A processual view on grammar: macrogrammar and the final field in spoken syntax

A processual view on grammar: macrogrammar and the final field in spoken syntax

Language Sciences 54 (2016) 77–101 Contents lists available at ScienceDirect Language Sciences journal homepage: www.elsevier.com/locate/langsci A ...

583KB Sizes 1 Downloads 53 Views

Language Sciences 54 (2016) 77–101

Contents lists available at ScienceDirect

Language Sciences journal homepage: www.elsevier.com/locate/langsci

A processual view on grammar: macrogrammar and the final field in spoken syntax Alexander Haselow University of Rostock, Germany

a r t i c l e i n f o

a b s t r a c t

Article history: Received 27 May 2015 Received in revised form 10 November 2015 Accepted 3 December 2015 Available online 18 January 2016

This paper presents an approach to linguistic analysis that distinguishes two cognitive serialization principles according to which speakers build up a unit of talk in real-time speech production, microgrammar and macrogrammar, and thus two different “category pools” from which speakers choose constituents. Based on empirical data of spontaneous spoken English, it will be argued that macrogrammatical elements serve different cognitive tasks that arise at particular points in the linear production of an utterance, such as finetuning epistemic value or modifying illocutionary force. This article focuses on one particular temporal slot for the production of macrogrammatical units, the potential end of a unit of talk or “final field”, providing an analysis of different unit types available for use in this slot and discussing the basic cognitive tasks these units serve in conversational interaction. Ó 2015 Elsevier Ltd. All rights reserved.

Keywords: Cognition Syntax Final field Microgrammar Macrogrammar Spontaneous speech

1. Introduction The ‘final field’ is a communicative space that follows a structural unit that has already reached a potential point of completion in real-time utterance production and which is available for the production of unit types with various metatextual and interactive functions, such as tag questions, comment clauses or vocatives. The final field is a crucial moment in the production of a unit of talk in so far as it allows speakers to deal with a variety of cognitive tasks before potential turn transition or continuation with a new idea. Exploring the kinds of linguistic units that may be produced in the final field and their function is of particular interest to cognitive science researchers as they are indicative of cognitive processes in on-line speech production. Principles of linguistic structure provide a window to cognitive structures since how we structure what we say may reflect how we structure thought, though not necessarily in a one-to-one relationship. At least, linguistic choices made at particular moments in the ongoing production of linguistic structure carry cognitive significance in the sense that they reflect the speaker’s conceptualization of what there is to do at a particular moment in the linear construction of linguistic structure, such as in the final phase. This linear temporal view on linguistic structure opens up avenues for addressing research questions on language and cognition that have not yet been sufficiently investigated, due to the predominance of static conceptualizations of structure, according to which the incremental growth of structure from the perspective of the acting speaker is backgrounded in favor of an analytic, product-based perspective. From an analytic perspective, linguistic structure appears a fixed, ready-made object available for linguistic description, due to the fact that the analyst is not forced

E-mail address: [email protected]. http://dx.doi.org/10.1016/j.langsci.2015.12.001 0388-0001/Ó 2015 Elsevier Ltd. All rights reserved.

78

A. Haselow / Language Sciences 54 (2016) 77–101

into the temporality of speech, but comes into contact with his/her object of investigation long after its production has been completed. The present study explores the kinds of linguistic units that are available for use at a moment at which a speaker is about to terminate an emerging unit of talk, and discusses the conclusions that can be drawn from the functions of these elements on the kinds of cognitive tasks that speakers consider relevant at this moment in real-time speech production. Based on the analysis of unit types in the ‘final field’ in a corpus of natural conversations in English it will be shown that the conceptualization of a structural unit as a configuration that emerges in the linear flow of time can provide insights into cognitive processes (or “phases”) that regulate speech planning and production. The starting point for the present study is the observation that in spoken English a limited set of linguistic units is used almost exclusively at the end of a unit of talk that is already potentially complete, such as final linking adverbials/final “particles” (e.g. then, though, actually, Biber et al., 1999: 791; Haselow, 2013), tag questions (e.g. can you?, Tottie and Hoffmann, 2009), or general extenders (e.g. and all, or something, Cheshire, 2007; Tagliamonte and Denis, 2010; Pichler and Levey, 2011). These elements represent sedimented patterns of expanding a unit of talk beyond a potential completion point and pose a challenge to grammatical theory and categorization as they are not integrated into the morphosyntactic dependency relations within a structural unit (e.g. a clause) and thus “outside” core grammar or “extra-clausal constituents” in terms of Dik (1997: 380). However, they are not autonomous either as they are not able to represent a unit of talk of their own, but require a structural unit to which they are linked in various ways, e.g. in terms of pragmatic scope, illocutionary force, and prosody. To the extent that these elements are systematically produced in the final phase of the production of a unit of talk they can be exploited for accessing what is going on in the speaker’s mind, that is, which tasks are relevant for the speaker at this point in time in the linear production of a unit of talk. Clause-based approaches to syntax in the structuralist tradition cannot adequately deal with elements that have no immediate constituent status: they are difficult to describe in terms of syntactic categories and their grammatical status is relatively unclear, which explains why they are often thrown into underspecified “wastepaper bin” categories such as “inserts”, “pragmatic markers”, and the like. I will propose an alternative model of grammar in which these elements find their proper place as grammatical units in a cognition-based syntax that distinguishes two cognitive serialization principles called microgrammar and macrogrammar. This approach attempts to link the processual or ‘emergent’ character of linguistic structures produced in real time to aspects of sedimentation and pattern formation, and relies on a productionbased rather than product-based description of syntactic structure. Thus, the structure of units of talk, which encompasses clausal but also non-clausal structures that represent a coherent intonation contour, is conceptualized as emerging in the linear flow of time (“on-line” in terms of Auer [2009]). Such structures consist of different “fields”, that is, optional time slots that may host a set of elements that serve a particular kind of cognitive task a speaker faces during the production of a unit of talk, e.g. (i) building up a pragmatic projection on the type of activity to follow, (ii) expressing a single ‘focus of consciousness’ in a core unit (Chafe, 1994), or (iii) retrospectively framing or modifying a core unit just produced. We may thus distinguish, for instance, an “initial field” (corresponding to Auer’s [1996a] “pre-front field” in German), one or more “medial field/s”, and a “final field”, or a “left” and “right periphery” (Beeching and Detges, 2014; Traugott, 2015). The grammatical status of the final field has received relatively little attention thus far and will therefore be discussed more in detail here. The term ‘field’ has not been established in syntactic terminology in English linguistics but is a central component in the approach developed here. A field is an optional communicative space outside the morphosyntactic dependency relations within a structural unit (e.g. a phrase or a clause) that provides an opportunity for the speaker to produce linguistic units that serve the expression of particular cognitive tasks that speakers need to attend to at the respective point in time in the ongoing production of a unit of talk. The focus of this study is on the function of the final field, based on data from spontaneous spoken English provided by the International Corpus of English (ICE-GB), as exemplified in (1a–c). The element in the final field in (1a) is analyzable as a “linking adverbial” or “final particle”, the one in (1b) as a tag question, and the one in (1c) as a general extender. (1)

a.

b.

c.

150 151 152 153 154

A: B: A: B:

the mother looked like Cher. yea:h I can’t stand >that kind of woman no-

241

A:

but you can’t just pick them up off the counter

242

B:

yeah you can.

009

B:

I’ve never understoo:d >and this is what’s caused a lot of problems

though<. [ICE-GB S1A-041]  can you .

n all h<. FINAL FIELD 

[ICE-GB S1A-079] [ICE-GB S1A-050]

A. Haselow / Language Sciences 54 (2016) 77–101

79

The final field of a unit of talk is interesting because it is important for understanding the cognitive tasks that speakers need to deal with at a point in time in ongoing utterance production at which they feel to have approached the end of a unit of talk and often also a point at which turn-transition becomes relevant. The structure of this paper is as follows. I will first discuss some basic cognitive principles according to which speakers serialize elements in emerging talk (Section 2), introducing the notions of “microgrammar”, “macrogrammar”, and “final field”, the latter being part of “macrogrammar”. Section 3 provides an overview of the types of elements available for use in the final field in English, which distinguishes seven different classes, based on common formal and functional properties: “extra-clausal” adverbs, linking adverbs, vocatives, general extenders, comment clauses, tag questions, and independent ifclauses. Each of these types will be discussed in terms of their syntactic properties and the tasks they serve in the linear, incremental production of a unit of talk, based on empirical data from the sections “S1A” (private direct conversations) and “S1B” (public direct conversations, e.g. broadcast interviews) of the ICE-GB. General aspects of the final field in a grammar of speech will be discussed in Section 4, the relation of the final field to other temporal slots in the linear emergence of structure is explored in Section 5. Section 6 is the conclusion. The main argument put forward in this paper is that elements in the final field are “external” to a structural unit (e.g. a clause) from a “microgrammatical” perspective, which is based on morphosyntactic dependency relationships, but “internal” and thus an integral part of the unit they accompany from a “macrogrammatical” perspective, which encompasses units that are integrated into the illocutionary force of a unit of talk in emerging discourse. 2. Microgrammar, macrogrammar and fields 2.1. Introducing microgrammar and macrogrammar In the present framework, ‘grammar’ is understood as a process emerging in the linear flow of time, based on moment-bymoment decisions of the speaker who needs to deal with various cognitive tasks while building up a structural unit. Grammar is, in this sense, both a knowledge system comprising conventionalized units and the conditions of their use, and an activity since a structural pattern is created by a process of incrementation, that is, by the addition of linguistic units to an already realized utterance part. Thus, grammatical description should be knowledge-based and activity-based rather than productbased. Under this ‘emergentist’ view in the sense of Hopper (1987, 1998, 2011), structure produced in spontaneous speech is not conceived of as a ‘static’ configuration and a ready-made product available for linguistic analysis, based on a priori categories and structural patterns, but as a temporal phenomenon that is always open at the leading end. This shift toward a temporal or ‘dynamic’ perspective requires rethinking the usefulness and adequacy of many of those terms and concepts that derive from purely analytic views on language and that disregard the on-line emergence of structure (see also Auer, 2009: 4). Among the most pervasive examples are terms like “(syntactic) position”, “periphery”, “movement”, “dislocation” or adjectives like “left” and “right” as in “left/right periphery”, which are used in both structuralist and functionalist approaches to syntax. Such notions reduce language and syntactic structure to a static, two-dimensional configuration of linguistic units with seemingly clearly definable starting and end points. The problem arising from such a perspective on syntax is that it ignores a basic condition under which structure comes into existence, which is the linearity of speaking in time: in real-time speech production, speakers and listeners are forced into the temporal emergence of structure, the speaker being at the leading edge of an unfolding structure (Auer, 2009; Auer and Pfänder, 2011). From a procedural perspective, speakers and listeners cannot jump back and forward or move from left to right and right to left in ongoing discourse, and they cannot survey the beginning and the end of an emerging structure simultaneously. It is also unlikely that speakers have a full structural plan at hand before a structural unit is produced, as is implied in assumptions such as the one that speakers are able to move or “dislocate” constituents from inside a structural unit to its “periphery”. What speaks against this assumption is that, for instance, speakers can revise the syntactic trajectory of a unit of talk underway (Hopper, 2011), or expand a unit that appeared completed at one moment in emerging discourse and thus make it part of a larger, more complex structure (Auer, 1996b, 2007; Couper-Kuhlen and Ono, 2007). Such phenomena suggest that speakers create structure in a piecemeal manner in the flow of time and are able to renegotiate it at any point in order to adapt rapidly to changing communicative demands (“local contingencies”) while producing a unit of talk. In spite of the relative uncertainty on the side of both the speaker and the listener as to the way in which a syntactic trajectory will finally end up (Hale, 2006), there can be no doubt that they are able to identify starting points, to predict possible continuations, and to anticipate potential points of completion, which can be more or less clearly identifiable. The identification of potential points of completion of a unit of talk or turn-constructional unit (TCU) is a complex issue (Ford and Thompson, 1996; Auer, 1996b) as it is based on various factors that interact in complex ways. Optimal or salient completion points are those at which a unit is syntactically, prosodically and pragmatically “complete” in the sense that it represents a “well-formed” syntactic unit (usually defined as a complete subject–predicate structure) produced under a coherent intonation contour with terminal pitch, and representing a clearly definable conversational activity. It is, however, not necessary that a unit of talk represents a “well-formed” syntactic unit in order to be interpretable as complete. Under a dynamic view of syntax, the structure of a unit-in-progress is emergent on a moment-by-moment basis and never predetermined or predeterminable (Hopper, 1987, 1998, 2004), which explains the general plasticity of structural patterns produced in speech. An initial noun phrase, for instance, can be interpreted as a subject at one moment, as it occurs in the

80

A. Haselow / Language Sciences 54 (2016) 77–101

canonical subject position in English, but may turn out as a so-called “left-dislocation”, or a hanging topic at a later point in time (Pekarek Doehler, 2011). Structure only appears fixed to linguists as they need to “freeze” their object of analysis for analytic purposes and are thus not forced into the linearity of speaking in time. Rather, they come into contact with a speech product long after its production has been completed. We should, therefore, not confuse the seemingly fixed character of the finished speech product with the openness of linguistic structure during its production, which is an incremental process in which the speaker has no “bird’s eye view” on his/her utterance (Hopper, 2011: 23). A grammatical approach that is able to explain the kind of structures produced in on-line speech production must be based on two principles. First, it must explain structure as being based on the cognitive conditions of real-time speech production, that is, the incremental, irreversible, process-related character of structure that is build up in the linear flow of time. Various approaches geared toward the linearity of language have been developed in the past decades, the most prominent of which are certainly On-line Syntax (Auer, 2009, 2015) and Emergent Grammar (Hopper, 1987, 1998, 2011; Auer and Pfänder, 2011). Secondly, the cognitive serialization principles speakers are assumed to be able to apply in spontaneous speech cannot be based solely on the knowledge of sentence grammar, that is, on the internal structure of one coherent morphosyntactic unit alone. We need to account for the large number of phenomena that are difficult to describe under traditional approaches to syntax, such as those which share the property of being “unintegrated” or “outside” a clausal unit from a morphosyntactic perspective and which do not express core grammatical meanings, but are linearly related to such a unit. Pervasive examples are so-called “dislocations”, parentheticals, free or unintegrated wh-clefts (e.g. what I want you to do – go down to the office, see Miller and Weinert, 1998: 120–127), or lexical elements such as discourse markers or stance adverbials. These phenomena can only be analyzed adequately under an approach that accounts for the piecemeal development of linguistic structure since structure is only retrospectively forming a complex configuration amenable to linguistic study. Serialization does not always coincide with hierarchization, as traditional grammars based on well-formed clause or sentence structures lead one to assume. In such grammars, syntactic structure is conceived of in terms of units in which all elements (“constituents”) are integrated into mutual morphosyntactic dependency relationships, such as agreement and government. However, a cursory glance at the data deriving from spontaneous speech suffices to show that, from this dependency-driven perspective, clausal units are continually “surrounded” or “interrupted” by extra-clausal material, as illustrated in (2). (2)

289 290 291 292 293 294

A: B: / A: /

you may think it wasn’t three weeks very well spent, WELL no no (.) not really¿ but uh (.) could’ve been yeah better ways of spending it you know [1.5 ] like the South of France [yeah] yeah well anyway give us uh you know a call when uh (2.0) when you know [ICE-GB S1A-095]

