Abstract
Because children with severe forms
of autism suffer attentional lability and are occasionally aggressive
or self-injurious, and because they do not naturally engage in
deictic pointing, they cannot cooperaye in standardized language
assessment protocols designed for normal children. Consequently,
such children may be rarely evaluated formally, and when evaluated,
their formal assessments may seriously underestimate their actual
linguistic competence. In this project, the receptive skills
of nonspeaking institutionalized autistic children in New York
State were assessed in a comparison of: 1) standardized testing
on the Peabody Picture Vocabulary Test-Revised, conducted by a
licensed speech/language pathologist, and 2) testing based on
activities developed within a social context. Children assessed
with the PPVT-R performed at chance level, failing to obtain basal
scores. The same children, given the opportunity to interact
with researchers in social activities designed to focus their
attention were more accurate at identifying common objects and
showed an understanding of simple abstract concepts. Results
indicate that standardized assessments for receptive language
do underestimate linguistic competence in nonspeaking autistic
children, and that alternate assessment protocols need to be developed.
Problems in Standardized Language
Assessment of Children with Severe Autism
Currently, there is quite a mixed view of the language abilities of nonspeaking persons with autism. The view that has prevailed for decades is that a lack of speech and language is associated with severe mental retardation (Rutter, 1978; 1983). This view has been completely contradicted by recent claims that through a technique called facilitated communication (FC), in which a facilitator supports the wrist of a nonspeaking person as he types words, nonspeaking persons with autism, formerly considered retarded, actually have normal language and cognition (Biklen, D. 1990; Crossley, 1988). Controlled tests of FC (cf. Jacobson, Mullick & Schwartz, 1995) have refuted such miraculous claims, and now the pendulum has swung back too far to the position that this autistic population has no language skills at all (Vazquez, 1995). The position we take here is in basic agreement with the traditional view that the absence of speech or any other productive form of communication (e.g. sign language) is indeed linked with profound retardation. We dispute, however, the notion that nonspeaking persons with autism are completely devoid of linguistic competence. We believe that the receptive skills of severely autistic children need to be examined much more carefully.
The place to start is to consider very early language acquisition in normal children. It is well-established that comprehension precedes production (Bloom, 1973;1974;
Huttenlocher, 1974), which is to say that a lack of productive speech does not necessarily imply a lack of comprehension. Before they utter single words, normal children indicate by direct gazing and by deictic pointing (two things that severely autistic children do not do) that they understand words that are said to them (Bruner, 1974/1975; Bates, 1976). Further, the development of reference (the concept that a given word applies to a specific object or referent) is initially contextually-bound and often tied to routines (Bates, 1979; Bruner, 1974/1975; Lock, 1980; McShane; 1980; Nelson, 1985; Reich, 1976). In other words, a child's early understanding of the word "slippers" may be limited to his mother's slippers, and only when they are near her bed, or only when she is putting them on; if the very same pair of slippers were left in another room, they might not be recognized (looked at or pointed at, in response to hearing the word). Later, children appear to realize that a word applies to a category of referents, and not just to a particular exemplar (Greundel, 1977; Nelson, 1985; Oviatt, 1982). At this point, children may overextend concepts, and every man becomes "Daddy". Eventually, normal children develop referential concepts that coincide with those of adults, but this takes years. Thus, the acquisition of referential skills is a complex process of concept formation that begins with the realization that things have names, and which continues with a developmentally-related pattern of using features of an object to determine its inclusion or exclusion from a particular category or label.
The emphasis on reference or naming skills, however, obscures the literature on the effects of context on the use and comprehension of language (Bates, 1976; Bruner, 1975; Nelson, 1981). Lund & Duchan (1988) regard context as fundamental to normal language use and comprehension, and they argue that assessment cannot be carried out in isolation of such considerations. In other words, assessment should be more naturalistic. If more naturalistic testing is recommended for normal children, shouldn't such considerations be of at least equal importance when assessing disabled children?
