About Me

My photo
Australian philosopher, literary critic, legal scholar, and professional writer. Based in Newcastle, NSW. My latest books are THE TYRANNY OF OPINION: CONFORMITY AND THE FUTURE OF LIBERALISM (2019); AT THE DAWN OF A GREAT TRANSITION: THE QUESTION OF RADICAL ENHANCEMENT (2021); and HOW WE BECAME POST-LIBERAL: THE RISE AND FALL OF TOLERATION (2024).

Wednesday, April 06, 2011

Is science strictly objective?

One of the objections that I often receive when I claim that moral judgments are not strictly objective is that the same reasoning can be deployed to demonstrate that science is not strictly objective either. Apparently that is meant to be a reductio ad absurdum, as if we possess transcendental knowledge that science is strictly objective.

Someone made this point on my Facebook profile overnight, I've seen Sam Harris use the argument on various occasions, and I see it frequently in debates on the internet.

It actually opens up huge issues in scientific epistemology, and I can't deal with those here. But there are some oddities about the argument.

First, note that the claim that moral judgments are not strictly objective is continuous with the claim that other kinds of evaluations are not strictly objective. They may be rational and non-arbitrary, but they are not binding on all rational beings. It looks to me as if the argument for this is compelling when it comes to non-moral evaluations.

So forget about moral judgments for the moment. Are we going to insist that all our practices of evaluation are strictly objective for fear that saying anything to the contrary will undermine the authority of science? I fully agree that the things we evaluate have objective characteristics - but the point is that there's going to be some slippage in what we require of these things (how do their objective characteristics conduce to meeting our desires, purposes, and so on?). Strictly speaking, there is no "we" here. Even if the people in a small group form a community with, for the scope of the argument, a unanimity of purposes, etc., they will not have a unanimity with all rational creatures in the universe ... or even all rational creatures on Earth.

Evaluations have an objective element, and they can be debated rationally. Eventually, however, evaluations are made relative to desires, etc., and in that sense have a subjective element. Outside of the sphere of moral judgment, this claim isn't especially controversial among educated people in culturally open societies. Do we now have to abandon this common understanding of non-moral evaluations for fear that we are undermining the authority of science?

That's one aspect that I haven't seen discussed. I'd like to know why a non-objective element in moral judgment is threatening to science but a non-objective element in evaluations more generally is not.

Here's another aspect. Imagine for the sake of argument that some kind of argument goes through to demonstrate that science is not strictly objective. Presumably the argument will have to be that our choices of scientific hypotheses and theories are actually something like evaluations. We evaluate scientific hypotheses against "values" such as explanatory power, parsimony, and track record in making novel predictions.

Ordinarily, we think of these as criteria for truth - or for likelihood of truth. Outside of science, we certainly do use them that way. But what if we do think of them as simply things that we happen to value? Well, yes - if you do think of explanatory power, parsimony, predictive record, and the like as things we just happen to value you'll turn science into something that is not objective.

What if you say that they are not things that we just happen to value but value for some other reason? Well, it might be that our reason is that they are indicators of likelihood of truth! In that case, the authority of science doesn't seem to be undermined after all. Sure, if we don't care about truth we won't use these criteria. But that in no way detracts from the fact that people who do use these criteria are discovering propositions that are likely to be true. And that gives science authority as a truth-seeking enterprise. Where's the problem?

Could we have some other non-arbitrary reason for valuing these particular things? Well, I can't immediately see what it would be. Nor do I see how we or any other rational beings in the universe really have much choice but to employ such criteria as explanatory power, parsimony, and predictive record. To call them "values", as if we have some real choice whether or not to value them, is misleading.

But all that said, there are many views in scientific epistemology. Perhaps there's some way of preserving science's authority, or much of it, while also claiming that it is not strictly objective, or that it only finds "empirically adequate" theories, not truths about the universe, or whatever. I don't really want to get into that today, and in all honesty I'd need to research the issue a lot further.

Without having done so, I take a straightforward scientific realist view. The criteria we use in science really are criteria of likelihood of truth and would be used by all rational beings that engage in science ... at least once they reach a certain stage of cultural evolution. That can be so without it following that our evaluations of cars, friends, people's characters, and so on are strictly objective.

But let's assume I'm wrong about science. Let's say, for the sake of argument, that my scientific epistemology is naive and false. In that case, perhaps science is also not strictly objective ... well so what? Perhaps that's an unwelcome conclusion. So?

You can't say that a metaethical view is false merely because a similar, but independently motivated, argument may (or may not) lead to an unwelcome conclusion in the field of scientific epistemology. The argument that moral objectivism is required on pain of undermining the authority of science is simplistic, illogical, and cuts no ice with me.

23 comments:

Alex SL said...

As long as nobody take up Sokal's offer to demonstrate that the laws of physics are social conventions by jumping out of a high window, and succeeds, I am not unduly worried about the objectivity of science.

Russell Blackford said...

lol, good point.

Thalamus said...

It was I who commented on your Facebook. This is what I said next:

Sam's analogy is indeed a recurring motif in these conversations, but this is not an accident. If science -being the quintessential instance of objective rationality- turns out not to be really objective, then no kind of objectivism could ever be conceivable.

As I have commented previously on your blog, this reminds me of Harris' remark claiming that religious moderates, in trying to simultaneously ingratiate fundamentalists and scientists, end up betraying both reason and faith all at once.
I think that in your laborious attempt to find a middle ground between what you call 'crude' relativism and moral realism, you've managed to obfuscate the basic principles of both, which in turn detracts from the clarity of your thesis.

I'm really interested in the kind of argument you would advance to convince a run-off-the-mill relativist that his/her views on morality are wrong.

DEEN said...

I would even use the argument the other way.

Science doesn't provide purely objective, absolute truths about nature either. And yet, science is eminently useful, and the best method we have found so far for understanding nature. It works, not because it converges to an absolute truth, but because it progresses away from ignorance. Most liberal believers would likely agree with this, and would not have much of a problem with it.

However, many people still want a moral system to yield purely objective, absolute truths. But that might just be unrealistic. We may have to accept that a moral system may be no more objective or absolute as science is. On the other hand, we also may find that such a moral system could be just as useful as science is. Not because it converges on an absolute morality, but because it progresses away from immorality.

Charles Sullivan said...

Could you explain "slippage" a bit more, Please? You've used this term a few times in relation to moral and non-moral evaluations.

Russell Blackford said...

Thalamus, what I find difficult about this is the implication in your question that I am engaging in some kind of contrivance or putting an argument in bad faith. Or that I am being "nonchalant" in some sense.

Whatever else I'm doing, I'm not engaging in any of this in bad faith and the arguments are not contrived for some ulterior purpose. As for being "nonchalant", that's certainly not something anyone who knows me would accuse me of. I give myself a great deal of anxiety by being, if anything, too conscientious.

The question about why a "run-of-the-mill" relativist is getting wrong is fair enough. I suppose I could write a blog post about that. I've given lectures on it in the past, to first-year university students, and have produced some consternation for some of them who've spoken to me afterwards. A lot of students aged about 17 or 18 seem to go into university with some kind of vague but rather crude moral relativism as their default position. They are then shocked when confronted with all the weaknesses of that position.

But the moral relativist positions argued by reputable contemporary philosophers - David Wong, Jesse Prinz, Gilbert Harman, and others - are far more sophisticated. They are not open to such devastating attacks.

That doesn't mean they're correct, of course. I think that they don't give enough weight to the arguments of error theorists like Mackie and Joyce. On the other hand, I also think they say enough to leave the line a bit blurred. Joyce, in particular, takes a line on the ubiquity of error that I think probably does go too far. I.e., I think that the semantics of moral discourse is more complicated than Joyce does.

