« A Curious Extrapolation | Main | On 'Homophobic' »

Friday, February 05, 2016

Comments

Feed You can follow this conversation by subscribing to the comment feed for this post.

>>The collars are, in themselves, senseless tags.
This seems remarkably similar to my description of Mill's theory of proper names, which I emailed last week :) Mill uses the analogy of a chalk mark upon a door, and notes that if we then put chalk marks on all the doors, then the original mark fails to identify any more.

However, Mill extends the analogy by saying that the mark is not upon the object itself, but on our idea of the object. Perhaps there is a grain of truth in this, but only a grain.

Ed,

In that same e-mail, you write, "(And please don’t confuse denotation with reference. For Mill, the name ‘man’ can denote). "

Part of the problem with this topic -- and every philosophical topic -- is the fluidity of the terminology. Apparently, you think that denotation is not a sort of reference.

Quine disagrees. He speaks of "two sorts of reference: designation and denotation." (Theories and Things, p. 43) Names and singular descriptions designate their objects. "A predicate denotes each of the objects of which it is true." (Ibid.)

True. Evans talks about ‘genuine reference’ to distinguish it from the lesser sorts of reference. So I think that denotation is not a kind of genuine reference. Nor is ‘lesser reference’ a genus of which genuine reference is a species.

But moving on. I think we all agree there is something to the idea that proper names are minimal. They don’t signify any attribute belonging to the bearer - except perhaps being the bearer - since we could always use the name to deny that the bearer had that attribute. For example, if Socrates meant ‘the greatest philosopher of ancient Greece’, we would not be able to say ‘Socrates might not have been the greatest philosopher of ancient Greece’. But we can’t say ‘Socrates might not have been Socrates’.

But you are familiar with these arguments. Given that ‘Socrates’ is minimal, and given that it cannot signify its bearer – which we also agree, against the majority of current philosophers of language – what does ‘Socrates’ mean?

If you think there are referring terms, do you think that only proper names are referring terms or do you include also Russellian definite descriptions such as 'the wisest Greek philosopher'?

Would you agree that a Russellian definite description is, by definition, a DD used attributively in Donnellan's sense? I would say yes. What about 'the man'? A DD? A Russellian DD?

'Signify' is one of those weasel words (slippery terms) you apparently picked up from the scholastics. Or so it strikes me. Please define 'signify.'

>>I think we all agree there is something to the idea that proper names are minimal. They don’t signify any attribute belonging to the bearer . . . For example, if Socrates meant ‘the greatest philosopher of ancient Greece’, we would not be able to say ‘Socrates might not have been the greatest philosopher of ancient Greece’. But we can’t say ‘Socrates might not have been Socrates’.<<

That there is something to the idea is accepted by all, including Lukas. Right Lukas? After all, one cannot simply dismiss Kripke's modal argument which you have just sketched.

But I should think that Fregeanism is better than Millianism -- which has unacceptable consequences.

And then there is the modal metaphysics at the back of Kripke's argument which is not exactly self-evident.

>>Given that ‘Socrates’ is minimal, and given that it cannot signify its bearer – which we also agree, against the majority of current philosophers of language – what does ‘Socrates’ mean?<<

I would put the puzzle like this.

1. 'Socrates' cannot have a reference-determining sense (a Fregean sense). For Kripke's reasons. If 'Socrates' refers, the reference is not routed through a Fregean sense.

2. 'Socrates' cannot be a senseless tag or label. Labels don't refer. If 'Socrates' refers, the reference is not direct.

3. 'Socrates' refers to Socrates.

4. Proper names are among the paradigmatically extralinguistic referring devices.

The London solution will be to deny (4) and maintain that 'Socrates' does not extralingusitcally refer. There is no extralinguistic reference. All reference is intralingusitic, intra-textual, back reference and cross reference. Eliminativism about extralinguistic reference!

Am I 'tuned in' to your way of thinking, Ed?

