Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.philol.msu.ru/~otipl/new/main/people/kibrik-aa/files/Communicative_intentions@Odessa_2008.pdf Äàòà èçìåíåíèÿ: Mon Mar 31 16:10:32 2008 Äàòà èíäåêñèðîâàíèÿ: Wed Jan 14 18:20:54 2009 Êîäèðîâêà:

Andrej A. Kibrik (Institute of Linguistics RAN) COMMUNICATIVE INTENTIONS AND THE STRUCTURE OF RUSSIAN TELEVISION INTERVIEW DIALOGUES1
. , , , . : ,

1. Levels of discourse structure One of the central goals of discourse analysis as a discipline is the study of discourse structure (van Dijk 1997). Understanding any natural object presupposes understanding what it's made of, what is its structure. For example, if biologists describe an animal, they would first view it from the perspective of anatomy and figure out its parts and organs (head, legs, skin, bones, blood etc.). After that, they can view their object from the perspective of physiology and understand why each part is there (legs help to move, skin protects, etc.); thus the structure receives an explanation in the functions it fulfils. Likewise, if we want to describe a particular discourse we need to be able, first, to figure out its structure, and, second, explain why each part is there. In this paper I attempt to provide an explanation of discourse structure and link the structure with underlying functional forces that shape it. In studies of discourse structure, one often distinguishes between the macro- and microstructures of discourse. Macrostructure consists of the largest discourse chunks, including immediate constituents of discourse2. Microstructure consists of the minimal units that can qualify as discourse units. There are also approaches that provide a unified framework for both macro- and microstructure of discourse, in particular theories of rhetorical relations that connect discourse units of any size (see Horowitz 1987). The

1 2

The study underlying this paper was conducted with support of the Russian Humanities Foundation grant # 08-04-00165a. The term "macrostructure" is used in this paper in its general sense, and not in the more technical sense of van Dijk (1980). Van Dijk understands macrostructure as a set of so-called macropropositions, i.e. as a summary of the text constructed by the addressee in the course of text interpretation.

1


present paper addresses issues in discourse macrostructure. Micro-units of discourse will only be briefly mentioned in section 5.3. In the discourse analysis literature, a number of terms have been proposed that designate macro-units of discourse. For dialogic discourse, such units usually refer to certain combinations of the participants' turns; cf. notions of adjacency pairs (Sachs, Schegloff, and Jefferson 1974) or minimal dialogues (Baranov and Krejdlin 1992). For monologic discourse, notions like paragraph (e.g. Longacre 1983) and episode (e.g. Tomlin 1987) have been used, as well as stages of narratives being parts of narrative schemas (e.g. Chafe 1994). In this paper I use the term "quantum" as a cover term for all macro-units of discourse. In this paper, I investigate one genre of dialogic discourse, namely TV interviews. The genre of interview has been selected because in interviews macrostructure is more clearly identifiable than in many other genres (see section 3 below), and it is always easier to start with simpler and more straightforward examples. Mutatis mutandis, the results of this study can be applied to other forms of dialogic communication. Spoken TV interviews rather than written media interviews have been selected for analysis here since in the latter it is difficult to discern the contribution of the original discourse participants from the later editorial stages. The present study addresses Russian TV interviews, but many points made here apply to other languages as well. Thus this paper focuses on the macrostructure of one discourse genre in one particular language, but the proposed analysis is intended to be generally applicable to explaining discourse structure. 2. Discourse macrostructure and the intentional structure Discourse is produced by speakers. Consequently, its structure is defined by the speaker's inner forces (Callow and Callow 1992). I will call these forces communicative intentions, or CIs. A CI is the original stimulus for the speaker to produce a discourse or its part, it is close to the folk notion of "thought". Many schools in linguistics and even
2


in discourse analysis have tried to restrict their attention to objective structure alone, and to disengage themselves from any inquiry into the level of thought or communicative intention. However, it is obvious that in order to understand the structure one needs to understand forces that shape that structure. An analogy from a more physical area is useful. In order to adequately describe a geological structure, a scientist needs to understand the tectonic processes that led to the formation of that structure. So, however elusive CIs may seem to us, if we strive for a realistic picture of discourse structure, we need to seriously inquire into the underlying communicative forces. In this paper I present the hypothesis that discourse macrostructure is a direct mapping of the intentional structure (for a number of similar approaches see Cohen, Morgan, and Pollack 1990). Discourse as a whole reflects the global communicative intention of the speakers, and macro-units of discourse reflect more local communicative intentions. Below I suggest that invisible communicative intentions can be fairly objectively studied and can be used to explain overt discourse structure. In section 3, I characterize interviews as a discourse genre. Section 4 lays out and details the proposed hypothesis of the connection between the discourse structure and the intentional structure. Section 5 is an extended analysis of a sample interview. Section 6 presents the conclusions of this study. 3. Interview as a discourse genre 3.1. Peculiarities of the genre So far, discourse analysis lacks an exhaustive typology (or classification, or calculus) of discourse genres. Biber (1989) proposed that a linguistically based exhaustive typology of genres is impossible, since genres are culturally based patterns and have no stable linguistic features; Biber proposed, instead, a typology of "text types" on the basis of objective morphosyntactic properties. Nevertheless, discourse genres are frequently identifiable, and some genres have relatively stable and delimited properties based on the pragmatics of usage. Interview is among such relatively
3


