This is the latest post in a series about Bertrand Russell's essay Why I am not a Christian, in which I run through his various arguments and show that for the most part (there are a few exceptions) they either don't apply to classical theism, or are just completely invalid in the first place. Having discussed the cosmological, physical law and design arguments, Russell concludes this section by discussing two moral arguments, firstly the direct moral argument, and secondly an argument from remedying injustice. The second of these arguments is not one I have seen much myself, although the moral argument is still frequently used. In this post, I discuss the moral argument for God, and in the next one I will discuss the argument from injustice.
Russell attributes the moral argument for God to Kant. He claims that Kant disposed of the three main arguments for God (the ontological argument, cosmological argument, which in Kant's view relied on the ontological argument, and the argument from design), but then invented a new argument, the moral argument for God.
I would heavily dispute both that Kant dismissed the three arguments and that these were the only pre-Kantian arguments for God (he missed the teleological argument, for example). The only way Kant managed to link the cosmological argument and ontological arguments was because he defined God from the ontological argument. The cosmological argument ends with an uncreatable creator; Kant argued that to get to his definition of God as the most perfect being conceivable, one had to use the ontological argument. Classical theists avoid this by defining God as the uncreatable creator. The ontological argument plays no role. One can then conclude that God is the most perfect being (only an uncreated creator can be a being of pure actuality; only a being of pure actuality cannot deviate from its state of maximal goodness; degrees of perfection are measured in terms of degrees from the maximally good state for that type of being), but this is a conclusion rather than a definition, and the argument (which I have only sketched out here) does not rely on the ontological argument. Kant was either unaware of, mistakenly discounted or seriously misunderstood the medieval arguments linking the cosmological God with the most perfect being; certainly he never addressed them (at least in any of the works of his that I have read).
Besides, Kant has the fatal problem that he relied on his own seriously flawed metaphysics. This undermines almost everything that he wrote, including his objections to the classical arguments for God. Before he answered them, he had to recast them into his own metaphysical language. Kant's metaphysics is incompatible with the classical metaphysics on which the traditional arguments are based. By recasting them, he altered them considerably. Thus the arguments which Kant attacked were not those wielded by the best theological philosophers. Thus even if Kant defeated the arguments he addressed, it says nothing about whether he would have defeated more conventional presentations of the arguments for God.
Anyway, onto the moral argument. This can be expressed as follows:
- It is logically inconsistent to believe that there are objective moral values and to disbelieve in God's existence.
- There are objective moral values.
- Therefore God exists.
The argument itself is straight-forward, and logically sound: if the premises are true then so is the conclusion. To dispute the argument one would have to attack one or other of the premises.
The most important arguments for the first premise were actually supplied by modern philosophy. Hume famously argued that it is impossible to move from a statement of fact to a statement of value. Moore coined the naturalistic fallacy as the belief that goodness is undefinable, that is to say that we can define goodness in one of two ways:
- A good action is that which we ought to do; a good being is that which we ought to be.
- It is good for us to act for some particular end; it is good for us to be in some particular state resembling a particular ideal.
Moore's point was that what people want to imply when they study moral goodness is the first definition; what they in supply practice is the second definition, and there is no way to show the equivalence of any example of the second definition with the first. Moore's argument for this principle was by elimination: he went through every definition of goodness that he mistakenly believed that had been proposed for the second definition, and showed that it could not be linked to what we ought to do or ought to be.
Hume's and Moore's arguments are generally regarded as sound; accept their premises, and the conclusions follow. So to evade these arguments, one has to evade at least one premise for each of them (or one premise that they have in common). The problem that atheists face is that the premises used by Hume and Moore in the construction of these arguments are the same premises that they adopt to justify their lack of belief in God.
Classical philosophy avoids this problem in the most obvious way: it states that the standard for goodness is more fundamental than being. The definition of goodness we are most used to when we are not thinking of moral goodness is that something is good if it is fit for purpose (the Aristotelian would phrase this slightly differently, discussing whether the natural tendencies or final causes of a being are hindered). For example, we impart upon apples the purpose that we want to eat them. A good apple is in this sense one good for consumption; a bad apple one that is not. (This definition is in Moore's second sense; I will link it to the first type of definition below.)
