Is perception cognitively penetrable? (2017-)
University of Wollongong (Australia.) University Postgraduate Award, and International Postgraduate Tuition Award,
While perceive and cognising sometimes inter-relate, perceiving has special properties that motivate thinking of it in terms of a perception-cognition boundary. The duck-rabbit (Fig. 1) illustrates how perceiving and cognising can relate at some point – when the picture becomes a ‘duck’ or a ‘rabbit’ – even if, at an initial perceptual stage, or on a pre-linguistic phase, the picture is yet neither a duck, nor a rabbit. The Müller-Lyer (Fig. 2)[1] illustrates a sharper case of perception-cognition boundary. Remarkably, for people susceptible to the sort of this ‘illusion’, it seems impossible to bring themselves to see what they think or believe to be the case (that the lines are of the same length). More than popular well-studied lab visual cases, these pictures denote that the interaction with the perceptual space is permeated by opportunities for ‘illusion’. Specifying how and the conditions under which perceive and cognising can and cannot relate is the explanatory challenge in cognitive science.
University of Wollongong (Australia.) University Postgraduate Award, and International Postgraduate Tuition Award,
While perceive and cognising sometimes inter-relate, perceiving has special properties that motivate thinking of it in terms of a perception-cognition boundary. The duck-rabbit (Fig. 1) illustrates how perceiving and cognising can relate at some point – when the picture becomes a ‘duck’ or a ‘rabbit’ – even if, at an initial perceptual stage, or on a pre-linguistic phase, the picture is yet neither a duck, nor a rabbit. The Müller-Lyer (Fig. 2)[1] illustrates a sharper case of perception-cognition boundary. Remarkably, for people susceptible to the sort of this ‘illusion’, it seems impossible to bring themselves to see what they think or believe to be the case (that the lines are of the same length). More than popular well-studied lab visual cases, these pictures denote that the interaction with the perceptual space is permeated by opportunities for ‘illusion’. Specifying how and the conditions under which perceive and cognising can and cannot relate is the explanatory challenge in cognitive science.
Duck-rabbit
|
Müller-Lyer illusion
|
Modular theories have been well placed to address this challenge by positing a strong perception/cognition boundary. They seek to explain the recalcitrant perceiving in such cases by positing an informational boundary between perceiving and cognising. Perception, on traditional modular accounts, is thought of as informationally encapsulated from more general forms of cognition. A defining feature of modules is that they are sensitive only to domain-specific kind of information. That is, they only have access to the information contained in inputs plus whatever information is available withinthe module. Thus Fodor (1983) characterises the module's job as being “to interpret transduced information and to make it available to central processes” (p. 45). The flow of information does not go in the other direction. On such an account, perceptual processing is informationally encapsulated from background cognising.
Those who believe that perceiving is, at least sometimes informed by cognising, have supplied philosophical arguments over the years to establish that perceiving is cognitively penetrable after all. These arguments have put pressure on the view that perception is, in fact, informationally encapsulated. It is argued that there has to be an intelligible link between the contents of perceptual experience and the contents of cognition (Macpherson 2012). These philosophical arguments are putatively supported by empirical research in neuroscience. For example, cognitive penetrability is assumed in research on the binding problem. Those working on the binding problem seek to explain how the diverse elements of perception and cognition are integrated or interact to generate or manifest a single, unified perceptual experience. One way of explaining how such binding occurs is to assume that top-down cognising binds perceptual experience using predictions or anticipations that parametrise or force conformity in lower-levels. Claiming that perceiving is informed by cognising, these theories reject the thesis that there is a strong perception/cognition boundary.
Taking perceiving as a form of predicting, instead, puts pressure on the possibility that perception is strongly independent of other forms of cognition. The view that perceiving and other forms of cognition inform one another can be traced back to Kant’s analysis by synthesis (Zahavi 2018; Swanson 2016, Srinivasan et al. 1982; Rao and Ballard 1999; Huang and Rao 2011) and Helmholtz, whose ideas led to perception as hypothesis testing (Helmholtz 1860/1962; Hohwy, 2012, 2013; Clark, 2016). But it was Rao and Ballard (1999) who recently revived the idea of perception as prediction, in ways that promote perception as permeable, by offering a treatment of perceptual processing using the tools of predictive coding. The theory is still developing, but in its standard version, basically, backward connections flow from higher to lower order perceptual areas trying to predict activity in lower order areas, while the counter-stream of ascending, forward connections convey prediction errors, i.e., the discrepancy between higher and lower order levels. Error detected at lower-levels adjusts the predictions generated at higher-levels. On this account, background cognising can inform even the lowest levels of perceiving.
Predictive processing theory re-draws the modular perception-cognition boundary. Although predictive processing’s boundaries are more flexible by contingently allowing the exchange of contentful information between higher and lower levels, it ultimately rests upon modular principles. Predictive processing has its explanatory limitations. Essentially, adopting old modular boundaries brings about new epistemological challenges grounded on the assumption that content has an explanatory job in perception.
In the light of visual illusions, there are a few puzzles for purely cognitivist predictive processing. My research hypothesis is that the radical views of cognition, by abandoning content as the explanatory realiser in perception, offer a more powerful tool for explaining how perceiving and cognising inter-relate.
[1]Other visual illusions could be used to illustrate the phenomenon, such as Checker shadow illusion or Ebbinghaus Illusion.
Hilbert 24th problem (2016-)
Team member
Mathematics Department, Nova University of Lisbon, Foundation of Science and Technology funded project
- In 2000, a draft note of Hilbert was found concerning a 24th problem for his Paris problem list.
- This problem concerns simplicity of proofs.
- The aim of the project to reassess Hilbert’s 24th problem as a philosophical challenge (rather than a purely formal exercise).
- We take a closer look at the specifically mathematical examples that concerned Hilbert himself, in particular syzygies (the modern understanding of which is subsumed by the notion of Gröbner bases).
- Syzygies are, however, just a starting point for looking at Hilbert’s problem in other suitable mathematical contexts.
- Special emphasis is put on the potential impact of Hilbert’s 24th problem on contemporary philosophy of mathematics.
The Notion of Mathematical Proof” (2014-2015).
Grant Research Fellow,
Centre for Artificial Intelligence, Nova University of Lisbon, at the Science and Technology Foundation (FCT) funded Project
In this project, we review the current state of the notion of mathematical proof, with special emphasis on recent developments on probabilistic proofs and computer-assisted proofs. The principle goal is to obtain a better philosophical understanding of the nature of proofs in general, and computerized proofs in particular.
The Cognitive Foundations of the Self” (2013-2014).
Grant Research Fellow,
Institute for Philosophy of Nova, Nova University of Lisbon, at the Science and Technology Foundation (FCT) funded Project
These are exciting times in what the studies regarding the notion of the ‘Self’ are concerned. Nevertheless, the transdisplinary approach (phenomenology, cognitive neuroscience, psychiatry, psychoanalysis) comes with a cost: a taxonomical confusion and fragmentation inherent to the proliferation of such distinct methodological approaches. Philosophy, given its general, far reaching, synoptic and conceptual approach, is specially suited to overcome this unfortunate scenario by providing conceptual clarification that facilitates the establishment of links between disciplines. This is the aim of the current research proposal: to adopt a general methodology suited to solve the problem of the conceptual fragmentation of the notion of ‘Self’.
Project Fund: PTDC/FIL-FCI/110978/2009
Wittgenstein’s Philosophical Investigations: Re-evaluating a project” (2013)
Grant Research Fellow,
Institute for Philosophy of Nova, Nova University of Lisbon, at the Science and Technology Foundation (FCT) funded Project
The project aims to trace the development of Wittgenstein’s second book project laying the foundations for a principled view of the origin of the Philosophical Investigations. The idea is to yield an adequate view of what belongs in the Philosophical Investigations corpus, while at the same time showing what its status is. In this way, the connections between the philosophy of culture and mind and the philosophy of psychology, on one hand, and between these and the philosophy of mathematics, on the other, as spread among Wittgenstein’s various projects for the Philosophical Investigations, will be re-evaluated. The majority of these sources have been available since 2000 in the Bergen Electronic Edition of Wittgenstein’s Nachlass, but we are still in need of a thorough examination of them. In addition, new items have come to light in the meantime which justify a close study. It is such a philological-philosophical study that the project aims to carry out, analysing the genesis of Wittgenstein’s Philosophical Investigations and also editing hitherto unpublished texts. Our understanding of the Philosophical Investigations project and its relation to other Nachlass parts and periods will be significantly improved, and this will make it possible to establish its potential for continued influence on 21st century thought in ways that are anchored to a secure text basis. This objective will be achieved in close cooperation with the Wittgenstein Archives at the University of Bergen (WAB).
