«Indicators and procedures : nevertheless and but1 DIANE BLAKEMORE University of Salford (Received August ; revised January ...»
In fact, it seems that Rieber does not intend to be construed in this sense, and that he is including but in that category of expressions which, according to speech act theoretic analyses, are conventionally performative, but which simply don’t happen to contain a performative verb – for example, expressions like thanks or pardon, which would be analysed as performatives I thank you and I apologize respectively.% From a speech act theoretic point of view saying that pardon is equivalent to the explicit performative I apologize is illuminating only in the light of the constitutive rules for performing a successful act of apologizing. Analogously, saying that but is equivalent in meaning to a performative of the form I suggest that P is illuminating only to the extent that we understand what it means to perform the speech act of suggesting. Rieber himself is doubtful whether suggest is the most appropriate verb here. However, this is not really the point, because it is clear that what he has in mind is something like signalling or showing or indicating. The fact that languages have expressions like but, claims Rieber, can be explained once it is recognized that not all linguistic communication consists in modifying the beliefs of the hearer : ‘ Sometimes a hearer simply wishes to call attention to something that the hearer would believe were it brought to her awareness ’ ( : ).
all she needs to do is to induce the hearer to notice something ’ ( : ).
This would seem to suggest that the communicative act associated with the use of an expression like but is analogous to the non-verbal act of, say, deliberately opening the fridge door in order to show someone how empty it is. If it is recognized that this behaviour was intentional, then it can be assumed that the communicator was intending to draw the audience’s attention to something, or, in other words, that the communicator intended by this behaviour to make it manifest that she intended to make certain assumptions manifest. Such behaviour, according to Sperber & Wilson ( : ) is a case of or .
In his discussion of indicating, Rieber links his notion of showing to Sperber & Wilson’s notion of ostension. However, it seems clear that what he means by (or indicating) only covers the sort of case in which  The assumption that hearers recover the information that the hearer is issuing a warning as part of the interpretation of the utterance is a speech act theoretic one rather than a Relevance theoretic one. See Sperber & Wilson (), Wilson & Sperber () for further discussion.
 I am grateful to the anonymous referee whose comments helped me clarify this point.
some of the assumptions that become manifest – for example, the assumption that the fridge is empty – might be manifest to the audience even if he had not recognized that the communicator had intended to make them manifest. In other words, it seems that what Rieber has in mind is the sort of case in which a communicative act provides for information rather than . For example, while the act of opening the fridge door may have provided direct evidence that it is empty, the act of producing the utterance in () can make the communicator’s intention to make an assumption manifest only if the audience has ﬁrst recognized the communicator’s intention to make this assumption manifest.
() There’s nothing in the fridge.
For the relationship between the evidence produced (the utterance) and the assumptions conveyed is arbitrary, and it is only by discovering the communicator’s intention to make particular assumptions manifest that the audience can discover, indirectly or inferentially, what these assumptions are.
This might seem to suggest that the distinction between showing (or indicating) and describing (or saying) the distinction between providing direct evidence for information and providing indirect evidence. However, Sperber & Wilson ( : ) have argued that there is not a sharp dividing line between showing and saying that, but rather that there is a whole continuum of cases ranging from cases of ‘ showing ’ to cases of ‘ saying that ’.
A of these cases are cases of ostension in the sense that they involve making one’s intention to convey information manifest, and they involve inference. Thus even if the act of opening the door provides direct evidence for the information that the fridge is empty, there are other assumptions which are made manifest only indirectly – for example, that I am trying to be relevant, that the communicator is aware that the fridge is empty. On this view, ostension is not just showing (to be contrasted with saying that), as Rieber seems to suggest, but a well-deﬁned domain which covers all cases of human communication.
Nevertheless, it seems that one can distinguish between diﬀerent types of ostensive communication, and that in particular, one can distinguish those cases in which the communicator provides evidence from which the audience is intended to derive assumptions inferentially from those cases in which the communicator’s intention is to the saliency of information by making it more accessible to those inferential processes. An example of the ﬁrst sort is the utterance in (a), which – given the appropriate contextual assumptions and the hearer’s recognition that the speaker has deliberately attracted his attention – would be taken as evidence for assumptions such as the ones in (b) and (c).
