Meta-ethics is the field in philosophy which attempts to answer the question of what goodness is.
This opens up a domain of philosophical investigation ‘above’ normative ethics (hence ‘meta-ethics’) in which competing meta-ethical theories argue for what they claim goodness actually is.
Normative ethics are ethical theories which attempt to devise a system which enables us to determine which actions are good and which are bad. Utilitarianism, Kanian deontology and Virtue ethics are all normative theories.
Normative ethical theories all have a Meta-ethical core. For example, some forms of utilitarianism claim that goodness = pleasure/happiness. That is a meta-ethical view about what goodness is and forms Utilitarianism’s meta-ethical core. Once that is established, Utilitarianism can go on to formulate the details of a system that enables us determine which actions are good and which bad. For example, in the case of Act Utilitarianism, that would include the hedonic calculus.
Normative theories typically require that goodness at least exists, though they argue over what it actually is. However, some meta-ethical theories (anti-realist theories) claim that goodness does not actually exist.
There are two main aspects to answering the question of what goodness is: metaphysical and linguistic.
Metaphysical: What is the nature of goodness? There are two opposing views on this:
Moral Realism: The view that moral properties (like goodness/badness) exist in reality.
Moral anti-realism: The view that moral properties (like goodness/badness) do not exist in reality.
Linguistic: What is the meaning of ethical language? There are two opposing views on this:
Cognitivism – ethical language expresses beliefs about reality which can therefore be true or false.
Non-cognitivism – ethical language expresses some non-cognition like an emotion, does not attempt to describe reality and therefore cannot be true or false.
The outcome of meta-ethical debate has the utmost importance for normative ethics. If anti-realism is true and there is no such thing as objective goodness/badness, then it seems difficult to construct a normative theory. If moral-realism is true and goodness does exist in reality, then what exactly goodness is will limit the normative theories which are valid to those based on that correct view of goodness.
There are mind-independent moral properties/facts.
This is the view that goodness is something real in the natural world – typically a natural property. The natural world is the physical world. A natural property is a trait or feature that a natural thing has. For example, temperature would be a natural property.
Bentham’s Utilitarian naturalism
Bentham’s Utilitarianism claims that goodness = pleasure. Pleasure is a natural property (at least if you don’t believe in a non-natural soul) of natural creatures. Meta-ethically, Utilitarianism is therefore a form of naturalism, moral realism and cognitivism.
“Nature has placed us under the governance of two sovereign masters, pain and pleasure. It is for them alone to point out what we ought to do, as well as to determine what we shall do … a man may pretend to abjure their empire: but in reality he will remain subject to it all the while”.
Bentham’s argument is that it is our human nature to seek pleasure and avoid pain, so that’s all there is for morality to be about. We just are the kind of thing which cannot help but find pleasure good and pain bad. Bentham claims we could try to pretend otherwise but cannot escape this nature. As this is a fact of our nature, it is therefore a fact that goodness = pleasure.
Virtue ethics & naturalism. Goodness for Aristotle meant living a good life; Eudaimonia, which means flourishing. There is a factual difference between a plant that is flourishing and one that is not. So too is it with humans, Aristotle claimed. If Aristotle’s function argument is correct then he has identified what makes humans flourish; using their reason well. In that case, since humans are natural beings and truths about us are factual truths, goodness is a matter of fact.
The linguistic claims of naturalism are straightforwardly that ethical language is cognitive as it functions no differently to expression of any other type of belief about reality. To describe the color of the table, I say ‘the table is brown’. This is a sentence expressing a belief about reality. The ethical language ‘stealing from that bank is good’ is no different for the naturalist. It is a sentence and a proposition about reality which will be either true or false depending on the sense in which that particular action of stealing involves whatever natural property the naturalist claims to be good.
Moral non-naturalism: Moore’s Intuitionism
G. E. Moore holds that when we observe or reflect on a moral situation, such as someone stealing, our intuition gives us the proposition ‘stealing is wrong’, depending on the consequences. This isn’t reducing morality to some subjective feeling however. Just as all humans have no choice but to perceive the color yellow when looking at a yellow thing, Moore thinks humans have no choice but to apprehend the truth or falsity of a moral proposition when observing or reflecting on the relevant moral situation. He thinks this occurs because we apprehend ‘non-natural properties’. Intuitionism is cognitivist as Moore thinks that ethical language expresses a belief about the non-natural reality, which is based on an intuition.
