Monday, October 12, 2009

Clever Hippocratics

Here’s a clever argument in the Hippocratic Nat. hom. 2.10–12:

Ἐγὼ δέ φημι, εἰ ἓν ἦν ὁ ἄνθρωπος, οὐδέποτ’ ἂν (10)
ἤλγεεν• οὐδὲ γὰρ ἂν ἦν ὑφ’ ὅτου ἀλγήσειεν ἓν ἐών• εἰ δ’ οὖν καὶ
ἀλγήσειεν, ἀνάγκη καὶ τὸ ἰώμενον ἓν εἶναι•

Man cannot be one (perhaps: a unity – the term hen is a bit slippery here but I’ll just breeze on…) since if he were he would not feel pain. (There would be nothing on account of which he would feel pain.) And if he were, per impossibile, to feel pain being such a unity then necessarily what cures would also be one.

I suppose there are two arguments here for a human being a complex of some kind. (By that I mean simply the claim that humans are in some way or other complex, either by being composed of more than one element, or being composed in some complex fashion such that not every part of a human is just like every other part of a human.)

The first argument is that humans feel pain. Why would something not feel pain if it were a unity? I suppose the assumption is that pain is to be understood as some kind of reaction to or sign of or indeed just is a change in the arrangement of whatever it is that humans are. A genuine unity could not undergo any such internal change at all and so a fortiori would not allow there to be any change of the sort that is related to pain.

There are some interesting links here with (i) Melissus’ idea that what is does not feel pain (B7.4) and (ii) Diogenes of Apollonia’s idea that in order for things to interact with one another there must be an underlying basic monism. I’ll have to do more to pursue those and there are no doubt other interesting strands to find, but those are the ones I thought of immediately. Interesting philosophers, these Hippocratics.

The second is also a good argument. Humans must be complex because what cures humans is not one or simple. I suppose the thought is this: if all humans were simple, then any disorder or pain-producing state would have to be such that there would necessarily be a single thing that would be capable of curing all humans (indeed curing all humans of all ailments). If a human were just a unified bit of some particular simple substance then, the thought goes, a single thing would be able to sure all ailments since, in order to have any effect on a person at all, it would also be able to effect all of the person (since the person in question is simple and a unity). The same thing would do to sure earaches as toothaches as kidney stones as…. Of course, as it is different things cure different ailments and different things might also cure the same ailment in different people. So people must be complicated.

Clever.

2 comments:

Matthew Duncombe said...

Hi James!

I like your post on the Hippocratics, and if I may, I would like to offer some thoughts on the second argument you present

As far as I understand the reasoning you attribute to them, it goes, as a reductio:

(1) If a man is a unity, then all 'parts' are token identical.
(2) If all parts are token identical, then if I change one part in one way, all parts change in that way (in this case, if I change a broken 'arm' so that it is cured, I also change a broken 'leg' so that it is cured)
(3) Therefore, only one token thing (e.g. a token plaster-cast) is needed to cure all parts of a man.
(4) But, in fact, one type of thing cannot cure all ailments, so one token thing cannot cure all ailments. Hence, man is not a unity.

One problem is that its not clear what 'all ailments' amounts to if all parts of a man are token identical. Does it mean that the man has the properties 'toothache' and 'kidney stone'? That doesn't work, as toothache is a property of teeth, and kidney stones are a property of kidneys, parts which they are supposed to be assuming don't exist. So if all parts of a man are token identical, 'all ailments' ceases to mean anything like 'ailments of different parts of the body': you are either (wholly, if you like) ailed, or not.

If they object that you could be 'ailed' and 'ailed*' and 'ailed**', they are correct. But then the argument collapses, as now there is no reason to assume that 'ailed' and 'ailed*' 'ailed**' will be treatable with the same type of treatment, let alone the same token treatment. Unless, you assume that a 'cure' will cure all types of ailments. But they deny this.

The second problem is that it doesn't follow from the argument that the same thing would be capable of curing all humans, unless you assume that in addition to each human having all parts as token identical, each human is type identical. But the latter doesn't obviously follow from the former. I could be a unity, but a completely different kind of unity from you. (e.g. 'up' and 'down' quarks).

Also the move from (2) to (3) looks fishy. Is the assumption being made that the thing causing a change (a cure) and the thing being changed (the human) must have the same mereological characteristics? It's not obvious that they are making this assumption. Maybe argument works without it, but in the text it is so condensed that its hard to tell. Consider this parallel:

(1') If a man is a unity, then all 'parts' are token identical.
(2') If all parts are token identical, then if I change one part in one way, all parts change in that way (in this case, if I change an 'arm' so that it is red, I also change a 'leg' so that it is red)
(3') Therefore, only one token thing (e.g. a token spray can) is needed to paint all parts of a man.

While the conclusion is true, it would also be true if man was a composite (unlike the first version of the argument). So the argument can't be valid as it stands.

I don't know what you think about this, but its a really interesting argument!

Hope all is going well in Cambridge. Paris is cold...

JIW said...

Dear Matt,

Thanks. I suppose I thought it was quite a simple point, but you're absolutely right that a lot will hang on the precise meaning of 'being one'. The second argument is perhaps the most interesting.

I'm not sure I follow the idea that token identity would preclude ailments being even localised in one area or another unless this too is part of a reductio: if the notion that a person is 'one' cannot account for the obvious fact that a person can be ailed here and not there then that is yet another reason to add to the list of those why man cannot be one. There is a chain of per impossibile reasoning, in that case:
If man is one he cannot fall ill.
If per impossibile a man who is one did fall ill he could not be ill in one part and not another.
If per impossibile such a man were to be ill in just his finger then we would still have to say that whatever cures the ill finger would also cure an ill kidney.

So I'm not sure that the inference from (2) to (3) is necessarily as you present it. It seems possible that there be multiple possible cures of any hypothesised ailment (and the argument need make no claims about the nature of the causation involved in curing) but whatever does cure an ailment will cure any ailment in the person who is one.

You're right that the argument does not insist that the same thing will cure all people, but the writer might well take there to be an analogous argument as follows: just as different things cure different ailments of the same person (and so that person is not one) so too since different things might cure the same ailment in different people then people are not 'one'...