To: Mark S.
From: Geoffrey Klempner
Subject: What is a law of nature?
Date: 11th January 2011 12:14

Dear Mark,

Thank you for your email of 1 January, with your essay for the University of London BA Methodology module, in response to the question, 'What is a law of nature?'

I haven't looked at the examiners' reports (as you know, I don't do this as a rule) but to my mind you haven't offered an answer to this question, but instead an answer to some question like, 'Evaluate the debate between the regularist and necessitarian accounts of the laws of nature.'

There is a possible clue that you have misunderstood what was asked for, in the fact that you offer a whole list of explanations of the laws of nature:

the methodological account (Cohen)
the inference-ticket account (Schlick, Ramsey)
the invariance account (Woodward)
the agency account (Menzies)
the capacities account (Cartwright)
the functional account (Lange)
the criterial account (Bird)
the dispositional properties account (Popper, Harre and Madden, Shoemaker)
('to mention only a few'!)

If the question were really asking for the *best philosophical explanation* for what makes a law of nature a law of nature, then surely you ought to consider every view and not just pick on the two favourites, or which most appeal to you. However, I don't think the question is asking this. It is asking, simply, What IS a law of nature? Having decided what a law of nature is, we can then go on to discuss explanations of what makes a law of nature a law of nature. You can say this, of course.

It is true that one has to go through some of the same moves. Generality, to take the most pertinent property, is one that surely belongs to laws of nature. But not all generalities are laws of nature. So what makes the difference? The point here, though, is to offer an analysis of 'what makes the difference' by looking at putative examples, and counterexamples. Having reassured ourselves that we have a good idea of what a law of nature is, then the next step is to offer a philosophical justification or explanation of this distinction.

It is also true that rival philosophical explanations will disagree over some cases of putative 'laws of nature'. This is a general feature of debates over explanations generally, not only in philosophy. You adjust the pool of data to fit the theory to some extent. In other words, a trade off. However, the first step is still to identify what we are talking about, what we are seeking to explain.

In my last email I talked about the idea of nature and 'putting questions to nature'. One aspect of the idea that nature is 'lawlike', or governed by or consists in 'laws' is the idea that answers to questions we put to nature will always be truthful. This does not go without saying. Why would the world be like this? Or why must it be?

According to my take on this question, your decision to 'focus on the laws of physics and chemistry' is a strategic error, because the question of what a law of nature is arises in its most acute form when we look at the 'special sciences', and what they regard (rightfully or wrongly) as 'laws'. Here's a test: suppose we are debating the question whether Freud's law of the 'return of the repressed' is a genuine law of nature. How is it helpful to have decided in advance that we are regularists about laws or necessitarians? (Maybe you have an answer to this: but I can't see what it would be.)

We have mentioned generality. What is generality? There are actually two questions here, which are highlighted in the way you lead into your discussion of Hume. There's the question how we decide whether or not a generalization is true, the grounds for asserting a general statement, and the question of what follows if a generalization is true, or what it's truth consists in. Humean scepticism about induction is based on the recognition that the statement that we make has consequences which far outstrip any grounds we have for making the statement.

This is a point I've made before: a generalization (or, at least an unrestricted generalization) makes a very strong claim. If I said that none of the the objects in my study contains over a kilogram of gold, that is a claim which is verifiable by exhaustive search (as if I didn't know it already). If I say that all objects in the universe attract one another with a force inversely proportional to the square of the distance between them, then there are all sorts of very good grounds which one could appeal to but none establishes the generalization exhaustively.

I don't know whether anyone has attempted to calculate the total mass of gold in the universe. I assume that an approximate figure could be produced by considering the relative distribution of various kinds of star. Then it is a law of nature that no lump of gold is more than n squared, where n is the maximum possible amount of gold that can exist, given the size and age of the universe, the mechanism for the creation of gold by nuclear fusion etc. If it just so happens that no lump of gold exists anywhere in the universe that weights more than 1000kg, then that deficiency could be rectified. There's a long time to go before the big crunch.

One particular point which you make in your essay which is relevant to the question, 'What is a law of nature?' (as opposed to 'What is the best explanation of what makes a law of nature a law of nature?') concerns instance-less laws and probability laws. These are examples of hard cases which we need to get clear about. But surely we don't need to decide on our explanation of laws of nature first: we know that there are instance-less laws and probability laws. You can't remove these from the pool of 'data'. If you defined a 'law of nature' in such a way to rule them out then (pace Quine) you'd just be changing the subject, talking about something that you've invented (schlaws of nature).

You asked me one question, with regard to your response to van Frassen's objection to Armstrong (p.8-9). The problem for me is that you haven't stated van Frassen's objection in a way that makes it look the least bit plausible. The Romeo example doesn't help. So, for all I know, you are just rebutting a straw man. If all cordates are necessarily renates then my being a cordate necessitates my being a renate, the two properties, in me, are causally linked. The are causally linked because of the kinds of property that they are. When you understand the properties that they are, then you understand why the causal linkage could not fail to be instantiated in any given individual, such as myself.

All the best,

Geoffrey