Manuel García-Carpintero got his PhD at the University of Barcelona in 1988, and has taught there since then. He works in the philosophy of language and mind, and related epistemological and metaphysical issues. Currently he is completing a book under contract with OUP on the nature of speech acts in general and assertion in particular, entitled Tell Me What You Know.
This week at The Junkyard, we're hosting a symposium on Greg Currie’s recent book: Imagining and Knowing: The Shape of Fiction. See here for an introduction from Greg. Commentaries and replies appear Tuesday through Thursday.
* * *
Commentary from Manuel García-Carpintero
Greg Currie’s Imagining and Knowing makes a compelling case for a mild skepticism about learning from fiction. Far from rejecting that fictions offer knowledge (which adds to their artistic merit), the book offers a partial defense. It argues that better support is needed than personal impressions, highlighting the significance of empirical research. The book is packed with original ideas and sharp arguments; I enjoyed every page.
I have disagreed before with the views on fictionality on which Currie relies (García-Carpintero 2013, 2019a, 2019b). What he needs here is rather ecumenical; but some earlier worries emerge. I’ll question the bottom-up (BU) approach, on which “[t]he bottom […] is the single utterance (of very variable length) intended to convey the intention that the audience imagine what the utterance expresses; I call this a fictive utterance” (30). Echoing his (1990, 49) earlier skepticism about the contrasting top-down (TD) approach, he declares to have “no ambition to move up from there to a definition of what it takes for something to be a work of fiction […] the conditions under which it is reasonable to call something a ‘work of fiction’ are too variable and context-dependent to allow of any such generalization” (30). I’ll say how I understand the BU-TD debate, and I’ll argue for TD on the basis that the fictionality of minimal utterances depends on their contribution to the full utterance constituting the work.
What is meant by part and whole when talking of fictive utterances? Utterances deploy the text by means of which a work is conveyed (Currie 1991). Currie (2014, fn. 12) proposes as a useful “simplification” that a fictive utterance is the utterance of a single sentence. He is clear however that he means to encompass filmic or theatrical fictions (10-11, 152 fn). What plays the “bottom” role for such “texts”? Let’s assume it is a shot, or a single thespian event. Currie’s definition of the “bottom” above, as his admission that fictive utterances are of “variable lengths”, then allows for the full “text” of a fictional work to be a fictive utterance. We may think that way of films that are, or pretend to be, a single continuous sequence shot, like Hitchcock’s 1948 The Rope or Sokurov’s 2002 Russian Ark. In such cases at least, BU and TD collapse, and Currie’s account should cover works.
In his highly contentious, but also highly influential, work on documentaries, Currie suggests that we think of the documentary category as a real kind, like water on the Kripke-Putnam view:[i] “Style and subject may be indicators of documentary status, as clarity and potability are indicators of water, but they are not defining features. Just as there can be fool's water, so there can be a fool's documentary. Very few things, it will turn out, are ideal documentaries, just as very few things in this world are pure samples of water” (Currie 1999, 285). This applies to the fiction/nonfiction divide in general, I’ll assume. Currie (2014) investigates whether the status of a work as fiction or otherwise supervenes on that of its minimal parts: a change in the former requires one in the latter. But supervenience is too weak a relation to define the debate; both BU and TD may be compatible with it. I’ll characterize it instead in terms of ontological priority (Rosen 2015), which I think fits Currie’s avowal quoted above: BU assigns explanatory priority to the classification of minimal utterances by themselves. TD rejects this, claiming at least equal explanatory significance for the utterance of the full text of which they are parts.
