Marja Etelämäki Department of Finnish, University of Helsinki
The paper deals with making demonstrative reference in Finnish conversation. The aim of the paper is to seek for the difference between the effects on reference caused by a) the basic meaning of the pronoun itself, and b) the sequential position and the construction of the turn and the position of the referential pronoun within the turn. The demonstrative expresses the relation of the referent to the actional context (Hanks 1990, 1992) while the construction of the turn and the position of, the pronoun within the turn are ways for speakers to express the information status of the referent in a given situation. The pronoun tämä (cf. "this") expresses immediate closeness of the referent to the context. It expresses that the interpretation of the pronoun's referent is crucial in interpreting the on-going context thus marking a change in both. The pronouns "tuo" and "se" (cf. "that") in turn are used in referring to given or accessible referents. They can be used even in referring to concretely new referents. In such situations the referent is introduced as a part of an on-going actional or sequential whole and it is also identifiable in it. (See also Laury 1997, Seppänen 1998.)
The dynamic, actional context of the speech situation has been shown to be fundamentally sociosentric (Hanks 1990). The identifiability of the referent is always based on a shared understandingof the on-going action, and the choice of the demonstrative has its effect on this action and its interpretation in the on-going situation. The referents introduced as new with the pronoun "tämä" are central to the interpretation of the on-going action, and such referents are introduced in turns which have an effect of starting a new action. The referent itself does not have to be new. It is the interpretation of the referent in a speech situation that makes a referent's status as new or given.
The construction of conversational discourse is based on how turns-in-interaction are connected to each other actionally. The "reading" of a turn - the interpretation of a turn's referential meaning - is constructed on the actional basis of the situation (Silverstein 1980, Schegloff 1995). The beginning of the turn is crucial to the interpretation of how the on-going turn is connected to previous discourse (Schegloff 1996). In terms of information structure, Finnish word order is discourse-configurational: clause-level constituent order is regulated by discourse-pragmatic factors rather than by grammatical restrictions (Vilkuna 1995). The topic is positioned in the beginning of the turn and the focus at the end position (see also Lambrecht 1988, 1994). In other words new information (focus) is being produced after it has been provided a place in the situational interpretation frame. The same is true for concrete referents: new referents tend to be introduced in the focus position, after the relevance of the referent and its relation to the on-going action is made known to the participants by the first part of the turn.
There are, however, constructions in which the pronoun "tämä" is positioned in the beginning - in the topic place - of the turn, and at the same time a new referent is being introduced to the conversation. This paper focuses on one such specific complex construction type which is a combination of two constructions, namely a recognizable but unfinished representative construction and a clause:
1 (2) 3 4 (5) 6 connective (+ adverb) + dem.pron TM + is (+adverb) + CLAUSE 1 2 3 4 5 6 mut sitte tää on taas. täst tulee kans tota samaa sarjaa. but then TM is in turn. TM-ELA becomes also that same series (but then there's this again. this is going to be a part of that same series.)
This paper attempts to analyze these complex constructions in a larger conversational context. The combination of the two constructions - the referent representing first construction and the claus - is not presented in any grammar as such, yet this kind of combination appears regularly in the data. I will propose that this complex construction is a semi-grammaticalized construction used for special purposes in conversation.
Sequential analysis of the data reveals that these constructions are used in places where the previous sequence has been completed, yet the turns they compose do continue the on-going basic frame. The continuation of the basic action which the conversationalists recognize makes it possible to introduce concrete new referents in turn-initial position. It is noteworthy that in this construction referents are being introduced with a single pronoun, not with a noun phrase. However, the referring pronoun is not in the very beginning of the turn, but the turn begins with a cluster of particles. The particles have a role on anchoring the turn in the on-going situation thus making the referent identifiable.
The aim of the study is two-fold: on the one hand I focus on the role of the demonstratives, especially on the role of the pronoun "tämä", in directing and interpreting conversational discourse, and on the other hand I hope to show how actional, sequential and sentential constructions are used by the conversation participants as a resource when they interpret each others' turns and co-construct a shared discourse on a certain topic. As for the newness or givenness of the referent, it is relative and superpositioned (see Agha 1996)to the pronouns meaning by the context. When the referent is concretely new it is introduced to the conversation by other ways, for example, by the construction of the turn using the conventionalized meanings of sentence-types as a resource.
Makoto Hayashi
University of Colorado at Boulder
This paper provides an interactionally grounded account of a seemingly 'deviant' grammatical practice recurrently employed by speakers of Japanese in naturally-occurring conversation. The grammatical practice examined here is termed as "postposition-initiated utterances," i.e., utterances initiated with some sort of postpositional item (e.g., postpositional particles). This practice of talking is 'deviant' from the perspective of traditional Japanese grammar, since virtually all versions of grammatical description of Japanese available so far state that postpositions should and do immediately follow a noun or noun phrase that they mark (cf. Kuno 1973; Shibatani 1990; Tsujimura 1996). Consider the following invented sentence:
(1) Taroo-ga Hanako-o kuukoo-de mita.
Taroo-Subject.Particle Hanako-Object.Particle airport-Locative.Particle saw
"Taroo saw Hanako at the airport."
As this example illustrates, postpositional particles follow the nouns that they mark, and indicate their grammatical roles in the sentence.
In real-time social interaction, however, speakers sometimes begin their utterances with a postposition. Consider (2), in which A asks B about her recent trip:
(2) A: de! nan'nichikan gurai °sore tte.°
and for.how.many.days about that Quotative.Particle
"And, for about how many days, that one?"
> B: ga tookakan.
Subject.Particle for.ten.days
"is for ten days."
B initiates her utterance with the postpositional subject particle ga to answer A's question. Note that, although the English translation may give the impression that B's utterance is a direct continuation of A's utterance, it is not, because the particle ga cannot follow the quotative particle tte to form a coherent utterance. Here B produces an utterance which, if taken by itself, is considered ungrammatical from the perspective of traditional grammar. This study examines what social action is accomplished through this seemingly 'deviant' practice of talking in Japanese conversation. Through this examination, it explores some ways in which the grammatical organization of utterances and the organization of social interaction in which those utterances are deployed co-organize each other. A larger goal of the study, then, is to contribute
to the growing body of research exploring the mutual bearing of grammar and social interaction (cf. Ochs, Schegloff, and Thompson 1996).
Among a range of social action that postposition-initiated utterances are deployed to do, this study reports on the three most recurrent types of interactional work that they are employed to accomplish. The paper argues that the common thread among these different types of action is that speakers employ postposition-initiated utterances to 'steer' or redirect the trajectory of the talk away from the direction in which it is projectably going.
One type of action achieved through postposition-initiated utterances is to answer someone's question preemptively, i.e., before the question is fully produced, so as not to allow the initiation of the question to lead into opening up a new sequence of talk. Consider the following partial transcript:
(3) A: Maiyaa sa[n ni? "By Mr. Myer?"
Myer Mr. by
> B: [ga choodo nottete "Happens to be on the elevator,"
Subject.Particle happen be.on
Fragment (3) occurs immediately after B delivers a punch line of a story that is supposed to be funny (i.e., some drunken guy behaving badly on an elevator was scolded by another guy named Mr. Myer). In such a sequential location in storytelling, an initiation of a question about the content of the story by a story-recipient is a systematic alternative to an expected, or perhaps 'preferred,' response -- good laugh. It could undermine the whole point in the storytelling by questioning the accuracy of the detail of the story, and hence, the credibility of the storyteller. Facing the initiation of such a move by a recipient (A), then, the storyteller (B) preemptively answers A's question with a postposition-initiated utterance, and thereby re-introduces the punch line of the story so as to provide recipients with another opportunity to produce a 'preferred' response. Note that B's utterance in (3) begins with the subject particle ga, which ties itself to the noun phrase Maiyaa san ('Mr. Myer') in A's utterance, and incorporates it into the construction of the answering utterance.
Another type of action achieved through postposition-initiated utterances is found in the interactional environment where two participants respond to a third party together. In such an environment, postposition-initiated utterances are deployed by the second respondent to construct his/her response parasitic on the first respondent's response, and at the same time shift the trajectory of the first response from the direction in which it is projectably going. Consider the following simplified transcript, where A and B are explaining to another coparticipant what B gave to their mutual friend as a wedding gift:
(4) A: ... yatsu [:: mitaina " like that stuff"
stuff like
> B: [no:: (.) Koburi mitaina:? "of a smaller kind."
