Review: Speech Acts

Title: Speech Acts
Author: John R. Searle
Publisher: Cambridge: Cambridge University Press, 1964 (1999)
ISBN: 0-521-09626-X

J.L. Austin wrote a fun, thought-provoking little book called, How to Do Things with Words. In it, he noted that some words had the strange property of doing something simply by being uttered in the correct context: "I promise" creates a promise, and "I declare unto you this day" declares something unto you this day. After much contemplation and some analysis, Austin realized that every statement is an act of some sort—every statement has an illocutionary force. If I say, "Beware the lion," I am performing the speech act of "warning" you. So-called performatives, such as "to promise," are special cases in which the illocutionary force of the word is identical to its general meaning, or locutionary force.

John Searle has received the mantle of speech act evangelist. Unfortunately, his book, Speech Acts, seems unsure of where it wants to take the Austinian theory of speech acts; it is, overall, more an explanation of John Searle's take on the philosophy of language in general than a vehicle to take Austin's theory to its next manifestation. Searle describes how he believes meaning is associated with language. He discusses what it means to reference, and makes a fine argument for distinguishing reference and predication. He outlines his own thoughts on proper names and referencing. Throughout the book he references speech acts. But Speech Acts does not provide a cohesive, coherent theory of why speech acts should be a cornerstone of knowledge expression through language.

One of the major problems with speech act theory, I believe, is that it inappropriately conflates the social and semantic analyses of language. Certain things we say and do have meaning only in a social context constructed through the interaction of individuals. Humans make laws, make promises, pass judgment, attempt to convince others of ideas, and defraud business associates, just to name a few. These are all social actions that depend on a certain set of social rules; "to defraud", for instance, depends on preconstituted social rules of in whom one can place trust in what circumstances, whether an contract is effected when an acceptance is placed in the mailbox or when it is received by the offeror, etc.

Austin correctly noted that certain words, such as "promise," have the unique capability of reaching up beyond the semantics within the sentence and referring to some higher social framework. Mundane non-performative words such as "walk," for instance, describe actions that do not depend on social context. As an example, assume that I ask a girl, in whom I'm interested, on 15 February, "Did your boyfriend buy you flowers for Valentine's Day?" What I'm really asking is, "Do you have a boyfriend?" While it can be said that, in terms of speech acts, the latter question is the illlocutionary force of the former, the semantics of the individual words of the former question—indeed, the semantics of the entire sentence as a whole—do not, absent some social context, mean anything close to what the second question conveys.

Searle's rendition of speech act theory seem to imply that, by recognizing that some statements have an inherent illocutionary force from the social realm, representation and meaning cannot be understood apart from these higher-level social frameworks (such as promising and convincing—as opposed to the social rules for language itself). The most obvious consequence of this error is Searle's faltering attempts in Chapter Eight to derive "ought" from "is"—that is, to derive an evaluative statement (e.g. "John ought to pay his bills") solely from one or more descriptive statements (e.g. "John made an agreement"). Apparently Searle believes he can now do what many philosphers of language felt logically impossible because somehow a social concept of "good" or "bad" has somehow become infused into the semantics of the descriptive statement itself through the wonder of performative speech acts. He isn't clear whether one can extrapolate away the distinction of descriptive statements and evaluative statements altogether.

The following represents my first reactions to Searle's ideas in Speech Acts. In many cases, they are my first reactions to certain philosophy of language concepts in general; many of those, such as my opinion of Searle's "expressibility" argument (17), may be naive oppositions to ideas held widespread throughout the philosophy of language community. In which instances I am disagreeing with Searle, in which instances I am disagreeing with well-established notions, and in which instances I am attacking from a position of ignorance, I currently cannot say. My overall feeling is that I'm not sure I believe that Searle is taking the theory of speech acts in the appropriate direction; more worrisome, I'm not sure in Speech Acts if Searle takes speech acts anywhere significant at all.

Notes

Searle tries to meet the objection that one might not be able to study the analyticity of a statement (whether it is analytic), because there is no in-depth understanding of a formal meaning of "analytic". Searle responds that if ones knows enough about the term to know if a certain formal rule for analyticity is ludicrous (e.g. "Every sentence beginning with "A" is analytic."), that means that we sufficiently understand what "analytic" means. The "borderline" cases of statements that cannot be categorized represent an ambiguity of the sentence or sentence subject, not in the definition of "analytic." (7)

Searle's response is naive. One can understand in coarse terms the gist of a term without knowing its details. Better phrased, one can understand to some extent the purpose of a term—a subset of the outcomes the term is supposed to produce without knowing the complete set of outcomes. Searle alludes to this: "Any criterion for analyticity must be judged by its ability to give certain results." (7). Just because one can determine whether certain sentences are analytic does not mean one fully (or even mostly) understands what "analytic" means.

