Utterance Interpretation and Actual Intentions

In this paper I argue, from the consideration of what I hope is the complete variety of a hearer’s approaches to a speaker’s utterance, that (1) the speaker’s intention does not settle the meaning of her utterance and (2) the hearer does not take a genuine interest in the speaker’s actual intention. The reason why the speaker’s intention does not settle utterance meaning is simply that no utterance meaning determination, as presupposed by intentionalists and anti-intentionalists alike, takes place. Moreover, in the regular course of interpretation the hearer does not care about the speaker’s actual intention, but only about what the speaker presents as her intention: the hearer’s goal is to come up with an interpretation which the speaker will accept rather than an interpretation which corresponds to the speaker’s intention. In cases of accountability, suspicion and lying, the speaker’s actual intention is irrelevant; it is at most the speaker’s hypothetical intention which is at stake.


Introduction
The question of the role of the speaker's communicative intention in the interpretation of a speaker's utterance is often conceived of in terms of whether or to which extent the speaker's intention settles or determines the meaning of her utterance. The notion of utterance meaning at stake in this debate is usually the semantic value of a sentence relative to a context or 'what is said' (Grice 1967, p. 25), i.e. the linguistic meaning of the sentence uttered supplemented by whatever is deemed required to obtain a fully truth evaluable proposition (disambiguation, reference assignment, pragmatic saturation, etc.). However, from the perspective which I will adopt here, there is no reason not to include also all sorts of pragmatic aspects of utterance meaning, so that by utterance meaning I understand 1 3 something akin to 'the total signification of the utterance of a certain sentence' (Grice 1967, p. 43).
According to intentionalists it is the speaker's intention which determines the meaning of her utterance. The intuition which is at the root of intentionalist approaches to utterance interpretation is that the purpose of language is to enable the speaker to communicate her beliefs and desires to the hearer; thus-intentionalists hold-what an utterance means depends fundamentally on what the speaker wants to convey by uttering it (Davidson 1986;Kaplan 1989;Neale 1992Neale , 2016Predelli 1998;Recanati 2005;Perry 2009; Korta and Perry 2010;Montminy 2010;Stokke 2010;Åkerman 2010;King 2013King , 2014Speaks 2016Speaks , 2017. Intentionalist approaches to utterance meaning are, as it were, speaker-oriented, they submit meaning to the speaker's authority. Here is a typical intentionalist statement: It seems to me that if we wanted to settle, for example, whether Nunberg's waitress had asserted that a sandwich had left, or that a person who ordered a sandwich had left, we might ask which (if either) of these propositions she meant to commit herself to. To answer this question is to settle what she asserted. MacFarlane (2011), p. 92. Cf. Nunberg (1995).
A different approach to utterance interpretation is taken by anti-intentionalists. The fundamental intuition which animates anti-intentionalist conceptions of utterance meaning is that the speaker's intention is private and inaccessible to the hearer, whereas meaning is essentially public. Consequently-anti-intentionalists think-the meaning of the speaker's utterance is determined by what is available to the hearer, either by means of shared conventions or of diverse sorts of contextual cues. What an utterance means depends fundamentally on what the hearer is in a position to understand by it (Kaplan 1978;McGinn 1981;Wettstein 1984;Dummett 1986;Travis 1989;Reimer 1991;Green 2001;Saul 2002;Gauker 2008;Mount 2008;Lepore and Stone 2015). Anti-intentionalist approaches to utterance meaning are, is it were, hearer-oriented, they submit meaning to the hearer's authority. Here is an anti-intentionalist statement: What the speaker says, strictly speaking, and thus his reference, is determined by the cues available to the addressee. Wettstein (1984), p. 74.
These diverging intuitions orient theorists in their accounts of utterance meaning. Intentionalists cannot hold that the speaker's intention unconditionally settles the meaning of her utterance. A speaker may deviate from accepted usage, misspeak and use faulty demonstrations. Surely an utterance does not mean whatever the speaker wants it to mean. In order to maintain an intentionalist position, intentionalists put constraints on the speaker's intention, either by a stipulated conformity to conventional meaning (Dummett 1986) or by the dependence of intention formation on reasonable expectations of being understood (Donnellan 1968;Davidson 1986;Neale 2016). For some cases, intentionalists have recourse to the fact that multiple intentions are involved in the making of an utterance (Bach 1987;Perry 2009). There are also intuitive limits to anti-intentionalism.
It is after all a very natural assumption that it is the speaker's intention which determines the meaning of her utterance, provided only that it be somehow made available to the hearer. In those case where the hearer cannot grasp the speaker's meaning because the speaker has failed to make it available, it may be thought that nothing is to be considered to be the meaning of the utterance (King 2014).
