It would be, indeed. But I take the view that the hedonic calculus should only be applied subject to the imperative to treat sentient beings as ends, and handing the kid over for torture would be treating him or her as a mere means, not an end.

The moral thing to do is what it has always been in such situations, e.g. in 1939 when we were made a not dissimilar offer by the Nazis: you fight the bastards. — Herg

I agree, but let's tweak the scenario a bit. Suppose that instead of wanting a kid to torture, the aliens really really want to play Magnus Carlsen in chess. If he agrees to play, humanity gets gifted technology. If he refuses, we all get sent to the salt mines, except Magnus. The aliens like him and will gift him a good life no matter what he decides. But Magnus refuses to play! His ego is such that he would rather the world burn then being coerced into a game. Should humanity force Magnus to play? Maybe by threatening to execute him if he doesn't? Wouldn't that be treating him as a means?

- Good post.

The point is that not turning the wheel is just as much a choice as turning it. — Herg

This is exactly right. As I said:

The problem with this analysis is that to discretely decide to continue sitting is an act. It is a decision to omit standing. If someone decides to omit an act then they have acted, by deciding to continue in the course they were already in. But again, I am going to leave this omission vs. commission question for another day. — Leontiskos

Now I would say that a decision is an act, but the larger point is that we are responsible for our decisions, and therefore we are responsible even for decisions that do not terminate in (external/transitive) "acts."

NO! You lack a distinction between letting something bad happen and doing something bad. My action was to give the first person water to save them; and the simultaneous deprivation of water (from that act) of the second person is an inaction: it is a negative counter-part to a positive. You do not seem to have fully fleshed out this kind of distinction, and instead insist on everything being an action.

I intentionally let, in this example, the second person die: I did not kill them. This is morally permissible because (presumably) there was no morally permissible way to save them. — Bob Ross

Strictly speaking I admit a distinction between commissions and omissions, but I would say your theory simply cannot support this distinction in our current context. This is because your only tool of analysis is conditional necessity. "I let them die," is (arguably) equally applicable to all of our other cases. For example the lever-puller will say that they merely let the single person die in order to save the five. If we analyze these cases according to conditional necessity, then each case is the same insofar as you perform an action in which you "intend" and "cause" the necessary death of someone (according to your own claims). To intend and cause the death of someone is to intentionally kill them. It makes no sense for you to say that you intend and cause their necessary death but you do not kill them. Some of this will hopefully clear up when I respond to your full post.

Please see me and Leontisk conversation which I am responding to here as well. I addressed this with their diagrams above. — Bob Ross

Herg's analysis was quite good, and should be addressed. He is making the same claim I am making but doing it in a more natural way. As he says, "whether Alan and Betty are present has no bearing on whether the car can swerve and hence save the lives of Charles and Dora."

Edit: If Bob Ross were right, then one would have a magic get-out-of-jail-free card for any moral quandary: just do nothing! For example, if a pilot (like Sully) found himself in a difficult situation, then he would just do nothing and apparently on Bob's theory he could never be blamed. This is incorrect because it is often wrong to do nothing (and the choice to do nothing is often the wrong choice to make).

Ok, good points.

By “means”, I mean “a necessary utility expended to produce an desired outcome”.
By “intention”, I mean “a purposeful or deliberate course of action”.

Because I accept the premise that “if I cannot achieve A without causing B, then I cannot intend A without intending to cause B”, I reject the premise I previously accepted (that “If I intend A and A causes effects P and Q, then I do not necessarily intend Q, even when I know that A causes Q”). — Bob Ross

Okay, so you are doubling-down and rejecting the premise.

Now I should say again that your definition of a "means" is not a good definition. I understand the word itself better than I understand the word-salad definition you are giving. Because your definition is less clear than the word, it is a bad definition. If you don't like 's definition then I suggest we take Merriam-Webster as a point of departure, "something useful or helpful to a desired end."

Correct; and I like your diagrams for explaining it. I understand that, from your perspective, P is not a means to Q in V because if one removes P from A then Q is still connected to A. The problem is that when Q and P are necessary utilities for bringing about A, then A cannot exist without being connected to Q and P: you cannot remove P and keep Q connected to A—A disappears. — Bob Ross

One of the fundamental problems here is that your account is unable to distinguish between the "V" and the "7" in my diagram. So then either 1) There are no relevant moral differences between "V" situations and "7" situations, and therefore your inability to distinguish them is unimportant, or 2) There are relevant moral differences between "V" situations and "7" situations, and therefore your inability to distinguish them is problematic. Note that in both "V" and "7" Q is conditionally necessary for P (i.e. P → Q).

I understand that, from your perspective, P is not a means to Q in V because if one removes P from A then Q is still connected to A. — Bob Ross

Yes, but let's continue referring to P as the good effect and Q as the bad effect. So it would instead be, "Q is not a means to P in 'V' because if one disconnects Q from A then P is still connected to A." Note that Q is not a means because it does not mediate A and P.

The problem is that when Q and P are necessary utilities for bringing about A, then A cannot exist without being connected to Q and P: you cannot remove P and keep Q connected to A—A disappears. — Bob Ross

Let's get back to fundamentals again. There are two basic principles in conflict:

• P_dfs: The principle of the diffusiveness of intention
• P_ndiv: The principle of the non-divisiveness of intention

The above account of indirect objects of intention seems to correspond to what Chisholm calls the principles of the ‘diffusiveness’ and ‘non-divisiveness’ of intention. The principle of the diffusiveness of intention is that if someone intends p, knowing that p entails q, then he intends the conjunction p and q. For instance, if a man intends to drive off in a car parked nearby, knowing that it belongs to someone else, then he intends this: to drive off in a car parked nearby that belongs to someone else. The principle of the non-divisiveness of intention is that if someone intends the conjunction p and q, he does not necessarily intend q by itself. The man may not intend just this: to drive off in a car that belongs to someone else.[22]

[22]: Roderick Chisholm, "The Structure of Intention." — Stephen L. Brock, Action and Conduct: Thomas Aquinas and the Theory of Action, Ch. 5: Praeter Intentionem, p. 204

(I take it that Chisholm's contemporary approach will be more accessible.)

You have a good grasp of P_dfs but you are letting it override P_ndiv. Let's go back to my example that you did not respond to:

Suppose an act has two effects: I press the accelerator and two things happen: my speed accelerates and my fuel is consumed. Do you think Aquinas is right, and it is possible to intend one effect without the other? When I press the accelerator do I intend to accelerate and do I intend to diminish my fuel? And even if so, do I intend them in the same way? Is the word "intend" being used in the same way for both of these effects? — Leontiskos

You seem to be committed to the position which says that when I drive to the grocery store I am intending to consume fuel. This isn't correct. I intend to get groceries, and the consumption of fuel is a side-effect that I would prefer not occur. I only assent to it because I don't know how to get to the grocery store without consuming fuel.

Brock uses the terms "direct intention" and "indirect intention." I directly intend to get to the grocery store; I indirectly intend to consume fuel. Key here is the relation of each to my purpose: if my direct intention does not come about then my purpose has not been achieved, but if my indirect intention does not come about it does not follow that my purpose has not been achieved. If for some reason fuel is not consumed during my trip, the trip is successful all the same. If I get back to my home and see that the fuel has not been diminished, I do not say, "Oh no, my purpose was not achieved!" Yet I do precisely this if I get home and realize that I have no groceries in the car. Direct and indirect intention are different, and any theory that cannot make sense of their difference is inadequate as a theory.

I would suggest reading Aristotle's account of the sailors who throw cargo overboard in Book III, Ch. 1 of the Nicomachean Ethics (link). It is very often referred to in this area of philosophy. A key element of that story is that the cargo is thrown into the sea even though doing so is undesirable and not directly intended.

P ← A → Q

The problem is that this doesn’t completely represent the relationship whereof Q and P are necessary for A. I don’t know how to represent it this way, but in logic it would be “A → (P & Q)”: you can’t say that A → P is true when Q is false given “A → (P & Q)”. — Bob Ross

Right: this is the antecedent of (2) of the argument I gave: < 2. (A → (P ^ Q)) → (A → (P → Q)) >.

This premise is false in the examples we have given: A entailing P and Q does not imply that A entails that P entails Q—you are thinking in terms of the 7 diagram instead of a reciprocated V diagram (where A cannot exist without being connected to both Q and P). — Bob Ross

On the contrary, the premise is tautological, and is therefore always true. It is nearly the logical representation of what you said just above, "you can’t say that A → P is true when Q is false given 'A → (P & Q)'." The point is that you cannot say that (P → Q) is false when (P ^ Q) (i.e. given A, both follow). Remember that the key is the relation between P and Q.

Take the car example, I am actually saying (as opposed to this premise 2):

{A → (P ^ Q)} → {![(A → P) && !(A → Q)]} — Bob Ross

Sure, that's fine. It is true that (A → (P ^ Q)) → ((A → P) ^ (A → Q)). But again, to know whether Q is a means to P we must understand the relation between P and Q, and the conditional necessity that you are betting all your chips on is represented by (P → Q).