delimited genres; for a detailed study of this genre see Jucker (1986). For the purposes of this paper, interview can be characterized by the following properties: · an interview inherently implies three roles: interviewer (Ir), respondent (Rt) (two interlocutor roles), and presupposed audience; the Ir asks question on behalf of the audience, the Rt answers them · questions raised by Ir must be of interest or relevance to the presupposed audience · Ir is the major shaper of the interview's form; it's his/her CIs that largely predetermine the course of an interview and its structure; Rt's CIs are mostly trivial: to supply the information requested by the Ir; below I assume that it is the Ir's CI alone that is responsible for the interview structure, and "CI" will mean "the Ir's CI" · interviews typically have a very hierarchical and well-organized structure (as compared to other genres of dialogues) 3.2. Russian TV interviews In this paper, we deal only with TV interviews recorded on Russian TV in the late Soviet years (turn of the 1980s/90s). These interviews have a number of features that make them simpler among the wider gamut of possible discourses belonging to the genre of "interview". First, these are spoken dialogues, unlike e.g. printed newspaper interviews. Spoken discourse is a more spontaneous and more basic form of verbal interaction since it involves fewer complications typical of written language (such as post-editing, polishing, etc.). Second, TV interviews are a case of face-to-face communication between Ir and Rt, unlike e.g. telephone interviews sometimes broadcast on the radio. Again, face-to-face communication is a more fundamental variety of discourse compared to interaction between spatially remote individuals that emerged only recently with development of technology. Third, interviews in question were all conducted in Russian in the late Soviet years. The time and country of the discussed interviews is crucial in one respect: Russian interviews of the late Soviet era are
4


primarily informational, that is, are oriented toward information retrieval from the respondent; this notion is elaborated in 3.3. Below the term "interview" will be used in a restrictive sense, that is, only interviews with the listed features will be considered. 3.3. Internal typology of interviews An internal classification of interviews can be based on several different parameters, including the following. · number of Irs: the prototypical number is 1; when there are several or multiple Irs the dialogue drifts to another genre, namely press conference · number of Rts: the prototypical number is 1; when there are several or multiple Rts the dialogue drifts to another genre, namely sociological interview or poll · relative "importance" of the Ir and the Rt: when Rt >> Ir in terms of social weight, one observes the "Soviet boss syndrome", that is, the Rt takes extra long turns and forwards his own message instead of responding the Ir's questions · the character of the global CI: retrieval of information vs. other; this latter parameter requires a longer commentary. During most of the Soviet era, no real interviews were held on television. Since all areas of public life were kept under the close control of the Communist party, very little spontaneous behavior could be allowed in the media. So even if an apparent interview took place, the whole of its content would be prearranged, and an interview could even be rehearsed, to make sure that nothing unpredictable is said by the Rt. In the late 1980s the social setting dramatically changed. Much of what had been banned before became quite possible, including spontaneous interviews. Since much information about many spheres of life were closed to the public before, there was sharp public interest to many issues. Watching TV (as well as reading newspapers) was a kind of obsession in Russia at that time since everything was completely novel and extremely interesting. The journalists working in the media at that time rediscovered interview as a genre, and
5


employed its form quite straightforwardly, in a fashion that can be called informational. That means that the overall goal of an Ir in a typical interview of the turn of 1980s/90s would be to retrieve some information from the Rt that would be highly interesting to the presupposed audience. That differs from the most common kind of an interview in the Western media, with its main purpose of disclosing or attacking the Rt's "face" (see Jucker 1986) (like for example talking to a politician running for an office and trying to reveal some black sides of his prior career; cf. interviews held on such American TV shows as Oprah Winfrey, Jenny Jones, Geraldo, or Jerry Springer). In this latter case, one can talk about an evaluational intent of the interview, in contrast to an informational intent. It could be mentioned that nowadays common Russian TV interviews are somewhere mid-way between the informational type described for the end of the Soviet era, and the evaluational and confrontational Western interviews. some Russian TV journalists mask an evaluational interview under the overt scenario of an informational interview. Thus Russian Irs of the late Soviet years, while conducting interviews, were led by a rather conscientious intent to retrieve some propositional (ideational) information from the Rt that the latter possessed while the audience did not. Below we are dealing with this type of informational interview. Of course, even in an informational interview there may be evaluational elements; this point will be developed in section 5.7. Informational interviews can be further classified depending on the kind of information related to the Rt and retrieved from the Rt in the dialogue. Obviously, the range of specific information types is open but the main distinctions are summarized in Fig. 1. The Ir may be interested in some information that Rt uniquely possesses (for example, having visited an exotic and remote country where few have been); otherwise the Ir is interested in some information about the Rt him/herself. In this latter case, the reason why the Rt is of interest can be twofold: s/he may be interesting as a member of a class (for example, representative of a profession poorly known to the broad public), or as an individual. Interesting individual properties can be very different, for example, a
6


Rt may have some unusual abilities (e.g. mnemonic), or interesting biography (e.g. a traveler), or perform some political activities of high public interest, etc.
information (a) accessible to Rt about Rt as an individual, in particular, his/her: (c) political activities ...............

