Empty names are names that fail to refer, like 'Santa,' 'Pegasus,' and 'Planet Vulcan.' 'Santa Claus' fails to refer because (on most semantics for empty names) there is no entity that is assigned to 'Santa' as its referent. This is clearly distinct from another view (e.g. Frege's) that 'Santa' should be assigned e.g. the empty set as its referent. That is, there is a difference from having no referent and referring to the empty set -- for my cat has no referent, but '∅' refers to the empty set.
So are there empty predicates? That is, are there predicates that do not signify properties (or extensions, kinds, intensions (= functions from possible worlds to extensions), or whatever your preferred semantic value for predicates is). There are of course predicates whose extension is the empty set (e.g. 'is not identical with itself') -- these predicates signify uninstantiated properties (assuming you think predicates signify properties). But they still signify a property.
There is a fairly massive literature on empty names. (I can recommend Ben Caplan's 2002 dissertation as a nice survey of the empty names landscape.) But there is no talk of empty predicates -- is this because somehow every predicate, unlike names, automatically refers?
Related issue: Philosophers of science often say things like 'phlogiston' and 'caloric' fail to refer. Often, in explaining their claim "The word 'phlogiston' does not refer", these philosophers will say things like "The extension of the predicate 'is phlogiston' (or 'contains phlogiston') is empty." But having the empty set for your extension is different from failing to refer. So when we say that 'contains phlogiston' fails to refer, it seems like we should be saying that it has no (determinate?) extension, not that its extension is empty.
So are there any empty predicates? Are such things even possible? And can the usage of the philosophers of science be defended?
Showing posts with label logic. Show all posts
Showing posts with label logic. Show all posts
4/03/2008
3/14/2008
Logical Pluralism, take 2
This post supersedes the previous one on logical pluralism; it's the post I would've written, had I bothered to do a bit of research before posting. I apologize in advance for how long it is...
Beall and Restall say we should be pluralists about logical consequence, because there are multiple acceptable ways of spelling out the notion of case in the standard criterion of validity:
(V) C is a logical consequence of P1 ... Pn iff in every case where all of P1...Pn are true, C is true also.
Different specifications of case, say B&R, yield different consequence relations.
My impulse is to address this issue 'from above'; that is, in general, when is any sort of pluralism an acceptable stance? Well, one case (though probably not the only one) in which it can be acceptable is when ambiguity is present. E.g., there are multiple, equally acceptable ways of spelling out 'Aaron is at the bank'. And Beall and Restall, on the 3rd page of "Defending Logical Pluralism," state that they think logical consequence is ambiguous; presumably the ambiguity is traced back to the word 'case' in (V) (what else in (V) could it be?). So if 'case' really is ambiguous, and (V) captures the core notion of consequence correctly, then we should be logical pluralists.
But I'm not sure 'case' is ambiguous -- prima facie, it doesn't feel like 'bank,' or 'duck'. It might just be 'general in sense' or 'lack specificity': for example, 'sibling' is general in sense between 'brother' and 'sister', the same goes for 'parent' and 'mother' and 'father.' And nobody wants to be a 'sibling-pluralist' or 'parent-pluralist.' (Note: 'thing', which seems to me closer to 'case', does not seem ambiguous either.)
Linguists, fortunately, have devised a test to distinguish ambiguous expressions from ones that are general in sense: the 'conjunction reduction' test. Suppose my friend Pat is making a monetary deposit, and my friend Tracy is sitting next to a river. Then the two sentences
'Pat is at the bank' and
'Tracy is at the bank'
are both true. However, these truths cannot be expressed in a 'conjunction reduced' sentence
'Pat and Tracy are at the bank,'
for this can only mean that both of Pat and Tracy are next to a body of water, or both are at a financial institution. No 'crossed reading' is possible. The impossibility of crossed readings in the reduced sentence suffices for the ambiguity of 'bank' (Jay D. Atlas, Philosophy without Ambiguity, OUP 1989, p.40). From the little I've seen, almost all linguists and linguistically-inclined philosophers accept this as an ambiguity test. And most also accept: If crossed readings are possible, then there is no ambiguity. E.g., 'Pat and Tracy are parents' can be read as saying that Pat and Tracy are both fathers, both mothers, or one of each -- and the possibility of that 'one of each' reading is what makes 'parent' non-ambiguous, but rather just general in sense.
With a test for ambiguity in hand, we can see if 'case' and 'consequence' are ambiguous (so pluralism is the right stance) or merely general in sense (in which case pluralism is not obviously the right stance). Let C0 be a construction (of the sort used in semantics for intuitionistic logic), and let S0 be a situation (of the sort used in semantics for relevance logic). (Cf. previous pluralism post if that doesn't make sense.) Then, I think, Beall & Restall would say
'C0 is a case'
is true, and so is
'S0 is a case'.
(If they don't, then 'case' is not ambiguous, but something else.)
Now the question is: Are 'crossed readings' possible for
'C0 and S0 are cases'?
That is: can that last sentence express that C0 is a construction and S0 a situation, or can it only be read as saying that C0 and S0 are both constructions or both situations? I lean towards the possibility of crossed readings, but I'm just not sure.
We might try the conjunction-reduction test with logical consequence as well:
not-not-A (relevantly) implies A [but not intuitionistically]
not-not-A (intuitionistically) implies 'If B, then B' [but not relevantly]
The conjunction reduced sentence is:
not-not-A implies A and 'If B, then B'
Here, a crossed reading DOES seem impossible, to me at least. So that makes it look like 'implies' is ambiguous. (Note: I'm not certain I've formed the correct conjunction reduction sentence.)
Now I'm stuck: it looks like 'implies' is ambiguous, while 'case' is not. If this is right, then we could deny (V) -- but what could we put in its place?
Final speculative thought: 'implies' is like 'before' in relativistic physics. If two events are spacelike related (= no signal can be passed between them), then 'e1 is before e2' is neither true nor false until a frame of reference is specified. In one frame of reference e1 will be before e2, and in another frame, e2 will be before e1. But once the frame (and a simultaneity convention) is chosen, then it becomes fully determinate which precedes the other. BUT no one frame is the 'right' one; each frame 'has equal rights'.
The analogy: picking a particular frame of reference is like picking situations over constructions or models etc. As there is no one right frame, so too there is no one right truth-maker. So 'A implies B' is then NOT like 'Aaron is at the bank', for I can use that second sentence meaningfully (= truth-valued-ly) without supplying some further information -- unlike 'e1 is before e2' for spacelike-related events. But once one picks a frame, all instances of 'e1 is before e2' become truth-valued; analogously, once one picks a specification of 'case', every instance of 'Does C follow from {P1, ... Pn}?' has a determinate answer.
In sum: ambiguity may not be the best way of thinking about the different consequence relations; rather, perhaps we should see how well we can make out this analogy with relativistic physics.
Beall and Restall say we should be pluralists about logical consequence, because there are multiple acceptable ways of spelling out the notion of case in the standard criterion of validity:
(V) C is a logical consequence of P1 ... Pn iff in every case where all of P1...Pn are true, C is true also.
Different specifications of case, say B&R, yield different consequence relations.
My impulse is to address this issue 'from above'; that is, in general, when is any sort of pluralism an acceptable stance? Well, one case (though probably not the only one) in which it can be acceptable is when ambiguity is present. E.g., there are multiple, equally acceptable ways of spelling out 'Aaron is at the bank'. And Beall and Restall, on the 3rd page of "Defending Logical Pluralism," state that they think logical consequence is ambiguous; presumably the ambiguity is traced back to the word 'case' in (V) (what else in (V) could it be?). So if 'case' really is ambiguous, and (V) captures the core notion of consequence correctly, then we should be logical pluralists.
But I'm not sure 'case' is ambiguous -- prima facie, it doesn't feel like 'bank,' or 'duck'. It might just be 'general in sense' or 'lack specificity': for example, 'sibling' is general in sense between 'brother' and 'sister', the same goes for 'parent' and 'mother' and 'father.' And nobody wants to be a 'sibling-pluralist' or 'parent-pluralist.' (Note: 'thing', which seems to me closer to 'case', does not seem ambiguous either.)
Linguists, fortunately, have devised a test to distinguish ambiguous expressions from ones that are general in sense: the 'conjunction reduction' test. Suppose my friend Pat is making a monetary deposit, and my friend Tracy is sitting next to a river. Then the two sentences
'Pat is at the bank' and
'Tracy is at the bank'
are both true. However, these truths cannot be expressed in a 'conjunction reduced' sentence
'Pat and Tracy are at the bank,'
for this can only mean that both of Pat and Tracy are next to a body of water, or both are at a financial institution. No 'crossed reading' is possible. The impossibility of crossed readings in the reduced sentence suffices for the ambiguity of 'bank' (Jay D. Atlas, Philosophy without Ambiguity, OUP 1989, p.40). From the little I've seen, almost all linguists and linguistically-inclined philosophers accept this as an ambiguity test. And most also accept: If crossed readings are possible, then there is no ambiguity. E.g., 'Pat and Tracy are parents' can be read as saying that Pat and Tracy are both fathers, both mothers, or one of each -- and the possibility of that 'one of each' reading is what makes 'parent' non-ambiguous, but rather just general in sense.
With a test for ambiguity in hand, we can see if 'case' and 'consequence' are ambiguous (so pluralism is the right stance) or merely general in sense (in which case pluralism is not obviously the right stance). Let C0 be a construction (of the sort used in semantics for intuitionistic logic), and let S0 be a situation (of the sort used in semantics for relevance logic). (Cf. previous pluralism post if that doesn't make sense.) Then, I think, Beall & Restall would say
'C0 is a case'
is true, and so is
'S0 is a case'.
(If they don't, then 'case' is not ambiguous, but something else.)
Now the question is: Are 'crossed readings' possible for
'C0 and S0 are cases'?
