The poll question was: If someone holds both (a) that an act is morally permissible if and only if it maximizes aggregate pleasure and (b) that nothing whatsoever (not even pleasure) is good, then s/he is what type of theorist?
And, as of 9:20 a.m. on 9/16/08, the results were as follows:
Total Votes: 73
A consequentialist: 75.3%
A nonconsequentialist: 24.7%
First of all, some of you have questioned why there’s no ‘not sure’ category for the second poll. Good question! The answer is that I don’t know squat about how to design or conduct polls properly. Sorry.
Now, let me explain why I’m interested in these results. In an interesting paper entitled “Consequentialism, Time, and Value”, David Killoren argues that, of the following two definitions given below, D1 offers the correct descriptive definition of ‘consequentialism’, where a good descriptive definition should match the usage patterns of competent speakers.
“D1 = whether an act is morally right depends only on consequences.”
“D2 = moral rightness depends only on the value (or goodness – I will use these terms interchangeably) of the consequences.”
In defense of his claim that D1 is the correct descriptive definition of ‘consequentialism’, Killoren says “it is hard to imagine a (competent, unbiased) user of philosophical jargon would not classify it [referring to the theory that consists of the conjunction of (a) and (b) in the poll question] as a consequentialist theory.” But now we see that even if it is hard to imagine, there are plenty of competent, unbiased users of philosophical jargon who would not classify it as a consequentialist theory. Indeed, the poll shows that roughly 25% of those polled classify it as a nonconsequentialist theory. Surely, there’s no good reason to think that all these respondents are either biased or not competent users of philosophical jargon. I think that we PEA Brains (i.e., participants on the blog) represent some of the best and brightest normative ethicists. Now I’m surprised that the number isn’t higher, but I think that it is high enough to indicate that there are competent, unbiased users of philosophical jargon who would not classify it as consequentialist.
What’s more, I would like to offer an explanation for the poll results: there is no canonical definition of ‘consequentialism’. This explains why there is such substantial disagreement over how to classify such a theory. And because there is no canonical definition of consequentialism, there is no correct or incorrect descriptive definition of ‘consequentialism’. I don’t think this because ‘consequentialism’ is a term of art, as Killoren suggests that I do in his paper. I readily acknowledge that there are many terms of art (in law, for instance) that do have canonical definitions. But it seems to me that a simple survey of the relevant literature shows that there is no settled or agreed-upon definition for ‘consequentialism’. There simply is no common usage pattern among all philosophers when it comes to the term ‘consequentialism’. Definitions of the term vary widely and in significant ways. For instance, quite a few philosophers think that agent-neutrality is definitive of act-consequentialism. Yet quite few others think that an act-consequentialist theory can be agent-relative.
Doug — Couldn’t one also say that there *is* a *canonical* definition of consequentialism (which does specify good consequences, though it admits of variation on other issues), but there’s still legitimate disagreement about whether to stick to it in using the term, i.e. whether it fits all the cases that the term might naturally cover.
Pat:
If that were the case, then why did 75% of the respondents classify the theory consisting in the conjunction of (a) and (b) as consequentialist?
It seems to me that Walter Sinnott-Armstrong is exactly right when he writes in his SEP entry: “In actual usage, the term ’consequentialism‘ seems to be used as a family resemblance term to refer to any descendant of classic utilitarianism that remains close enough to its ancestor in the important respects. Of course, different philosophers see different respects as the important ones. Hence, there is no agreement on which theories count as consequentialist under this definition.”
This would explain why so many respondents to the poll were reluctant to classify a theory that holds that an act is right iff it maximizes aggregate pleasure as nonconsequentialist even when it’s understood that the theory denies that pleasure is good. This theory too closely resembles classical utilitarianism, for many to be comfortable classifying it as nonconsequentialist.
Doug,
I agree with you and Walter that there is no canonical definition. (Or, maybe better yet: there are multiple, equally canonical definitions.) But I think Patricia is right that you could have 25% of people competently use a term in a “revolutionary” way, i.e., one that diverges from the canon. Also, I think you could have 25% of competent language users use the term simply incorrectly (without sacrificing their competence). Imagine if we took a poll as to what constitutes water, and 25% (or even more) said that it was clear, potable, liquidy stuff. That would be at best weak evidence that ‘water’ doesn’t mean H20, or that water cannot take non-liquid forms.
While I’m all for x-phi, I think this question–whether there is a canonical definition of a term of art like ‘consequentialism’–is one of those times where it’s best settled by seeing how the term is used in the literature of the art (e.g., how it is described theoretically, applied to paradigm cases, etc.). This, as you point out, will likely get you your “no canonical definition” thesis, anyway. Then again, I guess that having a poll to back up those data never hurts.
Josh,
You write: “I think Patricia is right that you could have 25% of people competently use a term in a ‘revolutionary’ way, i.e., one that diverges from the canon.”
But do you think that you could have 75% of people competently use a term in a “revolutionary” way, i.e., one that diverges from the canon? That, I take, is what Pat is suggesting in claiming that the canonical definition would have rightness be a function of goodness even though 75% of the respondents claimed that a theory that doesn’t make rightness a function of goodness is consequentialist.
