In a recent episode of the Reasonable Faith podcast, William Lane Craig offers his thoughts on a 2012 paper by Jeremy Koons, Can God's Goodness Save the Divine Command Theory from Euthyphro? Koons' paper is another in a growing number of critiques aimed at the divine command meta-ethics advocated by figures like Craig, Robert Adams, and William Alston. Though a simple sort of divine command theory (DCT) received a devastating blow centuries ago from the famous Euthyphro dilemma put forward in Plato, modern defenders have adapted the DCT to resist the challenge presented by the dilemma. If good actions are merely those in accordance with god's commands, then goodness is arbitrary, since god could command anything and it would be good. However, Alston and others who adopt a modified DCT argue against this arbitrariness on the basis of the perfectly good nature of god. God could no more command infanticide, they say, than he could make a rock too heavy for himself to lift, because it would be in contradiction to his nature as god.
Does this move work? Craig believes it exposes the Euthyphro as a false dilemma, presenting a third option that is not identical to the other two options. Yet adding a third possibility to a dilemma does not necessarily mean the challenge underlying it is broken. It could rather indicate that we actually face a trilemma, which could be just as problematic as the original dilemma. This, I think, is where Professor Koon's paper is of real value. The question behind it is whether or not this move of DCT works any better than the two options typically posed by the Euthyphro. Craig firmly contends that it is better, but his arguments don't seem to warrant such conviction.
One of Craig's main criticisms is that Koons sets up a new dilemma that is just as flawed as the original. He says:
What he will ask now is: are these properties like loving-kindness, impartiality, generosity good because God possesses them or does God
possess them because they are good? He imagines this as a dilemma. It seems to me there is no dilemma there at all. The divine command
theorist, and Alston in particular, is very clear. These properties are good because God possesses them.
No doubt, this is what theological non-voluntarists like Craig, Adams, and Alston want to assert. But in his paper, Koons provides a puzzling quote from Alston that almost seems to suggest the opposite:
Note that on this view we are not debarred from saying what is supremely good about God. God is not good, qua bare particular or undifferentiated thisness. God is good by virtue of being loving, just, merciful and so on.
Craig seems to interpret the attention Koons gives to this quote as an accusation of contradiction. I don't think is what Koons is getting at, though, especially since he clarifies shortly thereafter that "Alston’s particularism requires that God’s goodness be logically prior to the goodness of the moral virtues. And we will see that this view is incoherent". It looks more like Koons is spelling out where he intends to direct his critique, and he directs it precisely where it should be directed, according to Craig.
All the same, Craig tries to resolve the apparent conflict by reference to the distinction Koons draws between explanations-why and explanations-what. Koons uses the contra-factual example of how even if the electron's negative charge were a brute fact that could not be further explained, it would still be possible to explain what a negative charge is. Thus, explanations-why may run out, but it need not mean there can be no explanation-what. Coming off of this distinction, Craig attempts to argue that this is exactly what divine command theorists like Alston are saying:
All the same, Craig tries to resolve the apparent conflict by reference to the distinction Koons draws between explanations-why and explanations-what. Koons uses the contra-factual example of how even if the electron's negative charge were a brute fact that could not be further explained, it would still be possible to explain what a negative charge is. Thus, explanations-why may run out, but it need not mean there can be no explanation-what. Coming off of this distinction, Craig attempts to argue that this is exactly what divine command theorists like Alston are saying:
When you get to God you've reached the metaphysical and moral ultimate, the explanatory stopping point. But that doesn't mean you can't explain what goodness is or wherein the goodness of God consists. As Alston says, you can still explain to people that God is loving, kind, merciful, generous, and so forth.
You can keep asking why the good is good, but eventually a stopping point must be reached, for theists and atheists alike. But, says Bill, you can continue to talk about what the good is in relation to the characteristics of god. However, this is where Professor Koons really has a bone to pick with DCT.
Koons observes that when the divine command theorist poses this explanation-what - that god is, per Alston, "good by virtue of being loving, just, merciful and so on" - this reverses the order of explanation employed by defenders of DCT that gets them to knowledge of the goodness of god. Usually, one thinks of god's characteristics to derive the conclusion that he is the supreme good. It's because god is loving, just, merciful, and so on that he is perfectly good. Proponents of DCT argue the opposite, that we start by intuiting that god just is all-good, and then derive the goodness of his characteristics from there. The problem with this is that it leaves astoundingly little content to the goodness of god. How do we conclude that god is good before knowing anything about who he is?
Craig proceeds to call for a necessary distinction between moral semantics and moral ontology. DCT, he says, is not a semantic theory or a theory of the meaning of ethical sentences, but is rather about the ontological grounding of moral values. Koons has made a category mistake, Bill asserts, because insisting on the meaninglessness or unintelligibility of the good is not a successful way to refute a theory concerned with moral ontology.
It's well known that Robert Adams once took DCT to be a theory of meaning, but the sharp divide Craig often wishes to draw between moral semantics and moral ontology is something to which not all ethicists commit. Particularly when it comes to theistic meta-ethics, it seems that semantics and ontology are more bound up than modern defenders of DCT will admit. In his 2004 paper, A Semantic Attack on Divine-Command Metaethics, Stephen Maitzen objects strongly to this sharp distinction on both religious tradition and logical grounds:
Koons observes that when the divine command theorist poses this explanation-what - that god is, per Alston, "good by virtue of being loving, just, merciful and so on" - this reverses the order of explanation employed by defenders of DCT that gets them to knowledge of the goodness of god. Usually, one thinks of god's characteristics to derive the conclusion that he is the supreme good. It's because god is loving, just, merciful, and so on that he is perfectly good. Proponents of DCT argue the opposite, that we start by intuiting that god just is all-good, and then derive the goodness of his characteristics from there. The problem with this is that it leaves astoundingly little content to the goodness of god. How do we conclude that god is good before knowing anything about who he is?
Craig proceeds to call for a necessary distinction between moral semantics and moral ontology. DCT, he says, is not a semantic theory or a theory of the meaning of ethical sentences, but is rather about the ontological grounding of moral values. Koons has made a category mistake, Bill asserts, because insisting on the meaninglessness or unintelligibility of the good is not a successful way to refute a theory concerned with moral ontology.
It's well known that Robert Adams once took DCT to be a theory of meaning, but the sharp divide Craig often wishes to draw between moral semantics and moral ontology is something to which not all ethicists commit. Particularly when it comes to theistic meta-ethics, it seems that semantics and ontology are more bound up than modern defenders of DCT will admit. In his 2004 paper, A Semantic Attack on Divine-Command Metaethics, Stephen Maitzen objects strongly to this sharp distinction on both religious tradition and logical grounds:
According
to
a
tradition
whose
philosophical
expression
dates
at
least
to
Anselm,
God
exists
of
metaphysical
necessity,
i.e.,
in
all
possible
worlds,
and
he
possesses
his
intrinsic properties
not
accidentally
but
essentially. Moreover,
even
atheists
have
acknowledged
the
good
reasons
for
thinking
that
if
God
exists
then
he
exists
(and
possesses
the
same
intrinsic
properties)
in
all
possible
worlds;
indeed,
some
atheists,
such
as
J.N.
Findlay,
base
their
alleged
disproofs
of
God's
existence
on
the
plausible
assumption
that
God
exists
necessarily
if
he
exists
at
all.
If
these
Anselmian
assumptions
are
correct,
then
all
of
the
following
sentences
have
the
same
truth-conditions:
(S1) 'God exists.'
(S2) 'God is omniscient.'
(S3) 'God is omnipotent.'
(S4) 'God is morally good. '
Since S4 is an ethical sentence, an attribution of a moral property to an ob ject, it belongs to the domain of sentences DCM [Divine Command Metaethics] needs to explain. If DCM gives only the truth-conditions, and not also the meaning, of S4, then it tells us nothing about S4 that is not just as true of the other three, presumably non-ethical, sentences. What is worse, if DCM gives only the truth-conditions of S4, then some entirely non-metaethical theory - a theory, say, giving the truth-conditions for attributions of omniscience - would tell us all that DCM tells us about that ethical sentence, in which case it is hard to see what would make DCM a metaethical theory, at least with respect to the moral attributes of God. So DCM had better concern not just the truth-conditions of ethical sentences but also their meaning.
(S1) 'God exists.'
(S2) 'God is omniscient.'
(S3) 'God is omnipotent.'
(S4) 'God is morally good. '
Since S4 is an ethical sentence, an attribution of a moral property to an ob ject, it belongs to the domain of sentences DCM [Divine Command Metaethics] needs to explain. If DCM gives only the truth-conditions, and not also the meaning, of S4, then it tells us nothing about S4 that is not just as true of the other three, presumably non-ethical, sentences. What is worse, if DCM gives only the truth-conditions of S4, then some entirely non-metaethical theory - a theory, say, giving the truth-conditions for attributions of omniscience - would tell us all that DCM tells us about that ethical sentence, in which case it is hard to see what would make DCM a metaethical theory, at least with respect to the moral attributes of God. So DCM had better concern not just the truth-conditions of ethical sentences but also their meaning.
Here we see more of the vacuousness of god's goodness under DCT. As Koons seems to be driving at, Maitzen argues that divine command meta-ethics can only be trivial in what it accomplishes. If we begin by intuiting the goodness of god, establishing the goodness of any other characteristics of god from that basis looks bleak indeed. The goodness of god would not necessarily mean all god's attributes are good-making. Is immateriality good because god has it? What about timelessness? Omniscience? These attributes seem non-moral, yet it doesn't appear that one has any means for distinguishing between them and the allegedly good-making attributes of god. On DCT, we just are not able to talk sensibly of the good-making properties of god, or of how those properties ground moral values.
To an extent, Craig wants to bite the bullet here. Goodness, he explains in the podcast, "is one of these primitives that really ultimately can't be defined." This is addressed by Koons in his paper, though, when he notes that this view, which comes from G.E. Moore, "merely meant that one could not analytically reduce the Good to other non-normative or non-moral concepts." The good is not absolutely inexplicable, but it cannot be neatly reduced in terms of definition to a non-moral proposition. So, the question remains of how effectively Craig, Alston, and Adams have accounted for the goodness of god in their theory, and whether their account is better than any of the competing accounts.
It's interesting to note how tempting it seems to be for theists to explain the goodness of god in light of god's particular characteristics. Near the end of the podcast, Craig identifies why he thinks god is a plausible explanatory ultimate for a moral theory. God, he says, is "worthy of worship." But why is this anymore indicative of god's perfect goodness than is his immaterial nature, his omnipresence, etc? It would not be far-fetched for one to make the case that worship has a moral component to it, let alone what it means to be worthy of worship. So is it perhaps that Craig and Alston are intuiting the goodness of god from his good-making properties, their denials notwithstanding? It certainly looks like a more sensible way of conceiving of the goodness of god than what modern DCT advocates claim to be doing. The alternative essentially seems to rest entirely on the mere assertion of belief that god is good. Who would fault anyone for needing more than that to devote as intimate an act as worship to another being?
Bibliography
Craig, The Euthyphro Dilemma Once Again, ReasonableFaith.org (Jan 4, 2015).
Koons, Can God's Goodness Save the Divine Command Theory from Euthyphro? European Journal for Philosophy of Religion 4/1 (Spring 2012), pp. 177-195.
Maitzen, A Semantic Attack on Divine-Command Metaethics, Sophia Vol. 43, No. 2 (Oct 2004).