|
|
|
|
Primary: Russell (1867) ‘On Denoting’ Searle, J. (1958) ‘Proper Names’ Kripke S. (1972) ‘Naming and Necessity’.
Secondary: Lycan, W. G. (2000) Philosophy of Language, London; Routledge.
|
|
|
Russell's Theory of Names
|
|
|
I’ll start by introducing Russell’s theory about narmes so that we know where the discussion that follows is going to end up. It’s pretty easy to state.
Names are just disguised definite descriptions.
|
|
|
Justifying Russell's Theory of Names
|
|
|
Russell’s theory that we looked at earlier is limited to definite descriptions. But they are not the only singular terms: there are also names. In fact names – like Pegasus, Hesperus, Sherlock Holmes, etc. – were the most common singular terms that were referred to in Mill’s and Frege’s (and Meinong’s) theories. Russell’s solution to the problems of reference could not be taken as a general solution if it could not also be applied to names, so the general problem still exists. Without denying the theory of definite descriptions, there are two obvious solutions to this: first, one could come up with an independent solution just for names – in which case the general solution would be the combination of these two (inelegant, klugey); second, the Russellian solution could be expanded to include names. Russell naturally tried for the second
Recall that, for Russell, whereas the merely apparent grammatical form of a definite description was something like
The present King of France
which looked like an independently comprehensible and referring expression, in fact its (real) logical form was only determinable in a particular context such as a sentence like
The present King of France is bald
and a sentence like that had an overall logical form that we could paraphrase as
There is a thing, which is a present King of France, and for any other thing, if the latter thing is a present King of France then it is the same thing as the former thing, and that thing is bald.
no wonder that normal language needed an abbreviation.
In a similar way, Russell could claim that the logical form of names was quite different from their apparent grammatical form. Their grammatical form showed no logical structure at all – and thus we are fooled into thinking that their logical behaviour can only be governed by their direct referential role as a label for an object. Their logical form, however, was just the same as that of definite descriptions, so that names are best described just as disguised definite descriptions.
This means that ‘Pegasus’, for example, is logically identical to a description such as, say, ‘the mythological winged horse’, and the role that the name ‘Pegasus’ plays in determining the logical form of a sentence in which it occurs (in transparent contexts) is exactly the same as the role that the definite description ‘the mythological winged horse’ would play, according to Russell’s theory, if the sentence had that denoting phrase instead of the name. But let us be even more concrete: the logical form of
Pegasus lives on Mount Helicon
is identical to the logical form of
The mythological winged horse lives on Mount Helicon
which is to be read as
There is a thing, which is a mythological winged horse, and for any other thing, if the latter thing is a mythological winged horse then it is the same thing as the former thing, and that thing lives on Mount Helicon.
and this will doubtless be a surprise to anyone who has ever had cause to say that Pegasus lives on Mt Helicon. But do not be led by your surprise, and your unawareness of intending any such thing, to reject the analysis on those grounds. The theory of meaning that we are trying to get it is required by us just because our naïve or folk theory of meaning (if there is one at all) is unable to answer the questions that we need answered. If we are really looking for a theory of meaning that is of the same kind as theories about planets or chemistry or geography, then we should be prepared for it to be quite unlike our folk theories or intuitions. After all, our folk theories of astronomy, medicine, etc. are things that we laugh at now.
Anyway, it’s obvious that by identifying names and their logically equivalent definite descriptions Russell can solve the puzzles of reference that faced the Millian View without resorting to Fregean senses. Let’s just review that briefly.
1. There’s a puzzle about identity statements, because the sentences
i. Venus is Venus ii. Hesperus is Venus
appear to have different cognitive contents. It’s a different thing to know the first from knowing the second; and there shouldn’t be a distinction if the meaning is the same, and the meaning will be the same if all that the words Venus and Hesperus contribute to the meaning of a sentence is their reference.
Frege’s solution was to propose that the different senses of the names explained their different contributions to the sense/meaning of the sentence. Russell says that the logical form of these sentences is identical to the logical form of the sentences
iii. The evening star of the Romans is the evening star of the Romans. iv. The evening star of the Romans is the evening star of the Greeks.
And the logical forms are obviously quite different.
2. There’s a problem with non-referring terms. Our statement to the effect that Pegasus lives on Mt Helicon seems meaningful, but if the MV is right there’s no meaning because there’s no reference. Frege said that the meaning could be attributed to the sense. Russell says that the meaning can be equated to the meaning of a statement involving a definite description, and we’ve seen how that works out.
