Miguel wrote:

The problem is that if IPSR is false then there is no reason whatsoever as to why EPSR would hold. It would be a merely accidental fact. EPSR does not explain why we tend to find explanations for things or why things don't pop into existence, it merely *describes* such facts.

It depends what you mean by "explain why we tend to find explanations for things or why things don't pop into existence." If every contingent fact has an explanation, then everywhere we will look we will find explanations for contingent facts. Why do we find them? Because they are there. Why do I find a can of Coke in my fridge? Because it is there. (Not: because it has to be there.)

Miguel wrote:

EPSR would not explain why the universe does not cease to exist at any moment, it merely accomodates it.

EPSR just states that every contingent fact has an explanation. If it is contingent that when the universe has existed as our has up to t, then it continues to exist after t, then EPSR states that this fact has an explanation. So EPSR does not explain this fact itself, in the sense that EPSR is not the explanans of "when the universe has existed as ours has up to t, it continues to exist after t"; but EPSR implies that that proposition has an explanans.

But is that a problem? IPSR seems to be in the same boat. If IPSR, then the explanans of  "when the universe has existed as ours has up to t, it continues to exist after t" need not be, and probably is not, IPSR.

There are some odd questions that arise when one starts asking whether facts of the form "Fact f has an explanation" have explanations. If these are facts, then they are contingent facts. So if facts of that form are facts, then EPSR is a contingent fact and implies that it has an explanation too. And it might be thought that that is a problem for the hypothesis that EPSR & ~IPSR, for what can the explanation of EPSR be?

I am not sure what to say here. I am somewhat doubtful that "facts" of the form "Fact f has an explanation" are really facts at all. But it might also be said that their explanations are just the explanans of f. What explains the fact that "The door's closing has an explanation"? Well, the same thing that explains the door's closing: the blowing of the wind.

And what explains the fact that "The door's closing has an explanation & John F. Kennedy's assassination has an explanation"? Well, not some special fact that happens to explain both the door's closing and John F. Kennedy's assassination, but just the fact that the wind blew and someone shot John F. Kennedy.

So, reading the universal quantifier in EPSR as a big conjunction, it is open to suggest simply that EPSR is explained by all the facts which are explanans.

Miguel wrote:

If EPSR is explained by all the facts which are explanans, 1) still there would be no reason whatsoever as to why things have explanations if they don't need them, 2) it wouldn't be an explanation, merely an assertion of EPSR.

Just to clarify, the suggestion is not that facts of the sort "Fact f has an explanation" (of which EPSR is the conjunction) are explained by facts of the sort "Fact f has an an explanation" but rather, simply, by the facts which are the explanations. So the proposal is not that the explanation of EPSR is, essentially, EPSR, but rather that it is just the facts which explain the conjuncts of EPSR. It is not, then, a reassertion of EPSR, and on the hypothesis that those facts do explain the conjuncts, the explanation of EPSR is indeed an explanation.

Now, given that you seem to be granting for the sake of argument that "EPSR is explained by all the facts which are explanans," I don't understand your 1). On the hypothesis EPSR says that each thing has an explanation, and is explained by those explanations. So that things have explanations seems to be explained.

Of course, "things have explanations" may not be intended universally but generically. It may be, that is, "things have explanations" is meant in a sense in which it is not entailed by "each thing has an explanation": it means something more like "things, by virtue of being [contingent] things, have explanations." And it is true that that modal fact is not explained on this view; but on the hypothesis that EPSR & ~IPSR, that modal fact is not even true. It is, indeed, hard to see how it differs from IPSR itself, so its truth cannot count in favor of holding IPSR.

Miguel wrote:

(compare "everything has an explanation" to "everything just happens to have an explanation").

I should say that I do hold some version of PSR. I am partly playing devil's advocate, partly trying to register the absurdity of defending PSR without putting a concrete, determinate proposal of what explanation is on the table.

Say one thinks that explanation is just some dyadic relation between facts. Having just said that, which of course does not distinguish explanation from various other relations, we can state certain theses, such as that this relation we are talking about relates each contingent fact to some other fact (EPSR) while denying that necessarily this relation relates each contingent fact to some other fact (~IPSR).

If that's to be ruled out, one needs further principles to show why that is so. It isn't clear why, though, understanding explanation in this way, we would start looking for a place for EPSR and IPSR in the explanations themselves:

Miguel wrote:

it is still a fact that if IPSR were false then the universe could have ceased to exist at any point in time, since it would be possible for it to cease to exist. So IPSR is part of the explanation in this sense. EPSR by contrast it no explanation at all, it is merely a description.

Explanation is a relation between facts, which we refer to using sentences which describe them. There's prima facie no reason why EPSR's being a description, in the sense of stating that things have explanations, disqualifies it from being explanatory. Lots of explanations are descriptions.

What accounts for the inclination to say these things is that there is more meat on the bones of the explanation relation. You say that "if IPSR were false then the universe could have ceased to exist at any point in time, since it would be possible for it to cease to exist"; I take it that you are thinking that "if a explains b, then necessarily something explains b", the thought being that something's having an explanation must rule out the possibility that it does not; something would not count as an explanation if it were possible for the explanans* to occur without being explained. (*Or some fact of the same sort as the explanans, for if one accepts the essentiality of origins it might be essential to a fact that it has the explanation it does.)

But I think there are some reasons, at least, to query that thought. For suppose one thinks that facts about causation ground explanations, but that causation is not reducible to modal notions (say, to counterfactuals, and to truths about possible worlds). Then the fact that a causes b is just a fact about the actual world and grounds the fact that a explains b. But if facts about causality are not reducible to counterfactual facts about other possible worlds, then it hardly seems appropriate to say, at the outset, that a could not cause b (or, again, a fact like b) if in some other possible world b has no explanation.

I think there are other reasons just based on the ordinary contexts from which we acquire our notion of explanation. For consider coincidences, like the coincidence that I see my friend while vacationing remotely. The notion of a coincidence is that what seems to demand a unified explanation in fact lacks one. In this case, it seems that "I am here and my friend is here" must be explained by "one" fact, and not merely by the conjunction "I decided to take vacation this week in this location and my friend decided arbitrarily to take vacation this week in this location". And to be sure, sometimes what would otherwise be coincidences do have unified explanations; my friend may have a strange obsession with me, or perhaps we both came across the same promotion for vacations in this spot at this time, because it was advertised in a spot we both pass in going to work each morning.

At least in this context, there is no problem in saying that what has a unified explanation in fact could have lacked a unified explanation. One might say that this has nothing to do with PSR, for the fact is still explained by the coincidental, conjunctive fact. But the point is just that it is not part of our notion of explanation that we withdraw our appraisals that some fact has an explanation on account of its possibly not having one.

A similar but distinct case: the explanation by rationalization of action. It's possible to do something "for no particular reason", in the sense that if one were asked why one is doing what one is doing, one would say, "Oh, no particular reason." When I ask my friend what he is up to at the park, and he says that he is there because he was stressed and needed some fresh air, I am not deterred from accepting that explanation on account of the fact that it is possible that he could have been there for no particular reason. Again, there is some deeper explanation for people's actions when they say that they are doing something for no particular reason. But it is hard to see why our policies of admitting explanations in the face of their possible absence falls away when we throw away the ladder and deploy the purified metaphysical concept rather than the one at home in our ordinary practices.

I still believe that arguing with PSR as the start is a bad approach.

Instead, I'd divide between "facts which have an explanation" and "facts that have no explanation". Because, let's be honest, "facts which should have an explanation but don't" is an irrational definition.

I'd go with the PoI instead of the PSR : can we "rewrite" explanations into "descriptions of fact"?

aftermathemat wrote:

FrenchySkepticalCatholic wrote:

I still believe that arguing with PSR as the start is a bad approach.

Instead, I'd divide between "facts which have an explanation" and "facts that have no explanation". Because, let's be honest, "facts which should have an explanation but don't" is an irrational definition.

I'd go with the PoI instead of the PSR : can we "rewrite" explanations into "descriptions of fact"?

Are you talking about PNC based arguments for PSR, and/or Lagrange's formulation of the PSR which is based on the principle of identity?



 

 
I think he's talking about modally weaker PSRs, from "possibly has an explanation" to "has an explanation".

I still don't see any reason to accept that "everything just contingently happens to have an explanation when it could have failed to" without accepting "everything necessarily has an explanation". Nothing will convince me that the first can actually explain, rather than describe -- in a non-illuminating, non-elucidating manner, but only as "matter of fact" -- the fact that things tend to have explanations and not pop into existence, whilst the actual PSR gives us an actual explanation for why it is so: that it could not have been otherwise; that the nature of beings metaphysically require that their existence is due to something actual; that something in which essence is distinct from existence can only exist if it receives existence from something else, etc., however you formulate it. Anything else would 1) not be as explanatory or illuminating, and 2) would not be as good a candidate for a metaphysical principle (in fact, how would it translate as a metaphysical principle, if it doesn't necessarily hold for things by virtue of what they are?)

As for what explanations are, we have a solid intuitive grasp on them, but we can divide them into any of Aristotle's four causes (other divisions, such as conceptual explanations, scientific explanations, and agential explanations, can respectively be made sense of by reference to the four causes).

Miguel wrote:

I still don't see any reason to accept that "everything just contingently happens to have an explanation when it could have failed to" without accepting "everything necessarily has an explanation". Nothing will convince me that the first can actually explain, rather than describe -- in a non-illuminating, non-elucidating manner, but only as "matter of fact" -- the fact that things tend to have explanations and not pop into existence, whilst the actual PSR gives us an actual explanation for why it is so: that it could not have been otherwise; that the nature of beings metaphysically require that their existence is due to something actual; that something in which essence is distinct from existence can only exist if it receives existence from something else, etc., however you formulate it. Anything else would 1) not be as explanatory or illuminating, and 2) would not be as good a candidate for a metaphysical principle (in fact, how would it translate as a metaphysical principle, if it doesn't necessarily hold for things by virtue of what they are?)

You are using explanation in two different senses. The sense of 'explanation' which occurs in any formulation of PSR is a metaphysical one; it is the idea of a certain kind of relation that obtains between things and/or facts.

But then you are also talking of the sense in which a theory might explain a variety of phenomena: the notion of explanation as abduction. The thought is that "every contingent fact must have an explanation" better explains the phenomena of our tending to find explanations than does "every fact has an explanation".

This is a different notion of explanation. It is certainly not the case that every contingent fact needs an explanation in this sense; that is, that of every set of phenomena they have some unified explanation, the theory which explains them; and it is certainly not the case that the best explanation in this sense is the correct one.

I think that's probably true, and IPSR better explains (in the second sense) our tending to find explanations (in the first sense) than EPSR. But it should be appreciated that these two senses don't obviously have something to do with each other, such that there is an incoherence in the hypothesis that EPSR & ~IPSR (the supposed incoherence depends on an equivocation on the two kinds of explanation).

My original reason for stating IPSR and EPSR was not to suggest that EPSR is an adequate explanation (second sense) of our finding that things have explanations (first sense). It was just that 'PSR' is ambiguous; by that, one could mean IPSR, or EPSR, or neither. For instance, do explanations relate facts or things or both? I am baffled by the suggestion that we can start showing, confidently, the rejection of PSR to be incoherent or disastrously implausible without answering that question.

But, given that there are more than one sense of 'explanation', there is also a problem of saying which we mean when we formulate PSR. Skepticism of PSR might be motivated by a tendency to think that there isn't a form of explanation distinct from the abductive sort. In engaging with the world we are essentially trying to build a good theory, a theory which is the best explanation of the phenomena which confront us. There is nothing metaphysical about saying that a fact is explained; it is just to say that we have fit it into our theory, and our theory accounts for it. Then there's no fact of the matter about our finding facts typically to have explanations, and therefore no need to posit PSR as a best explanation of the phenomena.