Wolfgang Schwarz

Hi,

I am puzzeld, either you or I are missing something here, most probable it is me:
Every utterance has a time and a world (sometomes calles "a context of evaluation"), otherwise we could not find out its truth value.
Is it as simple as that? (I leave it open, if it is one context, several ones,or time and place and person that makes the utterance, index/indices , and what "a world" means here, just to get something to start with).

M.

Hey, not sure what your proposal is. The puzzle is how to complete: "what Frege said is still true today iff ...". Three proposals that don't work are:

1) ... iff the (secondary) proposition Frege expressed is true in the present context;

2) ... iff the primary proposition Frege expressed is true in the present context;

3) ... iff the sentence Frege uttered is true in the present context.

Shouldn't "I" in Frege's utterance be taken as rigid reference, so that evaluation in present context doesn't affect what "I" means there?

Wo,

I am kind of reluctant to assume anything about secondary/primary propositions, not sure what you allude to. (That is really my ignorance.)

However if the basic question is to find some permutation of
"What Frege said is still true today iff what Frege said is (still (?)) true today", I would go for (3) if the sentence (utterance, content of utterance (??)) Frege uttered is true in the present context (i.e. today).
What will be the problem?
Of course all this terminology would need some systematic cleaning up ... do you feel like doing this kind of work on "content", "possibilities", "at a world", A,B,C,XYZ-intensions, "2nd propositions", "sentences", their types, "utterances at t in w", contents of these utterances etc.?

M.

Dear Wo,

your first proposal is what Wolfgang Kunne in his "Conceptions of Truth" (Oxford, 2003) calls "eliminitavist eternalism". This is the position that the proposition which is expressed by a sentence which contains context-dependent elements, like "today", "here", "I", etc., is devoid of these elements.

For example, the sentence

1) Today I'm tired.

actually expresses the proposition:

1*) On the 21st of May M.L. is tired.

1*) is basically what Quine called an "eternal sentence": a sentence whose truth-value, or the truth-value of the propositions which it expresses, does not change throughout time (or in different worlds).

One of the problems with this view is that the relation between 1) and 1*) cannot be so strong as to say that they are equivalent (whatever that means). This is because I may well have opinions about the present day, and utter 1), but I may be oblivious as to what day it actually is today. So in that case, when I utter 1), it doesn't seem that I actually "mean" 1*), since I don't even know it is the 21st of may today.

Another objection is that the sentence

2) Today is the 21st of may.

seems to be informative, while the context-indepentent proposition it expresses:

2*) The 21st of may is the 21st of may.

is trivial.

A problem with the opposing view, namely that sentences (or what is expressed by them) is temporally indeterminate, Kunne called "the problem of wish fulfillment". I quote (p.312):

"Consider the following argument:

(W1) What Bertie wished at 10 a.m., 18 May 1906, was that it would soon stop raining in London.
(W2) What Ann wished at 10 a.m., 18 May 2000, was that it would soon stop raining in London, and what she wished came true 5 minutes later.
(W3) Therefore, at 10:05 a.m., 18 May 2000, what Bertie wished came true."

If we don't interpret Bertie's wish as specifically pertaining to the time and place it is "wished" about (i.e., eliminate the context-dependent elements from the sentence the wish is expressed in, and transform it into a proposition which does not change its truth value over time or place), we have to accept the above reasoning, which seems intuitively invalid.

M.L.

Tanasije: yes, "I" is rigid, but so is "today". For the sense in which what Frege said is no longer true, we have to keep the referent of "I" fixed, but not that of "today".

M: The problem I see is that none of the common notions of content can do the work needed here. On one notion (roughly corresponding to secondary intensions, and the first option M.L. mentions), we'd have to keep both the referent of "I" and of "today" fixed, which would make the content expressed by (1) *that more people on this-and-that say in 1867 died of tuberculosis than of cancer*, which is still true. On another notion (roughly corresponding to primary intensions, and perhaps the second option M.L. mentions), we keep neither fixed, so that (2) expresses the centered proposition true of anyone who is not particularly well-known among philosophers.

