Kania[1] has recently developed an argument which poses a serious challenge to the “ubiquity thesis” – the view that every literary narrative[2] necessarily has a fictional narrator.[3] Kania characterizes a fictional narrator as a (possibly non-human) agent who tells (or is responsible for) the narrative and who exists on “the same ontological level” as the characters and events described by the narrative.[4] Kania’s central claim is that there are numerous non-peripheral examples of literary narratives – in particular, those written in a “plain third-person omniscient narrative style” – for which there is no reason to believe they have fictional narrators.[5] Kania defends this claim by arguing that the central reason for thinking such works have fictional narrators – the ontological gap argument – does not establish any such thing.


As Kania reconstructs it, the ontological gap argument has a positive and a negative stage. The positive stage consists of the claims (i) that understanding a literary narrative requires an account of how we (the readers) acquire information about the fictional world it describes and (ii) that the only adequate account is one according to which this information is provided to us by an agent of some kind. And the negative stage consists of the argument that this agent must be a fictional narrator because otherwise she would not be related cognitively to the characters and events in the story so as to be able to coherently provide us with information about them “as if they were real.”[6] 


Kania criticizes both stages of this argument. He criticizes the positive stage by pointing out that if understanding a literary narrative requires an explanation of how we acquire information about the fictional world it describes, it presumably also requires an explanation of how an agent who provides us with it goes about acquiring this information and transmitting it to us. But in many works, no such explanation is forthcoming: the process by which the information is acquired and transmitted by the narrator is indeterminate or even incoherent.[7]  And he criticizes the negative stage by (i) emphasizing the continuity between literary narratives written in a third-person omniscient style and other fictional narratives – such as jokes and improvised bedtime stories – which lack fictional narrators and (ii) arguing that if there is a genuine epistemic gap between fictional worlds and the actual world then even if fictional narrators could acquire information about the worlds they inhabit, they could not transmit this information to readers (inhabitants of the actual world). Kania concludes with the suggestion that fictional narrators should be posited on a case by case basis when (and only when) so doing yields a better understanding of the work under consideration.


Although Kania’s conclusion may strictly speaking be correct, his argument for it runs into serious difficulties. Not only does his taxonomy of tellers of fiction blur a number of important distinctions, he equivocates between two versions of the ontological gap argument, one of which is immune to the criticisms he raises. Kania explicitly acknowledges only two categories of tellers – actual tellers and fictional tellers:

“[we]can thus distinguish between the actual telling of the fictional story, which John Barth engages in, and the fictional telling of the story engaged in by Horner.”[8]

There is, however, a pair of important (and orthogonal) distinctions that can be drawn among tellers – one tripartite and one bipartite – which together yield six categories rather than Kania’s two. First, following Kania, one can distinguish ontologically between tellers that are actual and those that are non-actual; but, departing from Kania, one can sub-divide the latter category into fictional and non-fictional (but still non-actual) tellers. By an actual teller, I mean a teller who performs her speech acts in the actual world; an actual teller is a real person whose act of telling really occurs. A fictional teller, in contrast, is a teller whose speech acts occur in a fictional world, that is, the world generated by a literary narrative wherein the events described in the narrative unfold; fictional characters who engage in dialogue, for example, count as fictional tellers.[9] A non-actual/ non-fictional teller is one whose act of telling occurs neither in the actual world nor in a fictional world, but rather in some alternative distinct from both of them. Figures who hypothetically or counterfactually perform acts of telling to that extent count as non-actual/ non-fictional tellers. Second, one can distinguish linguistically between narrators who engage in factual discourse – or “fact-telling” – and those who engage in fictional discourse – or “fiction-telling.” By fact-telling I mean what Searle calls “serious” discourse wherein speakers perform familiar sorts of illocutionary acts, such as assertion, and are held responsible for complying with the various semantic and pragmatic rules governing them.[10] Fiction-telling, in contrast, is what authors do when they tell their fictional stories. For present purposes, I will remain neutral regarding whether such tellers are best understood to engage in illocutionary pretense[11] – that is, pretending to perform familiar illocutionary acts – or to genuinely perform sui generis fictional illocutionary acts with their own semantic and pragmatic rules.[12] 


