I'll go bother a philosophy professor if I have to. But I'd rather ask Mefi first.
September 17, 2010 10:57 PM   Subscribe

I have a philosophy of language question. Or maybe it's actually a literary theory question.

I've been reading about naming and reference to try to get a handle on how philosophers treat referring terms with paradoxically "empty" referents, such as fictional proper names. I've read some Kripke, Russell, Pavel, and Donnellan or summaries thereof, and I think I understand most of the debate about rigid designators, definite descriptions, and the causal theory of names. I have some background in propositional logic and formal semantics, but I feel like that's not helping some of my bafflement here. I still have a couple of questions that I'd like to put to anyone who's familiar with these topics in the philosophy of language. If that's you, o brilliant Mefite:

1) is Russell's theory that proper names are definite descriptions in disguise really a theory about names being reliant on subjective states of knowledge? That is, does Russell think that the proper name X is tied to known properties of X because people can only refer to entities to which they have ascribed properties? Does this deny the possibility of an objective reality in which names are pure indices, not tied to anyone's state of knowledge or belief at all?

2) is there anything in the philosophical literature about referring terms having creative properties-- like, they presuppose entities into being ? Maybe this is an aspect of possible worlds theory? I can think of a literary theorist who has addressed something like this-- Derrida has this spiel about how the signers of the Declaration of Independence call "we the people" into being through the very speech act of self-naming/self-assertion involved in the act of signing. In other words, Derrida thinks that the signers who named themselves "we the people" actually used the power of naming to cause "we the people" to exist. But this isn't strictly applicable to proper names, and in any case I loathe Derrida and don't want to depend on him.

Anyway, this might be incoherent if I haven't actually understood anything-- but perhaps there's a philosopher out there who can set me straight.
posted by ms.codex to Religion & Philosophy (11 answers total) 10 users marked this as a favorite
not an answer, but if you don't find the answer you're looking for on AskMe, you might also try PhilPapers. I subscribe to a couple of their forums and it seems like they've got a fair number of professional philosophers willing to pounce on any question.
posted by nangua at 3:10 AM on September 18, 2010

Well, I'm not qualified to answer this, so I was hoping someone else would speak up. But as I remember things:

The reasons for Russell's theory are intuitions about how the meaning and truth value of sentences could depend on contingent facts.

For Mill's logic, as I've heard it explained, the meaning of a name was simply the thing it referred to. So if you had a sentence like "Thales lived in ancient Greece" and Thales turned out not to have existed, the name and the sentence would both end up being meaningless. This seems just wrong -- the sentence must have a meaning.

Frege tried to solve the problem by distinguishing between sense and reference. The sense is probably (but he was never very clear on the matter) some sort of description like: "the student of Plato" for the case of Aristotle. So for him sentences with empty names have meaning, but because they lack reference they lack truth-value. So if Thales turns out not to have existed, the sentence "Thales lived in ancient Greece" would be neither true nor false.

This was an improvement but still seemed counter-intuitive to Russell. He thought the sentence would obviously be both meaningful and false in that case. So he decided that names without guaranteed reference are actually descriptions. Thus "Aristotle was born in Stagira" can be divided into two different truth claims: 1) There was a thing that was the student of Plato, and 2) and that thing was born in Stagira. If either one of these are false, then the sentence is false. This way the sentence "Aristotle was born in Stagira" definitely has a truth-value. This, as far as I know, is the reason for his theory of descriptions -- just to preserve his intuitions about what sorts of sentences have truth-values.

There were also "real" names that were not definite descriptions -- but they needed to have some sort of guaranteed reference. I think he called this "direct acquaintance" and I don't remember anymore if ultimately the only two real names in the language were "this" and "that" or there were any other candidates. Maybe this is what you mean by a "pure index". I'm not sure how this met the need for "objective reality" but I think it did fulfill some kind of epistemological purpose for him. Anyway, even if turns out to be true that the meaning of my sentences depends on my own knowledge, this doesn't exclude the possibility of objective reality. Presumably my knowledge, which my meaning depends on, is knowledge of objective reality. So I think these are separate issues entirely.

Both Frege's and Russell's theories might have had the unfortunate result that they tied meaning to subjective states of knowledge. Frege says somewhere something like: if I only know Aristotle as the student of Plato and you only know him as the guy from Stagira, then we're speaking two different languages as far as it concerns the name Aristotle. I think this is an ugly and absolutely untenable consequence of the theory and certainly not the reason for the theory in the first place, and I'd guess that Russell probably felt the same way: he probably wasn't trying to tie meaning to subjective knowledge.