Genitive.Particle smaller.version like
A's utterance in (4) is the last part of her extensive description of B's gift (some type of bowl). When A has produced the noun yatsu ('stuff'), B comes in with a postpositional particle, no ('of'), and thereby constructs her utterance parasitic on that noun in A's utterance. At the same time, B, as the person who actually gave the gift, adds qualification to A's description of the gift; that is, what she gave was a smaller version of the kind of bowl that A has been describing. Here, a postposition-initiated utterance is used to build one's utterance parasitic on another's while adding some qualification to that other participant's talk and redirecting its trajectory.
The third type of action achieved through postposition-initiated utterances is to resume from self-initiated interruption. Consider the following.
(4) A: Moshi- rokujihan yatta? "If-, Was it 6:30?"
if 6:30 was
B: Un. "Yeah."
> A: ni tadoritsuitenakattara... "(if) I'm not there at Ø [=6:30]..."
Temporal.Particle have.arrived:if
In (4), A cuts of moshi ('if'), which projects a conditional clause, and asks for confirmation
about the time of an upcoming meeting. After receiving confirmation from B, A initiates a subsequent utterance with a postpositional temporal particle, ni ('at'), which connects itself to the noun rokujihan ('6:30') in A's prior utterance. By doing so, A resumes the construction of the interrupted conditional clause while incorporating the outcome of the interpolated confirmation-seeking sequence, i.e., the confirmed time of the meeting.
This study shows that interactional needs motivate speakers to exploit and stretch the usage of conventional grammar (i.e., utterance-initial use of postpositions), while a typological feature of Japanese grammar (i.e., postpositions attaching themselves to preceding noun phrases) provides resources to meet those interactional needs. This study thus documents another instance that demonstrates the mutual bearing of grammar and social interaction.
Eun-ju Kim
Pennsylvania State University
Many studies have examined the phenomenon of collaborative completion, whereby one on-going turn may change in the middle of a speaker's turn-in-progress and is then completed by a recipient of the utterance (Lerner, 1991, 1996; Lerner and Takagi, 1998; Ono and Yoshida, 1995; Hayashi and Mori, 1998). Lerner (1991, 1996) examines co-construction by focusing on grammatical factors, especially those which could constitute preliminary components, thereby creating room for a recipient to intervene and complete the other party's turn or utterance. Ono and Yoshida (1995) study co-construction in Japanese conversation and conclude, perhaps prematurely, that co-construction is "rare" in Japanese due to syntactic and pragmatic constraints. Hayashi and Mori (1998) reexamine the phenomenon, finding that co-construction actually occurs often in Japanese conversation, and that by engaging in such conversational activity, participants also simultaneously engage in a range of other interactional negotiations (e.g. displaying congruent understanding, managing alignment, etc.). Most of the studies (Lerner, 1991, 1996; Lerner and Takagi, 1998; Ono and Yoshida, 1995; Hayashi and Mori, 1998) focus on syntactic factors that make co-construction possible in the respective languages.
Using a database of approximately 60 minutes of audio and videotaped Korean conversations between female friends, the current study builds on previous work and will show that co-construction does occur frequently in Korean in spite of the syntactic and pragmatic constraints that Ono and Yoshida isolate for Japanese. However, this study will also propose that mutually shared experiences and/or ideas among participants of talk-in-interaction are primary to the occurrence of collaborative co-construction rather than particular syntactic features, as noted by Lerner (1991,1998), Lerner and Takagi (1998), Hayashi and Mori (1998), etc. In order to show systematically how co-constructed utterances are built during conversations between friends, this study will closely examine how shared ideas and experiences facilitate co-alignment between conversants. To this, this study will point out how syntactic features in Korean play certain roles in the co-construction activity. Then, the study will examine the exceptionally interesting phenomenon whereby a partial or even entire repeat of an utterance produced by one speaker appears in another speaker's contiguous turn.
Example (1) illustrates how utterances are collaboratively produced by different speakers across a number of turns. Further, it is indicative of an elaborate process of assessment (one turn), agreement (one turn), elaboration (one turn), agreement + elaboration (one turn), and agreement + ultimate sentence completion (one turn) across five turns of two speakers. What is evidenced here is a complete assessment sequence (Goodwin, 1992) whereby two participants collaboratively construct an evaluation of particular restaurant and systematically build on each other's comments and demonstrate mutual agreement.
(1) (T&M)
1. M: phondey katun tenun ice pwutamepsi mekul swu issnun tekwu,=
It is possible for us to eat in Ponderosa-type restaurant without much burden of money.
