A corpus-based study on contrast and concessivity of the connective ‑ciman in Korean

Abstract

Providing a corpus-based analysis of the contrastive connective -ciman in Korean, this study demonstrates that global connections are as frequent as local connections as opposed to previous literature. In representing various senses possible with a -ciman connection, this study adopts a fuzzy representation, where meanings range from conceptual to discoursal. The identified meanings include explicit contrast, denial of expectation, speech act hedges and idiomatic expressions. The fuzzy representation is supported for at least two reasons. First, categorization of some cases is often blurred. Second, it can better capture the relatedness of various meanings whose enduring sense concerns ‘contrast’. It is further revealed that interpretations of -ciman phrases are compositionally made with co-occurring linguistic units. The “pragmatic” meanings of -ciman constructions are explained in terms of different levels of representations in which the contrast occurs. We further investigate the possibility of -ciman as a stance/discourse marker with accompanying expressions.

Keywords:
Publication history
Table of contents

1.Introduction

Contrast consists of a range of phenomena cross-linguistically. Predominantly, these notions are realized by lexical devices such as the English words but or nevertheless. In addition to lexical methods, discourse contrast or concessivity can be achieved by use of a future tense in non-temporal use, especially in Romance languages (Baranzini and Mari 2019Baranzini, Laura, and Alda Mari 2019 “From Epistemic Modality to Concessivity: Alternatives and Pragmatic Reasoning per absurdum”. Journal of Pragmatics 142: 116–138. DOI logoGoogle Scholar). Moreover, contrastive markers strongly resonate with epistemic modality, because contrastive markers can help the speaker distance herself from the propositional content.11.Following the tradition in pragmatics literature (e.g., Sperber and Wilson 1986/1995Sperber, Dan, and Deirdre Wilson 1986/1995Relevance: Communication and Cognition (Second edition with postface). Oxford: Blackwell.Google Scholar), the speaker is referred to as she and the hearer as he, unless the genders of interlocutors are revealed from the context. At the same time, contrast involves a series of arguments in which one point of view is presented in a provisional way, and then necessarily followed by a segment “advancing the opinion favored by the speaker” (Zafiu 2018Zafiu, Rodica 2018Epistemic and Evidential Markers in the Rhetorical Context of Concession. Journal of Pragmatics 128: 116–127. DOI logoGoogle Scholar, 117).

Korean has a repertoire of contrastive devices, one of which is the use of the connective ending -ciman. As is the case with most lexical units, the ending -ciman can convey a variety of meanings that depend on the linguistic context in which it occurs. Mainly for that reason, the ending -ciman has been discussed as either a contrastive or a concessive marker in the Korean linguistics literature (e.g., Lee and Lee 1999Lee, Huyca, and Conghuy Lee 1999Theyksuthu Pwunsekcek Kwuke Emiuy Yenkwu [A Text-Analytic Approach to Korean Endings]. Seoul: HankwukmwunhwasaGoogle Scholar; Lee 2000Lee, Unkyeng 2000Kwukeuy Cepsokemi Yenkwu [A Study of Korean Conjunctive Suffixes]. Seoul: Thayhaksa.Google Scholar; Park 2014Park, Jae-Yeon 2014 “Hankwuke Yenkyelemi Uymi Hwakcangeyseuy Hwanyuwa Unyu [Metonymy and Metaphor in the Semantic Extensions of Korean Connective Endings].” Kwukehak [Journal of Korea Linguistics] 70: 117–155.Google Scholar; Lee 2018Lee, Keum-Hee 2018Pocosa manun kyelhaphyeng yenkyel phyohyeney tayhaye: -cimanun, -tamanun, -(u)kkamanun lyu cwungsimulo. [Final endings + particle Maneun on coupled forms focused mainly -jimaneun, -damaneun, -eulkkmaneun]. Kwukehak [Journal of Korean Linguistics] 86: 63–88.Google Scholar). Regarding the variety of meanings of the ending -ciman, opinions have been divided about whether it has a unitary sense (e.g., Choi 1937Choi, Hyenpay 1937Wulimalpon. [Korean Grammar]. Seoul: Top Publishing.Google Scholar; He 1977He, Wung 1977Wuliyeysmalpon [Grammar of Old Korean]. Seoul: Saymmwunhwasa.Google Scholar; Nam and Ko 1983Nam, Kisim, and Yengkun Ko 1983Phyocwunkwukemwunpeplon. [Grammar of the Standard Korean]. Seoul: Top PublishingGoogle Scholar) or dual/multiple senses (e.g., Lee and Im 1983Lee, Iksep, and Hongpin Im 1983Kwukemwunpeplon. [Korean Grammar]. Seoul: Tayhaksa.Google Scholar; Yoon 1989Yoon, Phyenghyen 1989Kwukeuy Cepsokemiey Tayhan Yenkwu. [A study of Korean Conjunctive Suffixes]. Ph.D. Dissertation. Cheonnam National University. Korea.; Lee 2000Lee, Unkyeng 2000Kwukeuy Cepsokemi Yenkwu [A Study of Korean Conjunctive Suffixes]. Seoul: Thayhaksa.Google Scholar). Following a pragmatic principle dubbed ‘Modified Occam’s Razor” (Grice 1989Grice, Paul 1989Studies in the Way of Words. Cambridge, MA: Harvard University Press.Google Scholar, 47), this study adheres to a unitary account of the semantic and pragmatic characteristics of the marker.

Even though the previous studies of the marker under issue provide insightful explanations of the meanings and uses of this marker, most of them are based on data constructed by the researchers. Departing from these existing approaches, this study aims to analyze the marker under consideration by means of authentic corpus data (i.e, the Sejong Corpus). Advocating a unitary approach, the current study will investigate what kinds of meanings are delivered by this marker compositionally with its contexts, and how the relatedness of diverse meanings of the marker can be accounted for.22.As one review correctly points out, the process of compositional interpretation needs to be empirically supported for it to be legitimately claimed. While agreeing with the comment, I assume that a general/agreed-upon interpretation of an utterance containing -ciman in a specific context can be derived by an “ordinary speaker” who is competent in understanding language and the context. In doing so, this study also hopes to uncover how the marker under discussion is actually deployed by ordinary speakers.

2.Background

2.1Contrastive connectives

Contrastive markers such as the English but are argued to deliver two major meanings, which are exemplified in (1) and (2).

(1)

Tom is tall but Bill is short.

(2)

Tom is tall but he is not good at basketball.

The first meaning in (1) is called semantic opposition (Lakoff 1971Lakoff, Robin 1971 “If’s, and’s and but’s about Conjunction.” In Studies in Linguistic Semantics, eds. by C. Filmore, and D. Langenden, 114–149. New York: Holt, Reinhard and Winston. 114–149.Google Scholar), external but (Halliday and Hasan 1976Halliday, M. A., and R. Hasan 1976Cohesion in English. London: Longman.Google Scholar), contrast use (Blakemore 1987Blakemore, Diane 1987Semantic Constrains in Relevance. Oxford: Blackwell.Google Scholar, 1989 1989 “Denial and Contrast: A Relevance Theoretic Analysis of But.” Linguistics and Philosophy 12: 15–38. DOI logoGoogle Scholar), or contrast in proposition (Schiffrin 1987Schiffrin, Deborah 1987Discourse Markers. Cambridge: Cambridge University Press. DOI logoGoogle Scholar)/ content (Sweetser 1990Sweetser, Eve 1990From Etymology to Pragmatics: Metaphorical and Cultural Aspects of Semantic Structure. Cambridge: Cambridge University Press. DOI logoGoogle Scholar) domain. The second meaning in (2), on the other hand, is dubbed denial of expectation (Lakoff 1971Lakoff, Robin 1971 “If’s, and’s and but’s about Conjunction.” In Studies in Linguistic Semantics, eds. by C. Filmore, and D. Langenden, 114–149. New York: Holt, Reinhard and Winston. 114–149.Google Scholar; Blakemore 1987Blakemore, Diane 1987Semantic Constrains in Relevance. Oxford: Blackwell.Google Scholar, 1989 1989 “Denial and Contrast: A Relevance Theoretic Analysis of But.” Linguistics and Philosophy 12: 15–38. DOI logoGoogle Scholar), internal but (Halliday and Hasan 1976Halliday, M. A., and R. Hasan 1976Cohesion in English. London: Longman.Google Scholar) or contrast in knowledge (Schiffrin 1987Schiffrin, Deborah 1987Discourse Markers. Cambridge: Cambridge University Press. DOI logoGoogle Scholar)/ epistemic (Sweetser 1990Sweetser, Eve 1990From Etymology to Pragmatics: Metaphorical and Cultural Aspects of Semantic Structure. Cambridge: Cambridge University Press. DOI logoGoogle Scholar) domain. Whereas Lakoff (1971)Lakoff, Robin 1971 “If’s, and’s and but’s about Conjunction.” In Studies in Linguistic Semantics, eds. by C. Filmore, and D. Langenden, 114–149. New York: Holt, Reinhard and Winston. 114–149.Google Scholar contends the distinction between (1) and (2) to be semantic, scholars such as Dascal and Katriel (1977)Dascal, Marcelo, and Katriel Tamar 1977 “Between Semantics and Pragmatics: The Two Types of ‘but’ – Hebrew ‘aval’ and ‘ela’.” Theoretical Linguistics 4: 143–172. DOI logoGoogle Scholar, Blakemore (1987)Blakemore, Diane 1987Semantic Constrains in Relevance. Oxford: Blackwell.Google Scholar, and Fraser (1998) 1998 “Contrastive Discourse Markers in English.” In Discourse Markers: Descriptions and Theory, eds. by A. H. Jucker, and Y. Ziv, 301–326. Amsterdam: Benjamins. DOI logoGoogle Scholar claim that the two are derived from a single sense, from which specific meanings are fleshed out in the context. For example, in her pioneering work, Blakemore (1987)Blakemore, Diane 1987Semantic Constrains in Relevance. Oxford: Blackwell.Google Scholar proposes that but in both uses functions as a semantic constraint by guiding the listener to interpret the proposition introduced by but as being contrasted to the previous utterance (in contrast use) or to a derived proposition (in denial-of-expectation use). In (1), the two propositions are contrasted to each other because the two involve lexical antonyms short and tall couched in syntactically same constructions. In (2), on the other hand, albeit without an overt contrast either lexical or structural, the interpretation of the second proposition would guide the hearer to infer a proposition from the first one, which can be paraphrased as (3).

