CC BY
48
Вестник Московского университета. Серия 9. Филология. 2019. № 1
Irina Kobozeva, Olga Ivanova, Leonid Zakharov
TOWARDS MULTIMODAL MODELLING OF VERIFICATION DISCOURSE MARKERS IN RUSSIAN DIALOG1
Lomonosov Moscow State University
1 Leninskie Gory, Moscow, 119991
The paper presents the results of a multimodal analysis of the most frequent Russian discourse markers (DMs) expressing positive verification — da 'yes', da-da 'yes, yes', nu da 'well, yes', konechno 'of course', razumeetsja 'sure', jestestvenno 'naturally'. In a dialogical turn of a given speaker they mark the proposition P of the previous turn of the interlocutor as true. DMs of this type have variable prosody and may redundantly be accompanied by nodding. The aim of the study is to identify factors that influence the paraverbal aspects of such DMs. The analysis of mini-dialogs from the multimodal sub-corpus of RNC revealed the correlations between various syntactic, semantic, pragmatic and sociolinguistic factors on the one hand and non-verbal behavior and prosodic features of DMs on the other. The results obtained may be useful for modelling the paraverbal behavior of embodied computer agents and robots.
Key words: multimodal speech analysis; discourse markers; verification; gesture; prosody; pragmatics; semantics; Russian.
THEORETICAL PREREQUISITES
Spoken discourse as the primary form of language has attracted in recent years an increasing number of researchers. In spoken discourse, three channels of information transfer are distinguished: verbal (segmental units of speech), prosodic (non-segmental units of speech: pitch variation, loudness, tempo, timbre, etc.), and visual (gestures, facial expressions, body language, etc.). Propositional information is coded mainly by verbal means of communication, prosodic means of communication express information of communicative and pragmatic nature (e.g., communicative structure of
Irina Kobozeva — Prof. Dr., Department of Theoretical and Applied Linguistics, Faculty of Philology, Lomonosov Moscow State University (e-mail: kobozeva@list.ru).
Olga Ivanova — BA in Linguistics, MA student, Department of Theoretical and Applied Linguistics, Faculty of Philology, Lomonosov Moscow State University (e-mail: wiola100@ rambler.ru).
Leonid Zakharov — Senior Engineer, Teaching and Research Computer Center, Faculty of Philology, Lomonosov Moscow State University (e-mail: leonid_zakharov@mail.ru). 1 The paper was supported by RSF (grant № 16-18-10004).
the utterance, its purpose, emotional state ofthe speaker), and gestures can convey both propositional and communicative-pragmatic information. All three types of sense coding tools are closely interrelated: they complement or duplicate the meaning transmitted through other channels, and sometimes contradict with each other, as in the case of irony, when intonation and facial expressions indicate that what Speaker had in mind was the opposite of what was expressed verbally. At the same time, the context, both verbal and extra-linguistic, plays an important role in conveying and understanding the meaning of the utterance. All this explains the increased attention of linguists studying spoken discourse to paraverbal means of communication, in particular, to the prosodic features of speech and to the accompanying gestures.
Usually researchers concentrate on investigating the interaction of verbal means (including discourse markers) with prosody (see, for example, [Bonnot & Kodzasov, 1998]) or with gestures (see, for example, [Cienki & Müller 2008, Richter, 2010]). Discourse markers (DMs) in speech, unlike full words and grammatical markers, often represent a multimodal cluster2. The program for describing DMs taking into account the information of all the three modalities of such clusters was proposed already in (Kobozeva & Zakharov, 2004), but there are few cases of its realization and they concern various non-connected elements of the DM system (see [Kobozeva & Zakharov, 2007], as well as [Grishina, 2012]). In this paper, we propose a unified multimodal frame for the analysis of a semantic group of the DM, which allows tracking correlations between verbal, prosodic and gestural phenomena, at the same time registering the factors that are presumably capable of influencing their realization. Discourse markers expressing verification were chosen as the object of modeling because their meaning in the spoken dialogue is transmitted either through the auditory channel (by means of pronouncing the word with a certain intonation) or through the visual channel (with the help of a nod) or through both channels simultaneously. In this connection, the question arises: what determines the way in which this meaning is embodied. In addition, the main markers of this type in Russian — da 'yes' and net 'no' — are represented in a number of prosodic variants reflecting the pragma-semantic differences between their usages (semantic-prosodic variability of da 'yes' was discussed in [Grishina, 2011; Kobozeva & Zakharov, 2012]).