The units in bold perform no definable function in the establishment of morphosyntactic relationships between the constituents of the clausal units that can be identified in lines 291 (“[there] could have been better ways of spending it”) and 294 (“give us a call when you know”). In line 291, these units are (i) the discourse marker but, which introduces a concession to the initial negation of A’s assumption that B might consider the “three weeks” as not “very well spent”, (ii) a cognitive planning marker (uh) creating a dysfluency and introducing a new “thought unit” (Kjellmer, 2003: 174), (iii) the self-reinforcing continuer “yeah”, (iv) the discourse marker “you know”, which refers to assumed mutual understanding and expresses putative plausibility, and (v) an increment (Schegloff, 1996: 59, Couper-Kuhlen and Ono, 2007), i.e. an expansion of a syntactic structure beyond an already existing potential point of completion, namely the prepositional phrase “like the South of France”, which itself consists of internal morphosyntactic relations based on embedding. The utterance in line 294 is similarly unintegrated, including e.g., a sequence of discourse markers in the initial phase of utterance production and cognitive planning markers within the unit-in-progress. Many of such units make no contribution to the propositional content of a unit of talk, but have procedural meanings in terms of Blakemore (1987), providing a cue on how to interpret the content of the unit they are part of, cueing speaker–listener relationship, or indicating cognitive planning. Such examples are, as analysts dealing with the syntax of spoken language know, by no means exceptional. They suggest that speech production and, in particular, the ways in which different kinds of linguistic units are combined into larger structures cannot be reduced to one monolithic system of language processing. Rather, linguistic activity seems to exhibit a dualistic organization: it is based on knowledge of morphosyntactic dependency relations, constituency and ways of syntactic embedding, on the one hand, and knowledge of how to assemble different kinds of units expressing information relevant on different levels of the general communicative system (interpersonal, pragmatic, discourse-structural level) into a coherent unit of talk, on the other hand. Without the assumption of this second kind of knowledge, units of talk such as the ones discussed above could only be analyzed as unintegrated conglomerates of structural fragments that are randomly assembled in the linear progression of talk. The hypothesis of a dualistic organization of language processing is not new, but in accordance with the findings deriving from different lines of research. It surfaces, for instance, in functionalist approaches to language, such as Dik’s (1997)

A. Haselow / Language Sciences 54 (2016) 77–101

81

Functional Grammar, which distinguishes clausal and “extra-clausal constituents” (ECCs), ECCs being “never essential to the internal structure of the clause with which they are associated” (Dik, 1997: 381) and units that can “only be understood in terms of pragmatic rules and principles” (Dik, 1997: 380) rather than in terms of grammatical/syntactic principles. Kaltenböck et al. (2011) distinguish between “thetical grammar” and “canonical sentence grammar”, both of which are part of “discourse grammar” (Kaltenböck et al., 2011: 854). 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, and include 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” (Kaltenböck et al., 2011: 853). A dualistic organization of human cognition is also assumed in neurolinguistic research on linguistic processing (see Heine et al., 2015 for an overview). Van Lancker Sidtis (2004, 2009), for instance, finds that the processing of “formulaic language” (e.g. interjections, discourse markers, comment clauses, expletives, oaths), which requires a holistic mode of processing, differs from that of “novel language”, which is processed in an analytic (compositional) way, the latter being associated with left-hemisphere (LH) activity whereas the first involves the right hemisphere (RH). This “dual process” model is in line with Code (1996), who shows that utterances produced by people suffering from LH damage consist primarily of automatic and nonpropositional speech, that is, they include basically “formulaic” speech. Prat et al. (2007) assume dualistic processing in text comprehension and distinguish between propositional representation and discourse model. Generally, recent studies on brain lateralization seem to converge to the following: while speech processing involves both hemispheres, there is a tendency for the LH to be stronger involved in structuring clauses (Brady et al., 2006), whereas the RH is heavily involved in pragmatic ability, inferencing (Brownell et al., 1986), the processing of non-literal, transferred meanings (Bottini et al., 1994; Rinaldi et al., 2004) and indirect speech acts (Stemmer et al., 1994). Pragmatic communication has been found to be difficult with patients suffering from RH damage (Shields, 1991; Cutica et al., 2006 and the references therein), who also suffer from difficulties to develop a unitary mental model of discourse (Marini et al., 2005; Marini, 2012). The RH is also involved in the processing of higher-level information on the macrostructure of discourse (Hough, 1990; Sherratt and Bryan, 2012) and the production and preservation of global discourse coherence by integrating multiple pieces of information into an organized whole (Myers and Blake, 2008). Discourse produced by patients with RH damage tends to be disrupted (lack of integration of textual units) and disfluent (e.g. pauses, repetitions) (Sherratt and Bryan, 2012). Moreover, social aspects of talk, such as maintaining exchange with the listener (involving the use of vocatives and formulae of social exchange) and appreciating the listener’s perspective, largely involve the RH (Myers and Blake, 2008; Hird and Kirsner, 2003). Formulae of social exchange, such as greetings or response/backchannel tokens like yeah, yes and no, are found in the speech of people with LH damage, whose verbal abilities are otherwise severely limited (Van Lancker and Cummings, 1999: 86–88). Principal correlations between linguistic structure and neural processing are discussed in Heine et al. (2015). These findings can be related to linguistic theory in the sense that the dualism of processing is reflected in the way language is used, more precisely, in the structure of units of talk, as exemplified in (2) above. On the one hand, language provides expressions dedicated to maintaining discourse coherence and relations to the addressee(s), many of which are conventionalized or “formulaic” and which are often found outside morphosyntactic dependency relations holding within a stretch of speech. Such elements require holistic processing and tend to be more associated with the RH. On the other hand, expressions can be embedded in morphosyntactic dependency relationships and require syntactic or analytic (compositional) processing, which involves the LH, which is associated with propositional thought. There seems to be a division of labor between two cognitive modes in which linguistic structure is produced and in which products of speech are mentally represented. While, as mentioned above, some more recent functional approaches to grammar imply a distinction into different components, they fall short in explaining recurrent serialization patterns of “extra-clausal” units. Thetical grammar, for instance, posits variability of placement of “thetical” elements. However, a closer look at the data shows that many elements that do not contribute to sentence grammar do not occur randomly within an utterance, but tend to be produced in particular temporal slots in the linear construction of an utterance. One such group will be discussed in this paper, namely elements that are produced in the final phase of utterance production, that is, which are typically final in a unit of talk. Thus, one important question for grammatical theory is how we can reconcile the assumption that, on the one hand, units that do not belong to sentence grammar should theoretically be able to occur anywhere in a structural unit as they are not subject to sentenceinternal distributional principles, with the observation that, on the other hand, many of these units do have a relatively regular and predictable distribution in the sense that their occurrence is restricted to particular time slots in the moment-bymoment emergence of linguistic structure. Next to the principles underlying the arrangement of constituents in hierarchical configurations based on dependency relations, there seem to be principles regulating the distribution of “extra-clausal” elements, which deserve to be investigated. I will use the term “microgrammar” to refer to first serialization principle, and the term “macrogrammar” for the second one. Both are aspects of “grammar”, which is defined as the ability of language users to produce text, that is, a piece of meaningful language that serves a particular purpose in an individual discursive and interpersonal context. It comprises the ability to create, but also to manipulate (revise, interrupt, reconfigurate) linear sequences of linguistic units with different kinds of meanings (conceptual, procedural). The terms “microgrammar” and “macrogrammar” are conceptually affiliated with the notions of microsyntax and macrosyntax as developed by different schools of Romance linguists (e.g. Berendonner, 1990; Blanche-Benveniste, 2003; see Avanzi, 2007 for an overview). However, the framework of micro-/macrosyntax is basically a discourse-based approach to syntax that seeks to provide descriptive tools for the analysis of spoken syntax and

82

A. Haselow / Language Sciences 54 (2016) 77–101

discourse relations from an analytic, product-based rather than procedural perspective and is not based on cognitive aspects of language production. Microgrammar is a serialization principle that refers to the formal means employed by speakers to structure a unit of talk based on internal hierarchization, embedding, constituency, and mutual dependency relations. The final product is a syntactic unit in which all elements, describable in terms of constituent types, form an integrated whole from a morphosyntactic perspective. Morphosyntactic integration is based on dependency relations, which are always binary, that is, in microgrammar each constituent is always morphosyntactically related to at least one other constituent in a hierarchical relationship. Macrogrammar refers to relational functions outside microgrammatical dependency relations, and is based on serialization principles that rest upon (i) speech planning, (ii) processibility, (iii) textual coherence and (iv) contextual embeddedness. Macrogrammatical elements can be lexical (e.g. discourse markers) or fragmentary, that is, consisting of pieces of formulaic speech that form more or less fixed chunks (e.g. the thing is – , – if you know what I mean). Since such elements typically serve particular routine tasks in ongoing discourse (see below), they tend to be conventionalized to a certain degree. Serialization on the macrogrammatical level surfaces in the loose sequence of units that are held together by discourse-structural and pragmatic, often also prosodic factors, all elements being part of the illocutionary force of a unit of talk, but not integrated into the morphosyntactic dependency relations of a microgrammatical unit. Units of talk such as the ones in line 291 and 294 in (2), for instance, are analyzed here as a coherent macrogrammatical unit as it forms an integrated whole on the semanticopragmatic level (the speaker performs a single conversational activity). The boundaries of macrogrammar are thus defined by the communicative intention of the speaker: elements that do not contribute to the illocutionary force of a unit of talk, but have an illocutionary force of their own form a new macrogrammatical unit. This view implies a certain “pragmaticization” of grammar that is, however, unavoidable if structure is seen as part of context-bound communication, rather than as an abstract, decontextualized system based on a priori patterns. Macrogrammatical elements are not in a morphosyntactic dependency relationship with any other element of a unit of talk and they do not change propositional content. Separation from a microgrammatical unit should, however, not be seen as an all-embracing criterion: while macrogrammatical units are morphosyntactically isolated, there is gradience e.g. with respect to prosodic separation ranging from full integration into the intonation contour including also a microgrammatical unit to separation from it by means of e.g. pitch reset, pauses and realizations as an intonation unit of its own. Gradience has been shown, for instance, for I think/I believe by Dehé and Wichmann (2010a, 2010b), allowing for different meaning interpretations of such units depending on their prosodic realization, and is at the core of Traugott’s (2015) understanding of clausal peripheries. Macrogrammatical elements may also have an effect on propositional content, e.g. modifying or specifying parts of it, such as epistemic certainty, but they do not alter propositional structure altogether. Generally, they indicate how content is to be interpreted or contextualized by the listener or how it is conceptualized by the speaker. Micro- and macrogrammar often alternate in the serialization of linguistic units, as illustrated in (2). It is precisely this alternation that requires much more research. Macrogrammar arises from the various cognitive tasks that become relevant at particular points in time in the linear production of an utterance, such as getting the addressee’s attention, gaining planning time, allocating a unit of talk in emerging discourse, or modifying various aspects of the content expressed in a microgrammatical unit, such as its epistemic value or illocutionary force. Linguistic structure is thus understood as a temporal phenomenon (Auer, 2009; Hopper, 2011; Deppermann and Günthner, 2015), the flow of speech in time being a continuous transition from one task to another and thus relevance-structured. The claim put forward in the present study is that the distribution of macrogrammatical elements is not random, but tends to be patterned as many of these occur in particular temporal slots or “fields”, depending on the moment at which the tasks they serve become relevant in the real-time emergence of a structural unit. The tasks speakers and listeners need to deal with at the beginning of a unit of talk (initial phase), for instance, differ from those at its possible end (final phase): in the initial phase, speakers need to deal with turn-taking issues and establish a new unit of talk as a meaningful contribution in ongoing talk, which involves getting the addressee’s attention (Deppermann, 2013), building up projections (raising expectations) on the trajectory of an utterance underway on various dimensions (above all syntax, semantics, pragmatics, prosody), and dealing with open projections from prior talk concerning the continuation of talk, e.g. providing the second-pair part to a first activity in “adjacency pairs” (e.g. question–answer, Schegloff and Sacks, 1973); at the potential end of a unit of talk in progress, however, the cognitive activity is not so much concerned with opening projections, at least as far as the current unit is concerned, but with the closing or resolution of projections, that is, the production of more or less predictable elements that complete a (syntactic, prosodic, pragmatic) gestalt, and making adjustments to the speech product that has just been emerging, e.g. to its illocutionary force, the kind of link to prior discourse, or preciseness of expression. The respective verbal means serving these tasks are not necessarily integrated into a coherent microgrammatical unit, but often linked to it by mere linearity. An overview of some important tasks that become relevant at different points in time during the step-by-step emergence of a unit of talk is given in Table 1. Table 1 shows the temporal dynamics of cognitive tasks that speakers need to deal with while constructing a unit of talk and thus while moving in time from one task to another. As suggested by the sequential order, there are temporal “phases” in the production of a unit of talk in which dealing with these tasks becomes highly relevant. The assumption of such “phases” implies that macrogrammar is embedded in the temporal structuring of cognitive tasks which, arguably, follow a particular temporal logics, given that some tasks are more relevant at one time in speech production than at another, as will be shown for the final field. Facilitating addressee–response (e.g. by means of a tag question), for instance, requires the

A. Haselow / Language Sciences 54 (2016) 77–101

83

Table 1 The temporal dynamics of cognitive tasks during the production of a unit of talk. Initial phase

Medial phase

 maintaining contact to the addressee  getting/claiming the attention of the addressee  dealing with turn-taking issues Discourse-organization  responding to prior talk (e.g. signalling  expressing a ‘focus of consciousness’ understanding or alignment) (an idea) (Chafe, 1994)  indicating link to prior talk  gaining planning time while keeping Cognition  building up projections the turn  dealing with projections (e.g. expectations, implications) from prior talk  establishing a new ‘focus of consciousness’ (Chafe, 1994) Interpersonal relation

Final phase  facilitating addressee–response

 making a particular kind of next activity relevant  providing a last procedural cue  pragmatic fine-tuning of the focus of conscious-ness just established

presence of something that the addressee may respond to, and is thus logically related to the final phase of the production of a unit of talk. However, there is no one-to-one correspondence between phases and tasks since speakers are free to address first what is their most urgent business and to postpose other tasks. Thus, for instance, the kind of link to prior discourse can be indicated initially, but it may also be postposed to the end of a unit of talk where it is expressed by means of final linking adverbs or “final particles”, such as then or though, which signal an inferential or contrastive relationship to a preceding unit of discourse (Haselow, 2013; Lenker, 2010: 198–200). However, given that the tasks in Table 1 are routine tasks, speakers have not only developed habitualized ways of how to express them, but also habitualized ways of when to deal with these tasks, as the relatively consistent distribution of particular unit types (e.g. those forming the “final field” as discussed below) shows. While the initial and final phase are starting and end points of a unit of talk that correspond roughly to “peripheries” under a gradient view (Traugott, 2015), given that the units produced in these time slots are typically morphosyntactically unintegrated, the medial phase is more difficult to determine as it may, in principle, refer to any time slot occurring in a stretch of speech after its beginning and before its anticipated end. Evidence comes from the use of units like actually, you know or “theticals”, which may occur in various medial time slots (see Aijmer, 1986; Kaltenböck et al., 2011), with differences in scope, though. Lack of space does not allow us to go into the details here. It would be worth investigating possible algorithms in the occurrence of macrogrammatical elements in the medial phase in order to discover distributional restrictions. Similar to what has been shown in Table 1, Beeching and Detges (2014) – as well as the authors in their edited volume – correlate what they call the “left” (LP) and “right periphery” (RP) of discourse units with particular functions that “pragmatic markers” are hypothesized to fulfil in these positions. Their approach is not task-based and thus not following the temporal logics of cognitive tasks speakers need to deal with in the linear progression of talk, but correlating structural slots with functional aspects of the kind of markers used in these slots. However, as Traugott (2015) shows, there is only weak evidence for many of the functions that are hypothesized to be characteristic of the LP or RP by Beeching and Detges (2014), such as “link to previous discourse” or expressing “subjective meanings”, which are not restricted to the LP, or “modalizing” and expressing “intersubjective meanings”, which are not restricted to the RP. Case studies on “peripheral” pragmatic markers (Traugott, 2012; Detges and Waltereit, 2014; Dehé and Wichmann, 2010a) suggest that the linguistic reality is too complex to allow for a binary opposition of the functional differences between LP and RP. It therefore seems more promising to focus on the sequential logics of cognitive tasks to be performed at particular time slots in utterance production, which can be assumed to be relatively stable across communicative contexts.

