Journal of Pragmatics North-Holland
8 (1984) 31-47
SPEECH ACTS, DISCOURSE CONNECTIVES *
31
STRUCTURE,
AND PRAGMATIC
Eddy ROULET
This paper demonstrates the need to reconsider the study of speech acts in a hierarchical model of discourse structure, integrating dialogal and monologal discourse. It emphasizes the decisive role of pragmatic connectives in the linking between discourse constituents at different levels.
Apart from a few noteworthy exceptions (in particular, the work of van Dijk, Ducrot, Edmondson, Franck, Labov, and Wunderlich), research on speech acts during the 70s has been confined to the description of isolated, fabricated examples, without reference to authentic discourse. Thus, in numerous works focusing on the relationship between form and function in indirect speech acts, the communicative context (often a fictitious one) is taken into account - in addition to the utterance -, but the co-text is neglected; the function of the act in terms of the modification of the relationship between the speakers (its illocutionary function, i.e. assertion, request, etc.) is described, but the syntagmatic relationship between this act and those which precede or follow it is ignored, except perhaps for elementary pairs of the ‘question/response’ type. By the same token, one finds many descriptions of the different types of markers of illocutionary function (cf. Roulet 1981) - especially for performative and modal verbs - but, except for the work of Ducrot et al. (1980) the question of pragmatic connectives (that is, of the markers of the relationship between acts in a given discourse) remains largely unexplored. When one attempts to describe the speech acts which constitute authentic (French) conversations and texts - as we have been doing in Geneva since 1979 - one inevitably is confronted with problems which go beyond the simple ‘question/response’ pair; in addition, one observes many forms - coordinative and subordinative conjunctions such as mais and bien que, adverbs and
* This research on speech acts and on pragmatic connectives on conversation was financed by the Swiss National Foundation for Scientific Research (Grant Nrs. 1.927.0.79 and 1.319.0.81). Author’s address: E. Roulet, Universitt de Gentve, Unitt de linguistique franqaise, CH 1211 Geneve 4, Switzerland.
0378-2166/84/$3.00
0 1984, Elsevier Science Publishers
B.V. (North-Holland)
32
E. R&et
/ Speech acts, discourse structure, pragmatic
connectives
adverbials such as pourtant and quand &me, interjections such as dkcidkment which, without being markers of illocutionary function, play a fundamental role in the linking of the speech acts which form real discourses, despite the fact that until now they have rarely been studied. * The study of authentic conversations rapidly reveals the existence of discourse units which are more complex than the speech act; on one hand, questions, requests, and responses can rarely be reduced to a single speech act; on the other hand, the units articulated by a mais or a finulement may be formed by several acts. The need to take these intermediary units into account (units that are situated somewhere in between the act and the discourse) has ultimately led us to postulate a hierarchical structure composed of at least three levels: exchange, move, and speech act. As this hierarchical model of discourse structure has already been presented in detailed form, with numerous examples of analyses of authentic conversations, in Roulet (1981), we will merely recapitulate here the essential hypotheses upon which it is based: 1. Each conversation can be analyzed into one or more exchanges (the existence of an intermediary level between the conversation and the exchange, and the nature of the relationship between exchanges are problems which will not be dealt with here). 2. Each exchange can be analyzed into moves, linked by illocutionary functions which are generally attributed to speech acts; thus, a question move might be followed by an information/response move, which might in turn be followed by an evaluation move. 3. Each move can be analyzed into a master act, possibly accompanied by exchanges, moves, or acts which are subordinated to it; the master act and the elements which are subordinated to it are linked by what we call interactive relations (preparation, justification, argumentation, etc.). The relationship between the constituents which form a discourse at different levels may be indicated by markers of illocutionary or interactive function which we call ‘pragmatic connective?, since they serve to articulate discourse units. If we apply this model to a brief, but relatively complex fragment of a conversation taken from the broadcast RADIOSCOPIE (a conversation between Jacques Chancel and Michelle Maurois concerning the latter’s father), the structure presented in table 1 emerges (where illocutionary relations are marked by solid arrows, interactive relations by dotted arrows). This analysis permits us to make the following observations: (a) the complex nature of the units which are generally considered as illocutionary acts of the question or response type is confirmed; here, in * Editor’s note (added in proof). See also the paper by S. Bruxelles et al., Dkidtment: classification dissimulke (Journal of Pragmatics 3 (1979): 127-50).
