(i) People tend to confuse (referential uses of) definite descriptions with complex demonstratives.

(ii) Many students need to get much clearer on the most crucial terms 'definite', 'description', 'refer', 'denote', 'true' 'sentence', etc. Moreover, many need to be clear about issues such as use and mention: definite descriptions exist whether or not they denote, contrary to what many have said. When writing a subject where a word is used in a variety of ways, particularly if different authors give the same term different meanings, be very clear about which meaning you intend when you use the term. In the present case, many students disregarded the variety of relevant ways that their use of words like 'meaning' might be taken.

(iii) Spatially separate each of your answers, make each answer completely self-contained, number them, and arrange them in the same order as the questions are asked on the examination, so it is clear which question a given answer is supposed to address. If you collaborate on an assignment, hand in a single assignment with all the collaborators' names on it.

1. [pp. 362-363] A1: The semantic function of a given definite description in a given context is determinate; i.e., once a definite description is selected, and it is embedded in a particular sentential context, its semantic properties are fixed. In particular, on any given occasion of use of the sentence in question, there is no contribution made to the semantic function of the definite description by the "way" it is used (in that sentence) on that occasion. A2: If nothing fits what a definite description describes, then the truth value of the containing sentence may be affected. (For Russell, such a sentence is false; for Strawson, it lacks truth-value.)

2. A speaker S makes an attributive use of 'the F' iff S uses 'the F' in a sentence in which he says something about the one and only thing that is F. S makes a referential use of 'the F' iff S says something about a thing x, and uses 'the F' with the intention that S's audience will be able to identify x. For example, consider the sentence "The winner of the marathon is quite an athlete". If I do not know who won the marathon, but only that the winning time was 2:07, I am making attributive use of 'the winner of the marathon', because I only mean to say something about the one and only winner of the marathon, whoever he is. On the other hand, if I think that G won the marathon, and I happen to know that G is an exceptional athlete, perhaps in many sports besides distance running, and I want to communicate this to my friends who are going to G's victory party, I might use the definite description referentially. Since I was only using it to pick out G and say something about him, my friends, who just learned that G had been disqualified, might still agree that I said something true of G. [Note: The most illustrative examples of referential uses of definite descriptions involve cases where the intended referent does not satisfy the description through which reference is made.]

3. (i)When using a definite description attributively, one might not know what thing one is denoting by means of the description; on the other hand, when using a description referentially, the speaker intends to make his audience aware of what thing in particular he is talking about. (ii) If there is not exactly one thing that is F, then any sentence in which 'the F' is used attributively is false. On the other hand, this is not so if 'the F' is used referentially. If there is an unusual abstract sculpture on the roof of my neighbor's house, and I have always mistaken it for a TV antenna, and one day after a storm I see that one of the metal rods is loose and is swinging and clanking in the breeze, I might tell my neighbor, "The TV antenna on your roof is broken." My neighbor might not have a TV antenna on his roof, but he will probably be able to tell that I am talking about his sculpture. Thus, I will have successfully referred to the sculpture, and I have told my neighbor something true of it. [In fact, I need not even believe it is a TV antenna to report something true: suppose I know it is a sculpture, but the person who is tending my neighbor's house while he is away doesn't, and I want to alert him that the storm damaged the thing on the roof. My referential use of 'the TV antenna on the roof' allows me to refer to the sculpture and to say something true of it.] (iii) When used referentially, even if nothing fits the description, orders can be obeyed ["Fix the TV antenna on the roof before my neighbor gets home"] and questions can be answered ["Did you fix your TV antenna yet?"]. (iv) Furthermore, while both uses usually presuppose that the speaker thinks that there is exactly one thing that fits the description, when used referentially, this fact is derivative from the fact that the speaker usually thinks that some particular thing fits the description. (v) Finally, as we have seen in the TV antenna case, one might refer to something with a referential use of a definite description, but upon learning that the referent was not the one and only F, admit that one had misdescribed the thing one had nonetheless successfully referred to. This cannot happen when a definite description is used attributively. [Note: If you point out that both uses of definite descriptions presuppose that something fits the description (367), you must explain how this claim is so, despite the fact that the whole point of distinguishing the two uses is that with respect to referential use, nothing, not even the referent, needs to "fulfill the description" in order for something to be truly predicated of the denotation of the referentially used definite description. Ideally, you would also explain how this presupposition is imputed to the referent of the referentially used definite description.]

