Euthyphro and Dicta

Euthyphro, The Argument from Moral Truth, and the Theory of Dicta

Andrew Jeffery

Read at the 64th Annual Northwest Conference on Philosophy, Oregon State University, November, 2012.
Update August, 2025

This paper explores the relationships between (1) the tacit dilemma Plato posed for theocentric accounts of right and wrong in the Euthyphro, (2) the inference that theism is the best explanation for the moral order; (3) and an objection A. N. Prior once offered against standard modal logics concerning what makes statements “truth apt.” The paper concludes by suggesting that a theistic "theory of dicta" could simultaneously address Timothy Williamson's meta-logical critique of propositional contingentism, while saving the Argument from Moral Truth for theism.

I.                 The Euthyphro Dilemma and the Argument from Moral Truth

In the dialogue Euthyphro, after Euthyphro embraces the thesis that what is loved by all the gods is holy, Socrates wonders whether, even if divine agreement were established, he would have has any closer to understanding what holiness is.[1]  He then asks, “Is what is holy holy because the gods approve it, or do they approve it because it is holy?”[2] The underlying dilemma in this context may be articulated as follows:

1. Either there is a reason why all the gods approve X, or there isn’t. 

2. If there is a reason why they all approve, this reason explains their attitude, and thus why X is to be done, and the fact of their approval is only a reliable indicator that this is so.

3. If there is no reason why they all happen to agree, then there is no telling what they will agree on, rendering divine approval useless for deciding what is to be done in subsequent cases.

4. So either divine approval is explained by a property of the thing or act approved, or divine approval is too accidental to serve as a guide for action.

5. Either way the Divine Approval Theory of Holiness must be abandoned.

It is widely recognized that Plato is posing a dilemma not only for Hellenic polytheism, but for any theory attempting to base moral truth on divine preferences (or commands). If the fundamental reason why something is morally right is that God approves or commands it, then the reasons for such divine attitudes or commands cannot, on pain of circularity, be likewise given in moral terms.  But if God approves of X for no reason at all, then morality would appear to be arbitrary.
Thus adapted to a monotheistic context, the dilemma for modern Divine Command theorists runs:

1.      Either God has reasons that justify his commands, or his commands have no reason behind them at all.

2.      If God has reasons backing his commands, then those reasons, not the commands themselves, are what really makes things right or wrong.

3.      If God has no reasons behind his commands, then God is arbitrary.

4.      If God is arbitrary, then God is imperfect.

5.      But if God is imperfect in any way, then God cannot be the supreme moral authority.

6.      Either way, it follows that God is not the foundation of morality.[3] 

Some also allege that if this difficulty for the Divine Command Theory goes unmet, it also undermines the so-called “Argument from Moral Truth” for the existence of God.  If it is already established that God cannot be the foundation of morality, then arguments that begin from a premise of moral realism cannot arrive at God via inference to the best explanation.  Rebecca Newberger Goldstein summarizes this criticism of the Moral Argument as:

1.      There exist objective moral truths. (Slavery and torture and genocide are not just distasteful to us, but are actually wrong.)

2.      These objective moral truths are not grounded in the way the world is but, rather, in the way the world ought to be. 

(Consider: should white supremacists succeed, taking over the world and eliminating all who don’t meet their criteria for being existence-worthy, their ideology still would be morally wrong. It would be true, in this hideous counterfactual, that the world ought not to be the way that they have made it.)

3.      The world itself—the way it is, the laws of science that explain why it is that way—cannot account for the way the world ought to be.

4.      The only way to account for morality is that God established morality (from 2 and 3).

5.      God exists.

. . . The major flaw of this argument is revealed in a powerful argument that Plato made famous in the Euthyphro. Reference to God does not help in the least to ground the objective truth of morality. The question is, why did God choose the moral rules he did? Did he have a reason justifying his choice that, say, giving alms to the poor is good, whereas genocide is wrong? Either he had a good reason or he didn’t. If he did, then his reasons, whatever they are, can provide the grounding for moral truths for us, and God himself is redundant. And if he didn’t have a good reason, then his choices are arbitrary—he could just as easily have gone the other way, making charity bad and genocide good—and we would have no reason to take his choices seriously. According to the Euthyphro argument, then, The Argument from Moral Truth is another example of the Fallacy of Passing the Buck.[4]

Thomists attempt to get around this difficulty, while preserving a theocentric account of morality, by grounding moral truth in God’s Intellect rather than in His Will.  Moral truths, so the Thomistic doctrine goes, are eternal ideas in the Mind of God, with which His Will necessarily coincides.[5]  It might be objected that this solution still reduces God to an explanatory “fifth wheel.”  After all, while an omniscient God would of course know all true propositions—but His knowing them would not be what makes them true—their being’ true would be what makes it the case that God knows them. 

