Philosophy of Language: Questions and Answers

Steve Hoenisch

Given the profoundly nuanced thought with which the philosophers below address the complexities inherent in language, these short answers may not be complete, or completely accurate. I suggest you address these questions with the full weight of your own critical thought.

These questions were posed by Prof. Jerry Katz in his course on Philosophy of Language at the Graduate Center of the City University of New York.

1 Frege

1.1 Frege, Kant, and Carnap Define Analyticity

Q: (i) State Kant’s definition(s) of analyticity. (ii) State Frege’s definition of analyticity. (iii) State Carnap’s definition of analyticity.

A: (i) For Kant, a proposition is analytic when the concept of the predicate is contained in the concept of the subject. Likewise, Kant conceived of an analytic statement as one that attributes to its subject no more than is already conceptually contained in the subject. In other words, analyticity is when a statement is true by virtue of meanings and independently of fact.

(ii) Frege’s definition of analyticity is that a proposition is analytic just in case it follows from the laws of logic plus definitions but without use of principles from special sciences.

(iii) Carnap holds that a statement is analytic if it is true by virtue of the intensions of the expressions occurring in it. More specifically, Carnap’s definition of analyticity is that a statement is analytic when it comes out true under every state description, which is any exhaustive assignment of truth values to the atomic, or noncompounded, statements of the language.

1.2 Frege’s Arguments for Sense

Q: State two of Frege’s argument for thinking that there is sense as well as reference.

A: One of Frege’s arguments for thinking that words have sense as well as reference is contained in the examples he gives for referring to Venus. “The morning star” and “the evening star” both refer to the same thing, Venus. They do not, however, mean the same thing; that is, they do not have the same sense. If they did have the same sense, the statement “the morning star is the morning star” (m=m) and “the morning star is the evening star” (m=e) would have the same meaning. But (m=m) is an uninformative, analytic fact about the world, while (m=e) contains an unobvious, empirical truth about the world. Therefore, Frege says, two expressions having the same referent do not necessarily have the same sense.

A second argument that Frege uses to establish that there is sense is well as reference appeals to the notion of indirect speech – e.g., when one speaks about words themselves or their sense or quotes what someone else said. In other words, Frege says, a mention-use distinction underlies speech. If there were only reference and no sense, such a view would leave the difference between direct and indirect speech unexplained. For example, if someone were to report, “He said the same thing that all criminals do: ‘I did not steal the money,’” Frege would say that it is clear the words do not have their customary reference but designate what is usually their sense.

A third argument is that there are expressions – such as “the least rapidly convergent series” – that contain sense without having a reference. Similarly: names without reference, like “Santa Claus,” though this is not Frege’s example.

1.3 Frege’s Argument for Sense and Reference

Q: State one of Frege’s arguments for thinking that words have a sense as well as a reference.

1.4 “The Morning Star Is the Evening Star”

Q: Frege says that in order for there to be a difference in the cognitive value of a=a and a=b (provided a=b is true) there must be a difference in the sense of the two statements. What is his argument for this?

A: If the equality of a=b were regarded only in terms of reference, it would seem that a=b and a=a could not differ in cognitive value. If the equality of a and b is seen only with regard to what they designate, the kind of relation expressed would only be that of a thing to itself – which does not explain the fact that they differ in cognitive value. If one persisted, however, and tried to explain the difference in cognitive value by appealing only to what each term designates, one would be forced to appeal to the nature of the signs themselves and to assert a relation between them. The problem with this approach, however, is that the relation between the signs themselves would hold only to the extent that they designate something. As such, the things designated by the signs a and b would be the same thing. And this, Frege argues, is arbitrary. For in such a case, a=b would no longer refer to the subject matter and would express no new knowledge. But by a=b we often seek to express new knowledge. If a=b only by virtue of its sign taken as an object, or by virtue that it refers to the same thing, the cognitive value of a=b can not be distinguished from a=a.

The difference in cognitive value, Frege says, can be explained only if the difference between the signs corresponds to a difference in the mode of presentation of that which is designated. This mode of presentation can surface only if it is contained in the sense of the sign, Frege argues. Thus one can explain how a=b differs in cognitive value from a=a only by appeal to sense.

To repeat all this in another way: Here’s what’s wrong with the notion that identity is a relation between things or objects in the world:

Consider:

(1) a=a "the morning star is the morning star"
(2) a=b "the morning star is the evening star"

The reference here is the same for each proposition (“the morning star” and “the evening star” both refer to the same star), yet the signs meaningfully differ. In other words, both should have the same cognitive value, but they do not. Thus it cannot be the reference, but the signs themselves, in which the identify relation resides.

Frege is arguing against thinking about meaning as being an extension of the object – that is, reference.

Frege’s solution to the morning star-evening star problem is to say that signs denote objects by their sense, or mode of presentation. Frege holds sense to be independent from us; it somehow has semantic value.

1.5 Frege vs. Russell on Logical Form

Q: Briefly, contrast Frege and Russell’s view on the logical form of sentences like “The present king of France is wise.”

A: The central difference between Frege’s and Russell’s view on sentences like “The present king of France is wise” (S) is that Frege distinguishes between a presupposition and an assertion while Russell does not.

In Frege’s view, (S) contains a presupposition that the subject, here the king of France, does in fact exist. If Frege’s presupposition condition is not fulfilled, the sentence would not be said to assert that the present king of France is wise. In fact, Frege would argue, if the presupposition requirement is not fulfilled, the sentence is neither true nor false. Further: For Frege, the presupposition is not automatically contained in the assertion; it does not follow that the sense of (S) contains the thought that king designates something.

