EPTThis thread is intended to give a primer on the most influential theories in contemporary analytic metaethics while also providing a place to discuss and debate disagreements and to voice opinions. It will also offer a framework for dividing up positions that can be used in many other debates, such as debates about whether artistic value is objective.
MORAL REALISM
There are many different ways of defining moral realism, but definitions will typically have something along the lines of the following three theses at their core:
(1) Ethical discourse is cognitive (or, fact-stating; or, truth-evaluable). That is, ethical assertions make claims about the world and are capable of being either true or false.[1]
(2) Some ethical statements are true.
(3) We know, or are capable of knowing, some ethical truths.
(2) Some ethical statements are true.
(3) We know, or are capable of knowing, some ethical truths.
Moral realism further divides into naturalistic and non-naturalistic variants.[2]
According to naturalistic moral realism, ethical facts simply are (or reduce to) a type of natural fact, not fundamentally different from facts about the properties of planets and animals.[3] Utilitarianism can be viewed as an example of this, since utilitarians identify ethical facts with naturalistic facts about what creates the most pleasure minus pain. You don't have to be a utilitarian to be a naturalist, however. Important proponents include Peter Railton, Richard Boyd, and Nick Sturgeon.
The most obvious kind of non-naturalistic moral realism is a divine command theory, wherein ethical facts are simply facts about what God wills.[4] However, the most prominent defenders of non-naturalistic MR are atheists such as Derek Parfit, Thomas Scanlon, and Tom Nagel. Their view doesn't commit them to any supernatural objects, just to the claim that reasons do not reduce to natural facts. For instance, they might say that it is wrong to torture babies for fun, but that this fact does not reduce to any natural facts.[5]
Dialectic:
+ Show Spoiler +
The most straightforward argument for each form of realism is that they preserve the appearances of moral discourses. More often than not, people seem to make moral utterances as though they endorsed (1)-(3), so these theses represent a starting point for discussion about morality.
Naturalistic realism can claim the further advantage that this benefit comes for no ontological cost. Their moral facts are nothing over and above the natural facts we already accept, so, unlike belief in a god, believing in them costs nothing in terms of Ockham's razor. Non-naturalists will typically counter that the natural facts are not enough to preserve the common understanding of morality, so if you want to do so you have to accept their view and it's additional costs.
The main argument's against realism are the arguments for the opposing views that I will outline below.[6]
Naturalistic realism can claim the further advantage that this benefit comes for no ontological cost. Their moral facts are nothing over and above the natural facts we already accept, so, unlike belief in a god, believing in them costs nothing in terms of Ockham's razor. Non-naturalists will typically counter that the natural facts are not enough to preserve the common understanding of morality, so if you want to do so you have to accept their view and it's additional costs.
The main argument's against realism are the arguments for the opposing views that I will outline below.[6]
EXPRESSIVISM
Expressivism is the contemporary version of non-cognitivism, the non-realist view you get when you deny thesis (1). That is, it claims that ethical statements are not truth-evaluable or fact stating.
The most simple version of this, emotivism, states that ethical statements are used to simply express emotions. So, 'Murder is wrong' would come out meaning the same thing as 'Boo murder!', which is not truth-evaluable. This view is associated with logical positivists such as A.J. Ayer and is widely regarded as refuted.
Expressivism brings in more complicated psychological states being expressed than simple emotions. Candidates for that psychological state include complex motivational states or commitments to complicated sets of plans. The most prominent recent expressivists have been Allen Gibbard and Simon Blackburn.
It's important to keep this view separate from the view that moral statements describe certain psychological states. On that view, 'Murder is wrong' might mean, 'I hate murder'. But notice that 'I hate murder' still has a truth value, which expressivists deny. That's why they say that your hatred is expressed, not described, by your moral utterances.
Dialectic:
+ Show Spoiler +
Pros
Traditionally, one of the most important arguments for non-cognitivism stems from G.E. Moore's open question argument, according to which it is impossible to define moral predicates using naturalistic properties. However, this argument is somewhat outdated since most varieties of realism do not attempt to give definitions anymore.
The other main argument for expressivism has been the supposedly essential connection between ethical attitudes/assertions and some form of endorsement or motivation. Expressivists will often claim that it is impossible to imagine someone thinking that stealing is wrong without also disapproving of stealing or being motivated in some part not to steal or to reduce the number of thefts (the details of the connection will vary with the specific account offered). It is easy to see why such a connection would exist if ethical attitudes just are states of endorsement or motivation, and ethical statements are just expressions of such states.
