Is the correlation between grounding and transitivity universal?

(1)

Is the correlation between grounding

Studies in Language 27:2 (2003), 221–244.

issn 0378–4177 / e-issn 1569–9978© John Benjamins Publishing Company <TARGET "chu" DOCINFO AUTHOR "Kawai Chui"TITLE "Is the Correlation between Grounding and Transitivity Universal?"SUBJECT "SL, Volume 27:2"KEYWORDS ""SIZE HEIGHT "220"WIDTH "150"VOFFSET "4">

and transitivity universal?

Kawai Chui

National Chengchi University

This paper studies the correlation between grounding and transitivity in Chinese narratives and conversations based on the ten Transitivity features proposed by Hopper and Thompson (1980), to examine the universality of the correlation across different spoken discourse and different languages. According to the percentages of high-transitivity features vis-à-vis ground-ing, highly transitive clauses in Chinese narratives, just like English written narratives, tend to be foregrounded. However, such correlation is not borne out in conversational discourse, in that highly transitive clauses are almost equally distributed in foreground and background. As conversations are more pervasive and reflect people’s habitual use of language, it is concluded that grounding is independent of the morphosyntactic and semantic mani-festations of transitivity.

1. Introduction

Grounding is a fundamental property of text organization. In past research, it was mainly studied in narrative discourse that was comprised of a main story line.

It is evidently a universal of narrative discourse that in any extended text an overt distinction is made between the language of the actual story line and the language of supportive material which does not itself narrate the main events. (Hopper 1979: 213)

Thus, foregrounded clauses refer to the topical development along the plot of the story, while backgrounded clauses “support, amplify, or COMMENT ON the narration” (Hopper 1979: 215). In daily conversation, however, topics are spontaneous and naturally developed by the participants at the moment of speaking. When speakers do not narrate and when what they talk about lacks a

(2)

clear story line, does grounding still legitimately characterize the various types of information in face-to-face interaction? Consider the conversational excerpt in example (1). The subject of talk is how speaker C interacted with her boyfriend in class by continuously looking at him, as expressed by the topical and foregrounded clause in IU8 to IU10 (see Appendix C for the deﬁnition of IU). The rest of the utterance conveys background information, orienting the topic with temporal information — ‘every time I attended the math class’ (IU2), the location of the boyfriend — ‘he sat on my right side’ (IU3), and a reason for the seat arrangement which accounts for the boy’s location — ‘girls sat on one side, and boys sat on the other side’ (IU4–6).

(1) 1 C: ..ranhou%,\

..then

2 …(1.2)mei ci wo shang shuxue,_

…(1.2)every time 1sg attend math class 3 …ta jiu zuo zai wo you bian,_

…3sg then sit at 1sg right side 4 …women shi%,_

…1pl cop

5 …nusheng yi bian,_

…girl one side 6 …nansheng yi bian ma,_

…boy one side prt 7 B: (0)<P uh= P>,/

BC

8 C: …<A ranhou wo zhe bian zheyangzi%,_ then 1sg this side like this 9 …yanzhu dou%,_

…eyeball all

10 …yizhi A> ^kan ta zheyangzi,_

…continuously look 3sg like this

C: ‘Then, every time I attended the math class, he sat on my right side. We had girls sitting on one side, and boys sitting on the other side.’

B: ‘Uh.’

C: ‘Then I kept looking at him from this side (i.e. the girls’ side).’ Example (1) clearly illustrates the fact that utterances in conversation do receive diﬀerent grounding status: some are pertinent to introducing a topic, while some support it in various ways. The pragmatic notion of grounding, which

(3)

Is the correlation between grounding and transitivity universal? 223

manages diﬀerent kinds of information in the two common types of spontane-ous speech, is thus universal and genre-independent. As Hopper and Thomp-son (1980: 280–281) remark,

in any speaking situation, some parts of what is said are more relevant than others. That part of a discourse which does not immediately and crucially contribute to the speaker’s goal, but which merely assists, ampliﬁes, or com-ments on it, is referred to as BACKGROUND. By contrast, the material which supplies the main points of the discourse is known as FOREGROUND.

Hopper and Thompson (1980: 295) have further proposed a discourse pattern-ing of transitivity in terms of groundpattern-ing, since “[g]roundpattern-ing itself reﬂects a deeper set of principles — relating to decisions which speakers make, on the basis of their assessment of their hearers’ situation, about how to present what they have to say”. The way to present what to say can be realized by the transi-tivity of the clause: “high Transitransi-tivity is correlated with foregrounding and low Transitivity with backgrounding” (Hopper and Thompson 1980: 251). They also believe that

the grammaticization of devices to indicate grounding in narrative begins in the more pervasive conversational genre, and is extended to other genres in a natural way; i.e., the same devices used to highlight the main points of a conversation are also appropriate in foregrounded parts of a narrative. (Hop-per and Thompson 1980: 283)

Since their study was based on English written narratives, the question remains: Do grounding and transitivity still correlate in diﬀerent spoken discourse and across diﬀerent languages? This paper will examine the universality of the correlation in two common types of Chinese discourse — oral narratives and daily conversations, based on the ten Transitivity features proposed by Hopper and Thompson (1980).

