Professional Documents
Culture Documents
ABSTRACT
I. INTRODUCTION
The present paper concerns the discourse marker puis,1 as it is used in spoken
French. The following example is representative:
(1) . . . on en a vu quand meme pas mal des Egyptiens puis on avait le personnel
Egyptien et tout sur le bateau et tout . . . (VE: 37)2
The analysis presented here forms part of a research project on the
distribution and semantics of discourse particles such as puis, alors, bon, ben,
eh bien, etc. in spoken French (see also Hansen, forthcoming a and b), and is
based almost entirely on authentic, non-elicited examples drawn from taped
1
In some of my examples, the marker is realized phonetically as [pi], which I consider a free
variant.
2
I have taken the liberty of 'simplifying' the transcriptions at various points, for the
convenience of the untrained reader: thus, phoneticized orthography has been changed to
conform to the standard, and many hesitations, repetitions and pauses have been deleted when
not in the immediate environment of puis. In all examples, the marker commented on is in
italics. The following symbols have been used: & and && indicate the beginning and end of
overlap, respectively (where several instances of overlap follow one another, § and §§ are also
used to distinguish them);, and ,, indicate short and somewhat longer pauses. I would like to
thank Professor Mary-Annick Morel of the University of Paris III, for allowing me access to
the corpora used.
31
M.-B. Hansen
32
Puis in spoken French
framing devices. Koch and Oesterreicher (1990: 51-71), seemingly the latest
manifestation of this line of thought, set up seven different categories of
'discourse words', and it seems that not only may one and the same form be a
member of several of these categories, but the members of each are
essentially interchangeable. Now, quite apart from the fact that the typology
postulated lacks a unifying principle (some categories seem to have a mainly
structural function, others an interactional, a cognitive, or a modal one), this
approach seems to be of little explanatory value, and it is my claim that (1)
different items do have distinct coded content, expressible, as noted above,
in terms of instructions, and (2) it is possible to formulate these instructions
so as to capture all uses of a single item, thereby avoiding an explanation in
terms of homonymy.
2. SYNTAX OF PUIS
Diachronically, puis originates in the Latin preposition/time adverb post.
According to the Tresor de la langue francaise (1990), French puis is derived,
not from post itself, but from the non-attested comparative form *postius.
Chevalier and Molho (1986: 29) question this received wisdom, and propose
instead to derive the particle from Latin postea, since, in their analysis, puis
involves an anaphoric element (= ea, surviving as the 'i' in puis). On phonetic
grounds, there is apparently little to choose between the two competing
etymologies (S. Hendrup, personal comment), and that proposed by Cheva-
lier and Molho would seem to fit my own analysis better (cf. infra). The item
apparently retained both its prepositional and its adverbial use into sixteenth-
century French (Gougenheim, 1951: 173-174). In Modern French, however,
the prepositional use has been entirely lost, and time adverbs being more
'lexical' than prepositions, one might say that puis has, in that sense, been
'degrammaticalized' (cf. Heine et al., 1991: 4). However, it has, in another
sense, been more strongly grammaticalized, in so far as its position is now
fixed clause-initially, whereas Old and Middle French adverbial puis could
occupy various positions in the clause. The particle has, in other words,
suffered a loss in syntagmatic variability (cf. Lehmann, 1985: 308). Both its
prepositional and non-clause-initial adverbial uses have in Modern French
been taken over by depuis.
Now, what is noteworthy is that in contemporary spoken French, the
temporal meaning ofpuis seems to have more or less disappeared, yielding to
a use of the marker as essentially an additive conjunct. Thus, of the 122
examples in my corpora, only 27 (i.e. less than a quarter) involve any kind of
succession in time, and what is more, in many examples the temporal
interpretation is only possible, but not necessary, whereas when it is
necessary, puis is often accompanied by a time adverb such as apres or ensuite.