To give an example, I might convince someone that any number multiplied by zero results in zero, and that person could then determine whether a given product is correct—without ever understanding what it means to multiply. More relevant, many (most?) speakers of Hungarian can identify that one should say "Látom a madarat" and not "Látok a madarat" ("I see the bird") without being able to explain (or even realizing) the formal rule that Hungarian verbs have different conjugations based upon whether the object is definite ("the bird") or indefinite ("a bird"). (See Törkenczy, Hungarian Verbs and Essentials of Grammar, (10).) Searle's assurance here that "analytic" is understood merely because we can correctly identify certain outcomes as being analytic ("We know [a certain test for analyticity give incorrect outcomes] precisely because we know what the word 'analytic' means." (7)) is oversimplistic, and only show that we know how the word "analytic" is used, not that we know what it means.
Studying language without studying speech acts, according to Searle, is "as if baseball were studied only as a formal system of rules and not as a game." (17). But is there a difference?
Searle makes the grand statement of expressibility: "I take it to be an analytic truth about language that whatever can be meant can be said." (17). If one accepts that meaning exists separately from language (Wittgenstein's pain, for instance, exists separately from any grimace or cry), then this statement is surely false. Searle acknowledges that a particular lanaguage may be "not adequate to the task," but for those situations one can "in principle at least enrich the language by introducing new terms or other devices into it." (20) But this doesn't help the situation or give usefulness to "expressibility"—this simply says that even if there are infinite number of meanings that could exist, one can always choose from an unlimited set of potential expressions. Computationally, one might start a counter at zero and, for every new meaning, use the next number from the counter and then increase the counter. Each number would then represent a distinct meaning. Easier yet, for every meaning one can take the next number and increase the counter, meaning that a single meaning could have several representations if used identically several times. But what has this shown? Simply that distinct meanings by definition can be counted. Searle has really only rephrased the definition of distinctness—that if two things are distinct, they can be counted, and that we have an infinite number of counting representations available to us. So how does that help us?
Searle distinguishes among making utterances, making propositions, and performing illocutionary acts. "I am not saying, of course, that these are separate things that speakers do, as it happens, simultaneously, as one might smoke, read, and scratch one's head simultaneously, but rather that in performaning an illocutionary act one characteristically performs propositional acts and utterance acts. ... They are not means to ends; rather, utterance acts stand to propositional and illocutionary acts in the way in which, e.g., making an "X" on a ballot paper stands to voting." (24). In technology, a direct analogy is the OSI model of network communication, in which each "layer" performs a different logical function in the communicational process. A special feature of such a "stack" is that orthogonally a layer may be interchanged with another variation, with the high-level communication continuing agnostic to the change in a layer. For example, the OSI "physical layer" can be interchanged between copper wire or fiber optics, while the communication protocol occurring above the physical layer continues normally. Analogously, an "utterance act" of Searle is a physical movement of vocal cords and air; it may be exchanged with another physical process, writing on paper, and the proposition acts and illocutionary acts remain the same (ignoring, for the moment, illocutionary acts that require other acts along with utterance, such as facial expressions).
Searle distinguishes between a proposition and an assertion. An assertion is a specific type of proposition, having the same referrants, but using them to assert the truthfulness of the proposition. (29). Searle's "proposition" seems to be the same as a "statement" in RDF.
At first, Searle's distinction of regulative and constitutive rules (34) hearken back to Hart's distinction between primary and secondary rules, respectively. Searle's distinction, however, is hard to maintain. He says that, "[W]here the rule (or system of rules) is constitutive, behavior which is in accordance with the rule can receive specifications or descriptions which it could not receive if the rule or rules did not exist." (35). In other words, constitutive rules make "new forms of behavior," which can get a new label. But can't a regulative rule become a constitutive rule if we give the action a new label? Searle says no: "[A]ppraisals [of regulative rules] are not specifications or descriptions [that would make them constitutive rules.]" (36). But is this really true? Is a rule that says, in baseball, the players change roles after either three strikes, four balls, or a valid hit regulative? If so, what if we call changing because of strikes "striking out?" We haven't created a "new form of behavior," but we've created a constitutive rule of baseball just as much as "castling" is a constitutive rule of chess—just by adding a label to some set of actions. Is there really a distinction?
Searle also tries to claim that "a slight change in a fringe rule," whatever that is, does not make football "a different game" (34), but this is a huge assumption, and ignores an entire discussion on the philosophy of identity.
While trying to explain the difference between constitutive and regulative rules, Searle asks what the difference is between fishing and making a promise. (37). This distinction doesn't help the clarification at all. Searle doesn't clarify that fishing and making a promise both involve what he calls constitutive rules. The distinction is that, in the case of making a promise, the constitutive rules of what we call "making a promise" are social rules made by and imposed by a community. The activity we call "fishing" is made up of constitutive rules that are imposed by nature. The only difference between fishing and asking a grocer for a fish is that in the former nature determines whether we get a fish according to its rules, while in the latter society (not just the grocer, but the entire commerce system, including the courts, for example, if a contract was made for a fish) determines whether we get a fish. Are those rules constitutive or regulative? It's hard to tell. Surely the rules of what we call "shopping" are more lenient than the rules for "playing chess," but don't those rules "constitute" what it is to shop? If one objects to this ("No, this is just a name we give to something that is done separate from the naming."), then what if we instead call "shopping" "making a commercial transaction? Does capitalism have no constitutive rules, and if so, what could be more central to capitalism than making a transaction?
In asking if all rules have penalties, and if there is a penalty for the rule that baseball is played with nine people on each side (41), he skips over all the positivist work of Hart, analyzing primary and secondary rules in the law.
Searle makes a good point: there are some rules in language that an agent may not even know about, yet when he/she sees a new case, he/she just "knows what to do." (42). This sort of analysis could be a subject in itself, even separate from language—what about color coordination?
Searle correctly points out that there is a difference between someone meaning something by an utterance, and the utterance to mean something, or to "have meaning". (42). Searle here examines meaning from the former viewpoint, ignoring how a sentence might mean something. He agrees with Paul Grice that meaning something has something to do with the speaker's intention of creating some sort of effect on the recipient. (43). "In the performance of an illocutionary act i the literal utterance of a sentence, the speaker intends to produce a certain effect by means of getting the hearer to recognize his intention to produce that effect; and furthermore, if he is using words literally, he intends this recognition to be achieved in virtue of the fact that the rules for using the expressions he utters associate the expression with the production of that effect. It is this combination of elements which we shall need to express in our analysis of the illocutionary act." (45). "Uttering 'Hello' and meaning it is a matter of (a) intending to get the hearer to recognize that he is being greeted, (b) intending to get him to recognize that he is being greeted by means of getting him to recognize one's intention to greet him, (c) intending to get him to recognize one's intention to greet him in virtue of his knowledge of the meaning of the sentence 'Hello'." (49). My initial impression is that this characterization, following Wittgenstein, inappropriately merges the independent concept of meaning with the concept of communication, only the latter of which has anything to do with any intention relating to the recipient of a message. For one "to mean something" may very well have to do with intention—the intention of creating a representation (whether written or verbal) of the meaning. There may be no intention to communicate that meaning to a second party. For an artist to paint with vivid red may be an instance of an artist "meaning" he or she is angry—the artist may have no intent to impart that meaning to a recipient. If the artist intends the recipient to understand this meaning by way of the color, the artist is communicating. If the medium utilizes the particular written and spoken rules described by Searle as the medium for communication (as opposed to the color of paint), the artist is communicating using language. This description correctly distinguishes among meaning, communication, and language; Searles seems to think that there is no meaning without communication using language.
Searle lays out the preconditions for an utterance to be a promise, including that the hearer would prefer the speaker's doing the thing promised to the speaker's not doing the thing promised, and that the speaker believes the hearer would so prefer (58). Searle says that without this condition, the utterance would be a threat. This seems incorrect, again relying too much on the interaction with the hearer in defining a particular speech act. Consider the unilateral promise that, "if you ever run out of money, I'll be happy to loan you some." At the time of the utterance, the hearer, a proud person with an independent streak, would prefer that the speaker never provide monetary assistance. Is this utterance not then a promise? What if, the very next week, the hearer is robbed of everything and, in a change or heart, relies on the speaker's utterance and buys a new car to transport the hearer to work. Has the hearer relied on something that wasn't a promise? Does the utterance become a promise only when the hearer at a later date decides that he/she prefers the object of the speaker's promise? I would propose that the utterance was always a promise, regardless of the hearer's desire.