() (a) An is coming.
(b) A number bus is coming towards the speaker and hearer.
(c) We should get on the bus that is coming towards us.
An example of the second sort is an act of pointing to an oncoming bus. This act cannot in itself be taken as evidence for assumptions like the ones in (b) and (c). However, it might make the bus more salient to the audience so that either it is accessible as a referent for the comprehension of a following utterance (for example, ‘ It’s coming ’) or it is accessible as a referent in an assumption which the audience constructs for himself (for example, (b) or (c)).
Pointing is, of course, a device rather than a linguistic one. The question is whether a linguistic expression like but might be said to ‘ point ’ in this sense. In the following section we will ask whether the use of but is consistent with Rieber’s tacit performative analysis of its role as an indicator.
. B U T Rieber himself points to one diﬃculty of his analysis of but as an indicator, namely, the fact that in attitude contexts like the one in () the speaker will not necessarily be taken to be suggesting that there is a contrast.
() Tom thinks that Sheila is rich but she is unhappy. However, I have always thought that all rich people are unhappy.
In a recent paper, Bach () has argued that the fact that but can occur in the that-clauses of indirect quotations shows that it cannot be construed as an indicator in any sense, and that it contributes straightforwardly to what is said. As Bach says, while it is true that an expression like but can be used to make ‘ an editorial comment on what he is reporting as being said ’, it can ‘ also contribute to is being reported ’ (Bach : , my emphasis).
Accordingly, he argues that an expression like but functions as an operator which combines with the rest of the sentence to yield a proposition which, although it is not part of the truth conditional content of the utterance, is nevertheless something which has truth conditions. The fact that but seems not to contribute to what is said, Bach argues, is due to the fact that this proposition, while truth conditional, is ‘ secondary to the main point of the utterance. Indeed, contrary to the common assumption of one sentence, one proposition, such utterances express more than one proposition ’ ( : ).
The idea that an utterance may express more than one proposition and that not every proposition expressed by an utterance contributes to its main relevance has also been explored in Relevance theoretic semantics (cf.
Blakemore , Ifantidou-Trouki ). However, paradoxically, this analysis was proposed for expressions like illocutionary adverbials, apposition markers and parenthetical verbs, which according to Bach () are like but since they do not contribute to the content of utterances. The diﬀerences between the ways these two approaches analyse expressions like illocutionary adverbials should be re-visited in the light of Bach’s paper.
However, the discussion here will be conﬁned to the question of whether it is legitimate to assume that an element following the complementiser in an indirect quotation contributes to something propositional.
It is generally recognized that what Banﬁeld () calls represented speech and thought may include expressions and constructions which, although they can be attributed to the person whose thought is being represented, cannot be easily analysed as contributing to something propositional. Consider for example, ah in (), and the reformulations and repetition in ().
() Ah ! that in itself was a relief, like being given another life.
(D. H. Lawrence, Lady Chatterly’s lover, p. ) () That was the way to live – carelessly, recklessly, spending oneself. He got to his feet and began to wade towards the shore, pressing his toes into the ﬁrm, wrinkled sand. To take things easy, not to ﬁght against the ebb and low of life, but to give way to it – that was what was needed.
It was this tension that was all wrong. To live – to live !
(Mansﬁeld, At the bay, p. ) It is true that the expressions and constructions which according to Banﬁeld, characterize free indirect speech cannot be indirectly quoted in embeddded clauses as but can. However, the phenomenon of free indirect speech or thought does raise the question of what it means to say that a writer or speaker is representing a . If we say that ah in () or the repetition in () is being used to represent a character’s thought, then it seems we cannot construe thoughts simply in terms of their (truth conditional) propositional content.