Doug — We may just disagree about what “canonical” implies. Josh’s reference to “how the term is used in the literature of the art” is what I had in mind — i.e. the official definitions. Though I didn’t vote, I’d actually agree with the 75% if the question were whether the term should be extended beyond the official definition.
I just thought to look up “canonical” on answers.com, and the relevant definition seems to be: “Conforming to orthodox or well-established rules or patterns, as of procedure.” Note the “or.” I’d say that the 25% minority went with the orthodox definition, but as Sinnott-Armstrong notes, it’s not that well-established.
Josh’s reference to “how the term is used in the literature of the art” is what I had in mind — i.e. the official definitions.
This is what I had in mind as well. The problem is that there is no one way that the term ‘consequentialism’ is used in the literature, and if there are any “official” definitions, then there is not just one, but many, official definitions, which conflict in fundamental and important ways. My claim is that there is no one canonical definition of consequentialism. If you think otherwise, tell me what the one canonical definition of consequentialism is, and I bet that I can find at least a half dozen uses of the term in the literature (in the canon) that conflict with that definition. Here are some instances of how ‘consequentialism’ is used in the literature:
“Consequentialism, as its name suggests, is the view that normative properties depend only on consequences.” Sinnott-Armstrong, SEP
(Note that there is deliberately no mention of value here, for he distinguishes consequentialism from evaluative consequentialism.)
“My use of the term ‘consequentialism’ is more expansive than most people’s. In my usage, egoism as theory of rationality, which gives each agent the sole ultimate aim of maximizing his own personal welfare, is a consequentialist theory.” Anderson 1993.
(Note that, on her definition, a consequentialist theory isn’t necessarily a moral theory, nor is it necessarily an agent-neutral theory. And since 1993, Anderson’s usage has become less and less idiosyncratic.)
“Consequentialism assesses the rightness or wrongness of actions in terms of the value of their consequences.” McNaughton, Routledge Encyclopedia of Philosophy
(Note that an appeal to value is essential in this definition — cf. Sinnott-Armstrong’s.)
“Consequentialism in its purest and simplest form is a moral doctrine which says that the right act in any given situation is the one that will produce the best overall outcome, as judged from an impersonal standpoint which gives equal weight to the interests of everyone. Somewhat more precisely, we may think of a consequentialist theory of this kind as coming in two parts. First, it gives some principle for ranking overall states of affairs from an impersonal standpoint and then it says that the right act in any given situation is the one that will produce the highest ranked state of affairs that the agent is in a position to produce.” Scheffler, Consequentialism and Its Critics
(Note that consequentialism is necessarily a moral theory and necessarily an agent-neutral theory on this definition.)
These different usages do not amount to minor quibbles such as those over actual or expected value or satisficing or maximizing functions. Whether ‘consequentialism’ is necessarily agent-neutral, necessarily about the value of consequences, necessarily about morality, etc. are very fundamental issues. Without resolving these sorts of questions it is difficult to having anything interesting to say. For instance, it’s just plain false that consequentialism must allow the killing of one to prevent five other comparable killings if the theory isn’t necessarily agent-neutral. So these issues go to the very heart of the debate over consequentialism.
I forgot to mention that there are philosophers like Norcross and Howard-Snyder who want to say that consequentialism needn’t offer a criterion of rightness at all, but instead needs only to offer us a scale by which to judge acts morally better or worse than others.
And I’m only getting started. If you want things to get even weirder, then we need only turn to how more fringe-elements like myself use the word ‘consequentialism’.
Doug — Work in “the canon” needn’t employ an “orthodox” definition (as in the answers.com reading). As I say, I’m willing to extend the definition in this case (largely because there doesn’t seem to be another category for the view you stated), but in general I have qualms about doing so. I’m reading something now, e.g., from the cognitive science literature that purports to defend “sentimentalism,” but takes the term to cover any view that assigns an *crucial role* to emotions in moral judgment, even if just in explaining its development. This is another case of referencing a term in isolation (and here perhaps in ignorance) of what was formerly an established use in the literature. It essentially wipes out the important middle-ground between sentimentalism and rationalism, and once the new use catches on — as it’s likely among people in cognitive science with little knowledge of the tradition of moral philosophy — the “canon” will be extended, and this use will then count as “canonical” in your sense. I find this sort of thing unfortunate, both as a disconnect from the tradition and insofar as it blurs over significant distinctions for the sake of a neat contrast.
Hi Pat,
I agree that someone’s use of a canonical term can stray from the canon and that work within the canon can seek to extend the use of some term beyond that of its canonical or orthodox usage. I’m just having a hard time seeing what the orthodox or canonical usage of ‘consequentialism’ is. Can you tell me what the canonical definition of consequentialism is and what makes it the canonical definition as opposed to just one of many definitions that we find in the canon? You seem to think that some usage patterns can be orthodox although not well-established. But when I look up the word ‘orthodox’, I find definitions such as these:
“1. of, pertaining to, or conforming to the approved form of any doctrine, philosophy, ideology, etc.