Searle later decides that this whole condition of the hearer's preference is pretty complicated and proposes that "a more elegant and exact formulation of this condition would probably require the introduction of technical terminology of the welfare economics sort." (59). Perhaps Searle means that the preference of the hearer should be determined, not by the actual subjective desire of the hearer, but by some objective utilitarian determination of what the should desire, or what, as the law might say, a reasonable person in the hearer's situation would desire. This does not help his case, because suddenly whether a promise was made depends on whether one would reasonably desire the outcome—preventing a valid promise from ever coming about for conditions that a minority of the population would subjectively want. Under this theory, in other words, if I'm the only one in the world that wants a car painted in some hideous color, no one would be able to successfully promise to paint it in that color for me.
Searle contrasts his (and, he says, Austin's and Wittgenstein's) "institutional" theory of communication, with what he calls the stimulous-response account of meaning, which collapses the illocutionary and perlocutionary aspects of communication. (71). It seems that, in defining the speech act of promising, Searle has done partly the same thing (see above) by makeing promises rely on the desires of the hearer.
In trying to explain reference, Searle notes the difference in uses of a word in referring to the referrant and to the word itself; his examples are, "Socrates was a philosopher," and "'Socrates' has eight letters" (73). He refers to a theory "by philosophers and logicians" that there are two words being used here, in the first of which "Socrates" is a name for the philosopher, and in the second of which "Socrates" is the name of the philosopher's name.

Searle laughingly dismisses this theory in a way that, while cute, is nonetheless misleading and slighly disingenuous. If the theory were true, states Searle, the correct name for "Socrates" is ""Socrates,"" and the last name just mentioned is """Socrates,""" and so on. Such an ad absurdum argument tries to make its point by cluttering up a sentence with unfamiliar syntax—of course a reader will not like all those quotation marks, so Searle must be right.

But take a step back and think about what is really happening here. The syntactical token "Socrates" can be a referrer or a referrant—that is, it can refer to Socrates, the man, or it can itself be analyzed as a word, a collection of characters. Because the same syntactical token is used in both cases, the common way to differentiate between the man and the word is to use quotation marks in the latter case, and when doing so "Socrates" with quotation marks is, in a sense, the name of the word—in any case, it puts the word "Socrates" in a different role: that of referrant rather than referrer. But what is the need for more quotation marks? What other option is there between man and word? Certainly, as is clear from the context of Searle's example, the second use of the word "Socrates" refers to the first word "Socrates," making the word itself a referrant, and its referrer a name of a name of a name of a man.

But, using Searle's own argument (76), there is no such convention to use multiple quotation marks in such an instance—the English convention is to use context when more than one level of indirection is being used. Searle then tries to use an unconventional syntax for indicating multiple levels of indirection to prove that there are not multiple levels of indirection by saying, in effect, "See, the syntax looks silly." But that doesn't mean the multiples of indirection are not present—it simply means that Searle is inventing a new syntax to represent the multiple levels of indirection.

To see illustrations of the unfairness Searle's argument, one need look no further than Searle's own statements. Only a few pages later (89), he uses the same syntactic construct he derided: he uses "Jones" with double sets of quotation marks: "The person referred to by my interlocutor as "Jones"". Indeed this example uses double sets of quotations marks to denote the same semantics as Searle discounted: multiple levels of indirection. Here ""Jones"" (ignoring the other characters in the sentence refers, not to Jones, nor to the name of Jones, but to one person's reference to the name of Jones.
Searle claims that a speaker cannot intend to refer to a particular object unless the speaker has the ability to provide an identifying description of the referent, should such a description be requested. (87). Again, Searle like Wittgenstein seems be confusing meaning with expression. What if I say, "that thing that I'm thinking of right now," but I can't think of the word—am I referring to something? I believe there is some underlying semantic reference going on, even if I can't describe the object—I know what I'm talking about. Certain people after brain damage may be able to write a description but not speak it; even those without brain damage have no doubt had an experience in which they could see an object but not be able to name it. Is reference possible in such a situation? Searle would say no. If I am in a field in the middle of the night with my arm around a tree and I say, "the tree by which I'm standing is very tall," have I successfully referred to the tree, even if I have no further description available? How is this different from the previous examples I've provided? And if this is reference through description, isn't it a very circular description ("the one which I'm standing by"), allowing any attempted reference to automatically be a true reference ("the one to which I'm referring")?
Searle tries to discount Strawson's explanation that a predicate "identifies" a universal, while a subject "refers to" a particular, by saying that this can be switched around so that a subject only "identifies" a universal as well. As examples, he gives as equivalent, "The thing which is a rose is red," and "The thing which is red is a rose," finding the meaning of these sentences equivalent, thereby causing problems for the idea that a subject refers and a predicate identifies. (117). But does this taken into account Wittgenstein's assertion that "is" in these propositions are really different words, with the former asserting a relationship between a particular and a universal and the latter asserting an equivalence between two particulars? Searle may have a point, but this may not be a good example to prove it.
Searle makes a distinction between a "property name," which he equates with a universal, and its corresponding "general term," although "general term" is not defined. He seems to imply that "general terms" are used only in predicate expressions (119). He then claims that property names depend on general terms, but not vice-versa. That is, a language could contain an idea of "kind", without necessarily having an idea of "kindness," but not the other way around—any language that talks about "kindness" will have to embrace the concept of "kind."