2.2. The final field The term ‘final field’ refers to a point in time at which the speaker has brought a unit of talk-in-progress to potential completion and expands it by the production of further units. What speakers and listeners consider “complete” is a contextsensitive issue and cannot be predicted out of a concrete speech context and the co-text (Auer, 1996b: 61–62) as it does not merely depend on syntactic “well-formedness”. The concept of “field” is more adequate for a dynamic conception of structure than that of “(left/right) periphery” since the latter implies a written-language-biased, analytic perspective on structure based on the visual form of language. Under this conception, linguistic structure is a bounded entity with “left” and “right” margins, which seduces us to assume that speakers have a bird’s eye view on the structural units they produce and are able to observe their beginnings and ends simultaneously.1 A “field” is defined by temporal sequentiality and as such related to other slots not in a structural (let alone hierarchical), but in a temporal ‘before–after’ relationship in incremental speech production. A field makes the production of certain linguistic units expectable and possible, based on ad-hoc decisions, but not necessary.

1 Recent studies on peripheries acknowledge the issue of linearity of speech, e.g. Detges and Waltereit (2014: 44), who explicitly address the unfolding character of speech. Nevertheless, the terminology used to refer to these temporal slots (“left”, “right”) is largely reminiscient of a product-based analysis.

84

A. Haselow / Language Sciences 54 (2016) 77–101

Crucially, the occurrence of the final field is not projectable for the listener and it is probably only arising in the moment-bymoment assemblage of structural units for the speaker him-/herself. It should be noted that the term “field” is also used in German syntactic analysis (Zifonun et al., 1997: 1502; Eisenberg, 2006: 384; Duden-Grammatik, 2009: 861–887), but conceptually not related to the use of the term in the present study. In German linguistics, a “field” basically refers to structural positions within a sentence; such positions are excluded from the analysis provided in the present study. A context-free definition of the potential end-point of a unit of talk can only be given on the basis of one major cognitive operation that guides language production and processing, namely the speaker’s and listener’s ability to draw mental projections, that is, to precalculate possible continuations on various dimensions of the communicative system (Auer, 2005, 2009). ‘Projection’ is a cognitive mechanism by which one (verbal or non-verbal) action or part of it prefigures the next, thus creating expectations on the side of the participants on the further trajectory of a configuration (syntactic, prosodic) or an activity underway. Projections are based on gestalts, which are schematic patterns created by speakers and identified by listeners that are perceived as an organized whole with the production of more or less well predictable elements in a gestaltconforming way. Gestalts can be syntactic (referring to linear structure and constituency), prosodic (referring to intonation patterns), semantic (referring to propositional content), and pragmatic (referring to the way a particular conversational activity is usually carried out).2 Speakers project and listeners are able to process projections based on their experience on how gestalts are typically structured and thus may have reached potential completion: syntactic projections (e.g. a saturated valence pattern), prosodic projections (e.g. a terminal pitch contour), semantic projections (e.g. a complete propositional unit) and pragmatic projections (identifiability of a conversational activity), all of which interact in complex ways (Auer, 1996b; Ford and Thompson, 1996). Projection as a major cognitive principle of speech production and processing requires us to focus on the temporality of speech since the trajectory of a projection refers to the temporal sequence over which it comes to completion. Due to the participant’s ability to project continuations, the chances to predict the further trajectory of a structural unit underway continually increase during the production of a unit of talk, until projections on relevant dimensions are complete. A “final field” arises when the speaker has reached a point in the production of a structural unit at which projections relevant for utterance interpretation are completed and at which the emerging unit is potentially independent of any following unit (“]1”). The final field expands a unit of talk and creates a new point of (potential) completion (“]2”), as illustrated in Fig. 1. The sequence produced up to a point in time “]1” is unit-of-talk-constitutive, that is, it can potentially represent a unit of talk of its own. The point of potential completion at “]1” is not to be understood in categorical terms, like the end of a sentence. It is a decision made by the participants, based on whether or not listeners are able to identify the illocutionary point of a unit of talk underway and whether or not speakers need to deal with the resolution of syntactic, semantic, pragmatic and prosodic projections. However, it is not necessary that projections on all levels are completed in order to understand “[ ]1” as complete. Syntactic completion, for instance, is often secondary, that is, speakers often do not complete projections on the syntactic level, and yet a unit of talk can be interpreted as complete when projections on other levels, e.g. the pragmatic one, are (almost) completed. The reason is that a syntactically “incomplete” structure can usually be interpreted against the context, co-text, or knowledge on how a particular activity is usually expressed, and thus allows for the identification of the kind of activity a speaker intends to perform without syntactic completion, as illustrated in (3). (3)

67 68 69 70 71

B: A: /

B: A:

I just wondered if you’d seen it. yes¼ ¼I have. and you- >OH you were the one who-< I was the one who’s seen it.

[ICE-GB S1A-006]

B’s sudden realization of the fact that of all people B knows it was A who had already seen the film is syntactically “incomplete” in the sense that it lacks the continuation of the relative clause that is normally projected by an initial relative pronoun. Yet, the unit is interpretable for A as the expression of an instantaneous change of the cognitive state of the speaker. Since there are not very many syntactic alternatives in play after the production of the relative pronoun “who” (a finite verb and an object, e.g. “has seen it”, are the most likely candidate items to follow) the continuation is highly predictable and thus not necessarily required to identify the activity carried out by B. Speaker A then confirms B’s deferred cognitive realization. Similarly, projections on the semantic level can remain unresolved when the unit-of-talk-thus-far makes the communicative goal of the speaker sufficiently clear. In (4), for instance, speaker B produces a unit of talk (line 152) that is semantically

2 Prosodic gestalts are difficult to describe as they are based on small-scale projections (Auer, 1996b: 69) and since it is problematic to define a bounded unit in prosody. Intuitively, listeners expect a possible completion point to coincide with the occurrence of the highest or lowest pitch protrusion in a globally, steadily rising or falling pitch contour. However, globally falling or rising contours are relatively infrequent – with more complex contours (e.g. rises and falls) it is impossible to predict how many and what kind of accent units or non-accented syllables will follow a final pitch accent. Nevertheless, projection in prosody is certainly possible to some extent, at least schematically, based on certain prosodic cues such as changes in pitch, intensity, duration, tempo and voice quality (Pierrehumbert and Hirschberg, 1990; Chafe, 1992). The end of an intonation unit, for instance, tends to be lengthened, decelerated, and is often characterized by “creaky” voice and a terminal pitch contour.

A. Haselow / Language Sciences 54 (2016) 77–101

[

]1

85

]2 time

PROJECTIONS (syntactic, prosodic, semantic, pragmatic)

FINAL FIELD

Fig. 1. Schematic representation of the final field.

(as well as syntactically) incomplete in the sense that its propositional content remains unfinished and is thus devoid of truthconditionality. (4)

150 151 152 153 154 155

B:

/ A: B:

but like the record company then refused to pay me without having seen the pictures¼ they said (.) Terry O’Neill’s done it  you know . (.) as if that would sort of [uh,] [so] have they used your pictures in the end of the day. (.) no they didn’t actually. which is uh, but don’t mind that¼uh, [ICE-GB S1A-052]

The unit of talk in line 152 semantically lacks the judgment of the speaker projected by the initial part, that is, B does not provide the part that expresses why the fact that “they said Terry O’Neill’s done it” is unreasonable and no justification for the state of affairs described by B. In spite of the fact that the unit is incomplete in several ways speaker A does not wait for B to continue, but adopts the turn in order to produce an information-seeking request. That the unit does not cause any trouble at this point in ongoing discourse (e.g. motivating A to initiate a repair sequence) can be explained with the fact that it is pragmatically interpretable as a particular kind of activity that can be accounted for: the unit can be heard as expressing B’s doubts about the plausibility of what “they” said. Note that the unit also leaves projections on other levels open: the unit “as if that would sort of” syntactically projects a predicate structure (e.g. “. play any role” or “.justify it”), which is not produced, and it is prosodically incomplete as it lacks a terminal contour, ending in comma intonation, i.e. with mid-level pitch plus a filled pause (uh), which projects more to come. Such examples illustrate that participants tolerate the incompletion of projections on different dimensions to a certain extent if the context and the co-text provide sufficient interpretive cues and if the unit-in-progress is advanced to such an extent that the remaining part can be more or less easily predicted. Referring back to Fig. 1, which illustrates the conceptualization of the final field, we can conclude that a unit “[ ]1” is not only perceived as complete if and only if the projections on all levels are resolved. It seems intuitively plausible that pragmatic projection is the crucial factor in deciding about the potential completeness of a unit of talk in the sense that unresolved projections on other levels can be tolerated as long as speakers have provided sufficient cues as to make the communicative goal identifiable. The relative weight of different kinds of projections in utterance interpretation requires further investigation. Identifying potential completion points is thus a complex issue, requiring the simultaneous tracking of several kinds of projections, and not tied to the rules of syntax alone. The role of syntax is important in so far as it is one important dimension on which projections are build up and completed, thus guiding the speaker’s planning activities and the listener’s anticipation of the further trajectory of a unit of talk-in-progress.3 The temporal slot between “]1” and “]2”, which is called the ‘final field’ here, represents the local reactualization of a sedimented structure that involves a fixed set of elements that are part of the participants’ knowledge store. Sedimentation is the result of routinization of communicative tasks that, in the case of the final field, are relevant at a point at which the speaker has reached the potential end of a unit of talk in progress. Elements in the final field are independent of the kind of structure produced until “]1” on all levels of projection, i.e. they are typically outside the projections within “]1”. However, they are interpretable only with reference to the unit they are attached to, they have no illocutionary force of their own, and do not represent a new conversational activity, but are part of the activity performed with the prior unit. Note that in spite of their integration into the illocutionary force of the prior unit, units in the final field can reach relative semantic, prosodic and syntactic independence. Grammatically, the final field is defined as a time slot available for the production of elements that are only loosely linked to the unit to which they are attached as they are not part of morphosyntactic dependency relations, lacking a microgrammatical constituent status, and are not potentially integratible into the preceding unit. This definition excludes adverbs with lexical meanings, which are constituents of a microsyntactic unit contributing semantic content, and all kinds of

3 Examples are syntactic rules that require the speaker to end with the production of a particular constituent, such as the German verbal brace (Auer, 1996b: 62–63), which requires speakers to close a structural unit with an infinite lexical verb with compound verb structures (the finite auxiliary forming the left brace, the infinite main verb the right one), or English verbal particles which, provided that they are movable, are expected to occur at the end of a structural unit when the particle which is not produced immediately following the verb (e.g. “he brought this question up”).

86

A. Haselow / Language Sciences 54 (2016) 77–101

postposed obligatory arguments of the verb. It also excludes so-called “right-dislocations”, which are linked to the prior unit by a co-referential pronoun, that is, they represent an argument of the verb that is realized twice, first as a pronoun and later as a full lexical unit. However, it encompasses lexical or phrasal units that have no such pronominal co-referent, but are only loosely attached to it, such as increments (e.g. “like the South of France” in (2)). These increments are not taking part in a projection, nor do they project themselves, and have no morphosyntactic dependency relation to any other unit. The following section (Section 3) provides an overview of the different types of elements in the final field and briefly discusses their functions in order to derive a common functional core. For practical reasons, I will only focus on conventionalized elements with procedural rather than conceptual meaning in the sense of Blakemore (1987) in English, excluding lexical expansions such as clausal increments.

3. Elements in the final field: an overview If we define the final field as being unrelated to the morphosyntactic dependency relations of the preceding structural unit and as an expansion of this unit, it seems natural that the final field does not only host single-word units, but also more complex ones, e.g. independent if-clauses (e.g. if you don’t mind me saying), general extenders (e.g. and stuff, and things like that), or comment clauses (e.g. I think, I believe). Such units do not obey preconceived ideas of syntactic units based on lexical compositionality and exhibit a reduced semantics where only in some cases residual components of the original semantic content of the individual parts are identifiable. They form constructions in the sense of Construction Grammar (Goldberg, 1995; Croft, 2001), that is, conventionalized form-meaning pairings with little or no internal compositionality serving, in our case, a variety of metatextual functions and showing different degrees of specificity ranging from more schematic to fully specified. As constructions, such units are complex signs that cannot be broken down into separate semantic, syntactic, pragmatic or functional modules, i.e. they lack analyzability. Since the form-meaning pairings always exhibit some degree of arbitrariness, their meaning (in Croft’s, 2001 sense) has to be defined in a wider way, comprising all the conventionalized aspects of the use of a construction (Croft, 2001: 19).4 The constructional character can be related to sedimentation in the sense that the respective units serve routine tasks for which speakers, over time, develop habitualized ways of expression. At least for the kinds of constructions discussed here one may follow the usage-based claim that sedimented patterns like the final field encode best what speakers (need to) do most frequently (Bybee, 2006), given that many communicative tasks are recurrent. Under this view, grammar emerges through practice as a by-product of frequent usage of particular forms for specific recurrent purposes in particular time slots.5 Historically, as discussed in Traugott (in press), units in the final field were once part of what I call ‘microgrammar’ here, which explains why there is at times syntactic and pragmatic ambiguity between one (the older microgrammatical) use or the other (newer, macrogrammatical), as shown for some individual unit types below (see also e.g. Brinton [2008] for comment clauses; Pichler and Levey [2011] for general extenders). The development of macrogrammatical units provides evidence for the assumption that the boundary between both domains of grammar is not always discrete. Based on the definition of the final field provided in Section 2, the types of elements found in this slot in the ICE-GB can be categorized into larger classes. However, it is not easy to decide whether a categorization based on formal or on functional aspects is appropriate. The problem of using formal criteria is that almost all of the elements in the final field violate many of the defining features that characterize the lexical or syntactic category they seem to correspond to according to their form, i.e. they “look like X”, e.g. like a clause or an adverb, but strictly speaking they are not X. Therefore, classifying units in the final field into formal categories of grammar writing in the structuralist tradition is often misleading, above all since these categories represent a long written-biased scholarly tradition (Linell, 2005). Units like if you want or if you don’t mind me saying, for instance, are formally affiliated with and thus “look like” if-clauses as they seem to exhibit a subject-predicate structure introduced by the subordinator if. However, the term is, in principle, misleading as such forms are not interpretable as genuine clauses introduced by a subordinator: they do not represent a free combination of individual constituents, but relatively specific constructions with their own pragmatic, rhetoric and discursive properties (see Brinton, 2008: 163–166); they are semantically not expressing a condition under which a state of affairs expressed in another clause is valid, and they are neither grammatically, nor semantically representing a subordinate clause (no embedding into a matrix clause, no semantic dependence on the proposition of the prior clause).6 Another example are units known under the label “extenders”, such as and stuff or and things like that, which appear to be interpretable as coordinated noun phrases from a structuralist perspective but, again, represent relatively specific constructions whose meaning cannot be generated by rules of linkage. They are semantically odd as there is often a mismatch between the morphosyntactic and semantic properties of the antecedent unit and that of the generic noun of the extender, e.g. you have to speak clearly and stuff (Pichler and Levey, 2011: 444–445).

