According to Ethical Response-Dependent Realism (ERDR), moral properties are response-dependent: for something to instantiate a moral property is for it to be disposed to elicit the right kind of response in the right kind of respondent. Most mornings I like ERDR. I like it mostly because it allows us to combine cognitivism and internalism. If a moral commitment is the cognizing of a disposition to elicit a motivational response in me, then both cognitivism and internalism are true of moral commitments.
Sadly, I just read a very interesting argument by Nick Zangwill against ERDR (in Erkenntnis 59: 285-290, 2003). The argument is that non-vacuous versions of ERDR are incompatible with the existence of action on the motive of duty. The argument is based on three ideas:
(DUTY) There is such a thing as acting on the motive of duty – doing the right thing for its own sake
(KNOWN) The essence of moral properties either is (or at least could be) known by us.
(TRY) If an agent A knows that E is the essence of properties P, then in trying to instantiate P, A tries to instantiate E.
Side note: Zangwill construes the “is” version of KNOWN as a commitment of analytic ERDR, according to which the response-dependence is part of the very concept of a moral property, and the “could be” version as a commitment of synthetic ERDR, according to which the response-dependence is the unobvious underlying nature of a moral property.
Zangwill’s argument can be formulated as a reductio. Assume for reductio that ERDR is true. The conjunction of ERDR and KNOWN entails the following:
1) Moral properties’ being dispositions to elicit the right responses in the right respondents either is (or at least could be) known by us.
The conjunction of 1 and TRY entails:
2) In trying to instantiate moral properties, we try to instantiate a disposition to elicit the right responses in the right respondents (or at least would upon discovering that moral properties are response-dependent).
However, acting on the motive of duty just means that one acts with disregard for the responses one’s action might elicit. That is, DUTY entails this:
3) We sometimes try to instantiate moral properties but do not try to instantiate any responses in any respondents (and would not try to instantiate any responses in any respondents upon discovering that moral properties are response-dependent).
It’s important to realize that giving up on 3 means giving up on the possibility of acting purely on the motive of duty.
The problem is that 2 entails the negation of 3:
4) We never try to instantiate moral properties without trying to instantiate any responses in any respondents (or at least would never so try upon discovering that moral properties are response-dependent).
Thus we reach an absurdity: the contradiction between 3 and 4.
I don’t know if there’s been discussion of this argument in the literature, but it seems pretty plausible on the face of it. The weakest links are probably those bridge principles that do not pertain to the meta-ethical substance of the argument, notably TRY. Suppose I try to buy gold. Is it true that I am thereby trying to buy atomic element 79? Perhaps the intensionality of trying is enough to secure that something like TRY is false. At any rate, that would be my first reaction in defending ERDR.
Suppose this is true of moral properties,
“. . .moral properties are response-dependent: for something to instantiate a moral property is for it to be disposed to elicit the right kind of response in the right kind of respondent.”
Let’s instantiate. Action A is morally right just in case it (tends to) elicit praise (approval, etc.) from well-informed, impartial respondents (or something of this kind).
Suppose it is essential to A that it elicits such responses and that TRY is true. It does seem to follow that,
1. In performing A I am trying to elicit praise (approval, etc.) from well-informed, impartial respondents.
Agreed. But how am I not performing A because A is the right thing to do? Part of what it means to say that A is the right thing to do is to say that A elicits certain sorts of responses. So, yes, given TRY, I am performing A because it elicits the right sorts of responses. But that is a long way from saying that I am not performing A for the sake of duty. Assuming TRY, I am not performing A for the sake of duty if (2) is true,
2. In performing A I am trying to elicit praise (approval, etc.) from well-informed, impartial respondents in order to try to elicit praise (approval, etc.) from well-informed, impartial respondents.
But I can affirm (1) and deny (2). In performing A I am trying to elicit praise from well-informed respondents, but that is just what it is to act from duty. I am not acting from duty in order to further elicit praise from well-informed, impartial respondents. So long as the iteration in (2) is false, I am acting autonomously (or for the sake of morality alone as in (1)) and not heteronomously as in (2).
Uriah,
Here’s a way to defend ERDR from Zangwill’s line, but it’s not a way you’ll like.
Even if there’s an equivalence between a certain kind of response having been elicited and the moral truth, Socrates taught us that there’s still an ambiguity to clear up. Either the response determines the moral truth, or the moral truth determines the response. The version of ERDR Zangwill is after supposes that the emotional response in the right kind of respondent does moral truth-determining work. But the ‘direction of fit’ could go the other way. In that case, the emotional response in the right kind of respondent is something like really good evidence that the something instantiates a moral property.
Take that route and Zangwill’s question towards the end of the paper — “in virtue of what is the response warranted?” — gets an answer: some response-independent moral property. This version of ERDR doesn’t give rise to the worrisome contradiction, and it’s compatible with DUTY, KNOWN, and TRY. However, internalism comes out false.
I wonder if anyone who is trying to go for ERDR would actually need to mind whether DUTY leads to problems in their view. They could say that this just explains our intuition that we should not be acting on the motive of duty because it also turns out to be an incoherent notion.