Classical philosophers regard moral goodness as a subset of this definition. Beings are defined in terms of their natural tendencies. For example, a living organism is defined in terms of its tendencies to grow, absorb nutrients, reproduce and so on. Because these are tendencies rather than capabilities, it does not cease to be a living organism if any of these faculties are damaged; the form of the organism (which I define as the set of all potentia connected by continuous changes of the underlying Hamiltonian) has the natural tendency to absorb nutrients, even if a particular animal might not be able to do so at some point in its life due to damage to the intestines.
Mankind, defined as a rational, living, social animal, can be thought of in terms of the tendencies that provide the definitions of having a rational nature, a living nature, a social nature and an animal nature. We can figure out what hinders these tendencies from being achieved, and consequently define the good state of an individual man (or men in general) as the set of potentia where these hindrances are minimised. We are thus led to an understanding of what a good man looks like, both in physical health and health of character. Divergences from this are various degrees of evil. The state of the being is key measure of goodness, so this leads to a virtue rather than duty or consequentialist ethics.
Moral goodness is that subset of goodness which focusses on our will and desires. The question is: Do our desires agree or disagree with the natural tendencies of whatever is their object? If they agree, then we desire for that being (which might be ourselves) to be good. If not, then we desire for it to be something which it isn't. For example, one could desire for somebody to be dead. But this is to desire that they cease to be a living being. But the definition of that somebody includes the tendencies of living organisms. A non-living rational social animal is a different type of creature. By desiring that they are dead, we desire that they are something other than what they are; not just to desire that they are a different potentia of the same form (which would still be rational), but a different form entirely (in my language marked by a discontinuous change, namely death). But this is a contradiction. I used the ending of life as the clearest example, but the same argument applies for any divergence from a state of goodness. Thus to desire evil is irrational.
If we want to remain as rational animals (that is, as human beings), then we ought to be rational in our thinking. But for us to not be rational animals entails a contradiction. In this way, the if clause is redundant, and we can simply say that we ought to be rational. Thus we ought not to desire evil, and conversely we ought to desire good. This entails both seeking to improve our own character, and practising good deeds (that is, actions consistent with good desires).
Thus, if you accept the premises of this argument, there is a definition of goodness which is both concrete and what we ought to be. We do not jump from fact to value, but a statement of what is good for that being is the most important part of the definition of that type of being; we reason from what is good to the possible forms of that type of being, which we then compare against observation.
The argument I used here required the Aristotelian concepts of final causality and formal causality. As Aquinas showed, these directly imply the existence of God. What if we reject these? Both Hume and Moore assumed implicitly that there are no formal or final causes while formulating their arguments, and this, to my mind, is why their conclusions are wrong. Any other concept of an objective goodness would require that there is an inherent purpose to any type of being. If we discuss tendencies instead of purpose, we are back with Aristotle. If we discuss purpose, then we are heading for a divine command ethics. Matter in itself does not have intrinsic purpose (on this atheists and classical theists agree); therefore if there is an underlying purpose that applies to some of the matter in the universe (i.e. us) it must be an extrinsic purpose, i.e. something something imposed on it from outside. Since that outside thing cannot be material (to avoid an infinite regress), it must be immaterial. Since the only immaterial beings which can impart a purpose to others are those with a will, this means that the purpose must be imparted by an immaterial will. Since the effects of the will are imparted on matter, it must be linked to the matter in some way, either as the guider, sustainer or creator of matter. From the unity and universality of ethics (if one person has one ethics applying to them, and another a different one applying to them, then it is not an objective ethical theory); we can see that ultimately there can only be one such source; as a termination of a chain (in this case the chain of purposes) it must be self-existent or self-sustaining. Since it directs beings which are in part an abstract form (which can be represented mathematically), it must contain an intellect. And we are well on the way to reaching God.
In summary, if we are to apply to moral goodness the same principles that we use in other (undisputed) uses of the word goodness, then there must be either an inherent tendency or an inherent purpose underlying moral agents such as ourselves. In an If/ought statement, the if clause can only be neglected if it represents a tautology; and since the if clause describes both a subject and a goal for that subject, it can only be a tautology if that subject is defined in terms of a tendency or purpose. If we select an inherent tendency, then we have Aristotelian metaphysics, which implies the existence of God. If we select an inherent purpose, then we are immediately into divine command ethics, which implies the existence of God.