Witness the following examples:
33
M.-B. Hansen
No succession in time:
(2) . . . on voit des politiciens de gauche qui vous disent c'est extraordinaire c'est un
vrai portrait sincere de Mitterrand etpuis elle a et elle a pas parle de la vie privee de
Mitterrand . . . (MP: 5)
(3) A: on est rentrees a la fac et puis y avait des livres sur les etageres
B: oui
A: tu te souviens
B: oui
A: etpuis le gars i/je lui ai dit mais je vous l'acheterai demain, et il a dit mais non
moije vous fais pas confiance, tu te souviens (CT: 11)
34
Puis in spoken French
35
M.-B. Hansen
And our problems do not end here: Nolke (1989, 1990) establishes three
categories of sentence adverbials, viz. 'connectors' (FR: 'adverbiaux connec-
teurs'), 'illocutionary adverbials' (FR: 'adverbiaux d'enonciation'), and
'locutionary adverbials' (FR: 'adverbiaux d'enonce'). The first group is said
to connect the host utterance with one or more other utterances (or utterance
acts). The second type of adverbials bear on various non-truth-conditional
aspects of the utterance act, without establishing a direct relation to the
co-text. And, finally, items of the third category comment on the propo-
sitional content of the utterance, for instance by indicating the speaker's
attitude towards the proposition expressed. The adverbials of each category
may then in principle be distinguished with the aid of formal tests, although
Nolke does note (1990: 26) that it is often difficult to separate connectives
from illocutionary adverbials in this manner, since, in a great many cases, the
former also bear on the act of utterance to some extent.
One of the tests mentioned in the article (1990: 23-24.) is a positional one.
In the (rather unlikely) sentence 'Entre nous, c'est done sans doute toujours
exactement a midi que le gardien quitte la banque', we find that the normal
order of adverbials is as follows: illocutionary adverbial, connector,
locutionary adverbial, frequency adverbial (not a sentence adverbial), and
'paradigmatic' adverbial (a contextual, but not a sentence adverbial).4 Now,
by this criterion, puts should be an illocutionary adverbial, and not a
connective, as I have claimed above, since, in the sentence quoted, puis could
replace entre nous, but not done. Although puis is intuitively felt to be a
connective, this does not necessarily invalidate the positional test. Two (not
necessarily mutually exclusive) explanations are possible: either puis strad-
dles the categories of connectives and illocutionary adverbials (this possi-
bility is implicit in Chevalier and Molho's (1986) analysis, cf. infra), or it is
closer than most conjuncts to being a true conjunction.
The latter explanation is supported by the fact that, in the collocation et
puis, the two elements cannot be separated, neither by a pause, nor by
another element, contrary to what is the case for ensuite, for instance (cf.
Blumenthal 1975: 324). This could be a sign of incipient coalescence, which,
according to Lehmann (1985: 308) is one of the formal characteristics of
grammaticalized elements. Moreover, if puis has some of the properties
usually associated with conjunctions, we can understand the inacceptability
of (13) above as being syntactic in origin, rather than having to do with
whether puis contributes to propositional content or not, since it would be
odd for a coordinating conjunction {puis) to follow immediately upon a
subordinating one (que) in the same clause. In fact, some grammarians do
classify puis with the coordinating conjunctions (Van Hout, 1974: 336-337).
On the other hand, such a classification is questionable, on the grounds that
coordinating conjunctions cannot normally be combined, either (cf. Dik,
4
For a definition and analysis of'paradigmatic adverbs' in French, cf. Nolke (1983).
36
Puis in spoken French
1968: 34), and puis, as already noted, is readily combined with et, whose
status as a conjunction is not in doubt. 5
What we should perhaps conclude, given these facts, is that the bound-
aries between word-classes may be fluid, allowing puis to behave in some
respects like a time adjunct, in others like a sentence adverb, and in others
again like a conjunction. Seen in the context of a grammaticalization hypo-
thesis, there is in fact nothing particularly strange about this, and the marker
at its present stage might be described as a 'hybrid form', that is to say that it
has reached a stage of evolution where it does not quite belong in either its
source category (time adjunct) or its target category (conjunction) (cf.
Heine et al., 1991:231-233).
Based on the examples in my corpora, I would like to argue that puis, when
functioning as an additive conjunct in spoken French, indicates, on the one
hand, that two and only two elements are being coordinated, and, on the
other hand, that these two elements should be interpreted as separately rele-
vant, in the sense of Sperber and Wilson (1986).
In support of the first part of this hypothesis, Vet (1980: 149-150) analy-
ses temporal puis in d'abord . . . puis as invariably introducing the last
element of an ordered series (in contrast to ensuite, which merely indicates
posteriority), with d'abord introducing the first. A construction such as
d'abord Si, puis S2, ensuite S3 would then, according to Vet, be semantically
anomalous. If the temporal aspect is left out of consideration, puis then
comes to mark the last element in a coordination. This, of course, entails
that if two or more utterances in a sequence are all introduced by puis, then
the relationship between them must be hierarchical, rather than linear, and
this seems, in fact, to be borne out by my data (cf. section 3.1, exx.