Moore’s ‘open question argument’
This is Moore’s main argument against naturalism. He argues that if naturalism were true, the result would be illogical.
IF: goodness = pleasure
THEN: (goodness = pleasure) = (pleasure = pleasure).
BUT: goodness = pleasure is informative, telling us about the world.
YET: pleasure = pleasure is not informative (tautology).
Therefore, goodness cannot = pleasure, or any other natural property. Therefore, naturalism is false.
A question is closed if it shows ignorance of the meanings of the terms involved to ask. A question is open if it does not display ignorance of those meanings to ask it. Since ‘Goodness = X natural property’ for a naturalist would be synthetic, one could be acquainted with the subject (goodness) but not the predicate (X natural property) and therefore would not necessarily be displaying ignorance of the terms involved to ask the question. Therefore, it will always be an open question whether goodness really is X natural property as we can always meaningfully and intelligibly ask the question ‘is goodness really X natural property?’
Mackie’s response: arguably Moore can at most prove that our linguistic concepts of goodness and pleasure are distinct concepts that cannot be identical. That doesn’t tell us anything about the actual metaphysical status of goodness in reality. Mackie made this kind of argument, claiming that in Moore’s time philosophers were too optimistic in thinking that linguistic analysis could tell us metaphysical truths.
“There are questions of factual rather than conceptual analysis: the problem of what goodness is cannot be settled conclusively or exhaustively by finding out what the word ‘good’ means, or what it is conventionally used to say or to do. Recent philosophy, biased as it has been towards various kinds of linguistic inquiry, has tended to doubt this.” – Mackie.
Mackie gives an analogy with direct realism to illustrate this view:
“[direct realism] about colours might be a correct analysis not only of our pre-scientific colour concepts but also of the conventional meanings of colour words, and even of the meanings with which scientifically sophisticated people use them when they are off their guard, and yet it might not be a correct account of the status of colours”. – Mackie.
Mackie’s point is that direct realism about colours could perfectly capture the meaning of our words about colours, and yet still be completely incorrect as a theory of colour perception. This shows that linguistic analysis is not a valid way to draw conclusions about what is true of reality. So, Moore’s rejection of naturalism fails. Mackie thinks his arguments against naturalism do not have such deficiencies, and they also apply to non-naturalism.
Moore’s naturalistic fallacy
Moore was influenced by Hume and went on to argue that goodness can’t be equated with any natural property (like happiness) as any attempt to do so commits the naturalistic fallacy. Moore claimed that we can’t define goodness. We can’t say what goodness is. It is like the color yellow – you can’t describe or define yellow, you just experience it and can only point to yellow things. What is yellow? What does it look like? Just yellow… Moore says the same is true for goodness. Therefore, goodness can’t be a naturalistic thing as naturalistic things can all be defined. So, we experience goodness, which Moore clams is due to a faculty of intuition.
Hume’s Fork and A J Ayer’s verification principle as issues for Naturalism
Hume claims that there are two types of knowledge and two corresponding methods by which such knowledge is arrived at. Mathematical and logical truths are ‘relations of ideas’ because their truth is not dependent nor affected by any matter of fact. The facts in the universe could change completely and yet 1+1 would still = 2.
- Analytic truths are relations of ideas which is arrived at a priori
- Synthetic truths are matters of fact which are arrived at a posteriori
Moral propositions do not seem either analytic nor synthetic.
However, if Bentham, Mill or Aristotle’s naturalism is correct, moral propositions seem to be synthetic; true because of the way the world is. E.g. if an instance of stealing causes the greatest happiness for the greatest number of people, Bentham would claim it is a synthetic truth that it is wrong.
Ayer’s verification principle was influenced by Hume’s fork. It holds that statements are either empirically verifiable, logical/analytic or meaningless. Since Ayer regarded ethical language as neither verifiable nor analytic, he concluded it must be meaningless.
However, Ayer’s verification principle has its own problems, such as the problem that it can’t verify itself.
Hume’s argument that moral judgements are not beliefs since beliefs alone could not motivate us
Hume’s theory of motivation holds that belief alone cannot motivate action as you need a desire for that – beliefs only providing knowledge of how to achieve that desire. This would rule out purely cognitive states as sufficient for motivating moral action and therefore ethical language must contain at least some non-cognitive element.
Beliefs could cause desires which then motivate action.
However, this doesn’t explain why some beliefs motivate desires in some people but not others.
McDowell responds that beliefs cause desires which then motivate action depending on the moral outlook and general understanding of how to live (virtue) of the person in question, which explains why the same belief causes a desire in some, namely their virtue, but not others who lack virtue.
Hume’s is-ought gap
Hume’s law criticises naturalism. Hume said philosophers talk about the way things are and then jump with no apparent justification to a claim about the way things ought to be.
There are various ways of phrasing Hume’s argument:
- You can’t get an ought from an is.
- You cannot infer values from fact.
- Just because something is a certain way, that doesn’t tell us anything about how it ought to be.
- Is-statements do not entail ought-statements.
One way of appreciating Hume’s argument is seeing that, for any moral proposition, you cannot give a factual justification for its truth. Take the example of “it is wrong to kill people” and try to figure out what the factual justification is for it. Why is it factually wrong to kill people?
You could certainly point out various factual consequences of killing people that we often take to be wrong, such as that it hurts or harms people or violates their preferences. It is indeed a fact killing people can involve such things. However, why is it a fact that it’s wrong to harm people? Why is it a fact that it’s wrong to violate people’s preferences?
Whatever answer someone presents as a fact (is-statement) from which they have inferred their values (ought-statements), it seems you can always question what their reason is for that inference.
Any move from a factual is-statement such as “killing people causes suffering” to: “therefore, killing people is wrong” simply assumes that causing suffering is wrong. It looks like we cannot infer values from facts.
Regarding Bentham & Mill, they point out that our nature finds pleasure good. Yet, Hume would point out that this only shows that our nature finds pleasure good, not that pleasure is good and that we therefore ought to maximise pleasure.
Hume concludes that you could be aware of all the facts about a situation, yet if you then pass a moral judgement, that cannot have come from ‘the understanding’ nor be ‘the work of judgement’ but instead comes from ‘the heart’ and is ‘not a speculative proposition’ but is an ‘active feeling or sentiment’. This looks like an argument against realism but also against cognitivism and for non-cognitivism, specifically emotivism.
Patricia Churchland argued for interpreting Hume as specifically targeting deductive reasoning from is to ought, so that that factual statements cannot entail moral statements. However, Hume doesn’t seem to say it’s impossible to reason from is to ought, just that philosophers have failed to do so thus far. Churchland proposes that Hume’s argument leaves it open for inductive reasoning to do that job.
We could take Bentham and Mill’s arguments for utilitarian naturalism as inductive. Their claim then would not be that our nature finding happiness good makes it good, but that our nature finding happiness good is evidence for happiness being good.
However, consider that we have strong evidence that human nature finding pleasure good is the result of evolution, in order to guide animals to evolutionary goals. So, we are not justified in regarding our nature finding pleasure good as evidence for pleasure actually being good since we have stronger evidence for it being the result of something else (evolution).
John Mackie’s argument from relativity and his arguments from queerness
Mackie’s anti-realist argument from relativity
P1. Descriptive moral relativism (that moral codes/views differ cross-culturally) is true
P2. Though descriptive moral relativism does not entail meta-ethical moral relativism, it offers indirect support.
P3. Disagreement over scientific matters does not show that there are no objective scientific truths
P4. The reason for scientific disagreement is variation in speculative inferences or availability/adequacy of evidence.
P5. The reason for variation in moral codes seems to reflect adherence to and participation in ‘different ways of life’, however.
C1. Variation in moral codes is better explained by their reflecting ways of life than that they express seriously inadequate and badly distorted perceptions of objective values.
The classic response to relativism is to highlight cross-cultural similarities. The basis on which objective values are placed is not specific moral codes but general basic principles which are recognized at least implicitly to some extent in all societies. The remaining differences in moral codes are then the result of differing concrete circumstances, social patterns or preferences.
