Evaluation of multilingual translation fundamentally involves measurement of meaning equivalences between the formally mapped discourses/texts of SL(source language) and TL(target language) both represented by a metalanguage called IL(interlingua). Unlike a usaal uni-directional MT(machine translation) model(e.g.:SL $\rightarrow$ analysis $\rightarrow$ transfer $\rightarrow$ generation $\rightarrow$ TL), a bi-directional(by 'negotiation') model(i.e.: SL $\rightarrow$ IL/S $\leftrightarrow$ IL $\leftrightarrow$ IL/T \leftarrow TL) is proposed here for the purpose of evaluating multilingual, not merely bilingual, translation. The IL, as conceived of in this study, is an English-based predicate logic represented in the framework of MRS(minimal recursion semantics), an MT-oriented off-shoot of HPSG(Head-driven Phrase Structure Grammar). In addition, a list of semantic and pragmatic checkpoints are set up, some being optional depending on the kind and use of the translation, so sa to have the evaluation of translation fine-grained by computing matching or mismatching of such checkpoints.

This paper proposes a semantics of the Korean NPI te isang 'anymore', which occurs in negative contexts. The NPI te isang is characterized, in Zwarts's(1990) term, as a strong NPI which requires an anti-additive function as its licensor. It is also noted that te isang is licensed by 'implied negative predicates' such as shilheha-'hate' and phokiha-'give up' only when it occurs within a clausal argument of the implied negative predicates. This fact is accounted for in terms of the function-argument structure and the scope relation between the NPI te isang and the implied negative predicates involved.

This paper aims to describe a German-Korean machine translation system for nominal phrases. Besides, we have two subgoals. First, we are going to revea linguistic differences between two languages and propose a language-informational method fo overcome the differences. The method is based on an integrated model of translation knowledge, efficient information structure, and concordance selection. Then, we will show the statistical results about translation experiment and its evaluation as an evidence for the adequacy of our linguistic method and translation system itself.

It has been suggested in the linguistics literature that quantification and distributivity are closely related phenomena and some linguists claimed that distributivity should be marked on quantifiers(determiners) as well as on predicates. I would challenge the claim that quantificational determiners should be classified in terms of distributivity. I suggest, on the empirical grounds, that distributivity is essentially a phenomenon over(one-place) predicates in a broad sense.

This paper is an HPSG approach to agreement phenomena involving the Korean honorific expressions. it is shown that the theoretical devices developed by the constraint-based theory of HPSG can be fruitfully used to capture the interactions between syntactic constraints and semantic of pragmatic factors in Korean honorific agreement. The HPSG's semantic feature 'referential index' plays a key rele in discribing the multiple interaction. The constraint-based theory of agreement proves successful in accounting for the phenomenon that may be called 'inconsistent' honorific agreement as well as 'consistent' regular honorific usages. However, this paper acknowledges its limit. Recognizing an important distinction between basic and 'coercive' honorific expressions, it is argued that a systactic-semantic-pragmatic approach such as the present one can only be applied to basic honorific agreement. Being sociolinguistic in nature, coercive honorific agreement is perhaps not amenable to formal linguistic investigation.

This paper claims that beneficiary adjuncts are best analyzed as involving external predication in a version of grammatical framework called Head-driven Phrase Structure Grammar. This paper also claims that verbal catefories need to include the attribute INDEX among their semantic components in order to account for the external predication proposed in this paper. This paper distinguishes between recipient and beneficiary reles and assumes that the former is a semantic argument of a verb-type relation and that the latter is an adjunct which makes a semantic contribution as a modifier. This approach achives a unified analysis of modification phenomena of nominal and verbal categories and it can also accomodate Parson's(1990) idea that a verbal category denotes a set of events, not just an event.

conflicting criterial used in identifying words have called the lexical integrity principle into question. That is, cases where the morphological word does not coincide with the syntactic word have notivated the syntactic view of word derivation, as pointed out by Bresnan and Mchombo(1995). Further, the implicit desire to make the clausal structure of Korean parallel to those posited for English(Chomsky 1991) and French(Pollock 1989) has also led most of the current literature on Korean morphology to claim that Korean verbal inflections head their own functional projections such as AgrP, TP, and MP im syntax. In this paper, I will first argue against such a syntactic view. After reviewing some basic properties of Korean verbal inflections, I will show that the evidence from mismatch phenomena supports the lexical integrity principle over the head-movement theories of word derivation. Then, I will propose a theory of lexical grammar which maintains the lexical integrity principle while retaining the effects of functional projections and syntactic movement.

Close examination of previous constraints on bound anaphora that are designed to directly constrain the distribution of referentially dependent(RD) items shows that no universal structural relation may exist that relates RD items to their antecedents. As an alternative to these constraints, this paper proposes an axiom of semantic interpretation, called the Principle of Referential Autonomy, which dispenses with any pretheoretical notion of grammatical functions or configurational notion like c-command. Together with certain English-specific facts, this principle enables us to infer the ungrammaticality of core examples of strong and weak crossover.