That is: can that last sentence express that C0 is a construction and S0 a situation, or can it only be read as saying that C0 and S0 are both constructions or both situations? I lean towards the possibility of crossed readings, but I'm just not sure.
We might try the conjunction-reduction test with logical consequence as well:
not-not-A (relevantly) implies A [but not intuitionistically]
not-not-A (intuitionistically) implies 'If B, then B' [but not relevantly]
The conjunction reduced sentence is:
not-not-A implies A and 'If B, then B'
Here, a crossed reading DOES seem impossible, to me at least. So that makes it look like 'implies' is ambiguous. (Note: I'm not certain I've formed the correct conjunction reduction sentence.)
Now I'm stuck: it looks like 'implies' is ambiguous, while 'case' is not. If this is right, then we could deny (V) -- but what could we put in its place?
Final speculative thought: 'implies' is like 'before' in relativistic physics. If two events are spacelike related (= no signal can be passed between them), then 'e1 is before e2' is neither true nor false until a frame of reference is specified. In one frame of reference e1 will be before e2, and in another frame, e2 will be before e1. But once the frame (and a simultaneity convention) is chosen, then it becomes fully determinate which precedes the other. BUT no one frame is the 'right' one; each frame 'has equal rights'.
The analogy: picking a particular frame of reference is like picking situations over constructions or models etc. As there is no one right frame, so too there is no one right truth-maker. So 'A implies B' is then NOT like 'Aaron is at the bank', for I can use that second sentence meaningfully (= truth-valued-ly) without supplying some further information -- unlike 'e1 is before e2' for spacelike-related events. But once one picks a frame, all instances of 'e1 is before e2' become truth-valued; analogously, once one picks a specification of 'case', every instance of 'Does C follow from {P1, ... Pn}?' has a determinate answer.
In sum: ambiguity may not be the best way of thinking about the different consequence relations; rather, perhaps we should see how well we can make out this analogy with relativistic physics.
2/28/2008
Logical pluralism and brother-in-law pluralism
In my philosophy of logic class this week, we discussed JC Beall and Greg Restall's version of logical pluralism. Our text was their 2000 Australasian Journal of Philosophy article, available on Restall's website here. I've been flipping through their fantastic book-length treatment (OUP, 2006) as well.
Here's their basic idea. The basic, accepted notion of logical consequence is adequately captured in the following:
(V) Consequence C is a logical consequence of premises P1, ... Pn = In every case in which P1, ... Pn is true, C is also true.
Beall and Restall further hold that the notion of case admits of a number of "precisifications" (2006, 88), that is, it can be 'spelled out' or 'fleshed out' in more than one way. [Note: I can't find the quotation now, but I think Beall and Restall said that 'case' is neither ambiguous nor vague (in the sense of having borderline examples). [CORRECTION (3/2/08): In their "Defending Logical Pluralism," Beall and Restall explicitly say that they think the concept of deductive consequence is ambiguous (p.3). This more or less vitiates the main point of this post. I say 'more or less' because there is a test accepted by linguists for distinguishing ambiguity from lack of specificity, and it's not clear that B&R's concept of 'case' passes the test; see my comment #6 in the comment thread.] Different spellings-out of 'case' give rise to different consequence relations (and thus different logics); as examples of cases, they give:
(i) Classical Tarskian models, (ii) possible worlds, (iii) constructions (which yield intuitionistic logic), and (iv) situations (which yield relevant logic).
Finally, because there are multiple ways of spelling out 'case', there is not one correct notion of consequence, since different consequence relations correspond to different ways of specifying the content of (V).
So if someone asks: "Does an arbitrary sentence p follow from a contradiction 'q and not-q'?", the pluralist answer is "Yes and No -- yes, it follows classically (when we take Tarskian models as cases), but no, it does not follow relevantly (when situations are the cases)." Similarly, the pluralist answers the question "Is 'p or not-p' a logical truth?" with "Yes and No -- yes, it is a classical logical truth (since it is true in all Tarskian models), but no, it is not an intuitionistic logical truth (since it is not true in all constructions)".
I find Beall and Restall's position attractive. But while thinking about it, I wondered about when, in general, pluralism is the right (or at least a reasonable) position to take. B&R's claim is the fact that 'case' can be precisified in more than one way -- the meaning of 'case' is somehow underspecified or indeterminate -- to justify being pluralists about 'case' and thereby via (V) about consequence. However, I wonder whether, if this rationale were accepted across the board, pluralism would be almost everywhere, and the appropriate answer to many, many questions would be "Yes and no".
Here's an example of what I mean. The meaning of the phrase 'my brother-in-law' is not completely specific; it is indeterminate between the brother of my spouse and the male spouse of my sibling. However, nobody is a "brother-in-law pluralist": When someone asks me "Is Leon your brother-in-law?", I shouldn't reply "Yes and No -- yes, he is the brother of my spouse, but no, he's not the male spouse of my sibling." And what holds for 'brother-in-law' holds for many, many other terms: lack of specificity is everywhere.
Hopefully the analogy is clear: 'case' and 'brother-in-law' can both be made (more) determinate in different ways. But if this underspecification in the notion of 'case' is all that is required to justify pluralism about consequence, then we should also be pluralists about 'brother-in-law', since there is underspecification there too.
How might someone sympathetic to logical pluralism (e.g. me) respond to this challenge? Well, we could find an example where pluralism seems like the right (or at least reasonable) attitude, and try to argue that 'case' is (more) like that example. For example, I think pluralism about the concept of 'thing' is reasonable: if someone holds out a deck of cards, and asks me "Are there 52 things here?", the right (or reasonable) answer should be "Yes and No -- yes, there are 52 cards, but no, there are far more than 52 molecules".
The question is then: What makes 'thing' different from 'brother-in-law'? And is 'case' (in Beall and Restall's use) more like 'thing' or 'brother-in-law'? The pluralist wants 'case' to be more like 'thing', but I haven't yet figured out how to draw a sharp line. Any ideas?
Here's their basic idea. The basic, accepted notion of logical consequence is adequately captured in the following:
(V) Consequence C is a logical consequence of premises P1, ... Pn = In every case in which P1, ... Pn is true, C is also true.
Beall and Restall further hold that the notion of case admits of a number of "precisifications" (2006, 88), that is, it can be 'spelled out' or 'fleshed out' in more than one way. [Note: I can't find the quotation now, but I think Beall and Restall said that 'case' is neither ambiguous nor vague (in the sense of having borderline examples). [CORRECTION (3/2/08): In their "Defending Logical Pluralism," Beall and Restall explicitly say that they think the concept of deductive consequence is ambiguous (p.3). This more or less vitiates the main point of this post. I say 'more or less' because there is a test accepted by linguists for distinguishing ambiguity from lack of specificity, and it's not clear that B&R's concept of 'case' passes the test; see my comment #6 in the comment thread.] Different spellings-out of 'case' give rise to different consequence relations (and thus different logics); as examples of cases, they give:
(i) Classical Tarskian models, (ii) possible worlds, (iii) constructions (which yield intuitionistic logic), and (iv) situations (which yield relevant logic).
Finally, because there are multiple ways of spelling out 'case', there is not one correct notion of consequence, since different consequence relations correspond to different ways of specifying the content of (V).
So if someone asks: "Does an arbitrary sentence p follow from a contradiction 'q and not-q'?", the pluralist answer is "Yes and No -- yes, it follows classically (when we take Tarskian models as cases), but no, it does not follow relevantly (when situations are the cases)." Similarly, the pluralist answers the question "Is 'p or not-p' a logical truth?" with "Yes and No -- yes, it is a classical logical truth (since it is true in all Tarskian models), but no, it is not an intuitionistic logical truth (since it is not true in all constructions)".
I find Beall and Restall's position attractive. But while thinking about it, I wondered about when, in general, pluralism is the right (or at least a reasonable) position to take. B&R's claim is the fact that 'case' can be precisified in more than one way -- the meaning of 'case' is somehow underspecified or indeterminate -- to justify being pluralists about 'case' and thereby via (V) about consequence. However, I wonder whether, if this rationale were accepted across the board, pluralism would be almost everywhere, and the appropriate answer to many, many questions would be "Yes and no".
Here's an example of what I mean. The meaning of the phrase 'my brother-in-law' is not completely specific; it is indeterminate between the brother of my spouse and the male spouse of my sibling. However, nobody is a "brother-in-law pluralist": When someone asks me "Is Leon your brother-in-law?", I shouldn't reply "Yes and No -- yes, he is the brother of my spouse, but no, he's not the male spouse of my sibling." And what holds for 'brother-in-law' holds for many, many other terms: lack of specificity is everywhere.
Hopefully the analogy is clear: 'case' and 'brother-in-law' can both be made (more) determinate in different ways. But if this underspecification in the notion of 'case' is all that is required to justify pluralism about consequence, then we should also be pluralists about 'brother-in-law', since there is underspecification there too.
How might someone sympathetic to logical pluralism (e.g. me) respond to this challenge? Well, we could find an example where pluralism seems like the right (or at least reasonable) attitude, and try to argue that 'case' is (more) like that example. For example, I think pluralism about the concept of 'thing' is reasonable: if someone holds out a deck of cards, and asks me "Are there 52 things here?", the right (or reasonable) answer should be "Yes and No -- yes, there are 52 cards, but no, there are far more than 52 molecules".
The question is then: What makes 'thing' different from 'brother-in-law'? And is 'case' (in Beall and Restall's use) more like 'thing' or 'brother-in-law'? The pluralist wants 'case' to be more like 'thing', but I haven't yet figured out how to draw a sharp line. Any ideas?
2/22/2008
Which came first: logical truth or consequence?
This term, I am teaching a philosophy of logic class. We've twice run across the following sentiment:
(CPT) Logical consequence is prior to logical truth.