>>'Signify' is one of those weasel words (slippery terms) you apparently picked up from the scholastics. Or so it strikes me. Please define 'signify.' <<
To communicate what one intends to say, using signs. To betoken. From the Latin, signum, mark, token etc. I have never got why you don’t like this. Is it because you believe that only people can signify, not signs? Surely it is the signs that signify. A red traffic light signifies that traffic must stop.

>> The London solution will be to deny (4) and maintain that 'Socrates' does not extralingusitcally refer. There is no extralinguistic reference. All reference is intralingusitic, intra-textual, back reference and cross reference. Eliminativism about extralinguistic reference! Am I 'tuned in' to your way of thinking, Ed?<<
Yep, which is why I am puzzled by your claim that the IL theory is ‘incomprehensible’. It may be totally wrong, but the fact you can express it reasonably well is proof that it is comprehensible, surely?

My only quibble is that you probably take this to imply that ‘ “Socrates” refers to Socrates’ is false. Not so. According to the IL theory, it is perfectly true. As is ‘“Frodo” refers to Frodo’. The latter fact is why I call it the ‘“Frodo”-Frodo theory’. You get the allusion, right?

The IL theory, correctly understand, is difficult to refute for the same reason that Johnson’s reply to Berkeley fails. As you know, Johnson kicked a stone, saying ‘I refute it [Berkeley’s theory] thus’. But as most philosophers acknowledge, this is a naïve refutation only. Berkeley of course is not saying that stones don’t exist, nor that they are hard, and can hurt your foot without strong shoes. By the same token, it is not enough to object that ‘Obama’ really really refers to Obama, that Obama is a really really actually existing object in the real world etc etc. This will not do. The IL theory does not deny any of this. Rather, the semantic mechanism that explains ‘Frodo’ referring to Frodo is the same as the one that explains ‘Obama’ referring to Obama. This is what you need to get your head round.

Ed,

I retract 'incomprehensible.' I was speaking loosely in violation of my own 'prickliness.' Mea culpa, mea maxima culpa.

More later. Now I am off to breakfast with Peter Lupu and two others. We will no doubt talk a little about you and this will involve referring to you using various devices. The references will all be extralingustic as befits your status as much more than a bit of text.

We will refer to you and say only good things about you.

>>We will refer to you and say only good things about you.

You refute it thus?

Have a good breakfast.

See also argumentum ad lapidem ‘a logical fallacy that consists in dismissing a statement as absurd without giving proof of its absurdity.’

Ed,

We agree that the Johnsonian refutation is lame. (Nice pun, eh?) Kicking a stone may hurt your foot but cannot refute Berkeley since he doesn't deny that there are stones or physical things generally. The good bishop is not an eliminativist about stones. But you are an eliminativist about extralinguistic reference. Or so it seems to me.

*Ad lapidem* is not in every case a fallacy. If you deny what obviously exists, then I don't need to argue against your 'theory'; it suffices to point out that this or that exists. For example, eliminativism in the phil of mind denies that there are beliefs and desires and mental states generally. I refute it thus: I am perplexed, angry, etc. that you should say such a thing.

Suppose someone claims that there are no pains. I don't argue with him: I kick him in the shins or higher up. Thus I refute him without saying a word.

It is a datum that there are pains. It is not a datum that realism about the physical world is true. Is it a datum that there is EL reference? Not as clear.

When you say 'I am hungry' you refer to E.B. How do you explain the reference of 'I' in terms of back-reference?

>>When you say 'I am hungry' you refer to E.B. How do you explain the reference of 'I' in terms of back-reference?<<
I think I explained this a while ago. The IL theory begins as a theory of narrative. It starts by explaining fictional narrative, and the semantics of back-reference in a fictional context. I think we agree that back-reference is the only economical explanation of this, right? Let me know if you think not, and give reasons.

