F I C T I O N AS A C T I O N

SARAH HOFFMAN

I Several accounts of the nature of fiction have been proposed that draw on the conceptual resources of speech act theory. Searle has argued that what characterizes fiction is non-deceptive pretending to assert.~ Currie at one time held that fiction is (the product of) a distinct kind of illocutionary act but more recently has developed an account in terms of Gricean intentions. 2 More recently still Martinich has offered a theory of fiction that turns on the suspension of the general prohibition against non-defective performance of a speech act, that is on the suspension of the conversational maxim of Quality? I argue that speech act theory is insufficient for the tasks of a theory of fiction, in that it does not provide the necessary resources for understanding what is most interesting and distinctive about the institution of fiction. This weakness can be illustrated in two ways. The first way arises in connection with Walton's account of the nature of fiction, which I argue can plausibly be reconstructed in terms of the concepts of speech acts and communicative action. Walton explicitly disavows speech act approaches to fiction, yet his own theory can be cast in its terms. Taken in combination with the main objections that Walton presents against communicative theories of the nature of fiction, this shows us the limits of understanding the nature of fiction from the perspective of communicative acts. However dependent the institutions of fiction are on the basic intentional structure of communication, focus on this structure diverts attention from works themselves in an unhelpful way.

513

SARAH HOFFMAN

The second way involves Searle's and Martinich's theories both of which are incomplete, I argue, in that they allow counterexamples and ignore the central role of the audience in the institution of fiction. In addition, Currie's earlier proposal fails to meet Searle's general argument that there cannot be distinctively fictional illocutionary acts. Speech act theoretic approaches to fiction, however, do enjoy some success. Currie's most recent proposal in terms of Gricean intentions captures an essential part of the nature of fiction. Even this account, however, suffers from the weakness that appears to be inherent in developing a theory of fiction from the perspective of speech acts or speaker intention. Such accounts do not provide the resources to account adequately for the most interesting processes of interpretation in our engagement with fiction, namely those by which we construct the content of fictions and through which fictions achieve or fail to achieve their intended effects. Ii Martinich's recently proposed theory of fiction turns on a version of Grice's theory of conversation in which Grice's maxims of conversation are generalized to Supermaxims. In normal nonfictional contexts, conversation is governed by four Supermaxims: Quality, Quantity, Relation, and Manner. Martinich has generalized Grice's maxims so that they cover alt speech acts and not just connative ones. 4 Quality constrains us to participate in speech acts only if we satisfy all the conditions for their nondefective performance. This rules out asserting statements I do not believe to be true, promising things I have no intention of doing, ordering people over whom I do not have the appropriate authority to do so, etc. There are contexts in which various of the maxims are for various reasons not operative. In fictional contexts, Martinich argues, the Supermaxim of Quality is not operative because it has been suspended. Martinich proposes that the suspension of a maxim consists in the failure to fulfill a maxim "in set circumstances in which the maxim does not always apply" (p.100). He identifies these circumstances as institutions, and fiction in particular as an informal institution in which Quality is suspended.

514

FICTION AS ACTION

Thus Martinich's general claim is that fiction differs from nonfiction in that fiction is language use in an (informal) institutional context in which participants in conversation are not always bound to participate only in speech acts that they can nondefectivety perform. For instance, because fiction suspends the Quality maxim I can assert "Sherlock Holmes lived in Baker Street" even though I do not believe that Sherlock Holmes lived in Baker Street (since he does not, and never did, exist). In addition, Martinich argues that in such contexts we do not have a failure of reference, again because the maxim of Quality is suspended. For reference to succeed all that is required is an intentional object. Thus even though Sherlock Holmes does not exist, when I utter "Sherlock Holmes lived in Baker Street" I manage to refer to the intentional object Sherlock Holmes. Martinich claims for his theory the virtues that any fully acceptable account of fiction should possess: it is "consonant with ordinary pretheoretic talk about fiction", "part of a general theory of language", and "consistent with a sensible ontology" (p.99). Martinich's theory certainly has the second virtue, but the price for the others is giving up the idea that we can only refer to objects that exist. Only by referring to nonexistent fictional objects can we make true claims about them (thus preserving the first virtue) without introducing "a special fictional world populated by fictional characters and events" (thus preserving the third virtue) (p.99). The key, then, is that, despite the fact that they do not exist, reference to fictional objects "does occur and is able to occur because the maxim of quality is suspended" (p. 104). Why suspension of Quality is required in order to secure reference to fictional objects, however, is not entirely clear. The way Martinich describes it, such reference actually looks rather like it is just a general feature of proficient language use. As he points out, once speakers have reached competence in their language they can easily talk about objects that are not or have never been present to a hearer. Moreover, "it is no more difficult to introduce real but unexperienced objects than it is to introduce fully fictional ones. Because referring is an intentional activity, only an intentional object is needed. Sometimes the intentional object is also an existing object, and sometimes

