In the last three decades there has been extensive work on elements at the “periphery” of linguistic units such as the sentence, clause, or utterance. One line of research, starting with Schourup (1982) and Schiffrin (1987) has focused on discourse markers and pragmatic markers more generally that bracket units of talk, express stance, and serve metatextual discourse functions. Data is usually, but not exclusively, conversational. Therefore, in this line of research, the term “periphery” usually applies to initial or final position in such units as the turn (e.g. Haselow 2012, Degand Forthcoming), or the utterance (e.g. Onodera and Higashiizumi 2013; Detges and Waltereit Forthcoming). Examples include elements like well, y’know:
Well you know, I was really interested in biofeedback.
(2009 NPR Science Program, How Stanley Milgram shocked the world [COCA])
Another line of research, starting with Rizzi (1997), has focused on syntactic structures and information-structuring elements at the periphery of the sentence, e.g. what resources are available in a language to distinguish between topicalization as in (2a) and contrastive focus fronting as in (2b), and how these structures relate to left-dislocation (2c):
Your book, you should give t to Kim (not Jane)
YOUR BOOK you should give t to Kim (not my book)
Your book/YOUR BOOK you should give it to Kim
Much of this work has been done in the framework of syntactic cartography (e.g. Cinque and Kayne 2004), and data is usually constructed.
In both traditions there has been active debate on how to define “periphery”. In both there has been a tendency until recently to focus on initial markers, except when researching Asian languages such as Japanese, Korean, and Chinese, in which roughly equivalent markers tend to occur finally. Recently elements at both peripheries have drawn considerable attention, especially in the literature on pragmatic markers and their historical development (see e.g. Onodera and Suzuki 2007, Beeching and Detges Forthcoming a). A key research question has been whether there is an asymmetry between initial (“left”) and final (“right”) periphery, and if so, what it correlates with. Because the term, “periphery” is used both for elements “inside” the clause as in (2) (the cartography project and most syntactic models) and for elements “outside” the clause as in (1) (research on pragmatic markers), another key question is how to define “periphery” as such and whether there is as sharp a divide between “inside” (“core clause”) and “outside” (“periphery”), as many models imply.
In this paper I address the question of how to define periphery, which requires that some attention be paid to the question about asymmetric functions. Section 2 introduces pragmatic markers. Section 3 reviews some of the models that have been developed in thinking about left and right periphery (LP and RP respectively) from a functionalist discourse perspective. In Section 4 I argue that there is a gradient, not rigidly categorical, distinction between “core clause” and periphery, exemplifying with an outline of the development of general extenders in English. Section 5 serves as a conclusion.
2 Introducing pragmatic markers
In her groundbreaking work on “discourse markers” (DMs), Schiffrin (1987: 31) defined them as “sequentially dependent elements which bracket units of talk”. She proposed that the following conditions have to apply for an expression to be used as discourse marker. It has to:
be syntactically detachable from a sentence,
be commonly used in initial position of an utterance,
have a range of prosodic contours,
be able to operate at both local and global levels of discourse.
In various writings (e.g. 1990, 2006, 2009) Fraser defined DMs as a subcategory of “pragmatic markers” (PMs) and sought to restrict their characterization in a more syntactic way. Fraser (2009) distinguishes DMs from three other classes of PM: those with illocutionary force (I promise, please), those that comment (frankly, reportedly), and those that organize discourse structure (in summary, look!). In all definitions, Fraser specifies that DMs signal sequential discourse organization and their basic structure is S1 – DM + S2, where S1 (sequential unit 1) has a separate illocutionary force from S2, and DM is part of S2.1 He further characterizes DMs as typically lexical (but, and) and non-truth-conditional; most are said to have a separate intonation contour set off by a pause (“comma intonation”). According to his definition, interjections (wow!), sentence adverbs (surely), focus particles (even) and several other pragmatic expressions are not DMs but other subtypes of PMs (p. 7).
