Standard items for English judgment studies: Syntax and semantics

The use of experimental methods in grammar research has gone from strength to strength and has established itself as one of the key ways to investigate linguistic patterning among words, phrases, and clauses up to the sentence level. This is strongly to be welcomed: many linguists have a feeling of unease about the thin ice of weak validity that work in syntactic and semantic theory sometimes skates upon when it is done without reasonable attention to its evidential base. In particular, if linguists can radically disagree about the underlying architecture of the grammar that they are attempting to describe without it being clear who is wrong and who is right, then this is an unmistakable sign that the data basis used is insufficient (either in quantity or quality or both) to uniquely determine the system to be described. With Popper, we can doubt that unfalsifiable claims are any scientific claims at all. In this paper we take the view that both hypothesis building and hypothesis testing can be improved by the use of more fine-grained data and the use of multiple lexical variants of structures. If linguists employ data sets with proper control of potential confounding factors then the range of analyses they will propose will be more constrained. But it is especially important that the data set permits sufficiently sharp descriptions and predictions to allow clear falsification of hypotheses. Another issue which is perceived to be problematic is the non-independence of the data source. When linguists give their own judgements and base their theory