Pronominal adverbs in German, which consist of da ‘there’, hier ‘here’, or wo ‘where’ as first element and a preposition as second element, like davor ‘before’, hierbei ‘hereby’, worin ‘wherein’, have often been explained by a movement of the first element out of the complement position of the preposition. This article points out some of the problems of movement analyses and presents an alternative account based on the diachronic development of pronominal adverbs. It is argued that the pattern after which pronominal adverbs are formed can be traced back to the univerbation of two adverbs with spatial meaning. This is accompanied by processes often associated with grammaticalization, such as semantic bleaching, phonological reduction, and a loss of separability in the standard variety. Some of the reduced forms are obligatory in phrasemes and particle verbs, thus constituting a split which can occur during grammaticalization. The reduction of the first element of pronominal adverbs and a doubling of the first element can be seen as part of a grammaticalization cycle.

In Standard Yiddish, -s and -ən are used as default allomorphs for plural word formation. It is argued here that the choice is left to the phonology, with -s acting as a default within a default. This status is used to explain the exclusive use of -s in the pluralization of proper names, which are claimed to be formed with no sensitivity to the phonological form of the base.

A subset of German control verbs allows for the discontinuous linearization of their infinitival complements, a word-order pattern known as the “third construction” pattern. Compared to alternative word-order options (notably, extraposition), third constructions are very rare in present-day German. Here we ask whether the third construction pattern’s low occurrence frequency can be accounted for by processing factors. We report the results from a self-paced reading task and a production priming task investigating whether third constructions are difficult to comprehend, difficult to produce, or both. Our results show that the third construction pattern’s local structural ambiguity impedes comprehension, and that the pattern is also resistant to priming. We conclude that this word-order pattern is an example of a “latent” construction that is grammatically licensed but strongly dispreferred in language use because easier-to-process word-order variants are available.