Advanced Psycholinguistics

Lexical priming is known to arise from phonological similarity between prime and target, and this phenomenon is an important component of our under-standing of the processes of lexical access and competition. However, the precise nature of the role of phonological similarity in lexical priming is understudied. In the present study, two experiments were conducted in which participants per- formed auditory lexical decision on CVC targets which were preceded by primes that either matched the target in all phonemes (CVC condition), in the first two phonemes (CV_ condition), the last two phonemes (_VC condition), the initial and last phonemes (C_C condition) or no phonemes (unrelated condition). Relative to the unrelated condition, all conditions except CV_ led to facilitation of response time to target words. The _VC and C_C conditions led to equivalent facilitation magnitude, while the CV_ condition showed neither facilitation nor inhibition. Accounting for these results requires appeal to processes of lexical competition and also to the notion that phonemes do not lend equivalent phonological similarity; that is, vowels and consonants are processed differently.

Phonological priming has long been used to study the mechanisms by which the acoustic signal is associated with the phonological representations of words in the mental lexicon. In this paradigm, a prime word is presented immediately prior to a target word, and some task – such as lexical decision or word shadowing – is per- formed on the target word. The type and extent of phonological similarity between the prime and target has been found to affect response latency for a wide variety of tasks. This paper presents the results of two auditory lexical decision experiments which investigate the nature of phonological similarity. There is evidence from a diversity of auditory tasks that overlap in the syllable rime between prime and target (e.g. hat priming cat) leads to facilitation, that is, faster or more accurate responses to the target (Dufour & Peereman, 2009; Dumay et al., 2001; Gray, Reiser, & Brinkley, 2012; McQueen & Sereno, 2005; Norris, McQueen, & Cutler, 2002; Radeau, Besson, Fonteneau, & Castro, 1998; Radeau, Morais, & Segui, 1995; Slowiaczek, McQueen, Soltano, & Lynch, 2000; Slowiaczek, Nusbaum, & Pisoni, 1987). This facilitatory effect is well-studied and it has been ar- gued to be due to overall prelexical excitation caused by the phonological similarity between the prime and target (Slowiaczek & Hamburger, 1992).

However, a good deal of the effects uncovered by early investigations have subsequently been demon- strated to be due, in part, to participant response bias (Dufour, 2008; Goldinger, 1998; Hamburger & Slowiaczek, 1996; McQueen & Sereno, 2005; Norris et al., 2002; Slowiaczek & Hamburger, 1992). This bias arises due to systemic patterns which may be present in the stimulus lists. For example, in a lexical decision task, if all nonword trials involve stimuli with no phonological overlap, then the presence of phonological overlap in some of the word trials suffices to cue the listener that the target is a word. The listener can then respond quickly, without actually accessing the target stimulus. This speedy strategy leads to an apparent facilitation for trials with phonological overlap, but is due entirely to strategic effects. For this reason, phonological priming experiments must be designed with carefully balanced sets of stimuli, with equal care taken over the non-target trials as over the target trials (Dufour, 2008). Overlap located at the beginning of the words (e.g. cab priming cat), often termed “onset overlap”, has been reported to lead to inhibition, that is, slower or less accurate responses to the target (Desroches, Newman, & Joanisse, 2008; Dufour & Peereman, 2009; Gaskell & Dumay, 2003; Goldinger, Luce, Pisoni, & Marcario, 1992; Hamburger & Slowiaczek, 1996). However, some studies failed to find such an effect (Gray et al., 2012; McQueen & Sereno, 2005;

Radeau et al., 1995; Slowiaczek & Pisoni, 1986), and indeed, early experimental forays into this phenomenon reported facilitation (Slowiaczek & Hamburger, 1992; Slowiaczek et al., 1987). After subsequent investigation, a general consensus has been reached that any facilitation observed can be attributed to expectancy-based strategies arising due to bias in the stimuli set; absent this bias and with a short ISI between prime and target, onset overlap leads to inhibition from cohort competition. That is, the presentation of cab causes lateral inhibition of cohort competitors, such as cat, can, candle, and so on. This competition is short-lived, though, such that for designs with a longer ISI (e.g. 500 ms), neither inhibition nor facilitation is observed (Dufour, 2008; McQueen & Sereno, 2005).

From the Lexique lexical database of French (New, Pallier, Ferrand, & Matos, 2001), 84 monosyllabic CVC words, including nouns, adjectives and verbs, were selected as critical target words. Each target word had five primes, not necessarily of the same syntactic class, associated with it, one for each of the five priming conditions. The five conditions are summarized in Table 1, and are described in terms of the extent and type of phonological overlap between the prime and target. In the CVC condition, the prime word and the target word are identical; in the _VC condition, the VC overlaps and the first consonant differs; in the C_C condition, the consonants overlap and the vowel differs; in the CV_ condition, the CV overlaps and the last consonant differs; and in the unrelated condition, all of CVC differ. While every effort was made to select critical target words with all five possible prime types, the presence of accidental gaps in the lexicon meant that 23 potential prime-target pairs could not be constructed. In total, 397 words (84 targets× 5 prime types − 23 impossible combinations) were chosen as primes.

Of these primes, 21 differed from their targets in subsyllabic structure: for example, style /stil/ was used as the onset competitor for ville /vil/, due to the paucity of possible onset competitors of shape CVC.1