Concerning utterance meaning determination, intentionalists and anti-intentionalists thus take opposing stances. However, the question of the role of intentions in interpretation may also be understood in a regulative rather than constitutive sense, namely: to what extent does the hearer take an interest in the speaker's communicative intention in interpreting her utterance? With respect to this question intentionalists and anti-intentionalists alike tend to take the same stance: even though the meaning of an utterance occasionally may deviate from the speaker's intended meaning, the speaker's communicative intention is in any case what the hearer ultimately wants to get at; the goal of interpretation is the recovery of the speaker's intention. Consider for instance what an anti-intentionalist like Wettstein says concerning the case of an utterance of that where the cues available to the hearer fail to determine a unique individual as its referent.
Wettstein asks rhetorically whether we should allow the referent in such a case to depend on the speaker's intention and answers as follows: The speaker, strictly speaking, has not asserted anything determinate, i.e., anything at all. The defect is, of course, negligible for the practical purposes of ordinary communication since the speaker when questioned easily supplies the information about which thing he meant. Wettstein (1984), p. 75. Cf. e.g. Davidson (1986), p. 99 andWilson andSperber (2002), p. 605.
Even though Wettstein holds that the speaker's intention is irrelevant for the institutional notion of utterance meaning, he nevertheless suggests that, as a matter of course, the hearer takes an interest in the speaker's intended meaning and also that this interest is concerned with factual information about the speaker's intention.
In this paper I will examine the question of intentions in interpretation in the two senses which we have just considered, i.e.: (1) Does the speaker's intention settle the meaning of her utterance? and (2) Does the hearer take an interest in the speaker's intention? As we have seen there is a considerable disagreement in the literature regarding the first question, whereas most theorists answer the second question in the affirmative. I will answer these questions by means of an inquiry into the interpretive interaction between the speaker and the hearer. From the viewpoint of the attitudes a hearer can take with respect to a speaker's utterance, it is my contention that both of the questions should be answered in the negative. However, this stance is not anti-intentionalist. I will attempt to show that the speaker's intention does not determine the meaning of her utterance, simply because no determination of utterance meaning takes place. Also I will attempt to show that the hearer does not take an interest in the speaker's intention, in so far as the hearer does not take a genuine interest in the speaker's actual intention. The view which I will put forward is in fact more speaker-oriented than most intentionalist accounts of utterance meaning.

3 2 Imagined Meaning
In some cases, the hearer acts upon a meaning which she takes to be no more than a possible meaning of the speaker's utterance. The hearer may imagine that the speaker had said something which she does not believe to be the speaker's intended meaning. Or she may imagine that she had taken the speaker to say that. Or the hearer simply imagines that someone had committed herself to the imagined meaning or that someone had taken the speaker to say so. The hearer's imagination may be carried quite far, but the imagined meaning is not without some connection to the linguistic meaning and contextual setting of the sentence uttered. A meaning without any such connection can hardly be entertained as the imagined meaning of the utterance in question. For what reasons does the hearer play with the speaker's utterance in this way?
Sometimes the hearer does so precisely for ludic reasons. Imagine that the speaker is Mrs. Malaprop and utters the phrase 'a nice derangement of epitaphs' (Sheridan 1775, p. 110). Thanks to her familiarity with the speaker's speaking habits and the context at hand, the hearer quite immediately understands that the speaker wants to say 'a nice arrangement of epithets'. Yet the hearer may entertain the linguistic meaning of the sentence as an imagined meaning of the speaker's utterance simply in order to have a laugh. In many conversational contexts imagined meaning serves the purpose of merriment.
Why should 'a nice derangement of epitaphs' be considered as the imagined meaning of Mrs. Malaprop's utterance and not as its actual meaning? 'A nice derangement of epitaphs' is after all the meaning of the phrase uttered. It is true that in order to settle the meaning of a phrase or a sentence it suffices to consider the lexical meaning of its parts and their mode of composition. However, settling the meaning of an utterance involves, if only because of indexicals, the consideration of various contextual parameters. Exactly which parameters and to which extent is a matter of dispute in the literature. One contextual factor in the case at hand is that it is clear to the hearer that the speaker does not want to say what the linguistic meaning of the sentence indicates, but wants to convey something only related to it. Is this information to be taken into account when settling the meaning of Mrs. Malaprop's utterance? There is hardly any definite answer ready at hand. The linguistic meaning of the sentence uttered is one thing, the context of the utterance is another. But this question need not perhaps be answered. What reason is there for the hearer to count the linguistic meaning of the sentence as the meaning of the speaker's utterance? What reason is there for the hearer to count anything as the meaning of the utterance? Does it not suffice to take 'a nice arrangement of epithets' to be what the speaker wants to say and the linguistic meaning 'a nice derangement of epitaphs' to be what the hearer imagines that the speaker says in order to have a laugh? Because linguistic meaning does not obviously constitute utterance meaning and because it is sufficient to consider linguistic meaning as imagined meaning, the question of the actual meaning of Mrs. Malaprop's utterance is better left aside. The communicative situation does not seem to be one where any notion of actual utterance meaning is called for.