Mackie defends his argument by claiming that people do not hold their moral judgements due to general moral principles but instead their views are better explained by social conditioning. So Mackie thinks our moral views come from our society, not perceptions of real moral properties. Evidence could be added to Mackie’s argument that moral views tend to fall for the most part along cultural boundaries. This suggests that the origin of moral views is better explained by social conditioning rather than some kind of apprehension of general principles.
Mackie’s anti-realist arguments from queerness
Metaphysical queerness. Mackie claims that since moral statements motivate us, for moral realism to be true there must be objective moral properties which motivate us, which requires ‘not-to-be-doneness’ to be somehow present in reality. Mackie is arguing that the impossibility of conceiving of what would have to exist in reality to make moral propositions motivating is grounds for thinking there are no such things as the concept is incoherent.
Epistemological queerness. Even if there were objective moral properties, how could we know them? Moore’s answer that we just have a mysterious faculty of intuition is arguably not an answer because it doesn’t explain how that faculty works.
There are no mind-independent moral properties/facts.
Mackie’s Error Theory
Mackie thinks that ethical language does attempt to describe reality, so he is a cognitivist. Mackie claims that objectivism about values has ‘a firm basis’ in ordinary thought and the meaning of ethical language. While he thought that metaphysically Moore was wrong to think ethical terms intuited some non-natural reality, nonetheless in moral contexts ethical terms ‘are used as if it were the name of a supposed non-natural quality’ (my emphasis).
Mackie illustrates his view with the case of a scientist doing research on bacteriological warfare who is in a state of moral perplexity, wondering whether it would be wrong of them to do such research. Mackie claims such a person would ultimately want to:
‘arrive at some judgement about this concrete case. While his emotions and prescriptions will be part of the subject of the judgement, no such relation between the scientist and their proposed action will be part of the predicate. What they want to decide is not whether they really want to do the work, whether it will satisfy their emotions, whether they will have a positive attitude towards it in the long run, or whether the action is one they can happily, sincerely and rationally recommend or prescribe in all relevantly similar cases. What they ultimately want to know is whether this action is ‘wrong in itself’.
Mackie thinks this is how ethical language is really commonly used. When the average person uses ethical language they mean to talk about reality. They talk as if reality were such that their ethical language was true.
Since Mackie’s arguments from relativity and queerness cause him to accept anti-realism, he concludes that although ethical language expresses cognitive beliefs about reality, there is nothing in reality which is described by it. Therefore, ethical language does express propositions so it can be true or false, but they are all false, hence the name error theory.
If someone says “stealing is good” that statement is false.
If someone says “stealing is bad” that statement is also false.
Any ethical claim involves the false assumption that there is such a thing as objective right and wrong. All ethical claims are false.
Ayer agreed with G.E Moore’s naturalistic fallacy argument, that ‘goodness’ could not be identical with any natural property. However, Ayer disregarded Moore’s ‘non-natural’ properties solution as unverifiable. Ayer thinks we are therefore left with the position that there are neither natural nor non-natural moral properties in reality, so anti-realism is true. Ayer’s anti-realism relies on the success of Moore’s arguments against naturalism therefore, however, as his own theory only specifically targets intuitionism.
Ayer accepted the fact-value distinction that Hume’s is-ought gap implied. Ayer also thought the connection between moral judgement and motivation and the connection between motivation and feeling, made it most plausible that emotions were the best candidate for explaining the psychological function of ethical language and its unverifiability. Ayer thinks we are expressing our feelings when using ethical language.
Boo/hurrah theory. Ayer concluded ethical language was meaningless according to his verificationist theory of meaning, since it can’t be empirically verified nor is it analytically true. Ayer proposed that rather than attempting to describe reality, ethical language really expresses emotion. Saying ‘X is good’ is really akin to hitting ur toe on a chair and saying ‘oww’. The meaning of ‘oww’ is that it expresses – it connects to – the part of your mind that feels pain. That feeling of pain is not a cognitive belief that could be true or false. It’s the same with ethical language says Ayer – it connects to and expresses non-cognitive emotions, not cognitive beliefs. So ‘X is wrong’ is really ‘boo to X’, or ‘X is good’ is really ‘hurrah to X’.
R. M. Hare agreed with Hume’s is/ought gap and with Moore’s rejection of naturalism. Although Hare was a non-cognitivist and thought ethical language didn’t describe reality and couldn’t be true/false, he nonetheless thought it could have meaning as an expression of commands/recommendations/prescriptions. So for Hare, ‘X is wrong’ means ‘don’t do X’. That is clearly not a description nor can it be true/false. We have choice and freedom to decide which prescriptions to make, whereas we have less choice over which emotions to feel. They are also more rational as they are the product of informed imaginative and consistent thought. They are still separate from truth however. Prescriptive moral statements prescribe how the world should be rather than describe how it is. Hare thought the word ‘ought’ means both a universal prescription and also reflected the speakers interests.
So, prescribing something imposes rational constraints on you because you could prescribe something that contradicts your previous prescription(s). This isn’t to say that you ‘couldn’t’, do that, but that you would be irrational if you did. Therefore Hare makes room for rationality and rational considerations to exist within ethics, avoiding the reductionism of emotivism.
Whether anti-realism can account for how we use moral language
including moral reasoning, persuading, disagreeing etc
Moore criticised non-cognitivism as being unable to explain moral disagreement. Ayer states that the impossibility of moral dispute follows from his emotivism too which means Moore’s argument from moral disagreement applies to his theory. Ayer admits that people do engage in disputes which are ‘ordinarily’ thought of as disputes about value and have what can sound like rational arguments on either side of what seems like a debate. If ethical language were really just an expression of emotion, that should not be possible.
Ayer responds by claiming that “one really never does dispute about questions of value”. Ayer claims that moral disagreements are either genuine disagreements about non-moral facts or not genuine disagreements. Ayer points out that when we disagree with someone morally, we ‘admittedly resort to argument’ to win them over to ‘our way of thinking’, but our arguments do not attempt to show that they have the ‘wrong’ ethical feeling towards a situation which they have ‘correctly apprehended’. Ayer claims that his analysis showing the impossibility of moral disagreement provides support for his claim that ethical language is not fact stating, since disagreement about facts is possible.
Ayer’s point could be further defended by building on Hume’s is-ought gap. It may appear that ethical language includes moral disagreement involving logic and argument that cannot reduce to mere expression of emotion. However, arguably that is a confusion. This appearance could be explained by people being unconscious of the arbitrary emotional associations they have with certain facts. We disagree about facts and then don’t realise that we have emotional associations with those facts, so we confuse the factual disagreement for a moral disagreement. We can feel strongly that a fact has an ethical implication. We might even be unconsciously tempted to fudge or obfuscate in order to deny a fact, because of the ethical implications we associate with it. As Hume said, reason is a slave of the passions.
Hume’s is-ought gap tells us that since we can’t derrive a value from a fact, our fact-value associations are arbitrary and thus nothing more than how we personally feel. We express our feelings as moral claims which can appear to disagree, but ultimately they only express emotions which cannot disagree but only conflict. The appearance of disagreement only arises if we fail to understand that our statements about ‘right/wrong’ or even moral ‘truth’, really reduce to expressions of emotional approval or disaproval. A lack of awareness of the emotional associations we have with facts could be causing people to confuse what is an emotional conflict for a factual disagreement.
R. M. Hare claims that he does not understand what is meant by “the objectivity of values” and that he has not met anyone who does. Hare provides a thought experiment to show this where has asks you to imagine two universes, one with objective values and one without. In each, Hare claims, people would behave and talk in the same way. This shows, Hare thinks, that the meaning of our ethical language is disconnected from anything like objective values, whether they exist or not. So, we should understand the meaning of ethical language as non-cognitive.
Mackie responds that although Hare is right that people would behave identically, nonetheless in the universe with objective values the people who happened to hold what would then be the correct moral judgements would have true beliefs.