“2. of, pertaining to, or conforming to beliefs, attitudes, or modes of conduct that are generally approved.
3. customary or conventional, as a means or method; established.” –dictionary.com
But when I look at the literature, I don’t find any use of the term ‘consequentialism’ that’s established, nor do I find patterns of usage that are generally approved of. Moreover, I don’t know of any official doctrine in philosophy (in the sense that the Bible represent the official doctrine for certain religions) that I can appeal to in determining what the canonical definition of ‘consequentialism’ is. The closest that I can come to for an official doctrine are certain classic texts, common textbooks, and prominent encyclopedias, but I don’t find any more agreement here than I do in the rest of the literature. It’s difficult enough to even get a straight forward definition of ‘utilitarianism’ from a text like Mill’s Utilitarianism.
Would you grant that for a definition to be canonical or orthodox, there has to be either some official doctrine that makes it so or some well-established common pattern of usage within the canon that makes it so. In the case of ‘consequentialism’, I find neither. If you find differently, I ask you to state what the canonical definition is as well as to state what doctrine or well-established pattern of usage makes it the canonical one.
Perhaps, if there was some well-established pattern of usage in the classic texts, common textbooks, or prominent encyclopedias, I might buy that there is some orthodox definition of ‘consequentialism’, but I don’t find any. Perhaps, though, I’m looking at the wrong texts, textbooks, and encyclopedias.
Hi Doug,
First, thanks for this insightful post. One thing we definitely agree about is that PEA Brains are among the best and brightest, so I’m really excited to have their attention drawn to my paper. (In fact, a great forum like this makes me wonder whether events like ISUS are really worth the costs of transporting, lodging and providing space for all those people. But I digress…)
A quibble: I didn’t mean to suggest that you endorse the “term of art” argument. What I say in the paper is that you “mention” it, and that’s really all I meant. The purpose of that footnote was just to show that I am not making up the line of argument to which I’m responding. (I do admit, however, that the footnote, as-is, doesn’t really belong, and in a final version should either be expanded or deleted.)
On to more substantive issues.
Here’s an alternative explanation for the poll results. Suppose the following were true: (1) D2 has a much better claim than D1 to being the “official” definition — i.e., when asked to provide a (descriptive) definition for “consequentialism,” most competent users would give something close to D2. However, (2) D1 better captures our unmediated intuitions. That is: If a competent user were asked to classify a given moral theory, and if she were to do so without (conscious) reference to any general definition like D1 or D2, her responses would fit the D1 pattern, not the D2 pattern.
If (1) and (2) are true, what should we expect when philosophers are asked to classify what I call Direct Utilitarianism (which is the conjunction of (a) and (b) in Doug’s post)?* We should expect that a certain number of them would rely on their unmediated intuitions, and would therefore vote “Yes,” whereas others would rely on the official definition, and would therefore vote “No.”
What proportions should we expect? I think that the weirder the theory one is being asked to classify, the more likely one is to worry about one’s unmediated intuitions and therefore resort to an official definition. Direct Utilitarianism is indeed very weird. So we should expect a relatively high percentage of respondents would resort to an official definition. In this light, 25% in the “No” column seems about right, perhaps even low.
[Aside: I am to some extent departing here from what I say in the paper, but I don’t think I’m departing from what I really meant. Anyway, I acknowledge that there is some mind-changing going on here.]
How would we test my explanation? We could try checking people’s classification of non-weird theories. For example, as I suggest in the paper, we should consider just (a) by itself. There’s nothing weird about (a). If people say (a) is sufficient for consequentialism, then they are speaking in accordance with D1; if they say (a) is not sufficient for consequentialism, then they are speaking in accordance with D2. I expect a properly conducted survey would show just about everybody on the D1 side of this division. (Of course, I can say this so confidently because there’s no way to prove me wrong with another poll, since a lot of PEA Brains have seen Doug’s post by now and I can always say that Doug spoiled their intuitions.)
Analogy: The explanation I propose has “consequentialism” be something like “bachelor.” Perhaps the official definition of “bachelor” is “unmarried male,” but if you describe an 8-year-old unmarried male to people, a lot of people will doubt whether he’s a bachelor. I suggest what’s going on in that case is a conflict between official definition and the definition implicit in unmediated intuition. I also suggest that if you present people with a really weird case of an unmarried male, a lot more of them will resort to the official definition.
In any case, I think that when you’ve got a case where official definition conflicts with unmediated intuition, something interesting is going on and it’s worth finding out more. But whether we’ve got a case like that here is something I certainly can’t claim to shown in this comment.
*=I am aware that a different theory often goes by the name “direct utilitarianism.” I still think “direct utilitarianism” is a good name for the theory under discussion here.
By the way, to support my claim that (1) D2 has a much better claim to being the official definition than D1, I think you can check our Nir Eyal’s ISUS X paper, “Non-Consequentialist Utilitarianism.” I believe he defends something close to that view. (I haven’t actually read the paper yet, so I may be wrong about this. But I did see the talk.)
