publisher colophon

The semantic issues that Saul Kripke addressed in Naming and Necessity overlap with those addressed by Michel Foucault in “What Is an Author?” The present essay examines this area of overlap. It shows that Foucault needs to take certain assumptions from Kripke’s theory of naming in order to secure his argument for treating authorial names as a special case. When placed on these Kripkean foundations, Foucault’s theory is a plausible one, and avoids the metaphysically peculiar commitments that are sometimes thought to be essential to it.


The semantic issues that Saul Kripke addressed in Naming and Necessity1 overlap substantially with those that were addressed by Michel Foucault in “What Is an Author?”2 The present essay examines their area of overlap, with a view to showing that each of these works affords a perspective on the other, from which facets that are usually obscure can be brought into view. It shows that Foucault needs to take some assumptions from Kripke’s theory of naming in order to secure one of his arguments for treating authorial names as special. It then shows that, once it has been placed on these Kripkean foundations, Foucault’s position avoids the metaphysically peculiar commitments that are sometimes thought to be essential to it.

These two works appeared at much the same time, but in rather different philosophical contexts. Naming and Necessity was first presented as a series of lectures, delivered at Princeton University in January 1970. [End Page 383] “What Is an Author?” was first presented in the United States that same year, although it had been given to an audience in France (and had been published in French) the year before. Their two authors belonged to quite different philosophical traditions. Kripke was then best known for his proofs of certain formal results pertaining to the completeness of modal logic.3 Foucault was best known for his historical work, tracing developments in our conceptualization of mental illness.4

Their two philosophical projects were, as one therefore might expect, quite distinct. Kripke’s lectures were ultimately concerned with questions about metaphysical possibility. They were concerned, in particular, with the question of whether it is possible for a mental event (such as a feeling of pain) to be identical with a physical event (such as a firing of some nerve cells). Foucault’s work was ultimately concerned with questions about interpretation. It was concerned, in particular, with the attempt to locate a source for some of the norms governing our interpretations of literature (where “literature” is to be construed broadly, as including the works of Freud and Marx, alongside those of Shakespeare and Stendhal).

Despite these fundamental differences of aim and of approach, the works of Kripke and Foucault both depended on points relating to the semantic content of proper names. Kripke’s concern with naming was announced in his title. Foucault’s was made explicit when he wrote:

The author’s name is a proper name, and therefore it raises the problems common to all proper names. . . . Obviously, one cannot turn a proper name into a pure and simple reference. It has other than indicative functions: more than an indication, a gesture, a finger pointed at someone, it is the equivalent of a description. When one says “Aristotle,” one employs a word that is the equivalent of one, or a series of, definite descriptions, such as “the author of the Analytics,” “the founder of ontology,” and so forth.

(Foucault, pp. 145–46)

In quotations such as this we see not only that Foucault shared Kripke’s concern with naming but that he, like Kripke, was concerned with the question of whether proper names are semantically equivalent to definite descriptions (such as “author of the Analytics,” or “the founder of ontology”). The answers that Foucault and Kripke gave to this question seem, on the face of it, to be irreconcilable.

Kripke’s answer, given in the first and second lectures of Naming and Necessity, was that the semantic content of a proper name is what [End Page 384] Foucault said it obviously cannot be: a “pure and simple reference.” This theory of naming—according to which “when you use a name its semantic function consists simply in referring to an object, and nothing else”5—originated in the 1843 work of John Stuart Mill.6 It contrasts with the theory given by John Searle in his 1969 book Speech Acts.7 Searle’s is the only theory that Foucault cited, and is apparently the theory that Foucault adopted in the quotation above. But it is one of the theories that Kripke refuted in his work of the following year. The arguments that Kripke gave were intended to strike at the heart of any theory according to which a proper name “is the equivalent of one, or a series of, definite descriptions.” They were intended to show “not simply that there’s some technical error here or some mistake there, but that the whole picture given by this theory of how reference is determined seems to be wrong from the fundamentals” (Kripke, p. 93).

Kripke’s arguments are, by now, very widely accepted. They are also easy to state: if “Aristotle” were indeed the semantic equivalent of such descriptions as Foucault mentioned, then the truth of some corresponding sentence—such as “Aristotle was the author of the Analytics or the founder of ontology”—would be a necessary truth, since that sentence would then be the equivalent of a tautology; but, since Aristotle might never have gone into philosophy at all, such sentences express truths that are instead contingent. It follows that the name “Aristotle” cannot be the equivalent of any such description, nor even of some disjunctive series of them (Kripke, pp. 30, 57). The theory that Foucault appeared to endorse must therefore be mistaken.

