Context and Content Lectures, Paris, 2010
I will use the term proposition for the contents of assertions, beliefs, and other attitudes like suppositions and conjectures.
The term content gets its meaning here from the force/content distinction. A content is an abstract entity that we can use to represent what is in common between attitudes or speech acts with different modes.
Propositions are also taken to be the “primary bearers of truth values.” Other things that have truth values, such as sentences, have them because of the propositions to which they are related. We will vindicate this claim.
I do not assume that we can do everything we want to do with just one kind of content. I want to leave it open, for example, that we might need sometimes to individuate belief contents as the temporalist suggests, and other times as the eternalist suggests.
I assume nothing here about what kinds of entities contents are (except that they are abstract)—whether they are structured or unstructured, for example.
Some semanticists take propositions to be the semantic values of sentences. I won’t assume this, but what I say about propositions should help us
The character (linguistic meaning) of an expression is a rule for determining its content in any given context.
The content of a sentence at a context is the proposition it expresses — what would be “said” if the sentence were uttered at that context.
The content of a subsentential expression is its contribution to the content of expressions of which it is a part (for a proper name, an object, for a predicate, a property).
So the character of “here” might be the rule:
Its content at a context c is the location of c.
Kaplan wants to say that a directly referential expression, like the indexical “here”, “designates the same object in all circumstances.”
But this can lead to confusion. Surely, you might say, there are circumstances in which “here” denotes Paris, and circumstances in which “here” denotes Berkeley!
We need a distinction:
We do not mean that the expression could not have been used to designate a different object. We mean rather that given a use of the expression, we may ask of what has been said whether it would have been true or false in various counterfactual circumstances, and in such counterfactual circumstances, which are the individuals relevant to determining truth-value. Thus we must distinguish possible occasions of use—which I call contexts—from possible circumstances of evaluation of what was said on a given occasion of use.
“Here” can denote different objects (have different contents) at different contexts of use.
But when we fix on a particular content expressed using, say, “It is cold here,” and ask whether this content would have been true if things had been different, the answer depends only on what the temperature would have been at the original location of utterance, and not on the counterfactual location of utterance.
Even if I’d been in the Sahara desert, it would have been cold here.
I do not exist.
There is no context at which this sentence expresses a truth.
Yet the falsehood it expresses is one that is true at some counterfactual circumstances.
By [“circumstances”] I mean both actual and counterfactual situations with respect to which it is appropriate to ask for the extensions of a given well-formed expression. A circumstance will usually include a possible state or history of the world, a time, and perhaps other features as well. The amount of information we require from a circumstance is linked to the degree of specificity of contents, and thus to the kinds of operators in the language.
Kaplan treats tenses as intensional operators.
Since he is doing two-stage semantics, these must operate on contents.
Will: Joe bakes a cake.
“Will” shifts the time, so the content of “Joe bakes a cake” must be time-variable.
This can be blocked by rejecting Kaplan’s syntactic assumptions and arguing that tenses are not operators (King).
Or it can be blocked by rejecting two-stage semantics (Lewis).
Kaplan is sometimes taken to hold that the presence of an operator that shifts X is both necessary and sufficient for the inclusion of X in circumstances.
But I only see a clear commitment to sufficiency.
Might there be other considerations, independent of operators, for including X in circumstances?
we may ask of what has been said whether it would have been true or false in various counterfactual circumstances
Sam is the tallest man in his town.
Would that have been true if his legs had been amputated?
Sam is the tallest man in his town.
Was that true last year?
But we might worry about overgeneration:
Sam is the tallest man in his town.
Is that true in your imagination?
Sam is the tallest man in his town.
Is that true for Harry too?
What we say about circumstances affects questions about the individuation of contents — when do we have one, and when do we have many?
Is there a single proposition that Kenneth Starr is alive, or just many of the form that Kenneth Starr is alive during interval i?
This might matter outside of semantics. (Consider, for example, an account of belief retention through time.)
Decisions we make about propositional truth have ramifications in propositional attitude psychology, and are therefore constrained in part by considerations from propositional attitude psychology.
Suppose we are considering people whose language originally lacked modal operators (world-shifters), but evolved to contain them.
Should we say that speakers of the language came to express different propositions with their non-modalized sentences after the modal operators were added?
Or that they expressed the very same propositions, but that truth for these propositions came to be relative to possible worlds, when before it was not?
Neither option seems attractive, but the idea that only world-shifting operators can justify relativizing propositional truth to worlds forces us to choose between them.
Could the circumstances of evaluation to which propositional truth is relative include coordinates besides worlds and perhaps times—for example, tastes, standards of precision, information states, moral codes, or epistemic standards?
Kaplan himself takes a fairly permissive stance:
What sorts of intensional operators to admit seems to me largely a matter of language engineering. It is a question of which features of what we intuitively think of as possible circumstances can be sufficiently well defined and isolated. If we wish to isolate location and regard it as a feature of possible circumstances we can introduce locational operators: ‘Two miles north it is the case that’, etc. … However, to make such operators interesting we must have contents which are locationally neutral. That is, it must be appropriate to ask if what is said would be true in Pakistan. (For example, ‘It is raining’ seems to be locationally as well as temporally and modally neutral.)