What is Currie’s reason for BU? The quoted avowal suggests prototypicality: the smaller the texts, the greater the chances that they are pure instances of fictionality. But this doesn’t explain Currie’s endorsement of BU. For the point applies as much to the non-fictionality that he takes to be specific to documentaries; but he rejects BU there. His view ascribes a crucial role to what he calls traces – belief-independent natural indicators of their contents; but any shot in a fiction film might play the role of a trace in a documentary – as shots of Kubrick’s movies in Ascher’s 2012 documentary Room 237 illustrate.[ii] This is a hermeneutic circle (1999, 292), on account of which Currie simultaneously defines documentary work and its filmic parts: a documentary predominantly consists of filmic parts that are traces of contents that contribute as such to the narrative it asserts (1999, 293). TD thus applies to documentaries on Currie’s account, belaying prototypicality: to establish whether a trace contributes as such to the narrative that the work asserts, we need to work out the latter. These aren’t just epistemic affairs of interpretation; they have metasemantic, ontological implications. On Currie’s Gricean view, authors must follow that path to fix their intent. On normative views, the determination of the beliefs prescribed by the work would incur the cycle.
Similar considerations apply to fictions; and we don’t need to depart from Currie’s work to find again support for them.[iii] In advancing his hypothetical intentionalist take on the interpretation of fictional works, he (1990, 76) correctly points out that this is a holistic affair. We will be misled if we just focus piecemeal on what looks like a minimal fictive utterance by the temporary or absolute unreliability or the narrator. Thus, the readership of Carroll’s Alice in Wonderland and the audience of Lynch’s 2001 Mulholland Drive are intended to temporarily imagine, respectively, that there are talking rabbits, and that Betty is a great actress, by utterances in those works; eventually they should get to know better. Those are “fictive utterances” only in an ancillary sense (García-Carpintero 2019a). Their contents don’t constitute the fictional world that one should imagine; rather, that those propositions are dreamt does. This is not established by any isolated part of the work, but by the whole of it.[iv]
Some nice points that Currie makes in the book equally support TD. He offers a Gricean, implicature account of how we come to learn from realist fictions location-setting contents that we are primarily meant to imagine. They “are established by the author in the first instance merely as true-in-the-fiction, and … there are processes of plausible inference available to suitably prepared readers which enable them to infer from ‘P is true in- the-fiction’ to ‘P is true’, for some propositions P and some fictions” (158).[v] These inferences would in some cases be thwarted; for instance, incautious spectators of Almodóvar’s 1999 All About My Mother may come to wrongly believe that the nice art nouveau flat where Sister Rosa’s parents live is in Duc de Medinaceli square in Barcelona, while it is in fact a long way from there. What disqualifies the inference in such cases? Currie appeals to this principle: “Deviations in fiction from truth are expected to be justified by their narrative payoff” (160). Although the Almodóvar example shows that other matters might also contribute to tolerating such deviations, this does explain many cases. Currie uses this point to support the implicature account: “Unless one is imaginatively engaged by the story one simply is not in a position to know whether there is any such narrative benefit accruing” (165). But the relevant narrative values, once more, cannot be adjudicated at the level of minimal fictive utterances, but ultimately at that of the full fictive utterance constituting the fictional work.
These considerations discredit BU for fictionality. They show that, given any putative minimal fictive utterance (sentence, image, event) taken with the content it has in isolation, its character as such might have to be rejected on the grounds that it doesn’t properly contribute to the fictionality of a bigger utterance of which it is part, ultimately the full fictive utterance constituting the relevant work. Our Currie-based considerations show that the fictive character or otherwise of minimal utterances can only be (metaphysically!) established relative to their contribution to that of the full utterance constituting the fictional work.
We have thus seen that prototypicality doesn’t support BU. We are thus left with the question, what is Currie’s reason for BU? The only one I find is an “explanatory” reason, not a “normative” one: to wit, that he is still attached to a necessary condition that he (1990) imposes on fictionality, that fictive content must be at most accidentally true.[vi] Here he officially assumes a weaker condition, that “to be worth the name, a fiction must contain some content which is not presented as true” (15), or just that including made-up content is a Waltonian standard feature of fictional works – one that defeasibly supports classification in the category (Currie 2014, 357).[vii] But the earlier view (21, fn) creeps in at places. The discussion in §1.3 suggests to me that the not-made-up material in Hugo’s Notre Dame de Paris, or “location setting” material in realist works like Gaskell’s Mary Barton, are not considered fictive utterances; cp. for instance Currie’s remark that what matters to classify these works is “that we identify the material that is fictive, the material that isn’t, and the relations between them” (25, my emphasis).