Verification discourse markers (VDMs) are discourse markers, which can be used independently as a verificative utterance, i.e. utterance that states truth or falsehood of a proposition introduced by the interlocutor (more formal definition is given below). So, da and net in Russian just as yes and no in English belong to the VDMs, since they are used not only as
2 The term was introduced by E.A. Grishina (see [Grishina, 2011: 243]).
initial components of a verificative statement, as in (1a) and (2a), but also form such a statement on their own, as shown in (1b) and (2b):
(1) Eto mesto svobodno? — a. Da, svobodno. / b. Da. Is this place free? — a. Yes, it's free. / b. Yes.
(2) Petya uzhe vyzdorovel? — a. Net, on yeshche ne vyzdorovel. / b. Net.
Petya has already recovered? — a. No, he has not recovered yet./ b. No.
Besides da the set of synonymous VDMs includes konechno 'of course', razumeetsya 'of course', yestestvenno 'naturally and a number of other words. There are also gestural synonyms for VDMs: nod is equivalent to da and head shake II [Grigor'eva, Grigor'ev, Kreidlin, 2001]3 is equivalent to net. These gestures, as well as VDMs, can be used independently as a complete communicative act of verification (gestural verificative statement), but they are often produced while pronouncing a verificative utterance with or without VDM.
The term "verificative utterance" was introduced in (Dobrushina, 1993). We shall use the definition offered by E.R. Dobrushina (1993) in a slightly revised form that is better suited for analyzing dialogues:
The utterance W is verificative if and only if:
1) the utterance W is a reaction of the speaker to some utterance U of the interlocutor;
2) the utterance U expresses a proposition 'P' in one of the three truth-value statuses4: "true" (T), "false" (F), "undefined truth value" (0);
3) the utterance W includes an assertion of the truth or falsity of the proposition 'P'.
From this definition it follows that verificative utterances can be either positive, as in (1), or negative as in (2). Next, only positive verificative utterances are taken into account. Let's pay attention to the fact that "verificative" is not an illocutionary characteristic of an utterance, since it refers only to its propositional content and is independent of the goal of the speaker and other felicity conditions of the corresponding speech act. Verificative utterances can have different illocutionary forces or functions (IF). Thus, the utterances in (1) have an IF of an answer to a question, and in (2) they are both answers to a question and rejections of the interlocutor's supposition. A number of other IFs are possible for a given semantic type of statements, for example, confirmation, consent, refusal and their pragmatic variations. Thus, IF of an utterance may be considered as one of the possible factors for varying the paraverbal accompaniment ofVDMs in verificative statements.
3 In [Kreidlin, 2004] these gestures are classified as regulators.
4 The truth value status of the proposition here is determined by the truth value attributed to it by the speaker, and not by its actual truth value in the world of discourse.
THE MATERIAL AND METHOD OF ITS ANNOTATION
The data for the analysis were taken from the corpus MURCO (Multimodal Russian Corpus) that is a sub-corpus of the Russian National Corpus (RNC). We extracted video records with examples of dialogical units containing a reactive utterance with one or more of the six VDMs prototypically expressing positive verification (see Table 1). For some of these VDMs it has been shown by Dobrushina (1983), and by Kisseleva in [Kisseleva & Payar, 1998] that they have subtle pragma-semantic differences, as
Table 1. The number of utterances with different VDMs included in the sample
da da-da5 nu da6 konechno razumeetsya yestestvenno
46 46 42 45 46 41
Each utterance (= dialogical turn) with one of those VDMs was annotated according to four groups of parameters.
Group 1 represents the formal characteristics, i. e. the observed form of the entire dialogical turn, or only of the VDM:
a) speech sequence (the written form of the utterance in the MURCO notation);
b) prosody of the VDM (transcription in the notation of S.V. Kodzasov (2009)7;
c) verbal description of the gestural series accompanying the speech, with fixation of the location of gestures in Cartesian coordinates after Grishina (2013, 2014);
d) other paraverbal characteristics of the utterance, especially the establishment of the eye contact of the verifier with the interlocutor when making an utterance with VDM.
Group 2 represents the syntactic characteristics of an utterance with VDM, which can presumably have an impact on the prosody of VDM or its gesture series. These characteristics include:
a) the syntactic function of the VDM which can be
— a parenthetical word, or sentence adverb (as in Of course, I will come; Yes, he is here);
5 In (Dobrushina, 1993) it was shown that the VDM with the iteration of da segment (reaching 5 times in our material) differs in meaning from VDM da.
6 2 Although VDM nu da is formally and semantically a combination of one of the prosodic-semantic variants of modal particle nu (see its description in [Baranov & Kobozeva, 1988]) is compositionally combined with the verificative meaning of the propositional particle da, for purposes of establishing correlations between verbal and gestural units, it was more convenient for us to consider this combination as a composite VDM.
7 In contrast to the works of E.A. Grishina, in which prosody of DM is described on the basis of subjective evaluation, in our study prosodic transcription is the result of instrumental analysis.
— a word equal to a sentence (e.g., Yes; Naturally, baby!)
— a sentence within the multy-sentential utterance (e.g, Of course! We invite only those / who agree to undergo a full survey.);
b) the presence in the utterance of other discourse units (DU) before or after the verificative DU;
c) the presence of other DMs in the utterance (e.g. A!Of course / we are very interested in it! Thank you!).
The characteristics (a) and (b) are important primarily for prosody, but also as a factor of possible gesture shift relative to VDM. As for (c), the presence of other DMs may help to clarify the pragmatic meaning of the verificative utterance, and that is important, because verificative utterances, the propositional part of which is often reduced to VDM, are illocutionally ambiguous.
The third group comprises pragmatic characteristics. They include:
a) the illocutionary force (IF) of the verificative utterance;
b) the IF of the stimulus utterance;
c) the pragmatic polarity of the utterance:
— positive (+) if the verifier is sincere and serious in stating the truth of the proposition P introduced by the interlocutor;
— negative (-), if the verifier uses irony, i.e. asserting the truth of P has in mind its falsity.
Parameter (b), i.e. the IF of the stimulus utterance is introduced in the annotation for the purpose of controlling the accuracy of the identification ofparameter (a), i.e. the attribution of IF to the utterance containing VDM. Parameter (c) of pragmatic polarity was included in the model based on the assumption that the nod, in contrast to the VDM, is an unambiguous although a secondary marker of positive verification, and therefore it will rather be present in utterances in which the speaker is sincere and serious about the positive verification of P, and will be absent when VDM is used in utterances with a phatic IF (e.g. the contact maintaining) or utterances of ironical agreement with the opinion of the Interlocutor.
The fourth group contains sociolinguistic characteristics:
a) social status relation of the Verifier (V) and the Interlocutor (I) and: V = I / V < I / V > I; b) the age relation of V and I: V = I / V < I / V > I;
c) the presence of conflict between V and I in the situation (+ / -);
d) sex of V and I: m / f.
Introduction of these parameters into the model is due to the fact that positive verification in utterances having IF of "confirmation" or "consent" is consistent with the Principle of politeness and the Compliance Rule (with the opinion and desires etc. of the Interlocutor) that is subordinate to it (Leech, 1983). Hence it can be assumed that the nod, being a sign that is informatively redundant in the presence of VDM, will function as a politeness marker, i.e. will be present in situations that presuppose an emphasized degree of politeness of V to I and will be absent when there
is no need for such an emphasis. The theories of politeness converge on the fact that the degree of politeness is influenced by the social distance between interlocutors, determined by their social position, age and gender. Conflictness of the situation can suspend the operation of the Principle of politeness, i.e. we can expect that in conflict situations there will be no nod. In addition, it was observed in gender linguistics that different sexes show different preferences in the choice of some discourse markers. We have assumed that the presence or absence of a nod in verificative utterances with VDM belong to such preferences.
Here is an example of an annotated fragment from the obtained database:
(3) [Trenton, Vladimir Basov, male, 56, 1923]
[Lady Vendeler, Elena Solovey, female, 32, 1947] Trenton: Minutku! U vas/ konechno/ yest' spisok dragotsennostey? 'Wait a minute! Of course you have a list ofjewelry, don't you?' Lady Vendeler: Razumeyetsya/ on u menya v kabinete. 'Of course / it's in my parlour.'