Why not, then, contents that are taste-neutral?
Let’s dismiss an argument that I have sometimes encountered:
By ‘proposition’ we just mean something that partitions the possible world-states into those at which it is true and those at which it is not true. So if your ‘propositions’ do not do that, they are not propositions properly so-called.
There is a real issue here that verbal stipulation can’t settle: do the contents of assertions, beliefs, and other so-called “propositional attitudes” have truth values relative to worlds only, or relative to worlds and something else?
If someone wished to cite, say, ‘The total number of inhabitants of the German Empire is 52 000 000’, as a counter-example to the timelessness of thoughts, I should reply: This sentence is not a complete expression of a thought at all, since it lacks a time-determination. If we add such a determination, for example, ‘at noon on 1 January 1897 by central European time’, then the thought is either true, in which case it is always, or better, timelessly, true, or it is false and in that case it is false without qualification.
By “thought” here, Frege means essentially what we mean by “proposition.” So this might seem to support opposition to time-neutral (taste-neutral, etc.) propositions. But that is doubtful.
Frege’s discussion seems to confuse two very different issues:
Before and after the passage just quoted, Frege is mainly concerned with (1).
Whereas ideas (in the psychological sense of the word) have no fixed boundaries, but are constantly changing and, Proteus-like, assume different forms, thoughts always remain the same. It is of the essence of a thought to be non-temporal and non-spatial.
But (2) is orthogonal to (1). A temporalist can accept that propositions are abstract, timeless entities, and endorse (1).
Moreover, Frege doesn’t really give a reason for thinking that a complete expression of a thought must include a time determination.
There’s a way to make the argument seem less confused, but it shows how useless Frege is for the modern opponent of time-relative contents. (I was helped here by Giorgio Volpe.)
Suppose we take Frege to be assuming that truth values are intrinsic properties of propositions, not properties propositions possess because of their relation to other things.
On this assumption, the only way a proposition can have different truth values at different times (2) is if the proposition changes in its intrinsic properties (1).
But then, a parallel line of thought would show that if propositions have different truth values at different possible worlds, propositions are contingent entities. This also conflicts with the idea that they are abstracta.
Frege held that sense determines reference. On a strong (and I think correct) way of reading the claim, it implies that in any case where you have the same sense, it has the same reference — reference is entirely determined by sense.
Since the sense of a sentence is a thought, and the reference is a truth value, this implies that thoughts have their truth values intrinsically.
So Frege is no friend to those who want to drive a wedge between worlds and times. (Indeed, he never took alethic modality seriously.)
Propositions are supposed to be the contents of beliefs and other propositional attitudes.
But a full specification of what someone believes must determine conditions for the belief to be accurate.
Thus, for example, if we don’t know whether the accuracy of Sam’s belief that it is 0 ∘ C depends on the temperature in London on Tuesday or the temperature in Paris on Wednesday, we don’t yet have the full story about what it is that Sam believes.
A condition for a belief to be accurate is a condition on a context: what must a context be like in order for a belief at that context to be accurate?
A standard possible-worlds content fixes accuracy conditions for belief as follows: a belief that p in context c is accurate iff p is true at the world of c.
But a time- and location- neutral content fixes accuracy conditions in the same sense: a belief that p in context c is accurate iff p is true at the world, time, and location of c.
Once we accept the relativity of propositional truth to worlds, we have accepted a kind of “incompleteness.” We have accepted the idea that both the content of an assertion or belief and its context must be taken into account in assessing it for accuracy. The question is just which features of which contexts must be taken into account, and how.
One might respond to these considerations by bringing the world of the context of use into the content of Sam’s thought (Schaffer). This allows us to say that the contents have absolute truth values.
This does violence to ordinary ways of individuating thought content. Intuitively, though, Sam could have had a thought with the very same content even if the world had been very different.
Moreover, bringing the world of the context into the content of Sam’s thought would make this content a necessary truth about this possible world, rather than a contingent truth about the weather in Paris.
We should not say, then, that Sam’s thought is about the world of use. It is not about any particular world. Acknowledging the fact that it depends for its accuracy on the world of use, we may adopt John Perry’s terminology and say that it concerns the world of use.
In ordinary talk, we don’t characterize claims as “true-in-w,” or as “true-in-w-at-t-on-s,” but as “true” (simpliciter).
But this no more shows that propositional truth is not relative to parameters than the fact that we normally say it’s “3 PM,” and not “3 PM Pacific Daylight Time,” shows that the time of day is not relative to a time zone.
More in a bit, when we get to monadic “true”.
How can we get sentence truth at a context out of a two-stage semantics?
If c is a context, then an occurrence of [a sentence] φ in c is true iff the content expressed by φ in this context is true when evaluated with respect to the circumstance of the context.
Dropping the assumption that there will always be a unique “circumstance of the assumption,” we can generalize:
A sentence S is true at context c iff the proposition expressed by S in c is true at all circumstances of evaluation compatible with c.
What “compatibility” amounts to must be worked out in detail for each semantic theory.
