ISSA Proceedings 1998 – Making A difference Or Not: Utterances And Argumentation

ISSAlogo1998As a linguist, I am limited, in the study of argumentation, to the linguistic traces of the argumentative process. Fortunately, they are numerous, and exactly like the relation between fossils and life forms, they present the advantage to be testable and that one can be sure that, even if some aspects of the argumentative process do not leave fossilized traces, most do.
Arguments are utterances and therefore they share certain characteristics of utterances (as opposed to propositions or phrases). To highlight what is probably the most important feature of utterances as far as understanding the relation between an argument and conclusion, is the aim of this paper.
I had the opportunity (Nemo, 1995) to present here a description and account of argumentative relevance, which I will quickly summarize, before introducing new evidence for my main hypothesis.

1. Utterances and argumentation
First of all, the distinction between proposition and utterance must be justified. If we consider the difference between proposition 1 and utterance (1):
1. Bill Clinton is alive.
(1) Bill Clinton is alive.
i.e. the difference between an unsaid proposition and an uttered proposition (the utterance), it must be remarked that 1 represents only the fact that Bill Clinton, is alive, whereas (1) represents both the fact that he is alive and the fact that this might (indexicaly and not theoreticaly) not have been the case. Consequently, the utterance (1) can represent only a moment when something has happened (an accident, an heart attack, an assasination attempt, etc…) whereas the proposition represents any moment in which 1 is true. In other words, the sentence is (only) an image of the reality whereas an utterance is the association of an image of the possible and an image of the reality: an utterance consists, minimally, of the association of a proposition with a modal frame, and hence receives the following description:
(1)
Bill Clinton may be alive – Bill Clinton is alive.
– may not be alive

The mere use of language implying a modal framing of reality. From this general standpoint a description of the argumentative value of utterances may be proposed. The constraints which have to be described in order to account for it are at least four, one accounting for the argumentative value itself, as opposed to informative value for example, another accouting for the argumentative orientation, and the others for the argumentative strength of utterances.

1.1. Argumentative utterances
To account for the argumentative value of an utterance, that is to account for the fact that we can say things (which are by no means informative) such as “ I’m your dad ” (this to say for example “ you should listen to me ” ) or “ I’m not three years old ” (this to say for instance “ I shouldn’t be treated like that ”), the existence of a scalarisation constraint must be hypothetized.
An utterance E may be used as an argument for an utterance R, if and only if it makes a difference for R that E is the case or not.

If we consider the exemple (2) for instance:
(2)
(S)he came but too late.

it is easy to observe both that what is meant is that when she actually came, it didn’t make a difference any more and that the meaning of “ P but Q” (the encoded meaning of ‘but’) is simply, as we shall see again later, to indicate that P is not making the difference it might have made because of Q.

1.2. Argumentative orientation
To account for the argumentative or scalar orientation of utterances, that is to understand how a certain reality can lead to opposite conclusions, the comparison constraint must be spelled out: Given the fact that the scalar (that is argumentative) value of an utterance depends on a comparison of the different possibilities which are introduced by the utterance, the scalar orientation of the utterance depends of the possibilities which are, or which are not, introduced.
Among the linguistic traces of the existence of this constraint are what Ducrot calls the argumentative operators, for instance peu (little), un peu (a little), presque (almost), à peine (hardly). I could add trop (too much ot too many) or seulement (only) but it can also be shown on operator free utterances. If we consider the utterance (3) and (4) and the surprising relevance of answering/retorting (3) to somebody who has said (4):
(3)
Il a peu souffert (He suffered little or He didn’t suffer much)
(4)
He suffered (He suffered)

First of all, it is clear that there is no need of any old information to understand what is going on: utterance (3) modal background consists in opposing suffering a little and suffering a lot, in which case suffering ‘little’ is not so bad. On the contrary, utterance (4) modal background consists in contrasting the fact of suffering with the possibility not to suffer at all, and therefore it presents the suffering as ‘bad’. Thus, by answering (3) to (4), or by opposing them with a mais (but), what (4) actually reminds the speaker of (4) is that the person in question might not have suffered at all, a possibility which the first utterance was simply not considering at all.
Hence, the relevance and interlocutive value of the answer (4) is completely dependent on the difference suffering or not suffering makes, and not at all on any new information (4) would convey. Yet, there are no reasons to believe that (4), because it is clearly uniformative and therfore violating Grice’s maxim of quantity, would be considered irrelevant: if it doesn’t not change anything about the representation of the world, it does change locally the set of possibilities to be considered, in other words what we shall call from now on the interlocutive image of what is possible.