2. T: = u: =
= yeh =
= yeh =
3. M: = salam yelessi kase, =
= person several/together-SUJ go-CONN=
= several people go together and =
4. T: = u yele myeng iss ul ttay ka kass kwu.=
= Yeh, many-unit of people exist-PRECED CONN go have-CONN =
= Yeh, when there are several people, (we) go- and =
5. M: = e [( yelemyeng ) pay pwulukey mek ulswu i ssnun teykwu,
= Right [(many-unit of people ) stomach fully eat-possibility is-ATTR place IE,
= Right [(many people ) there, it is possible (for us) to eat till fulfilling stomach,
6. T: [kunyang
[DM
[just
7. T: um.
Yes.
Yes.
The main assessment is launched at line 1 where M mentions the restaurant Ponderosa as an inexpensive place to eat. T immediately provides an agreement token at line 2, and M elaborates her original assessment at line 3, which itself is an incomplete utterance by virtue of the connective -se 'and.' Crucially, in line 4, T's uptake begins with an agreement marker u 'yeah', not with a syntactic completion of the previous utterance, by demonstrating her understanding of this as an assessment first, and an elaboration second. T continues her turn using the subordinate clause construction yele myeung issul ttay kakasskwu 'when there are several people, we go and-', also signaling an incomplete clause, this time with verbal connective ko (realized in speech as kwu). To this, in line 5, M then provides an agreement marker, again, not a syntactic completion, and again demonstrating her recognition of T's just said utterance as an assessment first and further elaboration second. M then continues her turn, which ultimately completes both syntactically and pragmatically the assessment sequence that she herself launched at line 1. This use of acknowledgement tokens in the middle of utterances-in-progress is possible only when paricipants share ceratain experiences in this case, going Ponderosa restaurant, and also share similar reaction to the experience, making co-alignment possible. These phenomena, however, have not been mentioned in the earlier work (e.g. Lerner, 1991, 1996; Ono and Yoshida, 1995; Hayashi and Mori, 1998)
Example (2) illustrates the use of a partial or full repeat of an utterance just produced by the other party and examines how the phenomenon of repetition undersocres the co-alignment of the participants and facilitates the collaborative construction of utterances in Korean.
(2) (M&T)
1 M: (kuleni)kka yosaynun kamyen Mongolian ka c(h)anh(h)a. hh=
(So ) nowadays-TOP go-COND Mongolian go-CONN;NE;IE.
(So ) if (we) go (out to eat), (we) go to Mongolian.
2 T: u:: koki mek ule.
Ye::h meat eat-PURP
Ye::h to eat meat.
3 M: u. ceil mwe ssako pay pwulu nikka kekika.
Yeh. They are cheapest and we can eat fully there.
4. T: u. ssako =
Yeh. (they are)cheap and =
5. M: =pap mek ulswu isskwu.=
=(We ) can eat steamed rice (there)
6. T: = u papmekulswu isskwu.
= Yeh. (We) can eat steamed rice (there)
7. T: Kule- paphako koki,
So- Steamed rice and meat,
8. M: mek[ki nu::n u mongkoliani ceil nass tela.=
Mongolian was the best to eat (steamed rice and meat)
9. T: [mekkinun e
[eating
10. T: = u.
=I agree
=I agree
In this example, M and T are discussing a Mongolian restaurant in the area. Actually, the entire sequence begins with a collaboratively produced utterance, similar to those just examined in example (1), whereby one portion of the utterance is launched by one speaker and the second speaker produces an agreement token followed by either an elaboration, an assessment, syntactic completion, or any combination thereof. Such is the case in lines 1 and 2. Notice the mention of the word kogi 'meat' in T's utterance at line 2, which also syntactically completes M's turn, begun at line 1.
What is interesting in this excerpt, however, is the use of repetitions across contiguous turns, which also represent co-constructed utterances. For example, at line 3, M first responds to T's completion with an agreement marker u 'yeah' and then launches what will later become an extended assessment sequence using the adjective ssata 'cheap' producing a syntactically complete utterance. T then recycles the identical construction ssako 'it is cheap and-' to which M then adds yet another component of the attractiveness of the restaurant pap mekulswuisskwu 'we can eat steamed rice there'. In the immediately contiguous utterance at line 6, T produces agreement marker u 'yeah' and then reapeats M's identical clause pap mekulswuisskwu. At line 7, T reintroduces her mention of koki, 'meat' originally launched in line 2, and adds the element pap 'rice' that was actually M's contribution to this feature-by-feature building up an assessment. In so doing, T produces only a compound noun phrase paphako koki 'meat and rice', which T ultimately completes at line 8 with a coda-like summary of her description of the Mongolian restaurant, coming full circle with her original topic initiated at line 1.