Let us consider a case where I also think that imagined utterance meaning is entertained, this time for security reasons. In 2016, at Minsk airport in Belarus, a Russian woman jokingly uttered 'If you are looking for something, the TNT is in my bag'. The security staff did not spend time on recovering the speaker's actual intention nor on settling the actual meaning of her utterance, but acted on an imagined meaning of the utterance to the effect that the speaker had seriously said that there was explosive material in her suitcase. The speaker was detained and her suitcase investigated. 1 Probably the hearers considered that whatever the speaker's actual intention or the actual meaning of her utterance was, the possible consequences of the merely possible meaning of the utterance had to be taken seriously.
Less frequent in conversation perhaps is opting for imagined meaning for aesthetic reasons. Performing artists imagine the meaning of the work they interpret to be this or that simply because it is in some way or other aesthetically rewarding. Literary scholars assign whatever meaning to the work makes it appear in the best or most adventurous way (cf. e.g. Borges 1944;Currie 1993). Also of less importance in conversation is the kind of exegetical interpretation philosophers and other scholars occasionally undertake. The interpreter exploits the implications and consequences of a meaning which she may know was not intended by the author, but which yields important conceptual insights (cf. e.g. Kripke 1982).
The hearer may thus entertain an imagined meaning for ludic, aesthetic, theoretical or security reasons. This play is justified on account of its beneficial effects. In such cases, the hearer disregards the speaker's actual intention and, further, she is best construed as making no claims as to the actual meaning of the utterance.

Most Reasonable Meaning
Another approach to the meaning of an utterance is taken by the hearer in case she wants to hold the speaker responsible for her utterance. The speaker's accountability typically becomes an issue if the speaker's utterance has normative consequences. The hearer may think that the speaker by her utterance committed herself to something's being the case or to performing a future action. Whether this is so may involve the speaker and the hearer in conflict concerning the content of the utterance. For some theorists utterance responsibility requires settling the speaker's intention and the meaning of her utterance, the speaker being responsible for what was actually said or what she actually meant to commit herself to. I think however that neither the speaker's actual intention nor the actual meaning of her utterance are at stake in cases of utterance responsibility.
In his exchange with Davidson on idiolects, Dummett says that 'a speaker holds himself responsible to the accepted meanings of words' (Dummett 1986, p. 462). Let us leave the idiolect issue aside for the moment and focus on what Dummett suggests about utterance responsibility: the speaker's responsibility is a matter of 1 3 what the speaker holds herself responsible for. I believe that the notion of utterance commitment should not be abstracted from the context in which it is at issue. The question of what the speaker has or has not committed herself to turns up only in the context of the hearer's holding the speaker accountable for her utterance. Outside such a context utterance commitment is not at issue. Consequently, utterance responsibility is not a matter of what the speaker holds herself responsible for, but of what the hearer holds the speaker responsible for. The hearer naturally holds the speaker responsible for the hearer's interpretation of the utterance.
That the speaker's responsibility parts from the hearer's interpretation does not imply that the hearer can hold the speaker responsible for any wilful interpretation of her utterance. Responsibility does not depend on how the utterance was in fact taken, but on how it was most reasonably taken. In order for the hearer to be able to enforce her interpretation, the hearer must advance arguments for the reasonableness of her interpretation. Indeed, the hearer's interpretation should ultimately be the most reasonable interpretation of the utterance. What constitutes the most reasonable interpretation of an utterance is certainly a matter of discussion. For our present purposes a few remarks will be sufficient.
The speaker's actual intention does not enter into consideration. What matters is what the hearer was in a position to grasp from the utterance. Accordingly only publicly available cues will be decisive. This does not imply that the speaker's intention is unimportant. In most cases the hearer's interpretation will not be the most reasonable one unless it corresponds to the best hypothesis about the speaker's intention, the question being what the hearer had the best reasons to assume the speaker's intention was, to judge from the diverse cues available to her. Though the speaker's actual intention does not matter, the hearer's interpretation is certainly not separated from any consideration about the speaker's intention. 2 What are the publicly available cues? Foremost, of course, conventional linguistic meaning. But since the question is what the hearer had the best reasons to understand from the utterance there is no reason to stop at conventional meaning. All sorts of contextual cues may be relevant: gesture, mimics, background knowledge, previous utterances, etc. What kind of discourse the utterance occurs in is an important factor. Colloquial, scientific and legal contexts have different standards of reasonableness. What the most reasonable interpretation is amounts to an all-things-considered judgment (Ariel 2002a, b;Gauker 2008) which must attend to the specific circumstances at hand in each case and cannot be regulated in advance. The notion of most reasonable meaning is thus very different from the formal and semantic notions of utterance meaning which occur in the literature.
The notion of most reasonable interpretation might seem hopelessly vacuous. It seems to me significant though that courts and ordinary people usually do come to agree on how an utterance was most reasonably interpreted according to the particular circumstances at hand. And it would hardly be possible to go further. Even if we advance some principles of semantic determination, these must be argued in terms of reasonableness, so that the ultimate form of interpretation would be the same. 3 Is there any reason for the theorist or the hearer to claim that the hearer's most reasonable interpretation represents, corresponds to or constitutes the meaning of the utterance in cases of accountability? What matters for responsibility is not what the meaning of the utterance actually was, but solely whether the hearer had good reasons to take the utterance the way she did. Nothing but the epistemic warrantability of the interpretation is at stake. To claim that the most reasonable interpretation constitutes the actual meaning of the utterance seems superfluous (cf. Kotátko 1998).
