We don't use this section enough, so here's an essay I wrote as Philosophy of Language practice. It's not being marked or anything so it doesn't matter if I put it up and I haven't really bothered to reference. I hope you enjoy it.



“Description Theory” is the name given to the theory of language put forwards by Bertrand Russell and added to by John Searle. It states that a name is essentially an abbreviation for a number of definite descriptions which the speaker believes to identify a particular object in the world. Saul Kripke has a number of arguments against the Descriptions theory of language, but I intend to show that all of his arguments either misrepresent description theory or are just philosophical sleight of hand.

According to Morris (2007), Kripke summarises description theory as follows. If a name is meaningful for the person speaking it then:
1. there are a number of things which that person believes about the named object.
2. the speaker must believe that some of the things they believe about the object are true only of that one object.
3. if most of the things a person believes to be true about the named object are true of only one particular thing then that thing is what is referred to by the name uttered.
4. if there is not exactly one object which most of the things believed about the named object are true for then the name does not in fact refer.
5. the person speaking knows a priori that if the object named exists most of what the speaker believes about the object is true about the object.
6. it is necessarily true that if the named object exists most of the things the speaker believes about the object are true.
Kripke goes on to attack each of these points individually, except for (1) which it seems any theory of language would find difficult to dispute.

Kripke’s first argument is that if (4) is true then there are lots of words which we think refer but which do not actually refer. Take, for example, Tony Blair, who could be described as “former British prime minister”. Since there is more than one former British prime minister it seems that, by (4), “Tony Blair” does not refer to anyone in this instance, but this seems absurd.

The intuitive response would be to point out that we know more than one thing about Tony Blair. We could narrow the list down with such things as “married to Cherie Blair”. Kripke would say that this is not enough however, because if it were the case that the only thing we knew about Tony Blair was that he was the former prime minister then the argument would still hold.

I think Kripke is wrong in this regard. If literally the only thing you know about the name is this one fact which does not pick out a specific individual then no, the name does not refer. If I choose to call a former (for variety) president Mr X, but the only information I give you about Mr X is that he is a former president of the United States, then you will not be referring to any specific individual if you use the name “Mr X” in conversation. You will have no idea in your mind who you are talking about out of the 44 candidates available to you. Statement (4) holds.

Next Kripke turns his fire on (2). Surely no one believes there has been only one former British prime minister, so if we define Tony Blair as “former British prime Minister,” (2) seems to be false. This argument is, however, susceptible to the same defence as the previous one. If the speaker doesn’t think their definition points to a specific person then they don’t really have a definite description.

In this case the wording of Kripke's definition is slightly different so there is a response available to him – it is possible to come up with a cluster of descriptions which can refer to only one person when taken together, but which individually each pick out a number of people. An example would be “Margret Thatcher is a former British prime minister and a woman”. Both “former British prime minister” and “woman” taken individually pick out literally dozens of objects each, but only one particular object sits in the middle of this Venn diagram.

Against this response I would argue that Kripke has his summary of description theory wrong. You can determine what you’re talking about by a conjunction of several descriptions which ultimately boil down to just one individual. Therefore there need not be any one individual description which picks out the object of reference. Or if you prefer, you could count a conjunction of several descriptions as being a single statement. In Russell’s formal logical formulation of statements one could replace the definite description symbol with a Γ signifying any set of well formed formulas denoting a single object. And if your entire collection of descriptions still fails to pick out just one person then, again, you’re really not referring to any one person in particular, however much you might hope you are.

On point (3), Kripke makes his best argument. He says that you could be wrong about the person to whom you are referring. If I define Bruce Willis as “the actor who played Rambo” then (3) would indicate that I was referring to Sylvester Stallone, but it seems absurd to say that I think Bruce Willis and Sylvester Stallone are the same person. Again, however, Kripke is presupposing some extra knowledge about the subject of my definition, which he denied I had in making the definition in the first place. If literally the only piece of knowledge I have about this “Bruce Willis” fellow is that he played Rambo then yes I am equating him with Stallone. In the same way, if you say that Mr X is the 43rd president of the US you have pinned him down as George W Bush. It is impossible to be wrong when you are defining a name with just one characteristic.

One could argue that Bruce Willis is obviously not Sylvester Stallone because no one who says “Bruce Willis” means “Sylvester Stallone,” but this breaks Kripke’s rule about not defining a word by how the word is defined (Morris 2007, p.78). You could not say, for instance, “Tony Blair is the person who is referred to by ‘Tony Blair’ in this sentence”. Doing so would obviously be circular. You could, on the other hand, talk about how other people tend to use a word for referencing. Kripke does not seem to explicitly allow this, but I think it would be another mistake of his not to. How other people use a word is an empirical statement. In fact, most of the examples of contradictions Kripke claims to come up with rest in some way on the idea that a word refers to a specific object, based on nothing more than common agreement that it does so. When Morris says that we do not think Gödel is the same person as Schmidt he is arguing from our notions of what we generally refer to when we say “Gödel”. If you leave aside the preconceptions then none of Kripke’s arguments so far really have any grounding.



