Controlling for homophone polarity and prime-target relatedness in the cross-modal lexical decision task

Abstract

Access of alternative meanings of ambiguous words was investigated with a cross-modallexical decision task. We considered two factors that may have been inadequately controlled in previous experiments, homophone polarity and prime-target relatedness. We used both nonpolar homophones with two nearly equiprobable senses and polar homophones with one clearly dominant sense. Polarity ratings were derived from homophone familiarity norms obtained by Kreuz (1987). We also equated prime-target relatedness across items for dominant and subordinate meanings of the homophones using norms that we collected. Sentence contexts biased the dominant meanings of the homophones. Under these conditions, there was a small but significant priming effect, and priming of the subordinate meaning was not significantly affected by homophone polarity.The cross-modal lexical decision task has been a primary procedure for investigating lexical access. A major issue addressed in previous research has been whether lexical access to alternative meanings of ambiguous words is context-dependent or initially exhaustive, with context operating on "postaccess" processes. Past cross-modal studies, however, may have failed to adequately control the polarity of ambiguous items (i.e., the degree to which one word sense dominated over the other) and the relatedness of the ambiguous items to the lexical targets (Holmes, Arwas, & Garrett, 1977;Onifer & Swinney, 1981;Seidenberg, Tanenhaus, Leiman, & Bienkowski, 1982;Swinney, 1979; Tabossi, 1988; Tabossi, Colombo, & Job, 1987; Tanenhaus, Leiman, & Seidenberg, 1979).To control polarity, for example, Onifer and Swinney (1981) presented subjects with homographs and homophones and asked them to describe the first meaning that came to mind. The words that elicited one common interpretation from over 75% of the subjects, and a different common interpretation from less than 25 % of the subjects, were deemed "polar" and used in the experim~nt. The problem with this procedure is that it may incorrectly label some nearly equiprobable (nonpolar) ambiguous words as polar. Consider a word with two senses of nearly equal frequencies. If one sense is just slightly more familiar than the other, it could be chosen as the dominant meaning by the 75 % of the subjects necessary in the Onifer and Swinney pretest, and would therefore be designated as "polar. " Priming of subordinate meanings of supposedly polar homographs would then provide an artifactual demonstration of exhaustive lexical access. If Onifer and Swinney labeled nearly equiprobable ambiguities as polar, their exhaustive access results would be less convincing.This research was supported in part by Nlli Grant I-ROI AG09253. We would also like to thank: Tina Loose for her help with statistical analyses and Sam Glucksberg for his helpful comments on an earlier draft. Correspondence may be addressed to Marcia K. Johnson. Department of Psychology. Green Hall. Princeton University. Princeton. NJ 08544-1010. 15In the present experiment, we sought to control polarity by...

Irony in conversation: salience, role, and context effects

“…The occurrence of irony in the conversations made its literal meaning available for further discussion and elaboration, as predicted by the graded salience hypothesis (Giora 1997(Giora , 2003 and indirect negation view (Giora 1995). However, they are partly inconsistent with the standard pragmatic model (Grice 1975;Searle 1979) and modular view (Onifer and Swinney 1981;Picoult and Johnson 1992;Seidenberg et al 1982;Swinney 1979), attesting that the literal meaning was not suppressed as irrelevant. Further, they are incompatible with the direct access view, demonstrating that irony did not avail the ironic meaning exclusively.…”

From Bye to Buy: Homophones as a Phonological Route to Priming

“…Polarized homophone pairs are important to consider in the current investigation. Bidirectional priming effects may not occur for polarized homophone pairs (Picoult and Johnson 1992). Reading "you" may not activate meanings associated with "ewe" to the same extent that reading "ewe" activates meanings associated with "you."…”

“…Bonin and Foyol (2002) found that in naming tasks, low-frequency homophones did not receive a priming/processing advantage from higher frequency homophone mates (but the reverse was true). Picoult and Johnson (1992) studied homophone priming effects in lexical decision tasks using polarized and nonpolarized homophone pairs. Priming effects were found to not always exhibit bidirectionality, but evidence did support that access to both meanings is increased (see also Pexman, Lupker, and Hino 2007).…”

Some acoustic indicators of verbal irony in semi-spontaneous contexts

“…De hecho, sin la presencia de señales contextuales sería imposible que el oyente descifrara el significado pretendido del hablante (Rockwell, 2000). Sin embargo, en la mayoría de estudios realizados ad hoc se concluye que el acceso a la información implícita irónica apenas se ve afectado por la información contextual (Vu, Kellas y Paul, 1998;Martin y otros, 1999), siendo muchas veces los procesos léxicos impermeables a los efectos del contexto y, por tanto, independientes del sesgo contextual (Picoult y Johnson, 1992;Hartung, 1998).…”