UPDATE!! SCIENTIFIC PROOF THAT A SUBSTANTIVE PROPERTY OF FINAL VALUE IS A PHILOSOPHER'S FANTASY (AND THAT EVEN PHILOSOPHERS DON'T TRULY BELIEVE IT).
(Sorry for the sensationalism, but suddenly one needs to compete for an audience around here! And I think I "buried the lead" in my original post.)…
I thought I’d take this opportunity to present one of the crazier ideas I’ve been working on. In the spirit of Philippa Foot’s “Morality as a System of Hypothetical Imperatives” (1972) and like-minded philosophers, I’ve argued (e.g. in my forthcoming book, Confusion of Tongues) that thin normative words like ‘good’ and ‘ought’ are essentially relativized to ends or goals (what I call an “end-relational” theory). So any logically complete sentence of the form ‘p is good’ is implicitly relativized to some relevant end: ‘p is good [for e]’, which I’ve interpreted as meaning roughly that p promotes/ raises the probability of e, or: e is more likely given p than given not-p.
This view encounters an obvious objection from final value: judgments about what is good “for its own sake”. What should an end-relational theory say about ‘good for its own sake’?
Whereas this locution is often taken as expressing a kind of goodness that is nonrelational, a compositional treatment of the expression itself suggests something different: a kind of goodness that is relativized reflexively. ‘For its own sake’ in general seems to mean for itself. For example, ‘He did it for his own sake’ means that he did it for himself, not that he didn’t do it for anybody. So might we suppose that to be good for its own sake is to be good for itself; i.e. p is good for its own sake iff p is good for p, or: p raises the probability that p?
This sounds crazy, of course: it’s true of any proposition p whatsoever that p is more likely if p than otherwise, but surely not everything is good for its own sake. But I suggest it becomes plausible with one further tweak. We don’t call something (instrumentally) “good”, simpliciter/sans phrase, just because it is good for something or other, but only if it is good for something that is relevantly desired or valued. I’ll capitalize (e.g. ‘p is Good’) to indicate this kind of use. Plausibly, something is said to be “good for its own sake” in case it is judged to be Good, and there is a question about the basis for its Goodness. Of things that are judged Good, some are Good because they promote something else that is relevantly desired, but others are Good because the relevantly desired thing they promote is themselves. So to be Good, for its own sake, is to be the object (or perhaps, constitutive of the object) of relevant intrinsic desire. Note that this fits with what Plato and Aristotle say: intrinsically valuable Goods are those which are desired for their own sake and not for the sake of anything else.
Most likely you’ll think this is still pretty crazy. But there’s some striking empirical evidence in its favor. An implication of this theory is that nothing can be Bad, Worse, Worst, Better, or Best for its own sake. (Because nothing can promote itself more or less than anything else promotes itself). Initially I thought this looked like a fatal problem, but then it struck me that those expressions did all seem bad to me. So I checked google usage data, and found the following:
‘It is ____ for its own sake’ ‘It is intrinsically ____’
‘good’ 4,600,000 992,000
‘bad’ 1 480,000
‘better’ 1 420,000
‘best’ 0 5
‘worse’ 0 27,000
‘worst’ 0 2
(Note: searching on the shorter string ‘bad for its own sake’ yields many more results, but I found that the majority of these seemed to be variants either of ‘desiring what is bad for its own sake’, where ‘for its own sake’ is qualifying the desiring rather than the badness, or of ‘good/bad for its own sake’, where ‘bad’ is plausibly only included for completeness/ as an afterthought.)
This is a striking pattern of use that demands an explanation. My hypothesis predicts it, and I take this as strong evidence that the hypothesis is correct. The usage for ‘intrinsically ___’ shows that it can’t be explained by a general focus on goodness over other kinds of relations. (I take it that to be “intrinsically bad” is to be Bad in virtue of intrinsic properties.)
So I’m interested in what people think. Is the evidence compelling, or are there other good explanations available for this pattern?
Thanks for a very interesting post. I don’t have a direct comment on whether this view helps to explain the data above, but was wondering if you could say something more about how you understand the notion of promoting. Let it be granted that those things that are good in themselves are objects of non-instrumental desires, things that are desired for their own sake. Is the way in which the object of desire promotes itself always the same kind of promoting? Is it always a matter of how the thing makes itself (or things of its kind) more probable? Or do some intrinsic goods promote themselves in different ways than others?
Hi Sven,
I’m using “promotes” here simply as a convenient shorthand for “raises the probability of”. I’d say that there are different ways of raising the probability of a proposition p–e.g. causally, evidentially, mereologically–and that each of these is a different way of being good for p. But in the case of p’s raising the probability of p, I’ve assumed (but am not committed to saying) that this is always the same kind of “promoting”, constituted by the relation of identity.
Thanks for your reply, Steve. That’s a very helpful clarification; I see what you have in mind.
Hi Steve,
I’m not sure I entirely grasp how the second proposal – that p is good for its own sake iff p is the object of relevant intrinsic desire – relates to your more fundamental proposal – viz., that p is good for its own sake iff p raises the probability that p. Do you think these are both analytic?
In any case, it occurs to me that one complication is that “good for its own sake” is only one expression in English that is used to pick out / express what philosophers call “final value.” (Personally, I dislike it, because it almost sounds like one is saying that something – e.g., knowledge – benefits itself, and that makes no sense.) I think “intrinsically good” is also used to express the property of final value, both by philosophers and by ordinary people. (True: some philosophers hold that if something is intrinsically good, then it is good on account of its intrinsic nature, but I’m not sure whether this provides the meaning of the phrase or whether it is a substantive proposal about which things have final value.) Additionally, “good as an end” is sometimes used to express the property of final value. So it might be that when we aim to pick out *negative* final value, we just do it using “intrinsically bad” or “bad as an end.” But if this is so, doesn’t your theory imply that these expressions mean that the thing in question lowers its own probability? And yet they do not seem unusual.
Hi Jason, thanks for the questions.
I’m not suggesting, strictly, that p is good for its own sake iff p raises the probability that p. Rather, that p is good for its own sake (as usually interpreted; i.e. Good for its own sake) iff p raises the probability of some relevantly desired end e, and e is p. I AM suggesting that the following is analytic: p is good for its own sake ONLY IF p raises the probability that p. That’s trivial, of course (the “only if” claim, not the analyticity claim). But I’d say that the substantive information communicated by these claims is carried pragmatically rather than asserted, so perhaps isn’t analytic for that reason. (Sorry, I realize this probably doesn’t clarify things much).
I agree that there are other ways of referring to final value, and that ‘intrinsic’ plausibly is often used this way, even if it isn’t always. It looks like you might be willing to grant my primary claim in this post, which is about the meaning of ‘for its own sake’. So I take you to be suggesting that this doesn’t touch the interesting and important notion of final value, which can be addressed with different expressions.
You ask “doesn’t your theory imply that these [other] expressions mean that the thing in question lowers its own probability?” I don’t think so. The claim for ‘good for its own sake’ is based on a compositional analysis, and there are no similar language-based reasons to read “intrinsically” or “as an end” in parallel ways. My broader view makes room for final/intrinsic badness: relative to a (relevantly) intrinsically desired end e, p is (intrinsically) bad in relation to e if p is not-e. Here the relation isn’t identity, but contradiction. This isn’t being “bad for its own sake”, but rather being bad for the sake of its negation. On the assumption that to desire that e is to be averse to not-e, what is finally Bad is what is an object of intrinsic aversion.
So the theory predicts that ‘intrinsically’ would be equally fine for ‘good’ or for ‘bad’, whereas ‘for its own sake’ will be unacceptable except for ‘good’–which is what the usage data seems to confirm.
As I should perhaps have stressed in my original post, I think these findings are of great significance because they imply that ‘good for its own sake’ doesn’t refer to a substantive property of intrinsic goodness that is prior to desire, and therefore suggest that ordinary evaluative thought doesn’t invoke any such property.
How does “bad talk” work on this view? It seems like the obvious answer is that to call something bad is to say it promotes something that is relevantly disvalued or “aversed.” (I’m not sure that that’s a word, but I’m using it so that ‘aversed’ is to ‘aversion’ as ‘desired’ is to ‘desire’.) In that case, if something is itself relevantly aversed, then isn’t it Bad for its own sake in virtue of promoting itself?
Hi Dale,
The idea is to take ‘bad/good for e’ as basic, where e is some end or outcome. To say that p is bad for e certainly isn’t to say that p promotes e, but rather that p demotes e, or decreases the probability of e.
So to call something Bad is on this view to say that it decreases the probability of some relevantly desired end. Hence the prediction that nothing can be bad (or Bad) for its own sake. But one way of being bad for an end e is to be not-e. Where e is an object of intrinsic desire, this is to be an object of intrinsic aversion.
Hi Steve
Sorry about being late in the party. Things are hectic here. I’m not sure I have much to say about the positive proposal or the data. I’m slightly sceptical about using google for this purpose (and for example ‘bad in itself’ gives over million results).
But, I just wanted to test a proposal of how to deal with the problem which is somewhat close to your own. It accepts the basic idea that all goodness claims are end-relative in the context. The idea then is that intrinsically good things are at least in part constitutive of the end relevant in the context whereas instrumentally good things are not – they are mere means to the end. Of course, both the means and what is constitutive of the end raise the probability of the end being realised – the constituents usually even more so. One advantage is that this proposal needn’t tie intrinsically good things to what we desire as the ends relevant in the context needn’t be the ones we desire.
Hi Jussi, thanks!
I agree with your proposal completely. (This was what I had in mind by my qualification, “or perhaps, constitutive of the object” in my post above.) This seems potentially more natural for ‘intrinsically good’ than for ‘good for its own sake’, by my proposed analysis of the latter (which is why I said “perhaps”). I also agree that this has the advantage you note.
The common use of ‘bad in itself’ seems entirely consistent with my hypothesis. There’s no reason to take ‘in itself’ as a reflexive relativizer, and it rather seems equivalent to ‘intrinsically’. As explained above, I recognize the existence of intrinsic badness.
I’m curious about why you’re skeptical about this use of google data. I assume you won’t disagree that the data clearly shows that people just don’t use ‘for its own sake’ with ‘bad’, ‘better’, ‘best’, etc. Do you doubt this datum has philosophical significance? (I’m going to assume you do, just because I want an excuse to make the following point).
This pattern (millions of ‘good for its own sake’, half a million ‘intrinsically bad’, basically no ‘bad for its own sake’) is so stark that it demands an explanation. Mark Schroeder suggests to me (as devil’s advocate) that the explanation might simply be that ‘good for its own sake’ is an idiom–in which case the datum wouldn’t be of much philosophical interest. But I’m skeptical about this explanation. I can see how ‘for its own sake’ could be an idiom, but not how ‘good for its own sake’ could only appear as an idiom, given that anything that is good for its own sake is also Good. Surely that would encourage detaching ‘for its own sake’ as an idiom in its own right, and then there shouldn’t be any barrier to attaching it also to ‘bad’, ‘better’, etc. This seems to point toward a compositional explanation for our aversion to these constructions, like the one my hypothesis provides.
Hi Steve
thanks for this interesting comment. I need to think about this more but just few quick comments.
Let me just introduce an alternative explanation. It’s one of Gricean maxims that your contributions to discourse should be brief and give just the required amount of information. In the case of good, there clearly are both instrumentally good things and things that are good for their own sake. So, when we talk about goodness, we must specify often which one of these values we are talking about.