Foucault offered no countervailing arguments for the equivalence of proper names and definite descriptions. He took the theory that postulates such an equivalence to be obvious (as the quotation above indicates). There is no reason to suppose that he anticipated the metasemantic considerations that Kripke’s work would introduce. It therefore seems rather as if Kripke got the better of Foucault here, by introducing a novel form of argument and so showing that what had seemed obvious to Foucault (and to Searle before him) was indeed “wrong from the fundamentals.”

I suggest that this reading captures only a part of the truth about Foucault’s position, and that the part it captures is not the most interesting part. [End Page 385]


In his remarks concerning the name of Aristotle, quoted above, Foucault does seem to be adopting the Searlean theory of names, which Kripke showed to be mistaken. In his subsequent remarks, concerning the name of Shakespeare, Foucault even seems to be considering the line of thought that would enable Kripke to demonstrate that theory’s mistake. And yet Foucault takes this line of thought to be pointing in a quite different direction.

When considering the case of Shakespeare, Foucault tells us that “if we proved that Shakespeare did not write those sonnets which pass for his, that would constitute a significant change, and affect the manner in which the author’s name functions” (Foucault, p. 146). Here, Foucault seems to be contradicting a central tenet of Kripke’s argument, more or less exactly. Kripke depends on the idea that authorial names are not caused to change their significance when new information becomes available about the works that those authors did not pen. He illustrates this with several examples, the most discussed of which is “Gödel.”

For this part of his argument, Kripke asks us to imagine that Gödel stole his proofs from a mathematician named Schmidt. He points out that, if this were the case, then the man we call “Gödel” would have turned out to be a thief. “Gödel” would not have been Schmidt’s name all along. Nor would that name shift so that, having previously denoted Gödel, it was now transferred to Schmidt (Kripke, pp. 83–87). From here, Kripke builds an argument that is fatal to any theory of names according to which the meaning of a name is equivalent to, or is somehow fixed via, the meaning of a description. This argument goes beyond Kripke’s earlier point about the possibility of Aristotle not going into philosophy. If the meaning of “Gödel” were fixed via a description then it would, presumably, be a description that characterizes that name’s bearer as the prover of Gödel’s theorems. But if that description fixed the name’s reference then (whether or not the semantic content of this description figures in the content that the name expresses), the name “Gödel” would refer to Schmidt in the scenario that Kripke has asked us to imagine. If the meaning of “Gödel” were given by some other description then we would need to imagine some other scenario, but the point could still be made.

By means of this argument—based as it is on the fixed reference of an authorial name—Kripke was able to show that no version of Searle’s theory can be made to work, whether we take it to be a theory of a [End Page 386] name’s semantic content or a theory of the way in which a name comes to be attached to its bearer. This gave Kripke his case for taking Searle’s “whole picture” to be “wrong from the fundamentals.”

The point that Kripke makes with the example of Gödel might seem to be exactly the point that Foucault misses when he suggests that the name of Shakespeare would undergo a “significant change” if it were shown “that Shakespeare did not write those sonnets which pass for his.” This would, however, be too quick. We can begin to see that something more interesting might be going on by realizing that Foucault is here taking Shakespeare’s name to be a special case. When he instead considers a more commonplace name, denoting some man in the street, his suppositions about reference cease to sound so Searlean. In the case of an everyday name, belonging to some Parisian doctor, Foucault remarks that the inaccuracy of associated descriptions does not compromise the link between a name and the individual designated by it. In these remarks he strikes a markedly Kripke-like note: “If, for example, Pierre Dupont does not have blue eyes, or was not born in Paris, or is not a doctor, the name Pierre Dupont will still always refer to the same person; such things do not modify the link of designation” (Foucault, p. 146).

Notwithstanding his earlier remarks about Shakespeare, Foucault here seems to be in agreement with something very much like Kripke’s point about “Gödel”—that the name will “still always refer to the same person,” even if the descriptions that we associate with that name turn out not to apply to this person. The non-Kripkean parts of Foucault’s theory come into play only when, having made this point in connection with the name of a Parisian doctor, he immediately adds that the “problems raised by the author’s name are much more complex” (Foucault, p. 146, emphasis added). His suggestion is not merely that the names of authors are more complex than the names of doctors. It is that, even within the class of authors, the name of Shakespeare is an especially complex case. (Foucault goes on to suggest that the names of Freud and Marx are more complex yet, but these “founders of discursivity” are not our present concern.)

