David Hume’s Muddle Part II

Is The “Is-Ought” Gap a Real Problem for Ethics?

This discussion of ethics (or meta-ethics), with all the previous postings, is intended to be one piece—one continuous argument. It started with the 20th century low point of emotivism, the argument that ethics is meaningless nonsense. We then trace the lineage of that point of view back to David Hume and his so-called “is-ought” or “description-evaluation” thesis. In many respects he’s a founding father of not only emotivism and its offshoots but much of contemporary relativism and subjectivism in ethics. Hume shares that distinction with G.E. Moore and his construct—the “naturalistic fallacy.” But it was Hume who blazed the trail that brought us to this point—this point of ethical uncertainty.

In Part I we pointed out that Hume had an austere concept of “fact” which got him into a muddled confusion. Hume limited facts only to things that can be perceived with our senses. Part I tried to show the serious mistake in such thinking. That discussion also hinted as to where this argument was ultimately leading. We’re now going to expand on the argument that a “fact” is a more complex concept than Hume ever thought. More importantly, we’ll debunk Hume’s dichotomy itself—his strict separation of facts from values. That is, we’ll attack Hume’s very set up of the issue.

I’m trying to make this series accessible even if, at times, I oversimplify things. I think philosophy, in particular ethics, is important. It shouldn’t be obscure but accessible to all. I’m a believer in what’s called public philosophy. I also submit that it’s important to think more modestly about ethics. Ethics is not other-worldly. Some philosophers paint a picture of ethics as beyond the everyday. Nothing could be further from the truth. I’ve suggested that we use science as a down-to-earth model—that is, we should look for reasonable and probable explanations and, like science, accept that our results will be limited, conditional and corrigible.

Incidentally, up to this point we have not discussed the source or sources of moral norms. As I said in Part I, making an attempt to determine the source of morality is a separate task. Our task here is limited—to debunk Hume’s dogmatic thesis. It’s a stubborn obstacle to any genuine debate about ethics. We need to put it aside before we can move on.

We are doing this through an analysis inspired by two analytic philosophers—an American, John Searle and an Australian, Julius Kovesi. Both men, by the way, attended Oxford at the same time and were mentored by J. L. Austin and a band of devoted Ludwig Wittgenstein disciples. Both have somewhat parallel arguments to displace Hume’s muddled thesis.

John Searle denies that he set out to do ethics. He insists that his focus is the analysis of language and the creation of institutional reality. He admits, however, that his analysis is a relevant tool for meta-ethics and ethics.¹ Searle nevertheless had no problem taking on Hume’s “is-ought” claim and has aggressively done so many times. We explored one such argument in Part I. We’ll explore a more fundamental critique here.

Julius Kovesi, on the other hand, clearly acknowledges that he was trying to put ethics on a more realistic footing. For him that also requires a convincing rebuttal of Hume. Both philosophers take a language oriented, analytic or concept formation approach.

Let’s start by pointing out that “language” is an indispensable part of reality. That may be difficult for some to fully grasp. Moreover, language itself is more than just a part of reality—language creates and structures much of our reality. To put it bluntly, some facts are created linguistically. We hinted at that, if you remember, in Part I with a very short discussion of “knives, forks and spoons.” We’re now going to shake things up and expand on that.

An insightful analysis was advanced in the past few years by John Searle. Searle applied his work on language and speech acts to the formation of factual concepts that create our complex institutional reality—that is, our civilization. We start with Searle.

“There is a fundamental distinction, in our basic conception of reality, between those features of the world that are observer independent such as force, mass, and gravitational attraction, and those features that are observer relative, such as money, property, marriage, and government. … [T]here are many observer relative features of the world that nonetheless can be objective matters of fact.” ²

“Institutional reality can only exist insofar as it is represented as existing. Something can be money, a football game, a piece of private property, a marriage, or a government only insofar as it is represented as such. In order for something to be any of these phenomena, it has to be thought of in a certain way, and these thoughts will represent it in a certain way. But the representation of these institutional facts always requires a language. … Now, this is the secret by which human beings can create an institutional reality and other animals cannot. Human beings have the capacity for seeing and thinking at a double level. We can both see the piece of paper and the dollar bill.”³

Searle uses the term “status function” to explain how we form descriptions of objects and states of affairs in order to structure our reality and greatly expand human capabilities. His favorite example is money. The “status function” of money has nothing to do with its physical characteristics—which is usually some sort of colored paper. Yet it is a fact that money exists. And the creation of the concept of money through language obviously expands our ability to do things. But what is important for this discussion is that, according to Searle, “there is a normativity attached to the notion of a ‘function.’”⁴

“We live in a sea of status functions, we are citizens of a country or a student or a professor, we are a holder of a valid driver-license and a beneficiary of an insurance policy. We live in a sea of invisible status functions, and this is the key to understanding human civilization.”⁵

According to Searle the relationships created by various status functions invariably come with duties, obligations, rights, and normative standings of various kinds. Searle argues that status functions actually impose upon us what he describes as desire-independent reasons for acting—reasons not controlled by personal inclinations. In short, actions we ought to undertake that are dictated by our status functions—our roles. For example, the status function of being President of the United States comes with a well-understood set of obligations. Searle goes on to label these institutional facts in contrast to brute facts. That is, in contrast to sense perceptions—which is Hume’s only understanding of facts. Thus, Searle can logically claim that facts—institutional facts—have the ability to impose obligations, directly contradicting Hume.

