RESUMEN
The purpose of this study was to explore the errors and revisions (i.e., repairs) that 3- and 4-year-old children who require augmentative and alternative communication (AAC) make as they construct 2-3-word utterances using graphic symbols. The current study involves supplemental analyses from a previously published manuscript. Various types of errors and revisions for four different semantic-syntactic structures-agent-action-object, entity-attribute, entity-locative, and possessor-entity-were analysed to explore patterns and differences across utterance types. Results indicated that the majority of errors were made during the baseline phase, and that error types varied depending on the utterance type. For example, inversions were common for agent-action-object utterances, but omissions were common for entity-attribute utterances. When the participants revised their utterances, the resulting messages were more accurate the majority of the time, regardless of utterance type. Past research has highlighted frequent word order errors within graphic symbol messages, but the current results indicate that error types are dependent upon utterance type. A more refined approach, then, is required to better our understanding of how children approach the task of learning to produce graphic symbol utterances.