Then we move to texts which have a narrative format, which are written as true, but whose truth, in particular whether any of the characters existed or not, is uncertain. For example the story of Tobias in the book of Tobit. The claim here is that the semantic mechanism underlying how we know which character each sense is ‘about’, is fundamentally the same, whether the narrative is genuine history or not.

Then to texts which we regard as ‘history’. History starts with a collection of primary materials, which the historian weaves into a narrative. The Latin historia means a narrative of past events, from the Greek ἱστορία, a narrative account of one’s enquiries.

Back-reference is the most economical and plausible account of narrative semantics. Narrare, to tell, relate, narrate, report, recount, set forth.

Moving now to indexicals, and the world-narrative or world-text. The world-text is a narrative whose language is accompanied by pictures and sounds and other signs that add to the semantic content of the propositions the world expresses to us. Think of how it works in films. The pictures speak, as it were. The claim of the IL theory is that all identification is re-identification, and that the first time we encounter a new person, the corresponding sentence in the world-narrative is indefinite only. Thereafter, when we ‘recognise’, i.e. re-cognise someone (re+cognoscere), the corresponding sentence is definite, i.e. involves back-reference.

Turning finally to your question. You asked “How do you explain the reference of 'I' in terms of back-reference”? Well, how do you understand the reference of ‘I’, when it appears, outside of quotation, in a comments post saying Posted by: London Ed?

Tobit begins his story I am Tobit and this is the story of my life. On our very first encounter with the story, how is the semantics other than indefinite? Someone is saying to you, through the medium of this story, that he is called ‘Tobit’.
(Weirdly, the narrative changes to the third person in 3:16 ‘As Tobit and Sarah were praying’. Thus ‘Tobit’ back-refers to the ‘I’ of the earlier narrative, and presumably ‘I’ now refers to whoever is ‘speaking’ to you from 3:16 onwards).

So the IL theory is not saying that the world doesn’t exist, or that all reality is linguistic. Not at all. The world-text is constantly bombarding you with signs expressing propositions that you mostly accept as true. However, the theory claims that all definiteness is linguistic only, and that all perception is essentially descriptive and thus indefinite in character. Such indefiniteness often masquerades as definite, such as in the illusory uniqueness of voices and faces. But this is an illusion.

>> But you are an eliminativist about extralinguistic reference.

I am an eliminativist about definiteness, if you like. It exists, just as 'demonic possession' exists, but is not what it appears to be.

I can't tell whether I disagree with London Ed's theory. I certainly wouldn't talk the way he does when he denies extralinguistic reference. But it's not clear to me that I'm actually disagreeing with the substance of what he says about how reference works.

"the semantic mechanism that explains ‘Frodo’ referring to Frodo is the same as the one that explains ‘Obama’ referring to Obama."

This I agree with completely.

But is Frodo a bit of language? I should think not. A Hobbit is a living, breathing creature, not a bit of language. Of course there are no such creatures, and there are bits of language about them, but the distinction doesn't disappear simply because one of the two things distinguished doesn't exist. If I seek a unicorn I am not seeking for a bit of language nor for an idea, but for a living creature. I won't find one. But if I did, it would be a physical thing that I found. If I seek and find a horse it is a physical thing I seek and find. The cases are the same except for the fact that in one case it's a non-existent living creature and in the other an existent one.

As with seeking, so with referring. When I refer to Frodo and Obama I am referring to persons, not bits of language. The semantic mechanism that makes this happen doesn't require the existence of the entities to which I refer.

Am I missing something?

Christopher: frodo is not a bit of language.

Dear Bill,

you write:

Suppose sense determines reference. And suppose the sense of 'Socrates' is specified by the definite description, 'the wisest Greek philosopher.' Used attributively as opposed to referentially (Donnellan), this definite description is non-rigid: it picks out different individuals in different possible worlds. So [...] the reference of 'Socrates' will be non-rigid.

I think this is – a very powerful – illusion.

