Linearity and tone in the unfolding of information

ABSTRACT

In this article, we show how speakers manage information flow in real time and signal their interactional expectations. Speech unfolds temporally, not only as a series of lexical items which are grouped into grammatical units, but also as a sequence of tone groups which have the potential to achieve an act of telling. While there is a general belief that speakers commence their discourse with information that is shared prior to telling information that updates the common ground, our analysis of a corpus of monologue and dialogue shows that matters are not so simple. Speakers’ informational needs are balanced moment by moment within and between increments which are themselves shaped by the interlocutors’ shifting apprehensions of communicative purpose and the extent of presumed shared information. In our analysis we combine (i) a speech functional analysis, (ii) a description of a hierarchy of informational foci and (iii) prosody in order to develop a detailed description of how speakers manage information flow in real time. This enables us to show how speakers simultaneously balance informational flow while signalling their interactional expectations. Our conclusion is that speakers manage information flow by balancing textual, interpersonal and ideational choices.

KEYWORDS:

• Intonation
• interaction
• information
• increment

Acknowledgments

Sterling work by the editors and grateful thanks to the critical comments from reviewers.

Disclosure statement

No potential conflict of interest was reported by the authors.

Key of Symbols used

Intonation

\=Falling tone /=Rising tone –=Level tone \/=Fall-rise tone

/\=Rise-fall tone syll=prominent syllable h=High key or high termination

l=Low key or low termination.

Grammar

N=Nominal V=Verbal V^l=Non-finite verbal A=Adverbial

E=Adjectival P=Prepositional d=determiner ex=exclamative

Lower case letters=suspension

Notes

¹ http://www.youtube.com/watch?v=EjxgnpVNjJQ [last accessed March 5 2019].

² In cases such as: | a \lot of houses got flooded | and it really damaged people’s \lives | it proved impossible to know definitely whether the and belonged to the first or second tone group. We decided to include such items in the following tone group as they are more naturally described as initial rather than final elements. In any case regardless of whether the item was initial or final it was by definition Given (Halliday 1967; Halliday and Greaves 2008).

³ Readers will note that the number of tones and hence tone groups in the conversation detailed in Table 3 is greater than that reported for all speakers in Table 1. This is because in the conversations speakers produced a number of tone groups either as (i) backchannels or (ii) as failed attempts to grab the floor which fell outside of increment structure.

⁴ The term increment is, as Auer (2007, 647) notes, a slippery beast but basically it represents an acknowledgement that speakers produce their language in a piecemeal fashion. In the Conversation Analytic framework the term refers to a stretch of speech which contains a prosodic boundary and is the grammatically structured extension of a turn completion unit (TCU) (Schegloff 1996, 59). In more recent interactional linguistics studies the increment is defined as a non-main clause continuation after a possible point of turn completion (Ford, Fox, and Thompson 2002, 16). Auer (1996, 2007) argues that a full typology of increments must incorporate syntax, semantics, prosody, semantics and action structure. His view is closest to the one proposed here, but differs in that we do not regard turn taking as a relevant factor.

⁵ As there are no example of Wh questions in the data, we have had to make one up.

⁶ The speaker was female and in the following examples the pronoun he or she is used to refer to the actual gender of the speaker.

⁷ By linear we simply mean the ordering of words as manifested in the speech signal. However, when we refer to the concept of linearity we are referring to the assumption that speech usually emerges with informationally given elements preceding informationally new ones.

⁸ Example 11 below is an example of an 8^th simple chain NVEA and hence Table 2 is not a complete list.

⁹ Declarative questions of the type you bought a house? make use of chains otherwise used for telling increments.

¹⁰ Suspensive elements are notated in lower case. The symbol ex stands for exclamative. Thus, in a full analysis a nominal element such as the big light is notated as d e N with the determiner and adjective suspending the articulation of the noun. The arrows signal that the earlier element prospects the later one or that a later element is required. In examples, such as 6, the use of double-headed arrows signals that the realisation of the later element does not fulfil the prospection or that the two elements joined by the arrow are interdependent and thus constitute one composite element in the prospection structure.

