統辭的 派生에 대하여 = On Syntactic Suffixation in Korean

다국어 초록 (Multilingual Abstract)

Obviously, suffixation is one of the morphological processes which makes lexical derivatives by attaching dependent forms to lexical bases. This means that morphological derivational suffixes should be attached to 'lexical' bases, and that the products of morphological derivation should be lexical units or other lexical bases. It is also self-evident that the semantic effects of morphological suffixes involve the semantics of the immediately preceding lexical bases. We accept these three conditions as defining characteristics of morphological derivation.
In Korean, however, we come to recognize the presence of some peculier suffix-like dependent forms showing different behaviors from typical morphological suffixes. Among such dependent forms are tul (a plural marker which performs the semantic function of individuating a generic entity), ssik (a marker adding distributive meaning to a plural entity), kkiri (a marker playing a semantic role of making a plural entity into some small inner closed circle), kkae (a marker denoting 'aroundness' after quantifying expressions), and a kind of -tap- (a derivative suffix which makes an adjective-like predication from NP and indicates that the subject has the quality of just described entity). These dependent forms especially verbal endings, prefinal endings, so-called case markers and delimiters, cannot be described as the ones participating in morphological processes. Firstly, they are not attached to lexical bases, but to syntactic constructions. Secondly, their products cannot rigidly be characterized as new lexical units or bases. Thirdly, their semantics sometimes does not operate on the immediately preceding elements, and sometimes the derivational results are not strictly independent words or lexical bases.
The procedure just described could be called 'syntactic derivation (SD)'. SD explains many strange phenomena involved in Korean syntax and sheds some light on linguistic typology. Especially, it is pointed out that the case marker in Korean should remain as a case marker after an NP. The case feature of an NP never percolates into the head noun, contrasted with the Indo-European cases. This means that, in Korean, cases are not abstract cases, not like the ones postulated in the Chomskyan school. A syntactic function carrier in nature determines the syntactic function of a syntactic construction, just as a morphological suffix determines the category of morphological derivatives concerned. The same is true for verbal endings and prefinal endings in Korean syntax. In addition, since the category of a syntactic construction is to be determined by syntactic derivational suffixes, which does not imply that the only means of determining the syntactic category is SD, X-bar theory plainly fails at least in Korean syntax. In the light of SD, the position that views the category of NPs with some postpositions (case markers) in Korean as PPs (postpositional phrases) also is not tenable. It is simply because that what determines the syntactic category of a construction is the nature of syntactic derivational suffixes.
How many and what syntactic categories should be set up in Korean syntax is an empirical question, which is to be determined by the consideration of syntactic function of suffixes and of explanatory convenience of syntactic descriptions. This process cannot operate without taking the semantic functions of the elements into consideration.