One can of course invent a new kind of content on the spot that (by stipulation) does the job, but that seems a little undermotivated to me. Though Dave Chalmers tells me that Brit Broogard's independently motivated theory of tensed propositions might help.

My own, somewhat tentative view is that when we say that a sentence or utterance is true at or for x, we only mean that a related sentence or proposition in which some parameter or other is replaced by x is true. Compare "that's true for me too" in response to "when I first read Kant, I thought I understood everything, but the closer I look at it, the less I understand" (or to that same sentence with "I" replaced by "Fred").

M.L.: Thanks for the Kuenne reference. Except that in the present case, neither of the two options gives the right result, I generally prefer the second. The problem of wish fulfillment looks similar to problems Scott Soames likes to raise about this view. The lesson I take from them is that the relation between that-clauses and contents (of either sentences or mental states) is much less straightforward than it might seem on first sight.

Notice that the same problem arises for the eternalist: suppose Bertie in world W1 and Ann in world W2 both wish at 10 a.m., 18 May 2000, that it would soon stop raining in London. And suppose that what Ann wished came true 5 minutes later. We don't want to conclude from that alone that Bertie's wish also came true.

The puzzle disappears if we move to a more explicit characterisation of the centered attitude contents, without that-clauses: the content of Bertie's wish is the set of centered worlds that are located at some place and time where it soon stops raining. This is also the content of Ann's wish. Ann's situation was actually of that kind, so her wish came true. Bertie's situation was not, so his didn't.

Dear Wo,

I think that maybe the whole endeavour of finding a single entity, which is rigidly designated by 'what is said' and which is sometimes true and sometimes false, is mistaken. Compare 'The Pope was Polish, but now he is German', where neither one man changed nationality, nor something was said about the holy see itself. Maybe we should read 'what was said was true but it is false now' as elliptical for something like 'that he (Frege) was not well-known among philosophers, was true then, but that he (Frege) is not well known among philosophers, is false now' - where the different occurrences of 'that he is not well known among philosophers' contribute differently to what is reported (o.k. - we have to change tenses, but I don't know if that is significant or merely an agreement-fact). To wheel out some ad hoc theory: 'what was said' goes proxy for a description of the expressed proposition - a description that could pick out different propositions depending on its syntactic embedding. Which description it goes proxy for should be a matter of pragmatics - there obviously is a reading according to which your sentence 'this was true in 1871, but it is no longer true now' comes out trivially false (in which 'this' picks out what was said as something like 'that in 1869 more people used to die of tuberculosis than of cancer'). I also think that there is no big difference between 'today' in (1) and 'I' in (2). What was said with (2) is false of Frege and true of you (according to yourself:).

Hi Bernd, I completely agree. But I think what is to blame is not only a naive conception of "what is said". One can just as well say that Frege's utterance is no longer true, or one could read aloud his note and comment that *this* is no longer true. It seems that whenever we evaluate an utterance for truth in any of these ways (or for truth-now, or for truth-actually, or for truth-for-Fred), there is this kind of indeterminacy, context-dependence and array of options.

So I'm inclined to put the blame on a naive conception of "true": we shouldn't assume that there is a close connection between whether or not an utterance is actually true and people's judgment about whether or not it is true (or true-now, or for true-actually, or true-for-Fred).

If that is correct, there are two quite different types of relativism: relativism about "truth"-as-used-in-English, and relativism about truth on a certain theoretically significant precisification. And while there is plenty of evidence for, say, relativism about epistemic modals in the first sense, this entails nothing about relativism in the second sense.

Hi Wo,

I am not sure, whether I have completely understood your attack on the naive notion of truth and Bernd’s attack on the naïve notion of what is said, but I think the puzzle could be solved by a distinction between what is said/is true simpliciter and what is said/is true of something. Let me explain.

I tend to think that when we say that Frege’s utterance is no longer true, what we mean is that what he uttered is no longer true. The same holds for the case with the note. (You cannot take it for granted that *this* refers to the note in the case you describe. Someone can read out a note that was written one minute ago and you can comment by saying ‘This is what I have told you for ages’). It is noteworthy that in both cases that constitute your puzzle we would specify what we mean by *this* in ‘This is no longer true’ by a that-clause, in the first case by ‘that more people die of tuberculosis than of cancer’ and in the second by ‘that Frege is not particularly well-known among philosophers’. So what do these ‘that’-clauses designate? Here is what I take Kuenne’s position to be (CoT 315 f.): When we say of something designated by a ‘that’-clause that it is no longer true, or that it was true at such-and-such time, the thing designated is an open proposition (or a ‘proposition matrix’). You get an open proposition when you kick out of an eternal proposition the part that specifies the time. If we predicate truth simpliciter we predicate it of an eternal propositions. Open propositions, however, are not true simpliciter but rather true of something (in this case a time). Hence the open proposition that Frege is not particularly well-known among philosophers can be true of (at) the year 1869 but false today. One could transfer Kuenne’s model to the ‘true of someone’-talk. A said to B: ‘You are a fool’, and B answers: ‘That’s also true of you’. If both are right, it’s true of A that he is a fool, and this is also true of B. The ‘that’-clause in the last remark designates the open proposition that you get by kicking out of the proposition that A is a fool the part that specifies the reference to A. (Such an open proposition is something very similar to a property. Note that it is not unusual to refer to properties by ‘that’-clauses, e.g. in ‘That he is curageous is his most valuable property’. However, properties do not seem to be true of something.)

Corresponding to the distinction between truth simpliciter and truth-of-something, one should distinguish between what-is-said simpliciter, and what is said of something. In the A-and-B case, what A said of B is what B said of A, and that obviously is not a complete proposition. In the Frege case, what Frege correctly said of the year 1869, namely that more people died of tuberculosis than of cancer then, would incorrectly be said of the year 2006. Again, what you can say OF something else are not complete but rather open propositions.
Wouldn’t that solve the puzzle?

Hi Tobias!

That's an interesting proposal, and I think it's not too far away from what Bernd said. If I understand it correctly, it would entail that terms like "what Frege said" and "this" (after reading his note) are systematically ambiguous between lots of things: all the open propositions one can get from the proposition actually expressed by kicking out various parts. Which of these is meant has to be determined pragmatically, taking into consideration whether the sentence continues with "...is no longer true" or "...is not true for me" or "...is not true for Diabetes and HIV today", etc.

I have a general understanding inhibition with respect to structured propositions, so I'm not sure how the proposal works out in detail. For my counterfactual utterance of "I'm actually a barber", I suppose we'd have to say that full propositions are not only eternal, but non-contingent, containing a parameter that fixes the world. Then we can kick out this parameter when we say that what I said (the resulting matrix) is not true in the actual world.

It also doesn't seem that the open propositions can always be denoted by a that-clause: what's the that-clause corresponding to [___ is a fool] or [more people die of ___ than of ___]? But that is probably not essential to the proposal.

Another worry: suppose I gave a talk yesterday to a group of people after an exhausting bike ride, and I read out the same presentation today to a different group. The talk contains the sentence: "I know that you are all tired and exhausted, so I might skip some of the proofs". After reading this today, I stop and comment, "well, this was true yesterday, but it is not true today, I suppose, so I will actually give all the proofs". What I claim was true yesterday, but not today, is that my audience is tired and exhausted. But how does "you" get kicked out of the proposition and re-evaluated at the other context? I didn't say, "this was true yesterday of my audience then, but is not true now of you".

Thanks for the reply, Wo! Two short comments:

(1) I think that there are ‘that’-clauses corresponding to [___ is a fool] and [more people die of ___ than of ___]. You can use them in the following sentences: ‘That he is a fool is true of John as well as of George’, ‘That more people die of the one than of the other is true of cancer and tuberculosis as well as of cigarettes and smarties’. ‘he’ and ‘the one … the other’ are anaphoric pronouns that function like free variables.

(2) In the tired-audience-case, the comment ‘this was true yesterday, but it is not true today’ strikes me as rather odd and I am not sure whether this is the correct way to say what you want to say here. After all, one possible reaction of today’s audience to this comment could be: ‘No, we were not tired yesterday! We are always as alert as we are today’. Here, the audience would understand you as saying something similar to ‘That you are all tired and exhausted was true yesterday but is no longer true today’, and this, I think, is the correct way to understand the comment. When you mean by your comment what you suggest in the example you would specify the thing which is no longer true by the clause ‘that my audience is tired’. Here the term ‘my audience’ should be understood as a definite description. Hence, in this case we get the proposition which is open with respect to time by starting from the general proposition that yesterday there was exactly one thing which was your audience and its members were all tired. That there is exactly one thing which is your audience and that its members are all tired, is something that was true yesterday but is false now.

Thanks Tobias, I got your first point. Not sure about the second though: can't one react in a similar way in all the other cases: "no, what Frege said is still true; when he said 'today', he obviously meant some time in 1869, not in 2007, and it is still true that in 1869, more people died of tuberculosis." The point is that there is also a natural sense in which one can say that what Frege said is no longer true. And that seems to me to apply to the tired audience case as well. But of course I'd be willing to ignore this intuition if it clashes with a good theory, especially if the intuition isn't generally shared.

(One reason why I'd prefer a more liberal explanation is that I'd like to apply it to candidates for relativism: epistemic modals, predicates of taste etc. That is, I'd like to say that "chili is tasty" is true iff the speaker likes chili despite that fact that people who don't like chili would not normally say "that's true".)

Salü Wo, are you back to Gemany? -
(written after an exhaustive workshop)
What about:

What Frege said is no longer true

is true iff:

The Kaplanian character of Frege's utterance is true at the context <Frege, time of Freges utterance, actual world> but is false at <Frege, time of *this* utterance, actual world>.

(Kaplanian character is true at c := its value for c is true at the world of c)

But try modal/counterfactual embeddings.

Ralf

Meister Ralf!

I see, you've enriched Kaplanian contexts with a speaker to hold fixed. I normally don't like "improper" contexts (where the speaker doesn't exist at the time and world of the context) in a Kaplanian setting, but I admit it gets around the problems here. However, we'd need even more context coordinates for other cases: when Hilbert pointed at Frege, saying "you're not particularly well-known among philosophers", again what he said is in some sense no longer true. So we need to hold fixed an addressee coordinate.

I think it's better just to say that when we judge whether an utterance is true at/for x, we re-evaluate some salient part of the sentence at x, and the rest in the utterance context.

I'll be back in Europe in early July. Looking forward already to being poor and unemployed!

Meister Wo!

Actually for Kaplan a context is not some n-tuple of entities (world, time, self...). In the informal presentation he takes contexts to be primitive entities and then adds that there are functions from contexts to relevant parameters, e.g.: the agent in c, the world of c, the time of c... ("agent" rather than speaker, because the agent need not speak.) This conception is very liberal with respect to relevant parameters. f(c) = the adressee of c is perfectly ok. Or so I remember.
I grant that it might happen that what we end up with on the Kaplan line and on your line might be two different formulations of the same theory. One thing I like to point out is that in the standard discussion about 2D-semantics only horizontal and diagonal propositions are distinguished. But there also are Kaplanian characters. (I learned my 2D-sem. from "Demonstratives" rather than from "Assertion", so that's why.)

I'm quite sure you'll be neither poor nor unemployed.

Ralf