(3)

If Toms is tall, he is good at basketball.

(3), which is an expectation inferable from the first conjunct in (2), is denied in the second conjunct of (2). Hence the use of but in (2) is a denial-of-expectation. Here expectations can come in many types, including presupposition, entailment and implicature. In the case of (3), the relevant expectation can be an implicature, more specifically, a particularized conversational implicature à la Grice (1989)Grice, Paul 1989Studies in the Way of Words. Cambridge, MA: Harvard University Press.Google Scholar.33.For various types of expectations, see Fraser (1998) 1998 “Contrastive Discourse Markers in English.” In Discourse Markers: Descriptions and Theory, eds. by A. H. Jucker, and Y. Ziv, 301–326. Amsterdam: Benjamins. DOI logoGoogle Scholar.

In addition to these two uses of but, it can also connect two segments that differ in their speech acts, as illustrated in (4).

(4)
A:

Don’t you see that the door is open?

B:

I’m not blind, but if you want me to close it, why don’t you say so straight out? (Dascal and Katriel 1977Dascal, Marcelo, and Katriel Tamar 1977 “Between Semantics and Pragmatics: The Two Types of ‘but’ – Hebrew ‘aval’ and ‘ela’.” Theoretical Linguistics 4: 143–172. DOI logoGoogle Scholar, 158–159)

In (4), a statement is followed by a question and these are connected by but. Such uses of but as in (4) are called speech-act contrast in Sweetser (1990)Sweetser, Eve 1990From Etymology to Pragmatics: Metaphorical and Cultural Aspects of Semantic Structure. Cambridge: Cambridge University Press. DOI logoGoogle Scholar. Still, the use of but in (4) appears to be licensed because a contrast relationship can hold between propositions derived from the two conjuncts, as spelled out in (5) and (6).

(5)

I am responding to your question in light of the literal content by saying “I’m not blind”.

(6)

I am challenging your question in light of the way the message was delivered by saying “why don’t you say so straight out?”

As proposed in most previous studies, this study takes ‘contrast’ to be the core and enduring meaning of such contrastive markers as but. That is, the ambiguity thesis will be abandoned on the grounds of the parsimony of the sense, spelled out by Grice (1989Grice, Paul 1989Studies in the Way of Words. Cambridge, MA: Harvard University Press.Google Scholar, 47) as “Modified Occam’s Razor: senses are not to be multiplied beyond necessity”. This principle states that senses are to be proliferated only when necessary as in lexical ambiguity, where unrelated meanings are derived from what is apparently the same word. Otherwise, the intuitive relatedness between the senses possible with but cannot be captured. This study adopts a unitary approach, suggesting that various senses possible with a -ciman construction carry contrast as an enduring or core sense. This core sense is then fleshed out as one of such various meanings in an ad hoc way in a specific context. That is, the intended meaning can be arrived at quite automatically or effortlessly by an interlocutor in a specific context, à la Jaszczolt’s Default Semantics (2005Jaszczolt, Kasia 2005Default Semantics: Foundations of a Compositional Theory of Acts of Communication. Oxford: Oxford University Press. DOI logoGoogle Scholar, 2009 2009Representing Time: An Essay on Temporality and Modality. Oxford: Oxford University Press.Google Scholar, 2016 2016Meaning in Linguistic Interaction. Oxford: Oxford University Press. DOI logoGoogle Scholar, and references therein).

2.2Contrastive connectives in Korean

Contrastive connection in Korean can be made in at least two ways: by conjunctive adverbials or by conjunctive verbal suffixes. In this study, the second method will be taken up and labelled “contrastive connectives,” because it pertains to the topic of the present study. In Korean linguistics literature, representative contrastive verbal suffixes include, but not limited to, -(u)na, -ciman, -ato, and -telato. A group of studies put forward that contrastive connectives can be distinguished between contrast and concession senses and classified accordingly by a range of terms (e.g., Lee and Im 1983Lee, Iksep, and Hongpin Im 1983Kwukemwunpeplon. [Korean Grammar]. Seoul: Tayhaksa.Google Scholar; Yoon 1989Yoon, Phyenghyen 1989Kwukeuy Cepsokemiey Tayhan Yenkwu. [A study of Korean Conjunctive Suffixes]. Ph.D. Dissertation. Cheonnam National University. Korea.; Lee 2000Lee, Unkyeng 2000Kwukeuy Cepsokemi Yenkwu [A Study of Korean Conjunctive Suffixes]. Seoul: Thayhaksa.Google Scholar). For example, Lee (2000)Lee, Unkyeng 2000Kwukeuy Cepsokemi Yenkwu [A Study of Korean Conjunctive Suffixes]. Seoul: Thayhaksa.Google Scholar dubs -(u)na and -ciman contrastive and -ato and -telato concessive. As discussed in Section 2.1 above, these bifurcate approaches are not supported for the theoretical reason of Modified Occam’s Razor. Also such taxonomic accounts do not offer clear-cut criteria to demarcate boundaries between categories. After all, in many cases, the same connective is proven to convey both contrastive and concessive senses. Nor can these approaches exhaust all meanings or functions possible with a connective in a specific context, because the meaning quite often arises from the context.

On the other hand, another group of studies suggests a unitary approach, contending that both contrastive and concessive senses are derived from a single overarching sense, which is labelled in a variety of ways (e.g., Choi 1937Choi, Hyenpay 1937Wulimalpon. [Korean Grammar]. Seoul: Top Publishing.Google Scholar; He 1977He, Wung 1977Wuliyeysmalpon [Grammar of Old Korean]. Seoul: Saymmwunhwasa.Google Scholar; Nam and Ko 1983Nam, Kisim, and Yengkun Ko 1983Phyocwunkwukemwunpeplon. [Grammar of the Standard Korean]. Seoul: Top PublishingGoogle Scholar). Nam and Ko (1983)Nam, Kisim, and Yengkun Ko 1983Phyocwunkwukemwunpeplon. [Grammar of the Standard Korean]. Seoul: Top PublishingGoogle Scholar propose ‘incompatibility’ as the core meaning shared by contrastive connectives, which is then fleshed out in specific contexts. On the other hand, the way specific meanings are figured out, extended, or changed has also been explored. According to Park (2014)Park, Jae-Yeon 2014 “Hankwuke Yenkyelemi Uymi Hwakcangeyseuy Hwanyuwa Unyu [Metonymy and Metaphor in the Semantic Extensions of Korean Connective Endings].” Kwukehak [Journal of Korea Linguistics] 70: 117–155.Google Scholar, the core meaning of -ciman (i.e., contrast) can be realized in an epistemic domain as in (7), or in a speech act domain as in (8) via metaphoric extension, which is carried out by means of a cross-domain mapping.

(7)

I have never worked with him-ciman, it seems that he is a nice person.

(8)

I need to ask my boss-ciman, why don’t you visit our office?

Even though previous research on contrastive connectives including -ciman has contributed to our deeper understanding of the phenomena under discussion, most studies have focused on constructed data drawn from researchers’ expertise on and knowledge of the Korean language. In contrast to previous studies, this paper suggests a fuzzy representation of senses of -ciman constructions surfacing from the corpus data. The identified senses include explicit contrast, denial of expectation/implicature, speech act hedges and idiomatic expressions. The rationale behind this fuzzy representation is at least two-fold. First, sometimes, the categorization of an example is rather fuzzy, suggesting that it can serve multiple functions simultaneously. Hence, the boundaries between categories in the current study will be analytical. Second, the fuzzy representation can better reflect relatedness among various senses of -ciman constructions. The so-called “pragmatic” senses of -ciman constructions, such as denial of expectations and speech acts, are accounted for by means of different levels of representations in which contrast occurs. Furthermore, the possibility of utilizing -ciman as a stance/discourse marker is pursued vis-à-vis certain idiomatic expressions.

3.Data and methods

The current study collected research data from the Sejong Corpus (https://​ithub​.korean​.go​.kr​/user​/main​.do) published by the National Institute of the Korean Language. The corpus site offers two different modes: written and spoken corpora. Whereas the written corpus comes with four different options (i.e., raw corpus, morphologically tagged corpus, morpho-semantically tagged corpus and syntactically tagged corpus), the spoken corpus has only the first two options. This study chose the morphologically tagged spoken corpus because it facilitates searching for target expressions.

The target expression, -ciman ‘but, though’, which is tagged as a connective ending in the corpus, was typed in, and then all relevant concordance lines were searched. This search returned 1,381 occurrences. Figure 1 presents the screen capture of the search.

Figure 1.Screen capture of the search for -ciman in the Sejong Corpus
Figure 1.

The culled concordance lines were copied and saved in an Excel file for ease of searching and cleaning the data. Among the collected results, 318 items were excluded from the list because the target form was employed sentence-finally or repeatedly. Because -ciman as a connective ending is always employed with a predicate, we examined predicates co-occurring with -ciman and found a total of 36 predicate types in the current data. Among them, the top 12 predicate types, which occurred more than 25 times, were targeted for this study. The chosen predicates are presented in Table 1, with their respective frequency counts. A total of 768 tokens were obtained for the current study.