In part because of the rather different scope of the definitions of DMs used by Schiffrin (fairly broad) and Fraser (more restrictive), many researchers have found “discourse marker” to be a problematic concept. Currently the larger category of “pragmatic markers” or “pragmatic particles” (Fischer 2006) is the main object of inquiry and the models that are the topic of this paper concern pragmatic markers in general, not discourse markers strictly defined.
Among PMs are some that occur primarily at RP, among them general extenders (e.g. or something, and stuff) and question tags (She left, didn’t she?). Like other PMs, general extenders are procedural; they may serve as backward-looking hedge, topic-closer, or turn-yielder. Aijmer (2002), for example, includes them in her discussion of “discourse particles”. Likewise, question tags are procedural: they are not used to ask yes-no questions, but may be attitudinal (challenging or emphatic) or confirmatory (asking for assent) (Tottie and Hoffmann 2009). The main difference between question tags and traditional PMs is that they do not occur clause-initially and do not derive from lexical items (Hoffmann 2006). Nevertheless, they are considered to be a sub-category of PMs by Fraser (1996) and Andersen (2011), among others. A third category of PM that occurs at RP is that of retrospective contrastive markers like then, however, though (Lenker 2010; Haselow 2012), as in (4):
I took lots of photographs
I don’t know if they are any good though (Haselow 2012: 193 [ICE-GBs1a-009 #002-003])
Haselow (2012: 193) says of (4) that it “retrospectively converts a preceding proposition/claim p into backgrounded information that is in some way incompatible with q”.
3 Some functionalist models of periphery structure
Much recent discussion about periphery has been concerned with whether the LP and RP are functionally symmetric or asymmetric. A structurally symmetric model of periphery has been used for several decades in Japan to account for phenomena in Japanese (see e.g. Minami 1974 and Hayashi 1983, updated in Shinzato 2007, and simplified in Onodera and Higashiizumi 2013). The Layered Structure Model of the Utterance is represented as a set of nested boxes (see Figure 1).
The “proposition” is at the center, “judgment/subjectivity” outside the center, and “communication/intersubjectivity” at the outermost layer. LP is associated with the two outer layers clause initially and RP with the two outer layers clause-finally. LP and RP are therefore represented as symmetric and “outside” the proposition. Each layer is represented as a box surrounding the one(s) inside, and this model allows for (inter)subjective elements to appear within the proposition (though this is not their default position). Although the model is structurally symmetric, the functions of elements at LP and RP are acknowledged to be rather different, e.g. at LP they tend to signal connective functions, but at RP they may express evaluation (Higashiizumi and Onodera 2013).
Focusing on the question whether function is different at LP and RP, Beeching and Detges (Forthcoming b) propose that asymmetry is expected because “discourse unfolds in time”, with the speaker taking up the right to speak and initiating a turn at the beginning and yielding it at the end. One asymmetric model developed by Detges and Waltereit (Forthcoming), draws on models of the French sentence by Morel (2007) and Danon-Boileau et al. (1991). It is a model of the utterance as part of a turn, and combines information structure, syntax and prosody (see Figure 2).
It is particularly designed to account for the use of strong personal pronouns (moi, toi, lui) both initially and finally in turns. A core, the “rheme”, specifies the event structure. The rheme, which is regarded by the authors as obligatory, may optionally be preceded by a “preamble” which is typically used to structure information coherently, to anchor and frame it. The preamble includes several subelements such as “binding” expressions with interactional function such as You see, viewpoint, temporal and spatial frames. Since the default vernacular French sentence is cleft (Lambrecht 1994), the preamble also includes the lexical topic. The rheme may be optionally followed by a “post-script”, which is interactional. In the preamble and post-script no substantial distinction is drawn between elements that are often in other work regarded as PMs (the “binding elements”), and those that would often be regarded as adjuncts (the temporal and spatial frames). While elements in the preamble and postscript may be prosodically detached, they need not be, a point to which I return in Section 4.
