Holden Härtl Semantic non-transparency in the mental lexicon: On the relation between word-formation and naming  1 Introduction The notion of Eigentlichkeit (‘authenticity’), when approached from a perspec- tive of language philosophy, is associated with the quality of linguistic entities of referring to things in the world in a truthful and a maximally transparent way, cf. Gardt (1995). Obviously, there are numerous expressions that do not seem to conduct themselves accordingly and display a rather “non-authentic”, i.e., non-transparent behavior instead, like indirect and ironic speech acts or idiomatic expressions like Cat got your tongue? Not unexpectedly, though, such expressions – despite their apparent lack of transparency – have merits in their own right. These can often be explained on grounds of Grice’s conversa- tional maxims, see Grice (1975). For example, indirect speech acts, as is known, come to the benefit of adhering to the politeness maxim and a certain metaphorical characteristic is added in the denotation of idiomatic expressions. So, we can hypothesize a systematic correlation to be functioning between au- thenticity, on the one hand, and expressivity, on the other. In the current paper, the interplay between these two factors is investigated in the domain of word-formation. In particular, I will focus on aspects of se- mantic compositionality and the supposed naming function of novel com- pounds and how these factors relate to the perceived novelty or “markedness” of novel compounds. For the analysis, we will concentrate on novel adjective- noun compounds (Blauschachtel ‘blue_box’, Schmalmesser ‘slim_knife’ etc.) with occasional glances at noun-noun compounds where relevant. We will start from a lexicalist perspective and the conventional assumption that word-for- mation is the preferred route for establishing a concept’s name in German, put this assumption to several tests and prove it to be correct. In the second part of the paper, I will argue that a systematic relation holds between the markedness of an A-N compound and its interpretation as a kind name as well as its affinity to be lexicalized. Specifically, this relationship will be traced back to a prag- matic principle, which holds that deviance from a conventionalized form (that  Acknowledgements: I am grateful to Sven Kotowski and Martin Schäfer as well as the audiences at the Linguistic Lunch Hour series (UW-Madison) and the workshop “Usage-based approaches to morphology” (DGfS conference 2013, Potsdam) for comments and fruitful discussion and also to the editors and Kim-Vivien Lichtlein for their support in the preparation of the manuscript. To appear in a special volume on transparency in language and communication. Berlin / New York: de Gruyter.