This study has examined how and when co-construction occurs in talk in interaction in Korean. Previous studies have focused on syntactic features as primary elements of collaboratively produced talk. However, while these syntactic features do indeed create room for a recipient's uptake and completion of another party's turn, the current study proposes that the participants' shared ideas and experiences are, in fact, primary factors, as they actively participate in the co-construction of elaborate assessment sequences, with syntactic elements being secondary. Finally, we hope that the current study of utterance co-construction in Korean can serve as a departure point for the examination of how positively shared experiences among human beings are interactively talked about, made relevant, and remembered.
Namhee Han
UCLA
In this paper, I will demonstrate three functions of mother's use of directives: (1) language socialization, which contributes to child's awareness of social interaction, (2) construction of structural continuity or cohesive textuality of the discourse that is collaboratively achieved by child-mom's reciprocal recycling, and (3) parental scaffolding, which helps a child develop extended discourse. I will present excerpts of audio-recorded conversation in Korean between 27-month-old boy, his mom and aunt in order to demonstrate a collaborative nature of directives.
One primary focus of directives used by caretakers has been their effect on child language development. For example, Newport, Gleitman & Gleitman (1977) have argued that imperatives in motherese inhibit child's development. Others such as McDonald & Pien (1982) and Olsen-Fulero (1982), in their discussion about mom's conversational style, have viewed the use of directives as arising from intention to control the child's physical behavior and thus as adversely affecting language learning as compared to the use of conversation-eliciting speech.
The arguments mentioned above, however, miss the importance of directives in helping a child to repeat and learn phonological forms (Scollon, 1979), lexical items (Stine & Bohannon, 1983), and syntactic-semantic relations of words (Corrigan, 1980). In fact, Snow (1977) has suggested that characteristics of child-directed speech result from the mother's intention to converse and maintain interaction with him. In fact, later studies show that not all directives can be viewed as negative. For example, Tomasello & Farrar (1986) have demonstrated that mom and child talk more and become engaged in longer conversation when mom uses directives within joint attentional focus, which produced greater word learning than when mom uses directives redirecting child's attention. Akhtar, Dunham & Dunham (1991) have also found that mom's use of commands following child's attention correlates a productive vocabulary measure at a later stage. In addition, as Pan, Imbens-Bailey, Winner & Snow (1996) have pointed out, pragmatic perspectives on examining parental communicative intents serves as an indispensable element in the child-mom interaction. Therefore, the earlier arguments with a focus on negative effect of directives fail to realize the significance of dialogic interaction between mother and child.
Directives in mother-child interaction have been reported as a language socialization device for young children. Ochs & Schieffelin (1984), for instance, have documented Kaluli and Samoan mothers who use directives to teach a child to be aware of contexts he is in and to act in an appropriate way. In other words, through providing context-relevant feedback as a form of directives, a mother shows her child how to solve certain problems and this experience ultimately leads the child to act in a socially appropriate way. A similar language socialization strategy is reported about Wolof community (Rabain-Jamin, 1998). I will show what kinds of social value are transmitted and taught by mom's use of directives in my Korean data.
This use of directives to give clear cues to what child may need to say in a certain context results in child's repeating or recycling of mom's model sentence, which is again followed by mom's recycling of child's utterance with different purposes. As a result, a stretch of action of 'reciprocal imitation', i.e. the mom's directives and the child's recycling of them and mom's further feedback (Folger & Chapman, 1978) produces a certain textuality, i.e. structural continuity. As Sacks (1967) has noticed, the connectedness in discourse is achieved largely by syntactic operations, which he calls 'tying techniques.' The use of particulars of prior wording in context is well documented in Goodwin's (1990) account of 'format tying' in argument and insult sequence. I argue that mom-child conversation is another interesting type of discourse where format tying is accomplished through reciprocal imitation or recycling.
The importance of parental conversational support has been recognized in relation to language development (Snow, 1985). Imbens-Bailey & Snow (1997) have suggested that parents' conversational support helps their child to develop extended discourse of narratives. As they have argued, the child is required to control over linguistic and pragmatic skills in order to make meanings for others. I will demonstrate how directives directed to a child help to produce an extended discourse and become engaged in social interaction as a social being.