Now, imagine that there only were things that were bad for their own sake. In this case, we could talk about just what’s bad. Adding that something was bad for its own sake would be redundant and thus violate Gricean maxims. I might think that things are actually like this. Things are just bad and so it’s redundant to say that they are bad for their own sake. These things are constitutive of the states of affairs we prefer not to obtain.
You might ask why there aren’t any instrumentally bad things. Well, it seems like if something lowers the probability of an end it just fails as a means – it’s not a mean at all. Because of this, this thing is not instrumentally bad (it’s not an instrument) but rather it just fails to have instrumental value.
About the googledata, I don’t quite know why I am suspicious. I just wonder whether what’s on the internet is reflective of what we would say and what is natural to say. Much of the hits might for example be philosophical works.
Thanks for the counter-proposal, Jussi. I’m not sure how seriously you’re putting it forward, but to treat it seriously: I agree that if all badness was final, then we’d find very few instances of ‘bad for its own sake’, for the Gricean reason you note. However:
(1) It seems independently implausible to me that all badness is final. E.g. having no money is bad, but not intrinsically.
(2) This hypothesis would explain why we don’t find many occurrences, but not why ‘bad for its own sake’ would strike us as improper.
(3) It doesn’t explain why we don’t find ‘better/ best for its own sake’.
(4) It would seem to extend also to use of ‘intrinsically bad’, but note that we have half a million hits of that.
(5) Finally, I don’t think it can explain the (almost) complete absence of ‘bad for its own sake’. By contrast, compare these google numbers for badly ungrammatical strings:
‘it is gooder’: 492,000
‘it is more gooder’: 264,000
‘it is badder’: 1,100,000
‘it is worster’: 73,000
‘it is worstest’: 8
‘it is worstester’: 4,600!
I think this clearly shows, BTW, that google data is picking up ordinary speech just fine. I did sample some of the pages of hits for ‘…for its own sake’, and I found that the majority of ‘good/bad for its own sake’ occurrences on those pages were philosophical, but not the majority of ‘it is good for its own sake’, which is a fairly widely used expression (3.6 million!)
The most plausible explanation of the complete absence of ‘it is bad/better/best/worse/worst for its own sake’, it seems, is that these are deeply incoherent in some way, and not just bad grammar or pragmatically infelicitous.
Hi Steve,
thanks. You are right, I haven’t yet decided how serious the proposal is. I’ve got some things to say to your objections to 1-5, but it’s better to leave it at this. I did want to ask about 3 as this is something I find the most curious. There are a couple of things to say about this. Mainly, I wanted to know how your proposal deals with this issue.
As I understand it, on your view, things that are good for their own sake are objects of the relevant intrinsic desires. I know that you say that nothing can promote itself more than anything else, but desires have strengths and a preference-ordering structure, and so you would assume that this would give us good for its own sake in degrees in a way that can be compared. Even if p and q equally promote themselves as objects of desires, I can desire p more than q ie. prefer p over q.
So, why don’t we say that x is better for its own sake than y to express that some things have more final value than others? Nothing in the structure of the theory of value seems to rule this out. So, I would just like to know why you think that the comparatives with ‘for its own sake’ are deeply incoherent?
I agree that things can be Better or Worse as ends/ in themselves. As you suggest, I account for this by appeal to degree of desire/preference. How the semantics accounts for this is a bit tricky; I tackle the problem in Chapter 6 of my book. Roughly the idea is that making an explicit comparison of value makes salient a preferred end, and whatever is better for that end is judged “Better”, simpliciter.
This means, in fact, that A is (intrinsically) “Better” than B just in case A is more good for A than B is good for A (where A is the saliently preferred end), making your question pressing. But I think the natural reading of “A is better for its own sake than B is” is rather that A is more good for A than B is good for B. That’s an impossibility, on my analysis.