| Autism is a neurobehavioral syndrome that has been characterized by abnormal development of language, social reciprocity, and social cognition, and a desire for sameness that is displayed as repetitive rituals, difficulty with transitions, and highly circumscribed interests (American Psychiatric Association, 1994). It is presently considered to traverse a spectrum of severity in each of these areas. Some controversy exists regarding whether primary language impairments, unrelated to social deficits, exist in autistic persons across the entire spectrum. To investigate this question, archived language samples of 7- and 9-year-old children who had been diagnosed at preschool age with high functioning autism (n = 16), developmental language disorder ( n = 16), and control subjects (n = 8) were examined for the presence of language disfluencies known as mazes. Hypotheses followed prior research examining specific maze types (filled pauses and content mazes) that suggested different patterns of frequency between the maze types within and between groups of children with Specific Language Impairment and normal controls.; Results of the current study showed a higher percentage of content mazes in each clinical group versus normal controls. Although differences in percentage of filled pauses could not be detected between groups, a within-subjects ANOVA demonstrated a significant interaction between group and specific maze type. These results lend support to the central hypothesis that even verbal HFA children do show evidence of higher-level language impairment that is more similar to DLD children than to normal controls. Regression analyses attempting to use additional language variables to predict maze types produced limited results. Most significantly, the percentage of topic-maintaining statements was found to be a predictor of percent content mazes. Integration of all results suggests that filled pauses may have served a pragmatic, as well as formulation, function for normal control participants that may have been less true for participants with language impairment or autism. Further research in this area will require more sensitive ways to define and measure language variables that are common in autism but have ambiguous origins. |
