Variations on what for in the history of English

Constructions that question kinds have been discussed widely in Germanic, Slavic, and Baltic, but English and earlier stages of English have mostly been overlooked. This paper presents data on kind questioning constructions in English, followed by an analysis of these constructions. The Old English what followed by a partitive genitive noun, labeled the ‘what partitive construction’ (WPC), questions the kind and, unlike in other languages, it also questions the number of the noun. I analyze the WPC as including an unpronounced noun that heads the DP, as is evident from the agreement on the verb. In Middle English, an overt light noun (e.g., kind and number) replaces the WPC which initially functions as the head of the entire DP but later grammaticalizes into a higher functional category. There is also a stage with a ‘what for construction’ (WFC) but, unlike in the history of German, the English WFC appears a few centuries after the loss of the WPC, seems unrelated to the WPC, and is not long-lived. The analysis of the English WFC involves a PredP, adapting ideas from earlier proposals. The main contribution of the paper is to present a new set of data and a synchronic analysis for them.