3. There’s a problem with negative existential statements. If Sherlock Holmes doesn’t exist then the statement that ‘Sherlock Holmes does not exist’ should be true. But on the MV it can be true only if it is meaningful, and it can’t be meaningful because it refers to a nonexistent. So if it’s true then its meaningless and not true. It can only be false. Frege says it’s meaningful but it can’t be true or false. Russell says that negative existential statements involving definite descriptions should be interpreted in terms of the scope ambiguity of the negation operator in the logical form of those statements. The same is therefore true of negative existential statements involving names.
To say that Pegasus does not exist is to say that the mythological winged horse does not exist; and this statement has the logical form
Ø ($x) [MWHx Ù ("y) [MWHy ® y = x] Ù Ex]
which you should read as
It is not the case that there is a thing, x, such that x is a mythological winged horse, and for any thing, y, if y is a mythological winged horse then y is the same thing as x, and x does exist.
4. There’s a problem with substitution into opaque contexts. Suppose we have the two sentences
i. John believes Cicero was an orator. ii. Cicero is Tully.
It is pretty clear that we can’t conclude that
iii. John believes Tully was an orator.
Frege talks about the reference in such contexts being the customary sense. Russell again talks about the scope ambiguity of natural language which is clarified in the logical forms that he gives. The same trick will of course work for names when they have been replaced in statements in which they occur by instances of the definite descriptions which they abbreviate.
Of course, the claim that names are definite descriptions, though it does solve these problems (if Russell’s theory of descriptions solves those problems), may be seen as motivated by more than just this. There is obviously some sort of connection between names and definite descriptions. Consider: whenever we are uncertain about a name that someone has used, like Pegasus for example, we ask for clarification, and the only clarification that we are likely to get is a statement such as ‘Oh, that’s the mythological winged horse who lived on Mount Helicon.’ Since this is all that we know about Pegasus (and how many here know anything more than that) it’s reasonable to suppose that whenever we use the name the cognitive content is indistinguishable from the cognitive content of that description. You’ll recall also that in ‘Sense and Nominatum’ when Frege was describing the sorts of things that sense could be for names he was inclined to give descriptions that fitted the objects that had those names. Now, it was no part of his theory that senses were descriptions, and he was explicit in providing examples elsewhere that showed the difference between senses and descriptions, but the fact remains that descriptions are what we naturally appeal to when we want to know what a name intends.
|
|
| Problems with Russell’s Theory of Names
|
|
|
So much in favour of Russell’s theory. There are, however, some serious difficulties with his view.
a. The Indeterminacy of Descriptions
Searle worried about the apparent claim that in any use of a name there is a single, unique definite description that the name is supposed to stand in for. There are a couple of ways in which this assumption offends our intuitions.
i. In the first place it will have the consequence that any statement involving a name as a denoting term is going to give rise to entailments which we will find quite unwarranted. For example, let us suppose that the denoting phrase ‘Sir Walter Scott’ is an abbreviation of the description ‘the author of Waverley and the author of Ivanhoe and the author of the Lay of the Last Minstrel.’ It would therefore be proper to say that
Sir Walter Scott was happy
is logically equivalent to the statement
The author of Waverley and Ivanhoe and Lay of the Last Minstrel was happy
and this is paraphrasable in the normal way to
There is a thing that is an author of Waverley and etc.
and so the statement with which we began must logically entail the statement
There is a thing that is an author of Waverley
which surely can’t be what was intended.
ii. Secondly, even if one accepts that in any particular usage a name abbreviates a definite description, there is no good reason to believe that one always has the same description in mind when one is using a name. There may be times when, as I speak of Sir Walter Scott I am thinking of him as the author of Waverley because what I am discussing is the origin of the historical romance as a major genre. At other times I think of him as the man who wrote himself out of debt, because I am thinking of the force of duty as a determinant of a virtuous character. Again, I may think of him as ‘the inventor of the so-called myth of the tartan’ because I’m talking about the ways in which nationalist justificatory myths are discovered. In all these cases, if you had asked me: who is this Scott of whom you speak so highly? I would have given a different description as an answer – and apart from the Russellian theory there seems to be no reason to believe that the conceptual content of the name that I used went essentially beyond the description that I give in response to your question. We may even suspect that the name can act as a placeholder for all sorts of conceptual content, and that at the time that I’m using the name there is no fact of the matter about which description I’m abbreviating. (Just because I know things about the bearer of a name doesn’t mean that any of that knowledge is logically involved in my use of that name.)
b. The Diversity of Descriptions
Furthermore, even if in my usage a name abbreviates a particular description, it does not follow from anything that Russell has said that in the usage of all other speakers the name abbreviates the same description. When I speak of Descartes I may be abbreviating the description ‘the X which is A, B, …, C, and died of pneumonia’, whereas when you speak of Descartes you are abbreviating the description ‘the X which is A, B, …, C, and was a soldier in the army of Prince Maurice of Orange’. The meaning of ‘Descartes’ for me is different from its meaning for you, and sentences using that name are not logically equivalent in my usage to the same sentences in your usage. Does this seem right?