Although Foucault thought that “if we proved that Shakespeare did not write those sonnets which pass for his, that would constitute a significant change and affect the manner in which the author’s name functions,” he does not seem to have thought that all authorial names behave in this way. There is therefore room for him to agree with Kripke about the case of Gödel, treating that case, not in the way that he treats “Shakespeare,” but in the way that he treats his imagined case of “Pierre [End Page 387] Dupont” or the actual case of Rimbaud: “When we discover that Arthur Rimbaud did not write La Chasse spirituelle, we cannot pretend that the meaning of this proper name, or that of the author, has been altered” (Foucault, p. 146).

This last example introduces a further complication (since La Chasse spirituelle was a forgery rather than a misattribution), but it does indicate that Foucault had some appreciation for the sort of consideration that Kripke would introduce. Rather than taking the Kripkean revolution in semantics to thoroughly undermine Foucault’s position, we can instead understand Foucault’s apparently non-Kripkean points as being restricted to the names of certain authors, which he treats as being exceptional—as being, he says, a “paradoxical singularity” (Foucault, p. 146)—precisely because of their non-Kripkean behavior.

This reading imposes a regimentation on some of the things Foucault says, but it does not violate their spirit. Nor would Kripke need to resist the idea that the semantics of authorial names might be somewhat special. He notes—albeit in passing—that the “exact conditions” that determine whether a name is used to designate a particular individual “seem in a way somehow different in the case of a famous man and one who isn’t so famous” (Kripke, p. 95).


I take the foregoing considerations to suggest that it may be possible to transplant Foucault’s points about authorial names from a Searlean to a Kripkean framework, without doing excessive violence to them. In effecting such a transplantation, we need to be careful about certain features of Kripke’s framework that are usually seen as being peripheral to it. Kripke is often read as if he were advocating a theory that leaves no room for the suggestion that authorial names might differ from others in the descriptive content that they contribute to the sentences in which they occur. This is because he is read as if he were claiming that no name ever contributes any descriptive content at all. Such a position would leave no room for some names to communicate descriptive content differently, but (contrary to what is sometimes suggested) this is not a position that Kripke asserts.

In some places, Kripke denies that he has any positive theory to offer (Kripke, p. 64, e.g.). Such remarks might be somewhat disingenuous, but it is true that the semantic points Kripke is most concerned to emphasize are negative ones: that names are not semantically equivalent [End Page 388] to descriptions, and do not attach to their bearers in virtue of any particular description applying to those bearers. Neither of these points entails that names must be entirely lacking in descriptive content. They are together compatible with a view according to which there is some rather unexciting descriptive information—such as the information that Jones is called “Jones”—that can be conveyed by the uses of a name. The Kripkean can allow that such trifling information is conveyed by a name, provided that this information is not responsible for attaching that name to its referent.

At the very end of his first lecture, and again at the beginning of his second, we see Kripke’s willingness to allow that information about naming itself can be conveyed (as a semantically peripheral matter) when a name is used. This idea is introduced in a rather compressed form, at a point when Kripke’s first lecture had overrun its allotted time. Kripke considers the idea that the uses of a name carry minimal information about naming, only in order to reject any attempt to build a semantic theory on the basis of such information. He rejects such attempts on the grounds that they would be circular. He never indicates that it would be false to suggest that information about naming is conveyed (Kripke, pp. 69–70). Kripke would insist that “Jones is late” cannot be the semantic equivalent of “The one called ‘Jones’ is late” (since the first could still have been true, whereas the second would have been false, if Jones had not been so named)—but he can maintain this much while allowing that any actual utterance of that first sentence will communicate the information expressed by the second, if only because Jones is here being so called. This point is clarified toward the beginning of Kripke’s second lecture:

Suppose we amend the thesis so that it reads: it’s trifling to be told that Socrates is called “Socrates” by us, or at least by me, the speaker. Then in some sense this is fairly trifling. I don’t think it is necessary or analytic. . . . As a theory of the reference of the name “Socrates” it will lead immediately to a vicious circle. If one was determining the referent of a name like “Glunk” to himself and made the following decision, “I shall use the term ‘Glunk’ to refer to the man that I call ‘Glunk’,” this would get one nowhere.

