Earticle

This paper discusses the main points of In"]Seok Yang"fs ground"]breaking work, Korean Syntax(1972), which can be viewed as an monumental work in Korean syntax within the early Chomskyan framework, and evaluates its significance within the recent minimalist theory. As it turns out, Yang"fs arguments about Multiple Case Constructions can be re"]illuminated as a cartographic structures of Specs within IP, similar to the cartographic structures recently suggested by Rizzi(1997, 2004). Other important aspects of Yang"fs theory concern sentential complementation constructions, in which he argued for clausal reduction and verbal compounding, and some of which provide significant dilemma in modern GB or minimalist frameworks. Some other Korean constructions involving so"]called Equi"]subject constructions offer other dilemmas, but some of them may in fact be shown to support little v analysis within recent minimalist framework.

This paper deals with the causativization of the verb 'run'. In section 2, I review Pinker's(1989) lexical rules on the causativization of the English verbs. Then some problems and disadvantages of his lexical rules are pointed out. In section 3, I will discuss the lexical semantic structures and the causativization and logical polysemy of the verb 'run"e in terms of Pustejovsky's(1995) Generative Lexicon Theory and the co-composition generative mechanism. And then some problems of Generative Lexicon Theory are pointed out. In section 4, I review the semantic network, profile shift, metaphor in Cognitive Grammar. Then I argue that the generative mechanism of Cognitive Grammar is more appropriate in dealing with the causativization of the verb 'run'.

Recently in the realm of generative grammar, some syntactitians including O'Neil (1995) and Hornstein (1999) among others have suggested the movement theory of control (henceforth MTC in short), which has established a diagnosis of obligatory control. Though the distribution of PRO in Korean has hardly been subject to proper consideration. the diagnosis is expected to enable us to judge the status of controlled subject. However, the fact that the diagnosis is helpful does not necessarily means that the MTC is doubtlessly correct. We can use the diagnosis provided by the MTC but deny its explanatory power on the basis of some conceptual problems to be pointed out in this paper. The main focus of this paper is put on these conceptual problems. In particular, it presents arguments against the recent movement-based approach to PRO, but in favor of case-theoretic account. In order to explore the nature of null subject in Korean, it is necessary to establish a definition that can be used to determine the possibility of the occurrence of PRO or pro as subject.

Kim, Joung Ran(2010) argued that the meaning of Korean topic and conditional sentences can be appropriately represented by the set inclusion relation. This article explores if the semantic structure of the set inclusion relation has the syntactic basis. It is first shown that the topic phrase and the conditional clause consistently have wider semantic scope than the negation in the sentence they appear. This fact suggests that they appear in the left branch occupying the higher position in the tree than the negation which is embedded in the right branch as the sentence form a bipartite structure. The Hua data reported by Haiman (1978) also suggest the surface bipartite structure for topic and conditional sentences in Hua. Then the author concludes that the set inclusion relation is derived from the syntactic bipartite structure, which is in turn translated into the set inclusion relation during the semantic processing. It is also argued that the Korean particle -nun seems to be the marker for the surface bipartite structure.

Korean lacks a contrast between anterior and posterior fricatives before the high front vowel /i/, as only posterior fricatives occur in this context. Yet the Korean loanword system distinguishes English /s/ and /./ using secondary labialization and the fortis-lenis contrast. The fact that English /s/ and /./ are mapped respectively to [.*] and [.w] in Korean loanwords may be related to the phonetic similarity. Experimental results show that Korean [.*] and [..] are acoustically and perceptually the closest to English /s/ and /./, respectively. This seemingly supports perception-based models of loanword adaptation (e.g., Peperkamp and Dupoux 2003), and argues against phonemic accounts of loanword adaptation (e.g., LaCharite and Paradis 2005), which predict the anterior-posterior contrast of English to be neutralized in the Korean loanword system. However, if loanword adaptation is purely based on phonetic similarity, we expect a labialized affricate to be used in place of English /t./, just as a labialized friative replaces English /./. The crucial difference between fricatives and affricates lies in the role of frication noise frequency, which is affected by the secondary labialization, in signaling the phonological contrasts of English. Whereas noise frequency difference is an essential cue for the contrast between /s/ and /./, it does not play a role in distinguishing the affricate /t./ from other English consonants. To capture this contrast-specific importance of phonetic similarity in loanword adaptation, I propose that cross-linguistic sound mapping occurs at the level of auditory feature representation containing only the phonetic details relevant to the perception of phonemic contrasts.

This paper discusses the determinism fallacy of the probe-goal checking theory of movement (Chomsky 2000, 2001) with respect to ambiguity of interpretive effects and proposes an alternative theory of movement that overcomes the determinism fallacy, i.e., the non-deterministic theory of movement. This paper proposes and motivates a non-deterministic theory of movement based on the notion of edge feature (EF) (Chomsky 2008). It will be claimed that the theory of movement based on EF, i.e., the theory of unmarked vs. marked internal merge (IM), is free from the determinism fallacy of the deterministic probe-goal checking theory of movement. Unmarked IM is characterized as inducing a default semantic effect, called a "D-effect", which simply results from the optionality of the operation of unmarked IM itself (Fox 2000), whereas marked IM is characterized as inducing a marked semantic effect called a "gNon-D-effect,"h due to the lexical properties of the head inducing the marked IM and the phrase undergoing the marked IM---leading to a theory of non-determinism.

I propose that the Korean morphological accusative case marker.(l)ul serves to function as an aspectual operator, telicizer, with a universal quantifying force that takes atelics turns them into telics. I will eventually provide the universal quantifying force of the Korean accusative case marker -(l)ul for the restricted distribution of Korean verbal noun objects. This proposal suggests that the so-called accusativus effectivus in Korean can arise from the presence of the morphological accusative case marker -(l)ul appearing on the aspectually associated adverbial constituents, in contrast to the fact that the accusativus effectivus in English arises from the presence of the indefinite article a(n).

Previous studies proposed that there is no CP ellipsis in English (Kennedy and Merchant 2000) and Korean (Ahn and Cho 2009, 2010, cf. Park 2009). In particular, Ahn and Cho (2009, 2010) suggested that apparent CP ellipsis in Korean be in fact DP ellipsis (null realization of DPs in more neutral terms), which is subsumed to pro-drop in Korean. Specifically, Ahn and Cho (2010) make it clear that there is no genuine DP ellipsis in Korean and apparent DP ellipsis is explained with pro. However, close examination of other data reveals that this is not always the case; considering sentences with reflexives, NPIs, indefinite DPs and free choice DPs, we claim that Korean has both pro with no internal structure and DP ellipsis with an internal structure in the ellipsis site, which explain frequent zero realization of argument DPs in many constructions. We propose that the DP contained in the target clause undergoes partial deletion - the DP is stripped of phonological and semantic features only - when it is identical with the one in the antecedent clause. That is, the internal structure remains intact in the ellipsis site. As to the interpretation of the null argument, we propose that the ellipsis site is filled in with semantic features by overt nominals in the antecedent clause and obtains interpretation at LF (cf. Oku 1998, Saito 2007).

This paper will show that the noun and verb of agreement languages such as English and Icelandic have their own independent distinct syntactic processes to determine their feature values with reference to quirky dative subject constructions. It will be claimed that the u[person] and the u[number] of the verb undergo individual probings to determine the verbal agreement values, while another late full-phi Agree is required to value the u[Case] of the noun. Agreement-less languages like Korean and Japanese do not have individual probing operations of the T except for the [person] probing for the honorification cases, so that they do not allow the agreement on the verb, but do the late full-phi Agree to determine the u[Case] of the noun alone. It will be argued that Case feature is distinct from phi-features in that Case feature cannot undergo the feature-copy procedure, unlike phi-features.

This paper focuses on discussing split object sharing, along with object sharing, in SVCs from Korean, and establishing the appropriate structure for this phenomenon. It first shows that this fact cannot be handled with any head-final structures, and then demonstrates that the fact can be appropriately represented with a head-initial structure. Incidentally, object sharing, whole or split, is shown to follow from movement into a theta-position. According to the current claim, then Korean starts out as SVO and ends up displaying SOV properties in large part, thereby supporting Kayne's (1994) universal SVO hypothesis and enhancing the possibility of eliminating the head parameter. Consequently, the results of this paper will require nontrivial modification in various areas of Korean syntax, for example, phrase structures, verbal morphology, etc.

This article develops a syntactic account that yields a parsimonious explanation of the head-final properties of verbal inflections in Korean. It addresses verbal reduplication patterns of so-called echoed verb constructions in the language, proposing that verbal inflectional morphology is derived syntactically from head-initial structures by phrasal movement. This analysis fits in well with the leading ideas of minimalism, in which human language is universal, in particular, with respect to base word order (Kayne 1994), and in which syntactic head movement is something of a conundrum, so it is restricted to an ancillary role.

Since Chomsky (1995, 2001) suggested that head movement be excluded from the core operations of narrow syntax, relegating it to the PF-branch of the computation, researchers have attempted to eliminate head movement from narrow syntax entirely or to find alternatives to the earlier narrow-syntactic approach to head movement. Even with these more refined proposals, however, it still remains unclear why head movement should not be part of narrow syntax. The question becomes even more acute when empirical facts suggest the other way around, that is, the existence of syntactic head movement. In this paper, I will argue that head movement is not necessarily a PF operation by showing that Korean does allow narrow-syntactic head movement, albeit partially, and explore how the observed syntactic head movement can be accommodated within the current minimalism. More specifically, it will be argued that the verb-inflection complex in Korean results from either V-to-T movement or V-to-C movement at narrow syntax, and that in the case of V-to-T movement, the V-T complex combines with the independently merged C via morphological merger at PF. Assuming so, I will further argue, along the line suggested in Suranyi (2005, 2007, 2008), that Korean syntactic head movement goes through reprojective processes (i.e., root merger) at syntax, hence obviating the violation of the Extension Condition and the c-command condition.

This paper reports on an investigation of the use of Korean concessives and contrastives based on Sejong corpus (2007), specifically that of most frequently used concessive and contrastive endings. Among the endings identified as concessive and contrastive in the literature, the concordance search revealed that -nuntey, -ciman, -ato, and -(u)na are used most frequently. Based on further thorough examination of meanings of selected sentences from the concordance results, the research demonstrates four main findings. Firstly, the meaning of concession is expressed much more frequently than the meaning of contrast. Secondly, concessive and contrastive endings show different patterns in their use: In terms of the degrees to which the endings are used concessively and contrastively, -ciman is highest, -nundey is lowest, and -una and -ato are in between. Thirdly, based on frequencies and degrees to which the endings are used as concessives and contrastives, -ciman is revealed to be the most frequently used concessive and contrastive ending: The meaning of concession is expressed in the frequency order of -ciman, -ato, -una and -nuntey The meaning of contrast is worded in the frequency order of -ciman, -nuntey, -una and -ato. Fourthly, the use of concessive meaning is most conspicuous in imaginative prose, which is un-informative and free to express the speaker in character, and that of contrastive meaning is comparatively apparent in spoken texts and discernible in news, in both of which the speaker/writer tries to draw the attention of the addressee/audience.

The purpose of this paper is presenting an overall picture of Interclausal Relations Hierarchy(IRH; Silverstein(1976), Givon(1980), Foley and Van Valin(1984)) between semantic and syntactic relations in regarding to Korean V+e+V constructions. To say, there appears to be some sorts of strong correlations between syntactic and semantic structure of V+e+Vs, which, in turn, help us capable predict what kind of syntactic structure will co-occur with what kind of semantic structure. For giving an overall picture of this correlation, in Figure 2, I present with IRH between semantic and syntactic structures in V+e+V constructions. Here the five morphosyntactic structures are ranked in accordance with the tightness of the syntactic link or bond among them, and the corresponding five semantic structures in accordance with the degree of semantic cohesion among them. In more detail, the lexeme in the morphosyntactic relation corresponds to the fusion of core schema into a lexeme in the semantic relation. Likewise, the auxiliary verb construction(AVC) corresponds to the inflectionalized of dependent verbs. And the compound verb construction(CVC) corresponds to Core schema event mapping restriction, which always splits Core Schema (Path+Motion) into separate clauses. Finally, the subordination clause construction(SCC) reflects the temporal sequence of its corresponding real world(Event linearization order restriction), together with 'Objective analyses of temporal event overlaps in-the-world', and 'Subjective degree of naturalness of temporal relations'.

This paper revisits the Generalized Immediate Scope Constraint (GISC) (Sells and Kim 2006) in Korean, which has been known as one of the Negative Polarity Item (NPI) licensing constraints holding at LF, and argues that the GISC needs to be restated as a more general Intervention Constraint strictly holding at S-structure. This line of reasoning will be supported by the fact that licensing of Korean NPIs is not constrained by the scope relation between an NPI and its licenser as the GISC defines, but rather constrained by the surface word order between them. The restrictions on the multiple occurrences of an NPI will be given as a theoretical consequence brought by the application of the Intervention Constraint.