Now, an easy way to understand the flaw in your reasoning is to reflect on your 7 diagram; because it implies that if you were to remove P that Q is still connected to A, which is not true, for example, in the case of the car swerving example. If A is “swerving to save four people” and P is “saving four people” and Q is “hitting two innocent bystanders”, then removing P does not result in A → Q. Viz, if the intention is to save four people, then the effect of hitting two innocent bystanders is no longer connected with the intention in the event that there are no four people to save. — Bob Ross

So first, A is not "swerving to save four people." Saving four people is impossible, and we do not intend the impossible. Here is what I already said:

A: Swerve right
P: Avoid hitting all four people
Q: Hit the two people on the right — Leontiskos

Now if we are aiming at Q then we don't need to achieve P (although we could). So we can remove P without removing Q. But in fact we are not aiming at Q; we are aiming at P, and Q is not a means to P as so helpfully demonstrated.

Fair enough--except for the killing part: it is not a killing. Yes, from what I have said, and I did not catch it on my last response, it follows that letting the second person die is a means towards saving the first person: I accept this. — Bob Ross

Okay, and I addressed your claim about "killing" in my post above. As I said, that seems to be a tangent. The central point has to do with whether Q is a means to P. If you want to say that Q is still intentional killing even in the "V" case then that is a separate argument from the claim that Q is a means to P.

The other absurdity that results from your view is that Q is a means to P and P is a means to Q, which is of course impossible. This is because in (2) (P ^ Q) is interchangeable with (Q ^ P), and what this means is that A leads both to the conclusion that Q is a means to P and to the conclusion that P is a means to Q. In other words: (A → (P ^ Q)) → (A → (P ↔ Q)). — Leontiskos

That’s exactly what it means for a thing to be dependent on two other things that are independent of each other. Operating this computer with two screens requires two different monitors. Monitor 1 and Monitor 2 are independent of each other analogous to the V diagram, but A (which is the operation of the computer with two screens) requires both. If you remove 1 or 2, then you cannot operate the computer with two screens (A); and it is not like the 7 diagram either: the monitors are independent of each other and mutually required for A. — Bob Ross

No, not at all. Some points:

• A "means" is asymmetrical. If A is a means to B, then B cannot be a means to A. Two things cannot be a means to one another.
• Monitor 1 and Monitor 2 are not a means to one another, nor is either one a means to the other.
• The operation of the computer does not require two screens.

Now you want to make A "Operation with two screens." Erm, okay, but then this would be the "two birds with one stone" scenario, not the (Car → (Groceries ^ Gas)) scenario. Remember that we are talking about double effect, where P is a good effect and Q is a bad effect. You are talking about a different thing, where there are two good effects (two effects directly intended). The computer scenario and the car scenario are both "V" scenarios, but the crucial point is that the intention is different in each of them. Your desire to have a second monitor is not parallel to my desire to consume fuel. You want a second monitor and I don't want to consume fuel.

The question of a means can be represented by a particular sort of counterfactual, "Supposing !(A → Q), would I still choose A?" On "7" scenarios the answer is no, whereas on "V" scenarios the answer is yes. Again, this is not to deny that (A → Q) might dissuade me from choosing A, but it does mean that whenever I answer 'yes' to the counterfactual it follows that Q is not a means to P.

I would prefer we just refer to it as intentional killing vs. killing: that is the most clear way of presenting it. — Bob Ross

Yes of course you would, for that would aid your position. Likewise, I would prefer if we just refer to it as killing vs. non-intentional killing. So again, I suggest we not use the word "killing" without a prefix.

Suppose that instead of wanting a kid to torture, the aliens really really want to play Magnus Carlsen in chess. If he agrees to play, humanity gets gifted technology. If he refuses, we all get sent to the salt mines, except Magnus. The aliens like him and will gift him a good life no matter what he decides. But Magnus refuses to play! His ego is such that he would rather the world burn then being coerced into a game. Should humanity force Magnus to play? Maybe by threatening to execute him if he doesn't? Wouldn't that be treating him as a means? — RogueAI

First of all, let's be clear: there is nothing wrong with treating someone as a means, provided you also treat them as an end.

I don't think the threat to execute him is your best option. Ex hypothesi, we are not going to follow through on that threat, and he may know that, so the threat may not work. Suppose we hypnotise him instead? Then his will is under our control, and we can make him play.

Clearly we are then using him as a means. The question is, can we also treat him as an end?

Many people would no doubt say no, because in effect we are treating him as a slave. I'm going to say something now that a lot of people would probably disagree with: treating a human being as a slave does not entirely preclude treating them as an end. This is because it is possible to treat a slave well out of a concern for their own feelings and welfare. It is even possible (though very rare) to treat someone as an end precisely by enslaving them (if, for example, the only alternative to slavery was that they be put to death). I am able to take this view because, as a strict hedonist, I do not regard freedom as an intrinsic good. Freedom is good only insofar as it conduces to greater net pleasure.

So my solution to your problem is to seek some way of getting Carlsen to play, such as hypnosis, that will still enable us to treat him well for his own sake, which is what I understand by treating someone as an end. This is a different solution from your original problem, because in that case we were treating the child as a means but very definitely not treating him or her as an end.

For example the lever-puller will say that they merely let the single person die in order to save the five

Which would be clearly false under my view, because them killing someone is not an inaction.

To intend and cause the death of someone is to intentionally kill them

No it is not: letting someone die is not an intentional killing—killing is an action. By “cause”, we could be saying that one’s action or inaction caused the person’s death.

It makes no sense for you to say that you intend and cause their necessary death but you do not kill them.

Do you not agree that “killing” is an action which results in the death of a living being? If you do, then it should be very apparent to you that letting someone die is not a killing—irregardless if you believe that, in some or all cases, letting someone die is morally impermissible.

Edit: If Bob Ross were right, then one would have a magic get-out-of-jail-free card for any moral quandary: just do nothing

This is either the biggest straw-man of my position I have had yet (to-date); or you have not grasped what I have been arguing. Letting someone die is morally permissible IF one cannot save them without committing an immoral act.

Now I should say again that your definition of a "means" is not a good definition.

You just keep asserting this, and I keep responding with the fact that I see no difference between Herg’s vs. my definition. Demonstrate why it isn’t good.

"something useful or helpful to a desired end."

That’s fine, and essentially the exact same as how I defined it.

One of the fundamental problems here is that your account is unable to distinguish between the "V" and the "7" in my diagram. So then either 1) There are no relevant moral differences between "V" situations and "7" situations, and therefore your inability to distinguish them is unimportant, or 2) There are relevant moral differences between "V" situations and "7" situations, and therefore your inability to distinguish them is problematic.

Like I said before, the V diagram is an incomplete representation of the examples we have had. Let’s take the car example: Q is required for P in this specific scenario S, although Q is not required to bring about P all else being equal (let’s call it E).

In S, one can represent it with the V diagram but with Q and P being both required for A; and one cannot remove Q without removing P.

So it would instead be, "Q is not a means to P in 'V' because if one disconnects Q from A then P is still connected to A." Note that Q is not a means because it does not mediate A and P.

Which is an inaccurate representation of, for example, the car example. One cannot save the two without running over the other two, so Q is required for A.

Let's get back to fundamentals again. There are two basic principles in conflict:
• Pdfs: The principle of the diffusiveness of intention
• Pndiv: The principle of the non-divisiveness of intention

I reject “P<ndiv>”. Why should one accept that?

“P<dfs>” seems directly contradictory to “P<ndiv>”.

You seem to be committed to the position which says that when I drive to the grocery store I am intending to consume fuel. This isn't correct. I intend to get groceries, and the consumption of fuel is a side-effect that I would prefer not occur. I only assent to it because I don't know how to get to the grocery store without consuming fuel.
…
if for some reason fuel is not consumed during my trip, the trip is successful all the same. If I get back to my home and see that the fuel has not been diminished, I do not say, "Oh no, my purpose was not achieved!"

Firstly, I am saying that if you intend to get groceries and you know that you have to consume fuel to do so, then you intend to consume fuel to get groceries.

Secondly, intending to consume fuel to get groceries does not result in the purpose being uncompleted if fuel didn’t end up getting consumed: consuming fuel was the means to the end (i.e., purpose) and not the purpose itself.

My point is that if one has a purpose X and knows that they need Y to achieve X, then they have the purpose of using Y to achieve X. You seem to be confusing this with the claim that Y is itself the purpose.

Remember that the key is the relation between P and Q.

In the car example, one cannot achieve P without Q, so Q is a means to achieving P because a means is something useful to a desired end, and Q is useful towards the end P. Not only that, but it is also necessary for achieving the end; so it is a necessary means.

But again, to know whether Q is means to P we must understand the relation between P and Q, and the conditional necessity that you are betting all your chips on is represented by (P → Q).

No I am not. P does not entail Q nor does Q entail P in the car example: A → P & Q. Q is a means towards A (A → Q), and A is the intention towards P (which can’t be expressed in the logic we have been using) and results in P (A → P).

So first, A is not "swerving to save four people." Saving four people is impossible, and we do not intend the impossible

That was a slight typo on my end, and you completely missed the point. Swap out “swerving to save four people” with “swerving to save two people” in this:

Now, an easy way to understand the flaw in your reasoning is to reflect on your 7 diagram; because it implies that if you were to remove P that Q is still connected to A, which is not true, for example, in the case of the car swerving example. If A is “swerving to save four people” and P is “saving four people” and Q is “hitting two innocent bystanders”, then removing P does not result in A → Q. Viz, if the intention is to save four people, then the effect of hitting two innocent bystanders is no longer connected with the intention in the event that there are no four people to save.