Doug, you ask:
Yes, I do. I think we could all decide tomorrow that the word “deontology” will henceforth refer to consequentialism. In so doing, we’d be diverging from the canon. But we’d be competent. Words’ meanings can change. (Also, meaning might be contextualized to different discussions, e.g., philosophy and cog sci, or sub-discussions within philosophy.) In a dovetail with David K’s post, the example I used during Nir’s talk was “awful.” “Awful” used to mean what it looks like it would mean: something like full of, or inspiring of, awe. Now it means something like, really bad. So when we use it to mean really bad, we don’t display incompetence. Instead, the meaning of the word has simply changed. And we could change the meaning of “awful” again. We could even make it mean consequentialism! (Okay, that’s a cheap shot from a Kantian.)
(FWIW, for that reason, I thought that Nir’s talk brought out that the canon includes at least two definitions of consequentialism — one from times of yore and one [among many, Doug and I and others would say] that is common today. Nir thought that the old one was better, but I think that is consistent with saying that there are two definitions there. And if we want to start evaluating them, I agreed with Alasdair Norcross’s point that, by today’s lights, a good characterization of consequentialism would have it say, not only that an act is right iff it produces good consequences, but also that it is right because it produces good consequences. I think Doug agrees with this, too, so I’m not sure that appealing to Nir’s paper will be convincing evidence for a big chunk of the interested parties.)
Hi David,
Regarding the quibble:
You write above: “I didn’t mean to suggest that you endorse the ‘term of art’ argument. What I say in the paper is that you ‘mention’ it, and that’s really all I meant.”
The ‘term of art’ argument goes as follows, quoting from your paper: “Since
‘consequentialism’ is a term of art – a bit of philosophers’ jargon – it is commonly
supposed that all useful or informative definitions for it must be merely stipulative.”
I don’t even ‘mention’ this argument. What I say in the paper that you refer to is: “I take ‘consequentialism’ to be a term of art and thus open to being defined differently by different philosophers.” To say that a term is open to being defined differently by different philosophers is not to say that “all useful or informative definitions for it must be merely stipulative.”
On to more substantive issues:
In your paper, you want to argue that “D1 represents a correct descriptive definition for consequentialism; and D2 represents an incorrect definition for consequentialism,” and you claim that the “criteria provided by a (good) descriptive definition should match the usage patterns of competent speakers.”
Now isn’t the claim that D2 doesn’t match the usage patterns of competent speakers an empirical claim? So don’t you need to provide some empirical evidence for this empirical claim? But where’s your empirical evidence? The one of the main pieces of evidence that you give in your paper is an appeal to “our” intuition that direct utilitarianism is to be classified as a consequentialist theory. But why think that this our intuition and not just the intuition that you and some others have? That is, why think that there is some common intuition held by the vast majority of competent users of philosophical jargon? The poll would suggest that not everyone shares your intuition. Also, the actual usage patterns as presented by the various quotes that I provide above in support of my contention that there is no common usage pattern to be found in the literature both suggest that your empirical claim is false. Now you want to explain away the poll data. You say that 25% classifying direct utilitarianism as nonconsequentialist is just what you would expect given that some people’s responses are mediated. But how do you know what to expect? Again, these sorts of expectations could only be established by gathering some empirical evidence. Where’s your empirical evidence?
Josh,
You’re right. I agree with everything you just said.
Doug — I meant to be suggesting that the usage of “consequentialism” was *no longer* “well-established,” i.e. that some authors weren’t adhering to the standard or orthodox definition: that rightness of an act depends on the worth of its consequences (in various formulations). If you agree with everything Josh just said, then I take it you agree with his statement that “by today’s lights, a good characterization of consequentialism would have it say, not only that an act is right iff it produces good consequences, but also that it is right because it produces good consequences.”
“by today’s lights, a good characterization of consequentialism would have it say, not only that an act is right iff it produces good consequences, but also that it is right because it produces good consequences.”
I hope everyone doesn’t agree with this. If this is what consequentialism is, then regular old maximizing utilitarianism fails to be a form of consequentialism. Acts can have good consequences without having the best consequences. I don’t know how to define it, and I suppose S-A’s “family resemblance” take is probably about right.
I’m curious about the claim that at some time in the past, there was a canonical definition of ‘consequentialism’. When was this, and what was the definition? Who invented the term?
Pat,
I’m not sure whether we’re disagreeing. Perhaps, we are, for I think that what Josh cites as a good characterization of consequentialism is not *the* standard definition, but just one of many definitions that philosophers offer. And I understood Josh’s point to be that a good characterization of consequentialism must not only identify which features of actions correlate with the rightness of actions, but also assert which features of action *make* it right. So, for instance, I think that, by today’s lights, a perfectly good characterization of consequentialism would say that an act is right just when, and because, it produces either the best outcome or the outcome that is best for the agent.
And, by the way, I’m not at all clear on what you mean when you say “the worth of its consequences (in various formulations).” Is ethical egoism a consequentialist theory on this definition? I wouldn’t think so, as I wouldn’t think that the consequences of my doing x have more worth than the consequences of my doing y just because the consequences of my doing x are better for me than the consequences of my doing y.
If you want to claim that Josh’s cited characterization is the standard one just because it is more common than any other alternative characterization, then okay fine; our disagreement is only verbal. But I would have thought that for some definition to be *the* standard one it would have to have more than a plurality of experts who endorse it, but rather a majority (at least) of experts who endorse it. And I doubt that the majority of experts do endorse the characterization that Josh gives.
Ben: Here, here. Unfortunately, I didn’t see your comment until, I posted mine. But you make some of the points that I wish that I had made.
Josh —
I seem to recall your ‘awful’ point being directed at Eyal’s def’n for utilitarianism, not for consequentialism. It’s there that the ‘old vs. new’ issue is more salient, I think. I don’t take Eyal to be arguing in favor of restoring or preserving an older definition for *consequentialism*. Instead, I think he is trying to show that a certain definition for consequentialism is extremely widespread NOW and therefore should be accepted by the recipients of his argument.
In the paper Eyal defines consequentialism as the view that “an act is morally right if and only if it maximizes the good.” That’s similar in the relevant respects to my D2 let’s call it D2′. (The difference between D2 and D2′, I take it, is that D2′ is subject to the ‘because’ objection that you mention, and also rules out non-maximizers, whereas my D2 does neither of these. These differences are important, but not important in the present context.) In Section III, he considers alternative definitions. He says “the alternative definitions are nearly idiosyncratic.” He says “Kagan, most contemporary philosophers, and I [i.e., Nyal]” define consequentialism in terms of the good. I take it that this indicates agreement with the claim that something like D2′ is the official definition for consequentialism.
David,
You write: He says “Kagan, most contemporary philosophers, and I [i.e., Nyal]” define consequentialism in terms of the good. I take it that this indicates agreement with the claim that something like D2′ is the official definition for consequentialism.
How does the fact that most (a majority) of philosophers define ‘x’ as ‘y’ indicate that ‘y’ is the *official* definition of ‘x’? Would that alone make it the *official* definition? Most philosophers might define ‘supervenience’ or ‘internalism’ a certain way, but it doesn’t follow from this that the majority are giving the *official* definition.
Moreover, from the fact that most philosophers define ‘consequentialism’ in terms of the good, it wouldn’t follow that most would define ‘consequentialism’ as D2.
Doug —
I’m sorry I misattributed a mention of a counterargument to you in a footnote.
Regarding the need for empirical evidence — I agree, there is a need for empirical evidence. I want to emphasize again that my alternative explanation for the poll results is just speculative; I didn’t show it to be the case, or come anywhere near doing that, simply by offering it.
Doug — a fair point about ‘official.’ As you point out above, I certainly couldn’t claim that there is some authority somewhere who has decided what ‘consequentialism’ means and thereby given it an official definition. I suppose ‘standard’ would have been a better word than ‘official’ — ‘standard’ in something like the way that standard definitions for ‘bachelor’ generally have something about being unmarried.
Doug —
“Moreover, from the fact that most philosophers define ‘consequentialism’ in terms of the good, it wouldn’t follow that most would define ‘consequentialism’ as D2.”
This is definitely true. But I think it does follow that their definition would (very likely) be similar to D2 in all the respects relevant to the argument of my paper.
Apologies in advance, for I have to be quick about this, but just a couple of things:
Ben, you’re right. I was shooting for a quick and sloppy, so rough and ready, characterization of the point that a good characterization of consequentialism will say not only *when* an act is right, but *what, fundamentally, makes it right*. Apologies for the sloppiness; hopefully it’s still clear, though, that some of us won’t find the referenced paper’s characterization compelling, since it only includes the former and not the latter.
But, that said, David K is right that I was misremembering my own objection to Eyal’s point (!). I should really stop doing that. In any case, I hope it’s still relatively intuitive that characterizations of doctrines can change over time. And they can vary even within one temporal location, eliminating the need for *one* official definition.
“…characterizations of doctrines can change over time. And they can vary even within one temporal location, eliminating the need for *one* official definition.”
This is definitely true, but it also seems reasonable to think that characterizations of doctrines can be correct and incorrect. If I tell you that Nazism involves a commitment to basic human equality, I am just wrong — provided I am talking about Nazism when I use the word “Nazism” (and not just using my own idiosyncratic definition for that word).
It’s a step beyond this to say that very widespread characterizations of doctrines can be incorrect. But this also seems right. Suppose it were widely believed (a) that Nazism involves a commitment to basic human equality and (b) that Nazism implies that Jews, but not Aryans, ought to be exterminated. Presumably, it can’t be that both (a) and (b) are correct. If we decide that (b) is correct, then we should decide that a characterization of Nazism involving (a) is incorrect.
For what it’s worth, in the paper, I was trying to argue that something like this is going on with consequentialism. Suppose, just for a moment, that Eyal and I are correct that (something relevantly like) D2 is widely regarded as the standard def’n for consequentialism. If that’s so, isn’t it odd that 75% of the responses to Doug’s poll are inconsistent with this definition (and consistent with D1)? It seems that those who fall into that 75% need to rethink either their response to the poll or the standard def’n.
Let me emphasize that I take the demand for empirical evidence seriously. I don’t claim that my (or Eyal’s) assertions about what is the standard definition are authoritative. What I’m doing is throwing out guesses, seeing where they lead, seeing whether others find them plausible, etc. etc. etc.
Another thing is that I haven’t said anything about why any of this is significant, even if it’s true. But that’s another conversation, I think.
David,
Why, as philosophers as opposed to linguists, should we be interested in whether a given definition of ‘consequentialism’ comports with the usage patterns of competent speakers? In your paper, you point to the fact that it is important to know whether the sort of moral theory which “we (i.e., the members of the philosophical community) have in mind [which, according to you, is D1] when we talk about consequentialism” is consistent with common-sense morality. But is this question any more important or interesting than the question of whether or not D2 is consistent with common-sense morality (whether or not D2 is what we mean by ‘consequentialism’)? In fact, might it be that the question of whether or not D2 is consistent with common-sense morality is much more important and interesting, for perhaps all theories that meet D2’s criteria, but not all theories that meet D1’s criteria, share with classical act-utilitarainism some very compelling idea. And, thus, showing that D2 is consistent with common-sense morality would show that we could have what’s most compelling about classical act-utilitarianism while avoiding its counter-intuitive implications.
When it comes to philosophical terms of art, shouldn’t we be most interested in which definitions bear the most philosophical fruit, not which definitions best comport with our usage patterns?
So why care what the correct descriptive definition of ‘consequentialism’ is? It’s not enough to say, as you do, that we want to know whether that theory is compatible with common-sense morality, for we want to know that about all moral theories (named or unnamed), for we think that an important desideratum for a moral theory is that it comports with our common-sense intuitions.
Doug — We don’t agree.
Doug —
I’m in the middle of writing a paper that tries to answer your question. My ideas are still kind of developing, but here’s what I think now. I’d be delighted to get your reactions.
Take “knowledge” to begin with. The story* goes that everybody thought JTB provides the correct descriptive def’n for knowledge. Then Gettier cases were provided. People consulted their unmediated intuitions and came back with the conviction that there is no knowledge in such cases. So they concluded that JTB isn’t adequate as a descriptive def’n for knowledge. Why did philosophers care?
[*= yes, this is a badly oversimplified version of the actual history here. I hope that doesn’t hurt its purpose as an illustration.]
Here’s one answer. There’s some property of beliefs we’re picking out when we call a belief knowledge. We care to know whether our beliefs have this property, partly because we want our beliefs to have this property. We think we can recognize this property in a belief if we are given enough information — i.e., we think we know it when we “see” it. So we trust our unmediated intuitions about whether a belief amounts to knowledge. Prior to the Gettier cases, we thought we could also provide a simple rule — a definition — by which to determine whether that property attaches to a belief. What the Gettier cases show is that it’s really no simple matter to find such a rule. We might eventually find it. Or we might never find it. Or maybe it’s already been found and just hasn’t won out in debate yet. But in any case, we’re unlikely to stop talking about knowledge; and we’re unlikely to stop thinking we can recognize knowledge in sufficiently specified cases.
I think “consequentialism” is something like this. There’s a property of a moral theory that we pick out when we call it a consequentialist theory. We care to know whether a theory has this property. Some care to know whether a theory has this property because they think any plausible moral theory has to have it. Others care to know whether a theory has this property because they think any plausible moral theory has to lack it. (There are other reasons to care, too.) Our unmediated intuitions seem to provide reliable guidance as to whether a theory has this property. In addition, we might have thought that D2 provides a simple rule by which to determine whether a theory has this property, just as we once thought JTB provides the rule for knowledge. I think the case of Direct Utilitarianism shows that D2 can’t do that job. So I think Direct Utilitarianism is a kind of Gettier case for moral taxonomy.
David,
There’s a big difference between ‘knowledge’ and ‘consequentialism’, right? ‘Knowledge’ is a term of ordinary English, whereas ‘consequentialism’ is purely a technical term.
More to the point, my question is *not* why should we care whether a moral theory has the property that the term ‘consequentialism’ picks out? We should, in general, be concerned with what properties moral theories have. My question is why should we care whether ‘consequentialism’ picks out the property that D1 picks out as opposed to the property that D2 picks out? What does it matter which definition matches the common usage patterns of competent speakers? It’s important to know which theories have the D1-property and it’s important to know which theories have the D2-property. And it may turn out that it’s more important to know which theories have the D2-property, because it may be that having the D2-property is important to assessing whether a theory is plausible whereas having the D1-property is not. So, again, I ask why care which property is picked out by the common usage patterns of competent speakers as opposed to just caring which properties of a moral theory make it plausible or implausible (whether or not our term ‘consequentialism’ picks out any of these properties)?
David,
You wrote, “This is definitely true, but it also seems reasonable to think that characterizations of doctrines can be correct and incorrect.”
Just to be clear, I agree with you that there are incorrect characterizations of doctrines (and I think this is consistent with everything I’ve said). I also, as I said, think that there are better and worse characterizations of doctrines.
You also write,
I agree with this too; one of the points of my initial comment to Doug was that linguistically competent people can misuse terms with which they are competent. But then you continue with something I think I disagree with:
I think both (a) and (b) could be correct, in principle. If Nazism really did subscribe to both commitments, then it would be contradictory, or paradoxical at best. But that doesn’t mean that it can’t subscribe to inconsistent propositions. (Indeed, I think that something like this is what is sometimes — not always, but sometimes — alleged when people accuse deontology of being paradoxical.) I think what you were getting at in your first stab is closer to the truth: we don’t accept (a) because ‘Nazism’ is defined such that (a) is false.
So our parallel question would be whether ‘consequentialism’ is defined such that certain propositions are false. Surely it is; but this doesn’t mean that there aren’t multiple, separate definitions being used by members of the same linguistic community. Thus I’m not sure I follow you when you say, “Suppose, just for a moment, that Eyal and I are correct that (something relevantly like) D2 is widely regarded as the standard def’n for consequentialism. If that’s so, isn’t it odd that 75% of the responses to Doug’s poll are inconsistent with this definition (and consistent with D1)? It seems that those who fall into that 75% need to rethink either their response to the poll or the standard def’n.” I haven’t read your paper yet, but I thought, first, that one of the live questions for the purposes of Doug’s post was just what you want us to suppose. Why isn’t the 75% just (non-decisive but suggestive) evidence against your supposition? A second question is supposed to be whether different members of the relevant linguistic community are working with different “standard” definitions. So more generally, I think we might have to give up the supposition that there is *one* standard definition. So just to sum up, this is consistent with saying that people can give incorrect characterizations of it; that there are better and worse characterizations of it; and that people can in principle widely mischaracterize it.
Doug —
So now you’re not only mentioning but also making the (or a) ‘term of art’ argument! I find this line of argument very puzzling, so I’m hoping you’ll fill me in. What is it about the difference between technical terms and ordinary English which makes the latter but not the former ripe for this kind of analysis? Are you assuming that technical terms all have, or ought to have, purely stupulative meanings?
Until we have an adequate definition for “consequentialism,” we will be somewhat in the dark about the property “consequentialism” picks out. Once we have the definition, we may find out that there’s no reason to care whether a theory has it; we may find out that having this property doesn’t make a theory better or worse. But the fact that philosophers have exhibited such longstanding interest in this property (whatever it is) should make us want to know about it. This is because, in general, when large numbers of philosophers are drawn in a certain direction over a long period of time, we have some evidence that there’s something interesting lying in that direction. But we won’t know for sure until we’ve got the definition.
Analogy: You’re hanging around in some public place and you notice that a lot of people are rushing down a certain corridor. Other people are staying as far away from the corridor as they can get. Don’t you wanna know what’s down that corridor?
David,
Please read my comment carefully. I’m not arguing along these lines at all:
(1) ‘Consequentialism’ is a term of art.
(2) All useful definitions of terms of art must be merely stipulative.
(3) Therefore, any useful definition of ‘consequentialism’ must be merely stipulative.
Indeed, I deny (2), as I admit that the correct descriptive definition of ‘consequentialism’ (if there is one in the case of ‘consequentialism’, which I highly doubt) might be an incredibly useful definition. But it might not be, right? Who cares, then, whether the most useful definition of ‘consequentialism’ is the “correct” one? What I care about as a philosopher is which definition is the most useful, whether it happens to match the common usage patterns of philosophers or not.
So why should I care about all your arguments about D1’s being the correct definition of ‘consequentialism’? Shouldn’t I instead be interested in whether D1 is a useful definition?
So I’m ask: “Why should we care about what property ‘consequentialism’ picks out?” And you answer: “Because that’s the property that many find interesting.” But you’re just assuming that “philosophers have exhibited such longstanding interest in this property (whatever it is).” I’m not so sure that it is the property that ‘consequentialism’ picks out that has been of interest. It’s true that philosophers have been interested in many theories which they have labeled ‘consequentialist’. It’s not at all clear, though, that this is because they’re interested in the property that’s definitive of ‘consequentialism’. It may be that the sorts of consequentialists theories that philosophers have been interested (e.g., various classical forms of utilitarianism) have had not only the D1-property, but also the D2-property, in common. And it may be that the property of interest is really just the D2-property, when in fact it’s the D1-property that’s definitive of ‘consequentialism’. And all this leads me back to my original question, which you still haven’t answered. Why be interested in which property is definitive of consequentialism as opposed to being interested in which property of the theories that philosophers have typically labeled consequentialist is the property that philosophers find to be of particular interest. Indeed, there’s a lot of evidence from the literature that what philosophers like Scheffler and Foot find compelling about utilitarianism is the idea that it could never be wrong to prefer a better state of affairs to worse state of affairs — or that it could never be wrong to bring about the best available state of affairs. This would suggest that D2 is the property that philosophers find most interesting.
Analogy: Yes, I want to know what’s of interest down the corridor, but I don’t care whether what is of interest to them down the corridor is something essential to that corridor. So I would be interested in a paper that argues that D1 is a property that makes a theory compelling and that explains why so many philosophers have found theories such as utilitarianism attractive. But I’m not so much interested in whether D1 matches the common usage patterns of philosophers except insofar as I want to be able to communicate with them without talking past them. But I can avoid miscommunication in the following manner: by first finding out what I take to be the most compelling property of act-utilitarianism and then just stipulate that, by the term ‘consequentialism’, I’ll mean ‘all and only those theories that have this property’. This is what I have done. And if you come along and tell me that that property is not in fact the property that matches the common usage patterns of philosophers, I’m going to ask: “So what?”
David,
Just to be clear. I’ve endorsed only the following two arguments:
Argument I:
(1) If a term is purely a philosophical term of art for which numerous definitions have been offered, then we should be interested in which of the various definitions that are on offer bears the most philosophical fruit, not which of the various definitions on offer best comports with philosophical usage. In either case, since there are various definitions on offer, it is important to stipulate which definition one is using whenever employing such a term.
(2) ‘Consequentialism’ is purely a philosophical term of art.
(3) Therefore, [fill in the consequent in (1)].
Argument II:
(1) There is a single correct descriptive definition of a term if and only if the vast majority of competent speakers exhibit some common usage pattern with regard to that term.
(2) In the case of ‘consequentialism’, it is false that the vast majority of competent speakers exhibit some common usage pattern with regard to that term.
(3) Therefore, there is no single correct descriptive definition of ‘consequentialism’.
Doug — I agree with a lot of what you’ve said. I think to a certain extent we’re just emphasizing different possible outcomes of the sort of analysis I’m trying to do. Insofar as I understand what it means for a definition to “bear philosophical fruit” I agree that this is the ultimate standard by which to judge the usefulness of a definition.
Anyway, this interchange has been extremely helpful to me. I have more to say, but I think I’m nearing the end of my present capacity to make good responses. I need to think about all this a little bit more. Maybe once my next draft is pulled together I can send it to you?
But I do have one further question for you. In Argument I, is there a reason why you’ve got “pure philosophical term of art” in the first premise? Would you endorse a version of the premise with the antecedent “If numerous definitions for a term have been offered…”? Also, what does it matter whether numerous definitions have been offered? If there’s just one def’n on offer, but another bears more “philosophical fruit,” why shouldn’t that be the one to use?
Josh — I have little to say other than that your response really helped clarify some issues for me. I think I agree with just about all of it. I’ll need to think more before I have anything useful to say in response.
Would you endorse a version of the premise with the antecedent “If numerous definitions for a term have been offered…”?
Probably not. I would, for instance, reject the following: If numerous definitions for a term have been offered, then we should be interested in which of the various definitions that are on offer bears the most philosophical fruit, not which of the various definitions on offer best comports with philosophical usage. Numerous definitions of knowledge have been offered, but I think that we should be interested in which definition best matches the common usage patterns of competent speakers.
what does it matter whether numerous definitions have been offered? If there’s just one def’n on offer, but another bears more “philosophical fruit,” why shouldn’t that be the one to use? It probably doesn’t. I was offering a sufficient condition, not a necessary condition, for its being the case that we should be more interested whether it bears any fruits than whether it comports with common usage.
I’m broadly in agreement with Doug on this one. I’d just like to add the following point, which seems relevant.
That some sentences using a given term are determinately true, and others determinately false, doesn’t imply that the term is unambiguous. For example, in maths, the term ‘countable’ is ambiguous. Sometimes it means being equinumerous with the natural numbers, and sometimes being equinumerous with some subset of the natural numbers. (By the way, ‘natural number’ is ambiguous too.) Nonetheless, some sentences using ‘countable’ are unequivocally true, e.g. ‘The natural numbers are countable’, and other unequivocally false. e.g. ‘The real numbers are countable’. Likewise, sentences using ‘Nazism’ or ‘consequentialism’ may be determinately true or false; but this doesn’t imply that either term is unambiguous.
Incidentally, I wonder what David would say about ‘countable’. Is there a uniquely correct definition of this word? And if so, is finding out the correct definition a matter of any importance?
Campbell — a useful point about ambiguity, thanks.
Since I am not a mathematician, I can’t say one way or the other whether it would be useful to go in search of a correct definition for ‘countable.’
I don’t think all terms of art can be profitably analyzed (in the way I’m trying to analyze ‘consequentialism’). But I think some terms of art can be profitably analyzed. I think that whether a term can be profitably analyzed has little to do with whether it’s a term of art.
Take ‘mammal,’ a biological term of art. There’s an interesting debate about how ‘mammal’ is to be defined.
‘Knowledge’ is a common-use word simply because its concept is useful to ordinary people in ordinary contexts. But it could have turned out (in a weird, weird world) that only a small handful of specialists had use for this concept. It seems odd to think that, if things had happened to turn out that way, there would be nothing philosophically interesting about an analysis of knowledge.