Not everything we find in a fiction artwork is part of the “fictive utterance” constituting it. García-Carpintero (2019b, 272) mentions for illustration authorial intrusions, as in Brontë’s Jane Eyre. Konrad (2017, 42 ff) mentions bits of Döblin’s Berlin Alexanderplatz, which she takes to be “irrelevant with regard to the events taking place in the fictive world”. But this doesn’t apply to the cases Currie discusses. The passages in Hugo’s and Gaskell’s works appear to be fictive utterances, “intended to convey the intention that the audience imagine what the utterance expresses”. They are relevant “to the events taking place in the fictive world”, to use Konrad’s wording; for they are to be inferentially combined with made-up content to fill up details of the fictional world. In fact, as we have seen Currie agrees with this, so I may have misunderstood him. As indicated above, he offers an implicature account of how we may come to know the truth of such contents in the actual world, on which they are first meant to be imagined; this seems to apply to the Hugo and Gaskell’s cases.
Be this as it may, the suggested explanatory reason cannot be a normative one: even the smallest utterances make trouble for the “made-up content” view of fictionality, as these cases illustrate. To conclude then, I fail to see any reason why Currie should still favor BU. The normative perspective I endorse certainly requires us to adopt TD; but, as Stock (2017) illustrates, TD is not incompatible even with arch-intentionalist views.
* * *
Financial support for my work was provided by the DGI, Spanish Government, research projects FFI2016-80588-R and FFI2016-81858-REDC, the award ICREA Academia for excellence in research, 2018, funded by the Generalitat de Catalunya. Thanks to Josep Corbí and Enrico Terrone for comments on previous versions, and to Michael Maudsley for his grammatical revision.
[i] Social kinds like speech acts and money may also be “real” in this sense.
[ii] Interestingly, shots of Eyes Wide Shut play the role of fictional reenactments in Room 237; they are fictive utterances there too, albeit not of the contents whose imaginings they prescribe in the original film. Shots of The Shining do play the role of traces in Room 237; they of course don’t play there either the same role as in the original fiction.
[iii] The paper just discussed has an almost explicit assertion of TD: “in a documentary meaning passes from image to narrative, while in nondocumentary meaning goes the other way” (1999, 296, my emphasis). This is qualified as “loosely put” (1999, 296), but it is also more precisely stated (ibid., 290). Note that the possibility of a fictive use of any shot in a documentary, illustrated in the previous footnote, also shows that the same applies there.
[iv] Cp. Stock’s (2017, 66) characterization of the content of fictive utterances, motivated by similar considerations.
[v] García-Carpintero (2016, 2019c) develops a similar account.
[vi] My points are very close to those that Stacie Friend has been raising over the years on the “patchwork problem”, to which I am very much indebted.
[vii] I have argued that only the latter is true (García-Carpintero 2016).
References
Currie, Gregory (1990): The Nature of Fiction, Cambridge: Cambridge University Press.
Currie, Gregory (1991): “Work and Text”, Mind 100, 325-340.
Currie, Gregory (1999): “Visible Traces: Documentary and the Content of Photographs”, The Journal of Aesthetics and Art Criticism, 57, 285-297.
Currie, Gregory (2014): “Standing in the Last Ditch: On the Communicative Intentions of Fiction Makers”, Journal of Aesthetics and Art Criticism 72:4, 351-363.
García-Carpintero, Manuel (2013): “Norms of Fiction-Making”, British Journal of Aesthetics, 53: 339-357, DOI: 10.1093/aesthj/ayt021.
García-Carpintero, Manuel (2016): ‘To Tell What Happened as Invention: Literature and Philosophy on Learning from Fiction’, in Andrea Selleri and Philip Gaydon (eds.): Literary Studies and the Philosophy of Literature: New Interdisciplinary Directions (London: Palgrave Macmillan, 123-147).
García-Carpintero, Manuel (2019a): “Normative Fiction-Making and the World of the Fiction”, Journal of Aesthetics and Art Criticism, DOI: 10.1111/jaac.12660.
