Thursday, September 08, 2005

Alchemy World: Why Water=H2O isn't metaphysically necessary

In this post I'm going to attack the view that water and H2O are identical in all metaphysically possible worlds. I was discussing this example with some of our new first-year students last week, and they seemed fairly impressed with this case, so I'll throw it out to whatever philosophical readers I have left.

Imagine Alchemy World. It isn't another planet in our universe, like Hilary Putnam's Twin Earth, but rather a planet in a different possible universe where instead of our 100+ elements on our periodic table, there are only 4. Call them E, F, A, and W. W makes up most of the rivers, lakes, and oceans, and it falls in rain. The people on Alchemy World drink W regularly. Their lives and the macrophysical phenomena surrounding them are identical to the lives and macrophysical phenomena of people on our world until, say, 1600 AD. Of course, things are different at the microphysical level, since our world has molecules of H2O where theirs has the indivisible element W. Suppose further that the elements of their world are so constituted that if H2O suddenly appeared in their world, it would annihilate everything.

Now the question is: In their world, is water H2O or W (or both or neither)? My intuition is that W, not H2O, constitutes water in such a world. Water, I'd say, is multiply realizable across the space of metaphysical possibility. It's probably only singly realizable across the space of physical possibility, however. Alchemy World contravenes physical laws, and I don't imagine you could get another reduction base for water without contravening physical laws.

Personally, it'd be a lot more convenient for me if I accepted the metaphysical necessity of the water-H2O identity. I think aggregate hedonic improvements constitute the good in all possible worlds, and I often have to refer back to the identity of H2O to water when I'm trying to make some point about how metaphysical necessity connects to issues in epistemology and science.


Richard Y Chappell said...

I think the standard account here would say (and I would agree) that when the inhabitants use the word "water", this refers to their watery stuff W, but W is not really water (i.e. what we call "water").

Then again, I'm vaguely sympathetic towards your position too.

Neil Sinhababu said...

Yeah, that's certainly the standard view.

Richard Y Chappell said...

Granted, W certainly plays the water role on Alchemy world. But it's a different substance isn't it? I'm not sure how much sense it makes to say that a substance can be multiply realizable. Is that really what you're wanting to suggest? Or do you think of water as a role rather than a substance?

Justin said...

So, I'm sticking to the conventional line that H isn't water, but I'm willing to grant that this example provides a more compelling case for you side than other examples do. So, for instance, I feel almost no inclination to say that XYZ is water, but I do lean a little toward saying H is. What I want to focus on in my post is, why is that? Why should your example be any different? Here are two hypotheses...

#1: Because the four element theory *really was* held in the past. No one ever actually thought that water was XYZ, but people really did think it was a basic element along with earth, air, and fire. And so, maybe your example ends up trading on this thought: what if the old theory was really right and the new theory was really wrong? Maybe this feature of the example inclines us to treat it counter-actually, in which case I and other conventionalists can account for its pull. This would need to be worked out more though.

#2: In the standard XYZ case, you imagine that the watery stuff has some sort of analyzable structure (akin to H2O), while in this case you are basically treating the watery stuff as a basic, un-analyzable substance. This seems to best fit with people's naive folk theory of water -- after all, there's a reason that people once thought it was a basic element. To the extent that people find this example more compelling than the XYZ one, maybe it's because this case engages that naive folk theory of water.

The question I'm raising here can be addrerssed even by people like Neil who don't share my standard-line take on things. Suppose that this example shows that water is multiply realizable. Still, why is this example somehow better able to show this than an XYZ example is? Because at least in certain ways, this case would seem to be not importantly different from an XYZ case.

Neil Sinhababu said...

I guess the way I'd put it, Richard, is that water seems more a functional term to me than a natural-kind term. (I guess I'm unused to the locution that some entity is a role. But if songs, pies, and desires are roles, that is what I mean.)

One thing I'd add to the list, Justin, is that the XYZ example is counter-actual where my W example is set in another possible world. There might be some intuitive pull to the idea that only one fundamental elemental/molecular constitution in any world can count as water. If all the watery stuff on Earth is H2O, then only H2O can be water in our possible world. Other thisworldly watery stuffs aren't really water. But when you go into another possible world and something else makes a credible bid for status as water, while H2O itself acts really weird, the other thing can become that world's lone kind of water.

Richard Y Chappell said...

Yeah, that's a more grammatical way of putting it :)

But if water is (something like) a functional term, then couldn't it be multiply realized within a single world? If we discovered the watery-stuff XYZ or W here on Earth, wouldn't you be committed to calling them "water" too?

Justin said...

First, a bit on terminology. A counter-ACTUAL is where you consider some other possible world *as actual*. So, imagine that the watery stuff around here *actually was* XYZ rather than H20. In that scenario, you're supposed to say that then water=XYZ and not water=H20. This differs from a case wherer you say, given that the actual watery stuff is H20, consider a world where the watery stuff is XYZ. There, you're supposed to say that water=H20 and not water=XYZ, so this possible world is just lacking water. From your comment I thought you might be clear on counter-actuals, but I wanted to add this just in case you weren't.

Anyhow, I don't think the factor you mention is influencing me much. When I mentioned XYZ scenarios in my original post, I meant to include scenarios where XYZ is the sole watery stuff, not just scenarios like the original Putnam one where the real world has both H20 and XYZ.

Here are how my own intuitions go. Consider a world where the sole watery stuff is XYZ (don't consider this world *as actual* though). I'm pretty comfortable saying that this is a world without water. On the other hand, consider a world where the sole watery stuff is W which is one of four basic elements along with E, F, and A. I think I still want to say that this is a world without water, but I admit I feel more inclination to go the other way now. If you have intuitions like mine (and maybe you don't) the question is, why should switching from XYZ to W mater at all? One might think that this should just be a trivial change, but I admit that for whatever reason, it seems to make some sort of difference.