Table 1.Top 12 predicates with -ciman in Sejong Corpus
Expressions Freq. Expressions Freq.
ha ‘do’+ciman 144 anh ‘not do’+ciman  36
iss ‘be’+ciman 143 eps ‘not be’+ciman  36
ani ‘not be’+ciman 130 toy ‘become’+ciman  32
molukeyss ‘may not know’+ciman  60 alkeyss ‘may know’+ciman  30
hayss ‘did’+ciman  58 ke ‘be that’+ciman  28
molu ‘not know’+ciman  45 anhass ‘was/were not’+ciman  26
Total 768

Some of the constructions in Table 1 are variants of the same predicate. For example, in ha ‘do’+ciman and ha+yss+ciman, the second form contains a past particle -yss right after the predicate ha ‘do’. Similarly in molu ‘not know’+ciman and molu+keyss+ciman, the latter has an additional suppositive particle -keyss. Even though variants are derived from the same basic form, the presence or lack of these elements (i.e., past particles and suppositive particles) may contribute to the meaning construction or retrieval. For that reason, variants were treated separately in this study.44.For example, the formulaic expression, mianha ‘be sorry’-ciman, barely allows the suppositive particle -keyss to co-occur with it. Of course, these two can be combined when intended literally. We then classified finalized lists of tokens according to categories emerging from extracted data, which are adopted and modified from previous research such as Sweetser (1990)Sweetser, Eve 1990From Etymology to Pragmatics: Metaphorical and Cultural Aspects of Semantic Structure. Cambridge: Cambridge University Press. DOI logoGoogle Scholar and Park (2014)Park, Jae-Yeon 2014 “Hankwuke Yenkyelemi Uymi Hwakcangeyseuy Hwanyuwa Unyu [Metonymy and Metaphor in the Semantic Extensions of Korean Connective Endings].” Kwukehak [Journal of Korea Linguistics] 70: 117–155.Google Scholar. When working on classification, we carefully took the relevant context surrounding -ciman constructions into account.

4.Analysis and discussion

We classified the senses of -ciman constructions compositionally with contexts, the results of which are presented in Table 2. We again caution that the demarcation is fuzzy rather than taxonomic, for at least two reasons. The first reason lies in blurred boundaries between categories, which suggest that a -ciman construction can carry multiple functions simultaneously. In addition, the fuzzy representation can better capture the relatedness between a range of senses possible with -ciman constructions, thus conforming to the parsimony of the sense (i.e., senses are not to be multiplied beyond necessity). The categories are aligned along a scale or cline, the ends of which are a conceptual connection and a discoursal connection (c.f., Lee 2002Lee, Hye-Kyung 2002 “Towards a New Typology of Connectives with Special Reference to Conjunction in English and Korean.” Journal of Pragmatics 34: 851–866. DOI logoGoogle Scholar). Here a conceptual connection refers to a connection on (elements of) propositional contents, while a discoursal connection indicates whatever function a connective performs apart from its conceptual connection. This second type of connection can recurrently be made by the so-called discourse connectives (Blakemore 1987Blakemore, Diane 1987Semantic Constrains in Relevance. Oxford: Blackwell.Google Scholar, 1989 1989 “Denial and Contrast: A Relevance Theoretic Analysis of But.” Linguistics and Philosophy 12: 15–38. DOI logoGoogle Scholar) or discourse markers (Schiffrin 1987Schiffrin, Deborah 1987Discourse Markers. Cambridge: Cambridge University Press. DOI logoGoogle Scholar; Fraser 1990Fraser, Bruce 1990 “An Approach to Discourse Markers.” Journal of Pragmatics 14: 383–395. DOI logoGoogle Scholar, 1998 1998 “Contrastive Discourse Markers in English.” In Discourse Markers: Descriptions and Theory, eds. by A. H. Jucker, and Y. Ziv, 301–326. Amsterdam: Benjamins. DOI logoGoogle Scholar), which encode ‘procedural’ information that guides the hearer to interpret an utterance (Blakemore 1987Blakemore, Diane 1987Semantic Constrains in Relevance. Oxford: Blackwell.Google Scholar, 1989 1989 “Denial and Contrast: A Relevance Theoretic Analysis of But.” Linguistics and Philosophy 12: 15–38. DOI logoGoogle Scholar). Although it is legitimately assumed that connectives can be distinguished between conceptual and discoursal ones, such bipartite approaches will be abandoned in this study, primarily because one connective can frequently perform both functions, depending on contexts.

Table 2.Classification of meanings of -ciman
Category Example in Sejong Corpus (in English translations)

Conceptual

tab2.svg

Discoursal

1. Explicit contrast Adults understand the situation-ciman, children do not.
2. Denial of expectation/implicature He was mentally retarded-ciman, he developed gradually.
3. Speech act hedges
     3.1. as Shown in the syllabus-ciman, we will study the relation between computers and corpora…
     3.2. as well as I didn’t study hard-ciman, the professor did not give a good grade.
     3.3 Topic introduction You can see from campaigns-ciman, the concept of campaigns has changed.
4. Idiomatic expressions I don’t know well-ciman, it might be this switch.

Examples belonging to 3.1 to 3.3 are sub-categories of speech act hedges. The term ‘speech act hedge’ therefore is employed as a cover/umbrella term that encompasses the three sub-categories together with typical speech act hedges.

Out of the categories presented in Table 2, some are labeled as “pragmatically” inferred meanings (e.g., Yoon 1989Yoon, Phyenghyen 1989Kwukeuy Cepsokemiey Tayhan Yenkwu. [A study of Korean Conjunctive Suffixes]. Ph.D. Dissertation. Cheonnam National University. Korea.; Lee 2000Lee, Unkyeng 2000Kwukeuy Cepsokemi Yenkwu [A Study of Korean Conjunctive Suffixes]. Seoul: Thayhaksa.Google Scholar), which include denial of expectation/implicature, speech act hedges, and idiomatic expressions. In contrast to such approaches, we will advocate a unitary analysis, wherein various senses of -ciman phrases bear an underlying or core sense, contrast. This core sense then manifests as one of these meanings presented in Table 2 in an ad hoc way in a specific context. To repeat, the meaning intended by a speaker is claimed to be fleshed out automatically or effortlessly by an interlocutor in a specific context (à la Jaszczolt’s Default Semantics 2005Jaszczolt, Kasia 2005Default Semantics: Foundations of a Compositional Theory of Acts of Communication. Oxford: Oxford University Press. DOI logoGoogle Scholar, 2009 2009Representing Time: An Essay on Temporality and Modality. Oxford: Oxford University Press.Google Scholar, 2016 2016Meaning in Linguistic Interaction. Oxford: Oxford University Press. DOI logoGoogle Scholar).55.One reviewers commented how my unitary approach can better explain the multiple meanings possible with -ciman than does the null hypothesis that it is polysemous and that various interpretations are derived from these multiple senses stored in the mental lexicon. Both types of approaches can be equally viable, differing in terms of viewing the multiple senses of an expression.

4.1Prevalence of global connection

Connections between linguistic units can be either local or global. The literature on contrastive connectives has revolved chiefly around local connections, i.e., ones between two adjacent linguistic units (c.f., Schiffrin 1987Schiffrin, Deborah 1987Discourse Markers. Cambridge: Cambridge University Press. DOI logoGoogle Scholar). However, examination of the research data of this study reveals that global connections prevail as much as local connections do. The excerpt in (9) illustrates the global nature of -ciman connections. (9) is excerpted from a conversation between two speakers who talk about a long road trip they had together. The speakers, who are college students, agree that the trip enabled them to become true friends. The speaker mentions that he is very close to a female friend, Senhuy, even after the speaker and Senhuy have spouses. Before talking about Senhuy, he prefaces his utterance with a -ciman clause in line 2, which means that closeness to friends can vary from person to person. Here the -ciman suffixed clause stands in a loose contrastive connection with the content of the very last clause (line 5). The two connected conjuncts are intervened by a portion of utterance, which renders the connection global.

(9)
  01
cikum
now
wuli
our
moim-ey
group-in
iss-nun
be-md
chinkwutul-hako-nun
friends-with-tc
02
etten
some
chinmilto-uy
closeness-of
chai-nun
difference-tc
iss-keyss-ciman
be-sup-though
  03
seywel-i
year-nm
cinaka-to
go-though
 
Senhuy-ka
Senhuy-nm
mwe
some
namphyen-i
husband-nm
  04
sayngki-ko
have-and
nay-ka
I-nm
anay-ka
wife-nm
sayngki-tunci
have-regardless
 
kaney
regardless.of
kelikkim-epsi
hesitation-without
  05
manna-l
meet-md
swu
possibility
iss-nun
be-md
sai-ya.
relationship-dc
    Even though there are differences in closeness with friends in our group, Senhuy and I can meet without any hesitation, regardless of whether she and I have spouses in the future.’ [#6CT_0051]66.For the transcription of the Korean data, the Yale Romanization was used. The abbreviations, which are borrowed and modified from Sohn (1999Sohn, Ho-Min 1999The Korean Language. Cambridge: Cambridge University Press.Google Scholar, 2013 2013Korean. Seoul: Korea University PressGoogle Scholar), are presented in the Appendix. Because of the limited space, morpheme-by-morpheme glossing is sometimes provided only for relevant parts. File numbers in the Sejong Spoken Corpus are given in square brackets after examples.

In global connections, the -ciman clause repeatedly serves the function of speech act hedges, modulating the force of the propositional content of the utterance, as will be addressed in Sections 4.4 below. (10) is excerpted from a conversation about movies. Speaker A comments that film versions of novels are mostly disappointing. The interlocutors then elaborate on the topic by talking about an example of a film adaption of a novel, thoymalok ‘Exorcist’s Journal’.

