On Possibility

 David Brightly comments:

The view I've arrived at is that sentences involving 'possibility' can be re-written into sentences involving just 'possibly', and that our modal notions arise from our encounter with inference. I'm happy to say, There is the possibility that the bulb will shatter -- we say things like that all the time -- provided it's understood to mean, Possibly, the bulb will shatter. I certainly don't want to commit myself to things called possibilities, unless they can be seen as constructions out of sentences, roughly, Possibly, S ≡ The truth value of sentence S cannot be determined from what we currently know together with deduction from known principles.

Can you persuade me otherwise? A 'big topic' I would imagine!

Let B be an ordinary light bulb. Light bulbs are typically fragile: they are disposed to shatter if suitably struck or dropped from a sufficient height onto a hard surface. I take Brightly to be saying two things.  He is maintaining, first, that there is no more to the possibility of B's shattering in circumstances C than the truth of the sentence, 'Possibly, B will shatter in C.' Second, he is offering an analysis of 'possibly' in such sentences.

First Claim

I take Brightly to be saying that there is nothing in B, and thus nothing in reality, that could be called B's disposition to shatter. In general, unrealized possibilities have no ontological status. But then what makes the sentence 'Possibly, B shatters in C' true? Presumably, Brightly will say that nothing makes it true: it is just true.  He would not, I take it, say the same about 'B exists.'  He would not say that nothing makes 'B exists' true, that the sentence is just true.  I would guess that he would say that it is B itself, or perhaps the existence of B, that makes 'B exists' true.  So there is something in reality that 'B' names, and this item is, or is part of, the truth-maker of 'B exists.'  

But if he says this, should he not also admit that there is something in reality that make 'B is disposed to shatter in C' true? 

To appreciate the point one must see that a disposition and its manifestation are different.  B is disposed to shatter at every time at which it exists.  But it needn't ever shatter.  It might remain intact throughout its career.  Therefore, the reality of a disposition cannot be identified with its actual manifestation.  The same goes for powers and potentialities.  If a man has a power he never exercises, it does not follow that he does not have the power.  The potentiality of a seed to sprout in the right conditions is something real even if the seed remains on a shelf and its potentiality is never actualized.  

There is an epistemological question that I want to set aside lest it muddy the waters. The question is: How does one know de re, of a particular light bulb, that it is disposed to shatter if it never does?  I am not concerned here with the epistemology of modal knowledge, but with the ontology of the merely possible, which includes the ontology of unmanifested dispositions.

A disposition, then, is real whether or not it is ever manifested. But doesn't this just beg the question against Brightly? I maintain that unmanifested dispositions are real.  Brightly denies this. If I understand him, he is eliminating unmanifested dispositions in favor of the truth of possibility sentences.

My objection to this invokes the Truth-Maker Principle: truths need truth-makers.  Or at least many classes of truths need truth-makers, one of these being the class of truths about the powers, potentialities, dispositions, and the like of concrete individuals. (I am not a truth-maker maximalist.) My point against Brightly is that the sentence, 'Possibly, B shatters in C,' if true, is true in virtue of or because of something external to this sentence, namely, the unmanifested disposition in B to shatter.

My view is consistent with the view that unmanifested dispositions reduce to the so-called 'categorical' features of things like light bulbs. Unmanifested dispositions can be real without being irreducibly real. What I have said above does not commit me to irreducibly real dispositions.  It commits me only to the reality of unmanifested dispositions, whether reducible or not.

Second Claim

" Possibly, S ≡ The truth value of sentence S cannot be determined from what we currently know together with deduction from known principles."

S in Brightly's example is 'The bulb will shatter.' True or false? I grant that the truth value cannot be known from what we currently know together with what we can deduce from known principles.  But this cannot be what the possibility that the glass will shatter consists in. Brightly is making the very real possibility that the glass shatter, the bomb explode, the round fire, the cat scratch, Hillary throw a lamp at Bill, etc., depend on our ignorance.  But then real possibility is eliminated in favor of epistemic possibility.

Suppose Sally knows that Tom is in Boston now and believes falsely that Scollay Square still exists.  I ask Sally: is it possible that Tom is in Scollay Square now?  She replies, "Yes, it is possible." But of course this is a mere epistemic possibility sired by Sally's ignorance. It is possible for all Sally knows. It is not really possible that Tom is in Scollay Square now given that the place no longer exists.

I don't think we should say that the possibility of the bulb's shattering consists in our igntrance as to whether or not 'The bulb will shatter' is true or false. Consider also that long before minded organisms arose in our evolutionary history, and thus long before there was knowledge or ignorance,  there we seeds and such with real potencies some of which were actualized and some of which were not.  

Related articles

Bertrand Russell: Empiricism is Self-Refuting. Is He Right?

Nothing is Written in Stone

Could There Have Been Nothing at All?

Invention of the Light Bulb

Comments

[David Brightly]

Good Morning, Bill, and thank you for following this up.

My brief earlier comment may have too much of an epistemic flavour. Let me see if I can offer an analysis of your light bulb experiment that reduces this. Firstly, yes, I am reluctant to allow a 'disposition to shatter' in B. This loads the modality onto B rather than the circumstances C where I think it lies. Let's think of dropping B onto a hard surface and suppose this system is deterministic and non-chaotic. That is, it displays no extreme sensitivity to initial conditions. Suppose circumstances C are specified almost exactly---floor smooth and homogeneous, height of dropping fixed, zero initial angular velocity---but that the orientation of the bulb is not specified. Suppose also we have circumstance CH in which the bulb is dropped on its head and circumstance CS in which it is dropped on its side as in your illustration. Suppose also that the structure of B is such that in circumstances C&CH the bulb does not shatter but in C&CS it does---something to do with the degree of departure from cylindrical symmetry of the impact, perhaps. Both C&CH and C&CS fix the initial conditions exactly. In C&CH it's impossible that the bulb shatter on pain of breaking the determinism assumption. Likewise in C&CS it's necessary that the bulb shatter. But circumstances C do not fix the initial conditions exactly: C&CH satisfies C and C&CS satisfies C. C is too 'coarse-grained'. It allows initial conditions under which B shatters and initial conditions under which it does not. Hence, under circumstances C, Possibly, the bulb shatters is true. What makes it true is the course- grainedness of C rather than anything that might be thought of as a constituent of B.

[Opponent]

Sartre gives the example of a nail which has been bashed hard and becomes bent. Then using a vice it is straightened out again, so it looks like normal. So pace there clearly is a disposition inside the atomic structure of the nail that makes it highly likely to bend if used as a nail.

However this is still a possibility: the nail might remain straight and work according to is proper function and essential nature qua nail. But I claim the rest is

>> 'Possibly, B shatters in C'
Where C is the force of the hammer striking the nail, plus the hardness of the surface it is driven into, plus the weakness of the nail caused by the previous bending and re-straightening.

I say there is really an underlying necessity. If we know all these factors, we know precisely what will happen. But we don’t. So ‘possibly p’ means ‘p is consistent with the limited information available’.

This ties with my point about the contingency of identity. All possibility is epistemic. Everything, given enough information, is necessary. There is a coin in this box. Possibly heads is face up, possibly tails. Given only that information. Given the information that heads is face up, it is necessarily face up.

[Opponent]

David: Possibly, the bulb shatters is true. What makes it true is the course- grainedness of C rather than anything that might be thought of as a constituent of B.

I think I am agreeing with this.

[BV]

Good morning gentlemen,

Do you say write 'vice'? We write 'vise' although we pronounce it like 'vice.'

That is a useful comment, Opponent. Your overall position is now clearer to me. All possibility is epistemic! You seem to be staking out a position between mine and Mr Brightly's. You seem to be granting me my point about dispositions being real ingredients in things while also holding that possibility is epistemic.

But you may be falling into confusion. If determinism is true, the present state of things could not have been otherwise GIVEN what went before. So the light bulb that broke when dropped had to break, which implies that its fragility (disposition to shatter) had to be realized or manifested. But aren't there deterministic possible worlds in which this same light bulb exists with its disposition to shatter, but does not shatter because the antecedent conditions are different? There are worlds in which the light bulb exists and shatters and worlds in which it exists but does not shatter. Its shattering is therefore contingent, not necessary. This is consistent with saying that in the actual world the bulb HAD to shatter. For now we are talking about a conditional necessity not an absolute necessity.

Absolutely speaking, it is contingent whether the bulb shatter, but conditionally necessary given what went on prior to its shattering in our world, which we are assuming is deterministic.

Suppose determinism is true and we are omniscient. Then the bulb must shatter given prior events and we know that it must. But we are not omniscient. So it seems to us to be possible that the bulb not shatter given what went before. This is a mere epistemic possibility sired by ignorance. But this does not show that all possibility is epistemic. For there is the absolute possiblity, grounded in the bulb, that it not shatter. This is real, not epistemic possibility.

[Opponent]

>> we pronounce it like 'vice.'
This is why we spell it ‘vice’ :)
Apparently the engineering type of vice comes from Latin vitis, a vine or corkscrew appearance. The other type (as in vice squad) comes from vitium (failing or defect). This is different again from as in ‘vice-President’, from vice ‘in the place of’. I have learned all this today.

>> All possibility is epistemic!
Change that to ‘all possibility is semantic’. The fact that p (the current evidence) is consistent with ~q, so ‘contingency’ means ‘logically consistent with existing evidence’. If more evidence p* comes to light, this may not be consistent with ~q, so q is no longer contigent.

>> But aren't there deterministic possible worlds in which this same light bulb exists with its disposition to shatter, but does not shatter because the antecedent conditions are different?
There are no possible worlds for us nominalists. Everything is actual, but there are modal propositions. The modal proposition ‘the antecedent conditions might have been different’ translates as its being p* is consistent with ~q’, where ‘being p*’ specifies a state which does not cause the bulb to break.

This is connected with a long and complex debate in probability theory about 'semantic probability'.

[BV]

David,

Excellent comment. Clearly, whether or not the bulb breaks will depend on the exact circumstances. And I grant you that underspecified circumstances will leave it open whether the bulb breaks or not.

>>Hence, under circumstances C, Possibly, the bulb shatters is true. What makes it true is the course- grainedness of C rather than anything that might be thought of as a constituent of B.<<

Your view seems close to that of the Opponent. Real possibility does not come into the picture at all. The shattering is either impossible or necessary. It is our ignorance of the exact circumstances that allows the illusion of possibility to arise.

But now I think my response to Opponent applies also to you.

[Opponent]

>>Real possibility does not come into the picture at all.

Of course ‘real possibility’ comes into the picture! It just means something different from whatever you think it means. ‘It is possible that q, given p’ means that q is consistent with p. ‘It is necessary that q, given p’ means that ~q is inconsistent with p, and so on.

>>The shattering is either impossible or necessary.

Or possible, as noted.

[BV]

G. Molnar puts Megaric Actualism like this:

MA. At time t an object x has the power to F iff x exercises the power to F at t.

Please tell me whether you accept or reject (MA) and why. I would like to hear from both of you.

[Opponent]

Ah right, Megarians say a thing can only do what it is actually doing. There is no inactive power.

Disagree. The example of the bent and straightened nail proves there is inactive weakness (active weakness would be the nail bending under the hammer). However this is perfectly consistent with the points I made above.

[David Brightly]

I'd say something like this: In so far as it is deterministic, the flux has no modality. 'Possible' and 'necessary' don't apply. From our human perspective, which partitions the flux into medium-sized objects, events, and circumstances, there is modality for the reason above of coarse-graining, it's just as real as the objects etc are, and, again from our perspective, the world is not deterministic. On top of this there is epistemic modality where we are ignorant of how the world is, even in the terms of our own perspective.

I read MA as can --> does and does --> can. I accept only the latter.

[BV]

Here is one question, David. You say >>In so far as it is deterministic, the flux has no modality.<<

By the flux you apparently mean space-time and its contents as viewed by physics. Now if the flux is deterministic, then each total state of the flux at a time could not have been otherwise given the preceding states. But this is just to say that each state is necessitated by the preceding states. But this causal necessitation is clearly a modal notion.

So I cannot read your quoted sentence above except as self-refuting.

[David Brightly]

No interesting modality. Just the trivial modality of necessity and impossibility with none of the merely possible that makes modality a puzzle.

[Astute opponent]

All the available information about the flux, absolutely everything, to the very finest granularity, is inconsistent with any other information. Whereas the truth of 'the phone is on the desk' is consistent with many fluxes.

Of course no one is denying 'modal notions', just as the materialist does not deny 'mental states'.

[BV]

Not so, David. Determinism is the view that past events plus the laws of nature make only one of the many merely possible futures actual, and thus (conditionally) necessary. You can't talk about necessity here without bringing in possibility.

>>In so far as it is deterministic, the flux has no modality<< is self-refuting.

You need to drop talk of determinism the definition of which involves modal notions.

[David Brightly]

Fair enough. Let me just require that the relation between the state of flux at an earlier time to that at a later time be functional. That is, at worst many-to-one, but never one-to-many. My intuition of an 'indeterministic' flux would allow a one-many relation. I can then revert to my original suggestion that there is no modality in the flux.