This publication is in the INCUBATOR section.
Roland Schäfer & Elizabeth Pankratz (in preparation) Function vs. interpretability of compound linking elements in German. To be submitted in Q4/2017.
In this paper, we take a closer theoretical and empirical look at those linking elements in German N1+N2 compounds which are identical to the plural marker of N1 (such as -er as in Häus-er-meer ‘sea of houses’). Often, the linking element is optional (e.g., Haus-tür ‘front door’). We discuss the theoretical positions w.r.t. these elements, wherein some authors claim that such linking elements do not have a grammatical function nor a semantic interpretation (as most linking elements in German), while others consider a plural meaning to be available. Our own study is based on a huge database (automatically extracted, but manually checked) database of German N+N compounds occurring in the DECOW16A corpus. We use the database to create a virtually exhaustive list of N1s occurring with a pluralic linking element and without a linking element, including extensive frequency analyses and analyses of the N1s’ productivity. Based on these lists, we show that there are conditions under which the pluralic linking element is indeed preferred, and that these conditions are highly plausible for requiring a plural interpretation of N1. In a split-100 experiment, we show how speakers’ intuitions are very much in line with the corpus findings. Our interpretation does not reference a specific morphological framework. Instead, we view our data as being in line with the general approach of Bresnan’s probabilistic grammar. Furthermore, in the spirit of Haspelmath’s framework-free grammar, we discuss the implications of our results for any morphological framework or theory as well as the distinction between function and interpretability.