The processing and representation of multi-word expressions (MWEs), ranging from noun compounds (e.g. hogwash in English, and Ohrwurm in German) to particle verbs (e.g. give up in English, and aufgeben in German) has remained an unsettled issue over the past 20+ years.
From a psycholinguistic perspective the question is how the semantic transparency of the constituents affects the processing and representation of the MWE they compose (Libben, 2006; Gagne and Spalding, 2009; Ji et al., 2011; Marelli and Luzzatti, 2012). For example, the semantic transparency of the head has been found to affect the processing of noun compounds in English but not in German, with similar findings contrasting the effects of semantic transparency in English and German particle verbs (Smolka et al., 2014).
From a computational perspective the question is how the different types of constituents (i.e., modifiers vs. heads, particles vs. verb stems) influence the automatic prediction of semantic transparency, as typically addressed by vector space models relying on the distributional hypothesis and empirical co-occurrence information from large corpora (Reddy et al., 2011; Bell and Schäfer, 2013; Salehi and Cook, 2013; Schulte im Walde et al., 2013; 2016).
In this workshop, we aim to exploit complementary evidence from the two very different types of MWE (noun compounds and particle verbs) to shed light on the interaction of constituent properties and compound transparency. We invite research contributions across languages and across research disciplines to provide a cross-linguistic perspective integrating linguistic, psycholinguistic, corpus-based and computational studies.
Relevant aspects include (but are not restricted to)
Sabine Schulte im Walde,