La
effect/ j’crois
qu’il a CtC heureux dans sa vie litteraire qu’il a eu de grandes joies /mais/ dans sa vie privee il a eu sa part de difficult& et de malheurs duns I’ensemble je n’sais pas si on peut qualifier sa vie de vie heureuse With en effect the interactive function of justification, tying the subordinated act to the master act of the embedded move, becomes explicit; mais articulates the entire embedded move and the act that follows it, as two opposed subordinated arguments, with the master act (marked by the now preposed duns I’ensemble) serving as a conclusion. (c) the presence of pragmatic connectives in a discourse, especially when they are grouped together as in the sequence et alors en effet, can mark the
34
E. Rot&t / Speech acts, discourse structure, pragmatic
connectrues
structure in a complex fashion at different levels. Our example gives rise to at least two questions: what is the reactive illocutionary function or the interactive function marked by each of these three connectives, and between which constituents (certainly not the same for the three connectives) do they create a bond? Although we do not yet have access to a systematic description of alors in its different uses as temporal connective, argumentative connective, and marker of conversational structure (see Auchlin 1981 for this concept), we can advance the hypothesis that afors (or done, which can be used as a substitute for it) serves to bind the subordinated moves, as arguments or preliminary conclusions, with the master act of conclusion which it introduces; in contrast to done (cf. Zenone 1982), alors does not present this conclusion as something which proceeds automatically from the arguments presented, but rather as the result of the position taken by the speaker. Alors is thus an interactive connective which operates at the level of the macrostructure of the move. A systematic study of the pragmatic functions of et is also lacking, despite the fact that it is without any doubt the most widely used connective of the French language. We hypothesize that et, in contrast to alors, does not tie the master act to the entire group of embedded moves which are argumentatively subordinated to it, but only to the curiously interrupted commentary of the preceding move (lines 13-15), or possibly to the whole of this embedded move (lines 11-15). The case of en effet is more complex; for this connective, however, we do have a preliminary description provided by Danjou-Flaux (1980). We can make a distinction between two possible functions: a reactive illocutionary function in dialogal discourse (in which case en effet may be replaced by effectivement):
Est-ce qu’elle a passe ses examens? and an interactive Vanessa
function
doit Ctre malade;
En effet (effectivement),
in monologal
discourse:
en effet, elle n’est pas venue travailler.
[l]
For the first use, en effet links two moves, of initiative (request for information) and reactive (confirmation) illocutionary functions, constituting an exchange; for the second, it marks an interactive relation of justification between the subordinated act and the master act of a move (this description of en effet remains provisional, because it does not take into account the interesting fact that it is the same form which assumes both of these functions). [l] I use the terms monologal, dialogal in order This distinction is explained below, p. 41.
to distinguish
them from monological,
dialogical.
E. Roulet / Speech acts, discourse structure, pragmutic
connectives
35
What is the function of en effet in the above example of a conversational fragment? The presence of an interactive marker of conclusion, introducing a master act, has already been noted, but does not seem to be compatible with the simultaneous presence of an interactive marker of justification introducing a subordinated act. This, in conjunction with the possibility of substituting effectivement (but not car) for en effet leads us to consider en effet as a marker of the reactive illocutionary function of confirmation, tying Michelle Maurois’ move as a whole to the initiating move by Jacques Chancel (to this latter one, a negative orientation - neither incompatible with the context nor with the intonation - must be attributed). Thus, the three connectives of the sequence et alors en effet function at different levels and articulate different constituents of the conversation. The same utterance may simultaneously be marked as a link with the preceding move of the speaker (et), as the conclusion of a more complex move (afors), and as a reactive confirmation of the preceding move of the other speaker (en effet). Our intention here is not to delve more deeply into the structure of dialogal discourse; numerous developments and illustrations of this type of structure have already been provided in the Special Issue of the Etudes de linguistique appfiquke (44, 1980) entitled; ‘L’analyse de conversations authentiques’, and, in fact, the vast majority of studies dealing with speech acts in discourse have taken conversation as their object (see Sinclair and Coulthard 1975; Labov and Fanshel 1977; Franck 1980; Edmondson 1981). We feel that it is more interesting to attempt to describe a type of discourse which has barely been studied until now: the monologal, written text. Taking as our starting point the hypothesis of Bakhtine (cf. Todorov 1981: 292), according to which monologal discourse is always to be considered within the framework of a dialogal structure, we view the monologal text as a move which is in fact a constituent of an exchange, even though this exchange may be implicit or fictitious. Thus, an editorial from Le Mutin such as the one entitled Tunisie: la politique de la France (reproduced as fig. 1) is a reactive move responding to an implicit question from the public or the political world about French politics, but also an initiative move which in itself calls for an evaluation on the part of the reader. We find, therefore, in this editorial a move structure with constituents at different levels which are tied by interactive functions - much like the one which emerged from the above analysis of Michelle Maurois’ response _ and this structure is in part marked linguistically by pragmatic connectives (these tend to be more frequent in written than in oral discourse, no doubt because of the need to compensate for the absence of prosodic indications). In order to extricate the hierarchical structure of a discourse of this length, with its numerous degrees of embedding, we have had recourse primarily to two types of heuristic tools: the insertion of a connective to render an unmarked interactive relation explicit, and the suppression of subordinated elements. The first of these means has already been used in the analysis of
VlJt
WLF
DU. MATIN l-
Tunisie : la politique de la France
B
-LE
POINT
DE WE
DU a MATIN
a~-
Les chances de Chapour Bakhtiar
IEN quclccalmc nc soil pastotalemcnl reven" dL
Gafsa, CDTunisie, il ne semble pas que I’c&alion du commando ait da consCquenar immtdntn sttr la rtabiltti du r&me. L’attaquc spcctaculairc de dimanche demier. qui cdncidait avow Ic se cond anmvcrsaire dc I’imeutr du 28 janvier 1978 d Tunis. map, aussi avec la p&mde rcligieuse du Moulcd (jaw anmvcrtirc de la naissance de Mahornet). prtsmtc certaincs similltudcs avcc la troubles graves qui cu. wnt lieu lc 20 novcmbrc deticr i La Mccquc e.? Arabic saouditc. La maniere dont Ic coup dc force a tti militaremcnr condut par Ic commando de Gafsa wait, Doton. trts proche de celle da rcbella de La Maque. II s’a&irtil d’une op&at~on~ ccrles planifitc et bicn prtparet. mais quelquc pcu suudwre, et dent le seul but wait de prcndre date. Au mien. il s’agirtit de contr6ler quelqun joun la ville pour dwutcr la conditions d’tm repli. II cst peu prnbable. en tout cas, quc la perronnc de Habib Bourpuiba, en vacanccs a fQ kilomitra au sud de Gal%. ut CtCla ciblc des arsaillants. fait, si le coup dc Cafsa confirme que da commandos tunisiens comprcnant da ressortissams Irb!ern s’cntraincnt cn Lihyc CI attendent. poour se manifcster. que la succession de Bourguiba sott ouvertc, on nc wit p&e pour I‘mswt qucl cowam pohtique ces commando, serancnt suweptibla de sou~cnir. Ccia bant. le gou\ernement franwis suit I‘affaire de trts pris lxs trcn~ navirm de guerre qti crcuscnt en MCd~tcrran& aura~ent, en effct, quttti Toulon &matthe SOLI.c’est-ldlrc au momcm mtme de l’attaque de Gafsa. S’agit-il sculemem d’une coincidcna ? D’autrc part. ar-il exactqur dew ations de transpon Transall et des hthzopt&es de combat Puma son, SW Ic point de gagner la Tumsie pour alder I’armk nationale ?
E
N
ECIDEMEh7, la iistc da questions au sujet de la * prtscnce franc&z H sur Ic continent affi&I e: ailleurs n’aura fait quc s’allongcr depuis I’mstallauon de Valtry &card d’lrtaing I I’Elyste. Quesdomqul resent d’ailleurs sans rtponsc ; ou.alors. lorsquc Ic gouvcrncmcnt ct I’Elys& acccptem de s’cxplrquer. il ut Uop tard pour cn debattre et I’opimon est placiz devant Ic ful accomplt. On relevera qur Lesitwrrogad~ns qut surgissent a propos de la Ttmisic coincident avcc d’autrcs questions. rat&s ella aussi sans rCponx. au sujet de I’mtcrvenuon dc w gcndarmn n frands I La Mecque. boxrc gouvcmcmcnt aurau-il inaugurC. saw oser I’avouer. une nouveUc poliuque de k canotmicre ?
D
Fig. 1.
A tentative d’assassinat, hier matin, dc I’ancien premier ministrc du chab, Chapour Bakhtiar revcndiqute par un groupe d’ewimistcz islamiques -, aura cu pour premier rCsultat dc nppelcr I I’opinion imernarionale I’existence d’unc opposition P I’imam Khomeini. Or, cet attcntat swicnt I un m* ment oil le rtgime en place. I TChiran. wait victime d’une s&ie de <‘cornplots n. Le sept%mc n’ttait-il pas annonci jcud, par le &nl du parti Toudch (communistes ozhodoxcs) ? Jour apr$s jour, enqutW%, arrestations. instructions des dossiers se suivent et Y ressem olent. i tel pcnnt que cettr chasse aw. comploteurs panit &ire la prmcipale. sinon I’uniquc acovtti dn au@ rites iianiennes.
D
A?8 ce climax d‘extrime contusmn. unc opposition e I’Cquipe de Khomeini pourraitxUe VOW ver Its bases d’une premiere Ctapc ? Chapour Bakhtw Ie pensc e, went de dtclarer P notrc confrtre le Mondr : N Lc chah a for@ de It&mime les armcs qti alIaicnt I’abartrc... Khomeinr agif cxactcmcnl de &me par son comportemcr~t : il cst name meillcurc propawde dam Je pays ! )>On a toutes Its raisons ntanmoins d’itrc sccpt~quo dcvant cct opiitime dc I’ancicn. et drrmer. prcmw mmistre du chah L’acdon d’un hommr cn cxil darn un appanement de Neuilly pcutellc peser lourd face aux clamcurs frCniliqua et fanatlqua qui montent drpuis bicntEt dix-huit mois da rue et dcs mosquees de TChiran ? Certes. I’Histoire offre “ucoup d’cxemples de brusques xtoumemcDts de Srtuatlon. Pourquoi ne pourrait-il pas y awn un coup d’mil i la rCvolutmn islamique ? Pour I’instant,, nen rx va darts cc sew ; au con&rc. La rCvolution tranienne est dc plus cn plus contap~euse. Ln gouvemcmenu da pays arabes du Prochc-Orient, comme cctn d’Eglple. dc Svrie ou d’Arabx saoudrte, dowent faire face 1 des diff&tCs intcrnes qui nc cessem de s’agpraver : panout s’Cl&ent des VOLX pour exiper le (( rctour i f’islam H. On imagmc mal surtout comment cc coup d’arrtt, s’il avail jamaz hcu. pourra~t dlrectement proftw P Chapour Bakhtiar.
E
on nc wit gukrc, I court tcrme. comma? I’lran pourrait SCdoter d’un rigime un wt 1011 peu dCmocratique. De lrois chows I’une. en cffct. Ou blen lc pays reste dam son Ctat actuel, soumis au riEorisme et au fanausme des n doccreurs du dogme relrg~eux ), Ou blen. deuxiemc h,pothtsc. les leaders de I’oppositmn ~nsullis a I’Ctranger retiss~sscni. derriere Chapour Bakhuar ou dernerc Ie eCntral Over i renverser le rtglmc khome~nlsre. cn s’appuyanr SW unr partie de I’armee. et dam ce cas ils ram6neront dam Icurs fourpons les hommcs du chah, et done Its mCthodn dlctatoriales dr CCdernler Duni& hypothtsc : leKremhn renouvellt avrc Teheran It u coup de Kaboul x Grm!ns considcrcni cettr hypothesc commc la piux plauslble. Dam I unmtdnt. il nc faudrait pas que la France. qui a fournr hw un marchepwd 6 l’lmam Khomcmi, f&w lujourd’hw la count echellc i Chapour Bakhtiu. (p. 10 et I1 : mm infurmatioro.) N hi,
E. Roulet / Speech acts, discourse structure, pragmatic
connectives
3-l
Michelle Maurois’ move. For the second, if we examine the beginning of the editorial, we see that the status of the act subordinated to the utterance introduced by bien que is manifested not only by its syntactic subordination, but also by the possibility of suppressing this utterance without modifying the coherence of the whole. If, on the other hand, we suppress the utterance introduced by il ne semble pas, the coherence of the text is destroyed; from this, we may deduce that the status of this utterance is that of a master act. As for the passage beginning with L’attaque spectaculaire.. . and ending with Les conditions d’un repli, its status as a subordinated move is attested to by the fact that we can suppress it and pass directly from il ne semble pas que l’optration du commando ait des conskquences immediates sur la stab&k du rkgime to II est peu probable en tout cas Taking into account the indications provided by these tools, our analysis of the structure of the first part of this editorial is schematized in table 2. At a basic level, the master act of conclusion il ne semble pus.. rests upon the subordinated act marked by the pragmatic connective bien que, the Table
2
si
‘__--->
$2 38 %.. ;y
le
coup
de
~afsa
conj’ime
que
des
cormandos
tunisieni:
---- imyvwuznt des ressortissants lybiens s ‘entrainen I et attendee, pour se manifester, que la succession I Bourguiba soi 1 ouverte, quel coumnt politique I on ne voit guBre pour Z’instant b--fcommandos seraient susceptibles de soutenir.
t en Lybie de ces
38
E. Rdet
/ Speech
acts, discourse
structure,
pragmatic
connectives
rejection of a negative argument, and the long subordinated move of justification which could equally have been introduced by en effet: L’attaque spectaculaire.. . . This move is itself formed by a master act, followed by a move of justification which could likewise have been introduced by en effet: la man&e dont... . This subordinated move is formed by a master act of conclusion (which could have been mared by done: II s ‘agirait d ‘une opkation.. .), which is preceded by a subordinated act of argumentation: LA manike dont.. . and followed by a subordinated act of restriction, marked by au mieux. The whole of the move beginning with Bien que and ending with b la cible des assaillants, whose structure we have just described and which could have served on its own to constitute a coherent text, is subsequently subordinated by the connective en tout cas to the conclusion II est peu probable que la personne de Habib Bourguiba.. ait 6tP la cible des assaillants. This whole, which could also have formed a coherent text, is in turn subordinated as a tentative conclusion to the master move constituted by the definitive conclusion introduced by En fait. This master move has its own internal structure, which is formed by a master act of conclusion, preceded by a subordinated act (marked by si) constituting a rejection of a negative argument. This first, constitutive move of the editorial forms a perfectly coherent text. In opposition to this first move, we find a second move, introduced by the pragmatic connective Celu &ant. This move is constructed around a master act of conclusion (the first utterance) which rests upon the subordinated move and act of justification, the interactive function of these being marked by en effet and d’autre part, respectively. This simple structure is represented in table 3. The last constitutive move of this editorial is more complex. It begins with an act which is clearly marked by dkidkment as a master act of conclusion resting upon several co-oriented arguments (cf. Ducrot et al. 1980). In this particular case, these arguments cannot be found simply in the co-text formed by the two preceding questions (had that been the case, one would have been
Table
3
Celc; itant,
et
des
Twisie
Le gout%r,ienent
h6licopLAws pour
aider
de
JFun~az’:;
combat
l’un?c’e
nxi:
Puma sent
nationaZe
:
1 *u:‘;ilirY2
SUE
de
Zc point
pIOL:s.
de
yagner
La
E. Roulet / Speech acts, discourse structure, pragmatic
connectives
39
led to regard this conclusion as the conclusion of the preceding move), but must be looked for in the context as well - that is, all the questions of the public about the government’s actions in Africa. Thus, pragmatic connectives serve not only to articulate discourse constituents, but also to articulate these constituents with that which is implicit, as the analyses of mais, mDme, d’ailleurs, and eh bien proposed by Ducrot et al. (1980) have demonstrated. What follows is a move formed by two coordinated acts, on the model of q ou alors q’, and introduced by the connective d’ailleurs. Ducrot et al. (1980) have shown that this connective articulates an argument, which is presented as a supplementary one, with a conclusion and with other arguments which may be implicit. Thus, this move presents two alternative, supplementary arguments to support a conclusion which appears explicitly in the text in the form of the syntagm sans oser I’avouer. The argument which these two supplementary arguments serve to reinforce is clearly the conclusion introduced by d&d& ment. This frequently occurring structure can be summarized as follows: an act of conclusion which, at the basic level, rests upon certain arguments, must be reinterpreted, at an immediately superior level, as an argument supporting a conclusion. The following act, introduced by On relkvera que.. , introduces a supplementary argument, which could have been marked by en outre, to support the conclusion suns oser I’avouer. We thus obtain table 4. We can now see how the three constitutive moves of the editorial that we have analyzed are articulated. The only marker of an interactive relation that we can observe at this superior level is celu dant. Unfortunately, we do not possess a systematic description of this connective, which, as it happens, does
Table 4
* VGE = ValCry Giscard d’Estaing.
40
E. Roulet / Speech acts, drscourse structure, prugmatic connectives
not even appear in French dictionaries. It is surely akin here to or, which introduces a new argument in a move, without indicating its orientation, and possibly to mais, which introduces an argument whose orientation is opposed to that of the preceding argument. The first move is implicitly oriented towards a conclusion such as ‘le gouvernement franqais n’a pas ci intervenir’, while the second is oriented by cela &ant towards the opposite conclusion (‘le gouvernement franqais semble vouloir intervenir militairement’), presented as stronger than the preceding one. These two implicit, opposed conclusions are combined with the final conclusion as two opposed arguments; this is verified by the possibility of inserting the pragmatic connectivefinalement, which has precisely this interactive function. This gives us table 5. This editorial, then, certainly has a move structure, formed by subordinated acts and moves at different levels which are linked together by interactive functions. Our model of hierarchical structure can thus be applied to monologal, as well as to dialogal discourse and, in both cases, allows a better understanding of the constituents at different levels, of the interactive functions which link these constituents to each other, and of the markers of these functions. Up until now, we have been working with a distinction between two basic types of discourse structure: _ the exchange, whose constituents are moves which are tied by illocutionary functions and may be articulated by markers of illocutionary function. This type of structure seems to correspond to what is traditionally called dialogue (or dialogal discourse); _ the move, whose immediate constituents are acts, moves, and exchanges subordinated to a master act. These elements are tied by interactive functions, and may be articulated by markers of interactive function; this type of structure seems to correspond to what is traditionally called a monologue (or monologal discourse). However, this distinction raises at least two problems: (1) there are discourses produced by a single speaker/writer whose immediate constituents are clearly tied by illocutionary functions; Table 5
,---________ I
on n?evoit gue‘repour l'instant quel courant politique les commxios intervenus en Tunisie pomraient souteniP.
yl 1\---- _._____. + ceh &ant, le gouvernement fmn&s au plan militaire. 41 I [Finalement]notre gouvenement imugumrait-il
!de la canonni&e
?
suit l'affaire de PBS
me muvelle
potitiqu-2
E. Roulet / Speech acts, discourse structure, pragmatic
connectives
41
(2) one frequently observes that in certain types of dialogal discourse (in particular, in interviews) moves are articulated by markers of interactive function. Concerning the first point, we found that in editorials in Le Matin produced by a single writer, the immediate constituents are indeed clearly tied by illocutionary functions. In the editorial entitled Les chances de Chupour Bakhtiar, reproduced on p. 36 above, we can distinguish three immediate constituents (a question move, an assertion/reply move, and an evaluation move); these are linked by illocutionary functions and form an exchange. Instead of enacting one move responding to the public’s implicit inquiry and calling for an implicit evaluation on the part of the reader (as in our first editorial), the author simulates an exchange by formulating, on his own account, not only the question, but also the assertion/reply and the evaluation. He does not do this through the use of reported speech which, by subordinating the words of others to his own discourse, would have created a move whose constituents are linked by interactive functions (compare the status of the subordinated move marked by dit-on in the first editorial). The three moves constituting the exchange are autonomous and could have been articulated by different speakers since they, in fact, have different addressers
PI. To account for this type of discourse, we propose that the traditional distinction between monologal discourse, produced by a single speaker, and dialogal discourse, produced by two (or more) speakers, be complemented by a distinction between monological discourse, whose immediate constituents are tied together by interactive functions, and dialogical discourse, whose immediate constituents are tied together by illocutionary functions. These two types of oppositions may appear in combination with each other, as table 6 demonstrates. Although we do not possess an authentic example of a dialogal, but
Table 6 Discourse
Monologal
MONOLOGICAL
Tunisie: la politique de la France.
DIALOGICAL
Les chances de Chapour Bakhtiar
Dialogal
Fragment de RA DIOSCOPIE
[2] In French, &zonciateur; we make a distinction between the speaker, who physically produces the utterance, and the addresser, who is the author of the message; one can make a similar distinction between the hearer of the utterance and the addressee of the message.
42
E. R&et
/ Speech acts, discourse strucfure,
prugmntrcconnectives
monological discourse, we can imagine the case of a board of examiners, with three professors successively introducing arguments oriented towards a common conclusion: A: L’exposC initial Ctait complet et clair. B: De plus, M. X a parfaitement rCpondu aux questions posCes. C: D’ailleurs, c’est le meilleur des candidats que nous matin. A: Done nous pouvons lui mettre la note maximale.
que nous lui avons avons
entendus
ce
In this example, the three speakers actually function as a single addresser, whose role is to evaluate the examination. It thus appears that, while the opposition between monologal and dialogal depends on whether there are one or two speakers present in the discourse, the opposition between monological and dialogical depends on whether there are one or two addressers. Until now, we have focused on moves which simulate an exchange and whose immediate constituents are therefore linked by illocutionary functions. But one also observes - especially in interviews - exchanges whose immediate constituents, while clearly linked by illocutionary functions (question, response, request of confirmation, etc.) are, at least at first glance, articulated by markers of interactive functions: mais, pourtant, quand mCme, en fait, au fond, alors, done, etc. An example of this is the following excerpt of an interview between a journalist (NO) and Professor Jacques RuffiC (JR), a specialist in demographic genetics, taken from Le Nouvel Observateur (Nr. 909, April, 9-16, 1982). We have condensed the interview, retaining only those passages which are necessary for an understanding of the articulation of the discourse. NO: JR:
NO: JR:
NO:
(...> Vous niez la selection naturelle? La notion d’adaptation est fondamentale dans la pens&e de Darwin. Mais, si la Glection se bornait SI Climiner tous les variants, il ne devrait . . que le type id&al. Or, l’observation plus subsister.. . , en definitive, montre que ce type idCal n’existe pas.. . Alors, comment expliquez-vous cette extreme variCtC face B la pression sClective? Pour le gCnCticien japonais Kotoo Kimura, la plupart des mutations n’apportent g I’individu ni dCfaveur sensible ni avantage particulier. Elles sont neutres. En ce qui concerne les gknes, plutat que d’Cvolution ordonnCe, il faudrait parler d’une sorte de dCrive 1aissCe au hasard. Kimura n’Climine pas complittement la sClection, mais lui confere le r61e d’un frein.. . . Dans votre livre, pourtant, vous ne vous ralliez pas entikrement B cette thbse du neutralisme.
E. Roule~ / Speech acts, discourse structure, pragmatic
connectives
43
JR:
NO: JR:
NO:
. Les genes constituent des elements d’un reseau extremement complexe.. . . Done, selon vous, la selection ne s’exerce pas directement sur les genes? Ce qui compte le plus, c’est la facon dont ces genes s’organisent entre eux pour former des systemes coherents, bien Cquilibres.. .Quand on a compris qu’elle [la selection naturelle] vise en realite toute une population, on s’apercoit qu’elle ne peut avoir pour effet que de maintenir sinon d’accentuer les differences entre les individus. [After two further contributions by JR, in which he develops additional arguments and furnishes other examples supporting his thesis, NO signals the end of the interview by saying:] En somme, pour vous le salut est dans la diversite? (...)
How can this exchange structure be reconciled with the preponderance of markers of interactive relations? To locate the problem, we will briefly recapitulate the process of expansion of the basic exchange described by Roulet (1981). In general, a reparative exchange is made up of three constituents: a move of initiative illocutionary function, a move of reactive and initiative illocutionary functions, and a move of reactive illocutionary function. However, if the second move does not respond in an entirely satisfactory fashion to the expectations of the first speaker (being a refusal or a refutation), that is, if it constitutes what Burton (1980) calls a challenging moue, it may set in motion a process of yielding to, respectively contesting the challenge, which entails the expansion of the exchange structure from three to five or seven constituents, or possibly even more. But it can also happen that the second move is neither supporting nor challenging, in the sense that it satisfies, albeit only partially, the initiator of the exchange (the response may be incomplete, or the conclusions to which it leads unclear, or the response may raise some objection). In the above exchange, an example of a reactive move which is unsatisfactory because it is incomplete is found in JR’s first reply, which introduces arguments tied together by or without drawing towards a conclusion; it forces the interviewer to ask JR, by means of a question move, to complete his reply. JR’s subsequent move is an example of a reactive move which is unsatisfactory because it raises an objection (the answer rests upon the hypothesis of Kimura, about which JR has expressed reservation); the interviewer is therefore led to ask JR, by means of a question introduced by pourtant, to respond to the objection raised by his reply. JR’s third move is unsatisfactory because the conclusion to which it leads, concerning the functioning of natural selection, is unclear; therefore, the interviewer is led to ask JR, by means of a move introduced by done, for a
44
E. Roulei / Speech acts, discourse structure, pragmatic
connectiues
confirmation of the conclusion that he has drawn from his reply. Finally, in the last move, introduced by the connective en somme, the interviewer, in order to close the interview, asks JR to confirm the conclusion which he, the interviewer, has drawn from the preceding moves. But how, in these different cases, can one account for the linking which, looking at the succession of moves constituting the exchange, seems to be due to linear illocutionary relations, whereas the presence of interactivity markers at the beginning of each reactive move marks interactive relations of subordination? For the case of objections introduced by mais, pourtant, or n&znmoins, the solution is quite simple and was already formulated by Ducrot in 1976: “Le schema general, selon lequel l’emploi de mais suppose un certain type d’accord avec un Clement precedent presente par P, s’applique aussi lorsque mais est en tCte de replique, et m&me lorsque le locuteur de mais contredit son interlocuteur. (Ce qui est accord& admis, reconnu, c’est alors l’acte de parole accompli par la personne a qui l’on s’adresse, acte consider+ comme un fait empirique que l’on enregistre; la coordination marquee par mais s’etablit a partir de cet enregistrement). Nous considerons done comme une forme d’accord le simple fait que la replique introduite par mais se presente comme la continuation d’un dialogue (le d&accord absolu Ctant le silence ou le coq-al’ane); il s’agit la Cvidemment d’un caractere qui n’est pas propre a mais et que l’on retrouve dans toutes les coordinations: il assure a la conversation le m&me type de continuite d’enchainement qui fait d’une succession de phrases un discours suivi.” (Ducrot et al. 1980: 126) In the interviewer’s move introduced by pourtant, we have an example of what is known as a ‘concessive’ structure: the speaker accepts the other’s move, but rejects the conclusion which can be drawn from it, according to the schema: JR: p
NO: (certes p), mais/pourtant/nCanmoins
4.
The journalist takes note of the explanation offered by the hypothesis of neutralism, but rejects the conclusion that could be drawn from it; his position that it is not an adequate explanation is based on the argument that JR himself does not adhere entirely to this hypothesis in the book that was the immediate occasion for this interview. The beginning of the reactive move in which the interviewer takes note of the other’s reply, can, as Ducrot suggests, be implicit, and it is for this reason that it figures in parentheses in the schema. It can nonetheless always be rendered explicit; after all, the links created by the connectives mais, pourtant, and nkznmoins ~ entirely consistent with their role as markers of interactivity are between two constituents of the same move, and not between two moves
E. Roulei / Speech acts, discourse structure, pragmatic
connectives
45
constituting an exchange. By analogy, as Ducrot suggests, we can treat the uses of alors, done, and en somme that appear in this interview in the same way. This global presentation of our approach to speech acts in discourse leads to the following observations: A. One cannot deal with the problem of the relations between speech acts in discourse without referring to other discourse units and to a hierarchical structure. Two contiguous speech acts in a discourse can support very different types of relations, often indirectly, depending on the structures into which they enter. One observes a direct link by interactive function only between the master act and the subordinated acts which immediately constitute a move. But, most often, contiguous acts belong to different moves: these moves either constitute a larger move and are tied together by interactive functions, or they constitute an exchange and are tied together by illocutionary functions; thus, the relation between the two contiguous acts is mediated by either interactive or illocutionary relations between constituents at a higher level. Hence, the problem of relations between acts must be treated as a specific aspect of the more general problem of relations between discourse constituents. Furthermore, the illocutionary functions of information, assertion, response, etc. which have traditionally been dealt with at the level of acts ~ must be restored to the level of relations between moves constituting an exchange. B. Ducrot (Ducrot et al. 1980) has clearly demonstrated that one of the major problems posed by the functioning of pragmatic connectives, even when there is an adequate description of the number and function of the constituents which they link together, is the recognition within the co-text (as well as within the context) of the units being articulated. Except for a few connectives, such as puisque and bien que, which belong to the grammatical category of subordinating conjunctions, the constituents articulated by the connective are not only two speech acts, but discourse units which are often much broader and more complex. The analysis of the hierarchical structure of the dialogue or the monologue, by determining the constituents which are possible at different levels, can contribute to the recognition and delimitation of linked units within the co-text. C. Of the illocutionary and interactive functions which can link discourse constituents, it is clearly the latter which are more important, in any event for monologal discourse. While we have had access, for some time, to several definitions and taxonomies of illocutionary functions, as well as to numerous descriptions of the principal markers of illocutionary functions in English, German, and French, we are only now beginning to define and classify the most common interactive functions: preparation, argument, justification, explanation, etc. The very terms used to designate these functions are at best intuitive and the distinctions between them are not always clear or sufficiently justified; this is the case, for example, for the distinction between the interactive relations of argument and justification. In addition, one wonders whether
46
E. Roulet
/ Speech
acts, discourse
structure,
pragmatic
connectives
one is dealing with the same type of interactivity in the relations between an argumentative act and a master act of conclusion, on the one hand, and between an act of tentative conclusion and a master act of definitive conclusion, on the other. The problem is aggravated by the absence of a complete inventory of the most commonly used connectives; besides conjunctions, whose status as connectives is entirely apparent, our languages possess a great number of adverbs, adverbials, and interjections, whose status as pragmatic connectives is completely obliterated by traditional descriptions (one has only to look at the definitions offered by dictionaries - even the most recent ones for words like d’ailleurs, finalement, and eh bien). As for systematic descriptions of the functioning of specific connectives, they are still extremely rare, for German and English as well as for French. We are only now beginning to understand the functioning of certain connectives which are commonly used in French and are important in argumentative discourse: quand mCme, pourtant, ainsi, aussi, apt-k tout, en tout cas, au fond, to cite only those examples upon which the Geneva research group is working at present (see Moeschler and de Spengler 1982; Schelling 1982; Zenone 1982; Roulet et al. 1984). If we are to arrive at a better understanding of the relations between speech acts in discourse, the moment has come to stop limiting our research to illocutionary functions, and to take into consideration the functions and markers of interactivity as well. (Translated
by Brenda
Bollag)
References Auchlin, Antoine, 1981. Reflexions sur les marqueurs de structuration de la conversation. Etudes de linguistique appliqued 44:888703. Burton, Deirdre, 1980. Dialogue and discourse. A sociolinguistic approach to modern drama and naturally occurring conversation. London: Routledge and Kegan Paul. Concession et consecution dans le discours. Cahiers de linguistique francaise 4. 1981. Geneva: Unite de linguistique francaise de 1’Universite. Danjou-Flaux, Nelly, 1980. A propos de de fait, en fait, en effet, et effecttuement. Le frangais moderne 48: 110-139. Differents (Les) types de marqueurs et la determination des fonctions des actes de langage en contexte. Cahiers de linguistique francaise 2-3, 1981. Geneva: Unite de linguistique francaise de TUniversitt. Ducrot, Oswald, et al., 1980. Les mots du discours. Paris: Minuit. Edmondson, Willis, 1981. Spoken discourse. A model for analysis. London: Longman. Franck, Dorothea, 1980. Grammatik und Konversation. Konigstein: Scriptor. Labov, William and Fanshel, David, 1977. Therapeutic discourse: psychotherapy as conversation. New York: Academic Press. Moeschler, Jacques, 1982. Dire et contredire. Pragmatique de la negation et acte de refutation dans la conversation. Bern: Lang.
E. Roule~ / Speech acts, discourse strucmre, pragmaric connectiues
41
Moeschler, Jacques and de Spengler, Nina, 1982. La concession ou la refutation interdite, approche argumentative et conversationnelle. Cahiers de linguistique fran$aise 4: 7-36. Roulet, Eddy, 1981. Echanges, interventions et actes de langage dans la structure de la conversation’. Etudes de linguistique appliqu& 44: 7-39. Roulet, Eddy et al., 1984. L’articulation du discours en fran$ais contemporain. Bern: Lang. Schelling, Marianne, 1982. Quelques modalit& de &ture. Les conclusifsfinalement, en somme, clu fond, de toute jaqon. Cahiers de linguistique franyaise 4: 63-106. Sinclair, John McH. and Coulthard, Ronald M.. 1975. Towards an analysis of discourse. London: Oxford University Press. Todorov, Tzvetan, 1981. Mikhail Bakhtine - Le principe dialogique. Ecrits du Cercle de Bakhtine. Paris: Seuil. Van Dijk, Teun A., 1980. Macrostructures. An interdisciplinary study of global structures in discourse, interaction, and cognition. Hillsdale, NJ: Erlbaum. Wunderlich, Dieter, 1979. ‘Was ist das ftir ein Sprechakt?’ In: G. Grewendorf, ed., Sprechakttheorie und Semantik. Frankfurt: Suhrkamp. pp. 225-324. Zenone, Anna, 1982. La cons&cution sans contradiction: done, par conskquenr, alors, ainsi, aussi. Cahiers de linguistique fran$aise 4: 107-141, 5: 189-214.
of Neuchltel (Switzerland). Eddy Roulet (b. 1939). Licence, Doctorat 6s lettres, University Professor of General Linguistics, University of Neuchltel, 1971-1977; since 1977, Professor of French Linguistics, University of Geneva. Editor of Cahiers de linguistiquefmn~aise, University of Geneva: Les diffkrents types de marqueurs et la dktermination des foncrions des acfes de langage en contexle (2 & 3, 1981). Concession et consCcuGon dam le discourse (4, 1982), Connecteurs pragmatiques et structures du discours (5, 1983). Co-author of Lhrriculation du drscours en franc& contemporain, Bern, Lang, 1984.