Abstract
On an ‘internalist’ picture, knowledge isn’t necessary for understanding the nature of perception and perceptual experience. This contrasts with the ‘knowledge first’ picture, according to which it’s essential to the nature of successful perceiving as a mental state that it’s a way of knowing. It’s often thought that naturalistic theorizing about the mind should adopt the internalist picture. However, I argue that a powerful, recently prominent framework for scientific study of the mind, ‘predictive processing,’ instead supports the knowledge first picture. Under predictive processing, it’s intrinsic to successful perceiving that it’s a state with the kind of modal robustness that’s distinctive of knowledge, which gives us reason to think of successful perceiving along knowledge first lines. Furthermore, I argue that the predictive processing framework encourages us to conceptualize experiences which don’t amount to knowledge along knowledge first lines, as states which by their nature fall short of knowledge.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
Philosophers of this tendency often trace their inspiration to Quine (1969). A related line of argument, also often traced to Quine, focuses not on truth but on normativity: knowledge is a normative notion because it entails notions like justification; since sciences of the mind deal in descriptive facts, normative facts are somehow non-naturalistic. This paper won’t directly address this line of thought; for resistance to it, see, e.g., Kim (1988) and Jenkins (2007).
For example: perceptual attention (Clark 2016, ch. 2); various perceptual phenomena and illusions (Hohwy 2013, ch. 5); the nature of consciousness (Clark, Friston, and Wilkinson 2019); hallucinations and delusions in conditions like schizophrenia (Benrimoh et al., 2018); the imagination (Kirchhoff 2018); emotions (Seth 2013); and various elements of social cognition, including mindreading in autism (Palmer et al., 2015) and stereotyping (Westra 2019).
For more examples and scientifically-informed discussion of such abstract expectations in perceptual processing (though not specifically in a PP framework), see Jenkin (2020). It’s a further, interesting question exactly where such priors come from; plausibly, some are innate while others are learned over time. On this question, see Clark (2013) and Kirchhoff (2018).
I don’t mean to saddle ‘internalist’ views with semantic internalism. Whether the contents of mental states depend on external factors is a distinct issue from whether types of states do (but see Williamson 2000, ch. 2 on why reasons paralleling those for the former may favour the latter).
Views of this sort can be further divided into at least two kinds. Some are more inferentialist, on which one infers beliefs about the external world from facts about how things seem or appear to oneself (e.g., McGrath 2017; Schiffer 2009). Others have it that our perceptual beliefs are non-inferential, but their basis is nevertheless still how things perceptually seem (e.g., Huemer 2006; Pryor 2000). For my purposes here, I’ll lump these two together (as in, e.g., White 2014), but see Barnett (2019) for discussion of important differences between them.
For views of this sort, see Williamson (2000) and Nagel (2013, 2021). I’m putting things a bit differently from these theorists here by referring to successful perceiving as a way of knowing, since others often write about genuine perceiving as a way of knowing—i.e., they hold that genuine perception is a state of knowledge and that, in the bad case, one doesn’t genuinely perceive. However, I take it that this is largely a terminological point, and that my talk of successful perceiving is more neutral between the knowledge first and internalist pictures (since both sides can agree that cases where perception provides knowledge are the cases where everything goes well perceptually). The key disagreement between them is whether, in the good case, perceiving literally is a way of knowing (where knowing is itself a mental state) or is merely a state which provides knowledge (where knowledge isn’t a mental state).
See Cassam (2007) for more detailed discussion of this point. (Cassam himself adopts a ‘mere entailment’ view of the relationship between successful perceiving and knowing, but he helpfully positions his view in relation to the stronger, Williamson-inspired one I’m adopting.).
It’s natural to ask whether the knowledge first picture is a kind of ‘disjunctivism’ about perception, though it’s difficult to give a straightforward answer due to the diversity of views under the disjunctivist label (see Soteriou 2016 for a detailed overview). If disjunctivism is merely taken as the idea that there’s a mental difference between subjects in the good and bad cases, then the knowledge first view is disjunctivist. However, some disjunctivists hold the stronger view that there’s nothing mentally in common between the good and bad case. The knowledge first view isn’t committed to this, since it can accept that a subject who knowledgeably perceives that P and one who hallucinates that P both occupy the mental state of seeming to perceive that P (see Williamson 2000, ch. 1). Furthermore, many disjunctivists are committed to the ‘naïve realist’ idea that perception is metaphysically direct in the good case. The knowledge first picture is neutral about this: what matters for the knowledge first picture is whether the connection successful perception establishes between a subject and the world is a knowledge-constituting one, not whether this connection is direct or indirect.
Whether or not all proponents of PP would accept this reading, it’s at least not idiosyncratic. It’s suggested by, for example, various remarks in Clark (2016), who distinguishes between ‘non-conscious (‘sub-personal’) prediction and the shape and flow of personal-level daily experience’ (78). I don’t mean here to deny that we also have conscious, personal-level credences in addition to outright beliefs (cf. Weisberg 2020), but, on my reading of PP, these would also be personal-level states that result from probabilistic sub-personal processing. An alternative to my favoured picture is that all personal-level beliefs and perceptual experiences inherit graded content from the probabilistic processing that produces them. This would require accepting that perceptual experiences have graded content, which, though accepted by some (e.g., Morrison 2016), I take to be the non-standard view.
My descriptions here might seem to over-intellectualize the notion of prediction involved in PP. It’s often argued that we should think of the generative model as a kind of structural representation, more akin to a map of the world than a set of sentences describing the world (Gładziejewski 2016; Wiese 2017; Williams 2018). By saying perceiving that P is a propositional attitude, am I committing to a conception of the brain’s predictions as having a more complex, linguistic structure? Not necessarily: we often describe and model non-linguistic representations propositionally (when describing the contents of pictures, episodic memory, sensory imagination, etc.). Assuming it makes sense to do so, we can do the same for perception in PP. For general discussion of the relationship between non-linguistic mental representations and propositional attitudes, see Langland-Hassan (2011).
For classic relevant alternatives accounts, see Stine (1976) and Goldman (1976). Theorists disagree about whether, and exactly how, farther-out possibilities become relevant as one moves between different conversational contexts—for example, whether merely entertaining a possibility makes it relevant (cf. Lewis 1996) or whether the practical importance of knowing P changes which possibilities are relevant (cf. Cohen 1999).
For internalists, this is because such modal robustness is a condition beliefs must meet to amount to knowledge, in addition to other conditions like truth. For knowledge first proponents, it’s because knowledge is a mental state which essentially or necessarily links one to truths, rather than one which merely happens to involve truth.
This is not to say that each particular brain’s ordering of the likelihoods of these alternative possibilities reflects their exact ordering in modal space, besides the fact that they’re modally nearby. What matters for the modal robustness of perception is whether alternative possibilities are in fact ruled out, not whether the probabilistic process by which they’re ruled out precisely models their modal distribution.
Or, one would have to change the world to bring it in line with one’s current leading hypothesis. For simplicity, I’ve left this kind of ‘active inference’ out of my account of PP in this paper, though it’s an important missing component (see Clark 2016, ch. 4). Regardless, this process would still help to ensure the modal robustness of one’s winning hypothesis, because it rules out an alternative to this hypothesis—this alternative being the way the world currently is, which gets ‘ruled out’ by changing the world.
This isn’t to say all possible bad cases result from a failure to rule out modally nearby possibilities. I claim that such failure is sufficient for being in a bad case, but there may also be other routes into bad cases. This section focuses on bad cases that illustrate that it’s necessary for the good to rule out modally nearby hypotheses, but Sect. 4 discusses other bad cases.
There’s one prominent kind of bad case I’m not touching on in this section: Cartesian skeptical scenarios like the brain-in-a-vat. Such scenarios might seem especially tricky for the knowledge first picture, since they involve a brain that, though it lacks knowledge, seems psychologically identical to its good case twin. However, my view is that PP is particularly conducive to the arguments of Chalmers (2005), according to which brain-in-the-vat-type scenarios aren’t actually bad cases after all; see Hohwy (2017) and Clark (2017) for discussion of this point. Unfortunately, discussion of these large issues would take me too far afield here. But adopting a Chalmers-style assessment of brain-in-a-vat cases allows us to hold both that the brain-in-a-vat and her good case twin are mentally identical and that both have perceptual knowledge, meaning such cases pose no counterexample to the knowledge first picture.
Hohwy himself doesn’t discuss the hollow mask example but applies his account to other illusions: the Müller-Lyer illusion (where manipulating the image in the right way allows one to veridically see the lines as equal) and multisensory illusions like the rubber hand (where the illusion persists only under conditions in which bodily movement is suppressed).
This is a very simplified account of hallucination, but it’s sufficient for my purposes. One obvious thing I’ve left out is why persistent, false error signals might be generated in the first place; see again Fletcher and Frith (2009). I’ve also overlooked the importance for PP accounts of hallucination of a mechanism known as ‘precision weighting,’ by which the brain adjusts the relative weightings of top-down predictions and bottom-up error signals.
The idea that hallucination is a kind of unsuccessful perception is especially stark in PP versus a bottom-up account of perception. On the latter, since perception involves receiving external sensory stimulation and extracting information from it, it’s not totally clear what the relationship between perception and hallucination is—hallucinations don’t seem very perception-like, given that they’re generated internally (cf. Wilkinson 2014, 149). At the very least, it’s not clear whether to think of hallucination as an unsuccessful version of perception, given their different aetiologies. Under PP, it’s more straightforward to see hallucination as unsuccessful perception, since it involves the same top-down predictions as in perception failing to be properly controlled.
Thank you to Jennifer Nagel for alerting me to the relevance of this sort of case.
As an anonymous reviewer helpfully points out, we can make it especially clear why one’s perceptual representation is unsuccessful here by considering how it fails to afford long-run prediction error minimization. The generative model encodes predictions about what would occur were one to interact with one’s environment, which would include, for example, that one will feel the flame’s heat if one reaches towards the candle. In this deviant case, though, reaching towards the candle will instead cause you to bump into the mirror, generating prediction error. This highlights why one’s generative model is sub-optimal when it fails to represent the mirror: it’s not optimized for prediction error minimization across possible interactions with one’s environment. One can only resolve this by incorporating the mirror into one’s model of the world.
For arguments that states involved in sub-personal perceptual processing can stand in epistemological relations to other states, including personal ones, see Jenkin (2020).
References
Armstrong, D. M. (1968). A materialist theory of the mind. Routledge & Kegan Paul.
Barnett, D. J. (2019). Perceptual justification and the Cartesian theatre. In T. S. Gendler & J. Hawthorne (Eds.), Oxford studies in epistemology (Vol. 6, pp. 1–34). Oxford University Press.
Benrimoh, D., Parr, T., Vincent, P., Adams, R. A., & Friston, K. (2018). Active inference and auditory hallucinations. Computational Psychiatry, 2, 183–204.
Byrne, A. (2009). Experience and content. The Philosophical Quarterly, 59(236), 429–451.
Cassam, Q. (2007). Ways of knowing. Proceedings of the Aristotelian Society, 107(3), 339–358.
Chalmers, D. J. (2005). The matrix as metaphysics. In C. Grau (Ed.), Philosophers explore the matrix. Oxford University Press.
Churchland, P. S. (1987). Epistemology in the age of neuroscience. The Journal of Philosophy, 84(10), 544–553.
Clark, A. (2013). Whatever next? Predictive brains, situated agents, and the future of cognitive science. Behavioral and Brain Sciences, 36(3), 181–204.
Clark, A. (2016). Surfing uncertainty: Prediction, action, and the embodied mind. Oxford University Press.
Clark, A. (2017). How to knit your own Markov blanket. In T. Metzinger & W. Wiese (Eds.), Philosophy and predictive processing. MIND Group.
Clark, A. (2018). Beyond the “Bayesian Blur”: Predictive processing and the nature of subjective experience. Journal of Consciousness Studies, 25(3–4), 71–87.
Clark, A., Friston, K., & Wilkinson, S. (2019). Bayesing qualia: Consciousness as inference, not raw datum. Journal of Consciousness Studies, 26(9–10), 19–33.
Cohen, S. (1999). Contextualism, skepticism, and the structure of reasons. Philosophical Perspectives, 13, 57–89.
Fantl, J., & McGrath, M. (2009). Knowledge in an uncertain world. Oxford University Press.
Fletcher, P. C., & Frith, C. D. (2009). Perceiving is believing: A Bayesian approach to explaining the positive symptoms of schizophrenia. Nature Reviews Neuroscience, 10(1), 48–58.
Fodor, J. (1980). Methodological solipsism considered as a research strategy in cognitive psychology. Behavioral and Brain Sciences, 3(1), 63–73.
Friston, K. (2005). A theory of cortical responses. Philosophical Transactions of the Royal Society b: Biological Sciences, 360(1456), 815–836.
Friston, K., Rigoli, F., Ognibene, D., Mathys, C., Fitzgerald, T., & Pezzulo, G. (2015). Active inference and epistemic value. Cognitive Neuroscience, 6(4), 187–214.
Friston, K., Schwartenbeck, P., FitzGerald, T., Moutoussis, M., Behrens, T., & Dolan, R. J. (2014). The anatomy of choice: Dopamine and decision-making. Philosophical Transactions of the Royal Society B: Biological Sciences, 369(1655), 20130481.
Frith, C. (2007). Making up the mind: How the brain creates our mental world. Blackwell.
Ganeri, J. (2010). The study of Indian epistemology: Questions of method—A reply to Matthew Dasti and Stephen H. Phillips. Philosophy East and West, 60(4), 541–550.
Gładziejewski, P. (2016). Predictive coding and representationalism. Synthese, 193(2), 559–582.
Goldman, A. I. (1976). Discrimination and perceptual knowledge. The Journal of Philosophy, 73(20), 771–791.
Hobson, J. A., & Friston, K. J. (2014). Consciousness, dreams, and inference: The Cartesian theatre revisited. Journal of Consciousness Studies, 21(1–2), 6–32.
Hohwy, J. (2013). The predictive mind. Oxford University Press.
Hohwy, J. (2016). The self-evidencing brain. Noûs, 50(2), 259–285.
Hohwy, J. (2017). How to entrain your evil demon. In T. Metzinger & W. Wiese (Eds.), Philosophy and predictive processing. MIND Group.
Hohwy, J., Paton, B., & Palmer, C. (2016). Distrusting the present. Phenomenology and the Cognitive Sciences, 15(3), 315–335.
Huemer, M. (2006). Phenomenal conservatism and the internalist intuition. American Philosophical Quarterly, 43(2), 147–158.
Jenkin, Z. (2020). The epistemic role of core cognition. Philosophical Review, 129(2), 251–298.
Jenkins, C. S. (2007). Epistemic norms and natural facts. American Philosophical Quarterly, 44(3), 259–272.
Kim, J. (1988). What is “naturalized epistemology?” Philosophical Perspectives, 2, 381–405.
Kirchhoff, M. D. (2018). Predictive processing, perceiving and imagining: Is to perceive to imagine, or something close to it? Philosophical Studies, 175(3), 751–767.
Langland-Hassan, P. (2011). A puzzle about visualization. Phenomenology and the Cognitive Sciences, 10(2), 145–173.
Levin, M. (2006). Gettier cases without false lemmas? Erkenntnis, 64(3), 381–392.
Lewis, D. (1996). Elusive knowledge. Australasian Journal of Philosophy, 74(4), 549–567.
Lu, Y., Stafford, T., & Fox, C. (2016). Maximum saliency bias in binocular fusion. Connection Science, 28(3), 258–269.
McGrath, M. (2017). Knowing what things look like. Philosophical Review, 126(1), 1–41.
Michaelian, K. (2016). Against discontinuism: Mental Time Travel and our Knowledge of Past and Future events. In K. Michaelian, S. B. Klein, & K. K. Szpunar (Eds.), Seeing the future: Theoretical perspectives on future-oriented mental time travel. Oxford University Press.
Mirza, M. B., Adams, R. A., Mathys, C., & Friston, K. J. (2018). Human visual exploration reduces uncertainty about the sensed world. PLoS ONE, 13(1), e0190429.
Molyneux, B. (2007). Primeness, internalism and explanatory generality. Philosophical Studies, 135(2), 255–277.
Morrison, J. (2016). Perceptual confidence. Analytic Philosophy, 57(1), 15–48.
Nagel, J. (2010). Epistemic anxiety and adaptive invariantism. Philosophical Perspectives, 24, 407–435.
Nagel, J. (2013). Knowledge as a mental state. Oxford Studies in Epistemology, 4, 275–310.
Nagel, J. (2017). Factive and nonfactive mental state attribution. Mind & Language, 32(5), 525–544.
Nagel, J. (2021). Losing knowledge by thinking about thinking. In J. Brown & M. Simion (Eds.), Reasons, justification and defeat (pp. 69–92). Oxford University Press.
Palmer, C. J., Paton, B., Kirkovski, M., Enticott, P. G., & Hohwy, J. (2015). Context sensitivity in action decreases along the autism spectrum: A predictive processing perspective. Proceedings of the Royal Society B: Biological Sciences, 282(1802), 20141557.
Parr, T., Corcoran, A. W., Friston, K. J., & Hohwy, J. (2019). Perceptual awareness and active inference. Neuroscience of Consciousness, 2019(1), niz012.
Parr, T., & Friston, K. J. (2017). The active construction of the visual world. Neuropsychologia, 104, 92–101.
Parr, T., & Friston, K. J. (2018). The anatomy of inference: Generative models and brain structure. Frontiers in Computational Neuroscience, 12, 90.
Paton, B., Skewes, J., Frith, C., & Hohwy, J. (2013). Skull-bound perception and precision optimization through culture. Behavioral and Brain Sciences, 36(3), 222.
Peacocke, C. (1979). Deviant causal chains. Midwest Studies in Philosophy, 4, 123–155.
Pernu, T. K. (2009). Is knowledge a natural kind? Philosophical Studies, 142(3), 371–386.
Pezzulo, G., & Nolfi, S. (2019). Making the environment an informative place: A conceptual analysis of epistemic policies and sensorimotor coordination. Entropy, 21(4), 350.
Plantinga, A. (1993). Warrant: The current debate. Oxford University Press.
Pryor, J. (2000). The skeptic and the dogmatist. Noûs, 34(4), 517–549.
Quine, W. V. O. (1969). Epistemology naturalized. In Ontological relativity and other essays (pp. 69–90). Columbia University Press.
Rao, R. P. N., & Ballard, D. H. (1999). Predictive coding in the visual cortex: A functional interpretation of some extra-classical receptive-field effects. Nature Neuroscience, 2(1), 79–87.
Schiffer, S. (2009). Evidence = knowledge: Williamson’s solution to skepticism. In P. Greenough & D. Pritchard (Eds.), Williamson on knowledge (pp. 183–202). Oxford University Press.
Seth, A. K. (2013). Interoceptive inference, emotion, and the embodied self. Trends in Cognitive Sciences, 17(11), 565–573.
Soteriou, M. (2016). Disjunctivism. Routledge.
Stine, G. C. (1976). Skepticism, relevant alternatives, and deductive closure. Philosophical Studies, 29(4), 249–261.
Swanson, L. R. (2016). The predictive processing paradigm has roots in kant. Frontiers in Systems Neuroscience, 10, 79.
Weisberg, J. (2020). Belief in psyontology. Philosophers’ Imprint, 20(11), 1–27.
Westra, E. (2019). Stereotypes, theory of mind, and the action–prediction hierarchy. Synthese, 196(7), 2821–2846.
White, R. (2014). What is my evidence that here is a hand? In D. Dodd & E. Zardini (Eds.), Scepticism and perceptual justification (pp. 298–321). Oxford University Press.
Wiese, W. (2017). What are the contents of representations in predictive processing? Phenomenology and the Cognitive Sciences, 16(4), 715–736.
Wilkinson, S. (2014). Accounting for the phenomenology and varieties of auditory verbal hallucination within a predictive processing framework. Consciousness and Cognition, 30, 142–155.
Williams, D. (2018). Predictive processing and the representation wars. Minds and Machines, 28(1), 141–172.
Williamson, T. (2000). Knowledge and its limits. Oxford University Press.
Williamson, T. (2005). Knowledge, context, and the agent’s point of view. In G. Preyer & G. Peter (Eds.), Contextualism in philosophy: Knowledge, meaning, and truth (pp. 91–114). Oxford University Press.
Acknowledgements
For comments on earlier drafts of this paper, thank you to David Barnett, Andy Clark, Liang Zhou Koh, Jennifer Nagel, Gurpreet Ratan, Mason Westfall, Evan Westra, and an anonymous reviewer for Philosophical Studies. This research was supported by the Social Sciences and Humanities Research Council of Canada.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
Munro, D. Perceiving as knowing in the predictive mind. Philos Stud 179, 1177–1203 (2022). https://doi.org/10.1007/s11098-021-01702-4
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11098-021-01702-4