Now if we are aiming at Q then we don't need to achieve P (although we could). So we can remove P without removing Q. But in fact we are not aiming at Q; we are aiming at P, and Q is not a means to P as
↪Herg
so helpfully demonstrated.

The point is that if we are aiming at P then we have to achieve Q: that we don’t need to achieve P if we aim at Q is irrelevant (at best).

Q is a means to P, because one cannot achieve P without Q. Q is a necessary, utility towards P.

The central point has to do with whether Q is a means to P.

This is the central point for our debate whether or not, in our examples, the person is intentionally killing anyone; but it is not the central point for our debate about whether or not letting a person die is always immoral.

If you want to say that Q is still intentional killing even in the "V" case then that is a separate argument from the claim that Q is a means to P

I would say they cannot be separate, because I hold, which you accept, that Q is intentional if it is a means towards the intentional end (that is being actualized).

• A "means" is asymmetrical. If A is a means to B, then B cannot be a means to A. Two things cannot be a means to one another.

Sure, that’s fine: maybe I misunderstood what you were originally saying. I am saying is that Q is a means towards A, and A is an intention towards P. Q is intentional then, if one accepts that Q is intentional if it is a means towards A (that is being actualized).

"Supposing !(A → Q), would I still choose A?"

What you are noting is the original intention, stripped of the necessary means towards A in the application of A in scenario S. That’s fine, but you can’t stop there: irregardless if I would still choose A if !(A → Q), in S A → Q and, therefore, Q is a means towards A in S.

I would prefer we just refer to it as intentional killing vs. killing: that is the most clear way of presenting it. — Bob Ross

Yes of course you would, for that would aid your position. Likewise, I would prefer if we just refer to it as killing vs. non-intentional killing. So again, I suggest we not use the word "killing" without a prefix.

How does that aid my position? And how is a distinction between killing vs. non-intentional killing and intentional killing vs. killing any different? They are both the same distinction as intentional vs. non-intentional killing.

I don't see how we're treating Magnus as an end by forcing him to play chess with the aliens. There is nothing at stake for him. He's not in any danger. We're not saving his life or bettering his condition in any way by forcing him to play/making him our slave. We're forcing him to play strictly for our own ends.

Don't know if this one's been mentioned ...

A "means" is asymmetrical. If A is a means to B, then B cannot be a means to A. Two things cannot be a means to one another. — Leontiskos

Sure, that’s fine: maybe I misunderstood what you were originally saying. I am saying is that Q is a means towards A, and A is an intention towards P. Q is intentional then, if one accepts that Q is intentional if it is a means towards A (that is being actualized). — Bob Ross

Egads. A is not an intention and has never been an intention. It is an action. In the trolley case it is to pull the lever, in the car case it is to swerve, etc. A is carried out for the sake of the end, P, but A is not an intention. It is an intended means to P. Similarly, your claim that “Q is a means towards A” makes no sense. It is not true in the trolley case that, “Killing the one is a means towards pulling the lever,” or in the car case that, “Killing the two is a means towards swerving” (even though both of these really would follow on your dubious claim that anything which is conditionally necessary is a means).

Now you still haven’t managed to recognize, “The other absurdity that results from your view is that Q is a means to P and P is a means to Q, which is of course impossible” (). For example, take your claim:

Q is a means to P, because one cannot achieve P without Q. Q is a necessary, utility towards P. — Bob Ross

As I have noted multiple times, (P ↔ Q). We cannot achieve P without Q and we cannot achieve Q without P. Again, therefore on your own reasoning, P must be a means to Q and Q must be a means to P, which is utterly impossible.

I gave the logic for this earlier: "In other words: (A → (P ^ Q)) → (A → (P ↔ Q))" ().

Feel free to check it to make sure that it is valid (link).

No I am not. P does not entail Q nor does Q entail P in the car example: A → P & Q. — Bob Ross

This premise is false in the examples we have given: A entailing P and Q does not imply that A entails that P entails Q — Bob Ross

Again, (A → (P ^ Q)) → (A → (P ↔ Q))

Note that contained within this is, for example, "2. (A → (P ^ Q)) → (A → (P → Q))" ().

At this point you are committing formal logical contradictions. You are affirming the antecedent and denying the consequent of a tautology.

Note too that given (A → (P ^ Q)) → ((A → P) ^ (A → Q)), your strange theory would have it that P and Q are both a means to A, as well as being a means to one another. This is yet another absurdity of your view.

In the car example, one cannot achieve P without Q, so Q is a means to achieving P because a means is something useful to a desired end, and Q is useful towards the end P. — Bob Ross

How is “hit two people” useful or helpful to the desired end of “avoid hitting all four people”? Or in the trolley case, how is the death of the one useful or helpful to the desired end of saving the five? Again, I would suggest you try to actually respond to ’s excellent analysis.

This is the central point for our debate whether or not, in our examples, the person is intentionally killing anyone; but it is not the central point for our debate about whether or not letting a person die is always immoral. — Bob Ross

You have all sorts of different, overdetermined arguments for your conclusion, like a castle with multiple walls. Whenever one suffers a defeat you silently switch to another. One has to do with intention via means, one has to do with direct intention, one has to do with omissions, etc. I am only interested in the question of intention via means, namely determining whether Q is a means to P. Your persistence even in the midst of formal logical contradiction makes me think that it will be a miracle if we ever settle that simple question.

Like I said before, the V diagram is an incomplete representation of the examples we have had. Let’s take the car example: Q is required for P in this specific scenario S, although Q is not required to bring about P all else being equal (let’s call it E).

In S, one can represent it with the V diagram but with Q and P being both required for A; and one cannot remove Q without removing P. — Bob Ross

Diagram “V” only makes sense relative to diagram “7”. Diagram “7” represents Q as a means between A and P, whereas diagram “V” represents Q as a non-means effect of A.

We can use a temporal argument to give another reason why Q_v is not a means. In the trolley example as soon as I switch the track the five people are saved, and yet the one person does not die until the trolley eventually runs over them. Now Q cannot be a means to P if P occurs before Q in time, and yet that is precisely what happens in the trolley scenario. By the time Q occurs, P has already been accomplished and completed. How could anything be “helpful or useful” to a desired end that has already occurred?

I reject “P<ndiv>”. Why should one accept that? — Bob Ross

I was hoping the grocery case would help illustrate why. As I said:

Suppose an act has two effects: I press the accelerator and two things happen: my speed accelerates and my fuel is consumed. Do you think Aquinas is right, and it is possible to intend one effect without the other? When I press the accelerator do I intend to accelerate and do I intend to diminish my fuel? And even if so, do I intend them in the same way? Is the word "intend" being used in the same way for both of these effects? — Leontiskos

Again, “Do I intend them in the same way?” That was the key question I was asking.

Secondly, intending to consume fuel to get groceries does not result in the purpose being uncompleted if fuel didn’t end up getting consumed: consuming fuel was the means to the end (i.e., purpose) and not the purpose itself. — Bob Ross

If the intention were, “Consume fuel to get groceries,” then the purpose would be uncompleted if fuel didn’t end up getting consumed. Then I would get back home, look at the fuel gauge, and say, “Crap! I didn’t manage to complete my intention of consuming fuel to get groceries! I got the groceries but I didn’t consume any fuel. What a failure of a trip!” The very absurdity of such a thing shows that the intention was not, as you say, “Consume fuel to get groceries.” But I grant that the fact that fuel is a causal means complicates the example, and that is why I didn’t press it when you ignored it. What it shows is that even known causal means can be indirectly intended, and this is probably too subtle.

A better example than fuel consumption might be emissions, in the case where my exhaust system is broken and therefore my car emits especially harmful pollution. If the direct intention before was not accurately captured by, “Consume fuel to get groceries,” then the direct intention now is certainly not captured by, “Emit pollution to get groceries.” The emission of pollution and the getting of groceries are not intended in the same way, and you are in need of a theory which accounts for this fact.

But if you want to see how Brock elaborates a bit, here is what he says:

The above account of indirect objects of intention seems to correspond to what Chisholm calls the principles of the ‘diffusiveness’ and ‘non-divisiveness’ of intention. The principle of the diffusiveness of intention is that if someone intends p, knowing that p entails q, then he intends the conjunction p and q. For instance, if a man intends to drive off in a car parked nearby, knowing that it belongs to someone else, then he intends this: to drive off in a car parked nearby that belongs to someone else. The principle of the non-divisiveness of intention is that if someone intends the conjunction p and q, he does not necessarily intend q by itself. The man may not intend just this: to drive off in a car that belongs to someone else.22

That is, he may not be doing what he is doing because it will deprive someone else of a car, but only because it will provide him with the quickest means of getting to his destination. Again, this is the difference between direct and indirect objects of intention: only the former are explanations of the action that fulfills the intention. But despite this difference, if the man knows that the car belongs to someone else, it would be ridiculous to say that he positively meant not to drive off in someone else’s car.