There are many ways one could argue for this. First is that duty here seems to mean all things considered duty, and acting from a duty is easy to understand as taking that something is your duty to be a reason for your action. These ideas do not seem to fit together. What our duty is just seems to be where the contributory reasons of the situation point – what is the final verdict of our reasons-judgments in the situation. It would be then odd to take the result of this process to be a further reason for your actions. For one, one would need to make another over-all judgment about reasons. This would lead to a regress. (Sorry, a bit of typical Reading philosophy there.)
Another point is that acting from a duty seems to come close to being a moral fetish. You don’t then care about the duty-making features, human suffering, disappointed intentionally creating expectations, well-being of others, and so on. If you act on the motive of duty, then you only get started when you notice that there is a duty based on these considerations which by themselves do not move you. This seems far of from our idea of virtuous actions. Duty by itself seems something too general, abstract and alianated to be something of which agents should be concerned about directly.
So, I think defenders of ERDR could do well by admitting that acting from a motive duty is an incoherent notion, or at least a motive the acting from which does and should not produce positive reactions. This is something which the view can accommodate. On the other hand, ERDRers can tell an interesting story of how virtuous agents can be motivated directly by the duty-making properties as a result of having the right kind of a response which in their view are extension determining for our moral notions.
Uriah,
Your first response to defending ERDR seems exactly right to me. If ‘tries to F’ is read extensionally, then TRY is true. If read intensionally, then TRY is false. On the true version of TRY, 4) above is true, so no reductio.
To keep it very simple just suppose the strong equivalence in (1) is true, where M is some moral action and E is some eliciting of response R.
1. M = E
To act for the sake of duty is to perform M for morality’s sake. Therefore, given (1), to act for the sake of duty is to perform E for morality’s sake. Seems to me perfectly coherent to say that x tried to elicit response R for the sake of morality alone. So I guess I can’t see where the reductio gets hold.
Uriah,
That strikes me as a pretty persuasive and interesting argument. In response to your doubt at the end:
“Suppose I try to buy gold. Is it true that I am thereby trying to buy atomic element 79? Perhaps the intensionality of trying is enough to secure that something like TRY is false.”
Certainly if you don’t know that gold is element 79, you need not be trying to buy element 79. In fact, as long as you have any *doubt* that gold is element 79, you need not be trying to buy element 79 (you can take the attitude: “Probably gold is element 79, but *whether or not it is*, I want some gold. If gold isn’t element 79, then I don’t want element 79.” But it’s doubtful that you can take this attitude if you are *certain* that gold is element 79, and *you’re bearing this in mind* at the time you get on the phone to buy some of it.
Of course, I just supposed stronger conditions than are stated in (TRY) above, but that’s okay. It’s plausible that, even if you know with certainty what duty is and are bearing it in mind, you can still act for the sake of duty. Acting from duty shouldn’t require one to forget or be uncertain about the nature of duty.
Hi Uriah,
I guess I might chime in here. I actually have a response to Zangwill’s paper (“Moral Response-Dependence, Ideal Observers, and the Motive of Duty: Responding to Zangwill”, Erkenntnis 60:3 (2004), pp. 357-69). In it, I argue (as many have in the comments above) that it is quite plausible to hold that one can act for the sake of duty even while recognizing that rightness is a matter of eliciting certain responses in ideal observers. [E.g., we could ask an agent whether she’d continue to seek the approval of ideal observers if she were to come to believe that rightness was not determined by such approval].
More tentatively, I considered in the paper whether the conjunction of (DUTY), (KNOWN), and (TRY) would pose a parallel problem for just about any rival to ERDR. Zangwill writes the following:
[B]eing motivated by things under their naturalistic guise is different from being motivated by things under the guise of the morally good, which is what the motive of duty requires. So if a moral response-dependent theory says that we are motivated to pursue the response of approval or the naturalistic features of things [which prompt such approval], the motive of duty is lost (Zangwill 2001, 273).
Zangwill only discusses response-dependent theories explicitly, but consider: he draws a firm distinction between being motivated by things under the guise of the morally good (required by the motive of duty), and being motivated by the naturalistic bases of moral properties, or approvals. But it seems the same distinction be drawn for any account of the morally good (or moral properties in general). And if so, won’t there similarly be a difference between pursuing the morally good simply for its own sake, and pursuing (for example) the morally good qua non-natural property? ‘I want to do what is right – the morally right action – because it is right, not merely because it possesses some non-natural property.’ If I explicitly understand moral goodness to be a non-natural property, can I no longer act from the motive of duty? As soon as we provide any account of the nature of moral properties it would seem that we would lose the pure motive of duty (if Zangwill were correct), because we’d now (assuming (KNOWN)) be pursuing whatever it is that moral properties amount to, be it naturalistic or not.
[Minor bit o’ trivia. Zangwill’s paper was actually published twice in Erkenntnis – once in 2001, once in 2003. The last time I checked, the 2001 version wasn’t available online. My response was to this 2001 printing.]
