It is proposed that in studying the development of children's speech, the findings in the data are heavily influenced by what is expected to be found on the basis of our theoretical preconceptions. This phenomenon is actually more widespread than has previously been acknowledged, and our expectations about how children learn language may have to be revised. The assumptions commonly made are: (1) children's speech can be analyzed using adult categories; (2) language development follows an orderly progression of stages; (3) all children's language development is similar; (4) the intelligibility of the child is an external criterion; and (5) the unintelligible utterances play little part in the intelligible speech of a child. Speech of the subject in this study was analyzed from age 7 months to 2 years and 3 months. The boy's speech was divided into two categories: (1) analytic, which deals with simple words; and (2) Gestalt, which aims at whole phrases and sentences. It is proposed that the Gestalt strategy of early language production is much more widespread than just this case reports, because children's sentences are often not understood as such. In addition to cognitive and sociolinguistic factors that affect a child's language learning, there may be a neurological factor related to the differentiation of the cerebral hemispheres and the eventual lateralization of language. This suggests that the subject's Gestalt strategy of language production may be related to the development of his minor hemisphere while his analytic strategy is related to major hemisphere development. (NCR)