of vocalization with respect to correct warm-up responses. Phase One consisted of (1) number of times the child looked in response to the five novel utterances (looks to researcher's face and sequential looks to researcher's face to direction of researcher's eye gaze), (2) percentage of the times that the child looked to these targets, (3) child choice in response to prompts to identify novel objects, (4) child responses to eye gaze shift, head turn, and show gesture interventions, and (5) responses to the find-it games.

A five point ordinal scale was used to score children's responses to identify the novel objects, the cues to look (eye gaze shift, head turn, show gesture), and the find-it games. Each of these prompts were presented twice. If the child looked both times, or chose correctly both times, the child received a score of two (+2). If s/he looked or chose once, but on the other prompt chose nothing,