Here’s a small point — not sure whether it’s relevant.
On ERDR, an act need not actually elicit approval from anyone in order for it to be morally right. It is sufficient that the act would elicit such approval in certain counterfactual circumstances. I’m not entirely sure what ‘acting from duty’ is, but perhaps it could be said that, although trying actually to elicit approval is inconsistent with acting from duty, trying to act in a way that would counterfactually elicit approval is not. In the former case, where it’s actual approval you’re after, we might say you only care about having a good reputation.
I guess Mike Almeida (are you Biff or Skippy?) made a similiar point, but here’s an analogy: I am a concert pianist who wishes to perform a sonata well. Furthermore, some sort of response-dependent theory of aesthetic value is true: A work of art has the aesthetic property beauty, say, iff it elicits the right aesthetic response in the right observers (dodging all the circularity worries for the moment). In wanting to perform the sonata well, I do try to get the audience to have the aesthetic response, but I also want them to have the aesthetic response because of certain features of how I played instead of other facts to which they might be aesthetically responsive. So I want them to note the cadence of my playing, changes in tempo, etc. rather than applauding because they like how shiny the piano is. So I can try to cause the response in them in the right way; hence, I’m trying to get them to have the right response “for its own sake,” one might say. Couldn’t the same be true under ERDR: that we can act trying to induce a response to morality “for its own sake”?
Jason,
I think non-naturalist intuitionism is the only theory that avoids the objection (which is unsurprising, because it is the truth. The truth avoids most objections.) It seems as though you’re treating the non-naturalist thesis as if it were another (rather strange) reductionist theory: ‘good’ reduces to ‘some non-natural property [with some further features, presumably]’. But that isn’t the view — the non-naturalist (such as Moore) thinks that good is undefinable and irreducible. Calling it “non-natural” is just a spooky way of saying that. (Moore actually had a strange view of what ‘naturalness’ is, but it’s irrelevant. Probably no one else would hold his strange view, and it’s unrelated to the Open Question Argument.)
Pretend that Moore also said that “right” and “duty” were indefinable (as he should have). Because Moore* simply refuses to analyze “right” at all, refuses to give an account of its nature, there are not two guises under which you can understand “right” on his view. You don’t have any two ways of being motivated to distinguish.
What is difficult in Moore is that the metaphysics of good gets mixed up with the semantics or meaning of ‘good’. So it is true that,
1. the property being good is identical to a simple non-natural property.
(1) makes a metaphysical claim about goodness. It obviously says nothing about the meaning of ‘good’. But Zangwill (Z) (as far as I can tell) is also saying something about the metaphysics of moral properties.
According to Z’s argument above.
(2) Moral properties just are dispositions to elicit the right responses in the right respondents.
Z runs his argument on (2), assuming TRY, and he gets this,
“In trying to instantiate moral properties, we try to instantiate a disposition to elicit the right responses in the right respondents”.
But then, as Jason suggests, run the argument on (1) and you get this,
“In trying to instantiate [the] moral property [is good], we try to instantiate a simple, non-natural property”.
But now Z objects that,
“We sometimes try to instantiate moral properties but do not try to instantiate any responses in any respondents”.
So now Jason might object that,
“We sometimes try to instantiate [the] moral property [is good] but do not try to instantiate a simple, non-natural property”.
So Jason’s argument seems to mirror pretty closely Z’s. Maybe I missed something that would make some serious difference.
Many very interesting comments here, folks. We got a bunch of ways to respond to the argument: from biting the bullet while showing it’s a perfectly edible one (Jussi), through rejecting the argument’s validity (Mike A. and Mike C.) or rejecting TRY (Robert and tentatively me), to claiming that the argument proves too much (Jason).
I’m particularly intrigued by Jason’s line of thought. It shows that Z’s argument doesn’t have anything in particular to do with ERDR. It would work the same for any non-primitivist theory. If you hold that the morally good is a primitive (irreducible, sui generis, and single-guised) property, you’re safe. But otherwise, you’re susceptible to the Z manoeuvre (as it shall henceforth be known). That’s basically what Mike H. is saying (except that, although intuitionism is the natural epistemological companion of ontological primitivism, there isn’t a necessary connection there as far as I can tell).
I’m a bit disconcerted by this because while I like ERDR only most mornings, I dislike primitivism all mornings. All this disconcert would be avoided, of course, if I could show that TRY is indeed false, or as Robert puts it more subtly, false in an extensional reading and not strong enough in an intensional reading. Unfortunately, Mike H.’s point seems to me right, on the face of it: there is an extensional variation on TRY that seems true and good enough to grease the argument. Hopefully Robert will post a refutation here soon!
I’m prepared to reconsider, but there is nothing so far that shows this is mistaken. In suggesting that,
“We sometimes try to instantiate [the] moral property [is good] but do not try to instantiate a simple, non-natural property”
I do intend (like Z) that we sometimes try to instantiate the good for its own sake, and that in such cases we are not intending to instantiate a simple, non-natural property. Granting for the moment the intensional reading of TRY, Z’s argument holds just as well against the position that moral properties are simple, nonnatural properties.
