The summer fun continues. Here’s a cool question raised by McT, quoted below:
"Both good and evil are quantitative. …. Good values then form a series, as do evil values. And these two constitute together the single series of values, of which the generating relation is "is better than" or "is worse than". Of any two values, good or evil, one will be better than the other, which will be worse than the first. This raises the question whether, after all, there are in reality two sorts of values, good and evil, or whether there is no such distinction, and no such positive qualities as good and evil but rather only relations of better and worse…. In a series of magnitudes each is larger or smaller than each of the others. But no magnitude is positively large or small. Is the series of values like this?" [Nature of Existence, pp. 409-410.]
This strikes me as a pretty cool question. Grant that there are objective facts concerning whether one thing is intrinsically better than another. Why go further and say that some things are good or bad?
Some side comments: First, McT probably should take the generating relation to be "is intrinsically equal to or better than" instead of "is intrinsically better than". Noting this doesn’t make the question he raised less interesting.
Second, I don’t want to give the impression that I think we have no reason to believe that things are good or bad. I’m curious what people take to be good reasons to hold that things have goodness or badness. McTaggart is also no skeptic, but I’ll hold off on his response until later.
Third, it’s helpful to note that someone who denies that, e.g., pain is bad or pleasure is good, can still say that every pleasure is intrinsically better than any pain, and any pain is intrinsically worse than a state of no sensation. And classical consequentialists, who hold that our sole duty is to maximize intrinsic value, can uphold their view while denying that anything is positively good or bad.
Cool.
I think the obvious reply is that some values will have both positive and negative valences, similar to the way that some experiences can include both pleasure and pain. An example might be the appreciation of a beautiful painting that celebrates sadism.
Of course, this raises some thorny issues about the individuation of values. For any proposed dual-valence value, what’s to stop us from saying that we simply have two separate values–e.g., the good one of appreciating beauty and the bad one of appreciating sadism?
John Broome discusses this (and comes down on one side, obviously) in “Goodness is reducible to betterness: the evil of death is the value of life”.
It can be kind of unclear exactly what the question is, I think. Broome’s most helpful way of putting it, in my opinion, is this. If you knew the complete ordering of, well, whatever the bearers of value are, so that you knew which things were better than which, could you still be missing some information about which things are in fact good and which are bad? He thinks not.
Hi Jamie,
Thanks for the reference. Broome’s way of putting the question is helpful.
Hi Justin,
That’s an interesting suggestion. But it seems to me that there is a good response to consider. It might be that there are “organic unities”, which we might have taken to be states that are say intrinsically good but with intrinsically bad parts. Suppose, for example, that being displeased that innocents suffer is intrinsically good. Call an instance of this state, D. You might think that a “component” of this state is a displeasure, which is itself intrinsically bad. Call this component C.
So it looks like D is igood whereas C is ibad. But perhaps in reality all that is going on is that D appears very late in the intrinsically better than series, whereas C appears much much earlier. Another way of putting that point is that the class of things that C is intrinsically better than is a proper subset of those things that D is intrinsically better than (and perhaps is a substantially smaller set). So I think we could still make sense of organic unities even on this picture.
Thanks, Kris. Yeah, that surely is a good response. I should have stipulated that I was setting aside the problem of organic unities.
This stipulation doesn’t seem completely unfair if we assume that–in at least some cases–we get the ordinal ranking of two or more values simply by comparing the sums of their relevantly different parts. Take out the artistic celebration of sadism and replace it with two other intrinsically bad things (I don’t know, admiration of Bush and love of Rumsfeld). To say whether the new whole (with its two intrinsically bad parts) is more or less valuable than the old whole (with its one really bad part), we would have to assign absolute magnitude to the three parts, or at least do more than say how each part ranks vis-à-vis the others.
OK, here’s one reason you might want to have the additional facts about goodness and badness.
A likes everything, including pleasure and torture. But he likes pleasure more than torture. The more pleasant a pleasure is, the more he likes it. He likes even a small pleasure more than any torture. He likes slight pains more than intense tortures.
B likes pleasure and hates pain. She likes intense pleasures more than mild ones, and hates intense pains more than mild ones.
Both A and B have attitudes that are proportionate to their objects. But A is crazy and B isn’t. One explanation for why A is crazy is that A likes the bad. That explanation isn’t available if we think all the value facts are exhausted by ‘better than’.
What do you think of that?