Let’s leave Searle for now and move on to an analysis by Julius Kovesi. By the way, this sketch of Searle’s work doesn’t scratch the surface of how the linguistic creation of institutional facts creates much of our reality. An excellent introduction would be Searle’s Making The Social World, Oxford (2010).

John Searle’s classmate at Oxford, Julius Kovesi, applies a similar analysis specifically addressed to ethics in his book Moral Notions.⁶ The late Julius Kovesi was not well-known outside his native Australia. However, the acclaimed ethicist, Alasdair MacIntyre, has praised his work saying that “Kovesi’s Moral Notions, is a minor classic in moral philosophy that has not yet received its due.”⁷ Kovesi’s work is dense and difficult. Nevertheless the effort pays off.

Kovesi starts with an insight about facts similar to Searle. From that he argues that we actually “evaluate” by way of or through factual descriptions—a deceptively simple distinction that restructures Hume’s dichotomy. As Kovesi puts it, “Evaluation is not an icing on a cake of hard facts.”⁸

“When we turn to our moral life we do not find that there is some raw material there waiting to be described and evaluated. … We do not enter the world at the level of brute objects. … It is for this reason that we do not move from ‘facts’ to ‘values’ or ‘derive’ them from ‘facts.’ By the time we have described the world it is not ‘brute’.” … [We] do not first keep uttering sounds like ‘I promise to pay you’ and then make a decision about keeping promises, but first find the need for the institution of promising and then make use of words to make promises.”⁹

Kovesi argues that concepts are formed with agreements about meanings and commitments—they embody implicit standards. This tracks Searle’s analysis of status functions including the normativity of functions. We form these descriptive concepts, says Kovesi, through constant discussion and debate over the purposes, wants, needs, ideals, desires and goals of everyone and anyone—in short, by all of us who need linguistic conventions in order to get along as a group or community. This process is, of course, a publicly conducted group enterprise. Over time descriptive concepts become ensconced in the language. So, for example, the concept of a “liar” became an established descriptive (and normative) concept.

Kovesi plainly shatters Hume’s dichotomy: “evaluation” is simply a function of “description.” In Kovesi’s paradoxical phrasing, “moral notions do not evaluate the world of description; we evaluate that world by the help of descriptive notions.”¹⁰ That is indeed paradoxical to anyone wedded to Hume’s thesis that “description” and “evaluation” are two completely separate parts of a dichotomy—like apples and oranges. Thus, we can now say that Hume’s basic mistake was limiting the concept of “fact” itself. His next mistake was in setting up his strict dichotomy. Both pushed him deep into a confused muddle.

According to Kovesi, some philosophers, trapped in Hume’s seemingly iron-clad dichotomy, needlessly anguish over finding ethical rules that are objective.

“A large number and surprising variety of moral philosophers seem to talk about our moral life and language as if each of us spoke a private language and yet paradoxically lived in society, as if our moral notions were private notions that we try to make universal.”¹¹

“[W]e form our notions from the very start from the point of view of anybody, these terms do not reflect my wants but anybody’s wants. The tension between my preferences and what is good, or between what I would like to do and what I ought to do enters into our life already with our language.”¹²

“This is not because we happen to be such nice people that we formulate our notions from the point of view of anyone, but because our language is public”¹³

In simple terms, Kovesi points out that Wittgenstein’s breakthrough argument disproving any sort of private language obviously must apply to language at all times—even language about ethics.

The best way to demonstrate these arguments of Searle and Kovesi is with a simple example. Starting simple is always the best way to start any ethical analysis. If we describe a statement made by President Donald J. Trump as a lie that would be a simple factual claim. As a factual claim it can be assessed for its truth or falsity—just like any factual claim. By way of assessing its truth or falsity we can conclude whether or not Donald J. Trump in fact told a lie. If so, then—in fact—he is a liar. A liar is a descriptive concept. Our use of the word liar, all things being equal, is a description of a person who broke a rule. He violated a built-in commitment to the truth in making a statement.

Now consider what just happened. To some, it may look like a move from description to evaluation. But it’s no such a move at all. It’s merely a statement of fact that can be assessed for its truth—a liar is one who tells a lie. If the fact is true, then Donald Trump is a liar. That is, he fits the description. It so happens that such a description also amounts to an evaluation of conduct. He broke an important rule about statement making—a basic rule about the function of language which is also an ethical rule. To tell a lie is wrong. Thus, calling someone a liar is both factually descriptive and a moral evaluation. Thus, it’s not a move at all from description to evaluation. It fits Kovesi’s claim that evaluation is a function of description. Or, it’s a factual description that has a well-known status function if we use Searle’s vocabulary.

