Perhaps the most popular accounts of authorial activity are those that invoke speech act theory.[1] Although speech act theory can been formulated in a number of different ways, I am going to follow Searle in taking there to be four basic types of speech acts: utterance acts, propositional acts, illocutionary acts, and perlocutionary acts.[2] Utterance acts involve the production of token expressions (paradigmatically sentences) of a language, such as “She is at the bank.” The product of a successful utterance act will be meaningful (in the sense of character), but it need not have propositional content. Someone might, after all, utter “She is at the bank” as an example of an ambiguous sentence. Propositional (or locutionary) acts are uses of linguistic expressions that that express propositions or have propositional contents. When successfully performed, contextual parameters (including, but not exhausted by, the speaker’s psychological states) disambiguate the deployed expressions and determine the extensions of any indexicals. In a given context, an utterance of “She is at the bank” might, for example, express that proposition that Mary is (on September 2, 2003) at the TD Canada Trust branch in Lethbridge, Alberta on 5th Street South. Illocutionary acts are communicative acts. They can be thought of as propositional acts performed with illocutionary force, roughly the combination of a basic purpose of some sort – to assert something, or get someone to do something, etc. – and number of presuppositions regarding such things as the psychological state of the speaker, the existence of the requisite sort of  conventional procedures, etc. Typically, illocutionary acts are successful when the listener(s) recognizes the speaker’s illocutionary intentions.[3] Finally, perlocutionary acts are acts performed by means of the performance of illocutionary acts. More to the point, a perlocutionary effect is one that stands causally downstream from the effects of successful illocution. So, for example, by means of an utterance of “She is at the bank,” a speaker might get a listener to recognize that she intends to assert that Mary is at the TD Canada Trust branch in Lethbridge on 5th Street South. And by means of this successful illocution she might cause the speaker to come believe that Mary is so located or, perhaps, to fear for Mary’s well-being, if the bank in question is being held up.

            Speech act analyses of authorial action can be fruitfully taxonomized in terms the type of speech acts they invoke. Given that authors do not believe the stories they write, not do they (typically) intend their readers to do so, the suggestion that authors assert the sentences contained in their texts is not generally thought to be tenable.[4] Nonetheless, many have modeled authorial action on the illocutionary act of assertion. Searle and Lewis, for example, have argued that although authors do not assert the sentences they produce, they pretend to do so.[5] Currie and others have argued, in contrast, that authors perform uniquely fictive illocutionary acts.[6] Hoffman has defended the related view, that what is characteristic of fiction-makers is that they perform uniquely fictive perlocutionary acts.[7] And Beardsley and Ohmann have defended the view that authors do not perform illocutionary acts at all.[8] Instead they perform propositional acts which lack illocutionary force.  In this chapter we will consider each of these approaches in turn. It is worth noting, however, that at the end of the day I will argue that the views Beardsley and Ohmann are headed in the right direction, although they don’t quite go far enough.  Instead of treating authorial action as the performance of propositional acts, it is better thought of as the performance of utterance acts.

            There is, however, a cautionary note that needs to be made before we begin. My focus, in this chapter, is on the speech act analysis of authorial action. My skepticism regarding such analyses stems a more general skepticism regarding written speech acts. Inscriptional activity by itself does not, in general, constitute the performance of an illocutionary act of any sort. Minimally, the products of this activity have to be deployed in a specific way, perhaps by waving a placard or passing a note, in order for an illocutionary act to be performed. And not only is there no characteristic behaviour authors engage in using their texts, even if there were, it is far from clear that by means of a single action an author would thereby perform each of the multitude of speech acts corresponding to the sentences that constitute her text.[9] Many of the views to be discussed in this chapter, however, gain a superficial plausibility as a result of a failure to properly distinguish between authorial activity and storyteller activity.[10] Storytellers, of course, do perform substantial speech acts, and the correct analysis of these acts is an interesting question in its own right. But not all authors ever perform (or even release) their works and storytellers often perform works they did not author. As a result, no conclusions regarding authorial activity can be drawn from an account of that of storytellers.

            One final comment: some of the views we will be considering analyze authors’ (or storytellers’) speech acts in terms the speech acts of the narrators – fictional characters – of the texts they produce. But narrator speech acts are of independent interest, regardless of any role they might play in the analysis of authorial action. Moreover, in chapter 4 below, I will be invoking these subjects as part of an account of reader/ listener attitudes toward fiction. As a result, in order to avoid repetition, I will postpone discussing narrator attitudes until the aforementioned chapter. 


I: Pretense


            Lewis and Searle both argue that authorial action is a type of pretense. In particular, they claim that what authors pretend to do is to perform illocutionary acts. Searle says, “… the author of a work of fiction pretends to perform a series of illocutionary acts, normally of the representative type.”[11] And Lewis says,

“Storytelling is fiction. The storyteller purports to be telling the truth about matters whereof he has knowledge. He purports to be talking about characters who are known to him, and whom he refers to, typically, by means of their ordinary names. But if his story is fiction, he is not really doing these things.”[12]

For simplicity I am going to focus here on assertive pretense. But given the wide variety of types of sentences that occur in fictional texts, authors must presumably be thought to pretend to perform other sorts of illocutionary acts as well.

