Are scientific theories full of truth-value gaps?

I'm almost done writing a paper that argues that if one accepts the pessimistic induction over the history of science, then one should be a semantic anti-realist about current science -- or should hold unorthodox views in philosophy of language. Here's the argument:

(P1) Certain claims that (a) contain terms with defective reference, or (b) exhibit presupposition failure, are neither true nor false.

(P2) Some fundamental theoretical claims of earlier science exhibit the type of (a) defective reference (e.g. 'phlogiston,' 'absolute velocity,' 'Vulcan') or (b) presupposition failure (e.g. 'Events A and B are simultaneous') described in (P1).

(C1) Therefore, some fundamental theoretical claims of earlier science are neither true nor false.

(P3) Present science probably resembles past science. [That's the step borrowed from the pessimistic induction]

(C2) Therefore, some fundamental theoretical claims of present science are neither true nor false.

The main objections, I think, are:
1. Sentences that contain non-referring terms or exhibit presupposition failure are false, not truth-valueless.
2. Specifically, natural kind terms that fail to pick out a kind/ property (e.g. 'phlogiston') do not generate truth-value gaps (even supposing proper names that fail to pick out an individual do generate truth-value gaps).

I offer replies to these objections, but none of them are absolutely decisive. So, my more tentative conclusion is that a proponent of the pessimistic induction either has to accept my original conclusion OR accept a currently unpopular position in philosophy of language -- e.g. one could justify objection 2 by appealing to a descriptivist account of natural kind terms, but that would fall afoul of the widely endorsed Kripke-Putnam arguments against such descriptivism.

I'll be giving this material as a talk at the University of Utah in a couple of weeks, so any feedback before then would be especially appreciated.


Chris said...

What account of presupposition do you appeal to? I am inclined to think that "A is simultaneous with B" is false, if the correct account of simultaneity is relative to a frame of reference. So, I guess I am sympathetic to the first objection you mention.

Greg Frost-Arnold said...

Hi Chris --

Good point. There are a couple of things to say.
My claim fits most naturally with the semantic (aka Fregean aka Strawsonian) account of presupposition, namely:

A presupposes B =df If A is true or A is false, then B is true.

Now I certainly do think that 'X and Y are simultaneous' (for spacelike related X, Y) is untrue, but not false. Following this line of thought, for such a sentence to be false, there would have to be such a (relational) property as simultaneity, and X and Y would determinately lack it.

Now, not everyone likes the semantic account of presupposition, preferring a pragmatic account. But many people who endorse a pragmatic approach claim that when a speaker utters a sentence with a failed presupposition, that utterance expresses no proposition at all (and thus cannot be either true or false). However, there is a camp of people who think (1) presupposition is pragmatic and (2) utterances with failed presuppositions are strictly speaking false, but infelicitous or otherwise unassertable.

So people in this camp could claim that every case of presupposition failure, such as 'X and Y are simultaneous', results in a truth-valued claim. So that is one escape route for the proponent of the pessimistic induction who wants to avoid semantic anti-realism. However, there might still be problems arising from reference failure in particular.

Fábio Fonseca said...


I'm sorry for this comment too out of the this theme. But i was reading your post about Descartes' position in relation of shaps and colors (11/11/2005).

I'm collecting informations about this kind of perception, but when one of them has necessary a function in relation to the other.

I'm from Brazil and it's hard to have informations about this.

If you could help me. I'll be glad for it. You can contact me by my blog or by e-mail: ffonseca_silva@hotmail.com

Sincerely thank you!

Chris said...

Greg, would you say the same thing about (uttered by a Newtonian) "The center of mass of the solar system is at rest"? That is, when the sentence explicitly involves a description? I guess there are more or less difficult cases to reach the falsity verdict, but this seems to me to be an even clearer case of falsity.

Greg Frost-Arnold said...

Fabio --

Unfortunately, I haven't done any further work on that topic, and Descartes on perception is not one of my areas of expertise. Good luck!

Chris --

Suppose counterfactually that the Sun was the center of the solar system. I don't feel a difference between
(1) The Sun is at absolute rest.
(2) The center of mass of the solar system is at absolute rest.
Why/How would there be a difference in the truth-value (or lack thereof) between (1) and (2)?