A more syntactic model, also designed for French, makes a further distinction between the center, utterance-initial and utterance-final elements, and turn-initial and turn-final elements (Degand Forthcoming, with conversational data based on Valibel (Dister et al. 2009)) (see Table 1).
It specifies positional slots, the central one being called “clause medial position”. This term avoids the evaluative implicature of “core” and the logical one of “proposition”, and highlights the fact that in Degand’s model, elements may appear both before and after the clause. Although the terminology used suggests the structure is symmetric, in fact the functions of elements in the slots are asymmetric: coherence markers play a far larger role in initial than in final position in English and French, for example.
Although each of the models outlined above is somewhat different, and is designed to account for different factors, all have in common the following: LP and RP are conceptualized as structurally symmetric, but functionally asymmetric, slots. There is little or no explicit discussion of how sharp the boundary is between the “outer” periphery and the remainder of the utterance/turn, although terms like LP and RP suggest discreteness.
The last model to be mentioned here is that of Kaltenböck et al. (2011). The authors propose that there are two grammars: one a “canonical sentence-grammar” similar to traditional syntactic grammars in which no attention is paid to PMs, the other a “thetical” grammar,2 both of them subdomains of “discourse grammar” (p. 854). The nucleus of the clause in sentence grammar is “the verb with its argument structure, optionally extended by peripheral participants (or adjuncts)” (p. 877). Thetical elements pertain to the situation of the discourse, e.g. text organization, source of information, attitude of the speaker, speaker-hearer interaction, and discourse settings. This category comprises comment clauses, tag questions, left-dislocations, afterthoughts, interjections, and other elements that are “positionally mobile” and not “licensed by any rule of canonical sentence grammar” (p. 853). Importantly for the discussion at hand, although each grammar has its own internal morphosyntax and prosody, the boundaries between the two are conceptualized as permeable. Structures from one grammar can enter into the other over time bidirectionally.3 The hypothesis that there are two grammars and that elements from thetical grammar can enter sentence grammar and vice versa is one of the main differences between this model and those mentioned above.
are positionally mobile,
do not form an immediate constituent of the clause in which they occur and are loosely connected syntactically with the anchor clause,
typically are set of by a pause (“comma intonation”),
have no impact on the truth value of an utterance,
do not add anything to the propositional content,
have wide scope (modify the whole utterance and not single segments),
cannot be interrogated, negated or focused.
There are counterexamples even in English to each characteristic except (g).
With respect to (a) neither general extenders nor question tags are highly mobile, and retrospective contrastive markers are not mobile at all. With respect to (b) and (c), both syntactic and prosodic integration is attested for many PMs. Indeed, Detges and Waltereit (Forthcoming) say that “a key finding” of recent work about French is that “the peripheries are not typically separated from the core by a pause”, in other words, “comma intonation” is not a cue to periphery, whether LP or RP. Retrospective contrastive though and then have been shown to have a falling, continuous prosody (Haselow 2012) and question tags may be prosodically continuous in many cases (Tottie and Hoffmann 2009; Dehé and Braun 2013). Figures 3 and 4 illustrate prosodically integrated and non-integrated question tags respectively.
In East Asian languages, PMs are typically integrated with the preceding clause (Yap et al. 2008) (d). As far as (e) is concerned, Ifantidou (1994) and others have pointed out that some evidential PMs like I hear, evidently are truth-conditional. Likewise Dehé and Wichmann (2010: 18) show that I think/believe are “semantically variable in their effect on truth conditionality” depending on whether they are prosodically stressed or not. Finally, although scope of PMs is typically over the whole clause (f), in some cases a phrase rather than the whole prior clause seems to be the scope domain, as in the case of the question tag in (6).