[Yevgeniy Tatarskiy et al. Klub samoubiyts, ili priklyucheniya titulovannoy osoby (Suicide club, or the adventures of a titled person), 1979]
Table 2. The annotation of the example (3)
Speech sequence Razumeyetsya / on u menya v kabinete.
Prosody razumeyetsya (\"— /")
Gestural sequence - Nod
Other paraverbalics Looks at the interlocutor Looks at the interlocutor
Syntactic function of VDM Parenthetical phrase
Presence of DU before / after VDM / +
Other DM in the utterance -
IF of the utterance with VDM Confirmation answer
IF of the stimulus turn Question with assumption
Pragmatic polarity +
Ratio of the social statuses of V and I V > I
Ratio of the ages of V and I V < I
Conflict -
Gender of V and I I (m), V (f)
The reliability of annotation was ensured by the fact that 20% of the data annotated by one analyst were independently annotated by another analyst. The degree of agreement for prosodic parameter was about 70%, for other paraverbal parameters it was 98%, for syntactic function it was 90%, for pragmatic parameters it was about 80%. Differences in annotation were resolved by discussion.
DATA ANALYSIS
The annotated data were analyzed. The distribution of nods over utterances with different VDM was calculated, see Table. 3.
Table 3. The ratio of utterances with a specific VDM, followed by a nod (%)
da da-da nu da konechno razumeetsya yestestvenno
69 60 55 75 66 9
Even from this simple calculation it is obvious that the VDMs differ in the frequency of supportive nodding: whereas for most VDMs presence of nods is more likely than its absence, for yestestvenno 'naturally' nod is not typical. This fact indicates that some semantic component that is part of the meaning of this VDM prevents the use of a nod. In this case we are talking about the special kind of the guarantee component the relevance of which for the semantic description of discourse markers has been proved proved in [Kisseleva & Payar, 1998]. In the case of yestesvenno, it is 'the nature of things' that is the guarantee of the truth of the proposition P, in addition to the speaker [Kisseleva, 1998: 65]. The choice ofVDM yestesvenno shows that the verifier wants his role as the guarantee to be backgrounded, and so attraction of attention to one's personal involvement into verification, achieved with the help of a nod, becomes undesirable.
In addition to the nods, the following types of gestures that accompany the VDMs and enrich them with additional meanings have been identified: "pinch", "forefinger", "opening the fingers up from the position ofthe tightly closed fist", "moving the arms away towards the interlocutor with the open palm up", "shrug one's shoulders", "raise one's eyebrows", "tilt one's head", "spread one's hands", etc. Moreover, in some contexts, there was a gesture "close one's eyes", which is associated with the meaning of the word yes [Kreidlin, 2004: 401] and as it can be used with verificative meaning independently, it must be included into the set of VDMs used in Russian dialogue.
Let us now turn to the revealed correlations.
It turned out that the presence / absence of an accompanying nod correlates with a selected discourse marker, see Table 4.
Table 4. Relationship between presence of a nod and choice of VDM
presence of a nod lack of a nod
da 32 14
da-da 22 20
nu da 28 18
konechno 33 12
razumeetsya 31 16
yestestvenno 9 32
X2 = 30.673, df = 5, p-value = 1.087e-05 Critical value of x2 with confidence level p<0.01 amounts to 15.086
Regardless of the marker used, the nod is usually absent if the pragmatic polarity of the verification is negative (irony, skepticism), as in (3), and / or the IF of an utterance is such that verification is not the goal of the speech act, as in (4), where Da-da was used not to confirm the truth of what the interlocutor said, but to inform him that his point was taken into account:
(3) Galya: Yesli chestno/ ya v Moskvu ne dlya togo priyekhala/ chtob poly zdes' myt'.
'To be honest / I did not come to Moscow to mop the floors.' Stasis: Nu konechno/ srazu zvezdoy/ da? Tol'ko ved'/ kak govoritsya/ vse tseli porazheny. Nado stoyat' v ocheredi. A ochered' dlinnyushchaya!
'Of course, celebrity at once, right? But in fact / as they say / all the targets are hit.
One must stand in line. And the line is so long!'
[Andrey Mikhalkov-Konchalovskiy, Avdot'ya Smirnova. Glyanets
(The Gloss), movie (2007)]
(4) Borya: Ey / alo /alo / patsany / yego shchas prezentovat' 'Hey / hey/ hey / guys / he's right about to be presented'. Lesha: Da-da. My shchas reshim vs'o. 'Yes, yes. We'll manage everything right away.'
[Oleg Fomin et al. Den' vyborov (The Election Day), movie (2007)]
There is also a correlation between the verifier's eye direction and the choice of VDM (see Table 4), as well as the presence of a nod (see Table 5).
Table 5. The presence of eye contact with different VDMs
presence of eye contact lack of eye contact
da 24 22
da-da 29 12
nu da 31 11
konechno 25 20
razumeetsya 40 7
yestestvenno 28 13
X2 = 12.413, df = 5, p-value = 0.1001 Critical value of x2 with confidence level p<0.05 amounts to 11.07
As our data show, VDM razumeetsya in most cases is accompanied by the eye contact with the Interlocutor, which can also be explained by the special type of the "guarantee" in the semantics of this VDM. In case of razumeetsya in addition to the verifier, any intelligent subject is also the guarantee of the truth of P (Kisseleva, 1998, p. 355], and thus the Verifier,
fixing his eyes on the Interlocutor as if endows the latter with the role of the guarantee of the truth of the proposition8.
The presence of eye contact between the interlocutors correlates with the presence or absence of a nod when pronouncing the VDM (see Table 6), and that is quite natural taking into account that nods are performed for the visual perception of the Interlocutor:
Table 6. Correlation between eye contact of V with I and nod
presence of eye contact lack of eye contact
presence of nod 115 31
lack of nod 71 57
X2 = 15.929, df = 1, p-value = 6.575e-05 Critical value of x2 with confidence level p<0.01 amounts 6.635
In the field of sociolinguistic variables, the following tendencies were revealed. The presence or lack of a nod turned out to be connected with the social distance between the interlocutors, see Table. 7:
Table 7. Correlation between social distance between V and I and nod
presence of nod lack of nod
V<I 33 16
V>I 39 20
V=I 76 74
X2 = 6.5893, df = 2, p-value = 0.03708 Critical value of x2 with confidence level p < 0.05 amounts 5.991
The data show that the verification is more often accompanied by a nod if there's a social distance between V and I. This can be interpreted as evidence that this gesture is a sign of positive politeness aimed at rapprochement with the Interlocutor rather than the negative politeness aimed at keeping the distance between the Speaker and the Addressee [Brown & Levinson, 1987].
The use of nods together with VDM also depends on the presence or absence of a conflict in the communicative situation, see Table. 8:
Table 8. Nodding in situations with or without a conflict
presence of nod lack of nod
presence of conflict 29 43
lack of conflict 121 74
X2 = 10.126, df = 1, p-value = 0.00234 Critical value of x2 with confidence level p < 0.01 amounts 6.635
As for the prosodic features of VDM, we observed the correlation between the illocutionary type of the utterance and such an integral prosodic
8 Such an interpretation of establishing visual contact of V with I in a dialogue was proposed by S.G. Tatevosov in personal communication.
parameter as the degree of general reduction of VDM: in cases where the verification of P is not the goal of the speech act, e.g. in the phatic reactions, exemplified in (4), VDM has a smaller duration and loudness, and is generally pronounced less distinctly. In this case, we can talk about a particular case of iconicity: central, prototypical uses of VDM are characterized by normal articulation, and peripheral — by the weakened one. Another correlation is between the presence / absence of conflict and loudness. Conflictual positive verification is characterized by abnormal level of loudness, but contrary to expectations, this level may be either higher than the normal level of sound intensity or lower than this level.
CONCLUSIONS
The connection between gestures and subtle differences in semantics of VDM is obvious, but its manifestation is more a tendency than a rigid rule. VDM is often accompanied by several simultaneous gestures. One and the same gesture can be realized in spoken discourse along with different discourse markers. The reason for such variability is the multidimensional semantic structure of VDM, complicated by the pragmatic component of each individual utterance, which determines the intersections of the sets of gestures accompanying different markers.
The presence or absence of a nod supporting the use of VDM depends on such sociolinguistic parameters as social status of interlocutors, their age, the presence of conflict in the situation. At the same time nodding is not connected with gender. The syntactic status of a VDM in the utterance proved to be irrelevant to nodding. VDM yestestvenno 'naturally' definitely constrains nodding, and this fact may be attributed to an independently established specific semantic feature of this word, namely the kind of the «guarantee» of the truth of the verified proposition, which is the nature of things. Eye contact with the Interlocutor is almost obligatory for VDM razumeetsya 'of course, lit. it's understandable', and a proposed explanation is again an independently established special kind of the «guarantee» of the truth of the verified proposition, which is the reasoning ability of the Interlocutor. At the same time there is a correlation between nodding and eye contact.