515

SARAH HOFFMAN

not" (p. 106). Successful reference requires only that the object referred to be thought of or mentally identified, not that it exist. If this is correct, then it is not really clear why the maxim of Quality needs to be suspended at all. We c a n refer to fictional objects (though they do not exist) and thus doing so would appear not to involve us in any violation of Quality. Some may find the contention that we can refer to nonexistent objects troubling. I will not, however, challenge this claim. Instead I will focus on a generalization of my point in the previous paragraph in order to show that Martinich's theory undermines itself. This problem arises because of the solution that Martinich proposes to the problem of truth in fiction. He tells us that "a statement of fiction is true if and only if some author or narrator asserts it" (p. 108). If reference to fictional objects is possible in the general case, then it does not require a suspension of Quality. So, by the same reasoning, if statements of fiction are true then asserting them does not require a suspension of Quality either. Consider a simple story, one composed in declarative sentences and in which no complications arise from narrator unreliability, internal inconsistency or authorial error. The story begins, "Sherlock Holmes lived on Baker Street." Martinicb's theory entails that the statements of this story are true, since the author asserts them. Why then do we need to invoke a suspension of Quality to explain what fiction is? The maxim of Quality constrains us to perform speech acts only if all the conditions for their nondefective performance hold. Fictional contexts can create a bar to nondefective assertion because assertion requires both that the speaker have some reason(s) or evidence for the truth of what is asserted and also that she believe what is asserted. If Sherlock Holmes never existed then what kind of evidence could I have that he lived in Baker Street? If I know he never existed then I will not believe that he lived in Baker Street. Thus I will not be in a position to nondefectively assert that he did. But if, as Martinich claims, statements in fiction are in fact true then why should either of these conditions fail? Having read the stories I will have evidence that Holmes lived in Baker St and I will believe that he did. And why, then, does the Supermaxim of Quality need to be suspended for me to assert that he did?

516

FICTION AS ACTION

Perhaps it is the speech acts of the author that become possible because of the suspension of Quality. This possibility is ruled out, however, by an argument parallel to the foregoing. Presumably the author has some character(s) and story in mind when composing her fiction. When asserting things, she refers to those characters and is in, as Martinich says, "a privileged position with respect to the events she reports" (p. 109). So the author is also already in a position to assert whatever she does assert nondefectively. She knows that what she is talking about is her story and she knows what is true in her story. So, again, there is no reason for the suspension of Quality. My conclusion here is that if Martinich is right about reference to fictional objects and truth in fiction, then he is wrong about the suspension of Quality characterizing fictional discourse. Yet whatever we make of the apparent conflict between Martinich's solution to the problem of truth in fiction and the feature he identifies as characterizing fiction, his account as it stands needs to be supplemented. We can see this from the fact that his own example of the suspension of Quality is an example where no fiction is produced: "In the United States, a member of the House of Representatives cannot be prosecuted for anything he says on the floor of the House. That is, the maxim of Quality for statements is suspended" (p.100). This illustrates that Quality can be suspended in a variety of ways and a variety of contexts--not all of which produce fiction. Suspension of a maxim of conversation requires an institution, but Martinich has not told us what differentiates the institution of fiction from other institutions which allow or require the suspension of Quality. To that extent, his account is incomplete and in need of supplementation. Adequate supplementation can in fact be found in the theories of fiction developed by Currie and Walton. For Currie and Walton the essential aim of the institution of fiction is to provoke the imaginative engagement of an audience with works through pretense or make-believe. I will take this up below. First, consider another account of fictional discourse that makes use of the concepts of speech act theory and falls short for something like the same reason as Martinich's. Searle proposes that what characterizes fiction is a particular style of failing to assert, namely, non-deceptively

