A comparative study of compound words in English, Japanese and mainland Scandanavian
The aim of this thesis is to propose a structure for compounds, specifically compound nouns in Japanese, English and Mainland Scandinavian within the framework of Chomsky's Minimalist Program and Bare Phrase Structure (Chomsky 1995). The purpose is to show that words are derived in Narrow Syntax as phrases and that words must have asymmetrical structure, i. e. a head of the word should be determined. The proposed structure of a compound noun in the languages in question is as follows: (1) P(X) root P(x) root P(x) Structure (1) is derived with the following assumptions in mind. 1. The place of Morphology within the Minimalist Program is argued to be outside the Lexicon and after the Narrow Syntax. This has led several linguists to argue that a word is derived in the same way as a phrase. Moreover, linear order is redundant in the Narrow Syntax, since the structure determines the word order. As a result, it is not the Right-hand Head Rule proposed by Williams (1981) which determines the head of a compound word but the structure does. The Right-hand Head Rule may have a place in the phonology, though, in stipulating how a word derived in the Narrow Syntax is spelled out. The rule is formulated by Williams to apply in Morphology. In most current minimalist theories morphology is after spell-out. But the head must be determined before spell-out, since it determines the LF as well as determining aspects of the PF. 2. Nothing prevents us applying Merge at the level of the word as well as the phrasal level. As Williams' (1981) Right-hand Head Rule cannot be used within the Minimalist Program, Collins (2002) definition of head is used for compound words. According to Collins, a head is a category which has one or more unsaturated features. Another stipulation taken from Collins (2202) is that when a lexical item is chosen from the lexical array and introduced to the derivation, the unsaturated features of this lexical item must be satisfied before any new unsaturated lexical items are chosen from the lexical array. The effect of these two assumptions is that when two categories a and ß are merged, only one of them, say a, can have an unsaturated feature (which is not saturated by ß), so a will be the head. The structure (1) shows the following. " First, a root without word class features is merged with a Property feature, the content of which is given by the root. " The Property feature is represented above as P(roperty) (x) where `x' represents the unvalued referential index. " There are two ways to check P(x): one is assigning xa value, that is an index, and the other is deleting x. Since the P(x) feature is unsaturated in the sense that it needs a referential index from either D or DP, it is a head, and as such it percolates to the dominating node. Then, another root is merged to form a compound word. As P(x) is the only unsaturated feature before and/or after the root is merged, it is percolated and it is the head of the whole compound. The present theory can account for the syntactic and semantic properties of a wide range of compounds, particularly noun-noun compounds in English, Japanese, and Mainland Scandinavian, within a syntactic theory based on minimalist assumptions.