This is a less technical version of a paper I wrote last year. It deals with the issue of meaning. That is, how a word gets its meaning and what this meaning represents. (In other words – what does “meaning” mean?).
Meaning is a huge subject in and of itself so this post only deals with one specific theory of meaning called “descriptivism” and explains why descriptivist theories of meaning are ultimately incorrect.
So what is “descriptivism”?
Descriptivists believe two things:
- The reference of a word (the things that word is used to pick out, for example the reference of the word “cat” is a real life cat) is determined entirely by a description of the word. This means our ability to use the word “cat” to pick out real life cats is solely dependent on our description of the world “cat”. (If something seems fishy about this – pun intended – you’re right, this position is highly problematic)
- The meaning of a word is associated with a description associated with the word. So the meaning of “cat” has something to do with our description of what a “cat” is (a sometimes living thing that most often has four-legs and a tail, and is a member of the feline family, etc.)
The important point is according to descriptivism, the purpose of description is to aid us in picking out specific referents (think: real life examples of the word)
In Reference and Definite Descriptions, Donnellan argues definite descriptions actually have two uses: attributive and referential. A definite description used in the attributive sense says something about “who or what-ever is the so and so.” It is essential in that there is no other way to refer to the referent except by the definite description.
To use one of Donnellan’s examples – if I come home to find my poor mother murdered, but don’t know who killed her, the only way I can refer to the person who committed this heinous act is by using the definite description “my mother’s murderer” or something to that effect. Further, if my mother wasn’t murdered but had somehow fallen and accidentally impaled herself with a kitchen knife, the definite description “my mother’s murderer” 1) wouldn’t apply to anyone, 2) any sentence in which I used that phrase would have no truth value. Again, according to Donnellan this is an example of a definite description used in the attributive sense.
Contrast that situation with the instance in which my mother’s murder is found and placed on trial. During recess I mention to a friend that my mother’s murderer smells awful. Here the definite description “my mother’s murderer” is not exclusive because I could use any number of definite descriptions (or proper names even) to refer to the person in the court room: John Doe, the man in handcuffs, etc. In this case “my mother’s murderer” is being used referentially because it refers to a specific person.
Further, even if the description is entirely wrong (imagine the man in court didn’t, in fact kill my mother) the sentences in which I used “my mother’s murderer” 1) make sense, 2) refer to the same subject, and 3) can be true (aside from the description itself). We can take this even further by imagining a case in which the definite description is entirely wrong, both parties are aware of this fact, but it still functions perfectly well to identify the subject the speaker is referring to. In fact, these kinds of cases occur quite often. Consider “the guy who discovered America” for Columbus or “the country with weapons of mass destruction” for Iraq.
Donnellan’s point is definite descriptions can be used to correctly identify a person or thing when the description(s) is accurate but also when the description(s) is inaccurate or even patently false.
Now, this may not directly contradict descriptivism (in fact it does not). However, it does create friction and provide cause for thinking descriptivism is at least incomplete and possibly incorrect.
Again, a descriptivist could believe something like (1) we identify things by using proper names (2) proper names are simply short-hand for a group of definite descriptions (3) some group of definite descriptions uniquely pick out every person/thing.
The problem is, with this view there does not appear to be an account for how definite descriptions can function perfectly well without providing any useful information about the referent what-so-ever, which is precisely what seems to be happening when definite descriptions are used referentially. If descriptivism is to be accepted at face value it appears that a definite description is useless if it is not accurate, and with the referential use of definite descriptions Donnellan show that this is not the case.
Thus, if Donnellan succeeds in proving that definite descriptions are and can be used referentially (which I believe he does) then the descriptivist position is at best incomplete and at worst incorrect.