ABSTRACTThis　paper　presents　the　results　of　a　preliminary　study　of　direct　quotation

愛知淑徳短期大学研究紀要 第3σ号 1991149

APreliminary Sociolinguistic Analysis of

      ．

Methods of Signaling Direct Quotations

David Clay Dycus

ABSTRACT

This paper presents the results of a preliminary study of direct quotation strategies used by guests on an American late night talk show． A classifica−

tion system for distinguishing types of direct quotation is proposed and non−

standard direct quotation signals are examined．

lNTRODUCTION

    Direct quotation， a common device in written English， is encountered in spoken discourse as well， although indirect quotation is used more often． Grammatically appropriate forms of the verb say are the generally accepted lexical signals used to indicate the presence of a direct quotation in speech． However， researchers using

sociolingustic information gathering techniques have shown， both directly（Butters，

1980）and indirectly（Wolfson，1978；1979）that options other than the strictly gram−

matica1 say are in use in American English speech． This paper， based on recorded

      グ

television talk show interviews， presents and examines five non−standard forms direct quotation signaling used in actual speech situations， and one standard signal， gramma−

tical say ．

DIRECT AND INDIRECT SPEECH

    Speech reporting is a universal language phenomenon（Coulmas，1985）．1｝The dis・

tinction between direct and indirect quotation is a long−standing one in the West：

Traditionally， two kinds of reported speech have been distinguished：ora tio rec ta（direct quotation）and oratio obliqtta（indirect quotation）． The former

evokes the original speech situation and conveys， or claims to convey， the ex−

act words of the original speaker in direct discourse， while the latter adapts

course（Coulmas，1985：142）．

       −149一

This distinction between direct and indirect quotation is widely accepted， which a sur−

vey of English grammars and language textbooks would clearly show．

    Direct quotation does not receive a great deal of attention in grammars， probably

      タ

due to the fact that the form of the originai utterance does not undergo any change when reported． Indirect speech， being grammatically complex， is covered in much more

detaiL Celce−Murcia and Larsen−Freeman（1983）devote a whole chapter to indirect

examples of direct speech in these sources signal the quotation with the verb say

and quotation marks． Sustitutes for say ， such as demanded ， asked ， pleaded ， etc．，

also appear， but these are far mOre common in written than spoken English． No non−

standard forms of signaling direct speech are given． However， sociolinguistic resear−

chers have indicated that at least two other non−standard forms are in use among na−

tive American English speakers（Butters，1980；Wolfson，1978，1979）．

lNFORMATION GATHERING TECHNIQUES

    The way information is gathered is always a major concern of any researcher． A

person dealing with spontaneous speech must necessarily deal with LaboV s famous

Observer s Paradox ． He presents it as follows：      ．

We are left then with the observeピs paradox ：the aim of linguistic research in the community must be to find out how people talk when they are not being systematically observed；yet we can only obtain this data by systematic

observation （Labov，1972：181）．

The paradox itself is real enough， but an extra dimension to the problem is added by

L・b・v ・i・・i・t・nce th・t th・verna・・1・・、2）a1・・k・・w・a・n・t・・a1・pee，h。r sp。nt、。，。。、

speech， the non−self−conscious language purely produced only when someone is not

（knowingly）being observed， is the most desirable kind for the observer to record for

study（Labov，1972）．

    While the vernacular is undoubtedly a rich and productive vein of speech， it is

certainly not the only variety of interest and value to researchers， and Labov s extoll−

ing of the virtues of the vernacular should not be interpreted as a call to abandon stu−

dies of other sorts of language． This is especially true according to the precepts of Dell Hymes ethnography of speaking ．

    As Hymes states in his pioneering work， The ethnography of speaking is con−

cerned with the situations and uses， the patterns and functions， of speaking as an activity in its own right （Hymes，1962， in Wolfson，1989：5）． He mentions no specific

APreliminary Sociolinguistic Analysis of Methods of Signaling Direct Quotations 151

pattern or style， such as the vernacular， as being ascendant over all others． Hymes provides as a framework sixteen components which can exert influence on speech． The

first two items on his list are Setting ， ＿which refers to the time， place， and physical

    To Labov the two are a barrier to be overcome in the quest for vernacular speech． However， according to the ethnography of speaking， the two simply provide a

new situation with language unique to it and， by its uniqueness， worthy of study．

    For the purposes of gathering information recorded television talk show inter−

views were used． Labov（1972：183）admits the possiblity of obtaining useful data

183）．

    This opinion has not gone undisputed． Stubbs（1983）， in a discussion of socioling−

uistic methodology， notes Wolfson s claim that， in an absolute sense， there is no such thing as  natural  speech． She feels there is only languag・e natural to  んθsituation in

which the conversational historical present tense（CHP）workS in narrative discourse

（Wolfson，1982）． She describes the narratives in her corpus as being ＿told by native speakers of American English仇naturally occuntng speech situations（emphasis added）

（P．55）、

    In line with Wolfson s implicit acceptance of the naturalness of most speech pro−

duced in talk show interviews， such interviews form the data base for this study． Such

tuition， it can be said that they generally produce language closer， relatively speaking，

to vernacular that formal． While guests tend to be restricted in terms of occupation（e．

9．entertainers and sports figures）they are of both sexes and have a variety of racia1，

ethnic，． educational， and religious backgrounds． Because of their fame much of their background is already known or can found by consulting readily available magazines and books． There appear to be no linguistic constraints on guests other than that they

speak fluent or near fluent English and avoid the use of profainity．

DESCRIPTION OF THE lNTERVIEWS

    The data this preliminary study is based on come from 8 tape recorded talk show interviews in which direct quotation strategies were used． These 8 interviews were drawn from a total of 13 interviews．3）Although the sample is not statistically random，

一151一

no attempt was made to be selective in recording interviews． After being recorded the interviews were reviewed for examples of direct quotation signaling and those contain−

ing examples of it were transcribed．

DESCRIPTION OF THE lNTERVIEWEES

   Most of the interviews were conducted one−on−one， but on occasion couples were questioned together． In such cases each member of a couple was considered a separate interviewee as long as the narrative produced was not cooperative， which was the case in all joint interviews encountered． Furthermore， only those who actually used direct

quotation are considered in this study． For that reason interviews 5 and 6， shown be−

low in Table 1．， are represented as having a single participant even though couples

were present and active in the interviews．

   The result was 8 interviews with a single interviewee for each． Six of eight（75％）

of the interviewees were men ranging in estimated age from 30 to 60 years． Five were

white and one was Hispanic−American． The two women， one white and one black，

were both approximately 25 years old． With the exception of a radio announcer， all were actors or actresses． Judged according to native speaker intuition， they all spoke standard English， although the black woman a．nd hispanic man sometimes slipped into non−standard pronunciation for comic effect．

 Table 1．

InterVieWeeS

∬nt． No．

  1．

  2．

  3．

  4．

  5．

  6，

  7．

  8，

肱MMMFMMFM

^Est． Age

  30   40   60   25   35   45   25   30

Racial／Ethnic Bacleground   White   White   White   White   White   White   Black

  OccuPation    Actor    Actor    Actor    Actress    Actor Radio Announcer    Actress    Actor

UNITS OF ANALYSIS

   Casual conversation can sometimes appear to be chaotic， filled with interruptions，

asides， and half−formed sentences trailing off into nothing． Such distractions， not at all tolerated in written discourse， do not mean that casual conversation lacks struc−

ture． The situation is quite the opposite， according to Stubbs（1983）：

APreliminary Sociolinguistic Analysis of Methods of Signating Direct Quotations l53

It is possible to show that long chunks of apparently casual conversation，

stretching over several minutes， have overall structural organization． There Is a very simple structural claim which is very relevant here：stories have beginnings， middles， and ends． For example， jokes and stories do not just

start． They are often introduced＿（and）lsl imilarly， stories do not just stoP：

they are ended．（p．23）

    Based on native speaker intuition interviews were divided into what were termed topic units ． A topic unit corresponds to the topic and supporting information of a

single story， anecdote， or shorter utterance． Its length can vary from a few sentences to a much longer narrative． Unless uncompleted， a topic unit has a recognizable begin−

ning and end and can be traced to its end regardless of asides and interruptions that may occur．