Serious (i.e., non-fictional) assertion is governed by a number of by a number of regulative and constitutive rules. One such (regulative) rule that will occupy us down the road is the sincerity rule which commits the speaker to belief in the truth of the expressed proposition. A propositional act which failed to satisfy to satisfy a regulative rule would be a defective assertion; a propositional act which failed to satisfy a constitutive rule would fail to be an assertion at all. What is important to note, however, is that the author who does not believe the propositions expressed by the sentences she produces does not thereby perform defective assertions. The explanation of this, according to Searle, is that authors of fiction (intend to) invoke conventions of fiction which suspend the aforementioned rules of assertion.[13] These conventions enable authors to use words with their ordinary literal meanings but without, for example, committing themselves to the truth of the propositions they express. An author (or anyone, for that matter) pretends to perform an assertion (or any other illocutionary act) just in case she actually performs the corresponding utterance act with the intention of invoking the aforementioned conventions.[14] It is worth noting that, while on Searle’s view, the act of pretense is performed in the very act of writing, Lewis seems to suggest that the pretense is performed by means of what is done with the product of the act of writing. He says, “he may type a manuscript and send it to his publisher, but in either case there is an act of storytelling.”[15] A related view is defended by Brown and Steinmann.[16]On this view, authors engage in illocutionary pretense, but what they pretend to do is make second-order reports of the speech acts of a fictional narrator, rather than pretending to perform the first-order speech acts themselves. Suppose, for example, a (rather dull) fictional text contained the declarative sentence “Snow is white.” According to Searle and Lewis, the author of the text pretended to assert that snow is white. According to Brown and Steinmann, in contrast, the author pretended to report that the fictional narrator asserted that snow is white.

            Both Searle and Lewis give accounts of metafictive discourse whose shared point of departure is authorial pretense, but since Lewis’s analysis takes us on a detour through narrator illocutionary action, my discussion of his view will be postponed until we arrive at the chapter on narrators and storytellers. The first thing to note, is that on Searle’s view, the critical attitude (i.e., the critic’s speech act) is neither straightforwardly assertion nor pretended assertion: “ …we need to distinguish not only between serious discourse and fictional discourse, as I have been doing, but also to distinguish both of these from serious discourse about fiction.”[17] Instead, what critics do is perform assertions and other speech acts while engaged in a “shared pretense,” an imaginative activity of a certain sort. Authors, by means of their pretended acts of referring (which in part constitute their acts of assertive pretense), create fictional characters, characters they pretend, or perhaps make-believe, exist; “…by pretending to refer she pretends there is an object she refers to.”[18] Sharing in an author’s pretense involves, in part, pretending, or again making-believe, that the fictional characters she created exist. And a critic, who, while engaging in this shared pretense, makes assertions, refers to the characters she pretends to exist: “I did not pretend to refer to a real Sherlock Holmes; I really referred to a fictional Sherlock Holmes.”[19]

            The now familiar idea that games of make-believe lie at the core of fiction[20] was at least foreshadowed by, if not fully present in, Searle’s theory. It will proof fruitful when we take up the theories of Currie and Walton and compare them Searle’s view to take them all to be based around this central notion. The differences among them are best understood in terms of how games of make-believe are thought to be generated by authorial activity and the distinct ways metafictive discourse is analyzed in terms of such games. But for now, I will focus my attention specifically on Searle’s view.

            Searle’s account of authorial activity has, in the literature, been subjected to number of sorts of criticism. One criticism has concerned the notion of pretense itself. Pavel, for example, has argued that, contra Searle, there simply is not a sharp distinction between serious action and pretense.[21] And Walton, while (apparently) conceding the cogency of this distinction, has argued that Searle’s analysis of pretense is askew.[22] As Walton points out, one can intentionally and non-deceptively play a harpsichord as if one were playing a piano without thereby pretending to play the piano. Another type of criticism has concerned whether or not authorial pretense can be used to draw fiction/ non-fiction distinction. The suggestion is that something counts as (a work of) fiction if and only if it is the product of assertive (or other illocutionary) pretense. Currie has argued, however, that since one can engage in non-deceptive pretence without thereby producing a work of fiction – for example, by imitating the conversational manner of a friend – being the product of assertive pretense is not sufficient for being fiction.[23]

            My present concern, however, is not so much with the nature of fiction as with the correct account of authorial creative activity. As a result, what is at issue is not the sufficiency of assertive pretense for fiction but with its necessity. That is, my concern is whether or not illocutionary pretense is essential to, or lies at the core of, authorial creative activity. Now Walton argues that the thesis that assertive pretense is necessary for fiction is easily refuted by the fact that creators of non-literary fictions cannot be plausibly thought to engage in such actions. As he puts it,

“Pierre-August Renoir’s painting Bathers and Jacques Lipchitz’s sculpture Guitar Player surely belong in the fiction category. But I very much doubt that in creating them Renoir and Lipchitz were pretending to make assertions.”[24]

The trouble here is that even if one were to accept Walton’s assimilation of painting and sculpture – and the representational arts, more generally – to the category of fiction, he has not made his case with respect to literary fiction. And literary fiction – fiction proper, as I would have it – is what is at issue here.

            There are two points in the creative process during which authors might be thought to engage in assertive pretense: during the process of text production itself, or by means of something that is done with the completed text. Consider, first, the process of text production. This activity is paradigmatically directed towards the production of a written text – an artifact consisting of a collection of sheets of paper partially covered with inscribed sentences or, increasingly, a number of word-processed computer files. Sometimes, I suppose, the goal is instead the production of an aural text, recorded on tape, compact disc, or, perhaps, even in memory. The question, then, is whether or not the production of these written or spoken sentences, which in part constitute texts, occurs essentially, or even characteristically, by means of assertive pretense. And the correct answer is pretty clearly the negative one. An author producing an aural text might pretend to assert a sentence if she were interested in soliciting input regarding whether it was something one of her characters, or her narrator, would say. But she would be unlikely to engage in such performances with every sentence in her text; it would certainly produce tedium if she did. Moreover, an author producing a written text would rarely, if ever, pretend to assert the sentences she inscribes. Even if it were part of the story that the narrator produced the written text, the author would be unlikely to “go through the motions” as if she were the narrator writing it. Few authors were interested in soliciting input as to whether her manner of producing written inscriptions is one that her narrator would have engaged in. In any event, authors often (usually?) work alone, and in such circumstances assertive pretense would have little point (unless the author was the “audience” of her own pretense). And finally, insofar as assertive pretense would have any point in the process of producing a text, the sentences utilized in the pretense would not thereby be included in the text, but only candidates for inclusion. The point of any such act of pretense would be to determine whether a given sentence ought to be included.

            Perhaps, instead, authors characteristically engage in assertive pretense not during the production of their texts but by means of something they do with their completed texts. There are three broad categories of things authors do with their completed texts: they keep them to themselves, they disseminate them, and they perform (i.e., publicly read) them. Now first, authors who choose not to share their texts with others, either by disseminating or performing them, can hardly be thought to have engaged in acts of assertive pretense. An author who tucks her completed text away in her bottom desk drawer has not pretended to do anything. And even if an author did intend to thereby non-deceptively go through the motions of asserting the sentences contained in her text, her attempt at assertive pretense would be an abject failure: she simply would have gone through the wrong motions.

Second, I am willing to concede that the notion of assertive pretense provides a plausible model of storyteller activity, although I will defend an alternative below. The problem is simply that performing their texts is not an activity that authors characteristically engage in. Many authors do publicly read parts of their texts, but unless their works are extremely short, they rarely read them from start to finish. It is true that many texts are performed – read aloud – from start to finish (although not normally in one sitting). But paradigmatically the storyteller in this case is someone other than the author. The kind of situation I have in mind is one in which a parent reads a book aloud to his children over the course of a number of nights. In light of such worries, Searle might retreat to his account of playwright activity: “…the author of the play is not in general pretending to make assertions; he is giving directions as to how to enact a pretense which the actors then follow.”[25] But whatever its virtues as an account of the activity of playwrights, taking the activity of authors novels to be directed towards producing instructions for performers is somewhat implausible. Not only do novels, unlike plays, (typically) lack explicit staging directions, this idea seems to suggest that the point or purpose of a novel is to be source of performances. But while the telos, if you will, of a play may be to be staged or performed, the telos of a novel is to be read (and I don’t mean aloud to an audience).

             Finally, an author might engage in assertive pretense by means of disseminating her text. What I have in mind here are actions such as giving a copy of her manuscript to a friend to read, sending it manuscript to a publisher, or perhaps even publishing and distributing it herself. Although I have grave doubts about the possibility of performing speech acts corresponding to all of the sentences contained in a text of any significant length by means of giving the text to someone, I am going to presuppose it is possible for present purposes. The model here would be that of passing a short note to someone: if I give you a note which reads “I promise I’ll meet you at Coco’s at 6 pm tonight,” I thereby promise to meet you (assuming, of course, that the various regulative and constitutive rules of promise-making are satisfied). The suggestion on the table, then, is that when an author of a work of non-fiction disseminates her work she thereby actually asserts the sentences contained therein, whereas an author of a work of fiction who disseminates her work thereby merely pretends to do so. And this suggestion is highly implausible. Consider the act of giving a fictional manuscript to a friend to read. It is simply bizarre to suppose that unless in so doing she is intentionally and non-deceptively going through the motions of giving a work of non-fiction to her friend, she is being insincere in some sense. Or consider sending a fictional manuscript to a publisher. The suggestion on the table seems to imply that an author who sends her manuscript to a publisher of fiction must thereby pretend to be sending it to a publisher of non-fiction, or, again, face the charge of insincerity. And this is just implausible. I don’t mean to suggest that an author could not engage in such pretense when disseminating her manuscript, only that doing so would be odd and have little point.  

            What remains to be done is to evaluate Searle’s account of metafictive discourse. Searle endorses the following five theses:

1. Fictional characters do not (actually) exist.[26]

2. A speaker can refer only to what exists.[27]

3. Critics can really refer to fictional characters when they share in the author’s pretense.[28]

4. Fictional characters exist in fiction.[29]

5. Authors create fictional characters by pretending to refer to them.[30]

The trouble is that the first three theses are prima facie inconsistent[31] and the last two are mysterious: it is unclear what “existence in fiction” amounts to and, as a result, it is unclear how authors, by means of their acts of pretense, can cause something to come to exist in fiction.

            Neo-Searleans, such as Seumas Miller and A. P. Martinich, have responded to the former worry by rejecting one or another of the first three theses. Miller, for example, rejects thesis (3). He argues that fictional characters are hypothetical objects – objects that would have existed had the relevant author’s pretend assertions been genuine and true. But, he claims, “…in discourse about fictional entities qua hypothetical entities these entities are not strictly speaking referred to. For in fact hypothetical entities do not exist…”[32] Martinich, in contrast, rejects thesis (2). He argues that fictional characters are intentional objects, to which critics can refer but which do not exist: “Because referring is an intentional activity, only an intentional object is needed. Sometimes the intentional object is also an existing object, and sometimes not.”[33] Miller’s view is, at bottom, quite similar to Lewis’s and, as a result, will be addressed in my discussion of Lewis below. The only comment I will make regarding Martinich’s view is that his notion of intentional objects which may or may not exist is no less obscure than Searle’s characters which do not exist but nonetheless exist in fiction.[34] 

            In my view, contra Miller, Martinich, et al., there is available a consistent and intelligible interpretation of Searle’s position that invokes the now familiar notion of make-believe. Central to many games of make-believe are props – objects which participants in such games make-believe have different properties than they in fact do (or make-believe are different kinds of things than they in fact are). So, for example, the participants of a sword-fighting game might make-believe that a stick is a sword. Moreover, it is not much of a stretch to take the participants themselves to be props in this sense. After all, the participants in a game could, for example, make-believe that one of their members is a monster (and she could, in turn, pretend to be one). Now one way to make sense of claim (5) is to suppose that by engaging in their assertive pretense, authors generate games of make-believe. After all, as long as we take pretending to refer to involve the indication some actual object or person, an author can be viewed as having created a fictional character in the following sense: she has made some object a prop in the game, which participants are to make-believe is the character. On this interpretation of Searle, the following sense can be made of theses (3) and (4): for those playing the game of make-believe the prop is the character, and since the prop exists in fact, the character exists in the game, or in fiction, if you prefer. Moreover, as long as a critic is a participant in the game – or “shares in the pretense” – she can really refer to the character by means of referring to the prop. But since in fact – that is, outside the game – the prop is not the character, the character does not exist, as thesis (1) would have it. Finally, the reason the author only pretends to refer to the character by means of indicating (what will become) the prop, rather than really doing so, is because the object she indicates does not become a prop in the game until she indicates it.

            Although this suggestion does make sense of Searle’s view, it has little to recommend it as an account of metafictive discourse. Authors simply do not generally, if ever, engage in behaviour which could plausibly be construed as giving a specific object a role as a prop in a game of make-believe. On a specific occasion, a storyteller might do so; for example, while reading a story, a parent might indicate to a child that he is to make-believe that his stuffed animal is a character in the story. But not only do storytellers fail to do this on every occasion of storytelling, it would be extremely rare for a storyteller to designate a prop corresponding to each character in the work she is performing. As a result, there just will not, in general, be enough fictional existents for even critics attending such performances to refer to in order that they might say what they want to say about the characters and events in the story.[35]


II: Fictive Illocutions


            An alternative account of authorial activity casts it as the performance of uniquely fictive illocutionary action. This fictive illocution is claimed to be distinguished from other sorts of illocutionary action in terms of the intended effects upon listeners. Whereas the goal of assertion is listener belief in the asserted proposition, the goal of fictive illocutionary action is that the listener make-believe[36] or imagine[37] the proposition expressed by her utterance. Currie, for example, gives the following (rather Gricean) analysis of fictive illocutionary action: a speaker who utters a sentence S thereby performs a fictive illocutionary action just in case (i) she intends that the audience recognize that S means some proposition P by means of their recognition that she intends S to mean that P and (ii) she intends that the audience will make-believe that P by means of the recognition that she intends them to do so.[38]

            Searle has argued against the idea of a uniquely fictive illocutionary action by pointing out that (i) “[in] general, the illocutionary act (or acts) performed in the utterance of the sentence is a function of the meaning of the sentence…”[39] and (ii) the very same sentences are used both in fictive and non-fictional discourse. As a result, if there were a uniquely fictive speech act, the sentences of our language would have to be ambiguous – differing in meaning as between their use in fiction and their use in serious discourse. Searle seems to be presupposing that the linguistic meaning of a sentence does not simply determine the proposition it expresses modulo various contextual parameters; it also determines the illocutionary action performed in its normal or, perhaps, literal use.