The only other possibility is if the requirement of self-consistency alone is enough to give us an objective moral code. It is, however, well known that this is not sufficient. The true theory of ethics (whatever it is, and assuming there is one) must be self-consistent; but many false theories are also self-consistent. Reason alone cannot be enough to determine a philosophical ethical framework (an application of Godel's theorem).
Thus I do not believe that the first premise of the moral argument for God can be questioned. This doesn't mean that atheists are necessarily immoral, or hold to an invalid moral code. It only means that if they do have a moral code, they cannot ground it in more fundamental principles while remaining consistent with their atheism. For the atheist, the moral code has to be seen as a brute fact, an arbitrary choice. Now, it might be that the atheist would by chance choose a code close to the true one (and, of course, many atheists, though less so today as society drifts further away from sound moral reasoning, have). But the atheist can not say that it is good for him to hold his particular favoured moral code is correct without contradicting his atheism.
What of the second premise of the moral argument for God? We can deny that there are objective moral values in two ways; firstly to deny that there are moral principles at all, and secondly to say that there are moral principles, but they are subjective, i.e. depending on the choices and viewpoints of the people involved.
To deny moral principles at all is to say that to not commit murder and to commit murder are of the same ethical standing. We see a hungry man in the street; we are just as justified to kick him in the face as we are to give him a sandwich. This is not a position which I think anybody holds. I do not believe that it is a rationally tenable position in view of what else we know about nature, but that's because I would reason from some other argument to God, and from God to an ethical position. Such an argument is obviously circular if we then go on to use the moral argument for God. I don't think that there is any contradiction in denying objective values; I just hope I never meet some such person who believes that and acts in accordance with that belief.
So most people who deny objective morality prefer adopting a subjective sense of ethics. For example, we might say that we have the right to do anything as long as we don't offend or harm others. That's a moral principle. But for it to mean something, we have to define what is meant by "offend" or "harm". Who determines what constitutes harm? This is where the one who believes in object ethics differs from the one who thinks that ethics is subjective. For the moral subjectivist to harm somebody is defined in terms of what they themselves consider harmful. But then, what makes us sure that that person is right? An alcoholic would consider it harmful if you deny him an extra bottle; to everyone sober it is evident that he is wrong. A pro-life activist, who believes that abortion is murder, is deeply offended by the actions of the abortionists; that there are people who commit such acts of sheer evil and yet freely walk the streets is an offence beyond reckoning. To them, the loss of each life is felt as painful as a stab in the heart. However, those in favour of abortion are deeply insulted by the insinuation that they are supporters of mass murder, and any attempt to constrain their deeds. They will be offended if told they can't commit abortion, and will thus go ahead and do it when circumstances are right for them. Whatever we do, we will offend one group of people. Do whatever you like as long as you don't harm or offend anyone is impossible to apply in practice, because when applied in practice the axiom leads to conclusions which are not self-consistent. And any true theory of ethics must be self-consistent.
Of course, all that is beside the point, because Do whatever you like as long as you don't harm or offend anyone is itself treated as an objective moral rule. What about those who don't believe it to be true? Either they can harm others as much as they like, or they will have the rule imposed on them. To impose moral rules means that you are either saying that morality is objective or determined by the arbitrary whim of whoever is strongest. But morality is meant to constrain our actions; what would constrain this strongest person if his will determines what is right and wrong? Thus in this case the moral rule is arbitrary; there is, in effect, no morality.
I have chosen one example; I could have chosen others. What I have said applies to any theory of subjective ethics. Either it collapses into an arbitrary amorality where at least somebody is able to bend the rules to their whim (and so can declare murder to be moral if they feel like doing a bit of murdering), or inconsistency, or some rule which is treated as though it were objective. Ethical relativism is simply not an option. A subjective ethics could only work if we never interacted with anyone else.
Thus, unless we are going to say that there is nothing reprehensible about murder, rape or child abuse, we have to accept the second premise of the moral argument for God. And the conclusion follows.
So how does Russell respond to this argument?
The point I am concerned with is that, if you are quite sure there is a difference between right and wrong, you are then in this situation: is that difference due to God's fiat, or is it not. If it is due to God's fiat, then for God Himself there is no difference between right and wrong, and it is no longer a significant statement to say that God is good. If you are going to say, as theologians do, that God is good, you must then say that right and wrong have some meaning which is independent of God's fiat, because God's fiats are good and bad not independently of the mere fact that He made them. If you are going to say that, you will then have to say that it is not only through God that right and wrong, but that they are in their essence logically anterior to God.
This line of thought dates back to Plato, though with one important difference: Plato applied it to the Greek gods, not to God. The gods, of course, are little different from super-humans; they are not the uncreatable and unchangeable creator of classical theology. And, indeed, I believe that this Euthyphro argument is valid, if applied to any authority except God.
The Euthyphro argument against divine command ethics (it is utterly powerless against Aristotelian based natural law ethics, the other ethical theory favoured by theists) has been refuted so many times that I am amazed that a philosopher as well read as Bertrand Russell was not embarrassed to wield it.
Now before we consider the argument in detail, we have to first decide whether we are applying it to a duty or virtue based ethical system. Virtue based ethical systems ask whether an individual being is good, usually by comparing it against some ideal. The basic idea is to focus on improving people's character, and then letting their deeds sort out themselves. A duty based system, on the other hand, labels individual actions as either right or wrong, and then lets the character of the actors sort itself out. For example a duty based system would see the act of murder as the primary evil, and treat the motivations of the murderer as secondary; the virtue based system regard's the murderer's motivations as the fundamental problem, and the action as merely a consequence of that evil. The act of attempted murder is evil primary on account of the motivations and character of the one committing it, and it does not matter if the attempt succeeds, fails, or, through no reluctance of his own, the potential murderer never carries the deed out.
It is not completely clear to me which of these two systems Russell has in mind when he frames his argument. Perhaps he thinks that it can be applied to both. However, when he says that it is no longer significant to say that God is good, he is implicitly adopting a virtue based system, or one that incorporates virtue based ethics. God is a noun, and virtue based systems are concerned about whether or not nouns are good. Duty based systems concern themselves with the goodness of verbs. In many respects this is good, because New Testament ethics is primarily a virtue system (albeit with elements of duty-based ethics mixed in; it is something of a synthesis between the two systems, but the emphasis is on virtue), with its calls for the building up of character, and focus on virtues such as agape and pistis. Subsequent Christian theology, which drew on both the insights of the Bible and the neo-Platonic philosophical system (whose biggest influences, Plato and Aristotle, were both different types of virtue ethicists), developed even more strongly towards virtue ethics. So we should consider whether the Euthyphro argument is valid in a virtue system or predominantly virtue based system of ethics.
Every virtue based system takes a number of assumptions for granted:
- Each being can exist in a number of different possible states (for example a man can be fat or skinny; he can be cowardly or reckless).
- Some of these states are identified as good, others as evil (evil being a lack of goodness and deviation from the ideal; the degree of evil is measured by the distance from the good state). There is always at least one good state.
- Our aim is to encourage goodness in both ourselves and others and avoid evil. So we should encourage both the coward and the reckless person towards the middle position, the virtue of courage.
The main difference between Plato and Aristotle is that Plato focussed on a single inflexible ideal as the defining standard of what is good, while Aristotle used the purposes and tendencies of the being to define a subset of the states as good, leaving open the possibility for some flexibility in what it is to be good.
The problem is, of course, in identifying which states are good and which are evil. And here the different schools of virtue ethics disagree. But which state is good isn't really relevant for the Euthyphro dilemma. All we need to understand is that some state is objectively good, and how that state is determined.
Now God is traditionally seen (and this is backed up both by various Biblical texts and by various philosophical arguments) as eternal and unchanging. Since He is unchanging, there is only one state which God can be in; and since every being has at least one good state, this means that God is good. This is simply a conclusion based on the definition of goodness and our knowledge and reasoning about God. Thus the goodness of God is not a matter of deciding which states are good and which are evil, but a consequence of the nature of God. To say that God is good is no more meaningless than to say that a man is good; to say that God must exist in a good state is not part of the definition of God, but a conclusion that logically follows from that definition. The two statements are not equivalent, since the standard definition of God is not a conclusion that can be drawn from the premise that God is good. If we labelled every conclusion of logical axioms as meaningless, then that would be to say (for example) that Cauchy's integral theorem is meaningless; an absurdity, since Cauchy's integral theorem is both useful and not self-evident from the basic definitions of complex functions and calculus. Thus a conclusion following from definitions need not be meaningless, and neither does the conclusion that God is good.
