If you’re looking for a test of a moral claim that does not rely on any background assumptions about morality, then I agree that I can’t give you an example. But that’s because your standard is way too high. When we test scientific hypotheses, the evidence is always interpreted in the context of background assumptions. If it’s kosher for scientific experiments to assume certain scientific facts (as they must), then why isn’t it kosher for moral experiments to assume certain moral facts?
Consider the analog of your position in the descriptive case: someone denies that there’s any fact of the matter about whether descriptive claims about the external world are true or false. This person says, “Show me how you’d test whether a descriptive claim is true or false.” Now you could presumably give all sorts of examples of such tests, but all of these examples will assume the truth of a host of other descriptive claims (minimally, that the experimental apparatus actually exists and the whole experiment isn’t a hallucination). If your interlocutor insisted that you give an example of a test that does not itself assume the truth of any descriptive claim, you would not be able to satisfy him.
So why demand that I must give an example of a test of a moral claim that does not assume the truth of any other moral claim? Why does the moral realist have this extra justificatory burden that the scientific realist does not? It’s fine to have problems with the specific assumptions being made in any particular experiment, and these can be discussed. Perhaps you think my particular assumptions are flawed for some reason. But if you have a general worry about all moral assumptions then you need to tell me why you don’t have a similar worry about non-normative assumptions. If you have a principled reason for this distinction, that would be the real basis of your moral anti-realism, not this business about untestability.
I do think moral and scientific reasoning are far less asymmetric than is usually assumed. But that doesn’t mean I think there are no asymmetries at all. Asymmetries exist, and perhaps they can be leveraged into an argument for moral anti-realism that is not also an argument against scientific realism. So I wouldn’t say it’s inconsistent to be a physical realist and a moral anti-realist. I will say that in my experience most people who hold that combination of positions will, upon interrogation, reveal an unjustified (but not necessarily unjustifiable) double standard in the way they treat moral discourse.
I don’t think it is a double standard. Empiricism admits the Problem of Induction, but says that the problem doesn’t justify retreating all the way to Cartesian skepticism. This position is supported by the fact that science makes good predictions—I would find the regularity of my sensory experiences surprising if physical realism were false. Plus, the principle of falsification (i.e. making beliefs pay rent) tells us what sorts of statements are worth paying attention to.
Moral reasoning seems to lack ay equivalent for either falsification or prediction. I don’t know what it means to try to falsify a statement like “Killings in these circumstances are not morally permissible.” And to the extent that predictions can be made based on the statement, they seem either false or historically contingent—it’s pretty easy to imagine my society having different rules about what killings are morally permissible simply by looking at how a nearby society came to its different conclusions.
In short, the problem of induction in empiricism seems very parallel to the is/ought problem in moral philosophy. But moral philosophy seems to lack the equivalent of practical arguments like accurate prediction that seem to rescue empiricism.
I do think one can offer a pragmatic justification for moral reasoning. It won’t be exactly parallel to the justification of scientific reasoning because moral and scientific discourse aren’t in the same business. Part of the double standard I was talking about involves applying scientific standards of evaluation to determine the success of moral reasoning. This is as much of an error as claiming that relativity is false because the nuclear bomb caused so much suffering. We don’t engage in moral reasoning in order to make accurate predictions about sensory experience. We engage in moral reasoning in order to direct action in such a way that our social environment becomes a better place. And I do think we have plenty of historical evidence that our particular system of moral reasoning has contributed to making the world a better place, just as our particular system of scientific reasoning has contributed to our increasing ability to control and predict the behavior of the world.
Now obviously there’s a circularity here. Our standards for judging that the world is better now that slavery is illegal and women can vote are internal to the very moral discourse we purport to be evaluating. But this kind of ultimate circularity is unavoidable when we attempt to justify any system of justification as a whole. It’s precisely the problem Hume pointed out when he talked about induction. Sure, we can appeal to past success as a justification of our inductive practices, but that justification only works if we are already committed to induction. Furthermore, our belief in the past success of the scientific method is based on historical data collected and interpreted in accord with this method. Somebody who rejects the scientific method wholesale may well say “Why should I believe any of these historical claims you are making?”