Currie, unsurprisingly, rejects this principle, which he refers to variously as the “determination principle”[40] and the “functionality principle.”[41] He correctly notes that a given sentence can be used to different types of illocutionary actions on different occasions of use. So, for example, “You will open the window” can be used to make an assertion or issue a command, among other things. And he argues that our ability perform different illocutionary actions with a given sentence does not stem from any ambiguity in meaning; instead it results from differences in “…what the speaker means by the sentence.”[42]  This is, of course, compatible with the claim that when, but only when, a sentence is used literally the illocutionary act performed is a function of its meaning.  After all, one might argue that when, for example, a speaker uses “You will open the window” to issue a command, she is using it non-literally. Currie goes on, however, to argue that an author will often use a declarative sentence he utters or inscribes literally “…in the sense that the proposition he intends to convey to his audience is exactly the proposition expressed by the sentence he utters.”[43] And since authors typically do not assert the declaratives they use, this gives us an example of a single sentence used literally (and with the same meaning) to perform distinct illocutionary actions on different occasions. The trouble with Currie’s manoeuvre here is that it simply presupposes that what authors do is perform illocutionary actions. But this is exactly what is at issue.

Even if the considerations Currie raises against Searle’s functionality principle do not suffice to undermine it, Hoffman offers a related argument against the existence of a uniquely fictive type of illocutionary action which is immune from these criticisms.[44] Instead of the functionality principle, Hoffman’s objection invokes what she calls the “expressibility principle”: for any illocutionary act, there is a possible linguistic expression whose literal meaning is such that its utterance in the right circumstances is the performance of that act.[45] Given that the same sentences are used in both fictive and non-fictional discourse, the expressibility principle implies that the sentences actually used in fiction do not literally express fictive illocutions (or, as above, that sentences of English are systematically ambiguous). Moreover, the expressibility principle implies that if the sentences used in fiction do not literally express fictive illocutions – and, hence, that fictive illocutionary acts are normally performed only indirectly – there must be distinct (possible) sentences which do literally express fictive illocutions.

            One potential worry is that the sentences which are likely candidates for literally expressing fictive illocutions do not express uniquely fictive illocutions. Hoffman has, for example, suggested that the kind of speech acts characterized by Currie could be literally expressed by sentences of the form “Make-believe that P!”[46] Other candidates include “Let us imagine that P!”[47] and “I invite you, whoever you are, to imagine that P.”[48] But commands, suggestions, and invitations are not uniquely fictive illocutionary actions – their occurrence is commonplace in non-fictional contexts. As a result, the candidates for expressing fictive illocutions are distinguished from other commands, suggestions, etc., in terms of their intended effects on listeners rather than in terms of their illocutionary force. The upshot, according to Hoffman, is that fiction-making is marked off from other sorts of activities in virtue of being a unique kind of perlocutionary action rather than in virtue of being a unique kind of illocutionary action.[49]

            There might, of course, be good reasons to resist both the functionality principle and the expressibility principle, or they might simply be ill-motivated. For present purposes, I am officially neutral on this question. Moreover, even if one or the other of them is true, given that our concern is with not so much with the nature of fiction as with the correct account of authorial creative activity, we still need to address whether authors make commands, suggestions, or invitations to their listeners. As a result, three distinct views of authorial activity remain on the table: authors perform uniquely fictive illocutionary actions; authors perform illocutionary actions that are not unique to fiction; and authors perform uniquely fictive perlocutionary actions. And I want to argue is that it is simply not the case that authors of fiction must, or even typically do, perform any of these sorts of actions.[50]

            As above, there are two points at which authors might be thought to perform one or another of these actions: during the writing process itself, or by means of something that is done with the completed text. First, it is implausible to suppose that authors typically perform any illocutionary or perlocutionary acts at all during the writing process for the same reasons it proved implausible to suppose they engage in acts of pretense: there would simply be little point in doing so. Moreover, even if a given author were to perform illocutionary and/or perlocutionary acts by means of her inscriptional acts, she could not be supposed to be performing the same kind of illocutionary act on each occasion. Perhaps in writing declaratives an author could (if she were so inclined) thereby command or suggest or invite a listener to imagine or make-believe the expressed proposition, but she could not be plausibly supposed to be doing the very same thing by means of her production interrogatives and other sorts of sentences. If authors perform illocutionary and/ or perlocutionary acts during the writing process, they would minimally have to perform a wide variety of different kinds of speech acts – fictional analogues of each of the acts we perform in ordinary speech.[51]

