Frege on Sense and Reference PDF

Preview

Summary

Lecturer Dr. Craig Bourne...

Description

Philosophy of Language Lecture 7 Frege on Sense and reference Summary of seminar 6: Donnellan on Attributive vs. Referential uses of definite descriptions Donnellan: that man’s insane Donnellan distinguishes two uses of descriptions: attributive and referential. For example, you hear that Smith has been murdered and that his body has also been horribly mutilated. You say: ‘The murderer of Smith is insane’. This translates as: the murderer of Smith, whoever it is, is insane. That is, you don’t know who is the murderer, but you still want to assert that whoever it is is insane. This is the attributive use. On the other hand, you see Jones get arrested, and you think he has been arrested for the murder of Smith. Jones is behaving wildly. You say: ‘The murderer of Smith is insane’ In this case, you refer to Jones, and your statement is true iff Jones is insane. This is the referential use.

Donnellan’s Treatment According to Donnellan, under certain circumstances a speaker may use a definite description to refer to something which doesn’t satisfy the description, but which he wants to single out. So assume as a matter of fact that Jones did not murder Smith. Still, on the referential reading, according to Donnellan, ‘The murderer of Smith is insane’ is true iff Jones is insane, i.e. even though Jones is not the murderer of Smith. Something is right in this. But something is wrong too. Don’t we want to say that strictly speaking, ‘the murderer of Smith is insane’ is false if the murderer of Smith is not insane (and also false if there is no murderer, or more than one murder), regardless of whether Jones is insane? (Russell’s view But don’t we also want to say that the speaker has been successful in using that description to refer to that individual? Could we concede this without having to agree with Donnellan that the sentence is true? Perhaps distinguish pragmatic (successful/unsuccessful) from semantic (truth/falsity) evaluation of sentence.

Mill’s theory of names Names (‘Socrates’ ‘London’) look to be paradigm cases of genuinely referring terms. According to Mill, the meaning of a name is nothing more than its bearer, or the object it denotes, or to which it refers. Names, unlike descriptions, do not have connotations.

Problems for the Millian account of names The problems we raised for (a certain account of) definite descriptions arise equally with this account: Names with no bearers: e.g. ‘Darth Vader’, ‘Father Christmas’. If the meaning of a name just is the bearer, then these names would be meaningless. But they’re not! So Mill’s account is lacking.  How can negative existential statements be meaningful let alone true? e.g. ‘Darth Vader does not exist’.  Frege’s puzzle: e.g. Harry Webb = Sir Cliff Richard  Substitutivity: Emily wishes that Cliff Richard would call round for tea.

Frege’s Proposal: sense and reference There must be more to the meaning of a name than its bearer Cliff Richard is Cliff Richard Cliff Richard is Harry Webb The first is trivial; the second isn’t. Frege concludes: the meaning of a name must be more than its reference: names not only have a reference, but also a sense.

Summary: Sense and reference sense Sense determines reference: if names agree in sense, they agree in reference. Sense is the ‘mode of presentation of the thing designated’: in grasping the sense, we think of the reference in a certain way (often using definite descriptions – so in this sense Frege is a descriptions theorist about names.) Sense is something objective for Frege. Communication requires an element that is constant across speakers. For Frege, this is the sense. Reference This is what determines the truth-value of statements. If there is no reference, for Frege, there is no truth-value – so there is similarity here with Strawson on DDs. (Russell says they’re plain false.) Frege’s theory avoids the problems of Mill’s theory. Truth and falsity is settled by reference, but cognitive value by sense. (Does it solve bearerless names problem?)

Frege’s wider project What does it take to UNDERSTAND various symbols of a language? Frege applies his sensereference distinction to all parts of language (see ‘On Sense and Reference’): Linguistic names predicates sentences

Sense Referent sense of a name object sense of a predicate concept thought (not idea) truth value

Concepts for Frege are functions. e.g. The concept of having a fat tongue is a function that takes objects, e.g. Jamie Oliver, Dasiy the cow, Simon the snake, as arguments. What does it deliver as the value of the function What the sentence ‘Jamie Oliver has a fat tongue’ refers to is the truth-value: True. Plugging in Simon the snake to the function would deliver the value = false. [All true sentences, then, refer to the same thing! It is their sense that distinguishes them.] The sense is the mode of presentation of the referent. e.g. there are lots of ways of presenting the True (it can be presented as grass is green or snow is white). It is something you grasp to direct your understanding to the right referent.

Extensional contexts Extensional contexts: you can swap co-referring terms without changing the truth-value of the original, e.g. At the level of sentence Grass is green and snow is white Grass is green and ravens are black At the level of predicates Triangles have three sides Triangles have three angles At the level of names Hesperus is bright Phosphorus is bright

The Trickier Case of intensional contexts Examples of so-called ‘intensional contexts’ (here, propositional attitudes): A believes that p John believes that: George Orwell wrote 1984 A hopes that p John does not believe that: Eric Blair wrote 1984 A fears that p Here, we cannot just swap co-referring terms in p and expect it to retain the same truthvalue. e.g. I might believe that Hesperus is bright without believing that Phosphorus is bright. Swapping 'p' in ‘A believes p’ for any True sentence 'q' does not guarantee that 'A believes q‘ is true.

So Frege concludes that the sentences after the ‘that’ clause do NOT have a truth-value as a referent, as they do in the standard extensional case. Frege says they have 'indirect referents’: senses are their referents. Why? Because if we think we should be able to swap co-referring terms in all contexts without changing truth-value, then swapping 'p' for another sentence 'r' that has the same sense as 'p' would keep 'A believes r‘ true. So the referent in these cases must be the sense. QED

Russell’s theory of names Russell distinguishes: logically proper names: genuinely referring expressions, such as ‘I’, ‘now’, ‘this’. These name objects with which we are immediately acquainted, namely the self, our location, something in our immediate vicinity. We do not associate any descriptions with these expressions (in Frege’s terminology, they would have no sense). ordinary proper names: those things having the grammatical form of proper names, such as ‘Johann Sebastian Bach’, ‘Bertrand Russell’. Russell thinks that although proper names, such as ‘Socrates’, appear to be paradigm cases of genuinely referring expressions, they are not. Russell thinks that proper names are disguised descriptions. Different people associate different descriptions with a given name, as does the same person at different times. For Russell, such differences in associated descriptions explain the problem about the differences in cognitive value in Frege’s puzzle, and related difficulties. So that’s why he treated names as definite descriptions. Russell then treats those associated definite descriptions as quantifier phrases to avoid the problem of bearerless DDs, etc. So, Russell would rephrase the name ‘Aristotle’ as: The teacher of Alexander the Great and pupil of Plato and… And then he goes on to give a quantificational treatment of that definite description. We’ll tackle Kripke’s objections to the description theory of names in the next lecture ...

Difference between Russell and Frege on names Russell’s solution is in some ways very much like Frege’s. But Russell cuts out the mysterious middle-man sense. Associate definite descriptions with a name. Then solve all of the problems that Mill encounters in the way DDs solve them. Advantage over Frege is that it avoids difficulty reconciling theory with bearerless names. Remember: according to Russell, DDs don’t refer; they quantify: ‘Father Christmas is fat’ says something like: ‘There is at least one object who rides in a sleigh, etc., and he is fat’. This is clearly meaningful but false since there is no such object. We don’t need to think that ‘Father Christmas’ refers to the object Father Christmas in order to be meaningful....