Moreover, it may even happen that what I abbreviate by a name has no overlap at all with what you abbreviate by a name. Suppose, then, that we are two ignorant persons. If you say ‘Moses wore sandals’, and by ‘Moses’ you abbreviate just ‘the fellow who received the tablets of the Law’, and I try to contradict you by saying ‘No, Moses did not wear sandals’, but by ‘Moses’ I abbreviate just ‘the fellow who led the Exodus from Egypt’, then it doesn’t seem that my contradiction succeeds because the logical form of my statement is not the negation of the logical form of your statement. Compare:
There is a thing, x, such that x received the law tablets, and for any thing, y, if y received the law tablets then y is the same thing as x, and x wears sandals.
It is not the case that there is a thing, x, such that x is a leader of the Exodus, and for any thing, y, if y is a leader of the Exodus then y is the same thing as x, and x wears sandals.
|
|
|
Searle's Theory of Names
|
|
|
|
In response to these problems Searle proposed that Russell’s theory could be solved by a slight modification; so that instead of a name being an abbreviation of a single definite description it is associated with a cluster of descriptions. Searle’s theory is for this reason known as the ‘cluster theory’. According to this theory each name has a set of descriptions associated with it, where these descriptions are to be thought of as ‘standard identifying statements’. The role of the name in a statement is to indicate that the object in question satisfies some appropriate but indeterminate number of the associated standard identifying statements. This is an idea with some little history, for Wittgenstein, in his Philosophical Investigations (sec. 79) must have had something like it in mind when he said:
If one says ‘Moses did not exist’, this may mean various things. It may mean: the Israelites did not have a single leader when they withdrew from Egypt – or: their leader was not called Moses – or: there cannot have been anyone who accomplished all that the Bible relates of Moses – … But when I make a statement about Moses, - am I always ready to substitute some one of those descriptions for ‘Moses’? I shall perhaps say: by ‘Moses’ I understand the man who did what the Bible relates of Moses, or at any rate, a good deal of it. But how much? Have I decided how much must be proved false for me to give up my proposition as false? Has the name ‘Moses’ got a fixed and unequivocal use for me in all possible cases? – Is it not the case that I have, so to speak, a whole series of props in readiness, and am ready to lean on one if another should be taken from under me and vice versa?
The fact that the number of descriptions in such a cluster is indeterminate is important for Searle since it gives a reason for the our using names rather than descriptions which is a little more convincing than our desire to save time. For Searle the role of names is to provide a stable referring device with which any number of descriptions (identifying statements) can then be associated. It allows us to actually get started on the business of talking about the object being referred to.
How does this help with the problems that we’ve looked at so far?
i. Because the name is associated with a cluster, no member of which need be necessary, it does not follow that any use of the name logically entails the existence of an object that satisfies some specific member of the cluster of standard identifying statements. To be more precise: because ‘Sir Walter Scott’ is associated with a cluster of concepts which includes ‘the author of Waverley’ and ‘the author of Ivanhoe’ and ‘the author of the Lay of the Last Minstrel’ and presumably very many others, and because the use of the name merely indicates that some object that satisfies enough of these descriptions, but not which of those descriptions exists, it cannot be concluded that there is an object that satisfies any of those particular descriptions.
ii. For the same reason, there is no difficulty in assuming that one may have different descriptions in mind when one is using a name. Since there is a cluster of concepts involved we can suppose that the particular selection of descriptions which are being considered is different in each usage.
iii. The same reasoning explains why it doesn’t matter much that different speakers may have different descriptions in mind when they are using the same name
|
|
Further Problems
|
|
|
|
Unfortunately, the problems which Searle thought that his proposal could solve turn out not to be all the problems for the description theory of names.
a. Modal Statements
Consider, for example, an argument that was put forward by Kripke, in Naming and Necessity, that the description theory in any form can not properly account for our intuitions about modal statements – by which I mean statements involving possibility and necessity. One of the examples that Kripke uses is the claim that ‘Nixon could have lost the 1968 election’. If all you know about Nixon is that he is the man who won the 1968 election, or, equivalently, if the description that is associated with the name Nixon is just ‘the man who won 1968 election’, then Kripke claims that the claim that was made about Nixon is logically equivalent to
It is possible that there is a thing, which is a winner of the 1968 election, and for any other thing, if the latter thing is a winner of the 1968 election then it is the same thing as the former thing, and that thing did not win the 1968 election.
