A linguist here, and a thousand times yes. Stop dragging our science in your philosophy (and anthropology, while we're on the topic), we've had enough of that.
I agree with everything but the section on non-maximal options. I think the question of "Should I block a mineshaft?" can be understood as "What is the deontic status of blocking a mineshaft?", which is an interesting question that I don't think is a merely semantic issue. I see two ways of answering the question, and both call for a serious reunderstanding of morality.
We can say non-maximal options have no deontic status, which (though my preferred view) is far from intuitive and implies that human beings never have any true thoughts about what they ought to do, or
We can say non-maximal options entailed by an optimal maximal option are obligatory, which means we have to reformulate our moral theory (standard moral theory would say you ought to do what's best, but blocking a shaft isn't best if you block the wrong one!).
> "We can say non-maximal options have no deontic status, which... implies that human beings never have any true thoughts about what they ought to do"
I'm not sure about that implication. I take it your thought is that we can never articulate an option with perfect precision. But we can refer to particular actions (e.g. "my blocking of that mineshaft", or even "what I would do were I to block a mineshaft"), and any particular action is the performance of a maximal option. So we can have true thoughts about the deontic status of the acts we do or would have done. Which all seems commonsensical enough?
Even in hypothetical cases, I think we can supervaluate over imprecision in description that makes no normative difference. Since *every* way of blocking shaft B has the same effect, we can speak with determinate truth about the deontic status of blocking shaft B, even though it is indeterminate what precise way of blocking shaft B we are talking about.
I suppose my thought was that moral deliberation in practice is rarely as simple as a philosophical thought experiment, and in most cases where we believe we ought to do something, there's actually a slightly more specific version of that option that would be better. The problem also arises if (a) our agency extends into the future and (b) maximal options cover all the options in an agent's control. It might seem that I ought to block shaft B, but what I really ought to do is block shaft B, make myself a burrito when I get home, do some journaling about that awful mineshaft situation I was just in, and go to bed early - and so if I don't do all of those, then I haven't done anything I ought to have. But perhaps there are cases where we do have thoughts of what we ought to do that do line up with our best maximal option; 'never' was a bit too strong.
I agree that limiting the information contained in a maximal act to only morally significant information makes the view more attractive, and that the view needn't be too unintuitive. Thanks for the lovely reply!
I agree with you. But would you distinguish the semantic arguments of Geach and Thomson from Foot's general line of argument in 'Utilitarianism and the Virtues'? As I understand the latter argument, it's not concerned with semantics but with relations of conceptual dependence (e.g. she says that the goodness or badness of consequences, or states of affairs, is “found. . .within morality, and forming part of it, not standing outside it as the ‘good state of affairs’ by which moral action in general is to be judged.”).
Sep 5, 2023·edited Sep 5, 2023Liked by Richard Y Chappell
'Just imagine a linguist trying to intervene in a heated debated between utilitarians and Kantians: “Guys, stop! You don’t need to argue any more: I just did a thorough empirical survey and it turns out that the English term ‘wrong’ picks out both pushing the guy off the trolley footbridge and telling the truth to the murderer at the door. Common sense ethics wins!”
Obviously this is daft. We’re not really interested in the English word ‘wrong’.'
Intuitively there seems to be something right about what your saying here, but it's hard to square disinterest in *all* semantic claims, with interest in the substantive claims we actually care about, given that very basic disquotational principles *seem* to take us from semantic claims to ones that aren't semantic. I.e. if I learn the semantic truth 'The sentence "failing to maximize utility is wrong" is true', then I learn something that entails 'failing to maximize utility is wrong' (given the t-scheme, which doesn't seem in doubt in this context.) Since "correctly noticing the entailments of claims I already know" is a way to get knowledge, it seems I can get knowledge of normative stuff via learning semantic stuff. Of course, you can refuse to speak English, and use the various words in "failing to maximize utility is wrong" in such a way that the sentence comes out false in your idiolect. But that doesn't seem to threaten the inference from the semantic claim about ordinary English to "failing to maximize utility is wrong" read with its ordinary English meaning. And presumably the latter is a normative claim, even if the different claim you make with "failing to max utility is wrong" is also normative and false. So insofar as your write that semantic stuff can't inform you about normative truths, that's actually pretty puzzling.