Some people have tried to replace one definite description with some sort of vague and variable cluster of description, and then you know how Kripke proposes a much different solution.

To your second point: it obviously relates vaguely to Austin's point about how certain speech acts create the reality they report on, like "I christen this boat so-and-so". But that's about whole speech acts and not individual names. As far as names creating their references: I'm not sure if it's true that this really happens in the case of "we the people". Do you have any other examples?
posted by creasy boy at 5:37 AM on September 18, 2010

I Am Not A Philosopher Of Language. I did a year of it at graduate level, but found myself agreeing with David Armstrong's dictum of "put semantics last". I'm a bit rusty on the literature. With that said, I'll try and say something vaguely intelligent which may prove useful in the event that someone who actually knows what they are talking about turns up and gives you a proper answer:

1. That's not a bad analysis. I think the "subjective states of knowledge" bit might be pushing in the wrong direction. Descriptive content doesn't rely on epistemic relations any particular knower has to that content. If Russell's answer to the reference problem worked, it would work just as well for a race of epistemically infallible androids as it would for epistemically fallible humans. Rather, I'd go in the other direction and say that the points made by Donnellan (with the 'man drinking champagne' example), Kripke and Gareth Evans show that Russell's analysis didn't take enough account of the interplay between language and subjective states of knowledge!

2. I can't think of anything directly relevant beyond Austin as creasy boy suggested.

And I feel similarly about Derrida: my initial reaction to the description he has simply shifted the problem here from the semantics of indexicals or whatnot into a metaphysical term, hanging the problem on what he means by 'existing'. Okay, yes, we say "we the people" and suddenly "we the people" thing magically comes into being. But what does that mean? If it is simply the set of those to whom the term refers, it existed before, and you are just now picking it out.
posted by tommorris at 6:49 AM on September 18, 2010

is there anything in the philosophical literature about referring terms having creative properties-- like, they presuppose entities into being ?

You'll want to look at John Searle's brand-new book Making the Social World. This idea runs throughout the whole book, but he especially zeroes in on it at pp. 109-15 ("Language is constitutive of social reality").
posted by John Cohen at 7:58 AM on September 18, 2010 [1 favorite]

1) is Russell's theory that proper names are definite descriptions in disguise really a theory about names being reliant on subjective states of knowledge?

If you take that approach, then there is nothing specific to proper names that relies on states of knowledge. Russell is specifically arguing his theory of descriptions to avoid the weird ontological commitments that Frege makes (i.e. he requires some kind of Platonic realm of senses). In a broad sense, all language is dependent on states of knowledge.

Does this deny the possibility of an objective reality in which names are pure indices, not tied to anyone's state of knowledge or belief at all?

Yes. Even if you are directly acquainted with some guy named "Bob" (assuming you will allow direct acquaintance to physical objects), the connection between the name "Bob" and the physical person is subjective, not objective. Russell is not claiming that the connection exists in any objective way - his entire approach here is subjective.
posted by ssg at 9:26 AM on September 18, 2010

1. In short, no. It's not a theory of psychological events, it's a theory of semantic content. Both he and Frege thought you had to differentiate between what someone might have been thinking (which Frege called the "idea" of an expression) and its meaning. Individuals might have any number of associations for an expression that do not arise to the level of meaning. For instance, if you say "Lehigh Ave." (a street in Pittsburgh), I immediately think of the dismal last two years of a relationship with my ex-girlfriend. But that's not what "Lehigh Ave." means by any stretch. If anything, you can think of what Russell was talking about as something that your subjective states needed to match up with or catch up to.

2. I know there are people who think about the metaphysics of fictional characters. (I know some personally, in fact, and saw one give a paper on this last weekend.) It makes a certain sense to say, "Sherlock Holmes was a detective" and if you say "Sherlock Holmes was a podiatrist," people will often unreflectively say that you're wrong or that it's false, when in fact there's no Holmes in the actual world to be right or wrong about. Shunting him off to another possible world is one way to handle this, but then again the reality of other possible worlds and the counterparts in them is at least as troubled and contentious as fictional entities. My friend who was giving the paper last weekend suggested the alternative (there may be earlier literature on this - I'll admit it's not my area) that we treat fictional characters as abstract objects - in particular, representations. If we say that something is an abstract object, then there is no claim that it is concretely instantiated somewhere. Asking where is Sherlock Holmes doesn't make any more sense than asking where is the number 37? The truth of claims about abstract objects is not a matter of some concrete object having properties that those claims represent. Instead, we have some series of formal methods for determining those truths - mathematics in the case of the number 37, and practices surrounding authorship and authorial intent when it comes to fictional characters. Whether abstract objects are real things in their own right or just conceptual constructions that we could handle differently is a matter of debate, too, but at least it makes fictional characters no more problematic than many other things we want to take seriously. And of course, there are those who think fictional characters are not real at all, just some sort of conceptual constructions, but you can imagine how those more strict views will go.