(b) as a representative of a class (c) personal properties

(b) biography

Fig. 1: Types of information of central interest/relevance in an interview 4. Knowledge frames, dynamics of communicative intentions, and the interview structure 4.1. The basic frame and the global CI In an informational interview, the main communicative intention of the Ir boils down to filling the gaps in his/her (and the audience's) knowledge base with the help of the Rt. Therefore, in order to understand the CI structure one needs to understand the underlying knowledge representation in the Ir. Remember that interesting/relevant information retrievable from the Rt can be very different. But in any case there is some set of data related to the Rt that the Ir originally has, and there are some elements missing that the Ir needs to complement his/her knowledge base. These missing elements predetermine the specific CIs of the Ir in the course of the interview. I will call the fragment of encyclopedic knowledge related to the Rt, the basic frame. The term "frame" was introduced into the analysis of knowledge and language by Minsky (1975). Frames are sets of knowledge associated with a particular fragment of reality. For example, a stereotypical frame of an apartment contains such elements as entrance, kitchen, bedroom, bathroom, etc. Many of us have a frame of our neighbour who, for instance, lives next door, has a cat, goes for a walk every morning, etc. Frames play a crucial role in the human information processing. Any incoming information
7


about a fragment of reality can be processed only vis-a-vis the already existing frame of that fragment. Examples of basic frames are: the Rt's experience as a cosmonaut; the Rt's professional life; the Rt's political program etc. etc. In order for an interview to take place, the basic frame related to the Rt should not be entirely empty. The global CI underlying the interview as a whole can be generalized in the following way: · to fill particular gaps in the Ir's basic frame related to the Rt This formulation is most general (for see Kibrik 1991 details), and can be specified in accordance with the nature of information being retrieved from the Rt. For example, frequently the Ir does not have a specific frame related to the Rt in advance, but rather has a stereotypical frame, which must be mapped onto a specific frame in the course of the interaction. One such example will be discussed in detail in section 5 below. 4.2. Local CIs and discourse structure The global CI is broken down into local CIs. Local CIs correspond to particular gaps in the Ir's basic frame, and thus are deducible from the global CI. The dynamics of CIs in discourse can be represented by a tree like in Fig. 2.
Global CI Local CIs: CI1 CI2 CIi1 ...... CIi2 CIi ..... CIij ...... ..... CIm CIin

Fig. 2: The dynamics of CI deduction in discourse More than one level of local CIs can be distinguished, as represented in Fig. 2: a local CI immediately deducible from the global CI (CIi) can give rise to a series of lower-level CIs. In principle more that two such levels may be relevant, but in the discussion below we will not need more than two hierarchical levels of local CIs.
8


Local CIs are directly mapped onto discourse structure. To each local CI a portion of discourse corresponds, that I call a quantum. "Quantum" is meant to be the basic notion in the realm of discourse macrostructure, and a cover term for such concepts as "paragraph", "adjacency pair", and the like. In interviews, a quantum contains at least a pair of the interlocutors' turns: a question by the Ir plus a reply by the Rt. After the Ir's local CI is satisfied, the corresponding quantum ends, and the Ir proceeds with the next local CI and the next quantum. Frequently it takes more than two dialogic cues to complete a quantum. Thus the sequence of the Ir's questions in an actual interview finds its explanation in the invisible, but powerful communicative intentions and knowledge structures. 4.3. The triad "knowledge representation -- CIs -- discourse structure" Let us take an example. In late 1980s, the first free (or quasi-free) parliamentary elections took place in the Soviet Union. (In the "classical" Soviet years the elections were purely fictitious since there was always only one candidate on the ballot.) Of course, a chance to choose the government for the first time was extremely fascinating to many people, and the public interest to the issues of election was very high. During and after the first campaigns, interviews with candidates and newly elected congress members were very popular in the media. While the business of democratic elections was new to the public, the general understanding of the basic procedure was already there. When an interviewer conducted an interview with a newly elected congress member (deputy), s/he had in mind the basic frame
(1) NEW DEPUTY a. b. c. Campaign Sphere of interest (or expertise) Proposed program
NEW DEPUTY

of the following design.

9


Slot (1a) of the basic frame can be further split into lower-level slots:
(1a) Campaign of the new deputy a1. a2. a3. location competitors scenario

This kind of knowledge representation was consistently reproduced in multiple interviews with elected deputies in late 1980s. An Ir, taking an interview from deputy X, would go through all slots in frame (1), including subframe (1a), and ask corresponding questions. After having received answers, the Ir would proceed with the following slot of the frame. Thus we can clearly see the triad "knowledge representation -- CIs -- discourse quanta" and the ways how knowledge representations are ultimately mapped onto discourse structure through the mediation of CIs. 4.4. Against circularity This approach should be used carefully in order to avoid a threat of circularity. Knowledge frames should be verified independently of the interview in question, otherwise for each interview a trivial "underlying" knowledge frame can be constructed ad hoc on the basis of the overt structure. The procedure I have been using is the following. From an interview as a whole, a discourse analyst gets a feeling of what is the global CI. After that, people belonging to the same cultural-linguistic group as the Ir (that is, appropriate representatives of the Ir's audience) are questioned on what are the important and interesting pieces of information needed to satisfy the global CI. All such pieces of information are incorporated into the hypothetical basic frame that supposedly was in the Ir's mind while s/he was planning the interview. Afterwards, the actual interview with local CIs, as displayed by the interview quanta, is compared to the
10