4. In a typical referential use of a definite description, one believes that there is some particular thing that (uniquely) satisfies the description. But it is possible to have communicative intentions such that one referentially uses a definite description even though one believes that the intended denotation does not satisfy the description. I might tell my friends that the police officer smacked me around even though I do not believe that the person who beat me up was a policeman. I might have been in a room with thugs dressed up as a police officer, a butcher, and a baker. My friends may have caught them all soon after, and they want me to tell them which one of the thugs beat me up. I thus use the description referentially, even though nobody thinks the guy is actually a cop.

Standardly, attributive uses of definite descriptions are such that the speaker does not believe of some particular thing that it satisfies the description. But one can make attributive use of a definite description even though there is some thing such that one believes that that thing satisfies the description. You might believe that G won the marathon, but before saying that the winner of the marathon is quite an athlete, you explain just how difficult it is to run a 2:06 marathon, and what sort of lung capacity and leg strength is needed for such an accomplishment. Thus, you show that the winner of the marathon, whoever he is (although you believe it is G), is quite an athlete.

Strawson's third claim, that the "reason he has said nothing true or false is that he has failed to refer," is true at most for certain extreme cases of referential uses of definite descriptions, such as when there is no object to which the speaker might be referring (and willing to "pick out" as such). If, for example, I am hallucinating, and I say 'The president of Rutgers is doing a jig', then I have simply failed to refer to anything at all. (Question: Will I have referred to anything if I had used some other device, such as 'Francis Lawrence', 'he', or 'that'?) This case is different from referential uses in which I do pick out something (such as Richard Foley, or a statue or a tree), even if it is not the president of Rutgers, or even a person. Where I have made reference to something, Strawson's third claim is false, because it purports to give the reason my utterance is neither true nor false, and the reason for the truth or falsity of my utterance will not be that I have failed to refer. Finally, Strawson's claim is false, or at least unclear, for all attributive uses of the definite description. Here too, regardless of whether my utterance is true or false or not, the reason will not have to do with whether or not the speaker has failed to refer, since when a definite description is used attributively, nothing is referred to, but only denoted.

6. Donnellan holds that Russell's theory of proper names is one according to which the only semantic contribution a proper name makes to a sentence in which it is used is to refer to its referent. This contrasts with (Donnellan's theory of) the attributive use of definite descriptions, which denote their denotations only by means of describing them and ascribing properties to them. However, Donnellan also thinks that the referential use of a definite description can function like a genuinely referring expression, since none of the descriptive components necessarily apply to the referent, and there is nonetheless "a right thing to be picked out by the audience". [Note: Donnellan brings out a distinction between definite descriptions and proper names by saying definite descriptions are general. To say that an expression is general is only to say (roughly) that it contains a predicative that is or could have been true of more than one thing. In the case of definite descriptions, of the form 'the F', the predicate 'F' might have been true of more than one thing, or at least it might have been true of something other than what it is actually true of. With that said, it can be seen that generality is not the same as vagueness or ambiguity. These three notions should be distinguished from one another.]

1. [p. 384] A definite description of the form 'the F' is improper when 'F' is not true of exactly one thing. The interesting cases of improper definite descriptions occur when they are seemingly unproblematically used to talk about some unique thing. For example, if A, B and C only use the Alexander library, and they know that the others use only that one too, and one night C asks B where A is, and B says "A is at the library," it is reasonable to suppose that C will understand that B intended to communicate that A is at the Alexander library. This is so even though the truth conditions for what B said, which entail that there is exactly one library and A is at it, are not met, and so strictly speaking, B's report is false. Such improper definite descriptions are very common: "The luncheon is at noon", "The conference is in December", "The cat is on the mat". Also 'the present king of Canada' is improper, since the contained description is true of nothing, hence it is not true of exactly one thing. [Notes: (i) There is no logical connection between the distinction between speaker's reference and semantic reference and the distinction between proper and improper (uses of) definite descriptions (although there may be an explanatory connection in particular instances of use). With respect to a given definite description of the form 'the F', what's at issue in this question is whether or not there exists exactly one thing that 'F' is true of. Even if 'F' is clear, unambiguous, and satisfied uniquely by some item x, it will always be possible to imagine a case where 'the F' has as its speaker's referent some item y (where x ≠ y). (ii) The impropriety of a definite description is distinct from whether the definite description is vague or ambiguous. These three notions should be kept distinct from one another.]

2. Kripke's distinction between speaker reference and semantic reference can be viewed as part of a larger distinction between the speaker's meaning and the semantic meaning of an expression. As Kripke notes, the account can be extended to proper names. For example, if two students are talking about "Dthat", and the one says to the other "Davidson discusses demonstratives in that paper," the student intends to refer to Kaplan, even though the semantics for 'Davidson' are such that it refers to Davidson and never refers to Kaplan. Thus, we might say that the speaker's meaning of his use of 'Davidson' is Kaplan (as the referent), and the semantic meaning is Davidson.

Non-referring expressions also may be said to have a semantic meaning and a speaker's meaning. With respect to predicates, for example, suppose a child is inviting his friend to stay for supper, and explains proudly that his mother is baking macaroni and cheese. In fact, no baking is involved in preparing macaroni and cheese, only boiling and melting on the stovetop. Nonetheless, the child is able to implicate a truth that will help his friend decide whether or not to stay.

With respect to quantification, we can see a divergence between speaker meaning and semantic meaning in the following example. After reading a book that has many particularly bad arguments for outrageous claims, X throws down the book and says "Every sentence in this book is false!" On the one hand, the semantic meaning of what X said is such that X's sentence is surely false. No doubt there is at least one sentence in the book that is true, even if only "This book was printed in Great Britain", and X's audience would realize this. On the other hand, what X meant to convey may have been true, which was that many, most, or even all of the central claims or assumptions the author made were false.

It is also possible for a similar distinction to occur with adverbs. Suppose I am a thug and I am explaining my most recent crime to my fellow thugs, and I say "At first he didn't want to give me his wallet, but after I pistol-whipped him ten or twenty times, he cheerfully handed it over." Here I did not intend to communicate that the victim was actually cheerful about handing over the wallet, and it is doubtful that any of the audience thought that he was cheerful. What I am communicating to my audience is that the victim handed over the wallet immediately upon demand and without resistance, as one would do if one were cheerfully handing something over.

Here is a case where the distinction occurs for prepositions. You were present when the Great War Hero died, and you remember vividly how the anti-tank grenade landed near his left foot, just fractions of an inch away. In telling the story to your grandchildren, you say to them that the grenade landed right on his left foot, and blew him and the whole surrounding area up. Here too, according to the literal meaning of your sentence, what you have said is false, but since what you wanted to convey was the proximity of the GWH to the explosion, the difference between the grenade landing on his foot or near his foot is unimportant, and so the proposition that is the speaker's meaning of this sentence is true.

Here is a possible example of the distinction as it applies to sentential connectives. You buy a ticket at the movie theater and present it to the ticket taker. The ticket taker takes your ticket, but pauses before letting you in, and says "I'll let you into this movie if you give me five dollars. Stunned, you give him the bribe, and walk in. What happened here? Unbeknownst to you, the ticket taker was actually an out-of-work semanticist, looking to make a bit of extra cash on the side. He would have let you in even if you had not given him the money. Notice that all he did was give a sufficient condition for letting you in: If you give him the bribe, then he lets you in. Since you had already given him your ticket, he would have let you in if you had not given him the bribe. However, the ticket taker knew that you would take him to mean (as is very common) that he would let you in if and only if you gave him the bribe. [Notes: (i) If you wish to answer 'yes' to this question, you need to consider some other types of expressions, such as predicates, adverbs, and quantifiers, and show how a distinction between speaker meaning and semantic meaning could be useful. If, on the other hand, you think that such a distinction will not be useful for non-denoting expressions, then you should show why Kripke's putative counterexamples to your claim, in e.g., section 3b and fn. 22, are faulty. (ii) Many of you were unclear about which distinction Lepore intended you to discuss in this question. In grading, I have assumed that it was your responsibility to interpret the question correctly, and to check with Lepore if you were unsure. I have scored the essays accordingly.]