What Mike just said…
Ok, so we’re on an offshoot here concerning the question whether Jason’s generalization of Z’s argument extends to primitivist accounts of moral properties. If it doesn’t, non-primitivists have to worry about the generalized Z argument. If it does, the generalized Z argument doesn’t cut against any ontological account of moral properties and is thus epidialectical (if you will) in the present context. Jason and Mike A. think that it does, Mike H. and I think it doesn’t. Mike H. is happy about this, I’m unhappy.
Perhaps the issue boils down to the question “Can a primitive property be multi-guised?” The way I (and perhaps Mike H.) thought about it, primitive properties are necessarily single-guised. If they are, then the Z maneuver can’t be applied to them. But if primitive properties are multi-guised, then it can. From what Mike H. and Jason are saying, I think you guys are allowing primitive properties to be multi-guised, and take “morally good” and “non-natural property N” to be two guises of the same property. My inclination, and I think Mike H.’s too, is to say that these are not really two different guises. In one mode of thought, the latter doesn’t seem to me to be a guise at all.
Of course, that’s not something we can just say. We need to make the case. Hopefully Mike will rise to the task… And I’m still waiting on Robert to refute the generalized Z argument once Mike shows that the argument doesn’t extend to primitive properties…
BTW, even if it doesn’t, it wouldn’t mean that the generalized Z argument is not interesting. It won’t favor any meta-ethical ontology over another, as Mike A. and Jason arguing, but it may still be taken as an argument for the incoherence of action on the motive of duty (or the basis of a “puzzle” about its coherence), which is surprising. These considerations may recommend Jussi’s original suggestion for how to handle Z’s argument. The bullet Jussi counsels us to bite would be eminently edible if everyone had to bite it anyway.
Hi Uriah,
At one point in your previous comment you suggest that being a non-natural property is not a ‘guise’ of a property at all (given one line of thought). Could you say a bit more?
I guess I’m still a bit puzzled. Couldn’t I be a person who wants to realize simple, non-natural properties? And wouldn’t realizing ‘primitivist’goodness be one way of doing this? Similarly, couldn’t I intend to realize goodness without intending to realize a simple, non-natural property (intensionally)? In your terms, it still seems to me that goodness, even on a ‘primitivist’ understanding, is ‘multi-guised’.
Uriah,
It looks like it’s sufficient to show that there are “guises” for simple, non-natural properties that substitutivity fails for at least some terms referring to that property in intensional contexts. And it looks like it is sufficient to show that there *might* be guises for such properties that substitutivity *might* fail.
So suppose Moore was exactly right. The property good is identical to the very simple non-natural property Moore discussed in Principia. Here’s a simple argument for substitutivity failure.
1. The property good (or ‘is good’) = The simple non-natural property Moore discussed in Principia.
2. Jones believes that giving to disaster relief is good.
3. Jones believes that giving to disaster relief has the simple, non-natural property Moore discussed in Principia.
Obviously (1) and (2) might be true and (3) false. Jones might have no idea who Moore is, for instance.
So I guess I’m wondering why the failure of substitutivity would not show that there are guises for that simple, non-natural property.
I’m still trying to get a handle on why, once I believe P = E, I now must be trying to instantiate E in trying to instantiate P. Perhaps it’s a psychological conjecture — now I just can’t help thinking of P under the heading of E? Sounds sort of pathological. Perhaps another principle, closer to the ‘open question’, I could see being true. Something such as:
If an agent A knows that E is the essence of property P, then if he tries to instantiate P, he will believe that in instantiating P he will be instantiating E.
This doesn’t deliver the conclusion Nick wants, though, since it doesn’t follow from the fact that he believes that in instantiating P he will be instantiating E that he is trying to instantiate E (cf. forseen but unintended consequences).
I want to try something desparate here in order to just understand how the objection is supposed to work. I take it that the crucial premise is the claim:
1. There is such a thing as acting on the motive of duty – doing the right thing for its own sake.
The argument then is that given ERDR and certain other plausible assumption the notion of acting on the motive of duty is an incoherent notion, and thus something’s got to go – either 1 or ERDR. I wonder. Consider:
1*. There is such a thing as drawing a square circle.
It is easy to derive a contradiction from this. One would draw a rectangle whose four sides are of equal length and a closed curve along which every point is the same distance from a fixed center point, but if one does one of these one is not doing the other. But does that mean that anything has got to give? I’m not sure. It may be that one can be doing the act of drawing a square circle, if ‘act’ here is not understood as a success term. Consider:
1**. There is such a thing as running a marathon.
Well, one can be running a marathon without completing it. People could be running marathons even if no-one ever finished or could finish one. Something less is required for being doing that. So, maybe one can be drawing a square circle without completing one (because it is impossible). Something less than success is required. And, maybe then one can be acting from the sake of duty without complete success (because again it is impossible). This would imply yet again that there is something like acting from the motive of duty, even though doing it succesfully would be logically impossible. Something less again is required.
Or maybe nothing counts as drawing a square circle because the end-result is impossible. But maybe something would still count as trying to draw a square circle. Many mathematicians have tried to prove theorems even though success here has then turned out to be impossible. So, maybe the intuition 1 is based on really is that
2. There are such things as tryings to act on the motive of duty.
Maybe our conviction of the truth of 1 is just conviction of 2, and nothing is lost if we abandon 1 after reflection and retain 2. If this was the case, then the derived contradiction from 1 would not be a worry.
But, maybe the reply to that is that as a square circle is a contradition, nothing counts as trying to draw one either. Trying requires an intention that must involve a representation of what is intended, and in the case of square circle none is available. Thus, there would be no tryings either. Maybe this objection would go for trying to act from the motive of duty too, and thus we couldn’t have 2 either.
But then, what is the evidence for 1 anyway? Yes, we often think we act from the motive of duty. However, maybe all the argument establishes is that this belief is false because there is no motive of duty. Maybe the experiences of acting from the motive of duty just are illusions of thought. This would imply that our experiences are not sufficient evidence for 1 and a further argument is needed that it is true.
Sorry about this. I think I’m too confused about the whole thing. But the Z objection seems to need more work in order to worry me.
Mike A,
You wonder what if anything differentiates Moore’s thesis,
1. Goodness is identical to a simple non-natural property.
from
2. Goodness is the disposition to elicit a response of approval (or whatever) in the appropriate respondents. (my paraphrase)
At first glance, one wants to say there is the important difference that (2) is a reductive theory of what goodness is, while (1) is not; (1) is a refusal to give a reductive theory. The phrasing of (1) is misleading, since the most perspicuous phrasing does not include any identity predicate — it’s like rephrasing “the sky is blue” as “the sky is identical with something blue.”
Why does this matter? (1) just says something that is true of goodness; (2) allegedly identifies its essence. If I want to bring about x, I may not care about some particular feature of x, even one that I am certain is a feature of it. But it’s less plausible that I may not care about the essence of x. While goodness might be necessarily indefinable, saying that presumably doesn’t explain the nature or essence of it.
Thanks Mike (H.). You suggest that the ‘is’ is better rendered as predication rather than identity. Suppose that’s right. Still all we need here is (1), and Z’s argument goes through the same way.
1. The property of goodness is essentially a non-natural property.
Now suppose I learn that the property of goodness is essentially non-natural. According to Z, if I intend to instantiate the property good, I must also intend to instantiate a non-natural property. But–as the old story goes–I may well intend to instantiate the good and not intend to instantiate any non-natural property at all. Same problem, so it seems.
Not sure how you’re understanding “essentially,” Mike. If you just mean (1) to say that goodness is necessarily non-natural, then I agree that even if you know (1) holds, and you intend to realize good, you needn’t intend to realize any non-natural properties (the reason is that while you know you’re going to realize a NN Prop., you might not care about the non-naturalness; you might care only about other aspects of what you’re going to bring about).
Without having read Z’s paper, I conjecture that he would say (and I think he should say, whether or not he would) that you have to take essentiality more robustly for (TRY) to be true. The essence of x has to mean not just something necessarily true about x, but something that explains *what x is*, something that constitutes x, that is the core of x-ness, etc.
Well, I’ll let Mike A. and Mike H. battle it out before I stick my nose in – seems like the judicious thing to do… Let me just say that Mike A.’s argument for the multi-guisedness of primitive good was effective in shaking my confidence in its single-guisedness (so, Jason, for now we can ignore my comments to the contrary).
I’ll also wait to see some reactions to Robert’s latest post. If Robert’s right, then neither Z’s argument nor the generalized argument go through.
Jussi: slowly slowly I’m warming up to your approach to Z’s argument. I like the idea that the intuitive pull we feel toward your 1 derives from that we feel toward your 2. Still, call me conservative, call me old-fashioned, I think I feel some extra pull toward 2. But certainly you make the bullet much more pleasantly bitable with this consideration. Nonetheless, I think other things being equal Robert’s approach would be dialectically safer if it worked.
“you have to take essentiality more robustly for (TRY) to be true.
I do understand by ‘x is essentially P’ that ‘x is necessarily P. There is some work on more or less interesting theses of essentialism in T. Parsons (et. al.) but I don’t see that it bears much TRY. As far as I can see TRY applies in every case where KNOWN is satisfied. Here’s KNOWN again,
(KNOWN) The essence of moral properties either is (or at least could be) known by us.
I know that being good is essentially a non-natural property. That seems to satisfy KNOWN.
(TRY) If an agent A knows that E is the essence of properties P, then in trying to instantiate P, A tries to instantiate E.
TRY seems to rely on the epistemic question alone (not on whether I care about the essential property in question or whether the property has other special features). It is knowing it’s essential that somehow (see Robert’s sobering observations on this bizarre point) gets me to intend to instantiate that essential property when I try to instantiate good. TRY tells me in this case that I am trying to instantiate a non-natural property every time I try to instantiate the good. And that of course (contrary to TRY + KNOWN) is untrue.
I can’t see any independent motivation for restricting TRY to essential properties that I care about or to the set of more robust properties, etc. Of course, maybe there’s some motivation for it that I’m missing.
Oh, Z would presumably say that “essence” in the whole argument, both in (KNOWN) and in (TRY), has to be understood in this robust way.
Note, btw, that Z uses the locution “the essence of x is…” whereas you use the locution “x is essentially…”. The latter is much more easily read as “x is necessarily… “. It sounds wrong to say that “the essence of goodness” is that it’s non-natural — that doesn’t tell you what goodness is!
Perhaps a more intuitively clear case is this: goodness is not a spinach quiche. That is a necessary truth. So you might say: “Goodness is essentially not a spinach quiche.” But I take it that no one would say: “Let me tell you what the essence of goodness is: not being a spinach quiche.”
Mike,
there is no important metaphysical difference expressed in the alternative locutions,
1. x is essentially P
2. The essence of x is P
Though the definite article in (2) makes it sound like it is specifying the exclusive essential property of x. Further (2) sounds like an identity statement, though you urge against that a few notes up.
Of course ‘non-natural’ does not sound right when substituted in (2), but that is a grammatical point. Your version is this,
“”the essence of goodness” is that it’s non-natural”
This sounds wrong because the grammar is awkward. Compare the equally wrong-sounding,
“the essence of goodness” is that it’s happy”
No one is claiming that the essence of goodness is non-natural just as no has claimed that the essence of goodness is happy.
The sentence doesn’t sound wrong because the property ‘being non-natural’ does not tell us what goodness is. It would be a miraculous coincidence if the bad grammar of this sentence were explained by Moore’s “discovery” being correct. Suppose it turns out that goodness is not a simple non-natural property. Would that sentence then sound grammatical?
In any case, we can render (1) and (2) this way,
1. Goodness is essentially non-natural.
2. The essence of goodness consists in non-natural properties.
That’s how I would put it. If (1) is true, then so is (2). Both sound fine. Moore specified his simple properties negatively, but he needn’t have done that. He might have dubbed them positively Schnatural properties, and (1) and (2) could be reformulated that way.
Thanks, Mike. I don’t think I successfully communicated my point. A couple of clarifications:
First, I don’t believe that
(2) The essence of x is P
in general cannot be an identity statement. I think whether it is an identity statement depends on what is plugged in for x and P, and perhaps what the speaker’s intentions are. So I’d say “The essence of being gold is being element 79” is probably an identity statement, whereas “The essence of gold is interesting” is not.
Second, when I said that
(SQ) The essence of goodness is that it’s not a spinach quiche.
sounded wrong, I didn’t mean that it was ungrammatical. In fact, it seems perfectly grammatical to me. Even if it isn’t, ordinary speakers could understand the sentence. (You can also rephrase it as “The essence of goodness is not being a spinach quiche,” and I’d say the same thing about that.) Rather, my point was *what the sentence is trying to say* (assuming “essence” is used in a standard English sense) seems wrong.
I know that “essentially” is often used to mean “necessarily” in metaphysics; but we should bear in mind that that is a technical usage. In any case, it shouldn’t distort our reading of “essence.”
Think about ordinary contexts in which someone talks about “the essence” of something, as in “The essence of teaching is…” or “You have captured the essence of Hamlet!” They don’t mean merely a property that necessarily applies to the thing, if it exists. They mean *something like* the most important aspect of the thing in question (never mind details now). Thus, no ordinary speaker would accept
(SQ’) The essence of teaching is not being a spinach quiche.
Nor, if you said, “Hamlet was not a spinach quiche,” would any ordinary speaker say, “Yes, you have captured the essence of Hamlet.”
Last point: Your proposed statement
(2′) The essence of goodness consists in non-natural properties.
doesn’t have the right meaning for my purposes, because it doesn’t identify the supposed essence; it says something that is true of the essence without saying what the essence is. If you specified the non-natural properties alluded to at the end of (2′), then I’d say you had identified the essence. Incidentally, I think Moore would say (2′) is false because goodness is a simple property. So, on the Moorean view, you could identify the essence of goodness by saying:
(2”) The essence of goodness is goodness.
That’s the essence-identifying statement that I think Moore would assent to. (Can properties be their own essences? If so, then the Moorean would say goodness is its own essence; if not, then goodness doesn’t have an essence.)
We’ve likely hit the fundamental disagreement here. You say,
“(2′) The essence of goodness consists in non-natural properties.
doesn’t have the right meaning for my purposes, because it doesn’t identify the supposed essence; it says something that is true of the essence without saying what the essence is.”
But now you’re asking me to skip the predication stuff and get back to the identification stuff! I was doing that above, saying
1. The property ‘is good’ is identical to a simple non-natural property.
But I was chided with,
“The phrasing of (1) is misleading, since the most perspicuous phrasing does not include any identity predicate. . .”