(10)
  01 A:
amwuthun
anyway
wulinala-twu
our.country-too
kulehciman
like.that-ciman
sosel-i
novel-nm
yenghwahwatway-se …
made.into.film-and
  02  
silmangha-n
be.disappoint-md
kyengwu-ka
case-nm
manhi
a.lot
iss-ess-ten
be-pst-md
ke
that
kath-ay.
like-dc
      ‘Anyway, although it’s the same in Korea, it looks like film adaptions of novels are pretty disappointing.’
  03 B: thoymalok.
      ‘Exorcist Journal’
  04 A:
thoymalok-un
exorcist.journal-tc
ccom
somewhat
kuke-n
that-nm
mwe.
what
      ‘Exorcist Journal is somewhat like that.’
  05 B:
cwuknu-n
die-md
cwul
possibility
al-ass-e
know-pst-dc
kuke
that
po-myense.
watch-while
      ‘I thought I was going to die watching that.’
06 A:
na-twu
I-too
mwe
some
thoymalok-uy
Exorcist.Journal-of
pisusha-n
similar-md
yu-nun
type-tc
pyellwu
much
an cohaha-ciman
not.like-ciman
  07  
ku
that
thoymalok
exorcist.journal
mwe
what
seykyeyphyen
world.part
mwe
what
ku
that
taumey
next
mwe.
what
      I don’t like such works as Exorcist’s Journal either, but there are parts about the World and the next’
  08 B:
honseyphyen
chaos.part
malseyphyen
the.end.of.the.world.part
      ‘There are parts about chaos and the end of the world.’
  09 A:
honseyphyen
chaos.part
malseyphyen
the.end.of.the.world.part
mwe
what
ilehkey
like.this
manhi
many
iss-ntey
be-but
      ‘There are parts about the chaos and the end of the world. Like these, there are many parts.’
  10 A:
com
Some
yuchihakey
childish
tway-ss-tula.
become-pst-dc
      ‘They [parts of Exorcist’s Journal] are a little childish.’ [#6CT_0030]

A’s utterance in line 6 begins with a -ciman clause, which expresses the speaker’s negative attitude/evaluation toward certain types of movies. Noticeable is that the content right after the -ciman clause (line 7) does not contrast with the -ciman clause, because A digresses from the main topic by enumerating a variety of versions of thoymalok. Furthermore, A’s utterance in line 7 cooperates with B’s utterance in line 8 by elaborating on the different versions of thoemalok. As can be seen in A’s utterances in line 10, A conveys his negative opinions about thoymalok. Then the contents in the -ciman clause in line 6 and the succeeding utterances by A do not contrast. What A wants to deliver by the -ciman clause is his propositional attitude to the topic, which can be rephrased as in (11) and (12).

(11)

If I don’t like such works as thoymalok, I should not talk about them.

(12)

I am talking about them.

The second clauses of (11) and (12) contrast with each other. The -ciman clause in line 6 performs a concessive function from A’s perspective, attenuating A’s act of stating his evaluation of certain types of novels.

4.2Explicit contrast

Explicit contrast consists of occurrences in which contrast is manifested by means of explicit lexical or structural contrast combined by a -ciman, as demonstrated by the first sentence in Table 2 above. To repeat, this type is dubbed semantic opposition (Lakoff 1971Lakoff, Robin 1971 “If’s, and’s and but’s about Conjunction.” In Studies in Linguistic Semantics, eds. by C. Filmore, and D. Langenden, 114–149. New York: Holt, Reinhard and Winston. 114–149.Google Scholar), external contrast (Halliday and Hasan 1976Halliday, M. A., and R. Hasan 1976Cohesion in English. London: Longman.Google Scholar), or contrast use (Blakemore 1987Blakemore, Diane 1987Semantic Constrains in Relevance. Oxford: Blackwell.Google Scholar, 1989 1989 “Denial and Contrast: A Relevance Theoretic Analysis of But.” Linguistics and Philosophy 12: 15–38. DOI logoGoogle Scholar). Another example can be seen in (13) below.

(13)
  01
kulayse
so
i
this
tungkanchekto-eyse-nun
interval.scale-on-tc
02
phullasu-nun
addition-tc
kyeysan-i
calculation-nm
toy-ciman
be.done-ciman
kophaki-nun
multiplication-tc
an
not
twayyo.
be.done
  03
kophaki
multiplication
ha-nun
do-md
uymi-nun
meaning-tc
ku
that
tases
five
pay-lanun
times-so.called
  04
uymi-ka
meaning-nm
an
not
toyn-ta-nun
become-dc-md
ke-yeyyo
point-dc
    ‘So, on this interval scale, addition can be calculated, but multiplication cannot be done. The logic of multiplication does not hold (on the interval scale).’ [#6CT_0002]

(13) is excerpted from a lecture on research methodology including quantitative measurement scales. In this specific part, the speaker explains the characteristics of interval scale with reference to other types of scales. What contrasts in this example are two lexical terms, addition and multiplication. Lexical contrast can be of various kinds (e.g., Lyons 1977Lyons, John 1977Semantics. Cambridge: Cambridge University PressGoogle Scholar) ranging from complementary antonyms (i.e., dead vs. alive and male vs. female) to taxonomic sisters (i.e., color terms and days of a week). In (13), the relevant contrast is that of taxonomic sisters that are from a set of paradigmatically contrasting elements. As Lyons (1977)Lyons, John 1977Semantics. Cambridge: Cambridge University PressGoogle Scholar points out, contrast here does not carry any implication as to the number of contrasting elements in a set.

In some cases, the concept of lexical contrast needs to be expanded to accommodate examples like (14), because semantic contents of two units do not stand in a strict contrastive relationship. Excerpt (14) is from a monologue of a student who has been to Afghanistan as a participant in a relief work project. The speaker provides an account about a tribe whose language lacks a writing system.

(14)
  01
callalapatu-ey
Zalarabad-in
sa-nun
live-md
mincok-un
people-tc
phasyai-lanun
Pashay-be.called
congcok-i-ketunyo.
tribe-be-dc
02
kuntey
but
i
this
mincokun
tribe
ku
that
mwullon
of.course
malun
spoken.language
iss-ciman
be-though
  03
kulca-ka
written.language-nm
eps-ess-ten
not.have-pst-md
ke-ya
that-dc
    ‘The people of Zalarabad are Pasha. They, of course, have a spoken language, but they didn’t have any written language.’ [#6CT_0013]

In (14), the two lexemes, mal ‘spoken language’ and kul ‘alphabet’, are contrasted. Then a set of contrastive lexemes is set up, in which spoken language and written language constitute the elements of the set. Structurally the two contrasted lexemes are couched in parallel syntactic patterns: malum issta (spoken language exists) vs. kulcaka epsta (written language does not exist). On the other hand, as will be discussed in Section 4.3 below, occurrences like (14) can be construed as an instance of denial of expectation because the clause suffixed by a -ciman evokes an inferred proposition that if a tribe has a spoken language, it commonly has a written language. This very expectation is denied in the second clause. This phenomenon supports the fuzzy representation of categories proposed in this study, because a -ciman clause performs dual functions at the same time.

4.3Denial of expectations

Examples in this category contain the use of -ciman which signals that the content of an upcoming talk is read as a denial of any expectations inferable from the content of an expressed clause (e.g., Blakemore 1987Blakemore, Diane 1987Semantic Constrains in Relevance. Oxford: Blackwell.Google Scholar; Schiffrin 1987Schiffrin, Deborah 1987Discourse Markers. Cambridge: Cambridge University Press. DOI logoGoogle Scholar). For that reason, contrast in this use is claimed to have a cancellative function (Bell 1998Bell, D. M. 1998 “Cancellative Markers: A Core/Periphery Approach.” Pragmatics 8: 515–541. DOI logoGoogle Scholar). Excerpt (15) is from a monologue about a movie, Beautiful Mind. The speaker comments that the main character of the movie was a brilliant man, although the speaker thinks he was beautified too much in the film.

(15)
01
nemwuna
too.much
ku
that
yenghwa-eyse
movie-in
mihwasikhye-ss-ta
glorify-pst-dc
lanun
like
02
sayngkak-i
thought-nm
tule-se
come.to-and
com
a.little
anthakkap-ki-n
sorry-nom-tc
hay-ss-ciman
do-pst-ciman
03
ku
that
salam
person
cachey-nun
self-tc
koyngcanghi
extremely
hwullyunghan
outstanding
salam-i-n
person-be-md
ke
that
kath-kwu.
see-dc
  Although I was sorry to think that he was glorified too much in the movie, he seemed to be a great man himself.’ [#6CT_0019]

In (15), there is no pair of contrasted lexical items in the two clauses connected by -ciman. Instead, the clause suffixed by -ciman can trigger an inference as in (16).

(16)

If I felt sorry to think that he was glorified too much, I think that he was not that outstanding.

However, this inferred information expressed in the second clause of (16) is denied in (15). By means of this, the contrast conveyed by -ciman is legitimized. Propositions derivable from the first clause can differ in their nature (Fraser 1998 1998 “Contrastive Discourse Markers in English.” In Discourse Markers: Descriptions and Theory, eds. by A. H. Jucker, and Y. Ziv, 301–326. Amsterdam: Benjamins. DOI logoGoogle Scholar; Lee 2002Lee, Hye-Kyung 2002 “Towards a New Typology of Connectives with Special Reference to Conjunction in English and Korean.” Journal of Pragmatics 34: 851–866. DOI logoGoogle Scholar). Specifically the choice of such propositions is constrained by the content of the second conjunct, as shown in (16). Blakemore (1987Blakemore, Diane 1987Semantic Constrains in Relevance. Oxford: Blackwell.Google Scholar, 1989 1989 “Denial and Contrast: A Relevance Theoretic Analysis of But.” Linguistics and Philosophy 12: 15–38. DOI logoGoogle Scholar) contends that the English discourse connective but constrains the way the utterance but introduces. However, Lee (2002)Lee, Hye-Kyung 2002 “Towards a New Typology of Connectives with Special Reference to Conjunction in English and Korean.” Journal of Pragmatics 34: 851–866. DOI logoGoogle Scholar proves that discourse connectives including but also contribute to the way the preceding utterance is construed, as illustrated in (16).

(17) is an excerpt from a conversation between two speakers about an array of topics such as cultures, movies, and life. Speaker A says that he had pizza with a mutual friend, Ciyeng. Responding to A, B comments that A should sometimes eat pindayttek77.“This is one of the mandatory dishes on traditional holidays or special festive occasions. Also called nokdujijim or binjatteok, bingdaetteok is made by peeling and soaking mung beans, grinding them, and then pan-frying with various vegetables.” (Retrieved on 27 August, 2019 from https://​terms​.naver​.com​/entry​.nhn​?docId​=5700793​&cid​=42701​&categoryId​=63067) (obviously as opposed to the representative western dish, pizza), because A is old enough to prefer Korean food to western food (line 7). As a response to B’s comment, A says that although he likes pindayttek, he eats pizza only rich with cheese (lines 8–9).