independently constructed basic frame. If there is a match between them (which is normally the case) it can be inferred, first, that the constructed basic frame coincides with the one that the actual Ir had in mind, and, second, that the whole model is working. 4.5. Spontaneous local CIs There is one complication in discourse structure and in the system of local CIs not mentioned above. The local CIs like those discussed above are planned, or deducible, from the global CI. In the course of interaction with the Rt, the Ir occasionally encounters pieces of information that are unexpected, puzzling, worldview-changing, inconsistent, contradictory, or otherwise disturbing. In reaction to such information Irs typically pose questions that are in no way deducible from the global CI. Such questions represent local CIs that are called spontaneous. Quanta resulting from spontaneous local CIs are linearly nested, or embedded, inside the planned, or deducible, quanta. After the Ir adapts the disturbing information, s/he resumes the planned local CI that was in the queue or interrupted at the time of digression. Some "worse" Irs may let the Rt to put them on the side track, and never return to the interrupted local CI, and, as a result, fail with their global CI. All components of the proposed model will be illustrated in detail in section 5 by the examples of one particular TV interview. 5. Example: dialogue with the speculator 5.1. Preliminaries The interview we are going to analyze in detail was recorded on June 14, 1989, from the Russian TV program "Legal channel". The interview was taken during the pending trial of a person who had been detained and accused for "speculation". In the Soviet legal terminology, speculation was essentially a synonym of free trade, and was prosecuted by law. Speculators would buy goods that were in short supply on the state market with its fixed prices, and then sell them on the black market at higher prices.
11


Many goods were sold only by speculators and could not be found on the legal market. An interview with a speculator on TV in 1989 was potentially interesting, because the majority of the population was not immediately familiar with the lifestyle of that professional group (although everybody would know that it existed). The reason for that lack of familiarity was that in the Soviet period the state policy was to conceal all "negative phenomena", and the existence of "speculators" was considered one of such. This interview elicits information about the Rt as a representative of a group (type (b) in Fig. 1 above). It is based on a generalized frame
GROUP LIFESTYLE OF A PROFESSIONAL

. For many professional groups, anyone belonging to the given language and

culture, has a corresponding concrete frame, but for the profession of speculator the Russian public of the late 1980s did not know many details of such concrete frame. Thus the global CI in this interview was to map the generalized frame
PROFESSIONAL GROUP LIFESTYLE OF A

onto a concrete frame

LIFESTYLE OF A SPECULATOR

.

In section 5.2, a transcript of the interview is provided. Section 5.3 contains a commentary on the principles of transcribing. (Note that transcribing spoken discourse is an enterprise that is far from elementary and straightforward; there is now a whole subdiscipline in discourse analysis developing consistent principles of discourse transcription, see e.g. Baker 1997.) The procedure of revealing the generalized frame, as well as some comments on culturally obscure points in the interview will be given in section 5.4. Sections 5.5 and 5.6 contain a discussion of the interview structure and its explanations. Section 5.7 introduces an additional layer of the intentional structure. 5.2. Transcript
1 1 11 2 3 Rt: Ir: Rt: ja spekuljant I'am a speculator (2) ty tak s gordost'ju Õto govoris' You say that with such pride nu ja gospodi radujus' ßto ja ne slesar' Well my Lord* I am glad that I am not a plumber

12


4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

Ir: Rt:

111 20 Ir: 21 22 23 24 Rt: 25 26

(1) ponjatno I see (1) ljudi rabotajusßie na zavode people working in a factory (1) vot OK i zarabatyvajusßie tam sto pjat'desjat dvesti rublej and earning there 150 to 200 rubles (0.5) oni ne v sostojanii pokupat' ( 1) vesßi they are not in a position to buy things ( 1) kotorymi ( 1) torguem my that we sell (1) vot OK (1) ponimaete** you understand? (0.5) Õto im prosto ne po karmanu i poÕtomu that's simply beyond their capacity and so (m 1) v sferu ( 1) moej dejatel'nosti vxodjat ljudi bogatye my area of activites includes rich people (0.5) finansovaja nezavisimost' podrazumevaet moral'nuju financial independence implies a moral one (0.5) vot OK (0.5) a esli (m .5) on moral'no nezavisim and if one is morally independent tak zaßem emu (m 1) vsevozmoÚnye prizyvy gospodi then what for would he need various slogans my Lord (1) on i sam (0.5) prekrasno (0.5) moÚet he himself can perfectly well (0.5) podumat' ßto emu nuÚno a ßto net ponder on what he needs and what he does not (0.5) sudja po vasim slovam judging by your words (0.3) vy v obsßem you in general (0.3) tak so (0.3) stremites' (0.3) k nekoj svobode lißnoj strive for certain personal freedom (0.5) da koneßno yes sure (0.5) a ja sßitaju ßto and I believe that (0.3) ljuboj normal'nyj ßelovek dolÚen stremit'sja k lißnoj svobode any normal person should strive for personal freedom

13


27 28 29 111 11 1 2 30 Ir: 31 32 33 34 35 36 37 Rt: 38 39 40 41 42 43 44 45 46 47 48
14

(0.5) potomu ßto kak skazal (0.3) linkol'n because as Lincoln said (0.3) xuÚe rabstva moÚet byt' tol'ko to worse than slavery can be only that (0.5) ßto kogda ßelovek znaet ßto on rab i ne xoßet stat' svobodnym a person knows that he is a slave and does not want to be free

(.3) xoroso all right a vot kak vy and how do you (m 0.3) nu well (0.3) svobodnoe vremja (0.3) provodite? spend your free time? ßto ljubite? what do you like? (0.3) ßto vam daet Õti (= 0.5) vot den'gi how do you take advantage of this money kotorye vy vot takim obrazom zarabatyvaete? that you earn in such a way (1) Õto mne daet gospodi vozmoÚnost' gospodi that gives me my Lord a chance my Lord pojti kupit' bilet v teatr to go and buy a ticket to a theatre (0.5) nu zaplatit' za nego tri ceny perhaps to pay a triple price for it (1) esli uÚ na to poslo i kupit' ego if that is not avoidable and to buy it i pojti posmotret' ßto and to go and see what (0.5) =ne dano (0.3) drugomu is not given to others (0.5) Õto daet mne vozmoÚnost' gospodi that gives me a chance my Lord (0.5) poest' po-ßeloveßeski ne toj kolbasy kotoruju to eat in a human way something besides the sausage that (1) daÚe koski ne edjat even cats do not eat (0.5) Õto mne daet vozmoÚnost' xot' that gives me a chance at least (0.3) odet'sja bolee menee prilißno to dress myself more or less decently (1) vot OK


49 50 51 52 53 54 55 56 2 3 57 Ir: 58 59 60 Rt:

61 Ir: 62 63 64 65 66 67 68 Rt: 69 70

(0.5) Õto mne daet vozmoÚnost' gospodi that gives me a chance my Lord (0.3) obsßat'sja (0.3) s bolee menee gospodi to socialize with more or less my Lord (1) kak by vam skazat' how to put it (1) ßtoby ne naxamit' tam not to be boorish (0.5) prijatnymi Úensßinami pleasant women (0.5) a ostal'nogo and in the rest (m 0.5) nißego mne Õto osobogo ne daet that gives me nothing special (0.5) osobennoj svobody no special freedom (1) tak OK a ßto dal'se? and what's next? (0.3) nu skol'ko Õto moÚet dlit'sja? I mean how long that can last? (2.5)(m .5) poka v oßerednoj raz (0.5) u kogo-to ne pojavitsja Úelanie posadit' until next time someone gets a will to imprison me (1.5) ne nu vse-taki no but still (m 0.5) gody idut years pass (0.5) uÚe n' || already n (= 0.5) tak skazat' molodost' proxodit so to speak the youth passes a vot dal'se dal'se ßto? but next, what's next? vot kak dal'se Úit'? how are you going to live further? (2) vy kopite ßto li na (0.3) ßernyj den'? do you put by for a rainy day? nikogda never (1) a zaßem? what for? (1) dlja togo ßtob u menja prisli Õto otnjali? to let someone come and seize it?

15


31

71 Ir: 72 Rt: 73

31 3

(3.5) to est' budet den' budet pisßa ja tak Õto|| that means there will be a day there will be food this is how I da yes (= 1) praktißeski gospodi (0.5) vse Úivut odnim (= 0.3) dnem in practice my Lord everybody lives this single day

5.3. Commentary on the transcript The discourse in question has a certain macrostructure that is marked by means of angular signs (, ) on the left side of the page. A discussion of the macrostructure is postponed till sections 5.5, 5.6. The transcript consists of lines, or, more precisely, pairs of lines. In each pair of lines the first line (in italics) is the original Russian text, and the second line is a free translation into English, to the extent possible imitating the semantic peculiarities of the original. No word-by-word interlinear translation was attempted here since it is not essential for the topic of this paper, and it would significantly complicate and lengthen the transcript. Each pair of lines is numbered with a small boldface number. There are 73 line pairs in the transcript. Lines of Russian text in the transcript correspond to elementary discourse units, coinciding with what Chafe (1994) calls intonation units. According to Chafe, discourse is produced in spurts, and such spurts can be defined prosodically (by pauses and intonation contours), cognitively (they correspond to one "focus of consciousness") and in terms of their informational content (a prototypical intonation unit is a clause). Since intonation units are the microstructure of discourse, they are not further discussed in this paper, but are simply accepted as given. The column after the line number column contains the designation of the speaker: either Ir or Rt. Of course, Ir and Rt are marked only at the beginning of their turns. The column after the speaker designation contains a digit in parentheses. This is the length of a pause preceding the current intonation unit. No refined pause length
16


measurements were necessary for this study, and only several degrees of length are distinguished. Extra short pauses are marked as 0.3 second. Longer pauses were rounded to half second (0.5, 1, 1.5, 2, etc.). A minority of intonation units is not separated from the preceding unit by a pause; in such cases no pause is marked. Some intonation units have a pause inside. Both unit-initial and unit-medial pauses can be absolute or filled with a certain phonetic signal. If a pause is filled, the nature of the filling signal is marked in front of the pause length; two types of filling is distinguished: non-nasalized shwa () and labial nasalized sonorant (m). In addition, there are some pauses that are filled with the last phonetic segment of the preceding word; this kind of filling is marked with the = sign in front of the length number. The = sign is also used in front of a phoneme that is drawled by a speaker. The || symbol indicates a truncation of false start. Regular scholarly transliteration from Cyrillic to Roman is used in the rendering of the Russian text. Punctuation is used sparingly, and no capitalization is used. The question of sentence boundaries is not essential for this study, and the corresponding decisions have been avoided. The asterisksmarks those spots in the transcript that require some special comment: * in line 3 and many times after that the Rt uses the Russian expression gospodi `my Lord' in contexts which are highly atypical in colloquial Russian discourse for this expression; my hypothesis is that the Rt was substituting "my Lord" for cursing that he found unacceptable in front of the camera; ** in line 11, the Russian word ponimaete? `do you understand?' is rendered in the record in a highly reduced form, approximately [pit']. 5.4. Revealing the generalized frame In order to obtain independent evidence on the possible basic frame underlying this interview, I have been asking several groups of students to construct such frame on

17


a purely deductive basis3. The students were told the following: "Suppose you are going to interview a representative of a profession of which you know nothing. What would be your questions?" The given replies, if summarized and somewhat reformulated, provide the following generalized frame.
(2) LIFESTYLE OF A PROFESSIONAL GROUP a. b. c. d. The nature of professional activity Income Life conditions Social security

In fact, these four large rubrics are a generalization of sets of more specific questions indicated by the students as essential. In other words, immediate slots of frame (2) have an internal structure and are frames themselves, which is particularly important for (2c).
(2c) Life conditions i. Consumption food clothing ii. Housing iii. Leisure socialization hobby

Some cultural comment is in order here explaining why certain slots of frame (2) should indeed be there. In 1989, when the interview was taken, the Soviet system was still in place, with its shortage of the most basic commodities and difficult access to

3

This part of the study was conducted in the course of my classes in Discourse analysis at the Linguistics department of Moscow State University. Over 100 students attended that class altogether and thus took part in the study. I express my gratitude to all of them.

18


goods and services. Therefore, the issues in consumption and other "life conditions" were not a mere function of a person's income. Different professional groups had better chances to get different commodities. For example, people working in construction industry frequently could get apartments quicker than others, and people working in food stores had easier access to food products. It is for these reasons that information about life conditions is a necessary part of the generalized frame
PROFESSIONAL GROUP LIFESTYLE OF A

.

As for the role of the journalist (the Ir), even though there was much more freedom in the media than, say, in early 1980s, in 1989 the communist party control was still there, and a journalist was supposed to at least coordinate his/her professional performance with the standards imposed by the communist rule. In particular, the Ir was supposed to act as a token member of the Soviet society, in accordance with the views that a Soviet citizen was officially expected to believe. 5.5. Discourse quanta based on deducible local CIs The global CI of the Ir in this interview is to map the generalized frame
OF A PROFESSIONAL GROUP LIFESTYLE

onto a concrete frame

LIFESTYLE OF A SPECULATOR

. Slot (2a)

of the generalized frame ("the nature of professional activity") is familiar to the Ir and the implied audience, since all speakers of Russian in 1989 knew what kind of activity speculators were involved in. Slot (2b) ("income") cannot be a subject of a discussion since it is taboo: if the Rt discloses information about his profits in front of the camera that could be used as evidence against him. Slots (2c) and (2d) generate two local CIs that are amply reflected in the text of the interview. Slot (2c) is responsible for the Ir's local CI to find out about the Rt's life conditions. The corresponding question of the Ir is formulated in lines 30 through 36. Literally, the Ir only touches a part of one subslot (2c-iii) "leisure" in his question. But a remarkable thing about this interview is that the Rt uses this cue to pull out most of the larger slot (2c) "life conditions" -- specifically, subslot (2c-i) "consumption" and
19


subslot (2c-iii) "leisure". He gives the Ir a much fuller report of the advantages acquired due to his profession than was actually requested. He sheds light on subslot "hobby" in lines 37-42; on subslot "food" in lines 43-45; on subslot "clothing" in lines 46-48; and on subslot "socialization" in lines 49-53. Interestingly, the Rt connects his better access to commodities with his higher income which is per se not quite typical of a Soviet citizen. Also, it is not accidental that the Rt does not mention housing ((2c-ii), the only remaining subslot): in the Soviet system even being rich did not guarantee better housing since there was no black market of real estate. The local CI related to slot (2c) is thus directly reflected in discourse structure as the line sequence 30-56. This is an example of a quantum, as introduced in section 4.2 above. In the transcript of the interview, quanta are marked on the left side of a page by means of angular signs: the sign marks the beginning of a new quantum, and the sign marks its end. Angular signs are followed by numbers. The same number is indicated at the beginning and at the end of a quantum. For example, the quantum embracing lines 30-56 and rendering the local CI related to slot (2c) of the basic frame has number 2. (Quantum number 1 will be discussed in section 5.6 below). Slot (2d) of the generalized frame is responsible for the Ir's local CI to find out about the Rt's social security. In the Soviet system social security was again partly dependent on professional affiliation, since some groups had better medical services, greater pensions, etc. than others. The Ir tries to ask a question about life prospects of the Rt in lines 57-59. As we will see below, this is an attempt to raise the issue of social security but an unsuccessful one. The Rt first cannot understand the question: the pause in front of line 60 is extremely long: it last for 3 seconds. Then the Rt chooses to interpret the question literally and provides a reply in line 60. In line 61, the Ir demonstrates that he is not satisfied and attempts another formulation of his question in lines 62-66. This time he comes much closer to an adequate formulation of the social security question. But the Rt fails to understand the question again: this is proven by another extra long pause after line 66. The reason why the Rt cannot understand the
20


social security question, perfectly natural for an average Soviet citizen represented by the Ir, is the following. The Rt already has the concrete frame
SPECULATOR LIFESTYLE OF A

, and there is no place for a social security slot there. Speculators are

outlaws and therefore they have no social guarantees. For a long time the Rt cannot make sense out of the Ir's questions. The Ir correctly interprets the long pause after line 66 as a failure to reply, and provides the third formulation of his question in line 67. This very specific formulation, finally, finds a clear response from the speculator in lines 68-70. Thus in this case there are not two but four, or even five (if 67 is separate from 61-66) turns in one discourse quantum. So far, we have inspected all slots of the generalized frame for their rendering in discourse structure. However, close to one half of the whole text has not been yet explained in terms of its underlying intentional function. 5.6. Quanta embedding As was pointed out in section 4.5, there is an important type of local CIs: spontaneous CIs. Remember that CIs are mapped onto quanta in discourse structure. As has been analyzed in detail above, main discourse quanta correspond to planned CIs. But what happens when the Ir gets a spontaneous CI induced not by his original global CI but by some disturbing information just received from the Rt? One could imagine that in such situation the whole interview structure would be broken, and the Ir would be completely off the track predetermined by his/her original global CI. But that does not happen: a normal Ir makes a temporary digression in which s/he realizes the spontaneous CI, and then resumes his/her order of actions predicted by the global CI. Such a digression is called here "quantum embedding" because the overt portion of discourse devoted to the spontaneous CI is embraced by the material devoted to the planned CIs. One example of quantum embedding is found at the very end of the interview. In lines 68-70 the Rt makes an explicit statement that social security is simply not in his frame of his own professional "career". This statement is very odd and surprising to the
21


Ir, and this is a typical situation in which a spontaneous local CI emerges. The resulting local CI is to verify the disturbing information, and it is reflected in the question formulated in line 71. By this time, the Rt has recovered from his earlier confusion (see end of previous section) and replies in lines 72-73 without a pause and even interrupting the Ir's question. The pair of turns in 71-73 is thus a separate discourse quantum resulting from a spontaneous CI. Since the stimulus for this spontaneous CI is found inside quantum 3, it is natural to assume that the quantum in 71-73 is embedded inside quantum 3. Embedded quanta are named by two-digit numbers, the first digit being the number of the embracing quantum, and the second digit being the ordinal number of the embedded quantum inside the embracing quantum. Thus the embedded quantum in lines 71-73 has number 31, as marked in the transcript. There can be two (or more) levels of embedding, and in that case embedded quanta are marked by three- (or more than three-) digit sequences. A more complex example of embedding than with quantum 31 is found in the first part of the interview preceding quantum 2. The Rt's turn in line 1 is a reaction to an Ir's question that was not recorded but meant something like "Why have you been imprisoned?" or "In what area do you work?" Such question represents the zero phase of an interview that can be called "establishing a contact"; establishing a contact is always necessary before an Ir can proceed with the realization of his/her planned CIs. Thus line 1 belongs to quantum 1 of the interview; quantum 1 started with the unrecorded Ir's question, and could have ended with the Rt's turn. However, it did not. In reaction to his question, the Ir probably expected an evasive reply of approximately the following content: "I sold one jacket, nothing special, and for some reason they have arrested me". But the actual reply found in line 1 is very straightforward and assertive: the Rt states that he is a speculator and does not even attempt to mask his professional affiliation. In the Soviet setting, this kind of an assertive statement of one's profession could be expected from a metal industry worker, or a military serviceman, or another group with high "reputation" in the Soviet ideology,
22


but not at all from a speculator. Hence the long pause of 2 seconds after line 1, and the Ir's spontaneous intention to verify such an unusual attitude; that intention is realized in line 2. Thus, line 2 is the beginning of the embedded quantum 11. As any normal interview quantum, 11 starts with an Ir's turn. Line 3 is the Rt's reaction; it sounds both ironic and defiant and demonstrates that the Rt is explicitly at odds with the official Soviet ideology (because apparently he is not respectful of the working class). At this point the Ir is taken aback and loses his initiative as the only participant of the interview who has independent CIs. He does not start a new quantum but simply accepts the Rt's point; the Ir's "I see" in line 4 sounds as helpless irony. The Rt, quite the contrary, takes the liberty to substantiate his position in lines 5-19, and thus continues quantum 11. The dynamics of ideas in that extract is quite interesting but its details are beyond the scope of this paper. The crucial point for us here is that in that extract the Rt again makes statements that are disturbing to the Ir: specifically, the Rt explicitly mentions personal freedom as having a high position in his system of values. The Ir, faithfully playing the role of a token representative of the Soviet society, gets another spontaneous CI: to test that foreign value system. Displaying his limited familiarity with the value of personal freedom, the Ir hesitantly formulates the question in lines 20-23. Thus a second-level embedding takes place, and quantum 111 starts. In lines 24-29, the Rt provides his reply, quite confident and apparently pre-rehearsed. After that the embedded quantum 111 ends, as well as the embracing quantum 11 and the highest level quantum 1. When quantum embedding takes place, a "good" speaker, after having finished with the spontaneous CI goes back to the interrupted deducible CI, and continues working on it. In terms of discourse macrostructure, this is represented, as a general rule, as an embedded quantum surrounded on both sides by the material of the planned quantum. In the particular example we have just analyzed, both the embedded and the embracing quanta end simultaneously.
23


