Abstract
The present study aimed at examining how semantically transparent and opaque compound verbs are accessed during language processing. In terms of long-term memory representations, transparent compound verbs (e.g. kaado-daadan ‘to give a gift’) that are compositional in their meaning are likely to be stored via their constituents, whereas opaque compound verbs (e.g. rokh-daadan ‘to take place’, literally ‘face-to give’) that are idiomatic must be stored as whole words. The study investigated how these two types of Persian noun-verb compound verbs are accessed during processing using an unmasked (Exp 1) and a masked (Exp. 2) priming technique. In priming experiments, the priming effect of the prime on the acceleration of identification of the target is compared to that of an unrelated control prime. In both experiments, a lexical decision task was used, where the participants are to decide as quickly and as accurately as possible whether a visually presented string of letters is a word in Persian or not. Forty-six native speakers of Persian participated in each experiment. Experiments 1 & 2 used the same 40 sets of items and followed the same procedure except for the prime duration that was 527 milliseconds in Exp.1 and 50 milliseconds in Exp.2. The results of the unmasked experiment (Exp 1) are summarized below (additional results will be discussed in the presentation).
In the unmasked experiment:
- Priming effect for the nominal constituent of opaque compound verbs
- Priming effect for both the nominal and the verbal constituents of the transparent compound verbs
In the masked experiment:
- Priming effect for the nominal constituent of opaque compound verbs
- No priming effect for the constituents of the transparent compound verbs
This means that in the unmasked experiment, where there is enough time for conscious processing, in the transparent condition, both noun and verb targets were recognized faster than in the unrelated condition supporting the idea that the meaning of transparent compound verbs is derived from the meanings of their constituents. In the opaque condition, there was facilitation in recognizing noun targets, suggesting that despite their non-compositionality opaque compound verbs are decomposed into their constituents during lexical access. [The fact that only the noun but not verb targets were facilitated is compatible with the results from aphasia studies (Nilipour, 2000; Semenza et al., 1997; Mondini et al., 2004)]. The findings suggest that both decompositional and whole-word routes are available to the parser and support a dual-access model of processing.
Posted in