The positive pragmatic polarity of verification is necessary, but not sufficient for supporting the verificative utterance with a nod. Prosodic variability of VDM correlates with the IF of the utterance, with the status of the verificative component in the semantic structure of the utterance (central or peripheral), with the pragmatic polarity of the utterance and the presence of conflict in the situation.
References
1. Baranov A.N., Kobozeva I.M. Modal'nye chastitsy v otvetakh na vo-pros. [Modal particles in answers to questions] Pragmatika i problemy intensional'nosti, 1988. Moskva: IYa AN SSSR, pp. 45-69.
2. Bonnot C., Kodzasov S.V. Semanticheskoe var'irovanie diskursivnykh slov i ego vliyanie na linearizatsiyu i intonirovanie (na primere chastits zhe i ved'). [Semantic variation of discourse markers and its influence on linearization and intonation (the case of particles zhe and ved')] In Kissel-eva K., Payar D. (Eds.). Diskursivnye slova russkogoyazyka: opytkontekst-no-semanticheskogo opisaniya. Moskva: Metatekst, 1998, pp. 382-443.
3. Brown P., Levinson S. Politeness: some universals in language usage. Cambridge: Cambridge University Press, 1987.
4. Cienki A., Müller C. (Eds.). Metaphor and Gesture. Gesture Studies, 3. Amsterdam — Philadelphia: John. Benjamins, 2008. P. 307.
5. Dobrushina E.R. Verifikatsiya v sovremennoi russkoi dialogicheskoi rechi. [Verification in modern Russian dialogical discourse] Diss. ... kand. filol. nauk. Moskva: MGU, 1993.
6. Grigor'eva S.A., Grigor'ev N.V., Kreidlin G.E. Slovar' yazyka russkikh zhestov. [The dictionary of Russian gestures] Moskva — Vena: Yazyki russkoi kul'tury; Wiener slawistischer almanach. Sonderband 49, 2001.
7. Grishina E.A. O mul'timodal'nykh klasterakh v ustnoi rechi. [Multimodal clusters in spoken Russian]. Komp'juternaja Lingvistika iIntellektual'nye Tehnologii, 2011, 10 (17). Moskva: RGGU.
8. Grishina E.A. Da v russkom ustnom dialoge. [Da in Russian dialog] Russian Linguistics, 2011, 35 (2), pp. 169-207.
9. Grishina E.A. Slovo i zhest korpusnye issledovaniya ustnoi rechi. [Word and gesture: corpus investigation of speech]. Saarbrücken: LAP, 2012.
10. Grishina E.A. Zhestikulyatsionnye profili russkikh pristavok. [Gesture profiles of Russian prefixes] Komp'juternaja Lingvistika i Intellektual'nye Tehnologii, 2013, 12 (19). P. 19.
11. Grishina E.A. Zhesty i pragmaticheskie kharakteristiki vyskazyvaniya. [Gestures and pragmatic characteristics of an utterance] In Mul'timodal'nye kommunikatsii: teoreticheskie i empiricheskie issledovaniya, 2014. Moskva, pp. 25-47.
12. Kisseleva KL. JESTESTVENNO, ili DE RERUM NATURAE. [JEST-ESTVENNO 'naturally' or DE RERUM NATURAE] In Kisseleva, K., Paiar D. (Eds.). Diskursivnye slova russkogo yazyka: opyt kontekstno-se-manticheskogo opisaniya. Moskva: Metatekst, 1998.
13. Kisseleva KL. RAZUMEETSYA, ili SAMI PONIMAETE. [RA-ZUMEETSYA 'of course' or YOU UNDERSTAND] In Kisseleva K., Paiar D. (Eds.). Diskursivnye slova russkogo yazyka: opyt kontekstno-se-manticheskogo opisaniya. Moskva: Metatekst, 1998.
14. Kisseleva K., Paiar. D. Diskursivnye slova russkogo yazyka: opyt kontekst-no-semanticheskogo opisaniya. [Discourse markers of Russian: a study in contextual-semantic description]. Moskva: Metatekst, 1998.
