Formulaic language: distribution, processing, and acquisition

Formulaic sequences are very frequently used in language as a preferred way to convey certain meanings. This thesis looks at distribution, processing, and incidental acquisition of formulaic sequences, by presenting four separate studies on different aspects of formulaicity. Study 1 investigated the distribution of four different categories of formulaic sequences (collocations, idiomatic phrases, lexical bundles, and phrasal verbs) and showed that those four categories vary considerably in terms of frequency. Also, register seems to affect the number of formulaic sequences used, as well as the categories of formulaic sequences preferred. Importantly, this study raised an issue of form variation of formulaic sequences (especially collocations) which seemed to be an under-researched area. Therefore, the following studies investigated the effect of form variation (focusing on non-adjacency) on collocation processing and their incidental acquisition. Studies 2 and 3 used an eye-tracking technique to investigate how native and non-native speakers of English process adjacent and non-adjacent verb-noun collocations. The results suggest that native speakers process both adjacent and non-adjacent collocations faster than matched control phrases, albeit the collocation effect seems to be larger for adjacent collocations. As for non-native speakers, there is a clear collocation effect for adjacent collocations and it is moderated by prior vocabulary knowledge. However, there seems to be almost no effect for non-adjacent collocations. This finding suggests that even advanced non-native speakers process non-adjacent collocations differently than native speakers. Finally, Study 4 tried to take the findings from the previous studies to a classroom. It investigated whether there is any difference between the chances of incidentally acquiring adjacent and non-adjacent collocations from reading. The results suggest low but durable gains for both adjacent and non-adjacent collocations, with no significant differences between these two groups of items. Overall, the results presented in the thesis support the idea that formulaic language is ubiquitous, but suggest that some of the criteria that have been widely used for identifying formulaic sequences might need to be reconsidered. It seems that collocations, at least for native speakers, retain their formulaic status even when presented non-adjacently. They seem to be successfully learned as non-adjacent dependencies as well. While this finding cannot be easily generalized to other types of formulaic sequences, it seems to suggest that a lot more research on form variation of formulaic sequences is needed in order to better understand the scope of the phenomenon