Discourse-pragmatic markers of (inter)subjective stance in Asian languages

Full Text

1. Introduction

The present special issue grew out of a workshop at the 18^th International Pragmatics Conference that took place in Brussels from July 9 to 14, 2023. Organized by Seongha Rhee, Reijirou Shibasaki and Wenjiang Yang, the objective of the workshop was to study the development of words of the same origin, all going back to Chinese, into discourse markers (DMs) having diverse functions in the Asian languages Chinese, Korean, Japanese, Vietnamese, and Thai. The contributors of this issue were asked to explore the extent of commonalities and differences displayed by developmental scenarios of the DMs with shared characteristics in the language or languages analyzed by them.

In accordance with this objective, the present issue deals with a number of words of Chinese origin used in Japanese, Korean, Vietnamese, and Thai, as well as Chinese. It investigates theoretically significant issues related to language contact, discourse-pragmatic aspects of language change, and socio-cultural influence on language development, among others, as exemplified in the development of DMs from their earlier lexical expressions originating from Chinese etyma.

As has been established in a body of research (e.g., Rhee et al. 2021, Higashiizumi & Shibasaki in preparation), a number of words of Chinese origin are used as DMs in Asian languages. The five Asian languages mentioned all belong to different language families, i.e., Sino-Tibetan (Chinese), Japonic (Japanese), Koreanic (Korean), Austroasiatic (Vietnamese), and Kra-Dai (Thai), and they have widely varied typological profiles in syntax and morphology. Notwithstanding their genealogical and typological differences, they share an important commonality: They all belong to the Sinosphere Space a fact that is reflected in a number of features of individual languages, especially after “Sinosphere” in their lexicon.¹

The Japanese lexicon, for instance, includes a large inventory of Sino-Japanese words; for example, they account for approximately 60 percent of the total word count in practical and popular science magazines (Shibatani 1990: 142–145). Similarly, Sino-Korean words account for 53 per cent of the headwords in an official dictionary (NIKL, n.d.) – according to Sohn (2001[1999]: 87), the number of Sino-Korean words amounts even to 60 per cent, thus being considerably larger than that of native Korean words, which account only for 35 per cent of the Korean lexicon.²

Vietnamese people have long had close interaction with Chinese and used Chinese characters until the current Latin-based script was adopted. Vietnamese has a massive inventory of ‘literary’ Sino-Vietnamese borrowings, but there are as well ‘colloquial’ Sino-Vietnamese borrowings (Alves 2007). Sino-Vietnamese words account for 60 to 70 per cent of Vietnamese lexicon (Nguyễn 1961, Alves 2001, 2009). Similar to Vietnamese, Thai people also have had interaction with Chinese for millennia, using Chinese characters before they began to use the current Thai script (Haarmann 2012[1986]). Their language is also known to have a large inventory of Sino-Thai words (SEALANG, n.d.).

The goal of this special issue is to demonstrate the substantial influence that Chinese had on the languages of the Sinosphere through written texts. The paper is structured as follows. Section 2 gives a concise characterization of the key notions of the present collection of contributions which are ‘discourse marker’, ‘grammaticalization’ and ‘cooptation’. Section 3 then provides a general overview of the contributions making up this special issue. Section 4 presents a discussion of the mechanisms underlying the changes examined, followed by some general observations in Section 5. And finally, the conclusions reached are summarized in Section 6.  

2. Discourse markers, grammaticalization and cooptation

For a better understanding of the analyses to be presented in the contributions to follow, three key notions employed there are briefly discussed in this section. The term ʽdiscourse markerʼ (DM) is used for a wide range of phenomena and is referred to with a variety of different terms. More than forty terms have in fact been identified (e.g., Dér 2010, see also Dér & Markó 2010), and nearly a dozen of them are presently in common use. DMs serve to monitor the production of texts and to provide processing instructions on how to interpret texts. A classic definition of them is the one in (1), taken up in a similar format in more recent work, as in (2).

(1)   [Discourse markers are] sequentially dependent expressions which bracket units of talk. (Schiffrin 1987: 31)

(2)   By a DM I mean a metatextual marker that signals some kind of relationship between clauses/utterances. (Traugott 2018: 27)

As has been shown more recently, however, text organization is neither the only function associated with DMs, nor is it always really a feature of them, as when we say in English, John is, well, a liar, where the item well, commonly classified as a DM, serves a function other than that of text organization. Instead of (1) and (2), the more complex definition in (3) has been proposed to take care of salient functions of these markers.