I'm conscious that I keep saying "I think". A lot of this is complex and some tentativeness is justified ... or so I think. :)

On "slippage", this is really a metaphor that is not much use if it's not enlightening. So ignore it if it doesn't help. But again, if you think of ordinary evaluative discourse as about whether certain things have features that enable to meet "our" requirements in certain things, you'll notice is that, strictly speaking, there is no "we". In some cases that won't matter - only a very few outliers will depart from our main requirements for a hammer. And that's not surprising because hammers were invented and are constructed for a quite specific, simple, easily-definable purpose.

In other cases, though, the notion of "we" and "our" is much more a fiction. It's not totally inappropriate to talk this way because we really do have something in common in what we want from a motor car, a novel, or a friend. But nor is it surprising that different people want rather different things from, say, novels, with the result the Booker Prize's jury can never bind the rest of us when it says that Midnight's Children was the "best" novel of the first few decades of the prize.

It's not just that it's very difficult sorting out whether Midnight's Children wasa better than, say, Oscar and Lucinda, as if this is just a complex measurement to be made. It's that the jury is using certain (largely tacit) criteria that other people are not forced, on pain of being simply wrong, to adopt as their own.

None of which detracts from the expertise of the Booker Prize jury. They do, in fact, have to take into account a lot of complicated things that require expertise in how novels work, how they relate to literary traditions, and similar issues that literary critics know a lot about. I don't think the practice of literary criticism is simply a sham any more than I think the practice of moral judgment is simply a sham.

That Guy Montag said...

Cheers for the chance to debate this Russell because the point you're raising is exactly the one that leads to my seeing morality as objective.

The basic problem arises from Neurath's Theory Ladenness of Observation. A simple way to put that would be I need to have a theory of tables in order to perceive what I do as a table. A more stark example would be a radio telescope. It's interesting that for instance pulsars were discovered because of a regularity in a radio frequency, not because we literally saw them. In order for that kind of an observation to lead to the conclusion that there are such things as pulsars you need to have a theory about what radio waves are and how they get emitted and received.

The obvious way around this problem seems to be the one you've hinted at, that it's possible to appeal to another concept in order to fix any problems. If we assume that truth is unambiguous for instance then it's not hard to see how truth can be used to fix any problems in our theory. Generally I think this is the right response but I can see why that's the kind of claim that needs a lot more of an argument to establish. What I'm pretty sure about is that science has this kind of a response available to it: if you don't like truth, you could always try reliability or some kind of theory of perception. So, seeing as I think there's an answer to this problem for Science, why can't I use the same answer for Ethics? The point is that I'm not degrading science "down" to the level of ethics, I'm in fact raising ethics to the level of science.

The only other puzzle will be the Humean one which I think is a slightly separate problem but one that I think is resolvable by the same means. If I conceive of moral motivation as moral perception, what I need is some kind of normative concept for when my motivation counts as perception. For disclosure I'm only slowly starting down this particular line of thinking but as I understand it John McDowell is saying a lot of the same things. I also think there's some parallels with Intuitionism but as I say, I'm still tentative on this.

Anonymous said...

Russell,

I think these arguments aren't meant to challenge your definition of "strictly objective", and from there to challenge your requirements for what an objective morality has to provide.

The initial starting point of the argument is that since pretty much everyone thinks that science is objective, using a definition of "strictly objective" that eliminates it just seems odd, and casts doubt that you've captured what the concept of objective really means. And much of the time, when presented with this counter people will revisit their definition of "strictly objective".

Some will bite the bullet and accept that, as you almost do here. But then the counter comes in that your definition of objective is not, in fact, what they MEAN when they use the term. If you're going to conclude that science is not "strictly objective", then I think it's clear that Sam Harris, for example, is going to reply that he never suggested that morality was "strictly objective" by your definition; he wants morality to be AS objective as science. If it can get that, he's happy. And a lot of objectivists will have similar objections, that your definition doesn't capture what they meant when they used the term, even if your definition is, strictly speaking, correct.

All of these lead to the real objection: that your standards for objectivity are stronger than they ever claimed to be able to or that they wanted to provide.