Anonymous said...

As a folklorist, I'm highly amused by the phrase "naive folk theory". I'm also intrigued by the problem. Not being a philosopher, I may not strictly understand the difficulty, though, so here's my naive folklorist theory...

It sounds to me like Richard is right -- you have a problem of terminology, and that problem does leave you basically having to agree that water is, or could be, multiply realized within a single world. However, if it is the case that your problem is one of terminology, then I think the solution lies in determining what connection you'd like to obtain between the term "water" and its referent(s). At the end of your post, you note:

I often have to refer back to the identity of H2O to water when I'm trying to make some point about how metaphysical necessity connects to issues in epistemology and science.

Probably this would be clearer to me if I were a philosopher, but as a naive folklorist, I react to this paragraph by being curious how exactly it is that you need to refer back to this identity in these cases. For example, if you need to say, "H2O equals water in all possible worlds, and therefore all possible environmentalists need to focus their efforts on preserving H2O," then it sounds like you should come up with a definition of water that isn't tied to the external properties of water (liquid, clear, tasty, etc.) but rather to the internal properties of water as it relates to life (provides energy during certain chemical processes, etc.). However, if you want to say, "H2O equals water in all possible worlds, and therefore all possible folklorists should investigate H2O in their researches on water mythologies," then the opposite seems to be true, since folklorists usually care about very different properties of water.

Apologies if all this was obvious already.

Anonymous said...

Addendum: My local philosopher tells me that what I've come up with is basically a (badly phrased) version of Lewis' view of this problem. Sorry to be redundant. - L.B.S.

Neil Sinhababu said...

Richard, what I said to you leaves room for water to be multiply realized within a single world. But I said to Justin that I think there's some intuitive barrier to this. A descriptivist view about proper names would probably have to incorporate some kind of uniqueness condition, so that in a world with two Caesar-like things, only the one that is the most Caesar-like counts as Caesar. I'd like to build such a condition into the semantics of "water".

Justin, isn't the original Putnam case one that involves what we'd regard as a counteractual? I seem to remember that the discovery of Twin Earth is presented as a possible future event, rather than as an alternative possibility. Since Twin Earth probably isn't actual, this is a counteractual case. (Though if Twin Earth really is out there orbiting Aldebaran or something, it's not.) I was thinking here that indicative conditionals address counteractuals while subjunctive ones address counterfactuals.

I think our intuitions are in fact triggered by different things. Replacing W with XYZ in my example doesn't change my intuitions in any significant way.

Kaitlin, I occasionally think of my dissertation as a contribution to folk psychology. (Folk science is commonsense belief on some issue that falls within the domain of science, and it's to be contrasted with real science that comes from real scientists who do real experiments.)

Is Ed Su by any chance the local philosopher over there? Good kid.

Justin said...

Let me say a little more about counteractuals.

Let's say that W is a non-actuall possible world where the watery stuff on Earth is XYZ, not H2O. I want to leave Twin Earth out of it -- there is no Twin Earth either in W or anywhere in the story that follows. Now, in thinking about W, there are 2 ways you might consider it.

#1: When you consider W, you might bear in mind what the actual world is like. Specifically, you bear in mind that the actual watery stuff is H2O, not XYZ. Given that this is so, you know that water=H2O and not water=XYZ. And so, you know that W is a world without any water; it just has some stuff that looks (tastes, etc.) a lot like water.

#2: When you consider W, you pretend that W is the actual world. This is what counterACTUALS are all about. Now, if W were the actual world, then the actual watery stuff would be XYZ, it would not be H2O. And, if the actual watery stuff were XYZ, then water=XYZ, and not water=H2O. You can think of it like this: if in the real world, scientists discovered tomorrow that the actual watery stuff was XYZ rather than H2O (maybe all previous scientists made a mistake in analyzing water), then this would show that what "water" rigidly designates is really XYZ rather than H2O.

#1 corresponds to metaphysical possibility while #2 corresponds to epistemic possibility. It's epistemically possible but metaphysically impossible that water is XYZ (assuming here that the actual watery stuff really is H2O). #2 captures the sense in which we can truly say, "well, water might turn out to be XYZ -- scientists might discover that tomorrow"; if all we had to work with was the notion of metaphysical possibility, then there would be no sense in which water might be anything other than H2O.

And finally, here's why this is relevant to the original question. If you give me some scenario to consider, and something about the scenario "tricks" me into thinking about it as a counteractual, then I might have intuitions which would *seem* to tell against Kripke but which actually don't. So for instance, suppose an anti-Kripke said to me,

"Imagine that tomorrow scientists discover that the actual watery stuff isn't H2O, it's really XYZ. Would that show that there really is no such thing as water? That's crazy. What it would show is that water would be XYZ. And this shows that water=H2O isn't metaphysically necessary."

That's all fine except the last sentence; it takes a wrong turn when anti-Kripke claims that this would show water=H2O isn't necessary. In my original post, I wondered whether something about the fact that people actually held the four element theory tricks us into thinking of your case as a counter-actual, in which case the intuition that water=W can be accomodated by Kripke.

Anonymous said...

My local philosopher is my dad (who says hi), but Ed is indeed a good kid.

Neil Sinhababu said...

Whoa! I didn't realize this before, and it's too cool to go without mention. Justin and Richard, the folklorist in our midst happens to be the daughter of a highly regarded metaphysician.

Neil Sinhababu said...

I mean, I knew he was Kaitlin's dad, but I didn't associate it with the fact that we were talking about issues of metaphysical possibility.