Kripke suggests that a common move is to suggest that Donnellan's distinction can be explained in terms of the above two kinds of occurrences of expressions. In particular, the suggestion is that attributive uses of definite descriptions are really just de dicto occurrences (i.e, the variable bound by 'the' in the logical form of the sentence is de dicto), and referential uses of definite descriptions are de re (i.e., the variable bound by 'the' is de re). He then offers a number of arguments against this suggestion. First, he argues that neither use of definite descriptions can be considered to be a de dicto occurrence of a definite description. The reason is that when definite descriptions have de dicto occurrences in "intensional" contexts (think of contexts created by 'believes', 'hopes', 'doubts', etc.), the definite description cannot be said to denote its ordinary denotation. 'The king of Canada is happy' is not true, because the definite description does not denote, but nonetheless, when the definite description is taken de dicto 'Bob believes that the king of Canada is happy' might be true. When reporting this belief of Bob's, we are not attempting to denote (either referentially or attributively) the king of Canada, because we know there is not one. Whatever the definite description means, Kripke concludes, it does not pick out an individual in the de dicto interpretation. For if it did pick out such an individual, and our report entailed that Bob believes that he is happy, then our report would be de re, contrary to hypothesis. Second, Kripke argues that the referential use of definite descriptions cannot be identified with a de re use of the definite description, because there are clearly attributive uses of a de re definite description. We can see this in the sentence 'The girl who stole my backpack, whoever she is, is at large, but the police know that she lives near the Edison Diner'. The definite description 'the girl who stole my backpack' is clearly being used attributively, since I don't mean to be picking anyone out in particular, but the second occurrence of 'she' is de re (with respect to the context created by 'know that'. Hence, Kripke claims, referential uses are not always de re uses. Third, although it might seem that Donnellan's distinction between uses of definite descriptions could be theoretically useful insofar as it could replace the Russellian notion of scope, Kripke argues that this is not so. The problem is that Donnellan's distinction says that descriptions are either referential or attributive, and nothing else, and when it comes to issues of scope (roughly, the contexts produced by such items as 'knows that', 'believes that', 'it's not the case that', 'possibly', etc) no twofold distinction can handle all the semantic data. For instance, consider the sentence 'Bill thinks that Mary suspects that Sally's landlord is insane'. Kripke argues that if we can only appeal to Donnellan's distinction, there will only be two possible interpretations of this sentence, one where the definite description is used referentially and one where it is used attributively. However, considering the definite description only, this sentence is three-ways ambiguous. It could mean (a) Sally's landlord is such that Bill thinks that Mary suspects he is insane, (b) Bill thinks that Sally's landlord is such that Mary suspects he is insane, or (c) Bill thinks that Mary suspects that Sally's landlord is insane. The truth of (a) entails that Sally's landlord exists. The truth of (b) entails only that Bill thinks that Sally has a (single) landlord, and that Mary suspects that he (the landlord) is insane. The truth of (c) entails even less: The landlord need not exist, and Bill need not think that he does; he only need think that Mary suspect that Sally has a (single) landlord who is insane. Note that (b), and even more vividly (c), are similar to 'the king of Canada' example above, in that when the definite description is de dicto, entailments regarding the existence of a denotation of the definite description are removed.

4. Kripke finds it hard to adjudicate between Russell and Donnellan because Russell is advancing a semantic theory, but, Kripke argues, it is unclear whether Donnellan's theory is semantic or pragmatic. Kripke offers two arguments for this latter claim. First, he argues that Donnellan is inconsistent about whether the distinction between referential and attributive uses of definite descriptions is in fact based on syntactic or semantic ambiguities of sentences containing definite descriptions. While the general "trend" of Donnellan's paper suggests that the thesis is semantic, at one point, Donnellan says that it is based on "pragmatic ambiguities", i.e., ambiguities concerning which use a speaker is making of a definite description, and which are (at least partially) a function of the speaker's intentions.