(23)-(26)). If puis, moreover, links only two elements, then it must be the
case that either whatever precedes or whatever follows the marker in a par-
ticular discourse move can be understood as forming a whole, even if it
consists superficially of two or more parts. In this, puis is unlike et, as shown
by the following examples (inspired by Laurandeau 1983: 20), where puis
seems to indicate that the conjoined adjectives belong to different concep-
tual categories, whereas et seems to place them in the same over-arching
category:
(14) Le drapeau franc,ais est bleu, blanc et rouge
(15) ?Le drapeau francais est bleu et blanc, et puis rouge
(16) ?Ce drapeau est bleu, blanc, rouge et effrange
(17) Ce drapeau est bleu, blanc et rouge, et puis effrange
5
On the other hand, it does not seem to accept combination with the other coordinating
conjunctions mats, oil and car (if the latter is, indeed, a coordinating conjunction).
37
M.-B. Hansen
38
Puis in spoken French
39
M.-B. Hansen
des livres a grand spectacle pour pouvoir faire vivre ses livres difficiles (h) et il
avait Malraux euh il avait des gens qui l'ont sauve (h) et puis en meme temps
dans le meme temps dans la meme collection (h) il a publie des collaborateurs
notoires et qui ont ete fusilles ou qui se sont suicides
A: ?? enfin il a ete mille fois plus habile que Bernard Grasset
B's utterances in (18) and (18') may be equivalent as to truth-conditional
meaning, but as the oddity of A's reaction in (18') shows, they are
pragmatically quite different.
The property of being able to mark an independently relevant act of
utterance can be seen to grow out of the temporal use of puis: linguistic
expressions being linear, such a new and independent act of utterance
normally succeeds the previous one in time. 7 The use of puis as an additive
conjunct may then perhaps be analysed as an example of a 'metaphor of
grammar' (Heine etal., 1991: 2O4ff), where a succession of events in the real
world is used as a vehicle for the expression of discourse structure, the
vehicle being of a less abstract nature than the topic.
The hypotheses developed here must now be applied to examples from
my corpora. The examples fall into three categories: occurrences of puis
between two semantically complete utterances by one speaker; occurrences
of puis at the very beginning of a turn at talk (where the marker connects its
host utterance to the turn of the previous speaker); and, finally, occurrences
of puis between constituents of a single sentence.
7
I use the qualifier 'normally', because the two utterance acts may sometimes be fused into one:
Chevalier and Molho (1986: 27) note that to the extent that an utterance such as Le marquis,
40
Puis in spoken French
There are two aspects to the argument here: the utterance preceding puis
achieves relevance by confirming part of speaker A's initial assertion, and
thereby increasing the strength of any assumptions derived therefrom. The
utterance following puis adds new information, which nevertheless takes its
point of departure in the previous utterance (this is marked explicitly by the
adverbial en mime temps). The example moreover demonstrates that conver-
sational puis does not have temporal succession as an inherent semantic
feature, since its combination with en mime temps appears to involve no
contradiction.
(19) A: . . . y a-t-il ou non deux lectures de cette fameuse declaration signee il y a
tout juste une semaine,, Roland Leroy
B: ben a mon avis y a pas deux lectures y a une seule lecture, et Georges
Marchais en parlant hier a la S. K. F. d'lvry s'est inscrit tout a fait dans le sens de
la declaration qui confirme les accords de 1981 parce que je rappelle que le
contenu de la declaration commune de la semaine passee, de la semaine demiere,
ce sont les accords de 1981, et: puis euh la declaration elle-meme indique
clairement de quoi il s'agissait. . . (VS2: 20-21)
Here we see an example of how utterances introduced by puis may clinch
an argument. I do not believe, however, that such a 'clincher'-effect should
be part of the semantic description of the marker, since it is not discernible in
all examples, and seems in any case to be simply a consequence of a natural
rhetorical strategy consisting in saving stronger arguments for last. Dis-
course markers like puis are not, in my view, inherently argumentative,
unlike items such as mais, parce que and others, which have been studied
primarily by Oswald Ducrot and his collaborators (cf. Ducrot et al. 1980;
Anscombre and Ducrot 1983), and which are said to mark utterances as
constituting either an argument or a possible conclusion in discourse.
(20) A: mais c'est vrai que pour la construction des pyramides c'est probable & que:
&&
B: & c'est possible && que ce soit uniquement le peuple § mais j'ai lu des
bouquins moi un petit peu parce que j'aime bien lire §§
A: § pas uniquement le peuple alors bon euh §§
B: c'est des bouquins qui parlent un peu d'Egypte tu vois j'en ai lu un peu en
rentrant parce que ca stimule un peu qa, (h) etpuiij'ai vu que: apparemment c'est
vrai ils disaient que c'etaient les paysans pendant les periodes de crues ouais (VE:
43-44)
The context of this example is a discussion between two women, one of
them (A) a history teacher, and the other one (B) having just returned from a
trip to Egypt, about whether or not the Ancient Egyptians had slaves. The
puis Charlotte entrerent is grammatical, it may only be understood as the answer to two
successive questions, as in: - Que fit le marquis? - II entra. - Puis que fit Charlotte? - Elle entra
(aussi). This may perhaps be made clearer by the following example: Jean et puis Marie sont
maries, which can only be understood to mean that they are not married to one another.
41
M.-B. Hansen
8
Wherever examples from my corpora are commented on, pronouns refer to the actual sex of
the interlocutors. Elsewhere, 'she' is used to refer to speakers, and 'he' to hearers.
42
Puis in spoken French
(22) . . . dans les raisons de cet echec il y a evidemment la position prise par la
Grande-Bretagne mais enfin on peut dire que ca fait des annees que c,a dure, il y a
effectivement comme le disait Amouroux tout a l'heure le fait, que entre le
Chancelier allemand et le President de la Republique franchise il n'y a plus les
memes rapports qu'il y avait autrefois, entre le chef d'etat allemand et le chef
d'etat francais, puis y a aussi probablement euh,, une faiblesse,, franchise,, une
faiblesse dans la negotiation (VS2: 5)
This is only apparently so, however. Of the three reasons cited for the
failure of the Athens summit of 1983 (referred to by cet echec), the first one is
actually rejected by the speaker as ultimately irrelevant, and only two are
retained as truly explanatory.
In all the following extracts, we find puis introducing several utterances in
a row, which, if my analysis of the marker is correct, is an indication that the
structure of the discourse is hierarchical:
(23) . . . moije suis absolument d'accord avecJean-Didier c'est vraiment la reussite
de ce livre c'est que on voit des politiciens de gauche qui vous disent ah c'est
extraordinaire c'est un vrai portait sincere de Mitterrand a. puis elle a et elle a pas
parle de la vie privee de Mitterrand elle a pas bon et puis a droite on nous dit ah
c'est extraordinaire qu'est-ce qu'elle a foutu a la gueule de Mitterrand (h) et c'est
le meme livre (MP: 5)
Figure 1
43
M.-B. Hansen
T1(A)
T2(B)
T3(A)
T4(B)
T5(A)
T6(B)
T7(A)
Figure 2
44
Puis in spoken French
T5 A: apres on va revoir tout qa au fur a mesure, puis apres l'apres-midi ils nous
ont emmenes a Saqqarah [. . .]
T6 A: et elle bon eux ils disaient le plateau de Guiseh, puis alors bon done on a
fait cette visite, et puis alors c,a e'etait le premier jour je crois, et le lendemain
matin on a du aller a Memphis
T 7 ® : oui
T8 A: etpuii a Saqqarah,, oui, et l'apres-midi on a visite Le Caire mais la partie
musulmane & parce que bon euh &&
T9 B: & oui d'accord &&
T i o A: y avait tellement de mosquees puis e'est interessant done on a fait cette
partie-la puis les souks, evidemment,, et puis apres on a quitte Le Caire pour
aller, a Thebes . . . (VE: 17, 20)
Memphis Saqqarah
bazar
mosques interesting
(for other reasons)
Figure 3
whether 'backchannel signals', such as B's oui in T4 here actually do constitute real turns at
talk.
45
M.-B. Hansen
which there is a rise in pitch and intensity, all suggesting that it should be
interpreted as parenthetical (cf. Wunderli, 1982).
