Unempirical Theories Do Not Survive
In the eighteenth and nineteenth centuries, biologists bitterly debated over the existence of an immaterial substance called the “élan vital.” Drawing from a philosophical tradition at least as old as writing,1 the vitalists argued that the existence of the élan vital was necessary to explain the existence of life. It was obvious to them that life could not arise from the merely artificial: there had to be something more at play. On the other hand, the mechanistic camp believed that life could be explained purely in terms of physical mechanisms. Today, this question is no longer seriously debated by biologists; the expert consensus is solidly against the vitalists.
Why was vitalism rejected? The theory was unfalsifiable,2 so it was not felled by an accumulation of opposing evidence as most theories are. Instead, as the underlying phenomenon became less and less mysterious, the idea of the élan vital became less and less appealing. Why invoke an invisible, ineffable substance when you can observe first-hand mechanisms that explain and predict how life behaves?
Consider another example: teleological theories of physics, which explain phenomena in terms of their supposed purposes (“teloi”) rather than their mechanisms. Aristotle believed that heavy objects fall because the ground is their “natural place.” Early biologists believed that species evolved according to an overarching goal, rather than via natural selection. Today, virtually no experts use these kinds of explanations as anything but a tool to aid comprehension. Attributing goals to phenomena can make them more intuitive, but these intuitions are ultimately flawed as proven by Newton or Darwin in the above examples.
TODO: I should probably replace this whole section tbh
Consciousness is Unempirical
And so to those of you who may be vitalists I would make this prophecy: what everyone believed yesterday, and you believe today, only cranks will believe tomorrow.
Francis Crick, discoverer of the structure of DNA
Today’s conception of consciousness will likely seem to future humans as telos or the élan vital seem to us today — a somewhat humorous, but ultimately understandable mistake born of bad intuitions.
Why? Because there is no empirical content to any of the discourse on consciousness. Any proposed test for consciousness is rejected as soon as it disagrees with intuition. For example, before the advent of LLMs, passing a rigorous Turing test was considered by many as sufficient evidence for consciousness. Now that this definition conflicts with the common intuition that LLMs are not conscious, though, this test has been rejected. Similarly, medical tests like the ability to recognize oneself in a mirror are rejected because they fail to match the intuition that young children and mentally disabled adults are conscious. Nobody can agree on a testable definition, because the concept itself is not testable.
Because they lack any empirical basis, statements like “it is possible for an artificial intelligence to be conscious” are ultimately meaningless despite being the subject of intense debate. One might as well discuss whether it is possible to synthesize the élan vital.
A note: the above argument applies only to the question of others’ consciousness, not to the question of one’s own consciousness, since that is observable. The very act of questioning one’s own consciousness seems to imply the existence of a mind performing the questioning. “Dubito, ergo sum,” as Descartes put it.
Mysteries of Consciousness
This does not, however mean, that consciousness will remain mysterious. The nonexistence of the élan vital has not prevented us from understanding life. Gravity is almost perfectly understood despite a distinct lack of natural places. Most of the mysteries we associate with consciousness today will be explained in the future by empirical research into the nature of the mind.
For example:
- How do humans model themselves and how does this influence our reported experiences?
- How do we model others, and how exactly is this related to the self-model?
- How are the many streams of sensory information with latencies ranging over hundreds of milliseconds integrated to create a single perceived “now”?
- Why do we feel that we have free will?
All of these are questions that, today, are often bundled together with philosophical discussions of consciousness. However, in the future, these phenomena will be demystified empirically, just as other natural sciences like cosmology or chemistry have grown apart from the philosophical and mystical over the course of history.