In a comment to the question Astronomical Waste @armand put

Utilitarianism is a moral theory trying to codify "what people should do" type of questions. As such it can't really be refuted (the infamous is/ought problem). But it can be critiqued, and has been extensively (how exactly are we supposed to measure and compare utility, or if it even makes sense, being a key point of debate)

I don't know the details, but if the moral theories cannot be really be refuted, then all of them must be equally valid, therefore Morality cannot exist in the first place. So, how this no have stopped the development of moral theories?

The following is a "formalization" of the argument:

Let L mean "is metaphysically logically valid"

Let M mean "is a moral theory"

Let N mean "defines morality"

Let E mean "Exists"

  | 0. ∀x(Nx → Mx)                     The definition of morality must be a moral theory
  | 1. ∀xy (Lx ∧ Ly → Lxy)             If x and y are methaphisically logically valid interdependently must also hold dependently
  | 2. ∀x  (Mx → Lx)                   Hume's Law                   
  | 3. ∀x  (Mx ∧ Lx → ∃y[My ∧ ¬Lxy])   Construction of contradictory Moral theories
  | 4. ∀x (Ex → Dx)                    Realism-Existence requires defineability
  | | 5. ∃x(Mx ∧ Lx)
  | | u 6. Mu ∧ Lu
  | | | 7. Mu ∧ Lu → ∃y(My ∧ ¬Luy)                        ∀E 3
  | | | 8. ∃y(My ∧ ¬Luy)                                  →E 6, 7
  | | | w  9. Mw ∧ ¬Luw
  | | | | 10. Mw                                          ∧E 9
  | | | | 11. Mw → Lw                                     ∀E 2
  | | | | 12. Lw                                          →E 11, 10
  | | | | 13. Lu                                          ∧E 6
  | | | | 14. Lu ∧ Lw                                     ∧I 12, 13
  | | | | 15. Lu ∧ Lw → Luw                               ∀E 1
  | | | | 16. Luw                                         →E 15, 14
  | | | | 17. ¬Luw                                        ∧E 9
  | | | | 18. ⊥                                           ¬E 16, 17
  | | | 19. ⊥                                             ∃E 8, 9-18
  | | 20. ⊥                                               ∃E 5, 6-19
  | 21. ∃!x(Mx ∧ Lx)                                     RaA 5-20 
  | | 22. ∃x(Nx)
  | | u 23. Nu
  | | | 24. Nu → Mu                                       ∀E 0
  | | | 25. Mu                                            →E 24, 23
  | | | 26. Mu → Lu                                       ∀E 2
  | | | 27. Lu                                            →E 26, 25
  | | | 28. Mu ∧ Lu                                       ∧I 25, 27
  | | | 29. ∀x(¬Mx ∨ ¬Lx)                                 Id 21
  | | | 30. ¬Mu ∨ ¬Lu                                     ∀E 29
  | | | | 31. ¬Mu
  | | | | 32. ⊥                                           ¬E 31, 25
  | | | | 33. ¬Lu
  | | | | 34. ⊥                                           ¬E 33, 27
  | | | 35. ⊥                                             ∨E 30, 31-32, 33-34 
  | | 36. ⊥                                               ∃E 22, 23-35
  | 37. ∃!x(Nx)                                          RaA 22-36

Thus there is no definition of morality, and because morality does no exists morality does no exists (I am no sure how to formalize this last).

