If a mind could be perfectly simulated, would the simulation actually be conscious?

[Craig90]

Imagine a computer simulation that reproduces every single neuron, every synaptic connection, and every relevant chemical process in a human brain down to whatever level of physical detail turns out to actually matter for producing a mind. The functionalist position in philosophy of mind says yes, that simulated system would be genuinely conscious, because what matters for consciousness is the pattern of information processing and its functional relationships, not the specific physical substrate carrying it out. On this view, consciousness is substrate independent the same way a piece of software is substrate independent, running identically whether it executes on one specific chip or a completely different one built from different materials.

Biological naturalists disagree sharply, arguing that consciousness is not just a pattern but a specific product of the particular biochemistry of living neurons, the same way digestion is not just an abstract pattern of breaking down food that could equally run on any substrate, but specifically depends on the actual biological, chemical processes involved. On this view, a perfect simulation of digestion running on a computer would not actually digest anything, no matter how detailed and accurate the simulation was, and consciousness might work the exact same way, remaining tied irreducibly to biology rather than to pattern alone.

The thought experiment gets sharper once you imagine a gradual replacement scenario, swapping out one biological neuron at a time for a functionally identical artificial one that produces exactly the same outputs given the same inputs. If consciousness fades gradually somewhere partway through that replacement process, that would seem to support the biological naturalist position. If instead the person reports feeling completely unchanged all the way through, with nothing ever noticeably different from the inside, that outcome would seem to strongly favor functionalism instead.

This is not a purely academic puzzle either. It bears directly on whether uploading a mind to a computer would actually preserve the original person's consciousness or would instead just create a separate, merely convincing copy, and on whether sufficiently advanced AI systems built entirely from silicon rather than biological neurons could ever possibly be genuinely conscious at all, or whether they are permanently locked out of experience no matter how sophisticated their behavior eventually becomes.