15. Kobozeva I.M., Zakharov L.M. Dlya chego nuzhen zvuchashchii slovar' diskursivnykh slov russkogo yazyka. [The sounding dictionary of discourse markers of Russian: what for?] Kompjuternaja Lingvistika i Intellektual'nye Tehnologii, 2004, 3 (10). Moskva: Nauka, pp. 292-297.
16. Kobozeva I.M., Zakharov L.M. «Kak mnogo v etom zvuke!..» (prosodiko-semanticheskie varianty russkogo mezhdometiya a). ["How much is in this sound!" (prosodic-semantic variants of Russian interjection a)] In Lingvisticheskaya polifoniya. Sbornik statei v chest' yubileya professora R.K. Potapovoi, 2007. Moskva: Yazyki russkikh kul'tur, pp. 609-627.
17. Kobozeva I.M., Zakharov L.M. Semantiko-prosodicheskie variatsii na temu «Da» (materialy k mul'timediinomu slovaryu russkikh diskursivnykh slov) [Semantic-prosodic variations of the theme "Da" (materials for the multimedia dictionary of Russian discourse markers)]. In Sredi nekho-zhennykhputei: Sbornik nauchnykh statei kyubileyu doktora filologicheskikh nauk, professora A.A. Kretova, 2012. Voronezh: Izd-vo VGU, pp. 52-62.
18. Kodzasov S.V. Kombinatornaya model' frazovoi prosodii. [Combinatorial model of phrasal prosody] In Kodzasov, S.V. Issledovaniya v oblasti russkoiprosodii. Moskva, 2009, pp. 13-47.
19. Kreidlin G.E. Neverbal'naya semiotika. [Non-verbal semiotics]. Moskva, 2004. P. 592.
20. Leech G. Principles of pragmatics. L.: Longman, 1983.
21. Richter N. The gestural realization of some grammatical features in Russian, In Proceedings of the 4th conference of the international society for gesture studies. Gesture: evolution, brain and linguistic structures. Frankfurt; Oder, 2010. P. 245.
И.М. Кобозева, О.О. Иванова, Л.М. Захаров
К МУЛЬТИМОДАЛЬНОМУ МОДЕЛИРОВАНИЮ ВЕРИФИКАТИВНЫХ ДИСКУРСИВНЫХ МАРКЕРОВ В РУССКОМ ДИАЛОГЕ
Федеральное государственное бюджетное образовательное учреждение высшего образования Московский государственный университет имени М.В. Ломоносова 119991, Москва, Ленинские горы, 1
Данное исследование посвящено мультимодальному анализу дискурсивных маркеров (ДМ) со значением позитивной верификации в русском языке, а именно выявлению корреляций между видом используемого ВДМ и особенностями его паравербального оформления. В работе были рассмотрены следующие ДМ — да, да-да, ну да, конечно, разумеется, естественно, которые в диалоге маркируют истинность пропозиции Р, введенной в рассмотрение в предшествующей реплике собеседника. Данные маркеры входят в состав верификативных реплик или же представляют собой са-
мостоятельные реплики, являющиеся полноценными речевыми ходами с иллокутивной функцией (ИФ) ответа на вопрос, подтверждения и т.п. В качестве материалов для исследования были использованы мини-диалоги из корпуса МУРКО, представляющие собой видеоряды с одновременной аудиодорожкой, сопровождаемые текстовой разметкой. В статье представлена аннотация ряда синтаксических, прагматических и социолингвистических признаков, статистический анализ которых позволил на уровне тенденций установить, какие из них влияют на жестовое и просодическое оформление рассматриваемых дискурсивных маркеров. Полученные выводы расширяют знания о мимико-жестикуляционном компоненте дискурсивных слов, что может способствовать дальнейшему развитию Мультимедийного корпуса русского языка, методик преподавания русского языка как иностранного или же найти применение при разработке анимированных компьютерных агентов и роботов.
Ключевые слова: мультимодальный анализ речи; диалог; дискурсивные маркеры; верификация; жест; кивок; просодия; семантика; прагматика.
