This is a preview of subscription content, log in via an institution to check access.
Notes
I use the term ‘allostasis’ instead of ‘homeostasis’ because, by understanding cognition in terms of brain-centered predictive regulation, it comes closer to capturing the material cash value of the distinctively representational dimension of cognition. Consider for example that organisms maintain homeostasis perfectly well without any need for cognitive representations of any kind (Schulkin & Sterling, 2019).
A similar point can be made about the notion of ‘semantic content’. In what follows, it will be argued that the ‘saying power’ of iconic representations comes in the form of affordances and solicitations to act which are devoid of the semantical distinctions between sense and reference, truth and falsity, the subject-predicate structure of claimings, the negation-based operation in understanding incompatible commitments and a view of the world as comprised of states of affairs that make our propositions true or false. In this respect iconic representations are clearly non-semantic. Yet, since it will also be argued that iconic representations as ‘cartographic maps’ can misrepresent and can specify a target under a description (as being a certain way), they can be said to exhibit rudimentary (sub-personal) intensionality, and in this sense, they can be said to have ‘semantic’ content. Thus, when I sometimes characterize iconic representations as non-semantic I mean it in the above sense, which does not exclude a rudimentary form of intensionality.
Perceptual hierarchies recapitulate the interconnected hierarchy of those (slow space/time-scale, fast space/time-scale) nested causal regularities in the world that affect the organism. The fast time-scale regularities are represented in low levels of the hierarchy, have small, detailed focused receptive fields of only a couple of degrees whereas later (higher) areas of processing have wider receptive fields. Fast regularities occur in perceptual inference in the shape of the variant aspect of experience, while slow regularities occur in the shape of the invariant aspect of experience. Low-level, fast scale regularities help choosing among hypotheses higher up and higher-level hypotheses about slower regularities work as control parameters on low-level regularities. Perceptual inference happens in this highly interconnected, cortical hierarchy, and can as such avail itself directly of its representation of myriad causal relations in its attempt to get the world right, in its construction of a first-person perspective, and in its ability to orient itself for action in the world (Hohwy, 2013; Friston, Kiebel and Daunizeau 2010).
Note that on this view, the upper level in each pair of levels only ‘knows’ its own expectations, it is only told how these expectations are wrong, and is never told directly what the level below ‘knows’. Similarly, the upper level never tells the lower level what message to pass on, it only ever anticipates what the lower level will tell it.
An analogous point can be made about the sense in which representations guide action. As the term is used in PP when we say that cartographic maps guide action, it is clearly different from the commonsensical one in which actions are understood as activities resulting from a subject’s desires or intentions. In the context of PP, actions are identified with a technical category of active inference, i.e. with endogenously-controlled activities that aim at prediction error minimization. This way of understating action does not presuppose personal-level intentional or representational categories (this is also needed for avoiding the homuncular fallacy) (Gladziejewski 2016, 574).
For a development of the view that the situatedness of neural representations is the key for solving the problems of a naturalistic determination of representational content see Piccinini, 2022.
There is a further question here over whether the concept of ‘function’ must be understood etiologically, in terms of contribution to fitness via past natural selection (Millikan, 1984; Neander, 2017), or as goal-constitutive, in terms of its present contribution to facilitating organismal goals such as survival, development, reproduction etc. (Moreno & Mossio, 2015; Piccinini, 2020). Hutto & Myin’s objections to teleosemantics are more relevant to the etiological account (but see Mann & Pain, 2022) and it needs to be shown whether they can be applied to teleosemantics based on a goal-constitutive account of function. On my part, though I am sympathetic to goal-constitutive accounts of function, I would prefer to understand function in terms of its contribution to the ‘allostatic balance’ of the brain-body-(natural-social)environment system and construe the ‘teleology’ involved in terms of convergence of the system into certain target states that function as ‘dynamic attractors’. One very general problem that I think plagues goal-constitutive accounts, such as Moreno & Mossio’s (2015), is the following: According to their view, close consideration of the conditions of existence of self-determined organized beings indicates what they ought to do on the grounds that otherwise the system would cease to exist. But a counterfactual argument regarding factual ‘internal conditions of existence’ cannot straightforwardly ground intrinsic normative ascriptions (about what the organism ‘ought to do’), unless it is presupposed on independent grounds that life must be itself valued as a goal in a normative sense (Corti, 2023). It might well be true that organic self-maintenance is a self-striving process, an ‘end in itself’, involving particular emergent forms of closed causal relations of dependence. This, I think, suffices for providing living organisms with an intrinsic directedness or ‘orientational’ structure (towards states that are conducive to its allostatic balance) which is necessary for making meaningful the attribution of malfunction or misrepresentation. But to accept that life is a thusly defined self-striving process does not amount to saying that it is something to be valued in the normative sense of ‘ought’ (as if self-preservation were normatively grounded in the very structure of being).
Here, it must be noted that from this PP viewpoint of understanding the normative-social dimension of our representational abilities, practice-based representations can be seen as a ‘technological’ improvement that permits us, at the level of language and conceptual thought, to perform ‘abstract generalizations’ concerning contents, action, and planning that are not possible in a simpler context (i.e. non-conceptual articulation) of PP representations. Recall (n. 5) that at the non-conceptual level, PP cartographic maps guide action in a way that does not presuppose the formulation of desires or intentions of a fully formed subject. In the context of PP, actions are identified with a technical category of active inference, i.e. with endogenously-controlled activities that aim at prediction error minimization, in a way that does not presuppose personal-level intentional or representational categories. Yet, from PP’s point of view, we can understand our normative conceptual capacities of intentional action as a socially induced technological development that, through the normative use of linguistic signs, better stabilizes, controls and channels the feedback-loopy process of active inference by distinguishing in it particular salience (affordance) patterns that make explicit (represent) what we call our practical reasoning capacities (‘intentions’, ‘purposes’, ‘goals’ etc.) And this obviously amounts to an authentic boost in expressive power and cognitive abilities.
Here I exploit Sellars’ distinction between two different kinds of representation: ‘signifying’ (conceptually articulated representation in the context of social practices) and ‘picturing’ (non-conceptually articulated representations of the environment by the nervous system). For similar uses of this pivotal Sellarsian distinction in discussions about the ‘representation wars’ in cognitive neuroscience see Sachs, 2019 and Sallis 2022.
Of course, it must be constantly borne in mind that at this basic, ‘atomic’ level of representing the world, the ‘descriptive’ part of the non-conceptual representation has direct affective power and which is in turn directly action-oriented. That is, the world at this level is directly perceived as affordance based and as soliciting certain kinds of action.
It must be noted, however, that affordances can be characterized by rudimentary intensionality even if a public language is necessary for subjects to grasp their intensions at the personal level (see also n. 13). Consider, for example, that treating something as food can yield different action possibilities and elicit different motor responses depending on who is considered as its most appropriate owner according to the hierarchy of the group.
This kind of primordial ‘saying power’ or ‘experience of error’ need not be considered as just a primitive level of our experiential contact with the world on top of which our conceptually and linguistically articulated capacities are erected. Indeed, the latter can themselves be understood as a modified and specialized kind of affordance-based embodied skillful coping with ‘signs’ whose conceptual-rational tools can stabilize, make explicit, criticize and thus ultimately alter and redirect our affordance and solicitation-based ‘web of motivations’ (Rouse, 2015). See also Sect. 6.
Interestingly, my view has affinities with Neander’s view to the effect that the subject need not grasp an intension for a content ascription to qualify as intensional (Neander, 2017, 38). In my terms, this amounts to saying that cartographic representation at the subpersonal level could be intensional, even if a public language is necessary for subjects to grasp their intensions at the personal level. One complication here is that, on my view, the ‘subpersonal’ content in question not thereby sub-experiential. It is an affordance-based experience with its own distinctive non-conceptual, non-linguistic content, which however is not Given (transparent) to the experiencer. Only when one becomes a member of our socially and conceptually articulated ‘space of reasons’ practices can one understand the intensionality involved in terms of the sense-reference distinction.
Indeed, proponents of radical enactivism have recently embraced neurally implemented similarity-based cognition, yet purging it of its robustly representational implications. For example, Segundo-Ortin and Hutto (2021) contend that the dynamic activity of the central nervous system can play a part in enabling intelligent behavior by temporally reconfiguring already existing neural structures in order to resemble specific aspects of relevant targets. In this way we can understand how the dynamics of the central nervous system can provide a causal explanation of the intelligent behavior of cognitive systems without assuming that the brain is in the business of contentually representing the external world.
Another important difference of my view from radical enactivism is that although in both views the bottom ‘contentless’ level of cognition is understood as non-conceptually articulated and comes in the form of a direct grasp of worldly affordances, in my view, which here is nearer to PP, this direct grasp of affordances is a characteristic of the generative models as projected in the world (i.e. it is a feature of what, after Kant, we might call ‘outer sense’) and, pace enactivism, it does not amount to a transparent immediate connection or ‘openness’ to the ultimate structure of reality, as it often presupposes categorization that, however elementary, is not just ‘given’ by (non-human or human) nature and is open to revision in the course of empirical inquiry. Recall, in this connection, that according to PP, which here seems to me to be in the right direction, the upper level in each pair of the hierarchy of levels, only ‘knows’ its own expectations, it is only told how these expectations are wrong, and is never told directly what the level below ‘knows’. Similarly, the upper level never tells the lower level what message to pass on, it only ever anticipates what the lower level will tell it. Note that this does not entail that the structure of the ‘really real’ world must come in a form that does not involve affordances. Hutto (2018) and others worry that since in PP minds are secluded from the really real world it is difficult to see how do they come by contents that refer to, or are about, inaccessible hidden causes that they putatively represent in the first place. In my view, PP does not imply a radical skepticism about reality, though PP needs to be more worked out philosophically to avoid such a conclusion.
It must be noted that there are other varieties of enactivism, namely autopoietic enactivism (Thomspon 2007), which, unlike radical enactivism of the Hutto & Myin variety, investigate cognition in a bottom-up fashion from an evolutionary perspective, yielding interesting hypotheses about the biological sources of intentionality, which is understood in terms of homeostatic regulation and reproduction of conditions conducive to self-maintenance. Although most proponents of autopoietic enactivism are anti-representationalists all the way up, there is room for arguing that since cognition is a more complex phenomenon than biological intentionality in that, unlike the latter, it clearly includes directedness towards non-existent things, we cannot exclude the possibility that some higher forms of cognition, e.g. those implemented in organisms with nervous systems, might turn out to be representational (best understood in PP terms). For a development of a view along these lines see Schlicht, 2018; Schlicht & Starzak, 2021. This potential compatibility between autopoietic enactivism and representationalism is surely an interesting theoretical possibility, yet it remains to be seen whether representationalism about cognition can be internally integrated to the enactivist framework or is merely an external appendage to it.
References
Brandom, R. (1994). Making it Explicit: Reasoning, representing and discursive commitment. Harvard University Press.
Brandom, R. (2019). A Spirit of Trust: A Reading of Hegel’s phenomenology. Harvard University Press.
Burge, T. (2010). Origins of Objectivity. Oxford University Press.
Cao, R. (2012). A Teleosemantic Approach to Information in the brain. Biology and Philosophy, 27(1), 49–71.
Christias, D. (2023). Normativity, lifeworld and science in Sellars’ synoptic vision. Palgrave Macmillan.
Clark, A. (2016). Surfing uncertainty: Prediction, Action and the embodied mind. Oxford University Press.
Corti, L. (2023). Organizational normativity and teleology: A critique. Synthese, 202, 96. https://doi.org/10.1007/s11229-023-04255-7
Crane, T. (2009). Is perception a Propositional attitude? Philosophical Quarterly, 59, 452–469.
de Oliveira, G. S. (2018). Representationalism is a dead end. Synthese, 198(1), 209–235.
Di Paolo, E., Buhrmann, T., & Barandiaran, X. (2017). Sensorimotor Life: An enactive proposal. Oxford University Press.
Fodor, J. (1991). Fodor’s guide to Mental representation. In J. Greenwood (Ed.), The future of Folk psychology (pp. 22–50). Cambridge University Press.
Friston, K., Daunizeau, J., & Kiebel, S. (2010). Action and behavior: A free-energy formulation. Biological Cybernetics, 102(3), 227–260.
Gladziejewski, P., & Milkowski, M. (2017). Structural representations: Causally relevant and different from detectors. Biology and Philosophy, 32(3), 337–355.
Gładziejewski, P. (2016). Predictive coding and representation. Synthese, 193, 559–582.
Haugeland, J. (1998). Having thought: Essays in the metaphysics of mind. Harvard University Press.
Hohwy, J. (2013). The predictive mind. Oxford University Press.
Hohwy, J. (2020). New directions in Predictive Processing. Mind and Language, 35, 209–223.
Hutto, D. (2018). Getting into Predictive Processing’s great guessing game: Bootstrap Heaven or Hell? Synthese, 195, 2445–2458.
Hutto, D., & Myin, E. (2013). Radicalizing Enactivism: Basic minds without Content. MIT Press.
Hutto, D., & Myin, E. (2017). Evolving Enactivism: Basic minds Meet Content. MIT Press.
Mann, S. F., & Pain, R. (2022). Teleosemantics and the hard problem of content. Philosophical Psychology, 35(1), 22–46.
McDowell, J. (1987). De re senses. Philosophical Quarterly, 34, 325–363.
Millikan, R. (1984). Language, Thought and other Biological categories. MIT Press.
Moreno, A., & Mossio, M. (2015). Biological autonomy: A philosophical and theoretical Inquiry. Springer.
Neander, K. (2017). A Mark of the Mental: In defense of informational teleosemantics. MIT Press.
Piccinini, G. (2020). Neurocognitive mechanisms: Explaining Biological Cognition. Oxford University Press.
Piccinini, G. (2022). Situated neural representations: Solving the problems of content. Frontiers in Neurorobotics, 14:16, 846979. https://doi.org/10.3389/fnbot.2022.846979
Price, H. (2013). Expressivism, Pragmatism and Representationalism. Cambridge University Press.
Ramsey, W. M. (2007). Representation reconsidered. Cambridge University Press.
Rouse, J. (2015). Articulating the World: Conceptual understanding and the scientific image. Chicago University Press.
Sachs, C. (2019). In defense of picturing: Sellars’s philosophy of mind and cognitive neuroscience. Phenomenology and the Cognitive Sciences, 18(4), 669–689.
Salis, P. (2022). The given and the hard problem of content. Phenomenology and the Cognitive Sciences, 1–26.
Schlicht, T. (2018). Does separating intentionality from Mental representation Imply Radical Enactivism? Frontiers in Psychology, 9, 1497. https://doi.org/10.3389/fpsyg.2018.01497
Schlicht, T., & Starzak, T. (2021). Prospects of Enactivist approaches to intentionality and cognition. Synthese, 198(Suppl 1), 89–113.
Schulkin, J., & Sterling, P. (2019). Allostasis: A brain-centered Predictive Mode of physiological regulation. Trends in Neuroscience, 42(10), 740–752.
Segundo-Ortin, M., & Hutto, D. (2021). Similarity-based Cognition: Radical Enactivism meets cognitive neuroscience. Synthese, 198(Suppl 1), 5–23.
Sellars, W. (1960). Being and being known. Proceedings of the American Catholic Philosophical Association, 34, 28–49.
Sellars, W. (1963). Abstract entities. Review of Metaphysics, 16, 621–671.
Sellars, W. (1981). Mental events. Philosophical Studies, 39, 325–345.
Thompson, E. (2007). Mind in life: Biology, Phenomenology and the sciences of mind. Harvard University Press.
Travis, C. (2004). The silence of the senses. Mind, 113(449), 57–94.
Van Es, T., & Myin, E. (2021). Predictive Processing and representation: How less can be more. In D. Mendonca, M. Curado, & S. Gouveia (Eds.), The Philosophy and Science of Predictive Processing (pp. 7–24). Bloomsbury Academic.
van Fraassen, B. (2008). Scientific representation: Paradoxes of perspective. Oxford University Press.
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
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Christias, D. Contentless Representationalism? A Neglected Option Between Radical Enactivist and Predictive Processing Accounts of Representation. Minds & Machines 34, 1 (2024). https://doi.org/10.1007/s11023-024-09659-z
Received:
Accepted:
Published:
DOI: https://doi.org/10.1007/s11023-024-09659-z