Thus, the hearer taking an interest in holding the speaker responsible for her utterance does not take the speaker's intention to determine the meaning of her utterance. She is not concerned with the actual meaning of the utterance at all. Nor does she take an interest in the speaker's actual intention.

Intended Meaning
We have so far considered the hearer's playing with the speaker's utterance and her holding the speaker responsible for her utterance. Most often, however, the hearer's attitude is different. She simply sets out, as it were, to find out what the speaker wanted to say. The hearer uses the linguistic meaning of the sentence uttered and all sorts of contextual cues-gestures, mimics, previous discourse, background knowledge, diverse features of the communicative situation-as evidence in order to come up with a hypothesis about the speaker's intention. 4 The notion of meaning which is at stake is 'the total signification of the utterance', i.e. the point the speaker wants to make, for the bare 'what is said' does not suffice to orient the hearer's response. 5 The reasons for taking an interest in the speaker's intention vary: the hearer may want to familiarize herself with the speaker's perspective or understanding out of pure curiosity or for the purposes of coordinating joint action.
The hearer setting out to find out what the speaker wanted to say may arrive at the speaker's intended meaning, or, in less felicitous cases, at a different meaning or no meaning at all. In these latter cases it is very natural for the hearer to ask the speaker 1 3 about her intention. Alternatively the hearer is by the speaker's further discourse in one way or other informed about the speaker's intention and adjusts her understanding accordingly. In all these cases, the hearer seemingly opts for the speaker's intended meaning in so far as she goes along with the speaker's intended meaning as the meaning in force. We may distinguish between cases where the speaker's intention is unlicensed by the linguistic and contextual material of the sentence and cases where the intention is licensed. Many have the intuition that the speaker's intention does determine the meaning of her utterance in the latter cases, but not in the former. And that in both cases the hearer takes an interest in the speaker's intention. But what does the hearer's adopting the speaker's intention really imply concerning utterance meaning and the hearer's interest in the speaker's actual intention? Does the hearer take the speaker's intention to constitute the meaning of her utterance? Does her asking for the speaker's intention amount to a genuine interest in the speaker's actual intention? I will now see what the implications for these questions are from the viewpoint of the hearer's attitude.

Unlicensed Intention
One reason why the hearer may not get at the speaker's intention is that linguistic and contextual cues may indicate a meaning different from the speaker's intended meaning. The speaker's intention is, as it were, unlicensed by the linguistic and contextual material of her utterance. Let us look at the case of Humpty Dumpty: […]-and that shows that there are three hundred and sixty-four days when you might get un-birthday presents-'Certainly,' said Alice. And only one for birthday presents, you know. There's glory for you! 'I don't know what you mean by "glory,"' Alice said. Humpty Dumpty smiled contemptuously. 'Of course you don't -till I tell you. I meant "there's a nice knock-down argument for you!"' 'But "glory" doesn't mean "a nice knock-down argument,"' Alice objected. 'When I use a word,' Humpty Dumpty said, in rather a scornful tone, 'it means just what I choose it to mean -neither more nor less.' 'The question is,' said Alice, 'whether you can make words mean so many different things.' 'The question is,' said Humpty Dumpty, 'which is to be master -that's all. ' Carroll (1872), p. 251.
Humpty Dumpty is often accused of having illegitimate meaning pretentions. Many theorists concur with Alice in pointing out that the conventional meaning of words determines what a speaker can mean by them and that in cases like this there is a distinction to be made between the speaker's intended meaning and the meaning of her utterance. But actually it seems that the lesson from Humpty Dumpty is quite different. Alice apparently takes an interest in Humpty Dumpty's intended meaning in saying 'I don't know what you mean by "glory"'. Nevertheless, upon Humpty Dumpty ready answer to her (indirect) question, she retorts 'But "glory" doesn't mean "a nice knock-down argument"'. Humpty Dumpty's arrogance can be construed as an implicit objection to Alice's procedure. Glory certainly does not mean 'a nice knock-down argument' according to the dictionary. Alice's concern, however, did not appear to be with lexical meaning, but with Humpty Dumpty's intention. In retrospective then, Alice's interest in Humpty Dumpty's intention seems to be severely restricted, as if she is prepared to go along with Humpty Dumpty's meaning on the condition that it is licensed by the dictionary. Humpty Dumpty's implicit point is that such a restricted interest in the speaker's intention does not really make sense. Either the hearer asks for the speaker's meaning and then there is no reason to tell what the dictionary meaning is. 6 Or the hearer takes an interest in dictionary meaning, in which case there is no reason to ask the speaker about what she meant. The lesson from Humpty Dumpty is not that the meaning of an utterance and the speaker's intended meaning must be kept apart, but that a hearer taking an interest in the speaker's intention in the sense of asking for and going along with it, has no reason to be concerned neither with linguistic nor utterance meaning. This position is fully compatible with the view that Humpty Dumpty may be legitimately criticized for not speaking intelligibly. And in case Alice would want to hold Humpty Dumpty responsible for his utterance, it would be sufficient for her to argue that she had good reasons not to get at his meaning. No such thing as the actual meaning of his utterance would have to be invoked.
Intuitions about cases of confusion play an important role in the debate between intentionalists and anti-intentionalists on utterance meaning. Theorists ask for example what the referent of a demonstrative is in case the speaker has a certain object in mind but happens to demonstrate a different object, as in the famous case invented by Kaplan: Suppose that without turning and looking I point to the place on my wall which has long been occupied by a picture of Rudolf Carnap and I say: (1) [That] is a picture of one of the greatest philosophers of the twentieth century. But unbeknownst to me, someone has replaced my picture of Carnap with one of Spiro Agnew. Kaplan (1978), p. 30 (layout by me).
For anti-intentionalists this is a case which clearly supports their view: the referent is the one determined by public means. Intentionalists however have elaborated accounts according to which the speaker in such cases has multiple intentions: in addition to an intention to refer to the object the speaker has in mind, there is an intention to refer to the object which is demonstrated. The referent of the demonstrative is determined by the latter intention. The multiplicity of intentions is plausible, but in the absence of any argument as to why it is precisely the latter intention which is determinative the strategy does not really solve the intentionalist's problem.
But could not the problem be dispensed with? Either the hearer wants to know what object the speaker wanted to refer to, in which case there is no reason to take any interest in the question as to which object was objectively referred to. Or the hearer wants to hold the speaker responsible for her utterance, in which case the question is whether the hearer had the best reasons to take the referent to be what she took it to be. Also in this case the question as to which object was objectively referred to serves no purpose. There is as little reason for the hearer to be concerned with the actual referent of a demonstrative as there is for Alice to be concerned with what Humpty Dumpty can mean by his utterance.
For the hearer opting for the speaker's intended meaning in cases where this intention is unlicensed the speaker's intention certainly does not settle the meaning of her utterance. The hearer may have the inclination to take linguistic and contextual factors to do so, but, in fact, she has no reason to be concerned with utterance meaning at all.

Licensed Intention
In most cases, presumably, the speaker's intended meaning is licensed by the linguistic and contextual material of her utterance. The speaker provides the adequate means for the hearer to recover or at least get sufficiently close to her intention, so that the hearer quite immediately arrives at the intended meaning. Some theorists account for successful communication by positing the actual meaning of the speaker's utterance as something which both the speaker's intention and the hearer's interpretation coincide with. It seems however that successful communication can be explained equally well by the simple convergence between the speaker's intention and the hearer's interpretation. 7 It happens though that the hearer does not immediately arrive at the speaker's intended meaning. The linguistic and contextual material of the utterance may be underdetermined with respect to the speaker's intention, the intended meaning being merely a possible meaning of the utterance. It is then natural for the hearer to ask for it and go along with it towards future interaction. Many theorists speak of the speaker's intention as determining the meaning of her utterance in those cases where the speaker's intention falls within the sphere of permissible meaning. In the case of an ambiguous utterance, for example, the speaker is thought both to take an interest in the speaker's intention and to take the intention to settle the meaning of the utterance. 8 This position was exemplified by the quote from MacFarlane in the introduction.
But does the hearer's asking for the speaker's intention and her going along with it really amount to its constituting or to her taking it to constitute the meaning of the utterance? For this to be the case, the hearer must ascertain that the speaker is veracious about her intention. The speaker may be confused, oblivious, insincere or may have simply changed her mind and an alleged intention cannot be considered to constitute the meaning of the original utterance. Typically though the hearer in such cases goes along with whatever the speaker wants her to go along with. For this reason it seems much more natural to construe the hearer's asking for the speaker's meaning not as a request for information about the speaker's intention, but as an invitation to make a novel utterance which is to be understood in its own right. It is true that many clarificatory utterances cannot be understood unless they are related to the original utterance; they may for example be elliptical. But that does not prevent our considering the novel utterance rather to erase and replace than supplement the original utterance. Even if the hearer has a strong intuition that it is the meaning of the original utterance which is at stake in her asking for the speaker's intention she must have some argument in favour of the idea that she goes along with the intended meaning because it is the meaning of the original utterance and not simply because she is invited to do so. In the absence of any decisive argument to this effect, it seems that opting for the speaker's intended meaning does not have any implications for utterance meaning.

Correspondence Versus Acceptance
We have so far considered cases where the hearer, as it were, opts for the speaker's intended meaning. For some reason or other, the hearer does not get at the speaker's intention and then asks the speaker for it and is prepared to go along with it. When adopting this attitude, there is no reason to say that the speaker's intention settles the meaning of the utterance nor to be concerned with settling utterance meaning at all. What about the speaker's actual intention? Does the hearer's opting for the speaker's intention imply a genuine interest in the speaker's actual intention?