II.       Lekta and Dicta

The above objection, however involves a tacit but disputable assumption about the ontology of propositions, namely, that they are platonic objects--eternal, mind-independent abstracta, true or false regardless, not only of what we think, but of whether anyone is there to think of them at all, in short, necessarily existing entities. This concept of propositions, often associated with Frege and the early Russell, can be traced all the way back to the Stoic lekton (assertible). Commenting on the Stoic definition of a lekton as a “self-complete sayable that can be stated as far as itself is concerned,” Barnes, Borzien and Mignuccie, in their contribution to The Cambridge History of Hellenistic Philosophy explain:

. . . But what then is the function of the remaining part of definition (1), the phrase (ii) ‘as far as itself is concerned’?  In fact it does not serve to narrow down the class of assertibles any further.  Rather, it is meant to pre-empt a misinterpretation: the locution ‘can be asserted’ could have been understood as too strict a requirement, that is, as potentially throwing out some things which for the Stoics were assertibles.  For there are two things that are needed for a statement of an assertible: first the assertible of itself, secondly someone who can state it.  According to Stoic doctrine, that someone would have to have a rational presentation in accordance with which the assertible subsists.  But there are any number of assertibles that subsist even though no one has a suitable presentation.  In such cases, one of the necessary conditions for the ‘assertibility’ of an assertible is unfulfilled.  Here the qualification ‘as far as the assertible itself is concerned’ comes in.  It cuts out this external, additional condition.  For something being an assertible it is irrelevant whether there actually is someone who could state the assertible.[6]

If instead one took the view that the fundamental bearers of truth were themselves contingent on the beings that used them, akin to what the Medievals called dicta, literally “things (that are) said,” the objection to the Thomistic solution above could be blocked.[7] Unlike mind-independent propositions, for something to be a dictum, there must be someone capable of (truthfully) saying it.  For an example of the contrast, consider Kenneth Konyndyk’s analysis of an objection A. N. Prior once offered against modal logic. Prior’s criticism was intended to apply to all systems that might hold “<>p ﬤ <> (p v q)” as a theorem:

. . . let p be ‘Only God exists’ and suppose this possible, and let q be ‘I don’t exist.’  Here p is possible ex hypothesi, but could the disjunction ‘Either p or q,’ i.e.,

     ‘Either only God exists or I don’t exist.’

possibly be true?  The peculiarity of it is that the disjunction is unstatable unless I exist, and is therefore only statable if both parts of it, ‘I don’t exist’ and ‘Only God exists,’ are false.  In this case, then, ‘Possibly either p or q’ is false although ‘Possibly p’ is true. [8]

Konyndyk continues,

Prior’s contention, notice, is that the proposition in question is statable (by anyone other than God) only if it is false, so it is not possible for it both to be stated and true.  Well, you might say, what does it matter whether anyone can state the proposition truly?  After all there is such a proposition, and if there is, then it is possibly true even if I cannot state it.  But this reply assumes just what Prior wishes to reject—that there is such a proposition whether statable or not.[9] 

Konyndyk adds, “It is worth noting here that Prior holds this not only for cases of pronominal reference, such as “I don’t exist,” but also for dicta using names which fail to refer—that is, names of non-existent objects.”[10] Why? Because on Prior’s preferred account of referential failure, no one can be in a position to truthfully state a proposition with an empty referent. 

Thus we have two distinct views on the fundamental bearers truth.  Both views assume the inadequacy of purely conventionalist or empiricist views.  Of the two, the necessitarian theory of propositions offers the more radically realist account in which propositions are mind-independent objects, possessing their truth or falsity independently of whether or not there is anyone to think of or discover them.  In contrast, the contingentist theory of dicta holds that, in order for anything to be a dictum, there must (at least) be a mind capable of truthfully asserting it. 

Critiquing Prior’s system Q, Timothy Williamson writes: “Since truth [on Prior’s view] requires statability, necessary truth requires necessary stability. Thus, a sentence of the form []A is true only if A contains no name of a contingent being, since otherwise the proposition A expresses is not necessarily statable.”[11] This will include seemingly tautological formulas containing non-referring terms, so the rule of necessitation fails.