Section 2 introduces the corpus of data. The structuring of topics in both conversational and narrative discourse will be discussed in Section 3, which is crucial to determining the grounding status and the transitivity of the clause in Section 4. Section 5 presents data analysis, followed by the discussion of the correlation between grounding and transitivity across diﬀerent discourse types and diﬀerent languages in Section 6.

(4)

224 Kawai Chui

2. Database

The research reported in this study rests upon the quantitative-qualitative analysis of four oral narratives and four casual conversations. The narrative discourse in the database refers to recountings of the story line of the movie Ghost produced in 1990. In the movie, the leading actor is killed accidentally by a hoodlum in a conspiracy organized by the leading actor’s best friend. Since the best friend needs to obtain a password from the leading actor’s girlfriend to launder money, the leading actor, through the help of a medium, returns in the form of a ghost to protect the girlfriend and avenge himself on the best friend. The four narrators, two females and two males, had already seen the movie before the taping sessions, which took place in the speech laboratory of their oﬃce, where they were each requested to recount the story of the ﬁlm without elicitation. Since the speakers and the interviewer were colleagues, the story was narrated in a natural way. The four narratives totaled thirty-six minutes.

The conversational data consist of four casual, unpremediated, multi-party conversations which took place from 1992 to 1995 among college students who knew each other. The first conversation had four participants; the second involved five; the third and the fourth were comprised of three speakers. The students were free to find topics of common interest. The data used in this study are twenty-minute excerpts from long recordings.

Subordinate clauses are not considered in the present study, because they function uniquely as the background in the corpus of data. Statements in the main clause, excluding copular clauses, presentative clauses, and clauses with nominal predicates, are the majority: 82.7% (464 out of 561) in storytelling, and 70.1% (1441 out of 2056) in everyday talk.

3. Structuring Topics in Discourse

To discuss the correlation between grounding and transitivity, the grounding status of a clause has to be ﬁrst determined. In narrative discourse, fore-grounded and backfore-grounded clauses can be identiﬁed within an episode — a thematic unit used to characterize the plot of the story (Tomlin 1987; van Dijk 1981). Without a main story line, how can the two types of clauses be distinguished in conversation? This section will propose that the notion ‘topic chain’ is useful for characterizing the advancement of conversational topics, as well as the progression of narrative episodes, so the results can be compared on the same basis.

(5)

Tsao (1979, 1987, 1988, 1990) has suggested that a ‘topic chain’ is a basic discourse unit, which can be characterized as a series of sentences sharing the same topic; the initial np of the ﬁrst sentence is the topic which, in turn, functions as the topic of the whole chain. Shi (1989), on the other hand, regards the ‘topic chain’ as a basic syntactic category, the largest unit in Chinese syntax. Li and Thompson (1979, 1981) use the notion ‘topic chain’ to interpret Chinese zero anaphors: the topic of the ﬁrst clause in a topic chain is the referent for the unrealized topics in the chain of the following clauses. You’s (1998) terminology is ‘topic continuity’, within which the referents of zero anaphors in written narratives can be recovered.

The scope of T[opic] C[ontinuity] can be deﬁned in terms of semantic propositions as follows: the previous proposition of the ﬁrst proposition of a topic continuity and the following proposition of the last proposition should be subsumed by other macropropositions, i.e. should belong to other discourse topics/themes. In the surface structure, the use of noun phrases…and changes in time or place usually indicate the end of the previous topic continuity and the beginning of a new topic continuity…. In addition, TC is a hierarchical discourse structure; a topic continuity may contain severalsub-topic continui-ties (sub-TC), which may contain their own sub-topic continuicontinui-ties.

(You 1998: 40)

The term ‘topic chain’ is still employed in the present study to analyze the structuring of conversational and narrative topics. Yet, the notion is defined in a different way: topic chains as discourse units characterize text organization based on the semantic relationship between utterances. A sequence of clauses about the same subject matter introduces a topic; a topic, or more commonly, a number of semantically-connected topics then form a topic chain; the topic chains sharing the identical theme further constitute a larger topic chain. In the corpus, three levels of topic chains were identified for conversations: ‘topic chain’, ‘sub-topic chain’, and ‘sub-sub-topic chain’; two levels for narratives.

3.1 Topic chain

A topic chain (TC) is typically comprised of a set of topics sharing the same theme, whose boundary represents major changes in time, place, event, or character(s). For general discussion of subject matter, events within the topic chain have no reference time or place. Table 1 illustrates three consecutive topic chains, namely ‘Seafood’, ‘Car Accident’, and ‘Doctors’. The topic succession does not follow the actual temporal sequence, since the car accident happened years before the seafood events.

(6)

A topic chain can be interrupted by a separate theme. The talk in Table 2

Table 1.Topic chains

TC Time Place Event Character(s) ‘Seafood’

‘Car Accident’ ‘Doctors’

one day in Boston another day in Oregon another day in Boston

Boston Oregon Boston quality of ﬁsh car accident medical diagnosis ﬁsh a boyfriend a doctor, a patient

Table 2.Interruption within a topic chain

TC Time Place Event Character(s) ‘Speaker C’s Romance’ junior-high-school time speaker C’s junior high school relationship with the ﬁrst boyfriend speaker C, the ﬁrst boyfriend

‘Long Hair’ high-school time generic wearing long hair speakers A, B, C ‘C’s Romance’ continued junior-high-school time speaker C’s junior high school relationship with the second boyfriend

speaker C, the second boyfriend

centers on the speakers’ past romance. While C communicates about her boyfriends in junior high school and just finished the story concerning the first boyfriend, another speaker interrupts and brings up a different topic on wearing long hair in high school. After which the original topic chain is resumed, and C continues telling about the second boyfriend.