    In most cases a topic unit is what would commonly be called a story ， but this

Considered in isolation， each smaller unit is also a story． To avoid this problem， the

（hopefully）more precise term topic unit was coined．

    Under this classification 28 topic units were recorded， which yielded 54 exam・

ples of direct quotation signaling in use．

TYPES OF DIRECT QUOTATIONS

    According to Celce−Murcia and Larsen−Freeman（1983：459）： When an English

directly quote the person＿or to use indirect speech（emphasis added） ．（p．459）

According to this definition， the reported speech is assumed to be a completely accu−

rate， or at least not purposely inaccurate， version of an utterance actually produced at some time in the past． While this is true of reported speech in its purest sense， it is not always true of direct quotation in casual conversation．

    Some of the interviewees used what， on the surface， was direct quotation when in

Consider the example below：

It s a thing stewardesses do， I think， before they go out． You know， they． 高≠汲

these really nice hairdos and then they go， How can I make em stupid to annoy everybody？ ． Then they stand there with this thing bobbin on their

head， going， Hi！How ya doin ？ ．（lnt．7）

一153一

・ノ

It is reasonable to assume that the speaker does not spend her time listening to ste−

wardesses as they do their hair． Although she gives the appearance of using direct speech it is obviously a ploy used to enhance her story and not an example of factual

direct quotation．

    Another problem is that direct speech can be used to recreate the language typical of a class of events and people． Strictly speaking， direct quotation can only apPly to the utterances of a single individual． Once applied to recurring or multiple speech acts

    Defining the relationship between a direct quotation and its source， then， is essen−

tial to fully understanding the function the quotation is actually serving， as Coulmas

（1985）clearly states：

Astyle of presentation is one thing， and the question of how a piece of

spokeq or written narrative relates to what a speaker actually said is

so difficult． The claim that part of what one says is an exact repetition of

       ロ

than in terms of truth functions relating an utterance to another．（p．141）

Examples of direct quotation encountered in this study support Coulmas observations．

To reflect percieved stylistic differences， examples of direct quotation were classified under one of the following headings：

1． Recounting a single speech event

2． Recounting recurring speech events 3． Imaginary speech event

In the first the speech is judged to reflect what was actually said． In the secound the speech is believed to be true of an occasion， situation， and／or gruop of people， but not of an individual speaking at a single point in time． This category includes instances of

quoting it）of a person or people of a given nationality， ethnic group， occupation， re−

gion， sex or sexual persuasion， religion， familial relationship（i、 e． the mother−in−

law ）， etc．， is（often unflatteringly）reproduced． The third category countains those ex−

amples of direct quotation which seem to be manufactured， not recalled． Of the 54 ex−

amples of quoting signals used，29（55％）recounted a single event，6（11％）dealt with recurring events， and 19（35％）were imaginary speech events．

    In all but one interview（lnterview 1）the strategy of recounting a single event

APreliminary Sociolinguistic Analysis of Methods of Signaling Direct Quotations l55

was used， making it the most widely distributed． It was the exclusive one used by speakers approximately 40 years old or older（Interviews 2，3， and 5）． It should be noted that although 30％of the occurrences were linked to imaginary speech events，

this strategy was used by only 3 people（lnterviews 1，7， and 8）， all approximately 30

direct speech for recounting recurring events（see Table 2．）．

SIGNALS OF D旧ECT QUOTATION IN AMERICAN ENGLISH

   The commonly accepted， or proper ， way of signaling the presence of direct

speech in both writing and speaking is by announcing it with saジ． In writing there is another signal， quotation marks， but in speech this is not an option（except for cumber・

some devices such as quote／end quote and I quote，＿ ． During this study other，

non−standard， signals were enco皿tered（see Table 2．）， most of which go unreported in grammars and dictionaries． Below is a description of the signals， both standard and

non−standard， used by the speakers in this study．

1．Grammatical say

       Here， the word say is used in the proper tense to announce a quota−

   tion． Since narratives are based on past events， the word usually takes a past

   Interview 7：

And they ask you where you re from and you say， Los Angeles ， and they say， Los Angeles！You（sic）got no culture．