A completely transcendental justification, one that would be normative to any possible mind in mindspace, is an impossible goal in both moral and scientific reasoning. Any justification you offer for your justificatory practices is ultimately going to appeal to standards that are internal to those practices. That’s something we’ve all learned to live with in science, but there’s still a resistance to this unfortunate fact when it comes to moral discourse.
And to the extent that predictions can be made based on the statement, they seem either false or historically contingent—it’s pretty easy to imagine my society having different rules about what killings are morally permissible simply by looking at how a nearby society came to its different conclusions.
Our scientific schemes of justification are historically contingent in the same way. There are a number of other communities (extremely religious ones, for instance) that employ a different set of tools for justifying descriptive claims about the universe. Of course, our schemes of justification are better than theirs, as evidenced by their comparative lack of technological and predictive success. By the same token, though, our moral schemes of justification are more successful than those of, say, fundamentalist Islamic societies, as evidenced by our greater degree of moral progress. In both cases, the members of those other societies would disagree that we have done better than them, but that’s because they have different (and I would say incorrect) standards of evaluation.
Empiricism has only one circularly justified position: You can (more or less) trust the input of your senses—which implies some consistency over time. Everything else follows from that. Modern science is better than pytolemic science because it makes better predictions.
By contrast, there’s essentially no limit to moral circularity. There’s the realism premise: There is a part of the territory called “moral rightness”. Then you need a circular argument to show any particular moral premise (these killings are unjustified) is part of moral rightness. And there are multiple independent moral premises. (When killing is wrong does not shed much light on when lying is wrong). It’s not even clear that there are a finite number of circularly justified assertions.
So I hold empiricism to the same standard as moral realism, and moral realism seems to come up short. Further, my Minimization of Circular Justification principle is justified by worry about the ease of creating a result simply by making in an axiom. (That is, the Pythagorean Theorem is on a different footing if it is introduced as an axiom of Euclidean geometry rather than a derived result).
If your principle is actually that circular justification must be minimized, then why aren’t you an anti-realist about both scientific and moral claims? Surely that would involve less circular justification than your current position. You wouldn’t even have to commit yourself to the one circularly justified position assumed by empiricism.
In any case, scientific reasoning as a whole does not just reduce to the sort of minimal empiricism you describe. For starters, even if you assume that the input of your senses is trustworthy and will continue to remain trustworthy, this does not establish that induction based on the input of your senses is trustworthy. This is a separate assumption you must make. Your minimal empiricism also does not establish that simpler explanations of data tend to be better. This is a third assumption. It also doesn’t establish what it means for one explanation to be simpler than another. It doesn’t establish that the axioms on which the mathematical and statistical tools of science are based are true. I could go on.
Scientific justification as it’s actually practiced in the lab involves a huge suite of tools, and it is not true that the reliability of all these tools can be derived once you accept that you can trust the input of your senses. A person can be an empiricist in your sense while denying the reliability of statistical methods used in science, for instance. To convince them otherwise you will presumably present data that you think establishes the reliability of those methods. But in order for the data to deliver this conclusion, you need to use the same sorts of statistical methods that the skeptic is rejecting. I don’t see how your shared empiricism helps in this situation.
Our schemes of justification, both scientific and moral, have developed through a prolonged process of evolutionary and historical accretion. The specific historical reasons underlying the acceptance of particular tools into the toolbox are complex and variegated. It is implausible in either case that we could reconstruct the entire scheme from one or two simple assumptions.
If you’d like to separate the axiom about the reliability of the senses from the axiom that sensory input will remain consistent, I won’t actively resist—I think reliability of the senses implies consistency of the sense, but I’m not certain my formulation is more technically correct.
Regarding Ockham’s Razor—I’m not sure that is a fundamental principle or a useful rule of thumb. If MWI and Copenhagen really are in evidentiary equipose, I’m not sure I should have a preference for one or the other (that’s obviously not the consensus position in this community).
It doesn’t establish that the axioms on which the mathematical and statistical tools of science are based are true.
I think deductive reasoning produces necessary truths—so in a sense, I get statistics “for free” as long as I accept the Peano axioms. Other than that, I don’t understand the quoted assertion.
More generally, empirical philosophy provides a place to stop the recursion. I don’t think circular justifications work at all, so I think a separate justification for using this stopping place is required—I have memory of consistent sensory impressions, and that is difficult to explain except by believing that consistency is true. One could object that I can’t justify reliance on my memory—so I’m being hypocritical to allow my memories to justify themselves. Maybe so, but there’s no other principled stopping place for recursion—and continuing recursion past this point devolves to the point that I don’t think coherence is a workable concept.
To return to the comparison with morality, I suggest that all the axiomatic assertions in the empirical program are at a fundamental level. When you start doing object level science, recursion goes away entirely. By contrast, object level morality never gets away from [EDIT: recursion]. As you noted, it is impossible to say whether we’ve made moral progress without referencing what moral position is better.
If progress (scientific, moral, etc) really is possible, we ought to be able to get away from recursive reasoning. That we can’t when dealing with moral reasoning is not a good sign that moral reasoning is talking about some objective fact.
When you start doing object level science, recursion goes away entirely. By contrast, object level morality never gets away from morality. As you noted, it is impossible to say whether we’ve made moral progress without referencing what moral position is better.
I don’t know what you mean by “object level morality never gets away from morality”. Read literally, that’s tautologically true, but I don’t see the relevance. Is this a typo?
Also, I’m not seeing the distinction here. When I’m engaged in object-level moral reasoning, or when I read examples of object-level moral reasoning on blogs or in newspapers, I very rarely come across recursion or circular justification. There’s usually an assumption that everyone in the community agrees that certain sorts of fundamental moral inferences are justified, and the debate is about whether those inferences can be made in a particular case. Here is a classic example of object-level moral reasoning. MLK offers a number of justifications for his moral stance on this particular issue. None of these justifications, as far as I can see, are circular. I don’t think this is atypical. Of course, if you think that every moral argument must also simultaneously justify the whole enterprise of objective moral evaluation, then every moral argument will have a circular component. But this places a disproportionately large burden on moral justification.
It’s true that if I want to argue that we have made moral progress I need to take for granted certain moral standards of evaluation, but if I want to argue that we have made scientific progress I need to take for granted certain scientific standards of evaluation. The only difference I can see is that the moral assumptions are as a matter of fact more contentious than the scientific ones, so perhaps moral debate breaks down on disagreement about foundational assumptions more often. But this is at least partly because most scientific debate is usually conducted in an institutional setting that has various mechanisms for consensus formation and weeding out sufficiently recalcitrant dissenters. Outside this setting, debate about descriptive issues is often just as contentious as moral debate. I know a number of new-agey people who have completely bizarre standards of epistemic justification. My discussion with them quite often breaks down on disagreement about foundational assumptions.
There’s usually an assumption that everyone in the community agrees that certain sorts of fundamental moral inferences are justified, and the debate is about whether those inferences can be made in a particular case.
That’s not my sense at all. Moral inferences are fairly easy (compared to cutting-edge scientific inferences). Toy example: If God wants us to attend church, the inference that church attendance should be compelled by the government follows quite easy. There are secondary negative effects, but the only reason to care about them is if the moral assertion that God wants church attendance is false.
When I read political arguments, they almost always operate by assuming agreement on the moral premise. When that assumption is falsified, the argument falls apart. Even for fairly ordinary moral disputes, the argument is usually based on moral principle, not facts or moral inference.
By contrast, equivalently basic scientific questions are fact and inference based. To decide how much weight a bridge can carry, knowing the strength of the steel and the design of the bridge is most of the work. In practice, those types of disputes don’t devolve into arguments about whether gravity is going to work this time.
