Online research in philosophy

This paper argues that the form of explanation at issue in the hard problem of consciousness is scientifically irrelevant, despite appearances to the contrary. In particular, it is argued that the 'sense of understanding' that plays a critical role in the form of explanation implicated in the hard problem provides neither a necessary nor a sufficient condition on satisfactory scientific explanation. Considerations of the actual tools and methods available to scientists are used to make the case against it being a necessary condition, and work by J.D. Trout that exploits psychological research on the hindsight and overconfidence biases is used to show that it is not a sufficient condition. It is argued, however, that certain intellectual and moral concerns give us good reason to still try to meet the hard problem's explanatory challenge, despite its extrascientific nature.

The “explanatory gap” is proposed to be the “hard problem” of consciousness research and has generated a great deal of recent debate. Arguments brought forward to reveal this gap include the conceivability of zombies or the “super-neuroscientist” Mary. These are supposed to show that the facts of consciousness are not a priori entailed by the microphysical facts. Similar arguments were already proposed by emergence theories in the context of the debate between mechanism and vitalism. According to synchronic emergentism, the property of a system is emergent, when it cannot - in principle - be deduced from a complete description of the system’s components. Here, I argue that apart from phenomenal properties there are many other properties that, even though they are clearly physical, are not reductively explainable either. The explanatory gap of consciousness is therefore only a part of a much more general problem.

According to one productive and influential approach to cognition, categorization, object recognition, and higher level cognitive processes operate on a set of fixed features, which are the output of lower level perceptual processes. In many situations, however, it is the higher level cognitive process being executed that influences the lower level features that are created. Rather than viewing the repertoire of features as being fixed by low-level processes, we present a theory in which people create features to subserve the representation and categorization of objects. Two types of category learning should be distinguished. Fixed space category learning occurs when new categorizations are representable with the available feature set. Flexible space category learning occurs when new categorizations cannot be represented with the features available. Whether fixed or flexible, learning depends on the featural contrasts and similarities between the new category to be represented and the individual's existing concepts. Fixed feature approaches face one of two problems with tasks that call for new features: If the fixed features are fairly high level and directly useful for categorization, then they will not be flexible enough to represent all objects that might be relevant for a new task. If the fixed features are small, subsymbolic fragments (such as pixels), then regularities at the level of the functional features required to accomplish categorizations will not be captured by these primitives. We present evidence of flexible perceptual changes arising from category learning and theoretical arguments for the importance of this flexibility. We describe conditions that promote feature creation and argue against interpreting them in terms of fixed features. Finally, we discuss the implications of functional features for object categorization, conceptual development, chunking, constructive induction, and formal models of dimensionality reduction. Key Words: concept learning; conceptual development; features; perceptual learning; stimulus encoding.

Abstra,ct— This paper will discuss learning in hybrid models that goes beyond simple rule extraction from backpropagation networks. Although simple rule extraction has received a lot of research attention, to further develop hybrid learning models that include both symbolic and subsymbolic knowledge and that learn autonomously, it is necessary to study autonomous learning of both subsymbolic and symbolic knowledge in integrated architectures. This paper will describe knowledge extraction from neural reinforcement learning. It includes two approaches towards extracting plan knowledge: the extraction of explicit, symbolic rules from neural reinforcement learning, and the extraction of complete plans. This work points to the creation of a general framework for achieving the subsymbolic to symbolic transition in an integrated autonomous learning framework.

We present a theoretical account of implicit and explicit learning in terms of ACT-R, an integrated architecture of human cognition as a computational supplement to Dienes & Perner's conceptual analysis of knowledge. Explicit learning is explained in ACT-R by the acquisition of new symbolic knowledge, whereas implicit learning amounts to statistically adjusting subsymbolic quantities associated with that knowledge. We discuss the common foundation of a set of models that are able to explain data gathered in several signature paradigms of implicit learning.

This commentary is an elaboration on Schyns, Goldstone & Thibaut's proposal for flexible features in categorization in the light of three areas not explicitly discussed by the authors: connectionist models of categorization, computational learning theory, and constructivist theories of the mind. In general, the authors' proposal is strongly supported, paving the way for model extensions and for interesting novel cognitive research. Nor is the authors' proposal incompatible with theories positing some fixed set of features.

The article criticises the attempt to establish connectionism as an alternative theory of human cognitive architecture through the introduction of thesymbolic/subsymbolic distinction (Smolensky, 1988). The reasons for the introduction of this distinction are discussed and found to be unconvincing. It is shown that thebrittleness problem has been solved for a large class ofsymbolic learning systems, e.g. the class oftop-down induction of decision-trees (TDIDT) learning systems. Also, the process of articulating expert knowledge in rules seems quite practical for many important domains, including common sense knowledge.The article discusses several experimental comparisons betweenTDIDT systems and artificial neural networks using the error backpropagation algorithm (ANNs usingBP). The properties of one of theTDIDT systemsID3 (Quinlan, 1986a) are examined in detail. It is argued that the differences in performance betweenANNs usingBP andTDIDT systems reflect slightly different inductive biases but are not systematic; these differences do not support the view that symbolic and subsymbolic systems are fundamentally incompatible. It is concluded, that thesymbolic/subsymbolic distinction is spurious. It cannot establish connectionism as an alternative cognitive architecture.

According to “imaginability arguments,” given any explanation of the physiological correlates of consciousness, it remains imaginable that all elements of that explanation could occur without consciousness, which thus remains unexplained. The O'Brien & Opie connectionist approach effectively shows that perspicuous explanations can bridge this explanatory gap, but bringing in other issues – for example, involving biology and emotion – would facilitate going much further in this direction. A major problem is the ambiguity of the term “representation.” Bridging the gap requires perspicuously explaining not just how we form “representations” in the sense of outputs isomorphic to what is represented, but also what makes representations conscious; I sketch briefly what this would entail.