            The alternative, again, is to suppose that illocutionary acts are performed by means of something that is done with completed texts, presumably acts of disseminating them. The suggestion, then, is that by sending their texts to publishers or giving them to friends to read they thereby command or suggest or invite a listener to imagine or make-believe the propositions that constitute the text, or following Currie, perform sui generis fictive illocutionary acts. There are two separate issues that need to be addressed here: whether or not authors intend readers, who come into possession of their texts, to imagine or make-believe something; and whether or not authors attempt to get readers to behave in this way by means of suggesting or commanding or inviting them to do so. For present purposes, I am willing to concede the intended effects of dissemination on readers, although a qualification will need to be made at the end of this section.

In order to address the second issue, we will have to look at the details of the processes by which authors disseminate their texts. First, consider an author who gives her completed manuscript to a friend. Normally, this act will be accompanied by an illocutionary act on the part of the author – a request, or perhaps a command, that the friend read the manuscript. Now one might argue that reading, or at least reading as fiction, is in part constituted by engaging in imaginative activities of the requisite sort, and, as a result, that the author thereby requests that the friend engage in these activities.[52] Short of this, one would have to either concede that the intended effect lies causally downstream of the author’s illocutionary action, or claim that she performed an additional  and distinct illocutionary action, perhaps by means of the behaviour she engaged in to get the manuscript into her friend’s possession. But in order for this behaviour to count as the performance of a speech act, the author would (roughly) have to intend by so behaving to produce the requisite effect in her friend by means of the friend’s recognition of this intention.[53] And although nothing prevents an author from, for example, handing her manuscript to a friend with this intention, doing so would typically be pointless. Given the wide variety of things one might be expected to do with a manuscript one has been handed – read it, put it somewhere for safe-keeping, hand it along to someone else, etc. – the recipient is simply unlikely to recognize the author’s intention. It is the accompanying request that reveals what she has in mind.

            Moreover, even if, as above, an author who requests (or suggests or commands) that her friend read the manuscript she has been handed thereby requests (or suggests or commands) that the friend imagine or make-believe the propositions contained therein, this will not help in the general case. Consider an author who disseminates her text by means of sending it to a publisher. Normally an author will send a cover letter with her manuscript requesting that it be published or considered for publication. But unlike reading a manuscript, publishing a manuscript – or considering it for publication – cannot plausibly be thought to be constituted by imagining or making-believe the propositions contained in the text. After all, these activities do not require reading the manuscript; rather they involve distributing it to others to read, either referees or the consuming public. As a result, no case can be made to the effect that by requesting that their manuscripts be published, or considered for the same, authors thereby request that imaginative activities be performed.

            One final comment. I conceded above that, although not performing the corresponding illocutionary acts, authors do intend that readers imagine or make-believe the propositions expressed by the sentences that constitute their texts. This concession needs to be qualified, however. In my view, what authors paradigmatically intend is that their texts be read as fiction. Now it might be true that what it is to read a text as fiction is to engage in specific sorts of acts of imagination or make-believe. And if this is true, there will be, of course, a sense in which authors paradigmatically intend that their readers engage in such acts. It is the same sense in which, for example, Lois Lane can be said to believe that Clark Kent can fly simply in virtue of believing that Superman can. But if we are concerned to individuate intentions in terms of how subjects conceive of their intended goals, we cannot attribute this intention to authors on this basis.   


III: Propositional Acts


            The last view I wish to consider in this chapter is one according to which authorial activity consists in the performance of propositional acts.  More to the point, on the view at issue authors perform propositional acts without illocutionary force. In other words, when they produce sentence tokens authors thereby express propositions, but they do not assert (or command, or request, etc.) these propositions. Although there are a number of theories I am inclined to characterize in this way,[54] this categorization is altogether satisfactory for a couple of reasons. First, it is not entirely clear that all of the theories in question are sensitive to the distinction between the performance of propositional acts and the performance of utterance/ inscriptional acts. Gale, for example, says the following:

“To refer in the locutionary sense is to utter a referring expression which has a certain sense or meaning, while to refer in the illocutionary sense is to refer in the locutionary sense with the intention of referring to some existent individual(s).”[55]

This seems to suggest that, on Gale’s view, the sentences produced by authors have linguistic meaning but lack propositional content.[56] But while it is uncontroversial that authors perform utterance/ inscriptional acts, it is the stronger claim that I wish to contest.