Axxyaan said...

I think bringing up science in this context is a category error.

Whether or not science itself is strictly objective is not the point. What is the point is that science tries to get objective results about what it examines.

But what exactly is examined may be a totally unobjective choice.

But that still means that if you want scientific answers about a subject, that subject needs to be objective. So if values are subjective, science can't produce them as answers.

David said...

"I fully agree that the things we evaluate have objective characteristics - but the point is that there's going to be some slippage in what we require of these things (how do their objective characteristics conduce to meeting our desires, purposes, and so on?)."

This is actually a key point to the constant evolution of scientific techniques with respect to study design. We use a variety of study designs, depending on the subject of study and intended outcomes, with recognised and unrecognised strengths and weaknesses. Combine this with the statistical reality of whatever a particular study (well designed) is actually presenting if a statistically significant effect or difference is observed.

"Eventually, however, evaluations are made relative to desires, etc., and in that sense have a subjective element."

By this do you mean, we apply a value or judgement to the observed difference or effect?

For example, I perform a study showing people of a lower socioeconomic status group are less likely to access a health service despite a higher rate of affliction.

Now, it is the application of values external to the scientific method which indicate what should be done about this observation, if anything at all (such as principles of equity and opportunity).

(This example is pretty common in my field of health services research and epidemiology)

Is that what you meant? If so I'm completely on board, otherwise I'm not sure I follow.

Ray said...

There are two issues here: objective truth of scientific/moral statements and objective bindingness of scientific/moral statements.

The analogy between scientific and moral statements with respect to truth is that neither can be declared "objectively" true or false unless the disputants agree on terminology. Resolving moral disputes requires an agreement on the exact meaning of "moral" and "immoral." Likewise resolving scientific disputes requires agreement on the meanings of such terms as "force", "position", and "time." Now in practice, people are much more likely to agree upon the meaning of the latter than the former, but even then there is always some ambiguity (this is a consequence of Godel's theorem, but a practical example from science arises in the interpretation of quantum mechanics, where a lot of the dispute is over the definition of the word "real.")

Then, there's objective bindingness. It is certainly true that it is not objectively binding on all rational agents to *act* morally. But, it is not so clear that it isn't objectively binding on all rational agents to *believe* all true moral statements (once the appropriate definitions have been sufficiently clarified to make said statements true.)Of course you may say, as common language often seems to say, that you do not truly believe moral statements unless you act in accordance with them. But, even if you accept that, science has a similar problem. What does it really mean to believe a scientific statement if you act the same in every conceivable way as someone who does not believe the statement, including passing a lie detector test when denying it?

Tony Newell said...

I happen to think the health/well-being analogy Harris uses to blur the distinction between science and morality is a bit weak, but not because neither is truly objective. Clearly, the objectivity of science is, if one works hard at it, questionable. Heck, propositional statements about motion such as, “that train is moving,” are clearly subjective. But, as many, including Mr. Blackford, have suggested… so what? If a rational consensus about fundamental moral judgments can be achieved, then an objective-as-possible statement about morality can be just as useful as a statement about motion. Anyone who crosses a street, or rides an elevator is clearly satisfied with the objectivity of statements about motion. So it’s relative objectivity, not absolute objectivity, that matters.

The problem, and indeed where Harris’s health/well-being analogy falls short, for me, is more a matter of practicality. In health, we have an objective zero-value point (excuse me for perverting another discipline’s term of art). In other words, with health we have an undisputable antithesis, i.e. death. There is no argument that I’m aware of that suggest there is anything less healthy than death. One may exist in a state of utter and constant agony, but so long as one remains alive, one remains healthier than if he were dead. This is important, I think, because it at the very least caps one end of the scale. Without being able to do is for well-being (there is no objective zero-value point) then the degree of subjectivity entailed in any statement about well-being is immeasurably greater. This is an understated weakness in approaching well-being under the pretenses of objectivity.