The point is subtle. Many authors disagree with the diffusiveness of intention principle.23 Donagan, following Bratman, while not explicitly addressing Chisholm, uses this example: ‘If you intend to run a marathon and believe that you will thereby wear down your sneakers, it does not in the least follow that you intend to wear down your sneakers.’24 This example is similar to one used by St Thomas. ‘If someone often or always gets his feet wet when he goes to a muddy place, then granted that he does not intend this, nevertheless it is not said to be by [bad] luck.’25 But what Aquinas means is that although he does not intend simply to get his feet wet, this still falls under his intention. What Chisholm’s critics seem to ignore is the other principle, the non-divisiveness of intention.26 When taken together with it, the principle of the diffusiveness of intention appears to fit very well with the doctrine that actions are individuals, i.e. that one and the same action may admit many true descriptions whose combination is logically contingent. — Stephen L. Brock, Action and Conduct: Thomas Aquinas and the Theory of Action, Ch. 5: Praeter Intentionem, p. 204-5

Also relevant:

Consider the frequent objection to the moral significance of the question of the directness or indirectness of some result, e.g. someone’s death. If what you do kills him anyway, what difference does it make whether or not you were aiming at his very death? It makes at least this much difference: if you did not aim at his very death, then, provided that you do succeed in bringing about what you did aim at, you will not mind if something happens to prevent him from dying after all.47 That is, even if his death is a nearly certain result of your succeeding in what you are doing, it is not you who are making certain of it. Of the things that might happen to thwart the killing of someone, at least some are things that someone bent on killing him would take into consideration and try to stop, while someone not having that aim would not. At any rate, his death cannot be as certain, on the supposition of your accomplishing your aim, as it would be if it were your aim, again on the supposition of your accomplishing your aim. The certainty of your getting what you aim at, if you get what you aim at, is plainly greater than the certainty of the occurrence of something other than what you aim at, if you get what you aim at. Still, this difference, which is only one of degree of certainty, is perhaps not very significant morally. At any rate the fact that you are not directly aiming at anyone’s death does not at all entail that no matter what you do, you cannot be guilty of murder.48 — Stephen L. Brock, Action and Conduct: Thomas Aquinas and the Theory of Action, Ch. 5: Praeter Intentionem, p. 210

"Supposing !(A → Q), would I still choose A?" — Leontiskos

What you are noting is the original intention, stripped of the necessary means towards A in the application of A in scenario S. That’s fine, but you can’t stop there — Bob Ross

Sure I can, because this is what it means for Q to not be directly intended.

So at this point in the conversation I am guessing you will have the good sense to reject your claim that "1. (P → Q) → (Q is a means to P)" (). Remember that earlier in our conversation you doubled-down against my proof and decided to reject our mutual principle instead of your false (1). The mutual principle was, "<If I intend A and A causes effects P and Q, then I do not necessarily intend Q, even when I know that A causes Q>. (the idea here is that P is clearly intended)" (). Now that you presumably understand that (1) is false, there is no reason for you to reject this principle. What you ought to do now is start from scratch and try to figure out what it means for Q to be a means to P, for it most certainly does not mean that (P → Q) (i.e. "You can't get P without Q").

No it is not: letting someone die is not an intentional killing—killing is an action. By “cause”, we could be saying that one’s action or inaction caused the person’s death. — Bob Ross

Then I would say that you need to refine what you mean by a cause.

Do you not agree that “killing” is an action which results in the death of a living being? If you do, then it should be very apparent to you that letting someone die is not a killing — Bob Ross

Sure, but I would hold that if I have not killed them then I have not caused their death. In general you are lacking distinctions regarding causation and intention. You are making these things more black and white than they are.

This is either the biggest straw-man of my position I have had yet (to-date); or you have not grasped what I have been arguing. Letting someone die is morally permissible IF one cannot save them without committing an immoral act. — Bob Ross

Okay that's fair, I retract and apologize for the claim that you think "letting happen" is morally indifferent.

You have a lot of good points, but I am, to be honest, losing track of the course of this conversation. So I am going to take your advice and wipe the slate clean. I am going to attempt to provide a more clear and robust analysis of what I am trying to argue: consider anything in here that is contradictory to what I have previously said as a concession.

I am going to keep it short and sweet, and let you navigate the conversation from here.

An intention is an ideal meant to be actualized.

An intention can have two aspects: what is essential and what is accidental.

The accidental aspect of an intention is the part of the ideal which is only in virtue of the particular circumstances in which it is being actualized.

The essential aspect of an intention is the part of the ideal which is not dependent on the circumstances whatsoever; and comprises the essence of it.

The course of action chosen in a situation to actualize the intention is a part of the accidental aspect of it. E.g., if I intend to quench my thirst and I decide that I should walk into the kitchen to get a water bottle (or fill up a glass with water), then doing so is something I intend to do but merely because in the particular circumstances it is best for actualizing the essential aspect of my intention (which is to quench my thirst).

A means is something useful for an intention.

A means is not necessarily necessary. E.g., if I have the intention of getting groceries, then using my car or my bike would be a means towards that end—but neither are necessary per se (since one could be used instead of the other).

Something that is necessary to accomplish an intention is necessarily a means towards that intention. E.g., if I have the intention of getting groceries and I can only get them by way of using my car, then my car is a necessary means towards that intention.

A means that is utilized for an intention always becomes a part of that intention. E.g., if I have the intention of getting groceries and I can use my bike or car and I choose to use my car, then using my car becomes a part of the ideal (but as an accidental aspect of it).

The foreseen consequences of a means always become a part of the intention which utilizes that means. E.g., if I have the intention of getting groceries and I choose to use my car of which I know will pollute the atmosphere, then my car polluting the atmosphere is a part of the ideal (but an accidental aspect of it).

It follows that in the V diagram, Q is not a means towards P; and P is not a means towards Q. A is a means towards Q and P.

It follows that swerving to kill two people to save two other people is a means towards saving the two people; but that the killing of the two people was not a means to saving the two people but, rather, an effect of swerving.

It follows that the effects of swerving are twofold: killing two people and saving two people; and that both are intentional, because the foreseen consequences of a means (which, in this case, is swerving) are always intentional (albeit a part of the accidental aspect).

It is always wrong to intentionally kill innocent people; so utilizing a means which has a foreseen consequence of killing an innocent person (thereby making it a part of the accidental aspect of the intention) is wrong.

What say you?

It is always wrong to intentionally kill innocent people; so utilizing a means which has a foreseen consequence of killing an innocent person (thereby making it a part of the accidental aspect of the intention) is wrong. — Bob Ross

Suppose the Allies can destroy a Nazi heavy water nuclear facility, and disable their nuclear weapons program (pretend the Nazi's have a robust program) by causing the death of one innocent German janitor. You're saying it's wrong for them to bomb the nuclear facility?

ETA: Forget the nuclear facility. Suppose the Allies know Hitler is touring the nuke facility and they can bomb it and kill him and only one innocent janitor gets killed. You're saying they shouldn't do it?

You have a lot of good points, but I am, to be honest, losing track of the course of this conversation. So I am going to take your advice and wipe the slate clean. I am going to attempt to provide a more clear and robust analysis of what I am trying to argue: consider anything in here that is contradictory to what I have previously said as a concession. — Bob Ross

Okay, fair enough, and this is a good post. I will offer a few more posts, but I don't know how long I will stick around. I was limiting myself to the "means" question in large part because I don't have time to get drawn into all of the other related topics.

An intention is an ideal meant to be actualized. — Bob Ross

Somewhat, but I would not call it "an ideal." That has connotations of optimality. We could say that an intention is an action or effect meant to be actualized, or that it is something that one plans to bring about, or something that one aims to bring about. So if I intend to build a computer, then I plan or aim to bring about a computer. Looking now at Merriam-Webster's 'intend', it gives, "to have in mind as a purpose or goal : Plan."

An intention can have two aspects: what is essential and what is accidental.

The accidental aspect of an intention is the part of the ideal which is only in virtue of the particular circumstances in which it is being actualized.

The essential aspect of an intention is the part of the ideal which is not dependent on the circumstances whatsoever; and comprises the essence of it. — Bob Ross

I think it is right to distinguish essential (per se) intention from accidental (per accidens) intention, but in all such cases we want the essential/per se to be primary (whereas you made the per se definition parasitic on the per accidens definition).

On an Aristotelian view intention is something like the force by which will/volition has an effect on things. It is how volitional beings exercise their volition. Because of this the essence of intention has to do with what is most central to the will/volition of such a being. So if I will the end of getting groceries, then it is natural to me to plan or aim to bring about the getting of groceries, namely to intend it. Once this intention is formed the first step in the execution of my will has been formed, and as Aquinas says, "The end is first in the order of intention; the means is first in the order of execution." After recognizing my will to eat, I first intend to get groceries (end), and then I think of driving my car to the grocery store (means). The end precedes the means in the order of intention. But in executing my plan I begin with the means by getting into my car, and finish with the end by purchasing the groceries. The means precedes the end in the order of execution. This structure is divisible: for example, I must turn the key in order to drive my car (and this is a means to the end of driving the car, which is a means to the end of getting groceries); and I must collect my key in order to start the car, etc. The intermediate ends are also a means, and each means is a means to each subsequent end.

My essential or per se intention is that which adheres most closely to the aim that my will has formed. As a metaphor, the fastest and most efficient route to the grocery store adheres most closely to my end, my intention to get groceries. A circuitous route to the grocery store does not adhere as closely to my end, my intention to get groceries. Thus if a road is closed and I am forced to take a detour, then although I intend to take the detour for the sake of my end, the taking of the detour falls under my intention in an accidental manner. The accidental part of my intention is that which falls away or fails to closely adhere to my proper intention (end). As you recognize, accounting for accidental or per accidens intention requires the explanation of something that is circumstantial in one way or another. It is worth noting that essential and accidental intention differ by way of a spectrum, and are not black and white logical categories (i.e. they are contraries rather than strict contradictories).