(17)
  01 B:
ciyeng-i-lang mwel mekesstakwu?
      ‘What did you say you ate with Ciyeng?’
  02 A: phisca.
      ‘Pizza.’
  03 B:
etten ciyengi
      ‘Which Ciyeng?’
  04 A:
kwupen a kwupeni anikwuna sonciyeng.
      ‘Student ID number 9, ah, not number 9, but Ciyeng Son.’
  05 B:
a sonciyeng.
      ‘Ah, Ciyeng Son.’
  06 A:
ung
toykey
kkwulkkwulhayse
phica
mekca
hay
kacikwu
ipmas
tolase
tto
kulen
ke
cohahacanha
tto.
      ‘Yes, I was so grumpy that I suggested having pizza. When I have a good appetite, I love that kind of food too.’
  07 B:
phicaman
cohahay
naika
myech
kayntey
pintayttektwu
com
mekkwu
kakkum
eng?
      ‘You only like pizza. I think you’re old enough to eat pintayttek often, huh?
08 A:
pintayttek-twu
pintayttek-too
cohaha-ciman
like-ciman
phica-twu
pizza-too
chicu-ka
cheese-nm
manhi
much
  09  
tulun
put.in
ke-man
thing-only
mek-e.
eat-dc
      Although I like pintayttek, I only eat pizza rich with cheese.’
  10 B: salccye
      ‘You’ll get fat.’ [#6CT_0039]

In lines 8 through 9, A conjoins two clauses using -ciman. Here we can observe a lexical contrast between pintayttek and pizza, whereby the use of -ciman is accommodated. At the same time, the contrast can be viewed to hold between derived/inferred propositions of the two conjuncts. From the first conjunct of A’s utterance in line 8, we can derive an assumption that A behaves according to his age (i.e., A likes bindayttek, as expected by ordinary Korean adults). From the second conjunct, we can infer a proposition that A does not quite follow this convention (because he eats the representative western food, pizza, even rich with cheese which is generally believed to be dispreferred by average Korean adults). The contrast here occurs between these two derived propositions.

4.4Speech act hedges

As discussed in Section 4.1 above, -ciman clauses can serve as a hedge of an assertion, marking the speaker’s concessive attitude toward the assertion (c.f. Schiffrin 1987Schiffrin, Deborah 1987Discourse Markers. Cambridge: Cambridge University Press. DOI logoGoogle Scholar). Even though definitions and usages of the term, ‘hedge’, may vary in the literature, a hedge usually refers to “a particle, word or phrase that modifies the degree of membership of a predicate or noun phrase in a set” (Brown and Levinson 1987Brown, Penelope, and Stephan Levinson 1987Politeness: Some Universals in Language Use. Cambridge: Cambridge University Press. DOI logoGoogle Scholar, 145). Hedges can also be defined as methods of mitigating the force of an assertion for pragmatic reasons such as displaying politeness or avoiding full commitment.

(18) is excerpted from a dialogue between two friends about employment. The speaker of (18) states that her employed friends complain about excessive working hours. Never having prepared for job hunting, the speaker feels envious of such friends despite their complaints about long working hours.

(18)
  01
chwiephan
aytulun
maynnal
sinipsawenila
yelhansiey
tuleka …
hamyemse
    ‘Those who are employed … complain that they leave office at 11 p.m. and come back at 5 a.m.’
  02
solcikhi tulekassunikka ku thwucengul hakwu issci kyayneytuli
    ‘Honestly, they can complain because they are employed.’
  03
acik nanun nayka cikcepcekulo an kyekkese
    ‘Because I’ve never experienced it directly yet,’
04
nay-ka
I-nm
mwe-lako
anything-as
ha-l
say-md
swu
possibility
iss-nun
possibility-tc
ipcang-un
position-tc
an-i-ciman
not-be-ciman
    I’m not in a position to say anything about it, but’
  05
pwuleptanun
sayngkakto
tulko
solcikhi
cenpwu
ta
chwiepi
toynun
palamey.
    ‘I feel envious of them, honestly, because they have jobs.’ [#7CT_0016]

In line 4, the speaker employs a -ciman clause, which asserts that she is not in a position to say anything because she did not go through the employment problem herself. Then she expresses her envy of her employed friends. The -ciman clause in this excerpt clearly conveys the speaker’s concessive attitude toward the upcoming assertion about her feeling about such friends, mitigating the force of this assertion.

When combined with a concessive marker, phrases corresponding to the English phrase I don’t know in different languages have turned out to function as a forward-looking marker revealing that “the speaker is not fully committed to what follows in his/her turn” (Weatherall (2011Weatherall, Ann 2011 “I Don’t Know as a Prepositioned Epistemic Hedge.” Research on Language and Social Interaction 44(4): 317–337. DOI logoGoogle Scholar, 317) in English and Helmer et. al. (2016)Helmer, Henrike, Silke Reineke, and Arnulf Deppermann 2016 “A Range of Uses of Negative Epistemic Constructions in German: ICH WEIß NICHT as a Resource for Dispreferred Actions.” Journal of Pragmatics 106: 97–114. DOI logoGoogle Scholar in German). In these cases, both lexical and structural properties compositionally contribute to engendering such concessive meanings. (19) is taken from the same corpus as (9) above, in which two interlocutors talk about a long road trip. In line 1, speaker A, who is a college student in his twenties, makes a verbal mistake. He must have wanted to say that he has not lived very long, but actually he said that he has lived long. Noticing this mistake, A corrects himself in line 3, but his mistake is laughed at by B in line 4. A repeats himself in line 5 by saying that because he has not lived long, he does not know very well. Then he gives his opinion about the subtlety involved in relationships between men and women.

(19)
  01 A:
kuntey
ku
namnyekwankyeylanun
key
cham
omyohanila
nato
manhi
sala
pwase.
      ‘So the relationship between men and women is so weird. Because I’ve lived long.’
  02 B: macayo.
      ‘Right.’
03 A:
manhi
long
sal-a
live-and
po-ci
see-nom
anh-ase
not.do-because
molu-ciman.
not.know-ciman
      Although I don’t know because I’ve not lived long,’
  04 B:
nato
manhi
sala
pwase.
      ‘(Did you say) because I’ve lived long?’
05 A:
ani
manhi
sala
poci
anhase
calun
moluciman
      ‘No, although I don’t know well because I’ve not lived long,’
  06  
ku
cham
mimyohaci
cham
selokaney
isengilan
concayka
cham
yomwulin
ke
kathay
      ‘It’s so subtle. It seems that the opposite sexes are weird.
  07  
issumyen
pwutamtoykwu
epsumyen
tto
pokwu
siphkwu.
      ‘If you have a partner, you feel pressured. If you don’t, you miss one.’ [#6CT_0051]

In (19), the predicate used in the -ciman clause is moluta ‘not know’, which is a representative epistemic modality marker. As in other languages such as English and German, the predicate alta ‘know’ and the ending -ciman collectively yield a concessive meaning, making the whole -ciman clause a hedge.

A few senses of -ciman clauses drawn on or related to the speech-act-hedge meaning are identified including comparison (i.e., as), addition (i.e., as well as), and topic introduction, which are discussed in Sections 4.4.1 through 4.4.3. Although the senses they carry are subtly distinguished, they still retain a common property of packaging the content in the -ciman clause as if it were given/old information or background knowledge. The -ciman clause in these cases thus functions as a hedge for the speaker to reduce the onus of presenting a new sub-topic. To repeat, these functions compositionally result from lexical meanings of the predicate, the marker -ciman, and the context. This is why the following three senses collapse under the overarching speech-act hedge category.

4.4.1Referring to a precondition or prior information (i.e., as)

A -ciman clause sometimes can carry the meaning of the English word as or like. Specifically, a -ciman clause can be employed to refer to a previously mentioned or presupposed assumption, simultaneously conveying the speaker’s hedging attitude that she wishes to be excused for repeating what has already been discussed or known to the interlocutors (Lee 2018Lee, Keum-Hee 2018Pocosa manun kyelhaphyeng yenkyel phyohyeney tayhaye: -cimanun, -tamanun, -(u)kkamanun lyu cwungsimulo. [Final endings + particle Maneun on coupled forms focused mainly -jimaneun, -damaneun, -eulkkmaneun]. Kwukehak [Journal of Korean Linguistics] 86: 63–88.Google Scholar). By doing so, the speaker can achieve a couple of discourse goals. She can remind her interlocutors of information already presented or presupposed, helping them comprehend what follows. In addition, the speaker can lessen the burden of repeating the same piece of information by adopting such hedging strategies.

(20) is an excerpt from a doctor’s lecture on children’s mental-health issues, including autism. In line 3, the doctor cites autism as the most representative mental disorder which was discussed with reference to a Korean movie, Marathon. The main character of the movie is an autistic young man who overcame adversity to be a great marathoner.

(20)
  01
i
cenpancekin
paltaley
mwunceyka
issnun
aytul
ceyka
chilyohanun
pangpepiketunyo
    ‘(This is) the way I treat the children who have a problem with overall development.’
  02
kuntey
i
cwungeyse
kacang
tayphyocekin
cangayka
yelepwun
sowi
malhanun
    ‘But among them, the most representative disorder is, ladies and gentlemen, the so-called, ’
03
akka
a.while
ceyka
ago
ku
that
malathon
Marathon
yayki-to
story-too
ha-yss-ciman
tell-pst-ciman
caphyey
autism
cangay
disorder
kyeythong-i-eyyo
type-be-dc
    as I mentioned the movie Marathon a while ago, it is a kind of the autism disorder.’
  04
ike
icey
caphyey
cangaynun
cenpancek
paltal
cangayla
kulayse
    ‘So autism is called an overall developmental disorder.’ [#9CT_0006]

The -ciman clause can also be used to represent the information available from the context in some form or other. (21) is from a medical lecture to the public by a doctor, in which the doctor starts his lecture by mentioning osteoporosis. As the excerpt is from the very onset of the talk, nothing was discussed before. However, it is inferable from the speaker’s utterances that the topic of the talk osteoporosis must have been presented as the title of the presentation material. The -ciman clause in line 4 informs the audience that the topic appears as the title. At the same time, it serves as a hedge, allowing the speaker to excuse herself to repeat the same information.

(21)
  01 yey yelepwun annyeng annyenghasyesssupnikka cehuy ku seypulansu pyengwenhako
    ‘Hello, everyone, how are you? Severance Hospital and,’
  02
hankyeleysinmwuni
kongtongulo
kaychoyhanun
wuli
a
kenkang
kongkaykangcwaey …
    ‘the Hankyoreh newspaper are co-holding this public lecture on health.
  03
a
onulun
ceyka
malssum
tulilyeko
ha-nun
kenun,
    ‘What I am going to present today is’
04
a
‘ah
yeki
here
ceymok-ey-to
title-in-too
ilehkey
like.this
nawa
appear
iss-ciman
be-ciman
    ‘um, as shown in the title here,’
  05
a
soli
epsnun
a
pamsonnim
koltakongcung
ilen
nayyongul
kaciko
    ‘the topic of a soundless thief, Osteoporosis.
  06
a
ceyka
malssumul
tulikeysssupnita.
    ‘I am going to give you a talk about it.’

4.4.2Adding information (i.e., in addition to)

Some -ciman clauses impart the sense of addition, roughly synonymous with the English expression in addition to. Still, they simultaneously convey a hedging connotation, reducing the weight of the information contained in -ciman clauses and thus bringing to the foreground what follows the -ciman clauses. Then the contrast can be made between different degrees of importance allotted to two compared parts. In examples classified as belonging to this category, the particle -to ‘too’ concurrently plays a significant role in constituting this information-adding interpretation. The compositional nature involved in the elucidation of a -ciman clause is again buttressed here.

In (22), the speaker talks about his close friends, one of whom is said to be especially cherished by the speaker. In line 2, the speaker states that he remembers that specific friend because he is an old friend of the speaker. In upcoming talks, the speaker adds another reason why he remembers that specific friend.

(22)
01
i
chinkwunun
mwel
hanun
salaminyamyen
hayyangtaylul
colephako
  ‘What this friend does is, he graduated from the Ocean University,’
02
kwuntay
taysin
cikum
paylul
sam
nyen
thako
isse
il
hang
sam
hangsalo
  ‘instead of serving in the army, he’s been on a boat for three years now as a first, a third mate.’
03
i
this
chinkwu-ka
friend-nm
way
why
kieknamnya-myen-un
be.remembered-if-tc
na-lang
I-with
olaytoyn
old
chinkwu-i-kito
friend-be-too
ha-ciman,
be-ciman
  ‘The reason I remember this friend is that in addition to being an old friend of mine,’
04
cengmal
nayka
kunkka
i
chinkwu
maley
i
chinkwuka
ilehkey
hayla
han
ke
  ‘(I can really follow) what he advised me to do.’
05
nay
insayngi
kellin
mwunceyil
ttay
nayka
ttalakal
swu
issul
mankhumey
chinkwuya.
  ‘He is such a friend that I can follow his advice on matters of my life.’ [#5CT_0017]

What the speaker intends to convey is that in addition to being an old friend, this very friend is memorable because he can be consulted with even about life-threatening problems. Therefore the -ciman clause here conveys information which is treated as primary, serving as a springboard for further crucial information.

This function of a -ciman clause is more conspicuous in (23). The speaker in (23) talks about how he won his girlfriend’s heart and eventually got married. The speaker and his wife first met in the U.S. when they were students. After a while, the speaker alone returned to Korea, but he went back to the U.S. again to see her with a special present. The ostensible reason was to pick up his luggage in the U.S.

(23)
  01 me senmwuley cip kathun ke issessnuntey me ssayngssu suthoanka?
    ‘There was a gift shop or something. Um, Thanks Store?
  02 kekise phocang com yeyppukey phocanghay talla kulayssnuntey,
    ‘I asked them to wrap the present beautifully there.’
  03 kuke kaciko mikwukey kasseyo.
    ‘I went to the States with the present.’
04
cey
my
yuilhan
only
mokcek-un
purpose-tc
cim
luggage
kaci-le
bring.back-to
o-n
come-md
kes-to
thing-too
iss-ciman,
be-ciman
    ‘The only purpose of my visit was to bring my luggage back, but in addition to that,
  05 tasi han pen e phyohyenhaki wihayse wase
    ‘Once again, um, I went there to confess my love.’
  06 mannacakwu kulayssteni kkamccak nollay. e hankwuk tuleon ke aninya kwu?
    ‘When I asked her to meet, she was surprised and said, “Huh, didn’t you go back to Korea?’ [#8CT_0031]

In line 4 in (23), the speaker mentions that the only reason to return to America was to bring back his luggage. Right after that, the speaker continues that the more important reason was to confess his love. In this narrative, the excuse contained in the -ciman clause is described as ostensible, whereas the reason revealed by the following utterance becomes far more prominent and relevant. In this respect, this function of -ciman is quite similar to another Korean contrastive connective -nuntey, which was proven to provide background information (e.g., Lee 1991Lee, Hyo Sang 1991Tense, Aspect and Modality: A Discourse-Pragmatic Analysis of Verbal Suffix in Korean from a Typological Perspective. Ph.D. Dissertation, UCLA.; Lee 1993Lee, Keedong 1993 “The Pragmatic Function of the Connective Nuntey.” Ene [Korean Journal of Linguistics] 5(1): 119–135.Google Scholar; Park 1999Park, Yong-Yae 1999 “The Korean Connective Nuntey in Conversational Discourse.” Journal of Pragmatics 31:191–218. DOI logoGoogle Scholar) or establish the situation.

4.4.3Introducing a sub-topic

The -ciman clause can also be deployed to introduce a sub-topic of a main topic, organizing the content of the -ciman clause to be already known to the interlocutors. Here the compositional nature involved in the construal of a -ciman clause is pronounced most prominently, because the predicate immensely contributes to this topic-introducing sense.

(24) is an excerpt from a dialogue between two persons who converse about politics focusing on recent presidential elections. Speaker A describes how the then presidential candidate of a minority party (Mr. Kwenyengkil) made fun of the then two major political parties. In line 4, speaker B introduces a sub-topic, which is election campaigns, utilizing a -ciman clause. He then elaborates on that sub-topic by commenting that the ways election campaigns are conducted have changed in a positive way. As well as bringing up a new sub-topic, the -ciman clause in (24) accomplishes a hedging purpose by rendering the sub-topic introduced in the -ciman clause familiar to the interlocutors.

(24)
  01 A: mak selo mak mincwutang ccokilang hannalatangi yaykihanikka
      ‘As the Democratic Party and the Grand National Party were talking to each other,
  02   kwenyengkil ssika hanun yaykika e hannalangtangun wencopwuphaytangiyo
      ‘Mr. Kwenyengkil said, “the Grand National Party is the original corruption party,’
  03   mincwutangun sincangkayeptangintey mwel hal mal issnyako …
      ‘and that the Democratic Party is a novice party, so what can you say?”’
04 B:
kunikkan
so
senke
election
wuntong-ul
campaign-ac
pwa-to
see-even
al-keyss-ciman
know-sup-ciman
      As you can notice from the election campaigns,’
  05   ma senke wuntongilan kaynyemto manhi tallacyesse.
      ‘the notion of election campaigns has changed dramatically.’ [#7CT_0038]

As discussed in 4.4 above, the predicate alta ‘know’ in (24) plays a substantial role in conveying the above-mentioned function. However, even non-epistemic predicates can denote a similar sense, as shown in (25), where the predicate nukkita ‘feel’ is used. In (25), a male speaker in his twenties recites a narrative about love. Prior to this excerpt, he talked about his first love, which failed badly. He adds that he had recently started reading a book, Men are from Mars, Women Are from Venus, which is about common relationship problems between men and women. He then brings up a sub-topic about the book, i.e., what he feels about the book in line 5.

(25)
  01 kulen ke (hethalkam) ttaymwuney swipkey isengeykey takakaci moshantanun cem
    ‘Because of that (despondency), (I) cannot easily get close to girls.’
  02 mwe mwe kathun namca chinkwutul kathun kyengwuya mwe
    ‘Um, um, in the case of male friends,’
  03 ilen yayki celen yaykihata pomyenun chinhaycil swu issciman waynci kelikami nukkyecinun
    ‘I can make friends with them, talking to each other. But, somehow I feel distant (with female friends).’
  04 mwe yocumey poki sicakhan chayki mwe hwasengeyse on namca kumsengeyse on yeca?
    ‘Um, the book I recently started reading is “Men Are from Mars, Women Are from Venus”?’
05
ku
that
chayk-ul
book-ac
ccokum
a.little
ilkepo-myense
read-while
nukki-nun
feel-md
ke-ciman ,
thing-ciman
    This is what I feel while I read the book.’
  06 cham pyelkayuy sayngmwulikwuna sayngkakhanun phaytheni ayey thullitanun ke
    ‘(Men and women) are totally different creatures. The ways they think are different.’ [#7CT_0010]

By couching this subtopic in a -ciman clause in line 5, the speaker can hedge against the burden of introducing this new sub-topic, thus making his utterance sound more modest.

4.5Idiomatic expressions

When -ciman is combined with some lexical hedges such as sorry, excuse me, and I don’t know, the whole -ciman clause constitutes an idiomatic expression. Indeed, it is well known that hedging can be achieved by a variety of devices ranging from particles to prosodic clues (e.g., Brown and Levinson 1989). In Korean, expressions such as I am sorry-ciman, excuse me-ciman, and I don’t know-ciman have been approached as stereotypical examples of -ciman-based idioms (e.g., Lee and Lee 1999Lee, Huyca, and Conghuy Lee 1999Theyksuthu Pwunsekcek Kwuke Emiuy Yenkwu [A Text-Analytic Approach to Korean Endings]. Seoul: HankwukmwunhwasaGoogle Scholar; Park 2014Park, Jae-Yeon 2014 “Hankwuke Yenkyelemi Uymi Hwakcangeyseuy Hwanyuwa Unyu [Metonymy and Metaphor in the Semantic Extensions of Korean Connective Endings].” Kwukehak [Journal of Korea Linguistics] 70: 117–155.Google Scholar). Among the data of the current study, a set of such idiomatic expressions was also identified.