Second, such systems, Williamson argues, will be incapable of distinguishing between seemingly contingent non-existence statements like:

~(Ex)(x = d),          

and explicit contradictions like:                     Kd & ~Kd 

because if d lacks a referent, neither formula expresses a proposition at all!

Thirdly, if modal necessitation fails for contingent propositions, it also follows that such systems will not be closed under tautological consequence. Williamson offers an example:

…from ~<>~A and ~<>~(A & B) we cannot conclude ~<>~B. For let A mean that David is the father of Solomon and B that David is a father. Then ~<>~A may be true because in any possible circumstance in which it is statable that David is the father of Solomon, David is the father of Solomon, since having David as his father is an essential property of Solomon, and ~<>~ (A ® B) may be true too because, in any possible circumstance in which David is the father of Solomon, David is a father, but ~<>~B may still be false because in circumstances where David lives childless it is statable but false that he is a father.[12]

Williamson goes on to argue that these problems can be generalized to any system with contingent propositions.[13] Propositional contingentism might thus seem to result in logical systems lacking significant desiderata for a robust system of inference. 

Nevertheless, if there is a God, necessary truth requiring necessary statability would not be a problem. The dicta theorist, then, who desires to hold both (1) that there are necessary or timeless moral truths, but also (2) that what can be said entails the existence of someone in a position to say it, seems to have a strong motivation to also embrace the idea of a God Who Speaks.

The Revised Argument from Moral Truth would look something like this:

1.  A statement can be true (or false) only if it could be truthfully asserted.  (Assertiblity is a necessary condition for a dictum’s existing at all.)

2.  A statement is assertible only if there is an agent in a position to assert it.

3.  Some basic moral statements are true necessarily and eternally.

4.  A statement can be necessarily true only if it is necessarily assertible.

5.  Something is necessarily assertible only if there is necessarily a mind in a position to assert it. 

6.  Therefore, there is at least one necessary and eternal Asserter of moral truths. 

Unlike in the Divine Command Theory, there is no implication here that God could have willed radically different moral truths from the ones we know.  God’s will plays no role in explaining why any particular moral proposition is true rather than false.  Rather, it is God’s existence eternally in a position to assert necessary truths that explains why there can be necessary Truths at all. 

III.               Faustian Bargain or Win-Win?

Would such a reformulation of the Moral Argument be a Faustian Bargain for theists? Besides simply seeming excessively Berkeleyan, would a theory of dicta, as compared to the theory of propositions, have disadvantages so great as to make abandoning this aforesaid family of theistic arguments a preferable option? To return to Prior’s counterexample to  àp ﬤ à(p v q), rejecting this theorem would rule out S5, the most powerful system of modal logic; and certain theists (Plantinga and those of a his persuasion) are rather fond of S5 in particular. Of course, both theistic and non-theistic platonists will probably see this entire argument as a reductio against its premises. 

But there are advantages as well. First, from a theistic point of view, a theory of dicta might offer a way to reconcile the tradition of propositional thought with the anti-platonic reservations of philosophers like William Lane Craig, who, motivated by the doctrine of divine aseity and quite un-Quinean criteria for ontological commitment, embraces a neutralist account of quantification together with a deflationary theory of truth.

Far too many philosophers, I think, are still in the thrall of a sort of picture theory of language according to which singular terms in true sentences must have corresponding objects in the world. . . A deflationary view of truth . . . need understand no more by the notion of truth as correspondence than that “that S” is true (or corresponds to reality) if and only if S. That is all there is to truth as correspondence, and it is wrong-headed to seek correlates in reality for all of the singular terms featured in S.[14]

On the theistic theory of dicta, God’s unique ultimacy is affirmed, while the intuitive (albeit derivative) necessity of logical, mathematical, and moral truths is retained.

Second, to any propositional contingentists disturbed enough by Williamson’s meta-logical critique, the theistic (version of the) theory of dicta offers a solution: accept the existence of a God Who Speaks. 

