Skip to main content

A Puzzle about Imagining Believing


Suppose you’re imagining that it’s raining hard. You then proceed to imagine, as part of the same imaginative project, that you believe that it isn’t raining. Such an imaginative project is possible if the two imaginings arise in succession. But what about simultaneously imagining that it’s raining and that you believe that it isn’t raining? I will argue that, under certain conditions, such an imagining is impossible. After discussing these conditions, I will suggest an explanation of this impossibility. Elaborating on the view outlined in Walton (1990), I will argue that the impossibility follows from the fact that imaginings ‘mimic’ beliefs in aiming at the fictionally true, just as beliefs aim at the true.

This is a preview of subscription content, access via your institution.


  1. Although belief-like imaginings are sometimes called ‘propositional’ or ‘attitudinal’ imaginings, using these terms here could be misleading, since desire-like imaginings and sensory (perception-like) imaginings can also be deemed propositional or attitudinal; see Kind (2013, 154-155); (2016, 5–6).

  2. See, e.g., Arcangeli (2019b), Currie and Ravenscroft (2002), Doggett and Egan (2007, 2012), Gendler (2003), Ichino (2019), Kind (2013), Langland-Hassan (2012), Liao and Doggett (2014), Liao and Gendler (2020), Nichols (2004, 2006), Nichols and Stich (2003), Van Leeuwen (2011, 2013), Walton (1990, 2015).

  3. See, e.g., Currie and Ravenscroft (2002), 12ff; Nichols (2006), §3.5; Weinberg and Meskin (2006), 180–181; Van Leeuwen (2013); Stock (2017), §6.2. Another, less discussed, similarity between imaginings and beliefs is that imaginings ‘aim’ at the fictional just as beliefs aim at the true (Walton 1990; see §4 below).

  4. That we cannot imagine an overt contradiction is usually accepted not only by those who take imagining to have an ‘imagistic’ or sensory component, but also by those who deny this; see Kind (2013, 151). Note too that I did not list the attitude of supposing, which on certain views (e.g., Currie and Ravenscroft 2002; cf. Balcerak Jackson 2016; Arcangeli 2019a) is regarded as (a type of) belief-like imagining. For if supposing is (a type of) belief-like imagining, the scope of belief-like imagining to which the puzzle in question applies should be narrowed, since without further qualification, we can suppose an overt logical contradiction (e.g., in reductio ad absurdum), and a fortiori, we can suppose content such as Moorean conjunctions. To be compatible with these views, my argument should be taken to apply only to imaginings that are constrained in certain ways, ways that render them subject to the puzzle (see my explanation below). Accordingly, my solution to the puzzle should be seen as revealing the specific constraints that render imagining subject to the puzzle.

  5. I will take an imaginative project to be the overall mental activity associated with imagining. More precisely, an imaginative project encompasses implicit and explicit imaginings, stipulations regarding the fictional world at which one’s imaginings are directed, patterns of inference that apply to one’s imaginings, various kinds of mental states that are functionally related to one’s imaginings (i.e., emotional responses, conative states, mental imagery, etc.), and so on.

  6. Some philosophers, e.g., Kind (2001), argue that imaginings must have an imagistic component; others (Walton 1990, 13; Van Leeuwen 2013, 222) disagree. My argument is neutral with respect to this debate. See also Arcangeli (2019b) for a recent discussion of the relation between mental images and sensory imagination.

  7. Defending a view on which belief-like imaginings are relatively unconstrained, Stock (2017) argues that “one can propositionally imagine that anything is the case, at least as long as one also imagines that there is some (good) explanation for it” (141). On Stock’s view, my account should be interpreted as an explanation of why it is impossible to imagine a Moorean conjunction by virtue of imagining a good explanation for it. Since my starting point is the tension between imagining p and imagining that one believes that not-p, the explanation for this tension can be framed in terms of different views of belief-like imagination, and specifically, views that impose minimal constraints on the scope of imaginative content.