It is also useful to define propositional truth relative to a context:
A proposition p is true at a context of use c iff p is true at all circumstances of evaluation compatible with c.
We can now rephrase our definition of sentence truth at a context as follows:
A sentence S is true at context c iff the proposition expressed by S in c is true at c.
The context of use plays two distinct roles in the definition of truth at a context:
A sentence S is true at context c iff the proposition expressed by S in c is true at c.
determining which proposition a sentence expresses
selecting which circumstances of evaluation are relevant to the truth of an occurrence of a sentence at the context.
So there are two distinct ways in which a sentence can be context-sensitive. It can be sensitive to a feature of context because that feature plays a content-determining role, or because it plays a circumstance-determining role.
A nonindexical contextualist about “tasty” holds that
No assessment sensitivity here.
To make room for asssessment sensitivity in our two-stage semantic framework, we need truth to be relative to both a context of use and a context of assessment:
A proposition p is true at as used at c1 and assessed from c2 iff p is true all circumstances of evaluation compatible with < c1, c2 > .
A sentence S is true as used at c1 and assessed from c2 iff the proposition expressed by S in c1 as assessed from c2 is true as used at c1 and assessed from c2.
The relation of “compatibility” now holds between circumstances and a pair of contexts—a context of use and context of assessment. As before, it must be defined on a per-application basis. Example:
A circumstance < w, s > is compatible with < c1, c2 > iff w is the world of c1 and s is the aesthetic standard relevant at c2.
An expression is assessment-indexical iff it expresses different contents relative to different contexts of assessment.
An expression is F-assessment-indexical iff the content it expresses as assessed from c depends on the F of c.
A proposition is assessment-sensitive if its truth value as used at c1 and assessed from c2 depends on features of c2.
A proposition is F-assessment-sensitive if its truth value as used at c1 and assessed from c2 depends on the F of c2.
There are assessment-indexical sentences.
There are assessment-sensitive propositions.
A truth-value relativist about “tasty” holds that
Three positions can be usefully compared using the terminology of “about” and “concerns”:
“True as used at c1 and assessed from c2,” like “true as used at c1,” is a technical term that has a role in our theories of language use, not a term in ordinary use.
The ordinary truth predicate is a monadic predicate that applies to propositions.
Let’s apply our semantic machinery to understand how it works!
The extension of “true” at a circumstance of evaluation e is the set of propositions that are true at e.
Given this clause, every instance of the Equivalence Schema will be true at every circumstance of evaluation, and hence also at every context of use and context of assessment.
If the language can express any assessment-sensitive propositions, “true” will also be assessment sensitive, since if p is assessment-sensitive, the proposition that p is true must be assessment-sensitive too.
This shows what is wrong with the thought that relativism about truth amounts to nothing more than an ordinary sort of contextualist (use-sensitive) semantics for “true.”
Suppose S1 is true in Ψ and S2 is false in Θ . And suppose that S1 and S2 express the same proposition, p.
S1 and S2 differ in truth value (by 1).
If two sentences differ in truth value, they differ in truth conditions.
If two sentences express the same proposition, then they have the same truth conditions.
So if two sentences differ in truth value, they do not express the same proposition (by 3, 4).
So S1 and S2 do not express the same proposition, contrary to our supposition (by 2, 5).
1. Suppose S1 is true in Ψ and S2 is false in Θ . And suppose that S1 and S2 express the same proposition, p.
We’re talking about sentences being “true in Ψ” and “false in Θ ,” so presumably we should think of Ψ and Θ as contexts. (N-S says: theories or social groups.) But are we to think of these as contexts of use or contexts of assessment?
Do they express the same proposition in some single context? Or the same proposition in their respective contexts?
Perhaps we can pin this down by looking at what the argument requires.
Like Socrates, Newton-Smith seems to have dropped the relativizing qualifiers (“in Ψ”). So we’d better restore these.
3. If two sentences have different truth values at (possibly) different contexts, they differ in truth conditions.
This means that we must construe (5) as:
5. If two sentences have different truth values at (possibly) different contexts, they do not express the same proposition.
(5a) If two sentences have different truth values as used at (possibly) different contexts, they do not express the same proposition.
(5b) If two sentences have different truth values as assessed from (possibly) different contexts, they do not express the same proposition.
Neither is true.
“It snows in Germany” as used in contexts in different worlds; or (on temporalist account) “Socrates is sitting” as used at different times.
“This chile is tasty” as assessed by people with different tastes.
No non-question-begging reason has been given to accept (5b), and (5a) is false even in non-relativist frameworks.
Newton-Smith’s argument exemplifies a general tendency in much of the literature on relative truth. Terms like “true in,” “true for,” and “truth-conditions” are deployed without any sensitivity to the various kinds of relativization of truth that can be used in semantics.
As we have seen, proper statement of a relativist position requires some care. A general argument against relative truth needs to take the same care.
At the end of Lecture 2, we had this list of tasks:
|Provide an account of propositions that|
allows them to be merely “relatively true.”
|Explain how the ordinary monadic truth predicate|
is related to the relativized truth predicate.
|Explain what one is doing in asserting|
propositions, if not putting them forward as true
|Explain what the relativized “true for/at” means.|