As for the Derrida comment, you might want to separate that from the creation of fictional characters. (Full disclosure: I'm not really a fan of Derrida.) What those who wrote and signed the DoI took themselves to be doing was not the creation ex nihilo of a new entity - THE PEOPLE - that would now start doing things. To appeal to the first person plural in that way is to assert some authority to make commitments for those you speak for and lay some corresponding obligations at the feet of others. Wilfrid Sellars used to talk about "we-intentions." These were not like my personal intentions to take some action, but rather were sort of proposals put forth to pursue some joint goal and establish norms that will move us towards that goal. It's more like saying, "This is how things are going to go around here from now on." Those can fail, of course, either because someone else has the authority to knock them down (or the guns to do so) or because whoever makes them doesn't have the traction with those around him or her to do so. (Imagine a crank putting a manifesto on the wall declaring - "WE WILL NOT STAND FOR THE GOVERNMENT'S LIES ABOUT UFOS ANYMORE!!") That's just one take on this subject of course, but note that there doesn't have to be another thing - THE PEOPLE - in the world or suddenly created to make sense of what the DoI means.
posted by el_lupino at 11:23 AM on September 18, 2010 [1 favorite]

I think some of the above comments about your "we the people" example are off-base. It might be counterintuitive to say that the phrase "we the people" had a role (along with a lot of other language/ideas) in "creating" the American people, since all those human beings exist no matter how anyone describes them. The language didn't literally give birth to us. But would you say that language doesn't create a husband and wife because, after all, the two human beings who constitute the married couple would exist no matter language was used to describe them? No, the two physical human beings exist, but we use language to create new institutional roles that they step into. Language confers status on pre-existing collections of molecules (including humans). By analogy, given that the 300,000 human beings within the US borders exist, they don't naturally possess the quality of "being Americans" or "being United States citizens." We have a national identity only because people have decided to do things like draw national boundaries, create a federal government, send persistent messages to the effect that all 300,000 individuals on this enormous area of land share a unified national culture, and so on.

Here's an inanimate analogy, which Searle often uses: you have a slip of paper -- part of a dead tree -- with ink printed on it in your wallet. Let's say it's a $5 bill. This is an objective fact -- it's a $5 bill whether or not anyone knows about or talks about that specific bill. (For instance, it could fall under your couch and be forgotten by everyone, but it'd still be a $5 bill as a matter of fact.) But the natural, physical thing (dead tree + ink) doesn't have any necessary social status. The object is defined by how people have chosen to use language to infuse it with institutional significance over time. I imagine that the $5 bill example is relatively uncontroversial; I don't see any reason to deny that an analogous process also applies to human beings with institutional/social roles.
posted by John Cohen at 1:24 PM on September 18, 2010

Thanks so much, everyone-- to creasy boy for the grand overview of Frege and Russell; to el_lupino, ssg, and tommorris for the clarification that Russell doesn't address the role of the knower in his account of semantic content (if I'm getting that right); to nangua and John Cohen for the valuable references. On your last point, John Cohen-- yes, on reflection it seems that the Derrida example wasn't a good analogy for the situation of fictional names, because, as you point out, Derrida was talking about using performative language to magick social relations into existence, not any referent in itself. Hmm. I'm going to look up the Searle book mentioned above to see if that comes closer to what I'm looking for; or maybe what I mean by "the creative properties of naming" is really a variant of what Austin says about the mutual entanglement of performative and constative language. I'll post here again if I have a breakthrough. Thanks again to all.
posted by ms.codex at 7:04 PM on September 18, 2010

Searle is very much standing on the shoulders of Austin. It's not like they represent two opposing schools of thought.
posted by John Cohen at 7:21 PM on September 18, 2010

I missed this over the weekend. (More philosophy AskMes plz!) I'm only really able to answer your first question. I think that some people above are saying false things about Russell. (Russell changed his views on these matters a lot though.)

is Russell's theory that proper names are definite descriptions in disguise really a theory about names being reliant on subjective states of knowledge?

