Cjep 60-4

The role of orthographically similar words (i.e., neighbours) in the word recognition process has been studied extensively using short-term priming paradigms (e.g., Colombo, 1986). Here we demonstrate that long-term effects of neighbour priming can also be obtained. Experiment 1 showed that prior study of a neighbour (e.g., TANGO) increased later lexical decision performance for similar words (e.g., MANGO), but decreased performance for similar pseudowords (e.g., LANGO). Experiment 2 replicated this bias effect and showed that the increase in lexical decision performance due to neighbour priming is selectively due to words from a relatively sparse neighbourhood. Explanations of the bias effect in terms of lexical activation and episodic memory retrieval are discussed. Résumé Le rôle de mots semblables d’un point de vue orthographique (c.-à-d. des mots voisins) dans le processus de reconnaissance du mot a été étudié de façon exhaustive à l’aide de paradigmes d’amorçage à court terme (par ex., Colombo, 1986). Ici nous démontrons que des effets à long terme à la l’amorçage du mot voisin peuvent également être obtenus. L’expérience 1 montre que l’étude préalable d’un mot voisin (par ex., TANGO) améliorait plus tard le rendement de décision lexicale pour des mots semblables (MANGO), mais diminuait le rendement pour des pseudomots semblables (LANGO). L’expérience 2 a répété cet effet de biais et montré que l’amélioration du rendement à la décision lexicale à cause de l’amorçage du mot voisin est sélectivement attribuable aux mots d’un voisinage relativement épars. Les explications de cet effet de biais en termes d’activation lexicale et de récupération en mémoire épisodique font l’objet de discussions. In the last decade, effects of orthographic similarity in visual word recognition have been studied extensively (see Andrews, 1997, for a review, and Grainger & Jacobs, 1996, for a model). It is generally accepted that in the initial stages of processing, a word stimulus will activate not only its own representation in lexical memory, but will also partly activate word representations that are orthographically similar to it (cf. Pecher, Zeelenberg, & Wagenmakers, 2005). The detailed consequences of the co-activation of orthographically similar representations have been subject to considerable debate. One of the reasons for this ongoing discussion lies in the fact that the impact of “neighbours” (i.e., words of the same length differing from each other in one letter such as MANGO-TANGO, Coltheart, Davelaar, Jonasson, & Besner, 1977; Landauer & Streeter, 1973) on the word recognition process has turned out to be subtle and task-dependent. That is, the influence of neighbours is generally considered to decrease performance in both auditory word recognition (e.g., Goldinger & Luce, 1989) and perceptual identification (e.g., Grainger & Segui, 1990). In contrast, neighbourhood effects in naming are generally facilitatory (e.g., Andrews, 1997), and results for lexical decision have been mixed. Two methods are generally used in lexical neighbourhood research, matching and short-term priming. With the matching method, neighbourhood characteristics are highly correlated with several other variables such as word frequency and bigram frequency that are known to affect word recognition performance. Experiments that use matching equate stimuli on all variables thought to affect performance except the variable of interest (i.e., the neighbourhood characteristic). The disadvantage of matching is that one can never be sure that the stimuli have been equated on every variable that may confound the results (e.g., imagineability, body-rhyme consistency, etc.). Also, for some variables, the quality of the matching procedure depends on the reliability of norms or frequency counts. Finally, matching is a procedure that is arguably sensitive to experimenter bias effects (cf. Forster, 2000). In short-term priming, also called form-related priming, the participant is usually presented with a briefly displayed “prime” word (e.g., TANGO or FLOOR) that is replaced by a target word (e.g., MANGO) to which a response is required (e.g., Colombo, 1986). This method circumvents the potential problems associated with the matching procedure, since every stimulus can be used in all conditions according to a counterbalEric-Jan Wagenmakers and Jeroen G. W. Raaijmakers University of Amsterdam Canadian Journal of Experimental Psychology, 2006, 60-4, 275-284 Canadian Journal of Experimental Psychology Copyright 2006 by the Canadian Psychological Association 2006, Vol. 60, No. 4, 275-284 DOI: 10.1037/cjep2006025 CJEP 60-4 12/4/06 9:05 PM Page 275