First, to avoid misunderstanding: I am not suggesting anything like Donellan's referential use of a description. In referential use, the sense does not determine the referent (as the Martini case illustrates). I am suggesting that a sense can determine the referent in two different ways, conveniently called de dicto and de re. For some reason, only the de dicto way is commonly acknowledged.

I would like to invite you to contemplate some examples first. Consider this statement:

(N) Cruz could lose in Iowa.

I think we all will agree that it is true, and that "Cruz" somehow rigidly refers to Cruz across the possible worlds, whereas "could lose in Iowa" predicates of Cruz that he lost in Iowa in some possible world.

Now I say that in natural language an equivalent statement can be made this way:

(D) The winner in Iowa [take it as a shorthand for as rich a description as to pick up Cruz uniquely] could lose.

Can it be denied that thus intended statements are commonly made (e.g. in a journalists' style)? If not, then we had better account for them in our theories. Donellan's "referential use" may (among other motivations) be an attempt to do so, but it is no good (for this purpose): in these cases it is essential that the referent satisfies the description in the actual world. So I say that there is one more kind of attributive use: viz. such that the description is used to pick up whatever satisfies it in the actual world, and to this it refers, across all the possible worlds.

Why should such kind of reference not be admitted? Can qualms about it perhaps stem from a kind of "Lewisian" understanding of the possible worlds (aptly criticised by Kripke in Naming and Necessity)? I.e. possible wordls as isolated parallel universes, any difference of their ontological status being either denied or, at least disregarded in modal contexts? In such a scenario it is indeed difficult to imagine that an individual be descriptively identified in one possible world and something predicated of it with respect to a completely unrelated world world where it is quite different.

However, the universe of possible worlds is just a technical description of the potentialities inherent in the actual world. Only the actual world is really there. There are no unsurmountable gaps between the worlds, no "trans-world identity" problem. The identity of individuals across the worlds is assumed from the beginning - since it's these actual individuals whose potentialities we wish to describe by means of the PW apparatus.

So to say that the description "the winner in Iowa" picks up Rubio in some other possible world is just to say that that description would pick Rubio if some counterfactual circumstances obtained - but since these do not actually obtain, it does not pick up Rubio, as a matter of fact. Morover, given that (i) the possibility to lose - i.e. the fact that he loses in some other possible world - pertains to Cruz in the actual world (and not to some completely distinct and unattainable individual in a parallel universe regarded as Cruz's counterpart), and that (ii) Cruz is the referent of "the winner in Iowa", it is true to say (D).

In this perspective it appears that it is the de dicto, non-rigid reference which looks complicated and needs some explanation. The "default" behaviour of a nominal expression in subject position seems to be just to refer to the individual which as a matter of fact satisfies its sense, i.e. to refer rigidly, de re. To let a expression vary its reference according to what would satisfy it were the given world actual requires a special context: the expression must appear within the scope of an (hyper)intensional operator, which only can "relativize" its reference - like this (I am loath to use these formalizations, since I don't think it is primarily a "matter of scope"; it manifests itself as a matter of scope when squeezed into the Frege-Russell procrustory) :

DE DICTO: Qw([(ix:D(x,w)](F(x,w))
vs. DE RE: [(ix:D(x,wa)]Qw(F(x,w))

Q - a quantifier
ix - definite description operator treated here like a restricted quantifier
w - variable for possible worlds
wa - constant denoting the actual world
D - a description-forming predicate
F - any predicate

Lukas,

>>(N) Cruz could lose in Iowa.

I think we all will agree that it is true, and that "Cruz" somehow rigidly refers to Cruz across the possible worlds, whereas "could lose in Iowa" predicates of Cruz that he lost in Iowa in some possible world.<<

I think what you want to say is that Cruz might have lost in Iowa. In fact, he won. But he might have lost where 'might have' is being used metaphysically not epistemically. We don't need to bring it possible worlds jargon at all.

Now suppose the sense of 'Cruz' is given by 'the winner in Iowa, etc.' Kripke's modal argument is essentially this:

It is true that Cruz might have lost in Iowa, but it is false that Cruz might not have been Cruz; ergo, the reference of 'Cruz' is not determined by the sense of 'the winner in Iowa.'

I am not understanding you. You say that we all agree that 'Cruz' refers rigidly? That's not true. Some of us are descriptivists.

To explain my point more fully, I hold all of the following are true:

(1) According the LOTR, Frodo is a hobbit
(2) In LOTR, Frodo is a living, breathing existing creature
(3) In LOTR it says that hobbits have large feet
(4) In the story of LOTR, hobbits are not linguistic items
(5) In the text of LOTR, ‘hobbit’ is a linguistic item
(6) In Lotr, ‘Frodo’ is the name of a hobbit
(7) In Lotr, ‘Frodo’ is the name of a character
(8) In Lotr, for some x, ‘Frodo’ is the name of x

Etc. But all of the following are false
(9) In LOTR , hobbits are linguistic items
(10) For some x, it says in LOTR that x is a hobbit
(11) For some x, ‘Frodo’ is the name in LOTR of x

Back reference is of course rigid. For example:

Obama was born in Hawaii, but he might not have been
I.e. if his father had stayed in Kenya, Obama would not have been born in Hawaii. What about
The current president of the US was born in Hawaii, but the current president of the US might have been born in New York
Here the second definite description does not back refer to the first. We are clearly supposing that a different person was born in New York. Or how about
The current president of the US was born in Hawaii, but he might not have been the current president of the US
The ‘he’ is rigid, ‘The current president of the US’ is not.

Bill,

I apologise for my garbled modal English. Also, I see I was misunderstood because I argued as if against Kripkean intuitions (most of my colleagues are Kripkeans, I, too, used to be).

I say that even if the sense of "Cruz" is given by "the winner in Iowa, etc.", there are two readings (with respect to the predicate) both of

(A) "Cruz might have lost in Iowa"

and

(B) Cruz might not have been Cruz".

The de re readings are both true, the de dicto readings are both false. Kripke choose to consider the de dicto reading of (A) and the de re meaning of (B).

There is a further complication in Kripke's statement of the argument, namely that he varies the predicate, not the subject, which for those wedded to the Fregean analysis of predication invokes the question whether these statements should be interpreted as identities or predications: I am trying to bracket this complication for the sake of simplicity (this is why I used examples where the subject is varied).

I agree that there is a refutable presumption that names are used de re whereas descriptions, at least in the predicate, de dicto. This is what makes Kripke's argument intuitively appealing. But there are cases of names being used de dicto ("Henceforth I never will be Romeo"; "Why did you become Jack the Ripper?"), which must be explained away by Kripke as not being genuine names, and there are descriptions used de re (in the subject position quite often) - not sure how Kripke handles this, but he explains away the semantic import of obviously "descriptive" names by divorcing "reference-fixing" from the sense (and, by implication, positing synthetic a priori truths, which for me is a kind of reductio argument by itself).

Now I say that whenever a subject refers de re, it refers rigidly; whenever it refers de dicto, it refers non-rigidly.
So any de re modal statement with a description at the subject place proves that descriptions can refer rigidly, which ruins Kripke's argument that names are non-descriptive because rigid.

Let me make an argument of it, to clear it up:

(1) In de re singular modal propositions (or more broadly: in de re non-extensional contexts), the subject refers rigidly.
(2) There are de re singular propositions (non-extensional contexts) with descriptions as subjects.
Ergo: descriptions can refer rigidly.
Ergo: the sense of names can be given by descriptions and still they can refer rigidly.

Lukas: >>The de re readings are both true, the de dicto readings are both false. Kripke choose to consider the de dicto reading of [Cruz might have lost in Iowa] and the de re meaning of [Cruz might not have been Cruz]).. <<

I am not following this at all.

What I am particularly struggling with is the ‘de dicto meaning’ of ‘Cruz might have lost in Iowa’, or ‘Cruz might not have been Cruz’.