517

SARAH HOFFMAN

pretending to assert. (EM,p.65) There are, however, at least two problems that undermine Searle's theory? First, it rules out strictly historical fictions in which authors actually do assert what they utter. If such fictions are possible, then the acts of fiction-making that create them are also acts of assertion. We would have to say that their authors were somehow both asserting and pretending to assert the very same propositions with the very same utterance acts. But this seems at least strange, if not necessarily unacceptable. Second, and more problematic, pretending to assert is not sufficient to create fiction. There are a variety o f circumstances where people behave in ways that are plausibly described as non-deceptive pretense at assertion yet which do not create fiction. Currie briefly describes two cases of this kind: verbal performances intended "to illustrate an idiotic line of reasoning," or "to imitate the conversational manner of an acquaintance" (NF, p. 17). These are cases of non-deceptive pretense of assertion. But these performances create fictions only in the sense that they create falsehoods. Something more seems to be necessary to produce fiction. Searle's account, like Martinich's, is missing an ingredient. One way of articulating the objection to Searle's theory is to note that it locates the important element of pretense involved with fiction in the actions of the author rather than in those of the intended audience. Readers, and audiences in general, are intended to make-believe or pretend that the content of a fiction is factual, but it is a mistake to think that an author merely pretends to perform speech acts. Pretense is essentially involved in fiction, but it is the pretense of audiences that matters, and this is the ingredient missing from both Martinich's and Searle's theories. This brings us to a third proposed speech act theory account of fiction, one that incorporates audience pretense or make-believe. Currie has claimed that "the writer of fiction engages in an act of speech that is distinctively fictional, and distinctively different from speech acts such as asserting, requesting, questioning, and so forth" ("WF," p.304). The writer's intention in performing these distinctively fictional speech acts is that her audience make-believes the sentences that she utters.

518

FICTION AS ACTION

Contrary to what Currie maintains here, however, there cannot be a distinctively fictional speech act. As Searle has argued, the illocutionary act performed in uttering a sentence is a function of the meaning of the sentence, and "if the sentences in a work of fiction were used to perform some completely different speech acts from those determined by their literal meaning, they would have to have some other meaning" (EM, p.64). But we know that sentences in fiction do not have novel fictional meanings--they have the same meanings that they have when used in non-fictional contexts. So fiction cannot be characterized by some unique speech act(s). Currie constructs an argument against Searle on this point, taking the ground of the objection to be the principle that there cannot be sentences with the same meaning that are nevertheless used to perform different speech a c t s . 6 This principle does entail that the sentences of fiction cannot be used to perform special fictional speech acts. If this principle is correct, then in order to be used to perform different speech acts the same sentence would have to have different meanings in fictional and non-fictional contexts. We know this is not the case and, further, we know that sentences with the same meanings can be used to perform a variety of distinct speech acts. The utterance "You can go now" could be an assertion, but could equally well be an order or a question. If so, the principle on which Searle's argument rests must be rejected and it would seem that distinctively fictional speech acts are possible. Currie's reconstruction of Searle, however, is misleading. The principle he identifies is not what underlies Searle's claim that there cannot be distinctively fictional speech acts. Searle allows more than one illocutionary act to be associated with one sentence meaning, saying that "in general the illocutionary act (or acts) performed in the utterance of a sentence is a function of the meaning of the sentence" (EM, p.64, my italics). In fact the ground of Searle's rejection of fictional speech acts is the principle that"for any meaning X and any speaker S whenever S means... X then it is possible that there is some expression E such that E is an exact expression of or formulation of X. ''7 This is the principle of expressibility and it entails that "for any possible speech act there is a possible linguistic element the meaning of which (given the context of

