1 Linguistic experience and productivity: corpus evidence for fine-grained distinctions

In this paper we show how qualitative and quantitative analyses of morphological productivity interact. A qualitative analysis alone does not match our intuitive notion of productivity. The quantitative analysis which must complement it, on the other hand, crucially depends on a detailed intensional understanding of the morphological process studied and its sub-processes. Morphological productivity (roughly: the readiness with which a word formation process forms new words) has long been one of the central mysteries of morphology. There are many detailed qualitative descriptions of word formation processes that list the restrictions for the possible bases for an affix (like the restriction that-able attaches to verbs in English) or the restrictions on a given compounding process (like the fact that noun-noun compounding in German is recursive while noun-verb compounding is not). Most of these works are not concerned with productivity or equate productivity simply with type frequency (see for example the standard descriptive works for German word formation Kühnhold, Putzer & Wellmann 1978, Fleischer & Barz 1992). Although we shall see below that qualitative studies of word formation processes are a necessary basis for all further research on productivity, they cannot account for our intuition that some word formation processes produce new words more easily or more frequently than others. How can we describe suc