If you prefer, you could re-state the central point as being that semantic interventions falsely assume that we have special reason to be more interested in ordinary English than in a subtly different philosophical idiolect.
In the various examples I give, it's assumed that English semantics has certain *structural* features that preclude the expression of a candidate philosophical view. I say that this sort of fact tells us nothing of normative interest.
> '...doesn't seem to threaten the inference from the semantic claim about ordinary English to "failing to maximize utility is wrong" read with its ordinary English meaning. And presumably the latter is a normative claim'
I don't think any genuinely normative claims are semantically transparent in the sense that one could determine their truth on purely linguistic grounds. So if one were to argue *on purely linguistic grounds* that the English sentence "failing to maximize utility is wrong" is true, then that would actually suffice to establish that the English word 'wrong' wasn't normative. Of course, we think that English 'wrong' is normative: but that's just to say that we're committed to denying that empirical linguistic data settles the truth of the claim in question.
'If you prefer, you could re-state the central point as being that semantic interventions falsely assume that we have special reason to be more interested in ordinary English than in a subtly different philosophical idiolect'
Surely we don't have to be *more* interested in the property picked out by the English "wrong" than in something else, we just have to be *interested* in it. It doesn't have to have *more* normative force than alternatives to be interesting but just *some*.
'I don't think any genuinely normative claims are semantically transparent in the sense that one could determine their truth on purely linguistic grounds. So if one were to argue *on purely linguistic grounds* that the English sentence "failing to maximize utility is wrong" is true, then that would actually suffice to establish that the English word 'wrong' wasn't normative. Of course, we think that English 'wrong' is normative: but that's just to say that we're committed to denying that empirical linguistic data settles the truth of the claim in question.'
I'm not sure exactly what it means to determine a truth on purely linguistic grounds. My point was more like "claims about truth are both semantic-truth is a semantic property-*and* substantive". But insofar as truth being a semantic property implies that semantic-y looking arguments are relevant to claims about what's true, I think its plausible the necessary extensional equivalence of "p" and " 'p' is true" is going to lead to semantic-y looking evidence bearing on "not semantic" looking claims. For example, data about what sentences people assent to in English can be evidence about what's true in English, surely! (Note, this is not saying they determine truth-values all by themselves!). The fact that people talk as if certain sentences involving "wrong" are true is exerts *some* semantic pressure on the meaning of wrong towards one where those sentences come out as true. (And on the other terms in those sentences, too plausibly.) Or at least, there will be a subset of meaning-determining contexts where this is the case. So evidence about how people use 'wrong' (possible caveat: in meaning-determining contexts) has got to be evidence (perhaps weak) about what sentences involving "wrong" are true. Which in turn bears on the normative facts via the t-scheme, unless the English wrong really isn't normative. But if "the stuff people do with the word (at least partly) determines what it applies to" is enough to ensure a word doesn't pick out a genuinely normative property, I don't see how any word could ever do that.
“If it turned out that I was a brain-in-a-vat and there was no English language—beyond the private language that I speak myself and hallucinate on the part of others—I cannot see how that would (or should) change any of my (fundamental) philosophical views in the slightest.”
Cheeky comment, but it should change your view about whether you’re a brain in a vat, which is a pretty fundamental philosophical view.
Nah, that's just a view about my contingent empirical circumstances, not really *about* a philosophical proposition as I'm thinking of it! (Philosophical questions in the vicinity could include questions like *what, if anything, I could still know, if in the circumstances of a BIV*. But the answers to those fundamental conditionals haven't changed.)
I’d say that words generally carry some of the moral intuition of people who happen to use them which is in fact a weak evidence for the claim on intrinsic value of word being associated with X and Y
There is something to be said about using lack of accessibility to increase accessibility. I think that when someone is struggling with exploring the philosophical understanding of something as more of a solo contemplative practice, they should entertain the method of going into a poetic direction with it at any cost of accessibility.. and then try to slowly take a small and manageable-enough piece of their poetic philosophical exploration and flesh it out from there; the maximally poetic is lower resolution in terms of being accessible, but sometimes that's exactly what's needed to break thinker's block.
Not much relevant to your main point, but surely you've missed the train re Geach and Thomson. It's not about the word 'good,' but the concept/thing that Moore insisted we should maximise—goodness, the good, etc.
And the claim is that goodness is a fundamentally attributive concept—and thus isn't the sort of thing it could make sense to maximise. Compare with the toy view: you ought to maximise the big / bigness. What sense could be made of that view? After all, to make, e.g., a big tennis ball is to make a small object; to make a big mouse is to make a small animal, etc. (Just as to make a good murderer is to make a bad christian, etc.) Bigness just isn't the sort of thing it's even possible to maximise. Geach and Thomson argue that the same goes for goodness.
Of course, their argument might go wrong (indeed, I've argued that it does), but to think it's just about words is confuse the subject of the argument (which does appeal to how we use words), with the subject of conclusion (not about the word 'good,' but about goodness).
Philosophers sometimes claim that they're talking about concepts when they're really just disquoting words. Using the word 'goodness' without the quotes doesn't mean that Geach and Thomson are talking about the concept that's of interest to consequentialists.
The problem is that disquoting doesn't remove the potential for ambiguity. Geach & Thomson's concept of goodness doesn't inherently matter any more than their word 'goodness' does. The question is whether SOME normative concept can play the role consequentialists need. And it's very obvious that some can. (My concept of goodness / desirability, for example.)
They obviously took themselves to be talk about the concept that was of interest to Moore. And they followed Moore's lead pretty closely (in the sense that PE made similar sorts of appeals to language).
And, plainly, they don't think it's very obvious that there's some other concept that can play the role, and they argue as such (By arguing that whatever goes for goodness similarly goes for them; just as whatever goes for bigness similarly goes for largeness or whatever). (Of course, they may well be wrong, but they knew what they were arguing.)
I mean, do you really think that two very smart philosophers really took themselves to be arguing against a normative view by just focusing on the English word 'good.' Come on, be serious.
I'm quite serious (and not interested in hosting rude interlocutors, so please mind yourself if you've any interest in continuing the conversation). Smart philosophers make terrible arguments all the time. It's hardly more charitable to attribute to them the view that there is no coherent thought that consequentialists could possibly be expressing when they claim that we should maximize the good.
Whether they realized it or not, it IS in fact very obvious that there is a coherent thought that consequentialists have in mind. So Geach, Thomson, and those who take them seriously are getting hung up on something -- whether you want to say it's a word, or an unduly limited cluster of concepts -- in a way that's detrimental to their ability to do philosophy well. I'm advocating for a more flexible mode of philosophy that avoids these pitfalls. The perspective I'm advocating for is one that makes immediately clear what a non-starter the Geach-Thomson argument is. And that's a virtue of this meta-philosophical perspective.
But it's open to one to follow Sider in thinking that there's a real, non-terminological question about the "logical structure" of reality that different mereological views disagree about.
Which of these meta-ontological views is right is, I think, a very interesting and difficult question!
They should do first-order philosophy -- including, e.g., proposing and evaluating claims and arguments about the topic in question (ethics, metaphysics, whatever) -- without any special concern for anyone's linguistic dispositions.
If you can't tell the difference between my examples of "semantic interventions" and ordinary philosophical argumentation, then I'm afraid I can't help you.
A linguist here, and a thousand times yes. Stop dragging our science in your philosophy (and anthropology, while we're on the topic), we've had enough of that.