Or are you thinking of specifying the property/kind (as opposed to an individual) via description (which I don't think your example does; though I could be convinced otherwise)? For example, replace 'phlogiston' with 'the stuff emitted in every combustion process'. And the semantic status of sentences containing empty definite descriptions is basically the Russell-Strawson dispute over 'The present King of France is bald' -- intuitions pull both ways.

(From my brief foray in the linguistics literature, it seems like many folks now think (i) the debate as it originally played out was a draw/ unresolvable, in part because (ii) speaker's intuitions about the difference between being untrue because truth-valueless and being false are difficult if not impossible to probe with experimental success. The current folks' approach to the Russell-Strawson King of France dispute is to say something like 'once we have a general theory (of presupposition, or description, or...) we'll accept whatever falls out of that theory about the 'The King of France is bald'. Sorry- that's probably both too long and too dense.)

But if you just mean the property/kind is given descriptively, then I completely agree that we'll get false sentences, e.g.
'There are 15-foot-tall humans'
is clearly false, because the (complex, descriptive) property fifteen-foot-tall human has no instances.

Kenny said...

I would have thought that the relativistic cases don't have to lead to either of these outcomes. On an account like Lewis' idea that many terms refer to the most natural candidate in the vicinity, I would think that terms like "at rest" or "simultaneous" would pick out the most natural candidate in the vicinity. That would mean, the context-sensitive term that means "simultaneous from my reference frame" or "at rest with respect to the center of mass of everything I care about" or something like that. This would allow these claims to be true. That's especially nice because we often even now say things like "the sun is at rest and the planets move around it" when talking about the solar system today, even though we know there is no such thing as absolute rest.

Greg Frost-Arnold said...

Hi Kenny --

On that Lewisian view, are there any terms that fail to refer? And as a corollary, are there any false existence claims? If we follow that Lewisian suggestion, then it seems like 'There is no such thing as absolute velocity' (or '... as phlogiston', etc.) turns out false -- which is not what modern physicists want to say, I'd guess.

Also, out of curiosity: is the 'most natural candidate in the vicinity' located by finding something that fits a description as well as possible, or by some sort of causal role considerations (Hardin and Rosenberg, Phil.Sci. 1982 take that route), or something else entirely?

Greg Frost-Arnold said...

Note to self, just so I don't forget:
A realist who is party to this debate probably cannot appeal to descriptivism for natural kind terms, because realists (think they) need a causal-historical account of reference-fixing to avoid the kind of Kuhn-Feyerabend worries about incommensurability/ changing the subject, a central anti-realist line of argument.

Hmm... Actually, now that I think about the dialectic, it seems like the respondent (= pushing objection 2) at that point will not be a realist, but rather an anti-realist trying to save the falsity of 'Wood contains a lot of phlogiston.' So maybe this note to self should be disregarded.

Anonymous said...

#2 isn't going to be quite sufficient, because of an example you yourself cite: Vulcan. That's not a kind term; it's a proper name.

I also think that #2 just generates some logical difficulties. Are you claiming that both the following statements are actually *false*?

1) Phlogiston has negative weight.
2) Phlogiston does not have negative weight.

Personally, I'd much rather be in the situation of gapping both those sentences than admitting that they're both false. No?

Greg Frost-Arnold said...

Hi Avrom --

Thanks for stopping by. I'll reply to your 2 comments in order.

First, the objection I have in mind is someone who admits that empty names generate truth-value gaps, but denies that empty natural kind terms do. This would be a problem for my bigger argument because there are actually very few empty names in the history of science -- but there are lots of empty natural kind terms. So this objector I'm imagining says 'Yes there are some truth-value gaps, but they are so very rare that realism in general is not threatened.'

Second: I am arguing for your favored position, and against the view that "Phlogiston has negative weight" is false.

One clarification: People who claim
"Phlogiston has negative weight"
is false, will always also claim
"Phlogiston does not have negative weight"
is true, just from the truth-table for negation. That is, the people I'm arguing against only say that ATOMIC sentences containing empty names (etc.) are false; compound sentences containing empty names just have their truth-value assigned in the usual way.

So, I apologize if my original post wasn't clear, but I'm pretty sure you and I are in complete agreement.

Anonymous said...

Yes, you're right of course, on all counts (including the fact that I fluffed the distinction between your charitable presentation of an objection and your actual argument). Thanks for reiterating, and sorry about that.