SPEAKERS' OBJECTIVE / SUBJECTIVE STANCE AS A SOURCE OF INTONATION VARIATION IN ENGLISH SPONTANEOUS MONOLOGUES
SPEAKERS' OBJECTIVE / SUBJECTIVE STANCE AS A SOURCE OF INTONATION VARIATION IN ENGLISH SPONTANEOUS MONOLOGUES
The study reveals one of the factors determining intonation variation in English spontaneous monologues – a speaker's objective or subjective stance on a theme. In a listening analysis, it has been observed that taking an objective stance a speaker generates an intonation pattern close to that of prepared speaking and reading, while a subjective approach to an issue and self-concentration leads to a number of deviations. The purpose of this study is to describe the intonation of the most deviating variety of spontaneous speaking and to identify one of the causes of its peculiarities. The intonation patterns of the two varieties of spontaneous speech were found to be different in the leading function of intonation, the sharpness of pitch contrasts and the most frequent types of nuclear tones. The findings suggest that the intonation of spontaneous speaking is not homogeneous. The research results can be beneficial to teaching English as a foreign language, particularly listening and speaking skills, as well as to improving the procedures of automatic spontaneous speech recognition and generation.
One of the difficulties in describing and modelling intonation is its variability under the influence of different factors: dialectal, situational, personal , , . A particular case of considerable disparity has been discovered between the intonation of reading, on the one hand, and that of spontaneous speaking, on the other , , , . Among many other differences between reading and speaking, it has been pointed out that a reader generates “surface” prosody “following only the roughest and simplest prosody-syntax mapping constraints” [9, P. 1221], while a speaker involves deeper cognitive processes. However, spontaneous speaking can hardly be viewed as an area of homogeneous intonation patterns. Listening observations on the material of English spontaneous monologues allow us to say that some realizations are not far from those of reading and prepared speaking, whereas others may deviate so far as to be considered a breach of the abstract model of English intonation associated with well-structured prepared speaking and reading , , . We hypothesize that speakers' objective or subjective approach to a discussion issue can be an important factor influencing intonation. Speakers can view a situation or a problem “objectively”, from the outside, without associating it with themselves or, on the contrary, they can subjectively describe their personal experience, feelings, emotions and attitudes related to the situation. In the latter case, speakers appear to be more involved in self-expression than in giving actual information. These kinds of stance taking are linguistically marked by the prevalence of 3d-person pronouns (it, he, she, they) or 1st-person pronouns (I, we) and create specific features both in the grammatical structure of speech  and its prosody . The intonation of spontaneous speech generated by speakers taking a subjective stance is of primary interest in this study.
2. Methods and principles
The research material was selected from a few podcasts of the BBC radio on-line broadcast over the past decade. The narrow corpus of spontaneously spoken extracts included ten samples of 1st-person monologues and ten samples of 3-d person monologues. The monologues were generated in British English by 11 men and 9 women. The preliminary listening analysis of the two kinds of spontaneous speech revealed considerable differences in their intonation organization. To describe these differences, we resorted to further auditory analysis and comparative evaluation of the most important intonation parameters. The selected extracts were provided with scripts which were annotated with the help of traditional symbols for nuclear tones (\ – falling, / – rising, \/ – falling-rising, > – mid-level), stressed syllables (') and pauses inside utterances ( | ). To provide more reliable listening impressions at some points, three auditors were invited to share their perception of particular intonation phenomena such as boundaries between intonation groups, sentences and phonopassages; complete and incomplete falling tones. So, the main methods of research applied in this study are auditory and comparative.
To begin with, it stands to reason to present annotated samples illustrating the 1-st and 3-d person spontaneous monologues and specify their overall psycholinguistic characteristics alongside their key intonational features. The third-person sample provided below is a Maths teacher's reflection on how Maths is taught at schools and what is wrong about the traditional approach to teaching it.
For \/me | the mathe'matics that's 'taught in 'most 'school \/classrooms | is a 'very \narrow … version of the subject. If you 'ask a mathema/tician | 'what they \/do | they'll 'say they 'all 'start by having a 'good \question. To \/pose a question | \/students | 'never \do this in maths classes. They may 'ask 'questions for \/help | but 'actually being i\/mmersed | in a mathe'matical situ\/ation | and coming 'up with a \/question | \doesn't happen. 'Here's a /teacher | 'showing you a /method, | 'now \follow | and repro\duce it. \/Some students | are 'very \good at that | but they just disen\gage | because it's ex'tremely \boring. So when we 'teach 'maths the 'way that mathe'matics is 'used in the \/world | and the 'way that mathema\/ticians work on it | \many more kids engage, | \many more kids do well .
This extract contains the presentation of a problem as an objective description and evaluation of what the speaker knows about the situation. The speaker chooses to distance herself from the discussion issue and to look at it from the outside. As a result, her attention is concentrated on 3-d person objects: mathematics, mathematicians, students, some students, they. Our research material shows that distancing is possible with any theme – even a speaker's inner world can be viewed and analyzed at a distance .
