Lecture 1 — Motivating Relativism

John MacFarlane (jgm@berkeley.edu)

Context and Content Lectures, Paris, 2010

What are we doing here?

Three ways to look at the work I’ll be describing in these lectures:

Theories of content

A theory of the contents of mental states and speech acts provides an answer to the following questions:

We don’t presuppose that there is a single kind of content for all speech acts and mental states. But there must be enough systematic connections to make a force/content distinction worth while.

Consumers of theories of content

I distinguish the theory of content proper from theories that use or presuppose a theory of content:

Constraints on theories of content

These consumers provide the constraints on theories of content. A familiar example:

Much analytic philosophy is theory of content.

What I’ll argue

I’ll argue that there are possible contents of thought and talk that go beyond those countenanced by orthodox theories. Theories of content should allow for contents that are assessment-sensitive. Otherwise their consumers will be forced to adopt distorting accounts of our thought and talk in certain areas.

Comparison to the usual debates

Viewed in this light—as a plea for more contents—my position has something in common with the “liberal” positions in the following familiar debates:

unstructured sets of conditionsstructured propositions

But the problem I’m focusing on is orthogonal to these debates. For none of the above liberal positions gives us assessment-sensitive contents. We’ll see why in Lectures 3–5.

What contents are assessment-sensitive?

A decent case can be made for the assessment sensitivity of the contents expressed by the following:

I develop this case in the second part of my manuscript, but here in Paris I’ll only have time for the first, theoretical part.

An example

Dulce de leche is a form of milk caramel that the Argentines particularly love. Not me. I don’t have such a sweet tooth. I say:

Dulce de leche is not tasty!

And I believe what I say. How should we understand the content of my belief and assertion?

How we use “tasty”

By what method did I decide that dulce de leche was not tasty? I followed this procedure:

If you know first-hand how something tastes, call it “tasty” just in case its taste is pleasing to you, and “not tasty” just in case its taste is not pleasing to you.

I think this pretty closely captures our use of “tasty.” “Tasty” is a pretty simple evaluative term — not the kind gourmands like to use.

How we use “tasty”

If you’re sceptical that TP guides our use of “tasty,” consider how odd it would be to say:

  1. I’m not sure whether espresso is tasty, but I hate how it tastes.
  2. I’m sure durian is tasty, but I can’t stand how it tastes.
  3. I love orange juice and hate tomato juice. But I’m told on good authority that tomato juice is tastier, and I believe it.


The fact that tastiness can be applied on the basis of our reactions does not, itself, tell against objectivism — the view that “tasty” has a standard possible-worlds intension.

Compare: “red”, “salty”

Against objectivism

Given the extensive and obvious disagreement in our tastiness judgements, the TP can be reliable at best for a small number of us. And none of us has any independent grounds for thinking herself better placed than these others to make tastiness judgements.

Objectivism must attribute unreflective chauvinism to everybody.

Compare our attitude in the face of disagreement about what is “red” or “salty.”

colorblind test

Educating tastes

The fact that you can “educate” your tastes is no argument for objectivism, since even when we know that our tastes are going to be changed as a result of a process of training, we do not refrain from applying “tasty.”

Apple tasting

Alternatives to objectivism

A good alternative must explain why “tasty” is governed by TP.

Two classic options:


assimilates “This is tasty” to “I/we like how this tastes”


assimilates “This is tasty” to “Yum!”


Girona is near BarcelonaDulce de leche is tasty to Sam
Girona is near ParisDulce de leche is tasty to Eduardo
Girona is near usDulce de leche is tasty to us
Girona is nearDulce de leche is tasty

Weak and strong

Weak contextualism

“X is tasty” can be used to express the proposition that X is tasty to Y, for some taster Y.

Strong contextualism

“X is tasty” is always used to express the proposition that X is tasty to Y, for some taster Y.

Strong Contextualism is required to block the case for relativism.


“Tasty” is a gradable adjective. Some kind of contextual sensitivity comes with gradability.

Even if everyone agreed about the extension of “tastier,” there would still be the question: how tasty does something have to be to fall into the extension of “tasty”? And this depends on context.

What is at issue here is whether there is an additional dimension of contextual sensitivity, affecting relative tastiness judgements.

Agreement and Disagreement

Eduardo: Dulce de leche is tasty.
John: No/I disagree/you’re mistaken, it’s not!

Explaining disagreement

The disagreement targets something other than the proposition asserted (e.g. a presupposition).

A: Your wife is very beautiful.
B: No/you’re mistaken. She’s not.
B: No/you’re mistaken. We’re not married.

Eduardo: Dulce de leche is tasty.
John: No/I disagree/you’re mistaken, it’s not!
John: No/I disagree/you’re mistaken, I don’t share your tastes.


John, age 10: Fish sticks are tasty.
Pedant, now: When you were 10, you asserted that fish sticks were tasty. Do you stand by that?

Answer A: No, I don’t. I retract that claim.

Answer B: Yes, I do. All I meant to assert was that fish sticks were tasty to me then. And they were.

The contextualist can readily explain Answer B, but not Answer A.

The Idealization strategy

Explain disagreement and retraction by taking the content asserted to involve an idealizing element:

“Dulce de leche is tasty” → that dulce de leche would taste good to me if my tastes were suitably improved and idealized.

But then TP again seems unreliable.


Explain disagreement and retraction by taking the content asserted to concern a group rather than an individual:

“Dulce de leche is tasty” → that dulce de leche tastes good to us (speaker + audience + relevant others).



Explain disagreement and retraction by taking the content asserted to concern a component of the conversationally shared “scoreboard,” which can evolve through accommodation.

“Dulce de leche is tasty” → that dulce de leche tastes good according to the standard of taste governing this conversation.

This explains why Eduardo might persist, even in face of my disagreement. It is his way of trying to “push” the standard.


Can’t make sense of continued, clear-eyed disagreements of taste.

A: France is hexagonal.
B: No it’s not. Look at these wavy lines.
A: I see the wavy lines, but that’s irrelevant. France is hexagonal.
B: It most certainly isn’t.
A: You haven’t convinced me.
B: Nor you me.

Classical Expressivism

In saying “this is tasty,” one is expressing, not asserting, one’s liking for the food.


The presence of an ethical symbol in a proposition adds nothing to its factual content. Thus if I say to someone, ‘You acted wrongly in stealing that money,’ I am not stating anything more than if I had simply said, ‘You stole that money.’ In adding that this action is wrong I am not making any further statement about it. I am simply evincing my moral disapproval of it. It is as if I had said, ‘You stole that money,’ in a peculiar tone of horror, or written it with the addition of some special exclamation marks. (Ayer, Language, Truth, and Logic)

Disagremeent again

Expressivism vindicates disagreement only in a weak sense. Ayer concedes this:

Another man may disagree with me about the wrongness of stealing, in the sense that he may not have the same feeling about stealing as I have, and he may quarrel with me on account of my moral sentiments. But he cannot, strictly speaking, contradict me. For in saying that a certain type of action is right or wrong, I am not making any factual statement, not even a statement about my own state of mind. I am merely expressing certain moral sentiments. And the man who is ostensibly contradicting me is merely expressing his moral sentiments.

Disagreement in attitude (towards dulce de leche), but not in claim.


The force-content distinction

Classical expressivism denies that “Dulce de leche is tasty” expresses a content at all. Instead of giving an account of the significance of asserting “Dulce de leche is tasty” by feeding the content of the sentence used into an account of assertion, it offers a direct account of the significance of this performance.

A significant loss of generality!

Loss of generality

  • “Is that tasty?”
  • “If only it were tasty!”
  • “He believes that licorice is tasty.”
  • “He wishes that licorice were tasty.”
Propositional anaphora
“This fish is tasty.” “Yes, that’s true/No, that’s not true/Sam said that too.”
  • “If that’s tasty, he’ll eat it.”
  • “That might be tasty.”
  • “There are no tasty cookies in that jar.”

Blackburn’s response

Blackburn attempts a compositional account of the attitudes expressed by compound sentences.

Mixed cases

So we end up with an expressive account of everything, supplanting truth-conditional semantics instead of supplementing it.

Gibbard’s response


In effect, reintroduces a force/content distinction, co-opting all the machinery of truth-conditional semantics.

The contents are sets of world/normative-system pairs — each is a “completely opinionated credal-normative state”.

“X is forbidden” holds at (w, n) iff “X is forbidden by n” holds at w.

We then use standard composition rules. For example conjunction is intersection, disjunction is union.

An argument is valid iff the conclusion holds at every (w, n) pair where the premises hold.

To assert that p (whether p is factual, normative, or mixed) is to express the state of mind of accepting or judging that p; and to judge that p is to “rule out” certain “combinations of normative systems with factual possibilities” - all those combinations not contained in content of p.

Whence expressivism?

We now have a uniform notion of content for factual, normative, and mixed claims. (The need for compositionality forced this on us.)

These contents can be asserted, believed, supposed, etc.

What started as a view that denied that “Dulce de leche is tasty” expresses a content has turned into a view that gives this sentence a content, but broadens the space of possible contents.

Once we’ve gotten to the point where we acknowledge the need for a new kind of content, we’ll be asking the kinds of questions that lead to assessment sensitivity. (Relativism is expressivism done right.)


But isn’t this pernicious nonsense?

Haven’t we learned that truth relativism is incoherent?

On Friday, we’ll look at general philosophical objections to truth relativism.

The draft chapters that are the basis of my lectures are available online, at: