Computational Theories

Computational theories of mind developed organically as the processing power of computers expanded exponentially to enable the emulation of mind-like capabilities such as memory, knowledge structure, perception, decision-making, problem solving, reasoning and linguistic comprehension (especially with the advent of human-like large language models like ChatGPT). The growing field of cognitive science owes its development to computational theories (Rescorla, 2020).

There is a reciprocal, recursive, positive-feedback relationship as computational theories of mind seek both to enhance the power and scope of computing and to advance understanding of how the human mind actually works. Classical computational theories of mind, which exemplify functionalism, are based on algorithms, which are routines of systematic, step-by-step instructions, and on Turing machines, which are abstract models of idealized computers with unlimited memory and time that process one operation at a time (with super-fast but not unlimited speed).

Classical Computationalism

Artificial intelligence adds logic, seeking to automate reasoning—deductive at first, then inductive and higher-order forms. Neural networks, with a connectionism construct, were a step-function advance. For example, chess computers have reigned supreme since 1997 when Deep Blue defeated the world chess champion, Garry Kasparov. But whereas the process has been literally massive brute-force calculations—hundreds of millions of “nodes” per second (a “node” is a chess position with its evaluation and history)—recent advances in algorithmic theory are dramatically improving capabilities. The implications go way beyond chess and are apparent.

Simulation and Modeling Theories

Philosopher-futurist Nick Bostrom espouses a computational theory of consciousness, which is consistent with his view that there is a distinct possibility that our world and universe, our total state of affairs, is a computer simulation (Bostrom, 2003, 2006). The logic is almost a tautology: A computer simulation would require, by definition, that our consciousness, and the consciousnesses of all sentient creatures, would be, ipso facto, computational consciousness.

Of course, Bostrom does not argue that we are living in a simulation, so his computationalism as a theory of consciousness is motivated by other factors, including computational neuroscience. In fact, one could make the case that the arrow of causal explanation points in the reverse direction: Consciousness as computational would need to be a condition precedent, necessary but not sufficient, for the simulation argument to be coherent.

Computer/AI scientist James Reggia explains that efforts to create computational models of consciousness have been driven by two main motivations: “to develop a better scientific understanding of the nature of human/animal consciousness and to produce machines that genuinely exhibit conscious awareness.” He offers three conclusions: “(1) computational modeling has become an effective and accepted methodology for the scientific study of consciousness; (2) existing computational models have successfully captured a number of neurobiological, cognitive, and behavioral correlates of conscious information processing as machine simulations; and (3) no existing approach to artificial consciousness has presented a compelling demonstration of phenomenal machine consciousness, or even clear evidence that artificial phenomenal consciousness will eventually be possible” (Reggia, 2013).

Critiques and Limits of Computationalism

Computer scientist Kenneth Steiglitz argues that all available theories of consciousness “aren't up to the job” in that “they don't tell me how I can know whether a particular candidate is or is not phenomenally conscious.” Moreover, he says, we will never be able to answer the question of AI consciousness—because “it is simply not possible to test for consciousness.”

This presents, Steiglitz worries, dangers of two kinds: (1) damaging or even destroying our own consciousness, and (2) bringing about new consciousness that will not be treated with proper respect and quite possibly suffer (Steiglitz, 2024).

Steiglitz states three principles of what we think we know about consciousness—the dual nature of mind and body, the dependence of mind on body, and the dependence of mind on computation—and he calls them all absurd, because “these do not follow from physics, biology, or logic.” He muses, “I wish I had a theory to account for consciousness—but I don't see how any theory could” (Steiglitz, 2024).

Philosophy-savvy attorney Andrew Hartford proposes an EP (Eternal Past) Conjecture such that “If there ever is something there always was something, because no-thing comes from Nothing,” and that “the always existor exists before all time, process or computation.” What follows, he says, is that while “it remains to be seen whether artificial consciousness is in the domain of all possibilities, we should not presume that we will necessarily build computational consciousness” (Hartford, 2014).

The mildly dismissive critique is that the computational theory of mind follows the historical trend of analogizing the mind to “the science of the day.”^[1]

Footnote

[1]. Closer To Truth videos on Computational Theory of Mind, including Rodney Brooks, Andy Clark, Donald Hoffman, Susan Greenfield, Peter Tse, Anirban Bandyopadhyay, Ken Mogi—https://closertotruth.com/video/broro-003/?referrer=8107.