  • 4
    Because "if the moral theories cannot really be refuted, then all of them must be equally valid" is false. Existence of invisible pink unicorns "cannot really be refuted" either, but they are not "equally valid" with ordinary trees and rocks. Moral theories strive to reproduce and codify prevailing morality, and prevailing morality is prevailing because acting on it tends not to lead to destructive consequences for those who do so. A single general ought, like survival and well-being, is enough get utilitarianism going. There is no need to "really prove" it, that most agree on it is enough.
    – Conifold
    Dec 13, 2022 at 8:22
  • @Conifold Philosophy is no Psychology, therefore it does no try to reproduce and codify prevailing morality, that is a work for Psychology. And demostrated by the fact that philosophers like Haare or Benjamin Bentham deny the use of common sense morality, even Kant does that. Invisible pink unicorns is by the way falsifiable (unless someone moves the goal post), but most importantly under what criteria could moral philosophy claim that Hume's emotive is no lees valid than Preference-Psycological egoism or vice-versa? Both rely in true facts, but the conclusions are no the same.
    – Rieke
    Dec 13, 2022 at 15:43
  • No cheesecake recipe can be the objectively right one. We cannot refute any cheesecake recipe as being the right one. This does not mean all cheesecake recipes are valid (some will taste horrible). And it does not mean cheesecake recipes cannot exist, instead many can try to come up with their own and sell it for profit.
    – tkruse
    Dec 13, 2022 at 16:46
  • @tkruse Your mistake is to say that the how much someone like a cheesecake means that the recipe is more valid than others have any significant analogy with ethics, I may not like Hume's Emotivism, but it does make it less right/valid that Preference Psychological Emotivism.
    – Rieke
    Dec 13, 2022 at 17:10
  • 1
    Psychology does not study idealizations, which prevailing morality is. Kant was not a utilitarian, and common sense blunders so much that it is best to track which prescriptions do, in fact, optimize utility independently. Proving/refuting is altogether meaningless for values and prescriptions anyway, it only applies to declarative statements. Values are judged by practice, some persist and others are abandoned. "Moral philosophy" does not claim anything, it isn't math or physics, there are many of them, all claiming different things. But some are more popular than others.
    – Conifold
    Dec 13, 2022 at 23:10

6 Answers 6


What counts as refutation (or proof) is different for different kinds of proposition. For example, one cannot logically refute the logical possibility that the sun will not rise tomorrow morning. But if someone predicts that, as a matter of fact, the sun will not rise tomorrow morning, their empirical prediction will be conclusively refuted by waiting until tomorrow morning.

The is/ought distinction is more complicated than most people recognize. Facts play an important part in moral judgements. For example, if you want to say that murder is wrong, you have to define murder and that means identify the criteria for classifying an action as murder, and that means identifying what facts constitute murder. The difficulty is to get agreement on which facts are relevant to particular judgements.

If all moral judgements are equally valid, then they are all also equally invalid. Which is a confusing way of saying that “valid” and “invalid” (as understood in the context of empirical or deductive judgments) do not apply to moral judgements. However, moral judgements can be criticized on various grounds. For example one moral rule might be incompatible with other moral judgements so that carrying out one moral judgement amounts to breaking another one and one would have to choose which one to follow. Another would be that they are impossible to carry out.

I can understand that you are doubtful about the distinction between metaethics and ethics, and I agree that the two are intertwined. But there is a point to distinguishing between consideration of specific moral judgements and consideration of morality as a whole – whatever judgements may be made about specific moral principles. For more discussion see https://plato.stanford.edu/entries/metaethics/

Morality is about values and hence about what we desire and pursue. One of the things we value is agreement about values (we are social animals, after all), which means agreement (at least on some matters) on what is desirable and what is worth pursuing. Discussion of these cannot be carried on in the way that deductive reasoning or empirical research are carried on; it is not constrained in the same way. But that does not mean that discussion of them is impossible. It just means that the rules are different – and that it is much more difficult to reach agreement.