4 As constructions, units in the final field form a dense network, sharing a set of features that link them into cluster: they are used in the same time slot, typically lack prosodic prominence, have no projective potential, and express metapragmatic meanings. 5 For further discussion of this view see Günthner (2011), Hopper (2011) and Pekarek Doehler (2011). 6 Note that the concept of ‘subordination’ has generally been subjected to considerable scrutiny (e.g. Matthiessen and Thompson, 1988; Thompson, 2002).

A. Haselow / Language Sciences 54 (2016) 77–101

87

An alternative way of classifying elements in the final field is to use functional features, but this is not unproblematic either. First, it is difficult to decide which functions have categorical character, and how fine-grained the functional classes should be. Secondly, categories based on functional aspects are not common in grammatical analysis, which tends to operate with formal categories. Attempts to develop an appropriate terminology based on functional aspects has often led to terminological confusion. Elements like final then, though or anyway have, for instance, been categorized as “conjuncts” (Quirk et al., 1985), “linking adverbials” (Biber et al., 1999), “connective adjuncts” (Huddleston and Pullum, 2002: 779), “adverbial connectors” (Lenker, 2010) or “final particles” (Haselow, 2013; Hancil et al., 2015) in the literature, and thus been labeled on the basis of a combination of formal (adverbial, conjunct, particle) and functional (connecting, linking) features. Units like and stuff or or anything have been labeled “referent-final tags” by Aijmer (2002), “extender tags” by Carroll (2008), and “general extenders” by Pichler and Levey (2011). It is not difficult to find other examples. The problem of using a structuralist approach to “extra-clausal”, macrogrammatical elements is thus that they usually do not fit into any of the existing “classical” categories of grammatical analysis (parts of speech or syntactic categories), neither formally nor functionally, since for many of the functions they fulfil no categories exist in sentence-based grammar. The continuous renewal of labels given to these elements is unsatisfying as it disintegrates the terminological inventory of the analyst. A further problem of using functional criteria is the one-to-many mapping between individual unit types and concrete functions. A single unit may, for instance, be used as a mitigator in one use, but as an intensifier in another (e.g. vocatives or tag questions), that is, a single unit is often used for various functions, and a single function is often expressed by more than one element. The fact that formal labels (e.g. if-clause, tag, adverb/ial) are more widely used for macrogrammatical elements than functional ones shows that traditional concepts are too deeply rooted in grammatical description to be abandoned, even if the element they denote is, at best, a marginal member of the respective class. The present study is based on a compromise solution that consists in using the traditional formal categories, which are important for identifying the different unit types among analysts, in combination with the specifying attribute “final” (e.g. final independent if-clause, final comment clause), which is supposed to indicate the particular functions that all unit types in the final field have in common. Since all of the elements discussed here are typically produced at the end of a unit of talk, we can assume that they are cognitively activated near a point in time at which the speaker feels to be approaching the end of a unit of talk. An overview of the different elements in the final field, based on the ICE-GB, is provided in Table 2.7 The order of elements reflects increasing internal complexity of the elements, ranging from low complexity (a single lexeme) at the top of the table to higher degrees of complexity (a more complex, “clause-like” construction) at the bottom. The functional values given in the table are discussed below. Elements in the final field are considered “macrogrammatical” elements, and thus part of the grammar of English, rather than being thrown into the unspecified category of “pragmatic markers”. As elements of macrogrammar (i) they are subject to particular serialization principles as they do not occur randomly in a unit of talk, (ii) they form conventionalized formmeaning pairs, (iii) they are organized paradigmatically, at least to certain degree, and (iv) they are integrated into the illocutionary force of the unit they accompany. Moreover, when two or more of these units co-occur their order is not random, which suggests that the final field has a syntax of its own (see Section 3.8). The different categories in the final field form a temporary subsystem of grammar (Hopper, 1998: 158) that is continuously re-organized and thus in flux, as the development and variation in the use of tags (e.g. tendencies toward invariant innit?), general extenders (see Pichler and Levey, 2011), or final particles (growth in the number of lexemes that develop uses as FPs, such as conjunctions or linking adverbs, Lenker, 2010; Kim and Jahnke, 2011; Haselow, 2013) shows. Elements in the final field are characterized by the following features: they (i) (ii) (iii) (iii) (iv) (v) (vi) (vii)

are used predominantly in spoken discourse; are not potentially turn-constitutive as they are backwards-oriented; make no contribution to the propositional content of an utterance; have procedural rather than conceptual meanings; have various functions on the metatextual and interpersonal level; are not integrated into the morphosyntatic structure of the unit they follow; are morphologically invariant and tend to have constructional character (forming schematic or specific constructions); are functionally variant when produced at other points in time in utterance production.

The following analysis of the unit types listed in Table 2 in their individual contexts seeks to identify core functions of the final field that reflect the cognitive tasks speakers deal with at the end of an emerging unit of talk.

3.1. Final adverbs This category includes units that “look like” adverbs, such as final uses of really or of course, but which violate major classdefining features: they do not depend on or modify another category (nouns, adjectives, verbs), and even postulating clausal

7

For a discussion of some of the elements in Table 2 from a historical perspective see Traugott (in press).

88

A. Haselow / Language Sciences 54 (2016) 77–101

Table 2 Elements in the final field. Form (“looks like .”)

Function

Example

(i) (ii) (iii) (iv)

Adverb (incl. response adverb) Linking adverb Noun phrase (vocative) Coordinated noun phrase

.of course, .really, .yeah .then, .though, .anway .Jim,.mate .and stuff, .or something

(v) (vi)

Clause (comment clause, matrix clause?) (elliptical) Interrogative clause

(vii)

(independent) If-clause

indicating shared knowledge, intensifying assertions utterance-connecting, modifying illocutionary force expressing listener-orientation, strengthening, mitigating reference to shared knowledge, completion point for lists, reports, indirect speech or illustrations, indicating vagueness expressing epistemic stance, marking an opinion strengthening illocution. force, reference to shared knowledge, facilitating involvement of addressee, expressing listener-orientation contact to listener, hedging, comment on adequacy of expression

.I think, .I believe, .isn’t it?, .can you? .if you like, .if I may say so

or sentential scope (as terms like “sentence adverb” suggest) is misleading in so far as they are not only attached to clausal or sentential units, but also to lexical, sub-clausal or larger textual units. It is important to note that adverbs in the final field are not the kind of adverb we find in the “end-position” of a sentence (Quirk et al., 1985: 490; Huddleston and Pullum, 2002: 575; Biber et al., 1999: 772), which is defined as the part of a sentence following all obligatory elements, and usually involves time, place and circumstance adverbs. The adverb-like elements referred to here are not integrated into the morphosyntactic dependency relations of a sentence and peripheral to the content of the unit to which they are attached in the sense that their deletion would not result in a loss of semantic content; when they are produced within the structural unit they accompany they adopt a different function. Adverbs in the final field are predominantly used to fine-tune epistemic stance, either indicating certainty (e.g. of course, yeah) or uncertainty/impreciseness (e.g. usually, probably, possibly), the latter of which may also mitigate an authoritative attitude, whereas others strengthen assertive force (really, absolutely). In this sense, they have procedural rather than propositional meaning in the sense of Blakemore (1987), that is, they provide an interpretive cue to the addressee guiding the processing of a unit of talk rather than expressing lexical content. Final really is an example for an adverb that retrospectively strengthens the assertive force of a proposition. In the final field, it has a broad scope over the unit it is attached to. (5)

174

B:

175

but (.) I did some volunteer work over the summer just one day a week (.) at a ((coughs)) are they called Minds or Drop In? (.) Day Centre? and that was quite successful really. (.) [ICE-GB S1A-035]

In (5), B first mitigates her assessment of her volunteer work in a day center (“quite successful”), probably in order to avoid self-praising, but eventually strengthens the core of the propositional unit (‘it was successful’) before she reaches a TRP by using final really. This way, B manages to follow the modesty maxim, on the one hand, and to emphasize the success of her work, on the other hand. As all units in the final field, really has a broad scope over the entire unit to which it is attached. This way, the emphatic meaning refers to the state of affairs expressed in a unit of talk as a whole. A strengthening effect on the illocutionary force of a unit of talk can also be reached by what I call echoing reinforcers, that is, items which normally serve as response adverbs expressing affirmation (yes, yeah) or negation (no) and which can occur as utterances of their own in this function, but which are also often used in the final field, where they underline agreement and certainty. They are usually attached to a responsive turn, where they reinforce an affirmation or negation and reassure information provided in the preceding discourse unit (e.g. that’s right yeah, ICE-GB S1A-016, 199-D; no he is not no, ICE-GB S1A015, 150-A). Thus, they indicate that a proposition is in accordance with an expectation or opinion of the addressee or with an assumption that had already been expressed or that had been “hanging in the air”, that is, a proposition that is expectable and thus plausible to the co-participants. In this use, they often accompany a unit of talk that merely repeats or “echoes” what the prior speaker has just said and reinforce its validity (e.g. B: it looks swollen that foot – A: it is yeah, ICE-GB S1A-047, 262–263). In such reiterative uses, final yeah/no indicate intersubjective understanding, a shared point of view, opinion or observation, and thus emphasize agreement. At the same time, the speaker retrospectively confirms his/her own utterance, signalling that the information can be added to the common ground with no further discussion. In some cases, speakers add final yeah to an utterance that expresses an inference, retrospectively indicating plausibility and implicit confirmation of what the prior speaker has left implicit, as in (6). (6)

242

A:

243

C:

yeah (.) I mean there’s quite a sort of overhead machine overhead on Windows isn’t there but I mean you you’ve got to have a big powerful machine to run it j or plenty of RAM. at the same time it’s not going to have Windows then yeah. [ICE-GB S1A-029]

Again, final yeah is a retrospectively oriented device for reinforcing the validity of the propositional content of the unit of talk it accompanies and expressing that the content is in line with the prior speaker’s point or argument. It also marks selfassurance on the side of the speaker that the content of the utterance is true and that it was worth saying. Final yeah is thus a linking device relating a unit of talk to the co-participant’s view with which it coincides.

A. Haselow / Language Sciences 54 (2016) 77–101

89

A similar function on the interpersonal level is fulfilled by final of course, which alludes to shared background knowledge or shared assumptions between speaker and addressee on the basis of which the propositional content of the unit it is attached to appears plausible and expectable (see also Simon-Vandenbergen and Aijmer, 2002). Only remnants of the original meaning of marking a state of affairs as the natural consequence of something (Lewis, 2003) can be found when of course implies that, based on what speaker and addressee know about the world and about each other, the proposition is (almost) self-evident and conforming to expectations, as in (7). (7)

97 98 103 104 105

B: [.] A: B: A:

i heard you speaking Welsh yesterday¼ ¼it was really nice. were you understanding it? no not a [WORD. j of [course ]. [no ]. ((laughs))

[ICE-GB S1A-069]

In (7), an evidential interpretation “as a natural consequence of me not knowing Welsh” is certainly not excluded. However, since B cannot assume A to know that she does not know Welsh, it seems more plausible to interpret the use of of course as marking the proposition as intersubjectively evidential, based on world knowledge (Welsh is a language that only few people are able to understand, let alone to speak) and degree of likelihood (B is not among these people). The function of of course is thus metatextual in that it serves as a rhetoric device to mark the content as plausible and expectable in given situation (‘you will understand that p’) and to background the news value of the unit it accompanies. Many adverbs pose serious problems for grammatical categorization, due to the various topological options, variable scope, and varying degrees of syntactic and prosodic integration. Of course, for instance, is classified as a subjunct, conjunct, and disjunct in Quirk et al. (1985: x9.5), as a pragmatic particle by Holmes (1988), and as a pragmatic marker by SimonVandenbergen and Aijmer (2002).8 By classifying the kinds of lexemes discussed above as adverbs/adverbials, the important distinction between conceptual (lexical) and procedural meaning is lost: in contrast to adverbs as defined in standard grammars, these lexemes have no effect on the semantic content and truth-conditionality of the preceding unit. Rather, they serve as a last-minute device to modify the illocutionary (usually assertive) force of a unit of talk, to indicate the epistemic status of the information expressed, and to indicate addressee-oriented meanings (e.g. shared knowledge) before turn transition or topical moves. Prosodically, adverb-like units in the final field show different degrees of prosodic integration into the intonation contour of the unit they accompany, ranging from separation by a micropause or prosodic cesura to full integration. They are typically produced with emphatic prosody when they strengthen illocutionary force, but otherwise tend to be prosodically unmarked, which is consistent with the observation that a loss of semantic weight and lexical meaning correlates with a loss of prosodic prominence (see the theory of intonational meaning in English by Pierrehumbert and Hirschberg, 1990). 3.2. Comment clauses The final field may host a fixed set of configurations that look like matrix clauses with a subject (usually a first person singular pronoun) and a mental verb that requires a complement clause as an object, such as I think, I believe, I guess, I mean, also you know or it seems. Such units, as well as forms like to-infinitive clauses, such as to be honest, or finite clauses introduced by as (e.g. as I see it), which usually function as adverbial clauses, are analyzed as parenthetical disjuncts by Quirk et al. (1985: 1112), which suggests non-clausal status. They are prosocially marked by increased speech and lowered volume. For practical reasons, I will restrict the category of comment clauses to subject-cum-mental verb combinations. The syntactic status of comment clauses is often open to interpretation. In English, for example, utterance-initial I think or I believe may be ambiguous between a main clause taking the rest of the sentence as object, and a comment clause or discourse marker. In medial and final position, the use as a comment clause is more likely, but it has still been assumed that a main clause analysis is possible and that medial/final comment clauses are the result of a syntactic movement operation (most recently Newmeyer [2015]).9 Since the idea of movement operations is excluded in the present approach, which is based on the linear, emergent nature of speech, it will be argued that final comment clauses no longer function as matrix clauses with an embedded, preposed subordinate clause. Rather, they form the final field of the unit they accompany and thus follow a structural unit without entering a microgrammatical relation with it, that is, without hierarchization and embedding. From a macrogrammatical perspective, the syntactic structure involving a comment clause should be analyzed as consisting of an independent (non-subordinate) assertion (in any syntactic format) and a metatextual unit in the format of a fixed construction (“pragmatic tag” [Weinert, 2012: 237] or “epistemic parenthetical” [Thompson and Mulac, 1991]) that is integrated into the illocutionary force of the unit it accompanies. The informational import lies on the assertion, which also

8 As a subjunct, of course is relatively integrated into the clause and belongs to the group of ‘emphasizers’, expressing that what is being said is true; as a disjunct it is also emphasizing, but has a wider scope and occurs at the boundaries of a clause, usually initially; as a conjunct it is syntactically peripheral and has a linking function, marking resultative and concessive relationships. 9 In languages such as Icelandic and German, utterance-medial and -final comment clauses are verb-first (e.g. held ég, glaub ich ‘I believe’) rather than verb-second and thus not used with main clause syntax, so that a movement analysis becomes less likely (see Dehé, in press).

90

A. Haselow / Language Sciences 54 (2016) 77–101

typically contains the topical subject, which is why “conversationalists are normally oriented to the material in the complements and independent clauses and not to the CTP-phrases” (CTP ¼ ‘complement-taking predicate’) (Thompson, 2002: 154). The case of comment clauses is congruent with the observation that speakers favor arrangements of clauses that are not complements syntactically, but combinations of evidential or epistemic “phrases” and declarative clauses (Thompson, 2002) linked by loose discourse-pragmatic dependency relations. In the final field, the mental verbs involved in unintegrated comment clauses are no longer complement taking and projecting, but express an epistemic-evidential evaluation relating to the prior message. A semantic reading of these verbs cannot be obtained. Final I believe, for instance, does not indicate the speaker’s belief in the truth of an assertion, but expresses an epistemic evaluation or an opinion to which the speaker is committed, but which is mitigated for some reason, e.g. due to a lack of epistemic certainty, as in (8). (8)

68 69 70

A:

yeah he had some sort of disease. undiagnosed  I believe so he had these big big cracks in his face.

[ICE-GB S1A-015]

The use of I believe in (8) lends the adjectival increment (Schegloff, 1996: 59; Couper-Kuhlen and Ono, 2007) “undiagnosed” the character of an assumption rather than a fact, based on some piece of evidence that the speaker might have (e.g. from what “he” has told him), on an inference, or on a vague memory that cannot be fully recovered at the moment of speech. An epistemic meaning is, however, often absent with comment clauses. Final I think, for instance, is predominantly used as a mere opinion marker, as in (9). (9)

150

B:

151

you can’t communicate with people (2.5) solely (2.0) through uh you need- (.) you need physical contact too (1.0) in everyday things not (1.0) uhm (3.0) not just once in a while. (3.5) j I think. [ICE-GB S1A-003]

The use of I think in (9) is not epistemic since the speaker, who is a dancing teacher working with disabled people, is not talking about a fact for which one needs objective evidence, but about his subjective experience with people at his job, from which he learned that an individual needs more than “words” in everyday interaction with other people. It is thus based on subjective experience, that is, B’s assertion is an opinion that is not (supposed to be) epistemically challengeable, but only by diverging subjective experience. I think is also often used as a hedging and face-saving device on the interpersonal level, weakening or mitigating the force of a claim or an immediate reaction in such a way that it “leaves room for intervention by the interaction partner” (Nuyts, 2001: 165). Another, functionally different kind of clause-like element in the final field is I mean. Final I mean, which in uses other than in the final field indicates that the speaker modifies his/her own talk (Schiffrin, 1987: 299) by expanding or reformulating an idea or opinion just expressed, is not projecting, but marking the preceding utterance or discourse segment as potentially expandible by a reformulation or an explication of an inference, but being potentially interpretible by mutual understanding (‘you know what I mean’). I mean in the final field is thus ambiguous between concluding a unit of talk, but leaving the speaker the option to continue (typically with a reformulation or expansion of a point or an idea), and signalling readiness to cede the turn and prompt the addressee to take over the turn, leaving the projected continuation or an inferred meaning pragmatically “hanging”, as in (10). (10)

32

B:

33 34 35

A: B:

how detailed do you want me to get really because I could: (.) go off (.) I could talk quite a long time about my father on a question like this and in what kind of way (3.0) do you want him [described ] as a as a person? (..) [well ] well uh- bearing in mind that I know nothing about him at all¼ I mean mhm [ICE-GB S1A-076]

In (10), A answers B’s request for information as to how detailed he should get in talking about his father with an assertion that only indirectly answers the question, that is, from which B needs to infer that he should offer a detailed account. By using final I mean, A indexes a shared understanding of the implications of her assertion, which are left hanging, but could be explicated in a continued turn. The fact that B displays understanding indicates that he is able to interpret the inference and understands I mean as an offer for turn-transition. A further frequently occurring comment clause in the final field is you know, which has an interpersonal meaning similar to that of final I mean. Its main function is to indicate that the speaker considers the addressee able to grasp the implications of an utterance, based on shared knowledge and on what s/he has said thus far, and that there is no significant discrepancy between the mental world of the speaker and that of the addressee (Schoroup, 1985: 102). The use of you know in (11), for instance, is clearly an appeal to enrich the information provided by A by drawing inferences in order to understand the reasoning behind the decision to “wait a couple of days”. (11)

97 98

A:

I’ve got her phone number right here. >the thing is what’s the etiquette of this¼

A. Haselow / Language Sciences 54 (2016) 77–101

99 100

C:

91

¼you’re then meant to wait a couple of days before you ring them up or else it appears uncool you know?< [ICE-GB S1A-020]

A can assume that his indecisiveness and his preference for waiting before making the call is plausible for his addressee and that the “etiquette of this” is part of a collective behavioral code that the participants can be assumed to share. Final you know indexes shared knowledge about such situations and its risks (calling too early might be interpreted as intimidating) and underlines the speaker’s assumption that, based on the relevant knowledge and experience, his line of thinking is comprehensible for his addressee. As has been shown, comment clauses do not always mark epistemic-evidential stance, but classify an utterance as an opinion (e.g. I think, I believe) or refer to shared understanding (you know). Research on the prosodic realization of comment clauses (Dehé, 2009, in press; Dehé and Wichmann, 2010a, 2010b) has shown that a variety of intonational phrasing patterns is possible with comment clauses, allowing both for prosodic separation such that the comment clause is phrased in its own Intonational Phrase (IP), and prosodic integration, such that the comment clause is phrased in one IP with surrounding material. The relation between meaning in context (procedural, propositional) and prosodic realisation has been investigated in a number of studies. Dehé and Wichmann (2010a: 18), for instance, show that I think/I believe are “semantically variable in their effect on truth conditionality” depending on whether they are prosodically stressed or not. There is evidence that comment clauses whose primary function is one of mitigation are typically unstressed and integrated (Dehé, in press). 3.3. General extenders General Extenders (GEs) are expressions that typically occur final to a clause or a unit of talk, such as and stuff like that, and that, or something or or whatever. The term General Extender refers to the fact that these expressions are non-specific (“general”) and that they extend an otherwise complete “grammatical” unit (Overstreet, 1999: 3). Pichler and Levey (2011: 448–449) use the following schema to describe the prototypical structural configuration of GEs:

e.g.

(connector) and and or

(modifier) all,

(pro-form/generic noun) things whatever.

(similative)

(deictic),

like

that,

From a Construction Grammar perspective, GEs are semi-fixed, complex, and unbound constructions, that is, form-meaning pairs that occur in a specific or schematic format and whose meaning is not compositional in the sense that it can be generated by rules of linkage of semantics and syntax. The combination of individual elements is variable, as indicated by round brackets in the schema. The pro-form includes generic referents such as anything, stuff, everything, something, or things, and is often combined with a connector and/or an optional similative-cum-deictic combination. The connectors and/or provide the basis for a structural subdivision between two main types of GEs, adjunctive variants (e.g. and stuff, and things) and disjunctive variants (e.g. or something like that) (Overstreet, 1999: 3–4). The presence of a connector is common, but absent in more fixed structural configurations, e.g. things like that, anything like that, type of thing. GEs differ greatly in syntagmatic length, and sometimes shorter variants derive from longer ones, having undergone a loss of lexical material, e.g. or something < or something like that (Cheshire, 2007; Pichler and Levey, 2011). Moreover, some variants seem to be more advanced in terms of phonological reduction (e.g. or something > o’ summin’) than others (Cheshire, 2007; Tagliamonte and Denis, 2010). Variationist research on GEs in Canadian English (Tagliamonte and Denis, 2010) and British English (Pichler and Levey, 2011) has shown that there is considerable synchronic variability in terms of the formal, morphosyntactic and functional properties of GEs, and instability in the social conditioning of variant choice of GEs. The system is highly fragmented with a large number of relatively infrequent variants and a small number of highly productive (e.g. and that, or something) ones. The overall system of GEs is continuously in flux, with ongoing renewal of individual variants and synchronic variability in the use of variants (e.g. and things like that coexisting with and things) and no clearly discernible preferences for or tendencies toward the use of longer or shorter variants. No stable system has as yet evolved (Pichler and Levey, 2011: 462).10 Only some individual variants have become synchronically fossilized in terms of their form, function, contextual uses and social embedding, such as and that, which typically follows nominal referents and is currently becoming the default variant among young speakers (Pichler and Levey, 2011: 462). A strong indicator for the microsyntactically unintegrated, macrogrammatical character of GEs is, next to their constructional form and thus lack of an internal syntax, the typical mismatch between the morphosyntactic and semantic properties of the generic noun (e.g. thing, stuff) and those of the antecedent phrase (NP or AdjP) to which GEs are attached, as mentioned above and illustrated in (12).

10 Assuming that GEs are the result of grammaticalization, this phenomenon is consonant with observations that, once initiated, none of the processes constituting grammaticalization necessarily has to go to completion or to move into an identifiable direction (Traugott and Trousdale, 2010: 28).

92

A. Haselow / Language Sciences 54 (2016) 77–101

(12)

a.

44 45 46

A: B: A:

b.

277

B:

this person over in the Survey of English Usage is transcribing it. oh right. but the names are changed and things like that so, (.) then they won’t know. [ICE-GB S1A-068] they sent one to my mother after she di:ed or something. [ICE-GB S1A-007]

In (12a) there is a mismatch between the morphosyntactic and semantic properties of the antecedent unit and the properties of the generic noun things in the GE: the generic noun is not attached to an NP with one or more countable common nouns, but to a clausal unit, and it does not semantically refer to entities, but to an activity (‘changing names’). The GE and things like that leaves it to B to think of further activities related to the activity of transcribing talk. In (12b), the indefinite pronoun something in generic use is not appended to a noun phrase with a countable, singular, inanimate noun, but has indeterminate scope, which is either over the prepositional phrase (“after she died”) or over the utterance as a whole, either of which is retrospectively marked as a vague approximation to how or when exactly the event occurred in reality. Or something indicates that there are semantic alternatives to the proposition over which it has scope. As macrogrammatial units, GEs are not analyzable as coordinated noun phrases or adjuncts, and thus not involved in the syntactic dependency relations of the prior unit. They form the final field, contributing to the way in which of the unit of talk to which they are attached is to be interpreted. Four functions can be distinguished: (i) referential, (ii) turn-taking, (iii) interpersonal, and (iv) cognitive functions. In referential uses they “cue the listener to interpret the preceding element as an illustrative example of a more general case” (Dines, 1980: 22), that is, the preceding referent is marked as a member of a more general set of things or of a particular conceptual category, or it represents a part of a larger, unspecified context. The turntaking function consists in marking a transition-relevant place (TRP), thus legitimating turn transition. GEs often mark the end of a list of items or of a quotation that could be expanded, but is suspended for some reason, indicating the speaker’s desire to close a unit of talk and probably the topic altogether and to “punctuate” or bracket units of discourse (Cheshire, 2007). The interpersonal function of GEs is to express informality and solidarity (Cheshire, 2007) which, one should add, is a pragmatic effect deriving from the implicational character of GEs: marking a unit of talk as potentially expandible, lacking precision or epistemic certainty presupposes that the speaker can rely on the addressee’s ability and, above all, his/her willingness to accept the conversational vagueness and the indication of mutual understanding implied in the use of GEs. In cognitive terms, GEs appeal to shared knowledge/common ground, where an exhaustive set or a detailed expression is not required for mutual understanding and one item is sufficient to evoke a mental frame of utterance interpretation. GEs serve as a retrospective hedge indicating vagueness and potential expandibility of a conversational activity, but in a more global sense referring to how something is expressed by the speaker and should be processed by the addressee, namely as an approximation and as potentially expandible, leaving open how exactly the speaker would continue. This implies that the speaker refers to shared background knowledge and understanding on how what was said thus far can be continued in a given context. GEs indicate or allude to intersubjective understanding similar to other elements in the final field discussed above (you know, I mean, of course) and signal that a higher degree of precision is not required at a particular point in discourse. In this sense, they further the progression of talk when further examples or more precise expressions are cognitively unavailable at the moment of speech or unnecessary, or when accuracy of information would unnecessarily threaten the smooth flow of discourse. GEs give the addressee “a rough but sufficiently exact idea about a certain state of affairs for the general purpose of the conversation” (Erman, 1995: 144). 3.4. Independent if-clauses This type is represented by a fixed set of routinized expressions that formally “look like” if-clauses but are neither subordinated to another clause, nor necessarily combined with another clausal unit. They are conventionalized expressions without syntactic dependency. The most common types in the ICE-GB are illustrated in (13a–c). (13)

a.

83

B:

84 85 86 87 88 89 90

A:

A:

b.

156

B:

c.

345

B:

B:

I’ve got the other- (.) the other one to that one Vicky and Tom Pitts has got his profile. oh have you? that’s it. that’s the one. yeah. (..) profile here. very short skirt on  if you don’t mind me saying . ((laughing)) with all my hair on one side of my head. [ICE-GB S1A-040] uhm (0.5) so I wouldn’t say that I actually looked on religion as a BAD thing Yif you see what I mean. [ICE-GB S1A-076] I told you I only read- (.) text books

A. Haselow / Language Sciences 54 (2016) 77–101

346 347

that’s all. (0.5) uh reference books if you li:ke.

93

[ICE-GB S1A-016]

Unlike “genuine” conditional clauses, if-clauses in the final field show no signs of syntactic embedding: they are not integrated hypotactically into a complex sentence, but loosely appended to a potentially complete unit (a lexical, phrasal or sentential unit). Semantically, these if-clauses do not express a conditional protasis that needs to be accompanied by an apodosis. Thus, unlike most “genuine” conditionals, they do not set up a mental space (a condition) within which the state of affairs expressed in the another clause is valid (Dancygier and Sweetser, 1997), i.e. the if-clause does no longer evoke a hypothetical frame in the strictest sense, but expresses metaconditional meanings that require a speech-act related interpretation, relating to different aspects of an utterance. Some of them refer to the adequacy of expression (if you like, if you will), others serve as a rhetorical device to mitigate the illoctionary force of speech acts that represent a potential face-threat to the addressee or the speaker him/herself (if you don’t mind me saying), or facilitate listener involvement (if you see what I mean). The if-clauses represent fixed, unbound (free) constructions. Final if you like, for instance, is to be analyzed as a chunk since its meaning cannot be broken down into separate semantic, syntactic and pragmatic modules: it does not express the literal meaning of a condition of the kind ‘in case you find sth. agreeable/enjoyable, q’ and does not require embedding into a main clause, but forms a “symbolic unit” in the sense of Langacker (1987), integrating various conventionalized aspects relating to its use and function. Moreover, “if” is no longer analyzable as marker of subordination and conditionality as it has extended its scope from syntax to connected discourse and from conditionality to pragmatic felicity. The conditional element surfaces only indirectly in that the felicity of the speech act is, at least rhetorically, expressed as depending on a pragmatic condition expressed in the if-clause, e.g. the condition that ‘you don’t mind’, which is the addressee’s consent. It thus expresses addressee-orientation and serves the involvement of the addressee in the dialogic activity (see also Brinton, 2008: 163–166). Prosodically, final if-clauses may represent an intonation contour of their own and be separated from the host unit by a pause, but typically they combine with the latter into one coherent intonation contour. In any case, such units are produced without pitch movement, that is, the pitch remains on the same low level it has reached with the last syllable of the preceding unit, and they are prosodically unmarked. 3.5. Linking adverbials/final particles Linking adverbials (Biber et al., 1999: 891) or final particles (FPs, Haselow, 2013) are non-inflecting, monomorphemic units that are prosodically integrated into a host unit, but have no constituent status (they are not integrated into microgrammatical dependency relations), and are consistently produced at the end of a pragmatically and intonationally completed unit of talk (Haselow, 2013: 389–390). Examples from English are final actually, then, or though, all of which lack conceptual meaning, do not contribute to the propositional content of a unit of talk, and have not effect on its truth value. They are produced with low-key intonation and prosodically unmarked. The most important class-defining functional feature of FPs is their relational function: they link the unit of talk they accompany to an aspect of the preceding discourse unit and mark it as a non-initial contribution to ongoing talk. Their categorical function is thus to signal that the unit of talk is a reactive turn within a dialogic sequence and thus motivated by prior talk, such as the content (or the speaker’s interpretation of the content) of a preceding turn or an implicature, serving the contextual fit of an utterance in ongoing discourse. FPs have a retrospective orientation, the pointing process being strictly backwards. For instance, as an FP then marks the unit it accompanies as an inference drawn from a preceding utterance or discourse segment, as illustrated in (14). (14)