Moreover, it seems that there are devices which do not contribute to propositional content that can be indirectly quoted in an embedded construction. Thus while the marked stress on walking in () would be understood as an editorial comment on the thought being reported, it seems that (a) can be interpreted in much the same way as (b) and hence that the emphasis on needed is being attributed to someone other than the speaker\narrator.
() She says she is WALKING to the station, for God’s sake. It’ll take at least an hour and the train leaves at .
() (a) John pointed out that they couldn’t really aﬀord a holiday. But no, she said that she NEEDED to get away.
(b) John pointed out that they couldn’t really aﬀord a holiday. But no, she NEEDED to get away.
These sort of phenomena suggest that Bach’s argument that the use of but in indirect quotations is evidence that it contributes to the (propositional) content of utterances can be maintained only if either it can be shown that expressions like ah and devices like repetition and emphasis do not encode any aspect of a represented thought at all or it can be shown that they contribute to something with truth conditions. It would seem that the ﬁrst option is diﬃcult to reconcile with the way that examples of represented thought and speech are interpreted, while the second is diﬃcult to reconcile with the notoriously vague (and sometimes poetic) eﬀects of these stylistic devices. In Relevance Theory these devices are viewed as means for directing or guiding the hearer\reader towards a particular line of processing, or in other words, towards a particular range of contextual eﬀects, or inferential strategy or context.& On this approach such devices may be used in the representation of an attributed thought not just in the sense of a thought , but rather in the sense of a thought . In particular, they can be used in the attribution of interpretation processes so that the hearer\reader is shown how the person whose thought is being reported sees its relevance.
As Bach stresses, non-truth conditional analyses of but assume that the conjunctive import of but is expressed in a proposition which is distinct from the conjoined proposition asserted. Rieber’s () tacit performative account maintains this assumption in the sense that the contrastive import of but is expressed in a conjoined but parenthetical performative, and I suggest that this contrasts, which is ‘ comma’d oﬀ ’ from the conjoined proposition which is asserted. Bach’s argument is that we do not need this extra proposition if we treat but as an operator on the rest of the sentence and hence as contributing to what is said. However, although his analysis does not treat the contrastive import of but as something expressed in an implicated proposition (cf. Grice ) or a conjoined parenthetical proposition (cf. Rieber ), it does nevertheless treat it as a constituent of a proposition.' That is, Bach assumes that the hearer of an utterance containing but recovers a proposition with something like as a constituent. My argument in the rest of this section is that this assumption cannot be maintained and that the use of but, including its use in examples like (), can be better explained if but is treated as a linguistically encoded means for constraining the interpretation process.
First, let us consider Rieber’s assumption that the utterances containing but express propositions whose conjuncts contrast. This assumption is diﬃcult to maintain in examples like ()–(), where but is used to ‘ conjoin ’ two diﬀerent types of speech act.
() There’s a pizza in the fridge, but leave some for tomorrow.
() I know that this bus goes to town, but does it go to Picadilly Gardens ?
() He’s ﬁnished the decorating, but what a mess.
 For further discussion, see Sperber & Wilson ( : –) and Blakemore ().
 This analysis could be compared with Mann & Thompson’s () rhetorical structure theoretical treatment of coherence relations in which it is claimed that the interpretation of an utterance includes a relational proposition about the coherence relation which it bears to the preceding segment in the discourse.
() A : If we get oﬀ the bus at St Peter’s Square, it’s only a few minutes’ walk to Kendalls.
B : But I wanted to go to Debenhams ﬁrst.
() [The speaker has just found the hearer eating the last slice of pizza.] But I told you to leave some for tomorrow.( The assumption that an utterance containing but expresses a conjoined proposition is also diﬃcult to maintain given the interpretive discrepancy between and and but noticed by Kitis (). Thus whereas in her example in () below, the speaker will be understood to be communicating an emotional attitude of outrage or surprise, the utterance in () can only be taken to be suggesting that the inference which is assumed to have been drawn on the basis of the ﬁrst segment (for example, that the woman isn’t having a lot of fun) is illegitimate.
() Her husband is in hospital and she’s seeing other men.