1.3. Argumentative strength
Two last constraints on scalar value account for argumentative or scalar strength. The first one is the scalar slope constraint: Given the fact that an utterance E is an argument for an utterance R if it makes a difference for R if E is the case or not-E, then the argumentative strength of the utterance E depends basically on whether the difference that E makes for R is small or big.
The second is the modal slope constraint and operates within the scalar slope constraint:
Given the fact that it makes a difference for an utterance R if E is the case or not, the more not-E will be possible (likely), the biggest the difference the fact that E is the case will make. And hence, the stronger the arugumentative or scalar value of the utterance E will be.

Linguistic traces of the existence of these constraints can be found in the use of words such as ‘même’ (even). If we consider utterance (5):
(5)
Même Pierre est venu (Even Pierre came) the fact that “ Pierre came ” is the strongest argument to prove the success of a meeting is due to the fact that Pierre was the most unlikely to come.

1.4. Some examples
We shall illustrate this description with the dialog (6), a dialog which includes the discourse marker ‘tout de même’ (‘even so’) and which is taking place in a shop between a customer C and a seller S, should be considered.
(6)
C’est cher !
– It’s expensive !
C’est de la qualité !
– It’s quality !
Tout de même !
– Even so !

It would be possible to say, as the first Ducrot for example would have said, that “ It’s expensive ” is an argument for a conclusion and that “ It’s quality ” is an argument for the opposite conclusion. But that’s not what is really at stake in this dialog. The meaning of the answer “ it’s quality ” in (6) is that “ it cannot be inexpensive ”, which, according to the modal slope constraint, weakens the first utterance scalar value: if it cannot be inexpensive because it’s quality, then the fact of being expensive cannot make any longer a difference.
Hence, to bring back again some scalar value to the initial utterance, the customer will have to reply “Even so”, this to say that “Even for quality it’s expensive”.

2. The difference it makes and the semantics of utterances
The next point I want to make clear is that the scalarisation constraint and its insistance on the importance of the difference what is said is supposed to make is not an adhoc and commonsensical hypothesis, nor something specific to argumentation. What is quite clear on the contrary it that even if what it means exactly has yet to fully explored, it should be considered as a linguistic discovery.
Why should it be so ? Mainly because the scalarisation constraint (SC) is shared by all utterances and appears in the most different contexts and speech acts. And because it is a key to the interpretation of utterances, either in the understanding the implicit of utterances or in the understandings of what argumentative connectives or operators actually encode.

2.1. The difference it makes and the implicit content of utterances
For instance, only the SC accounts for the fact that utterances children such as (7) may be uttered even by a mother to one of her own children:
(7)
I am not your dad

to say something such as “go and see your dad directly, I am not the relevant person for this”. But this is also why saying (8):
(8)
I’m your dad.

sometimes mean things such as “Don’t talk to me like that” and sometimes things such as “You can talk to me”.
And this is still why it can be guessed that what the utterance (9):
(9)
On est Alsacien ou on ne l’est pas (one is Alsatian or one is not) is talking either about the difference it makes to be Alsatian or not, or about the difference the whole utterance makes, namely that there is no middle between the two possibilities. Similarly, as was observed within the Relevance Theory framework, this is why answering
(10)
He is French

to the question:
(11)
Does he know how to cook ?

may be explained by the sole hypothesis that this answer must be interpreted through the question “What difference does it make for cooking abilities to be or not to be French ?”. This is also why, an even more subtle implicit of such utterances, one cannot answer:
(12)
It’s right around the corner

to somebody looking for a gas station and asking:
(13)
Do you know where the closest Gas station is ?

if(s)he knows that the station is actually closed.