(Kripke, pp. 72–73)

Although Kripke would insist that this semantically peripheral conveyance of information can be of little importance when explaining how it is that names refer, we can begin to understand Foucault’s semantic [End Page 389] treatment of authorial names by noticing that even this trifling conveyance of information breaks down in those cases where proper names are functioning as the names of authors. Seeing Foucault’s account of this will enable us to situate his theory of “the author function” within a Kripkean framework.


It is central to modern practices of literary attribution that the name by which the author of a work is identified will typically be some version of the name by which that author is addressed in everyday life. In a prior tradition, of much longer standing, works were more typically published pseudonymously, if they had any name attached to them at all. Several of Foucault’s predecessors (most notably Roland Barthes8) had drawn attention to this. Like them, Foucault takes it to be emblematic of the postmodern moment that this practice of real-name attribution has become unstable: he concludes his essay by noting that, “as our society changes, at the very moment when it is in the process of changing, the author function will disappear” (Foucault, p. 160).

Since he takes writing in one’s own name to be associated with the attempt to “ward off death,” Foucault’s interpretation of this instability is given in somewhat melodramatic terms. He writes:

Our culture has metamorphosed this idea of narrative, or writing, as something designed to ward off death. . . . The work, which once had the duty of providing immortality, now possesses the right to kill, to be its author’s murderer, as in the cases of Flaubert, Proust, and Kafka.

(Foucault, p. 142)

It is not enough, however, to repeat the empty affirmation that the author has disappeared. For the same reason, it is not enough to keep repeating (after Nietzsche) that God and man have died a common death.

(Foucault, p. 145)

This theme of death has been emphasized in several of the discussions that have been prompted by this part of Foucault’s work. Those inclined to make much of it should remember that it was Barthes, rather than Foucault, who put the word mort at the forefront of these discussions, and that he did so partly for the sake of a pun.

If we set aside any thanatological consequences that it might have, we can still see that the tradition of pseudonymity does have consequences [End Page 390] concerning the information that authorial names communicate. The most basic of these is that, whereas the sentence “Jones is late” typically does give one the information that someone who goes by the name “Jones” is late, the sentence “Stendhal is the author of Le rouge et le noir” does not give one the information that someone who goes by the name “Stendhal” is the author of that book. It could not do so, since the person responsible for that book was not called Stendhal. The result of this is, in Foucault’s account, that “To say that X’s real name is actually Jacques Durand instead of Pierre Dupont is not the same as saying that Stendhal’s name was Henri Beyle” (Foucault, p. 147).

Because uses of “Stendhal” do not convey the information that some person goes by the name of Stendhal, the name “Stendhal” cannot be used to abbreviate the description “person who goes by the name of Stendhal.” This is reflected in the figurative uses of such a name. Suppose I were to tell you that there are two Henri Beyles in a class I am teaching. I would be telling you that there are two people who go by that name, and I could then describe a student by telling you that he is one of the Henri Beyles. I would here be using “Henri Beyle” to abbreviate “person who goes by the name of Henri Beyle.”9

Authorial names prefer not to behave in this way. The case in which I talk about one of the Henri Beyles therefore contrasts with the case in which I describe some student by telling you that he is the Stendhal of the class. On the most natural interpretation of that last sentence, I would not be telling you what name the student goes by. I would instead be telling you something about the student’s literary style. Whether that name is used on the title page of a book, or in a somewhat figurative usage (as when we describe someone as being a Stendhal) an authorial name does not tell you what the person in question is known as. Foucault does not elaborate on any examples, but it is clear that he appreciates this point. He notes, “In current usage . . . the notion of writing seems to transpose the empirical characteristics of the author into a transcendental anonymity” (Foucault, p. 144).

Since uses of “Stendhal” preserve the author’s anonymity, and do not convey even the minimal information that some person has been called “Stendhal,” such names might be thought to have a semantic profile that is ultra-Millian; they might, that is, be thought of as cases in which the semantic contribution of a name is absolutely nothing but the individual named. But Foucault immediately goes on to introduce considerations that enable us to see that this ultra-Millian view would face a problem. [End Page 391]