But such a connection between beliefs and motivation/endorsement would be hard to make sense of for a naturalistic moral realist, since it is hard to see why merely believing some fact could have motivational or endorsing effects independent of a persons contingent desires. Naturalists thus typically dispute the connection, whereas non-naturalists typically dispute that such action or endorsement guiding belief states are impossible.
Cons
One of the most important arguments against non-cognitivism in general has been the Frege-Geach problem. This is easiest to understand using emotivism, according to which 'Murder is wrong' just means 'Boo murder!'. The problem with this view is that it has a hard time making sense of a compositional semantics for moral terms. 'Murder is wrong' doesn't simply appear in straightforward assertions, but also in more complicated embedded contexts such as: 'Johnny believes murder is wrong' , 'If murder is wrong, Sally wouldn't do it', and 'What if murder is wrong?'. None of these sentences make sense if 'Murder is wrong' has no meaning beyond its expression of disapproval for murder, since that disapproval is in no way built into the more complicated sentences.
This objection is widely regarded as having refuted the most simple forms of non-cognitivism, and one of the main tasks of more recent work in the area has been trying to find ways around it. It is still a matter of debate whether any such proposal has succeeded.
Another argument against expressivism attempts to refute the claimed essential connection between endorsement/motivation and ethical beliefs. Many naturalistic realists claim that amoralists are conceivable. Amoralists are perfectly competent users of moral language and can answer most everyday questions about morality correctly, but they remain completely indifferent to moral considerations. They will happily agree with you that setting cats on fire is wrong, but ignore that factor when deciding whether they will set a cat on fire today.
Traditionally, one of the most important arguments for non-cognitivism stems from G.E. Moore's open question argument, according to which it is impossible to define moral predicates using naturalistic properties. However, this argument is somewhat outdated since most varieties of realism do not attempt to give definitions anymore.
The other main argument for expressivism has been the supposedly essential connection between ethical attitudes/assertions and some form of endorsement or motivation. Expressivists will often claim that it is impossible to imagine someone thinking that stealing is wrong without also disapproving of stealing or being motivated in some part not to steal or to reduce the number of thefts (the details of the connection will vary with the specific account offered). It is easy to see why such a connection would exist if ethical attitudes just are states of endorsement or motivation, and ethical statements are just expressions of such states.
But such a connection between beliefs and motivation/endorsement would be hard to make sense of for a naturalistic moral realist, since it is hard to see why merely believing some fact could have motivational or endorsing effects independent of a persons contingent desires. Naturalists thus typically dispute the connection, whereas non-naturalists typically dispute that such action or endorsement guiding belief states are impossible.
Cons
One of the most important arguments against non-cognitivism in general has been the Frege-Geach problem. This is easiest to understand using emotivism, according to which 'Murder is wrong' just means 'Boo murder!'. The problem with this view is that it has a hard time making sense of a compositional semantics for moral terms. 'Murder is wrong' doesn't simply appear in straightforward assertions, but also in more complicated embedded contexts such as: 'Johnny believes murder is wrong' , 'If murder is wrong, Sally wouldn't do it', and 'What if murder is wrong?'. None of these sentences make sense if 'Murder is wrong' has no meaning beyond its expression of disapproval for murder, since that disapproval is in no way built into the more complicated sentences.
This objection is widely regarded as having refuted the most simple forms of non-cognitivism, and one of the main tasks of more recent work in the area has been trying to find ways around it. It is still a matter of debate whether any such proposal has succeeded.
Another argument against expressivism attempts to refute the claimed essential connection between endorsement/motivation and ethical beliefs. Many naturalistic realists claim that amoralists are conceivable. Amoralists are perfectly competent users of moral language and can answer most everyday questions about morality correctly, but they remain completely indifferent to moral considerations. They will happily agree with you that setting cats on fire is wrong, but ignore that factor when deciding whether they will set a cat on fire today.
ERROR THEORY
Error theory agrees with realism on thesis (1); they take moral statements and beliefs to be truth-evaluable and not mere expressions of feelings. They differ by denying thesis (2), claiming instead that all ethical claims are false.[7]
The common strand to error theorists is that they think that moral assertions are all committed to some proposition that is false. A good analogy is what atheists think about claims about God, such as 'God loves dogs'. They think such claims are systematically false because they require the God's existence to be true.
What's the false proposition that all ethical claims are committed to? Different theories give different answers, but a popular answer stemming from J.L. Mackie is that all ethical claims are committed to the existence of categorical imperatives.[8] A categorical imperative is something you have a reason to do regardless of any of your contingent desires or goals.[9]
So such an error theorists agree with non-naturalistic realists about what ethical claims require of the world. They just disagree about what is present in the world. Contrastingly, naturalistic realists and error theorists typically agree that there are no categorical imperatives, but they disagree about whether or not moral claims require them.
Dialectic:
+ Show Spoiler +
Pros
Mackie gave two arguments for the claim that there are no categorical imperatives. The argument from “queerness” simply states that such things would be very metaphysically odd and hence are unlikely to exist unless significant evidence can be mounted for them. How could some properties be inherently reason giving, independent of how people value such properties? The argument from disagreement states that if there were such properties then we would expect there to be more agreement about them.
He focuses much less on the claim that ethical statements actually require categorical imperatives, apparently finding it obvious.
Cons
Non-naturalistic realists will want to downplay the “queerness” of categorical imperatives and the presence of disagreement. On the latter front, they may attempt to point to more general moral principles that end up giving different results in different contexts, explaining some seeming disagreement. They can also point to how widespread divergence in empirical beliefs affects what we think is right. For instance, maybe two groups agree that if God wanted the murder of non-Israelites then that would be a good thing, but disagree about whether God wants it.
Naturalistic realists will dispute the claim that ethical statements require categorical imperatives. The amoralists can once again be considered an argument against such a requirement, since an amoralist would plausibly not have reason to be ethical.
Mackie gave two arguments for the claim that there are no categorical imperatives. The argument from “queerness” simply states that such things would be very metaphysically odd and hence are unlikely to exist unless significant evidence can be mounted for them. How could some properties be inherently reason giving, independent of how people value such properties? The argument from disagreement states that if there were such properties then we would expect there to be more agreement about them.
He focuses much less on the claim that ethical statements actually require categorical imperatives, apparently finding it obvious.
Cons
Non-naturalistic realists will want to downplay the “queerness” of categorical imperatives and the presence of disagreement. On the latter front, they may attempt to point to more general moral principles that end up giving different results in different contexts, explaining some seeming disagreement. They can also point to how widespread divergence in empirical beliefs affects what we think is right. For instance, maybe two groups agree that if God wanted the murder of non-Israelites then that would be a good thing, but disagree about whether God wants it.
Naturalistic realists will dispute the claim that ethical statements require categorical imperatives. The amoralists can once again be considered an argument against such a requirement, since an amoralist would plausibly not have reason to be ethical.
WHAT ABOUT RELATIVISM?
Relativism is a very popular view among philosophically inclined non-philosophers. However, it hasn't received nearly as much support in (recent analytic) philosophy. Short answer is that if you think you are a relativist you are probably better described as an error theorist, so you may want to put that in the poll.
Long answer:
+ Show Spoiler +
On its face, relativism sounds contradictory. If something is wrong just in case a society disapproves of it and not wrong if a society doesn't disapprove of it, then the presence of disagreeing societies would seem to force us to accept 'X is wrong and X is not wrong', a contradiction.
Gilbert Harman's way around this was to introduce hidden indexicals into moral terms. An indexical is a term whose meaning varies with the context of utterance. 'I' is an indexical, since 'I am frogrubdown' has a different meaning when I say it than when you say it. Harman's claim is that there are indexicals in moral terms that pick out the speakers culture. So S's utterance of 'X is good' would mean that S's culture approves of X.
This view can be either hermeneutic or revolutionary. On the hermeneutic variant, relativism is a claim about language as we currently speak it. The revolutionary variant asks us to change our use.
The problem with the former is that it seems wildly implausible. Whether or not realism is true, people generally speak as though the moral claims meant more than just 'my culture things this'. If that were what they thought they were saying, then they would offer wholly different sorts of evidence (such as opinion polls) in moral debates than they actually do.
Of course, people aren't always right, but it's easier to be right about meanings than most facts. Linguistic intentions largely determine linguistic meanings, and you need a strong argument if you want to claim that people are systematically misunderstanding their own languages.[10]
Harman accepts this and advances a revolutionary relativism. But revolutionary relativism is basically just error theory. It's the view that we should replace present moral language because that language is horribly misguided in its attempt to describe non-existent facts. Error theory agrees with the second part and is simply silent on what we should do about the problem with moral language.
So relativism just takes error theory and adds the additional claim that a relativistic language would be a good replacement for our current one. This further claim can be disputed and has been by the likes of Paul Boghossian.