Searle supports his idea by noting that syntactically English derives his "propery names" from his "general terms"—"kindness" is derived from "kind." He goes as far as to say that, "[O]ne cannot comprehend the notion of a given universal without first understanding the corresponding predicate expression," that to understand "any given universal" one must first understand "how to use the general term from which the universal is derived." (120-121). To Searle, a universal is just the name of a property.
Searle thinks that "the tendency to construe predication as a kind of, or analogous to, reference is one of the most persistent mistakes in the history of Western philosophy." (122).
Searle gets down to his point: a proposition contains a reference and a predicate, and the whole thing is a speech act—the predication is not a separate act. However, the proposition's reference contains a separate speech act (123)—so does that speech act contain a predicate? Searle must mean that there can be speech acts without predicates, else the speech sub-act of referring within a proposition would have a predicate that does not refer.
Searle tries to disprove J.O. Urmson's contention that it is impossible to derive evaluative statements from descriptive statements. In his arguments, Searle makes good points—but they may not have anything to do with what Urmson was talking about in at least one instance. Searle gives a definition of a "valid deductive argument" as one in which the premises of the argument entail the conclusion of the argument. He then says that meeting that description entails that the evaluation that the argument is valid (133). The problem is, Urmson (from the quotes Searle provides) seems to have a different definition of "valid," one that relies on opinion. Urmson clearly states, "to call an argument valid is not merely to classify it logically, as when we say it is a syllogism or modus ponens; it is at least in part to evaluate or appraise it; it is to signify approval of it." (133). When Searle responds, "It is not a matter of opinion that the argument 'all men are mortal and Socrates is a man; therefore, Socrates is mortal' is a valid deductive argument" (134), all he thereby succeeds in doing is asserting that he uses a different definition of "valid" than does Urmson—not that, if both were to use the same definition of "valid" so as to include opinions, that one could entaiol such an evaluative label as "valid" from any description of an argument. Searle's later analysis of evaluative labels that follow from definitional descriptions seem valid, however, in pointing out that a difference in illocutionary force does not preclude entailment of underlying propositions, apparently in contradiction to Urmson. (135-136).
Searle explains that many times an assertion may seem out of place because the underlying proposition is obvious, such as in "I remember my name." He says that some philosophers (notably Austin and Wittgenstein) believed this phenomenon to reflect something about the meaning of the words (e.g. "remember") themselves, rather than to reflect something about when it is appropriate to make assertions. This is no doubt true, but when Searle applies this to the sentence, "He knows that he feels pain," he is missing Wittgenstein's point: that there may be some connection between the "pain" that presupposes "know", and saying, "He does not know that he feels pain" may not be simply strange, but nonsensical, if having pain requires some sort of awareness. That case would be completely different from the counter-example used by Searle that, "He has five fingers on his left hand." (143). (At least Searle uses the occasion to introduce one of the rare occasions of humor in the book: "[W]e would not say "He has five fingers on his left hand" unless there is some abmormal feature of the situation, e.g., if he has six fingers on his right hand, or if we wish to free him of suspicion of being the four fingered left-handed murderer.")
Searle makes the distinction between "meaning" and "use" in explaining how an analytical description can be different from an evaluative conclusion. For example, says Searle, a man might provide criteria for a good car, and then describe his own car using terms that match the description of a good car. "[H]aving such a commitment [that the car is in fact a good car] is not at all the same as actually having asserted that it is a good car." (155). As I point out later, his description of the car does not in any way entail that the car is good—it only allows one to derive "good car" from the provided criteria or transformation function. (Searle also mentions the "People v. Goldberg et al. unpublished trial court case, Berkely Superior Court, California, 1965" in which "if Cx then Ox" may be analytic, but asserting the proposition "Cx" may be different from asserting "Ox". "For asserting 'Ox' in public you can go to jail." (155).
"It does not matter much whether we say of the assertion 'The king of France is bald' that it is false or pointless or what not, as long as we understand how it goes wrong." (158-159).
"'He is a Napoleon' means 'He is like Napoleon in many respects'…." (163). Yes, but only indirectly. "He is a Napoleon" means that "What 'Napoleon' has come to connote is also descriptive of him." Two things follow from this distinction: (1) that the real Napolean may not necessarily have been a "Napoleon" (whether he was is contingent upon what connotation "Napoleon" has acquired, and (2) that "Napoleon" at one time did not not connote, only denote.
"We use proper names in existential propositions, e.g. 'there is such a place as Africa', 'Cerberus does not exist'. Here proper names cannot be said to refer, for no such subject of an existential statement can refer. If it did, the precondition of its having a truth value would guarantee its truth, if it were in the affirmative, and its falsity, if it were in the negative. (This is just another way of saying that 'exists' is not a predicate.')" (165). This doesn't ring true. If dereferencing is a separate step from the existential assertion, why can't this work? "There is such a place as that large continent you've heard about from which all human life has sprung," or whatever the dereferencer happens to consider "Africa" to refer to.
Searle's explanations of the connotative power of proper names (170) offers several important insights. However, several questions remain. Do proper names refer through their related descriptions (the disjunction of true identifying descriptions, according to Searle (169)), or do proper names connote through their referants? Put another way, which happens first: connotation or denotation? Searle seems to be saying that proper names are substitutions for identifying descriptions of a speaker, and if the disjunction of corresponding true descriptions of two speakers are identical, then the proper names have the same referant—first connotation, then denotation. That may be true for some descriptions, but not for others. For example, it may be that a proper name denotes its referant through the disjunction of common connotations, but connotate other not-commonly-held descriptions through the referent, after the reference has been established.
Searle find denotative power of proper names through the disjunction of shared true descriptions of the referent. ("[I]f none of the identifying descriptions belived to be true of some object by the users of the name of that object proved to be true of some independently located object, then that object could not be identical with the bearer of the name." (169).) But do such shared descriptions have to be analytically true? If Alice says that Jesus was born in Nazareth, but Bob says that Jesus was born in Bethlehem, do both instances of "Jesus" have the same referent? What if Jesus was not born in either location—would Alice and Bob still be speaking of the same person? Perhaps these descriptions are not the "disjunction" of common analytically true descriptions Searle requires for a proper name to denote. Perhaps Searle means such descriptions as, "The individual worshipped by the Christian religion," which would be commonly true regardless of whether Jesus even existed. But on this level, does "Jesus" become a universal, like the color red, separated from the need for any inherent attributes of the referant?
Searle claims that a negative existential statement using a proper name, such as "Aristotle never existed," asserts that "a sufficient, but so far unspecified, number of the descriptive backings of 'Aristotle' are false." (171). This is incorrect—the statement makes an assertion about the existence of the referent of "Aristotle," and such a statement is not connotative. Take, for instance, the following story:

