Johannes Kizach (Aarhus): Complexity, givenness and g-maze

When we produce sentences we encode meaning by means of a syntactic structure, and when we interpret sentences, we decode meaning via the syntactic structure. This suggests that the syntactic structure is central in sentence processing, and indeed, it has been argued that the complexity (i.e. the length) of the phrases is crucial for processing: Short-before-long order ensures rapid recognition of the sentential structure (who does what to whom), as compared to long-before-short orders which delays the understanding (in head first languages such as English or Danish). Phrasal length is, in this perspective, always relevant for sentence processing: For all sentences it is important that processing is as rapid and effortless as possible.
Other factors which have been argued to be crucial for sentence processing and word order choice, such as givenness and animacy, do not have this property of being relevant in all cases. In fact, givenness is only relevant for the processor in cases where two referential DPs that differ in terms of givenness, need to be ordered. The same goes for animacy which is relevant only when two referential DPs differing with regard to animacy, need to be ordered.
The potential benefit of heeding complexity (having short-before-long wherever possible) is therefore much larger than the potential benefit of heeding e.g. givenness. So if the human processer is clever, it should care more about complexity than about givenness.
In this talk, I will present results from corpus studies and experiments on word order and show how these results support the idea that length/complexity is more fundamental for the processor than other factors seem to be.