Friday, June 22, 2007

Modal and temporal irrelevance

One objection you sometimes hear against Lewis’s modal realism (from van Inwagen, Chihara, Jubien, among others) is that what goes on at concrete spacetimes is irrelevant to what is necessary or merely possible. The objection, I take it, is this. We can grant for the sake of argument that there are the many cosmoi Lewis would have us believe in. But even on the assumption that there are these things, it’s not clear what they would have to do with modality. When we think of what is merely possible we are thinking of what could have been the case but isn’t (that’s the work the ‘merely’ is doing); but Lewis tells us that the merely possible is the case, it just isn’t the case here – at the sub-portion of all that there is that is spatio-temporally related to us: the portion that Lewis calls ‘actuality’.

The gist of the modal irrelevance objection, I take it, is that to say that something is merely possible demands that it not be the case – not simply that it not be the case in our surroundings, but that it not be the case at all. If Lewis is right about what there is then, the thought goes, it simply turns out that what is actually the case is a lot more complex than we thought. Actuality, the thought goes, is everything that (unrestrictedly) is the case: if there are talking donkeys that aren’t spatio-temporally related to me then there are actually talking donkeys that aren’t spatio-temporally related to me. Lewis can choose to use the term ‘actually’ as he wishes, of course; likewise with the terms ‘possible’ and ‘necessary’. But nevertheless the point remains that, using those words as we use them, Lewis is asking us to accept that actuality is a lot bigger than we supposed; he’s not asking us to accept the existence of the merely possible. The latter request is unfulfillable: you can’t accept the existence of the merely possible, because if something exists then it’s not merely possible – it’s actual.

Does anyone raise the analogous objection against eternalism? We can imagine someone arguing as follows:

“Just as I have the intuition that to say something is merely possible demands that it not be the case, so I have the intuition that to say that something is merely past or future demands that it not be the case. And yet the eternalist says that past and future events are the case: they’re not the case in my surroundings – the portion of what there is that the eternalist calls ‘the present’ – but they are the case nevertheless. But just as I can’t see why the presence of talking donkeys would result in it being merely possible that there are talking donkeys as opposed to it being actual just because the talking donkeys aren’t spatio-temporally related to me, so I can’t see why the presence of dinosaurs would result in it being the case that there were or will be dinosaurs as opposed to it being the case that there are presently dinosaurs just because the dinosaurs happen to be in a direction I can’t point!”

The objections are totally analogous. Just as the modal irrelevance objection says that you can’t have an ontology of the non-actual, you can only make actuality more complicated, so the temporal irrelevance objection says you can’t have an ontology of the non-present, you can only make the present more complicated.

The temporal irrelevance objection claims that if it was true that there are dinosaurs but that there are not presently dinosaurs then this demands that there are no dinosaurs; likewise, if it will be true that there are Martian colonies but that there are not presently Martian colonies then this demands that there are no Martian colonies. The eternalist will claim that something’s being merely past or future demands not that it not be the case, but only that it not be the case presently. But that is, seemingly, no more convincing than Lewis’s claim that something’s being merely possible demands only that it not be the case actually. The actualist can agree that something’s being merely possible demands only that it not actually be the case provided that actuality is understood as encompassing everything (unrestrictedly). Likewise, the presentist can agree that something’s being merely past or future demands only that it not presently be the case provided that the present is understood as encompassing everything (unrestrictedly).

So, here are some questions:

(1) Has anyone made the temporal irrelevance objection, or anything like it, in the literature?

(2) Do the modal irrelevance and the temporal irrelevance objections really stand or fall together or is there some disanalogy between them?

(3) If they are analogous, is this so much the worse for the eternalist or so much the worse for the modal irrelevance objection?

(2) is the question I’m most interested in. I guess one reason you might think that the temporal irrelevance objection is worse off is that the eternalist can point to certain relations between the dinosaurs and us that justify our claim that the dinosaurs existed before us, whereas there is nothing analogous the Lewisian can do w.r.t. the talking donkeys.

Those relations would, presumably, be causal relations. But will that help? Can’t the presentist who is impressed by the temporal irrelevance objection simply reply that if some of the things that there are stand in causal relations to some of the other things that there are then – unless we have independent reason to think that the former things are past entities and the latter things present entities – we should conclude that causal relations can hold between presently existing entities, not that there is in fact non-present ontology?

Anyway, those are my ramblings for today. Thoughts?

In other news, I see that David Cameron has proposed that the salaries of GPs be tied to the health of their patients and patient satisfaction. Ignoring the obvious problems with such a stupid idea (such as it making it even harder to get GPs to work in deprived areas), I wonder if he’s going to take the obvious next step: to link MP’s salaries to the general level of Eudaimonia and the approval ratings of them by their constituents. Somehow, I doubt it.


Ian and Gilda said...

My answers to your questions would be:

(1) Not directly, as far as I can see. The closest thing I know of is the old objection that a B-theory of time makes time just another dimension of space and so not really time at all.
(2) It looks like we intuitively think of past and future times as more actual than other possible worlds even if sometimes people seem to treat them as less actual than the present. That all by itself should give us some pause, I think, when comparing the objections. One could also point the objection back at the presentist and argue that whatever truthmakers the presentist provides for claims about past or future, though indeed present, will still be of unclear relevance to what did or will happen.
Whether that would work or not, we could also point to the following disanalogy between times and worlds: One could argue that all these other times couldn't be present since they are not simultaneous and hence must really past or future depending on their B-relations to the present, but there's nothing similar for worlds (this, of course, could get us into tensed/tenseless territory, which would greatly complicate issues).
In addition, eternalist times (arguably) play the appropriate role common sense gives times and does so in a fairly-straight forward way. And because of the causal relations issue and other cross-temporal relations we seem to be committed to, we are already directly commited to such other times' existence (as well as past and perhaps future objects) in a way we are not for other possible worlds or merely possible individuals. So to insist that eternalist times are irrelevant to the temporal facts where they have all these features is just to be pig-headed about matters - what more could we possibly want to show that X is relevant to Y than is had here?
(3) If they are indeed analogous (which I don't think they are), I think I would treat this as so much the worse for the modal irrelevance objection (whose initial plausibility for various reasons has been lessened for me recently anyway).

Unknown said...

Nice comparison, Ross! Perhaps one thing that makes the temporal objection slightly worse is that talk about past objects in moodless present tense is more familiar than such talk about merely possible objects. Imagine a language where the only way to talk about past objects involves tense; "there are well-known Greek philosophers" could then only be interpreted true iff there are such philosophers now. So it is true by the rules of grammar (i.e. analytic) that if something existed or happened yesterday, then it does not exist or happen (simpliciter).

A disanalogy: unlike my concept of the past, my concept of the possible comes with strong extensional constraints; the idea that the world came into existence only 5 minutes ago is coherent, the idea that there are only 5 possible worlds is not. When one thinks of the Lewisian pluriverse, it is very tempting to regard it as a kind of extended actuality in the sense that it is somehow contingent whether those disconnected universes exist, and if so, how many. But then these universes really can't have anything to do with modality, because what is possible is not contingent.

This is of course not quite the modal irrelevance objection, but it may be something in the background that drives it. That would explain why the irrelevance objection is always directed at Lewis, and not at ersatzers -- even though merely possible worlds do exist on the ersatzer's view just as they does on Lewis's.

Ross Cameron said...


I guess I don't see how the eternalist can argue that these locations (neutrally described) are not simultaneous without begging the question. What did you have in mind?


Thanks - that disanalogy seems important.

I guess the presentist, if pushing the analogy, would have to claim that, even if there are (unrestrictedly) dinosaurs, for all we know there might not be in the future - in which case the unrestricted presence of dinosaurs can hardly be a consequence of the fact that there were dinosaurs. But that looks a lot less intuitive than the thought that even if there are all these cosmoi, there might not have been.

Steve said...

Hello Dr. Cameron. I'm a non-philosopher who enjoys reading philosophy papers. Related to your (1) I wanted to mention I saw this parallel discussed in a draft paper by Takeshi Yagisawa (links on this post of mine).
Best regards,
- Steve Esser

Louis said...

Hi Ross,

I think there are lots of irrelevance objections in the literature on
Lewis; but I would reconstruct the one you give here as follows.

Intuition supports:
(M1) If P is merely possible, then P is not the case.

