We hear a lot about logical possibility - as in ‘it is logically possible that…’ - in contemporary analytic philosophy. For example, David Chalmers is always saying that it's logically possible that, say, zombies could exist, etc. What do these claims amount to? Do they amount to much and should we be put out by them? Bertrand Russell thinks not:
"No logical absurdity results from the hypothesis that the world consists of nothing but myself… and that everything else is mere fancy. But although this is not logically impossible, there is no reason whatever to suppose that it is true; and it is, in fact, a less simple hypothesis, viewed as a means of accounting for the facts of our own life, than the commonsense hypothesis that there really are objects independent of us, whose action on us causes our sensations." (Russell, 1912:22.)
When I wake up, it's logically possible that I'm still asleep. When I wake up, instead, it's logically possible that all my family are now dead. When I move over to the tap, it's logically possible that poison, not water, could come out of it. If it is water, it's logically possible that I could choke on the water. Then I look out of the window. It's logically possible that the town I see it a simulation of what I saw the day before. And so on.
I could think of ten logical possibilities before each breakfast.
But what is the point, even from a philosophical perspective?
Perhaps there is a big difference between the merely possible and the logically possible.
It's the logically possible that excites the sceptics and philosophers like Chalmers. Should I conclude, from all these logical possibilities, in the same way as Russell above? That is, should I say that "although [they are] not logically impossible, there is no reason whatever to suppose that [they are] true"?
Put simply. Something that's logically possible still may not be true. Indeed something that's logically possible probably won’t be true – won’t be the case. So why contemplate the logical possibility at all, even philosophically? Where will it get me?
Russell’s next point is that the logically possible hypothesis is ‘less simple’ than the everyday one. He writes that "it is, in fact, a less simple hypothesis, viewed as a means of accounting for the facts of our own life". In my examples, this means that it is a less simple hypothesis to believe that I'm currently dreaming. It's a less simple hypothesis to assume that all my family is now dead. It's a less simple hypothesis to assume that poison will now come out of my tap rather than water. It's a less simple hypothesis to assume that my window-view is a mere simulation of the facts.
However, perhaps it doesn't really matter if my everyday hypothesis - if it is a hypothesis - is simpler than the logical possibility. Simplicity may not equal truth. A logical possibility may not automatically equal a falsehood. Perhaps the simplicity or complexity of each hypothesis simply doesn't matter to that hypothesis’s truth or falsehood. Perhaps simplicity is only psychologically appealing – and that’s it.
This is interesting because simplicity, or the simplicity of theories or hypotheses, is given a lot of credence in science – or so many philosophers of science have told us.
Russell, however, thinks that the simplest explanation is the ‘best’:
"We can then ask: what is the best explanation of these experiences? Russell answers that the simplest explanation is the best; and the simplest explanation is the commonsense explanation. So we know that there is an external world because we seem to experience an external world; and the best explanation of our seeming to experience it, is that we really do experience it. Russell’s argument is an instance of the form of argument known as abduction, or inference to the best explanation." (Chappell, 79)
Again, why is the simplest explanation seen as the ‘best’? Is there a necessary or logical relation between simplicity and truth? Certainly many philosophers and scientists have thought so. But why? Because simplicity is more aesthetically appealing? In that case, what is the logical or necessary relation between aesthetic appeal and truth?
The example of "the best explanation of our seeming to experience the external world, it that we really do experience it" (79) doesn't really seem like an explanation at all. It certainly doesn't ‘seem’ like a philosophical explanation, let alone a justification of one’s belief that the external world really exists! The fact that I ‘seem’ to be rich doesn’t make me rich. The fact that I think or seem to be attractive to all women doesn’t mean that I am attractive to all women. I seem to experience goblins when drunk; though that doesn’t mean that they exist. Indeed if I seem to experience goblins when sober, why should I assume that goblins really exist and not, instead, explain my experience in some other way? So, again, why is the simplest explanation the best explanation?
In addition, when we make an abductive inference, is it necessarily the case that it must be the simplest inference or explanation? That is: does abduction = inference to the best explanation? They are indeed often connected; though are they synonymous? Chappell gives is own example of why the simplest explanation may not be the true explanation:
"Suppose sixteen very similar murders occur in one city in one month. On the facts we have, the best (and simplest) explanation of the murders might well be that the murders were all committed by the same criminal. It obviously doesn’t follow that this explanation is true: there will be many facts that we are not explaining by this hypothesis, because we don’t know them." (80)
For example, eight of the murders might have been carried out by a ‘copycat’ killer – or even fifteen of them. Perhaps two killers are working in conjunction. Perhaps because all the killings were done with a handgun may have something to do with the fact that a thousand handguns were found and sold in the city to a thousand people – that’s why all the killings were handgun killings, not because they were all committed by the same person. Perhaps the ‘fact’ that all the deaths were similar may turn out to be a false in the end. Someone who died of poisoning might in fact have gone on a mushroom hunt the day before. What do we mean by ‘similar’ as in ‘similar murders’? For example, aren’t many murders similar to each other anyway? For example, in New York I would guess that 90% of killings are done with handguns of some description. So all these murders will be similar. It doesn’t follow that they were all done by the same person. If there were no guns, perhaps most murders would be strangulations. Again, not every case of strangulation would then point to the same suspect.
Despite all these possible explanations, it's still the case that an inference to the best explanation may be, well, the best explanation:
"Inference to the best explanation is obviously a useful intellectual strategy, perhaps an indispensable one." (80)
Indeed the simplest explanation of every murder in every city in the last ten years would be that the same man committed all of them – all 100,000 of them! Actually, perhaps that wouldn’t be a simple explanation because we would need to explain how he found the time and the skill to kill 100,000 people! We could settle on the simple hypothesis that all these murders were carried out by the same gang. Surely that would be a simple hypothesis.
The problem with using inference to the best explanation as "a useful intellectual strategy" (80) in these examples at least, is that innocent people may be convicted of murder simply because their committing the murders is the simplest explanation on offer. That doesn’t or wouldn’t mean that they actually did commit the murders! However, the simplest explanation may be ‘useful’ in the sense that if this weren't an ‘intellectual strategy’ the detective, or indeed anyone in any walk of like, may be overburdened with data. And some of that data would of course be irrelevant or bogus. The problem is, of course, of deciding what facts or data are irrelevant or bogus. Perhaps some complex pieces of data, or some complex facts, aren't irrelevant or bogus. Perhaps they are really explanatory and even true. Still, a detective, or anyone else, can't have all the facts. And even all the facts he knows can't be thoroughly accounted for. We need to select what we think is relevant. Perhaps the way to do this is to select the simple facts or explanations of the matter at hand.