Igor Mikhailov

The paper is dedicated to particular cases of interaction and mutual impact of philosophy and cognitive science. Thus, philosophical preconditions in the middle of the 20th century shaped the newly born cognitive science as mainly based on conceptual and propositional representations and syntactical inference. Further developments towards neural networks and statistical representations did not change the prejudice much: many still believe that network models must be complemented with some extra tools that would account for proper human cognitive traits. I address some real implemented connectionist models that show how ‘new associationism ’ of the neural network approach may not only surpass Humean limitations, but, as well, realistically explain abstraction, inference and prediction. Then I stay on Predictive Processing theories in a little more detail to demonstrate that sophisticated statistical tools applied to a biologically realist ontology may not only provide solutions to scientific problems or integrate different cognitive paradigms but propose some philosophical insights either. To conclude, I touch on a certain parallelism of Predictive Processing and philosophical inferentialism as presented by Robert Brandom.

According to some critics, if biology is a kind of reverse engineering for the nature, it is quite poorly prepared for the task. Thus, the issue is more likely with its ontology. Multiple hypotheses and conjectures found in papers on methodological issues claim that living systems should be viewed as complex networks of signal-transmitting paths, both neural and non-neural, that feature modularity and feedback circuits and are prone to emergent properties and increasing complexity. If so, we are on the eve of a new stage in computer models development where not only computers are used to emulate life, but life itself is construed as a complex network of interacting natural computers.

The article covers such issues as the relevance of the theory of perception as a multi-level information processing, the methodological role of the concept of representation and the relation of neurodynamic structures to subjective experience. The author critically reviews the philosophical presumptions underlying the various concepts of “local rationality,” the core of which is constituted by the belief that large ethnic cultures generate or are based on their own rationality and their own logic. Three statements are successively considered: thinking is based on the inherent rules of rationality, logic is an extract of rationality, and types of rationality are geographically consistent with large ethnocultural communities. Empirical arguments are presented that demonstrate equivocality of the first two theses. In particular, firstly, it is shown that the genuine rationality of thinking lies not in following the rules that are immanent to it, but in the development of thinking and, more broadly, cognitive operations towards optimization of certain indicators of the cognitive or motor system that are important for survival and adaptation. In addition, this rationality is multivariate, and the choice between variants is often weakly determined or even random. Thus, the first statement turns out to be refuted. Secondly, by reference to the well-known experiments, it is shown that most people do not explicitly follow some declared logical rules in solving even logical or mathematical problems, and yet there is reason to consider their behavior rational. The third thesis, as shown with some limited empirical material, appears to be partially confirmed. Nevertheless, the demonstration of the doubtfulness of the first two theses makes the conclusion that different nations have different logics insufficiently substantiated.

Cognitive research can contribute to the formal epistemological study of knowledge representation inasmuch as, firstly, it may be regarded as a descriptive science of the very same subject as that, of which formal epistemology is a normative one. And, secondly, the notion of representation plays a constitutive role in both disciplines, though differing therein in shades of its meaning. Representation, in my view, makes sense only being paired with computation. A process may be viewed as computational if it adheres to some algorithm and is substrate-independent. Traditionally, psychology is not directly determined by neuroscience, sticking to functional or dynamical analyses in the what-level and skipping mechanistic explanations in the how-level. Therefore, any version of computational approach in psychology is a very promising move in connecting the two scientific realms. On the other hand, the digital and linear computational approach of the classical cognitive science is of little help in this way, as it is not biologically realistic. Thus, what is needed there on the methodological level, is a shift from classical Turing-style computationalism to a generic computational theory that would comprehend the complicated architecture of neuronal computations. To this end, the cutting-edge cognitive neuroscience is in need of а satisfactory mathematical theory applicable to natural, particularly neuronal, computations. Computational systems may be construed as natural or artificial devices that use some physical processes on their lower levels as atomic operations for algorithmic processes on their higher levels. A cognitive system is a multi-level mechanism, in which linguistic, visual and other processors are built on numerous levels of more elementary operations, which ultimately boil down to atomic neural spikes. The hypothesis defended in this paper is that knowledge derives not only from an individual computational device, such as a brain, but also from the social communication system that, in its turn, may be presented as a kind of supercomputer of the parallel network architecture. Therefore, a plausible account of knowledge production and exchange must base on some mathematical theory of social computations, along with that of natural, particularly neuronal, ones.