But, according to Lewis's view, if P is merely possible, then P is the
case (elsewhere). So Lewis's analysis is incorrect (assuming that
there are merely possible P).

Interestingly, I think only de dicto necessities can be at issue here,
since Lewis would agree, e.g., that it is not the case anywhere that
Gore won the US presidential election in 2000. Hence your emphasis on
talking donkeys instead.

At any rate, I will focus on your question (2). If the reconstruction
above is correct, then the analogous argument against some forms of
eternalism would be:

Intuition supports:
(T1) If P is merely past, then P is not the case.

But, according to the Eternalist's view, if P is merely possible, then
P is the case (elsewhen). So the Eternalist's analysis is incorrect
(assuming that there are merely past P).

If I were an Eternalist, I would respond to the objection as follows:

"(T1)'s plausibility stems from the fact that "is not the case" is
naturally read in a tensed way. So read, (T1) can be paraphrased by:

(T2) If P is merely past, then P is *presently* not the case.

But my view does not require denying (T2). So any reading of (T1)
which is intuitively plausible is one which does not conflict with
my Eternalism."

This response seems to me very plausible, since, after all, "is the
case" is grammaticaly present tense, and paraphrases of present tense
sentences by insertion of "presently" is typically adequate. The
analogous response on Lewis's behalf seems to me much less plausible,
since there is no obvious modal analogue of the grammatical present
tense. A defender of Lewis might claim that there is a hidden "modal
index" in (M1), so that it is paraphrased as:

(M2) If P is merely possible, then P is *actually* not the case.

But, even granting the presence of a "modal index", the paraphrase
does not seem plausible, in view of the odd reading the Lewisian
needs to give "actually". (As you note, this isn't how we use the

None of this is to suggest that Lewis's view succumbs to the relevance
objection; I only mean to suggest that the view cannot be defended by
a certain way of accomodating the recalcitrant intuition (M1). The
Eternalist view, however, has a good shot at accommodating (T1).

Louis deRosset

Anonymous said...

Hi Ross,

First a question to Lewis who writes,

(M1) If P is merely possible, then P is not the case.

How does Lewis's modal metaphysics not support the intuition in M1? Denying M1 amounts to claiming that MP -> P.


The objection to Lewis is that everything that exists, (intuitively) actually exists. Plantinga and other ersatzists fuss over that. But is there any intuition that what exists also obtains? Plantinga certainly wants to say that every world exists at the actual world (indeed, at every world) and yet deny that every world obtains at the actual world.
So, for him, possible worlds as maximal and consistent states of affairs exist at the actual world but do not obtain at the actual world. Is the untutored intuition that P2 is true but not P1? It is not immediately obvious what it means for a state of affairs to exist but not obtain. But I guess this would show only that exchanging Lewis worlds for ersatzs worlds would not help much on this issue.

P1. For any maximal state of affairs P, if P exists then P obtains.

P2. For any maximal state of affairs P, if P exists then P actually exists.

Anonymous said...

Hi. Just a belated comment from someone passing by.

This is a good comparison being made between a complaint against Lewis' possible worlds and an eternalist view of time.

Louis made a good point when he said that only de dicto necessities are at issue for Lewis. Lewis uses counterpart relations to replace trans-world identity. So, it is, indeed, the case that Gore could not have won the 2000 election in any sense but that his counterpart could have.

I think all this comes down to Lewis' view that 'actual' is like an indexical term, meaning 'this world'. This means that there can be no counterfacutals for things whose reference is fixed by an indexical.

Translating that into the theory of eternalism, if 'now' is defined indexically as 'this time' then the eternalist can similarly privilege present times without suggesting that this is an ontological privileging. So, I think the comparison is apt but will meet the same response in each case.

(for the record, I am a presentist)

As for Wolgang's disanalogy, I think that the criticism of Lewis in mind here is precisely that there is no principled basis for saying that 'there are only 5 possible worlds' is incoherent if possible worlds are just closed space-times. I think that for Lewis and the eternalist, the objection can still be pressed that there is an intuitive confusion of categories going on: possibility with existence and space with time, respectively.

Again, this confirms the parallel. Though, I would not try to argue against these views in this way.