Mackie regarded emotivism and prescriptivism as ‘part of the truth’ as they explain why ethical language is motivating of action, but claims it’s a ‘very natural reaction’ to non-cognitivism to “protest that there is more to ethics than this, something more external to the maker of moral judgements, more authoritative over both him and those of or to whom he speaks, and this reaction is likely to persist even when full allowance has been made for the logical, formal constraints of full-blooded prescriptivity and universalizability. Ethics, we are inclined to believe, is more a matter of knowledge and less a matter of decision than any non-cognitive analysis allows.” Mackie claims Naturalism satisfies this ‘knowledge’ demand. Linguistically, however, naturalism only seems to allow for a purely descriptive and ‘inert’ ethical statements, which cannot explain their motivating force. So Mackie concludes that to properly account for how we use ethical language both as involving more than mere non-cognitions and yet also being motivating, anti-realism would have to be combined with cognitivism.
The problem of accounting for moral progress
Moral progress is when a society improves its moral views and practices. Human rights and rights for minorities and workers are examples of this. Actual improvement presupposes an objective standard in which progress takes place, however. If anti-realism is true then actual progress is not possible as there is no objective spectrum on which to improve or worsen.
Ayer can be defended with the counter-claim that there is no such thing as actual progress. The fact that women can now vote, for example, Ayer would regard as nothing more than that a sufficient number of people were persuaded to have a certain emotional reaction which was what society happened to require for the law to be changed.
Hare would regard moral progress as the increasing rational coherence of our prescriptions. For example, racism is not universalizable because a racist prescription against another race cannot apply to the speaker or the speakers race. There is ultimately no rational reason to prescribe racism nor to think one’s own race superior to another. Hare would explain the history of moral progress as the gradual erosion of irrational prescriptions and their replacement with rational ones.
Mackie acknowledges that there have been moral reformers who sought to instigate moral progress but argues that this does not come from their somehow having figured out objective moral facts. He instead suggests that ‘progress’ in fact resulting from thinking through the already held moral doctrines in a new way or recommending some new action because consistency of it with previous doctrines was desired. For example, the American constitution claims that all men have inalienable rights, and this was used by Martin Luther King to argue that black people should also have equal rights. So, King was merely expanding already existing doctrines not discovering objective moral progress. Mackie thus thinks our moral views and the changes they undergo originate from what a society happens to value due to contingencies of history and evolution, not some real objective standard within which their change could count as objective progress.
Whether anti-realism becomes moral nihilism
Nihilism is the view that we should abandon all moral principles because they can never be justified. Anti-realists claim that there are no objective moral values. It follows that there is no objective truth nor falsity in ethics. Different people are not good or bad, they just have different emotional associations. Towards Jews, Hitler merely had a particular emotional association, prescription or cognitive belief that is in error. There’s nothing more that can be said, no way to say Hitler was ‘really wrong’, just that one might have a different emotional reaction than him. Anti-realism seems to lead to moral nihilism.
Anti-realists could respond that they are not committed to total abandonment of morality. Emotivists hold that we can still morally condemn people, just that those condemnations are non-cognitive expressions of our negative emotional reaction to them. Prescriptivists claim we can condemn irrational things like Nazism and prescribe universalizible things like equality, though those are non-cognitive prescriptions. Hare gave the example of a fanatical Nazi who found out his ancestors were Jewish. He admitted this to other Nazis and was killed. Hare would argue that it is irrational as Nazism as a prescription can’t be universalised, though there is nothing objectively wrong about being irrational, for Hare. Mackie could say that although ethical language is in error, on the anti-realist view there’s obviously nothing ethically ‘wrong’ with using erroneous ethical language, so there’s no need to abandon it. Nor is there any sense to be made of the idea that we ‘should’ abandon morality. In fact that is an incoherent claim, as ‘should’ is an evaluative word.
Ultimately, however, whether anti-realism leads to Nihilism or not is irrelevant to the truth of the theory. This criticism in a way misses the point of meta-ethics however. Meta-ethics is just trying to determine what rightness and wrongness are. We may not like the result, it may indeed lead to the destruction of the world. But that doesn’t mean it’s incorrect. The science behind nuclear bombs may well end up destroying the world, but that doesn’t mean it’s incorrect. The behavioural consequences of everyone believing something doesn’t not count against its truth.