  8. This description of third-person belief-ascription is largely theory-neutral. Specifically, it is compatible with both ‘theory-theory’ and simulation accounts of mindreading, since on both sorts of accounts, ascribing a belief from the third-person perspective is based on behavioral evidence. See Barlassina and Gordon (2017).

  9. More precisely, Walton characterizes fictional worlds as “sets of propositions-as-indicated-by-a-given-work” (1990, 67). Two fictional worlds may thus be composed of the same set of propositions, yet differ because they are set forth in different works. Since Walton takes fictional worlds to be stipulated not only when engaging with fiction, but also in playing games of make-believe, daydreaming, etc. (Walton 1990, pp. 44–45), this characterization also applies to imaginative projects more generally: a fictional world is the set of propositions stipulated to be true in a given imaginative project. That is, a fictional world is associated with the mental activity—the imaginative project—in the context of which it is set up. Hence it is also individuated by the relations between imaginings and other kinds of mental states that are part of the imaginative project (emotional responses, conative states, mental imagery, etc.).

  10. On this characterization, fictional worlds almost never overlap the real world, since they usually include propositions that are false simpliciter. Furthermore, unlike the real world, fictional worlds are usually highly indeterminate.

  11. More specifically, taking ‘Bp’ to denote one’s first-person belief that p, the idea is that, modulo the axioms of doxastic logic, B(p & B(not-p)), as well as B(p & not-Bp), are inconsistent. Regarding the omissive form, given that belief distributes over conjunction, B(p & not-Bp) entails Bp and B(not-Bp). On the axiom according to which Bp entails B(Bp) (commonly referred to as axiom 4), this yields B(Bp) and B(not-Bp), which violates another axiom of doxastic logic (axiom D), according to which a contradiction cannot be believed. The inconsistency of the commissive form is derived from another axiom (4c), according to which B(B(not-p)) entails B(not-p), which is inconsistent with the first conjunct, i.e., Bp; see, e.g., Rieger (2015, 217-218). Other logical analyses of the paradox have been suggested by Rieger (2015, 218ff), Van Benthem (2004), and others (Van Benthem discusses the paradox in the context of the paradox of the knower, analyzing it in terms of dynamic logic).

  12. The claim that Moore’s paradox does not involve a contradiction is occasionally made by those who generally do analyze epistemic paradoxes in terms of doxastic logic. E.g., Yalcin (2007) differentiates between doxastically-paradoxical conjunctions such as ‘it is raining and it might not be raining,’ and Moorean conjunctions, arguing that unlike those doxastically-paradoxical conjunctions, Moorean conjunctions are not contradictory, as their conjuncts do not “have incompatible truth-conditions” or “mutually entail each other’s falsity” (984).

  13. On the approach that analyzes Moore’s puzzle in terms of doxastic logic (see note 11), the property of being committed to truth does not straightforwardly explain why beliefs are subject to the puzzle. On this approach, the specific property of beliefs that renders them subject to the puzzle is described by axiom D (which indeed follows from the ‘commitment to truth’ feature), on which a contradiction cannot be believed.

  14. Note that, since we can obviously consider (assume, etc.) a contradiction, a fortiori we can consider (assume etc.) the conjunction ‘p and I consider (assume, etc.) that not-p.’ In fact, we can also imaginep and I imagine that not-p,’ but in this case, the possibility of imagining this conjunction ensues since the real-world imagining and the imagined imagining are directed at two different fictional worlds.

  15. That it is possible to consider (assume, hypothesize, etc.) Moorean conjunctions, but not to imagine them (in the sense discussed in this paper, and under the said conditions) is often implicit in various philosophical arguments. E.g., in discussing the ‘brains in a vat’ thought experiment, we can obviously consider the conjunction ‘I’m a brain in a vat, but I believe that I’m not.’ Yet without further qualification, imagining this conjunction under the said conditions is impossible: we cannot imagine, e.g., that we are not really eating a delicious red apple, while simultaneously imagining that we believe, from the first-person perspective, that we are eating that apple. In trying to imagine the second conjunct, it seems that we must set aside our imagining of the first conjunct, ‘detach’ ourselves from it, and adopt a different perspective from which it is the case that we are eating the apple.

  16. Note that Walton uses ‘fictionality’ and ‘fictional’ to refer to fictional truth. Note also that, as explained in §3, the set of propositions stipulated to be true in an imaginative project constitutes a ‘fictional world’; see Walton (1990, 66-67), and note 9 above.

  17. On the ‘doxastic logic’ approach (see note 11), formalizing the inconsistency that arises from Walton’s thesis requires revising certain axioms. Assuming that, like beliefs, imagining propositions in a given imaginative project distributes over conjunction, IB(p & B(not-p)) entails IB(p) and IB(B(not-p)) (where ‘IB’ denotes belief-like imagining). However, applying an ‘imaginative’ version of 4c is problematic, since IB(B(not-p)) obviously does not entail B(not-p) (i.e., imagining a belief does not entail that the belief arises in the real world; moreover, even if it entailed this belief, we wouldn’t arrive at the required contradiction). One way to derive a contradiction from IB(p) and IB(B(not-P)) is to argue that, since the belief in question is imagined under the first-person mode of presentation, it is ‘transparent’ in the sense that IB(B(not-p)) entails IB(not-p). The idea is that if one imagines (first-person-) believing a proposition, one also imagines, due to the alleged transparency, that very proposition. We then get that IB(not-p) is inconsistent with IB(p), by an ‘imaginative’ version of axiom D. A different way to derive a contradiction is by accepting another thesis of Walton, i.e., that every imagining is de se (Walton 1990, §I.4), or specifically that imagining something involves “imagining (oneself) believing or knowing it” (1990, 214). On this thesis, the first conjunct, IB(p), entails IB(B(p)). And since B(p) and B(not-p) are inconsistent (i.e., as per the axiomatic analysis of Moore’s paradox, i.e., regarding belief), their conjunction cannot be imagined. Note, however, that since I proceed to reject Walton’s thesis, and argue that the impossibility in question ensues from a different kind of commitment, I will not develop or defend these ideas here.

  18. As mentioned in §1 and note 7 above, on certain accounts of imagining, to say that it is constitutive of belief-like imaginings that they are posited to be fictional first-person beliefs might be too strong. For on some accounts of imagining, imaginings are belief-like only in circumscribed respects, in specific contexts, or under specific constraints. To align with these views, my claim should be modified as follows: in the specific cases where imaginings are subject to the puzzle, they are subject to the puzzle because they function, or are constrained so as to function, as fictional first-person beliefs.

  19. This claim applies only to mental images that arise in the course of imagining, and not to those that arise in other contexts, e.g., in recollection. However, even in the context of imagining, some would dispute this claim. Arcangeli (2019b), e.g., argues that mental images are not a distinct kind of attitude, but rather a type of (sensory) content that is common to different attitudes (remembering, belief, desire, imagining). Hence it appears that, on Arcangeli’s view, mental images (qua mental content) cannot serve as props. However, even on Arcangeli’s view, real-world visual experiences (and, similarly, actions, desires, intentions) can serve as props. E.g., it is fictionally true of Greg and Eric’s visual experience of a tree stump that it is a visual experience of a bear. In this vein, my claim is that imaginings themselves ipso facto serve as props: they are posited to be fictional first-person beliefs.

  20. On the said formalistic approach, this should be analyzed in terms of a commitment to follow the principle that imaginings are posited to be fictional (first-person) beliefs. This idea can be expressed by assuming that an ‘ideal’ imaginer must also imagine the fictionally-true proposition that follows from this principle, namely, she must imagine B(p & B(not-p)). However, since B(p & B(not-p)) is contradictory (according to the standard analysis of Moore’s paradox in terms of doxastic logic; see note 11), it cannot be imagined (as per the impossibility of imagining a contradiction).

  21. On the first-person perspective in the context of imagining, see Nichols (2008).

  22. In this paper, I defend the claim that imaginings are ipso facto posited to be first-person beliefs by arguing that it provides a cogent and plausible explanation of the impossibility in question. I have presented other arguments for this claim elsewhere (see Chasid 2020, Forthcoming). Let me sketch the contours of one such argument here. Consider the following example. In reading a work of fiction and imagining that A is the villain (as the work’s first chapters implicitly direct us to do), and then shifting to imagine that B, not A, is the villain (as per the final chapter’s guidance), we ascribe a sense of ‘error’ to our first imagining, we feel a surprise-like or discovery-like feeling, etc. Such reactions arise even if we read the work a second time, i.e., while knowing the pertinent fictional truths. I argue that these reactions—i.e., ascribing ‘error,’ feeling ‘surprised,’ etc.—show that we posit a fictional doxastic point of view vis-à-vis the propositions we imagine. More generally, our responses to imagining demonstrate that, when we imagine, we posit ourselves to be fictional believers vis-à-vis what we imagine, although we do not necessarily imagine ourselves believing anything.


  • Adler, Jonathan E., and Bradley Armour-Garb. 2007. Moore’s paradox and the transparency of belief. In Moore’s paradox: New essays on belief, rationality, and the first person, ed. Mitchell S. Green and John N. Williams, 147–162. Oxford: Oxford UP.

    Google Scholar 

  • Arcangeli, Margherita. 2019a. Supposition and the imaginative realm. New York: Routledge.

    Google Scholar 

  • Arcangeli, Margherita. 2019b. The two faces of mental imagery. Philosophy and Phenomenological Research (Early view).

  • Balcerak Jackson, Magdalena. 2016. On the epistemic value of imagining, supposing, and conceiving. In Knowledge through imagination, ed. Amy Kind and Peter Kung, 41–60. Oxford: Oxford UP.

    Chapter  Google Scholar 

  • Barlassina, Luca and Gordon, Robert (2017). Folk psychology as mental simulation. The Stanford Encyclopedia of Philosophy, Edward N. Zalta (ed.), URL = <>.

  • Byrne, Alex. 2005. Introspection. Philosophical Topics 33 (1): 79–104.

    Article  Google Scholar 

  • Chasid, Alon. 2016. Imaginatively-colored perception: walton on pictorial experience. The Southern Journal of Philosophy 54 (1): 27–47.

  • Chasid, Alon. 2020. Imagining in response to fiction: unpacking the infrastructure. Philosophical Explorations 23 (1): 31–48.

  • Chasid, Alon. Forthcoming. Belief-like imagining and correctness. American Philosophical Quarterly.

  • Coliva, Annalisa. 2015. How to commit Moore’s paradox. Journal of Philosophy 112 (4): 169–192.

    Article  Google Scholar 

  • Currie, Gregory, and Ian Ravenscroft. 2002. Recreative minds: Imagination in philosophy and psychology. Oxford: Oxford UP.

    Book  Google Scholar 

  • Doggett, Tyler, and Andy Egan. 2007. Wanting things you Don’t want. Philosophers’ Imprint 7 (9): 1–17.

    Google Scholar 

  • Doggett, Tyler, and Andy Egan. 2012. How we feel about terrible, non-existent Mafiosi. Philosophy and Phenomenological Research 84 (2): 277–306.

    Article  Google Scholar 

  • Fernández, Jordi. 2013. Transparent minds: A study of self-knowledge. Oxford: Oxford UP.

    Book  Google Scholar 

  • Gendler, Tamar. 2003. On the relation between Pretense and belief. In Imagination, philosophy, and the arts, ed. Matthew Kieran and Dominic Lopes, 125–141. London: Routledge.

    Google Scholar 

  • Green, Mitchell S., and Williams, John N., (eds.) (2007). Moore’s paradox: New essays on belief, rationality, and the first person. Oxford UP.

    Google Scholar 

  • Hintikka, Jaakko. 1962. Knowledge and belief. Ithaca: Cornell UP.

    Google Scholar 

  • Ichino, Anna. 2019. Imagination and belief in action. Philosophia: 1–18.

  • Kind, Amy. 2001. Putting the image Back in imagination. Philosophy and Phenomenological Research 62: 85–109.

    Article  Google Scholar 

  • Kind, Amy. 2013. The heterogeneity of the imagination. Erkenntnis 78 (1): 141–159.