Gilbert Harman's way around this was to introduce hidden indexicals into moral terms. An indexical is a term whose meaning varies with the context of utterance. 'I' is an indexical, since 'I am frogrubdown' has a different meaning when I say it than when you say it. Harman's claim is that there are indexicals in moral terms that pick out the speakers culture. So S's utterance of 'X is good' would mean that S's culture approves of X.
This view can be either hermeneutic or revolutionary. On the hermeneutic variant, relativism is a claim about language as we currently speak it. The revolutionary variant asks us to change our use.
The problem with the former is that it seems wildly implausible. Whether or not realism is true, people generally speak as though the moral claims meant more than just 'my culture things this'. If that were what they thought they were saying, then they would offer wholly different sorts of evidence (such as opinion polls) in moral debates than they actually do.
Of course, people aren't always right, but it's easier to be right about meanings than most facts. Linguistic intentions largely determine linguistic meanings, and you need a strong argument if you want to claim that people are systematically misunderstanding their own languages.[10]
Harman accepts this and advances a revolutionary relativism. But revolutionary relativism is basically just error theory. It's the view that we should replace present moral language because that language is horribly misguided in its attempt to describe non-existent facts. Error theory agrees with the second part and is simply silent on what we should do about the problem with moral language.
So relativism just takes error theory and adds the additional claim that a relativistic language would be a good replacement for our current one. This further claim can be disputed and has been by the likes of Paul Boghossian.
WHAT DO YOU THINK?
Poll: What's your metaethical view?
Moral Realism (Naturalist) (38)
34%
Error Theory (30)
27%
Moral Realism (Non-Naturalist) (19)
17%
Expressivism (15)
14%
Other (9)
8%
111 total votes
Error Theory (30)
Moral Realism (Non-Naturalist) (19)
Expressivism (15)
Other (9)
111 total votes
Your vote: What's your metaethical view?
(Vote): Moral Realism (Naturalist)
(Vote): Moral Realism (Non-Naturalist)
(Vote): Expressivism
(Vote): Error Theory
(Vote): Other
According to the largest survey of contemporary philosophers, 66% are cognitivists (i.e., anti-expressivists) and 56% are realists.
The above is a small part of just one set of debates in metaethics. Feel free to bring in whatever relevant discussions you want.
Footnotes
+ Show Spoiler +
[1] I didn't state it explicitly, but you should interpret this as ruling out the assertions being speaker relative or having some hidden indexical. More on this in the relativism section.
[2] There really aren't any very good definitions of 'naturalism' and 'natural' out there. For now, let it suffice that no naturalistic theory can countenance stereotypically supernatural objects such as gods and ghosts. They are happy with quarks, guitars, solar systems and societies, though.
[3] There's an important distinction here between reductions and analyses. Most naturalists don't think you can analyze moral terms using only natural object terms. For instance, they're unlikely to find the simple definitions found in utilitarianism plausible. Their reduction is metaphysical, not conceptual.
[4] This runs into the Euthyphro problem.
[5] Non-naturalistic realism is also sometimes characterized as saying that moral facts are intrinsically motivating, like desires. So, if you believe that lying is wrong you will necessarily have some motivation not to do it. Many people find such a connection plausible when it comes to desires, but at least since Hume it's generally been felt that beliefs cannot have these motivating features. Naturalists do not think that they have them.
[6] I'm only discussing the anti-realist views that come from rejecting thesis (1) or (2). Rejecting only thesis (3) gets you moral skepticism, which is a less prevalent view.
[7] Astute readers will probably object, “How can both 'Murder is wrong' and 'It is not the case that murder is wrong' be false? If one is false, the other must be true?” One potential way around this objection is to limit the claims of error theory to logically atomic sentences, letting the logically complex sentences inherit their truth values in the standard way. So, 'Murder is wrong” would be false and 'It is not the case that murder is wrong' would be true.
[8] Other supposedly false supposed commitments of ethical claims could include the presence of libertarian or agent-causal free will or the existence of God. The former claim would drag you into a compatibilism debate and the latter claim would drag you into a Euthyphro debate (and an atheism debate). I wouldn't recommend it.
[9] This is extremely similar to the claim that moral beliefs are inherently motivating. But there are complicated debates in the vicinity about whether or not reasons are always (or should always be) motivating that I don't want to get into. You might want to pretend for now that having a reason to do something and having some motivation to do something are the same thing.
[10] The problems are even worse than I've indicated here. Some of them are arguably lessened by using a recent, alternative tool for dealing context sensitivity developed by John Macfarlene. Interestingly enough, this is known as relativism. The differences involve some complex formal semantics that I won't get into here, but if people have questions I can discuss it in the thread.
as a philosopher, I have to be happy!