There is also a problem with fact that health is an independent condition. My health is not dependent in any way on yours. The same cannot be said for well-being. But this is a different problem for a different time.

I think Harris is right to advocate an objective approach to morality. We have little to gain in worrying about weaknesses that exist in only the most extreme, absolute sense. The important problems people like Harris need to address involve practicality.

Tony Newell said...

I happen to think the health/well-being analogy Harris uses to blur the distinction between science and morality is a bit weak, but not because neither is truly objective. Clearly, the objectivity of science is, if one works hard at it, questionable. Heck, propositional statements about motion such as, “that train is moving,” are clearly subjective. But, as many, including Mr. Blackford, have suggested… so what? If a rational consensus about fundamental moral judgments can be achieved, then an objective-as-possible statement about morality can be just as useful as a statement about motion. Anyone who crosses a street, or rides an elevator, is clearly satisfied with the objectivity of statements about motion. So it’s relative objectivity, not absolute objectivity, that matters.

The problem, and indeed where Harris’s health/well-being analogy falls short, for me, is more a matter of practicality. In health we have an objective zero-value point (excuse me for perverting another discipline’s term of art). In other words, with health we have an undisputable antithesis, i.e. death. There is no argument that I’m aware of that suggests there is anything less healthy than death. One may exist in a state of utter and constant agony, but so long as one remains alive, one remains healthier than if he were dead. This is important, I think, because it at the very least caps one end of the scale. Without being able to do is for well-being (there is no objective zero-value point) then the degree of subjectivity entailed in any statement about well-being is immeasurably greater. This is an understated weakness in approaching well-being under the pretense of objectivity.

There is also a problem with fact that health is an independent condition. My health is not dependent in any way on yours. The same cannot be said for well-being. But this is a different problem for a different time.

I think Harris is right to advocate an objective approach to morality. We have little to gain in worrying about weaknesses that exist in only the most extreme, absolute sense. The important problems people like Harris need to address involve practicality.

Kirth Gersen said...

If I may, Russell, you are almost never contrived or nonchalant, and you are one of very few people that I never get a sense of bad faith from. You do, however, tend to approach things almost too-diligently from the philosophical end rather than from the pragmatic side of things -- carefully creating several layers of intellectual latticework between the principle in theory and the principle in action -- and that sometimes can potentially lead to impatience in more direct-minded readers.

As Alex SL points out, if something works the same way for everybody, that's a good enough working definition of "objective," even if it fails on some philosophical level.

Thalamus said...

I find interesting and rather bizarre that the default position on morality among first year university students be moral relativism, especially when we know that their default position is also theism. It's just somewhat paradoxical and I think it sadly speaks to the pervasive hypocrisy/ignorance of these students. Hopefully this issue will be introduced into the fray of our blogging discussions sometime in the future.

Now, I never meant to accuse you of being deliberately dishonest. I think I know enough about you to at least presume that your a sincere and principled fellow. I actually agree with you on most issues. I simply have not been able to discern the shades of gray on this debate on morality. I think an issue that must be adjudicated in absolute terms: either we see morality as an objective natural phenomenon consisting of facts about the brain and therefore amenable to scientific study, or we see as merely the product of our own subjective and often quixotic idiosyncrasies; something like aesthetics. By my lights, any attempt to blur the line delineating this dichotomy would eventuate in an intellectual cul de sac.

I suppose your position lacks tone. Maybe a direct question will prompt a direct answer from you. The manner in which you respond to this will dispel many of my doubts.

Do you think that the Muslim practice of genital excision is morally justified? And if it is not, do you think their poor moral judgment is due to a lack of facts about the world, i.e., to an abject ignorance about the biological and psychological nature of suffering, or it's emotions which are the ink coloring their views?

pboyfloyd said...

"If science -being the quintessential instance of objective rationality-..."

And, of course, "If a wood-chuck, could chuck wood.."

I believe that Mr. Russell is trying to remove this very muddy reasoning which you seem desperate to reinsert.

If he removes the, "If this then that.", and replaces it with, "It doesn't matter much if it's this or that because the conclusion of the continued argument is dependent on an unnecessary premise which colours the first argument.

Marshall said...

Jeremiah said on the other thread "I think that people only retreat to the position of requiring some sort of hard objectivism when they are challenged on some belief that they happen to be passionate about."

Which I think is pretty right, morality is prior to rationality. People want their "moral beliefs", their sense of right and wrong, to be a real stake in the ground, unquestionable. In the case of people who like Sam Harris, John Searle, and John Rawls, that means that when challenged they go in the direction of "how things are", aka Objectivity. You could argue that the Theists are more honest, since they go directly to Faith in their own righteousness.

I think the threat to science is just as much a red herring as the theistic worry about the threat of science. What is really bothering people is all this postmodern loss of certainty.

Russell Blackford said...

Thalamus, the default position of the students I'm talking about is not theism. Most of them think theism is pretty much bullshit.

I live in Australia, remember. Theism is not the default position here in the way that it is in the US.

But, yes, it would be interesting to teach a comparable group of US students.

Russell Blackford said...

Oh, and of course I don't think the practice of female genital mutilation is morally justified. But that doesn't take things a lot further. I might think that because I'm a moral error theorist and I think nothing is "morally justified".

I realise that that's a cute response, but the point is that there really are these traps to watch out for when we throw around terms like "morally justified". What do those terms actually mean? Do they, strictly speaking, mean something coherent?

As for whether I think female genital mutilation is a good practice or a bad practice, I think it's a bad practice. Very much so.

bad Jim said...

Science works well when the subject of our inquiries is unaware of our investigations. An electron doesn't care what we think about it, and we can characterize its behavior with exquisite precision. The same is not true about another person, a business or a state.

To a first approximation, a human observer and an inanimate recorder will record certain sorts of events identically and reliably, and other sorts of events otherwise. To the extent that we can characterize the events of the first sort in relatively simple terms, we call such characterizations "science". We also label our efforts to describe the second sort using similar means as scientific, but the results so far have not proven equally reliable.

That Guy Montag said...

Just another point that struck me: as heated as this discussion about Metaethics gets, as far as I can tell nobody disagreed on what the practical implications are which is really surprisingly little as far as I can see. In fact the most stark consequence that I can conceive of from all this debate is it removes what I think is an illegitimate barrier on discussion that places ethics as somehow being outside the scope of ordinary human inquiry. Realism is one way to do this but so too is Russell's nuanced Subjectivism.

Jeremiah said...

@ Thalamus
Well my two cents on this notion that it must be either one or the other is as follows. 'Right' and 'wrong' clearly do have meaning to us as social animals in explaining a somewhat broad but similar class of behaviors and desires which Russell explained well in his recent article I think. I don't think this middle ground betrays the principles of moral relativists in that it agrees that there is no external objective standard to measure morals against but disagrees with relativists in that this fact means we can't have meaningful conversations about what human beings consider moral behavior and acting to fulfill the desires that drive our morals. On the other side it agrees with realists that there is a sort of moral knowledge out there to be discussed, made manifest by our desires, but disagrees with them that there is anything that you could consider an objectively 'right' answer or 'true'. Things can be true in a generalized consensus sort of way, like Russell's example of a 'good car' but not in the hard and fast way such as the answer to 3+3 is 6.

So I would say that morality is subjective, but not arbitrary, and because it is not arbitrary we can have meaningful conversations about it without being confined to a position of having no moral authority at all which some relativists adhere to.

That Guy Montag said...

Jeremiah:

How is that necessarily subjectivist and not merely particularist? "Aim for answers that are appropriate to the question" is a pretty obvious principle and one that's clearly used in science in say distinguishing between biology and astonomy. In fact it appear to have a far broader effect: I might be tempted to suggest it's a necessary feature of language more generally that it has to be appropriate for its subject. So apart from a particularly stark reductionist realism it can't really be an argument against objectivism that we use terms like good in a nuanced way.