ausblenden:
Schlagwörter:
-
Zusammenfassung:
INTRODUCTION. Previous research on the automaticity of initial syntactic structure building during sentence comprehension employed paradigms involving conscious processing (Friederici, 2011; Hahne & Friederici, 1999). In order to better understand the rapid nature of early phrase structure building, in the present behavioral study we addressed syntactic processing in the absence of conscious awareness. It has been argued that the automatic nature of early syntactic processes might be strongly connected to humans’ predictive capacity to generate syntactic expectations upon the upcoming material (Lau et al., 2006). Following a recent study by Berkovitch and Dehaene (2019), we employed subliminal syntactic priming to test the connection between our predictive syntactic capacity and the automaticity of syntactic analysis.
METHODS. We run three behavioral experiments — one pilot (N = 19) determining the adequate effect size, followed by two main experiments (N = 43, N = 40) — to explore the early influence of congruent primes (er “he”, ein “a”) on the processing of German verbs and nouns. Conscious perception of the primes was induced using unmasked syntactic priming, while unconscious perception was induced using subliminally presented masked primes. The first main experiment included nouns and verbs with the ending -t, which is part of the stem for nouns but an inflectional suffix for verbs. Given that numerous studies reported a rapid decomposition of morphologically complex words into the stem and affixes (Beyersmann et al., 2011; Beyersmann et al., 2016), to exclude a possible influence of the suffix “-t” on grammatical categorization, in the second experiment we only included nouns and verbs with various endings and eliminated overt inflectional morphology by using irregular past tense verbs. The (un)awareness of the primes was controlled in a prime visibility task in both studies.
RESULTS. In the first main experiment, we found a strong prime × category interaction, which was independent of masking (p < 0.001), and which was further confirmed by a congruency effect in both masked and unmasked conditions. Thus, the pronoun facilitated the processing of verbs, whereas the determiner facilitated the processing of nouns even when syntactic context was not consciously perceived. The second main experiment confirmed the prime × category interaction, together with a main effect of category — pointing towards late processing mechanisms at work by irregular verbal forms without overt inflectional morphology.
CONCLUSION. Overall, both experiments suggest that the pronoun and the determiner pre-activate an appropriate abstract syntactic category, facilitating the processing of upcoming nouns and verbs. The results indicate that abstract syntactic representations might be accessed unconsciously in an automatic fashion, thus providing empirical basis for future imaging studies focusing on the neural behavior of early structure building processing in the human brain.
References: Berkovitch & Dehaene, Cogn Psychol, 2019; Beyersmann et al., Psychon Bull Rev, 2011; Beyersmann et al., Psychon Bull Rev, 2016; Friederici, Physiol Rev, 2011; Hahne & Friederici, Journal of Cogn Neurosc, 1999; Lau et al., Brain and Language, 2006.
