Talking About What's Not There: Reference, Meaning, and Empty Names
Before I continue I would like to make a distinction: there may be in fact two problems when talking about the problem of empty names. First and foremost, the proper name itself seems not to refer, so the use of the name, in itself, is a problem. I will refer to this problem as ENP1. However, an additional problem arises when using these empty names in sentences to express propositions. This shall be ENP2. In what comes I hope to spell out the consequences of both of these problems, some strategies used to solve either or both, and additionally attempt to give reasons why no account I've encountered so far is satisfactory.
Here is a brief, simple statement of ENP1: Usually, when using proper names, we are referring to objects. However, we sometimes use words that don't refer to existing objects. How does a theory of language explain this? Historical development of the problem rose to reactions against a theory of names proposed by John Stuart Mill. His theory was that a name's only purpose was to designate that which is being referred to. This seems to be the intuitive, pre-philosophical position. When uttering "Michael Jordan", we're simply talking about Michael Jordan. But the position gets complicated and downright contradictory very quickly: what if we're talking about a non-existing object? When I talk of "Frodo" I cannot be talking about an actual Frodo because "Frodo" does not denote any physical being; so either
(a) I am referring. By referring to 'Frodo' i am indeed referring to something that does not exist physically. This has the unfortunate consequence of a metaphysical explosion, where every and anything cognizable does exist and that which is physical just happens to subsist. Further, this view will force me to explain why or how any not materially-existing entity can have properties, if i wish to say anything about this non-existing entity. I will put this view aside for fear of ontological commitment. (From purely secondary sources, I understand this is the approach Meinong took). OR
(b) I am attempting to refer yet I am not. However, the name also carries with it some set of descriptions or a manner in which it was presented (a sense). The name, then, has no referent yet is not meaningless either (this is the approach Frege takes). OR
(c) I am attempting to refer yet i am not. Any name is replaceable with a definite description that insinuates its own existence. So any predicate attributed to this failed proper name would not hold because there's no subject to latch on to (this is the approach Russell takes). OR
(d) I am attempting to refer yet I am not. This makes the sentence meaningless. Attempting to attribute properties to nothing is impossible. (This is the approach Kripke et al. take)
(b) I am attempting to refer yet I am not. However, the name also carries with it some set of descriptions or a manner in which it was presented (a sense). The name, then, has no referent yet is not meaningless either (this is the approach Frege takes). OR
(c) I am attempting to refer yet i am not. Any name is replaceable with a definite description that insinuates its own existence. So any predicate attributed to this failed proper name would not hold because there's no subject to latch on to (this is the approach Russell takes). OR
(d) I am attempting to refer yet I am not. This makes the sentence meaningless. Attempting to attribute properties to nothing is impossible. (This is the approach Kripke et al. take)
The Causal theory of names gives the intuitive account of what a name means: its reference. In other words, 'meaning' and 'reference' are words which refer to the same concept. If this theory is true, then empty names are meaningless. 'Frodo', 'Santa Claus', and 'Pegasus' have no meaning. But surely we have some kind of ideas and mental properties associated with those words, even though they fail to refer to physical objects in the world. A person who uses the term 'Frodo' attempts to refer to some small hobbit; 'Santa Claus' attempts to refer to a jolly old man who gives presents in December; 'Pegasus' attempts to refer to a horse with wings. This theory has its problems, because if 'Frodo' fails to refer in the same way 'Pegasus' does, then they should be interchangeable within sentences where they're used without losing any significance. so 'Pegasus has the ring' and 'Frodo has wings' should be equally meaningless. However, this doesn't seem to go well with our intuitions.
The Descriptivist theory of names deals better with empty names than the causal theory, yet it has startling problems in other areas. It deals better with empty names because it equates these empty tags with descriptions of the intended properties of those tags, such that 'Santa Clause' would be translated as E(x) (Cx& Px & Jx & Rx) where C=is Chubby, P= gives Presents, J= is Jolly, R= is Red. The existential qualifier automatically implies 'Santa Claus' exists. Kripke pointed out that in effect these Descriptions which could be used as a strategy against saying empty names are meaningless would also make the properties attributed to the names necessary. Although I don't have much of a problem accepting this consequence, a lot of people do.
Neither of these two theories are satisfactory. The Descriptivist program ignores the fact that any proper name does have a causal link to that which is referred to, or at least attempted to refer. This is obviously and almost trivially true when the proper name accurately designates an individual in the real world. However, to properly use a name it seems as if we have to know something about exactly what we're talking about, not only have the ability to causally apply its use.