As a matter of fact, it seems that the scalar maxim: Do not say something which makes no difference (to what is at stake) would probably be the most direct description of cooperativeness. The same constraint is also present in indirect speech acts, such as:
(14)
It’s hot in here.
(15)
the bin is full.

In which it combines with another feature, the X-dependency feature (Nemo, 1998), to produce the directive effect.

2.2. To make or not to make a difference: fossilized tracesmof the SC
When linguists try to describe discourse connectives, the main problem is to understand what exactly is at stake in the use of a connective. I have mentioned earlier that the meaning of ‘but’ was not to oppose but to indicate that something is not making the difference which it would be expected because of what follows (or as regards what follows). This description, which applies to the normal oppositive use of ‘but’, also account for all conversational uses and reinforcing uses such as (16):
(16)
He is stupid but stupid

in which what is said is both that there is stupid and stupid, as we shall see later, and that the person concerned is of the second kind, which refers to the scalar slope constraint (How important is the difference something makes).

Another example of the importance of the SC will be provided by the discourse marker De toute façon (often translated by ‘anyway’) and its various uses, all examples borrowed from Corinne Rossari’s work on reformulative discourse markers (1994, 66-67). It must be noticed that in all the utterances of the form ‘A de toute façon B’, the utterance B imply that it makes no difference whether A or not A . So that with ‘A de toute façon B’, to use Rossari’s phrase, “ Il ne sert à rien de dire A puisque de toute façon B ” (“ It is not worth saying P as anyway Q ”). Let us show this with a few examples:
(17)
A – Où as-tu trouvé ce sac ?.
B – De toute façon, c’est un modèle qui ne se fait plus.
A – Where did you find (buy) this bag ?
B – ‘Anyway’, it’s a model which is not made any more.

In this dialog, what de toute façon means is that the question is not worth answering, because it wouldn’t make any difference knowing where the bag was bought, as it is not made any more. Thus, this example must be related to example (19)
(18)
A – Quand on veut, on peut.
B – De toute façon, je ne veux pas.
A – If you want to, you can.
B – ‘Anyway’, I don’t want to.

In this dialog, what ‘de toute façon’ means is that the first conditionnal utterance makes no difference, as its premise is not true, which is to say that it doesn’t matter that “ if you want to you can ” when you actually don’t want to do (something). Similarly, in:
(19)
Avec un type comme Ackley, si on levait les yeux du livre, on était foutu. De toute façon, on était foutu.
With a guy like Ackley, if you just lifted your eyes from the book, you were in deep trouble.
‘In any case’, you were in deep trouble. The monological context gives ‘de toute façon’ an autocorrective dimension: it is the conditionnal ‘if you …’ which is presented as incorrect as it actually makes strictly no difference to lift your eyes from the book or not, being in trouble in both cases.

Other examples are even more interesting:
(20)
Écoute, c’est un bon prix, et de toute façon il n’est pas négociable.
Listen, it’s a good price, and de toute façon it is not negociable. Because what is said is not that saying A is not worth, but that saying not-A, or arguing on A, wouldn’t be worth. Or still because ‘de toute façon’ may apply its scalar disappointment value to whole discourses, discussions and conversations, either backwards, and to say that what was said makes no difference for the present or the future, as in utterance (22):
(21)
De toute façon, tout ça, c’est du passé !
‘Anyway’, all this is history !
or forward, as when (22) is uttered to say in advance that whatever could be said or asked, it would not and should not make any difference to the performative reality of the speaker not being there:
(22)
De toute façon, je ne suis pas là !! C’est clair ?
‘Whatever they could say makes no difference’, I’m not here !! Is that clear ?
All those examples showing, as so many other examples with other connectives would, the importance of the scalar dimension of utterances Example (23) finally, which combines the two discourse markers mais and de toute façon, is a good example of the way all the constraints interfere one with another:
(23)
L’équipe de France est une très bonne équipe mais, de toute façon, en finale il n’y a que des très bonnes équipes.
(The French team is a very good team, but anyway in a final, there are only very good teams)

The first utterance, uttered by a Brazilian player just before the final, is given as an argument for « we should respect the French team », but as it can be interpreted too as « we should fear them », the but indicates that the fact that the French is a very good team is not making the difference it might have made (i.e. to impress the Brazilian team for instance) because as in final there are only very good teams (things may not be otherwise, a modal slope development), playing the French team or another very good team makes actually no difference (as is indicated by de toute façon): because it is not possible to play in a final a team which wouldn’t be very good, the fact of playing against a very good team loose all scalar value.