We already observed that the hearer does not ascertain that the speaker's reported intention is the actual intention behind the original utterance and used this observation as an argument against the view that the speaker's intention determines the meaning of her utterance. Even though I have spoken of convergence, intention recovery and opting for intended meaning, it may be more accurate to conceive of the hearer's opting for the speaker's intended meaning as her setting out not to come up with an interpretation which corresponds to the speaker's intention, but with an interpretation which the speaker will accept. Agreement, neither correspondence nor correctness, is what ultimately matters. It is as if the hearer does not ask the speaker whether her interpretation matches her intention, but asks the speaker to accept her interpretation: 'Are you okay with my taking your utterance this way?'.
Here are two examples of conversational interaction in support of this view. Schegloff et al. provide the following piece of dialogue: Lori But y'know single beds'r awfully thin tuh sleep on Sam What? Lori Single beds.//They're-Ellen Y'mean narrow? Lori They're awfully narrow//yeah. Schegloff et al. (1977), p. 378 When Ellen here asks Lori 'Y'mean narrow?', she is not asking about Lori's intention, but proposing an understanding of Lori's utterance and wants to know whether Lori is willing to accept it, no matter the original intention.
Clark considers the phenomenon of 'elective construal': Case Susan's question could be understood in different ways, notably as a yes/no question or as a request to mention the credit cards accepted. The managers do not ask for Susan's intention, but propose by their different responses a way of taking her question. Clark comments: 'In each case, the manager's construal was validated by Susan and became their joint construal' (Clark 1996, p. 217). Once the hearer's interpretation is validated, there is no further issue as to what the speaker's intention really was (cf. also Cavell 1967). Now these kinds of cases should not make us neglect the fact that in very many cases the best way for the hearer to end up with an interpretation which will be accepted by the speaker is precisely to set out to recover the speaker's actual intention. However, this fact does not contradict that the ultimate goal of the hearer opting, as it were, for the speaker's intention is rather to obtain the speaker's agreement than a match with the original intention. Therefore, the hearer's opting for the speaker's intended meaning is most accurately to be understood as the hearer's opting for what the speaker presents as her intended meaning.

Two Possible Countercases
I have so far argued that the hearer's different approaches to the speaker's utterance do not support the idea that intentions settle utterance meaning nor that the hearer takes a genuine interest in the speaker's actual intention. In case the hearer does not want to play with the speaker's utterance nor wants to hold the speaker responsible for it, the hearer goes along with whatever meaning the speaker wants her to go along with. However, the hearer may ask the speaker for her intended meaning with a less casual attitude than the one I have hitherto considered. The hearer's tone in 'What do you mean?' might be inquisitive rather than inviting, the hearer seemingly being far from prepared to go along with whatever the speaker wants her to go along with. These are cases which intuitively could be conceived as cases where the hearer does set out to ascertain what the speaker's intention actually was.

Suspicion
In some cases the speaker makes an utterance which on a possible or even most reasonable interpretation reveals things which ought to be concealed. The speaker may have said too much about how things are in the world or in the speaker's mind. For clarificatory purposes, the hearer asks for the speaker's intention and the speaker delivers it. But the hearer suspects that the speaker has become aware that she said too much, wants to conceal it and pretends to have had a different communicative intention than the one she actually had. It seems to the hearer that there is a distinction to be made between the speaker's alleged intention and her actual intention. Is this not a case where the hearer wants to know the real meaning of the utterance or takes a genuine interest in the speaker's actual intention? Let us consider an example from Cervantes' Don Quixote. Sancho Panza pretends to be returning from a visit to Dulcinea. A little later, the conversation turns on the merits of Dulcinea compared to those of Dorothea. Sancho Panza says: […] As to beauty, I have nothing to do with it; and if the truth is to be told, I like them both; though I have never seen the lady Dulcinea.' How! never seen her, blasphemous traitor!' exclaimed Don Quixote; 'hast thou not just now brought me a message from her? 'I mean,' said Sancho, 'that I did not see her so much at my leisure that I could take particular notice of her beauty, or of her charms piecemeal; but taken in the lump I like her.' 'Now I forgive thee,' said Don Quixote; 'and do thou forgive me the injury I have done thee; for our first impulses are not in our control.' Cervantes (1605), i. 30.
Sancho Panza says that he has never seen Dulcinea which does not seem to be compatible with his having returned from a visit to her. This is why Don Quixote asks Sancho to explain himself. Sancho now suggests that what he meant to say in uttering 'I have never seen the lady Dulcinea' was something along the lines of 'I have never seen the lady Dulcinea so much at my leisure that I could take particular notice of her beauty', so that he cannot really tell who is the most beautiful of Dulcinea and Dorothea. As we can see, Don Quixote is reassured by Sancho's explanation. Don Quixote opts for what Sancho Panza presents as his intended meaning without further ado. But let us suppose that the hearer does not trust the speaker's veracity. Would the hearer then take a genuine interest in the speaker's actual intention?
First, it must be observed that the hearer is not in a position to make any definite claims as to the speaker's actual intention. As Fricker says: Lies about my own intentions and other mental states may be suspected, but cannot be refuted. Fricker (2012), p. 88.