Of the six categories of tellers generated by this pair of distinctions, three – actual fact-tellers, actual fiction-tellers, and fictional fact-tellers – are quite familiar. Actual fact-tellers are simply real people who actually perform ordinary sorts of speech acts; authors of non-fiction, among numerous others, are included in this category. Actual fiction-tellers are real people who engage in fiction-telling, and include, most prominently, authors of fiction.[13] Fictional fact-tellers are inhabitants of fictional worlds who engage in serious discourse; this typically includes several of the characters that appear in a given literary narrative including the explicit narrator, if there is one. Less familiar are fictional and non-actual/ non-fictional fiction-tellers. Fictional fiction-tellers are simply inhabitants of fictional worlds who engage in fiction-telling, such as characters who are themselves authors of fiction. A potentially problematic example is a story-telling narrator – an explicit narrator who clearly indicates that the story being presented is fictional.[14] Non-actual/ non-fictional fiction-tellers are tellers of fictional stories whose speech acts occur neither in the actual world nor in fictional worlds. The primary examples of such tellers are idealized author figures of the sort invoked for various theoretical purposes by hypothetical intentionalists.[15] For present purposes, however, the important category is the least familiar one – that of the non-actual/ non-fictional fact-teller. Such tellers, like fictional narrators, make assertions describing the goings on at fictional worlds and perform other familiar sorts of illocutionary acts; but unlike fictional narrators, they do inhabit those worlds, nor do they inhabit the actual world. In contrast to the other categories, there are no well-known examples of non-actual/ non-fictional fact-tellers – part of the goal of this paper is to establish the theoretical need for such figures. Nevertheless given that it is generated by the pair of distinctions developed above, it remains a possible category into which a teller might fall. And even if the ubiquity thesis is false, it still could be true that every literary narrative must have a non-actual fact-telling narrator, that is, either a fictional narrator or a non-actual/ non-fictional fact-teller.


The reason that this is important is because Kania equivocates between two distinct ontological gap arguments – an epistemic argument and a doxastic argument. And while the epistemic argument putatively establishes that a literary narrative must have a fictional narrator, the doxastic argument, if successful, establishes only that it must have a non-actual fact-teller. Moreover, the arguments he marshals undermine only the epistemic variant of the argument and leave the doxastic variant untouched. Kania is addressing the epistemic argument when he says,

“…the artist cannot present the fictional world to us because he or she does not stand in the right relation to it … [this] problem is solved by positing an agent at the fictional level, who thereby does stand in the right epistemic relation to the fictional world.”[16]

This argument can be fairly reconstructed as follows:

1E) Understanding a literary narrative requires the supposition that it has a teller who is so cognitively related to the fictional world as to be able to provide readers with the information about it she in fact provides

2E) The only kind of teller who is appropriately cognitively related to a fictional world is a fictional narrator

CE) Understanding a literary narrative requires the supposition that it has a fictional narrator.

And as should be clear, Kania’s positive stage criticisms bring (1E) under suspicion and his negative stage criticisms undercut (2E).   


In contrast, Kania is addressing the doxastic argument when he says, 

“… though at first glance it might seem that actual authors narrate their novels, this is incoherent, because they cannot make the assertions about the fictional world that are clearly being made … Graham Greene is clearly not asserting this; he of all people should know that neither Wilson nor the Bedford Hotel exist … [anyone] prepared to make the assertion about Wilson and the Bedford must be at the fictional level and Graham Greene is simply not.”[17] 

This argument can be reconstructed as follows:

1D) Understanding a literary narrative requires the supposition that it is told by a fact-teller

2D) There are no actual fact-tellers of literary narratives

CD) Understanding a literary narrative requires the supposition that it is told by a non-actual fact-teller

Given his taxonomy of tellers, Kania is forced to conclude that if a teller is non-actual then it must be fictional. According to the more fine-grained taxonomy developed above, a non-actual teller can be either fictional or non-actual/non-fictional. One might worry about the coherence of a non-actual/ non-fictional fact-teller having beliefs and making assertions about a fictional world it does not inhabit. But it is worth noting that having beliefs and making assertions about worlds one does not inhabit is commonplace; consider, for example, assertions about Heaven and the afterlife.[18]


As should be obvious, Kania’s criticisms of the ontological gap argument do not undermine its doxastic variant. As long as the motivation for (1D) is not that a fact-teller is required to explain how we acquire information about a fictional world, Kania’s positive-stage criticisms do not get any purchase here. Moreover, the motivation for locating fact-telling narrators of literary narratives outside the actual world is simply that the actual authors of such narratives tell them as fiction rather than as fact, rather than to ensure that these narrators stand in the right kind of cognitive relations to fictional worlds. As a result, Kania’s negative-stage criticisms are impotent here as well. After all, if it is not assumed that fact-telling narrators need to occupy the worlds they describe in order to have beliefs or make assertions about them, then there is no reason to assume that such narrators need to occupy the actual world to communicate their beliefs to readers.[19]   


Even if the doxastic argument is immune to Kania’s criticisms, the question remains whether there is any positive reason to endorse it. More to the point, although 2D seems incontestable, as it stands 1D is sorely lacking in motivation. There are, however, compelling grounds for accepting this premise. Understanding a literary narrative requires determining what fictional truths it generates. But the fictional truths generated by a narrative extend beyond the propositions expressed by the sentences which constitute it.[20] Now the obvious suggestion is to suppose that unexpressed fictional truths are generated by pragmatic implicature from the utterances of (one of) a narrative’s teller(s). But the conversational principles by means of which implicatures are generated presuppose that speakers are engaged in fact-telling rather than fiction-telling discourse. Consider, for example Grice’s maxim of quality: do not say what you believe to be false.[21] This maxim does not govern fiction-telling discourse because there is simply no expectation that fiction-tellers believe what they say; thus fictional-tellers simply cannot pragmatically impart anything by means of it. Only a fact-teller – who is obligated to say only what she believes – could generate implicatures by such means.[22] As a result, if unexpressed fictional truths are generated by pragmatic implicature – and what other option is there? – then they can be discerned only by supposing (or imagining) that the text is told by a fact-teller.[23]   