Now there is no question that I cannot provide a reductive essence for goodness. Indeed in specifiying that the essence of goodness is the simple non-natural property described in Moore, I am agreeing that goodness is *essentially non-reducible*. But to say of goodness that it is essentially simple or essentially non-natural is obviously to say much more than that goodness is essentially goodness. So by any reasonable test it gives us a guise for goodness.
Perhaps you want to use essence in some more restrictive or technical way. That’s fine. My purpose is to run Z’s argument on goodness as Moore describes it. Well, actually, that’s Jason’s purpose. But I think he’s right.
Maybe there’s nice reason why Z might balk, but he’d have to say something more than that the essence I am specifying is non-reductive. That much I know, but it looks like the argument works anyway.
In any case, we seem to have reached the basic disagreement.
Now Uriah can tell us where we’re both wrong……
I’d keep a close eye on that spinach quiche if I were you. There’s a lunch thief at large, and I doubt that he’ll stop at sandwiches.
About these two statements–
1. Goodness is identical to a simple non-natural property.
2′. The essence of goodness consists in non-natural properties.
–my complaint about both of them is that they don’t identify the essence of goodness. My complaint about the phrasing of (1) was that it is misleading, since it superficially seems like it might be an identity statement, but it’s really just a predication. It’s of the form “(Ex) (Nx & Px & x=g)” (for some x, x is non-natural, x is a property, and x = goodness), which is really just a lengthy way of saying “Ng & Pg”. It’s not of the form “g=a”.
(2) is of the form (ignoring the plural form at the end) “(Ex) (Nx & Px & x=e)” (for some x, x is non-natural, x is a property, and x = the essence of goodness). It’s not of the form “e=a” either.
“Well, so what?” you might wonder, “How does this show there aren’t guises of goodness?” Indeed, there are guises of goodness, including the one you mentioned earlier. “The simple, non-natural property discussed by Moore in Principia Ethica” identifies goodness by one mode of presentation; “goodness” identifies it by another.
But, as I’m understanding Z’s argument, you need more than just any old guises of a thing to run the argument. The intensionality of “S is trying to bring about instances of _____” means that you’re only going to be able to guarantee substitutivity given some further assumptions. It’s going to be controversial that you get substitutivity even with any further assumptions — but I think it’s plausible. So the claim, as far as I understand it, is that
3. S is trying to bring about x.
4. x = y.
5. Therefore, S is trying to bring about y.
is invalid (due to opacity); yet, if you add further assumptions, you get a valid inference, like:
6. S is trying to bring about goodness.
7. S knows that the essence of goodness is y, and S is bearing this fact in mind at the time.
8. Therefore, S is trying to bring about y.
(7) gives three special conditions. First, it’s not just any old identity we have, but an identity that is “essence-identifying”. Second, S is aware of the identity (as an essence identification). Third, S is thinking of this fact as he tries to bring about goodness. Also, you probably need to add an assumption about the nature of the concept “good”, maybe that it also has to be essence-identifying.
So the claim is that the argument doesn’t work if you put in “the non-natural property Moore talked about in P.E.” for “y”, since that mode of presentation doesn’t identify goodness by its essence. If Moore is right, I think, there’s only one mode of presentation of goodness that actually identifies it by its essence.
Ok, you say,
“1. Goodness is identical to a simple non-natural property.
My complaint about the phrasing of (1) was that it is misleading, since it superficially seems like it might be an identity statement, but it’s really just a predication. It’s of the form “(Ex) (Nx & Px & x=g)”.
I have no idea what the forgoing means. Anything of the form of (3),
3. (Ex)(Nx & Px & x=g)
is an identity statement. Indeed (3) logically entails the paradigmatic identity statement,
4. (Ex)(x = g)
Adding conjuncts to (4) does not make it a non-identity statement. Same goes for (2′). So I’m lost there.
Further it’s evident that (1) is not equivalent to the conjunction (Ng & Pg) since (1) entails (4) and the conjunction plainly does not. So the conjunction is not shorthand for (1). It just doesn’t mean the same thing.
You note this,
“if you add further assumptions, you get a valid inference, like:
6. S is trying to bring about goodness.
7. S knows that the essence of goodness is y, and S is bearing this fact in mind at the time.
8. Therefore, S is trying to bring about y.
(7) gives three special conditions. . . So the claim is that the argument doesn’t work if you put in “the non-natural property Moore talked about in P.E.” for “y”, since that mode of presentation doesn’t identify goodness by its essence.”
I’m guessing that (7) is your interpretation of (TRY). I don’t see any special conditions in (TRY), nor do I see them in (7), but its your principle. If you say they’re there, good enough. And I doubt that the inference is valid, but let’s table that. What I flatly deny is this,
“So the claim is that the argument doesn’t work if you put in “the non-natural property Moore talked about in P.E.” for “y”, since that mode of presentation doesn’t identify goodness by its essence.”
When I refer to goodness as a simply, non-natural property I am specifying all there is to the essence of goodness. There is nothing more to specify; this is Moore’s allegedly discovery. Goodness just is a simple, irreducible, non-natural property, and there is nothing more to goodness. It is Moore’s motto (via Butler), and the motto of Principia, that “everything is what it is and not another thing”. And that is exactly what I’m stating. What is the metaphysics of goodness? Goodness JUST IS a simple, irreducible, non-natural property, and not another thing. There is nothing more to say in describing the essence of goodness. That description specifies exactly what goodness is.