It is true that the hearer may occasionally trust her own hypothesis about the speaker's intention more than the speaker's self-report. In the light of the evidence available to the hearer, the speaker's self-report may simply not be credible. But, second-in the cases of suspicion that we are here considering-the speaker's intention, whatever it may be, would in any case not serve the hearer's actual purpose. The hearer's real concern is with how things are. The speaker's intention, whatever it is, will at most be an indication of how things could be. In the case at hand, the hearer would want to settle whether Sancho Panza had been to Dulcinea or not. Whether Sancho Panza's actual intention was that he had not seen Dulcinea at all or not seen her appearance in an adequate way does not have any direct implications for this question. A less credulous hearer than Don Quixote would not take Sancho Panza's word about his intention to settle whether Sancho Panza had been to Dulcinea or not, but would proceed at an independent inquiry into this matter. For the purposes of this inquiry, there is no reason to attempt to get hold of Sancho Panza's actual intention.
The hearer's asking for the speaker's intention in cases of suspicion is in the service of an inquiry in which the speaker's intention is ultimately irrelevant. Therefore the interest taken in the actual intention behind the utterance is apparent only. The hearer rather entertains the possibly actual intention as an imagined meaning of the utterance and proceeds at a factual investigation into how things really are. The possible interpretation of the utterance serves as a mere springboard for further investigations. For the immediate purposes of the conversation the hearer may well go along with the possibly alleged intention.

Lying
In cases of suspicion of the kind we have just considered the hearer suspects that the speaker is not accurate about her original communicative intention. But the hearer's interest is not focused on whether the speaker lies or not about her intention, but whether the state of affairs which the speaker possibly tries to conceal by means of her alleged intention obtains or not. There are of course also cases where the hearer does want to establish whether the speaker lied about a state of affairs, i.e. whether she stated something which she believes to be false. In such cases of lying, the actual utterance plays a more important role than in cases of suspicion where it serves as a mere indicator of a possible state of affairs. But does this imply that such things as the actual meaning of an utterance or the speaker's actual intention matter in cases of lying?
According to many accounts of lying the speaker lies to the hearer if and only if 1. The speaker asserts that p to the hearer, 2. The speaker believes that p is false.
The first condition seems to require that the meaning of the speaker's utterance be settled, i.e. what the speaker asserted should be ascertained as an objective matter of fact. For many theorists a condition for the speaker's asserting something is that she intended to assert it, so for these lying depends on the speaker's actual intention.
I believe that cases of lying fall under the hearer's holding the speaker responsible for her utterance and should be dealt with accordingly. I have already argued that the speaker's responsibility parts from the hearer's most reasonable interpretation of her utterance, which implies that it is independent of such things as the actual meaning of the utterance and also the speaker's actual intention. The issue at stake is wholly epistemic and amounts to whether the hearer had the best reasons to take the speaker's utterance the way she did. Thus the first condition quoted above should run along the following lines: the hearer has the best reasons to believe that the speaker wants to convey p. I take legal practice to support this conception of lying and in order to show this I will consider the United States Supreme Court decision in Bronston v. United States which has become the legal standard of perjury.
The movie producer Bronston was convicted of perjury in the District Court because of his answers to a lawyer as a witness at a bankruptcy hearing.

Lawyer
Do you have any bank accounts in Swiss banks, Mr. Bronston? Bronston No, sir Lawyer Have you ever? Bronston The company had an account there for about six months, in Zurich. Burger (1973) As a matter of fact, Bronston himself had had an account in a Swiss bank for some years. The District Court noted that the last utterance is truthful on its face. Nevertheless, they argued that it constitutes a lie, since a reasonable implicature is: I have never had any bank accounts in Swiss banks.
For the District Court the last utterance was clearly intended to mislead the lawyer. The Supreme Court, however, reversed the sentence. Concerning Bronston's intentions they say: It may well be that petitioner's answers were not guileless, but were shrewdly calculated to evade. (Burger 1973).
The Supreme Court concedes that Bronston possibly intended to mislead. Nevertheless they object that the jury should not be permitted to engage in conjecture whether an unresponsive answer, true and complete on its face, was intended to mislead or divert the examiner […] (Burger 1973).
From this one might be tempted to think that the reason why the Supreme Court reverses the sentence is that they consider that there is insufficient evidence as to Bronston's communicative intention: Bronston cannot be convicted of perjury because it cannot be established that he intended the implicature in question.
However, the Supreme Court actually considers speaker intentions to be irrelevant. This is brought out by their account of an example given as an illustration by the District Court. The example runs as follows: [I]f it is material to ascertain how many times a person has entered a store on a given day and that person responds to such a question by saying five times when in fact he knows that he entered the store 50 times that day, that person may be guilty of perjury even though it is technically true that he entered the store five times. (Burger 1973).
The Supreme Court comments: The illustration given by the District Court is hardly comparable to petitioner's answer; the answer "five times" is responsive to the hypothetical question, and contains nothing to alert the questioner that he may be side-tracked.