It’s a mixture isn’t it of original instruments. (Dehé and Braun 2013: 131 [ICE-GB])
Focusing on the role of PMs in the flow of speech, Beeching and Detges (Forthcoming b) hypothesize that the correlations in Table 2 hold between position and pragmatic function in conversation.
|Link to previous discourse||Anticipation of forthcoming discourse|
|Focalizing, topicalising, framing||Modalizing|
While LP is confirmed in Beeching and Detges (Forthcoming a) as the locus for information-structuring such as linking and focalizing, the evidence is weaker for other factors. In particular, Detges and Waltereit (Forthcoming) point out that the strong first person pronoun moi in French may be used to signal illocutionary commitment and therefore it can be modalizing at LP:
|‘I can assure you that I shall make them for you, these sleeves’ (Detges and Waltereit Forthcoming [ELICOP, Moi 11])|
Traugott (Forthcoming a) also shows that some PMs may be intersubjective at LP, e.g. surely may be used at LP to invite a response. The generalizability of a link between LP and subjectivity on the one hand and RP and intersubjectivity on the other has been further questioned by work on languages other than English, e.g. on rhetorical interrogative forms in Korean (Rhee 2013).
The models discussed above were all originally designed to account for synchronic characteristics of PMs. They have also been used as the basis of hypotheses about historical developments, given assumptions that over time it is more likely for clause internal or adjunct arguments to develop into PMs, and for subjective meanings to develop into intersubjective ones than vice versa (e.g. Traugott and Dasher 2002). Both these hypotheses have been supported, but with different degrees of strength in different languages and with respect to different data points.
4 Arguments for a continuum between core clause and periphery
Each of the models discussed in Section 3 except for that in Kaltenböck et al. (2011) makes reference to slots, whether functional (“intersubjective”, “viewpoint”) or structural (“LP” and “RP”). A question that arises is whether functional slots and their structural boundaries are discrete or gradient. In this section I argue that they are both gradient.
Referring to PMs that occur primarily at LP, Fraser (2006: 168) claims that they are “separate and distinct from the propositional content of the sentence”, i.e. functionally discrete. This is implausible from a historical perspective since it is widely recognized that PMs originate within the proposition (e.g. Traugott and Dasher 2002).
The fact that LP and RP slots do not have structurally or prosodically absolutely discrete boundaries is illustrated by work on English I think, one of the most extensively studied PMs in English.4 One line of investigation has been whether it originated in a main clause, as proposed by Thompson and Mulac (1991). Brinton (2008 and elsewhere) has argued that it did not, but rather arose as a comment clause (as I think), often evidenced in Middle English (c.1100-1500) in post main clause position. In either case, some examples are undecidable with respect to whether they are structurally main or adjunct clauses or are used pragmatically as comment clauses. Another line of inquiry has been whether, when used in initial position, there is ambiguity even in spoken data (specifically that recorded in ICE-GB) between use as a main clause, a comment clause, and an adverbial “discourse marker”5 used as a hedge (Dehé and Wichmann 2010). Dehé and Wichmann (2010: 62–63) conclude that distinctions are typically made as follows: (i) if the pronoun I is accented, then the sequence serves as a main clause, but if (ii) the verb is prominent, the sequence serves as a comment clause, and (iii) an unstressed sequence is used as a “prehead” with the discoursal hedge function. Main clause and comment clause uses may prosodically either be separated from or integrated with the following clause, whereas the “discourse marker” use is integrated with it (Dehé and Wichmann 2010: 49–50). Furthermore, that may appear in all three uses. In main clause uses that functions as a complementizer, but it may also be an optional routinized part of both uses (p. 65). However, distinct prosodic boundaries between the initial sequence and what follows cannot be clearly established, even with automatic acoustic analysis of contemporary spoken data. They disappear almost completely in writing. “Comma intonation” is not required or even characteristic of PMs. By hypothesis, distinct boundaries, whether structural or prosodic, between clauses and subparts of clauses were also not identifiable as new uses developed.
The development in English of categories of PMs used at RP provides confirmatory and indisputable evidence from general extenders, question tags, and retrospective contrastive markers that boundaries between core clause and periphery are not always discrete (Traugott Forthcoming b). Here I exemplify with a brief outline of the history of general extenders (for largely synchronic studies of extenders see e.g. Dines 1980; Cheshire 2007; Pichler and Levey 2010, 2011).
Pichler and Levey (2010) hypothesize that general extenders may have developed in steps such as those in (8).
- Stage 0:
final, indefinite, member of a set
- Stage 1:
textual marker of a set, implicating a larger category; backward-looking and topic closing
- Stage 2:
interpersonal, backward-looking hedge; turn-yielding
It appears that only Stage 0 is attested in Old English (c.750-1100). Carroll (2008) shows that extenders denote a set (indefinite, but specific) and usually collocate with a modifier:
|‘One ought to remind married couples and also every person that …’|
|(850–950 HC CoCura 397 [Carroll 2008: 8])|
This passage follows a list of instructions to exhort and remind people of various duties and behaviors important for the Christian life. Here and eac gehwelnce mon is not “semantically empty” but denotes a category equivalent to ða gesinhiwan. All the same, it introduces “an element of vagueness in the proposition” (Aijmer 2002: 13), which is crucial to the later development of general extenders.
By Middle English some examples appear that suggest that formulae were developing which imply a larger set, without necessarily denoting a specific (but indefinite) set, e.g. or some other, which can have an ellipsis interpretation:
or he may passe to Ieen or Vinice or some oþer.
or he may go to Genoa or Venice or some other
(a1425 Mandeville’s Travels 214, Hamelius, ed. 1919 [Carroll 2008: 13])
Such ambiguity is the likely site for the development from specific to general extenders (Pichler and Levey’s Stage 1) since it is a “critical” context (Diewald 2002) for change, pragmatically and also syntactically ambiguous between the older and newer meaning. Among the clearest examples of general extenders are et cetera and and so forth, neither of which allows an ellipsis interpretation.
A study of the development of extenders like and stuff (like that), and everything, or something (like that), and or anything in COHA (1800–2010) suggests that the potential has persisted for ambiguity of function between extension of a list (specific extender) and more metatextual uses such as topic closure (general extender). With regard to scope over the preceding text, examples of and stuff are often ambiguous. While stuff originally referred to equipment, supplies belonging to an army, etc., OED cites it as being used in the sense “matter of an unspecified kind” (stuff n1, III) from the sixteenth century on and in the routine and stuff from the seventeenth, often in a pejorative context, e.g.:
She turned to me and said, ‘Lewis, I find you pretend to give the Duke notionsof the mathematics, and stuff.’
(?1697 J. Lewis Mem. Duke of Glocester (1789) 66 [OED stuff n1, III. d])
Here, as in many examples, including contemporary ones, it is unclear whether and stuff has the NP mathematics as its antecedent, or the non-finite clause to X. It is only in the twentieth century that examples appear where and stuff has unambiguous backward-looking scope over an activity because it no longer requires an NP with listing characteristics as an antecedent:
they’re all the time hollering about how brave I am and stuff. You know what makes me brave? It’s because I’m so happy I got you. (1941 Wolff, Whistle Stop [COHA])
In short, general extenders do not have many of the alleged characteristics of PMs listed in (3) and(5):they do not occur utterance-initially (3b), they are minimally mobile (5a), they are relatively integrated structurally and prosodically with the core clause (3a, 5b, c), and they usually do not have wide scope over the clause (5f). Although distributionally atypical PMs, nevertheless, they have prototypical functions.