3. Making differences or not: the semantics of tautological and other anomalous utterances
It is not easy to account for the actual semantic interpretation of tautological utterances (Wierzbicka, 1991: 391-451), which is hardly linkable with the so-called propositional content or logical form that could be expected to be the fundamental meaning of the sentence. Nor to account for their pragmatic and conversational relevance: after twenty years of considerable focus on relevance, we still have almost nothing to say which could account for it.
However, the fact that neither semantics nor pragmatics could actually fully account for such utterances has something to do with our way to understand the semantics/pragmatics interface: tautological utterances, among others, actually falsify the idea that there would be what is said on one side (the explicature) and what is inferred from what is said on the other side (the implicatures)[i]. As a matter of fact, it seems clear on such examples that accounting for the meaning of what is said and accounting for the relevance of what is said is exactly the same task. Let us consider first apparently tautological utterances such as:
(9) On est Alsacien ou on ne l’est pas.
(One is Alsatian or not).

As soon as (9) is interpreted as a representation, it is tautological one, because saying P or not-P is always true. But, if we consider that utterances are comparisons, and not representations, then the semantic meaning of (9) may be obtained directly: (9) refers to the difference it makes to be Alsacian or not, as far as something is concerned. Therefore, (9) is normally used to point to a DP (Distinctive properties) of Alsatians (compared implicitly to other French people), such as drinking a lot of beer, in order for instance to present as normal such ort such attitude. It must be noticed that what is observed here in a tautological utterance is not specific to tautological utterances. The semantic interpretation of utterances such as (24):
(24)
Les Alsaciens boivent de la bière.
(Alsatians drink beer).
Is a problem too, because to be meaningful it is not necessary that all Alsatians actually drink beer, because the referent of Les Alsaciens is also underdeterminated and finally because the fact that boivent de la bière (drink beer) must be interpreted as boivent beaucoup de bière (drink a lot of beer) remains equally unexplained. But here, once again, it is clear that as soon as (24) is not treated as a representation but as a comparison, all those semantic difficulties disappear.

The comparison versus representation thesis that we shall support as a starting point to understand tautological utterances also apply to all utterances of the form Det N est Det N (Det N is Det N), tautological double characterizations being precisely of the form Det1 N1 est Det1 N1 (Det1 N1 is Det1 N1), but also to utterances of the forms Det1 N1 est Det2 N1 (Det1 N1 is Det2 N1), ”) or to paradoxal utterances of the Det1 N1 n’est pas Det2 N1 (Det1 N1 is not Det2 N1) form. For all this last kind of utterances, it must be remarked first that they escape the excluded middle constraint: things may be N and not-N in the same time, a situation which may be called the included middle.
(25)
Mes vacances n’ont pas été des vacances.
(My holidays were no holidays)
(26)
Ses vacances n’en ont pas été.
(His holidays just were not holidays)
(27)
Son père n’était pas un père.
(His (Her) dad was not a dad). Therefore, it is easy to understand that the relevance of tautological or paradoxal utterances is linked with the existence of this internal negation, which leaves many linguistic traces, for instance hedges:
(28)
La guerre est la guerre.
(War is war)
(29)
La guerre n’est pas toujours la guerre
(War is not always war)
(30)
Cette année, j’ai pris des vraies vacances.
(This year, I took real vacations) Hence, tautological and paradoxal utterances may be described as double comparisons: they both mobilize the DP of a class on one hand – the fact of not working for holydays for instance – and in the same time they either advance that no difference should be expected between the members of the class (about those DP) or on the contrary advance that a difference should be made![ii]