[…] Moreover, it is very doubtful that an answer which, in response to a specific quantitative inquiry, baldly understates a numerical fact can be described as even "technically true." Whether an answer is true must be determined with reference to the question it purports to answer, not in isolation. (Burger 1973).
The lie in the imagined situation is technically speaking an implicature. However, the Supreme Court evidently holds that the innocence of literal meaning does not acquit the speaker from liability and that what is asserted depends on the context of the utterance. It would seem then that Bronston could well be convicted of perjury on account of his implicature, since neither the speaker's possibly innocent intention nor the innocent literal meaning of the utterance prevents the speaker from telling a lie. So why is not Bronston convicted of perjury after all? The Supreme Court's reversal of the sentence depends on the consideration that [i]f a witness evades, it is the lawyer's responsibility to recognize the evasion and to bring the witness back to the mark, to flush out the whole truth with the tools of adversary examination. (Burger 1973).
In other words, the problem with District Court's decision is that it depends on an interpretation of Bronston's utterance, which may well be reasonable, but which is not the most reasonable interpretation in the present context. Here it should be pointed out that what is to be considered the most reasonable interpretation in one context is not necessarily the most reasonable interpretation in another context. The special circumstances and standards of the particular context affect what the most reasonable interpretation is. The standards are naturally more demanding in a courtroom context than in everyday casual conversation. In any case, Bronston is acquitted from perjury, not because it cannot be excluded that he did not want to make the implicature in question nor because of the literal meaning of his utterance, but because his utterance is not most reasonably taken to implicate that he never had any bank accounts in Swiss banks. The conception of lying which I have here outlined contrasts with most philosophical accounts where literal meaning and intentions play a dominant role. A recent theorist of lying says e.g.: One can avoid lying, while still being misleading, if one can exploit which question the hearers can be expected to accommodate, while remaining in a position to claim that one intended to address a different QUD [question under discussion]. Stokke (2018), p. 128.
What the speaker presents as her actual intention is here taken to be decisive, as if it sufficed for the speaker in order to avoid liability to point out that her alleged intention is licensed by the linguistic meaning of the sentence uttered. Bronston and the fictive store enterer are certainly in a position to claim that their intentions were different from the hearer's interpretation. Nevertheless, the Supreme Court does not let them escape liability on that account. Legal practice as exemplified by Bronston v. United States lends support to the view that whether a speaker lies or not does not depend on what is the possibly intended meaning, but on what is actually the most reasonable interpretation. Lying is thus also a case where neither objective determination of utterance meaning nor interest in the speaker's actual intention occurs.

Conclusion
I have argued, from the consideration of what I hope is the complete variety of a hearer's approaches to a speaker's utterance, that (1) the speaker's intention does not settle the meaning of her utterance and (2) the hearer does not take a genuine interest in the speaker's actual intention. The reason why the speaker's intention does not settle utterance meaning is simply that no utterance meaning determination, as presupposed by intentionalists and anti-intentionalists alike, takes place. This result is in accordance with what some theorists have already argued notably in the case of demonstrative reference (see, e.g., Ariel 2002a, b;Pagin 2002;Smit 2012;Heck 2014;Gauker 2015;Åkerman 2015;Bach 2017). Further, in the regular course of interpretation the hearer is concerned with what the speaker presents as her intention. For this reason it seems that the hearer's goal is to come up with an interpretation which the speaker will accept rather than an interpretation which corresponds to the speaker's intention. Also in the case of accountability, suspicion and lying, the speaker's actual intention is irrelevant. But even though the speaker's actual intention does not play a constitutive or determinative role in interpretation, the speaker's hypothetical intention nevertheless plays an important regulative role: in most cases nothing will be a reasonable interpretation unless it corresponds to what the hearer had good reasons to take as the speaker's intended meaning. Thus the anti-intentionalist theses which I have argued for do nothing to diminish the fundamental role of speaker intentions in interpretation.
The picture of the speaker's and the hearer's interpretive practice which these considerations suggest is what could be called the conversational model of utterance interpretation. Faced with an utterance of the speaker's the hearer has basically three options which differ in frequency and sphere of employment, but in a fundamental sense are on a par. The hearer may opt for 1. The speaker's intended meaning or, more precisely, what the speaker presents as her intended meaning; 2. The hearer's interpretation or the most reasonable interpretation of the utterance, i.e. what the speaker is most reasonably taken to say; or 3. An imagined meaning of the utterance, i.e. what the speaker could be imagined to have said.
One might now object that this conversational model of utterance interpretation has bearing, at most, on the anthropology or sociology of communication and the epistemology of interpretation, but no bearing at all on the issue which philosophers of language are or ought to be concerned with, namely the metaphysics of meaning (Harris 2016;Neale 2016). What counts as meaning for practical purposes may have nothing to do with meaning as such. What counts as a lie in court may have nothing to do with the philosophical concept of lying. But is it really necessary to have notions of utterance meaning or lying in addition to the conceptions of meaning and lying which are actually operative in our interpretive practice? If not, there is perhaps no metaphysics of meaning.