¹¹ Had the speaker produced a suspensive Adverbial element signalling time or place such as yesterday, e.g., yesterday it had been on the news and stuff the suspensive a element would have signalled the angle with which the speaker intended to orient the hearer towards the unfolding proposition.

¹² In Section 3.2 alone we have presented the examples with both the intonation and chains shown. We have not done so elsewhere in the paper because of space and readability issues. But all of the proposed increments presented in the paper do represent fulfilment of the grammatical criterion.

¹³ See Section 2 for discussion of the measuring of the relative pitch height of onset and tonic syllables.

¹⁴ The NV elements (presumably they are) are ellipted.

¹⁵ In another context where the speaker had not felt that overt mention of the fortitude of the people was required, she could have realised her message simply by stating that people were continuing to go to the pub and trusted her hearer to infer that this meant that they were “still surviving”.

¹⁶ Tone groups can be equated to what ToBI theorists (e.g., the chapters in Jun 2007, 2015) label intonational phrases, see Ladd (2008).

¹⁷ Esser uses the term key, though his transcriptions indicate that he is actually describing termination. In tone groups with only a single prominence key and termination are realised on the tonic syllable (Brazil 1997). O’Grady (2010, 157–200) shows that in such cases linear position determines which system predominates; key in increment initial position and termination in increment final position.

¹⁸ Esser’s data were monologic telling and hence did not require any active intervention from the hearer.

¹⁹ For instance when Mary was describing how the images on the YouTube clip compared to her personal experiences of the flood, she produced the following stretch of eight tone groups containing three level tones. The three dots (…) indicate non-junctural pausing and are further evidence of processing issues or in the terms we employed in Section 3, Mary was adopting an oblique orientation. | – erm | – yeah | \and | just … it \/was quite | like \/horrible weather | and it kind of does show you that |I think like … where I – live | it was really bad as /well |.

²⁰ Heritage’s work draws upon Kamio’s (1994) theory of informational territories which argues by analogy with studies of animal behaviour that if a piece of information is closer to the speaker that it belongs to the speaker’s territory.

²¹ In the data set there is only one example of an asking increment and as a result, the following discussion focuses only on telling increments.

²² Halliday has been consistent that focal items are not objectively new but are those which the speaker predicts as New, e.g., (Halliday 1967; Halliday and Greaves 2008; Halliday and Matthiessen 2014). Thus, we do not expect the relation between tonic accents and freshly introduced items to be 100% predictive, but we do expect that more often than not tonic items will correspond with freshly introduced items. See O’Grady (2016) for a fuller discussion.

²³ We conducted two χ² tests. The first checked to see if there was a significant difference between the informational status of the final tonic items in congruent A- and AB-events. The second checked to see if there was a difference in the syntactic position of the final tonic item, i.e., whether it was in final or non-final position between congruent A- and AB-events. Only the former was positive and achieved a significance value greater than P = 0.05. The results were that the former being χ² = 8.9351, df = 1, p-value = 0.002797 and the latter being χ² = 2.6845, df = 1, p-value = 0.1013. This suggests that there may be both significant grammatical and informational reactances in how speakers produce tone groups encoding A- and AB-events. However, to be certain we would need to confirm these tentative results by examining a far larger data set.

²⁴ Though as this example is taken from monologue there was no space for the hearer to produce an overt response.

²⁵ Despite the absence of the past tense morpheme, we have interpreted and then he show me as a grammatically complete structure because it is immediately followed by the conjunction but and a further proposition which expressed her disinterest in such types of videos.

Additional information

Funding

This work was supported by Cardiff University CURL grant AH3350E893 - RLS 1819.

Notes on contributors

Gerard O’Grady

Gerard O’Grady is a Reader at the Centre of Language and Communication Research at Cardiff University. He is the author and editor of 5 books including A Grammar of Spoken English Discourse, London: Continuum 2010. His recent work examines differences between language as process and product and the information structure of spoken discourse.

Tom Bartlett

Tom Bartlett is Reader in Functional and Applied Linguistics at Glasgow University. His research is centred around language as a social phenomenon. He specialises in Systemic Functional Linguistics (SFL) and Critical Discourse Analysis (CDA), with a particular focus on functional descriptions of Scottish Gaelic and Discourses of Sustainability.