Individual Differences in the Perception of Entrenchment of Multiword Units

While for a long time words and grammatical rules were regarded as the basic units in language, it has become increasingly clear that we have a much more varied set of units at our disposal, including multiword chunks. How such chunks are represented, which factors contribute to their entrenchment as units and to what extent this differs from person to person are questions that require investigation. In this paper we report on the results of a magnitude estimation task investigating variation in entrenchment, both between units and across individuals. Interesting discrepancies were found between corpus frequencies and the participants' assessments of unit status, as well as substantial differences between individuals. These results are important with respect to usage-based theories of language, as they contribute to a more precise account of basic assumptions regarding the relative importance of the factor frequency in shaping linguistic representations and in giving rise to variation across speakers.