519

SARAH HOFFMAN

utterance) is sufficient to determine that its literal meaning is a performance of precisely that speech act" (SA, p.21). The idea behind the expressibility principle is that there are no speech acts that cannot in principle be performed directly. The principle does not, however, entail that any particular possible speech act can only be performed directly. Unlike the Currie's principle, expressibility is not compromised by the fact that speech acts can be performed indirectly. On the contrary, the expressibility principle is worth formulating in part precisely because the same speech act can be performed by producing different utterances and because the same utterance can be to perform different speech acts. In other words, expressibility is worth noting because of the possibility of indirect speech acts. It points to a methodology for constructing a theory of speech acts in this rather complicated situation. Searle's principle of expressibility tells us that to understand the nature and structure of the various speech acts it is enough to look at instances where they are performed directly (SA, p.21). How then does the principle of expressibility show that fiction-making cannot be a special kind of illocutionary act? As Searle's argument indicates, it entails that if there were speech acts unique to fiction, the sentences in fictions would have to have different meanings from those that they have in non-fictional contexts. Why? Because the principle of expressibility asserts that for any possible speech act there is a possible linguistic element with a literal meaning such that its utterance (in the right context) constitutes the performance of that speech act. In particular, if there is a unique fictive speech act there must be some possible linguistic element that fits this requirement. But in fictions what we get are the regular linguistic elements that get used in non-fictional contexts. The fiction-maker produces linguistic elements whose utterance in the right circumstances constitutes the performance of the regular, non-fictional speech acts of assertion, direction etc. Both the novelist and the journalist may write, "It was a dark and stormy night." If these utterances have a different illocutionary force than their literal meaning dictates, then the principle ofexpressibility is false and Searle's program of systematizing speech act theory fails.

520

FICTION AS ACTION

More recently Currie has dropped the language of illocutionary force, but maintains nevertheless that an author creates fiction by performing a particular fictional orfictive communicative act. (NF, p. 11) He provides the following formal characterization of fictive utterances (NF, p.31 ): (Do) U's utterance of S is fictive iff U utters S intending that the audience will: (1) Recognize that S means P (2) Recognize that S is intended by U to mean P (3) Recognize that U intends them to make believe that P (4) Make believe that P And further intending that: (5) (2) will be a reason for (3) (6) (3) will be a reason for (4) 8 An utterance that satisfies these conditions may superficially look like an assertion but has fictive rather than assertive force. It is an act that involves the characteristically complex Gricean intention the recognition of which secures illocutionary uptake. It is a genuinely communicative, fictive speech act. What Currie has identified is not, however, a distinctively fictional speech act, but an ordinary one that also appears in non-fictional contexts. It is a directive, albeit one that aims at a particular (and particularly fictional) perlocutionary effect, namely getting its audience to imagine or make-believe the propositional content of the directive. Once we see that it is this intended effect on an audience that characterizes fictional discourse, it become clear that it is the notion of make-believe that is doing the essential work for the account of fiction. Moreover, even if we accept that Currie's theory of fiction is basically correct, it is clear that the business end of the communicative act that characterizes fiction is really the audience end. This becomes apparent when we consider what Cmxie has to say about truth in fiction. It turns out that here the author drops out of the picture. 9 Authorial intention does not determine the content of fictions. And this, I believe, shows the limitations of any account of fiction that makes communicative speech acts central. Such accounts put too much emphasis on authorial intention. What matters for fiction is how it is read far more than how it is produced.

521

SARAH HOFFMAN

Fictions may well be dependent on the intentional structure of communicative acts. And this ought not to surprise us, after all, many fictions take linguistic form. But it does not follow from this that what is especially characteristic of fiction is something about its communicative structure. This claim is strengthened by considering a final account of fiction which makes audience pretense or make-believe central to fiction, the theory that Walton presents in Mimesis as Make-Believe. Walton expressly disavows speech act theory as an approach to understanding the nature of fiction, yet the description of the nature of fiction we find in his theory turns out to have striking similarities to Currie's fictive intent account. Indeed we can recast Walton's theory as a speech act theory of fiction. By itself this should further alert us to the limitations of speech act theory as a framework for understanding fiction. According to Walton's account, a proposition's fictionality consists in a prescription to imagine it. Such prescriptions are embodied in ganaes of make-believe--these are the generators of fictional propositions. Fictional works and fictional propositions are related through games of make believe: works generate fictional propositions by virtue of their use in games of make-believe. Any game of make-believe is constituted by a set of rules, implicit, explicit, or both. Games of make-believe involving works of fiction have rules which, in concert with the work, generate propositions that are to be imagined. The last point has a crucial implication. The prescription to imagine that constitutes the fictionality of propositions entails that works of fiction are vehicles for directive illocutionary action. Prescriptions are a kind o f directive. Thus works of fiction are utterances used to perform directive speech acts, aimed at getting us to make-believe their contents. So we have here a case of indirect illocutionary action, exactly as we had in Currie's theory. These accounts both then make fiction the vehicle for the same kind of speech a c t - - a directive aimed at a particular perlocutionary effect, namely getting an audience to imagine or makebelieve. Currie's and Walton's accounts are in this respect substantially the same.

