Event-related Brain Potentials Elicited by Morphological, Homographic, Orthographic, and Semantic Priming

Abstract

The morphological structure of words, in terms of their stem morphemes and affixes, could influence word access and representation in lexical memory. Three experiments were carried out to explore the attributes of event-related potentials evoked by different types of priming. Morphological priming, with pairs of words related by their stem (hijo/hija [ son/ daughter]), produced a sustained attenuation (and even a tendency to positivity) of the N400 shown by unrelated words across the three experiments. Homographic priming (Experiment 1), using pairs of words with a superficially similar stem, but without morphological or semantic relation (foco/foca [floodlight/seal]), produced an initial attenuation similar to the morphological pairs, but which rapidly tended to form a delayed N400, due to the impossibility of integration. However, orthographic priming (rasa/rana [flat/frog]) in Experiment 2 does not produce attenuation of the N400 but an effect similar to that of unrelated pairs. Experiment 3 shows that synonyms advance more slowly than morphological pairs to meaning coherence, but finally produce a more positive peak around 600 msec.

ERP Evidence of Morphological Analysis from Orthography: A Masked Priming Study

“…structure occurs at the orthographic level, with little or no contribution from semantics. Previous ERP research investigating long SOA priming in Spanish had given some support to this idea (Dominguez et al, 2004;Barber et al, 2002). These studies found a reduction in the early part of the N400 component of target recognition that was equivalent for stem homograph pairs (which had an apparent morphological relationship) and transparent morphological pairs.…”

“…Like Rastle et al (2004), we compared masked priming effects for semantically transparent pairs (e.g., darkness-DARK ), semantically opaque pairs (e.g., corner-CORN ), and nonmorphological form pairs (e.g., brothel-BROTH). Based on the suggestive findings of Dominguez et al (2004) and Barber et al (2002), our expectation was that the semantically transparent and semantically opaque primes would yield equivalent attenuation of the N400 component in response to their respective targets (relative to unrelated control primes), and that this reduction in the N400 component would be greater than that yielded by nonmorphological form primes. The pattern of data predicted in this study would constitute the first unambiguous neurophysiological evidence for a form of morphological decomposition that is insensitive to semantic information.…”

“…However, intriguing effects were found in the N400 component of the ERP waveform, previously shown to be sensitive to lexico-semantic information (e.g., Holcomb, Reder, Misra, & Grainger, 2005;Kiefer, 2002;Kutas & Federmeier, 2000). Specifically, Dominguez et al (2004) and Barber et al (2002) found that in the early part of the N400 temporal window, both morphologically related pairs and stem homograph pairs were associated with an attenuated N400 relative to the N400 elicited by unrelated words ( pavo-META, turkey-goal). Later in the ERP segment, the N400 for morphologically related pairs remained reduced, whereas the N400 for stem homographs increased, reaching the same amplitude as that in the unrelated condition.…”

“…There is already some ERP evidence consistent with the proposal that there is a form of morphological decomposition that arises early in visual word processing, prior to the activation of semantic information. Two recent studies conducted in Spanish compared priming effects using fully visible primes (250 msec) in stem homograph pairs (e.g., rata-RATO; rat-moment) with priming effects in morphologically related pairs (e.g., loca-LOCO; madwoman-madman) (Dominguez, de Vega, & Barber, 2004;Barber, Dominguez, & de Vega, 2002). Behavioral effects were consistent with other long SOA visual priming studies (e.g., Rastle et al, 2000): Robust priming effects emerged only for the semantically related morphological pairs.…”

Electrophysiological evidence for a natural/artifactual dissociation

“…This component is a negative wave with an onset latency between 200 and 250 ms post-stimulus, with maximum amplitude at around 400 ms. This wave has been related to semantic processing, since its amplitude is enhanced with semantic incongruence between word and context (Kutas and Hillyard, 1980;Neville et al, 1986) and decreases in conditions in which the context enables the subject to predict the word (Holcomb and Neville, 1991) or when semantically related stimuli have previously been presented (Bentin, 1987;Bentin et al, 1985;Holcomb, 1988;Kutas and Van Petten, 1988;Holcomb and Neville, 1990;Dominguez et al, 2004;Kiefer, 2005). The presence of the N400 has also been tested in studies in which nonverbal materials are used, and, just as with words, semantic priming in picture processing tasks decreases the amplitude of this component (Barrett and Rugg, 1990;McPherson and Holcomb, 1999).…”