Saturday, August 30, 2014

Preemption of laws of nature

I've been thinking about cases of preemption of laws of nature. Normally, an instance of a law is explained by the law. Let's say some dose of cyanide is lethal to chickens within minutes. Then it seems it's a law that a chicken is dead ten minutes after ingesting that dose (take "that dose" to be an abbreviation for the actual dose) of cyanide. Normally, then, if a chicken ingested the cyanide more than ten minutes ago and now is dead, that's because of the law.

But what if the law is preempted? Suppose that a few seconds after taking the cyanide, the chicken is eaten by an unfortunate fox. Then the case is still an instance of the law, but is not explained by the law. (This is a tweak of an improvement by Brad Rettler of a case I used in class today.) The law is true, but preempted by the fox, and hence it is not explanatory.

So it seems that for an event to happen because of a law is something more than for it to fall under the law.

Consider a way of getting out of this. Perhaps it's not a law that

  1. a chicken is dead ten minutes after ingesting that dose of cyanide.
Maybe instead it's a law that
  1. if meanwhile there are no other causes of death and it's not the case that the chicken dies causelessly, then a chicken is dead from the cyanide ten minutes after ingesting this dose of cyanide.
If so, then we have to suppose lawhood not to be both closed under conjunction and relevant entailment, since (1) follows from the conjunction of (2) with the law that dead chickens stay dead. That's an interesting result. Moreover, we have to deny that laws are universal generalizations that support counterfactuals, since (1) supports counterfactuals every bit as well as (2) does.

All this can be done. Still, (1) looks much more like a law than (2) does, and it seems to me that rather than fooling with something like this, it's better to allow that laws can be preempted. An instance of a law need not be true because of that law.

There might even be laws where in fact typically instances of the law are not true because of the law. Maybe, because of biological clocks in our cells, it's a law that nobody survives to be 130 years of age. But most people don't die because of this law. They die of heart attacks, car accidents, cancer, etc.


Heath White said...

I'm not an expert in this field but I think it is a pretty standard view that many (most?) laws of nature come with implicit "ceteris paribus" clauses.

Alexander R Pruss said...

Good point. One thinks of the ceteris paribus clauses as added to limit when the antecedent happens. But perhaps one should also add them to limit when the antecedent is explained by the law.

For instance, barring miracles, it might be that it's exceptionlessly true that chickens are dead ten minutes after taking such a dose. (Maybe not in our world, as we could imagine some science fictional device that sucks all poison out of their bodies.) But while it's exceptionlessly *true* that chickens are dead ten minutes after the dose, it's not a *law* that chickens are dead ten minutes after the dose. Instead, it's a law that ceteris paribus chickens are dead ten minutes after the dose.

Heath White said...

It's tough being a ceteris paribus chicken. :-)

Alexander R Pruss said...

It's tough being a merely ceteris paribus chicken. Being a ceteris paribus chicken takes nothing more than just being a chicken. :-)