The utterance (28) would be a good example of the first case, as utterance (31):
(31)
Une voiture est une voiture
(a car is a car) which is used most of the time to say that all cars are the same, that iI n’y a pas voiture et voiture (there is no car and car). It seems, nevertheless, that contrasting (31) with the utterances (28) and (32):
(32)
Boys will be boys
(les garçons seront toujours des garçons) leads to observe the presence of an X-dependency feature in utterances (28) and (32), as thay both convey the idea that “there is nothing anybody can do about it”, a feature which is not present in all tautologies, but in very different kind of utterances. The utterance (31) on the contrary may perfectly be used as an answer to a question of the form Do you want this or that car model ? to assert that it makes no difference to him (her). With (31), it must be noticed, it is not the DP of the class which are focused on (the fact that wars are cruel or that boys are unruly), but what may distinguish cars one from antother (being big, confortable or fast) and thus properties which are neither common nor distinctive.

If we consider finally examples such as (33):
(33)
Lui, c’est lui, moi, c’est moi.
(He is he, I am I) it is clear first that it is the necessity not to consider two people as one single entity[iii] which is at stake here, but also that ‘considering’ two people as one entity concerns one’s attitude torward those people, and not inner properties of these persons.

What is at stake in tautological utterances hence is the necessity or not to make a distinction between things of the same type (or which belong together). And as regards finally the pragmatic or contextual dimension of the interpretation of such utterances, it appears to be important but very limited: in some contexts – i.e. in contexts where a difference has been made – tautological utterances will be used to remind that no difference should be made, while in other contexts – i.e. in contexts where no difference has been, or could, be made – tautological utterances will be used to insist on the necessity for things to be kept separated, and neither altered nor confused[iv].
The contextual dimension of these utterances is hence undisputable but limited to the determination of which of the two possible interpretation will be contextually valid.

4. Conclusion: utterances as implicit comparisons and the study of argumentation
That utterances convey implicit comparisons is of course very important to understand argumentation in general and enthymemes in particular.
In the first case because the most simple pieces of deductive or inductive reasoning cannot simply be described without taking into account these implicit comparison sets, as may be observed in such simple examples as (34) and (35):
(34)
He wasn’t going far. Hence he took his bike.
(35)
He was going far. Hence he took his bike. the first utterance (34) implicitly comparing going by bike or by car (or train, etc..) while the second supposes that the choice to be made was between going by foot and going by bike.

And in the second case because if utterances are simply not representations – as may be observed again with (36):
(36)
Nadia n’est pas sa soeur.
(Nadiai is not her sister) an utterance which is not the assertion (and representation) that Nadiai is not heri (own) sister, but actually a comparison between Nadia and Nadia’s sister – then the role of utterances in argumentative processes must be reconsidered.

What is actually important to notice hence is that:
– comparing is a way to present a reality in contrast with another;
– comparison is the process of highlighting differences;
– differences are not inferences;
– differences are not objective stimuli but realities which do not exist outside of the comparing process.

And that if utterances do consist of an association of an image of the reality with a modal frame, then what is needed in the study of argumentation is to take fully into account this modal framing.

NOTES
i. As A. Wierzbicka remarks (1991: 400), despite Levinson’s agreement (1983: 110-111) about the fact that “exactly how the appropriate implicatures in these cases are to be prediceted remains unclear”, “context” appears to be “an excuse for analytical failure”.
ii. A case which can be found in Chinese ‘concessive’ tautologies, for which, according to A. Wierzbicka (1991, 423), “The subordinate clause states an ‘undeiable truth’but the main clause contradicts this truth with respect to a specific instance : since this particular entity (X) belongs to a certain kind, one might expect that it will have certain properties, generally seen as characteristic of that kind: and yet, the speakers point out, this particular X (X) doesn’t have the properties in question”. But which must be considered together with all the numerous cases for which it is the existence of the necessity to make a difference which is stated : it might be the case that there are culture-specific interpration of such or such formula,, but the semantic content of these formula seems to be potentially universal.
iii. A. Wierzbicka’s (1991, 431) example of the (Chinese) statement that “husband is husband”, in a situation in which what is at stake is the way a group of housewives should behave with Mrs Tanaka, whose husband has just been gaoled, works the same way : it point to relationships with people, and insists on the necessity not to consider them as ‘going together’.
iv. As for instance the Chinese tautologies of irreductible difference (Wierzbicka, 1991, 427).