But now we come to everything except God; everything created by God. And here things are different. Because the beings are created (that is they come into existence), they must have at least two possible states, one set of states where they are non-existent and the other set where they exist (this allows them to change from non-existence to existence). One particular superposition of these states must be actual; others remain potential, but the being itself is a mixture of actuality and potentiality (while God is pure actuality). Since there are multiple states, it is now possible to label some states as good and others as evil. Obviously, those who argue that God is required for there to be an objective ethics will involve God somewhere in the reasoning used in this labelling, either taking God's fiat as a premise, or the existence of God as an additional consequence of the premises. For these beings there is such a thing as wrong (which there isn't for God), and our problem is to determine which states are right and which are wrong.
So here we come to the first flaw in Russell's argument. It assumes that what is right and wrong for God is determined in the same way as what is right and wrong for us. But this is not so. God is a fundamentally different type of being to anything created. To determine that standard for God to be good is a matter of reason alone. Every Christian theist would say that God is rational, so His nature satisfies the law of non-contradiction and cannot do impossibilities such as conceive of a pentagonal circle. To say that God is bound by His own nature and the laws of logic is thus not to place a restriction on God and call to some being prior to Him to determine His standard of goodness (since we are discussing virtue ethics and virtue ethics only applies to beings, it must be some being). For everything else, determining what right and wrong mean for it is not a matter of reason alone without also some understanding of the nature of the being. It cannot be. For example, we cannot determine the existence of rice pudding through reason alone, let alone decide what makes a good rice pudding.
Thus for every being except for God, an objective moral principle (that is a statement of which of its possible states are good and which evil) must be built on two sets of premises:
- The basic assumptions behind logic and reason.
- The nature of the being in question.
The first of these does not imply that there is some being higher than God in the chain of providing ethical meaning, since the postulates behind logic and reason are not beings. But the nature of the being does depend on the will of God, either directly or indirectly through the processes of physics; and as part of that determination of its nature is a decision about which of its possible states are good and which are evil.
Thus we come to Russell's second mistake. He assumed that there was a contradiction between the arbitrary nature of God's fiats and an objective process of logical reasoning that leads us to understand ethics. Every chain of reasoning requires premises that cannot be proved by that chain. In the case of ethics, those premises go back to the being's nature. But the nature is due to God's fiat. God could have decided that rabbits have large white tails (either to make them easy for huntsmen to shoot, or for some other reason), or He could have chosen otherwise. But once He makes that choice, the nature of rabbits is fixed, and consequently also the standard of goodness for rabbits. Since God is immutable, that standard cannot change in time; rabbits, of course, as a species can change and evolve, but if the do so too much then they cease to be rabbits, become something else, and the standard of goodness appropriate for that species will apply to them.
So the standard of goodness for everything except God is determined as the nature of the beings is fixed by God's fiat (and so logically follows and can be deduced from knowledge of that nature) so is both objective and down to God's arbitrary fiat. God, however, is an entirely different type of being, the only being whose standard of goodness is fixed by the demands of reason alone. Since reason is not a being, this does not mean that there is some being which God depends on to determine what is right and wrong for him; thus we avoid a contradiction (since God is defined as the being which does not depend on anything else for His existence or nature) and an infinite regress.
In this way, the Euthyphro dilemma fails to show that there cannot be an objective ethics based on God's fiat, and Russell's objection to the argument from moral reasoning fails.
However, this response only applies to God. The Euthyphro dilemma does show that no authority except God can determine what is right and wrong by fiat. Ethical principle cannot be determined by government, philosophers, mobs, individual strong men, councils of the United Nations, constitutions, declarations, bills, statements or gods. Such things have no authority to legislate morality, any more than they have authority to legislate science (for example, by declaring pi to be 3, or that the sun goes round the earth, or that all same sex attraction is an immutable trait immune to counselling; all governments that have tried to write such absurdities into law have ended with egg on their faces). Governments can and should respect natural law ethics (and science), but not claim to be above it. They should not stray beyond the bounds of uncertainty in computing the conclusions of natural law.
Having said that, I will admit that the moral for God is not as strong as the cosmological, or teleological arguments, or the argument from physical law. One could avoid it by denying that there are any ethical principles. I'm just thankful that atheists are generally inconsistent in this regard.
All fields are optional
Comments are generally unmoderated, and only represent the views of the person who posted them.
I reserve the right to delete or edit spam messages, obsene language,or personal attacks.
However, that I do not delete such a message does not mean that I approve of the content.
It just means that I am a lazy little bugger who can't be bothered to police his own blog.