Ku as a Demonstrative

강남길 한국중앙영어영문학회 2010 영어영문학연구 Vol.52 No.1

In this paper, we argue, in support of the revised version on Principle A, that demonstrative binding is possible if one or the other occurrence of the third person is put forth as if it were a demonstrative. We contend that Principle A must be understood as a condition of demonstrative binding in Korean, thereby suggesting that Principle B does not regulate demonstrative binding. Under Higginbotham’s (1992) proposal, whereby demonstratives can stand apart from the disjointness conditions, Principle B functions so as to exclude the possibility of common reference. The central idea of our analysis, however, is that in the case of Korean demonstrative binding, Principle A functions so as to include the possibility of common reference. Even if two elements have the same coindex, one cannot determine that they have the same value without the speaker’s intention. We present evidence from the previous approaches that kare ‘he’ and ku ‘he’ cannot be bound by a quantifier irrespective of domains. However, the previous approaches are undermined by the fact that Principle A restricts demonstrative binding in Korean. The crucial evidence against the previous approaches is based on the observation that the demonstrative ku ‘he’ admits a bound variable reading and that Principle A constrains demonstrative binding through common reference arising from the speaker’s intention. In conclusion, while variable binding and demonstrative variable binding seem to be considered as having the same semantic relations, the latter is not claimed to be subject to Principle B.

Quantifiers, Pronouns, anda Bound Variable Interpretation

강남길 한국중앙영어영문학회 2007 영어영문학연구 Vol.49 No.1

 Abstract Quantifiers, Pronouns, and a Bound Variable Interpretation Kang, Nam-Kil (Far East University) In this paper, we state the necessary conditions on a BVA reading of pronouns. For this, we argue that there are four instances of a BVA reading in Korean and English and further argue, following Ueyama’s (1998) insight, that there are configurations in which a BVA reading between a quantifier A and its dependent B does not obtain. More simply put, a BVA reading is not available if A is one of A-type QPs and B is a long NP. B-type QPs, unlike A-type QPs, can in fact yield a BVA reading, irrespective of the lexical choices. This fact thus provides confirming evidence for Ueyama (1998)’s insight that the availability of a BVA reading is affected not only by a quantifier A, but also by its dependent term B. In addition to this observation, we maintain that there is evidence that common reference arising from the speaker’s intention is a necessary condition on a BVA reading. In this paper, we also examine the case of epithets where a BVA reading is available. In fact, anaphoric R-expressions such as epithets are incomplete definite descriptions. They induce common reference only if c-command does not hold. Hence, it seems reasonable to entertain the hypothesis that anaphoric R-expressions cannot be linked in common reference with a c-commanding argument.

An Analysis of the Predictive Power of Four Models in Machine Learning and Deep Learning: Focusing on the Reviews of the Movie Noryang

강남길 한국중앙영어영문학회 2024 영어영문학연구 Vol.66 No.1

This article is two-fold. The ultimate goal of this article is to provide a big data analysis of 330 reviews of the movie Noryang and to evaluate the Naive Bayes model, the Random Forests model, the DNN model, and the LSTM model in machine learning and deep learning. A point to note is that the name Yi, Sun-shin was the most widely used by viewers, followed by the word movie, and the word general, in that order. A major point of this article is that the name Yi, Sun-shin and the word movie showed up twice as the first keyword. This in turn implies that these keywords are the most noteworthy ones. The sentiment analysis argues that about 75% of viewers think of the film as well-made and that they were highly satisfied with it. In this paper, we used the Naive Bayes model, the Random Forests model, the DNN model, and the LSTM model and made them predict whether each review is positive or negative. The Random Forests model works well for our data, whereas the Naive Bayes model does not. When learning took place 25 times, the DNN model worked well for our data (its accuracy rate is 82.76%). When it comes to the LSTM model, its accuracy did not improve even though learning took place 9 times. Yet, the LSTM model is slightly better than the DNN model with respect to the accuracy rate of test data.

Coindexation-Based Binding Theory and Linking Theory

강남길 대한영어영문학회 2006 영어영문학연구 Vol.32 No.2

Kang, Nam-kil. “Coindexation-Based Binding Theory and Linking Theory.” Studies on English Language and Literature. 32.2(2006): 187-205. In this paper, our discussion is based on the evaluation of the coindexation-based binding theory and Linking Theory. In Linking Theory, binding relations are meant to be captured through linking. Linking is partly similar to and partly different from coindexing. In this paper, we demonstrate that such differences between linking and coindexing are empirical rather than notational. We also provide several semantic and syntactic evidence for Linking Theory. Those cases are really empirical and hence prove important arguments in favor of Linking Theory. As further evidence for Linking Theory, we consider the possibility of dispensing with reconstruction in BVA. There are some tenable reasons to argue for it. (Far East University)

Crossover and Reconstruction

강남길 한국중앙영어영문학회 2014 영어영문학연구 Vol.56 No.1

The main goal of this paper is to show how Korean crossover cases with scrambling are accounted for without recourse to reconstruction. To achieve this goal, the properties of crossover in Korean are investigated within the conception of connectivity and an agreement-based approach without assuming the reconstruction hypothesis. One major point of Korean crossover cases is that they are instances of connectivity and they fall within our reference agreement principles. With respect to connectivity and agreement, it is important to note that ku ‘he’ can display connectivity with A-type QPs, whereas ku-tul ‘they’ can display connectivity with B-type QPs. That is, if a QP is an A-type QP, ku ‘he’ can display connectivity since its referent consists of a single entity. However, if a QP is a B-type QP, ku-tul ‘they’ can display connectivity since its referent consists of more than a single entity. On the other hand, Korean reflexives, regardless of the types of QPs, can display connectivity with A-type QPs or B-type QPs since their referents can consist of both a single entity and more than a single entity. Additionally, this paper provides some empirical evidence which does not entertain the reconstruction hypothesis that a dependent term after reconstruction must be c-commanded by the QR trace of a QP.

Referential Opacity in Propositional Attitudes

강남길 한국중앙영어영문학회 2013 영어영문학연구 Vol.55 No.3

This paper has attempted to show how referential opacity is licensed within the system of the mental spaces approach. Our analysis is motivated by the fact that we can capture referential opacity in Korean in terms of mental spaces and links between them within the mental spaces approach. One major point of propositional attitudes in Korean is that the embedded subject position with past tense induces only a de re reading, despite the well-known generalization that propositional attitudes are referentially opaque. We take this fact as support for the assumption that subject positions in the complement of propositional attitude verbs in Korean are transparent. Most importantly, we have discussed the cases of referential opacity in Korean propositional attitude sentences and proposed that they fall within the system of the mental spaces approach. No special devices are needed to account for referential opacity in Korean propositional attitude sentences. An interesting point with regard to a de re reading within the system of the mental spaces approach is that there is a referential link between the reality space (R) and the belief space (B). With respect to a de dicto reading, on the other hand, it is important to note that there is no referential link between the reality space (R) and the belief space (B). Finally, we have proposed in this paper that in Korean propositional attitude sentences pronouns induce interpretations that are looser than reflexives and pro yields interpretations that are looser than pronouns.

Null Arguments

강남길 한국중앙영어영문학회 2011 영어영문학연구 Vol.53 No.3

One of the most well-known features of linguistic theory is that there exist various empty categories at various levels of mental representation. Languages such as Korean and Japanese are well-known for the fact that they allow arguments of a predicate to be freely dropped. Such null arguments have been a challenge to linguistic theory in that their contents are not recovered from the verbal agreement system. Simply put, null arguments occur in languages such as Korean, Mandarin, and Japanese, which has no agreement system whatsoever. If so, why should we allow the existence of pro? What types of empty categories appear in Korean? In order to give a plausible answer to these queries, we discuss the issue of assuming ECs such as PRO and pro. The present study assumes that both PRO and pro are allowed in Korean and that PRO appears as the subject of both tensed and untensed clauses, and pro as an empty pronominal in subject and object positions. Previous approaches assume that overt plural pronouns in Korean have a definite reading, whereas pro does not. Unfortunately, the picture turns out to be even more complicated. We have demonstrated in this paper that Korean pro can be more broadly used than the Korean overt pronoun ku in that pro can bear definite as well as indefinite interpretations, while overt pronouns have only definite interpretations. Finally, we have proposed that pro can function as a free variable depending on the speaker’s intention since the referent of pro is always influenced by the discriminating use of a demonstrative by the speaker. We have shown that this fact results from extending the spirit of Higginbotham’s (1992) analysis to pro.

Referential Number Agreement and Disagreement of Pronouns and Anaphors: Why are pronouns and anaphors different?

강남길 한국중원언어학회 2014 언어학연구 Vol.0 No.33

This paper was intended to investigate the difference in the referential number agreement between Korean pronouns and reflexives. It was postulated that the difference is dependent on the fact that unlike Korean reflexives, Korean pronouns have a number feature. One major point to be noticed in this paper is that Korean pronouns such as ku 'he' and ku-tul 'they' include their antecedents only by number agreement. More specifically, ku is a structurally or semantically singular NP, requiring a single entity as its referent, while ku-tul 'they' is a structurally or semantically plural NP requesting more than a single entity. Another major point which is worth emphasizing is that caki'self' and caki-casin 'self-self' are linked to antecedents by either the number agreement or number disagreement. Put differently, their referents can consist of either one entity or more than one entity. Finally and perhaps most importantly, the difference in the referential number agreement between pronouns and anaphors can be accounted for by the fact that pronouns have their own perfect number features, whereas Korean reflexives do not. This fact leads to the hypothesis that the referents of 'caki' and 'caki-casin' consist of one entity or more than one entity since they are semantically uncountable NPs.