    Article  Google Scholar 

  • Kind, Amy (ed.) (2016). Introduction: Exploring imagination, in Amy Kind (ed.), Routledge handbook of philosophy of imagination. New York: Routledge, 1–11.

  • Langland-Hassan, Peter. 2012. Pretense, imagination, and belief: The single attitude theory. Philosophical Studies 159 (2): 155–179.

    Article  Google Scholar 

  • Liao, Shen-yi, and Tyler Doggett. 2014. The imagination box. Journal of Philosophy 111 (5): 259–275.

    Article  Google Scholar 

  • Liao, Shen-yi and Gendler, Tamar (2020). Imagination. The Stanford Encyclopedia of Philosophy, Edward N. Zalta (ed.), URL = <>.

  • Lycan, William. 1996. Consciousness and experience. Cambridge MA: MIT Press.

    Google Scholar 

  • Nichols, Shaun. 2004. Imagining and believing: The promise of a single code. Journal of Aesthetics and Art Criticism 62 (2): 129–139.

    Article  Google Scholar 

  • Nichols, Shaun. 2006. Introduction. In The architecture of the imagination, ed. Shaun Nichols, 1–16. New York: Oxford UP.

    Chapter  Google Scholar 

  • Nichols, Shaun. 2008. Imagination and the I. Mind & Language 23 (5): 518–535.

    Article  Google Scholar 

  • Nichols, Shaun, and Stephen Stich. 2003. Mindreading: An integrated account of pretence, self-awareness, and understanding of other minds. Oxford: Oxford UP.

    Book  Google Scholar 

  • Rieger, Adam. 2015. Moore’s paradox, introspection and doxastic logic. Thought: A Journal of Philosophy 4 (4): 215–227.

    Google Scholar 

  • Stock, Kathleen. 2017. Only imagine: Fiction, interpretation and imagination. Oxford: Oxford UP.

    Book  Google Scholar 

  • Tuna, Emine Hande (2020). Imaginative resistance. The Stanford Encyclopedia of Philosophy, Edward N. Zalta (ed.), URL = <>.

  • Van Benthem, Johan. 2004. What one may come to know. Analysis 64 (2): 95–105.

    Article  Google Scholar 

  • Van Leeuwen, Neil. 2011. Imagination is where the action is. Journal of Philosophy 108 (2): 55–77.

    Article  Google Scholar 

  • Van Leeuwen, Neil. 2013. The meanings of ‘imagine’ part I: Constructive imagination. Philosophy Compass 8 (3): 220–230.

    Article  Google Scholar 

  • Walton, Kendall. 1990. Mimesis as make-believe. Cambridge MA: Harvard UP.

    Google Scholar 

  • Walton, Kendall. 2015. In other shoes: Music, metaphor, empathy, existence. New York: Oxford UP.

    Google Scholar 

  • Weinberg, Jonathan, and Aaron Meskin. 2006. Puzzling over the imagination: Philosophical problems, architectural solutions. In The architecture of the imagination, ed. S. Nichols, 175–202. New York: Oxford UP.

    Chapter  Google Scholar 

  • Yalcin, Seth. 2007. Epistemic modals. Mind 116 (464): 983–1026.

    Article  Google Scholar 

Download references


I am grateful to those who commented on this paper and helped sharpen the argument. In particular, I’d like to thank Anna Ichino, Peter Langland-Hassan, Julia Langkau, Nessa Olshansky-Ashtar, Alik Pelman, Eric Peterson, and Assaf Weksler for their input. An earlier version of this paper was presented at the 2019 SSPP annual meeting, where I received valuable feedback. I also want to thank two anonymous referees, and the ROPP Editor, for their insightful comments. This research was supported by the Israel Science Foundation (grant No. 939/16).

Author information

Authors and Affiliations


Corresponding author

Correspondence to Alon Chasid.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Chasid, A. A Puzzle about Imagining Believing. Rev.Phil.Psych. (2021).

Download citation

  • Accepted:

  • Published:

  • DOI: