Thick Rules and Fat Concepts: A User’s Guide to Philosophical Fog

Or: How to turn “it’s complicated” into an academic paper and get cited for it

There is a venerable tradition in philosophy of not solving problems but reformulating them until they dissolve in a cloud of terminology like a sugar cube in a particularly pretentious cup of tea. The theory of thick rules and thick concepts is among the most elegant specimens of this genus. It sounds approximately the way it feels when someone in a meeting says “Well, it’s not quite that simple” — except that this time it takes three hundred pages and results in a publishing contract with Oxford University Press.

But let us proceed in order.

What are thick concepts?

Bernard Williams introduced a distinction in Ethics and the Limits of Philosophy (1985) that has since enjoyed so much philosophical attention one might almost forget how little it actually explains. The idea is this: there are thin concepts — terms like “good,” “right,” “ought” — that are purely evaluative, meaning they only assess without saying much about the world. And then there are thick concepts — terms like “cruel,” “courageous,” “generous” — that allegedly describe and evaluate simultaneously. When I call someone “cruel,” so the theory goes, I am describing their behaviour (they deliberately cause suffering) and negatively evaluating it at the same time. Description and norm are inseparably fused, like milk and coffee in a latte macchiato that nobody ordered.

Hilary Putnam enthusiastically picked this up in The Collapse of the Fact/Value Dichotomy (2002) and announced that it proved the entire distinction between facts and values had collapsed. The is-ought gap, that notorious obstacle against which philosophers have been banging their heads since Hume? Settled. Because look: “cruel” contains both. Problem solved. Applause.

Which, upon closer inspection, is roughly as convincing as claiming to have solved the problem of gravity by inventing a word that simultaneously means “falling” and “undesirable.”

The miracle of normative self-pollination

For here lies the first problem, and it is not a small one: suppose “cruel” does indeed contain a fusion of description and evaluation. Wonderful. But who says I ought to avoid cruelty?

This may sound like an absurd question, but that is precisely the point. The is-ought problem does not ask whether we have words that somehow sound like both. It asks whether from a description of what is, it logically follows what ought to be. And the mere existence of a word containing both descriptive and evaluative elements answers this question no more than a word simultaneously meaning “cat” and “cute” proves that cats are objectively cute (even though they obviously are, but that is a different debate).

Williams himself was more aware of this than his successors would like. He explicitly warned against selling thick concepts as a solution to the is-ought problem. Putnam took the step anyway, with the confidence of a man who opens an umbrella and declares the weather problem solved once and for all.

To put it in Popperian terms: a genuine solution would need to show why the evaluative component of “cruel” is binding. Instead, it is simply packed into the word and then triumphantly pulled out again, like a magician finding the rabbit he himself hid in the hat. This is not bridging the is-ought gap. It is a petitio principii with linguistic décor.

Which thick concept shall it be, then?

Let us suppose we swallow the theory anyway. Thick concepts contain norms. Fine. But which thick concepts are the right ones?

For different cultures, eras, and communities have decidedly different thick concepts. What Victorian society considered “indecent” — a thick concept par excellence, descriptive and evaluative at once — is a basic right in another society. What a Spartan called “cowardly” was, for a Quaker, a moral imperative. The Japanese concept of giri (duty/loyalty) contains normative elements that in a Western context would partly count as subservience.

If thick concepts deliver norms along with themselves, then different thick concepts deliver different norms. Which means we are standing exactly where we started: we need a criterion to adjudicate between thick concepts. A criterion that cannot itself be a thick concept, because that would be circular. In other words: we need a thin principle to sort the thick concepts. Which is roughly like having a magnificent collection of compasses, each pointing in a different direction, and now needing another compass to work out which one is correct.

And here — like an uninvited guest at a dinner party who turns up at precisely the moment you thought you had finally got rid of him — the theory of thick rules enters the stage.

Thick Rules: The art of turning “it depends” into a theory

The idea of thick rules is essentially this: thin, formal rules — “do not lie,” “keep your promises,” “maximise utility” — do not work because they ignore context. Real moral competence, the theory holds, consists in following rules enriched with contextual knowledge, experience, and practical wisdom. Rules that are, as it were, thick. (The lineage runs from Aristotle’s phronesis through Ryle’s knowing how (1949) to the Dreyfus brothers and their model of expertise.)

A thin rule says: “Do not lie.” A thick rule says: “Do not lie, except when it appears appropriate given the circumstances, where appropriateness arises from the contextual interplay of relevant factors that an experienced moral agent grasps intuitively.”

If you feel a slight unease at this point: yes. Exactly.

For what is happening here is the following: the theory of thick rules claims that moral rules only work when enriched with an indeterminate, not fully articulable, context-dependent layer of practical wisdom. Which operationally means: you can never know whether you are applying the rule correctly unless you are an expert, and the expert is the one who applies the rule correctly.

Hubert Dreyfus and Stuart Dreyfus developed an influential stage model of expertise in Mind over Machine (1986) that has exactly this structure: the novice follows thin rules; the advanced practitioner considers context; the expert acts intuitively and often cannot even explicitly state their reasons. Which sounds like an empirical theory about learning processes — and has genuine merits as one — but as a normative theory produces a remarkable problem.

The expert problem, or: The man who issues his own driving licence

Who is the expert? Well, the one who correctly applies the thick rules. And how do we recognise that they are applying them correctly? Because they are an expert.

If this argumentative structure seems familiar: it is. It is a circle, and not a hermeneutic one (which at least has the decency to be embarrassed about itself), but a cheerful, shameless, tautological circle. It has the logical structure of a man who claims to be an excellent driver and, as proof, presents his own driving licence, which he issued to himself, because he is, after all, an excellent driver.

Jonathan Dancy, the foremost proponent of moral particularism, goes even further and argues in Ethics Without Principles (2004) that there are no general moral rules at all — each situation is unique, and the moral agent must judge anew each time. Which pushes thick rules to their extreme: the rule has become so thick that it encompasses every single case individually, which by definition is no longer a rule but an enumeration. Dancy at least has the honesty to think through what the thick rules theory actually implies.

The immunisation strategy, or: Why the theory can never be wrong

And here we arrive at the actual core of the problem, and it is — to stay with the culinary imagery — rather fat.

The theory of thick rules is not falsifiable. It is not slightly unfalsifiable, it is not unfalsifiable under certain conditions — it is unfalsifiable in principle, and this is not an accident but its central function.

Let us observe the mechanics: someone applies a thick rule and the outcome is bad. What does the theory say? “Then the rule was not thick enough.” Or: “Then the context was not properly grasped.” Or: “Then the agent was not yet an expert.” In no conceivable case can the theory fail, because every counterexample is immediately converted into a confirmation — the case in question simply shows why the rules needed to be even thicker, the context even finer, the expertise even deeper.

Hans Albert coined the term immunisation strategy for this kind of manoeuvre in Treatise on Critical Reason (1968), and he would have had great fun with thick rules, because they are the textbook example. A theory that books every conceivable outcome as confirmation says nothing about the world. It is, to use Popper’s terminology, empirically vacuous. It has the explanatory power of a horoscope and the precision of a weather forecast that says: “Tomorrow there will be weather.”

The great reveal, or: What the theory actually says

Let us set aside the entire terminological apparatus for a moment and ask what the theory of thick rules, stripped of all academic packaging, actually claims. The answer can be summarised in two sentences:

1. Moral decisions are context-dependent and complicated.
2. People with more experience tend to make better decisions.

This is — I hope you are sitting down — not wrong. It is even obviously true. It is so obviously true that my grandmother knew it without ever having heard of Aristotle’s phronesis, Ryle’s knowing how, or the Dreyfus brothers. My grandmother would have said: “Well, you have to judge each case on its merits” and “Ask someone who knows about these things.” Which is exactly the same, only without the three hundred pages and the publishing contract.

The entire theory of thick rules and thick concepts can thus be summarised as follows: there is an old, deep, stubborn philosophical problem (the is-ought gap, the question of how to ground moral norms), and the solution consists in inventing very clever-sounding words that do not solve the problem but act as though they have, while delegating the actual work to a mysterious entity called “experience” or “expertise” that is never specified precisely enough to be testable.

Which, when one thinks about it, is a rather thick trick.

The perfect circuit: How thick concepts and thick rules save (and drown) each other

Let us summarise the beautiful interplay once more, for it has the elegant structure of an Escher staircase — every step looks as though it goes upward, and at the end you are standing right where you started.

Thick concepts are supposed to bridge the is-ought gap by fusing description and evaluation. But which thick concept is the right one depends on context. Grasping the context correctly requires thick rules — rules enriched with practical wisdom. But who applies the thick rules correctly is determined by an expert. The expert is recognised by their correct deployment of thick concepts. Which presupposes knowing which thick concept is the right one. Which depends on context. And so on.

It is a circle within a circle within a circle, like a Russian matryoshka doll that you open only to find inside — the same doll.

Why all this still makes careers

One might wonder why a theory that essentially says “it’s complicated” while absorbing all counterexamples like a philosophical black hole is so popular. The answer is simple: precisely because of that.

A theory that can never be wrong is immensely comforting. It offers intellectual comfort without requiring intellectual labour. One can say, in any seminar, at any panel discussion, in any moral dilemma: “Well, that requires phronesis” or “That depends on the thick concept” or “An expert would judge differently here” — and in every case it sounds clever, is irrefutable, and requires no one to actually commit to a position.

It is, in other words, the perfect academic theory.

Why this is not harmless

One might shrug at this point and say: so what? Thick concepts, thick rules — sounds a bit woolly, but where’s the harm? It’s just academic philosophy, who reads it anyway. This is roughly the argument one hears about religion too: it’s nice, gives people comfort, and we wouldn’t want to abolish Christmas trees either.

The problem is: pseudo-solutions are not neutral. They are actively harmful, on at least two levels.

First: whoever believes the question is settled stops looking. If academic philosophy contents itself with the notion that “cruelty” as a thick concept somehow settles the is-ought gap and the rest is a matter of phronesis, then no one seriously searches for a solution that meets the minimum standards of rational theory-building — that is, consistency, openness to criticism, empirical testability, and explicit falsification conditions. It is not ruled out that such a solution exists. But it is buried under the debris of pseudo-solutions, much like a map that might actually lead to the treasure disappears in a library full of beautifully drawn fantasy maps. The result is not merely academic stagnation. It is the active prevention of progress — because every approach that might point in the right direction must first fight its way through shelf-metres of literature that declares the problem either solved or unsolvable in principle.

Second — and this is the practical damage: thick rules delegate moral judgments to experts without providing a testable criterion for who counts as an expert. In practice, this means: whoever passes as a moral expert determines the moral rules. And who passes as a moral expert depends on institutional power, social status, rhetorical skill, and the ability to deploy context-saturated formulations in such a way that disagreement appears as a sign of insufficient phronesis. This is not a theory of morality. It is a theory of authority with a philosophical veneer. The wrong expert, whom one follows because one cannot know which is the right one, does not merely deliver a different opinion — they deliver a prescription for action that by definition cannot be criticised, because every criticism is booked as evidence of one’s own incompetence.

Third: the pattern is not confined to ethics. Wherever pseudo-solutions occupy the place of real solutions, the same effect arises. The history of science is full of epochs in which a comfortable consensus blocked the search for better theories — not because the better theory was unknown, but because the willingness was lacking to recognise the consensus for what it was: a collective immunisation strategy against uncomfortable questions.

The theory of thick concepts and thick rules is such a strategy. It does not immunise a particular norm against criticism, but something more fundamental: the question itself. It transforms “How can the is-ought problem be solved?” into “The is-ought problem is solved (or unsolvable), and whoever doesn’t see this lacks sufficient phronesis.”

It would be time to reopen the question.

References:

• Albert, H. (1968). Traktat über kritische Vernunft [Treatise on Critical Reason]. Tübingen: Mohr Siebeck.
• Bunge, M. (1974–1989). Treatise on Basic Philosophy, 8 vols. Dordrecht: Reidel.
• Bunge, M. & Mahner, M. (2004). Über die Natur der Dinge. Stuttgart: Hirzel.
• Dancy, J. (2004). Ethics Without Principles. Oxford: Oxford University Press.
• Dreyfus, H. & Dreyfus, S. (1986). Mind over Machine: The Power of Human Intuition and Expertise in the Era of the Computer. New York: Free Press.
• Putnam, H. (2002). The Collapse of the Fact/Value Dichotomy and Other Essays. Cambridge, MA: Harvard University Press.
• Ryle, G. (1949). The Concept of Mind. London: Hutchinson.
• Williams, B. (1985). Ethics and the Limits of Philosophy. Cambridge, MA: Harvard University Press.