Sally was a unicorn with a pure white mane, a happy disposition, and one feature unique among all unicorns: she had pink eyelids. Unfortunately, Sally never existed.

The existential statement, "Sally never existed" makes no assertion about the descriptive backings of "Sally." In fact, after this existential statement, all of the descriptive backings of "Sally" remain true.
Searle thinks that identity statements using proper names is just another way of stating "that the descriptive backing of both names is true of the same object." (171). But this is not quite correct—such statements are statements of identity, not description, that relate referents after denotation has occurred, even if each denotation occurs through some disjunction of connotations, as Searle proposes. Take, for instance, Alice saying to Bob, "Jesus is really Satan." It is not clear if Alice is asserting the Jesus has a red tail and a pitchfork, or if Alice is asserting that Satan is part of the Christian Holy Trinity. What is clear, however, is that Alice is asserting that, the referent of "Jesus" is the same referent of "Satan," whatever descriptive backing that referent might originally be determined to have.
Combining the discussion of existential statements using proper names and identity statements using proper names gives more evidence of how proper names act as universals. Consider the statement, "Jesus was really a woman named Beth who wandered along the Mediterranean Sea collecting sea shells circa 200 A.D." This may be communicated, following Searle's theory, because "Jesus" has a common descriptive backing between speaking and listener of "the individual Christians consider to be God incarnate." Consider, however, the statement, "Jesus was only human and was not divine." This can be communicated if speaker and listener have a common but distinct descriptive backing—namely, that Jesus was a male who lived in the middle east around 0 A.D. The same proper name can therefore, in different contexts, survive with different existential backings. When a minister refers to "Jesus" in a church service, which common descriptive backings are being used? Does it matter? Proper names, therefore, seem to be merely a way to refer to sets of descriptions. These sets of descriptions can be very dynamic—may they even allow dereferencing without resolution of the actual backing descriptions? See, for example, Wittgenstein's explanation of a beetle in a box.
"[W]e never get referring completely isolated from predication for to do so would be to violate the principle of identification, without conformity to which we cannot refer at all." (174).
Searle seeks to disprove the long-held philosphical idea that one cannot derive an evaluative statement soley from descriptive statements (i.e. one cannot derive "ought" from "is"), and offers the following example:
- 1. John uttered the words "I hereby promise to pay you, Smith, five dollars".
- 2. Jones promised to pay Smith five dollars.
- 3. Jones placed himself under (undertook) an obligation to pay Smith five dollars.
- 4. Jones is under an obligation to pay Smith five dollars.
- 5. Jones ought to pay Smith five dollars. (177).
Searle spends several pages making no interesting point whatsoever, and the only reason his line of argument is even presentable is because it is clouded in vaguaries. For example, Searle states the to get from step 2 (promise) to step 3 (obligation), "[P]romising is, by definition, an act of placing oneself under an obligation. … Therefore, I think 2 entails 3 straight off …." (178). Proud of himself, Searle adds the footnote: "At this point we have already derived an 'evaluative' statement from 'descriptive' statements since 'obligation' is an 'evaluative' word." (179 fn 1).

But whatever does Searle mean by "obligation?" Does he mean that Jones feels as if he/she should pay Smith five dollars? Does he mean that Jones is legally obligated, in that a court of law in that jurisdiction would compel Jones to pay Smith five dollars because of the promise? The first, a feeling of compulsion on the part of Jones is not evaluative—it merely describes the psychological state that Jones is left with, based upon his experience of society's treatment of promises. The second, a legal compulsion to pay, is not evaluative in the least—it merely makes a prediction of the outcome of a court case based upon a set of conditions. (Note that the second meaning of "obligation" does not even naturally follow from making a promise. US courts, for example, require consideration before recognizing a legal obligation based upon a promise. Donative gifts, therefore, normally impose no legal obligation.)

If Searle intends to say that some sort of moral obligation (or one of several jurisprudential definitions of "law" that say, for example, that "the law says Jones should pay", separate from any prediction of the outcome of a court case), then it is not evident in the least that 2 entails 3. The only solution is that Searle's definition of "promise" contains the notion of "obligation" (which Searle seems to admit), in which case Searle has proved nothing. (More on this later.)

Furthermore, if "obligation" is evaluative in the sense of "ought", we have already moved to step 5 and have to need for the intermediate steps. In order to progress to 5 here, however, let's assume that 3 is not evaluative and that it implies merely a feeling of compulsion on the part of Jones, for instance. Searle thinks that "…there is here the tautology that if one is under an obligation to do something, then, as regards that obligation, one ought to do what one is under an obligation to do." (180). This is a huge assumption, and an erroneous one.

What if ten-year Jones old promises to pay Smith five dollars? Legally, Jones is not old enough to appreciate the consequences of promising, and the law will not impose on infant Jones an obligation to pay. Ought Jones to pay? What if Jones is 21 years old, but promises to pay Smith five dollars for a stolen television? Jones may be obligated to pay, but should he/she pay? What if Jones promises a slave trader in a third-world country five dollars to purchase and free a slave. After the slave is freed, is Jones obligated to make the slave trader? One could make the argument that, true to his/her word, Jones should make good on his promise. But ought Jones to pay the slave trader five dollars, knowing that this, in classical economic terms, raises the demand for slaves and thus encourages other slave traders to engage in the pracice of slave trading.

There is obviously many opposing considerations in determining "ought" from "is", but Searle irritatedly brushes aside such arguments by ignoring "conflicting and overriding obligations at the same time…." (180). Searle wishes to consider merely one primary obligation from a single promise. Searle would probably add that he made it quite clear that this promise is that made under conditions "C", "the necessary and sufficient conditions for the utterance of the words … to constitute the successful and non-defective performance of the act of promising. This includes the input and output conditions, the various intentions and beliefs of the speaker, and so on…" (178). And that's exactly why Searle's entire argument proves nothing, for "C" in this case is a transformation function, the set of conditions necessary for a statement to be a promise. "C" defines the social conditions, the legal conditions, an all other necessary conditions for promise to entail "obligation" and to entail "ought."

