Unified schemas which allow for deriving multiply complex word-formation products are a central concept in Construction Morphology (CxM). Based on examples such as un-V-able formations in English, Booij (2007: 38) argues that unified schemas (in this case: [un[Vable]A]A) can be conceived of as "short cuts in coining new complex words". In this paper, we explore three prospective cases of schema unification and discuss what kind of evidence supports the assumption of unified schemas. The first two case studies are diachronic in nature. Drawing on corpus analyses of data from the Early New High German period (1350−1650) and from the early stages of New High German, we show how the developments of the complex patterns diverge from the developments of their counterparts. To this end, we analyze the frequency and productivity of the (sub-)constructions and assess the semantics of the word-formation products. Firstly, nominalization with the suffix -ung has been shown to undergo a diachronic decrease in morphological productivity (Demske 2000; Hartmann 2016). However, unified schemas such as [Be-X-ung]N or [(PREF)-X-ierung]N are shown to be still productive, e.g. Beplankung, Belaberung, Vercomedianisierung (from www.wortwarte.de, a collection of neologisms). In a similar vein, complex derivation of the type [un-V-lich]ADJ 'un-V-able' is shown to have remained productive for a longer period of time than its simplex parent schema [un-V-lich]ADJ. Moreover, many un-V-lich derivatives historically precede their unprefixed counterparts, or lack them altogether (unwiderstehlich 'irresistible', but *widerstehlich). Our third case study explores present day German pseudoparticiples (bebrillt 'bespectacled') using web data. The complex pattern can be shown to diverge stylistically from its parents schemas and also to provide semantically more uniform derivatives. Overall, our results show that the concept of unified schemas can help explain important differences in the development of the individual subpatterns in terms of morphological productivity and in terms of semantic aspects of the word-formation constructions.

Extravagance can be conceived of as a (more or less) deliberate deviation from established norms that evokes surprise or attention. In this paper, we present an empirical case study of a rare but quite remarkable morphological pattern, namely German pseudo-participles – forms that look like past participles but lack a verbal counterpart, e.g., be-brill-t ‘be-glassed’ (derived from the noun Brille ‘glasses’). In our study, we investigate the extravagance of the pattern using data from the webcorpus DECOW16A. We combine a detailed qualitative analysis of singleton coinages with various explorative quantitative approaches, taking into account the complexity of the formations and their anchoring in the context. In addition, we examine metalinguistic indicators of extravagance such as double quotes and meta-comments. We argue that German pseudo-participles are a prime example of an “extravagant” phenomenon at the interface of morphology and pragmatics, and as such can be instructive for our understanding of extravagant morphology on a general level, especially with regard to the role of surprise in generating extravagant effects.

The ellipsis of parts of words (be- und entladen ›load and unload‹) is fairly common in contemporary German. This article takes a diachronic perspective, thus revealing how the phenomenon was even more frequent around 1700. Many of the examples are no longer grammatical, e. g. schwest- und brüderlich ›sist- and brotherly‹. It is shown how the phenomenon originated in two distinct domains: In compounds, the ellipsis arose in the 16th century from constructions like frouwen gelust und list ›women's lust and Ø cunning‹. In the domain of inflection, the earlier data cannot readily be analyzed as ellipsis. The two conjuncts (von feindt vnd freunden ›of enemyØ and friends‹) could be perceived as a unit, thus inflected only once. The crucial step was the merge of the two domains in the early 17th century. This event is likely to have caused the extensive type and token frequency of the phenomenon around 1700.

No abstract

No abstract