3.2 Sub-topic chain

Within each topic chain may be found more than one subsidiary topic chain of the same theme; that is, the ‘sub-topic chain’ (sub-TC). The sub-TCs may not be inherently ordered. Nor do they necessarily share the same temporal and spatial information with their host TC. The speakers in Table 3 are discussing the departmental gathering to be held at the end of the semester. Since the place of the party has not yet been decided, the seven sub-TCs about the important elements to hold the gathering consistently refer to the same time but lack a speciﬁc location. The space and time in other sub-TCs may be distinct from those of their host TC. In the narrative TC on the leading actor (i.e. the man) thwarting his friend’s conspiracy (see Appendix A) in Table 4, a narrator ﬁrst focused on the friend who tries to launder money by transferring it to a bank account (‘Friend

(7)

Launders Money’). Then, the recounting was shifted to the leading actor who

Table 3.Sub-topic chains

TC: ‘Departmental Gathering’

Sub-TCs Time Event Character(s) ‘Food Arrangement’ end of semester arrange for food food, utensils ‘Place’ end of semester the place holding the

gathering

place, expense

‘Gas’ end of semester gas for cooking gas, food ‘Food Containers’ end of semester choose the right kind of

food containers

‘Helpers’ end of semester check the schedules of helpers

students

‘Entertainment’ end of semester arrange for perform-ances at the party

a teacher

‘Making Tables’ end of semester make tables for the party

tables

transfers the money to another account before the friend does (‘Man Transfers Money’). The last sub-TC goes back to the friend when he ﬁnds the money is gone (‘Friend Discovers Money Being Transferred’).

The sub-TCs do not necessarily proceed sequentially; a new sub-TC of a previous topic chain can be brought up after other unrelated topic chains. For instance, the last sub-TC of ‘Departmental Gathering’ in Table 3 is ‘Making Tables’, after which the speakers discuss ‘Playlet Presentation’, ‘Art Exhibition’, and then ‘Departmental Directories’. After ‘Departmental Directories’, the participants return to ‘Departmental Gathering’ and resume the old topic of entertainment. Afterward, ‘Playlet Presentation’ is also resumed to introduce a new sub-TC about audience.

Non-sequential sub-TCs are not interruptions, since no one initiates a different topic in the middle of the talk. A real instance of interruption within an ongoing sub-TC can be found in Table 5. The speaker is sharing her own experience in seeing a cockroach in the library, but she stops and discusses with other speakers whether cockroaches can fly or make noises, after which she resumes and finishes the interrupted topic.

(8)

3.3 Sub-sub-topic chain

Table 4.Sub-topic chains of diﬀerent times and places

TC: ‘Man Thwarts the Conspiracy’

Sub-TCs Time Place Event Character(s) ‘Friend Launders

Money’

one day oﬃce the friend starts to launder money

friend, illegal money ‘Man Transfers Money’ another

day

bank the man transfers the money to the medium’s bank account

man, illegal money

‘Friend Discovers Money Being Transferred’

another day

oﬃce the friend discovers his money being transferred

friend, illegal money

Table 5.Interruption within a sub-topic chain

TC: ‘Cockroaches’

Sub-TCs Time Place Event Character(s) ‘A Cockroach in the

Library’

some time in the past

library speaker A found a cockroach in the library speaker A, a male student, a cockroach ‘Cockroaches’ Attributes’

generic generic speakers discuss whether cock-roaches can ﬂy and make noises

cockroaches

‘A Cockroach in the Library’ continued

some time in the past

library speaker A’s reac-tion

speaker A

The sub-topic chain can be further divided into ‘sub-sub-topic chain’ (sub-sub-TC) whose content is directly subsidiary to its host sub-TC, but indirectly related to the TC as a whole. Table 6 indicates two sub-sub-TCs on speaker B’s first boyfriend (sub-TC1) which is part of the speaker’s romance in senior high school (TC). Like the TCs in Table 1 and the TCs in Table 4, the two sub-sub-TCs here — ’ Meeting the Boy’ and ‘A Girl Confiding about the Boy’ — occur at different times and places.

A topic chain boundary may coincide with its ﬁrst sub-topic chain bound-ary, as a sub-topic chain boundary does with its ﬁrst sub-sub-topic chain

(9)

boundary. It is counted, respectively, as an instance of TC and an instance of

Table 6.Sub-sub-topic chains

TC: ‘Speaker B’s Romance at Senior High’ Sub-TC1: ‘First Boyfriend’

Sub-Sub-TC1 Time Place Event Character(s) ‘Meeting

the Boy’

ﬁrst year in senior high

a meeting place discuss how spea-ker B met the boy

speaker B, the ﬁrst boy Sub-Sub-TC2 Time Place Event Character(s) ‘A Girl

Con-ﬁding about the Boy’

third year in senior high

cram school the boy wrote about speaker B

a girl, speaker B, the ﬁrst boy

Sub-TC2 Time Place Event Character(s) ‘Second

Boy-friend’

senior-high-school time

street discuss how spea-ker B met and got along with the boy

speaker B, the second boy

Table 7.Frequency distribution of TCs, sub-TCs, and sub-sub-TCs

narratives conversations n % n % TCs Sub-TCs Sub-sub-TCs total: 38 11 0 49 77.6 22.4 0.0 100.0 37 57 14 108 34.2 52.8 13.0 100.0

sub-TC in the present tabulation. The frequency distribution of the various levels of topic chain in the two types of discourse is presented in Table 7.