The objective approach to a discussion issue entails an intonation pattern which should point out those distant objects through the falling-rising tone. Therefore, falling-rising nuclear tones are quite frequent in monologues with the objective stance. What is also important for the speaker is declaring something about the objects singled out. This task is carried out with the help of complete falling tones reaching a low pitch level. Thus, the objective stance in spontaneous speaking brings the speaker to use the same basic intonation pattern revealed in reading and prepared speaking. This generalized model of English intonation involving the falling-rising referring tone and the falling declaring tone is identified by many linguists including D. Brazil and M. Halliday, who highlight its connection with the theme-rheme structure of utterances , . It is consistently realized in 3-d person spontaneous monologues, although the conditions of spontaneous speech production occasionally break it.
Many more deviations from the general model of English intonation are found in the spontaneous variety with the speaker taking the subjective stance. In the sample below the speaker shares her personal experience of being married to a person of a different religion. She concentrates on her inner thoughts and feelings rather than on people and objects outside.
When we \met, | we met \online | and then we corres\ponded | for quite some \time | before 'meeting in \person | in \Oxford | and I think we >always | … we were 'always a\ware | of the re'ligious 'aspects of our \lives. It was \never anything | that cre'ated any \tension | rather I think we 'always 'focused upon how 'much of a 'shared \history there was | between 'Islam and Ca\tholicism ...| and I think that kind of \influenced our decision | to sort of 'honour \both sides in our /marriage. Getting .married in Ma/laysia | re\quired you know | 'Joe con\verting, | 'doing some con\version ceremony | which he was 'happy to \do | and on \/my part | it re'quired \/me | kind of a\greeing to |… to 'certain >principles | with the >priest, | so I think we 'always a>pproached it | with 'quite an 'open \heart | and I think per\/haps | 'maybe 'slightly \/arrogantly | I think that you know we per'haps still >think | that we can 'marry the \/two | in some sort of 'perfect \way | but I >think you know | I su>ppose | 'not 'everyone will 'probably a\gree, | you know we might up\set some people | a'long the \way. But for the \/most part | I have 'not found that \difficult | I have 'not found that 'difficult to \do | but \/maybe | it's just \arrogant .
In this kind of 1-st person communication the pronouns I, we naturally predominate and represent the unmarked theme of the text. As a result, intonation does not contribute to the theme-rheme distinction, and their boundary is blurred. Even if the theme is expressed by the third person, it is often pronounced as something known or predictable from the context and, therefore, requiring no special effort in pronunciation. Thus, in the sample text above, the theme “getting married in Malaysia” is pronounced at a low pitch level, at a faster speech rate, with a rising nuclear tone signaling only non-finality and making it difficult to grasp the theme. A similar observation about the vague intonation division between the theme and rheme in spontaneous speech has been made by other researchers, for example .
Not only is the theme-rheme boundary blurred due to the subjective stance but the whole syntactic organization of the text appears to be unstable. Speakers seem to find it difficult to divide their speech into well-formed sentences. Aiming to signal continuation a speaker adds new chunks to what has already been said with the help of such linking devices as and, but, I think, rather, I mean, you know, like. The speaker's intention to continue is also rendered through the types of nuclear tones among which incomplete falling, low rising and mid-level are the most frequent. The invited auditors found it a complicated task to identify adequate locations for full stops delineating grammatically complete sentences. Consequently, it is not easy to turn such a spoken text into a written one.
Since a speaker's subjective stance entails more deviations from the general intonation model, it is worth while viewing these deviations in more detail drawing on the auditors' perception of the extracts. We concentrate mainly on the pitch component of intonation: nuclear tones, pitch contrasts at the boundaries and within utterances.
The main function of nuclear tones in texts with the speaker's subjective stance is the indication of continuation intention. Various tones are employed to serve this purpose, the most common being incomplete falls, low rises and mid-level types. The falling-rising tone appears rarely and, if used, indicates contrasts rather than themes. In the monologue presented above, for instance, the speaker compares her husband's actions with hers, which leads to the use of the fall-rise on my and me (… and on \/my part | it re'quired \/me...). Incomplete falls predominate in long utterances as an indication of non-finality and are opposed to complete falls terminating utterances and phonopassages. The auditors found it a feasible task to tell between incomplete and complete falling tones. Complete falls provided locations for full stops unanimously identified by the auditors, while incomplete falling tones caused a great deal of hesitation about punctuation marks in the written script. Mid-level tones often emerged in the subjective reflection not only as a means showing continuation, cohesion or hesitation but also as a way of presenting an enumeration .
Unlike grammatical sentences, intonation groups and larger semantic chunks (phonopassages) were easily identified by the auditors. Besides nuclear tones, pitch contrasts at their boundaries contributed to the identification of phonopassages. Thus, the monologue below consists of three ideas – phonopassages, separated by complete falling tones and the pitch contrast between the low-level ending of one chunk and the high-level beginning of the next. The three phonopassages in the monologue are separated with the sign ||. In them, the speaker presents three reasons why his marriage to a person of different religious beliefs has been a success: the support of their parents' families, his natural curiosity and interest in other religions, viewing religion as a cultural background rather than daily routine.