5.7. Informational and evaluational intentions Thus the discussion of the interview macrostructure and its connection to the underlying intentional structure and to the knowledge representation is over. One additional point needs to be made here. As has been pointed out above, even in an informational interview such as the analyzed one, there is a layer in the intentional structure that is not exactly reducible to the intention to fill the gaps in the knowledge base. Such layer, distinct from the purely informational component, can be called evaluational. In the analyzed interview, both the Ir and the Rt are willing to evaluate the elements of the basic frame they are discussing. To put it simply, the Ir evaluates the elements of the basic frame
LIFESTYLE OF A SPECULATOR

negatively, while the Rt

evaluates them positively. In this particular example, as well as most interviews of the late Soviet era, evaluations are overbuilt on top of the informational elements. But in other kinds of interviews the evaluational component may be as important as the informational one. The interaction between the informational and evaluation intentions in interviews is one of the directions of the future studies. 6. Conclusion In this paper, I attempted to demonstrate that discourse macrostructure cannot be explained without reference to communicative intentions of discourse participants. Communicative intentions, as the speakers dynamically unfold them in time, shape the overt discourse form and predetermine the discourse macrostructure. In order to see this connection more clearly, a relatively regulated discourse genre was selected, namely the interview. Interviews unlike e.g. ordinary conversations are essentially controlled by only one participant, namely the interviewer. Consequently, normally there is little or no conflict between separate sets of communicative intentions, and a connection between the intentional structure and the discourse structure can be more clearly seen.

24


Discourse macrostructure consists of discourse chunks that I propose to call quanta. Each quantum can be attributed an underlying local communicative intention. Most local intentions are realizations of the global intention that is the initial stimulus for the interviewer to enter interaction. The global communicative intention, in the case of an interview, amounts to constructing a certain knowledge representation connected to the respondent: the basic frame. The setup of the basic frame can be verified independently of particular interviews. For the interviewer, the basic frame is the basis for breaking the global communicative intentions into the local ones. Therefore, such local intentions can be called deducible, or pre-determined. Another kind of local communicative intentions are spontaneous. They cannot be deduced from the global intention but rather emerge in an ad-hoc manner. When the respondent provides an information that is in some way disturbing to the interviewer, the latter gets a spontaneous intention for checkup, verification or clarification. Spontaneous local intentions result in quantum embedding: after realizing the spontaneous intention, the interviewer returns to the point of the interrupt. The present model is intended as an explanation of the structure of interviews, but also of a broader range of discourses. As was pointed out above, interview was selected for analysis because it is a simple genre in the sense that discourse structure can be explained by the CIs of one discourse participant. Other dialogic genres can be analyzed by the same basic model, but a greater number of complicating factors should be taken into account. In particular, the problem of interacting (and often contradictory) CIs of two or more discourse participants should be addressed. When different participants have their separate CIs, the resulting discourse structure is a result of complex negotiation and compromise. Such complicated issued remain for future research.

25


REFERENCES Baker, Carolyn D. 1997. Transcription and representation in literacy research. In: J. Flood, S.B.Heath, and D. Lapp (eds.) Handbook of research in teaching. Literacy through the communicative and visual arts., p. 110­120. Newark, Delaware: International Reading Assn. Baranov, Anatolij A., and Grigorij E. Krejdlin. 1992. Illokutivnoe vynuzhdenie v strukture dialoga. [Illocutionary causation in dialogue structure.] Voprosy jazykoznanija, N. 2, p. 84­99. Biber, Douglas. 1989. A typology of English texts. Linguistics, vol. 27, p. 3­43. Callow, Kathleen, and John C. Callow. 1992. Text as purposive communication: A meaning-based analysis. In: W.C.Mann and S.A.Thompson (eds.) Discourse description: Diverse linguistics analysis of a fund-raising text, p. 5­38. Amsterdam: Benjamins. Chafe, Wallace. 1994. Discourse, consciousness, and time. Chicago: University of Chicago Press. Cohen, P.R., J.L.Morgan, and M.E.Pollack (eds.) 1990. Intentions in communication. Cambridge: MIT Press. Horowitz, Rosalind. 1987. Rhetorical structure in discourse processing. In: R. Horowitz and S.J.Samuels (eds.) Comprehending oral and written language, p. 117­159. San Diego and London: Academic Press. Jucker, Andreas H. 1986. News interviews: A pragmalinguistic analysis. Amsterdam: Benjamins. Kibrik, Andrej A. 1991. O nekotoryx vidax znanij v modeli estestvennogo dialoga. [On certain types of knowledge in a model of natural dialogue.] Voprosy jazykoznanija. N 1, p. 61-68. Longacre, Robert. 1983. The grammar of discourse. New York: Plenum Press. Minsky, Marvin. 1975. A framework for representing knowledge. In: P.H.Winston (ed.) The psychology of computer vision, p. 211­280. New York: McGraw-Hill. Sacks, Harvey, Emanuel Schegloff, and Gail Jefferson. 1974. A simplest systematics for the organization of turn-taking for conversation. Language, vol. 50, p. 696­735. Tomlin, Russell. 1987. Linguistic reflections of cognitive events. In: R.Tomlin (ed.) Coherence and grounding in discourse, p. 455­480. Amsterdam: Benjamins. van Dijk, Teun A. 1980. Macrostructures. Hillsdale, NJ: Erlbaum. van Dijk, Teun A. 1997. The study of discourse. In: T. van Dijk (ed.) Discourse as structure and process. Vol. 1 of: Discourse studies: A multidisciplinary introduction, p. 1­34. London: Sage.

26