Take, for example, a case of mistaken identity in babies. Suppose a given woman has identical twins, and the mother has decided to name them 'Rex' and 'Fred'. At the time of birth, 'Rex' means Rex and 'Fred' means Fred. however, almost immediately afterwards a malignant nurse puts them in their cribs and switches their wristbands so that Rex's says 'Fred' and Fred's says 'Rex'. The entire rest of their lives Rex and Fred suffer from mistaken identity. But something seems wrong about this example, because it seems as though at some point the names that were originally intended for their persons becomes irrelevant, not only because of their repeated anchoring with a different name in such a way that the new name does begin to refer in other people's minds (as Evan's 1973 example of 'Madagascar'), but also because of the independent properties each individual twin begins to acquire and become associated with their new, supposedly incorrect names. That is to say, save talk of possible worlds, some properties do seem to have more of an anchoring in the individual than others: if Fred once saw a particular musician or Rex a particular sporting event, then them having done that doesn't seem to influence who they are very much. However, if Fred decided in grade 7 to become a physicist and Rex decided at a similar age to dedicate his life to practice of Fencing, then these bits of information seem indispensable to any moment of framing reference of the person at some point in the future. So there are in fact two ways here that the Causal Theory of Reference seems to make things a bit uneasy.
So how does this tie back in to the discussion of empty names? I guess in some ways I'm attempting to scale back on the sharp distinction Kripke made between the metaphysics and epistemology of reference, because I believe in order to accurately use any name, especially when it is an empty one, a certain knowledge of some of its properties in necessary in order to do away with the conclusion that the name is meaningless. This may be more in tune with the Frege's or Evans's Theory of Names more than Russel's or Kripke's.
ENP2 creates even more of a problem because empty names are being used in sentences, some of which are intuitively true and some aren't. Both 'Pegasus exists' and 'Pegasus does not exist' seem to, at a glance, express some proposition that is either true or false. Presumably, the former is false while the latter is true. Many of the same problems as in PEN1 return, yet there are some additional ones: If it is required of a proposition to include both a subject and a predicate for it to have meaning (and it seems to be so), then statements purportedly asserting some proposition which include a subject-term that does not refer do not actually express any proposition, under a strict causal theorist's view. So neither of the two statements previously mentioned actually express a proposition. Furthermore, if a statement does not express a proposition, it seems to be meaningless (at least according to Davidson), another conclusion we don't want to reach, both because of the examples of negative existential statements about empty names and from examples dealing with less technical examples such as 'Frodo exists.' We want to say that this statement is not meaningless yet not true either.
Perhaps what would be best as of this moment is to give a brief sketch of Davidson's theory of meaning, as it is an important one that elucidates the problem of empty names when applied used in declarative sentences. According to his theory of meaning, any proposition P stated in object language L by statement S has the following meaning: 'S' is true iff P. This seems to be true when applied to formal languages without self-reference and empty names and other sorts of messy stuff, which is why Tarski preemptively limited his theory of truth to those types of languages. But attempting to apply it to natural languages is not a false start either. Both 'S' and P must express some proposition with a subject and a predicate (ideally both 'S' and P share the subject, predicate and their appropriate relation). The given language is comprised of a set of axioms (reminiscent of Kripkean semantics for proper names) in such a way that any particular term only refers. But as soon as a term which includes an empty name is considered, or evaluated for its meaning, the biconditional is not true and thus the proposition fails to have meaning.
Let's use a concrete example: take into consideration the proposition that 'Santa Claus is jolly.' Let us name this proposition "Santa Claus is jolly." so that, to find its meaning, we must apply it in Davidson's meaning schema: "Santa Claus is jolly" is true iff Santa Claus is jolly. Now, if 'Santa Claus is jolly' is an illegitimate proposition, deemed so from the fact that Santa Claus doesn't refer to a physical object in the real world, then the proposition is meaningless and has no truth value. However, the left side of the biconditional asserts that that which we had just determined to have no truth value is in fact true, which turns out to be a false assertion (because the truth value of said proposition is undetermined). This makes the entire biconditional false and we have no meaning for the sentence. This is an undesired consequence because we do have certain cognitive associations with the proper name 'Santa Claus' that might give us reasons to believe he's either jolly or not.
Braun attempts to clean this problem up by proposing two possible solutions a supporter of the causal theory may adopt. The first option is ultimately to bite the bullet and say that this alleged proposition is in fact meaningless because of the fact that it isn't a proposition. Along with this option he proposes that there may in fact be a pragmatic replacement to what he calls the No Proposition view. If I understood him correctly, where a statement like 'Vulcan does not exist' may be uttered, the real content of the statement is a conditional along the lines of "If we were to check space at this location and time, there would not be a planet there." I am interested in this solution, yet he does not go into detail, so I will not consider it past this moment.