General extenders show that the function of PMs that may be used at RP is gradient. They are not always “separate and distinct from the propositional content of the sentence” as Fraser (2006: 168) proposed. They also show that the structural boundary between core clause and RP is not always decidable, and therefore not discrete. Synchronically, functional gradience has been found to be characteristic of the other categories of PMs that occur at RP, including comment clauses (Brinton 2008), question tags (Tottie and Hoffmann 2009) and retrospective contrastive then (Haselow 2012). It has also been found to be characteristic of PMs that may appear at LP. Such synchronic functional gradience is the typical output of gradual micro-step changes (Traugott and Trousdale 2010). Likewise the boundaries between the “core” clause and clause-initial and clause-final slots (LP and RP respectively) show structural and prosodic variability.
Focusing on PMs that are used clause-finally, I have provided additional evidence that the characteristics of PMs that are cited repeatedly are weak tendencies associated with a restricted set of PMs in a restricted set of languages and cannot be used as diagnostics.
Furthermore I have suggested that the answer to the question “periphery of what?” is complex (see also Rhee 2013). From the perspective of PMs, the following distinctions can be made:
Given a model of grammar which identifies a core, nucleus, or other structure associated with the proposition, there will be elements that are preferred at the beginning or end of the clause (or utterance), among them adjunct structures.
“Outside” these can be various PMs with different functions, such as:
linking to prior discourse and coherence-marking, focalizing, topicalizing, attention-getting, and hedging, most of which are preferred clause/utterance-initially in English, and
response-inviting, turn-yielding, and anticipating forthcoming discourse, most of which are preferred at RP in English.
Whether used in initial or final position, these PMs can be said to be at the periphery of the clause.
However, the boundaries between I and II are not sharp. They are gradient and permeable, as evidenced by synchronic functional, structural, and prosodic analysis and by gradual reinterpretation over time of elements available in I as elements available in II.
A question for further research is the sequential relationship among PMs in II. For example, in English information-structuring PMs such as topic resumptive as far as can be preceded by hedging and coherence markers, a kind of complexity that is reflected in Table 1 in Section 3.
I am grateful to Kate Beeching, Liesbeth Degand, Ulrich Detges, Alexander Haselow, Yuko Higashiizumi, Noriko Onodera, Heike Pichler, Richard Waltereit, and several colleagues at IPra 13, New Delhi, for discussion of issues of “periphery”. Many thanks to Alexander Bergs and an anonymous reviewer for helpful comments on the final draft.
Data bases cited
The Corpus of Contemporary American English. 2008–. Compiled by Mark Davies. Brigham Young University. http://corpus.byu.edu/coca/.
Corpus of Historical American English. 2010–. Compiled by Mark Davies. Brigham Young University. http://corpus.byu.edu/coha/.
Brosens, Veerle et al. Etude linguistique de la communication parlée. http://bach.arts.kuleuven.be/elicop.
Helsinki Corpus of English Texts. 1991. Compiled by Matti Rissanen (Project leader), Merja Kytö (Project secretary), et al. University of Helsinki. http://www.helsinki.fi/varieng/CoRD/corpora/HelsinkiCorpus/index/html.
International Corpus of English-Great Britain. International http://ice-corpora.net/ice/index.htm.
Oxford English Dictionary. http://www.oed.com/.
Base de données du français oral de Belgique. http://www.uclouvain.be/260466.html.
