engleski

Arguments and adjuncts in synthetic compounds – evidence from English and Croatian

Within the framework of Construction Morphology (CM) (Booij 2010a, 2010b, 2015, 2018 ; Gaeta 2010, Gaeta and Zeldes 2017) and Relational Morphology (Jackendoff and Audring 2016, 2018), creation of new lexemes is licensed by word-level constructions, i.e. morphological schemas, which perform a dual function in the mental lexicon – they connect related complex words and serve as a basis for formation of new words. One of the key properties of schemas is that they can be unified with one another, which may result in increased productivity of the resulting schema. The aim of this paper is to examine the formation of deverbal synthetic compounds (SCs) in English and Croatian using corpus data and native speaker intuitions. Specifically, I aim to look at the interaction of two issues related to the formation of deverbal SCs that have been raised in previous works dealing with SCs – the order of morphological operations in the creation of [NV-er]N compounds, and the range of possible thematic relations between the deverbal head and the nominal complement. Traditionally, two logical possibilities have been established for the order of operations: i) verbal NV compound is created first, followed by -er suffixation (Lieber 1983, Ackema and Neeleman 2004) ; and ii) agentive deverbal noun is formed first, followed by incorporation of the nominal complement (Selkirk 1982). A third option has recently been put forth by Booij (2010), who claims that formation of SCs in English involves unification of two independent word-formation schemas – the unproductive [NV]V schema for verbal compounds, and the highly productive [Ver]N schema for agentive nouns. This approach resolves the shortcomings of both approaches by placing the burden of formation on schemas, thus nullifying the need for accounting for the intermediate steps in the formation. With regards to this issue, my goal is to use corpus data to test the justifiability of this proposal. The second issue regarding the types of thematic relations has been present since the earliest works on SCs in the generative tradition (cf. Roeper and Siegel 1978). Whereas some authors adopt a restrictive approach and allow only internal arguments (in the sense of Williams 1981) as complements (Selkirk 1982, Grimshaw 1990, Oshita 1994, and more recently Gaeta & Zeldes 2017), others allow the so-called semantic arguments to take part in the formation of SCs (cf. Lieber 1983, Miller 2014). The goal of my paper here will be to examine the hypothesis that SCs can also take non-internal arguments as complements and show certain empirical inadequacies of the restrictive approaches that dismiss cases with these types of arguments as root compounds.

arguments, adjuncts, synthetic compounds, unified schemas, thematic relations

nije evidentirano