There are problems in interpreting the naming behavior of children. Children may misname a word because the word is absent from their vocabulary, because it is not yet firmly established, or because of a word retrieval or lexical assessing problem. Preliminary results are reported of an experimental technique designed to account for these differences and to assist the examiner in making inferences about the development of naming skills in children. A 50- item confrontation naming task was designed with pictures chosen from 5 semantic categories. Each category contained five words judged to be in a 9-year old's vocabulary and five words judged to be of rare occurrence. Each child saw each picture twice. Responses were divided according to accuracy, and errors were placed into one of two groups. The first group contained errors in words known to be in the child's vocabulary; the second contained errors for words that were unavailable to the child. Examination and comparison of data suggest that the naming process used for corrections immediately after a response and for single errors might be the same process as that used for words that are either unstable or unavailable to the child. The data also show differences in use of naming strategies according to age group. This suggests that strategies for retrieving words change in time. (AMH)