The course of action chosen in a situation to actualize the intention is a part of the accidental aspect of it. E.g., if I intend to quench my thirst and I decide that I should walk into the kitchen to get a water bottle (or fill up a glass with water), then doing so is something I intend to do but merely because in the particular circumstances it is best for actualizing the essential aspect of my intention (which is to quench my thirst).

A means is something useful for an intention. — Bob Ross

The danger here is attaching intention too strongly to ends such that it becomes unattached to means. It is true that intention is most perfectly found in the end, and that all means are "accidental" to one degree or another. Nevertheless, we do intend the means to an end. You have an intention to quench your thirst but you also end up forming an intention to walk into the kitchen. The obtaining of the water bottle is also something you plan or aim to bring about, even though it is subordinated to your aim to bring about the quenching of your thirst. Both are intended, but the force of will that intention serves is always stronger and more direct in the case of the end, and therefore the quenching of thirst is more fully intended than the obtaining of the water bottle.

A means is something useful for an intention. — Bob Ross

Means and ends are both kinds of intentions, and both must be appropriated by an agent. There may be a cup of water sitting nearby that you are not aware of. This cup of water is "something useful for [the intention of quenching thirst]," but it is in no way an actual means to that end if you are not aware of it. This is tricky, because we can talk about means and ends apart from appropriation by an agent, for example, "You could aim at that end. You could achieve that end by such-and-such a means," but what is being proposed is the possibility of appropriation; the possibility of intentionally using the means in order to achieve the end. We could say that an unknown means is a kind of per accidens intention ("he would intend to use it if he knew about it"), and in relation to this a known means is a kind of per se intention. But it is simultaneously true that, in relation to one another, the end is a per se intention and the (chosen) means is a per accidens intention.

This is part of the subtlety of Aristotle's thought: that essential/accidental constitute a relational dyad more than two distinct and clear-cut categories. Intention is one of those things that modern modal logic can't really capture, for example:

A means is not necessarily necessary. — Bob Ross

Yes, exactly, and what this means is that the nature of a means is neither necessary nor merely possible. It is not modally representable. It is the actuation of a tendency or final cause; to intend is a kind of tending towards, not a merely logical reality.

A means that is utilized for an intention always becomes a part of that intention. E.g., if I have the intention of getting groceries and I can use my bike or car and I choose to use my car, then using my car becomes a part of the ideal (but as an accidental aspect of it). — Bob Ross

Yes, but:

The foreseen consequences of a means always become a part of the intention which utilizes that means. E.g., if I have the intention of getting groceries and I choose to use my car of which I know will pollute the atmosphere, then my car polluting the atmosphere is a part of the ideal (but an accidental aspect of it). — Bob Ross

So we have it that a (chosen) means becomes a part of the intention of the end, and the foreseen effects of a (chosen) means become a part of the intention of the end. But the same crucial question that I asked earlier arises here: do they become equal parts in the intention of the end? Do they become parts in the same way? They do not, because whereas the means is a kind of accident of the end, the foreseen effect is a kind of accident of an accident of the end (i.e. it is an accident of the means). What follows is that the foreseen effect does not adhere as closely to the formal intention (the intention of the end) as the means does, and therefore the foreseen effect and the means fall under the intention of the end in different ways. (Aquinas talks about the essence/accidents of the intention as well as the matter/form of the intention, which provides some additional resolution).

This is what Brock means by direct intention vs. indirect intention. A means is directly intended whereas a foreseen effect is only indirectly intended.

It follows that in the V diagram, Q is not a means towards P; and P is not a means towards Q. A is a means towards Q and P. — Bob Ross

A is a (chosen) means to P, and Q is a foreseen effect of A, but does it follow that A is a means to Q? Are all causes means to their foreseen effects? With respect to Q, A is the cause of an effect, not the means to an end. Only with respect to P is A a means to an end. This is where it becomes important to recognize that a means must be appropriated by a volitional subject in order to truly be a means. Consider the question, "Why did he cause A?" There are two main answers on offer:

1. "He means to achieve P"
2. "He means to achieve Q"
3. (both)

I would say the correct answer is (1). If Q is merely a foreseen effect then he neither means to achieve Q, nor does he mean to not-achieve Q. The fact that he does not mean to achieve it shows that it is not directly intended (in Brock's language). The fact that he does not mean to not-achieve it shows that it is indirectly intended on the presumption that he chooses/intends A. If he means to not-achieve Q then he would not choose/intend A. As I said above, "Supposing !(A → Q), would I still choose A?"

But the same sort of conclusion would seem to follow on your own definition:

A means is something useful for an intention. — Bob Ross

If A were a means to Q, then you would say that A is useful for some intention (with respect to Q). But what intention is that? The intended end is P, and A is only intended because it is a means to P.

A is "a means" to Q in the abstract sense, but it is not a means that you appropriate via intention. A is a cause of Q, not a means to Q. If we intend an end, we also directly intend any (chosen) means to that end; but if we are using A as a means to P, it does not follow that we are also using A as a means to Q. The intention of an end flows into the means, but the intention of a means does not flow into all of its various effects. Q is indirectly intended given A, but we are not using A as a means to Q.

It follows that swerving to kill two people to save two other people is a means towards saving the two people; but that the killing of the two people was not a means to saving the two people but, rather, an effect of swerving. — Bob Ross

Throughout you have been making dozens of very minor mistakes which I overlook for the sake of time. Here is an example of that. Your first clause is technically true, but the description is inaccurate. Action A (swerving) is not done "to kill two people to save two other people." This is like the claim that driving to the grocery store is done to emit pollution to get groceries. The bolded "to" indicates that Q is a means to P, which is exactly what you deny in your second clause. That "to" should attach to "swerving," and, "to kill two people" should be omitted altogether. One is swerving to save two people, not to kill two people to save two people. Throughout this thread you have been committing these minor infelicities which subtly bolster your position.

A driver could do what you represent here, but if I were the driver I would not do it that way. My actions would only be properly represented if those four words were omitted. A is a means to the intended end, and that intended end is P, not Q. A is done for the sake of P, not for the sake of Q. Therefore I am "swerving [for the sake of] saving people," not for the sake of killing people. A is a means to saving, not killing. Only if I intend Q would A be a means to killing people, and in that case I would be committing murder. In that case I would be using A as a means to my intended end of Q. If I were the driver I could perhaps be charged with manslaughter, but not murder. Your approach can't seem to distinguish the two.

It follows that the effects of swerving are twofold: killing two people and saving two people; and that both are intentional, because the foreseen consequences of a means (which, in this case, is swerving) are always intentional (albeit a part of the accidental aspect). — Bob Ross

So the whole point of my conversation with you is to demonstrate that a means and a foreseen effect are both intended, but in different ways! As Brock says, the means is directly intended and the foreseen effect is indirectly intended. If a foreseen effect were a means then there would be no difference between a means and a foreseen effect. For insight into the import of this, see the most recent quote from Brock that I provided.

It is always wrong to intentionally kill innocent people; so utilizing a means which has a foreseen consequence of killing an innocent person (thereby making it a part of the accidental aspect of the intention) is wrong. — Bob Ross

Is it always wrong to accidentally-intentionally kill innocent people? More precisely, is it always wrong to indirectly intend to kill innocent people? Is voluntary "manslaughter" always negligent? Those are the questions that need to be answered.

P.S. Another thing that we haven't directly delved into is the difference between causal necessity and logical necessity. For example, in the trolley case we can think of two logical propositions, 1) If I pull the lever then the track will shift, and 2) If I pull the lever then the one will die (A → Q). (1) is causal and (2) is merely stipulative. Because the lever is physically connected to the track, to pull the lever is to shift the track. The movement of the lever has a per se causal ordination to the shifting of the track. Track-levers always shift tracks. (An even more direct and per se cause is that to pull the lever is to cause the lever to move spatially.) But (2) does not have a per se causal ordination to the death of an individual. (2) is merely a per accidens cause, and this is why it needs to be stipulated within the problem itself (i.e. we need no explanation that track-levers shift tracks, but we do require an explanation for why pulling a track-lever would result in someone's death). This relates to intention because indirect intentions and per accidens causality tend to go hand in hand. The closer to per se causality an act approaches, the less plausible is the idea that the effect was not directly intended. For example, it makes some sense to say that you pulled the lever without (directly) intending to cause the person's death, but it makes no sense at all to say that you pulled the lever without intending to switch the track. Pulling the lever and switching the track are very nearly the same thing.

I don't see how we're treating Magnus as an end by forcing him to play chess with the aliens. — RogueAI

We aren't; that's the bit where we're treating him as a means.

There is nothing at stake for him. He's not in any danger. We're not saving his life or bettering his condition in any way by forcing him to play/making him our slave. — RogueAI

You don't have to save someone's life or better their condition to be treating them as an end. If you're playing chess and enjoying it, then my leaving you alone to enjoy yourself is treating you as an end.

We're forcing him to play strictly for our own ends. — RogueAI

Of course, but you said this:

The aliens like him and will gift him a good life no matter what he decides. — RogueAI

So we can treat him as an end by letting him go to the aliens. Of course he didn't want to play, but we have hypnotised him so that he now does, and presumably he will enjoy doing so. If you want to tweak the scenario so that he suffers while he's playing, then in effect we're complicit in torturing him, in which case the situation is essentially the same as it was with the kid, and so we shouldn't hand him over, we should fight the aliens instead.

Correct. If they know that they are going to kill an innocent person by intentionally killing hitler; then they are intentionally killing that innocent person to kill hitler.

I want to put this train of thought in your mind:

If you kill someone, then you are not treating them as an end (unless it is mercy killing).

If you let someone die when you could save them, the same is true.

Therefore, if you let your car kill four people instead of two, you are not treating the additional two as ends.

Why did you let your car kill the additional two people? Because you believed in the principle (which I will label IP) that it is wrong to kill innocent people by positive action.

So you had an end in view when you let your car kill those additional two people, the end being that your behaviour should conform to IP. And letting your car kill those two additional people was a means to that end. We have already said that you did not treat those additional two people as ends; so in fact you treated them merely as means, which according to the principle that we should treat people as ends rather than just as means (which I shall label EP) is morally wrong.

We now have two conflicting principles: IP and EP. Which should we obey?

I think we should obey EP. For one thing, it is applicable in all situations, whereas IP is only applicable in situations where someone dies. For another, the range of persons (or sentient beings, if you want to cast the moral net wider) is as wide as it could be: in fact it is coterminous with the entire class of persons (or sentient beings). And this is unsurprising, because if beings have moral status, then it is arguably a tautology to also say that they should be treated as ends. There is also the point that there is a fighting chance that we may be able to derive, if not IP itself, then something very like IP, from EP, while clearly the reverse is impossible.

I would say that EP is really the entire point of morality. It is the principle that we should try as best we can to respect (and, where necessary, look after) the interests of others. We do not respect their interests by letting our cars plough into them. And if there are supposed principles where, when we obey them, we find ourselves trampling over the interests of others rather than respecting them, I think we should at the very least be very suspicious of those principles.

↪RogueAI Correct. If they know that they are going to kill an innocent person by intentionally killing hitler; then they are intentionally killing that innocent person to kill hitler. — Bob Ross

Surely the point here is that if let Hitler live, he will continue to fail to treat millions of people as ends by murdering them; and our only way of treating those people as ends is to fail to treat Hitler, and unfortunately the janitor, as ends. So we either fail to treat two people as ends, or we are complicit in the failure to treat millions as ends.

Okay, fair enough, and this is a good post. I will offer a few more posts, but I don't know how long I will stick around. I was limiting myself to the "means" question in large part because I don't have time to get drawn into all of the other related topics.

Fair enough.

Somewhat, but I would not call it "an ideal."

It is an ideal insofar as it is an idea about how reality should be. When one takes on a purpose, they are implicitly conceding that they believe reality is not the way it should be. To your point, it may not be an ideal insofar as it is ultimately how reality should be.

"The end is first in the order of intention; the means is first in the order of execution."

:fire:

My essential or per se intention is that which adheres most closely to the aim that my will has formed.

Not quite. The per se intention would be the ideal which you are trying to actualize, and any means of achieving it would be accidental. That which adheres most closely to that aim is just a means that adheres most closely to that aim—and so it is accidental, not per se.

So:

As a metaphor, the fastest and most efficient route to the grocery store adheres most closely to my end, my intention to get groceries.

Getting groceries is the per se intention; and the fastest and using the most efficient route to the grocery store is the per accidens intention.

I think we are distinguishing per se and per accidens differently, perhaps?

The danger here is attaching intention too strongly to ends such that it becomes unattached to means

I would say a means is NOT an intention; but means can be intentional.

. You have an intention to quench your thirst but you also end up forming an intention to walk into the kitchen

But a means is not identical to an intention: there are means which one doesn’t not intend (e.g., they are not aware of them). The intention towards walking in the kitchen is a separate intention from quenching my thirst, but they are closely connected: the latter is the essence of what I am intending to do, whereas the former is just an accidental means towards it. I still intend both, just differently.

Likewise, if I did not intend the former, it would not change the fact that it is a means to the latter.

Means and ends are both kinds of intentions

Not in the sense as I defined it. I guess, what is your definition of a ‘means’? It can’t be how I defined it, because that definition does not preclude unintentional means.

do they become equal parts in the intention of the end?

Not in the sense that I think you mean it. I only intend to walk to the kitchen because I intend to quench my thirst.

This is what Brock means by direct intention vs. indirect intention. A means is directly intended whereas a foreseen effect is only indirectly intended.

Ok, that’s fine then.

A is a (chosen) means to P, and Q is a foreseen effect of A, but does it follow that A is a means to Q?

A is a means to Q because A is useful for facilitating Q—even if one accidentally intends or doesn’t intend at all Q.

You are right that Q is a foreseen effect; but that doesn’t absolve A from being a means toward that effect.

This is where it becomes important to recognize that a means must be appropriated by a volitional subject in order to truly be a means.

Then you reject the definition I gave, and don’t know what definition you are using.

1. "He means to achieve P"
2. "He means to achieve Q"
3. (both)
[4. “He per se intends P, and per accidens intends Q” ]

I can appreciate what you are going for: he “means” to achieve P and not Q in the sense that P is the main, primary goal and Q is not at all. But this does not absolve him, in this scenario, from intending Q; because a foreseen effect of a chosen means is intentional.

If Q is merely a foreseen effect then he neither means to achieve Q, nor does he mean to not-achieve Q. The fact that he does not mean to achieve it shows that it is not directly intended (in Brock's language). The fact that he does not mean to not-achieve it shows that it is indirectly intended on the presumption that he chooses/intends A. If he means to not-achieve Q then he would not choose/intend A. As I said above, "Supposing !(A → Q), would I still choose A?"

It seems like, then, you are agreeing with me with different words: Q is, in this example, intentional—but indirectly. Then it is not true that “He means only to achieve P”: he means to achieve P, and this requires him to achieve Q as well (simply because A is a means towards both and never one or the other).

If A were a means to Q, then you would say that A is useful for some intention (with respect to Q). But what intention is that? The intended end is P, and A is only intended because it is a means to P.

It is useful for achieving Q—which has not direct relation to any intention. There could be no intention towards Q, and it would still follow that Q can be facilitated by way of A.

A is "a means" to Q in the abstract sense, but it is not a means that you appropriate via intention

It is appropriated via intention if one accepts my premise that “if one foresees an effect (Q) of a means (A) and chooses to use that means (A) to achieve their intention towards another effect (P), then they thereby intend Q”.

Throughout you have been making dozens of very minor mistakes which I overlook for the sake of time. Here is an example of that. Your first clause is technically true, but the description is inaccurate. Action A (swerving) is not done "to kill two people to save two other people."

“to” in that sentence does not imply that it is a means: it is referring to an intentional effect facilitated equally in order to achieve the other effect. I think you may be splitting hairs here a bit.

So the whole point of my conversation with you is to demonstrate that a means and a foreseen effect are both intended, but in different ways!

That’s fair and I agree now.

Is it always wrong to accidentally-intentionally kill innocent people? More precisely, is it always wrong to indirectly intend to kill innocent people? Is voluntary "manslaughter" always negligent?

This is why I was wary to call it “per accidens”, because I am NOT referring to a colloquial usage of the term “accident”: the latter is used commonly to refer to something someone didn’t intend to do. Manslaughter is when someone unintentionally kills someone: having a per accidens intention to kill someone for the sake of a per se intention to save someone else is NOT an unintentional killing.

The accidental aspect I am referring to, is the part of the intention, which is still an intention, that is required in the specific circumstances to achieve the original, per se, intention. Both are intentional, even if they are intended in different ways. That’s not how manslaughter works.

Another thing that we haven't directly delved into is the difference between causal necessity and logical necessity.

Yes, that’s fine; it was stipulated in the trolley problem to avoid that kind of conversation.

This relates to intention because indirect intentions and per accidens causality tend to go hand in hand. The closer to per se causality an act approaches, the less plausible is the idea that the effect was not directly intended. For example, it makes some sense to say that you pulled the lever without (directly) intending to cause the person's death, but it makes no sense at all to say that you pulled the lever without intending to switch the track

I would say that what you are truly getting at, is that knowledge one should have about what they are doing is tied to what we believe they intend. This is absolutely true, and can ride, pragmatically, on a per se vs. per accidens causality.

If you want to tweak the scenario so that he suffers while he's playing, then in effect we're complicit in torturing him, in which case the situation is essentially the same as it was with the kid, and so we shouldn't hand him over, we should fight the aliens instead. — Herg

I agree with you up to this point. I think we have to force him to play. Being forced to play a game of chess is not even close to turning over a little kid to be tortured. However, threatening to shoot him is an empty threat because he knows you need him alive to play the chess game. I think we should at least try that and see if it convinces him to play. You can't let the selfish actions of one person doom humanity.

Correct. If they know that they are going to kill an innocent person by intentionally killing hitler; then they are intentionally killing that innocent person to kill hitler. — Bob Ross

But your position is absurd. If the Allies could end the war in one bombing raid and the collateral damage is only one innocent person, of course they should do that.

Good post.

It is an ideal insofar as it is an idea about how reality should be. — Bob Ross

A problem with that definition actually manifests here:

Not quite. The per se intention would be the ideal which you are trying to actualize, and any means of achieving it would be accidental. — Bob Ross

Here you seem to confuse an intention with an ideal. An intention is not "the ideal which you are trying to actualize." This is a common slip given the way you defined an intention. It is like saying, "A hamburger is the thing which you are trying to eat." To put it starkly, 'intend' is a verb whereas 'ideal' is a noun. When we talk about an intention we are really talking about something that is being intended. Intention therefore needs to be centrally defined in terms of intending, in terms of acting, in terms of verbs. It is the difference between eating-a-hamburger and a hamburger. Your construal here makes the specific noun, 'ideal', the grammatical subject of the clause, and this gives the impression that what is at stake is a noun with some accidental properties. The problem is that the "doingness" of intention is not accidental, it is central. Ideally I would want to talk about what it means to intend, not what it means to have an intention. Everything depends on what it means to intend, as this is the more fundamental reality.

My essential or per se intention is that which adheres most closely to the aim that my will has formed. — Leontiskos

That which adheres most closely to that aim is just a means that adheres most closely to that aim—and so it is accidental, not per se. — Bob Ross

The idea with that detour metaphor was that intention of the end must be understood in relation to the power of will/volition, and intention of the means must be understood in relation to the intention of the end, and foreseen effects must be understood in relation to the intention of the means. Each stage is more accidental/per accidens, but it must simultaneously be recognized that within each stage there are differences between what is more essential or more accidental. An intention of the end can be more or less essential depending on how it relates to my telos and my will; an intention of the means can be more or less essential depending on how it relates to my intended end, etc.

Getting groceries is the per se intention; and the fastest and using the most efficient route to the grocery store is the per accidens intention.

I think we are distinguishing per se and per accidens differently, perhaps? — Bob Ross

Yes, you are using them as if they are strict logical categories, which is what I opposed in my last post. Essentialism always has a muddy boundary between what is essential and what is accidental, and we essentialists think of this as a feature rather than a bug. An essential property vs. an accidental property is not quite the same as a necessary property vs. a non-necessary (contingent) property, although there is some overlap here. The basis of this is the idea that the forms of natural things are not logically circumscribable in that strict manner. We are trying to understand the whatness of intention rather than to merely enumerate its properties or define its boundaries.

I would say a means is NOT an intention; but means can be intentional. — Bob Ross

So if you intend to quench your thirst and you begin filling your glass at the faucet, you would say that you did not intend to fill your glass? Filling your glass is a means.

But a means is not identical to an intention: there are means which one doesn’t not [sic] intend (e.g., they are not aware of them). The intention towards walking in the kitchen is a separate intention from quenching my thirst, but they are closely connected: the latter is the essence of what I am intending to do, whereas the former is just an accidental means towards it. I still intend both, just differently.

Likewise, if I did not intend the former, it would not change the fact that it is a means to the latter. — Bob Ross

First, we agree that you "intend both" (the means and the end).

Second is this problem of appropriation. You are claiming that an unappropriated means is a means. You are claiming that the nearby cup of water that you are not aware of is a means to your end of quenching thirst.

Colloquial we use "means" in two related senses: actual means and potential means. Suppose I am about to play a tennis match and I have three racquets in my bag. Each racquet is a potential means to playing tennis. But once I choose the Wilson racquet it becomes an actual means (or an appropriated means, a means that is intended). I have been using the word 'means' to mean 'actual means', and I think this is the more primary and precise sense of the word. You have been using the word ambiguously to refer to either actual or potential means.

Not in the sense as I defined it. I guess, what is your definition of a ‘means’? It can’t be how I defined it, because that definition does not preclude unintentional means. — Bob Ross

I think we agree that an actual means is intended whereas a potential means is not. In my last post I was referring to an actual means as "a (chosen) means."

So we have it that a (chosen) means becomes a part of the intention of the end, and the foreseen effects of a (chosen) means become a part of the intention of the end. But the same crucial question that I asked earlier arises here: do they become equal parts in the intention of the end? — Leontiskos

Not in the sense that I think you mean it. I only intend to walk to the kitchen because I intend to quench my thirst. — Bob Ross

Walking into the kitchen is an actual means, and quenching your thirst is an end, and both are intended. I was asking about the relation between the actual means and the foreseen effect, not the relation between the end and the actual means.

Part of the difficulty is that you are tying up foreseen effects with potential means. A foreseen effect is a potential end, and that which causes the foreseen effect is a potential means to that potential end; but because the foreseen effect is not an actual end that which causes the foreseen effect is not an actual means (to that end). What follows is that a foreseen effect is the effect of a potential means, but not an actual means.

A is a means to Q because A is useful for facilitating Q—even if one accidentally intends or doesn’t intend at all Q.

You are right that Q is a foreseen effect; but that doesn’t absolve A from being a means toward that effect. — Bob Ross

A is a potential means but not an actual means to Q. Because A is not an actual means to Q, A is not intended vis-a-vis Q. A means is only intended if it is being used to achieve some intended end, and therefore A is not an actual means to Q because Q is not an intended end.

I can appreciate what you are going for: he “means” to achieve P and not Q in the sense that P is the main, primary goal and Q is not at all. But this does not absolve him, in this scenario, from intending Q; because a foreseen effect of a chosen means is intentional. — Bob Ross

I would not say, "a foreseen effect of a chosen means is intentional," although this is true. The problem is that you are falsely implying that A is an actual means to Q. I would rather say, "a foreseen effect of a chosen act is indirectly intentional." Aquinas would say that Q is not intended, but that Q does fall under the agent's intention (as an accidental part of that intention). Q is not intended because only means and ends are intended, and Q is neither. Nevertheless, Q falls under the agent's intention because it is accepted as a known consequence of his action. But sticking with Brock's language, we can simply say that Q is indirectly intended.

It seems like, then, you are agreeing with me with different words: Q is, in this example, intentional—but indirectly. Then it is not true that “He means only to achieve P”: he means to achieve P, and this requires him to achieve Q as well (simply because A is a means towards both and never one or the other). — Bob Ross

We agree that Q is indirectly intended, but I would not say that he means to bring about Q, nor that he is required to achieve Q. I would not say the first because A is not an (actual) means to Q, and I would not say the second because it is not an accurate use of the word 'achieve'. I would say that he is required to accept Q, not achieve Q.

It is useful for achieving Q—which has not direct relation to any intention. There could be no intention towards Q, and it would still follow that Q can be facilitated by way of A. — Bob Ross

It is useful for achieving Q but it is not being used to achieve Q, and therefore it is not being used as a means to Q. It is not an actual means to Q, but it is a potential means to Q.

It is appropriated via intention if one accepts my premise that “if one foresees an effect (Q) of a means (A) and chooses to use that means (A) to achieve their intention towards another effect (P), then they thereby intend Q”. — Bob Ross

A key problem is your idea that the word "means" does not indicate a relation. To talk about a means only makes sense in relation to some end, and therefore a means always involves a relation to an end. You are saying that A is a means and then you are equivocating on the end. A is a means qua P but A is not a means qua Q.* This is subtle, but Q is not an effect of a means, it is an effect of an act. A is simultaneously an act and a means, but it is only a means in relation to P. We should not say, "Q is the effect of a means," we should say, "Q is the effect of an act," or, "Q is the effect of a cause." Similarly, we should not say, "P is the effect of a means," but rather, "P is the end of a means," or, "P is the goal of a means."

This is why I was wary to call it “per accidens”, because I am NOT referring to a colloquial usage of the term “accident”: the latter is used commonly to refer to something someone didn’t intend to do. Manslaughter is when someone unintentionally kills someone: having a per accidens intention to kill someone for the sake of a per se intention to save someone else is NOT an unintentional killing.

The accidental aspect I am referring to, is the part of the intention, which is still an intention, that is required in the specific circumstances to achieve the original, per se, intention. Both are intentional, even if they are intended in different ways. That’s not how manslaughter works. — Bob Ross

Indirect intention does involve a kind of absence of intention. Involuntary manslaughter does work that way. Negligence is a form of indirect intention. The trolley lever-puller might be charged with involuntary manslaughter, but they would not be charged with murder.

So the first wall of your castle was the idea that a necessary condition indicates a means, and we have overcome that wall. The second wall is the idea that A is a means to Q, and I think we are close to overcoming that wall. The third wall is now in play, which is the idea that A is impermissible because Q is indirectly intended.

Regarding this third wall, suppose there is an evil and it is morally impermissible to directly intend this evil. Does it follow that it is impermissible to indirectly intend this evil?

I think not. Take the matter of the especially bad car emissions due to a faulty exhaust system. Is it impermissible to directly intend those emissions? For example, to allow your car to idle for the sake of the emissions? I think so. Does it follow that it is impermissible to get groceries in the car, even when you know it will produce those emissions? No, I don't think so.

* Or if you like: A is an actual means qua P but A is only a potential means qua Q. The point is that a potential means has no effect qua means, for it is not actualized as a means.

Firstly, my argument, which has been refined quite a bit with the help of @Leontiskos, does not hinge on a principle of “never treat a person as a mere means, but always simultaneously an end-in-themselves” (EP) because I don’t think it is enough: I am thoroughly convinced, by the help of @Leontiskos, that swerving is the means to saving the two and killing the other two is not a means towards that. However, killing the two is still immoral; because I maintain that it is an intentional killing of two innocent human beings.

I hold a much stronger principle than EP, which is something like “never treat a person without the dignity which they innately deserve”. We can call this DP. IP, as you called it, stems from this principle (ultimately).

Secondly, what exactly do you mean by “treating someone as an end” vs. not <...>? I would say that, although one is not treating the sacrificed two as a means towards saving the other two, one is, by intentionally swerving with the knowledge that it will kill them, violating those two sacrificed people’s right to not be killed when innocent and, in virtue of that, are thereby not treating them as an end-in-themselves. Your distinction rests on the assumption that not treating someone as a means entails that they are treated as an end-in-themselves—which is a false dichotomy in my view.

Thirdly, letting something bad happen to someone because one intends to not do immoral acts is not a means towards that intention. A really easy way to explain why is by example: if I intend to abide by your EP principle and I am faced with either (1) doing nothing or (2) treating them as a means, then by your own logic if you choose #1 you have actually treated them as purely a means towards EP and so #1 collapses into #2.

The reason this absurdity occurs, is because you are using the term “means” entirely too loosely (which I have done as well—so you are not alone in this [; ). Nothingness cannot facilitate anything, so it cannot be a means; that is, not acting can never be a means because it cannot facilitate anything (since it is purely negative: it is nothing) nor can whatever is inacted upon as a means be a means because nothing was acted upon.

What you are implicitly doing, is thinking of non-activity like activity. If I act in a manner where I pull a lever with the intent to save the five, then the lever and (arguably the) action is a means towards that intention; but if I don’t pull the lever (which is purely a negation of what could have been done) with the intent to not sacrifice the one (or simply to not save the two), then there is nothing being used, by the agent which did not act, all else being equal, which facilitated that intention. Instead, in the latter case, what facilitates the intention is already in play--and that is the whole point of not doing something for the benefit of an intention.

It follows that the consequences of not doing something can never be a means, nor can the inaction itself. So I reject that when I let the four die, that I am using them as a means towards my intention.

However, I would like to note that because I do not use them as a means by letting them die it does not immediately follow that my inaction is morally permissible; because, although I am not intentionally killing them, it may be wrong that I am letting them die (e.g., if I could save them easily without doing something immoral, then letting them die is morally imperissible). You simply do not have this sort of distinction in your view, because you just view it through the lens of means vs. ends.

So we either fail to treat two people as ends, or we are complicit in the failure to treat millions as ends.

Again, my whole view hinges, in this case, on whether one can save the millions without committing anything immoral; and, to keep things simple, this hinges on whether or not they are intentionally killing innocent people. We are just approaching it totally differently. Your conclusion here is utterly consequentialistic.

I disagree. You have fallen into the consequentialist trap. You think it is ok to use people as a means towards (or at as sacrifices for) good ends.

[1] If you kill someone, then you are not treating them as an end (unless it is mercy killing).

[2] If you let someone die when you could save them, the same is true.

[...]

...the principle that we should treat people as ends rather than just as means (which I shall label EP)... — Herg

Similar to what has said, I don't think (1) or (2) violate EP. (1) and (2) fail to treat someone as an end, but they do not treat that person as a means. EP requires that we "treat people as ends rather than just as means." (1) and (2) violate the separate principle that we must always treat everyone as an end, which is not a commonly accepted moral principle.

Surely the point here is that if let Hitler live, he will continue to fail to treat millions of people as ends by murdering them; and our only way of treating those people as ends is to fail to treat Hitler, and unfortunately the janitor, as ends. So we either fail to treat two people as ends, or we are complicit in the failure to treat millions as ends. — Herg

This is the same equivocation between EP and the separate principle. Kantian morality does not admit of perplexity, where there are cases where we must decide who to treat as an end. It is not true that "our only way of treating those people as ends is to [treat Hitler and the janitor as mere means]."

The reason I edited my last substantial post to you is because I went back and read more of your posts (). I am not sure how well Benthamite utilitarianism mixes with the second formulation of Kant's Categorical Imperative. It is a bit of an odd mixture, and this movement from (Kant's) EP to the separate principle is a case in point. As I see it, the difficulty is that EP can't really fit the role of a "subordinate end," to use Bentham's language. Bentham's approach seems opposed to Kant's, and Kant seems directly opposed to consequentialism.

Leontiskos: I haven't forgot about your original response: I will respond later to it, as I don't have enough time right now.



I just wanted to clarify something about my response; because I think I slightly blundered in a couple spots.

If I let something facilitate my goal, it does not follow that I used that something to facilitate my goal: the latter is an action, but the former is about inaction. However, something which facilitates my goal, even if it is already in play (and so I didn't actually use it), is something which is a means (and, to Leontiskos other point, an actual means) towards my end; because that something is facilitating my goal. So it does follow that my inaction of not pulling the lever, which is to say that the lever not being pulled, is the means by which I achieve my intention of having no effect on the situation.

In the example you gave, letting the five die nor killing the one is a means to my desired non-effect: not pulling the lever, or more precisely the lever not being pulled, is that means. The five people dying is an effect of that means being utilized, which in this case is already being utilized because it is already in play. So, to recap:

1. I did not use the five, by letting them die, as a means because there is no action I took which leveraged a means towards that end; and

2. Letting the five die is not a means towards the end of doing nothing: it is, rather, an effect, in some cases, of doing nothing.

Check my post for the use of hypotheticals.

In short, it is mostly a private matter for us to meditate upon and tinker with. Some will find use in tinkering and others will be repulsed by what they find and run away screaming.

disagree. You have fallen into the consequentialist trap. You think it is ok to use people as a means towards (or at as sacrifices for) good ends. — Bob Ross

Consequentalism gives the proper result in the scenario I was talking about: You bomb Hitler and end the war, even if it means the death of one innocent person.

1. I did not use the five, by letting them die, as a means because there is no action I took which leveraged a means towards that end — Bob Ross

In the case of Alan, Betty, Charles and Dora, where the driver let Charles and Dora die by not turning the wheel, can we at any rate agree that you consider the lives of Charles and Dora to be less important than obedience to the rule that you should not kill an innocent person by positive action?

[1] If you kill someone, then you are not treating them as an end (unless it is mercy killing).

[2] If you let someone die when you could save them, the same is true.

[...]

...the principle that we should treat people as ends rather than just as means (which I shall label EP)...
— Herg

Similar to what ↪Bob Ross has said, I don't think (1) or (2) violate EP. (1) and (2) fail to treat someone as an end, but they do not treat that person as a means. EP requires that we "treat people as ends rather than just as means." (1) and (2) violate the separate principle that we must always treat everyone as an end, which is not a commonly accepted moral principle. — Leontiskos

It's 50 years since I read Kant, so I am horribly rusty. But when I look up Kant's second formulation of the categorical imperative, I find that it reads as follows:
"Handle so, dass du die Menschheit sowohl in deiner Person, als in der Person eines jeden andern jederzeit zugleich als Zweck, niemals bloß als Mittel brauchest“
Google translate renders this as:
"Act in such a way that you use humanity both in your person and in the person of everyone else at all times (1) as an end, (2) never just as a means.”
I have inserted the numbers here. I take (1) to be complete as it stands, and (2) to be entailed by (1). Because (1) is complete and is not dependent on (2), I question your statement that "the separate principle that we must always treat everyone as an end... is not a commonly accepted moral principle," because that is in fact (1). (Unless you mean that few people accept Kant's second formulation. But I don't think you mean that.)

Whatever Kant meant by what he wrote, the emboldened rendering above is what I was aiming for (except that I think all beings capable of pain and/or pleasure should be treated as ends, not just humans: "The question is not, Can they reason nor Can they talk, but, Can they suffer?" ).

Surely the point here is that if let Hitler live, he will continue to fail to treat millions of people as ends by murdering them; and our only way of treating those people as ends is to fail to treat Hitler, and unfortunately the janitor, as ends. So we either fail to treat two people as ends, or we are complicit in the failure to treat millions as ends.
— Herg

This is the same equivocation between EP and the separate principle. Kantian morality does not admit of perplexity, where there are cases where we must decide who to treat as an end. — Leontiskos

I'm sorry, I don't know what you are getting at in your second sentence here. Can you put it another way?

I am not sure how well Benthamite utilitarianism mixes with the second formulation of Kant's Categorical Imperative. It is a bit of an odd mixture, and this movement from (Kant's) EP to the separate principle is a case in point. As I see it, the difficulty is that EP can't really fit the role of a "subordinate end," to use Bentham's language. Bentham's approach seems opposed to Kant's, and Kant seems directly opposed to consequentialism. — Leontiskos

I don't see the EP as a subordinate end, and I apologise if I gave the impression that I did. It's rather the other way round: I see the EP (in my two-part formulation) as primary, and the hedonic calculus, if we need it all, as secondary.

In fact I am no longer sure whether we need the hedonic calculus. I am a hedonist, and so I think that treating people as ends must in the end be a matter of trying to give them more net pleasure: but I don't think this necessarily commits us to the traditional utilitarian hedonic calculus. But I must confess that I only recently stopped being a utilitarian, and my ideas in this area are still somewhat in flux.

This just begs the question.

In the case of Alan, Betty, Charles and Dora, where the driver let Charles and Dora die by not turning the wheel, can we at any rate agree that you consider the lives of Charles and Dora to be less important than obedience to the rule that you should not kill an innocent person by positive action?

I believe so (if I am understanding correctly). What you are saying is that, under my view, saving Charles and Dora is less important than not doing immmoral acts; and that I certainly agree with (and I think so should you).

I am a virtue ethicist, so I think a moral compass is the most vital and important aspect of normative ethics--it is the kernel so to speak. Being a moral agent, in the sense of embodying what is good and not what is bad (by doing at least morally permissible and obligatory actions), is of central and paramount importance. Any theory that posits otherwise seems to be missing the point of normative ethics entirely (IMHO).