Thursday, March 19, 2020

The four causes and supertasks

Suppose I make a geologist’s hammer out of a chunk of steel and break a rock with the hammer. Then the chunk of steel is the material cause of the hammer, and the hammer is the efficient cause of the rock breaking.

The hammer then is explanatorily prior to the rock breaking and the chunk of steel is explanatorily prior to the hammer.

Admittedly these are different kinds of explanatory priority. But they do nonetheless combine: it is clearly correct to say that the chunk of steel is explanatorily prior to the rock breaking. (I am not claiming that transitivity holds across all the kinds of explanatory priority, though I suspect it does, but only here.) But now notice that this instance of explanatory priority does not correspond to any of the four causes: in particular the chunk of steel is neither the material nor the efficient cause of the rock breaking (it is only insofar as the chunk was shaped into a geologist’s hammer that it broke the rock). Hence, the four causes do not exhaust all the types of explanatory priority.

Other examples are possible. I push a rock with my hand, and consider the conjunctive state HM of there being a hammer and a rock moving. Then HM is explained by the chunk of steel and my hand. But the chunk of steel and my hand constitute neither a material or not an efficient (nor any other) cause of HM. Thus, again, we have explanatory priority not corresponding to one of the four causes.

The above examples do, however, permit one to hold the following view:

  1. All fundamental instances of explanatory priority are instances of the four causes.

Thus, the four causes would be like Aristotle’s four elements or three types of friendship: they combine to provide all the cases.

But now an interesting bit of heavy-duty metaphysics. Suppose that dense causal sequences are possible, i.e., causal sequences such that between any two items in the sequence there is an intermediate one. Then no instance of causation in the dense sequence will be fundamental. And hence (1) won’t tell us as much as it seems to. Indeed, given dense causal sequences, weakening the four cause thesis to (1) eviscerates the four cause thesis.

Thus we have an argument that if we want to take the four cause thesis seriously, we need to accept (1), and hence we need to reject dense causal sequences.

But if supertasks are possible, it seems like dense causal sequences should be possible. So, if we want to take the four cause thesis seriously, we need to reject supertasks.

It is, by the way, interesting to think about supertasks where the items in the task alternate between different types of causation.

Note that the above point applies to other sparse pluralisms about causation besides the four-cause one.


Martin Cooke said...

Do you need to reject supertasks independently though? Why not just say that, given the truth of the four cause thesis, dense causal sequences are not possible? It seems to me that any attempt to rule out supertasks would have to make some pretty substantial metaphysical claims to begin with...

Philip Rand said...

Martin Cooke


Concluding causality must be covariant. BUT, solutions under particular conditions need not be.

If one knows a non-covariant solution A in one frame of reference and wants it in another this can be implemented by use of a particular transformation rule.

Alexander R Pruss said...


Here's a better argument against supertasks than the one in the post: The outcome of a supertask is caused, but is not directly (and hence not fundamentally) caused by anything, since between the outcome and every cause there are infinitely many intermediate causes. And if we want to maintain a robust four-cause story, we need causation to reduce to fundamental causation.