            A second difficulty with this categorization stems from the fact that authorial activity consists of more than merely the performance of propositional acts lacking illocutionary force. As Walton puts it, “[fiction] is not just language stripped of some of its normal functions; it is something positive, something special.”[57] In light of such worries, the views I have in mind typically combine the shared denial of illocutionary force with a variety of different positive accounts of what authors do over and above the performance of propositional acts. Ohman, for example, argues that what authors do is “purportedly imitate” a series of speech acts.[58] Gale defends the related view that they engage in assertive (and other illocutionary) pretense.[59] Beardsley argues that authors “represent” the fictional acts of fictional narrators.[60] And, finally, Eaton claims that authors perform “translocutionary” acts whereby they attribute illocutions to fictional narrators, or as Eaton would have it, dramatic speakers.[61] Despite these differences, the views to be considered here share one important feature: the actions they attribute to authors are (or can be) performed in the production of each individual sentence that constitutes a fictional text.

            The first thing I wish to establish is that authors cannot, in general, be taken to perform propositional acts. Insofar as we wish to apply the resources of speech act theory to authorial activity, we can only view authors as performing utterance or inscriptional acts. That is, authors can be viewed only as producing a series of (meaningful in the sense of character) sentences which lack propositional content. The reason for this is that the proposition expressed by means of a propositional act is a function of the linguistic meaning of the uttered sentence and features of the context in which it is uttered. And so, if authors perform propositional acts, the propositions they thereby express will be determined by features of the contexts in which they utter or inscribe the sentences which constitute their texts or, perhaps, features of the contexts in which they disseminate the their texts. But propositions so determined are, in general, irrelevant to a correct understanding of fictional texts and, hence, cannot be thought to serve as the propositional contents of the sentences contained therein. Consider, by way of illustration, a novel set in the future that contains the sentence, “When they arrived, they discovered the colony was deserted.”  If the author performed a propositional act when she produced this sentence, the proposition she expressed would make reference to some time prior to its utterance, not to some future time. Or consider any (unquoted) sentence in the first person that occurs in one of the Sherlock Holmes stories. If Doyle performed a propositional act when producing such a sentence, the subject of the proposition he expressed would have been would have been Doyle himself, rather than Dr. Watson. The lesson here is that the relevant context for interpreting a work of fiction is that in which the narrative voice is understood to occur rather than the context(s) in which an author produced it. And although these contexts could in principle coincide, such cases are quite rare.

            What remains to be done is to evaluate the various accounts of what authors do over and above the performance of propositional acts. Gale and Ohmann are pretty clearly endorsing versions of the pretense theory, and as a result, their views are inadequate for the reasons discussed above. Beardsley’s notion of representation is harder to pin down. At times its sounds like pretense:

“Waiting on a table is a complex action that involves various elements, including the actual handling of food. But the gifted mime can represent the action with the most minimal gestures – without food, or a dinner to serve, or even a table or tray.”[62] (my emphasis)

At other times, however, it sounds more like Eaton’s notion of translocution:

“When Tolstoy writes that Vronsky put up the roof of the carriage, he is producing text that could be used in describing an event. But since the name “Vronsky” does not refer to any actual person, no illocutionary act of describing occurs; it is only represented – as the fictional act of a fictional narrator.”[63] (my emphasis)

Insofar as representation is pretense, it has already been discussed. The notion of translocution requires some comment however. There is, in my view, something right headed about the idea that works of fiction should read or interpreted as a series of illocutionary actions performed by a fictional narrator. Moreover, if doing so is part of what it is to read a work as fiction, I would be happy to concede that that they be so read is a characteristic (de re if not intensional) object of authors’ intentions, at least of those who disseminate their texts. Where I balk is at any suggestion that this must be traced back to corresponding sentence by sentence intentions on the part of authors. One might, of course, recommend that an author writing a text with an explicit narrator think of each sentence she writes as an illocutionary action of her narrator. After all, doing so might improve the chances of producing a text which reflects the intended personality of the narrator. But an author need not heed such advice. Moreover, the value of the advice becomes negligible as we move from texts written in the first person whose explicit narrators have fully developed personalities to texts written in the third person whose implicit narrators “fade into the background.”


IV: Conclusion – Utterance Acts


            The upshot of this discussion is that insofar as authors are thought to be performing speech acts, they can only be thought of as performing utterance or inscriptional acts – acts of producing meaningful sentences tokens, but tokens lacking propositional content. Of course, this is not by itself sufficient for fiction-making, but the distinguishing mark of fiction is not to be found in authors’ communicative purposes. Fiction-making is the process of creating an artifact of a certain kind – a novel, or a novella, or a short story. The speech act approach to authorial activity erroneously models it on the activity of speakers engaged in communication rather than that of artists producing art objects. The mistake is a natural one given that art objects in question consist of sentences, which are primarily used for communication. Nonetheless, the relation of author to text should not be thought of as the relation of speaker to utterance, but rather as word-sculptor to word sculpture.


[1] See, e.g., Beardsley, 1981, Currie, 1990, Eaton, 1969, Gale, 1971, Lewis, 1978, Ohmann, 1971, and Searle, 1975. It is worth noting that many of these accounts are focused primarily on drawing the fiction/ non-fiction distinction rather than on providing an account of the semantics of fictional discourse.

[2] See Searle, 1969.

[3] See Grice (1957) for a more elaborate account of the intentions involve in illocutionary action.

