Open peer commentary on the article “Missing Colors: The Enactivist Approach to Perception” by Adrián G. Palacios, María-José Escobar & Esteban Céspedes. Upshot: Palacios, Escobar and Céspedes consider misrepresentation and comparability in the context of the enactivist approach of colour perception. This consideration leads them to propose the introduction of a weak form of representationalism to account for internal representation of “reality” and “shared experience” and to accommodate the Bayesian principle of prior information used in machine vision. The weak representationalism is not limited to brain states but may include embodied factors to be compatible with the enactivist framework. My commentary will essentially consider the misrepresentation and comparability arguments used by the authors to introduce the notion of representation.
Hutto D. D. (2009) Mental representation and consciousness. In: Banks W. P. (ed.) Encyclopedia of consciousness. Volume 2. Academic Press, New York: 19–32.
Intentionality and consciousness are the fundamental kinds of mental phenomena. Although they are widely regarded as being entirely distinct some philosophers conjecture that they are intimately related. Prominently it has been claimed that consciousness can be best understood in terms of representational facts or properties. Representationalist theories vary in strength. At their core they seek to establish that subjective, phenomenal consciousness (of the kind that involves the having of first-personal points of view or perspectives on the world – perspectives that incorporate experiences with specific phenomenal characters) is either exhausted by, or supervenes on, capacities for mental representation. These proposals face several serious objections.
In this paper, we consider how certain longstanding philosophical questions about mental representation may be answered on the assumption that cognitive and perceptual systems implement hierarchical generative models, such as those discussed within the prediction error minimization (PEM) framework. We build on existing treatments of representation via structural resemblance, such as those in Gładziejewski (Synthese 193(2):559–582, 2016) and Gładziejewski and Miłkowski (Biol Philos, 2017), to argue for a representationalist interpretation of the PEM framework. We further motivate the proposed approach to content by arguing that it is consistent with approaches implicit in theories of unsupervised learning in neural networks. In the course of this discussion, we argue that the structural representation proposal, properly understood, has more in common with functional-role than with causal/informational or teleosemantic theories. In the remainder of the paper, we describe the PEM framework for approximate Bayesian inference in some detail, and discuss how structural representations might arise within the proposed Bayesian hierarchies. After explicating the notion of variational inference, we define a subjectively accessible measure of misrepresentation for hierarchical Bayesian networks by appeal to the Kullbach–Leibler divergence between posterior generative and approximate recognition densities, and discuss a related measure of objective misrepresentation in terms of correspondence with the facts.
This paper starts by considering an argument for thinking that predictive processing (PP) is representational. This argument suggests that the Kullback–Leibler (KL)-divergence provides an accessible measure of misrepresentation, and therefore, a measure of representational content in hierarchical Bayesian inference. The paper then argues that while the KL-divergence is a measure of information, it does not establish a sufficient measure of representational content. We argue that this follows from the fact that the KL-divergence is a measure of relative entropy, which can be shown to be the same as covariance (through a set of additional steps). It is well known that facts about covariance do not entail facts about representational content. So there is no reason to think that the KL-divergence is a measure of (mis-)representational content. This paper thus provides an enactive, non-representational account of Bayesian belief optimisation in hierarchical PP.
This article offers a critique of constructivism in science education which is an attempt to define and identify not only the weaknesses, but the successes of constructivism. Its success has been to generate a significant body of empirical data which has contributed to our knowledge and understanding of difficulties in the learning of science. That knowledge has also enabled the development of some innovative methodology for the pedagogy of science and a greater awareness of the learner. However, as a theoretical referent, it suffers from a flawed instrumental epistemology which is a misrepresentation of the views and practice of science and scientists. Further it has confused the manner in which new knowledge is made with the manner in which old knowledge is learned, assuming that the two are one and the same thing. The result of these failings is that it offers no guidance on adjudication between theories, the organization and sequencing of content within the science curriculum and rejects any value for didacticism. The failure of constructivism to recognize its own limitations has lead to it enjoying a hegemony in the research community which is undeserved. Instead, it is proposed that an alternative of modest realism offers, not only a better representation of the practice of science, but additionally some value in determining issues of pedagogy.
Palacios A. G., Escobar M.-J. & Céspedes E. (2017) Authors’ Response: Is a Weak Notion of Representation not Compatible with a Contextualist and Enactivist Account of Perception? Constructivist Foundations 13(1): 135–140. Fulltext at https://cepa.info/4418
Upshot: We argue that the notion of basic perception could help to develop a general enactivist account of perception, without compromising the compatibility between our approach to this theory and the notion of weak representation. To support this, we elaborate on the contextual and normative aspects of our enactivist proposal, on perception, and on how these aspects may be crucial for understanding misrepresentation and comparability.
Palacios A. G., Escobar M.-J. & Céspedes E. (2017) Missing Colors: The Enactivist Approach to Perception. Constructivist Foundations 13(1): 117–125. Fulltext at https://cepa.info/4412
Context: Part of Varela’s work focused on the study of visual perception, particularly on the grounds of an enactivist theory of vision. Problem: Varela held that the problem of misrepresentation and the comparability of visual experience were crucial. We live with other creatures in sensory worlds that are not tractable, so could we share color-similar experiences? We are still missing an integrative enactive framework to tackle the problems of misrepresentation and comparability related to animal color experience. Method: We carried out a literature survey to draw attention to the status of the enactivist theory of vision and to explore how the problems of misrepresentation and comparability may be tackled. Results: As shown, philosophy and computational science have recently incorporated concepts from neurobiology that close gaps between disciplines and support aspects of the enactivist approach of vision. Implications: Epistemological problems related to perception are here tackled, considering some controversial assumptions related to vision. We argue that an enactivist theory of visual perception may not only clarify the problematic consequences of those assumptions, but also fruitfully guide future philosophical and empirical research on this topic. Constructivist content: The presence of singular “visual channels”, as well as physical, sensorimotor and evolutionary factors, constrains our own perceptual experience as proposed by enactivism.