Churchland's work is in the school of
analytic philosophy in
western philosophy, with interests in
epistemology and the
philosophy of science, and specific principal interests in the
philosophy of mind and in
neurophilosophy and
artificial intelligence. His work has been described as being influenced by the work of
W. V. O. Quine,
Thomas Kuhn, Russell Hanson,
Wilfrid Sellars, and
Paul Feyerabend. Along with his wife, Churchland is a major proponent of
eliminative materialism, the belief that where by folk psychology is meant everyday mental concepts such as beliefs, feelings, and desires, which are viewed as theoretical constructs without coherent definition, and thus destined to be obviated by a scientific understanding of human nature. From the perspective of
Zawidzki, Churchland's concept of eliminativism is suggested as early as his book
Scientific Realism and the Plasticity of Mind (1979), with its most explicit formulation appearing in a
Journal of Philosophy essay, "Eliminative Materialism and the Propositional Attitudes" (1981). Churchland believes that beliefs are not ontologically real; that is, that a future, fully matured
neuroscience is likely to have no need for "beliefs" (see
propositional attitudes), in the same manner that modern science discarded such notions as legends or witchcraft. According to Churchland, such concepts will not merely be
reduced to more finely grained explanation and retained as useful proximate levels of description, but will be strictly
eliminated as wholly lacking in correspondence to precise objective phenomena, such as activation patterns across
neural networks. He points out that the history of science has seen many posits that were considered as real entities: such as
phlogiston; caloric; the
luminiferous ether; and
vital forces that were thus eliminated. Moreover, in
The Engine of Reason, The Seat of the Soul, Churchland suggests that consciousness might be explained in terms of a
recurrent neural network with its hub in the
intralaminar nucleus of the
thalamus, and feedback connections to all parts of the cortex. He acknowledges that this proposal will likely be found in error with regard to the neurological details, but states his belief that it is on the right track in its use of recurrent neural networks to account for consciousness. This has been described as a
reductionist rather than eliminativist account of consciousness. ==Personal life==