In short, topic chains, whether narrative or conversational, display a hierarchical organization similar to the organization of ‘supertopics’, ‘topics’, and ‘subtopics’ proposed by Chafe (1994), and to the structuring of ‘topic continuity’ in You’s (1998) framework. Most of the topics proceed sequentially, especially in narrative discourse. The conversational events, however, may not be organized in a temporal sequence. Moreover, an on-going conversational topic, be it part of a topic chain or of a (sub)-sub-topic chain, can be interrupt-ed by a new distinct theme, after which the speakers can go back to the inter-rupted topic; or without interruption, an old topic can be brought up again

(10)

after other unrelated topic chains. Thus, topic development is spontaneous and, in daily chat, can be negotiated among the speakers.

4. Identiﬁcation of foregrounded and backgrounded clauses

The purpose of studying the topic chain structure is to identify foregrounded and backgrounded clauses in TC, sub-TC, or sub-sub-TC. Foregrounded clauses introduce diﬀerent topics within their local (sub-(sub-))TC domain. Backgrounded clauses, on the other hand, are orientations, descriptions, elaborations, and digressions concerning one particular topic. This section will illustrate the identiﬁcation of these two types of clauses in context.

One of the conversational topic chains is ‘Handwriting and Computer Documentation’ that consists of three sub-topic chains: ‘the person Z’s hand-writing and documents’, ‘the person Lu’s handhand-writing and documents’, and ‘the person Gi’s handwriting and documents’. All the utterances in the second sub-topic chain are presented in example (2). It has two foregrounded events moving the topics from computer documents (IU2–3) to Lu’s handwriting (IU6–7). The clause in L’s second turn ni kan ta de zi ‘you look at his hand-writing’ (IU5) is backgrounded, because it functions to orient the major character from ‘Lu’s document’ to ‘Lu’s handwriting’. So is the last clause (IU8–10), elaborating that Lu’s pretty writing has many diﬀerent styles.

(2) 1 L: …(.7)<A meiyou A>.\ neg 2 ..xiang Lu o=.\

..like Lu prt

3 …(1.)ta de <L2 document L2> dou nong de hao

…(1.)3sg assc document all make compl very piaoliang dui bu dui,_

pretty right neg right 4 Z: …<PP hm= PP>.\

BC

5 L: (0)ni kan ta de zi,_

(0)2sg look at 3sg assc character 6 ..ta nage zi o=,_

(11)

7 …zhende hao piaoliang.\

…really very be.pretty

8 …(1.1)erqie ta ziji benshen de zi=,_

…(1.1)moreover 3sg SELF per se assccharacter 9 ..benshen ^jiu hen you

hen%--..per se empvery have very

10 …(.7)you henduo ^butongde <L2 style L2>.\

…(.7)have many diﬀerent style

11 H: …<PP oh= PP>?/ BC

L: ‘No, like Lu, his documents are all very pretty, right?’ Z: ‘Hm.’

L: ‘You look at his handwriting. His handwriting is really very pretty. Moreover, his own writing…he can write in many diﬀer-ent styles.’

H: ‘Oh.’

Example (3) illustrates the identiﬁcation of foregrounded and backgrounded clauses in a narrative topic chain about the murderer breaking into the leading actress’s house (‘Murderer Breaks into Woman’s House’, see Appendix A). The foregrounded clauses move the topical events from the woman changing clothes at home (IU8–9), to the hoodlum breaking into the woman’s house (IU10–11), then to the man scaring a cat (IU24–25), and ﬁnally to the cat scaring the hoodlum away (IU26–27). The rest of the excerpt is background information. First, the preceding topic chain, ‘Man Stays in the World after Death’ (see Appendix A), already mentions that the man does not go to heaven after death but follows and stays beside the girlfriend. Thus, the clause in IU4 indicates the man’s habitual behavior of following his girlfriend, which accounts for his presence in the topic chain. The man’s recognition that the hoodlum is the murderer (IU12–15), and his nervousness and worry (IU16–19), together explain why the man has to save the woman by scaring the cat. Finally, the clause in IU21–22 is a pre-announcement for resolving the predicament.

(3) 1 W: …(H)jieguo=?/ …(H)as a result 2 …(1.3)youyitian=?/ …(1.3)one day 3 …(.8)youyitian=,_ …(.8)one day

(12)

4 …ta gen zhe ta nupengyou.\

…3sg follow DUR 3sg girlfriend 5 …zai= ta nupeng,\

…be:at 3sg REPAIR 6 …zai jiali deshihou.\

…be:at home when 7 …(.8)uh=,_

…(.8)pf

8 …ta% —

…3sg

9 …nupengyou zheng zai huanyifu,\

…girlfriend just progchange clothes 10 …(.7)ranhou you yige ^huairen,\

…(.7)then there:be one bad guy 11 …^huairen chuang jinlai.\

…^bad guy break into a house come in 12 …(H)jiu ta yi ^kan,\

…(H)then 3sg as soon as look 13 …jiushi= zai= jieshang,\

…that is on street 14 …shasi ta de nage=,_

…kill 3sg assc that 15 …nage= hunhun.\

…that hoodlum 16 …(H)ranhou ta jiu=,_

…(H)then 3sg then 17 …ta jiu hen jinzhang._

…3sg then very be:nervous 18 …(.8)hen pa ta=,_

…(.8)very be:afraid sg

19 …zai qu shanghai ta nupengyou.\

…again go hurt 3sg girlfriend 20 …(.9)(H)suoyi=,_

…(.9)(H)so 21 …(1.2)ta jiu=,_

…(1.2)3sg then

22 …xiang le yige banfa.\

(13)

23 …jiu=,_

…then

24 …(.9)jingxia le= nage=,_

…(.9)scare pf that

25 …ta nupengyou= suo yang de mao.\

…3sg girlfriend SUO raise assc cat 26 …ranhou ba nage huairen gei gei=,_

…then BA that bad guy GEI GEI 27 …(1.)gei ^xiazou le.\

…(1.)GEI scare away pf

W: ‘As a result, one day, one day, he follows his girlfriend. At her…at home, uh…his girlfriend is changing clothes. Then, there is a bad guy….a bad guy breaking into the house. He takes a look (at the bad guy). The bad guy is the…the hoodlum who killed him on the street. He, then…then, he is very nervous. He is very afraid that the bad guy will hurt the girlfriend again. So, he thinks of a solution. (He) scares the cat that his girlfriend raises. Then (the cat) scares away the bad guy.’ Table 8 presents the frequency distribution of foregrounded and backgrounded clauses in the narrative and conversational texts.

Table 8.Frequency distribution of foregrounded and backgrounded clauses

narratives conversations n % n % Foregrounded clauses Backgrounded clauses total: 161 303 464 34.7 65.3 100.0 375 1066 1441 26.0 74.0 100.0 5. Data analysis

Identifying foregrounded and backgrounded clauses is the first indispensable step to examine the correlation between grounding and transitivity. The discourse patterning of transitivity that “the foci of high Transitivity and low Transitivity correlate with the independent discourse notions of foregrounding and backgrounding respectively” (Hopper and Thompson 1980:294) is based on ten defining properties: ‘kinesis’, ‘affirmation’, ‘mode’, ‘punctuality’, ‘aspect’, ‘participants’, ‘agency’, ‘volitionality’, ‘affectedness of O’, and ‘individuation of O’.

(14)

234 Kawai Chui

These features can be categorized into two types: the ﬁrst half are concerned with the ‘speech event’; the second half with the ‘participant’. Their respective frequency distribution vis-à-vis grounding in Chinese spoken discourse will be presented in this section.

5.1 Speech event

‘Kinesis’ is concerned with the action-state distinction. Action and state verbs constitute the majority in the discourse data examined here: 72.6% (337 out of 464) in narratives, and 88.8% (1279 out of 1441) in the conversations. The minor verb types, categorized as ‘others’ in Table 9, are mental act, mental state, and process verbs. ‘Aﬃrmation’ and ‘mode’ refer to the aﬃrmative-negative and the realis-irrealis dichotomy, respectively. Their frequency distribution in foreground and background is given in Table 9.

The feature of punctuality “refers to the suddenness of an action, or the

Table 9.‘Kinesis’, ‘aﬃrmation’, ‘mode’, and grounding

narratives conversations

foreground background foreground background

n % n % n % n % Action verbs State verbs Others total: 134 11 16 161 83.2 6.8 10.0 100.0 113 79 111 303 37.3 26.1 36.6 100.0 271 86 18 375 72.3 22.9 4.8 100.0 535 387 144 1066 50.2 36.3 13.5 100.0 Aﬃrmative Negative total: 154 7 161 95.7 4.3 100.0 258 45 303 85.1 14.9 100.0 326 49 375 86.9 13.1 100.0 878 188 1066 82.4 17.6 100.0 Realis Irrealis total: 136 25 161 84.5 15.5 100.0 191 112 303 63.0 37.0 100.0 169 206 375 45.1 54.9 100.0 482 584 1066 45.2 54.8 100.0

absence of a clear transitional phase between onset and completion” (Hopper and Thompson 1980: 286). ‘Aspect’, being realized as ‘perfectivity’ in this study, is concerned with “an action viewed from its endpoint” (Hopper and Thomp-son 1980: 252). These two factors have to do with action verbs only, because states are inherently continuous, undiﬀerentiated, and boundless. The frequen-cy distribution of punctual and non-punctual, perfective and imperfective clauses with respect to grounding is shown in Table 10.

(15)

5.2 Participant

Table 10.‘Punctuality, ‘aspect’, and grounding

n % n % n % n % Punctual Non-punctual total: 27 107 134 20.1 79.9 100.0 17 96 113 15.0 85.0 100.0 17 254 271 6.3 93.7 100.0 25 510 535 4.7 95.3 100.0 Perfective Imperfective total: 66 68 134 49.3 50.7 100.0 42 71 113 37.2 62.8 100.0 70 201 271 25.8 74.2 100.0 130 405 535 24.3 75.7 100.0

The ﬁrst feature related to participants is ‘number of participants’ taking part in

Table 11.‘Participant’, ‘agency’, ‘volitionality’, and grounding

n % n % n % n % One participant Two participants total: 50 80 130 38.5 61.5 100.0 117 89 206 56.8 43.2 100.0 166 162 328 50.6 49.4 100.0 502 310 812 61.8 38.2 100.0 Human A Non-human A total: 104 2 106 98.1 1.9 100.0 11 7 18 61.1 38.9 100.0 117 14 131 89.3 10.7 100.0 130 20 150 86.7 13.3 100.0 Volitional A Non-volitional A total: 63 41 104 60.6 39.4 100.0 5 6 11 45.5 54.5 100.0 58 59 117 49.6 50.4 100.0 59 71 130 45.4 54.6 100.0

speech events. Clauses taking sentential or verbal arguments were excluded in the present tabulation; those including three participants were calculated as two-participant clauses, since only four instances were found in the corpus — three occurrences of gei ‘give’ and one occurrence of jiao ‘teach’. ‘Agency’ refers to the participants in role A (for Agent, see Dixon 1979). This study deﬁned it in terms of ‘humanness’; thus, two kinds of participants were distinguished: ‘human’ and ‘non-human’. Finally, the property of ‘volitionality’ only refers to human participants in role A performing an action purposefully (Hopper and

(16)

236 Kawai Chui

Thompson 1980: 252). Non-human referents in this particular syntactic position are irrelevant to this semantic distinction. Table 11 presents the statistics for these features in foreground and background.

The last two features are pertinent to the referents in role O (for Object, see

Table 12.‘Aﬀectedness’, ‘individuation’, and grounding

narratives conversations foreground background foreground background

n % n % n % n % Totally aﬀected O Partially/not aﬀected O total: 50 21 71 70.4 29.6 100.0 36 15 51 70.6 29.4 100.0 45 77 122 36.9 63.1 100.0 67 85 152 44.1 55.9 100.0 Human O Non-human O total: 52 19 71 73.2 26.8 100.0 6 45 51 11.8 88.2 100.0 24 98 122 19.7 80.3 100.0 33 119 152 21.7 78.3 100.0 Concrete O Abstract O total: 70 1 71 98.6 1.4 100.0 24 27 51 47.1 52.9 100.0 116 6 122 95.1 4.9 100.0 97 55 152 63.8 36.2 100.0 Generic O Particular O total: 3 68 71 4.2 95.8 100.0 24 27 51 47.1 52.9 100.0 71 51 122 58.2 41.8 100.0 84 68 152 55.3 44.7 100.0

Dixon 1979). First, the ‘affectedness’ of O referents has to be determined by the entire clause, in that “[t]he completion of the action of a clause naturally involves the whole O. When the action is viewed not as completed, but as under way, the O is less likely to be completely affected” (Hopper and Thompson 1980: 287). There are many other situations in which the events affect O referents only to a certain degree or they have no effect on them, as in hypothet-ical or deontic situations. In the present tabulation, referents which are com-pletely affected by the events are classified as ‘totally affected’; partial and no affectedness are grouped together as ‘partially/not affected’. Finally, the ‘individuation’ of referents in role O is determined by the human-nonhuman, concrete-abstract, and generic-particular distinctions. Their respective frequen-cy distribution is given in Table 12.

(17)

6. Transitivity and grounding

Based on the results in Section 5, this section will compare Hopper and Thompson’s (1980) results with the ﬁndings for Chinese discourse to examine the universal correlation between transitivity and grounding across various discourse types and diﬀerent languages.

Hopper and Thompson (1980:283–284) note that “in languages like English, the audience infers grounding not from a single morphosyntactic feature, but from a CLUSTER OF PROPERTIES, no single one of which is exclusively characteristic of foregrounding”. Many studies also regard the ten transitivity features as a ‘cluster concept’ (Wallace 1982; Reinhart 1984), in that even though some of the semantic distinctions do not distinguish the foreground from the background, the transitivity of a clause, be it foregrounded or backgrounded, should depend on the overall distribution. I thus follow Hopper and Thompson’s calculation of the average based on the percentages of high-transitivity features in both foregrounded and backgrounded clauses. Table 13 lists the percentages of high-transitivity features vis-à-vis grounding from Tables 9, 10, 11, and 12.

In English written narratives, the average of high Transitivity in fore-grounded clauses is 78%; the percentage is reduced to thirty-nine in back-grounded clauses (see Table 2 in Hopper and Thompson 1980: 288). As recapitulated by DeLancey (1987: 53),

[t]he canonical transitive clause has two participants, reports a kinetic event, is punctual and perfective, and a definite, referential, individuated, and wholly affected patient and a volitional agent which ranks high on the animacy hierarchy, and is affirmative and realis.

Clauses bearing these highly transitive features “also turn out to predominate in the foregrounded portions of discourse” (Hopper and Thompson 1980: 292). The averages in Chinese narratives are close to those in English, though the degree of transitivity in the background is higher (49.5%). However, such correlations are not borne out in the conversational discourse, in that highly transitive clauses are almost equally distributed in foreground and background (51.4% vs. 46.5%). In other words, imperfective or irrealis events, for instance, do not often function as background in everyday talk, as they do in narration (see Tables 9 and 10). These results are contrary to Wallace’s (1982: 205) claim that people are more interested in what is factual or real, designating “speciﬁc, actual, deﬁnite, bounded acts”. Unbounded events in the past, the present, and the future can be interesting topics of conversation.

(18)

The claim that the more high-transitivity properties a clause includes, the

Table 13.Grounding and high transitivity

foreground background foreground background Participants Kinesis Aspect Punctuality Volitionality Aﬃrmation Mode Agency O-aﬀectedness O-individuation average: 61.5% 83.2% 49.3% 20.1% 60.6% 95.7% 84.5% 98.1% 70.4% 89.2% 71.3% 43.2% 37.3% 37.2% 15.0% 45.5% 85.1% 63.0% 61.1% 70.6% 37.3% 49.5% 49.4% 72.3% 25.8% 6.3% 49.6% 86.9% 45.1% 89.3% 36.9% 52.2% 51.4% 38.2% 50.2% 24.3% 4.7% 45.4% 82.4% 45.2% 86.7% 44.1% 43.4% 46.5%

more likely it is to be foregrounded seems to be universal across languages, yet only in narrative discourse. Why does high transitivity correlate with fore-ground information only when the speaker tells a story, be it English or Chinese, written or spoken? The answer lies in the content of the discourse. In storytelling, the foregrounded clauses are, in general, highly transitive, because the topical events along the plot of the movie are overwhelmingly factual actions performed by human participants. Speakers in everyday talk, however, are interested in a wide variety of topics about the real and the unreal; the speciﬁc and the general; the human and the non-human. Thus, the overall average of high Transitivity as determined by the cluster of properties is low. For instance, in the topic chain on a departmental gathering (see Table 3), the speakers who are in charge of this party discuss food arrangement, the location of the party, gas, food containers, helpers, entertainment, and tables. Since it is a prospective gathering at the moment of speaking and there is no inherent order among the seven topics, most of the speech events are imperfective and have not yet happened. Another topic chain centers on a car accident, so the speech events are mostly factual in the past. Such a wide variety of topics concerning the present and the future are not available in the narrative data because of the restriction on the choice of content. The narrators do not communicate a lot about the immediate environment or about the future; their narration follows the movie plot to a large extent. As a result, foregrounded narrative events are mainly realis and perfective. Moreover, since the movie story is human-centered, the topic chains are mainly comprised of concrete,

(19)

particular, human referents in the foreground, including the leading actor, the leading actress, the man’s friend, a hoodlum, and a medium. In daily talk, human referents are not the overwhelming majority. Speakers can bring up any topics, so humans and non-human entities are equally the center-of-interest. In the topic chain on patents, for instance, it is the non-human ‘patent’ that is the major participant when speakers discuss patent writing.

Hopper and Thompson’s (1980: 294) claim that “SEMANTIC AND GRAMMATICAL PROPERTIES WHICH ARE IRRELEVANT TO FORE-GROUNDING ARE ALSO IRRELEVANT TO TRANSITIVITY” is not univer-sal. Transitivity does not correlate with grounding in daily talk. As conversa-tions are more pervasive and reﬂect people’s habitual use of language, the correlation found in narrative discourse is merely a coincidence; the events in the stories happen to be mostly actions which center around and are maneu-vered by human participants. To claim that speakers use highly transitive clauses to characterize foregrounded events just when they narrate the human-centered type of stories is inexplicable. The ten semantic features lend them-selves only to transitivity. Organizing texts into foreground and background is independent of the morphosyntactic and semantic manifestations of transitivity.

References

Chafe, Wallace L. 1994. Discourse, consciousness, and time. Chicago: The University of Chicago Press.

DeLancey, Scott. 1987. “Transitivity in grammar and cognition”. In: Tomlin, Russell S. (ed.),

Coherence and grounding in discourse, 53–68. Amsterdam/Philadelphia: John Benjamins.

Dixon, R. M. W. 1979. “Ergativity”. Language 55: 59–138.

Du Bois, John W.; Schuetze-Coburn, Stephan; Cumming, Susanna; and Paolino, Danae. 1993. “Outline of discourse transcription”. In: Edwards, Jane A. and Lampert, Martin D. (eds.), Talking data: transcription and coding in discourse research, 45–89. Hillsdale, N.J.: Lawrence Erlbaum Associates.

Fleischman, Suzanne. 1985. “Discourse functions of tense-aspect oppositions in narrative: toward a theory of grounding”. Linguistics 23: 851–882.

Givón, Talmy. 1987. “Beyond foreground and background”. In: Tomlin, Russell S. (ed.),

Hopper, Paul. 1979. “Aspect and foregrounding in discourse”. In: Givón, Talmy (ed.),

Discourse and syntax, 213–241. New York: Academic Press.

Hopper, Paul; and Thompson, Sandra A. 1980. “Transitivity in grammar and discourse”.

(20)

240 Kawai Chui

Li, Charles N.; and Thompson, Sandra A. 1979. “Third-person pronouns and zero-anaphora in Chinese discourse”. In: Givón, Talmy (ed.), Discourse and syntax, 311–335. New York: Academic Press.

Li, Charles N., and Sandra A. Thompson. 1981. Mandarin Chinese: a functional reference

grammar. Berkeley: University of California Press

Reinhart, Tanya. 1984. “Principles of Gestalt perception in the temporal organization of narrative texts”. Linguistics 22: 779–809.

Shi, Dingxu. 1989. “Topic chain as a syntactic category in Chinese”. Journal of Chinese

Linguistics 17: 223–261.

Tomlin, Russell S. 1985. “Foreground-background information and the syntax of subordina-tion”. Text 5: 87–122.

Tomlin, Russell S. 1987. “Linguistic reﬂections of cognitive events”. In: Tomlin, Russell S. (ed.),

Tsao, Feng-fu. 1979. A functional study of topic in Chinese: the ﬁrst step toward discourse

analysis. Taipei: Student Book Company.

Tsao, Feng-fu. 1987. “A topic-comment approach to the ba construction”. Journal of Chinese

Linguistics 15: 1–54.

Tsao, Feng-fu. 1988. “Topics and clause connectives in Chinese”. Bulletin of the Institute of

History and Philology, Academia Sinica 59: 695–737.

Tsao, Feng-fu. 1990. Sentence and clause structure in Chinese: a functional perspective. Taipei: Student Book Company.

van Dijk, Teun A. 1981. Episodes as units of discourse analysis. In: Tannen, D. (ed.),

Georgetown University Round Table on Languages and Linguistics, 177–195. Washington,

D. C.: Georgetown University Press.

Vetters, Carl. 1992. “Foreground and background: Weinrich against Labov”. In: Kefer, Michel and van der Auwera, Johan (eds.), Meaning and grammar: cross-linguistic

perspectives, 367–381. New York: Mouton de Gruyter.

Wallace, Stephan. 1982. “Figure and ground: the interrelation-ships of linguistic categories”. In: Hopper, Paul (ed.), Tense-aspect: between semantics and pragmatics, 201–223. Amsterdam/Philadelphia: John Benjamins.

You, Yu-ling. 1998. Interpreting Chinese zero anaphors within topic continuity. Illinois: Ph.D. dissertation, University of Illinois at Urbana-Champaign.

Appendix A: All the topic chains in Ghost

TC Time Place Event Character(s) ‘House Moving’ one day an apartment the friend helps the man

and the woman move into a new place

man, woman, friend

(21)

TC Time Place Event Character(s) ‘Friend’s

Conspir-acy Against Man’

one day hoodlum’s place

the friend asks a hood-lum to steal the man’s bank-account password

friend, hoodlum, password

‘Murder’ after a show an alley the hoodlum robs the man and kills him acci-dentally

hoodlum, man

‘Man Stays in the World after Death’

after the man dies

woman’s house the man does not go to heaven but stays beside the woman man, woman ‘Murderer Breaks into Woman’s House’

one day woman’s house the hoodlum breaks into the woman’s house to get the password, but he is driven away by the man through a cat

man, woman, hoodlum, cat ‘Man Discovers Friend’s Conspiracy’

the same dayhoodlum’s place

the man follows the hoodlum and discovers the friend has organized the robbery

man, hoodlum, friend

‘Friend Gets Close to Woman’

one day woman’s place the friend tries to get close to the woman to get the password

friend, woman

‘Man Meets Medium’

one day medium’s place the man meets a medi-um, and tries to get her help

man, medium

‘Medium Meets Woman’

one day woman’s place on behalf of the man, the medium tells the woman about the friend’s plot medium, wom-an, man ‘Man Learns Tricks from a Ghost’

one day subway a ghost at subway teach-es the man to play tricks

man, ghost

‘Man Plays Tricks on Friend’

one day friend’s oﬃce the man punishes the friend by playing tricks on him

man, friend

(22)

TC Time Place Event Character(s) ‘Man Thwarts the

Conspiracy’

one day bank the friend tries to laun-der money, but the man transfers the money to the medium’s account

man, friend, illegal money

‘Fight between Man and Friend’

one day woman’s place the friend threatens the man to retrieve the money, so they ﬁght, and then the friend dies

man, friend ‘Man Goes to Heaven’ after the ﬁght

woman’s place the woman can see the man through the medi-um, then the man goes to heaven

man, woman, medium

Appendix B: Abbreviations in the interlinear glosses

1pl first person plural 1sg first person singular 2pl second person plural 2sg second person singular 3pl third person plural 3sg third person singular assc associative morpheme ba the morpheme BA bc backchannel compl complementizer cop copula verb dur durative aspect emp emphatic adverbial exp experiential aspect gei the morpheme GEI neg negative morpheme pf pause filler prf perfective aspect prog progressive aspect prt discourse particle repair repair phoneme(s) self reflexive morpheme suo the morpheme SUO

(23)

Appendix C: Transcription conventions

The transcription system was proposed by Du Bois et al. (1993). ‘Intonation unit’ (IU) is deﬁned as a stretch of speech uttered under a single coherent intonation contour, which tends to be marked by a pause, a change of pitch, and a lengthening of the ﬁnal syllable.

Units

{carriage return} intonation unit

-- truncated intonation unit {space} word - truncated words : identity/turn start [ ] speech overlap Transitional continuity . ﬁnal , continuing ? appeal

Terminal pitch direction

\ fall

/ rise

_ level

Accent and lengthening

^ primary accent = lengthening Pause …(N) long … medium .. short (0) latching Vocal noises (H) inhalation % glottal stop Quality

<A A> allegro: rapid speech <P P> piano: soft

<PP PP> very soft

Specialized notations

<L2 L2> code switching from Mandarin to English (( )) transcriber’s comment

(24)

244 Kawai Chui

Author’s address

Kawai Chui

National Chengchi University

P. O.Box 1–322, Mucha, Taipei, Taiwan 116 Email: kawai@nccu.edu.tw

(25)