It's no news that one can define a function "C" for each descriptive statement P and a resulting evalutative statement Q. If I say, "you ought to buy a pizza (Q) when I plant blue flowers in my garden (P)", I've defined a transformation function for P and Q—I've defined the conditions under which P entails Q. This does not mean that a descriptive proposition entails an evaluative proposition, it simply means that a semantic-less transformation has taken place. I can say that "if foo then bar under conditions bleh," and one can make transformations from foo to bar with no knowledge of meaning—but one must have the transformation function bleh. Without bleh, foo does not entail bar.

Here Searle has simply defined a transformation function C that, from the moment statment 1 is given, provides statement 5—not from any intrinsic qualities of the descriptive statement 1 or even 2, but because Searle said, "let's assume C." As Searle defined C to provide "ought" from "is," there is no intrinsic entailment going on—merely transformation. For all those words, Searle essentially proved, "If we assume that this descriptive statement entails this evaluative statement, then if I give you this descriptive statement you'll find that it entails this descriptive statement." Magic!

It seems to be that "evaluative statement" should be definied as a statement that is not entailable from any descriptive statement in the absence of a transformation function. For example, is it good to raise taxes? One tranformation function in this case might be "a policy decision is good if, after five years, the average household income has been raised." Without such a transformation function, "to raise taxes" cannot entail good or bad—evaluative cannot be derived from descriptive. Provided a transformation function, a descriptive input can yield an evaluative result, but the only entailment taking place is in the processing of the transformation function, not in the semantic relationship between the result and the input. Entailment has to do with semantic relationships between input and output. All that's happening here is transformation.

Searle seems to eventually recognize something like the transformation function I'm speaking of, when he says, "No set of brute fact statements can entail an institutional fact statement without the addition of at least one constitutive rule." (185). (I think I gave a better explanation, though.) Throughout the chapter, though, he seems not to recognize the distinction between entailment and transformation. It is true that, given "systems of constitutive rules", one can "derive" an evaluative statement from descriptive statements (186). I would propose that entailment is only a subset of all transformation functions Derivation implies some sort of transformation, but the transformation being done here is simply following the "constitutive rule[s]," not logical qualities contained in the statement itself.

Let's look at the "constitutive rule[s]" more closely. Such a set of rules—a transformation function—comes in two parts: a definition section, and an evaluative section. This is all directly analogous to law. For example, we might look at the fictitional Earthling Code of Obligations (an analogue of the ALI's Restatement of Contracts) which describes how Earthlings make promises. It would contain a definition section: "when one makes a future prediction of one's actions, intending to undertake those actions, and attempting to communicate this prediction to a second party, shall be referred to as a promise." Another section of this Code, the evaluation section, defines the actual transformation functions which use the definitions: "one who makes a promise ought to perform the promised action." One cannot derive the results of the evaluative section through entailment—the purpose of the evaluative section is to provide a transformation function from facts (identified by the definitions) to evaluative conclusions."

Searle confuses the distinction between definition and evaluation.

When we assert "He made a promise" we commit ourself to the proposition that he undertook an obligation. In exactly the same way, when we use the word "triangle" we commit ourself to its logical properties. So that when we say, e.g. "X is a triangle" we commit ourselves to the proposition that X has three sides. And the fact that the commitment in the first case involves the notion of obligation shows that we are able to derive from it an 'evaluative' conclusion, but it does not show that there is anything subjective (matter of opinion, not a matter of fact, or a matter of moral decision) in the statement "He made a promise", any more than the fact that the statement "X is a triangle" has logical consequences, shows that there is a moral decision involved in the committed use of the word "triangle". (194)

This reasoning is completely incorrect. Saying "X is a triangle" is merely to say that X is something that meets the definition provided for triangle. "X has three sides" logically follows from the definition of triangle, through the syllogism, "anything that has three sides shall be called a triangle; X is a triangle; therefore X has three sides." On the other hand, "Smith should pay five dollars" does not logically follow from the fact that we have, using the definition of promise, identified a particular utterance that comprises a promise. It requires the application of the separate evaluative function, "one ought to do what one promises to do." For triangles, similar evaluative statements might be, "X is more beautiful than Y [a circle]," or "one should not place a X above Y [a circle] in the painting", derived from evaluative functions, "triangles are more beautiful than circles" and "circles are the holiest symbols, and should always appear above all other shapes."

Still confused, Searle mumbles something about whether one can derive moral statements from descriptions. The answer is straightforward: yes, given the appropriate transformation function, supplied by society, religion, personal beliefs, a utilitarian formula, or whatever. Descriptive statements do not and cannot entail moral conclusions but, given an appropriate transformation function (and there are many to choose from), one can derive moral conclusions. Morality is not inherent in facts, just as ethics are not inherent in economics.