The Morphosemantic Transparency/Opacity of Novel English Analogical Compounds and Compound Families

This study deals with novel English analogical compounds, i.e. compounds obtained via either a unique model (e.g. beefcake after cheesecake) or a schema model: e.g., green-collar based on white-collar, blue-collar, pink-collar, and other X-collar compounds. The study aims, first, to inspect whether novel analogical compounds maintain the same degree of morphosemantic transparency/opacity as their models, and, second, to find out the role played by the compound constituents in the constitution of compound families, such as X-collar and others. To these aims, the study proposes a scale of morphosemantic transparency/opacity for the analysis of compound constituents. In particular, the compound constituents in our database (115 examples) are analysed in connection with: 1) their degree of transparency (vs. opacity, including metaphorical/metonymic meaning), linked to their semantic contribution in the construction of the whole compound’s meaning, and 2) their part-of-speech. Against the common assumption that productive word-formation rules mostly create morphosemantically transparent new words, or that rule productivity is closely connected with transparency, the study of our database demonstrates that novel analogical compounds tend to maintain the same transparency/opacity degree as their models. It also shows that, in nuclear families and subfamilies of compounds, the part-of-speech of the constituents, their degree of transparency/opacity, and their semantic relation are reproduced in all members of the analogical set.