We question the common practice of using standardized assessments such as the Peabody Picture Vocabulary Test-Revised (PPVT-R, Dunn & Dunn, 1981) to gauge receptive language skills of severely autistic children. While such measures may indeed provide an acceptable assessment for normal children, they may be contraindicated for use with an autistic population for several reasons. To begin with, the validity of any assessment is dependent upon the availability of normative reference data for the population being tested; there are no such data for severely autistic persons (Dunn & Dunn, 1981). Further, tests such as the PPVT-R and the Receptive One-Word Picture Vocabulary Test (Gardner, 1985) provide measures of referential functioning only, ignoring a broader spectrum of receptive skills; recognizing names of individual objects is not the only way children normally demonstrate understanding. More importantly, standardized assessment protocols which restrict re-testing, and which require children to point at the correct picture in a series of black and white line drawings, make no accommodation for the attentional lability and the lack of deictic or proto-declarative pointing that are pathognomonic of severe autism (Baron-Cohen, 1989; Curcio, 1978; Ricks & Wing, 1975; Sigman, Mundy, Sherman, & Ungerer, 1986). Thus, it is possible that standardized assessments, which are commonly included in official school records, and which are frequently reported in professional journals, may seriously underestimate true receptive competence of severely autistic children. We suggest that more naturalistic testing, designed to focus attention and avoid pointing problems, may result in more accurate assessment of receptive skills. At the very least, we suggest that a more natural approach to assessment is better than no assessment at all.
This is an ongoing project designed primarily to explore the receptive language of autistic children, just as other investigators have broadly mapped out the receptive skills of normal children. We examine receptive competence in severe autism by focusing on those children who actually fail to obtain basal scores on the PPVT-R and those who are considered untestable. Using social activities designed to focus attention, we examine the children's performance in identifying common objects and following simple instructions, and we compare it to formal evaluations made by speech/language pathologists. We propose that such naturalistic evaluations will reveal greater receptive
competente than indicated by formal
evaluations.
Method
Subjects
All subjects are institutionalized
nonspeaking autistic children at the Anderson School for Autism
and Developmental Disabilities in Staatsburg, New York State,
USA. In formal assessments conducted by the schoolls speech/language
pathologist, the first three children (SE, SL and BP) failed to
obtain basal scores on the PPVT-R, performing at chance level.
SE, the only female, is now 11:7 years old; she knows a few signs,
but she engages in delayed echolalia of a single word or nonsense
utterances; she is taking Ritalin. SL is currently 13:8 years
old; he is completely mute and is mildly self-injurious and very
easily distracted. BP, 11:1 years old when last tested, is completely
mute; he frequently stares into space and appears unresponsive;
he is self-injurious and occasionally very aggressive. The following
two children (GR and MG) were never given any sort of formal standardized
test because, in the opinion of the speech/language pathologist,
they are untestable. GR, 15:7 years old, is completely mute;
he is mildly self-injurious, and he is taking Trexan (naltrexone).
MG, 8:10 years old, is completely mute; he typically has a calm
demeanor. Informed consent was obtained from parents or guardians,
and children were treated in accordance with the ethical principles
of the American Psychological Association.
Materials & Apparatus
All test objects used in assessment were exemplars of things commonly used by or with these children in the classroom or in residential areas. These included a large ball, keys, sponges, combs, hairbrush, toothbrushes, shoes, crayons, paper, cup, bowl, teddy bear, socks, sunglasses, flowers, hat, watch, pumpkin, garbage can, spoon, book, bell, (coffee) can, wall switches for electric lights, and candy. Large full color photographs of common objects were mounted on 15 X 20 cm index cards.
A small portable Panasonic cassette
tape recorder, with a variety of musical selections, was used.
All sessions were videotaped with a Panasonic VHS camcorder.
Procedure
We interviewed the school's speech/language pathologists who had done their own evaluations of each of our children. We discussed old PPVT-R scores that were already on file for some of the children, and in two cases we were able to videotape their formal re-testing on the PPVT-R. We also established that there were no formal assessments on file for two of the children, and that there were no plans to conduct any.
All children were initially desensitized to the assessment procedure through multiple visits with the evaluator, a 37-year old female graduate student (the second author, C.G.). Over several weeks, she established a rapport with the children, beginning as a classroom observer and building to a one-on-one level of interaction with each child. Once this relationship was in place with each child, formal sessions began. On each day that a child was to participate in a session, the first and or second author went to the classroom to greet the teacher and child, and to chat briefly. Then each child was asked "Do you want to come play with us today?" In general, we found the children quite eager to follow us (including on those days when we did not intend to test them, and they were not supposed to be following us). In fact, we occasionally encountered some difficulty getting children to return to the class after a session was over.
We started by getting an informal baseline performance on receptive tasks by placing four objects or four pictures at a time on the table and asking the child to "Point to the (object/picture) or "Give me the (object/picture)". We decided to use four items because the PPVT-R uses four line drawings at a time.
Subsequently, all sessions were begun with an attempt to focus the child's attention. As this process was individualized for each child, it was necessary to determine the means by which each child would attend. Initially, the evaluator played a variety of musical selections on the tape recorder, and responses to the music were noted. The tape recorder itself became a means of focusing attention, as some of the children became interested in the machine and how to push the buttons to make music. SL, who was easily distracted, became more agitated by any type of music, and was most attentive when he was allowed to walk around and throw individual items in a garbage can.