  • I am sure that codifying common sense morality is a work for the science of Psychology, no for Philosophy. And maybe in compability, but the existence of at least two independent moral systems that is self consistent that may be realizable, should be good enough to invoke Hume's Law and say that morality does no exist?
    – Rieke
    Dec 13, 2022 at 15:17
  • @Rieke. Psychologists would disagree with you; as scientists, concerned with what is and the the is/ought distinction means that they cannot codify a morality or say anything about values. If two moral systems existed as you describe, the is/ought distinction would not be able to ban either of them. It would compel you to say either that two moral systems exist (fact) or that one is better than the other (value), but these are separate unconnected judgements. But what I said was not a definition of a moral system; those were just examples of criticisms that might be made of any proposal.
    – Ludwig V
    Dec 13, 2022 at 15:43
  • 1
    @Rieke. By the way, it seems that you are attracted to moral nihilism, which is a respectable view in philosophy. Follow this link for more... en.wikipedia.org/wiki/Moral_nihilism
    – Ludwig V
    Dec 13, 2022 at 15:49
  • Is to note that even the claims of science must be Wertfrei, this no prevents they to make experiments and theories about what is thought to be wrong by their subjects of study, just they can prescriptive morality. Is also to note that is a scientific labor discovering the rules of common sense morality, including the possible simplification of any scientific theory. By the way how you chosse between Hume's emotivism and Preference-Psychological Egoism?, both have in fact are based in true facts but have diferent conclusions.
    – Rieke
    Dec 13, 2022 at 15:50
  • @Rieke. Yes, discovering what common sense morality is would be a scientific project. Hume's emotivism and Preference Psychological Egoism suggest how one might set about that project, but couldn't make an assessment whether the resulting morality is right or wrong.
    – Ludwig V
    Dec 13, 2022 at 15:56

It's the difference between "Normative Ethics" which is the Utilitarianism, Deontology, and Virtue Ethics, and "Metaethics" which is about what should classify as an ethical system. For instance you can define what is ethical as what provides the most amount of good for society, in which you can then have a debate over which Normative Ethical system is best. However, you can alternatively claim that what constitutes ethics is what is "good" and then proceed with the argument that terms such as "good" are essentially meaningless, as they have zero truth value. So really in this question you are alternating between Normative and Meta Ethics.

  • I don't believe in a separation between metaethics and ethics. In any case metaethics would be any thing that is no the a derivation from a system or the definition a system.
    – Rieke
    Dec 13, 2022 at 7:37
  • @Rieke, what do you mean by "don't believe in a separation"? That sounds similar to denying that there's a difference between general algebra and natural-numbered arithmetic, for example. Dec 13, 2022 at 13:19
  • @KristianBerry It is no. Branches of the Mathematics have well defined areas of study. The separation between Metaethics and Ethics is really blur, as a example one could give a Theory of Ethics that says that every ethic may be a valid ethic as long it comply with some metaethical principles and is being enforced in a socielty. And by the way the results of a Branch of math always an be used in another ones.
    – Rieke
    Dec 13, 2022 at 15:07

Good quesition

This question is one that is widely asked. However, it starts with a number of assumptions which are suspect.

Invalid assumptions

  1. Hume's is/ought barrier is recognized to be difficult to cross, but that does not make it uncrosssable. The most common methods to cross is are a) intuitionism -- we can intuit morality thru direct access, and b) indirect realism -- we infer the reality of useful theoretical entities, and moral realism is highly useful.

  2. "the moral theories cannot be really be refuted, then all of them must be equally valid". This is an "all or nothing fallacy". https://www.academia.edu/21565174/Fallacy_All_or_Nothing

Theories about our world, be they either about the physical, mental, social, or moral aspects of our world, can be BOTH supported AND counter-indicated, simultaneously, and basically can never be either "refuted" or "certain". See Quine -- theories are always underdetermined by evidence, so no evidence an ever refute a theory. We build up confidence in a proposition about our world by the accumulation of support, and the minimality of counter-indications.

There CAN actually be multiple mutually contradictory theories about our world, which are all more supported than not. A classic example from engineering is aerodynamic flow, where turbulent and laminar flow models involve mutually contradictory assumptions. They can't both be true, probably nether are, but we can solve most aerodynamic problems using one or the other, or sometimes a combo of the two.

Apply this to morality -- there can be multiple moral theories that are highly supported, but which contradict each other.

  1. "therefore Morality cannot exist in the first place". This is a fallacy of false inference. If we are currently not able to definitively describe something, does not support that it therefore does not exist.


You seem to lean strongly toward a black/white view of the world, which is common for people who are attracted to rationalism and are seeking certainty.

However, Kant, in The Critique of Pure Reason, laid the groundwork for the near total repudiation of rationalism. Our world is CONTINGENT. It cannot be pre-specified by any rationality-based claims. Instead, we have to observe it, and draw uncertain inferences.