But this can’t be true, because it involves an obvious contradiction, and thus the description theory can’t be true. This may seem to apply only to the unrevised theory because the cluster theory may have a large range of different descriptions that it can appeal to as characterizing Nixon when the election result is actually in dispute, but this is not the case, because for any description that is supposed to be a member of the cluster of descriptions associated with Nixon we can imagine the situation – or talk about the situation anyway – in which the Nixon-labelled thing does not satisfy the description. And the same is true for all subsets of descriptions from that cluster.
A better response to Kripke is to propose, quite plausibly, that there is an ambiguity in the scope of the possibility operator in the logical form of the statement. This is exactly the same sort of scope ambiguity that Russell appealed to in his interpretation of negative existentials – although his terminology for the phenomenon was that of primary and secondary occurrences. The interpretation that Kripke has preferred is that the possibility operator has wide scope, and applies to the entire statement, for which reason we sometimes refer to this sort of modality as de dicto, meaning ‘about the statement’. But the contentious claim is also interpretable as involving the possibility operator with merely narrow scope, so that the possibility applies only to the object about which the statement is being made. That sort of interpretation is referred to as having modality de re, meaning ‘about the thing’.
It’s a little bit easier to see where the ambiguity comes in if we use the more formal way of writing. In this style the wide, de dicto, Kripkean interpretation turns out like
à [ ($x) [Wx Ù ("y) [Wy ® y = x] Ù ØWx] ]
and the narrow, de re, Dummettian interpretation looks like
($x) [Wx Ù ("y) [Wy ® y = x] Ù à [ØWx] ]
and that could be read as
There is a thing, which is a winner of the 1968 election, and for any other thing, if the latter thing is a winner of the 1968 election then it is the same thing as the former thing, and that thing could have lost the 1968 election.
which really seems quite unproblematic.
b. False Descriptions
Kripke has another problem for description theories of all stripes, and that is that some times descriptions are OK for referring even when they are actually false about the object that they’re supposed to describe. If the descriptions get their referring power through the accuracy of the descriptions – which is certainly the assumption that has been generally made in description theories – then it is a mystery how there can be successful reference in such cases.
Kripke gives the example of Gödel, who is known to us as the man who proved the Incompleteness Theorem. But what if Gödel wasn’t the author of that proof, but simply stole it from the unfortunate Schmidt? In that case, all that we know about Gödel, and all that lies behind our use of the name ‘Gödel’ is actually false of him but true of Schmidt. And yet ‘Gödel’ refers to Gödel and not to Schmidt. How is it so?
|
|
Kripke's Theory of Rigid Designators
|
|
|
|
I’ve given a few general objections that Kripke has made wrt descriptivism about names, but Kripke also has a much more general positive doctrine about the nature of reference for names, which we’ll now look at. Kripke’s theory of reference actually comes in two parts. There’s a theory about the nature of reference and there’s a theory – on independent grounds – about the way reference actually works. I’ll start by introducing Kripke’s theory about the nature of reference. In fact I’ll start by saying just what his theory has to say about the reference of names and then I’ll explain how he came to hold that idea.
He thinks that a name normally refers to the same thing in all possible worlds where that thing exists.
|
|
Kripke's Modal Argument
|
|
|
|
Possible Worlds
Kripke has an argument against the descriptivist view of reference known as the ‘modal argument’, because it appeals to our intuitions about modal statements – statements which involve attributions of necessity or possibility, which are called modal properties. Now, a way to think about modal statements is to imagine that they are talking about situations in a range of possible worlds. Exactly what we are to take possible worlds to be is a bit of a mystery, but for our purposes it’s enough to think of them just in some intuitive sense as corresponding to ways that the world could be; and, in fact, corresponding to some description of how the world could be. Thus when we make a statement that
Bacon could be the author of Hamlet
we interpret this as meaning the same thing as
There is a possible world in which Bacon is the author of Hamlet
and the possible world in which this is the case is one in which it is truly stated that
Bacon wrote Hamlet.
This statement is true in that world but it is false in the actual world. The point is just that we accept that some statements can be true in some worlds and false in other worlds. Just so, the things to which a definite description provides reference can change from possible world to possible world, because the things which are their truth-makers change. The author of the best-selling book in Australia today might be Bob, although it is actually Fred. The highest tide last year might have been the 2m tide, but it was actually the 1.5m one, the tastiest cake might have been the one made by Melissa, yet in fact it was the one made by Lisa. And so on.