(3)   Discourse markers are (a) invariable expressions which are (b) semantically and syntactically independent from their environment, (c) set off prosodically from the rest of the utterance in some way, and (d) their function is metatextual, relating a text to the situation of discourse, that is, to the organization of texts, the attitudes of the speaker, and/or speaker-hearer interaction. (Heine et al. 2021: 6)

Of all the terms used instead of ‘discourse marker’ it is ‘pragmatic marker’ that has received the widest currency. The latter term can be found in a number of different usages, most of all the following three:³ (a) as an equivalent of DM, (b) as referring to phenomena other than the ones covered by a DM, and (c) as a more general category that includes DMs. To the extent that the relevant authors do not make it quite clear which of these three usages is intended in their work, the term has at times given rise to confusion.

This does not apply to the study of Higashiizumi and Shibasaki (in preparation), who use ‘pragmatic marker’ in the sense of (b). For them, ‘discourse markers’ (in their sense) can connect both preceding and following information textually, whereas ‘pragmatic markers’ show the speaker’s attitude to the preceding or the following utterance, not necessarily connected to the following discourse. On this view, corpus-based and text-based evidence from East Asian languages examined in Higashiizumi and Shibasaki (in preparation) suggests that diachronically, some ‘discourse markers’ have the potential to become ‘pragmatic markers’, thereby moving from the textual to the interpersonal level. According to Traugott (1982), a change in the opposite direction seems unlikely in language change in general. Of course, it is hard to determine at some point in history whether or not, or to what extent, a given marker shows or does not show the attitudes of a speaker (or writer). Such examples can be considered as cases of ‘layering’ (Hopper 1991) or ‘bridging context’ (Heine 2002).

In contrast, in the study of Traugott (2022: 5, Figure 1.2), reviewed in contribution (11) (see Section 3), ‘pragmatic marker’ is used in the sense of (c), in that it includes discourse structuring markers, which again include what she calls ‘multifunctional DMs’. In accordance with the definition in (3), both pragmatic markers and discourse structuring markers are treated here within the general category of DMs.

While research on DMs is relatively young, that on grammaticalization has a distinctly longer history, going back to the 19^th century. Its employment for describing the development of DMs, however, is of more recent origin (see especially Traugott 1982, 1995).

In more general terms, grammaticalization is based on the context-induced manipulation of linguistic expressions in discourse (cf. the context extension model of Heine 2002). A wealth of definitions of it has been proposed, and the one in (4) is in the spirit of most of these definitions. As stipulated in this definition, grammaticalization is a unidirectional process, which means that cases not conforming to the definition are not within the scope of grammaticalization theory, to be accounted for in terms of alternative mechanisms of linguistic change.

(4) Grammaticalization is defined as the development from lexical to grammatical forms and from grammatical to even more grammatical forms. Since the development of grammatical forms is not independent of the constructions to which they belong, the study of grammaticalization is also concerned with constructions and with even larger discourse segments. (Heine & Kuteva 2002: 2, see also Narrog & Heine 2021)

As for the criteria used to identify instances of grammaticalization, there are especially three proposals, namely those of Lehmann ([1982] 2015: 132), relying on six criteria or parameters, of Hopper (1991: 25–9), defining five principles of change, and of Heine and Kuteva (2007: 33–44), using four parameters for identifying instances of grammaticalization. Since the last proposal addresses most of the concerns of the other two approaches, it is adopted here. The parameters distinguished are listed in (5). With the exception of (5a), they have a focus on loss of linguistic substance in the process; their main purpose is to assist in the identification of instances of grammaticalization rather than defining it; for the latter, see (4). Note, however, that in the same way as there is loss in the process of grammaticalization, there are also gains (cf. (5a)). In fact, the present contributions bear witness to the observation made throughout the contributions and elsewhere that these gains can be and commonly are indeed fairly rich.

(5) The parameters of grammaticalization proposed by Heine and Kuteva (2007: 33–46)

1. 1. Context extension: The rise of new meanings when linguistic expressions are extended to new contexts, leading to context-induced reinterpretation.
   2. Desemanticization (ʽsemantic bleachingʼ): Loss or generalization of meaning content or functions in such contexts.
   3. Decategorialization: Loss of morphosyntactic properties characteristic of lexical or other less grammaticalized forms.
   4. Erosion (ʽphonetic reductionʼ): Loss of phonetic substance, including prosodic features.

The third notion figuring in the contributions to follow is cooptation (e.g., Heine et al. 2013, 2017, 2021: 26–27, Furkó 2014). Introduced only recently, this notion addresses a feature of DMs that is not within the scope of grammaticalization theory, namely their metatextual function of discourse management, addressed in the definitions of (2) and (3). Grammaticalization and cooptation thus have starkly contrasting effects: Whereas the former typically entails increasing integration of linguistic items within the sentence or the word, the latter enables language users to extend the use of such items beyond the confines of a sentence. For example, English beside(s) experienced a process of grammaticalization from adverbial phrase to adverb, all happening within a sentence, but was later on coopted as a DM whereby it now serves discourse organization beyond the sentence (Traugott & Trousdale 2013: 109–12, Heine et al. 2021: 29).

Cooptation is described as a cognitive-communicative operation enabling speakers or writers to switch their perspective from the level of reasoning anchored to the meaning of sentences to the metatextual level of reasoning immediately anchored to the situation of discourse (Heine et al. 2021: 67, cf. Rhee 2013). Its main effects on the text segments coopted are summarized in (6).

(6) Common effects of cooptation (Heine et al. 2021: 68)

1. 1. Meaning: From meaning as part of the sentence to meaning outside the sentence
   2. Function: From sentence-structuring function to metatextual function
   3. Syntax: From syntactic constituent of the sentence to syntactically unattached status
   4. Prosody: From prosodically integrated to unintegrated or less integrated status
   5. Semantic-pragmatic scope: From more restricted to wider scope
   6. Placement: From positionally constrained to less constrained placement

Beyond the three terms discussed in this section, there are other terms that are equally used commonly for describing the rise and development of DMs, especially terms like pragmaticalization and constructionalization. However, since such terms seem to be less relevant to the analyses proposed in the contributions to follow, we are not discussing them in this section.

Furthermore, it should be emphasized that, on the one hand, the development of DMs from lexical material presumably constitutes the predominant pathway characterizing the history of DMs. On the other hand, this constitutes by no means the only way. Other fairly common sources of DMs are provided, for instance, by straightforward borrowing, involving neither grammaticalization nor cooptation. Discussing a wide range of cases from languages across the world, Heine et al. (2021: 215) note:

It would seem in fact that DMs are amongst the first grammatical items that speakers borrow or code-switch in situations of intense language contact, frequently but not always from the language of the more dominant or ʽprestigiousʼ group involved (Matras 1998, Grant 2012).

Furthermore, there are also two other pathways whereby DMs can arise, namely either via the grammaticalization of interactive categories such as vocatives, interjections, directives and attention signals (Heine 2023: 277–297), or via general extenders (Brinton 2017: 272–283, 2024, where ‘pragmatic markers’ are used instead of ‘discourse markers’). For the former, English provides a few examples, like the interjection oh, as observed, for example, by Jucker (2002: 218): “To this extent, the Early Modern English oh is a proper discourse marker”. Note, however, that in a process of grammaticalization, the earlier function tends to be retained side-by-side with the new grammaticalized function, and this is exactly what appears to have happened with oh: While it developed into a DM in certain contexts, in many other contexts its function as an interjection is still alive and well.

For the latter, Brinton (2017: 281–282, see also Brinton 2024, Brinton & Shibasaki forthcoming) identifies the following pathway: the parenthetical use of or whatever becomes the general extender or/and/like whatever, finally giving rise to the stand-alone ‘pragmatic marker’ whatever.⁴ However, cases of interactive categories and general extenders are not dealt with in the contributions to follow and are therefore not considered further in this paper.

3. The contributions to this special issue

The contributions to this special issue cover a wide range of topics, and they are now looked at in turn.

The first contribution, by Wenjiang Yang, provides a contrastive diachronic analysis of the Chinese form yuánlái ‘originally, previously’ and Japanese ganrai ‘originally, inherently’, tracing their development from temporal adverbials to DMs. The Japanese form was borrowed from the Chinese one through written texts, both having been temporal adverbials meaning ‘originally, from the beginning’, and this meaning has persisted till today. The development of the two into DMs followed contrasting pathways: Chinese yuánlái turned into a DM encompassing mirative, background and justificational functions, whereas Japanese ganrai eventually evolved into an elaborative marker. This case study relates to several topics, to be discussed in Section 4. First, borrowing was restricted to lexical material and it involved the written rather than a spoken mode. Second, the shift from sentence grammar into the domain of discourse organization appears to have happened independently in the two languages. And third, the change can be viewed as one of parallel grammaticalization rather than replica grammaticalization, in that it took quite different directions and lines of semantic-pragmatic change in the two languages.

