In this article, we investigate the peculiar distribution of the auxiliaries have and be in Earlier English and its consequences for theories of the perfect and auxiliary selection. We argue on the basis of a large-scale corpus study that the periphrastic construction with be was restricted to a stative resultative interpretation, whereas that with have developed a wider range of uses, crucially including the experiential perfect in addition to resultatives. Support comes from comparing the Earlier English patterns with related ones in Norwegian and German for which native-speaker judgments are available. On the basis of this insight, we propose distinct formal analyses for the two constructions and show how they account for the attested patterns and changes in Middle and Early Modern English. Of particular theoretical relevance is the premise that what has been called the “perfect” is not a homogeneous, monolithic category, and that certain kinds of variation can only be understood by teasing apart the pieces involved. Earlier English and German auxiliaries have distinct distributions because their “perfects” have disinct syntactic and semantic makeups.
Stored only when both papers are in the corpus — 101 citations exist upstream.