The segment in (26) is excerpted from a multi-party conversation of three close friends, addressing an array of topics including movies. In (26), the participants exchange their thoughts about a movie, The Rainmaker. This movie features stories and scenes of trials, in one of which the main character’s (lawyer of a plaintiff) skillful performances stand out prominently, quieting the court and even influencing the jury’s decision. In (26), the three interlocutors share their opinions about such clichés of Hollywood movies’ court scenes. In line 9, participant C, prefacing his utterance with I am sorry-ciman, states that court scenes are all the same. Indeed the same speaker (speaker C) utters the same assertion in line 2 and repeats it in line 9. The only difference is the presence of I am sorry-ciman in line 9. The speaker does not utilize this -ciman clause because he literally feels sorry. Rather, he uses it as an idiomatic expression, emphasizing his opinion about such a typical unfolding and simultaneously showing and seeking consent among the speakers. The other speakers agree and use the same expression, as demonstrated in lines 3, 8, and 10.

(26)
  01 A: mak mwela kulenikka cangnay coyonghayciko mak ilen ke anya?
      ‘(The lawyer) says something. The court becomes quiet. Something like that?’
  02 C: tay taypwupwun ta kulay. Haha.
      ‘It’s almost like that. Ha-ha.’
  03 B: taypwupwun ta kulay. mwela kulemyen pepceng ta coyonghayciko.
      ‘It’s almost like that. If the lawyer says something, the whole court becomes quiet.’
  04 C: haha kuchi.
      ‘Ha, ha, right.’
  05 A: mak mwela kule.
      ‘The lawyer says something.’
  06 B: paysimwen neme kako mak.
      ‘The jury turns to support the lawyer.’
  07 A: pepceng coyonghayciko mak kkullye nakako ike anya?
      ‘Isn’t it that the courtroom becomes quiet and somebody is just dragged out?’
  08 B: ta kulay.
      ‘It’s all like that.’
09 C:
ta
all
kul-ay
like.that-dc
ta!
all
mianha-ciman
be.sorry-ciman
ta
all
kulehta-n-ta.
like.that-in-dc
      ‘It’s all like that! I am sorry, but it’s all like that.’
  10 B: ta kulay.
      ‘It’s all like that.’ [#7CT_0029]

Another -ciman-based idiom recurrently addressed in the literature is I don’t know-ciman. Of special interest is that this construction involves the so-called epistemic predicate know in its negative form. Examination of contexts of this idiomatic phrase frequently indicates that users do not lack knowledge about the issue in question. Rather this idiomatic expression serves as a prepositioned epistemic hedge (Weatherall 2011Weatherall, Ann 2011 “I Don’t Know as a Prepositioned Epistemic Hedge.” Research on Language and Social Interaction 44(4): 317–337. DOI logoGoogle Scholar), showing that the speaker is not completely committed to what follows.

The monologue in (27) is taken from a seminar on language learning and errors. While explaining redundancy in spoken language, the speaker mentions a scholar called George who criticized the methodology of oral auditory language teaching with great acrimony. The expanded context indicates that the speaker is quite knowledgeable about redundancy in language use and its relevant literature. Nevertheless, he moderates the force of his argument by using an I don’t know very well-ciman in line 4.

(27)
  01 mwe wulika yaykilul hamyense ipmaleysenun panpokul manhi hakey toycanhayo
    ‘Well, as we talk, we repeat ourselves a lot in spoken language.’
  02 ku thukhi ipmaleyseuy kulen ku kanthwusa sayongina mwe phyohyenuy panpokina,
    ‘Especially in spoken language, use of interjections or repetition of expressions,’
  03 kulen kel uymihanun ken anin ke kathkwuyo yeylul tulese ku, swuep sikaney,
    ‘I don’t think these are what is meant. For example, in class,’
04
um
um,
cal
well
molu-keyss-ciman,
not.know-sup-ciman
    ‘um, I don’t know very well, but,
  05 thukhi i cocilanun salami kwutwuchengkaksik ku pangpepul
    ‘particularly, this scholar named George (criticized) the oral auditory methodologies.’
  06 acwu sinlalhakey piphanhako isstanun kulen nayyongi issketunyo?
    ‘It is said that he criticized this methodology very harshly?’ [#6CT_0003]

This ciman clause as an idiomatic chunk fulfills a politeness purpose, marking the speaker’s concessive attitude to the audience rather than conveying its literal meaning.

As such, these rather fixed chunks such as I am sorry-ciman and I don’t know-ciman can serve as stance or discourse markers (Carter and McCarthy 2006Carter, Ronald, and Michael McCarthy 2006Cambridge Grammar of English: A Comprehensive Guide. Cambridge: Cambridge University Press.Google Scholar),88.Stance markers are classified as a sub-category of pragmatic markers by Carter and McCarthy (2006)Carter, Ronald, and Michael McCarthy 2006Cambridge Grammar of English: A Comprehensive Guide. Cambridge: Cambridge University Press.Google Scholar. In linguistics literature, the same marker has been dubbed in different terms, including, but not limited to, discourse markers (e.g. Schiffrin 1987Schiffrin, Deborah 1987Discourse Markers. Cambridge: Cambridge University Press. DOI logoGoogle Scholar; Fraser 1990Fraser, Bruce 1990 “An Approach to Discourse Markers.” Journal of Pragmatics 14: 383–395. DOI logoGoogle Scholar), discourse particles (e.g. Schourup 1985Schourup, Lawrence 1985Common Discourse Particles in English Conversation. New York: Garland.Google Scholar; Mosegaard Hansen 1998Mosegaard Hansen, Maj-Britt 1998The Function of Discourse Particles: A Study with Special Reference to Spoken Standard French. Philadelphia: Benjamins. DOI logoGoogle Scholar), and discourse connectives (e.g. Blakemore 1987Blakemore, Diane 1987Semantic Constrains in Relevance. Oxford: Blackwell.Google Scholar). which express the speaker’s stance/attitude towards the following message much more than delivering propositional content meanings, largely because their literal meanings are bleached, conveying instead the speakers’ concessive stance or attitude toward the upcoming utterances and simultaneously meeting any desired politeness expectations. It is seldom the case that function words (i.e., pronouns, prepositions, and conjunctions) or affixes (i.e., suffixes and prefixes) serve as stance markers. Cross-linguistically, most stance markers developed out of content words, such as nouns, adjectives, or adverbs (e.g. Gray and Biber 2014Gray, Bethany, and Douglas Biber 2014 “Stance Markers.” In Corpus Pragmatics, eds. by Karin Aijmer, and Christoph Rühlemann, 219–248. Cambridge: Cambridge University Press.Google Scholar). However, as discussed in Lee (2015) 2015 “A Corpus-Pragmatic Analysis of Wuli .” Discourse and Cognition 22(3): 59–78. DOI logoGoogle Scholar, the Korean first-person plural pronoun wuli ‘we, our’ functions as a stance marker in contemporary spoken Korean. This argument was supported by the fact that the pronoun wuli demonstrates the properties of common nouns such as pluralization (e.g., wuli-tul ‘we-plural’) or suffixation by a possessive marker -uy (e.g., wuli-uy ‘we-of’). In a similar vein, the ending -ciman may function as a stance marker together with certain co-occurring idiomatic expressions, even though it is not a content word or a free morpheme.

5.Conclusion

Departing from most previous studies on the Korean contrastive ending -ciman which have revolved around constructed data, in this paper we conducted a corpus-based analysis of this ending. Connections between linguistic units can be made either locally or globally. Previous literature on contrastive endings has primarily focused on local connections, i.e, two adjacent linguistic units. However, we demonstrated that global connections are frequently entertained by ordinary speakers. Adopting a fuzzy representation, we identified a range of meanings possible with -ciman constructions emerging from corpus data. The identified meanings include explicit contrast, denial of expectation, speech act hedges, and idiomatic expressions. Under speech act hedges, several sub types were further suggested. The fuzzy representation can be supported for at least two reasons. First, taxonomic categorization of a substantial number of examples is often fuzzy, strongly indicating that they can bear multiple functions simultaneously. Second, the fuzzy representation can better reflect the commonalities of various senses possible with -ciman constructions, whose underlying or enduring sense concerns ‘contrast’. We also demonstrated that construal of a -ciman-based construction is compositionally solicited together with other accompanying linguistic units, principally the predicate and certain particles. This was illustrated by a range of examples, where the predicate or a specific particle mutually contributes to the construal of a -ciman clause.

The so-called “pragmatic” meanings of -ciman constructions, such as denial of expectations and speech acts were accounted for by means of different levels of representations in which the contrast occurs. Furthermore, we speculated that the ending -ciman can be employed as a stance/discourse marker together with certain co-occurring idiomatic expressions. Some fixed expressions suffixed with -ciman can be used to express speakers’ concessive attitude toward an assertion and thus allow them to moderate the burden of asserting the message in a proper way.