This sentiment is also expressed as 'The real subject matter of logic is the notion of consequence, not a special body of truths.' (References: We've seen this in Stephen Read's Thinking about Logic Ch.2, and in John Etchemendy's work (1988, p.74) too.)
Seeing (CPT) surprised me, since logical truth is (in most cases -- see below) definable in terms of logical consequence, and vice versa: If C is a consequence of P1 ... Pn, then 'If P1 and ... and Pn, then C' is a logical truth. And if T is a logical truth, then T is a consequence of the null set of premises. This is well-known: Beall and Restall, in their recent Logical Pluralism, make exactly this point.
So, in light of the interdefinability of logical truth and consequence, what would prompt someone to say consequence is somehow prior to logical truth? Stephen Read appeals to valid arguments that ineliminably use infinitely many premisses: A(0), A(1), ... Therefore, ∀x Ax. We can't turn this into a logical truth ('If A(0) and A(1) and ..., then ∀x Ax') in standard languages, because standard languages don't allow for infinitely long sentences. This seems like a fair point in favor of (CPT), but it does assume that (i) you accept arguments with infinitely many premises, and (ii) reject languages with infinitely long expressions. [Edit: as Shawn correctly notes in the comments, these two assumptions are fairly widely held. But I have always been a bit suspicious (perhaps for no good reason) about the idea of an argument with infinitely many premises.]
Here's another argument for (CPT), from extremely weak languages. Imagine we have a propositional language with sentence letters p, q, ..., and only two sentential connectives: 'and' and 'or' specified in the usual way. In this language, there are no logical truths (because we don't have 'If... then...' or anything equivalent), but there are still logical consequences: A is still a logical consequence of 'A and B', and 'A or B' is a logical consequence of A. So here is a case where we have logical consequence without logical truth.
But both of these arguments (Read's and mine) rely on somewhat unusual cases. Are there other reasons to accept (CPT) that do not appeal to unusual circumstances? Is there a big literature out there that I don't know about? And does anything really hinge upon whether we think logical truth is prior to consequence, vice versa, or neither?
(CPT) Logical consequence is prior to logical truth.
This sentiment is also expressed as 'The real subject matter of logic is the notion of consequence, not a special body of truths.' (References: We've seen this in Stephen Read's Thinking about Logic Ch.2, and in John Etchemendy's work (1988, p.74) too.)
Seeing (CPT) surprised me, since logical truth is (in most cases -- see below) definable in terms of logical consequence, and vice versa: If C is a consequence of P1 ... Pn, then 'If P1 and ... and Pn, then C' is a logical truth. And if T is a logical truth, then T is a consequence of the null set of premises. This is well-known: Beall and Restall, in their recent Logical Pluralism, make exactly this point.
So, in light of the interdefinability of logical truth and consequence, what would prompt someone to say consequence is somehow prior to logical truth? Stephen Read appeals to valid arguments that ineliminably use infinitely many premisses: A(0), A(1), ... Therefore, ∀x Ax. We can't turn this into a logical truth ('If A(0) and A(1) and ..., then ∀x Ax') in standard languages, because standard languages don't allow for infinitely long sentences. This seems like a fair point in favor of (CPT), but it does assume that (i) you accept arguments with infinitely many premises, and (ii) reject languages with infinitely long expressions. [Edit: as Shawn correctly notes in the comments, these two assumptions are fairly widely held. But I have always been a bit suspicious (perhaps for no good reason) about the idea of an argument with infinitely many premises.]
Here's another argument for (CPT), from extremely weak languages. Imagine we have a propositional language with sentence letters p, q, ..., and only two sentential connectives: 'and' and 'or' specified in the usual way. In this language, there are no logical truths (because we don't have 'If... then...' or anything equivalent), but there are still logical consequences: A is still a logical consequence of 'A and B', and 'A or B' is a logical consequence of A. So here is a case where we have logical consequence without logical truth.
But both of these arguments (Read's and mine) rely on somewhat unusual cases. Are there other reasons to accept (CPT) that do not appeal to unusual circumstances? Is there a big literature out there that I don't know about? And does anything really hinge upon whether we think logical truth is prior to consequence, vice versa, or neither?
1/13/2008
Boghossian on ('metaphysical') analyticity
I've been thinking recently about an objection Paul Boghossian (and many others) make against the Tractarian/ Carnapian conception of an analytic truth, viz. a sentence that is true solely in virtue of the meaning of the sentence. (Boghossian calls this kind of analyticity 'metaphysical analyticity,' which I think is potentially misleading, given the staunch anti-metaphysical tastes of the logical empiricists. Oh well.)
Boghossian considers the notion of metaphysical analyticity untenable. Why? He asks a rhetorical question:
Boghossian is not alone in this view: the basic idea can be found in Quine's "Carnap and Logical Truth," and is developed by Gilbert Harman, Elliott Sober, and Margolis & Laurence. How should we interpret this rhetorical question? Boghossian appears to be claiming that the truth of a sentence of the form 'S means that p' is never a sufficient condition for the the truth of a sentence of the form 'S is true'---that appears to be intended force of the rhetorical question in the quotation immediately above. And that is certainly one reasonable way of cashing out the notion of the truth of a sentence being `fixed exclusively by its meaning.'
If we do understand Boghossian's view in this way, then I think his claim is either misleading or incorrect. Consider a standard material biconditional of the form
(1)p iff q
If such a biconditional is true, we usually say thatq is a necessary and sufficient condition for p . But as we teach undergraduates in Introduction to Logic classes, if this biconditional is true, then (within the classical propositional calculus) so is
(2)p iff [q and (r only if r )]
(Any other logical truth of the classical propositional calculus could be substituted for r only if r.) If we simply read off the surface structure of sentence-schema (2), one might think thatq was no longer sufficient for the truth of p--because there appears to be a second condition that has to be met in order for p to be the case, namely that r only if r. Of course, strictly speaking, this is true: every sentence of the propositional calculus presupposes the truth of all the logical truths of the propositional calculus. However, it seems seriously misleading to me to say that the truth of q is not a sufficient condition for the truth of p in our original biconditional--for that is not the way we standardly understand sufficient conditions.
Hopefully the direct parallel with Boghossian's claim is clear. I certainly agree that his 'truism' quoted above is true. However, when a logical truth--which, as Carnap and Quine agree is a paradigmatic case of analytic truth (if there are any)--is substituted for p in his schema, then that instance of the truism will have (almost) exactly the form of the second biconditional (2). Then, in the usual sense of 'sufficient condition,' we will have a case in which (contra Boghossian) an instance of 'S means that p' is sufficient for 'S is true.' To say otherwise, we would have to give up either classical logic (specifically, the idea that (2) follows from (1)) or the usual understanding of sufficient conditions.
However, one could object that neither classical logic nor our standard view of sufficient conditions is sacrosanct. I think there are reasonable replies to these objections (telegraphically: for whatever non-classical logic you choose, you can substitute some other logical truth for 'r only if r' in (2) above, and the point carries); but I'll leave matters here since this post is too long already.
Comments and criticism from any angle are very welcome, but what I personally go back adn forth on with the above argument is whether it's a 'cheap point' or not... superficial logic-chopping, or genuine insight?
Boghossian considers the notion of metaphysical analyticity untenable. Why? He asks a rhetorical question:
"Isn't it in general true---indeed, isn't it a truism---that for any statement S,
S is true iff for some p, S means that p and p?
How could the mere fact that S means that p make it the case that S is true?" (Boghossian 1996, "Analyticity Reconsidered," Nous [p.364]
Boghossian is not alone in this view: the basic idea can be found in Quine's "Carnap and Logical Truth," and is developed by Gilbert Harman, Elliott Sober, and Margolis & Laurence. How should we interpret this rhetorical question? Boghossian appears to be claiming that the truth of a sentence of the form 'S means that p' is never a sufficient condition for the the truth of a sentence of the form 'S is true'---that appears to be intended force of the rhetorical question in the quotation immediately above. And that is certainly one reasonable way of cashing out the notion of the truth of a sentence being `fixed exclusively by its meaning.'
If we do understand Boghossian's view in this way, then I think his claim is either misleading or incorrect. Consider a standard material biconditional of the form
(1)
If such a biconditional is true, we usually say that
(2)
(Any other logical truth of the classical propositional calculus could be substituted for r only if r.) If we simply read off the surface structure of sentence-schema (2), one might think that
Hopefully the direct parallel with Boghossian's claim is clear. I certainly agree that his 'truism' quoted above is true. However, when a logical truth--which, as Carnap and Quine agree is a paradigmatic case of analytic truth (if there are any)--is substituted for p in his schema, then that instance of the truism will have (almost) exactly the form of the second biconditional (2). Then, in the usual sense of 'sufficient condition,' we will have a case in which (contra Boghossian) an instance of 'S means that p' is sufficient for 'S is true.' To say otherwise, we would have to give up either classical logic (specifically, the idea that (2) follows from (1)) or the usual understanding of sufficient conditions.
However, one could object that neither classical logic nor our standard view of sufficient conditions is sacrosanct. I think there are reasonable replies to these objections (telegraphically: for whatever non-classical logic you choose, you can substitute some other logical truth for 'r only if r' in (2) above, and the point carries); but I'll leave matters here since this post is too long already.
Comments and criticism from any angle are very welcome, but what I personally go back adn forth on with the above argument is whether it's a 'cheap point' or not... superficial logic-chopping, or genuine insight?
