Neuroscience Letters 408 (2006) 10–15 Event related potentials reveal differences between morphological (prefixes) and phonological (syllables) processing of words Alberto Dom´ ınguez a,b,∗ , Maira Alija c , Fernando Cuetos c , Mauel de Vega a a Facultad de Psicolog´ ıa, Universidad de La Laguna, 38205 La Laguna, Tenerife, Spain b Instituto Universitario de Ling¨ u´ ıstica Andr´ es Bello, C/Juan de Vera, 13, 38201 La Laguna, Tenerife, Spain c Facultad de Psicolog´ ıa, Universidad de Oviedo, Plaza Feijoo s/n, 33003 Oviedo, Spain Received 23 February 2006; received in revised form 23 May 2006; accepted 23 June 2006 Abstract Behavioral measures in visual priming tasks show opposite effects for syllables and morphemes, which indicate that they are processed by two independent systems. We used event related potentials (ERPs) to explore two priming situations in Spanish: prefix related words (reacci´ on- REFORMA [reaction-reform]), in which prime and target words shared a first syllable that was also a prefix, and syllable related words (regalo- REFORMA [gift-reform.]), in which the shared first syllable was a pseudoprefix in the prime word. Prefix related pairs, unlike syllable related pairs, evoked a very early positivity in reaction to the target (at 150–250 ms window), suggesting that the prefix information is immediately available, at a prelexical stage. By contrast, syllable related pairs showed a larger N400 effect. This late negativity may be caused by lateral inhibition among lexical candidates activated in the lexicon by the prime’s first syllable. © 2006 Elsevier Ireland Ltd. All rights reserved. Keywords: Morphological priming; Prefixes; Syllables; Phonological priming; Morphemes; Inhibitory hypothesis; N400; Early positivity In visual word recognition, readers use letters, syllables, or mor- phemes as sublexical clues for activating an appropriate lexical representation in memory. Whereas letters are perceptual units with clear limits (at least in their printed form), syllables and morphemes have no limits marked in the text. Instead, they are second-order units, obtained by converting clusters of letters into phonological units (the syllables), or meaningful units (the morphemes). Letters, syllables, and morphemes have been proven to influ- ence word recognition but how are these units processed? In a simple serial model, the reader would first recognize let- ters, then group them into syllables, and finally recognize the morphemes that form the word and serve as the basis for its meaning. The results described in the literature on syllable and morpheme processing do not, however, fit a serial model. For when participants perform the lexical decision task on targets preceded by a syllable related prime (e.g., nor-ma/NOR-TE [norm/north]), reaction times increase significantly in compar- ison with a condition in which the words are orthographically ∗ Corresponding author. Tel.: +34 922317637; fax: +34 922317461. E-mail address: adomin@ull.es (A. Dom´ ınguez). related but do not share the initial syllable (e.g., no-ria/NOR- TE [waterwheel/north]) [10,4]. In contrast, decision times for target words are faster when preceded by a word sharing the ini- tial root morpheme (e.g., ni-˜ no/NI- ˜ NA [boy/girl]) compared to when targets are preceded by a morphologically unrelated word [11]. These effects have also been obtained with reaction times in languages other than Spanish, and they are independent of other factors, such as orthographic neighborhood, bigram frequency, or the semantic or formal overlap between words [1,6,18], and not only in visual word recognition but also in word production [5,8]. One possible explanation for the disparity between syllabic and morphological effects is the inhibitory hypothesis. Accord- ing to this, in syllabic priming tasks, the initial syllable of a word would activate all lexical entries corresponding to words starting with that syllable. When the prime is presented, the target – among other possible candidates – will be initially activated in the lexicon, but later on it should be inhibited to accomplish prime recognition. Thus, when the target stimulus appears its recognition is delayed. By contrast, in morphological priming tasks, there is a facilitatory semantic priming because the root morpheme of the prime preactivates a meaning shared by the target. 0304-3940/$ – see front matter © 2006 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.neulet.2006.06.048