522

FICTION AS ACTION

But this may be a bit hasty. After all Walton himself argues that speech act approaches are "remarkably unhelpful in explaining fiction". (MM, p.76) It may seem odd to make this claim while at the same time offering what is essentially a speech-act theoretic theory of fiction. When the content of his objections are appreciated, however, we can see that there is not in fact a serious conflict here. Walton's rejection of a speech act account of fiction has two aspects. The first relates to the scope of the concept of fiction: Walton believes that there are authorless fictions and this doesn't seem to be consistent with a theory that characterizes fiction as essentially the vehicle for intentional action. The second relates to the appropriate role of authors in interpretive practice. Walton appears to think that speech act approaches to fiction wrongly make the action of making fiction too fundamental. Walton maintains that we cannot take sentences to be actually asserting anything except insofar as we understand them to be the product of a speaker (or inscriber) who intends to assert something with them. We can, however, take sentences to be telling a story even if we do not take them to be the product of a speaker (or inscriber) who intends to tell a story. In other words, Walton claims that there can be fictions without authors. But fictions without authors cannot be taken to be the products of intention. Imagine a naturally occurring story. There is a large rock on which there are cracks in the shape of sentences that tell a story. Walton says of such a case that the "realization that the inscription was not made or used by anyone need not prevent us from reading and enjoying the story in much the way we would if it had been... Some dimensions of our experiences of authored stories will be absent, but the differences are not ones that would .justify denying that it functions and is understood as a full-fledged story" (MM, p.87). When being read as a story, the cracks in the rock are a work of fiction, but they are not the product of an agent who is using them to perform an illocutionary act. This appears to be a problem for any claim that what characterizes works of fiction is that they are vehicles for a particular kind of directive illocutionary action.

523

SARAH HOFFMAN

In this case there is no agent, so there cannot be any action. But look at this case a little more closely. Walton wants to claim that you can take something to be the vehicle for a story even if you do not also take it that there is an agent using the object as a vehicle. But this can't be right. By Walton's own theory, to read something as a story is to take it that there are prescriptions to imagine certain propositions. A bunch o f s c r a t c h e s in a r o c k c a n resemble the w o r d s that are c o n v e n t i o n a l l y used to get you to i m a g i n e certain p r o p o s i t i o n s . Recognizing this is true of a bunch of scratches is not the same as reading them as a story however. If you go beyond noting the resemblance to actually attributing fictionality to the propositions, you imply that there is a set of prescriptions to imagine the propositions that make up the story. In reading the scratches as a story you are taking them to be the vehicle for a request, a request to imagine certain propositions. And it is not the case that you can take something to be the vehicle for a request without also taking it that there is an agent intending that the thing is such a vehicle. A directive is an illocutionary act just like an assertion. If we cannot have one without an agent then how can we have the other without an agent? Walton's claim seems more plausible than it really is because of the way he phrases it, but the realization that an agent did not produce the cracks in the rock does, I think, prevent us from simply reading the story. I do not dispute that we can read scratches in rocks as telling stories. We can. I do not mean to imply that if we do such a thing we believe that there really is some agent who inscribed the story. We don't. I do think that what is happening if we 'read the story' is that we are at least pretending that there is a directive to imagine the content of the story. In fact we are not merely pretending that there is a directive; we are making it the case that there is one, one generated by our own intentions. We are in effect the authors of the story. Treating a natural object as if marks on it relate a story is a way of turning it into an artifact that expresses certain intentions. This is one way o f making a prop, as Lamarque has pointed out. We do not make the rock or inscribe its markings, but we take it as we find it and make it into a prop in our game of make-believe, t~ In order even to read a work as having propositional content, we must be