[4] But see Martinich, 2001. The view I defend in my positive section below, however, does head in this direction.

[5] Searle, 1975, and Lewis, 1978.

[6]See, e.g.,  Currie, 1990, Niniluoto, 1985, and Grant, 2001.

[7] Hoffman, manuscript.

[8] See, e.g., Beardsley, 1981, Ohmann, 1971, and Gale, 1971.

[9] This line of criticism will be developed more fully in chapter three below.

[10] This is most explicit in Lewis (1987) p. 39.

[11] Searle (1975) p. 325.

[12] Lewis (1978) p. 40. Note: Lewis (1978, p. 39) explicitly equivocates between the activities of authors and the activities of storytellers.

[13] Searle (1975) p. 326.

[14] Searle (1975) p. 327.  It is not clear whether or not on Searle’s view one’s intention to invoke the conventions of fiction can misfire, resulting in infelicitous assertive pretense.

[15] Lewis (1978) p. 39 (my emphasis). Note: presumably Lewis would have to require that the author send the manuscript to his publisher with the intention of invoking the conventions of fiction.

[16] Brown and Steinmann, 1978.

[17] Searle (1975) p. 329.

[18] Searle (1975) p. 330.

[19] Searle (1975) p. 330.

[20] See, e.g., Currie, 1990, and Walton, 1990.

[21] Pavel (1981) pp. 171-2.

[22] Walton (1990) pp. 81-2.

[23] Cruuie (1990) pp. 17-8.

[24] Walton (1990) p. 82.

[25] Searle (1975) p. 328.

[26] Searle (1975) p. 329.

[27] Searle (1975) p. 330.

[28] Searle (1975) p. 330.

[29] Searle (1975) p. 329.

[30] Searle (1975) pp. 329-330.

[31] Margolis, 1983.

[32] Miller (1992) p. 3.

[33] Martinich (2001) p. 107. It is worth noting that, strictly speaking, Martinich does not follow Searle and defend the view that authors engage in assertive pretense. Instead they perform genuine assertions (and  other illocutionary acts) while suspending the Supermaxim of Quality (which requires that one not perform a speech act unless all conditions for its non-defective performance are satisfied).    

[34] One could, I suppose, give a straightforward Fregean interpretation of Martinich’s view according to which intentional objects are Fregean senses, only some of which have referents. But taking intentional objects, so construed, to be propositional contribution of names would presumably run afoul of the many of the difficulties Kripke (1980) raised against descriptivism.

[35] One might also worry that only critics attending storyteller performances – relatively few critics indeed – would be in a position to refer to the fictional existents storytellers create. Such a worry might be alleviated, however, by adapting Kripke’s (1980) causal-historical theory of referring to allow pretended reference to play a role analogous to that of initial baptisms. I expect, however, that Searle would disapprove of some any such manoeuvre. See Searle (1983) pp. 231-261.

[36] Currie, 1990.

[37] Niniluoto, 1985 and Grant, 2001.

[38] Currie (1990) p. 31. Currie goes on to present a number of revisions to this account in order to make it more general.

[39] Searle (1975) p. 324.

[40] Currie (1985) p. 385.

[41] Currie (1990) p. 14

[42] Currie (1990) p. 15.

[43] Currie (1990) p. 15.

[44] Hoffman, manuscript.

[45] Hoffman (manuscript), p. 6. It is worth noting that Hoffman claims that the expressibility principle is what really lies behind Searle’s objection to Currie’s position. Although this may be true, the functionality principle is certainly more commensurate with what Searle actually says.

[46] Hoffman (manuscript), p. 8.

[47] Niniluoto (1985), p. 220.

[48] Grant  (2001), p. 400.

[49] Hoffman (manuscript), p. 10.

[50] Walton (1990, p. 87) argues against speech act accounts of fiction by arguing that “…the basic concept of a story and the basic concept of fiction attach most perspicuously to objects rather than actions.” My concern in this section, however, is not with the correct account of fiction but with the correct account of authorial activity.

[51] One advantage of the pretense view is that to address such worries they need only claim that authors can pretend to perform each of the illocutionary acts we perform in ordinary speech.

[52] It is, of course, far from clear that a speaker who, for example, requests that a listener perform an action thereby requests that the listener perform all or any constituent actions. 

[53] Grice, 1957.

[54] E.g., Beardsley, 1981, Ohmann, 1971, Gale, 1971, and, perhaps, Eaton 1969, 1979.

[55] Gale (1971), p. 328.

[56] The interpretation of Gale on this point is tricky because he seems to endorse a Fregean account of propositional content according to which sentences containing empty names can express propositions. 

[57] Walton (1990), p. 78.

[58] Ohmann (1971), p. 14.

[59] Gale (1971), p. 336. 

[60] Beardsley (1981), p. 301.

[61] Eaton (1969), p. 167.

[62] Beardsley (1981), p. 294.

[63] Beardsley (1981), p. 301.