Lastly, a similar revision could be supplied, mutatis mutandis, to The Argument from Mathematical Truth.  If necessary truth requires necessary statability, then the existence of any necessary truth seems to require an ontological commitment to a God Who Speaks. In fact, it might be a stronger argument than the Argument from Moral Truth.  In philosophical circles, mathematical realism is embraced by many who reject moral realism, thus the status of mathematical truths as necessary is widely accepted both within, and outside of, the narrow confines of academic philosophy.  This alternative argument avoids the usual causal objections to treating mathematical truths as explananda. Explanation, in scientific practice, typically involves accounting for states of affairs in terms of previous states of affairs together with causal generalizations.[15]  Thus, to suppose that mathematical truths could be explanada at all thus might seem to commit one to supposing, for example, there could have been a time prior to  which the number 17 did not exist--surely not even a conceivable state of affairs! [16]   However, just because an explanation cannot be given in terms of prior events does not mean no explanation is called for, nor does it mean that none can be given. On the theistic version of the theory of dicta, the intuitive necessity of necessary truths, albeit a derivative sort of necessity, is preserved.

In summation, the theistic theory of dicta, in which an assertion is necessarily true iff there is a necessarily existing Asserter, solves the problem presented by Plato’s dilemma, as well as the meta-logical problems for propositional contingentism presented by Williamson. While it requires an alternative system of modal logic, it will preserve most modal intuitions about necessity and closure under tautological consequence, and so do justice to the intuitions of moral and mathematical realists.  Finally, there is the surprise that embracing or rejecting the soundness of certain theorems of modal logic, even on the most innocent-seeming formal grounds, has rather substantive implications for our ontological commitments.

Andrew V. Jeffery

Green River College, Olympic College

[1] Euthyphro, 9c. That is, even if divine unanimity picks out the extension of that which should be done or not done, Socrates would want to know why the gods all agree.

[2] Ibid., 10a.

[3] From the penultimate draft of Paul Herrick’s Think With Socrates (OUP, 2015), although this version of the summary didn’t make into print.

[4] Rebecca Newberger Goldstein, 36 Arguments for the Existence of God (Random House, 2010), 371-2. 

[5] Stump and Kretzmann make an even stronger claim, that absolute Goodness is identical with God’s essence. Although absolute Goodness is an abstract object, this is not inconsistent with conceiving of God as a particular being, insofar as the essence of any particular being, it’s haecceity, is an abstract object. See Eleonore Stump and Norman Kretzmann, “Being and Goodness,” in Thomas V. Morris, ed., Divine& Human Action, Cornell University Press, 1988; 281-312. 

[6] Jonathan Barnes, Susanne Borzien, and Mario Mignucci, “Logic”, in Keimpe Algra, ed., The Cambridge history of Hellenistic philosophy (CUP, 2005), 94.

[7] The originator of the theory of dicta seems to have been Peter Abelard. See Catarina Dutilh Novaes, “Medieval Philosophies of Truth” in The Encyclopedia of Medieval Philosophy, (Springer, 2020).

[8] Quoted in Konyndyk, Kenneth, Introductory Modal Logic (U. of Notre Dame, 1986), 61.  The original comes from A.N. Prior, Time and Modality (Oxford, 1957), 49. 

[9] Ibid.  Konyndyk goes on (62-63)  to sketch a brief outline of what a modal logic of dicta might look like, which necessitates a fourfold distinction between “strong” and “weak” necessity and possibility. 

[10] Ibid.

[11] Timothy Williamson (2013) Modal Logic as Metaphysics, OUP, 69ff). To be clear, Willamson is not arguing for theism; he is rather arguing, on the naturalistic assumptions, against contingentism.

[12] Ibid., 70.  Williamson follows Prior in using the formula “~<>~ A” (it is not possible that A is not the case) instead of the putatively equivalent [] A (necessarily A is the case). Prior eschewed the necessity operator [] in Q because its inclusion would have made the Barcan Formula derivable, a result Prior found unacceptable.

[13] Ibid., 72-75.

[14] Craig (2025) “Propositional Truth: Who Needs It?” Reasonable Faith
(https://www.reasonablefaith.org/writings/scholarly-writings/divine-aseity/propositional-truth-who-needs-it).

[15] “A scientific explanation is an adequate description of underlying causes helping to bring about the phenomenon to be explained.”  Richard Miller, Fact and Method: Explanation, Confirmation and Reality in the Natural and Social Sciences (Princeton, 1987); 60.

[16] The appendix to Goldstein’s novel mentioned earlier also includes an “Argument from Mathematical Reality” (387-388).  While space considerations preclude a full discussion of that argument here, it is not clear to me how the flaws she finds in that argument would apply to the argument I am suggesting in this paper, unless, perhaps, she would say that I am committing the “Fallacy of Using One Mystery to Explain Another.”