Productive use of the English past tense in children with focal brain injury and specific language impairment Virginia A. Marchman, a, * Cristina Saccuman, b and Beverly Wulfeck c a School of Behavioral and Brain Sciences, The University of Texas at Dallas, P.O. Box 830688 GR 41, Richardson, TX 75083-0688, USA b San Diego State University/University of California, San Diego, USA c San Diego State University, San Diego, USA Accepted 2 April 2003 Abstract In this study, 22 children with early left hemisphere (LHD) or right hemisphere (RHD) focal brain lesions (FL, n ¼ 14 LHD, n ¼ 8 RHD) were administered an English past tense elicitation test (M ¼ 6:5 years). Proportion correct and frequency of over- regularization and zero-marking errors were compared to age-matched samples of children with specific language impairment (SLI, n ¼ 27) and with typical language development (TD, n ¼ 27). Similar rates of correct production and error patterns were observed for the children with TD and FL; whereas, children with SLI produced more zero-marking errors than either their FL or TD peers. Performance was predicted by vocabulary level (PPVT-R) for children in all groups, and errors did not differ as a function of lesion side (LHD vs. RHD). Findings are discussed in terms of the nature of brain–language relations and how those relationships develop over the course of language learning. Ó 2003 Elsevier Science (USA). All rights reserved. Keywords: Specific language impairment; Focal brain injury; Morphology; English past tense; Plasticity; Overregularization 1. Introduction English-speaking children typically begin to mark plural or past tense forms before their second birthday and often do so appropriately for both regular (e.g., ‘‘Daddy walked’’) and irregular forms (e.g., ‘‘Johnny took my blocks,’’ ‘‘I won!’’). Later, inappropriate uses of inflectional morphemes (e.g., taked, winned) begin to be observed. These errors persist well into the school-age period, however, their frequency gradually diminishes as childrenÕs production of both regular and irregular forms approaches an adult-like pattern. It is generally assumed that these errors reflect progress in the devel- opment of productive language use, i.e., the hallmark human ability to generate words or sentences that have not been heard in the input (Berko, 1958; Bybee & Slobin, 1982; Cazden, 1968; Kuczaj, 1988). In recent years, the details of this achievement and the precise mechanisms guiding its development have been the subject of considerable study, refinement, and debate. It has become clear that a simple stage-like ac- count is inadequate and does not account for the com- plex developmental pattern that has emerged in more recent studies (e.g., Marchman, 1997; Marcus, Pinker, Ullman, & Hollander, 1992; Plunkett & Marchman, 1991). Children do not enter a period in which the reg- ular rule is applied across-the-board. Instead, past tense forms of some irregular verbs are produced correctly at the same time that others are being overregularized. Although it is rare to find a child who never produces overregularizations (Marchman, 1997), errors typically reflect only a small portion of childrenÕs irregular verb use (e.g., less than 15% reported by Marcus et al., 1992). Finally, while overregularizations are the most oft-cited evidence that children have abstracted systematicities that are inherent in the language, other types of pro- ductions also occur, including zero-markings (e.g., ‘‘he sit’’) and vowel changes (e.g., ‘‘she brang’’). Analyses have shown that these errors are systematic (Marchman, Brain and Language 88 (2004) 202–214 www.elsevier.com/locate/b&l * Corresponding author. E-mail address: vamarch@utdallas.edu (V.A. Marchman). 0093-934X/$ - see front matter Ó 2003 Elsevier Science (USA). All rights reserved. doi:10.1016/S0093-934X(03)00099-3