Whenever I read a question like this, I try to strip out the words 'objective' and 'subjective' and see if the question still makes sense. They usually muddle things up more than help. Creasy boy is right, up above, in saying that states of knowledge are objective things. You could very easily have a semantic theory in which the meaning of my words is determined by my mental states. My mental states are objective, so meanings on such a theory could be objective as well.

That is, does Russell think that the proper name X is tied to known properties of X because people can only refer to entities to which they have ascribed properties?

This is a good question. I think the reason you're having issues here is that Russell's theory of language was intimately interwoven with his theory of epistemology. He did not care quite so much about how English sentences carried meaning: he wanted to know how thoughts carried meaning. He was after a language of thought; a language of reason. (And he tried to formalize this language of thought and reason in Principia Mathematica-style first-order logic and set theory.) Contemporary philosophers of language are concerned more with the workings of languages as they are spoken. They are not after enormous theories about the nature of thought and logic---they just want to know how English works. So they are more careful about respecting the line between philosophy of language and epistemology.

Anyway, it seems to me that there are two questions here. Firstly, can we refer to X only if we have knowledge of X? And secondly, is knowing the properties of X the only way we have of knowing X? (That is: can we only know X by description?)

The second question is easier to answer: no. Russell held that an object can be known either by acquaintance or as the unique satisfier of some description. However, the problem of error led Russell to believe that the world is hidden behind a veil of sense data, and thus, sense data are the only things we can immediately be acquainted with. This means that there are very few actual proper names. All physical objects are only known by description. There are deictic pronouns like "that" (only if it's used to refer to a patch of sense data, though), perhaps "I", and that's about it. All other names are definite descriptions. An apparent name like "Barack Obama" secretly has the structure "the current president of the US" or something like that. (Kripke eventually killed this theory.)

The answer to the first question is 'yes'. This is the troublesome bit that I think is what's causing you consternation: if the reference of a proper name depends on knowledge by acquaintance, then it will only be applicable to a private language. Now, what about the names that are only apparent proper names---that are actually covert definite descriptions? Are they also limited to private language? Maybe not---philosophers tend to strip out the epistemological bit and say that Russell thought that names were "associated" with definite descriptions or something, instead of saying that the objects must be known by description. But in truth, Russell just had very little to say about the semantics of natural language. His was a theory of thinking, not a theory of saying. It's wholly plausible that he was just giving the semantics of private thought, and would balk at definite descriptions being "associated" with names in a public language.

So, yes, you are right to think that Russell's theory is "subjective" in a certain way. He isn't trying to give an account of public language, so Russell felt free to pull on all sorts of epistemological relations to build his reference relation. This let him do all sorts of things that semanticists today would view with horror: for instance, mangle the surface structure of English in order to account for the semantics of 'the'. After all, who cares about the structure of English sentences? What matter is the structure of thought, and that can be very different from English. (Semanticists today treat 'the' as a quantifier, which respects its syntactic category in English. It very much bugs me that most intro logic books still teach a Russellian semantics for 'the'.)

If you want to read a little more about about why Russell tied his semantics so tightly to his epistemology, Gareth Evans discusses it a little in his chapter on Russell in Varieties of Reference.


Does this deny the possibility of an objective reality in which names are pure indices, not tied to anyone's state of knowledge or belief at all?

I'm not too sure what you're asking, but nothing about an objective reality should be threatened by any of this. Russell thought that all of us are trapped behind our veil of sense data, and that we cannot know anything by acquaintance in common with one another, but he did think that an objective world existed and that we could both refer to the same things. We can both know the same object by different descriptions.

Of course, he later devised neutral monism and decided that everything is a logical construction of sense data, but I can't begin to pretend I know what was going on with that.

Oh wait, I thought of something you might want to read with regard to your second question: check out Ian Hacking on "looping kinds." These are natural kinds such that our system of classification has altered the kind, such that we've altered classification, which has altered the kind, etc. The Social Construction of What? is a good book to look at. It's not really philosophy of language though.
posted by painquale at 11:38 PM on September 21, 2010

Hey painquale, just came back and found your amazing response-- thanks!
posted by ms.codex at 6:32 PM on October 4, 2010

« Older Help me navigate the world of camping purchases...   |   Trollface is my bridesmaid. Newer »
This thread is closed to new comments.