Compounds and Syntactic Phrases in Modern Irish

Abstract

1. Introduction (1) A subject which has attracted much attention in recent studies of morphology is manner in which formation of compounds should be handled. For a representative spectrum of opinions see Roeper and Siegel 1978, Selkirk 1982, Lieber 1983, Fabb 1984, Sproat 1985, Di Sciullo and Williams 1987, Borer 1988, Roeper 1988, Spencer 1991. starting point for this article is a set of data in Irish. (2) In traditional grammars and descriptions of this language, category compound is not recognised as such, but at same time there is a large group of expressions which bear all hallmarks of words rather than phrases. Briefly, problem is as follows. There exist in Irish phrasal groups consisting of a noun followed by another noun in genitive, e.g., mac Sheain John-Gen. -- John's son. We wish to claim that many of these are in fact compounds, on basis of criteria familiar from other studies of this sort, namely, lexical integrity, semantic idiosyncracy, and non-specificity. Having established existence of this category we turn to a more intriguing problem. This is existence of phrases which differ from compounds in that they exhibit a lower degree of lexical integrity and are specific rather than generic. At same time, they seem to behave more like words than syntactic phrases. An example is phrase fear an ti -- man house-Gen. -- the master, which looks very much like any other syntactic phrase, but on closer inspection turns out to possess a number of lexical properties. As we shall see, Irish data is very similar to Modern Hebrew analysed by Borer (1988). We try to point out parallels between two systems and ascribe them to properties of universal grammar. 2. Compounds 2.1 Identifying compounds Some authors of Irish grammars claim that compounding is not a productive word-formation process in this language. de Bhaldraithe (1953: 254) can serve as an example: for him a compound is an expression that resembles an English compound in that second member is head and whole is treated as a single phonological word, e.g. muic-fheoil [mik'o:l'] -- pigmeat -- pork, carn-fholt [karnolt] -- heap-hair -- heaped hair. This is an uncontroversial statement, and we will not be taking issue with it. de Bhaldraithe also mentions (254: footnote) phrases consisting of Noun+Genitive, which, he says, function as semantic units and correspond to compounds in other languages. It is not clear what status of these phrases is for him. We will argue that on criterion of lexical integrity these behave more like words than phrases and as such should be regarded as being produced by word-formation component. There is no obvious formal difference between syntactic phrases and what we claim are compounds. It is as if in English we only had phrases of sort of wood, master of school, son of John and had to decide whether they were syntactic combinations or lexical units. However, by taking two phrases consisting of Noun+Genitive from Irish and performing some elementary syntactic tests on them, it is possible to achieve at least a crude distinction between compounds and phrases. Below we examine two phrases, cos adhmaid -- wood-Gen. -- wooden-leg and seol an bhaid -- sail boat-Gen. -- the sail of boat. first proves resistant to any kind of syntactic operations. (1)a. One-substitution A. Ca bhfuil an chos adhmaid? is wood-Gen. Where is leg? B. *An ceann adhmaid? one wood-Gen. The one? b. Wh-movement A. Chonac an chos adhmaid I-saw wood-Gen. I saw leg B. *Cen chos? Which leg? c. Co-ordination *cos agus lamh adhmaid and wood-Gen. wooden and hand cf. …