I’m cutting it off here because you need too much terminology going into his arguments against (5) and (6).

Reading it now, I really enjoy that you didn't use the generic uber-complex lingo most people use to write essays like this at the expense of understanding.

That "uber-complex lingo" is used for the sake of understanding. These are very complex and abstract concepts and using specific syntax to avoid misinterpretation is pretty important. If you can't read it it's your fault, not theirs.

Unless it's a legal document. Law school is basically a 4 year program that teaches you how to read that Legalese gibberish.


Anyway, is there a link to Kripke's lectures? I can't find much, I think I *may* have found the first lecture but I can't find the actual arguments you're responding to. If I have, your counter-argument for (4) seems weak, but I'd rather refer to whatever you used.

Kripke is a bitch to look at because he never actually writes stuff down or publishes anything. I've been going on what Morris wrote in An Introduction To The Philosophy of Language, but I can't find a free online way to look at that. 3.2 and 3.3 here contain most of the things I've argued against in one form or another.

Nevermind, I found a good source. Well the only thing I can think to point out is that a lot of your examples involve knowing only one specific thing about the person you're referring to, when that violates (1). Don't you have to know more than one thing about that person? If you don't it's a circular reference and inherently meaningless, isn't it? I could be wrong.

Also,

Quote from Zed

If I define Bruce Willis as “the actor who played Rambo” then (3) would indicate that I was referring to Sylvester Stallone, but it seems absurd to say that I think Bruce Willis and Sylvester Stallone are the same person.

I don't think that's what he meant to say with his rebuttal to (3). If I incorrectly believed Columbus was the first person to discover America, and that he discovered the earth was round, does that mean whenever I say "Columbus" I'm actually referring to whoever did those things? Of course not, because these names are not rigid designations determined solely by what characteristics are true. They're referential according to what the speaker is trying to say about his personal conceptualization of the object being referenced.

"Kripke concludes that thesis (2) is false, since we don’t use names in ways such that the corresponding descriptions lead to uniqueness or non-circularity. We can’t prop up uniqueness by reference to encyclopaedias where theories or famous deeds are listed"

That was his rebuttal for (2), so when you say "Therefore there need not be any one individual description which picks out the object of reference" as a defense for it, it seems like you two are in agreement more than anything.

But, again, I could be wrong.

A lot of Kripke's arguments rely on you only knowing one thing about the object. I think the circularity comes into it if you're counting knowing the name of something as knowledge. I don't think there's any problem with only knowing one thing about something under description theory. The only thing I know about King Harald is that he's king of Norway. It seems ok.

Quote from Exilement

I don't think that's what he meant to say with his rebuttal to (3). If I incorrectly believed Columbus was the first person to discover America, and that he discovered the earth was round, does that mean whenever I say "Columbus" I'm actually referring to whoever did those things? Of course not, because these names are not rigid designations determined solely by what characteristics are true. They're referential according to what the speaker is trying to say about his personal conceptualization of the object being referenced.

I can't work out whether I agree with you or not. I think what you've written here is what Kripke says about incorrect descriptions, but what I wrote was what Kripke says description theory says about incorrect descriptions, which he then argues is wrong. Or am I misunderstanding you?

"Kripke concludes that thesis (2) is false, since we don’t use names in ways such that the corresponding descriptions lead to uniqueness or non-circularity. We can’t prop up uniqueness by reference to encyclopaedias where theories or famous deeds are listed"

That was his rebuttal for (2), so when you say "Therefore there need not be any one individual description which picks out the object of reference" as a defense for it, it seems like you two are in agreement more than anything.

But, again, I could be wrong.

Kripke is saying that the individual descriptions don't pick out something uniquely, and I agree with him there. Where we disagree is on whether or not it's necessary for there to be a description which picks out something uniquely. I think it's allowable under description theory to have five or six descriptions which, taken together, only pick out one object. Kripke doesn't seem to allow this. I think his reasoning is that even if you can give examples of things you know lots about, there will be some things/people that you only know one thing about and that one thing won't be enough to pin down what you're referring to. My argument against that is that in situations like that you really are not referring to a specific thing.

Oh, alright. Honestly I'm not sure where he argues that last point about not allowing individual, non-unique descriptions, but if that's his argument then I can't really agree with that.

I guess the wording of (2), "the speaker must believe that some of the things they believe about the object are true only of that one object" is a little unclear. Does that mean the speaker must believe certain specific descriptors must be true of only that one object? Or that the sum of those descriptions fits only one object? I guess the wording might not allow the latter, but it seems entirely valid either way.

It is a bit ambiguous. I've found an essay in one of my textbooks which is meant to be by Kripke, but I haven't read it yet. If it clarifies his point I'll let you know.