8 10 11

B: A: B:

my grandmother was really uhm (.) helpful. you told her all about it then. yeah I di:d

[ICE-GB S1A-049]

The information provided by speaker B leads speaker A to infer that the state of affairs expressed in the unit of talk (“you told her all about it”) is likely to be true. A thus rank-shifts B’s utterance pragmatically into a conditional protasis whose truthvalue is, however, not hypothetical, but taken as a premise since B said it (‘if, as you say, p, q then?). Moreover, with declaratives final then modifies the illocutionary force of the unit to which it is attached from an assertion to a request for confirmation. It is important to note that the relational function is not restricted to the textual domain or the propositional level, but includes aspects of the communicative situation in the widest sense, for instance in terms of the speaker’s knowledge state and recipient design. Final then, for instance, typically marks an utterance as diverging from the speaker’s expectations, as in (14), final actually indicates potential divergence from the listener’s expectations. In this sense, FPs establish a link between two subsequent textual units, but are also deployed to structure dialogic interaction in a general way, encompassing non-propositional, metacommunicative aspects of the preceding discourse unit. They provide information on how to integrate the unit of talk they accompany into “a developing mental model of the discourse” (Hansen [2006: 25] on discourse markers). Syntactically, FPs have been analyzed in a variety of ways, e.g. as conjuncts, adverbial connectors, linking adverbials, or connective adjuncts. Classifying these elements as adverb-like is, however, problematic: they create a two-place relation in discourse rather than merely modifying an individual unit of talk, and thus have no “classical” constituent status. As macrogrammatical elements they structure language not within, but beyond a single structural unit. FPs form a paradigm of

94

A. Haselow / Language Sciences 54 (2016) 77–101

elements with the same core function (retrospective-relational), but expressing individual relational meanings. While the FP then indicates an inferential relationship, for instance, final though indicates dissonance between the content of two subsequent units of talk, while also expressing a weak acknowledgment of the plausibility or the validity of the prior utterance. This function can be observed in (15), where speaker B acknowledges the propositional core of his prior assertion, but corrects a possible inference that could be drawn from it, namely that the photographs he took are “any good”. (15)

1 2 3

A: B:

oh you know when you were in France did you uh take any good photographs. (..) I took lots of photographs. I don’t know if they’re any good though. [ICE-GB S1A-009]

FPs express the speaker’s perspective on the relation between an aspect of a prior unit of talk (its content or implied meanings) and a current one, i.e. the speaker functions as the origo in which a specific value, such as dissonance (final though) or inference (final then), is localized. FPs thus have a relational structure that qualifies them as macrogrammatical elements with two different pointing fields, namely connectivity (between discourse units) and subjectivity (pointing to the speaker’s perspective). 3.6. Tag questions Quirk et al. (1985: 810) define tag questions as a type of yes-no question that conveys a positive or negative orientation and that is appended to a sentence expressing a statement. If the statement is positive, the tag is generally negative (16a), and vice versa (16b).11 (16)

a. b.

372 373 305 306

A: A: B:

and he thinks he’s SO: cool Y doesn’t he . (.) oh he’s a WINDOW CLEANER [ICE-GB S1A-041] you haven’t seen her¼have you. no- no- no- I have not made the effort [ICE-GB S1A-098]

Formally, tag questions consist of an operator and a subject (in that order), the operator usually being the same as the operator of the preceding sentence or, if the latter has no operator, the auxiliary do, as in (16a). The subject of the tag is usually a pronoun that is coreferent with the subject in the preceding sentence and agreeing with it in person, number and gender. Note that many varieties of English have invariant tags, e.g. invariant isn’t it or innit. The syntactic status of tags is unclear: Quirk et al. (1985) do not use a specific terminology, but treat tags as units that are “appended” to a sentence expressing a statement and that are in a paratactic relation to this sentence (Quirk et al., 1985: 919) which, according to the authors’ definition of ‘parataxis’, means that they are constituents at the same level of constituent structure. As “appendices”, they are “outside” the clause or sentence, but unable to occur on their own, that is, without a syntactic “host”. Tags are functionally variant: they can express a genuine information- or confirmation-seeking request eliciting a response by the addressee, but they may also serve other interactional functions. As genuine requests for information/confirmation, tags are used to invite the addressee’s response to the statement to which they are appended. In such uses, they are typically produced with rising intonation if the addressee is about to decide the truth of the proposition (genuine question). The statement to which the tag is attached thus acquires the character of an assumption (for which the speaker may nevertheless have strong evidence). In (16b), for instance, the speaker elicits confirmation of a statement to which he is not fully committed and attributes epistemic power to the addressee. However, tag questions are not always used to answer a question and thus to elicit a response, but often merely imply the addressee’s consent, i.e. they mark a statement as already agreeing with the addressee’s view, based on assumed shared knowledge. In such “facilitative uses” (Holmes, 1983; Tottie and Hoffmann, 2009: 146), the speaker is committed to the truth of the proposition and the tag merely implies a particular expectation on the side of the speaker, namely that the addressee will agree, as in (16a), where a response by addressee is not required for negotiating the validity of a proposition. Tags in this function facilitate interaction in that they offer the addressee to engage in ongoing interaction by eliciting a minimal affirmative response (e.g. hmm, yeah), i.e. they invite the addressee to contribute to the discourse. Tags are often “attitudinal” (Tottie and Hoffmann, 2009: 145) as they indicate the speaker’s attitude such as disapproval, solidarity, and they may be pragmatically challenging, softening, hortatory, or emphatic (e.g. I’ve told you not to touch it, haven’t I?). Some tags are ambiguous between a genuine question, a mere invitation for confirmation, and a rhetorical device with the same speaker continuing his/her turn. Tags thus align differently on a scale from representing a TRP in a maximal way (genuine question, rising tone, pause) and a mere form of address inviting for a minimal contribution to ongoing talk, where the opportunity for the addressee to respond is restricted to minimal tokens (e.g. a head-nod or continuers like “mhm”) overlapping with the current speaker’s ongoing turn.

11 There are instances where both the statement and the tag are positive (e.g. so you have a car have you?), which occur in contexts where speakers arrive at a conclusion by inference or recall what has been said (Quirk et al., 1985: 812), often with a sarcastic undertone (you are having fun have you?).

A. Haselow / Language Sciences 54 (2016) 77–101

95

3.7. Vocatives (address terms) The category of vocatives or ‘address terms’ encompasses proper names, that is, the name of the addressee, but also forms of endearment (e.g. dear, darling, honey) and other conventionalized forms (e.g. mate), as in (17). The use of the last mentioned type is less common in English than in other languages, e.g. Spanish (e.g. tío, tía, hombre, see Stenström and Jørgensen, 2008: 4) or German (e.g. Alter, Digger), and more typical for adolescent speakers than for speakers of other age groups. (17)

280

B:

284

A:

he was standing outside in front of the National Gallery going-isn’t it beau:tiful we just don’t have anything like this in America. (.) [.] because you haven’t got a history mate  you should’ve said to him . [ICE-GB S1A-006]

The different kinds of address forms have no morphosyntactic dependency relation to the unit they accompany, but the status of isolated noun phrases in traditional syntactic analysis. From a macrogrammatical perspective, they are analyzed as integrated into and contributing to the illocutionary force of an utterance, situating a unit of talk into the concrete interpersonal context: they express an attitude toward the addressee (e.g. positive affection with endearments) and, particularly with conventionalized forms such as mate, control or maintain the contact to the addressee (Briz, 2001: 225), monitoring the attention of the addressee. On the illocutionary level, they often indicate emphasis, underlining the (alleged) importance of a unit of talk. Proper names are particularly common with illocutionary types appealing the attention of the addressee, above all directives (e.g. warning, advice), in which they express insistence or strengthen the expectation of a response to a directive, e.g. to an information-seeking request or a request (e.g. Can you help him, Bill?). 3.8. Interim conclusion Units in the final field follow a potential point of completion in the sense that they are produced when the unit they accompany leaves no projections on any relevant level open, expanding a unit of talk under production in real time. The different unit types in the final field exhibit different degrees of syntactic complexity, ranging from single lexemes (e.g. FPs, vocatives) to clause-like units (e.g. if-clauses). All of them serve as retrospective framing devices that indicate how the unit of talk they accompany is to be interpreted and how this unit is linked to various aspects of the communicative situation, such as the turn-taking system or orientation toward the knowledge state or attention of the addressee. Units in the final field can occur in combination. A tag question can, for instance, be followed by a general extender, as in (18a), or an FP by another FP, as in (18b). (18)

a.

B:

b.

A: 175 176

[well he has this stupid girl he falls in love with Y doesn’t he or something . NO:: [ICE-GB S1A-006, 112] B: we should¼maybe just leave a message here saying head over. (.) A: she won’t bother coming then though. [ICE-GB S1A-039]

The combinability of final field elements is not unconstrained: some combinations seem to be excluded as they do not occur in the ICE-GB and intuitively appear unnatural, such as an independent if-clause followed or preceded by a question tag. Moreover, the serialization of units in the final field is often constrained: an FP can, for instance, be followed by a comment clause, but is usually not preceded by it (e.g. *He’s not invited I think then). This indicates that the final field has a syntax of its own that requires further investigation. The functions of the different unit types in the final field are very similar, which suggests that speakers deal with recurrent cognitive tasks at the potential end of a unit of talk in the linear flow of time. These tasks are all related to the transfer of an idea that is first a subjective message to the intersubjective level: they display addressee-orientation in that they are used to fine-tune epistemic or pragmatic aspects of a message in order to facilitate utterance interpretation, and to establish contact to the addressee in a variety of ways, e.g. facilitating interaction by prompting the addressee to engage in ongoing interaction. They provide a last procedural cue guiding the addressee’s interpretation of a message before (potential) turn transition. The functions can be summarized as follows: (i) fine-tuning a message in terms of epistemic value (e.g. comment clauses, adverbs), illocutionary force (e.g. FPs, adverbs) or adequacy of expression, (ii) integrating a message into the developing mental model of discourse (marking its communicative function within ongoing discourse, relative to prior discourse, or the rhetorical relationship between two neighboring units of discourse), (iii) intersubjective understanding: referring to the co-participant’s knowledge state (e.g. displaying shared knowledge), (iv) organizing turn-taking (preparing continuation, marking a potential TRP, preparing a next activity, requesting or providing an opportunity for listener response, facilitating turn transition, checking listener’s attention),

96

A. Haselow / Language Sciences 54 (2016) 77–101

(v) serving interpersonal rapport (keeping the communicative channel open). Some of these tasks are more relevant than others in an individual local discourse context, depending on how many of these tasks have already been attended to in the utterance-so-far and on whether the context already provides sufficient interpretive cues that allow for successful (intended) interpretation of the respective unit of talk. Turn-taking functions, for instance, can be fulfilled with the syntactic design of a unit of talk (e.g. by means of an interrogative format); retrospective fine-tuning of the epistemic value of a message is not required when the preceding structural unit already includes indicators of epistemic stance (e.g. modality). The final field is thus a resource that is used when the interactive past has not already contributed to a solution of the tasks listed above or when the speaker considers the conditions required for successful processing as not fulfilled. Elements in the final field enrich a message by expressing fine-grained metatextual meanings. The functions reflect the specific set of tasks that speakers typically need to deal with at the end of a unit of talk emerging in real time: fine-tuning a message, seeking a response or preparing next turns are tasks which, in the default case, speakers can attend to only after a message has been produced and is manifest to the participants. For instance, speakers can often judge the adequacy (e.g. in terms of semantic precision, epistemic value, pragmatic effect) of expressing their thoughts only after they have been produced in unprepared speech, where possible unwanted effects may occur on different dimensions. The final field provides a room for “communicative manoeuvre” for the speaker who is always situated at the leading end of an unfolding utterance, where new ideas and effects can only be integrated by a process of incrementation, that is, by adding new “bits and pieces” to an utterance (Hopper, 2011: 23). From this dynamic, temporal perspective on utterance production the final field is the last opportunity to make whatever adjustments are necessary according to the speaker’s intuition or contextual demands before possible turn transition or continuation with a new idea.

4. The final field in a grammar of speech: micro- and macrogrammar The final field is problematic for grammatical theories because neither its occurrence, nor the kind of element chosen by the speaker are in any way predictable in a context-free grammar. This leads to the dilemma that while the final field is an interesting phenomenon with a high degree of recurrence in speech data, it is hardly or only schematically describable in a context-free grammar. The interest in the final field for linguistic analysis lies in the fact that it is an example for linguistic patterns that are the result of the processual character of language and ad-hoc decisions of the speaker involved in real-time speech production. A structuralist analysis of the units in the final field cannot adequately deal with this kind of “unintegrated” syntactic material as it misses the fact that the final field emerges during the step-by-step creation of structure rather than being based on a priori syntactic planning. A description of the structural principles of speech therefore needs to rest upon a dynamic view of speech production and must be based on the main condition that determines the cognitive processes of the conversationalists involved, which is the linearity of speaking in time. Distinguishing between microgrammar and macrogrammar offers a wider perspective on syntactic structure by postulating two different serialization principles in on-line speech production, which are either based on morphosyntactic dependency relations, hierarchization, constituency, and embeddedness (microgrammar), or on speech planning and different cognitive tasks that are related to the communication process in the widest sense and that become relevant at a particular time in utterance production (macrogrammar). Both alternate in speech production, the details of which need to be explored further by studying typical serialization patterns involving macrogrammatical elements. As has been shown above, macrogrammatical units in the final field serve a specific set of tasks that have not been attended to during utterance production, or which cannot be attended to as they are not relevant at other points in the linear emergence of structure. Their functions refer to the macrostructure of discourse as they provide cues for listener to integrate parts into a coherent whole, or to speaker–addressee interaction, that is, the social context of communication. These functions correlate with those discussed in neurolinguistic research discussed in Section 2, that is, there seems to be a correspondence between the observations made in neurolinguistics and those reflected in the macrogrammatical component of language. Since macrogrammatical elements in the final field express functions that are associated with RH activity, they can be assumed to be processed differently from microgrammatical elements. The idea that some aspects of grammar activate participation of the RH has also been developed for the domain of “situation of discourse” in Heine et al. (2015), who discuss phenomena of “thetical grammar” that correspond with RH activity. Since macrogrammatical elements are not describable on the basis of a priori patterns of concatenation, there is only one way in which they can be approached in a context-free description of linearity, namely by means of a probabilistic account of the point in time in the production of a unit of talk at which their production is likely and thus considered acceptable by the co-participants. Thus, temporal (or serial) regularities will determine distributional probabilities. From a usage-based perspective, the unit types listed in Table 1 and discussed in this paper are prototypes of a probabilistic syntax whose distribution is relatively consistent: they occur at the end of a unit that leaves no projections on relevant levels open. The difference to microgrammatical elements is that their occurrence in the linear flow of time is not determined by dependency relations, where e.g. the occurrence of a lower-ranked element makes the prediction possible that the higher-ranked element will follow, but by cognitive constraints related to speech production in real time: the functions they serve are highly relevant

A. Haselow / Language Sciences 54 (2016) 77–101

97

at the point in time at which the content of the unit they are attached to is manifest to the participants and the unit is potentially complete. Determining the distributional potential of a macrogrammatical unit requires a kind of grammar writing that is based on how language is produced step-wise in time. In order to understand temporal probabilities it is necessary to understand the function of the respective unit, which in turn requires an understanding of the temporal sequence of different kinds of cognitive tasks or requirements in producing a unit of talk in a spontaneous way. It seems that the main factor determining the distribution of macrogrammatical units is the task they serve: many of these units occur in particular temporal slots or “fields” because this is the moment at which the task they serve becomes relevant in the real-time creation of a structural unit. Under this “emergentist view” in the sense of Hopper (1987, 1998), language structure is not a fixed configuration of a priori categories and fixed slots, but “subject to the exigencies of communication” (Hopper, 1998: 157). The final field as an optional communicative space shows that structure is a continual movement along a vaguely defined trajectory, and that the end of a unit of talk is always provisional and never determined or determinable. Speakers create structure as they go along in the moment-by-moment configuration of talk whereas listeners, for whom the final field is not projectable, need to process structure as “‘mobile observers’, travelling along with the stream of speech and on the spot producing hypotheses of understanding which change and vary with the point the utterance has reached” (Franck, 1985: 238). 5. The final field vs. other temporal slots The definition of the functions fulfilled by elements in the final field does not answer the question if the same functions could be fulfilled by using the respective elements in other time slots, e.g. before the projections built up in a unit of talk have been completed. I argue that the relation between the final field and the production of the set of elements discussed above is relatively strong since there are restrictions with respect to the use of elements in the final field in other temporal slots in the linear emergence of a structural unit: many of these elements are typically not produced in the initial or medial phase of the production of a unit of talk, that is, before projections are completed since the tasks they serve are not relevant in the process of building up projections. This concerns, for instance, tag questions or independent if-clauses, which facilitate turn-transition and fine-tune the pragmatic import of a message. There seem to be cases where these and other final-field elements appear to occur utterance-medially from an analytic perspective, that is, when considering the finished speech product. However, a closer look at such cases shows that they tend to occur at a point in the linear emergence of structure at which the speaker has come to an intermediate end of a projection, using this moment as a mental resting point for speech planning only to continue with a part that is expected to complete the projection, or to allow for turn transition. The first case is illustrated in (19). (19)

a.

B:

b.

A:

B:

when we had the talk on clinical psychology and stuff (.) you know you need a lot of experience not just with people with problems and stuff [ICE-GB S1A-035, 047] I mean in in that piece we’ve just heard from The Revenger’s tragedy it’s a mixture isn’t it of original instruments.j (.) and kind of- what sound to me like modern [ (.) ] trumpets. [ye:s ] [ICE-GB S1B-023, 140]

In (19a), the speaker projects a complex turn which consists of a temporal adverbial clause introduced by the subordinator when which, upon its production, raises expectations on a bi-clausal structure expressing a temporal relation between cooccurring or successive events. The general extender and stuff in (19a) is produced after the first unit can be heard as complete, that is, after the syntactic projections of the first clausal unit have been completed. At this point the speaker begins to hesitate, adding the addressee-oriented marker you know, before he continues. However, the continuation is semantically odd as it does not complete the cognitive projection initiated by when in the sense that it matches the a priori semantic template offered by syntactic when.(then). structures. Rather, the speaker is creating a continuation on the fly, improvising as the unit of talk unfolds. A similar strategy can be observed in (19b), where A adds the question tag isn’t it to an intermediate completion point of a unit-so-far expressing a core idea (“it’s a mixture”), alluding to the addressee’s agreement. The idea is specified by a postmodifying prepositional phrase after this potential end point, based on the speaker’s impression that the nominal reference “mixture” might be semantically too unspecific to be identifiable for the listener. The ad-hoc character of continuations of a unit of talk-so-far is also reflected in the addition of increments to a potentially completed unit, which may follow units in the final field, as in (20). These increments are, like units in the final field, nonprojected units that merely expand a structural unit beyond a point of completion (Couper-Kuhlen and Ono, 2007). (20)

B:

so she’s got about four thousand books now I think. j (.) or three thousand [ICE-GB S1A-025, 319]

98

A. Haselow / Language Sciences 54 (2016) 77–101

Continuations beyond final-field elements are thus based on the speaker’s ad-hoc decision to expand a unit of talk, which is motivated e.g. by post-hoc ideas, associations or repairs that need to be integrated into an ongoing unit before TRPs, or by a lack of turn uptake and responsive activities by the addressee, which requires the current speaker to continue. Thus, the final field is a time slot occurring at a particular moment in speech production at which the speaker feels to have reached the end of a projection, but sometimes continues recompleting a unit of talk with unit types whose production was not fully planned when the utterance had been initiated. For the analyst working with the finished speech product, cases like (20) appear as utterance-medial uses of elements that otherwise occur in the final field. This interpretation, however, misses an important point: speakers are free to reanalyze their utterances in the moment-by-moment unfolding of structure since “the leading edge of an utterance is always open to renegotiation through increments which [.] may work to recreate an original grammatical project and (seen retrospectively) transform it into a new construction.” (Hopper, 2011: 32). It is at such crucial points in the on-line emergence of structure – intermediate between completion of a (sub)unit and different alternatives as to how to continue – that elements representing the final field are produced. A further aspect providing evidence for the assumption that the final field serves a specific set of functions is that those units that can be produced at earlier points in the production of a unit of talk are not used with the same function in the final field. Lexical units that appear flexible, such as final particles or adverbs, differ semantically and functionally from the use in the final field when they are produced at an earlier point in time. Moving an FP like actually, for instance, to a position within a structural unit leads to an alternative semantic interpretation since the scope of the particle is narrowed to the subsequent phrase (he did not actually invite her; he actually invited his parents.). As an FP, however, actually is typically interpreted as having scope over the entire unit of talk it accompanies. Moreover, it expresses mirativity, that is, the speaker’s sudden realization in on-line speech production that a state of affairs violates an expectation, a function that is absent with non-final uses. Changes in scope and function also occur with adverbs, particularly evaluative (e.g. of course) and epistemic adverbs (e.g. perhaps), which have broad scope over the entire utterance in the final field and are used to indicate shared knowledge (of course) and epistemic stance (perhaps) in a post-hoc way. Within a structural unit, they have narrow scope over a particular constituent and are less likely to be interpreted as a deferred comment on an aspect of the unit of talk just produced. Those units that could also be produced within a unit of talk tend to have a higher communicative weight in the final field as they are separated from the core message through serialization, which underlines their metadiscursive function: they do not contribute to the emergence of a contentful unit, being part of an emerging syntactic gestalt, but follow a contentful unit in time and thus acquire a framing function guiding the interpretation process of the listener or linking it to aspects of speaker–addressee interaction. Rather than being embedded in a message, in the final field they embed the message they accompany in a particular discursive or interpersonal context. 6. Conclusions This paper has provided an analysis of an important time slot in the linear production of a unit of talk, the so-called “final field”, which hosts elements that are not morphosyntactically integrated into a prior structural (usually clausal) unit, but loosely appended to it. Since elements in the final field cannot be analyzed in terms of constituency and embeddedness, their analysis requires a grammatical approach that is based on the linearity of speech and its incremental character. According to this view, the final field is a time slot that emerges during the step-by-step creation of structure in real time, rather than a fixed “position” available to a priori syntactic planning. The units available for use in this field cannot occur on their own as they are integrated into the illocutionary force of a unit of talk and provide a processing cue. Based on the function of unit types in the final field it has been shown that this field plays an important role in spontaneous speech as it allows speakers to fine-tune various aspects of a unit of talk and to establish contact to the listener in different ways. Even though the use of the final field is typical for spoken genres, it should not be concluded that this is a recent syntactic development that has just begun to diffuse into different text types. Many of the current developments have their traces in earlier periods of language development, as e.g. Traugott’s (in press) recent study on the increasing use of “right-peripheral” elements in the history of English shows. Thus, while the existence of the final field itself is not a recent phenomenon (given that the cognitive tasks in the final phase of utterance production are certainly stable over different periods of language development), linguists have just begun to direct their attention to this long neglected part of the structural design of a unit of talk and to understand its functional properties. The main reason is certainly to be found in the ‘written-language bias’ that has long characterized linguistic analysis (Linell, 2005), where preference has been given to well-formed, integrated syntactic configurations rather than to natural speech products deriving from conversational interaction. Existing studies on different kinds of elements in the final field suggest that the final field is a system in flux as speakers continuously update the set of elements available for use (again, see Traugott [in press] for a historical overview). This is not unique to English, where the rise of some unit types such as FPs has been incremental over time, probably triggered by the sequential character of dialogic interaction (Haselow, 2014), including relatively recent developments such as final uses of though (Barth-Weingarten and Couper-Kuhlen, 2002; Lenker, 2010: 200), even (Kim and Jahnke, 2011) or but (Mulder and Thompson, 2008; Hancil, 2014), and the rapid changes in the morphosyntactic and functional characteristics of general extenders in many varieties of English (Cheshire, 2007; Tagliamonte and Denis, 2010; Pichler and Levey, 2011). Many of the contributions to a recent volume on “final particles” (Hancil et al., 2015), for instance, document the emergence of utterancefinal uses of lexemes deriving from microgrammatical (clause-internal) units in various languages of the world (e.g. Dutch, German, Japanese, Korean, Russian), and also some contributions in Beeching and Detges’ (2014) edited volume deal explicitly

A. Haselow / Language Sciences 54 (2016) 77–101

99

with RP elements in languages other than English. The total range of lexemes available in final position in different languages, their combinability, and the constraints on their order remain to be investigated. The study on the final field has shown that we need to take a new look at elements that have been categorized as “macrogrammatical” here and try to determine the point at which they are likely to be produced in the moment-by-moment emergence of structure, based on the cognitive tasks they serve and the time at which these tasks become relevant in the linear construction of a unit of talk. The task is then to correlate different time slots, e.g. the initial and the final one, with sets of elements that are produced preferrably in these slots. The result of such a procedure – a description of macrogrammar – is strictly deductive, empirically-based, and can never be deterministic and based on a priori rules. Macrogrammar is thus a probabilistic, cognitive grammar describing the use of macrogrammatical units not in terms of (quasi-)obligatoriness and a priori patterns, but in terms of sequentially organized cognitive tasks accompanying the creation of linguistic structure in the linear flow of time. Acknowledgments I would like to express my gratitude to an anonymous reviewer as well as to Elizabeth Traugott, Bernd Heine, Gunther Kaltenböck, Lachlan Mackenzie, and Graeme Trousdale, who have been of help in writing this paper. Transcription conventions

[] ¼ (.) (2.0) (( )) :, :: rea(hh)lly ABsolutely really [ Y j  word . , ? ¿ >words<

overlap and simultaneous talk latching micropause measured pause transcriber’s descriptions segmental lengthening according to duration laugh particles within talk strong, primary stress via loudness stress via pitch or amplitude shift to especially high pitch fall in pitch prosodic cesura produced softer than surrounding talk falling intonation (terminal pitch) continuing intonation rising intonation a rise stronger than mid-level but weaker than high-terminal pitch talk is compressed or rushed

References Aijmer, K., 1986. Why is actually so popular in spoken English? In: Tottie, G., Bäcklund, I. (Eds.), English in Speech and Writing: a Symposium. Almqvist & Wiksell, Stockholm, pp. 119–127. Aijmer, K., 2002. English Discourse Particles. Evidence from a Corpus. John Benjamins, Amsterdam. Auer, P., 1996a. The pre-front field in spoken German and its relevance as a grammaticalization position. Pragmatics 6 (3), 295–322. Auer, P., 1996b. On the prosody and syntax of turn-continuations. In: Couper-Kuhlen, E., Selting, M. (Eds.), Prosody in Conversation. Interactional Studies. Cambridge University Press, Cambridge, pp. 57–100. Auer, P., 2005. Projection in interaction and projection in grammar. Text 25 (1), 7–36. Auer, P., 2007. Why are increments such elusive objects? An afterthought. Pragmatics 17 (4), 647–658 (Special Issue: Turn continuation in cross-linguistic perspective, ed. by Couper-Kuhlen, E., Ono, T.). Auer, P., 2009. On-line syntax: thoughts on the temporality of spoken language. Lang. Sci. 31, 1–13. Auer, P., 2015. The temporality of language in interaction: projection and latency. In: Deppermann, A., Günthner, S. (Eds.), Temporality in Interaction. John Benjamins, Amsterdam, pp. 27–56. Auer, P., Pfänder, S. (Eds.), 2011. Constructions: Emerging and Emergent. De Gruyter Mouton, Berlin. Avanzi, M., 2007. Regards croisés sur la notion de macro-syntaxe. Trav. Neuchâtelois Linguist. 49, 39–58. Barth-Weingarten, D., Couper-Kuhlen, E., 2002. On the development of final though: a case of grammaticalization? In: Wischer, I., Diewald, G. (Eds.), New Reflections on Grammaticalization. John Benjamins, Amsterdam, pp. 345–361. Beeching, K., Detges, U., 2014. Introduction. In: Beeching, K., Detges, U. (Eds.), Discourse Functions at the Right and Left Periphery: Crosslinguistic Investigations of Language Use and Language Change. Brill, Leiden, pp. 1–23. Berendonner, A., 1990. Pour une macro-syntaxe. Trav. Linguist. 21, 25–36. Biber, D., Johansson, S., Leech, G., Conrad, S., Finegan, E., 1999. Longman Grammar of Spoken and Written English. Pearson Education, Harlow. Blakemore, D., 1987. Semantic Constraints on Relevance. Blackwell, Oxford. Blanche-Benveniste, C., 2003. Le recouvrement de la syntaxe et de la macro-syntaxe. In: Scarano, A. (Ed.), Macro-syntaxe et pragmatique. L’analyse linguistique de l’oral. Bulzoni, Firenze, pp. 53–75.