The second argument Kripke gives is that Donnellan does not clearly assert that statements containing definite descriptions can have truth conditions which differ from those that Russell would assign to the statement. To see this, consider a case of a wedding, where a middle-aged man wearing a distinctive black suit with a white collar is staggering around the altar reeking of gin. Person A sees this and utters "The priest is drunk." Let us suppose that we know that the man staggering around the altar is not the priest, but a reveler from a costume party next door. Let us also suppose that we know that there is exactly one priest, and he is sober. Donnellan would contradict the Russellian analysis if he said that (in the present circumstances, where A uses 'the priest' referentially) "The priest is drunk" is true iff the reveler is drunk. But instead of saying this, Donnellan remarks that it is hard to identify the statement that A has expressed; the most Donnellan gives us the claim that A has said something true of the reveler. This, however, is not the same as saying that the proposition that the reveler is drunk is true, nor is it the same as saying that the proposition that the priest is drunk is true. Without such a commitment regarding what proposition is true in this case, it is hard to see whether Donnellan's theory ever conflicts with Russell's.

5. The semantic meaning of an utterance of a sentence is the interpretation(s) of the sentence in its context of utterannce that is (are) generated by a semantic theory of the language of that sentence. So, for instance, a semantic theory for English might say that the sentence 'the man to my left is noisy' semantically means (very roughly) that the (unique) man who is to the left of the speaker at the time of utterance is noisy. When we consider the context of an actual use of this sentence, say my use of it last Sunday, the theory tells us that that utterance semantically means that the man who was to my left last Sunday is noisy. (Thus, in a sense there are two forms of semantic meaning, varying according to whether the meaning of the utterance is relativized to a context -- i.e., speaker and time -- of use). The speaker's meaning of an utterance is the meaning that the speaker intended to convey to her audience in virtue of uttering that sentence (at that particular time). Now let's suppose that I made a mistake and the person to whom I intended to ascribe noisiness is in fact a woman. In this case, then, although the semantic meaning of my utterance is determined independently of the mistake I made, the speaker's meaning of my utterance is something (roughly) like: that entity which I intended to denote with this utterance of 'the guy to my left' is noisy. In this case, I had the general intention to use 'the guy to my left' to semantically mean the guy to my left at the time of utterance, and I also had the specific intention to denote a particular person by means of my utterance of this definite description at that time. (Part of this specific intention was (roughly) the belief that the person to whom I wanted to ascribe noisiness was the semantic denotation of that use of 'the guy to my left'; cf. also Kripke's footnote 22). The crucial thing to note here is that speaker's meaning is that speaker's meaning varies according to a given speaker's desires to talk about some entity; semantic meaning does not vary in this way. Given that people are capable of making lots of descriptive mistakes, if speaker's meaning were considered part of the semantic meaning of an utterance, it would turn out that theories about the meanings of our words (i.e., semantic theories) would be virtually vacuous, because almost any singular term would be capable of denoting almost every entity. But this conclusion seems patently false: if you describe me as 'the jerk who broke my window', and then learn that I did not break your window, in normal circumstances you would withdraw this description, and find some other means for describing me, because you would no longer believe that the meaning of the definite description accurately described me. It is for this reason that speaker's meaning is generally considered to be part of pragmatics, a different area of research into language use than semantics.

6. Kripke seems to have five distinct arguments against Donnellan. I will present what I think are valid and charitable reconstructions of all five arguments. I will not assess any of them. The first argument is:

What Kripke wants to do in this argument is this. He wants to show that we would continue to use definite descriptions in the different ways Donnellan discusses even if Russell's analysis was correct. But if we would use descriptions in these ways even if Russell was right, then the fact that we do use descriptions in these ways does not show that Russell is not right. Obviously, the work of this first argument is in (1). On closer examination, we can see that (1) has a structure of, roughly:

The second argument is:

The third argument is:

To create a valid argument, I have introduced a certain kind K of evidence that one has just in case a certain condition is met. If the argument had been constructed only in terms of evidence in general, and had followed Kripke, it looks like it would involve a denial of the antecedent. (This part of Kripke's argument is, roughly, "If Donnellan said that he had the right kind of intuition, then he would have evidence against Russell's thesis; but he doesn't say anything like this.")

The final two arguments are not criticisms of Donnellan's attacks on Russell as such, but are direct attacks on Donnellan, and I include them for the sake of completeness. The fourth argument is:

This argument is at best weak support, since Kripke admits he has no evidence for (2) and one might question (1) as well; maybe the ambiguity is a product of facts about human psychology or evolution, and so no actual natural languages have made the diambiguation.

Finally, the fifth argument is:

[Notes: (i) Very few papers discussed more than one argument, despite the wording of the question. Many papers replaced the articulation of an argument with a cursory survey of some of Kripke's premises, and then a statement of the final conclusion. In many of these there was some interesting argumentation in between, and understanding it is perhaps the main point of the question. (ii) A number of papers addressed no arguments against Kripke's interpretation of Donnellan's theory, but instead considered other arguments, such as the one Kripke gives for seeing Donnellan's distinction as one instance of a general distinction.]

1. The distinction between natural and non-natural meaning (meaning_NN) is designed to distinguish the meaning of 'means' in pairs of sentence such as 'those hoof prints mean that deer have walked by' and 'my utterance of "you were charming" means that you were charming then'. Concerning the former use of meaning (natural meaning) (a) sentences of the form 'x means that p' "entail" (Grice's term) 'p', and (b) the fact that x means that p does not entail that someone meant that p by x. With respect to the second kind of meaning (meaning_NN), (a') sentences of the form 'x means that p do not entail 'p', and (b') the fact that x means that p does entail that someone meant that p by x. A necessary condition for meaning_NN that is not necessary for natural meaning is that meaning_NN requires a number of complex intentions on behalf of the speaker and an audience (for example, that the speaker U intended that (1) audience A produce some response Russell, and (2) A think that U intends (1), and (3) A fulfills (1) on the basis of his fulfillment of (2). In fact, x can naturally mean that p even if there are no intentions whatsoever (provided, of course, that neither the existence of x nor the truth of p entail that intentions exist).

2.

3. The Co-operative Principle is

Grice claims that the CP governs our most central and standard talk exchanges, which are those that have such goals as giving and receiving information, and influencing and being influenced by others. We use the CP to communicate more than we have said (in Grice's sense of 'said') by assuming in various contexts that the CP is governing our conversations. Then, when a conversational contribution is made, and what is said in the contribution seems to conflict with the CP, we generally need to suppose that something other than (or in addition to) what was said was intended by the speaker to be communicated. Often this something more can be narrowed down to a further proposition (or family of propositions) that is different from the one literally expressed by the speaker, and which is the only thing he could be reasonably supposed to have been thinking of if he was observing the CP when he made his contribution. Moreover, we assume that the speaker believes that his audience is able to grasp the need for this further proposition (and he believes that they will believe that he believes it is needed, and that they will believe that he believes that they will believe it is needed), and so this further proposition will be communicated. (For a nice argument form of how conversational implicatures are derived via CP, see Neale's paper, p. 432 in Ludlow; be sure to note fn. 23, as well.)

As an example of how the CP is used in this way, consider the following conversation. During flu season, X sees a sick-looking Y and asks her how she is doing, and Y responds by saying "I am going to make out my will today." Here, the accepted goal of the talk exchange was for X to receive information from Y about Y's physical health. What Y said violates the CP, since it is not the right kind of response, which would be something like "I am feeling very bad", or "I'm feeling quite good". Instead of offering such a response, what Y said gave a report about her intentions to prepare a certain legal document. But, we may suppose, Y meant her utterance to be an appropriate response to X's question. Furthermore, it is likely that Y's utterance respected the CP only if she felt very ill and wanted to communicate that she felt so ill that she felt (almost) as if it would be advisable for her to prepare for her immanent death from illness. It is also reasonable to suppose that Y believed: (i) that X could figure this out, (ii) that X believed that Y has figured this out and intends to communicate something by it, and (iii) that X believed that Y believed that X could figure this out. Thus, the proposition that Y is feeling very bad is expressed because of the assumption that Y's response to X's question was governed by the CP. [Note: The question asks you to explain how the Co-operative Principle functions in an explanation of how speakers are able to communicate more than what is said. It is not an explanation to simply assert the Co-operative Principle and then say that there are conversational implicatures; the point is to show how this happens. In general, such an explanation includes, but is not limited to, stating that (i) there is a common assumption that the speaker is observing CP, (and that everyone knows this of everyone, etc), (ii) what the speaker said (in Grice's sense) violates CP as a contribution to a conversation, (iii) the speaker would have uttered the sentence in question only if she was attempting to implicate something other than what was said, and that (iv) the implicatum preserves CP.]

4. Grice's maxims fall into four categories, Quantity, Quality, Relation, and Manner. In the first category, there are the two maxims

The supermaxim of Quality, "try to make your contribution one that is true", is (roughly) covered by these two specific maxims:

The category of Relation has only the maxim

Finally, the category of Manner has only the single supermaxim

(M6) Be perspicuous,

which entails various more specific maxims, of which Grice lists four:

(a) Avoid obscurity of expression.

(b) Avoid ambiguity.

(c) Be brief (avoid unnecessary prolixity).

(d) Be orderly.

Grice notes four ways in which these maxims might be broken. First, one may simply violate a maxim. Such a case would occur if, e.g., one answered a question without realizing that one had no evidence to support one's answer. Second, one may opt out of obeying a maxim; this would happen if one was deliberately attempting to deceive or conceal (in the latter case, one might be as cagey as possible about giving true answers, as when a defendant is on the witness stand, being cross-examined by the prosecution). Third, one may violate a maxim in order to respect another because of a clash between the two. It often happens that the maxim to avoid obscurity of expression is violated in order to respect the maxims of Quality, as when one is asked by someone untrained in the field to explain a difficult theory. Finally, and most importantly, one may deliberately flout a maxim in order to generate a conversational implicature. Here are some examples of how one might deliberately and openly flout a maxim in order to communicate more than what was said.

(1) Over coffee, A asks B to describe her past relationship with C. B responds by saying only "We dated, it's over, and that's that." Here (M1) is being flouted by B. Instead of offering a full description, she gives only this curt reply, implicating to A that she does not want to discuss the subject.

(2) You are asked by the boss what you thought of a job candidate that had just been interviewed by you and several others, and you reply "I was fairly unimpressed, but Mary thought he was great, and Betina thought he was great, and Bob thought he was excellent." Here one violates (M2) in order to implicate that while you did not think very highly of the candidate, your opinion is in the minority, and perhaps that there are reasons for doubting your opinion.

(3) Looking forward to a relaxing weekend, you are suddenly handed a stack of work that must be completed over the weekend, and you say "Oh, boy, just what I was hoping for." Here you are saying what you know to be false -- this stack of work is just what you had hoped to avoid. It is so obvious that this is not just what you were hoping for that you are able to implicate to your audience that you are unhappy at receiving the work.

(4) I flout (M4) when I have had a bad day in which many things have gone wrong, and as I trudge to the parking lot, my companion asks how my day was. I reply, "Terrible. The way things are going, I'm sure there will be eggplant for dinner tonight." I despise eggplant, and I realize that misfortunes and so forth in the office does not provide evidential support for tonight's dinner menu. I have inadequate evidence for what I said, but my utterance expresses that the misfortunes of the day have left me pessimistic about the coming evening.

(5) Betina and Plumpton are having coffee, and Betina asks, "So, who are you voting for in the upcoming election?" and Plumpton responds with a smile "Arkansas is a truly wonderful state." What Plumpton said is irrelevant to Betina's question, but since Plumpton knows that Betina knows (that he knows that she knows) that Clinton is from Arkansas, this positive statement about Arkansas implicates a favorable attitude regarding Clinton, which is relevant to Betina's question.