Rigid and Non-Rigid Designators
The fact that definite descriptions point at different things in different possible worlds is a fact that Kripke describes as their having non-rigid reference (floppy reference). It contrasts with what Kripke believes is the typical characteristic of proper names, that they have rigid reference. Proper names point at the same thing in any possible world. When Kripke makes this claim about most proper names he’s really appealing to an intuition we have that when we are talking about the possibility that a thing could have properties other than the properties that it does have, that we are actually talking about that very thing having different properties.
We have to notice a couple of qualifications here.
Contrariwise, there are times when things that look like descriptions really aren’t non-rigid, but they are used as names. Jack the Ripper is the classic example. The tag looks like description but really refers to just one person.
A Rule of Thumb
That second qualification may make us wonder whether there is really a principled distinction to be made here. If some apparent names can be non-rigid (and must be classed with descriptions) and some apparent descriptions can be rigid (and should be classed with names), then is there a way – other than ideological convenience – by which we can judge in which category one these things should be placed? Kripke says there is – or rather he has a rule of thumb that he uses to determine whether something is being used as a rigid or a non-rigid designator. This rule of thumb is to put the phrase X into the statement frame
X might not have been X.
How does this work in general? Well, for the description of Aristotle it might tell us that
The teacher of Alexander might not have been the teacher of Alexander
and that seems to be reasonable enough: we can certainly imagine the situation where the thing that is the teacher of Alexander in our actual world is not the teacher of Alexander in some other world. On the other hand for the name of Aristotle it would claim that
Aristotle might not have been Aristotle
and this seems much less likely. In fact we think it is absurd. Note that we don’t mean by such a sentence that Aristotle might have been called something other than Aristotle. That we think is true but beside the point. The claim is that the very person that we know as Aristotle might have been some other person – or perhaps might not have been any person at all.
Rigidity and Necessity
Looking at this from another direction Kripke can be seen to be using the phenomena noted in justifying his rule of thumb to justify his claim that names must be rigid designators. He appeals to our intuitions about necessity and contingency to demonstrate that the descriptivist view cannot be the correct way to interpret the ‘meaning’ of names.
Some statements we take to be necessarily true, by which we just mean that they cannot be false. In terms of possible worlds we think that such statements are true in all possible worlds (or they are false in none.) Statements which are necessarily true are statements like ‘2 + 2 = 4’ or ‘no bachelors are married’. Most significantly for our purposes, any statement like ‘X is X’ is necessarily true. So, in particular,
(a) the teacher of Alexander is the teacher of Alexander
is necessarily true. By contrast, those statements which are true, but not necessarily true are said to be contingently true. They might have been false. They are true in possible worlds other than the actual world. Such statements are much more common. They include statements like
(b) Aristotle is the teacher of Alexander
Aristotle, we feel, might very well have rejected that job offer and yet still have been Aristotle. Accordingly it seems that using one designation of the object a statement about it turns out to be a necessary truth, and using another designation it turns out to be a merely contingent truth. This can’t be right. Therefore it seems to be a necessary conclusion that whereas the designator ‘teacher of Alexander’ does not refer to the same person in every possible world, the designator ‘Aristotle’ does do so. Proper names like ‘Aristotle’ we conclude are therefore the sort of thing that we use to fix the identity of an object across possible worlds.
However, it’s worth making a note here that when we talk about ‘fixing the identity’ of Aristotle, I don’t mean – and Kripke wouldn’t mean – that we could view the name as an abbreviation for just the essential properties of Aristotle (the properties that belong to Aristotle in all the possible worlds where he exists). We have no idea of what those properties might be, and we certainly don’t use them to identify objects. Just imagine if you had to know the essential properties of me before you could refer to me. How that would cripple your conversations! Kripke is quite clear that the possible worlds in question are not the sorts of things that one can see through a telescope, the identification of Aristotle in such a world is not a matter of matching some criteria to a discovered object , it is a matter of stipulation. When we say that Aristotle might not be the teacher of Alexander we are stipulating that there is a world in which Aristotle exists and in which he is not the teacher of Alexander.
|
|
Same Old Problems with Direct Reference
|
|
|
|
According to Kripke then, it can’t be that the meaning of names is given by a definite description or a cluster of the same. In fact it rather looks like the normal function of proper names is purely referential. But this position, if we were to adopt it, would just bring us back to the position of Mill, and would open Kripke up to all the same objections that were made against Mill’s system. Perhaps we could retreat to Frege’s system, but if we were to do so we would need to have something to say about exactly what sense was supposed to be.
|