Pronouns, Reflexives, and Sloppy Identity

강남길 한국중앙영어영문학회 2009 영어영문학연구 Vol.51 No.3

The goal of this paper is to capture the behavior of pronouns and reflexives based on sloppy identity. One major point of surface anaphora in Korean is that Korean pronouns, unlike English pronouns, cannot give rise to the strict/sloppy ambiguity. An interesting property of the Korean overt pronoun ku ‘he’ is that it can never induce the sloppy reading, due to the same restriction prohibiting a bound variable interpretation, unlike English pronouns. This, in turn, suggests that the Korean overt pronoun kunye ‘she’ cannot also exhibit the sloppy reading, unlike English pronouns. This asymmetry between English and Korean may be accounted for by assuming that the quantificational force applies to English with regard to sloppy identity, while it cannot apply to Korean. Another central idea of this paper is that Korean reflexives share common properties with English reflexives with regard to the obligation of the sloppy reading, but ku-casin ‘he-self’ may constitute further support for its pronominal status. That is, it may give rise to the strict/sloppy ambiguity under VP ellipsis, regardless of local binding and non-local binding. It is worth mentioning that deep anaphora in Korean admits the strict/sloppy ambiguity. A demonstrative may in general be understood as a context of copying or as a context of anaphora. Under copying, sentences which include a demonstrative behave like the comparable VP-deletion structure, regardless of locality, whereas under a context of anaphora, deep anaphora gives rise to the strict reading without a linguistic antecedent. The goal of this paper is to capture the behavior of pronouns and reflexives based on sloppy identity. One major point of surface anaphora in Korean is that Korean pronouns, unlike English pronouns, cannot give rise to the strict/sloppy ambiguity. An interesting property of the Korean overt pronoun ku ‘he’ is that it can never induce the sloppy reading, due to the same restriction prohibiting a bound variable interpretation, unlike English pronouns. This, in turn, suggests that the Korean overt pronoun kunye ‘she’ cannot also exhibit the sloppy reading, unlike English pronouns. This asymmetry between English and Korean may be accounted for by assuming that the quantificational force applies to English with regard to sloppy identity, while it cannot apply to Korean. Another central idea of this paper is that Korean reflexives share common properties with English reflexives with regard to the obligation of the sloppy reading, but ku-casin ‘he-self’ may constitute further support for its pronominal status. That is, it may give rise to the strict/sloppy ambiguity under VP ellipsis, regardless of local binding and non-local binding. It is worth mentioning that deep anaphora in Korean admits the strict/sloppy ambiguity. A demonstrative may in general be understood as a context of copying or as a context of anaphora. Under copying, sentences which include a demonstrative behave like the comparable VP-deletion structure, regardless of locality, whereas under a context of anaphora, deep anaphora gives rise to the strict reading without a linguistic antecedent.

Sloppy Identity, VP Ellipsis, and Null Anaphora

강남길 한국중앙영어영문학회 2007 영어영문학연구 Vol.49 No.3

One of the central doctrines of VP ellipsis and null anaphora is that they relate sloppy identity. The way they are hypothesized to do so is by postulating ambiguity of interpretation in the complete clause, or by treating null objects as a kind of VP ellipsis. What we have been arguing here is focussed on such semantic questions. However, it should be pointed out that there are problems at the very heart of the two analyses we have dealt with here. Attractive though the two analyses might be, they are empirically incorrect. Several problems are worth mentioning regarding them. First, the lambda abstraction-based approaches should always postulate ambiguity of interpretation in the complete clause that serves as a source for the elliptical clause. To put it differently, they require hidden ambiguity and the derivation of their rules is too procedural. Second, as stated in the course of our discussion, the lambda abstraction-based approaches and the equation-solving approach cannot account for the existence of the mixed reading under VP ellipsis. Third, the parallelism between the availability of the sloppy reading and that of a bound variable interpretation for a pronoun is not consistent. The sloppy reading can in fact arise independently of bound variable anaphora. Fourth, the NOC in Korean does not necessarily yield sloppy readings in the contexts where English VP ellipsis does. A further problem is that the NOC in Korean seems to require a different treatment. The results observed in this paper suggest that the NOC in Korean cannot be treated as analogous to VP ellipsis in English. Apart from theoretical considerations, there are empirical problems with the analysis treating the NOC in Korean as analogous to VP ellipsis in English.