100

A. Haselow / Language Sciences 54 (2016) 77–101

Bottini, G., Corcoran, R., Sterzi, R., Paulesu, E., Schenone, P., Scarpa, P., Frackowiak, R., Frith, C., 1994. The role of the RH in the interpretation of figurative aspects of language. Brain 117, 1241–1253. Brady, M., Armstrong, L., Mackenzie, C., 2006. An examination over time of language and discourse production abilities following right hemisphere brain damage. J. Neurolinguist. 19 (4), 291–310. Brinton, L.J., 2008. The Comment Clause in English: Syntactic Origins and Pragmatic Development. Cambridge University Press, Cambridge. Briz, A., 2001. El español coloquial en la conversación. Esbozo de pragmagramática. Ariel, Barcelona. Brownell, H.H., Potter, H.H., Bihrle, A.M., Gardner, H., 1986. Inference deficits in right brain-damaged patients. Brain Lang. 27 (2), 310–321. Bybee, J., 2006. From usage to grammar: the mind’s response to repetition. Language 82 (4), 711–733. Carroll, R., 2008. Historical English phraseology and the extender tag. Selim 15, 7–37. Chafe, W., 1992. Intonation units and prominences in English natural discourse. In: Proceedings of the IRCS Workshop on Prosody in Natural Speech, Institute for Research in Cognitive Science, Report No. 92-37. University of Pennsylvania, Philadelphia. Chafe, W., 1994. Discourse, Consciousness, and Time. The Flow and Displacement of Conscious Experience in Speaking and Writing. University of Chicago Press, Chicago. Cheshire, J., 2007. Discourse variation, grammaticalisation and stuff like that. J. Socioling. 11 (2), 155–193. Code, C., 1996. Speech from the isolated right hemisphere? Left hemispherectomy cases E.G. and N.F. In: Code, C., Wallesch, C.-W., Joanette, Y., Lecours, A.R. (Eds.), Classic Cases in Neuropsychology. Lawrence Erlbaum, Mahwah, NJ, pp. 319–336. Couper-Kuhlen, E., Ono, T., 2007. Increments in cross-linguistic perspective: introductory remarks (Special issue). Pragmatics 17 (4), 505–512. Croft, W., 2001. Radical Construction Grammar: Syntactic Theory in Typological Perspective. Oxford University Press, Oxford. Cutica, I., Bucciarelli, M., Bara, B.G., 2006. Neuropragmatics: extralinguistic pragmatic ability is better preserved in left-hemisphere-damaged patients than in right hemisphere-damaged patients. Brain Lang. 98 (1), 12–25. Dancygier, B., Sweetser, E., 1997. ‘Then’ in conditional constructions. Cogn. Linguist. 8 (2), 109–136. Dehé, Nicole, 2009. Clausal parentheticals, intonational phrasing, and prosodic theory. J. Linguist. 45 (3), 569–615. Dehé, N., The prosodic phrasing of parenthetical comment clauses in spontaneous spoken language: evidence from Icelandic held ég. Stud. Linguist.(in press). http://www.academia.edu/11087933/The_prosodic_phrasing_of_parenthetical_comment_clauses_in_spontaneous_spoken_language_Evidence_ from_Icelandic_held_%C3%A9g_to_appear_. Dehé, N., Wichmann, A., 2010a. The multifunctionality of epistemic parentheticals in discourse: prosodic cues to the semantic-pragmatic boundary. Funct. Lang. 17 (1), 1–28. Dehé, N., Wichmann, A., 2010b. Sentence-initial I think (that) and I believe (that): prosodic evidence for use as main clause, comment clause and discourse marker. Stud. Lang. 34 (1), 36–74. Deppermann, A., 2013. Turn-design at turn-beginnings: multimodal resources to deal with tasks of turn-construction in German. J. Pragmat. 46 (1), 91–121. Deppermann, A., Günthner, S. (Eds.), 2015. Temporality in Interaction. John Benjamins, Amsterdam. Detges, U., Waltereit, R., 2014. ‘Moi je ne sais pas vs. Je ne sais pas moi’: French disjoint pronouns in the left vs. right periphery. In: Beeching, K., Detges, U. (Eds.), Discourse Functions at the Right and Left Periphery: Crosslinguistic Investigations of Language Use and Language Change. Brill, Leiden, pp. 24–46. Dik, S.C., 1997. The Theory of Functional Grammar. Part 2. Mouton de Gruyter, Berlin. Dines, E.R., 1980. Variation in discourse – ‘and stuff like that’. Lang. Soc. 9 (1), 13–31. Duden-Grammatik, 2009. Die Grammatik. Published by the Dudenredaktion. Dudenverlag. Zürich, Mannheim, Wien. Eisenberg, P., 2006. Grundriss der deutschen Grammatik. Vol. 2: Der Satz. Metzler, Stuttgart. Erman, B., 1995. Grammaticalization in progress: the case of ‘or something’. In: Moen, I., Simonsen, H.G., Lødrup, H. (Eds.), Papers from the Fifteenth Scandinavian Conference of Linguistics, Oslo 13–15 January 1995. Department of Linguistics, University of Oslo, Oslo, pp. 136–147. Ford, C., Thompson, S., 1996. Interactional units in conversation: syntactic, intonational, and pragmatic resources for the management of turns. In: Ochs, E., Schegloff, E.A., Thompson, S.A. (Eds.), Interaction and Grammar. Cambridge University Press, Cambridge, pp. 134–184. Franck, D., 1985. Sentences in conversational turns: a case of syntactic ‘double blind’. In: Dascal, M. (Ed.), Dialogue. An Interdisciplinary Approach. John Benjamins, Amsterdam, pp. 233–245. Goldberg, A., 1995. Constructions: a Construction Grammar Approach to Argument Structure. University of Chicago Press, Chicago. Günthner, S., 2011. Between emergence and sedimentation: projecting constructions in German interaction. In: Auer, P., Pfänder, S. (Eds.), Constructions: Emerging and Emergent. De Gruyter Mouton, Berlin, pp. 156–185. Hale, J., 2006. Uncertainty about the rest of the sentence. Cogn. Sci. 30, 643–672. Hancil, S., 2014. The final particle ‘but’ in British English: an instance of cooptation and grammaticalization at work. In: Hancil, S., König, E. (Eds.), Grammaticalization – Theory and Data. John Benjamins, Amsterdam, pp. 235–255. Hancil, S., Haselow, A., Post, M. (Eds.), 2015. Final Particles. De Gruyter Mouton, Berlin. Hansen, M.-B.M., 2006. A dynamic polysemy approach to the lexical semantics of discourse markers (with an exemplary analysis of French toujours). In: Fischer, K. (Ed.), Approaches to Discourse Particles. Elsevier, Amsterdam, pp. 21–41. Haselow, A., 2013. Arguing for a wide conception of grammar: the case of final particles in spoken discourse. Folia Linguist. 47 (2), 375–424. Haselow, A., 2014. Sequentiality in dialogue as a trigger for grammaticalization. In: Hancil, S., König, E. (Eds.), Grammaticalization – Theory and Data. John Benjamins, Amsterdam, pp. 203–233. Heine, B., Kuteva, T., Kaltenböck, G., Long, H., 2015. On Some Correlations between Grammar and Brain Lateralization. Oxford Handbooks Online. Oxford University Press, Oxford. Hird, K., Kirsner, K., 2003. The effect of right cerebral hemisphere damage on collaborative planning in conversation: an analysis of intentional structure. Clin. Linguist. Phon. 17 (4–5), 309–315. Holmes, J., 1983. The functions of tag questions. Engl. Lang. Res. J. 3, 40–65. Holmes, J., 1988. Of course: a pragmatic particle in New Zealand woman’s and men’s speech. Aust. J. Linguist. 2, 49–74. Hopper, P.J., 1987. Emergent grammar. Berkeley Linguist. Soc. 13, 139–157. Hopper, P.J., 1998. Emergent grammar. In: Tomasello, M. (Ed.), The New Psychology of Language. Laurence Erlbaum Associates, Mahwah, NJ, pp. 155–175. Hopper, P.J., 2004. The openness of grammatical constructions. Chic. Linguist. Soc. 40, 239–256. Hopper, P.J., 2011. Emergent grammar and temporality in interactional linguistics. In: Auer, P., Pfänder, S. (Eds.), Constructions: Emerging and Emergent. De Gruyter Mouton, Berlin, pp. 22–44. Hough, M., 1990. Narrative comprehension in adults with right and left hemisphere brain damage: theme organization. Brain Lang. 38, 253–277. Huddleston, R., Pullum, G., 2002. Language Description: the Cambridge Grammar of the English Language. Cambridge University Press, Cambridge. Kaltenböck, G., Heine, B., Kuteva, T., 2011. On thetical grammar. Stud. Lang. 35, 852–897. Kim, M.-J., Jahnke, N., 2011. The meaning of utterance-final even. J. Engl. Linguist. 39, 36–64. Kjellmer, G., 2003. Hesitation. Engl. Stud. 84, 170–198. Langacker, R., 1987. Foundations of Cognitive Grammar Vol. 1. Stanford University Press, Stanford. Lenker, U., 2010. Argument and Rhetoric. Adverbial Connectors in the History of English. De Gruyter Mouton, Berlin. Lewis, D., 2003. Rhetorical motivations for the emergence of discourse particles, with special reference to English of course. In: van der Wouden, T., Foolen, A., van de Craen, P. (Eds.), Particles, Special issue of Belgian Journal of Linguistics, 16, pp. 79–91. Linell, P., 2005. The Written Language Bias in Linguistics. Routledge, London. Marini, A., 2012. Characteristics of narrative discourse processing after damage to the right hemisphere. Seminars Speech Lang. 33 (1), 68–78. Marini, A., Carlomagno, S., Caltagirone, C., Nocentini, U., 2005. The role played by the RH in the organization of complex textual structures. Brain Lang. 93, 46–54.

A. Haselow / Language Sciences 54 (2016) 77–101

101

Matthiessen, C., Thompson, S.A., 1988. The structure of discourse and ‘subordination’. In: Haiman, J., Thompson, S.A. (Eds.), Clause Combining in Grammar and Discourse. John Benjamins, Amsterdam, pp. 275–329. Miller, J., Weinert, R., 1998. Spontaneous Spoken Language: Syntax and Discourse. Oxford University Press, Oxford. Mulder, J., Thompson, S.A., 2008. The grammaticization of ‘but’ as a final particle in English conversation. In: Laury, R. (Ed.), Crosslinguistic Studies of Clause Combining: the Multifunctionality of Conjunctions. John Benjamins, Amsterdam, pp. 179–204. Myers, P.S., Blake, M.L., 2008. Communication disorders associated with right hemisphere damage. In: Chapey, R. (Ed.), Language Intervention Strategies in Aphasia and Related Neurogenic Communication Disorders, fifth ed. Lippincott Williams & Wilkins, Baltimore, pp. 963–987. Newmeyer, F.J., 2015. Parentheticals and the grammar of complementation. In: Schneider, S., Glikman, J., Avanzi, M. (Eds.). Parenthetical verbs. De Gruyter Mouton, Berlin, pp. 13–38. Nuyts, J., 2001. Epistemic Modality, Language and Conceptualization: a Cognitive-pragmatic Perspective. John Benjamins, Amsterdam. Overstreet, M., 1999. Whales, Candlelight, and Stuff like that. General Extenders in English Discourse. Oxford University Press, New York. Pekarek Doehler, S., 2011. Emergent grammar for all practical purposes: the on-line formatting of left and right dislocations in French conversation. In: Auer, P., Pfänder, S. (Eds.), Constructions: Emerging and Emergent. De Gruyter Mouton, Berlin, pp. 45–87. Pichler, H., Levey, S., 2011. In search of grammaticalization in synchronic dialect data: general extenders in Northeast England. Engl. Lang. Linguist. 15 (3), 441–471. Pierrehumbert, J., Hirschberg, J., 1990. The meaning of intonational contours in the interpretation of discourse. In: Cohen, P.R., Morgan, J., Pollack, M.E. (Eds.), Intentions in Communication. MIT Press, Cambridge, MA, pp. 271–311. Prat, C.S., Long, D.L., Baynes, K., 2007. The representation of discourse in the two hemispheres: an individual differences investigation. Brain Lang. 100 (3), 283–294. Quirk, R., Greenbaum, S., Leech, G., Svartvik, J., 1985. A Comprehensive Grammar of the English Language. Longman, London. Rinaldi, M.C., Marangolo, P., Baldassarri, F., 2004. Metaphor comprehension in right brain damaged patients with visuoverbal and verbal material: a dissociation (re)considered. Cortex 40, 479–490. Schegloff, E.A., 1996. Turn organization: one intersection of grammar and interaction. In: Ochs, E., Schegloff, E.A., Thompson, S.A. (Eds.), Interaction and Grammar. Cambridge University Press, Cambridge, pp. 52–133. Schegloff, E.A., Sacks, H., 1973. Opening up closings. Semiotica 8, 289–327. Schiffrin, D., 1987. Discourse Markers. Cambridge University Press, Cambridge. Schoroup, L., 1985. Common Discourse Particles in English Conversation. Garland, New York. Sherratt, S., Bryan, K., 2012. Discourse production after right brain damage: gaining a comprehensive picture using a multi-level processing model. J. Neurolinguist. 25, 213–239. Shields, J., 1991. Semantic-pragmatic disorder: a right hemisphere syndrome? Br. J. Disord. Commun. 26, 383–392. Simon-Vandenbergen, A.-M., Aijmer, K., 2002. The expectation marker ‘of course’. Lang. Contrast 4, 13–43. Stemmer, B., Giroux, F., Joanette, Y., 1994. Production and evaluation of requests by right hemisphere brain-damaged individuals. Brain Lang. 47, 1–31. Stenström, A.-B., Jørgensen, A.M., 2008. La función fática de los apelativos en la conversación juvenil de Madrid y Londres. In: Actas del III Congreso EDICE. Universidad de Valencia, pp. 1–14. Tagliamonte, S., Denis, D., 2010. The stuff of change: general extenders in Toronto, Canada. J. Engl. Linguist. 38 (4), 335–368. Thompson, S.A., 2002. “Object complements” and conversation: towards a realistic account. Stud. Lang. 26, 125–164. Thompson, S.A., Mulac, A., 1991. The discourse conditions for the use of the complementizer ‘that’ in conversational English. J. Pragmat. 15, 237–251. Tottie, G., Hoffmann, S., 2009. Tag questions in English – the first century. J. Engl. Linguist. 37, 130–161. Traugott, E.C., 2012. Intersubjectification and clause periphery. In: Brems, L., Ghesquière, L., Van de Velde, F. (Eds.), Intersections of Intersubjectivity, Special issue of English Text Construction, 5(1), pp. 7–28. Traugott, E.C., 2015. Investigating “periphery” from a functionalist perspective. Linguist. Vanguard 1 (online journal). Traugott, E.C., On the rise of types of clause-final pragmatic markers in English. To appear in: J. Hist. Pragmat. (in press). http://www.academia.edu/ 11087933/The_prosodic_phrasing_of_parenthetical_comment_clauses_in_spontaneous_spoken_language_Evidence_from_Icelandic_held_%C3%A9g_ to_appear_. Traugott, E.C., Trousdale, G., 2010. Gradience, gradualness and grammaticalization. How do they intersect? In: Traugott, E.C., Trousdale, G. (Eds.), Gradience, Gradualness and Grammaticalization. John Benjamins, Amsterdam, pp. 19–44. Van Lancker Sidtis, D., 2004. When novel sentences spoken or heard for the first time in the history of the universe are not enough: toward a dual-process model of language. Int. J. Lang. Commun. Disord. 39, 1–44. Van Lancker Sidtis, D., 2009. Formulaic and novel language in a “dual process” model of language competence: evidence from surveys, speech samples, and schemata. In: Corrigan, R., Moravcsik, E.A., Ouali, H., Wheatley, K.M. (Eds.), Formulaic Language, Acquisition, Loss, Psychological Reality, and Functional Explanations, Vol. 2. John Benjamins, Amsterdam, pp. 445–470. Van Lancker, D., Cummings, J.L., 1999. Expletives: neurolinguistic and neurobehavioral perspectives on swearing. Brain Res. Rev. 31, 83–104. Weinert, R., 2012. Complement clauses in spoken German and English: syntax, deixis, and discourse-pragmatics. Folia Linguist. 46 (1), 233–265. Zifonun, G., Hoffmann, L., Strecker, B., 1997. Grammatik der deutschen Sprache, vols. 1–3. Mouton de Gruyter, Berlin.