Introduction

The question of how consciousness arises and what, if any, effect it has on our behaviour is clearly both fascinating intellectually and of great practical and ethical significance^[1]. One very common view, perhaps the prevailing view, on how consciousness arises seems to be that consciousness is evoked whenever a physical system carries out a suitable computation. I don't know if this position has a name so I will call it the computational view of consciousness for the rest of the post. I am in no way a philosopher nor am I familiar with the great deal of thought that people have no doubt put into these issues but the computational view of consciousness seems to me to be quite a strange position to hold. Therefore, I would be very interested to learn about what others have had to say on the lines of argument I will describe here against this view.

The Causal Influence of Mental Events

My first reason for being skeptical of this view is that it seems to me to be at odds with the idea that mental events can have a causal influence on the physical world. That our behaviour is influenced by mental events seems to me hard to deny. Surely when I talk about being conscious, the fact that I am conscious is at least part of the cause of my talking about it. How could we have any conception of "consciousness" but for our direct experience of it? However, this seems to sit poorly with the claim that consciousness is only evoked by computations being carried out. After all, the operation of a computer can be explained in terms of logic gates, which can be understood in terms of simple mechanical laws, apparently leaving no place for any mental states evoked by the computer to influence the computer in turn^[2].

I have tried to formulate counter-arguments to the above but haven't yet succeeded in finding anything I find particularly convincing. The claim that mental events can have a causal influence on the physical world seems particularly hard to dispute. It just doesn't seem plausible to me that our brains could "know" that they had brought about mental events without being influenced by them somehow. However, the idea that the computational view of consciousness is incompatible with mental events influencing the physical world seems more open to attack. Perhaps we could argue that the brain carries out computations differently to a digital computer such that its operation can't be explained in terms of pieces governed by mechanical laws, like logic gates, so there is room for the influence of the mental events evoked by the brain's computational action to form an important part of the brain's operation in turn. However, by introducing a fundamental difference between the causal role played by mental events in the brain and the role they play in a digital computer, we seem to have at least violated the spirit of the computational view of consciousness and it would seem quite strange that brains were influenced by the mental events they bring about but digital computers weren't.

How Is "Carrying Out" a Computation Defined?

The other concern I have with the computational view of consciousness is that it's not clear to me that we can objectively say which physical systems are carrying out computations or which computations they are carrying out. When we say that a computer is performing a particular computation, is that not merely a matter of perspective? Imagine a physical system consisting of a clock and some number of beans. The only thing that changes as the system evolves in time is the number on the clock, which counts the number of seconds that have passed since some initial time. Then, given a Turing machine, $M$, could we not simply identify the state of the physical system with $n$ beans and $t$ seconds displayed on the clock with the state that $M$ would be in after $t$ steps if it was given the $n$th allowed input and then claim that the system was a physical instantiation of $M$ because its dynamical evolution matched the evolution of the Turing machine's computation? It would therefore seem that, depending on our perspective, this very simple system could be seen as an implementation of whatever Turing machine we liked. Is it possible to come up with a sensible definition of what it means for a physical system to carry out a computation such that silly arguments like this don't work?

Questions

I have only begun to think about these issues so I would be keen to hear what people who have thought about these questions much more thoroughly have had to say. How do people who adhere to the computational view of consciousness respond to these kinds of arguments? Are they generally thought of as serious challenges?

Also, it would seem that nothing in our current understanding of the laws of physics that would predict the occurrence of mental events, let alone their having a causal influence on the physical world. Therefore, if mental events can indeed exert a causal influence on the physical world, it would seem that our current physical theories fail to predict not just mental phenomena but also physical phenomena that depend on the influence of mental events. Is there any consensus on whether this should be seen as an inadequacy of our current theories of physics and whether investigating the laws that govern such phenomena might be a promising route to better understand consciousness?

1. ^{^}

   For instance, the question of whether we might create conscious entities via computational simulation and what their experiences would be like seems very important ethically. These questions also seem very relevant to the capabilities future AIs might have.

2. ^{^}

   I suppose you might claim that those mechanical laws subtly depend on the influence of mental events. However, given that you could, for example, build a NAND gate out of dominoes and account for its operation in terms of gravity and Newton's laws, you would have to draw the somewhat far-fetched conclusion that basic mechanical laws such as gravity or Newton's laws depend on the influence of mental events.