   Grammatical say was by far the most commonly encountered signal in

speaker． It was also the preferred strategy of 50ut of 8 speakers．

2．CHP say

       Nessa Wolfson has done a great． deal of Work on what is termed the con−

   versational historical present（CHP）． CHP is typified by tense switching with−

   in conversational narrative． It functions， she says， ．．． to organize the story

   sent tenses highlighting the significant section of a story， narrative or joke

   or not a verb is used in its CHP form requires looking at an entire topic unit

      −155一

narratlve．

    CHP say is not well represented in the corpus， accounting for only

11％of the signals．

3． Go−as−saジ（Go say ）4）

       Ronald Butters（1980）first encountered examples of the use of go as a    substitute for say during fieldwork begun in the mid−1970 s． It is used in    informal narrative style and may be used to signal gestures that accompany    performed narratives as well as direct quotations． He concluded in 1980 that    tt was ln currency among whites and blacks under 35 years of age， indicat−

   ing that it is a relatively recent phenomenon， although his informants were

   not conscious of it being new in any way． The fact that no definition for go    as a substitute for say in direct speech apPears in any dictionary supPorts，

   in Butter s opinion， the view that it is a recent phenomenon．

      Table 2．

Number of Quotation Strategies Selected

Signals ＃1 ＃2 ＃3 ＃4 ＃5 ＃6 ＃7 ＃8 Total

Grammatica1 say 7 5 3 2 2 5 5

CHP saプ

^{3 3 6}

Go−as−saジ 4 2 4

CHP Go−as−saジ

¹

Zero lexical signa1 2 5 7

Other 1 1

4

5 3 5 2

    Another important point is that， while go−as−say can be used in

direct discourse（ He went， You must be kidding ．）， it doesn t appear in in−

direct discourse（ He went that I was kidding． ）or interrogatives（ What did he go next？ ）（Butters，1980；Shourup，1982）． Go−as−say clearly func・

tions to signal direct quotations and has an additional advantage， according

to Shourup（1982）of being less ambiguous that saジin certain situations．

Take， for example， the sentence（spoken），ノbhn said theαticident was my fault．

To the listener there are two possible interpretations：（1）it is indirectly re−

ported speech（ John said the accident was my（the speaker s）fault）， or，（2）it is a direct quotation（ John said， The accident was my（John s）fault ．）． By

 ゜      ！1   ．     

slnce go−as−say，1n lts present usage， cannot apPear in indirect discourse．

    There are 10 instances of go−as−say signaling in the corpus， all in tenses grammatically appropriate to the time frame of the narrative， account一

APreliminary Sociolinguistic Analysis of Methods of Signaling Direct Quotations l57

ing for 19％of the totaL． lts use was restricted to speakers of approximately 30years of age and under（Interviews 1，7， and 8）of all ethnic backgrounds，

which supports the contention that it is a relatively recent linguistic develop−

ment（Butters，1980）．

4．CHP go−as−saジ（Go say ）

       In this case the go−as−say signal described above appears in a CHP

    form（see CHP say above）；that is， in the present tense form in a past tense     narrative． Butters notes that the narrative go is most often used in the pre−

    sent tense， even when past time is indicated （1980：305）． In other words， it     most often appears in his transcripts as part of CHP narrative．

       CHP go−as−say appears only once， accounting for 2％of the total sig・

    nalS．

5．Zero lexical signal

       Acommon device in literature to avoid excessive use of saジto signal

   direct speech is to develop the context in which the utterance or exchange is    to take place and then enter immediately into it， omitting lexical signals and

   assumed understanding of these conventions of writing）to delineate the    quotation from the rest of the text． In speech， where quotation marks are not

    non−verbal signaling is also possible． Encountered in this study were exam・

    ples of direct quotation in which no lexical signals were used． In these cases

    （and， occasionally， decreased）volume of voice， and attempts to imitate the     voice of another person．

       Zero lexical signal appeared 7 times， accounting for 13％of all signals． It

    7and 8）and each case signaled an imaginary speech act．

6．Other Signals

       An apParently idiosyncratic use of

    interview（Interview 8）：

№盾煤h @or  have got  apPeared in one

We could adopt． We could see Sally Struthers． You know，

she s＿she s always got， Please help help little Manuel from Costa

Rica． Please help him because he s going to steal your hubcaps．

一157一

   This usage of got was only seen once，

quotation signals recorded．

making up 2％of the direct

      Table 3．

Relationship Between Direct Speech Type And Signal

∫ 劉α1 A 8 c ^Tofα∫

Grammatical saジ

22（41％） _{3（06％） 4（07％）} 29（54％）

CHP saジ

6（11％） 6（11％）

X0−as−say ^{3（06％） 7（13％）} 10（19％）

CHP go−as−saピ 1（02％｝ 1（02％）

Zero lexical signal 7（13％｝ 7（13％）

Other

_{1（02％） 1（02％）}

Total _{29（54％） 6（11％） 19（35％） 54（100％）}

A．Recounting a single speech event B．Recounting a recurring speech event C．Imaginary speech event

DIRECT SPEECH TYPES AND CHOICE OF SIGNAL

   Table 3 shows the frequency of quotation signal relative to the type of direct quotation Used． Grammatical say is found in all three types， but was used most often ln recountlng single speech events． It is interesting to note that signals other than

grammatical say tended to be chosen when speakers were producing imaginary

speech events． It should be noted that 50f 7 examples of zero lexical signal were from asingle interviewee， and the use of it may be idiosyncratic．

   Grammatica1 say and CHP say together accounted for 65％of the quotation

Though go−as−saジand CHP go−as−say were far less commonly used， they still

part ．of the repertoire of speakers approximately 35 years old or under、

DISCUSSION

   The sample in this stady is small and it would be premature to make any sweep・

ing conclusions from it． However， some preliminary conclusions can be drawn to guide future research． Butters （1980）observation that go−as−say （by projecting his ori・

ginal age estimate to the present time）is used by speakers under 45 years of age is

supported by this study． Further sociolinguistic research into this direct quotation sig・

nal is warranted to determine how deeply ingrained it has become in American En−

glish speech、

   It is interesting to note the relationship between direct quotation signals and the

APreliminary Sociolinguistic Analysis of Methods of Signaling Direct Quotations l59

narratives in which they appeared． All of the instances of go−as−saジ， CHP go−as−

saジ， and zero lexical signal in this study appeared in what Wolfson calls the per−

formed storジ． The performed story is defined by the the presence of some or all of

the following features：direct speech， asides， repetion， expressive sounds， sound effects，

and motions and gestures． It functions to ＿structure the experience from the point of view of speaker and to dramatize it （Wolfson，1978：216）． Since the interviewees in this study were performers， presumably selected in part for their ability to entertain，

it is not surprising that they delivered many of their narratives as performed stories．

    Discussion is also necessary regarding the classification system for direct quota・

tions developed during the course of this study． It is provisional and in need of refine−

ment， but points out important distinctions that need to be made when examining

（1982）points out， such intuition is not always a reliable indicator of the discourse rules guiding language． However， the other extreme， a truth−functional analysis of direct speech presents a completely new set of hazards， since language can be used to invent any number of fictions masquerading as reality， can carry meanings not at all apparent in surface structure alone（Stubbs，1983）， and also allows a range of stylistic possibilities that affect interpretation（Coulmas，1985）．

    Perhaps the best way out of the dilemma is to recognize both the strengths and limitations of a researcheピs intuitive knowledge of a language and cultural system and use it， with a discerning eye， to discover and develop more precise formulations

（Stubbs，1983）． Such intuitions are， in a way， hypotheses that can be explored logical−

ly and scientifically． They can be a good place to start sociolinguistic inquiry， as is the case in this study． Hopefully， future research will lead to a comprehensive finish．

    If further studies bear out the findings presented here there are positive implica−

tions for TESL instructors and for those involved in materials development． Gramma−

tical say accounts for the majority（54％）of the direct quotation signals in this pilot

study and was used by all of the interviewees． If a larger sample of interviews yields similar results in the future， it would indicate that teaching direct quotation according

guage）and what is actually spoken（the sociolinguist s understanding of the language）．

The non−standard signals probably do not pose great problems for students since they are， on the surface， single slot substitutions in an otherwise fixed and easily rec−

ognizable pattern． The complexities underlying their use， most notably the ability to recognize， interpret， and possibly produce， performed stories， would， however， be an

important part of the communicative competence upper−intermedicate and advanced

fessionals， in turn， would find explict knowledge of the structure of performed narra・

tives and the function of CHP usefuいn instructing higher−level students．

       −159一

CONCLUSION

    Direct speech， deceptively simple when viewed at the sentence level， becomes quite complicated when studied in social and discourse contexts． The best way to gain

true understanding of its many facets is by observing it in use in social context， as a function of discourse， not as a grammatical artifact examined out of context． Television

dominated．

    To deal with apparent differences in content and function， a three−part classifica−

tion system for direct quotations was devised． It is tentative but will hopefully lead to discussion and progressin our understanding of the complexities involved in direct

       Notes

1） Coulmas main point， however， is that the realization of this language universal is highly    variable in terms of mood， tense， word order， and complementation．

2） Other terms， noted by Stubbs（1983）， are unplanned and casual 、 At the oppsite end of    the spectrum is langage variously termed as contrived ， artificial ， invented ， intro・

   spective ， and hypothetical ．

3） Originally there were 14， but one was excluded because the interviewee spoke British，

   not American， English．

      ACKNOWLEDGEMENTS

This article is a revised version of a paper presented at the Northern New England TESOL Conference in Lebanon， New Hampshire， November 3−5，1989． I am indebted to Prof． Wil・

liam Perry for his patient guidance and instructive comments on the original research that led to the first version of this paper． I also wish to thank Seiko Ono for comments on the present article． Any shortcomings which remain in spite of their efforts are， of course， my own．

      References

Butters， Ronald R．（1980）． Narrative Go Say ． American Speech，5（4），304−307

Celce−Murcia， Marianne， and Diane Larsen−Freeman．（1983）、 TんθGrantmar Booカ． Newbury    House， London．

Coulmas， Florian．（1985）． No One Dies in Shangrila ． In D． Tannen and J． E． Altais（Eds．），

  ．」Linguistics（Georgetown University Roundtable on Languages and Linguistics）． George−

   town University Press， Washington， D． C．

Labov， W．（1972）． The Study of Language in its Social Context ． In J． B． Pride and J． Holmes

   （Eds．）， Sociolinguistics． Penguin Books， New York．

APreliminary Sociolinguistic Analysis of Methods of Signaling Direct Quotations 161

Quirk， Randolph， and Sidney Greenbaum， Geoffrey Leech， and Jan Svartvik．（1972）． A Contem・

   porary English Grammer． Longman， Longman， London、

Schourup， Lawrence．（1982）． Quoting With Go Saジ、 American SPeech，55（2），148−149．

Stubbs， Michae1．（1983）． Discourse、Analysis． Basil Blackwell， Oxford．

Wolfson， Nessa．（1978）． A Feature of Performed Narrative：the Conversational Historical     Present ． Language in Society，7（1），215−237．

Wolfson， Nessa．（1979）． The Conversational Historical Present Alternation ． Language，55     （1），168−181．

Wolfson， Nessa．（1982）．（1982）． On Tense Alternation and the Need for Analysis of Native     Speaker Usage in Second Language Aquisition ． Languag・θLeaming，32（1），53−69．

Wolfson， Nessa．（1989）． Perspectives： Sec励nguistics and TESOL． Newbury House， New York．

一161一