The Generalized Ought-Implies-Can Principle and Novák's Objection

This entry is an addendum to my Prague paper (see link below) in which I deploy a principle I call GOC, a principle that comes under withering fire in the ComBox from Dr. Lukáš Novák.  Here is my reformulation of his objection.  You will have to consult my Prague paper to see what I mean by 'really possible.' Neither of us are metaphysical naturalists, but we are assuming naturalism to be true for the sake of this discussion. The burden of my Prague paper is to show that metaphysical naturalism is not logically consistent with David Benatar's claim that "while some lives are better than others, none are (noncomparatively or objectively) good." (The Human Predicament  67)

1) Necessarily, if a state of affairs S ought to be, then S is really possible. (GOC)

2) That no child starves is a state of affairs that ought to be. (Novak's plausible premise. It is supposed to hold whether or not naturalism is true.)

Therefore:

3) That no child starves is really possible. (1, 2)

But:

4) That no child starves is not really possible on naturalism. (Premise I share with Novak:  e.g., a child who is the sole survivor of a shipwreck washes ashore  on a deserted island where there is no food.)

5) (3) and (4) are mutually contradictory.

Therefore, by reductio ad absurdum,

6) Either (1) is false or (2) is false or (4) is false.

7) (2) and (4) are both true. (Novak assumes)

Therefore

8) (1) is false.

How might I respond? Well, I agree that (4) is true.  And I have a separate argument for (1). So I argue that, on naturalism, (2) is false.  Thus I argue:

1) Necessarily, if a state of affairs S ought to be, then S is really possible. (GOC)

4) That no child starves is not really possible on naturalism.

Therefore

~2) It is not the case that on naturalism no child's starving ought to be.

This is the analog of the cases of the ought-to-do in which an agent cannot do X. If an agent cannot do X, then it is not the case that he ought to do X. 

Related articles

Is the Quality of Life Objectively Evaluable on Naturalism?

Comments

[Lukáš Novák]

Bill,

for simplicity's sake, let's not talk of any child's starving but of this child's starving.

I say that it is obvious that it would be better if that child did not starve. It is the fact that it would be better that renders rational sustained attempts of any other person involved to obtain some food for the chiled -- irrespective of whether as a matter of fact theese attempts are doomed or not. I would say that anyone who doubts this suffers from a serious malfunction of value perceptivity (Wertblindheit).

Therefore, I say that on this reply of yours, BOB fails, as not everything that obviously would be better to be ought to be.

Now you might say: I concede that as a matter of fact the child ought to be fed: but I say that on naturalism it must be denied - but this is not a problem for me but for naturalism.

I don't think such a reply would help your case. For the reason why on naturalism it is not the case that the child's starving ought not be is that on naturalism there is no God (for God, if He existed, is the only agent that could feed the child). So, your position would entail that we can very easily deduce the existence of God (or a similarly powefur being) from the mere obviousness of the objective desirability of naturalistically unachievabe states of affairs. But this is very doubtful: it shows that you are proving too much, or that your principles are too strong, or at least not so obvious as to be accepted without argument. Moreover, to imply that it is the existence of God is what makes the child's being fed truly objectively desirable is quite strange: for our insight that to feed the child is what ought to be done in this situation does not and need not involve God at all: we can know what ought to be done without knowing that God exists. This suggests that it is not God's existence/non-existence what makes the difference, after all, i.e. that objective desirability of a state of affairs does not in fact require an actual agent actually capable of actualizing it.

Or you might bite the bullet and deny that it would be better that the child be fed. Then I say: there certainly is some sense of "better" according to which it would be better; and most ordinary people would prefer this sense as the natural one. And according to this, arguably natural sense of "better" and "worse", Benatar would have his point; because for this sense of "better" BOB would fail, and so your entire objection.

[BV]

Lukas,

Thanks for the discussion.

What you may be failing to grasp is that I distinguish the ought-to-do from the ought-to-be. This is no innovation of mine. You can find the distinction in Nicolai Hartmann and Max Scheler.

There are states of affairs that ought to be even if there are no finite agents with the power to bring them about, which implies that there are no finite agents who are morally obligated to bring about these states of affairs.

It would be objectively better for the child not to starve than to starve. From this I infer that the child's not starving objectively ought to be instead of the child's starving.

For some reason you balk at this. Why? I suspect it is because you use 'ought' only in connection with the ought-to-do.

[Vlastimil]

Greetings from my vacation, gentlemen. I am grateful for the discussion.

Bill made a distinction btw different senses of 'ought' also in
http://maverickphilosopher.typepad.com/maverick_philosopher/2017/02/the-ethical-dilemma-of-sebastian-rodrigues.html?cid=6a010535ce1cf6970c01bb097585a5970d#comment-6a010535ce1cf6970c01bb097585a5970d

Bill did so after I quoted there Brentano against something like GOC.

Apart of GOC, already Bolzano, in his Wissenschaftslehre, distinguishes what (i) ought to be/done in general (which includes, inter alia, what is supererogatory) from that which (ii) ought to be done because it's obligatory.

More later, I hope.

[Lukáš Novák]

Bill,
there must be some huge misunderstanding. No problem with the ought-to-be concept (and its distinction from ought-to-do) with me. This is not my problem at all. Let me try to put it as clear as possible:

Your argument is based on this chain of implications [SoA = state of affairs]:

[B>O]   For any SoA X, if X is better, then X ought to be.
[O>A]   For any SoA X, if X ought to be, then X is actualizable.

where "better" means "better than the actual state of affairs" and "S is actualizable" means "there actually is an agent that is actually capable of bringing about S".

I say: there is a SoA S (e.g., the starving-child-on-the-island SoA) that is evidently better but, at the same time, evidently actualizable only by God. Therefore, either we have an easy generator of evident proofs of God's existence, or there is something wrong with the implication chain (note that this objection has nothing to do specifically with naturalism!).

I opt for the latter. I say that either there is a fallacy of equivocation in the term "ought to be" (it has a weaker meaning in [B>O] but a stronger meaning in [O>A]), or one of the implications is false (viz., [B>O] on the stronger reading of "ought to be", [O>A] on the weaker reading of "ought to be").

Or, if you tried to stipulate that [B>O] is true on the stronger meaning, I retort: this amounts to stipulating the meaning of "better" against common usage. The stipulated "better" either induces God's existence or vastly reduces the domain of "better" states of affairs (exluding the feeding-of-the-starving-child SoA and the like). I don't question the legitimacy of this stipulation (any stipulation is legitimate), but I question the legitimacy of imputing thus stipulated concept of "better" to Benatar. Because there always is the common notion of "better" that is meaningful and does not induce God's existence in case of contingently unactualizable states of affairs; and Benatar can always make use of it. Moreover, I deny that in this stipulated strong sense of "better" it is evident that it is better that the child be fed (i.e. the theistic-proof-generator cannot be stipulated to function).

In other words, the stipulated meaning of "better" would be irrelevant: what is relevant is the meaning of "better" which makes statements like "I't would be better to feed the child", "It would be better for Peter not to have cancer" etc. evidently true; and this meaning, precisely because it makes this kind of statements evidently true irrespective of God's existence or our knowledge thereof, does not imply God's existence in any easy way. Moreover, these statements are also evidently true irrespective of the existence of any agent capable of actualizing the respctive SoA, or our knowledge thereof; therefore, they don't imply thus easily the existence of any capable agent either.

[BV]

Vlastimil,

Here is the Brentano passage from The Origin of our Knowledge of Right and Wrong. Brentano writes in the note On the Charge of Excessive Rigorism:

"It is certain that no man can entirely avoid error. Nevertheless, avoidable or not, every erroneous judgement is a judgement that ought not to have been made, a judgement in conflict with the requirements of logic, and these cannot be modified. The rules of logic are not to be given up merely because of the weakness of our powers of reasoning. Similarly, the rules of ethics are not to be given up because of weakness of will. If a man is weak willed, ethics cannot cease to demand from him that he love what is known to be good, prefer what is known to be better, and place the highest good above all else. Even if one could show (and one cannot) that there are circumstances under which no one could remain true to the highest good, there would not be the slightest justification for setting aside the requirements of ethics. The one and only correct rule would remain evident and unalterably true: Give preference in every case to that which is better."

Are you taking this as an argument against Kant's principle of ought implies can? As follows. No erroneous judgment ought ever to be made despite the fact that we sometimes cannot avoid error. If so, then my inability to avoid all error does not ever ansbolove me of the moral obligation alway to avoid error.

Now that is absurd by my lights.

I say it forces a distinction between the ought-to-do and the non-agential ought to be.

Brentano is right when he says, "every erroneous judgement is a judgement that ought not to have been made," but only if we read the 'ought' as expressing a non-agential ought.

Here is another example to think about: It ought to be that God exists.

This is true, but it is in no one's power, not even God's, to bring it about that God exists.

This is an example of a non-agential ought-to-be.

[BV]

As for the supererogatory, that falls with the sphere of what an agent can do. So it is not part of the non-agential oguth-to-be.

Another example: it ought to be that there are free agents. No agent, not even God, has the power to bring that about.

[Lukáš Novák]

Bill and Vlasta,

there are two distinctions going across each other:

1. (a) the ought of obligation vs. (b) the ought of mere counsel (i.e., of the superrogatory)
2. (c) the agential ought vs. (d) the non-agential ought.

It seems that all the four combinations are possible:

(c.a) A man ought not to fornicate.
(c.b) A Catholic ought to receive the Eucharist frequently.
(d.a) Totalitarian regimes ought not to exist.
(d.b) Mistakes in logic ought not to exist.

[Vlastimil]

Bill,

Your reply to Brentano is reasonable. To test it further we'd need to explore what 'agential', 'moral', and 'obligatory' mean.

Yes, the supererogatory is agential ought. Your distinction is different from the one I mentioned. (Also not sure Bolzano puts it as I did. Will check when back at home.) Perhaps you should clarify, however, whether agential and non-agential oughts can overlap, or whether they are mutually exclusive. Just consider: it ought to BE that I do what ought to DO.

[BV]

Lukas mounts a powerful objection:

>>Your argument is based on this chain of implications [SoA = state of affairs]:
[B>O] For any SoA X, if X is better, then X ought to be.
[O>A] For any SoA X, if X ought to be, then X is actualizable.
where "better" means "better than the actual state of affairs" and "S is actualizable" means "there actually is an agent that is actually capable of bringing about S".
I say: there is a SoA S (e.g., the starving-child-on-the-island SoA) that is evidently better but, at the same time, evidently actualizable only by God. Therefore, either we have an easy generator of evident proofs of God's existence, or there is something wrong with the implication chain (note that this objection has nothing to do specifically with naturalism!).<<

We agree that the child's not starving is better than the child's starving. And we agree that there is no finite agent that can bring about the child's not starving. We also agree that on naturalism there is no God. So we ought to agree that on naturalism there is no agent that can bring about the child's not starving.

I then argue that if S ought to be, then S is (really) possible, and if S is possible, then S is possibly actual, i.e., actualizable, and if actualizable then there exists an agent with the power to actualize S.

Lukas points out, correctly, that this amounts to an argument for the existence of God if we assume that the child's starving ought not be.

But that is hardly an objection to my chain of reasoning. Lukas will agree that, in general, one cannot disqualify as invalid a chain of reasoning just because it issues in the proposition that God exists.

So I put the question to Lukas: Do you object to the move from ought-to-be to possibility? Or the move from possible to possibly actual? Or the mover from possibly actual to the existence of an actualizer.

How does this relate to Benatar? His thesis is that the quality of human life, for all, is objectively bad. He tries to show this by pointing out all sorts of limitations we must live with. For example, we cannot 'breathe water.' He thinks it would be objectively better if we could both breathe air (via lungs) and breathe water (via gills). But that is not possible given the laws of nature. And so I infer by GOC that it is not the case that it ought to be. Therefore our limitation in respect of getting oxygen cannot be part of a case that shows that human life, for all, is objectively bad, so bad, in fact, that anti-natalism is the upshot.

Up to his point, all my argument needs in the GOC principle. So I ask Lukas: do you have an argument against it?

[BV]

Vlastimil writes,

>>Perhaps you should clarify, however, whether agential and non-agential oughts can overlap, or whether they are mutually exclusive. Just consider: it ought to BE that I do what ought to DO.<<

They overlap. From 'you ought to feed your kids' one can infer 'it ought to be the case that you feed your kids.'

For every ought-to-do there is a corresponding ought-to-be. But it is not the case that for every ought-to-be there is a corresponding ought-to-do where the agent is finite.

But things change if we count God among the agents.

[Vlastimil]

"things change if we count God among the agents"?

To a degree, I gather. For, Bill, you also said here, "it ought to be that there are free agents. No agent, not even God, has the power to bring that about."

[Vlastimil]

"all my argument needs in the GOC principle."

Bill, you also need AA or PPA. That's one crucial issue for your paper.

Besides the in/compatibility of pessimism and naturalism, you also need to explain why Benatar cannot be a Platonist like Wielenberg is, or even a nominalist like Swinburne is. That's another crucial issue.

Anyway, your paper is a great illustration of how analytic and even scholastic metaphysical themes are relevant to existential and even emotionally pressing worries.

[Vlastimil]

Going back to the in/compatibility of naturalism and pessimism, you also need to explain why even given PPA Benatar cannot be a PPA-naturalist that alludes to quantum chances (and the quantum statistical nature of natural laws that allows for various flukes). Lukáš pointed this out in one of his comments to the previous post of yours.

[BV]

Vlastimil,

I said "Up to this point, all I need to the GOC principle."

If it ought to be that we have both lungs and gills, then it is possible. But it is not possible given the laws of nature. Therefore, it is not the case that it ought to be that we have both lungs and gills.

But then it is no argument against the quality of our lives that we don't have both lungs and gill.

One way you could help me is by sketching how Platonism and nominalism could help Benatar.

[Lukáš Novák]

Bill,

my reply will be inevitably complicated.

Of course, given that I believe that God exists and that He exists necessarily, I must concede that, as a matter of fact, possibility implies the existence of a capable actualizer -- for the trivial reason that a necessary proposition is implied by anything, and the proposition that an omnipotent actualizer exists is necessary.

I do not question the logical nexus here but the epistemical nexus. Although God's existence is indeed implied in this easy way, it cannot be made evident this way. It is not immediately evident that there must be an actualizer for any evidently objectively desirable state of affairs. I think it can be made evident by means of extensive argumentation that will demonstrate the necessity of God's existence through the analysis of metaphysical (broadly logical) possibility as such. But without such an enterprise, GOC is simply unsubstantiated (gratis assertum gratis negatur): one can be epistemically justified in accepting objective desirability of a SoF as evident and yet doubting or rejecting God's existence. The evidence of the desirability is not immediately liked to the existence of a capable actualizer; rather, it stems from the insight into the natural powers and inclinations in the things themselves. A child has a nature that requires to be sustained by means of food; therefore, when starving, the child is deprived of what is its natural good. This is what we perceive and why we understand that a starving child ought to be fed, irrespective of circumstances. The existence or non-existence of an agent capable of bringing about that desirable end is wholly extraneous to this consideration -- unlike the purely metaphysical possibility of such end, i.e. the metaphysical compatibility of "being fed" with the child, which is, indeed, inextricably connected with that consideration even on the epistemic level. And as I said: it is possible to derive actualizability from metaphysical possibility, but in an extremely complicated and non-trivial way, which amounts to a demonstration of God's existence.

You see where the problem is: it seems to me that epistemically, you cannot substantiate GOC without an independet demonstration of God's existence; and to say that it is obvious just begs the question. But once you have demonstraded God's existence, there is no need to be bothered about antinatalism any more.

Put in more general terms: I am skeptical about the possiblity to argue out such fundamental existential questions like whether life is worth living at all or not, with the question of God's existence or non-existence bracketed. It is like trying to foretell who will be the next president of Russia without knowing whether Putin is going to run.

[Vlastimil]

Lukáš,

"once you have demonstrated God's existence, there is no need to be bothered about antinatalism any more"

Not so fast. Even given God, perhaps there is no sufficiently good afterlife or the chance of achieving it isn't sufficiently high. Cf.
https://drive.google.com/file/d/0B3Ad4KbZcW-rcGNhZV9KZDZRODg/edit
http://maverickphilosopher.typepad.com/maverick_philosopher/2017/05/prudential-anti-natalism.html

To address this worry requires a complicated argument, too.

[Vlastimil]

Check also this, Lukáš. I invited the author to the conf, but he could not fund his travels.
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2616888

[Vlastimil]

Bill,

I guess Benatar will reply to you at the conf like this:
- my pessimist argument does not commit me to any particular modal metaphysics, it is simply silent about modal metaphysics
- perhaps you showed that it commits me to avoiding this particular combination: GOC, AA/PPA, no quantum flukes, pessimism, naturalism
- so what?, there are many other epistemically possible combinations (some including Platonism or nominalism), and the burden is on you to show why none is plausibly available to me

[BV]

V,

I am not trying to refute Benatar since I don't believe one can refute a philosophical theory, except for a couple of silly and sophomoric ones. 'Refute' is a verb of success: If I refute your theory, then I show conclusively that it is untenable.

You probably don't accept my metaphilosophy.

I am merely trying to show ONE way in which Benatar's view can be rationally rejected, or reasonably opposed.

Also, as you may recall, I reject burden-of-proof considerations in philosophy. They are crucial in courts of law nd perhaps in formal debates. But as I have said many times, philosophy is not debate.

[Vlastimil]

Bill,

You need not try to render your conclusion strictly conclusive. You can just try to bolster it by explaining whence your focus on your combination of premises (GOC, AA/PPA, no quantum flukes, perhaps even non-Platonism?) -- why it is more plausible or probable than rival combinations.

[BV]

Lukas writes,

>>I am skeptical about the possibility to argue out such fundamental existential questions like whether life is worth living at all or not, with the question of God's existence or non-existence bracketed.<<

But if you don't bracket the ex/non-ex of God, then you beg the question.

I think you hold that if God exists then life is worth living and propagating, and that ff God does not exist then it is not worth living and propagating. Right?