Partly this is because I did not know who Cruz was. Having referred here I now understand that there is a person called ‘Cruz’ who is an American politician and a junior U.S. Senator from Texas. He is a Republican candidate for President of the United States in the 2016 presidential election. Right. I also learn that he, i.e. Cruz, won in Iowa. But I am unable to read ‘Cruz might have lost in Iowa’ as meaning anything but that it might have been the case that he, i.e. that guy I just heard about, i.e. Cruz, might have lost in Iowa, although he actually didn’t lose. Note the italicised referring expressions.

As for ‘Cruz might not have been Cruz’, that means that it might not have been the case that he was himself. That’s clearly false. What is this true reading?

If I may, can I link to my post on Intentional Identity, concerning a famous counterexample by Geach, which apparently resists either a de re or de dicto interpretation.

Ed,

First note that we are talking the (grammatical) predicates not subjects here. In order that Kripke's argument be relevant for the names vs. descritpions issue, we must consider the grammatical predicates, in Fregean analysis, as singular terms, and therefore the propositions must be analysed as predications of identity relation of the ordered quasi-pair of individuals. So the de re / de dicto distincion applies to the grammatical predicates, or subjects #2 in the Fregean analysis.

So we must in fact interpret "Cruz might have lost in Iowa" as "Cruz might not be the winner in Iowa" - which, unfortunately, is not a natural interpretation of the English sentence, but otherwise the example would be irrelevant.

Now the modalized predication of identity can apply to "the winner in Iowa" either de re or de dicto. If it applies de re, then the statement is about the actual-world referent of "the winner in Iowa", i.e. Cruz, of which it predicates some property (here "not to be identical with Cruz") with respect to other possible worlds. If it applies de dicto, the statement is, so to speak, about the descriptive content expressed by "the winner in Iowa", of which it says that some of its various referents in various possible worlds have not certain property (namely "to be identical with Cruz"). (I have ignored here the complication that the same distinction could be made about the first argument of the identity relation, "Cruz", i.e. that the property "to be identical with Cruz" might also be construed ambiguously, see below. This is why the Kripkean examples muddy the waters).

Now what about "Cruz"? It is a name, so it is not usually interpreted de re - the purpose of names is precisely to enable easy de re reference, so using them de dicto goes against the primary motivation of their existence. But it can be done, and sometimes it is done. "Cruz", IMHO, means something like "the man named 'Cruz'"; so it can be meaningfully said, de dicto as regards the latter occurrence of "Cruz", e.g. "It is not necessary that Cruz is Cruz" (as, e.g., at some point in the past Spanish names might have been banned for some reason, and so the Cruzes might have been Crosses by now) - like Romeo says that henceforth he never will be Romeo, not meaning, of course, that henceforth he never will be himself, but that he will cease to satisfy the identifying descriptive content associated with "Romeo".

Much less exceptionally you can use the "descriptive" names de dicto: "Jack the Ripper (de re) need not have become Jack the Ripper (de dicto)" (argues the prosecutor at court). And quite the same holds with descriptions: "Tomorrow, the King (de re) might not be the King (de dicto) any more" (said ominously the revolution leader) - i.e. "The person who is the King in the actual world does not have the property (or satisfy the descriptive content) of being the King tomorrow in some possible world". The first occurrence refers rigidly, the second non-rigidly.

So the names differ from descriptions in defeasible presumptions about their usual usage, but are not strictly bound to either rigid or non-rigid reference; and so expressing or not expressing descriptive content is independent from the issue of (non)-rigidity.

The comments to this entry are closed.

My Photo
Blog powered by Typepad
Member since 10/2008

Categories

Categories

December 2019

Sun Mon Tue Wed Thu Fri Sat
1 2 3 4 5 6 7
8 9 10 11 12 13 14
15 16 17 18 19 20 21
22 23 24 25 26 27 28
29 30 31        
Blog powered by Typepad