Each child was assessed in individual sessions. When the child was calm and focused, we began the trials. The evaluator and the child were seated next to each other at a table. The first author videotaped the sessions and observed unobtrusively. For the object identification task, five to eight objects at a time were placed on the table. In order to maximize comprehension, a variety of directives was used. Each set of trials began with the evaluator instructing the child "Give me the [object]." If the child did not respond, the directive was rephrased, (e.g. "Can you give me the [object]").
Additionally, the child was asked to use and manipulate a particular object, for example, "Show me how you brush your hair", "Clean the table with the sponge", "Turn the music on/off", or "Ring the bell". The environmental context was also used in assessment. Children were asked to "Turn the lights off" as they were leaving the room, "Throw the [object] in the garbage", or "Show me how you dance to the music". We then moved on to tasks which involve the comprehension of prepositional phrases, such as "in" or "on," with instructions such as "Put the comb in the cup".
Finally, we attempted to assess comprehension of abstract concepts such as "other" or "another", by using multiples of an item randomly arranged on the table along with several unrelated objects. The child was then asked "Give me the sponge", and then (if accomplished), "Give me the other one," or "Give me another one".
During all tasks, the evaluator was carefully monitored (by the first author, C.V.) to prevent the use of cueing. When the directive "Give me ... " was used, the evaluator extended her hand toward the child, being careful not to look at or gesture toward the requested object. M & M candies and praise ("Thanks for playing the game, you're doing so wellll) were used as rewards for cooperation; in some cases, the playing of a musical selection served as a reward.
All sessions were conducted in either
the unoccupied school library or the child's unoccupied classroom.
Sessions were generally 30 minutes in length, and rarely had
to be terminated earlier because the child became disturbed or
tried to leave the room.
Results
Videotapes were transcribed and scored separately by the first and second author. Our results may be briefly outlined as follows:
1- The baseline measures we took resulted in chance level performance (approximately 25% correct).
2- The videotaped formal re-testing of two children (SE and SL) on the PPVT-R by the speech/language pathologist resulted in chance performance (approximately 25% correct).
3- Efforts to focus the children's attention and to conduct assessment in a social context led to substantially better performance than that obtained on our baseline measures or on formal PPVT-R assessment. All children, including those who were assumed to be untestable, demonstrated functional receptive skills.
4- Different children have different skills.
5- A childls performance can vary enormously from one day to
the next. Decrements in performance appeared to be caused by: a change in medication; a recent home visit; a recent dental visit; medical problems; excessive sleepiness; and for no obvious reason.
6- Asking children to use or to manipulate
an object led to better performance than simply asking them to
hand over an object, i.e., "Show me how you drink out of
a cup" versus "Give me the cup".
Discussion
We suggest that nonspeaking autistic children comprise a heterogeneous group with a range of functional receptive language skills and cognitive abilities that are not reflected in standardized assessment instruments and protocols designed for normal children. If we take at face value our baseline measures, or the PPVT-R scores, or the speech/language pathologists' assumptions that children are untestable, we should conclude that these nonspeaking autistic children, and by extension severely autistic children in general, have no receptive competence. Thus, it appears that when techniques for assessing severely autistic children are simply borrowed from those used for normal children, the results are misleading.
Within this group, the child who shows the greatest range of receptive skills is the girl, SE. She was successful at most of the tasks, and her performance was fairly consistent. One quirk we found with her is that when asked for a shoe, she grabs her own shoe, ignoring the shoes on the table, and when asked for "another one", she grabs her other shoe, again ignoring the shoes on the table. This may be an example of the sort of underextension errors commonly made by very young normal children. The other children showed a variety of receptive abilities which manifested themselves in unusual ways. For example, while SL generally refused to hand objects to the evaluator, we were able to capitalize on his penchant for throwing things in the garbage by asking him to throw specific objects in the garbage can. And while his receptive skills were not as strong or as reliable as those of the girl, SE, he seemed to have a better grasp of the preposition "in", as when we asked him to "put the comb in the cup". Even the children who were classified as untestable demonstrated understanding by turning off the light switches when told to do so, and by quickly learning how to operate a tape recorder they had never seen before. They responded appropriately, for example, to "Push the big button. You have to push it harder". At the other extreme, BP repeatedly failed all but one of the tasks ("Show me how you brush your hair"), and he became so aggressive that we decided to exclude him from our study. Thus, we acknowledge that the advantages of naturalistic assessments may be limited.
