Brain and Language 81, 79–88 (2002) doi:10.1006/brln.2001.2508, available online at http://www.idealibrary.com on Stem Homographs and Lemma Level Representations Mark Allen and William Badecker Johns Hopkins University Published online November 21, 2001 Previous studies have found that priming a target like cerr-o (‘‘hill’’) with an unmasked stem-homograph (e.g., cerr-ar, ‘‘to close’’) or with an unmasked allomorph of a stem-homo- graph (e.g., cierr-a, ‘‘he/she/it closes’’) has an inhibitory effect on target recognition. We report evidence from stem-homograph priming studies that implicate both morphological pars- ing and lexical selection processes at the lemma level. We argue that stem-homograph inhibi- tion is the product of lexical selection mechanisms that are engaged just when a fully ambigu- ous stimulus (here, an inflectional stem) must be assigned one of its viable interpretations in order to afford conscious perception of lexical identity.  2001 Elsevier Science (USA) Key Words: inflection; morphological parsing; stem-homograph priming; lexical inhibition. In the course of lexical recognition, a stimulus like park will tend to activate multi- ple word forms that are close matches to it (e.g., part, pork, and dark); these forms compete for selection, a winner is determined, and the other candidates are sup- pressed. Regarding this last step, lexical decision studies have shown that orthograph- ically similar words (e.g., park–part) tend to inhibit each other in immediate priming tasks (Colombo, 1986; Dominguez, de Vega, & Cuetos, 1997; Drews & Zwitserlood, 1995; Grainger, 1990; Henderson, Wallis, & Knight, 1984; Stolz & Feldman, 1995). It is reasoned that if park inhibits the target part, then part was a rejected candidate during the recognition of park. Processes that map orthographic stimuli onto stored lexical forms (via multiple activation and selection by inhibition) will not suffice to disambiguate every form semantically or syntactically, though: Whether park is a noun or a verb cannot be determined by its spoken or written form alone. Thus, mechanisms of candidate acti- vation and selection appear necessary not only at the lexeme level (i.e., at the level of orthographic form), but also at the lemma level (i.e., the meaning/syntax level). The difference between selection at the lexeme level and selection at the lemma level, though, is that a winner at the lexeme level may be determined simply in terms of goodness of fit with the stimulus form itself, whereas selection at the lemma level must draw on other information, such as linguistic context and/or how often a particu- lar word form is used with one sense as opposed to another (Simpson, 1994). We raise these points because we view the inhibitory effect of stem-homograph (SH) priming as support for parsing models of lexical comprehension (Allen & Ba- decker, 1999; Laudanna, Badecker, & Caramazza, 1989, 1992; Orsolini & Marslen- Wilson, 1997). Stem-homographs, e.g., the Spanish verb cerr-ar (‘‘to close’’) and Address correspondence and reprint requests to Mark Allen, Department of Psychology, Box 351525, University of Washington, Seattle, WA 98195. E-mail: mkallen@u.washington.edu. 79 0093-934X/01 $35.00  2001 Elsevier Science (USA) All rights reserved.