(6a) A prestigious scholar gives a talk on a theory of hers, and afterwards takes questions from the audience. A beginning student jumps up and down, waving his hand. When called on, he proceeds to offer an extremely naive and unjustified criticism of the scholar's theory. The scholar responds icily to the student's criticism in a vocabulary that ranges over the entire technical apparatus of the field. The student doesn't understand the response, due to the obscurity of expression, but sits down and remains quiet for the rest of the Q & A session. Here one of the scholar's motives was to implicate by means of the technical vocabulary that the student was out of line in trying to attack a theory of which he had only a superficial understanding.

(6b) Flouting the maxim of avoiding ambiguity is common in sexual or romantic situations, where one wants to suggest (i.e. implicate) a certain message without being so bold as to actually say it. Consider the following situation. X has been eyeing Y in the mall, and they have exchanged several flirtatious smiles. Y goes into a Taco Bell and buys a burrito. As she is putting a lot of hot sauce on the burrito, X comes up to Y and says "You must really like spicy food. Maybe sometime you would like to try the food at the restaurant down the street with me." Y replies "Well, I do like a lot of spice in my life." Y's reply here is clearly ambiguous: It would seems odd for her to utter this sentence if all she meant was that yes, she does enjoy spicy food. The other available meaning, that she is adventurous, seems important here. By making this ambiguous utterance, Y implicates that she is interested in meeting X.

(6c) When asked how her shopping trip with her friend went, X responds, "Well, first we went all through mall A, and then we went all through mall B, and then we went all through mall C, and then we went all through mall D, and then...." Instead of just saying "We did a lot of shopping" or "We went to a lot of malls", X offers this description, which in the appropriate context will implicate that the shopping trip was unpleasantly long or tedious or exhausting.

(6d) One can implicate that the speech of another was disorderly by reporting a conversation with him in an obviously disorderly fashion: "It was interesting talking to him about fish, because he knows a lot about freshwater bass -- well, actually, he knows a lot about fishing rods and reels -- well, bass are caught with bait, and boy oh boy does he know about bait. The best bait for bass is -- by the way, did you know that bass are related to the stegosaurus? They died in the tar pits, and chemically, that tar is the same stuff that goes into your lungs when you smoke, so you shouldn't smoke. But anyway, about bass...." [Note: I have included Grice's four ways of violating a maxim; however, I did not dock any answers that considered only flouting maxims for the purpose of generating conversational implicatures.]

5. Attributive definite descriptions are taken to be true. Utterances containing referentially used definite descriptions (whether proper or not) can be treated as implications of propositions that replace the definite description in question with a proper name or a demonstrative. Cf. Neale, pp. 442-43.

1. The crucial thing you need to show is that once the referent of a demonstrative or indexical is determined, that referent is all that is contributed to the content of the utterance; i.e., no Fregean sense is contributed. Ideally, you would do this by articulating Kaplan's distinction between meaning and content, and using this distinction to argue that while the meaning of a demonstrative or indexical may have a sense, the content of an utterance containing such an expression does not contain such a sense, but only the referent. This appararatus will allow us to account for two important facts about demonstratives and indexicals. These facts are: (i) once the conent of a demonstrative or indexical is determined in an utterance, it acts like a proper name, as we see by considering how such expressions behave in nonextensional contexts (e.g., 'I could have been the king of France', 'John believes that that man is happy'); (ii) like definite descriptions, the meaning of a demonstrative or indexical does not determine a unique content in every world (and in the case of demonstratives and indexicals, this content is not even determined for every time). Just as 'Bob Dole is the President of the USA in 1998' could have been true, so also, the sentence 'Bob Dole is that man' could be true, if the utterer of the sentence has an associated demonstration of Bob Dole with her use of 'that man'. (Of course, if the utterer demonstrates someone else by accident or mistake, the utterance is untrue.)

2. From the class 11 notes: "An expression is directly referential iff its referent, once determined, is taken as fixed for all possible circumstances. We need them to explain the truth of sentences such as 'I could have been the first person on Mars'. Presumably this is true (in some sense of 'could'), and what makes it true is something like: there is a possible world in which I, KJ, am the first person on Mars. [Note: Since this definition is given in the class notes, and since the clause 'once determined' is absolutely crucial for arguing that demonstratives and indexicals directly refer, it was essential to include the clause in your characterization of direct reference.]