Tema: No desenvolvimento típico há um movimento constante em direção à maior complexidade lingüística com o decorrer do tempo, mas esta evolução nem sempre é observada nos distúrbios de linguagem. Especificamente nos Transtornos do Espectro Autístico, os estudos têm apresentado resultados conflitantes a respeito da evolução lingüística. Objetivo: Caracterizar a linguagem oral de crianças autistas estudando as variáveis gramaticais, vocabulares e pragmáticas em três momentos distintos: avaliação inicial, após seis e após doze meses de terapia fonoaudiológica. Método: participaram desta pesquisa dez crianças do sexo masculino, com idades entre 2.7 e 11.2 anos. Todos foram diagnosticados por médicos como portadores de Transtornos do Espectro Autístico. Foram realizadas filmagens de 30 minutos de interação entre terapeuta e paciente, em três momentos diferentes (inicial, após 6 e 12 meses de terapia). De cada filmagem, foram transcritos os quinze minutos iniciais para análise do Perfil Funcional da Comunicação. Para a análise dos aspectos gramaticais foram transcritos cem segmentos de fala e estes também foram utilizados como corpus de análise dos aspectos vocabulares. Os dados foram analisados quanto à funcionalidade, aspectos gramaticais e vocabulares e foram comparados entre si longitudinalmente. Resultados: não houve diferença estatisticamente significativa nas variáveis entre os três momentos de coleta, exceto com relação ao número de segmentos frasais não contados em outras categorias. Foram encontradas diferenças significativas entre diferentes variáveis ao longo do período de coleta. Conclusão: as variáveis estudadas mostraram consistência no desempenho do grupo estudado

Language is a symbolic process that allows intra and inter-personal communication. Several authors consider it as an integrated system comprised of sub-systems that must work harmonically and interdependently. These functioning can be observed in normal development but not in cases of language disorder. Among them, the disorders of the autism spectrum (ASD) are complex conditions that include disorders in the areas of cognition and socialization, as well as communication. The unbalance among these areas has been described by many authors. Nevertheless there is no agreement whether it can be also observed among language components. The general purpose of this research was to study vocabulary, grammatical and pragmatic variables on the language of subjects with autism spectrum disorders on three different moments: language therapy onset, after six and twelve months. Subjects were ten individuals with diagnosis included in the Autism Spectrum Disorders (ASD) with ages varying between 2:7y. and 11:2y. (mean 7:2y.). All participants were of the male gender and had their diagnosis performed by psychiatrists according to the DSM-IV criteria. All were filmed at the start of weekly language therapy process (moment 1), after six months (moment 2) and twelve months latter (moment 3). The method used on the three studies includes the analysis of a spontaneous speech filmed samples of 30 minutes interaction sessions of each child and a language therapist. To the analysis of the pragmatic variables the first 15 minutes of each sample were transcribed. Data refer to number of communicative acts per minute, communicative means used, amount and type of communicative functions expressed, proportion of interpersonal communicative acts and occupation of the communicative context. To the vocabulary analysis the assessment identified psychological state terms (including desire, emotional, cognitive and physical) and designative terms (of action, natural and cultural entity, special and temporal localization, artifacts, names of body parts and persons). Grammatical variables referred to morphemes, grammatical classes, mean length of utterance in morphemes and words, unintelligible segments and non-categorized phrasal segments. To allow the discussion of data the research was subdivided in three studies. Study 1 assessed vocabulary, grammatical and pragmatic variables of the sample, characterizing the group performance along the studied period. Study 2 analyzed the associations between the types of variables along 12 months of language therapy. Study 3 investigated which of the variables predicted linguistic development after twelve months of language therapy. Statistical analysis showed that the item that better identified the changes that occurred during the twelve months of language therapy was the number of non-categorized phrasal segments. Generally the associations among variables were relatively uniform, showing consistent performance. The third study showed that some grammatical variables predicted pragmatic development while vocabulary variables predicted grammatical development. These findings highlighted the language peculiarities of the studied subjects