524

FICTION AS ACTION

positing an agent with certain kinds of intentions who makes use of the work to get us to recognize those intentions, and thereby create certain effects on us as readers. It may be that we are ourselves that agent, but an agent there is nevertheless. What then of the contention that speech act theory approaches to fiction make the action of fiction-making central in a way that only works of fiction ought to be? According to Walton, speech act theories propose that fiction ~ to be understood in terms of the actions whereby works of fiction are produced... But [this] is exactly backwards. The basic notion is that of works of fiction, or rather, that of things, whether human artifacts or not, which function as works of fiction do, not the notion of acts of fiction making" (MM, p.77). The issue here could still be one of demarcation. Walton thinks that speech act accounts of fiction get wrong which things really are fictions. They do not give fiction a wide enough scope. If this is really the nature of the disagreement, then it is mainly a matter of emphasis: Walton sees interaction with works in games of make-believe as o f primary importance in determining a work's status as (non)fictional, while a speech act approach insists on conditions of origin as determinative. If we focus, as Walton does, on the role that objects play in our institution of consuming and interpreting fictions, rather than the intentions with which they were produced, then certain works will count as fiction even though they were not created with fictive intention. Such fictions could be of two kinds: those that are authorless and those that are read as fiction but created with non-fictional intentions. The authorless case has already been discussed: An author is always implicitly involved in our interactions with naturally occurring stories and other such examples, even on Walton's own account. We can admit Walton's point that we interact with these as fictions without conceding that this makes fiction entirely independent of fictive intent. On the other hand, we have works treated as fictions that have authors, but whose authors' intentions in creating them were not fictive. An example of this type, one that that Currie discusses, is the Bible. It can be read and enjoyed as fiction, as indeed it is by many people. If this became the main way in which the Bible was read would it follow that the Bible wag

525

SARAH HOFFMAN

fiction? On Walton's view that interpretive practice is the fundamental determiner of fictional status it would be. Yet Currie objects to this. According to Currie, the most we should say that the Bible in such circumstances is read as fiction, but is not really fiction because it was not produced with fictive intent. So again we have a dispute about which things really are fictions. Currie points out that when we treat works as if they are fictions (even if they were not created with fictive intent), then that is all there is to it: We are treating them as if they were fiction, though they in fact are not. (NF, p.36) Walton wants to maintain that treating these works as fiction makes them fiction, independent o f the intentions with which they were produced. I ' m not sure that we have to take sides in this particular debate, or that we should even want to. In fact we have in place all the conceptual resources needed to understand the situation without choosing. We can refuse to endorse the claim that all there is to fiction is intent and a't the same time refuse to endorse the claim that all there is to fiction is being interpreted as fiction. The central cases of fiction are those that are both created and read as fiction, the cases where these come apart serve to highlight the fact that the institution o f fiction has both a front and a back-end. The Bible is aptly chosen by Currie to make his case against Walton. To call the Bible fiction would be inflammatory to Jews and Christians who take it more seriously than the word 'fiction' might suggest. But it nevertheless seems clear that intuitions about this sort of case can be accommodated within the speech act framework. We can distinguish the intentions with which a work is produced from the intentions that are implicitly attributed to its creator (or Creator) when it is interpreted in a certain way. If we take ourselves to have good reason to believe that a work is largely or completely false, then there would not seem to be much point in reading it as non-fiction even if we recognize that it was intended to be so read. Nevertheless, there may still be very good reasons to engage with the work through the kinds o f games of make-believe that Currie and Walton make fundamental to fiction. Works, both fictional and non-fictional, are to some extent independent of their authors. This is not a new point but it seems important to make it again in the