The following might be thought a counter-example to my analysis, in
that it appears to involve a reformulation, which may not only be thought
irrelevant (for an account of the relevance of reformulations, see however,
Blakemore, 1993, 1994), but which would seem to preclude a hierarchical
understanding of the passage:
(26) A: mais sinon est-ce que t'es timide qu&and tu connais pas les gens
B: & moi
B: ouais
A: moi aus
B: pourquoi
A: parce que moi aussi
B: pourtant t'as pas ete trop timide avec Nadia
A: ben parce que t'etais a cote hein sinon mais c'est different avec Nadia
B: nonj'ai prefere
A: parce que je la connaissais avant de la voir si tu veux
B: ah d'accord, non parce que laj'ai prefere Isabelle c'etait mieuxje trouve,, tu
comprends
A: c'etait mieux comment ca
B: c'etait mieux t'etais a l'aise et tout et puis elle t'a bien aimee, et puis en plus
Nadia elle t'a pas du tout mis mal a l'aise
A: mais elle MET les gens mal a l'aise quand elle les aime pas (CT: 7-8)
It is clear from the context and from the speaker's choice of words that
the last utterance following puis is more than just a reformulation,
however. First, we have the adverbial en plus following the second occur-
rence of puis, suggesting that to the speaker the upcoming utterance is not
simply another way of saying the same thing, but that new information is
being added. Secondly, the subject of the utterance is dislocated to the left,
where it is represented by a lexical NP, even though it does not represent a
new discourse entity, but is in fact co-referential with the pronominal
subject of the preceding utterance. This indicates that marked topic shift is
taking place.10 According to Barnes (1985: 65), the most frequent type of
left-detachment of lexical NPs is where 'the referent has just been men-
tioned but is not yet established as a topic of the discourse'. In the passage
quoted here, the discourse topic is A's claim to be shy, and her meeting
with Nadia is evoked as a possible counter-example. So, although Nadia is
the subject of the utterance following the first puis, she is still not a topic in
her own right. Once she has been established as such, however, a lengthy
discussion of her personality and behaviour ensues, whereas the original
shyness-topic is not referred to again.11 In other words, subsequent occur-
10
Cf. also Fox's (1987: 69) suggestion that in English conversation, the use of a full NP to refer
to an already established discourse entity may be a way of demarcating a new structural unit.
11
The passage as a whole is an example of what conversation analysts call 'stepwise transition
of topic' (cf. Schiffrin 1994: 261).
46
Puis in spoken French
Figure 4
47
M.-B. Hansen
12
Interestingly, in the original transcription, made by a native French speaker, the verb
parlai(en)t is in the singular. One cannot actually hear the difference, of course, but
syntactically only the plural could be correct. It seems that the transcriber has simply
overlooked the anomaly here, and has interpreted the antecedent of the relative clause as
being the particular guide mentioned in turn 1, which is no doubt what the speaker in fact
intended.
48
Puis in spokeh French
(30) A: parce que qu'est-ce qu'il reste de Thebes, de Thebes & la ville elle-meme y a
rien, parce que c'est &&
B: & de:, rien non il ne reste que la partie && des: religieuses § quoi §§
A: § oui les §§ absolument
B: oui oui
A: parce que bon l'architecture civile pour eux c,a n'avait strictement aucune
importance
B: puis c'etait construit en pierre en briques seches done c'est detruit hein (VE:
22-23)
(31) A: elle explique que Mitterrand etait pas de gauche originellement et qu'il Test
devenu justement grace a des rencontres comme celle de sa femme
B: oui etpuis enfin politiquement y a pas de mal a etre plus a gauche que Mollet
je veux dire qui a represente toute la: le socialisme sous la Quatrieme et les debuts
de la Cinquieme (h) done je veux dire c'est a celui qui sera moins a gauche que
l'autre pour devenir chef de la gauche (MP: 9-10)
49
M.-B. Hansen
group, others resist an analysis of this kind. This again supports the
hypothesis that (et) puis in spoken standard French may be acquiring the status
of a true conjunction, as is apparently already the case in Quebecois (cf.
Laurandeau, 1983). The number of examples in my corpora for which only a
non-elliptical interpretation is possible is, however, not large enough to
warrant the conclusion that conjunctive puis is fully grammaticalized, the
more so since at least some instances may be explained by other factors.
As regards the semantic properties o(puis, the bit of discourse following
puis in the non-elliptical examples cannot easily be understood as constituting
a separate act of utterance. It can, however, still be considered to have con-
textual effects of its own and to be the second of two and only two elements.
