I have lately come to seriously consider the view that the everyday notion of consciousness doesn’t refer to anything that exists out there in the world but is rather a confused (but useful) projection made by purely physical minds onto their depiction of themselves in the world. The main influences on my thinking are Dan Dennett, (I assume most of you are familiar with him) and to a lesser extent Yudkowsky (1) and Tomasik (2). To use Dennett’s line of thought: we say that honey is sweet, that metal is solid or that a falling tree makes a sound, but the character tag of sweetness and sounds is not in the world but in the brains internal model of it. Sweetness in not an inherent property of the glucose molecule, instead, we are wired by evolution to perceive it as sweet to reward us for calorie intake in our ancestral environment, and there is neither any need for non-physical sweetness-juice in the brain – no, it's coded (3). We can talk about sweetness and sound as if being out there in the world but in reality it is a useful fiction of sorts that we are "projecting" out into the world. The default model of our surroundings and ourselves we use in our daily lives (the manifest image, or ’umwelt’) is puzzling to reconcile with the scientific perspective of gluons and quarks. We can use this insight to look critically on how we perceive a very familiar part of the world: ourselves. It might be that we are projecting useful fictions onto our model of ourselves as well. Our normal perception of consciousness is perhaps like the sweetness of honey, something we think exist in the world, when it is in fact a judgement about the world made (unconsciously) by the mind.
What we are pointing at with the judgement “I am conscious” is perhaps the competence that we have to access states about the world, form expectations about those states and judge their value to us, coded in by evolution. That is, under this view, equivalent with saying that suger is made of glucose molecules, not sweetness-magic. In everyday language we can talk about suger as sweet and consciousness as “something-to-be-like-ness“ or “having qualia”, which is useful and probably necessary for us to function, but that is a somewhat misleading projection made by our world-accessing and assessing consciousness that really exists in the world. That notion of consciousness is not subject to the Hard Problem, it may not be an easy problem to figure out how consciousness works, but it does not appear impossible to explain it scientifically as pure matter like anything else in the natural world, at least in theory. I’m pretty confident that we will solve consciousness, if we by consciousness mean the competence of a biological system to access states about the world, make judgements and form expectations. That is however not what most people mean when they say consciousness. Just like ”real” magic refers to the magic that isn’t real and the magic that is real, that can be performed in the world, is not “real magic”, “real” consciousness turns out to be a useful, but misleading assessment (4). We should perhaps keep the word consciousness but adjust what we mean when we use it, for diplomacy.
Having said that, I still find myself baffled by the idea that I might not be conscious in the way I’ve found completely obvious before. Consciousness seems so mysterious and unanswerable, so it’s not surprising then that the explanation provided by physicalists like Dennett isn’t the most satisfying. Despite that, I think it’s the best explanation I've found so far, so I’m trying to cope with it the best I can. One of the problems I’ve had with the idea is how it has required me to rethink my views on ethics. I sympathize with moral realism, the view that there exist moral facts, by pointing to the strong intuition that suffering seems universally bad, and well-being seems universally good. Nobody wants to suffer agonizing pain, everyone wants beatific eudaimonia, and it doesn't feel like an arbitrary choice to care about the realization of these preferences in all sentience to a high degree, instead of any other possible goal like paperclip maximization. It appeared to me to be an unescapable fact about the universe that agonizing pain really is bad (ought to be prevented), that intelligent bliss really is good (ought to be pursued), like a label to distinguish wavelength of light in the brain really is red, and that you can build up moral values from there. I have a strong gut feeling that the well-being of sentience matters, and the more capacity a creature has of receiving pain and pleasure the more weight it is given, say a gradience from beetles to posthumans that could perhaps be understood by further inquiry of the brain (5). However, if it turns out that pain and pleasure isn’t more than convincing judgements by a biological computer network in my head, no different in kind to any other computation or judgement, the sense of seriousness and urgency of suffering appears to fade away. Recently, I’ve loosened up a bit to accept a weaker grounding for morality: I still think that my own well-being matter, and I would be inconsistent if I didn’t think the same about other collections of atoms that appears functionally similar to ’me’, who also claim, or appear, to care about their well-being. I can’t answer why I should care about my own well-being though, I just have to. Speaking of 'me': personal identity also looks very different (nonexistent?) under physicalism, than in the everyday manifest image (6).
Another difficulty I confront is why e.g. colors and sounds looks and sounds the way they do or why they have any quality at all, under this explanation. Where do they come from if they’re only labels my brain uses to distinguish inputs from the senses? Where does the yellowness of yellow come? Maybe it’s not a sensible question, but only the murmuring of a confused primate. Then again, where does anything come from? If we can learn to shut up our bafflement about consciousness and sensibly reduce it down to physics – fair enough, but where does physics come from? That mystery remains, and that will possibly always be out of reach, at least probably before advanced superintelligent philosophers. For now, understanding how a physical computational system represents the world, creates judgements and expectations from perception presents enough of a challenge. It seems to be a good starting point to explore anyway (7).
I did not really put forth any particularly new ideas here, this is just some of my thoughts and repetitions of what I have read and heard others say, so I'm not sure if this post adds any value. My hope is that someone will at least find some of my references useful, and that it can provide a starting point for discussion. Take into account that this is my first post here, I am very grateful to receive input and criticism! :-)
- Check out Eliezer's hilarious tear down of philosophical zombies if you haven't already
- [Video] TED talk by Dan Dennett http://www.ted.com/talks/dan_dennett_cute_sexy_sweet_funny
- Reading “The Moral Landscape” by Sam Harris increased my confidence in moral realism. Whether moral realism is true of false can obviously have implications for approaches to the value learning problem in AI alignment, and for the factual accuracy of the orthogonality thesis
- For anyone interested in getting a grasp of this scientific challenge I strongly recommend the book “A User’s Guide to Thought and Meaning” by Ray Jackendoff.
Edit: made some minor changes and corrections. Edit 2: made additional changes in the first paragraph for increased readability.