One of our most intriguing findings is that asking children to use or manipulate an object leads to better performance than asking them to give you the object. First, a little change in phraseology can lead to a big difference in performance and in evaluation. Also, this effect may be related to how young normal children acquire vocabulary faster when words are part of routines such as getting dressed; this phenomenon has been described as affecting encoding or acquisition. what we have found is that for autistic children, the benefit accrues to retrieval, i.e., they seem to "know" the word better if it is put in the context of some action or behavior.
Ours is an ongoing project. we are
at different stages with different children. A couple of children
(SE and SL) have worked with us for months, and others are relatively
new to us. We are now in the process of expanding our subject
sample to include more children. In working with these children,
we feel that we have learned things about severe autism that are
of general theoretical interest, and that we have caught on to
strategies that might make assessment more accurate. Ultimately,
we hope our work will lead to the development of alternase assessment
protocols that are designed to meet the needs of severely autistic
children.
References
Baron-Cohen, S. (1989). Perceptual role-taking and proto-declarative pointing in autism. British Journal of Developmental Psychology, 7, 113-127.
Bates, E. (1976). Language and context: The acquisition of pragmatics. New York: Academic Press.
Bates, E. (1979). The emergence of symbols: Cognition and communication in infancy. New York: Academic Press.
Biklen, D. (1990). Communication unbound: Autism and praxis. Harvard Educational Review, 60, 219-314.
Bloom, L. (1973). One word at a time. The use of single word utterances before syntax. The Hague: Mouton.
Bloom,L. (1974). Talking, understanding, and thinking. In Schiefelbusch, R. L. & Lloyd, L. L. (Eds.), Language perspectives: Acquisition, retardation and intervention. Baltimore, MD: University Park Press.
Bruner, J. (1974/1975). From communication to language: A psychological perspective. Cognition, 3, 255-287.
Bruner, J. (1975). The ontogenesis of speech acts. Journal of Child Language, 2, 1-19.
Curcio, F. (1978). Sensorimotor functioning and communication in mute autistic children. Journal of Autism and Childhood Schizophrenia, 8(3), 281-292.
Crossley, R. (1988). Unexpected communication attainments by persons diagnosed as autistic and intellectually impaired. Paper presented at International Society for Augmentative and Alternative Communication, Los Angeles, CA.
Dunn, L. & Dunn, L. (1981). The Peabody Picture Vocabulary Test-Revised. Circle Pines, MN: American Guidance Service.
Gardner, M. (1985). Receptive One-Word Picture Vocabulary Test. Novato, CA: Academic Therapy.
Greundel, J. M. (1971). Referential extension in early language development. Child Development, 48, 1567-1576.
Huttenlocher, J. (1974). The origins of language comprehension. In R. L. Solso (Ed.), Theories of cognitive psychology. Hillsdale, NJ: Lawrence Erlbaum Associates.
Jacobson, J., Mullick, J. & Schwartz, A. (1995). A history of facilitated communication. American Psychologist, 50(9), 750-765.
Lock, A. (1980). The guided reinvention of language. London: Academic.
Lund, N. & Duchan, J. (1988). Assessing children's language in natural contexts. Englewood Cliffs, NJ: Prentice-Hall.
McShane, J. (1980). Learning how to talk. Cambridge: Cambridge University Press.
Nelson, K. (1985). Making sense: The acquisition of shared meaning. New York: Academic Press.
Oviatt, S. (1982). Inferring what words mean: Early development in infants' comprehension of common object names. Child Development, 53, 274-277.
Reich, P. (1976). The early acquisition of word meaning. Journal of Child Language, 3, 117-123.
Ricks, D. & Wing, L. (1975). Language, communication and the use of symbols in normal and autistic children. Journal of Autism and Childhood Schizophrenia, 5(3), 191-221.
Rutter, M. (1978). Language disorder and infantile autism. In M. Rutter & E. Schopler (Eds.), Autism: A reappraisal of concepts and treatments. New York: Plenum Press.
Rutter, M. (1983). Cognitive deficits in the pathogenesis of autism. Journal of Child Psychology and Psychiatry, 24, 513-531.
Sigman, M., Mundy, P., Sherman, T. & Ungerer, J. (1986). Social interactions of autistic, mentally retarded and normal children and their caregivers. Journal of Child Psychology and Psychiatry, 27(5), 646-656.
Vazquez, C. (1995). Failure to confirm the word-retrieval problem hypothesis in facilitated communication. Journal of Autism and Developmental Disabilities, 25(6), 597-610.