Hi Ben,
That’s interesting. I’m definitely attracted to the view that you shouldn’t like what’s bad!
But maybe there is a response available. Let me first introduce the concept of a negative state. Suppose that S is a state in which x has feature F. Then ~S is a negative state in which x has feature ~F.
Example: Let s1 be a state in which John feels pain; ~s1 is a state in which John does not feel pain.
Suppose that, for some given F, for all S, S is intrinsically worse than ~S. In a case like this case, let’s call F “schlecht”. Suppose that, for some given F, for all S, S is intrinsically better than ~S. In a case like this, call F “gut”.
Could we say that what’s messed up A’s attitudes is that A likes it when things have schlecht properties, and what is right about B’s attitudes is that she dislikes it when things have schlecht properties? Is this a worse explanation?
(Reflecting on this case suggests to me that one might simply try to define “is intrinsically good” as “is a state in which a gut property is exemplified”.)
Hi Kris,
Yeah, that would work, I think. But your original question was “why go further and say that some things are good or bad?” And I take it you agree that there is a reason to say the additional things about goodness and badness. (Even if ‘good’ and ‘bad’ can be defined in terms of ‘better.’)
I’m curious about what he says about magnitude in the end. I changed the quotation by replacing size for value and got:
*Both large and small are quantitative. …. Larger objects then form a series, as do small objects. And these two constitute together the single series of sizes, of which the generating relation is “is larger than” or “is smaller than”. Of any two objects, large or small, one will be larger than the other, which will be smaller than the first. This raises the question whether, after all, there are in reality two sorts of sizes, large and small, or whether there is no such distinction, and no such positive qualities as large and small but rather only relations of larger and smaller…. In a series of magnitudes each is larger or smaller than each of the others. But no magnitude is positively large or small. Is the series of values like this?”
I’m not sure what I would say about this. It’s true that size is relational in some way. As positivists tried to argue, it might not make sense to say that everything doubled in size last night. Things can double in size only if some other things to which they can be compared remain the same size. On the other hand, I’m not sure I want to accept that it follows from this that there are not two sorts of sizes – large and small. Would the value case be different?
I also have couple of naive worries. What is it that makes one object better than another? I’m tempted to say that it is its value – how good it is. But, if there are not positive qualities like goodness, how could this be?
Second, if there are objective facts about intrinsically better and things that are neutral and there are facts about how much better some objects are than others and the neutral things, don’t we get amounts of goodness and badness for objects as a result?
I’m sure that Broome is right about this though. He usually is.
Kris,
I’m intrigued by your response to Ben, but I wonder if your proposal might be better expressed without the notion of a negative state. Something like the following might work.
Say that a property F is a complement of a property G iff, necessarily, a thing exemplifies F iff it doesn’t exemplify G. For simplicity, assume every F has a unique complement ~F. (You might want to call ~F a “negative” property. But I think that would be misleading. The complement relation is symmetric; F is the complement of ~F just as much as ~F is the complement of F. So it seems arbitrary to call one positive and the other negative.)
Now say that a property F is gut iff for all states S and S’, if F is exemplified in S and ~F is exemplified in S’, then S is better than S’. And say that a state S is good iff a gut property is exemplified in S.
Is that what you had in mind?
Hi Jussi,
I tend not to take positivististic worries too seriously. But waiving them aside, doesn’t it seems like claims about an object being large or small are always relative to at least a comparison class, as in large for a human baby, small for an elephant baby? And this suggests that claims about an object being large or small really are to be understood in terms of “larger than”. McTaggart thinks that we ought to say something similar about claims about goodness or badness if fundamentally there is just the betterness than relation. He writes:
“We certainly make judgment that the value in a self or aggregate of selves is positively good or evil. Now if there is in reality no positive good or positive evil, then we must treat such judgments not as being objectively true, but as being true only with reference to some arbitrary standpoint. If I say that the value in A is a good value, this will have to mean that it is better than the average value of persons with whom I am acquainted … or of something of a similar nature.”
(This isn’t an idea McT develops, because he does think that there are “positive qualities” of good and bad.)
Granted, if there are objective facts about which things are intrinsically neutral, then it is easy to see that there will also be facts about which things are good. But the position under consideration would deny that there are objective facts about which things are intrinsically neutral. There is a series of things ordered by the “is better than” but there is no distinguished neutral point in the series.