Сведения об авторах: Кобозева Ирина Михайловна — доктор филологических наук, профессор кафедры теоретической и прикладной лингвистики филологического факультета МГУ имени М.В. Ломоносова (e-mail: kobozeva@list.ru); Иванова Ольга Олеговна — бакалавр, студент магистратуры при кафедре теоретической и прикладной лингвистики филологического факультета МГУ имени М.В. Ломоносова (e-mail: wiola100@rambler.ru); Захаров Леонид Михайлович — старший инженер Учебно-научного компьютерного центра филологического факультета МГУ имени М.В. Ломоносова (e-mail: leonid_zakharov@mail.ru).
^исок литературы
1. Баранов А.Н., Кобозева И.М. Модальные частицы в ответах на вопрос // Прагматика и проблемы интенсиональности. М., 1988. С. 45-69.
2. Бонно К., Кодзасов С.В. Семантическое варьирование дискурсивных слов и его влияние на линеаризацию и интонирование (на примере частиц же и ведь) // Дискурсивные слова русского языка: опыт контекстно-семантического описания. М., 1998. С. 382-443.
3. Григорьева С.А., Григорьев Н.В., Крейдлин Г.Е. Словарь языка русских жестов. Языки русской культуры, 2001.
4. Гришина Е.А. О мультимодальных кластерах в устной речи // Компьютерная лингвистика и интеллектуальные технологии: По материалам ежегодной Международной конференции «Диалог». 2001. № 10 (17).
5. Гришина Е.А. Да в русском устном диалоге // Russian Linguistics. 2011. № 35 (2). С. 169-207.
6. Гришина Е.А. Слово и жест. Корпусные исследования устной речи. Saarbrücken, 2012.
7. Гришина Е.А. Жестикуляционные профили русских приставок // Компьютерная лингвистика и интеллектуальные технологии: По материалам ежегодной Международной конференции «Диалог». 2013. № 12. С. 19.
8. Гришина Е.А. Жесты и прагматические характеристики высказывания // Мультимодальная коммуникация: теоретические и эмпирические исследования. М., 2014. С. 25—47.
9. Киселева К.Л., Пайар Д. Дискурсивные слова русского языка: опыт контекстно-семантического описания / Под ред. К. Киселевой и Д. Пайара. М., 1998.
10. Добрушина Е.Р. Верификация в современной русской диалогической речи: Дисс. ... канд. филол. наук. М., 1993.
11. Киселева К.Л. РАЗУМЕЕТСЯ, или САМИ ПОНИМАЕТЕ // Дискурсивные слова русского языка: опыт контекстно-семантического описания / Под ред. К. Киселевой и Д. Пайара. М., 1998.
12. Киселева К.Л. ЕСТЕСТВЕННО, или DE RERUM NATURAE // Дискурсивные слова русского языка: опыт контекстно-семантического описания / Под ред. К. Киселевой и Д. Пайара. М., 1998.
13. Кобозева И.М., Захаров Л.М. Для чего нужен звучащий словарь дискурсивных слов русского языка // Компьютерная лингвистика и интеллектуальные технологии: Труды международной конференции «Диалог». М., 2004. С. 292-297.
14. Кобозева И.М., Захаров Л.М. «Как много в этом звуке!..»: просодико-семантические варианты русского междометия А // Лингвистическая полифония: Сборник в честь юбилея проф. Р.К. Потаповой / Отв. ред. В.А. Виноградов. М., 2007. С. 609-627.
15. Кобозева И.М., Захаров Л.М. Семантико-просодические вариации на тему «Да» (материалы к мультимедийному словарю русских дискурсивных слов) // Среди нехоженных путей: Сборник научных статей к юбилею доктора филологических наук, профессора А.А. Кре-това. Воронеж, 2012. С. 52-62.
16. Кодзасов С.В. Комбинаторная модель фразовой просодии // Код-засов С.В. Исследования в области русской просодии. М., 2009. С. 13-47.
17. Крейдлин Г.Е. Невербальная семиотика. М., 2004.
18. Brown P., Levinson S. Politeness: some universals in language usage. Cambridge, 1987.
19. Cienki А., Müller C. Metaphor and Gesture. Gesture Studies, 3. Amsterdam; Philadelphia, 2008. P. 307.
20. Leech G. Principles of pragmatics. L., 1983
21. Richter N. The gestural realization of some grammatical features in Russian // Gesture: evolution, brain and linguistic structures. Proceedings of the 4th conference of the international society for gesture studies. Frankfurt; Oder, 2010. P. 245.