9/26/2007
Logic job at Alberta
From my man at the University of Alberta, Ingo Brigandt, comes news of a logic job in his department:
The Department of Philosophy, University of Alberta, invites applications for a tenure-track position in Philosophy, with a specialization in Logic. Other areas of research and teaching specialization and competence are open. The appointment will be made at the rank of Assistant Professor, effective July 1, 2008. Responsibilities include undergraduate and graduate teaching and maintaining an active research programme. Tenure stream faculty normally teach four one term courses per year. Candidates should hold a PhD in Philosophy and provide evidence of scholarly and teaching excellence. Salary is commensurate with qualifications and experience, and the benefit package is comprehensive. Applicants should arrange to send a letter of application indicating the position applied for and describing areas of research interest, curriculum vitae, all university transcripts, a sample of written work, letters from three referees, and, if available, a teaching dossier and teaching evaluations to Bruce Hunter, Chair, Logic Search, Department of Philosophy, University of Alberta, Edmonton, Alberta, CANADA, T6G 2E5. CLOSING DATE: November 10, 2007. The University of Alberta hires on the basis of merit. We are committed to the principle of equity in employment. We welcome diversity and encourage applications from all qualified women and men, including persons with disabilities, members of visible minorities, and Aboriginal persons. All qualified candidates are encouraged to apply; however, Canadian citizens and permanent residents will be given priority. For further information concerning the Department, please consult http://www.uofaweb.ualberta.ca/philosophy/.
Ingo also tells me that Alberta will be advertising a postdoc and an open Associate professor position this year, so all you Oilers fans should start polishing your CVs.
The Department of Philosophy, University of Alberta, invites applications for a tenure-track position in Philosophy, with a specialization in Logic. Other areas of research and teaching specialization and competence are open. The appointment will be made at the rank of Assistant Professor, effective July 1, 2008. Responsibilities include undergraduate and graduate teaching and maintaining an active research programme. Tenure stream faculty normally teach four one term courses per year. Candidates should hold a PhD in Philosophy and provide evidence of scholarly and teaching excellence. Salary is commensurate with qualifications and experience, and the benefit package is comprehensive. Applicants should arrange to send a letter of application indicating the position applied for and describing areas of research interest, curriculum vitae, all university transcripts, a sample of written work, letters from three referees, and, if available, a teaching dossier and teaching evaluations to Bruce Hunter, Chair, Logic Search, Department of Philosophy, University of Alberta, Edmonton, Alberta, CANADA, T6G 2E5. CLOSING DATE: November 10, 2007. The University of Alberta hires on the basis of merit. We are committed to the principle of equity in employment. We welcome diversity and encourage applications from all qualified women and men, including persons with disabilities, members of visible minorities, and Aboriginal persons. All qualified candidates are encouraged to apply; however, Canadian citizens and permanent residents will be given priority. For further information concerning the Department, please consult http://www.uofaweb.ualberta.ca/philosophy/.
Ingo also tells me that Alberta will be advertising a postdoc and an open Associate professor position this year, so all you Oilers fans should start polishing your CVs.
4/19/2007
Can a sentence without a truth-value ever be approximately true?
I am curious to hear people's thoughts on the question in the title. There has been a lot of philosophical work done on the idea that a sentence can be strictly speaking false, yet nonetheless approximately true (or 'truthlike' or 'verisimilar'). For example: I am 5'11", but if someone said 'Greg is 6 feet tall,' we want to say that that claim is approximately true or something like that. But what if the claim was (strictly speaking) neither true nor false? (Readers may insert their own favorite truth-valueless sentence here.)
I ask because, as I mentioned in an earlier post, I'm toying with the idea that the Pessimistic Induction over the history of science plus something like Kuhnian incommensurability (esp. untranslatability) will lead us not to the conclusion that current science is likely to be false, but rather is likely to lack a truth-value. For if we cannot translate the claims of a pre-revolutionary language into the post-revolutionary one, then the pre-revolutionary language (from our current point of view) is truth-valueless, not false.
I ask the question in the title because one common realist response to the Pessimistic induction is: "Well, yes, our current scientific theories are probably not exactly true, but they are approximately true." If truth-valueless sentences cannot be approximately true, then this response is not available to the realist.
I ask because, as I mentioned in an earlier post, I'm toying with the idea that the Pessimistic Induction over the history of science plus something like Kuhnian incommensurability (esp. untranslatability) will lead us not to the conclusion that current science is likely to be false, but rather is likely to lack a truth-value. For if we cannot translate the claims of a pre-revolutionary language into the post-revolutionary one, then the pre-revolutionary language (from our current point of view) is truth-valueless, not false.
I ask the question in the title because one common realist response to the Pessimistic induction is: "Well, yes, our current scientific theories are probably not exactly true, but they are approximately true." If truth-valueless sentences cannot be approximately true, then this response is not available to the realist.
4/02/2007
semantic pathology spotted in the wild
Perhaps the most famous instance of a sentence that exhibits semantic pathology is the Liar: 'This sentence is false'; if the 'this' strikes you as problematic --
(1) (1) is false.
But there are many other types of semantic pathology, such as the 'heterological' paradox and the so-called 'truth-teller': 'This sentence is true.' My colleague James Woodbridge is doing a lot of interesting research in this area; check out his work if you are interested.
This is not a serious post about semantic pathology, but rather just a field report. I sometimes wonder whether these examples like the truth-teller are all that important, since it's hard to imagine circumstances under which speakers might utter it. But I think I may have found a couple of instances of something akin to the truth-teller "in the wild":
(i) A couple of weeks ago, I went to see Spamalot, the musical adaptation of Monty Python and the Holy Grail. One of the songs contained (something close to) the following line: "This is the song that goes like this."
(ii) In the instructions for the Pennsylvania state tax forms, I found (something like) the following:
"You are eligible for the Tax Forgiveness credit if you meet the following requirements:
1. ...
2. ...
3. You meet the eligibility requirements for the Tax Forgiveness Credit.
4. ..."
(1) (1) is false.
But there are many other types of semantic pathology, such as the 'heterological' paradox and the so-called 'truth-teller': 'This sentence is true.' My colleague James Woodbridge is doing a lot of interesting research in this area; check out his work if you are interested.
This is not a serious post about semantic pathology, but rather just a field report. I sometimes wonder whether these examples like the truth-teller are all that important, since it's hard to imagine circumstances under which speakers might utter it. But I think I may have found a couple of instances of something akin to the truth-teller "in the wild":
(i) A couple of weeks ago, I went to see Spamalot, the musical adaptation of Monty Python and the Holy Grail. One of the songs contained (something close to) the following line: "This is the song that goes like this."
(ii) In the instructions for the Pennsylvania state tax forms, I found (something like) the following:
"You are eligible for the Tax Forgiveness credit if you meet the following requirements:
1. ...
2. ...
3. You meet the eligibility requirements for the Tax Forgiveness Credit.
4. ..."
3/02/2007
Scott Soames' "Actually" in Vegas
One of my favorite philosophers, Scott Soames, was in Vegas last weekend; he gave a great talk entitled "Actually", and he and his wife Martha took a trip with some of the folks in my department out to the beautiful Valley of Fire state park.
One of the main points Soames pushed was that "actually" plays two very different roles in philosophical semantics. 'Actually p', spoken in a given world W1, means 'p is true in world W1' (sorry, I can't do corner-quotes in Blogger). Now, Soames says there are two ways to specify this world W1:
(1) by picking it out purely indexically, in the way 'I' picks out the speaker of the token, 'now' picks out the moment of utterance, etc. -- so on this way, 'actually p' means 'p is true in THIS world' (or '...in OUR world').
(2) by (roughly- see Soames' paper for the gory details) giving the proposition associated with the Carnapian state-description of W1. (A state-description assigns a value of true or false to every atomic sentence in a language.)
Scott pointed out that if we pick out world W1 in the second way, then 'Actually p' is a priori: once you have a state-description of the world in which that sentence is uttered, then you have all the information you need to figure out its truth-value. This runs contra the conventional wisdom that says 'Actually p' is a posteriori, since we only learn whether p is true in this world via experience; Soames's point was that 'Actually p' is learned via experience when we use way (1) of picking out the world W1, but since there is this other way (2), 'Actually p' is in fact knowABLE a priori, if not knowN a priori in practical cases.
This is very clever, and I need to think more about it, but I think my esteemed colleague James Woodbridge had the best question/ objection of the afternoon. He said 'Actually p' does not just mean 'p is true at world W1,' but rather 'p is true at W1 AND W1 is actual/ instantiated.' Someone could give me the entire state-description of the world we currently inhabit, and then I could calculate from that description that p is true in a world satisfying such a description -- but I still wouldn't know that 'Actually p' is true, because I wouldn't know that the state-description matched this world. (And I was too slow on the uptake to grasp Soames's reply to James.)
My favorite part of Soames's whole talk, though, was the following approximate quotation:
"Perhaps the contingent a priori is just an old wives' tale."
I know I can't get my grandmother to stop talking about Kripke...
One of the main points Soames pushed was that "actually" plays two very different roles in philosophical semantics. 'Actually p', spoken in a given world W1, means 'p is true in world W1' (sorry, I can't do corner-quotes in Blogger). Now, Soames says there are two ways to specify this world W1:
(1) by picking it out purely indexically, in the way 'I' picks out the speaker of the token, 'now' picks out the moment of utterance, etc. -- so on this way, 'actually p' means 'p is true in THIS world' (or '...in OUR world').
(2) by (roughly- see Soames' paper for the gory details) giving the proposition associated with the Carnapian state-description of W1. (A state-description assigns a value of true or false to every atomic sentence in a language.)
Scott pointed out that if we pick out world W1 in the second way, then 'Actually p' is a priori: once you have a state-description of the world in which that sentence is uttered, then you have all the information you need to figure out its truth-value. This runs contra the conventional wisdom that says 'Actually p' is a posteriori, since we only learn whether p is true in this world via experience; Soames's point was that 'Actually p' is learned via experience when we use way (1) of picking out the world W1, but since there is this other way (2), 'Actually p' is in fact knowABLE a priori, if not knowN a priori in practical cases.