The second contribution deals with an issue of a different kind. In their diachronic analysis of Chinese shènzhì ‘even’ and bùguò ‘however, but’ in their various usages, Haiping Long and Lei Wang classify these two forms as discourse structuring markers, that is, as a weakly grammaticalized type of DMs. The authors show that the two forms experienced different ways of development. Whereas shènzhì was originally used in clause-initial position, expressing speaker-oriented meanings, it acquired clause-medial usages in later stages of its development. These features are not shared by bùguò: While also emerging in clause-initial position, it neither conveyed speaker-oriented meanings nor did it acquire clause-medial usages. Based on this observation, Long and Wang identify an interesting correlation between meaning and syntactic change in grammaticalization, in that change from initial to medial position can be accounted for with reference to the presence of early speaker-oriented meaning. The authors find further support for this hypothesis in developments to be observed in English, also confirmed by similar developments in Korean and Hungarian. This case study differs from the other contributions mainly in two respects. First, it does not involve language contact and, second, it focuses on DM-internal developments rather than on the question of how discourse structuring markers or, more generally, DMs arise. But its conclusions are noteworthy, demonstrating that the source meaning of a DM can determine the course of its subsequent grammaticalization and usage.

The topic of the third contribution, by Seongha Rhee and Lin Zhang, is devoted to a systematic comparison between a Korean DM and its etymological source in Chinese. The Korean DM cincca, literally meaning ‘a true thing’, is historically a noun phrase which was grammaticalized into an item carrying adjectival and adverbial functions of adding genuineness or excellence in quality to a modified noun, or an adverbial function of adding emphasis to an adjective or a predicate. Turning into a DM, it retained its intensifying function but developed various new functions. Its corresponding cognates in Chinese are provided by the DMs zhende and zhenshi, having similar functions, but zhenshi differs in one important respect, marking the speaker’s negative evaluation of the person or event it refers to. The authors demonstrate that the Korean and the Chinese DMs underwent massive grammaticalization processes, resulting in both similarities and differences with regard to their functional distribution, prosody, and relative degree of desemanticization, that is, their loss of earlier meaning features. While both the Chinese and the Korean items are DMs, the evidence available suggests that the rise of Korean cincca as a DM was a language-internal process taking place some time after the turn of the 20^th century.

In the fourth contribution, Yuko Higashiizumi, Reijirou Shibasaki and Keiko Takahashi deal with a pathway of grammaticalization that appears to also have occurred in some other languages, as shown in the World Lexicon of Grammaticalization (Kuteva et al. 2019: 443). The authors trace the development of the Japanese adverb shinni ‘truly’, containing the Sino-Japanese noun shin ‘truth’. The adverb serves, for example, to intensify the illocutionary force of an apology or a regret, akin to English very. While the development of Japanese shinni was restricted to grammaticalization, the authors note that similar processes can also be observed in Chinese, Korean and Thai but leading one step further, giving rise to DMs, now having interactional functions of discourse processing.

The fifth contribution by Jiyeon Park is about semantic-pragmatic manipulation of the Japanese adverb zenzen ‘completely, entirely, not at all’, borrowed from Chinese. The author argues that zenzen evolved as a DM (‘pragmatic marker’) indicating the speaker’s epistemic stance and viewpoint. Starting as a lexical item with objective meanings, zenzen underwent cooptation and turned into a DM having subjective and intersubjective functions. In spoken discourse, it is also found in a specific construction where it combines with a copula. Functioning as an adjectival noun, its syntactical independence from the sentence gives it the appearance of a complex DM or, more generally, of a thetical (Kaltenböck et al. 2011). Park’s main concern is with semantic change in grammaticalization leading to the expression of subjectivity and intersubjectivity, but she also addresses features in development of zenzen that are suggestive of cooptation as it is characteristic of the rise of DMs (see Heine et al. 2017, Heine et al. 2021).

The growth of DMs from lexical sources is also the subject of Hyun Sook Lee in the sixth contribution. Originally appearing in the 15^th century as a noun for ‘the world’ or ‘the world or society people live in’, the item was extended semantically, receiving a range of additional meanings, and it grammaticalized into an adverbial degree modifier (‘very’) and negative polarity item (‘at all’). When combining with the locative particle -ey (seysangey), it became a DM in Modern Korean, being syntactically independent and enjoying positional freedom.

This study of seysang and seysangey discusses a number of issues, illustrating in particular the growth of subjective and intersubjective meanings and other innovations characteristic of the shift from material of sentence grammar to status as a DM.