García-Carpintero, Manuel (2019b): “On the Nature of Fiction-Making: Austin or Grice?”, British Journal of Aesthetics, DOI: 10.1093/aesthj/ayy054.
García-Carpintero, Manuel (2019c): “Assertions in Fictions: An Indirect Speech Act Account”, Grazer Philosophische Studien, DOI: 10.1163/18756735-09603013..
Konrad, Eva-Maria (2017): “Signposts of Factuality: On Genuine Assertions in Fictional Literature”, in Sullivan-Bisset, E., Bradley, H. and Noordhof, P. (eds.), Art and Belief, Oxford: OUP, 42-62.
Rosen, Gideon (2015): “Real Definition”, Analytic Philosophy 56(3), 189-209.
Stock, Kathleen (2017): Only Imagine, Oxford: OUP.
Response from Greg Currie.
Thanks to Manuel García-Carpintero for comments that focus on how I characterise fiction in Imagining and Knowing. I said my approach to understanding what fiction is works from “the bottom up”. He thinks that’s wrong. He has a point, but I’ll develop it in my own way.
I said that we can distinguish between somebody telling us that it is raining and someone encouraging us to imagine that it is, the latter being a clear case of fictive utterance. I said that larger stretches of discourse are often a mixture of the two with no clear way to dissect the result into fiction and nonfiction; calling such discourse fiction or nonfiction is often then a matter of discerning, if we can, a primary motivation on the part of the speaker.
And here’s the point: it would be wrong to say that we can take individual sentences from a text and label them “wholly fictional/nonfictional”. If we read
(1) Giulia emerged from Milan’s oppressively magnificent cathedral
we may have the sense that while what is said is presented as true in the fiction (Giulia being a merely fictional person), it is also implied (weakly or strongly) that the cathedral (really) is an aesthetic failure; the sentence itself seems to have a fictional and a nonfictional aspect, something Manolo notes early in this commentary. We generally could go through the text and locate a range of things offered to us to imagine, as well as a range of things intended to be believed, or perhaps to be considered as serious candidates for belief, though these things would not correspond in any systematic way to the sentential division of the text. And I take it that it is our general sense (without doing a thought-by-thought calculation) of what is offered to be imagined and what offered for belief, that helps us make or defend a judgement that the work is fiction or that it is nonfiction. In the example I took from Stacie Friend—a newspaper article which makes vivid a certain real danger by asking us to imagine how things may be in a few months time—there might be a greater quantity of fictional material than there is stuff that is asserted (p.23). Yet we recognise that the primary purpose of the piece is assertive, and judge it to be nonfiction as a result.
You might say there is holism in this: we have to have a sense of the author’s overall intention before we can sensibly label the text fiction or otherwise. True, but the situation is worse than that from the point of view of a bottom up approach. For it now turns out that the elements we identify as fictional or as nonfiction are not given independently of the whole, as they would be if it were sentences we were identifying. To identify the elements we need to see the global relations between all this stuff; they do not exist there in the text of the work, able to be identified by the fact that they begin with a capital and end with a stop. How we interpret a textual unit like (1) depends on our sense of the tone of the whole. I took 1 to suggest that the cathedral really is oppressive. But what if (1) is free indirect discourse? In that case “oppressively magnificent” represents Giulia’s way of thinking, and is no longer exportable to the real world. Decisions like that can be hard to make and may indeed have no resolution. If there is an answer it depends on our sense of the whole. There is a hermeneutical circle here; we determine the parts and the whole together in a process of mutual adjustment.
Manolo puts it like this:
the fictive character or otherwise of minimal utterances can only be (metaphysically!) established relative to their contribution to that of the full utterance constituting the fictional work – whether or not it coincides with the total amount of sentences or shots in it.
I take it he is saying that the circle is not simply epistemic; the process of interpretation generates the parts and the whole together. I am not sure; I like to think of interpretation as discovery, though it will often be the discovery of indeterminacy. But the problem is at least an epistemic one and it shows that “bottom up” is not the right way to put it. But then “top down” would not be right either, since neither top nor bottom has priority. “Mutual adjustment” sounds better to me.