The conclusion of this piece – that every literary narrative necessarily has a non-actual fact-telling narrator – is one in which both Kania and his opponents can take some solace. It is compatible with the contention of Kania’s opponents that every literary narrative has a teller distinct from the actual author. And since this figure is a fact-teller rather than a fiction-teller, this conclusion cannot be identified with hypothetical intentionalism, a view that even Kania finds tolerable.[24] And it is compatible with Kania’s contention that whether or not a given literary narrative has a fictional narrator is an interpretive question: a non-actual fact-teller should be located in the world of the fiction only if doing so yields a better understanding of the narrative. The oversimplified taxonomy of tellers that both Kania and his opponents presuppose makes their dispute appear much less tractable than it actually is.


Peter Alward

Department of Philosophy

University of Lethbridge




Alward, Peter (2006), “Leave Me Out of It: De Re but not De Se Imaginative Engagement with Fiction,” Journal of Aesthetics and Art Criticism 64: 451-460.


---- (manuscript), “Truth in Fiction: The Story Concluded.”


Byrne, Alex (1993), “Truth in Fiction: The Story Continued,” Australasian Journal of Philosophy 71: 24-35.


Currie, Gregory (1995), “Unreliability Refigured: Narrative in Literature and Film,” Journal of Aesthetics and Art Criticism 53: 19-29.


---- (1990), The Nature of Fiction, Cambridge: Cambridge University Press.


Grice, Paul (1975), “Logic and Conversation,” in Syntax and Semantics, vol. 3: Speech Acts, P. Cole and J.L. Morgan (eds),  New York: Academic Press.


Kania, Andrew (2005), “Against the Ubiquity of Fictional Narrators,” Journal of Aesthetics and Art Criticism 63: 47-54.


Levinson, Jerrold (1996), “Intention in Interpretation and Literature,” in The Pleasures of Aesthetics: Philosophical Essays, Ithaca, NY: Cornell University Press, pp. 175-213


Searle, John (1975), “The Logical Status of Fictional Discourse,” New literary History 6: 319-332.


Stecker, Robert (1987), “Apparent, Implied, and Postulated Authors,” Philosophy and Literature 11: 258-271.




[1] Kania, 2005.

[2] Note: “literary narratives,” as Kania uses the expression, are a sub-class of fictional narratives.

[3] The necessity operator should be understood to have wide scope. That is, it is properly represented as,

ð("x)(if x is a literary narrative then x has a fictional narrator),

rather than as,

("x)(if x is a literary narrative then ð(x has a fictional narrator)).

[4] Kania (2005), p. 47.

[5] Kania (2005), p. 50.

[6] Kania (2005), p. 48.

[7] Kania (2005), p. 49.

[8] Kania (2005), p. 49. Kania does hint at a third category – non-actual but non-fictional tellers – in his discussion of hypothetical intentionalism (pp. 52-3).

[9] Given that real persons can occur as characters in literary narratives, a real person can be a fictional teller. What is essential, however, is that her speech act occur in the fictional world and not the actual world.

[10] Searle (1975), p. 320 ff.

[11] Searle, 1975.

[12] Currie, 1990.

[13] For present purposes, I will ignore the distinction between the activities of composition and performance (e.g., public reading) that authors might engage in. In my view, only the latter essentially involves the performance of illocutionary action (or illocutionary pretense).

[14] See Kania (2005), p. 52.

[15] See, e.g., Levinson, 1996. This category is the primary stalking horse of Stecker, 1987.

[16] Kania (2005), pp. 50-1.

[17] Kania (2005), p. 50.

[18] One might also worry about how non-actual/ non-fictional fact-tellers could acquire beliefs about fictional worlds, or whether their beliefs could be justified. But as Kania (2005, p. 49) points outs, similar worries sometimes arise even when a work has an explicit fictional narrator, and there may simply be no determinate answer.

[19] Of course, I have not addressed Kania’s continuity argument against the negative stage. I do not find this argument to be particularly persuasive, however, because I think that some of the examples he appeals to – improvised bedtime stories, e.g. – do have non-actual fact-telling narrators. I do acknowledge that this does not seem plausible in the case of jokes, but such discourse may require special treatment. 

[20] In cases of narrative unreliability, not all of the propositions expressed by these sentences are included among the fictional truths. For a discussion of how the approach sketched here handles unreliability, see Alward (2006) and Alward (manuscript). 

[21] Grice, 1975.

[22] One might, of course, argue that there are a separate set of conversational principles governing fiction-telling-telling. But without independent evidence for such principles, this manoeuvre is ad hoc.

[23] One might worry that this conclusion cannot be reconciled with the existence of fictional works with story-telling narrators, who clearly indicate that the story being presented is fictional. Such works are best treated as telling stories within stories. In the “outer story” there is a fictional fiction-teller – the storytelling narrator – while in the “inner story” there is a fictional fact-teller.

[24] Kania (2005), pp. 52-3.