This is very clever, and I need to think more about it, but I think my esteemed colleague James Woodbridge had the best question/ objection of the afternoon. He said 'Actually p' does not just mean 'p is true at world W1,' but rather 'p is true at W1 AND W1 is actual/ instantiated.' Someone could give me the entire state-description of the world we currently inhabit, and then I could calculate from that description that p is true in a world satisfying such a description -- but I still wouldn't know that 'Actually p' is true, because I wouldn't know that the state-description matched this world. (And I was too slow on the uptake to grasp Soames's reply to James.)
My favorite part of Soames's whole talk, though, was the following approximate quotation:
"Perhaps the contingent a priori is just an old wives' tale."
I know I can't get my grandmother to stop talking about Kripke...
6/27/2006
Analyticity in model-theoretic languages
Part of why I am drawn to philosophy of science and logic is that I like to operate with clean and neat formulations of apparently messy concepts -- and these two sub-disciplines of philosophy embrace such tastes more than other sub-fields. Of course, I am not claiming that ethicists and metaphysicians are muddle-headed; most think about their sub-discipline's topics with far more clarity and rigor than I can. I am merely expressing a personal preference for studying deontic logic instead of the most recent form of consequentialism.
Enough autobiography -- I mention it only to explain my motivation for this post. And my point is this: if we adopt the usual formalization of an interpreted language (viz., the model-theoretic one), then we apparently cannot capture the notion of analyticity -- at least in the way Carnap, who is widely recognized as the champion of analyticity, conceives of it.
Conceiving of a language in model-theoretic terms is one widely-used way of introducing precision into a philosophical endeavor. Most readers probably can recite the definition of a model-theoretically understood langauge by heart, but for the innocent:
A language L consists of a ordered triple, where
- L carries grammatical information: which symbols belong to the language, which strings of symbols count as sentences, which grammatical category each symbol belongs to, etc.;
- M is a model =< D, f >, where the domain of discourse D is a set of individuals, and f is an interepretation function, which assigns an individual in D to each proper name in L, sets in D to one-place predicates, sets of ordered pairs drawn from D to two-place predicates, and so on; and
- r specifies the truth-values of certain compound sentences, given the truth-values of their components -- in other words, r basically specifies the truth-tables.
So much for the model-theoretic conception of language; what about analyticity? Carnap, throughout his career, identifies the analytic truths as those sentences that are true merely in virtue of the language one speaks. That is, if we specify that I am speaking a particular language, in the course of that specification, I might present enough information that the truth-values of certain sentences within that language are fixed. (For an obvious example: if I specify what 'and' and 'not' mean in my language via the usual truth tables for those words, any sentence of the form 'p and not-p' comes out false merely in virtue of the rules governing the language I am using.)
Now, after all that rehearsal of material most readers probably know well, I can get to my point. In a model-theoretically characterized language, the truth-values of ALL sentences are determined by the specification of that language. For example, the truth-value of atomic sentences such as 'Fb' are true iff the individual named by 'b' is in the extension of the set associated with 'F' (i.e. 'Fb' is true iff f(b) is an element of the set f(F)). And Carnap certainly never wanted every sentence of a (non-contradictory) language to be analytic.
The problem then is: one of my favorite tools for 'precisification' in philosophy -- model-theoretic languages -- apparently affords no way to characterize one of the concepts I'm most interested in: analyticity. What to make of this? The first, obvious thing to say is: "Of course there couldn't be any explication of analyticity in such languages, because such languages are extensional, and Carnap and Quine (who represent opposing positions in debates over analyticity) both basically agree that analyticity is an intensional notion."
This is right, but I think there is something further to note: in a straightforward sense, every sentence in a (classical) model-theoretic language has its truth-value determined by the specification of the language. That is, by specifying the language, we fix the truth-values of all the sentences in such a language. That seems odd -- the model-theoretic way of specifying a language that has proved very useful in certain situations, but it likely cannot be a fundamental and/or universally applicable one.
One further point: Carnap, Quine, and the other primary antagonists in battles over analyticity all agree that if there is any such thing as analytic truth, then the (so-called) logical truths are paradigm instances of analytic truths, i.e., truth in virtue of meaning (if you are thinking of "Two Dogmas" and don't believe me, look at Word & Object, sec. 14, fn.3, p.65). But the model-theoretic conception of language characterizes the logical truths as a class of sentences that are true across a set of related langauges. That is, to know whether a sentence is a logical truth in one model-theoretic language, you have to check whether that sentence is true in a bunch of other model-theoretic languages that share certain features with the first one.
So, one might think that the way to cash out analyticity in the idiom of the philosophical logician is to use something like Kripkean possible world semantics (which are used, with some variations, in modal, deontic, epistemic, and temporal logics). But these are usually not given linguistic interpretations, and it's not clear to me that it's possible to give a decent one... though I'd love to be wrong. Any thoughts?
Enough autobiography -- I mention it only to explain my motivation for this post. And my point is this: if we adopt the usual formalization of an interpreted language (viz., the model-theoretic one), then we apparently cannot capture the notion of analyticity -- at least in the way Carnap, who is widely recognized as the champion of analyticity, conceives of it.
Conceiving of a language in model-theoretic terms is one widely-used way of introducing precision into a philosophical endeavor. Most readers probably can recite the definition of a model-theoretically understood langauge by heart, but for the innocent:
A language L consists of a ordered triple
- L carries grammatical information: which symbols belong to the language, which strings of symbols count as sentences, which grammatical category each symbol belongs to, etc.;
- M is a model =< D, f >, where the domain of discourse D is a set of individuals, and f is an interepretation function, which assigns an individual in D to each proper name in L, sets in D to one-place predicates, sets of ordered pairs drawn from D to two-place predicates, and so on; and
- r specifies the truth-values of certain compound sentences, given the truth-values of their components -- in other words, r basically specifies the truth-tables.
So much for the model-theoretic conception of language; what about analyticity? Carnap, throughout his career, identifies the analytic truths as those sentences that are true merely in virtue of the language one speaks. That is, if we specify that I am speaking a particular language, in the course of that specification, I might present enough information that the truth-values of certain sentences within that language are fixed. (For an obvious example: if I specify what 'and' and 'not' mean in my language via the usual truth tables for those words, any sentence of the form 'p and not-p' comes out false merely in virtue of the rules governing the language I am using.)
Now, after all that rehearsal of material most readers probably know well, I can get to my point. In a model-theoretically characterized language, the truth-values of ALL sentences are determined by the specification of that language. For example, the truth-value of atomic sentences such as 'Fb' are true iff the individual named by 'b' is in the extension of the set associated with 'F' (i.e. 'Fb' is true iff f(b) is an element of the set f(F)). And Carnap certainly never wanted every sentence of a (non-contradictory) language to be analytic.
The problem then is: one of my favorite tools for 'precisification' in philosophy -- model-theoretic languages -- apparently affords no way to characterize one of the concepts I'm most interested in: analyticity. What to make of this? The first, obvious thing to say is: "Of course there couldn't be any explication of analyticity in such languages, because such languages are extensional, and Carnap and Quine (who represent opposing positions in debates over analyticity) both basically agree that analyticity is an intensional notion."
This is right, but I think there is something further to note: in a straightforward sense, every sentence in a (classical) model-theoretic language has its truth-value determined by the specification of the language. That is, by specifying the language, we fix the truth-values of all the sentences in such a language. That seems odd -- the model-theoretic way of specifying a language that has proved very useful in certain situations, but it likely cannot be a fundamental and/or universally applicable one.
One further point: Carnap, Quine, and the other primary antagonists in battles over analyticity all agree that if there is any such thing as analytic truth, then the (so-called) logical truths are paradigm instances of analytic truths, i.e., truth in virtue of meaning (if you are thinking of "Two Dogmas" and don't believe me, look at Word & Object, sec. 14, fn.3, p.65). But the model-theoretic conception of language characterizes the logical truths as a class of sentences that are true across a set of related langauges. That is, to know whether a sentence is a logical truth in one model-theoretic language, you have to check whether that sentence is true in a bunch of other model-theoretic languages that share certain features with the first one.
So, one might think that the way to cash out analyticity in the idiom of the philosophical logician is to use something like Kripkean possible world semantics (which are used, with some variations, in modal, deontic, epistemic, and temporal logics). But these are usually not given linguistic interpretations, and it's not clear to me that it's possible to give a decent one... though I'd love to be wrong. Any thoughts?
5/18/2006
Quine on logical truth, again
In a previous post, I asked about the relationship between Quine's definition of logical truth and the now-standard (model-theoretic) one. Here's the second installment, which I decided to finally post after sitting on it for a while, since Kenny just posted a nice set of thoughts on the very closely related topic of logical consequence.
The standard definition is:
(SLT) Sentence S is a logical truth of language L = S is true in all models of L.
Quine's definition is:
(QLT) S is a logical truth = "we get only truths when we substitute sentences for [the] simple [=atomic] sentences" of S (Philosophy of Logic, 50).
(Quine counts open formulas, e.g. 'x burns', as sentences; specifically, he calls them 'open sentences.' I'll follow his usage here.)
So what does Quine think is the relationship between the standard characterization of logical truth and his own? He argues fot the following equivalence claim:
(EQ) If “our object language is… rich enough for elementary number theory ,” then “[a]ny schema that comes out true under all substitutions of sentences, in such a language, will also be satisfied by all models, and conversely” (53).
In a nutshell, Quine argues for the 'only if' direction via the Löwenheim-Skolem theorem (hence the requirement of elementary number theory within the object language), and for the ‘if’ direction by appeal to the completeness of first-order logic. I'll spell out his reasoning in a bit more detail below, but I can sum up my worry about it here: in order to overcome Tarski’s objection to Quine’s substitutional version of logical truth, Quine appeals to the Löwenheim-Skolem theorem. However, for that appeal to work, Quine has to require the object language to be rich enough to fall afoul of the incompleteness results, thereby depriving him of one direction of the purported equivalence between the model-based and sentence-substitution-based notions of logical truth.
The 'only if' direction
Quine presents an extension of the Löwenheim-Skolem Theorem due to Hilbert and Bernays:
“If a [GFA: first-order] schema is satisfied by a model at all, it becomes true under some substitution of sentences of elementary number theory for its simple schemata” (54).
A little logic chopping will get us to
If all substitutions of sentences from elementary number theory make A true, then A is satisfied in all models.
-- which is what we wanted to show. (I think this argument is OK .)
The 'if' direction
Quine takes as his starting premise the completeness result for first-order logic:
(CT) “If a schema is satisfied by every model, it can be proved” (54).
(Quine then argues that if a schema can be proved within a given proof calculus whose inference rules are “visibly sound,” i.e., “visibly such as to generate only schemata that come out true under all substitutions” (54), then such a schema will of course ‘come out true under all substitutions of sentences,’ Q.E.D.) This theorem is of course true for first-order logic; however, Quine has imposed the demand that our object language contain the resources of elementary number theory (explicitly including both plus and times, so that Presburger arithmetic—which is complete—is not in play). And once our object language is that rich, then Gödel’s first incompleteness theorem comes into play. Specifically, in any consistent proof-calculus rich enough for elementary number theory, there will be sentences [and their associated schema] that are true, i.e., satisfied by every model, yet cannot be proved -- providing a counterexample to (CT). So the dialectic, as I see it, is as follows: to answer Tarski's objection to the substitutional version of logical truth, Quine requires the language to be rich enough to capture number theory. But once Quine has made that move, the crucial premise (viz., CT) for the other direction of his equivalence claim no longer holds.
I think I must be missing something -- first, Quine is orders of magnitude smarter than I am, and second, while Quine is fallible, this does not seem like the kind of mistake he's likely to make. So perhaps someone in the blogosphere can set me straight.
And I have one more complaint. Quine claims that his "definition of logical truth agrees with the alternative definition in terms of models, as long as the object language is not too weak for the modest idioms of elementary number theory. In the contrary case we can as well blame any discrepancies on the weakness of the language as on the definition of logical truth" (55). This defense strikes me as implausible: why would we only have (or demand) a well-defined notion of logical truth once we reach number theory? Don't we want a definition to cover all cases, including the simple ones? If the model-theoretic definition captures all the intuitive cases, and Quine's only when the language is sufficiently rich, isn't that a good argument against Quine's characterization of logical truth?
------
(And for those wondering what Quine thinks is the advantage of his characterization of logical truth over the model-theoretic one, the answer is: Quine's uses much less set theory. "The evident philosophical advantage of resting with this substitutional definition, and not broaching model theory, is that we save on ontology. Sentences suffice,… instead of a universe of sets specifiable and unspecifiable. … [W]e have progressed a step whenever we find a way of cutting the ontological costs of some particular development" (55). Quine recognizes that his characterization is not completely free of set theory, given the proof of the LS theorem, so he says his "retreat" from the model-based notion of logical truth "renders the notions of validity and logical truth independent of all but a modest bit of set theory; independent of the higher flights" (56).)
The standard definition is:
(SLT) Sentence S is a logical truth of language L = S is true in all models of L.
Quine's definition is:
(QLT) S is a logical truth = "we get only truths when we substitute sentences for [the] simple [=atomic] sentences" of S (Philosophy of Logic, 50).
(Quine counts open formulas, e.g. 'x burns', as sentences; specifically, he calls them 'open sentences.' I'll follow his usage here.)
So what does Quine think is the relationship between the standard characterization of logical truth and his own? He argues fot the following equivalence claim:
(EQ) If “our object language is… rich enough for elementary number theory ,” then “[a]ny schema that comes out true under all substitutions of sentences, in such a language, will also be satisfied by all models, and conversely” (53).
In a nutshell, Quine argues for the 'only if' direction via the Löwenheim-Skolem theorem (hence the requirement of elementary number theory within the object language), and for the ‘if’ direction by appeal to the completeness of first-order logic. I'll spell out his reasoning in a bit more detail below, but I can sum up my worry about it here: in order to overcome Tarski’s objection to Quine’s substitutional version of logical truth, Quine appeals to the Löwenheim-Skolem theorem. However, for that appeal to work, Quine has to require the object language to be rich enough to fall afoul of the incompleteness results, thereby depriving him of one direction of the purported equivalence between the model-based and sentence-substitution-based notions of logical truth.
The 'only if' direction
Quine presents an extension of the Löwenheim-Skolem Theorem due to Hilbert and Bernays:
“If a [GFA: first-order] schema is satisfied by a model at all, it becomes true under some substitution of sentences of elementary number theory for its simple schemata” (54).
A little logic chopping will get us to
If all substitutions of sentences from elementary number theory make A true, then A is satisfied in all models.
-- which is what we wanted to show. (I think this argument is OK .)
The 'if' direction
Quine takes as his starting premise the completeness result for first-order logic:
(CT) “If a schema is satisfied by every model, it can be proved” (54).
(Quine then argues that if a schema can be proved within a given proof calculus whose inference rules are “visibly sound,” i.e., “visibly such as to generate only schemata that come out true under all substitutions” (54), then such a schema will of course ‘come out true under all substitutions of sentences,’ Q.E.D.) This theorem is of course true for first-order logic; however, Quine has imposed the demand that our object language contain the resources of elementary number theory (explicitly including both plus and times, so that Presburger arithmetic—which is complete—is not in play). And once our object language is that rich, then Gödel’s first incompleteness theorem comes into play. Specifically, in any consistent proof-calculus rich enough for elementary number theory, there will be sentences [and their associated schema] that are true, i.e., satisfied by every model, yet cannot be proved -- providing a counterexample to (CT). So the dialectic, as I see it, is as follows: to answer Tarski's objection to the substitutional version of logical truth, Quine requires the language to be rich enough to capture number theory. But once Quine has made that move, the crucial premise (viz., CT) for the other direction of his equivalence claim no longer holds.
I think I must be missing something -- first, Quine is orders of magnitude smarter than I am, and second, while Quine is fallible, this does not seem like the kind of mistake he's likely to make. So perhaps someone in the blogosphere can set me straight.
And I have one more complaint. Quine claims that his "definition of logical truth agrees with the alternative definition in terms of models, as long as the object language is not too weak for the modest idioms of elementary number theory. In the contrary case we can as well blame any discrepancies on the weakness of the language as on the definition of logical truth" (55). This defense strikes me as implausible: why would we only have (or demand) a well-defined notion of logical truth once we reach number theory? Don't we want a definition to cover all cases, including the simple ones? If the model-theoretic definition captures all the intuitive cases, and Quine's only when the language is sufficiently rich, isn't that a good argument against Quine's characterization of logical truth?
------
(And for those wondering what Quine thinks is the advantage of his characterization of logical truth over the model-theoretic one, the answer is: Quine's uses much less set theory. "The evident philosophical advantage of resting with this substitutional definition, and not broaching model theory, is that we save on ontology. Sentences suffice,… instead of a universe of sets specifiable and unspecifiable. … [W]e have progressed a step whenever we find a way of cutting the ontological costs of some particular development" (55). Quine recognizes that his characterization is not completely free of set theory, given the proof of the LS theorem, so he says his "retreat" from the model-based notion of logical truth "renders the notions of validity and logical truth independent of all but a modest bit of set theory; independent of the higher flights" (56).)
5/02/2006
Knowledge via public ignorance
Yesterday Rohit Parikh gave a very interesting talk at Carnegie Mellon on a kind of modal epistemic logic he has been working on recently with several collaborators, cleverly called topologic, because it carries interesting topological properties. The first thing Parikh said was "I like formalisms, but I like examples more." In that spirit, I wanted to describe here one simple example he showed us yesterday, without digging into the technicalia, because it generates a potentially philosophically interesting situation: someone can (under suitable circumstances) gain knowledge merely via other people's declarations of ignorance.
Imagine two people play the following game: a natural number n>0 (1,2, ...) is selected. Then, one of the players has n written on his or her forehead, and the other player has n+1 written on his forehead. Each player can see what is written on the other's forehead, but cannot see what is written on their own. The game allows only two "moves": you can either say "I don't know what number is on my forehead" or state what you think the number on your forehead is.
So, for example, if I play the game, and I see that the other person has a 2 written on her forehead, I know that the number on my own forehead is either a 1 or a 3, but I do not know which. But here is the interesting part: if my fellow game-player wears a 2, and on her first move says "I don't know what my number is," then I know what my number is -- at least, if my fellow game-player is reasonably intelligent. Why? If I were wearing a 1, then my interlocutor would say, on her first move, "I know my own humber is a 2" -- because (1, 2) is the first allowable pair in the game. Thus, if she says "I don't know what my own number is" on her first move, then I know my number can't be 1, so it must be 3. This same process of reasoning can be extended: by playing enough rounds of "I don't know" moves, we can eventually successfully reach any pair of natural numbers, no matter how high. We just have to keep track of how many rounds have been played. (This may remind the mathematically-inclined in the audience of the Mr. Sum-Mr. Product dialogue.)
What is interesting to me about this is that the two players in such a game (and the other examples Prof. Parikh described) can eventually come to have knowledge about the world simply via declarations of ignorance. These cases prompt two questions for me:
(1) Is this type of justification for a belief different in kind from the others philosophers busy themselves with? Or is this just a completely normal/ standard/ etc. way of gathering knowledge, which differs only superficially from other cases? (I'm not qualified to answer this, since I'm not an epistemologist.)
(2) Are there any interesting real-world examples where we achieve knowledge via collective ignorance in (roughly) this way? (Prof. Parikh suggested that there might be a vague analogy between what happens in these sorts of games and game-theoretic treatments of evolution, but didn't have much further to say.)
Imagine two people play the following game: a natural number n>0 (1,2, ...) is selected. Then, one of the players has n written on his or her forehead, and the other player has n+1 written on his forehead. Each player can see what is written on the other's forehead, but cannot see what is written on their own. The game allows only two "moves": you can either say "I don't know what number is on my forehead" or state what you think the number on your forehead is.
So, for example, if I play the game, and I see that the other person has a 2 written on her forehead, I know that the number on my own forehead is either a 1 or a 3, but I do not know which. But here is the interesting part: if my fellow game-player wears a 2, and on her first move says "I don't know what my number is," then I know what my number is -- at least, if my fellow game-player is reasonably intelligent. Why? If I were wearing a 1, then my interlocutor would say, on her first move, "I know my own humber is a 2" -- because (1, 2) is the first allowable pair in the game. Thus, if she says "I don't know what my own number is" on her first move, then I know my number can't be 1, so it must be 3. This same process of reasoning can be extended: by playing enough rounds of "I don't know" moves, we can eventually successfully reach any pair of natural numbers, no matter how high. We just have to keep track of how many rounds have been played. (This may remind the mathematically-inclined in the audience of the Mr. Sum-Mr. Product dialogue.)
What is interesting to me about this is that the two players in such a game (and the other examples Prof. Parikh described) can eventually come to have knowledge about the world simply via declarations of ignorance. These cases prompt two questions for me:
(1) Is this type of justification for a belief different in kind from the others philosophers busy themselves with? Or is this just a completely normal/ standard/ etc. way of gathering knowledge, which differs only superficially from other cases? (I'm not qualified to answer this, since I'm not an epistemologist.)
(2) Are there any interesting real-world examples where we achieve knowledge via collective ignorance in (roughly) this way? (Prof. Parikh suggested that there might be a vague analogy between what happens in these sorts of games and game-theoretic treatments of evolution, but didn't have much further to say.)
4/29/2006
Modal logic workshop at CMU
I spent yesterday at a workshop devoted to modal logic at Carnegie Mellon University. Rather than rehearse everything that happened, I'll simply point interested parties to the workshop webpage, which has abstracts for the talks. Mostly local folks presented their work, but Johan van Bentham from Amsterdam and Stanford was here, along with Rohit Parikh, who'll be presenting on Monday as well.
I certainly learned a lot, and even though my brain was hurting afterwards, I enjoyed myself too.
I certainly learned a lot, and even though my brain was hurting afterwards, I enjoyed myself too.
4/04/2006
Semantics and necessary truth
I have recently been reading, with great profit, Jason Stanley's draft manuscript Philosophy of Language in the Twentieth Century, forthcoming in the Routledge Guide to Twentieth Century Philosophy. The paper's synoptic scope matches the ambitious title.
I'm curious about one relatively small claim Jason makes. On MS p.17, he says:
Here's my worry: as a preliminary, recall (as Tarski taught us) that semantic vocabulary should always be indexed to a particular language -- e.g., we must say 'is a true sentence of English' or 'x refers to y in Farsi' etc. in the full statement of sentences like (7). But then I am not so sure that such sentences are not true in all possible worlds. Is it really the case that, in English, "is smart" could have expressed the property of being from Mars? We specify a particular language (in part) by specifying the semantic values of the words of that language (at least, if we are not proceeding purely formally/ proof-theoretically). Wouldn't we be speaking another language at that point, that was similar to English, but not the same?
My intuitions lean towards saying that this would not be English, but those intuitions aren't firm. I think the question boils down to: "Is 'English' a rigid designator (i.e., does 'English' refer to the same thing(s) in all possible worlds)?", but I'm not sure about that, either. Which way do your intuitions run?
I'm curious about one relatively small claim Jason makes. On MS p.17, he says:
intuitively an instance of Tarski's schema T [GF-A: '...' is a true sentence if and only iff ...] such as (7) is not a necessary truth at all:This certainly has (as Stanley says) an "intuitive" ring. But now I'm not sure it's correct.
(7) "Bill Clinton is smart" is a true sentence if and only if Bill Clinton is smart.
(7) is not a necessary truth, because "Bill Clinton is smart" could have meant something other than it does. For example, "is smart" could have expressed the property of being from Mars, in which case (7) would be false.
Here's my worry: as a preliminary, recall (as Tarski taught us) that semantic vocabulary should always be indexed to a particular language -- e.g., we must say 'is a true sentence of English' or 'x refers to y in Farsi' etc. in the full statement of sentences like (7). But then I am not so sure that such sentences are not true in all possible worlds. Is it really the case that, in English, "is smart" could have expressed the property of being from Mars? We specify a particular language (in part) by specifying the semantic values of the words of that language (at least, if we are not proceeding purely formally/ proof-theoretically). Wouldn't we be speaking another language at that point, that was similar to English, but not the same?
My intuitions lean towards saying that this would not be English, but those intuitions aren't firm. I think the question boils down to: "Is 'English' a rigid designator (i.e., does 'English' refer to the same thing(s) in all possible worlds)?", but I'm not sure about that, either. Which way do your intuitions run?
4/03/2006
Tarski, Quine, and logical truth
The following must have been addressed already in the literature, but I'm going to mention it anyway--perhaps a better-informed reader can point me in the direction of the relevant research. W.V.O. Quine offers the following characterization of logical truth:
Now some historical questions: did Quine think his condition was sufficient, or just necessary? (I quickly checked "Truth by Convention," and I didn't find any conclusive evidence that he considered it sufficient.) If Quine does consider this a proper definition of logical truth, how does he/ would he answer Tarski's objection? -- and/or why doesn't Quine simply adopt Tarski's definition of logical truth, viz. 'truth in all models'?
(You might think Tarski's objection shouldn't count for much, since I used a very contrived language to make the point against Quine. In Tarski's defense, however, (a) assuming that every object has a name in our language also seems somewhat artificial, and (b) Tarski proved (elsewhere) that a single language cannot contain names for all (sets of) real numbers (See "On Definable Sets of Real Numbers," reprinted in Logic, Semantics, Metamathematics).)
The logical truths are those true sentences which involve only logical words essentially. What this means is that any other words, though they may also occur in a logical truth (as witness 'Brutus,' 'kill,' and 'Caesar' in 'Brutus killed or did not kill Caesar'), can be varied at will without engendering falsity. ("Carnap and Logical Truth," §2)All I wanted to mention here is that Alfred Tarski had already shown, in 1936's "On the Concept of Logical Consequence," that Quine's characterization is a necessary condition for a sentence to be a logical truth, but not a sufficient one. For example, if one is using an impoverished language that (i) only has proper names for things over five feet tall, and (ii) only has predicates applying only to things over five feet tall, then the sentence 'George W. Bush is over five feet tall' will be a logical truth -- because no matter what name from this impoverished language we substitute for 'George W. Bush' or what predicate we substitute for 'over 5 feet tall' in this sentence, the resulting sentence will be true.
Now some historical questions: did Quine think his condition was sufficient, or just necessary? (I quickly checked "Truth by Convention," and I didn't find any conclusive evidence that he considered it sufficient.) If Quine does consider this a proper definition of logical truth, how does he/ would he answer Tarski's objection? -- and/or why doesn't Quine simply adopt Tarski's definition of logical truth, viz. 'truth in all models'?
(You might think Tarski's objection shouldn't count for much, since I used a very contrived language to make the point against Quine. In Tarski's defense, however, (a) assuming that every object has a name in our language also seems somewhat artificial, and (b) Tarski proved (elsewhere) that a single language cannot contain names for all (sets of) real numbers (See "On Definable Sets of Real Numbers," reprinted in Logic, Semantics, Metamathematics).)
2/22/2006
Quantum logic question
I've been thinking about quantum logic (QL) recently, and in particular about the usual semantics for 'or' in QL. I've become puzzled, and hopefully someone out there in the blogosphere can help me clear up my confusion.
For the uninitiated: In QL, propositions are represented by/ interpreted as subspaces in a Hilbert space -- including one-dimensional subspaces, i.e. rays. There are multiple ways of formulating in colloquial language what these subspaces are to represent (see final paragraph below), but (atomic) sentences are usually taken to have the form:
'The value of observable O is o1'
where 'observable' just means any physical quantity (e.g., position, momentum, energy, spin), and o1 is just a particular value (or range of values) of that observable. (E.g. 'The energy of this system is between 4 and 6 Joules.') Such sentences are true iff the state-vector of the system lies within the subspace.
Now, think of a particle P in a superposition of spin up and spin down along the y-axis. This particle's state is of course represented by a different vector (call it V_s) than particles in the spin-up state (represented by V_up), or particles in the spin-down (V_down) state. However, because the usual QL semantics assigns to 'p or q' the linear span, instead of the union, of the rays associated with p and q, the claim 'P is spin-up or P is spin-down' will be true -- because V_s is in the linear span of the spin-up ray and the spin-down ray. Each of the disjuncts is false, but the whole disjunction is true. (To me, this feature of QL is even more striking than the failure of the so-called distributive law, i.e., [p&(q or r)] iff [(p&q) or (p&r)], which commentators on QL tend to focus on.)
This seems intuitively wrong to me (or at least as 'wrong' as something can be in formal semantics). In 2-D Euclidean space, suppose we have a unit vector V at a 45-degree angle to the x-axis. I don't think anyone would consider the sentence 'V lies along the x-axis or V lies along the y-axis' to be true. V is not a unit vector on the x-axis or on the y-axis, but a distinct third thing. I don't see why we would change policies in the quantum case, which appears analogous to me.
So now I can ask my question: could we change the semantics for 'or' to avoid these apparent problems? In particular, in the usual semantics for quantum logic, why must all propositions be represented as subspaces on a Hilbert space? -- why not also allow subsets (which might not be closed under linear combinations)? For then we could allow 'or' to mean the union of rays, and 'P is spin-up or P is spin-down' will come out false.
One further note: some people (e.g. R.I.G. Hughes, "Quantum Logic and the Interpretation of Quantum Mechanics," PSA 1980) take the atomic QL propositions to have a different correlate in colloquial language. Instead of
'The value of observable O in system S is within o1,'
they take the subsets of Hilbert space to mean
'The result of a measurement operation for observable O in system S is within o1.'
Under this understanding, my above worries disappear -- for the result of a spin-y measurement surely will be either spin-up or spin-down. However, QL then becomes much less interesting, because it is just about measurement outcomes, instead of about these supremely odd things, superpositions.
For the uninitiated: In QL, propositions are represented by/ interpreted as subspaces in a Hilbert space -- including one-dimensional subspaces, i.e. rays. There are multiple ways of formulating in colloquial language what these subspaces are to represent (see final paragraph below), but (atomic) sentences are usually taken to have the form:
'The value of observable O is o1'
where 'observable' just means any physical quantity (e.g., position, momentum, energy, spin), and o1 is just a particular value (or range of values) of that observable. (E.g. 'The energy of this system is between 4 and 6 Joules.') Such sentences are true iff the state-vector of the system lies within the subspace.
Now, think of a particle P in a superposition of spin up and spin down along the y-axis. This particle's state is of course represented by a different vector (call it V_s) than particles in the spin-up state (represented by V_up), or particles in the spin-down (V_down) state. However, because the usual QL semantics assigns to 'p or q' the linear span, instead of the union, of the rays associated with p and q, the claim 'P is spin-up or P is spin-down' will be true -- because V_s is in the linear span of the spin-up ray and the spin-down ray. Each of the disjuncts is false, but the whole disjunction is true. (To me, this feature of QL is even more striking than the failure of the so-called distributive law, i.e., [p&(q or r)] iff [(p&q) or (p&r)], which commentators on QL tend to focus on.)
This seems intuitively wrong to me (or at least as 'wrong' as something can be in formal semantics). In 2-D Euclidean space, suppose we have a unit vector V at a 45-degree angle to the x-axis. I don't think anyone would consider the sentence 'V lies along the x-axis or V lies along the y-axis' to be true. V is not a unit vector on the x-axis or on the y-axis, but a distinct third thing. I don't see why we would change policies in the quantum case, which appears analogous to me.
So now I can ask my question: could we change the semantics for 'or' to avoid these apparent problems? In particular, in the usual semantics for quantum logic, why must all propositions be represented as subspaces on a Hilbert space? -- why not also allow subsets (which might not be closed under linear combinations)? For then we could allow 'or' to mean the union of rays, and 'P is spin-up or P is spin-down' will come out false.
One further note: some people (e.g. R.I.G. Hughes, "Quantum Logic and the Interpretation of Quantum Mechanics," PSA 1980) take the atomic QL propositions to have a different correlate in colloquial language. Instead of
'The value of observable O in system S is within o1,'
they take the subsets of Hilbert space to mean
'The result of a measurement operation for observable O in system S is within o1.'
Under this understanding, my above worries disappear -- for the result of a spin-y measurement surely will be either spin-up or spin-down. However, QL then becomes much less interesting, because it is just about measurement outcomes, instead of about these supremely odd things, superpositions.
8/25/2005
Underdetermination and equivalence modulo p
Since the description of this blog states that it deals with "issues in logic" related to philosophy of science, I figure that, for the sake of truth in advertising, I should post something logical. (Though I don't feel particularly rushed: there are already a fair number of smart logicians actively participating in the blogosphere -- check my blogroll. For reasons I don't understand, the situation is different in philosophy of science. Any armchair anthropologists have an explanation?)
Underdetermination arguments occur in many quarters of philosophy: Descartes' demon is perhaps the most famous, but they have also played a leading role in discussions about scientific realism during the last few decades. In this post I want to characterize a particular sort of underdetermination using elementary logical notions. (This form of underdetermination either is -- or is closest to -- the Quine-Duhem variety, I'm not sure which at the moment.)
Consider two sets of sentences, A, B such that neither set is a logical consequence of the other. Now suppose there is a third set of sentences C such that:
If C then (A iff B).
That is, if we assume that C is true, then A and B are logically equivalent. (In all models where C is true, either both A and B are true, or both are false).
Then we say A and B are equivalent modulo C.
How does this relate to underdetermination? We can have two theories that are not logically equivalent (and thus are not 'the same theory'), but do become logically equivalent if we make some further assumptions (C above) -- and these further assumptions can be taken to be "auxiliary hypotheses" or "background knowledge" (or whatever one wishes to call the other claims a theory uses, besides its own, to make predictions). If we are committed to the truth of the background knowledge, then we cannot decide between the two theories.
(If this is a bit abstract, here's a toy example:
A = M and (if p then q) [assume M says nothing about p or q]
B = M and (if p then not-q)
C includes the sentence 'not-p';
so neither of A and B implies the other; if C is true then A and B are logically equivalent, while if C is false then A and B are inconsistent.)
My question: is anything philosophically interesting going on here? If we hold r to be true, do we really need to choose between (r or s) and (r or not-s)? I think not -- though they differ in logical content, they are not rivals (or are they?). At least, if we take r to be true, then they are definitely not rivals, though they might be considered rivals ‘on their own’. They certainly are genuine competitors when we hold r false -- though then they are no longer equivalent in any sense.
Comments:
1. This is not the usual sort of underdetermination situation. First, the notion of "empirical content" (or "empirical equivalence," i.e. identity of empirical content) does not appear, so the much-maligned observable/ unobservable distinction is never mentioned. Second, and more importantly, the two theories A and B are not incompatible: the Cartesian demon, on the other hand, is either deceiving us or not (at least on the usual interpretation). The demon-hypothesis is incompatible with the 'real-world' hypothesis. On the other hand, 'if p then q' and 'if p then not-q' are not inconsistent -- we need simply hold that 'not-p' is true.
2. How does this relate to a ‘real’ example, e.g. Bohmian mechanics? It is empirically equivalent to standard quantum mechanics as long as absolute position is undetectable -- but not if absolute position is detectable. In other words, the standard theory and the Bohmian theory are empirically equivalent modulo the claim that absolute position is undetectable. And that is structurally similar to the toy example above. (Of course, there is the difference in this case that the two theories are 'empirically equivalent modulo p,' not 'logically.')
3. Lastly, it is probably considerations akin if not identical to the above that prompted philosophers to move to ‘total theories’ (i.e. theories PLUS all their auxiliary assumptions) as the proper objects of epistemic evaluation. See e.g. (Leplin, Erkenntnis, 1997).
Underdetermination arguments occur in many quarters of philosophy: Descartes' demon is perhaps the most famous, but they have also played a leading role in discussions about scientific realism during the last few decades. In this post I want to characterize a particular sort of underdetermination using elementary logical notions. (This form of underdetermination either is -- or is closest to -- the Quine-Duhem variety, I'm not sure which at the moment.)
Consider two sets of sentences, A, B such that neither set is a logical consequence of the other. Now suppose there is a third set of sentences C such that:
If C then (A iff B).
That is, if we assume that C is true, then A and B are logically equivalent. (In all models where C is true, either both A and B are true, or both are false).
Then we say A and B are equivalent modulo C.
How does this relate to underdetermination? We can have two theories that are not logically equivalent (and thus are not 'the same theory'), but do become logically equivalent if we make some further assumptions (C above) -- and these further assumptions can be taken to be "auxiliary hypotheses" or "background knowledge" (or whatever one wishes to call the other claims a theory uses, besides its own, to make predictions). If we are committed to the truth of the background knowledge, then we cannot decide between the two theories.
(If this is a bit abstract, here's a toy example:
A = M and (if p then q) [assume M says nothing about p or q]
B = M and (if p then not-q)
C includes the sentence 'not-p';
so neither of A and B implies the other; if C is true then A and B are logically equivalent, while if C is false then A and B are inconsistent.)
My question: is anything philosophically interesting going on here? If we hold r to be true, do we really need to choose between (r or s) and (r or not-s)? I think not -- though they differ in logical content, they are not rivals (or are they?). At least, if we take r to be true, then they are definitely not rivals, though they might be considered rivals ‘on their own’. They certainly are genuine competitors when we hold r false -- though then they are no longer equivalent in any sense.
Comments:
1. This is not the usual sort of underdetermination situation. First, the notion of "empirical content" (or "empirical equivalence," i.e. identity of empirical content) does not appear, so the much-maligned observable/ unobservable distinction is never mentioned. Second, and more importantly, the two theories A and B are not incompatible: the Cartesian demon, on the other hand, is either deceiving us or not (at least on the usual interpretation). The demon-hypothesis is incompatible with the 'real-world' hypothesis. On the other hand, 'if p then q' and 'if p then not-q' are not inconsistent -- we need simply hold that 'not-p' is true.
2. How does this relate to a ‘real’ example, e.g. Bohmian mechanics? It is empirically equivalent to standard quantum mechanics as long as absolute position is undetectable -- but not if absolute position is detectable. In other words, the standard theory and the Bohmian theory are empirically equivalent modulo the claim that absolute position is undetectable. And that is structurally similar to the toy example above. (Of course, there is the difference in this case that the two theories are 'empirically equivalent modulo p,' not 'logically.')
3. Lastly, it is probably considerations akin if not identical to the above that prompted philosophers to move to ‘total theories’ (i.e. theories PLUS all their auxiliary assumptions) as the proper objects of epistemic evaluation. See e.g. (Leplin, Erkenntnis, 1997).
Subscribe to:
Posts (Atom)