There are secondary negative effects, but the only reason to care about them is if the moral assertion that God wants church attendance is false.
Unless the secondary effects were that people are more likely to eat bacon for breakfast that day now that they aren’t able to sleep in and it also happens that God doesn’t want people to eat pigs.
If you’re looking for a test of a moral claim that does not rely on any background assumptions about morality, then I agree that I can’t give you an example. But that’s because your standard is way too high. When we test scientific hypotheses, the evidence is always interpreted in the context of background assumptions. If it’s kosher for scientific experiments to assume certain scientific facts (as they must), then why isn’t it kosher for moral experiments to assume certain moral facts?
Consider the analog of your position in the descriptive case: someone denies that there’s any fact of the matter about whether descriptive claims about the external world are true or false. This person says, “Show me how you’d test whether a descriptive claim is true or false.” Now you could presumably give all sorts of examples of such tests, but all of these examples will assume the truth of a host of other descriptive claims (minimally, that the experimental apparatus actually exists and the whole experiment isn’t a hallucination). If your interlocutor insisted that you give an example of a test that does not itself assume the truth of any descriptive claim, you would not be able to satisfy him.
So why demand that I must give an example of a test of a moral claim that does not assume the truth of any other moral claim? Why does the moral realist have this extra justificatory burden that the scientific realist does not? It’s fine to have problems with the specific assumptions being made in any particular experiment, and these can be discussed. Perhaps you think my particular assumptions are flawed for some reason. But if you have a general worry about all moral assumptions then you need to tell me why you don’t have a similar worry about non-normative assumptions. If you have a principled reason for this distinction, that would be the real basis of your moral anti-realism, not this business about untestability.
You position suggests that one cannot consistently be a physical realist and a moral anti-realist. Is that a fair summary of your position?
I do think moral and scientific reasoning are far less asymmetric than is usually assumed. But that doesn’t mean I think there are no asymmetries at all. Asymmetries exist, and perhaps they can be leveraged into an argument for moral anti-realism that is not also an argument against scientific realism. So I wouldn’t say it’s inconsistent to be a physical realist and a moral anti-realist. I will say that in my experience most people who hold that combination of positions will, upon interrogation, reveal an unjustified (but not necessarily unjustifiable) double standard in the way they treat moral discourse.
I don’t think it is a double standard. Empiricism admits the Problem of Induction, but says that the problem doesn’t justify retreating all the way to Cartesian skepticism. This position is supported by the fact that science makes good predictions—I would find the regularity of my sensory experiences surprising if physical realism were false. Plus, the principle of falsification (i.e. making beliefs pay rent) tells us what sorts of statements are worth paying attention to.
Moral reasoning seems to lack ay equivalent for either falsification or prediction. I don’t know what it means to try to falsify a statement like “Killings in these circumstances are not morally permissible.” And to the extent that predictions can be made based on the statement, they seem either false or historically contingent—it’s pretty easy to imagine my society having different rules about what killings are morally permissible simply by looking at how a nearby society came to its different conclusions.
In short, the problem of induction in empiricism seems very parallel to the is/ought problem in moral philosophy. But moral philosophy seems to lack the equivalent of practical arguments like accurate prediction that seem to rescue empiricism.
I do think one can offer a pragmatic justification for moral reasoning. It won’t be exactly parallel to the justification of scientific reasoning because moral and scientific discourse aren’t in the same business. Part of the double standard I was talking about involves applying scientific standards of evaluation to determine the success of moral reasoning. This is as much of an error as claiming that relativity is false because the nuclear bomb caused so much suffering. We don’t engage in moral reasoning in order to make accurate predictions about sensory experience. We engage in moral reasoning in order to direct action in such a way that our social environment becomes a better place. And I do think we have plenty of historical evidence that our particular system of moral reasoning has contributed to making the world a better place, just as our particular system of scientific reasoning has contributed to our increasing ability to control and predict the behavior of the world.