Aarts, Bas. 2007. Syntactic gradience: The nature of grammatical indeterminacy. Oxford: Oxford University Press. Google Scholar
Aijmer, Karin. 2002. English discourse particles. Evidence from a corpus. Amsterdam and Philadelphia: John Benjamins. Google Scholar
Andersen, Gisle. 2011. Pragmatic markers and sociolinguistic variation: A relevance-theoretic approach to the language of adolescents. Amsterdam and Philadelphia: John Benjamins. Google Scholar
Beeching, Kate & Ulrich Detges (eds.). Forthcoming a. Discourse functions at the left and right periphery: Crosslinguistic investigations of language use and language change. Leiden: Brill. Google Scholar
Beeching, Kate & Ulrich Detges. Forthcoming b. Introduction. In Kate Beeching & Ulrich Detges (eds.), Discourse functions at the left and right periphery: Crosslinguistic investigations of language use and language change. Leiden: Brill. Google Scholar
Brinton, Laurel J. 2008. The Comment clause in English: Syntactic origins and pragmatic development. Cambridge: Cambridge University Press. Google Scholar
Carroll, Ruth. 2008. Historical English phraseology and the extender tag. Journal of the Spanish Society for Medieval English Language and Literature (SELIM) 15. 7–38. Google Scholar
Cinque, Giulielmo & Richard S. Kayne (eds.). 2004. The Oxford handbook of comparative syntax. New York: Oxford University Press. Google Scholar
Degand, Liesbeth. Forthcoming. “So very fast very fast then”: Discourse markers at left and right periphery in spoken French. In Kate Beeching and Ulrich Detges (eds.), Discourse functions at the left and right periphery: Crosslinguistic investigations of language use and language change. Leiden: Brill. Google Scholar
Dehé, Nicole & Anne Wichmann. 2010. Sentence-initial I think (that) and I believe (that): Prosodic evidence for use as main clause, comment clause and discourse marker. Studies in Language 34. 36–74. CrossrefWeb of ScienceGoogle Scholar
Detges, Ulrich & Richard Waltereit. Forthcoming. Moi je ne sais pas vs. Je ne sais pas moi. French disjoint pronouns in the left and right periphery. In Kate Beeching & Ulrich Detges (eds.), Discourse functions at the left and right periphery: Crosslinguistic investigations of language use and language change. Leiden: Brill. Google Scholar
Diewald, Gabriele. 2002. A model for relevant types of contexts in grammaticalization. In Ilse Wischer & Gabriele Diewald (eds.), New reflections on grammaticalization, 103–120. Amsterdam and Philadelphia: John Benjamins. Google Scholar
Dister, Anne, Michel Francard, Philippe Hambye & Anne Catherine Simon. 2009. Du corpus à la banque de données. Du son, des textes et des métadonnées. L’évolution de banque de données textuelles orales VALIBEL (1989–2009). Cahiers de Linguistique 33. 113–129. Google Scholar
Fischer, Kerstin. (ed.). 2006. Approaches to discourse particles. Amsterdam: Elsevier. Google Scholar
Fraser, Bruce. 2006. Towards a theory of discourse markers. In Kerstin Fischer (ed.), Approaches to discourse particles. Amsterdam: Elsevier, 189–204. Google Scholar
Haselow, Alexander. 2012. Subjectivity, intersubjectivity and the negotiation of common ground in spoken discourse: Final particles in English. Language and Communication 32. 182–204. Web of ScienceCrossrefGoogle Scholar
Hayashi, Shiro. 1983. Nihongo no bun no katachi to shisei [The structure and modality of a Japanese sentence]. In Danwa no Kenkyuu to Kyooiku (Nihong kyooiku shidoo sankoosho 11), 43–63. Tokyo: National Language Research Institute. Google Scholar
Higashiizumi, Yuko & Noriko O. Onodera. 2013. Introduction. Workshop on Form-function-periphery mapping in Japanese: “Exchange structure” and “Action structure”. IPra 2013, New Delhi, 8–13 September. Google Scholar
Hoffmann, Sebastian. 2006. Tag questions in early and late modern English: Historical description and theoretical implications. Anglistik 17. 35–55. Google Scholar
Ifantidou, Elly. 1994. Evidentials and relevance. PhD thesis. University College London. Google Scholar
Lambrecht, Knud. 1994. Information structure and sentence-form: Topic, focus and the mental representation of discourse referents. Cambridge: Cambridge University Press. Google Scholar
Lenker, Ursula. 2010. Argument and rhetoric. Adverbial connectors in the history of English. Berlin: De Gruyter Mouton. Google Scholar
Minami, Fujio. 1974. Gendai Nihongo no Koozoo (The structure of modern Japanese). Tokyo: Taishukan Shoten. Google Scholar
Onodera, Noriko O. & Yuko Higashiizumi. 2013. Emerging pragmatic elements at right periphery in Japanese. Paper presented at IPra 13, New Delhi, 8–13 September. Google Scholar
Pichler, Heike & Stephen Levey. 2010. Variability in the co-occurrence of discourse features. University of Reading Language Studies Working Papers 2. 17–27. Google Scholar
Pichler, Heike & Stephen Levey. 2011. In search of grammaticalization in synchronic dialect data: General extenders in north-east England. English Language and Linguistics 15. 441–471. Web of ScienceCrossrefGoogle Scholar
Rhee, Seongha. 2013. LP and RP in grammaticalization of rhetorical interrogative forms in Korean. Paper presented at IPra 13, New Delhi, 8–13 September. Google Scholar
Rizzi, Luigi. 1997. Parameters and functional heads: Essays in comparative syntax. Oxford: Oxford University Press. Google Scholar
Schiffrin, Deborah. 1987. Discourse markers. Cambridge: Cambridge University Press. Google Scholar
Schourup, Lawrence. 1982. Common Discourse particles in English conversation. Working Papers in Linguistics 28. The Ohio State University. Google Scholar
Schwenter, Scott A. 2000. Viewpoints and polysemy: Linking adversative and causal meanings of discourse markers. In Elizabeth Couper-Kuhlen & Bernd Kortmann (eds.), Cause – condition – concession – contrast: Cognitive and discourse perspectives, 257–281. Berlin: Mouton de Gruyter. Google Scholar
Shinzato, Rumiko. 2007. (Inter)subjectification, Japanese syntax and syntactic scope increase. In Onodera and Suzuki, Eds. Journal of Historical Pragmatics 8(2). 171–206. Web of ScienceCrossrefGoogle Scholar
Thompson, Sandra A. & Anthony J. Mulac. 1991. A quantitative perspective on the grammaticalization of epistemic parentheticals in English. In Elizabeth Closs Traugott & Bernd Heine (eds.), Approaches to grammaticalization, Vol. 2, 213–329. Amsterdam and Philadelphia: John Benjamins. Google Scholar
Traugott, Elizabeth Closs & Richard B. Dasher. 2002. Regularity in semantic change. Cambridge: Cambridge University Press. Google Scholar
Traugott, Elizabeth Closs. Forthcoming a. On the function of the epistemic adverbs surely and no doubt at the left and right peripheries of the clause. In Kate Beeching & Ulrich Detges (eds.), Discourse functions at the left and right periphery: Crosslinguistic investigations of language use and language change. Leiden: Brill. Google Scholar
Traugott, Elizabeth Closs. Forthcoming b. On the rise of types of clause final pragmatic markers in English. Journal of Historical Pragmatics. Google Scholar
Traugott, Elizabeth Closs & Graeme Trousdale. 2010. Gradience, gradualness and grammaticalization: How do they intersect? In Elizabeth Closs Traugott & Graeme Trousdale (eds.), Gradience, gradualness, and grammaticalization, 19–44. Amsterdam: Benjamins. Google Scholar
Yap, Foong Ha, Charles Tsz-kwan Lam & Jiao Wang. 2008. Clausal integration and the emergence of sentence-final particles in Chinese: A comparison of ER YI YI and YE YI YI. http://www.sfu.ca/gradlings/SFUWPL/ICEAL2/Yap_EtAl.pdf
He does, however recognize that some DMs, especially but, and, so may occur without a preceding S1, especially in responses (Fraser 2009: 9).
Structural and prosodic boundaries may not coincide. The point here is that structural boundaries appear not to be discrete or “subsective” in the sense of Aarts (2007), and that prosodic boundaries are not discrete either.