Tuesday, September 2, 2014

Consciousness and the Fallacy of Composition

I wrote some time ago that I consider the Just World Fallacy to be perhaps the most pernicious fallacy there is, but lately it seems as if the Fallacy of Composition is popping up in discussions everywhere. Its influence is certainly not as destructive socially and politically, but it seems awfully widespread, and it seem to easily confound the thinking of otherwise smart and reasonable people.

What is the Fallacy of Composition? It is the mistake of concluding that the whole must have some property that its parts have individually or, perhaps more relevantly for present purposes, that the whole cannot have any properties beyond those of its parts. For example, one would be mistaken to conclude that birds cannot fly from the fact that individual feathers or leg bones are incapable of powered flight. So far, so obvious.

Where I increasingly notice the effect of this fallacy is in discussions of mind, consciousness and artificial intelligence. The funny thing about it is that the same mistake appears to be made by people on the most extreme ends of the spectrum, that is mind-body dualists and religious believers on one side and straight-laced rationalists and physicalists on the other side.

Perhaps the best known is the least sophisticated case, that of a believer arguing as follows: we are made of atoms/meat, atoms/meat alone cannot feel, think and reason, therefore there must be something magic about our ability to do so. Ergo, there must be supernatural souls and the god(s) who bestowed them upon us. Of course, this argument is problematic in several interestingly distinct ways - one might mention God of the Gaps-thinking, for example - but the Fallacy of Composition is surely part of it.

Pretty much the same argument is made in slightly more intellectual terms when the nature of consciousness is discussed by philosophers who remain sceptical of physicalism or perhaps even only certain brands of reductionism. Common key words here are for example the Chinese Room, philosophical zombies or the Swamp Man. This may be simplifying a bit and perhaps not fully represent the original intentions of the people who came up with these thought experiments, but in many discussions of human consciousness and the feasibility of artificial intelligence the argument goes like this: Imagine some entity that is completely indistinguishable from a standard human in its behaviour but is merely the material meat-human (or, in the case of the Chinese Room, an apparatus following rules), without some additional quality X that is really important, e.g. true understanding, true recognition or true consciousness. From this we must conclude that there is some (magical?) extra that a real human has beyond being the material meat-human.

Again, it should be immediately obvious that this argumentation is problematic on several levels, perhaps most evidently in being a case of circular reasoning. However, the Fallacy of Composition plays its role in supplying the intuition that a human only made of meat cannot possibly be conscious because a piece of meat isn't conscious, or that a complex machine cannot possibly have understanding because its parts don't have understanding. The thought experiments are set up to invite the reader to fall prey to precisely this fallacy.

Strangely enough, nobody who argues in this way against, say, the possibility of computers gaining consciousness, would make the same argument about other properties of computers. The machine currently in front of me can display fantastic pictures, play enticing music, allow me to communicate with distant people, and with a bit of luck even beat me at strategy games. Looking inside of it when it is turned off, that may seem like magic. Indeed my ancestors of 5,000 years ago would surely have made the same argument as a dualist philosopher would make about my own brain. We have never so far found a magical extra ingredient to be a good explanation for anything, so why for consciousness?

But as mentioned above, for some weird reason the same error of reasoning appears to be found on the other side of the divide. When discussing compatibilist Free Will, one finds fellow physicalists writing that "we don't make decisions, our neurons make them for us", or "we are the slaves of our genes". But who are "we" if not precisely our neurons and our genes? And now I read that "our egos are illusions, diffuse products of brain activity". Well, there is one brain in my body; why is its activity to be considered more "diffuse" than the activity of my car, for example, which is also made up of parts interacting with each other to produce motive force?

So some dualists and some physicalists may potentially have an equally hard time of reconciling the cognitive with the physical. The difference is only that where the dualist concludes that because our minds/consciousness/selves exist, they must obviously be magical, the physicalist instead concludes that they cannot really exist. The true solution to the apparent conundrum might lie not in some nebulous middle but in the realisation that a whole may be able to do things that its individual parts cannot, such as "make up a self", "make decisions" or "be conscious".


  1. Fair enough, but what happens once the whole (the body) has made up a self? If I was such a self, I'd believe that I (the self) existed, and that I had a body, rather than that the body had me. And it's hard for the self to decide this, without being conscious at the time. But it's nicer to call the self 'me', and several selves 'us'.

    If you take a monist (or physicalist) view of the universe, why divide it up in particular into the bodies that you equate with 'us'? To regard some part or other as being anything demands an observer. So when we say "we", we don't mean our bodies so much as the part of us that observes ourselves. It's impossible to avoid circularity.

    As in this discussion!

  2. Sorry, but I disagree: Yes, people desperately want to believe that they survive death and therefore make up stuff like souls, but in practice, when the chips are down, they actually see themselves as their bodies. As I have argued before, it is "argh, you shot me into the foot", not "oh dear, you shot the body I am currently using as a vehicle into the foot".

    But of course, what is relevant is that we are our bodies, whether people like to believe it or not. Those who say that our neurons or genes make decisions for us are basically smuggling dualism in, because such a sentence only makes sense if "us" is envisioned as a soul-thingy riding along in the body. The only difference to religious dualists is that those consider the soul to be in charge while the confused physicalists consider it to be a helpless observer.

  3. People disown parts of their bodies all the time, when it suits them. Few people are going to deny that they've been shot, but they may feel they should have been the opposite sex. If the self is a sub-system of the body, it can also hold conversations with other selves. Someone's self-image may be invested in their gorgeous hair, but we feel no remorse following a haircut. Yet we may keep a diary that is highly personal but part of an extended embodied cognitive system rather than being our bodies. Being purely physical doesn't mean that we exactly correspond with our bodies, we're something our body acts out, and essentially social. I'd emphasise the processes that constitute ourselves over the exact location. It's misleading to say we're helpless observers, we're involved, active observers

    In a very circular kind of way.

    I do like your plant pictures!

  4. I am not sure how much we disagree anyway, not least because I don't understand how the things you say are a problem for what I am saying. My concern is not with the fact that we have an extended self-image, but with people denying that I make decisions, for example. I make decisions to the same degree that a pocket calculator calculates or that a stomach digests, no more and no less; the fact that we are part of the universe and of a complex cause-effect network doesn't change that, and I would argue that one cannot coherently deny the first without likewise denying the second and third, turning all our language into useless mush.