Notes

1.Following the tradition in pragmatics literature (e.g., Sperber and Wilson 1986/1995Sperber, Dan, and Deirdre Wilson 1986/1995Relevance: Communication and Cognition (Second edition with postface). Oxford: Blackwell.Google Scholar), the speaker is referred to as she and the hearer as he, unless the genders of interlocutors are revealed from the context.
2.As one review correctly points out, the process of compositional interpretation needs to be empirically supported for it to be legitimately claimed. While agreeing with the comment, I assume that a general/agreed-upon interpretation of an utterance containing -ciman in a specific context can be derived by an “ordinary speaker” who is competent in understanding language and the context.
3.For various types of expectations, see Fraser (1998) 1998 “Contrastive Discourse Markers in English.” In Discourse Markers: Descriptions and Theory, eds. by A. H. Jucker, and Y. Ziv, 301–326. Amsterdam: Benjamins. DOI logoGoogle Scholar.
4.For example, the formulaic expression, mianha ‘be sorry’-ciman, barely allows the suppositive particle -keyss to co-occur with it. Of course, these two can be combined when intended literally.
5.One reviewers commented how my unitary approach can better explain the multiple meanings possible with -ciman than does the null hypothesis that it is polysemous and that various interpretations are derived from these multiple senses stored in the mental lexicon. Both types of approaches can be equally viable, differing in terms of viewing the multiple senses of an expression.
6.For the transcription of the Korean data, the Yale Romanization was used. The abbreviations, which are borrowed and modified from Sohn (1999Sohn, Ho-Min 1999The Korean Language. Cambridge: Cambridge University Press.Google Scholar, 2013 2013Korean. Seoul: Korea University PressGoogle Scholar), are presented in the Appendix. Because of the limited space, morpheme-by-morpheme glossing is sometimes provided only for relevant parts. File numbers in the Sejong Spoken Corpus are given in square brackets after examples.
7.“This is one of the mandatory dishes on traditional holidays or special festive occasions. Also called nokdujijim or binjatteok, bingdaetteok is made by peeling and soaking mung beans, grinding them, and then pan-frying with various vegetables.” (Retrieved on 27 August, 2019 from https://​terms​.naver​.com​/entry​.nhn​?docId​=5700793​&cid​=42701​&categoryId​=63067)
8.Stance markers are classified as a sub-category of pragmatic markers by Carter and McCarthy (2006)Carter, Ronald, and Michael McCarthy 2006Cambridge Grammar of English: A Comprehensive Guide. Cambridge: Cambridge University Press.Google Scholar. In linguistics literature, the same marker has been dubbed in different terms, including, but not limited to, discourse markers (e.g. Schiffrin 1987Schiffrin, Deborah 1987Discourse Markers. Cambridge: Cambridge University Press. DOI logoGoogle Scholar; Fraser 1990Fraser, Bruce 1990 “An Approach to Discourse Markers.” Journal of Pragmatics 14: 383–395. DOI logoGoogle Scholar), discourse particles (e.g. Schourup 1985Schourup, Lawrence 1985Common Discourse Particles in English Conversation. New York: Garland.Google Scholar; Mosegaard Hansen 1998Mosegaard Hansen, Maj-Britt 1998The Function of Discourse Particles: A Study with Special Reference to Spoken Standard French. Philadelphia: Benjamins. DOI logoGoogle Scholar), and discourse connectives (e.g. Blakemore 1987Blakemore, Diane 1987Semantic Constrains in Relevance. Oxford: Blackwell.Google Scholar).

Abbreviations

ac

accusative particleending

dc

declarative sentence-type ending

in

indicative mood suffixending

md

pre-nominal modifier suffix

nm

nominative particleending

nom

nominalizer suffix

pst

past tense suffixending

sup

suppositive mood suffixending

tc

topic-contrast particle

References

Baranzini, Laura, and Alda Mari
2019 “From Epistemic Modality to Concessivity: Alternatives and Pragmatic Reasoning per absurdum”. Journal of Pragmatics 142: 116–138. DOI logoGoogle Scholar
Bell, D. M.
1998 “Cancellative Markers: A Core/Periphery Approach.” Pragmatics 8: 515–541. DOI logoGoogle Scholar
Blakemore, Diane
1987Semantic Constrains in Relevance. Oxford: Blackwell.Google Scholar
1989 “Denial and Contrast: A Relevance Theoretic Analysis of But.” Linguistics and Philosophy 12: 15–38. DOI logoGoogle Scholar
Brown, Penelope, and Stephan Levinson
1987Politeness: Some Universals in Language Use. Cambridge: Cambridge University Press. DOI logoGoogle Scholar
Carter, Ronald, and Michael McCarthy
2006Cambridge Grammar of English: A Comprehensive Guide. Cambridge: Cambridge University Press.Google Scholar
Choi, Hyenpay
1937Wulimalpon. [Korean Grammar]. Seoul: Top Publishing.Google Scholar
Dascal, Marcelo, and Katriel Tamar
1977 “Between Semantics and Pragmatics: The Two Types of ‘but’ – Hebrew ‘aval’ and ‘ela’.” Theoretical Linguistics 4: 143–172. DOI logoGoogle Scholar
Fraser, Bruce
1990 “An Approach to Discourse Markers.” Journal of Pragmatics 14: 383–395. DOI logoGoogle Scholar
1998 “Contrastive Discourse Markers in English.” In Discourse Markers: Descriptions and Theory, eds. by A. H. Jucker, and Y. Ziv, 301–326. Amsterdam: Benjamins. DOI logoGoogle Scholar
Gray, Bethany, and Douglas Biber
2014 “Stance Markers.” In Corpus Pragmatics, eds. by Karin Aijmer, and Christoph Rühlemann, 219–248. Cambridge: Cambridge University Press.Google Scholar
Grice, Paul
1989Studies in the Way of Words. Cambridge, MA: Harvard University Press.Google Scholar
Halliday, M. A., and R. Hasan
1976Cohesion in English. London: Longman.Google Scholar
He, Wung
1977Wuliyeysmalpon [Grammar of Old Korean]. Seoul: Saymmwunhwasa.Google Scholar
Helmer, Henrike, Silke Reineke, and Arnulf Deppermann
2016 “A Range of Uses of Negative Epistemic Constructions in German: ICH WEIß NICHT as a Resource for Dispreferred Actions.” Journal of Pragmatics 106: 97–114. DOI logoGoogle Scholar
Jaszczolt, Kasia
2005Default Semantics: Foundations of a Compositional Theory of Acts of Communication. Oxford: Oxford University Press. DOI logoGoogle Scholar
2009Representing Time: An Essay on Temporality and Modality. Oxford: Oxford University Press.Google Scholar
2016Meaning in Linguistic Interaction. Oxford: Oxford University Press. DOI logoGoogle Scholar
Lakoff, Robin
1971 “If’s, and’s and but’s about Conjunction.” In Studies in Linguistic Semantics, eds. by C. Filmore, and D. Langenden, 114–149. New York: Holt, Reinhard and Winston. 114–149.Google Scholar
Lee, Huyca, and Conghuy Lee
1999Theyksuthu Pwunsekcek Kwuke Emiuy Yenkwu [A Text-Analytic Approach to Korean Endings]. Seoul: HankwukmwunhwasaGoogle Scholar
Lee, Hye-Kyung
2002 “Towards a New Typology of Connectives with Special Reference to Conjunction in English and Korean.” Journal of Pragmatics 34: 851–866. DOI logoGoogle Scholar
2015 “A Corpus-Pragmatic Analysis of Wuli .” Discourse and Cognition 22(3): 59–78. DOI logoGoogle Scholar
Lee, Hyo Sang
1991Tense, Aspect and Modality: A Discourse-Pragmatic Analysis of Verbal Suffix in Korean from a Typological Perspective. Ph.D. Dissertation, UCLA.
Lee, Iksep, and Hongpin Im
1983Kwukemwunpeplon. [Korean Grammar]. Seoul: Tayhaksa.Google Scholar
Lee, Keedong
1993 “The Pragmatic Function of the Connective Nuntey.” Ene [Korean Journal of Linguistics] 5(1): 119–135.Google Scholar
Lee, Keum-Hee
2018Pocosa manun kyelhaphyeng yenkyel phyohyeney tayhaye: -cimanun, -tamanun, -(u)kkamanun lyu cwungsimulo. [Final endings + particle Maneun on coupled forms focused mainly -jimaneun, -damaneun, -eulkkmaneun]. Kwukehak [Journal of Korean Linguistics] 86: 63–88.Google Scholar
Lee, Unkyeng
2000Kwukeuy Cepsokemi Yenkwu [A Study of Korean Conjunctive Suffixes]. Seoul: Thayhaksa.Google Scholar
Lyons, John
1977Semantics. Cambridge: Cambridge University PressGoogle Scholar
Mosegaard Hansen, Maj-Britt
1998The Function of Discourse Particles: A Study with Special Reference to Spoken Standard French. Philadelphia: Benjamins. DOI logoGoogle Scholar
Nam, Kisim, and Yengkun Ko
1983Phyocwunkwukemwunpeplon. [Grammar of the Standard Korean]. Seoul: Top PublishingGoogle Scholar
Park, Jae-Yeon
2014 “Hankwuke Yenkyelemi Uymi Hwakcangeyseuy Hwanyuwa Unyu [Metonymy and Metaphor in the Semantic Extensions of Korean Connective Endings].” Kwukehak [Journal of Korea Linguistics] 70: 117–155.Google Scholar
Park, Yong-Yae
1999 “The Korean Connective Nuntey in Conversational Discourse.” Journal of Pragmatics 31:191–218. DOI logoGoogle Scholar
Schiffrin, Deborah
1987Discourse Markers. Cambridge: Cambridge University Press. DOI logoGoogle Scholar
Schourup, Lawrence
1985Common Discourse Particles in English Conversation. New York: Garland.Google Scholar
Sohn, Ho-Min
1999The Korean Language. Cambridge: Cambridge University Press.Google Scholar
2013Korean. Seoul: Korea University PressGoogle Scholar
Sperber, Dan, and Deirdre Wilson
1986/1995Relevance: Communication and Cognition (Second edition with postface). Oxford: Blackwell.Google Scholar
Sweetser, Eve
1990From Etymology to Pragmatics: Metaphorical and Cultural Aspects of Semantic Structure. Cambridge: Cambridge University Press. DOI logoGoogle Scholar
Weatherall, Ann
2011 “I Don’t Know as a Prepositioned Epistemic Hedge.” Research on Language and Social Interaction 44(4): 317–337. DOI logoGoogle Scholar
Yoon, Phyenghyen
1989Kwukeuy Cepsokemiey Tayhan Yenkwu. [A study of Korean Conjunctive Suffixes]. Ph.D. Dissertation. Cheonnam National University. Korea.
Zafiu, Rodica
2018Epistemic and Evidential Markers in the Rhetorical Context of Concession. Journal of Pragmatics 128: 116–127. DOI logoGoogle Scholar

Address for correspondence

Hye-Kyung Lee

Department of English

Ajou University

206 Worldcup-ro

Suwon, Kyunggi-do 16499

Republic of Korea

[email protected]

Biographical notes

Hye-Kyung Lee is Professor of Linguistics at Ajou University, Korea. She completed a doctoral degree in Linguistics at the University of Cambridge. Her research interests include pragmatics, corpus linguistics, discourse analysis, Korean linguistics and English education. Her current focus is on using corpus data and tools in investigating topics in pragmatics. Her recent articles have appeared in Journal of Pragmatics, Language and Linguistics, and Discourse and Cognition. She served as the editor-in-chief of Discourse and Cognition.