The >difficulties | are 'only lo\gistical | and bureau\cratic, | they are 'not necessarily like on the \principled | or e\motional side, | and in \/part | it's just because 'both of our \families | have been in'credibly su\pportive | and they are 'not dog'matic about re\ligion | and I think 'both \families | 'deal with rel>igion | with a 'good >dose of … | \humour | as \well. || And \yeah, | it's about 'shared curi\osity as well. I'm 'actually \interested | in 'getting to 'know \more about Islam | and I think \you've likewise | been 'interested to 'know about the 'vagaries of Ca\tholicism. || And I 'say >that | 'not being in 'any way a 'practising >Catholic, you know, | \no | it's just 'almost a 'cultural \background thing .
Another intonation phenomenon based on pitch changes is pitch downstep which implies that a few consecutive intonation groups, semantically connected, follow a physiologically determined downward direction, each of them beginning on a slightly lower level than the preceding intonation group. Pitch downstep is more vivid in monologues with the speaker's objective stance. In the subjective stance extracts it occurs rarely. Instead, we notice intonation groups pronounced here approximately within the same low register with no downward shift. Thus, the subjective stance elicits pitch leveling-off within a chunk, which can be characterized as the frequent use of “the low-level scale and mid-level ending” [1, P. 188]. Like prepared speech and reading, the objective stance involves relatively steeper slopes in declination and stronger pitch resetting at the boundaries .
Comparing the objective and subjective stance monologues in relation to the length of the intonation group measured in the number of stressed words, we have found no differences. On average the intonation group includes two stressed words: 2,08 with the objective stance-taking and 1,85 with the subjective stance. Apparently, this kind of span is appropriate for the realization of the psycholinguistic mechanisms of attention and working memory underlying the production and perception of the basic unit in spontaneous speech. However, the number of unstressed words surrounding stressed ones seems to increase with the subjective stance, which can be partially accounted for by a greater number of unstressed linking and pause-filling devices such as you know, like. Even the phrases I think (that) and I believe (that) used as discourse markers remain unstressed . Longer unstressed spans can lead to increased speech rate with the subjective stance-taking, but this has to be experimentally tested.
The listening observations presented above allow us to point out some systematic differences in the intonation organization of speech generated with the speaker's objective and subjective approach to a theme.
The first difference lies in the leading function of intonation characterizing the two types of monologues. Speakers taking the objective stance primarily aim at the clear informational structuring of their speech, at marking the theme and rheme of their utterances. In case of the subjective stance speakers use intonation to signal the continuation of their self-expression. In accordance with the leading function of intonation, different types of nuclear tones predominate. The objective stance is marked by a high proportion of falling-rising and complete falling tones while the most frequent tones with the subjective stance are incomplete falls, low rises and mid-level types.
Secondly, the overall pitch changes are more contrastive with the objective stance determining noticeable downstep within and among related intonation groups. Pitch leveling-off with the subjective stance increases the impression of speech monotone and deviation from the generalized model of English intonation. This has an impact on the rhythmical structure of speech. In monologues with the objective stance-taking, rhythm as a regular alternation of stressed and unstressed syllables is also supported by regularities in pitch changes, whereas the subjective stance generates monologues with rhythm mainly based on sentence stress.
Thirdly, both the intonation and the syntactic organization of objective-stance monologues make it easy to parse them into sentences, in contrast to a subjective-stance text which can be successfully divided into intonation groups and phonopassages but not into grammatically well-formed sentences. This feature may be a reflection of the less discrete character of the subjective reality which the speaker tries to represent spontaneously.
Despite the differences, the two kinds of spontaneous monologues realize the same intonation system and are characterized by similarities in intonation as well. Thus, the length of the average intonation group comprises two stressed words in both varieties. Besides, the types of nuclear tones are contrasted in a quantitative, not in a qualitative way. For example, in the objective-stance texts, falling-rising tones make up 27,5% of all the tones used, while in the subjective-stance monologues, their number drops to 12%.
The auditory and comparative analysis of spontaneous speech generated by speakers taking an objective or a subjective stance has revealed a few differences and similarities in their intonation organization. The former kind of speech tends to concentrate on structuring factual information and presenting argumentation, but the latter often involves the speaker's self-expression and emotional evaluation. This passionate personal involvement and self-concentration determine some deviations from the generalized model of English intonation. It has been discovered that the subjective stance is marked by the prevalence of nuclear tones serving to signal continuation, as well as smoothed pitch contrasts and a greater number of unstressed words. It can be assumed that a similar distinction of intonation is possible in reading, which may depend on the type of information prevailing in the text: factual or emotional. So this research has presented some evidence proving that spontaneous speech can vary in its intonation organization and that it is too vast a generalization to view it as a homogeneous area.