Three quick clarifications/observations:
a) By an “identity statement”, I meant a statement of the form “a=b” (with “a” and “b” singular terms).
b) The proof that “Ng & Pg” is logically equivalent to “(Ex)(Nx & Px & x=g)” follows. (I use “>” for “if … then” and “<->” for “if and only if”.)
1. Ng & Pg. (Assumption)
2. (x)x=x. (Axiom–the “law of identity”)
3. g=g. (2, UI)
4. Ng & Pg & g=g. (1,3, Conj.)
5. (Ex)(Nx & Px & x=g). (4, EG)
6. (Ng & Pg) > (Ex)(Nx & Px & x=g). (1-5, CP)
7. (Ex)(Nx & Px & x=g). (Assumption)
8. Na & Pa & a=g. (7, EI)
9. a=g. (8, Simp.)
10. Ng & Pg & g=g. (8,9, Leibniz’ Law)
11. Ng & Pg. (10, Simp.)
12. (Ex)(Nx & Px & x=g) > (Ng & Pg). (7-11, CP)
13. (Ng & Pg) <-> (Ex)(Nx & Px & x=g). (6,12)
c) It is not the case that “Goodness just is a simple, irreducible, non-natural property, and there is nothing more to goodness,” because there can be other simple, irreducible, non-natural properties that aren’t goodness. For example, badness might be simple, irreducible, and non-natural, without being the same as goodness. Perhaps qualia are also simple, non-natural, and irreducible, but they aren’t indistinguishable from goodness, nor are they indistinguishable from each other.
a. Concerning your proof, I said that the conjunction did not entail (Ex)(x=g). And it doesn’t. There is nothing in the conjunction that remotely entails that there is an identity relation in your language, or the law of identity is validated, or Leibniz Law. On the other hand, there is an identity statement in (1) (that I referred to in the note above), and that’s why I said that, given the *explicit* identity statement, it does entail (Ex)(x=g). There is the difference. I never said that there aren’t logics strong enough to represent and validate the additional assumpions you make and I never said that in these logics that conjunction does not entail (Ex)(x=g). Of course, as the assumptions increase so do the equivalencies.
b. “It is not the case that “Goodness just is a simple, irreducible, non-natural property, and there is nothing more to goodness,” because there can be other simple, irreducible, non-natural properties that aren’t goodness”
This is false. First, the fact that there *might be* other properties that are non-natural, simple, etc., does not entail that goodness is not actually just a simple, non-natural property. Second, even if there are other simple non-natural properties (and not merely might be other such properties) we could more strictly describe goodness as the simple, non-natural, irreducible property possessed by everything with intrinsic value.
c. Concerning the claim, “By an “identity statement”, I meant a statement of the form “a=b” (with “a” and “b” singular terms)”.
What I said was that (Ex)(x = g) is an identity statement (and so are propositions that entail it). It obviously follows from instantiation that a = g. So there’s the identity statement you seem to be after. And any logic that has EG also has EI, so there is no worry there. So (Ex)(x = g) is also an identity statement by your particular standards.
That’s an interesting argument, Nick.
FWIW, I think the following (this may recapitulate several of the points made above)
To perform an action for its own sake is to accord performance of the action final value (and perhaps no instrumental value). That is, the action is valuable (wholly) non-instrumentally.
Let us say, with ERDR, that for the action to be valuable is (essentially) for it to be disposed to elicit a certain valuing response in some suitable respondent S.
So in accordance with the spirit of ERDR assume that for the action to be finally valuable is (essentially) for it to be disposed to elicit a certain finally valuing response in some suitable respondent S.
Say I grant that (in some suitably extensional sense) in trying to perform a finally valuable action, I am trying to perform an action that is disposed to elicit a certain finally valuing response in some suitable respondent S.
This doesn’t show that I am treating the action as valuable insofar as it is a means to eliciting such a finally valuing response. That would presuppose that the finally valuing response was itself valuable, or at least that I believe it to be. But there seems no reason to think that it need be the case that:
S’s finally valuing action A is itself disposed to elicit a valuing response from S
and if I am acquainted with the essence of value as outlined in ERDR, I need not believe that it is so disposed. Thus S finally valuing action A need not itself be valuable, and I needn’t believe that it is. So there seems no reason to think that I am pursuing the action (in part) as an instrumental means to achieving such a response.
So there seems no reason to think that when I treat the action as valuable, I am treating it as valuable even in part as an instrumental means to the achievement of a valuing response from S. So there’s no problem about my performing actions for their own sake.
So I deny that Duty, properly understood, entails (3). What it entails is
3*) We sometimes try to instantiate moral properties, (which essentially involves instantiating properties that are disposed to elicit responses from certain subjects) yet we do not do so as a means to instantiating responses in any respondents.
which isn’t inconsistent with (4).