In the same way that individual words may be acquired, collocations can be learned incidentally through repeated encounters in context (Webb et al., 2013), or by more explicit methods such as using concordancing tools. Graded readers can be used for both of these approaches, either through gaining exposure to vocabulary through extensive reading of graded readers (Horst, 2005) or concordancing using a graded corpus (Cobb, 2019). However, it is unclear whether graded readers offer sufficient exposure to frequent collocations to facilitate acquisition. Word frequency and structural complexity are taken into account when simplifying texts, but the effect of such grading on the phraseology of the language used has not been widely explored.
One corpus study of graded readers at two levels (Allan, 2016) showed that restricting grammatical structures and vocabulary range intensified lexical patterning in simplified texts and thereby led to a higher concentration of common lexical bundles. However, this was found to be optimal with a higher level corpus of graded readers (2,300 headwords), and diminished with more restricted readers (1,700 headwords).
The present study further explores the effects of text simplification, focusing on verb noun collocations. Three sets of texts are examined, consisting of an unabridged original version and corresponding versions graded to three levels of difficulty. Corpus software is used to identify the frequency and contextual use of the target collocations at the different levels. Tracing their use in this way can shed light on both the number of encounters offered and their transparency to the learner (e.g. through contiguous and non-contiguous use), features which are relevant both to incidental and explicit learning approaches.