The inference to moral realism, and the applicability of morality to us and our world, is never CERTAIN. It cannot be. Just as the reality of our world at all cannot be certain. The justification to accept physical realism, AND moral realism, is PRAGAMATIC, not rationalistic.

  • Firstly what is Rationalism?. Secondly, assuming that Hume's Law is true, there is no fallacy of all or nothing, if there is no possible to properly refute or falsify a moral theory (neither locally), the there is no way of claiming any of these theories as logically false, there is no way to claim that any of these is no more valid than other, you may be able to use other kind of critiques to try to invalidate a moral theory, but you would never be able to undermine the logical validness of a moral theory, just change the popularity of such theory.
    – Rieke
    Dec 14, 2022 at 1:16
  • Is to note that when two or more different theories that contradict each other are used in science and engineering, all of them are false, but approximate the reality good enough such use is practical and their predictions are correct if they are given a margin of error.
    – Rieke
    Dec 14, 2022 at 1:21
  • Thirdly there is no fallacy of false inference, because (1) for something to exist it must be definable (at least in my metaphysical beliefs), (2) therefore if morality exists therefore it must exist a moral theory that correctly represents morality. (3) But because all moral theories are (logically) equally valid, there cannot be a moral theory that correctly represents morality or morality is a meaningless term, because admits everything. (4). Therefore morality is no definable or is an meaningless term, (5) therefore Morality is nonexistant or is a meaningless term.
    – Rieke
    Dec 14, 2022 at 1:28
  • And no matter that something is to be the more practical concept ever, practicity is no a valid proof method. It only can say that something at most is an uselfull fiction
    – Rieke
    Dec 14, 2022 at 1:35
  • @Rieke I have created a chat forum for you to continue to argue your point. Comments on an answer are to point out errors in the answer, in order to improve it, not to argue against its conclusion. Here is the chat forum: chat.stackexchange.com/rooms/141285/…
    – Dcleve
    Dec 14, 2022 at 1:52

At first glance, the formal argument seems valid enough (if there are gaps/wrong angles, they could easily be filled in and adjusted, and I will assume the validity of the intended argument accordingly). So the dialectic focus shifts to: which premises are sound or unsound?

Since the framing of the OP question is the is/ought problem, I will not question the soundness of that premise (that wouldn't be fair to the OP). I will observe that the is/ought problem might be interpreted as the syntactic case of that which the open-question argument is a semantic example of, so the is/ought problem arguably stands or falls with its semantic counterpart. (In formal logic, syntax and semantics are routinely adjoined in the thematics of formal theories of logic.)

The premise that stands out, to me, as most unstable is, "For some x to exist, x must be definable." The SEP article on definitions lists at least six kinds of definitions that terms admit of. Offhand, we usually think of defining a term as listing other terms to which the defined one can be "reduced," e.g. "bicycle" more or less reduces to "vehicle with only two wheels." So the premise at issue could be interpreted as saying that "morality" must have some reductive definition, yet the argument goes on to invoke the fact that incompatible definitions of "morality" seem to arise at will, with the result that it's hard to tell what exactly the word "morality" refers to.

That's not so much an existence question, though (undefinable things exist by the by; in the foundations of mathematics, the question of undefinable subsets is relatively prominent, for example). And the solution is at least twofold: a moral theory can then be evaluated modulo a local definition, a "for the purpose of the occurrent argument" moment in reasoning; or then we can say that when variety of definitions arises less tractably, there's still a Quinean "change the logic, change the subject" remainder.

It's the problem of "talking past each other," then, more than a "does the subject actually exist" problem, again. Of course, as per (4) in the OP argument, this makes talk of "refuting" a moral theory slippery; many moral theories involve assertions that aren't testable by observation, this lack-of-testability goes back to the is/ought issue, "and we are done." However, some moral theorists do a good job of opening their questions up to other flavors of disconfirmation per the structure of their theories, e.g. John Rawls has a fairly malleable higher-order justification in play in A Theory of Justice. And Rawls even has subarguments in play that do make use of empirically falsifiable premises.

One possibility, then, is that some moral theories, and the self-definitions they depend on, have higher-order virtues of presentation that bypass the "talking past each other" issue pretty well, and hopefully it's these theories that get more attention as time goes on.

  • Your answer is untrue in its first sentence. There can be multiple valid but contradictory logic systems. The formal argument invalidly assumes that there can be no more than one logically valid moral system, and this is untrue in all logic applications. This is an unstated assumption required by the false assertion in line 8. See derivations of non-euclidean geometries for an example. All geometries are logically valid, yet contradict.
    – Dcleve
    Dec 14, 2022 at 15:44
  • @Dcleve, validity is relative to a given inference system. Talk of metavalidity is acceptable but that's not the level of assessment I was engaged in, yet neither did I mean to imply that such talk is entirely irrelevant in principle. As I said, any technical gaps/misinferences in the OP's argument can easily be corrected for by the principle of linguistic charity. Dec 14, 2022 at 15:45
  • The "formal argument" depends on this invalid reasoning. Its conclusion cannot be retained, no matter how charitable one is.
    – Dcleve
    Dec 14, 2022 at 15:49
  • @Dcleve, based on the OP's apparent level of familiarity with philosophy, I am guessing that they are working in first-order logic. FOL might not be sound modulo logical pluralism, and your answer has great merit on that score, but my own answer approached the OP more on their own terms as such. It isn't fair to expect every newcomer to know in advance all the details of higher-order logical soundness; wouldn't be fair for me to expect that, anyway. Dec 14, 2022 at 15:54
  • "undefinable things exist by the by; in the foundations of mathematics, the question of undefinable subsets is relatively prominent, for example" Citation needed.
    – Rieke
    Dec 14, 2022 at 16:53

It depends on what you mean by the existence of morality. If you take morality to mean a poorly defined set of beliefs, some mutually incompatible, held by humans, then morality clearly exists and your argument becomes irrelevant.

If you assume morality is some absolute set of values that exists outside of the human mind, then Hume's rule suggests your assumption is untenable.

  • What assumption?
    – Rieke
    Dec 14, 2022 at 9:41
  • My sentence began...'If you assume morality is...' That assumption. Apologies if my sentence was inadvertently hard to follow. Dec 14, 2022 at 11:05
  • How is untenable?
    – Rieke
    Dec 14, 2022 at 16:38

The fact a person cannot derive an 'is' from an 'ought' does not in any way prevent a person from formulating an ought or collection of oughts (morality).

In other words, morality can exist regardless of whether or not an ought can be derived from an is.

To suggest otherwise is to commit a non-sequitur.


To reflect the new question title and the OP's comment to the initial answer:

If some kind of objective morality is said to exist, then it would simultaneously constitute an 'is' and an 'ought'. It would constitute a fact about the world; a fact which tells us how to morally conduct ourselves. It would straddle the is/ought gap and in doing so provide us with a set of oughts from a description of what is. In other words, there would be no is/ought problem.

This likely goes some way to explaining why so many believe that the is/ought gap does prevent an objective morality: we currently have no evidence that such a moral 'is' exists, and without it we cannot derive a universal ought.

  • 1
    I mean objective morality. No informal moral claims.
    – Rieke
    Dec 14, 2022 at 16:47
  • I have added an edit to reflect your clarification. Dec 15, 2022 at 5:27
  • I would say that a morality exists if people adopt it, in the only way that a morality can exist. Numbers don't exist in the way that physical objects exist. If you demand that I demonstrate the existence of a number in the same way that I demonstrate that a table or chair exist, I will be unable to satisfy you. That doesn't prove that numbers don't exist.
    – Ludwig V
    Dec 16, 2022 at 13:11
  • Agreed. Only I would call that sort of morality a subjective morality (and, if widely adopted, a consensus subjective morality). Even a God-drawn morality would be subjective, if divine. Dec 16, 2022 at 13:15
  • @Futilitarian Well, it depends a bit how you define "subjective". I think it is a polar concept and gets its meaning, at least partly, by being the opposite of "objective". But it isn't possible to give a meaning to "objective morality" except "per impossibile", as per Hume's distinction. To put the point another way, can you define any other kind of morality than the ones that you describe as "subjective"?
    – Ludwig V
    Dec 16, 2022 at 14:24

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .