Summarization is a complex cognitive task and its simulation is very difficult for machines. This paper presents an automatic summarization strategy that is based on a discourse categorization of the textual information. This categorization is carried out by the automatic identification of discourse markers in texts. We defend here the use of discourse methods in automatic summarization. Two evaluations of the summarization strategy are presented. The summaries produced by our strategy are evaluated with summaries produced by humans and other applications. These two evaluations display well the capacity of our application, based on EXCOM, to produce summaries comparable to the summaries of other applications.

Since countability is a grammatical notion, the distinction between count and mass nouns may not reflect countability in the real world. Based on this, Chierchia (1998a; 1998b) provides a typological study of plurality and genericity, which does not account for countability in Korean. Nemoto (2005) revises Chierchia's analysis to deal with count and mass nouns in Korean and Japanese. This study discusses problems with the previous analyses and proposes that the semantic feature of humanness is the main criterion for countability in Korean.

This paper investigates ordering alternation phenomena in Korean using the dative construction data from Sejong Corpus of Modern Korean (Kim, 2000). The paper first shows that syntactic weight and information structure are distinct and independent factors that influence word order in Korean. Moreover, it reveals that heaviness and givenness compete each other and exert diverging effects on word order, which contrasts the converging effects of these factors shown in word orders of right-branching languages like English. The typological variation of syntactic weight effect poses interesting theoretical and empirical questions, which are discussed in relation to processing efficiency in ordering.

English floating quantifiers (FQ) are both limited and complex in the sense that they are introduced by a limited set of words, all, both, and each, and display free distributional possibilities. This paper provides a non-movement approach to the syntax of English floating quantifier constructions. The non-movement analysis we develop here is different from stranding movement analyses in that all the FQs are base-generated while the linkage with their antecedent refers to grammatical features such as SUBJ and PRD. The analysis avoids the postulation of abstract levels as well as empty elements in capturing the flexibility of English FQ constructions, making the grammar of English simpler.

Geis (1973) has provided various properties of the subjects and by + Gerund Phrase (GerP) in English causative constructions. Among them, the two main issues of Geis's analysis are as follows: unlike Lakoff (1965; 1966), the subject of English causative constructions, including causative-inchoative verbs such as liquefy, first of all, should be acts or events, not persons, and the by + GerP in the construction is a complement of the causative verbs. In addition to these issues, Geis has provided various data exhibiting other idiosyncratic properties and proposed some transformational rules such as the Agent Creation Rule and rule orderings to explain them. Against Geis's claim, I propose that English causative verbs require either Proper nouns or GerP subjects and that the by + GerP in the constructions as a Verbal Modifier needs Gerunds, whose understood Affective-agent subject is identical to the subject of causative verbs with respect to the semantic index value. This enables us to solve the two main issues. At the same time, the other properties Geis mentioned also can be easily accounted for in Head-driven Phrase Structure Grammar (HPSG) by positing a few lexical constraints. On this basis, it is shown that given the few lexical constraints and existing grammatical tools in HPSG, the constraint-based analysis proposed here gives a simpler explanation of the properties of English causative constructions provided by Geis without transformational rules and rule orderings.

We examine attributive possessive constructions in Korean in examples like wuli manwula 'my wife' and wuli nala 'my country', where an apparently plural pronoun wuli 'our' is used to convey the singular meaning of 'my.' An example like wuli manwula has been a frequent topic of joke, because it clearly contains the literal sense of 'shared wife.' This type of use and the frequent use of wuli in an exclusive my context led many to claim that these are products of Korean culture which emphasizes groups and group solidarity over individuality, the example wuli manwula being the extreme case. Drawing on the work of Taylor (1989; 1996), Croft (2001; 2003), Haspelmath (1997; 2003), and 윤재학 (2003), we compare the meanings of Korean and English possessive constructions and seek a more linguistically based account for the data. In particular, it is shown that (i) wuli in question is not really the plural form of nay 'my', (ii) the possessive constructions in Korean are more conservative and limited in possible relations than the English counterparts are, and that (iii) these two facts interact to produce the afore-mentioned uses.

Tough-construction is, undoubtedly, one of the most puzzling syntactic problems in the history of transformational grammar. Various approaches have been proposed including Deletion approaches (Akmajian, 1972; Lasnik and Fiengo, 1974) and movement approaches. Among the movement approaches, Chomsky (1977; 1981) argues for movement of null Operator, and Hornstein (2001) argues for a two-step movement equipped with Sideward movement. Most recently, Schueler (2004) and Hartman (2009) each have also argued for a kind of movement approach. With the development of the Minimalist syntax (Chomsky, 1995; Chomsky, 2000; Chomsky, 2001), tough-construction, an age old problem in the description of grammar, turns into another round toward to a more satisfactory answer. By examining the most recent competing analyses of tough-constructions, this paper defends and extends Schueler's (2004) analysis, rather than Hartman's (2009) two step movement approach. Furthermore, this paper proposes that tough-subject originates from the intermediate CP internal Spec-TopicP position rather than from the iterated CP layer (Authier, 1992). This approach has more descriptive power than it was originally argued for in Schueler (2004) and is a step closer toward Minimalism insofar as the conception of government is no longer utilized.