Here’s another analogy. Imagine an infinite vertical line, the points of which are ordered by the “is higher up than” relation. This is a series in which there is no distinguished center point — any point you pick as the zero point would be arbitrary. The value series, on the hypothesis we are entertaining, is like this.
Hi Campbell,
That’s the sort of thing I had in mind. Is this sort of analysis of intrinsically good and bad in terms of intrinsically better than plausible?
I haven’t read the Broome paper yet, so I’m not sure what analysis he proposes.
So the Broome/Dreier Question goes:
“If you knew the complete ordering of, well, whatever the bearers of value are, so that you knew which things were better than which, could you still be missing some information about which things are in fact good and which are bad?”
I agree it’s a nice question. Now here is the Bradley response:
“A likes everything, including pleasure and torture. But he likes pleasure more than torture. The more pleasant a pleasure is, the more he likes it. He likes even a small pleasure more than any torture. He likes slight pains more than intense tortures.
B likes pleasure and hates pain. She likes intense pleasures more than mild ones, and hates intense pains more than mild ones.
Both A and B have attitudes that are proportionate to their objects. But A is crazy and B isn’t. One explanation for why A is crazy is that A likes the bad. That explanation isn’t available if we think all the value facts are exhausted by ‘better than’.”
Which is a nice response. But is there not a problem thereby invited in that the same issue as arises with good/bad and better also seems to arise with likes/dislikes and prefers.
“If you knew Lenman’s complete preference ordering for, well, whatever the objects of preference are, so that you knew which things I prefer to which, could you still be missing some information about which things I like and which I dislike?”
Certainly some familiar standard economist’s and decision theorist’s ways seem to take it that all “utility information” is representatable as information about preference orderings.
When it’s a question of pain and pleasure it’s natural to think there’s an obvious baseline of “hedonic neutrality” above which we have positive pleasure, below which positive pain (that not being the same as just less and less pleasure). But it’s hard to generalize that thought beyond narrowly hedonic contexts.
It seems to me that there are a few theses here that should be kept distinct, because they can’t all be true.
One is the thesis that goodness and badness are reducible to betterness, in something like the way Kris and Campbell have suggested. Call this the reduction thesis.
Another is the view that there is no such thing as goodness or badness. (This seems to be what McT actually suggests in the quoted passage.) Call this the no-good thesis.
Another is the thesis that there is no non-arbitrary zero point or “neutral” value. Call this the no-zero thesis. Kris alluded to this thesis in his last comment.
The no-zero thesis goes with the no-good thesis, but neither is compatible with the reduction thesis. If the reduction thesis is true, there is such a thing as goodness – for something to be good is for it to be better than its negation, or complement, or whatever. And then obviously we can get neutral value out of goodness and badness.
If the example in my first comment was convincing then we should prefer the reduction thesis to the others, I think.
Jimmy, yeah, I was anticipating a similar sort of response. You’d just have to deny that there is any non-arbitrary distinction between pro-and con-attitudes in general. For some purposes (maybe economics) this might be harmless. But suppose we want to know how good a person someone is, for example. It seems like we need to know more than their preference orderings. We need to know whether they like the bad, or hate the good. Otherwise A and B in my example come out as equally good people.
Also, doesn’t it just seem like there *is* a non-arbitrary difference between pro- and con-attitudes? Some things we are attracted to, and some we are repulsed by.
So can we suppose A has the very same preference ranking as B but, as it were, zero, is in a different place? (Or in effect as with the Kelvin scale for temperature, zero for A is at the bottom.) But because B has the same preference ranking, B is willing to pay such-and-such in order to avoid torture (he hates giving up $10,000 but not as much as he hates a afternoon of torture) and so is A (A LIKES giving up $10,000 and likes being tortured but likes being tortured a bit less.) And on general if the preference ranking is the same, A is disposed to make all the same choices B makes. Saintly A will move heaven and earth to save me from torture and so will B even though B likes (let’s suppose) the idea of me being tortured but not quite as much as he likes the idea of moving heaven and earth in order to prevent this.
So I’m not sure goodness is the problem. But the “in order to”s are certainly odd. I would like you to give me $100. But not as much as I would like a date with Reese Witherspoon. So it makes sense to say I would prefer the date with RW to the $100 if I had to choose one or the other. But it would hardly say, given the positive valence of both, that I was taking the date with RW IN ORDER TO avoid having to accept the $100. And it seems similarly odd to say of A that, if he likes both, he is giving up the $10,000 dollars IN ORDER TO avoid torture.
But that too looks capturable by preference ranking info. If you offer me a three way choice, $100 or date with RW or BOTH, I’ll take the both please. That’s why it would be odd to say in the two way choice I choose the date with RW in order to avoid the cash. And if A is to have the very same preference ranking as B, A must not only prefer giving up $10,000 to avoid the torture but have the full preference ranking:
Not give up the cash, not get tortured.
Give up the cash
Get tortured.
Give up the cash and get tortured.
Once we add this sort of structure it can look as if it would just be incoherent to say of him that he likes being tortured. (Assuming, like B, he prefers more money to less.) But that would be to move towards the kind of reductive view Kris and Campbell suggest.
I’m tempted towards a weak form of no-zero thesis. Compare temperature again. There’s no deep reason to put zero where the centigrade scale does instead of where the fahrenheit scale does. Neither scale is “right”. But the freezing point of water is quite an interesting point for various purposes so it makes sense. Other values for zero may make sense for other purposes. So with lining and disliking, there’s the cost-benefit analyst’s zero which is whatever lies on an indifference curve alongside no change in my monetary endowment (I pay nothing and receive nothing). Then there is the hedonistic neutrality zero. These are quite interesting places to stick the zero point for various theoretical purposes we might have and there may be others. But as with centigrade and Fahrenheit there’s no very deep philosophical question where zero should really be…
Kris,
But here the analogy breaks down. This is from the second quote you give:
“Now if there is in reality no positive good or positive evil, then we must treat such judgments not as being objectively true, but as being true only with reference to some arbitrary standpoint.”
Even if size was always relative to a comparison class (which I’m still doubting), it does not follow that size judgments are not objectively true or that they are true only with a reference to some arbitrary standpoint. Even from the view from nowhere some babies are wee in comparison to ther human babies and some elephant-babies are chubby in comparison to others. And, objectively, truly so. So, the analogy he suggests first seems to work just against his claim.
I think Jimmy’s last comment is right. The label ‘non-arbitrary’ is not exactly the right one to choose here, since it some specific context there may be a very good reason to choose one zero-point rather than another (just as in choice of temperature scale). Similarly with good: it’s convenient in many contexts to fix a zero by choosing a state that is no better or worse than its complement. Still, someone who wasn’t aware that we had fixed that point, but was aware that the state in question was no better or worse than its complement, would not be missing any facts.
Try this. Suppose someone, “our friend”, knew exactly which sticks were longer than which other sticks. You and I have additional information: we know that there is a zero-length, and we know what it is. There is something we know, a fact, that she doesn’t. It reveals itself (in case one is skeptical that that knowledge itself amounts to anything) in our further knowledge of which sticks are twice as long as which others. This is plainly something that our friend can’t figure out from what she knows already.
Now, Ben may think that there really is absolute intrinsic goodness. Maybe he thinks that some things are indeed twice as good as others, and that the Broomean who discovers all the facts about which things are better than which will still be missing the ratio facts. I doubt this — I do not believe that there are any such absolute ratio facts of value — but at least this fixes something for me and Ben to disagree about. (I don’t know that Ben does hold that view — I’m inferring that he does from past postings and discussions, which I remember very imperfectly.)
Right Jamie, I do think all that stuff. So we have something to disagree about, which is nice indeed. But it’s not just ratios that the Broomean misses out on. It’s also intervals or differences. For example, it seems like there’s a much *smaller difference* in value between a pleasant 5-minute massage and a pleasant 6-minute massage than between a pleasant 5-minute massage and a 10-year torturous pain. You can believe in those different-sized intervals without believing in an absolute zero point. But you don’t get intervals from just an ordering.
It’s also possible to believe in an absolute zero point but deny that there are different-sized intervals. (Joe Mendola does that, I think.) So all of these questions come apart really, except that of course you do need an absolute zero to make sense of ratios. I have more to say about this, but it’ll have to wait until later.
But you don’t get intervals from just an ordering.
There is a well-known method for deriving a cardinal scale from an ordering. The objects of the ordering need to have a certain structure; they need to be “lotteries”. And the ordering needs to satisfy the so-called axioms of expected utility theory, e.g. transitivity, continuity, etc. Though I understand it is a matter of controversy just what this formal result shows.
Jamie,
Which ratio facts do you not believe in? It’s important to distinguish two sorts. There are ratios of levels, e.g. X is twice as good as Y. And there are ratios of differences, e.g. X is better than Y to twice the extent that Y is better than Z.
It might be worth noting that if there are no facts of the latter kind, then utilitarianism is down the toilet, and that if there are no facts of the former kind, then utilitarianism is down the toilet unless population size is fixed.
The ratios I’m skeptical about are the ratios of levels. Those are the ones that require a zero on the scale.
I don’t mind ratios of differences, though I think there are different ways to ‘cardinalize’ good (one is with lotteries). So in a sense I guess I’m also somewhat skeptical about ratios of differences: it’s not that there is no way to make sense of the idea, but that there may be many incompatible ones. So that’s a lot like the worry that fixing a zero is ‘arbitrary’.
The ratios I’m skeptical about are the ratios of levels. Those are the ones that require a zero on the scale.
Well, what do you think of the Kris’s proposal for fixing a zero (i.e. the thing about negative states, which I recast in terms of complementary properties)?
Incidentally, doesn’t Jeffrey say something quite similar in The Logic of Decision?
Jamie, one more question:
Suppose there are two possible worlds X and Y such that the population of X is greater than that of Y, and everybody in Y has greater wellbeing than everyone in X. Do you say there’s no fact of the matter as to which world has the greater total wellbeing?
Campbell,
I said here what I think about that, in the first paragraph. I think it’s fine, but I don’t think that zero is picked out by constraints already imposed by our concept, or anything like that.
Yes, similar. I think he tends to use the necessary proposition as the zero, which for him is like saying that zero is the status quo.
Yes. Don’t you? That one seems easy.
Yes. Don’t you? That one seems easy.
I think there might be a fact of the matter. Of course, I don’t mean that the information I’ve given already is sufficient to settle what the fact is. But I think, in principle, there might be information that would settle it. Are you saying no such information is possible?
Campbell, good point about deriving cardinality from an ordering. (I guess you are referring to von Neumann and Morganstern?) So, I guess I should take back that claim.
Let me see if I can fumble around for a reason not to take betterness as primitive. I think betterness is an internal relation. Betterness holds between two things X and Y just in virtue of the intrinsic natures of X and Y. If so, then it’s not appropriate to take it as primitive. The relata and their natures – their values – determine that the betterness relation holds between them. (I think maybe Jussi was gesturing at something along these lines in an earlier comment.) For the same reason, you wouldn’t want to take hotterness as primitive with respect to temperature, because hotterness is an internal relation.
Of course, some of you may want to deny that betterness is an internal relation. I don’t know what McTaggart would say, which is of course what really matters here.
Right, my position is that the answer has to be relative to the fixing of a zero (which isn’t done by the facts). I did think you meant that you had given enough information to answer — if not, then why did you include the info you did include? I’m confused about the example.
Anyway, why don’t you give information that settles what the fact of the matter about the sum of welfare is, and either you’ll have convinced me or I can tell you why I think it doesn’t settle it.
Oops, Ben’s comment intervened. I mean that last one to address Campbell.
Jamie,
I wanted to give a case in which there couldn’t be a fact about which had the greater sum of wellbeing unless there were facts about ratios of wellbeing levels. There are cases in which that isn’t so. For example, if the population size is held fixed across X and Y, then ratios of wellbeing differences are sufficient to settle which has the greater sum.
I guess my point is just that if there’s no ratio scale for wellbeing, then variable population ethics, at least of a broadly utilitarian sort, is hopeless. For example, Parfit’s “repugnant conclusion” doesn’t make any sense. Perhaps that doesn’t bother you.
I had in mind information like this: the ratio of the average wellbeing in Y to that in X is greater than the ratio of population size in X to that in Y. But I suppose you think that’s either meaningless or false. Which is it, by the way?
McTaggart’s question is indeed very cool, as demonstrated by the number of complicated issues it has raised just in this thread alone. I want to add another complication to the pot, by way of response to an argument of Ben’s:
“Let me see if I can fumble around for a reason not to take betterness as primitive. I think betterness is an internal relation. Betterness holds between two things X and Y just in virtue of the intrinsic natures of X and Y. If so, then it’s not appropriate to take it as primitive. The relata and their natures – their values – determine that the betterness relation holds between them. (I think maybe Jussi was gesturing at something along these lines in an earlier comment.) For the same reason, you wouldn’t want to take hotterness as primitive with respect to temperature, because hotterness is an internal relation.”
Here’s another position worth considering, which definitely addresses Ben’s worries without conceding that there are “objective facts” about goodness and badness. On this view, there is a property of being intrinsically valuable, and it comes in quantities, and one thing is better than another thing in virtue of the first having more of that quantity than the second. But although there is this property, there aren’t the additional properties of being good and being bad.
I think a view like this will be “axiologically equivalent” to the view that McTaggart is considering, and Jamie is advocating. Yet this view doesn’t take “is intrinsically better than” to be fundamental, nor does it deny that it is internal, supervening on the value properties had by its relata.
Hm, now why do you say that the Repugnant Conclusion doesn’t even make sense without absolute ratios? Is it that “barely” in “barely worth living” sounds like it must be something like a small fraction? I suppose that might be right, but it seems to me we could make do with an intuitive, unquantized notion there.
As to ‘meaningless or false’: I don’t know which to call it, but you know the view, right? How about ‘incomplete’? It awaits the specification of a zero. Then it will have truth conditions (or be a truth condition).
Darn, again someone intervened — that was again addressed to Campbell.
Kris, quick question about the view you describe. What are the possible quantities of value? Can something have zero or negative value?
Ben:
In order to make sense of negative or positive value, I think I’d have to posit a distinguished 0 point. On this view, we don’t do that. Things have value, to lesser or greater degrees.
Kris,
OK, but some things can lack value altogether, right? Would those things have zero value?
Jamie,
Isn’t it assumed in just about every discussion of the Repugnant Conclusion that the total wellbeing in the Z-world is greater than that in the A-world? According to you, that assumption is faulty. You say it’s indeterminate which world has a greater total. Granted, that’s not to say that the RC itself makes no sense; but it does suggest that most discussions of it are at least a bit confused.
Ben,
Consider ‘lateness’. One thing is later than another in virtue of their temporal location (which isn’t intrinsic, so that’s a failure of the analogy). Similarly, things are better or worse in virtue of their location in value space, which is in turn fixed by their intrinsic properties. But without a coordinate system, there is no way to say which things have negative value, just as there is no way to say which things have negative lateness, though once we fix a zero point in time, we’ll see that in that system something are ‘not late at all’.
As I said, the analogy isn’t perfect because lateness doesn’t seem to be intrinsic to events, but I like it because ‘later than’ is a comparative but in some good sense prior to ‘late’.
Campbell [10:51], I don’t see that that assumption is needed. Is it? Maybe I’m wrong.
A+ is at least as good as A; B is better than A+; and so on. So, Z is better than A.
Does one of the steps require that total well-being in Z is greater?
Broome doesn’t assume that in Weighing Lives, does he? That would require him to contradict “Goodness is reducible to betterness”.
Jamie,
Right, I can see why you’d like that analogy. Of course, I don’t like it since I think betterness is an internal relation. The view Kris was suggesting in his last comment was supposed to be compatible with betterness being internal, and I was trying to figure out how that view is supposed to go.
Ben [11:06], I was hoping the internalness didn’t matter. (I do think ‘better’ is internal.)
Okay, then consider ‘redder than’. This is defined over pairs of colored objects, and the ordering is given by the spectrum (so blue is redder than violet, contrary to the phenomenology). This is an intrinsic relation, but the definite facts of comparison are prior to the property of being red (maybe reddish is better?).
I like this one better. Uh, more.
Jamie,
I’m not sure. It might depend on what you’re trying to do. If you want to show that certain views imply the RC, then the assumption might be needed. Also, as I recall, Parfit’s argument for the claim that A+ is at least as good as A may involve comparisons of total wellbeing between worlds with different population sizes.
In any case, the assumption is often made. For example, from the SEP entry on the RC:
You say this:
I don’t see how that follows. Couldn’t there be facts about ratios that are reducible to betterness?
OK Jamie, that’s helpful, thanks. I’ll have to think about that example. I’m wondering, though, if you would agree with Kris that the view he describes is equivalent to yours. On the view he describes, value comes in quantities. This suggests (doesn’t it?) that something can be twice as valuable as something else. And you want to deny that, I thought.
Campbell,
I agree that showing that certain views imply the repugnant conclusion might demand assuming that one total welfare was larger than another, and plainly the SEP entry does take that comparison to be relevant. I was more worried when you said that the repugnant conclusion makes no sense without absolute ratios. I’m pretty sure it does.