I agree with everything but the section on non-maximal options. I think the question of "Should I block a mineshaft?" can be understood as "What is the deontic status of blocking a mineshaft?", which is an interesting question that I don't think is a merely semantic issue. I see two ways of answering the question, and both call for a serious reunderstanding of morality.
We can say non-maximal options have no deontic status, which (though my preferred view) is far from intuitive and implies that human beings never have any true thoughts about what they ought to do, or
We can say non-maximal options entailed by an optimal maximal option are obligatory, which means we have to reformulate our moral theory (standard moral theory would say you ought to do what's best, but blocking a shaft isn't best if you block the wrong one!).
> "We can say non-maximal options have no deontic status, which... implies that human beings never have any true thoughts about what they ought to do"
I'm not sure about that implication. I take it your thought is that we can never articulate an option with perfect precision. But we can refer to particular actions (e.g. "my blocking of that mineshaft", or even "what I would do were I to block a mineshaft"), and any particular action is the performance of a maximal option. So we can have true thoughts about the deontic status of the acts we do or would have done. Which all seems commonsensical enough?
Even in hypothetical cases, I think we can supervaluate over imprecision in description that makes no normative difference. Since *every* way of blocking shaft B has the same effect, we can speak with determinate truth about the deontic status of blocking shaft B, even though it is indeterminate what precise way of blocking shaft B we are talking about.
I suppose my thought was that moral deliberation in practice is rarely as simple as a philosophical thought experiment, and in most cases where we believe we ought to do something, there's actually a slightly more specific version of that option that would be better. The problem also arises if (a) our agency extends into the future and (b) maximal options cover all the options in an agent's control. It might seem that I ought to block shaft B, but what I really ought to do is block shaft B, make myself a burrito when I get home, do some journaling about that awful mineshaft situation I was just in, and go to bed early - and so if I don't do all of those, then I haven't done anything I ought to have. But perhaps there are cases where we do have thoughts of what we ought to do that do line up with our best maximal option; 'never' was a bit too strong.
I agree that limiting the information contained in a maximal act to only morally significant information makes the view more attractive, and that the view needn't be too unintuitive. Thanks for the lovely reply!
I agree with you. But would you distinguish the semantic arguments of Geach and Thomson from Foot's general line of argument in 'Utilitarianism and the Virtues'? As I understand the latter argument, it's not concerned with semantics but with relations of conceptual dependence (e.g. she says that the goodness or badness of consequences, or states of affairs, is “found. . .within morality, and forming part of it, not standing outside it as the ‘good state of affairs’ by which moral action in general is to be judged.”).
Yes, for sure, Foot's methodology seems perfectly respectable to me (though I disagree on the substance).
'Just imagine a linguist trying to intervene in a heated debated between utilitarians and Kantians: “Guys, stop! You don’t need to argue any more: I just did a thorough empirical survey and it turns out that the English term ‘wrong’ picks out both pushing the guy off the trolley footbridge and telling the truth to the murderer at the door. Common sense ethics wins!”
Obviously this is daft. We’re not really interested in the English word ‘wrong’.'
Intuitively there seems to be something right about what your saying here, but it's hard to square disinterest in *all* semantic claims, with interest in the substantive claims we actually care about, given that very basic disquotational principles *seem* to take us from semantic claims to ones that aren't semantic. I.e. if I learn the semantic truth 'The sentence "failing to maximize utility is wrong" is true', then I learn something that entails 'failing to maximize utility is wrong' (given the t-scheme, which doesn't seem in doubt in this context.) Since "correctly noticing the entailments of claims I already know" is a way to get knowledge, it seems I can get knowledge of normative stuff via learning semantic stuff. Of course, you can refuse to speak English, and use the various words in "failing to maximize utility is wrong" in such a way that the sentence comes out false in your idiolect. But that doesn't seem to threaten the inference from the semantic claim about ordinary English to "failing to maximize utility is wrong" read with its ordinary English meaning. And presumably the latter is a normative claim, even if the different claim you make with "failing to max utility is wrong" is also normative and false. So insofar as your write that semantic stuff can't inform you about normative truths, that's actually pretty puzzling.
If you prefer, you could re-state the central point as being that semantic interventions falsely assume that we have special reason to be more interested in ordinary English than in a subtly different philosophical idiolect.
In the various examples I give, it's assumed that English semantics has certain *structural* features that preclude the expression of a candidate philosophical view. I say that this sort of fact tells us nothing of normative interest.
> '...doesn't seem to threaten the inference from the semantic claim about ordinary English to "failing to maximize utility is wrong" read with its ordinary English meaning. And presumably the latter is a normative claim'
I don't think any genuinely normative claims are semantically transparent in the sense that one could determine their truth on purely linguistic grounds. So if one were to argue *on purely linguistic grounds* that the English sentence "failing to maximize utility is wrong" is true, then that would actually suffice to establish that the English word 'wrong' wasn't normative. Of course, we think that English 'wrong' is normative: but that's just to say that we're committed to denying that empirical linguistic data settles the truth of the claim in question.
'If you prefer, you could re-state the central point as being that semantic interventions falsely assume that we have special reason to be more interested in ordinary English than in a subtly different philosophical idiolect'
Surely we don't have to be *more* interested in the property picked out by the English "wrong" than in something else, we just have to be *interested* in it. It doesn't have to have *more* normative force than alternatives to be interesting but just *some*.
'I don't think any genuinely normative claims are semantically transparent in the sense that one could determine their truth on purely linguistic grounds. So if one were to argue *on purely linguistic grounds* that the English sentence "failing to maximize utility is wrong" is true, then that would actually suffice to establish that the English word 'wrong' wasn't normative. Of course, we think that English 'wrong' is normative: but that's just to say that we're committed to denying that empirical linguistic data settles the truth of the claim in question.'
I'm not sure exactly what it means to determine a truth on purely linguistic grounds. My point was more like "claims about truth are both semantic-truth is a semantic property-*and* substantive". But insofar as truth being a semantic property implies that semantic-y looking arguments are relevant to claims about what's true, I think its plausible the necessary extensional equivalence of "p" and " 'p' is true" is going to lead to semantic-y looking evidence bearing on "not semantic" looking claims. For example, data about what sentences people assent to in English can be evidence about what's true in English, surely! (Note, this is not saying they determine truth-values all by themselves!). The fact that people talk as if certain sentences involving "wrong" are true is exerts *some* semantic pressure on the meaning of wrong towards one where those sentences come out as true. (And on the other terms in those sentences, too plausibly.) Or at least, there will be a subset of meaning-determining contexts where this is the case. So evidence about how people use 'wrong' (possible caveat: in meaning-determining contexts) has got to be evidence (perhaps weak) about what sentences involving "wrong" are true. Which in turn bears on the normative facts via the t-scheme, unless the English wrong really isn't normative. But if "the stuff people do with the word (at least partly) determines what it applies to" is enough to ensure a word doesn't pick out a genuinely normative property, I don't see how any word could ever do that.
“If it turned out that I was a brain-in-a-vat and there was no English language—beyond the private language that I speak myself and hallucinate on the part of others—I cannot see how that would (or should) change any of my (fundamental) philosophical views in the slightest.”
Cheeky comment, but it should change your view about whether you’re a brain in a vat, which is a pretty fundamental philosophical view.
Nah, that's just a view about my contingent empirical circumstances, not really *about* a philosophical proposition as I'm thinking of it! (Philosophical questions in the vicinity could include questions like *what, if anything, I could still know, if in the circumstances of a BIV*. But the answers to those fundamental conditionals haven't changed.)
I’d say that words generally carry some of the moral intuition of people who happen to use them which is in fact a weak evidence for the claim on intrinsic value of word being associated with X and Y
There is something to be said about using lack of accessibility to increase accessibility. I think that when someone is struggling with exploring the philosophical understanding of something as more of a solo contemplative practice, they should entertain the method of going into a poetic direction with it at any cost of accessibility.. and then try to slowly take a small and manageable-enough piece of their poetic philosophical exploration and flesh it out from there; the maximally poetic is lower resolution in terms of being accessible, but sometimes that's exactly what's needed to break thinker's block.
Not much relevant to your main point, but surely you've missed the train re Geach and Thomson. It's not about the word 'good,' but the concept/thing that Moore insisted we should maximise—goodness, the good, etc.
And the claim is that goodness is a fundamentally attributive concept—and thus isn't the sort of thing it could make sense to maximise. Compare with the toy view: you ought to maximise the big / bigness. What sense could be made of that view? After all, to make, e.g., a big tennis ball is to make a small object; to make a big mouse is to make a small animal, etc. (Just as to make a good murderer is to make a bad christian, etc.) Bigness just isn't the sort of thing it's even possible to maximise. Geach and Thomson argue that the same goes for goodness.
Of course, their argument might go wrong (indeed, I've argued that it does), but to think it's just about words is confuse the subject of the argument (which does appeal to how we use words), with the subject of conclusion (not about the word 'good,' but about goodness).
Philosophers sometimes claim that they're talking about concepts when they're really just disquoting words. Using the word 'goodness' without the quotes doesn't mean that Geach and Thomson are talking about the concept that's of interest to consequentialists.
The problem is that disquoting doesn't remove the potential for ambiguity. Geach & Thomson's concept of goodness doesn't inherently matter any more than their word 'goodness' does. The question is whether SOME normative concept can play the role consequentialists need. And it's very obvious that some can. (My concept of goodness / desirability, for example.)
They obviously took themselves to be talk about the concept that was of interest to Moore. And they followed Moore's lead pretty closely (in the sense that PE made similar sorts of appeals to language).
And, plainly, they don't think it's very obvious that there's some other concept that can play the role, and they argue as such (By arguing that whatever goes for goodness similarly goes for them; just as whatever goes for bigness similarly goes for largeness or whatever). (Of course, they may well be wrong, but they knew what they were arguing.)
I mean, do you really think that two very smart philosophers really took themselves to be arguing against a normative view by just focusing on the English word 'good.' Come on, be serious.
I'm quite serious (and not interested in hosting rude interlocutors, so please mind yourself if you've any interest in continuing the conversation). Smart philosophers make terrible arguments all the time. It's hardly more charitable to attribute to them the view that there is no coherent thought that consequentialists could possibly be expressing when they claim that we should maximize the good.
Whether they realized it or not, it IS in fact very obvious that there is a coherent thought that consequentialists have in mind. So Geach, Thomson, and those who take them seriously are getting hung up on something -- whether you want to say it's a word, or an unduly limited cluster of concepts -- in a way that's detrimental to their ability to do philosophy well. I'm advocating for a more flexible mode of philosophy that avoids these pitfalls. The perspective I'm advocating for is one that makes immediately clear what a non-starter the Geach-Thomson argument is. And that's a virtue of this meta-philosophical perspective.
What do you think about mereology?
I'm personally skeptical, for the reasons set out here:
https://www.philosophyetc.net/2006/06/what-is-existence.html
But it's open to one to follow Sider in thinking that there's a real, non-terminological question about the "logical structure" of reality that different mereological views disagree about.
Which of these meta-ontological views is right is, I think, a very interesting and difficult question!
This seems so obvious to me that I am probably missing something.
I resonate with the spirit of this but what other evidence do you think philosophers should appeal to (and how is it non-linguistic)?
They should do first-order philosophy -- including, e.g., proposing and evaluating claims and arguments about the topic in question (ethics, metaphysics, whatever) -- without any special concern for anyone's linguistic dispositions.
If you can't tell the difference between my examples of "semantic interventions" and ordinary philosophical argumentation, then I'm afraid I can't help you.
Thank you :)
I blame Wittgenstein.