Some will argue that this final step, that Trump is morally blameworthy, is an extra move. I completely disagree. There is nothing tricky going on here. There was no illegitimate leap from fact to value or description to evaluation. It’s all in plain view. As Wittgenstein would say; “Nothing is hidden.”¹⁴

But let’s note two very important things. First, being a liar is objective. It’s not subjective. It’s a moral concept to be sure. For those invested in Hume’s dichotomy, that’s paradoxical. But quite simply a lie is a descriptive concept with an agreed meaning. The heavy moral lifting has already been done elsewhere.

So, I don’t need to import from the outside an extra “ethical” premise to complete an evaluation. In summary, being a liar is factual, objective, and moral. This may feel wrong because we have been under the spell of Hume’s linguistic framework. I suggest we think outside that box.

If you still object to the conclusion—that Donald Trump acted unethically—there are ways to question or even defeat that evaluation. However, you are required to do one of three things to do so.

The first option, which is the easiest and best option in such a case, is to argue that Donald Trump is indeed a liar. However, his lie was not unethical because he was compelled to lie in order to preserve a higher moral value. Let’s say he lied to protect someone’s life. Obviously that option requires you to supply additional facts and reasoning—what is that higher value and how does his lie preserve it.

Your second option is a little more difficult but it’s also a good option. You must reinterpret Trump’s statement under a different factual description. In short, you must find another concept that correctly describes his statement. This is a technique common to politics. The difficulty in doing this explains why political spin doctors are in such high demand and are paid so much money.

Your third option is the most difficult to pull off. Under this option you must provide a convincing argument that being a liar is not wrong—that it’s not immoral. That’s quite a feat to pull off because it contradicts the established meanings in our language.

But our language is always a work in progress. Concepts—ethical and otherwise—are created, refined and re-defined all the time. Such a feat as described in this final option is not a complete impossibility. For example, an “adulterer” has lost much of its moral stigma over time. In fact, the word is seldom used to describe such behavior nowadays.

As a final note, ethicists claim that moral determinations should be action-guiding. Again, action based on that description of Trump can be taken without resort to any extrinsic factors—like the “passions” of Hume’s own “sentimentalist” ethics. Someone could charge Trump with perjury if his lie was under oath or someone could decide to vote for someone else. So what more would we need? We have a descriptive concept, its truth can be verified so it’s objective, and it alone is enough to rely upon as action-guiding. At the end of the day, the description—Trump is a liar—if true, is objective, morally blameworthy and action-guiding. When we see ethics this way it appears down to earth. Again, consultation with Hume’s “sentimentalism” (or some other source) is not needed. However, I must admit that all moral problems are not this easy.

So that’s my quick and dirty refutation of Hume’s confused muddle. I admit that Searle and Kovesi were simplified for this forum—maybe oversimplified. But I hope it inspires some further thoughts on the matter.

1. Searle, John R., Philosophy in a New Century, Cambridge University Press, (2008) p. 166.

2. Searle, John R., Philosophy in a New Century, Cambridge University Press, (2008) p. 166-167 (my italics)

3. Searle, John R., Philosophy in a New Century, Cambridge University Press, (2008) p. 35-36.

4. Searle, John R., The Normative Structure of Human Civilization, Werner Gephart Ed., V. Klostermann, (2017) p. 24. In fact, according to Searle;

5. Searle, John R., The Normative Structure of Human Civilization, Werner Gephart Ed., V. Klostermann, (2017) p. 24

6. Kovesi, Julius, Moral Notions, Routlege & Kegan (1967 ed.); Cybereditions (2004 ed.).

7. Erwin and Tapper, “MacIntrye on Kovesi”, juliuskovesi.com,(2025).

8. Kovesi, Julius, Moral Notions, Routlege & Kegan (1967 ed.) p. 24-25, Cybereditions (2004 ed.) p. 21. (My italics.)

9. Kovesi, Julius, Alan Tapper, ed., Values and Evaluations, Peter Lang Publishing, (2004) p. 36, 38-39.

10. Kovesi, Julius, Moral Notions, Routlege & Kegan (1967 ed.) p. 161, Cybereditions (2004 ed.) p. 115.

11. Kovesi, Julius, Moral Notions, Routlege & Kegan (1967 ed.) p. 57, Cybereditions (2004 ed.) p. 43.

12. Kovesi, Julius, Moral Notions, Routlege & Kegan (1967 ed.), p. 59; Cybereditions (2004 ed.) p. 43. (Emphasis added)

13. Kovesi, Julius, Moral Notions, Routlege & Kegan (1967 ed.), p. 111, Cybereditions (2004 ed.) p. 81.

14. Wittgenstein, Ludwig, Philosophical Investigations, Prop 435, 4th ed., p. 136, Blackwell Publishing Ltd (2009).