The following examples show a structure which may plausibly be
understood as elliptical:
(32) A: . . . par exemple y avait euh comment elle s'appelait Myriam, qui jouait
dedans qui jouait son role
B: mm mm
A: tu vois de Myriam, Collomb et puis un autre petit un tout petit Marseillais
qui faisait craquer toutes les petites nanas je sais pas si tu te rappelles (CV: 9)
(33) . . . ils etaient l'un et l'autre des prisonniers de guerre, De Gaulle a ete prisonnier
32 mois en 1914-18 et a essaye de s'echapper et Mitterrand a ete prisonnier plus
d'un an douze ou dix-huit mois et lui il s'est echappe et De Gaulle trouvait tres
bien qu'on s'echappe en 14-18 mais tout a fait scandaleux qu'on s'echappe en
39-40 ca prouve qu'on avait ete prisonnier et done la y a et comme il a commence
sa carriere comme representant et ca elle le montre tres bien Catherine Ney (h)
des evades car e'est comme ca qu'il a commence il est arrive en disant voila je
represente tous les gens (h) qui se sont evades et puis tous ceux qui ont et£
prisonniers (MP: 7-8)
(34) A: . . . oui e'est des & bas-reliefs &&
B: & un bas-relief &&
A: un bas- § relief §§
B: § oui §§ e'est possible hein
A: et c,a represente e'est une bande dessinee, enfin & nous on appellerait c,a une
bande dessinee &&
B: & oui qa fait une bande dessinee ouais &&
A: et e'est une dispute de bateliers sur le Nil § alors §§
B: § ah oui §§
A: t'as l'equivalent done t'as la succession des scenes, et puis alors sortant de la
bouche des belligerants, euh ce qui correspond aux bulks & d'aujourd'hui
B: & e'est vrai (VE: 51-52)
50
Puis in spoken French
(33') . . . c'est comme ca qu'il a commence il est arrive en disant voila je represente
tous les gens (h) qui se sont evades et puis je represente tous ceux qui ont ete
prisonniers
(34') . . . t'as l'equivalent done t'as la succession des scenes, et puis alors sortant de
la bouche des belligerants, euh t'as ce qui correspond aux bulles & d'au-
jourd'hui
In (35), however, ellipsis is highly unlikely, since (35') seems to have a
different meaning:
(35) A: alors pour eux les Coptes c'est en meme temps que Isis Osiris
B: au oui & ben oui &&
A: & ils ont beau && il ont beaucoup de mal (h) alors mais les Chretiens ils
vivaient comment a l'epoque
B: oui il ne § ils n'existaient §§ pas aussi
A: § mais c'est §§
A: alors c'est difficile de leur faire comprendre que les:
B: ah oui
A: les Coptes c'est tres proche de & nous par rapport a la religion &&
B: & nous, oui, bien sur par rapport a la &&
A: de l'Egypte an- § cienne §§
B: § cienne ouais §§
A: mais quand tu leur dis que c,a a quand meme pres de deux mille ans, et puis
que c'est tres proche
B: de nous
A: de nous, euh: c,a les fait rever un petit peu (VE: 64)
(35') A: mais quand tu leur dis que c,a a quand meme pres de deux mille ans (a les
fait river un petit peu, et puis quand tu leur dis que c'est tres proche
B: de nous
A: de nous, euh: c,a les fait rever un petit peu
Finally, in (36) and (37), expansion seems to be out of the question:
(36) A: au oui entre ce qu'on enseignait y a vingt ans,
B: ah bon
A: et puis ce qu'on enseigne aujourd'hui y a des differences parce que y a des
acquis, nouveaux, et puis y a des choses qui se sont revelees fausses
B: ah bon
A: oui
B: ah ben je sais pas ca parce que bon nous on a les notions qu'on a entendues
cet ete, c'est tout . . . (VE: 12)
(37) . . . done on en a vu un petit morceau par avion, puis on est alles jusqu'a Abou
Simbel parce que y a les deux temples-la: euh:, de: comment Ramses euh:,
Ramses II, oui, etpuis sa femme qui sont cote a cote quoi (VE: 31)
Now, in (35), the fact that we are dealing with an example of indirect
speech goes some way towards providing an account for the use of puis,
since, in the corresponding direct speech, the two conjuncts would be
highly likely to constitute separate utterances, for instance as follows:
51
M.-B. Hansen
(38) La religion copte, $a a quand meme pres de deux mille ans; mais c'est tres proche
de nous par rapport a la religion de l'Egypte ancienne
52
Puis in spoken French
4. CONCLUSION
To sum up the preceding discussion, I have argued that the coded content of
conversational puis should be seen in procedural terms, i.e. as consisting in a
set of instructions from the speaker to the hearer, directing him on the one
hand to search for two and only two elements to be connected on the same
level of the discourse, and on the other hand to understand these elements as
being individually relevant to some 'common integrator'.
I have presented evidence that puis in contemporary spoken French is no
longer to be analysed as a time adverb, and have argued that it has, in fact,
become grammaticalized as an additive conjunct, with the important proviso
that such an analysis be based on semantic, rather than the traditional
syntactic criteria, since the syntactic behaviour of this particle is unorthodox.
I have, moreover, hypothesized that puis may be undergoing further
grammaticalization tending towards its becoming a true conjunction, and
that it may not therefore be possible to place it squarely in any one particular
word class. If it is nevertheless possible to avoid an analysis in terms of
homonymy, it is because the basic semantic instructions associated with this
marker seem to remain the same in its different syntactic uses, and because,
even when it functions below the level of the sentence, the units connected
can be considered as acts in themselves, rather than mere constituents of acts.
13
An anonymous reviewer suggests that (40) above might more appropriately be transcribed as
follows:yean, et puis Marie, soul maries; and that it is the syntax, rather than the occurrence of
puis, which is responsible for the non-ambiguity of this sentence. Now, if the commas are to
be understood as indicating brief pauses in the speech stream, I disagree with the proposed
transcription. If, on the other hand, they are taken to indicate that etpuis Marie is somehow on
a different level of structure, this seems to be entirely in line with my proposal that units
connected by puis be understood as separate discourse acts. The same reviewer also suggests
that Jean, et Marie, sont maries (the commas presumably intended to indicate pauses in the
speech stream) would similarly be unambiguous. 1 am not convinced that this is necessarily
the case, but even so, it merely shows, as has often been noted, that discourse markers are
usually optional, in the sense that they can only reflect (and thereby guide the hearer towards)
an already possible interpretation, which may also be indicated by other means: then in
English If. . . then . . . constructions is a clear example of such optional marking.
S3
M.-B. Hansen
Author's address:
Institut d'Etudes Romanes,
Faculte des Lettres et Sciences Humaines,
Universite de Copenhague,
Njalsgade 80,
DK-2300 Copenhague S,
Denmark
REFERENCES
Anscombre, J.-Cl. and Ducrot, O. (1983). L'Argumentation dans la langue. Brussels:
Mardaga.
Auchlin, A. (1981a). Mais, hein, pis bon, ben voild, quoil Marqueurs de structuration
de la conversation. Cahiers de linguistique francaise, 2: 141-160.
Auchlin, A. (1981b). Reflexions sur les marqueurs de structuration de la
conversation. Etudes de linguistique appliquee, 44: 83—103.
Barnes, B. K. (1985). The Pragmatics of Left-Detachment in Spoken Standard French.
Amsterdam: John Benjamins.
Berrendonner, A. (1990). Pour une macro-syntaxe. Travaux de linguistique, 21:
25-36.
Blakemore, D. (1987). Semantic Constraints on Relevance. Oxford: Blackwell.
Blakemore, D. (1993). The relevance of reformulations. Language and Literature,
2(2): 101-120.
Blakemore, D. (1994). Relevance, poetic effects and social goals: a reply to
Culpeper. Language and Literature, 3(1): 49-59.
Blumenthal, P. (1975). Zur kommunikativen Funktion von Adverbien und
Umstandsbestimmungen im Franzosischen. Romanische Forschungen, 87:
295-332-
Brown, P. and Levinson, S. C. (1987). Politeness. Cambridge: Cambridge
University Press.
Chevalier, J.-Cl. and Molho, M. (1986). De Timplication: esp. pues fr. puis.
Travaux de linguistique et de litterature, 24(1): 23-24.
Dik, S. C. (1968). Coordination: Its Implications for the Theory of General Linguistics.
Amsterdam: North-Holland.
Ducrot, O. et al. (1980). Let Mots du discours. Paris: Editions de Minuit.
Ek, J. A. van and Robat, N. J. (1984). The Student's Grammar of English. London:
Longman.
14
I would like to thank Peter Harder, Henning Nelke and Ebbe Spang-Hanssen for helpful
comments (not all of which have been heeded) on earlier versions of this work. Needless to
say, they are not to be held responsible for any remaining errors in the presentation.
54
Puis in spoken French
55
M.-B. Hansen