Now obviously there’s a circularity here. Our standards for judging that the world is better now that slavery is illegal and women can vote are internal to the very moral discourse we purport to be evaluating. But this kind of ultimate circularity is unavoidable when we attempt to justify any system of justification as a whole. It’s precisely the problem Hume pointed out when he talked about induction. Sure, we can appeal to past success as a justification of our inductive practices, but that justification only works if we are already committed to induction. Furthermore, our belief in the past success of the scientific method is based on historical data collected and interpreted in accord with this method. Somebody who rejects the scientific method wholesale may well say “Why should I believe any of these historical claims you are making?”
A completely transcendental justification, one that would be normative to any possible mind in mindspace, is an impossible goal in both moral and scientific reasoning. Any justification you offer for your justificatory practices is ultimately going to appeal to standards that are internal to those practices. That’s something we’ve all learned to live with in science, but there’s still a resistance to this unfortunate fact when it comes to moral discourse.
Our scientific schemes of justification are historically contingent in the same way. There are a number of other communities (extremely religious ones, for instance) that employ a different set of tools for justifying descriptive claims about the universe. Of course, our schemes of justification are better than theirs, as evidenced by their comparative lack of technological and predictive success. By the same token, though, our moral schemes of justification are more successful than those of, say, fundamentalist Islamic societies, as evidenced by our greater degree of moral progress. In both cases, the members of those other societies would disagree that we have done better than them, but that’s because they have different (and I would say incorrect) standards of evaluation.
Yes, there’s inherently a certain amount of unsatisfying circularity is everything. But that’s a weakness that calls for minimization of circularity.
Empiricism has only one circularly justified position: You can (more or less) trust the input of your senses—which implies some consistency over time. Everything else follows from that. Modern science is better than pytolemic science because it makes better predictions.
By contrast, there’s essentially no limit to moral circularity. There’s the realism premise: There is a part of the territory called “moral rightness”. Then you need a circular argument to show any particular moral premise (these killings are unjustified) is part of moral rightness. And there are multiple independent moral premises. (When killing is wrong does not shed much light on when lying is wrong). It’s not even clear that there are a finite number of circularly justified assertions.
So I hold empiricism to the same standard as moral realism, and moral realism seems to come up short. Further, my Minimization of Circular Justification principle is justified by worry about the ease of creating a result simply by making in an axiom. (That is, the Pythagorean Theorem is on a different footing if it is introduced as an axiom of Euclidean geometry rather than a derived result).
If your principle is actually that circular justification must be minimized, then why aren’t you an anti-realist about both scientific and moral claims? Surely that would involve less circular justification than your current position. You wouldn’t even have to commit yourself to the one circularly justified position assumed by empiricism.
In any case, scientific reasoning as a whole does not just reduce to the sort of minimal empiricism you describe. For starters, even if you assume that the input of your senses is trustworthy and will continue to remain trustworthy, this does not establish that induction based on the input of your senses is trustworthy. This is a separate assumption you must make. Your minimal empiricism also does not establish that simpler explanations of data tend to be better. This is a third assumption. It also doesn’t establish what it means for one explanation to be simpler than another. It doesn’t establish that the axioms on which the mathematical and statistical tools of science are based are true. I could go on.
Scientific justification as it’s actually practiced in the lab involves a huge suite of tools, and it is not true that the reliability of all these tools can be derived once you accept that you can trust the input of your senses. A person can be an empiricist in your sense while denying the reliability of statistical methods used in science, for instance. To convince them otherwise you will presumably present data that you think establishes the reliability of those methods. But in order for the data to deliver this conclusion, you need to use the same sorts of statistical methods that the skeptic is rejecting. I don’t see how your shared empiricism helps in this situation.
Our schemes of justification, both scientific and moral, have developed through a prolonged process of evolutionary and historical accretion. The specific historical reasons underlying the acceptance of particular tools into the toolbox are complex and variegated. It is implausible in either case that we could reconstruct the entire scheme from one or two simple assumptions.
If you’d like to separate the axiom about the reliability of the senses from the axiom that sensory input will remain consistent, I won’t actively resist—I think reliability of the senses implies consistency of the sense, but I’m not certain my formulation is more technically correct.
Regarding Ockham’s Razor—I’m not sure that is a fundamental principle or a useful rule of thumb. If MWI and Copenhagen really are in evidentiary equipose, I’m not sure I should have a preference for one or the other (that’s obviously not the consensus position in this community).
I think deductive reasoning produces necessary truths—so in a sense, I get statistics “for free” as long as I accept the Peano axioms. Other than that, I don’t understand the quoted assertion.
More generally, empirical philosophy provides a place to stop the recursion. I don’t think circular justifications work at all, so I think a separate justification for using this stopping place is required—I have memory of consistent sensory impressions, and that is difficult to explain except by believing that consistency is true. One could object that I can’t justify reliance on my memory—so I’m being hypocritical to allow my memories to justify themselves. Maybe so, but there’s no other principled stopping place for recursion—and continuing recursion past this point devolves to the point that I don’t think coherence is a workable concept.
To return to the comparison with morality, I suggest that all the axiomatic assertions in the empirical program are at a fundamental level. When you start doing object level science, recursion goes away entirely. By contrast, object level morality never gets away from [EDIT: recursion]. As you noted, it is impossible to say whether we’ve made moral progress without referencing what moral position is better.
If progress (scientific, moral, etc) really is possible, we ought to be able to get away from recursive reasoning. That we can’t when dealing with moral reasoning is not a good sign that moral reasoning is talking about some objective fact.
I don’t know what you mean by “object level morality never gets away from morality”. Read literally, that’s tautologically true, but I don’t see the relevance. Is this a typo?
Also, I’m not seeing the distinction here. When I’m engaged in object-level moral reasoning, or when I read examples of object-level moral reasoning on blogs or in newspapers, I very rarely come across recursion or circular justification. There’s usually an assumption that everyone in the community agrees that certain sorts of fundamental moral inferences are justified, and the debate is about whether those inferences can be made in a particular case. Here is a classic example of object-level moral reasoning. MLK offers a number of justifications for his moral stance on this particular issue. None of these justifications, as far as I can see, are circular. I don’t think this is atypical. Of course, if you think that every moral argument must also simultaneously justify the whole enterprise of objective moral evaluation, then every moral argument will have a circular component. But this places a disproportionately large burden on moral justification.
It’s true that if I want to argue that we have made moral progress I need to take for granted certain moral standards of evaluation, but if I want to argue that we have made scientific progress I need to take for granted certain scientific standards of evaluation. The only difference I can see is that the moral assumptions are as a matter of fact more contentious than the scientific ones, so perhaps moral debate breaks down on disagreement about foundational assumptions more often. But this is at least partly because most scientific debate is usually conducted in an institutional setting that has various mechanisms for consensus formation and weeding out sufficiently recalcitrant dissenters. Outside this setting, debate about descriptive issues is often just as contentious as moral debate. I know a number of new-agey people who have completely bizarre standards of epistemic justification. My discussion with them quite often breaks down on disagreement about foundational assumptions.
Yes, typo corrected.
That’s not my sense at all. Moral inferences are fairly easy (compared to cutting-edge scientific inferences). Toy example: If God wants us to attend church, the inference that church attendance should be compelled by the government follows quite easy. There are secondary negative effects, but the only reason to care about them is if the moral assertion that God wants church attendance is false.
When I read political arguments, they almost always operate by assuming agreement on the moral premise. When that assumption is falsified, the argument falls apart. Even for fairly ordinary moral disputes, the argument is usually based on moral principle, not facts or moral inference.
By contrast, equivalently basic scientific questions are fact and inference based. To decide how much weight a bridge can carry, knowing the strength of the steel and the design of the bridge is most of the work. In practice, those types of disputes don’t devolve into arguments about whether gravity is going to work this time.
Unless the secondary effects were that people are more likely to eat bacon for breakfast that day now that they aren’t able to sleep in and it also happens that God doesn’t want people to eat pigs.