Russell, on the other hand, makes no distinction between assertion and presupposition. A sentence, on Russell’s view, is either true or false. And, he would say, a presupposition requirement like Frege’s arises because Frege confuses the grammatical subject with the logical subject.

On Russell’s view, (S) contains 3 separate propositions; one of them is the logical subject. If any propositional component of a sentence is false, the whole sentence is false. For example, if a proper noun does not refer to a single object – as in (S) – it is a logically false proposition. Russell takes a different explanatory track than Frege becomes he wants to preserve the law of excluded middle in the analysis of language.

2 Strawson

2.1 Strawson on the King of France

Q: What is Strawson’s reason for thinking that someone who said “The present king of France is wise” is not asserting that there is at present a king of France?

A: The reason lies in Strawson’s distinction between types and tokens, or in his words, between a sentence and the use of a sentence. While for Strawson significance, or meaning, lies with mentioning a sentence, truth and falsity are a function of the use of a sentence. Someone who said (in our time) that “The present king of France is wise” (S) would be merely mentioning the sentence, not actually using it to make a true or false assertion about the king of France.

Thus, Strawson argues, if someone actually said to you that (S) and asked whether you thought it was true or false, you would be inclined to say you thought it was neither. It has, for Strawson, no truth value. Because there is no such person as the king of france, because it is about no one, the question of the truth or falsity of the sentence does not arise.

Cf. Donnellan, in “Reference and Definite Descriptions,” who after distinguishing between the referential and attributive uses of definite descriptions argues against Strawson that the truth or falsity is affected differently depending on whether the definite description is referential or attributive. (See pp. 196-197.)

2.2 Strawson: Russell’s Twice Wrong About the King of France

Q: Strawson thinks that Russell says two false things about the sentence “The king of France is wise.” State both of them.

A: Answer: Strawson thinks that Russell would first say that anyone (now) uttering the sentence “The king of France is wise” (S) would be making either a true assertion or a false assertion. Second, Russell would say that part of what the person uttering (S) would be asserting would be that there at present existed one and only one king of France. Both of Russell’s claims are, Strawson believes, false.

2.3 Russell Confuses Meaning with Mentioning

Q: Why does Strawson think Russell wrong to claim that part of what is asserted when someone asserts “The present king of France is wise” is that there exists at present one and only one king of France? Is his reason adequate?

A: Russell is wrong to claim that part of the assertion “The present king of France is wise” (S) entails that there is a “king of France” (D) because, Strawson argues, Russell has failed to distinguish between a sentence and the use of a sentence – that is, between a type and a token. For Strawson, truth lies with a token and significance, or meaning, with the type.

Russell says D, though it is the grammatical subject of S, is not the logical subject of S. Strawson counters that D is not the logical subject of S. In fact, Strawson says, S is not logically a subject-predicate sentence at all. Russell fancies that he is talking about sentences when he is in fact, Strawson says, talking about the use of sentences.

Meaning is a function of the sentence; referring and truth or falsity are functions of the use of the sentence. Russell’s mistake was that he thought referring must be meaning. Russell, that is, confused meaning with mentioning.

3 Searle

3.1 Searle’s Distinction Between Regulative and Constitutive Rules

Q: State Searle’s distinction between regulative rules and constitutive rules.

A: Regulative rules regulate a pre-existing activity, an activity whose existence is logically independent of the existence of the regulative rules. Regulative rules characteristically take the form of imperatives.

Constitutive rules, on the other hand, constitute (and also regulate) an activity the existence of which is logically dependent on these rules; that is, constitutive rules create or define a new activity or form of behavior.

In football, an example of a regulative rule would be the etiquette or sportsmanship to which opposing players customarily adhere as they play the game. A constitutive rule would be that a touchdown is scored when a player crosses the opponents’ goal line in possession of the ball while play is in progress.

Searle’s distinction between constitutive and regulative rules is philosophically significant because, Searle argues, the semantics of language can be regarded as a series of systems of constitutive rules and illocutionary acts are acts performed in accordance with these sets of constitutive rules.

3.2 Searle vs. Kripke on the Semantics of Names

Q: What is Searle’s account of the semantics of names? State Kripke’s best argument against Searle’s account and explain why you think it is the best.

A: In general, Searle’s account of the semantics of names, adopted in part from Wittgenstein, holds that names are logically connected “in a loose sort of way” with the characteristics of the object to which they refer. [Cf. Wittgenstein’s theory of family resemblances in Philosophical Investigations. ]

More specifically, Searle’s account of proper names culminates in his view that it is a necessary fact that a proper name, e.g. Aristotle, has the “logical sum, inclusive disjunction, of properties commonly attributed to him.” Any individual not having at least some of the properties attributed to him could not be Aristotle, Searle argues.

Kripke, however, destroys Searle’s account of proper names with his counterfactual of Jonah, which is that all the descriptive information we have about him – gleaned from the Biblical story – turns out to be false. Yet Jonah nevertheless existed and is referred to.

The Jonah case is the best of Kripke’s arguments against Searle because it is a case in which all the known descriptive information about Jonah is presumed false. Searle would thus be forced to say – counterintuitively – that Jonah did not exist.

4 Kripke

4.1 Kripke on Nixon

Q: Give Kripke’s best argument for thinking that “Nixon” does not mean something like “the individual called ‘Nixon’.” Explain why you think it is the best of his arguments.

A: Kripke’s best argument combines his notion that a name is a rigid designator with the use of possible worlds to distinguish contingent and necessary truths. First, in Kripke’s view, a proper name is a rigid designator, which means that it designates the same object in every possible world. The rigidity of ‘Nixon’ stems from the stipulation that the token of the proper name `Nixon’ is being used to speak of the same individual in every possible world.

This leads to the second point, the one that gives Kripke’s argument its greatest force. Once we have stipulated that we are talking about the same individual in every possible world, it is merely a contingent truth that Nixon is called `Nixon.’ It is easy to imagine another world in which Nixon is not called Nixon. Yet it is a necessary truth that the stipulated individual is Nixon. Given that Nixon is Nixon, he may have been called something else in another possible world, but he still would have been the same individual. As Kripke puts it: “It is not the case that he might not have been Nixon (though he might not have been called `Nixon.’” Naming and Necessity, p. 49.)

This argument is Kripke’s best because it combines the strong intuition that a proper name is a rigid designator with the powerful logical device of possible worlds, which can be used to show whether a truth is contingent or necessary.

4.2 Kripke on the Meaning of “Nixon” and “Godel”

Q: State one of Kripke’s arguments against the view that “Godel” means “discoverer of the incompleteness of arithmetic,” and explain why he thinks that it is wrong to say that ‘Nixon’ means “the man called ‘Nixon.’”

A: One of Kripke’s arguments against saying that ‘Godel’ means “the discoverer of the incompleteness of arithmetic” is that it could turn out that Godel was in fact not the discoverer of the incompleteness of arithmetic. Assume that, instead, a man named Schmidt was. Then, under the cluster-of-descriptions theory against which Kripke is arguing, when someone says ‘Godel’ was “the discoverer of the incompleteness of arithmetic,” he is in fact referring to Schmidt – because Schmidt is the unique person satisfying the description. The only problem, of course, is that he is referring not to Schmidt but to Godel. The example devastates the logic of the description theory; see Naming and Necessity, p. 84. A proper name, as Kripke’s Godel-Schmidt example demonstrates, does not mean any properties or descriptions associated with the name.

Kripke thinks it’s wrong to say that ‘Nixon’ means the man called ‘Nixon’ for similar reasons. In Kripke’s view, a proper name is a rigid designator, which means that it designates the same object in every possible world. The rigidity of ‘Nixon’ stems from the stipulation that the token of the proper name ‘Nixon’ is being used to speak of the “same contextually specified individual in every possible world,” as Katz summarizes it in “Names without Bearers” (p. 14-15).

Once we have stipulated that we are talking about the same individual in every possible world, it is merely a contingent truth that Nixon is called ‘Nixon.’ Another world in which Nixon is not called ‘Nixon’ can easily be imagined. Yet it is a necessary truth that the stipulated individual is Nixon. Given that Nixon is Nixon, he may have been called something else in another possible world, but he still would have been the same individual. As Kripke puts it: “It is not the case that he might not have been Nixon (though he might not have been called ‘Nixon’” Naming and Necessity, p. 49).

4.3 Kripke Contra Kneale

Q: Kneale thought it “obviously trifling” to tell someone that Socrates is called Socrates, and he thought this a reason to think that “Socrates” means “The individual called ‘Socrates’.” Kripke does not agree it is a good reason. Why?

A: Kripke thinks

1. Socrates may not have been called Socrates, etc.
2. Socrates in Greek may be different 
from Socrates in English.

It’s wrong to say that ‘Nixon’ means the man called ‘Nixon’ for similar reasons. In Kripke’s view, a proper name is a rigid designator, which means that it designates the same object in every possible world. The rigidity of ‘Nixon’ stems from the stipulation that the token of the proper name ‘Nixon’ is being used to speak of the “same contextually specified individual in every possible world,” as Katz summarizes it in “Names without Bearers.” [p. 14-15.]

4.4 Undesirable Consequences of Kripke’s Referential Account of Names

Q: Kripke thinks that names do not have a sense – and he has an account of how they are used referentially in spite of having no sense. Can you think of any undesirable consequence of his view about names that cannot be handled on his account of how they are used referentially?

A: Kripke’s referential account of names presents at least three undesirable consequences. It fails to resolve conundrums posed by negative existential statements, identity, and substitution in opaque contexts. I expand on each in turn below.

First, Kripke’s referential account of names leaves the meaning of negative existential statments like “Santa Claus does not exist” (S) unexplained. If names, as Kripke believes, have no sense but only reference, then (S) would be meaningless. Yet (S) is meaningful and is no doubt used frequently during December to express true statements about the world.

The problem carries even more weight philosophically when, for instance, Kripke’s account of names is pressed into formulating an athiestic position, for how could Kripke assert within his referential theory of names that “God does not exist”?

A second undesirable consequence of Kripke’s theory is that it fails to demonstrate why “Mark Twain is Mark Twain” (m=m) is a trivial truth, while “Mark Twain is Samual Clemens” (m=s) is an unobvious, informative truth about the world. In other words, the problem of identity remains. The (m=s) example lends support to the view that even names can have some degree of sense, even if it isn’t the discriptive content which Kripke dispenses through his powerful counterfactuals.

A third undesirable consequence is that Kripke’s referential account of names leaves unresolved the problem of opaque contexts. In other words, if, as Kripke claims, names have semantic value only in terms of their reference, how can Kripke account for

1. (T) Lois Lane believes Superman flies
2. (F) Lois Lane believes Clark Kent flies

where 1. is true and 2. is false even though they refer to the same bearer?

Yet if Kripke’s approach fails to deal in a straightforward way with the above three cases – negative existential statements, identity, and substitution in opaque contexts – we have a deeper problem. For even though Kripke cannot adequately handle these cases within his referential theory, we must grant that he, with his counterfactuals, has demonstrated that names are not synonymous with the descriptions or properties of traditional intensionalist theory. Thus the failure of Kripke’s theory to deal adequately with the above three cases entails an unfortunate result for the philosophy of language: How to account for the meaning of names – and even such intuitions about their sense that arise from (m=s) above – without on the one hand requiring that they carry descriptive content and without saying, on the other hand, that they have meaning only if they have a bearer.

5 Donnellan

5.1 Donnellan on Analytic Sentences

Q: Donnellan says the following: “Given present circumstances, the correct thing to say is that all whales are mammals. But whether this is, as we intend it, a necessary truth or contingent is indeterminate.” What is his reason for thinking it indeterminate?

A: Donnellan’s reason for thinking it indeterminate is that our present use of such an analytic sentence, while correct now, should not be expected to hold for all hypothetical cases. It is true that, at present, whales are mammals. But whether this is an analytic or a contingent truth is indeterminate because the decision rests upon our being able to predict the outcome of all hypothetical cases, and this we cannot do. In other words, the criteria of the application of the term mammal to the total possible class of whales is indeterminate and thus not permanently fixed in advance of possible further empirical discoveries.

5.2 “Whales Are Mammals”

What is Donnellan’s basic reason for thinking that an alleged analytic sentence like “Whales are mammals” does not express a necessary truth?

Donnellan’s basic reason is that our present use of such an analytic sentence, while correct now, should not be expected to hold for all hypothetical cases. It is true that, at present, whales are mammals. But whether this is an analytic or a contingent truth is indeterminate because the decision rests upon our being able to predict the outcome of all hypothetical cases, and this we cannot do. In other words, the criteria of the application of the term “mammal” to the total possible class of whales is indeterminate and thus not permanently fixed in advance of possible further empirical discoveries.

Donnellan has other, related reasons why “whales are mammals” is not analytic, the primary one being that one must either be taught or learn, perhaps from a dictionary, that a whale is a mammal.

5.3 Can Millians Be Atheists?

Q: Do Millians like Donnellan and Kripke have a problem in formulating the claim of atheism? If so, what is it? Can it be overcome?

A: Yes, such Millians as Donnellan and Kripke do have a problem in formulating the claim of atheism. The problem stems from their direct reference account of meaning. In the Millian view, proper names have denotation, but not connotation. Thus, since for Millians a name acquires its semantic value solely from its referent, Millians cannot explain the meaning of negative existential statements and other statements in which the referent is uncertain. For example, the sentence, “God does not exist” poses a problem for Millians because on their account, since the referent of “God” is uncertain, they cannot explain how the sentence nevertheless has meaning, as it intuitively does. Millians would be forced to say the sentence has no meaning when in fact it does.

The problem can be overcome only by introducing a notion of sense into the Millians’ theory. By introducing sense, the direct reference theory, of course, collapses into a broader account of meaning, one that reconciles the strong intuitions of the Millians – that proper names derive meaning not from the descriptions or properties associated with them – with the power that the notion of sense gives a theory in accounting for such problems as negative existential statements. The trick is that the notion of sense must be construed so narrowly that it does not associate a name with descriptive properties of the bearer. One possibility, expounded by Katz, would be to define the sense of an expression as the aspect of its structure that determines its sense properties. On this view, the sense of a proper noun would have the form The thing which is a bearer of `N’. The theory, that is, must be a pure metalinguistic one.

6 Putnam

6.1 Putnam’s Twin Earth Argument

Q: What is Putnam’s Twin Earth argument that meanings are not in the head (or that my Doppelganger on Twin Earth and I here cannot both have the word water with the same meaning?

A: Putnam’s Twin Earth counterfactual argues that the extension of the term water in the idiolect of my Doppelganger on Twin Earth is different from the extension of the term water in the idiolect of me here. For example, on Earth I can point to a particular liquid, composed of H20, and call it water. Meanwhile, my doppelganger on Twin Earth can point to a similar liquid – one used in the same ways on Twin Earth and having the same superficial properties as water on Earth but composed of XYZ, not H20 – and call it water. Yet, by the scientifically determined nature of water on Earth, the liquid on Twin Earth, when a sample of it is brought back to Earth, will not be water. The extensions of the two words are different, and hence they have different meanings. Otherwise, upon returning to Earth with the water sample for Twin Earth, we would find ourselves in the paradoxical situation of saying that, based on its superficial descriptive properties, it is water while, because it is not H2O, it is not water.

Putnam’s Twin Earth argument further shows that the sense of the term water is not enough alone to fix the extension of the natural kind term because the extension of the term is determined by a scientific appeal to the natural world. Thus, extension, Putnam says, is not determined by psychological state. The meaning of natural kind terms is not in the head.

6.2 Putnam and Burge on Meaning as Psychological State

Q: Why does Putnam think that knowing the meaning of a word is not just a matter of being in a certain psychological state? What does Burge add to Putnam’s conclusions? What might be said against their position? What side are you on? Explain.

A: Putnam believes that knowing the meaning of a word is not just a matter of being in a certain psychological state because, as he demonstrates with the help of his Twin Earth water counterfactual in “The Meaning of Meaning,” it is possible for two speakers to be in exactly the same psychological state even though the extension of the term “water” in the idiolect of the one is different from the extension of the term “water” in the idiolect of the other. For example, on Earth I can point to a particular liquid, composed of H20, and call it “water.” Meanwhile, my doppelganger on Twin Earth can point to a similar liquid – one used in the same ways on Twin Earth as water is on Earth but composed of XYZ, not H20 – and call it “water.” Yet, by the scientifically determined nature of water on Earth, the liquid on Twin Earth, when a sample of it is brought back to Earth, will not be “water.”

Putnam’s example shows that the sense of the term “water” is not enough alone to fix the extension of the natural kind term because the extension of term is determined by a scientific appeal to the natural world. Thus, extension, Putnam says, is not determined by psychological state. Meaning is not in the head.

Burge takes Putnam’s argument a step further. While Putnam confines his argument to natural kind terms, Burge puts forth a counterfactual that expands the scope of terms for which meanings are not in the head. Burge believes that his counterfactual of people with arthritis demonstrates that the meanings of such terms as “arthritis” are not in the head, but are socially determined. Terms like “brisket,” “contract,” and “recession” probably provide analogous cases, Burge says.

6.3 Cats Ain’t Necessarily Animals

Q: Explain Putnam’s reason for thinking that cats ain’t necessarily animals.

A: In a nutshell: The criteria of application of a term rests on science.

Putnam, in “It Ain’t Necessarily So,” argues that the “analyticity” of “cats are animals” depends upon the fact that the word “animal” is the name of a semantic category and the word “cat” is a member of that category. It might turn out that cats are in fact not animals at all, but robots. Such a discovery would remove the word “cats” from the semantic category of “animal.” In other words, if all cats are discovered to be something other than animals – say, highly sophisticated automata controlled by Martians - then we will no longer say that cats are animals, even though we may continue to call the robot cats “cats.”

Moreover, Putnam distinguishes truths of the sort “all cats are animals,” which are true because they are based on a rather advanced body of scientific knowledge, from truths of the sort “all bachelors are unmarried,” which are true by definition. The former, Putnam says, are “less necessary” than the latter; they are contingent upon scientific knowledge. Aspects of that scientific knowledge, such as that all cats are animals, could turn out to be wrong. On the other hand, scientific knowledge has no bearing on whether bachelors are thought to be unmarried or not.

6.4 Putnam on Analyticity and Semantic Categories

Q: Why does Putnam think that so-called analytic sentences (e.g., “Cats are animals”) are not necessary truths.

A: Putnam, in “It Ain’t Necessarily So,” argues that the analyticity of “cats are animals” depends upon the fact that the word “animal” is the name of a semantic category and the word “cat” is a member of that category. But it might turn out that cats are in fact not animals at all, but robots. Such a discovery would remove the word “cats” from the semantic category of “animal.” In other words, if all cats are discovered to be something other than animals – say, highly sophisticated automata controlled by Martians – then we will no longer say that cats are animals, even though we may continue to call the robot cats “cats.”

The placement of cats in the category of animals in turn rests upon a rather advanced body of scientific knowledge, which makes such sentences as “All cats are animals” seem like necessary truths. Scientific knowledge is subject to change, however, and we may in fact discover, for example, that cats are not really animals. That cats are animals is contingent upon our body of scientific knowledge.

Moreover, Putnam distinguishes truths of the sort “all cats are animals,” which are true because they are based on scientific knowledge, from truths of the sort “all bachelors are unmarried,” which are true by definition. The former, Putnam says, are “less necessary” than the latter. Scientific knowledge has no bearing on whether all bachelors are thought to be unmarried or not.

6.5 Putnam and Natural Kind Terms

Q: What is Putnam’s twin earth argument? Is it related to his robot cat argument? How? What does the twin earth argument show?

A: Putnam’s Twin Earth argument shows that the extension of the term “water” in the idiolect of my Doppelganger on Twin Earth is different from the extension of the term “water” in the idiolect of me here and, as such, that meanings are not in the head.

The argument goes like this: On Earth I can point to a particular liquid, composed of H20, and call it “water.” Meanwhile, my Doppelganger on Twin Earth can point to a similar liquid – one used in the same ways on Twin Earth and having the same superficial properties as water on Earth but composed of XYZ, not H20 – and call it “water.” Yet, by the scientifically determined nature of water on Earth, the liquid on Twin Earth, when a sample of it is brought back to Earth, will not be “water.” The extensions of the two words are different, and hence they have different meanings. Otherwise, upon returning to Earth with the water sample for Twin Earth, we would find ourselves in the paradoxical situation of saying that, based on its superficial descriptive properties, it is water while, because it is not H2O, it is not water.

Putnam’s Twin Earth argument further shows that the sense of the term “water” is not enough alone to fix the extension of the natural kind term because the extension of the term is determined by a scientific appeal to the natural world. Thus, extension, Putnam says, is not determined by psychological state; the meaning of natural kind terms is not in the head.

Besides its similar counterfactual form, Putnam’s cat robot argument is related to the twin earth argument by showing that what we believe to be the extension of a term is tied to an advanced body of scientific knowledge about the world. That we say “all cats are animals” is contingent upon our knowledge about the world. If it turns out that all cats are in fact robots, we will no longer say they are animals, though we will continue, Putnam believes, to call them “cats.” Thus, the cat-robot example, like the twin earth water argument, shows that the meaning of “cat” is not in the head.

Both arguments show that meanings of such natural kind terms as “water,” “animal,” and “cat” are not in the head and that the criteria of application for such terms is based in an advanced body of scientific knowledge. The criteria of applying the word “animals” to cats, for instance, is contingent upon our knowledge that cats are animals and not, say, robots controlled by Martians. Likewise, the criteria for applying the word water to a substance that has all the superficial phenomenal properties of water is that the substance’s nature is in fact H20 and not, say, XYZ.

That is, the (current) core criteria by which we call a liquid substance “water” is that it is H20. The essential nature of water is that it is H20, and knowledge of this essential nature rests of an advanced body of scientific knowledge.

In “It Ain’t Necessarily So,” Putnam is broadly concerned with, among other things, defending the traditional synthetic-analytic distinction against Quine’s attack of it. More specifically, though, Putnam is interested in the criteria for a term’s application. As such, he finds himself asking “what would we say if … .” Putnam attempts to use such hypothetical questions to discern not only whether people apply a term on the basis of a particular characteristic, but also whether or not people would continue to apply the term to the set if it were to lack the characteristic in question.

7 Grice

7.1 Grice on Meaning and Intention

Q: What does Grice think is the proper way to explain the locution “A meant_nn something by x”?

A: For Grice, the proper way to explain that ‘A meant_nn something by x’ is to say that “A intended the utterance of x to produce some effect in an audience by means of the recognition of this intention.” To ask what A meant, Grice says, is to ask for a specification of the intended effect.

7.2 Grice vs. Searle on Convention

Q: Can Grice reply to Searle’s counterexample that the American soldier’s utterance “Kennst du das Land…” means “I am German”? If so, why, and if not, why not?

A: Searle’s counterexample of a American soldier trying to convince his Italian captors that he is German officer by uttering “kannst du das Land … ?” is intended to demonstrate that Grice’s account of meaning does not show the connection between what a speaker means and what the words he utters mean. Grice’s account of meaning, Searle says, overlooks the fact that what we mean with an utterance is also a matter of convention. Searle thinks his counterexample satisfies all Grice’s conditions for meaning – yet the utterance still fails to have meaning, he says.

Grice can reply to Searle’s counterexample in a straightforward way. Grice can merely say that since the American officer does not really know German, what he is saying by uttering in German “do you the know the land where the lemon trees bloom?” does in fact mean “I am a German officer” given the context, the speaker’s intention, the recognition by the hearer of that intention and the effect he hopes it will have. In the situation of Searle’s counterexample, the token of the type in question does mean “I am a German officer.”

One shortcoming of Grice’s reply, however, is that it is limited to the particular utterance in question, though Grice can respond with his own example against Searle: “Welcome to my shop,” says an Arab shopkeeper to a Brit. The sentence, however, is a English token of an Arabic type that means “you are a pig” or some such. In such a case, the meanings are the same, Grice would say.

Major point: Both the examples demonstrate that Grice’s account is one of tokens of utterances or sentences.

7.3 Grice’s Conception of Utternace Meaning

Q: What is Grice’s conception of utterance meaning? How can Grice’s conception be defended against Searle’s alleged counterexample?

A: Grice’s conception of utterance meaning rests on the notion of intention. To say that an utterance meant something is, for Grice, to say that a person intended the utterance of X to produce some effect in an audience by means of the recognition of the intention. Thus, to ask what A meant, Grice says, is to ask for a specification of the intended effect.

Searle’s counterexample of a American soldier trying to convince his Italian captors that he is German officer by uttering “kannst du das Land, wo die Zitronen bluhen?” is intended to demonstrate that Grice’s account of meaning does not show the connection between what a speaker means and what the words he utters mean. Grice’s account of meaning, Searle says, overlooks the fact that what we mean with an utterance is also a matter of convention. Searle thinks his counterexample satisfies all Grice’s conditions for meaning – yet the utterance still fails to have meaning in the way Grice intended it to, Searle says.

Grice can reply to Searle’s counterexample in a straightforward way. Grice can merely say that since the American officer does not really know German, what he is saying by uttering in German “do you the know the land where the lemon trees bloom?” does in fact mean “I am a German officer” given the context, the speaker’s intention, the recognition by the hearer of that intention and the effect he hopes it will have. Grice’s response appeals to the token-type distinction, with the token of the sentence bearing the burden of meaning in this case. Grice could accuse Searle of failing to perceive such a mention-use distinction.

8 Austin

8.1 Austin on Performative Sentences

Q: Austin thinks that performative sentences do not have truth conditions. Does Austin have a reason for thinking this? Does that reason undermine Davidson’s conception of the meaning of sentences?

A: Performative sentences do not have truth conditions, Austin says, because they do not describe or report anything. Further, uttering a performative sentence like “I promise to pay you back next week” is not so much saying something as doing something. If nothing is said, it can be neither true nor false. More: The truth values of performatives cannot be analyzed without taking into account the speaker’s intentions and the hearer’s understanding of the use of such a form.

In broad terms, Davidson’s holistic conception of meaning seeks to replace Frege’s intensionalist position that manifests itself as “‘p’ means ‘q’” with an extensionalist approach based in the truth conditions of sentences. More specifically, Davidson would define meaning by replacing “p means q” with “S is true if and only if p,” with ‘S’ standing for sentence.

Davidson’s definition of meaning works, he says, because it gives necessary and sufficient conditions for the truth of every sentence. To give truth conditions, Davidson says, is a way of giving the meaning of a sentence.

But if Austin’s view that performative sentences have no truth value is correct, then Davidson’s definition of meaning cannot give the conditions of truth, and hence the meaning, for performative sentences – because they have none.

Austin’s view, if right, undermines Davidson’s conception of the meaning of sentences because, for it to work, all sentences must have truth conditions to have meaning. If Austin is right, Davidson finds himself having to explain meaning without appealing to truth conditions.

8.2 “I hereby wish you a happy new year.”

Q: “I hereby wish you a happy new year.” Why does Austin think such utterances are not true or false.

A: Austin thinks that such performative sentences as the one above do not have truth values because they do not describe or report anything. Further, uttering a performative sentence like “I hereby wish you a happy new year” is not so much saying something as doing something. If nothing is said, it can be neither true nor false. More: The truth values of performatives cannot be analyzed without taking into account the speaker’s intentions and the hearer’s understanding of the use of such a form. These three interconnected reasons lead Austin to think that performative utterances are neither true nor false.

9 Davidson

9.1 Davidson Contra Traditional Theorists on Belief Sentences

Q: What problem might Davidson have in mind when he says that traditional theories of meaning cannot handle the semantics of belief sentences? How might a traditional theoriest try to solve the problem?

A: The problem that Davidson has in mind is the one introduced by Mates in “Synonymity” and explored by Church in “Intensional Isomorphism,” both of whom are broadly concerned with investigating Carnap’s notion of synonymy and, more generally, meaning. Consider: (1) Whoever believes that D, believes that D. (a week) (2) Whoever believes that D, believes that D’. (seven days)

The problem, as Mates points out, is that if anybody even doubts that whoever believes that D believes that D’, then a Carnapian explanation of synonymity is wrong. The doubt indeed arises because, Davidson believes, (1) and (2) may have different truth conditions.

The specific problem that Mates’ remark elicits in Davidson’s eyes is that one cannot account for the meaning of sentences, let alone their truth conditions, on the basis of the senses of the words in them. Knowing the meaning of the words in a sentence does not, for Davidson, add up to knowledge of what the sentence means. Specificially, Mates’ example demonstrates, Davidson believes, that knowledge of a sentence’s syntax combined with knowledge of its words does not alone yield knowledge of what the sentence means. This in turn leads Davidson to theorize that the only way to give the meaning of a sentence is to provide a matching sentence that gives its meaning.

A traditional theorist might try to rescue the sentence’s meaning by putting forth a notion of sense instead of relying on the truth conditions of the terms. The doubt beween 1 and 2 arises because the two terms have a slightly different sense, even though they are extensionally synonymous.

9.1.1 Comments

Compare Mates’s problem, of which Davidson is thinking: The problem emerges when senses are the same. Examples:

(i) Whoever believes that B, believes that...fortnight.
(ii) Whoever believes that B, believes that...two weeks.

Davidson would say that in a case like this you have the same syntax and the same sense yet the truth conditions are different.

The problem with Davidson’s line of criticism is that it fails to grant Frege’s theory with the explanatory power that it has with respect to other important cases, such as: - negative existential references - analyticity - identity statements - opaque contexts and substitution.

9.2 Davidson on Sentential Meaning

Q: Explain why Davidson thinks that a dictionary does not “touch the standard problem, which is that we cannot account for even as much as the truth conditions of [belief] sentences on the basis of what we know of the meaning of the words in them.”

A: One cannot account for the meaning of sentences, let alone their truth conditions, on the basis of the senses of the words in them, even when combined with knowledge of the sentence’s syntax, Davidson argues. Knowing the meaning of the words in a sentence does not, for Davidson, add up to knowledge of what the sentence means.

The problem can be demonstrated by a paradigm introduced by Mates in “Synonymity”:

(1) Whoever believes that D, believes that D.
(2) Whoever believes that D, believes that D'.

The problem, as Mates points out, is that if anybody even doubts that whoever believes that D believes that D’, then a traditionalist’s explanation of synonymity is wrong. The doubt indeed arises because, Davidson believes, (1) and (2) may have different truth conditions even though the only difference between the sentences lies in two synonomous expressions. If the sentences have different truth conditions, despite the synonymy of the only different expressions in them, then knowing the meaning of the words, as provided for instance by a dictionary, cannot account for the differences in truth conditons.

9.3 “Snow Is White” Is True If and Only If Grass Is Green

Q: Davidson says, “The grotesqueness of (`Snow is white’ is true if and only if grass is green) is in itself nothing against a theory of which it is a consequence, provided the theory gives the correct results for every sentence.” Why does he think this? Is his reason a good one? Explain.

A: Answer: First off, Davidson thinks that such an absurdity is not party to his theory. Indeed, he believes that such a consequence is nothing against it. For Davidson, that “Snow is white is true if and only if grass is green” (S) is not part of the apparatus of his theory, but a consequence of it. In other words, Davidson believes that the grotesqueness, or absurdity, of (S) doesn’t apply to the theory and, as such, is not of valid criticism of it.

Davidson’s reason for thinking this, however, is not a good one. First, if Davidson’s theory just matches truth conditions with meaning, he fails to provide a sufficient answer to the objection because Davidson’s theory does not account for our intuitive notion of meaning. In other words, it leaves something out of its explanation of our intuitive notion of meaning precisely because it produces this absurd result. The theory is too sparse; it does not properly specify meanings.

Second, Davidson’s response is inadequeate because, even granting that Davidson’s theory by itself is not absurd and that such simple material truths about the world as “grass is green” and “snow is white” have the same truth value, the conjuction of the two is absurd. Thus the problem seems to be with Davidson’s theory even though Davidson says he’s not committed to such a consequence. We can, however, generate exactly the same kind of absurdities using the formal truths of logic, and to that Davidson cannot respond that he is not committed to the truths of formal logic, because they are built into his theory.

10 Quine

10.1 Quine’s Argument Against Explaining Synonymy Using Substitution

Q: State Quine’s argument against the possibility of explaining synonymy on the basis of substitution procedures. Would there be any bad consequences of accepting the form of argument that Quine uses here? Explain.

A: First, Quine would argue that in general, there appears to be a class of so-called analytic statements typified by “No bachelor is married” (1). It can be turned into a logical truth (a statement that is true and remains true under all reinterpretations of its components other than logical particles) by putting synonyms for synonyms. Thus (1) can be rendered as the logical truth “No unmarried man is married.” Yet such a procedure leans on the notion of synonymy, which, Quine says, requires as much clarification as analyticity. After finding that attempts to define synonymy rest on pre-existing synonymies, Quine turns to interchangeability to explain it.

Interchangeability in all contexts without change in truth value, however, will not explain synonymy because, discounting the drawback of appealing to a prior notion of “word,” it requires that we rely on the notion of analyticity. That is, interchangeability salve veritate assumes we have already made satisifactory sense of analyticity, making the argument circular.

More specifically, we are forced to rely on necessary truths because truth values cannot pick out synonymous pairs – otherwise you fail to account for coextensive terms. Thus you need necessary truths to get around problems with coextension. But the appeal to necessary truths begs the question because it requires analyticity. If you have to use `necessarily’ you have analyticity already, Quine argues.

10.2 Quine Contra Carnap

Q: Quine makes two arguments against Carnapian meaning postulates providing a basis for drawing the analytic-synthetic distinction. Briefly state them. Is there a linguistic reason to think that one of Quine’s argument in this connection has to be right?

A: Besides the possibility that the semantical rules of meaning postulates may themselves contain the word ‘analytic’ or that analyticity may be attempted to be explained by the semantical rules themselves, meaning postulates explain analyticity only relative to a particular language. Analytic truths, however, are supposed to be language independent.

Anyway, how would one know that languages are not being missed? Languages might form an open set, in which case the examples would not be complete and the definition would not capture the abstract property of analyticity.

Another argument against Carnapian meaning postulates is that they provide merely a list of examples of analytic sentences without actually giving a definition of analyticity. Even if you suppose there is only one language, the list would not define the heading “analytic”; it would just give its extension. Carnap’s meaning postulates theory, Quine argues, is just a fancy version of using examples for definitions. Meaning postulates are not about the sense of words but about their referents. As such, meaning postulates are not about synonymy and analyticity at all, but about the broader concepts of necessary truth and necessary equivalence.

10.3 Quine’s Indeterminancy Thesis

Q: State Quine’s indeterminacy thesis. Explain (i) why he thinks indeterminacy is different from the evidential underdetermination of scientific theories and (ii) why he thinks that the study of translation is “worse off” than physics.

A: Quine’s indeterminacy thesis, founded on the finding that radical translation reveals no language-independent propositions, holds that in the study of intensional semantics there is no proper scientific subject – no facts against which to verify a theory. In radical translation, the absence of indepedent controls in the matching of one seemingly synonymous term for another causes indeterminancy. Thus, interderminancy makes it misleading to say that an individual statement has empirical content, and it further becomes folly to seek a boundary between analytic and synthetic statements.

Indeterminancy is different from the evidential underdetermination of scietific theories because with indeterminancy, there is no fixed evidence from which to draw a sample in order to formulate a hypothesis, while scientific theories suffer from only underdetermination because there is a body of fixed evidence from which a sample can be drawn and a hypothesis developed, even though the sample is incomplete. Thus, the physicist is forced to project his theory toward elements outside the sample.

The study of translation is worse off than physics because it suffers from indetermination while physics suffers from underdetermination. In physics, a theory can be verified against a body of fixed evidence; the underdetermination is epistemic – a limitation on knowledge.

In the study of translation, however, indeterminacy prevails: there is no fact of the matter to be right or wrong about – because there are no language-neutral propositions. In other words, there is no body of evidence against which a semantic theory could be verified. As such, semantic theories are, Quine would say, theories about nothing.

10.4 Quine’s Argument Against the Analytic-Synthetic Distinction

Q: What is the structure of Quine’s argument against the analytic-synthetic distinction in “Two Dogmas of Empiricism”? Does the argument play any role in his case for the indeterminancy of translation?

A: The structure of Quine’s argument in “Two Dogmas of Empiricism” is a proof by cases. There are, Quine says, three places that we could reasonably look to make objective sense of meaning: definition, logical theory, and linguistics. Quine asks whether the methods for explaining in any of the three areas can explain analyticity and synonymy.

The proof by cases dispels each one in turn. Definition falls short because it either assumes prior synonymy relations or else has nothing to do with meaning. In the area of logical theory, Carnap’s meaning postulates and semantic rules are shown to be of no help in revealing the nature of analyticity and synonymy; with their extensionalist nature, they pertain more to necessary truths and necessary equivalence than analyticity. In linguistics, Quine demonstrates that the methods for defining concepts are unable to define synonymy and analyticity without being circular. Thus, Quine concludes, there are no methods by which one can clarify snynonymy and analyticity and hence there are no linguistically neutral meanings.

The argument from “Two Dogmas” supplies the “missing” argument in the case for the inderminancy of translation. The argument plays a role in the indeterminacy thesis because Quine’s reason for thinking that independent controls do not exist in translation takes its force from the argument that there are no linguistically neutral meanings. The absence of linguistically neutral meanings is a prerequiste for the indeterminacy of translation.

Interpretation and Indeterminancy Throughout the essay, I will argue a hard line: the exact meaning of a speaker’s utterance in a contextualized exchange is often indeterminate. Within the context of the analysis of the teacher-pupil exchange, I will argue for the superiority of interactional linguistics over speech act theory because it reduces the indeterminacy and yields a more principled interpretation, especially when the interactional approach is complemented by elements from other sociologically influenced methods, namely the ethnography of communication and Labovian sociolinguistics. Read more …

Hermogenes Was Right, Socrates Wrong In Harris and Taylor’s chapter on Plato’s “Cratylus” in Landmarks in Linguistic Thought, Cratylus takes the position that the form and meaning of a word are inextricably related. For Cratylus, “everything,” including Hermogenes, “has a right name of its own, which comes by nature” even though some people, like Hermogenes, are named incorrectly (Cratylus 383, as quoted in Harris and Taylor, p. 1). Read more …