REFERENCES
Anscombre, J.-C. & Ducrot, O. (1983), L’argumentation dans la langue, Bruxelles: Mardaga.
Bakhtine, M. (1977), Le marxisme et la philosophie du langage. Essai d’application de la méthode sociologique en linguistique, Paris: Editions de Minuit.
Cadiot, P. & Nemo, F. (1997), Propriétés extrinsèques en sémantique lexicale, Journal of French Language Studies. Cambridge: Cambridge University Press.
Cadiot, P. & Nemo, F. (1998), Analytique des doubles caractérisations et logique de conformité, Sémiotiques, 13.
Carston, R. (1988). Implicature, Explicature, and Truth-Theoretic Semantics. In Kempson, R. (éd.) (1988), Mental Representations: The Interface Between Language and Reality, Cambridge: Cambridge University Press, 155-181.
Chauviré, S. (1989), Wittgenstein, Paris: le Seuil.
Davis, S. (éd.) (1991), Pragmatics, A Reader, Oxford: Oxford University Press.
Ducrot, O. (1972), Dire et ne pas dire,. Principes de sémantique linguistique. Paris: Hermann.
Ducrot, O. (1979), Les lois de discours, Langue française, 42: 21-34.
Ducrot, O. (1980a), Les mots du discours, Paris: Editions de Minuit.
Ducrot, O. (1989), Logique, structure, énonciation, Paris: Editions de Minuit.
Gazdar, G. (1979), Pragmatics, Implicature, Presupposition and Logical Form, New York: Academic Press.
Grice, P. (1975). Logic and conversation. In Cole, P. & Morgan, J.L., (éd.) (1975), Syntax and Semantics 3: Speech Acts, New York: Academic Press.
Jacob, P. (1997). Pourquoi les choses ont-elles un sens ?, Paris: Odile Jacob.
Jacques, F. (1985), L’espace logique de l’interlocution, Paris: P.U.F.
MacCawley, J.D. (1981). Everything that Linguists have Always Wanted to Know about Logic, Chicago: The University of Chicago Press.
Mey, J.L. (1993), Pragmatics. An Introduction, Oxford: Blackwell.
Meyer, M. (1982), Logique, langage et argumentation. Paris: Hachette.
Nemo, F. (1988), Relevance, Journal of Pragmatics, 12: 5-6.
Nemo, F. (1995), A description of argumentative relevance. In F.H. Van Eemeren, R. Grootendorst, J.A. Blair, and C.A. Willards (eds.), Analysis and evaluation (Proceedings of the third ISSA Conference on argumentation). Amsterdam: Sic Sat International Centre for the Study of Argumentation, 226-232.
Nemo, F. (1997), The Modal frames of Speech. In P. Weingartner, G. Schurz, G. Dorn (eds.), The Role of Pragmatics in Contemporary Philosophy. Kirchberg am Wechsel: Verlag-Hölder-Pichler-Tempsky, volume 2, 700-708.
Nemo, F. (1998), The Pragmatics of Signs, the Semantics of Relevance and the Semantics/Pragmatics Interface. In The Semantics/Pragmatics Interface from different point of views, CRISPI Series, Vol 1, Elsevier Science.
Sapir, E. (1944). Grading: A Study in Semantics. In Philosophy of Science, (1944), II, Mandelbaum: 122-149.
Searle J.R. (1996), The Construction of Social Reality, New York: The Free Press.
Sperber, D.& Wilson, D. (1986a), La pertinence, Paris: Editions de Minuit.
Wierzbicka, A. (1991). Cross-Cultural Pragmatics, New-York-Amsterdam: Mouton de Hueyter.