Another process leading from lexical item to DM is discussed by Sunhee Yae in the seventh contribution. Ultimately a noun in Late Middle Korean, the Sino-Korean item iltan ‘one morning’ was grammaticalized into an adverb and a connective, acquiring new meanings associated with priority, short duration and conditionality. The author goes on to also analyze the Chinese equivalent yīdàn of Korean iltan. Differences between the two relate most of all to the following: Chinese yīdàn seems to be less strongly grammaticalized. While it acquired uses to express brief duration and conditionality, it does not express priority.

The development of iltan displays the whole gamut of processes to be expected in the evolution of DMs, that is, grammaticalization and cooptation as well as semantic-pragmatic developments into the expression of subjectification and intersubjectification. At the same time, it does not confirm the hypothesis of a functionally motivated asymmetry between the left periphery and the right periphery of DM placement – an observation also made by Hyun Sook Lee in contribution 6. Korean iltan and Chinese yīdàn seem to provide an instance of parallel grammaticalization, where two items having the same lexical source underwent a development in the same direction, even though this development was more pronounced in Korean.

Moving from Korea to Thailand, the eighth contribution by Kultida Khammee deals with the Thai lexeme ciŋ ‘true’, borrowed from Middle Chinese cin ‘true, real’. Grammaticalization had the effect that a lexical item denoting ‘true, truth, real, genuine’ assumed adverbial functions as an intensifier denoting ‘surely, certainly, definitely’. But it also appears to have acquired a number of DM functions expressing intersubjective meaning. The development of ciŋ was apparently not restricted to grammaticalization; ciŋ also must have undergone cooptation, giving rise not only to a DM but also to response signals (Heine 2023), in that it turned into the interactive forms ciŋ ‘yes’ and its negative counterpart mây ciŋ ‘no’, expressing agreement and disagreement, respectively. All evidence that there is suggests that these changes of Thai ciŋ were essentially language-internal, not shared by its Chinese etymological source item cin.

The ninth contribution by Mayumi Adachi deals with a topic related to several other contributions, but in another language of the Sinosphere, namely, Vietnamese. Vietnamese thật has its origin in a Sino-Vietnamese lexeme meaning ‘full, rich, fruit’, but this meaning was extended to also express ‘truth, reality, and fact’. Occurring as an adjective and adverb in the 17^th century, it is now found in various spoken and written contexts, including uses as an adjective ‘real’ and an adverb ‘really’. But it is found also in DMs in utterance-initial position in the forms thật ra (thật + ‘out’), quả thật (‘fruit’+ thật), and kỳ thật (‘its’+thật), all of which are glossed ‘in fact’. These combinations seem to be vital to attracting the attention of potential interlocutors, thus constituting a case of intersubjectification.The author argues that thật experienced a process of grammaticalization leading from denoting obvious facts to expressing the speaker’s affective or evaluative stance. Once again, we seem to be dealing with a language-internal change involving grammaticalization but presumably also cooptation as a DM.

The remaining contributions 10 and 11 are of a different kind: They present reviews of books that have appeared more recently and provide new theoretical perspectives on the analysis of DMs and their development.

The first book, which is the subject of the tenth contribution, is The Rise of Discourse Markers (Heine et al. 2021), reviewed by Fangqiong Zhan. The authors of the book note that some previous works on DMs have suffered from not accounting for salient properties of DMs, such as their relatively independent semantic, syntactic and prosodic status. Studying DMs in languages spoken in various parts of the world, but especially in English, French, Japanese and Korean, they argue that the presence of such properties can be explained in terms of cooptation, a cognitive–communicative operation whereby text segments such as clauses or parts of them are transferred from the level of sentence organization to the domain of discourse organization. The development of DMs thus is explained by a combination of two distinct mechanisms, namely grammaticalization and cooptation. In her careful analysis of the book, Zhan draws attention in particular to the fact that the book could have benefitted from taking into account findings made on constructional change, especially as it is described in the framework of Diachronic Construction Grammar (e.g., Traugott & Trousdale 2013, see also Hilpert 2013 and Barðdal et al. 2015).

The final paper, contribution 11, is a review by Foong Ha Yap and Mikyung Ahn of the book Discourse Structuring Markers in English (Traugott 2022). In the book, the author turns to construction grammar, offering a diachronic perspective on how new functional categories are constructionalized, and suggests how a network model can account for the interlocking patterns in language change. The main theoretical question asked is how pragmatic/contextual factors can best be incorporated in Discourse Construction Grammar, the framework proposed by the author. A constructionalist perspective is provided based on reconstructions of a number of different types of English DMs, more specifically of discourse structuring markers (see also the review of Heine & Long 2022). Whereas Heine et al. (2021) focus on two specific phenomena in the development of DMs, namely directionality and cooptation, the author of this book provides a wider framework for the reconstruction of DMs. Given the relevance of the framework expounded in Traugott (2022), the reviewers suggest that the book calls for new studies investigating discourse structuring markers also in other languages. Unlike the book reviewed in the preceding contribution (Heine et al. 2021), this book is restricted to looking at one particular language, that is, English.

4. Mechanisms of change

As the preceding summary of the various contributions suggests, the development of DMs that have their origin in Chinese etyma, as observed in Japanese, Korean, Thai and Vietnamese, provides uniquely valuable insights into the role played by conceptual and cognitive operations as well as socio-cultural and discourse-pragmatic strategies in language change. With regard to this development, a couple of more general distinctions need to be made.

First, the development of a number of the DMs examined involve two main stages. At stage 1, there was contact-induced transfer, more precisely borrowing of Chinese lexical material into the other four languages examined. And at stage 2, there was internal development within each of the languages. Obviously, developments within Chinese only involved the second stage (e.g., Long & Wang 2024). The meanings of the borrowed items at stage 1 tended to be close to those of the source language Chinese. Subsequently, at stage 2 then, the borrowed items underwent diverse changes in their form and meaning within each of the languages concerned, a process known as ‘internally-motivated’ change (Heine & Kuteva 2005: xii). And second, with regard to their status as DMs, two different kinds of development need to be distinguished. On the first development, the relevant item was already a DM in the donor language Chinese and borrowing simply involved transfer of the DM from the donor language to the recipient language. Various lines of research have in fact established that borrowing of DMs is very common in the languages of the world (see Matras & Sakel 2007, Heine et al. 2021: Chapter 7). Developments in the recipient languages Korean, Japanese, Vietnamese, and Thai then would have concerned mainly grammaticalization, whereby a borrowed DM was extended to new contexts and acquired novel grammatical functions.

Some particular meanings appear to have their own potential of triggering grammaticalization independently from language contact and regional areas (e.g., ‘fact,’ ‘result,’ ‘issue,’ to name but a few; see Higashiizumi & Shibasaki in preparation for details). As was shown in various contributions to the present issue, the developments of the DMs of Chinese origin exhibit such scenarios at varying degrees.  

On the second development, borrowing entailed the transfer of lexical material from Chinese, and the rise and development of the DM took place subsequently within the donor language. And here again, two different pathways need to be distinguished: Either the development from lexical item to DM took place simultaneously with its borrowing, that is, the development into a DM coincided with the process of borrowing. Alternatively, a lexical item of Chinese was borrowed as lexical material in the recipient language and the development from lexical item to DM took place subsequently in the recipient language. In the latter case, the development can follow two contrasting pathways: It either involves ‘replica grammaticalization’, namely a process whereby a grammaticalization process in the recipient language is realized based on a model process in the donor language. Alternatively, grammaticalization takes place independently of one another in the two languages, that is, there is ‘parallel grammaticalization’ (Heine & Kuteva 2005: 92–100). More recent research suggests that replica grammaticalization is crosslinguistically not highly common.

Another important issue relates to the nature of contact between languages. Typically, DMs are used in spoken discourse (Jucker 2002). Therefore, if a DM is borrowed from one language to another, it would imply borrowing through spoken contact (e.g., you know in Clyne 2003). However, spoken language is for the most part not accessible to historical analysis via written documents. Situated in a geographically connected land mass, China, Vietnam and Thailand have long maintained contact. Direct contact between Chinese and Vietnamese and between Chinese and Thai has continued from historical times, involving migration and intermarriage, though the documentation of early contact is scanty.

On the other hand, the heavy lexical borrowing of Chinese words in Japanese and Korean was channeled extensively through written texts, as these peninsular and insular countries are more distantly located for contact as compared to Vietnam and Thailand. The Japanese and Korean cases, thus, present a somewhat special contact situation not widely found in the languages of the world (see, e.g., Norman 1988, Sohn 1999, Irwin & Zisk 2019, Narrog et al. 2018). It is widely assumed that the rise and development of DMs is a matter of the spoken mode of speech. While this is presumably true most of the time across the world, such processes can be channeled as well through the written media, as can be seen in the contributions to follow (see also Shibasaki and Higashiizumi, in preparation).

Another significant theme addressed in this special issue relates to (inter)subjective stance-marking and stance-taking, often heavily influenced by the culture of the speech community (Englebretson 2007). Situated in interactive discourse, DMs signal both a subjective stance, one that is based on the evaluation of a subject, and an intersubjective stance, based on some form of alignment with the interlocutor (Traugott 1982, Du Bois 2007, Kaltenböck et al. 2011). The grammaticalization process of DMs may be widely variable across individual forms and languages, but it has the following shared feature in common: The DMs, by virtue of being markers functioning at the level of metadiscourse, or macrostructure, are formed by discourse-pragmatic strategies to fulfil the speaker’s needs to be efficient, persuasive, and expressive (cf. Hyland 1998, Heine et al. 2021).

5. Discussion

As the observations made in Section 3 suggest, there are two main themes in this special issue. One concerns diachronic reconstruction. Based on the analysis of historical text data and linguistic reconstruction work, the contributions to follow all pursue much the same goal, namely contributing to a deeper understanding of grammatical development in the languages of East and Southeast Asia. The primary incentive underlying most of the contributions is to present a thorough description of language-internal processes as they can be accounted for especially with reference to principles underlying grammaticalization.

A topic to be addressed in more detail in future research concerns the stage of transition from lexical coding anchored to the structure of sentences to coding immediately anchored to the situation of discourse. This transition has been described in terms of cooptation (see Section 2) but more historical data are needed to understand the exact nature of the process. For example, what induced speakers or writers of Chinese, Japanese, Korean, Thai and Vietnamese to introduce new ways of structuring their texts, as well as of expressing their attitudes and their relations with their hearers or readers, respectively? And what exactly were the semantic, syntactic, and prosodic effects of this process?

The second theme in the collection of contributions is language contact. All the linguistic items examined can ultimately be traced back in some form or other to Chinese. Their extension from the Chinese donor language to the four recipient languages, i.e. Japanese (Yang 2024, Higashiizumi et al. 2024, Park 2024), Korean (Rhee & Zhang 2024, Lee 2024, Yae 2024), Vietnamese (Adachi 2024) and Thai (Khammee 2024) involved borrowing, that is, transfer of form-meaning units from one language to another; replication, that is, the transfer of meaning and/or structure but not form, seems to have played only a minor role, if at all. As we saw in Section 3, the process following borrowing was apparently for the most part one of internal grammaticalization, no clear case of replica grammaticalization has been reported. That the overall development is one of borrowing of lexical material from Chinese, which subsequently undergoes grammaticalization and cooptation in the recipient language, not involving replica grammaticalization, has also been observed in studies beyond the present contributions (Higashiizumi & Shibasaki in preparation). These studies also confirm another observation made here, namely that this process seems to always have taken place independently in each of the recipient languages.

As Section 4 showed, the nature of borrowed items can be of two kinds: Either that item was already a DM in the donor language and as such underwent grammaticalization and cooptation subsequently in the recipient language (Japanese, Korean, Vietnamese and Thai), or borrowing consisted of lexical material from the donor language (Chinese). While borrowing of DMs is
a worldwide common process, there seems to be no clear case of it in the studies examined here; rather, the process seems to have been generally one where Chinese provided lexical material to be changed subsequently in the recipient languages.

In addition to their focus on detailed description and reconstruction work, the contributions also deal with general theoretical issues relating to the growth of discourse structuring material. For example, some authors (Rhee & Zhang 2024, Lee 2024, Yae 2024, Khammee 2024) take issue with the claim made in various earlier studies according to which the development of DMs entails structural asymmetry between two kinds of usages -- a hypothesis that has received quite some attention in work on discourse organization (Beeching et al. 2009, Detges & Waltereit 2014, Beeching & Detges 2014, Beeching et al. 2018, Pons Bordería 2018). On this hypothesis, the left periphery of an utterance tends to be associated with subjective functions while the right periphery is dedicated to intersubjective functions. While being appealing in some way, this hypothesis must be taken with care, as the studies by Hyun Sook Lee on Korean seysang and seysangey (Lee 2024) and Sunhee Yae on Korean iltan (Yae 2024) suggest. That the hypothesis is problematic has also been observed by other researchers (e.g., Pons Bordería 2018).

6. Conclusions

What makes this collection of contributions an invaluable document is especially its richness in the documentation and analysis of processes characterizing the history of the languages of East and Southeast Asia.

The combination of languages featuring in this special volume provides an optimal basis for comparative linguistic work. On the one hand, the languages have a number of features in common. They share a long history of language contact resulting in massive borrowing. This history accounts for the presence of a substantial volume of lexical items being similar in form and meaning to be found in the five languages. What makes this also a noteworthy case is the fact that contact-induced change and its effects on the present-day spoken language strongly involved the written mode – thereby contrasting with situations in most other parts of the world where language contact was restricted to the spoken medium (see also Higashiizumi and Shibasaki in preparation).

On the other hand, the five languages analyzed also differ remarkably from one another. For one thing, they are – as far as we know – all genetically unrelated. And for another, they exhibit highly contrasting typological profiles relating to both their morpho-syntactic organization and their phonology. In particular, Chinese, Thai and Vietnamese have predominantly isolating-analytic structures, and all three are tonal languages. Japanese and Korean, by contrast, are characterized by the presence of agglutinating and inflectional word structures, and neither disposes of distinctions in lexical or grammatical tone. Thus, these Asian languages present a close to ideal testing ground for theories of diachronic linguistics.

The testing achieved in the contributions to follow is most of all of the following kind. The findings presented suggest that there is a general line of linguistic development leading from lexical material, such as nominal and adverbial text segments, to the emergence of grammatical units serving new ways of expressing the attitudes of speakers or writers and functions of interaction between speakers and hearers, or writers and readers. But the primary function of the units is to organize the overall structure of discourse, that is, the way texts are constructed and anchored to the situation in which discourse takes place. These units, referred to here as discourse markers (DMs), have been described in previous work also in terms of a variety of alternative theoretical notions, such as pragmatic markers or discourse structuring markers.

The present special issue investigates the extent of commonalities and differences displayed by developmental scenarios of DMs with shared characteristics in multiple languages. It offers a good opportunity to broaden our knowledge on various issues on DMs, and it deserves the attention of researchers working on contact linguistics, interactional linguistics, cultural linguistics, and language change at large. It is hoped that the approach adopted will encourage further comparative research in other regions of the world to contribute to a more comprehensive understanding of how our knowledge of discourse processing in the languages as they are spoken and written today can be explained with reference to the way in which this knowledge evolved over time.

 

_{¹ Thailand was strongly influenced by China early in its history, but at a later time it was more strongly influenced by India, leading to substantial borrowing of words from Sanskrit and Pali. Thus, even though Thailand is normally regarded as belonging to the Indosphere (Haarmann 2012[1986], Kulke & Rothermund 2004), the influence of Chinese on Thai is strongly felt in the modern Thai lexicon (see below).}

_{² The last two figures are adopted from Shibasaki and Higashiizumi (in preparation).}

_{³ See, for example, Heine et al. (2021: 11, fn. 17). See Brinton (2017: 2–10) for a comprehensive definition of pragmatic markers.}

_{⁴ In the same manner, Brinton and Shibasaki (forthcoming) explore another complex case, no matter (what), in the history of English. See Overstreet and Yule (2021) for detailed discussions of general extenders.}

About the authors

Bernd Heine

University of Cologne

Author for correspondence.
Email: bernd.heine@uni-koeln.de
ORCID iD: 0000-0002-9304-7814

Emeritus Professor at the Institute of African Studies (Institut für Afrikanistik und Ägyptologie), University of Cologne, Germany. He has held visiting professorships in Europe, Eastern Asia, Australia, Africa, North America, and South America. His research interests embrace African languages, grammaticalization theory, linguistic typology, language contact, and interactive grammar. He has authored and co-authored numerous books, the most recent of which are The Rise of Discourse Markers (CUP, 2021), and The Grammar of Interactives (OUP, 2023).

Cologne, Germany

Wenjiang Yang

University of Macau

Email: wenjiangyang@um.edu.mo
ORCID iD: 0009-0000-1802-0630

Associate Professor of Japanese linguistics at University of Macau, Macao SAR, China. He received his PhD in Japanese linguistics at Peking University in 2014. His research interests include morphology, syntax, semantics, historical linguistics, and contrastive linguistics, with a special focus on tense, aspect, evidentiality and discourse markers in Present-Day Japanese as well as their diachronic evolution.

Macao SAR, China

Seongha Rhee

Mahidol University; Hankuk University of Foreign Studies

Email: srhee@hufs.ac.kr
ORCID iD: 0000-0003-0312-0975

Professor of Linguistics at Mahidol University, Thailand and Professor Emeritus at Hankuk University of Foreign Studies, Korea. He received his PhD in linguistics from the University of Texas, Austin in 1996. His primary research interest is to identify cognitive and discursive mechanisms that enable language change from the crosslinguistic and typological perspectives. He published World Lexicon of Grammaticalization (co-author, 2019, CUP); book chapters in The Cambridge Handbook of Korean Linguistics (2022, CUP); and research articles in Nature, Journal of Pragmatics, Language Sciences, Russian Journal of Linguistics and Lingua, among others.

Nakhon Pathom, Thailand; Seoul, Republic of Korea

References