The alternative view, one that Braun himself prefers, is what he calls the Unfilled Proposition strategy. In this view, any statement using an empty name is in reality an attempt to attribute a predicate to something that is not there. However, it still saves the structure of propositional content and thus explains how we might be fooled into thinking that we are in fact expressing a proposition. Again, to give an example: 'Vulcan is a planet' is structurally similar to the assertion that 'Venus is a planet'. But where the latter has a propositional composition of <Venus, being a planet>, the former has a propositional composition of [ ____, being a planet]. However, even Braun himself brings up an objection to his theory, that give another empty name, any other empty name, the unfilled proposition would look identical. the statement 'McX is a planet' would also look like [ ____, being a planet] if McX doesn't refer.
At least on the face of it, Braun's argument in favor of the Unfilled Proposition view is at the very least susceptible to the very same arguments against Kripke-Millian views on proper names: if proper names only refer, then the use of empty names shouldn't make any difference in cognitive content, and statements about them should be equal in propositions expressed.
Again, the descriptivists run into big problems (outlined by Kripke) when dealing with positive propositions that assert properties onto things that obtain because these descriptions are doomed to be necessary. There is also the problem of circularity or reference, where the descriptivist might fall into a vicious circle if all he knows of a referent is the same property he uses to provide a definite description for the same referent. If 'Albert Einstein' is defined as 'The discoverer of the Theory of Relativity' and 'the Theory of Relativity' is defined as 'the theory discovered by Albert Einstein' and nothing else, then those two names would not be able to singularly pick out Albert Einstein and the Theory of Relativity, but would designate any two individuals with the relationship of discovery of a type of theory. These and many other problems make pure descriptivism inadequate for reference in most cases, yet for some of the same reasons this makes desciptivism strong explanatory-wise when it comes to empty names.
Is there a go-between, an intermediate position? Could it be that there is some theory that falls between these two arguably troubled enterprises that takes the best of both sides while being strong enough to defend against criticisms from both angles? I sure haven't found a perfect fit yet, but there is much to ask for a theory of meaning and reference. If there is one aspect I believe must be saved is the distinction between meaning and reference. Or if one aspect of meaning is just reference, maybe there is also an additional aspect to it that cannot be ignored. Even though descriptivism as described in Russell (1905) has been shown to be largely false, there's something in there that can be recovered. Likewise with Kripke, as his position saves many intuitions for non-empty names. Frege almost incorporated bits from both, but he also either left a lot of questions unanswered (how does a proper name tag an individual, or fail to?) or drew some unintuitive conclusions (that a sentence's reference is its Truth value). Either way, the role of fiction, meaning in context, metaphors, malapropisms, and countless other linguistic and semantic phenomenons have been left unaccounted for, and surely a complete theory of reference and empty names will have a lot of work to do.
Bibliography
David Braun. 1993. "Empty Names." Nous 27: 449-69.
-----. 2006. "Names and Natural Kind Terms." Handbook of Philosophy of Language. Lepore and Smith (eds). Oxford, EN; Oxford University press. Note: Found via Braun's website at URL: http://www.ling.rochester.edu/~braun/Papers/names.pdf
Donald Davidson. 1967. "Truth and Meaning" The Philosophy of Language. A.P. Martinich (ed). 3rd Edition, Oxford, EN; Oxford University Press.
Martin Davies. 2005. "Gareth Evans." The Encyclopedia of Philosophy. Donald Borchert (ed). 2nd Edition, McMillan Reference USA. Note: Found via Davies's website at URL: http://philrsss.anu.edu.au/~mdavies/papers/evans.pdf
Gareth Evans. 1973. "The Causal Theory of Names." The Philosophy of Language. A.P. Martinich (ed). 3rd Edition, Oxford, EN; Oxford University Press.
Gottlob Frege. 1892. "On Sense and Nominatum." The Philosophy of Language. A.P. Martinich (ed). Herbert Feigl (trans). 3rd Edition, Oxford, En; Oxford University Press.
Saul Kripke. 1980. Naming and Necessity. Cambridge, Ma.; Harvard University Press.
Marta Reimer. 2003. "Reference." Stanford Encyclopedia of Philosophy. Note: Found vie SEP at URL: http://plato.stanford.edu/entries/reference/
Bertrand Russell. 1905. "On Denoting." The Philosophy of Language. A.P. Martinich (ed). 3rd Edition, Oxford, EN; Oxford University Press.
Alfred Tarski. 1944. "The Semantic Conception of Truth and the Foundations of Semantics." The Philosophy of Language. A.P. Martinich (ed). 3rd Edition, Oxford, EN; Oxford University Press.
This work is licensed under a
Creative Commons Attribution-No Derivative Works 3.0 United States License.
No comments:
Post a Comment