Could there be facts about ratios that are reducible to betterness? Hm, I thought we had more or less agreed that there couldn’t be. Take all the betterness facts. They constitute an ordering, which in turn represents them fully. Now put them on a scale. The ordering and the expectation property are preserved under affine transformation to a new scale. But ratios are not. (Ratios of differences are, of course.) So the ratios are not fixed by the betterness facts.
Ben,
Well, no, I don’t think what Kris said really implies that. I do pretty much agree with what Kris said, though I might be interpreting details differently from the way he intended.
That a property comes in quantities doesn’t mean there is a definite sense to one thing having it twice as much as another. Temperature is a good example – pretend we don’t know that there is a real ‘absolute zero’ (in fact, a quantum physics grad student once told me that systems can in principle have negative Kelvin temperatures, though I don’t vouch for that). Hotness comes in quantities, and hotter than is intrinsic and supervenes on those quantities. But tagging real numbers to the properties requires us to fix a zero, which is arbitrary. So although today is twice as warm by Fahrenheit numbers as April 6th was, say, it is six times as warm by Celsius numbers.
‘Comes in quantities’ leaves some room for precisification, I guess. The properties have some of the structure of the real numbers, maybe that’s a more precise way to put it.
Jamie,
I’m not sure the temperature case really supports your view. The temperature of a substance is the mean kinetic energy of its molecules. Only the Kelvin scale accurately represents ratios of MKE. F and C scales are more convenient to use, but don’t accurately represent the ratios. So in an important sense K, F and C are not on a par. There’s an objective, absolute fact of the matter about the MKE of some substance, and whether it’s exactly twice that of some other substance. So I think pretending we don’t know what absolute zero is means pretending we don’t know what temperature is, or pretending we don’t know what it is to have zero energy.
Is there another example of something that comes in quantities (in some recognizable sense) where there are no ratios? Your color case doesn’t really seem to work either, since the relevant quantity will be wavelengths and those stand in ratios. I’m not smart enough to think of one.
Huh. Well, we didn’t know a priori that temperature had an absolute zero (and as I mentioned, I’m not even sure it’s true according to quantum physics). This means, I think, that we can’t know a priori that quantifiable properties stand in absolute ratios. If it’s true, it’s a posteriori. But it’s pretty dubious that we do know this a posteriori. So I thought that was a pretty good example.
I didn’t say in the color example that redness was wavelength, though, so that one is still in good shape. From my specification, redness could perfectly well be the logarithm of the wavelength (as loudness, in Bels or dB, is the log of sound pressure).
That example raises doubts about ratios of intervals. I can’t think of a good example with all these features:
1. The ordering relation is an intrinsic relation.
2. The property is a familiar, real-world one that we talk about a lot.
3. The choice of zero, but not the rest of the cardinalization, is pretty clearly arbitrary, in the relevant sense.
It’s kind of surprising that none comes to mind. I can’t see why that should be. It’s not hard to think of examples that fit any two of the criteria. Maybe one will just come to me, or, more likely, a reader will come to the rescue.
Jamie,
What are we counting as betterness facts? Is the fact that a thing is better than its complement a betterness fact? I was thinking that it was.
Campbell, sure, the fact that something is better than its complement is one of the betterness facts. But to fix the zero, you need to add that anything no better or worse than its complement has value zero, and that is not a betterness fact. Isn’t this Broome’s point? He contrasts the reducibility of good to better with the irreducibility of length to longer, on the grounds that knowing all the longerness facts wouldn’t tell you which length is zero. There is a fact about which length is zero, but no fact about which value is zero. That’s the contrast. Isn’t that Broome’s view?
But perhaps we are simply lost in what amounts to no more than a verbal dispute. Broome definitely endorses (in “Goodness is reducible to betterness”) the claim that there is no natural zero for goodness. Agreed?
Jamie,
Maybe it is merely verbal. I took another look at the Broome article. He discusses various ways of deriving absolute degrees of goodness from betterness by appealing to “obvious standards”, including Jeffrey’s method of comparing a proposition with its negation. He concludes:
I think I agree with all of that, as do you. If there’s any residual disagreement it might over whether some standard of comparison might be in some sense uniquely correct, or whether it’s a merely matter of context which standard is the relevant one. I gather you accept the latter. I’m not sure which to accept.