526

FICTION AS ACTION

circumstances; failed authorial intentions are not restricted to fiction. This raises an interesting aspect of the concern Walton expresses about speech act theory approaches to fiction. Rather than the fairly narrow question of scope, the spirit of his objection may really be closer to worries about the intentional fallacy. Accounts of the nature of fiction that make real authorial intention basic rather than works and their properties--such as speech act theoretic approaches--run the risk of importing irrelevant information into the process of interpretation. Making fiction a function of authorial intention puts us at risk of basing our interpretations of works on unachieved authorial intention rather than intentions achieved and present in the work. We should really want works to be at the center of the institution of fiction in our account rather than authors or creators. This is in general a legitimate worry, but it is not a consequence of Currie's a c c o u n t as I have read it, nor of the speech act theory reading of Walton. Indeed, what Currie says about truth in fiction shows that authorial intention (independent of features present in a work) all but disappears from interpretation. In describing the process of determining what is true in a fiction, Currie posits an implied author who is the teller of the story who is distinct from both the narrator and the historical author. The contribution that the intentions of the historical author make to our interpretation is limited to whatever is achieved in the text itself, though in some cases it may be a bit more than this. Currie argues that the fictional author that is created in our interpretive practices has the conventional beliefs of the community of the historical author, unless the text itself provides evidence that they are unconventional in some way. If the historical author had conventional beliefs then their intentions would on this scheme turn out to contribute more than just what is achieved in the text." The important point, however, is that this description of the process makes it clear that Currie's account does not make acts of fiction-making more basic than works of fiction themselves. When we engage fictions and figure out what their content is we look to the features of the work, not to authorial intentions regarding content that have not been realized in it. This certainly seems like the right strategy. The main evidence that we have for determining authorial intent, and thereby the content of a work, is

527

SARAH HOFFMAN

the work itself. But this is not undermined by the idea that fictions are what authors with fictive intentions produce. What is shown by this move away from authorial intent and towards the work itself is the weakness of speech act theory as a tool to address what may well be the more interesting questions about fictions, those about our engagement with them and the ways in which they achieve and fail to achieve certain effects--whether intentionally or not. We have seen that the speech act theoretic accounts of fiction provided by Martinich and Searle are not adequate. They do not properly identify the nature of fiction. But we have now also seen that even Currie's most successful theory in terms of Gricean intentions is incomplete. This does not show that speech act theory is irrelevant to understanding the nature of fiction. It does appear, however, that speech act theory either makes the authorial action of fiction making too central or it does not by itself provide a full account of fiction. Even if it provides part of the answer to the question of what fiction is, speech act theory cannot begin to address the evaluative and aesthetic questions that art and fiction raise. UNIVERSITY OF SASKATCHEWAN S A S K A T O O N , SK S7N 5A5 CANADA

e-maih [email protected]

2

3

528

NOTES Searle, John, "The Logical Status of Fictional Discourse," Expression and Meaning: Studies in the Theory of Speech Acts, (New York: Cambridge University Press, 1979) pp. 58-75; hereafter abbreviated EM. The earlier account is found in Currie, Gregory, "Works of Fiction and Illocutionary Acts," Philosophy and Literature 10 (1986): 304-308; hereafter abbreviated "WF". The more recent theory is found in The Nature of Fiction, (New York: Cambridge University Press, 1990); hereafter abbreviated NF. A.P. Martinich, "ATheory of Fiction," Philosophy and Literature 25 (2001): 96-112.

FICTION AS ACTION

See Martinich "Conversational Maxims and Some Philosophical Problems," Philosophical Quarterly 30 (1980): 215-228. 9 Currie (NF, pp. 12-18, and "WF") makes these and other objections to Searle's account. See also Walton, Kendall, Mimesis as Make-Believe: On the Foundations of the Representational Arts, (Cambridge: Harvard University Press, 1990); hereafter abbreviated MM. 6 NF, p.14. Currie also discusses Searle's argument in "What is Fiction?'" Journal of Aesthetics and Art Criticism 43 (1985): 385-92. 7 John Searle, Speech Acts, (Cambridge: Cambridge University Press, 1969) p. 20; hereafter abbreviated SA. s I have given the first of his three formulations. He revises this to accommodate non-literal language, and the possibility of the author not having a particular audience in mind. Using either would, however, only serve to unnecessarily complicate matters. NF, pp. 75-80. ,o Peter Lamarque, "Essay Review of Mimesis as Make-Believe," Journal of Aesthetics and Art Criticism 49 (1991): 161-166. H NF, pp. 75-80. The success of this strategy and Currie's general approach to truth in fiction has been challenged by David Davies in "Fictional Truth and Fictional Authors," British Journal of Aesthetics, vol. 36 (1996): 43-55.

529