If authorial names gave us absolutely nothing but a person—a “real and exterior individual” (Foucault, p. 147)—then, in a true sentence of the form “Author1 is Author2,” we would have been given absolutely nothing but that one individual (mentioned twice). No predicative content would have been introduced. The proposition expressed by such a sentence could then only be that this individual is herself. A general problem exists regarding how such propositions can be informative. That problem is not what concerns us here. We are concerned with a more particular problem, which arises only when the names in question are authorial. This more particular problem can be seen by noticing that, if authorial names gave us nothing but a person, then the relation expressed by “Author1 is Author2” would be the relation of identity, and so it would, necessarily, be an equivalence relation. The problem that Foucault raises, immediately following his remark about “Stendhal” and “Pierre Dupont,” is that it is not: “One could also question the meaning and functioning of propositions like ‘Bourbaki is so-and-so, so-and-so, etc.’ and ‘Victor Eremita, Climacus, Anticlimacus, Frater Taciturnus, Constantine Constantius, all of these are Kierkegaard’” (Foucault, p. 147).

We can expand on this point by noticing that, while it is true that “Climacus” is one of Kierkegaard’s authorial names, and also true that “Anticlimacus” is, the books that Kierkegaard writes as Climacus are quite different from those that he writes as Anticlimacus: Climacus is, roughly speaking, concerned with doubt, Anticlimacus with faith. It would be quite wrong to say that Climacus is Anticlimacus. In order to avoid being committed to such a statement as an immediate consequence of our saying that both of these are Kierkegaard, we must acknowledge that the “is” in “Climacus is Kierkegaard” cannot be the “is” of identity. But since it would have to be the “is” of identity if the semantic import of authorial names were nothing other than a bare individual, with no predicative content, we can conclude that this ultra-Millian theory of authorial names cannot be correct. This leads us to look for a Foucauldian alternative.


When we view it against the Kripkean background indicated above, we see why the problem raised by Foucault’s examples is one that requires a semantic response. The Kripkean background supports the claim that names (contrary to the Searlean view) are devices of “pure and simple reference,” and not “the equivalent of a description.” This, when taken [End Page 392] together with the now-destabilized tradition of pseudonymous attribution, leaves authorial names without any apparent predicative content. It is that which forces us to read the “is” in “Climacus is Kierkegaard” as being the “is” of identity. We then face a problem when it turns out to denote a nontransitive relation.

We cannot give an adequate answer to this problem merely by giving a metaphysical theory, according to which authors depend for their existence on the texts that are attributed to them (or according to which they are in some other way “dead”). Foucault is very much concerned with the elaboration of such a theory. He tells us that “these aspects of an individual which we designate as making him an author are only a projection, in more or less psychologizing terms, of the operations we force texts to undergo”(Foucault, p. 150). He also notes that “the author does not precede the works” (Foucault, p. 159). But Foucault recognizes that this metaphysical phase of his theorizing cannot, by itself, address the problem created by the apparent nontransitivity of authorial identities. Such nontransitivity would remain problematic, whatever the metaphysical peculiarities of an authorial name’s referent: identity with a thing needs always to be transitive, whatever the metaphysical category of the thing in question. The problem that Foucault’s examples introduce can therefore be addressed by his metaphysical theory of what authors are, only because that theory is arrived at via a semantic route, in which he says that “the links between the proper name and the individual named and between the author’s name and what it names are not isomorphic and do not function in the same way” (Foucault, p. 146). By understanding Foucault’s theory of authorship to be built on a primarily semantic foundation, we are able to avoid a metaphysically peculiar interpretation of that theory.

This route to the avoidance of metaphysical peculiarity is not entirely straightforward. At times, Foucault does seem to be presenting us with a metaphysically peculiar theory, according to which an author is never a person but only ever a creature of ideology. This is most clearly suggested when he writes:

The author is not an indefinite source of significations that fill a work; the author does not precede the works; he is a certain functional principle by which, in our culture, one limits, excludes, and chooses; in short, by which one impedes the free circulation, the free manipulation, the free composition, decomposition, and recomposition of fiction. One can say that the author is an ideological product, since we represent him as the [End Page 393] opposite of his historically real function. . . . The author is therefore the ideological figure by which one marks the manner in which we fear the proliferation of meaning.

(Foucault, p. 159)

Passages such as this might seem to suggest that Foucault is advancing a theory according to which it is somehow impossible for a person to write a book, since no book ever has a person as its author, but only “a certain functional principle.” One tradition of poststructuralist thought embraces such a conclusion (and takes itself to be following Foucault in doing so), but the resulting position runs contrary to a strong line of commonsense thinking. It is not what Foucault’s arguments require, when the semantic stages of those arguments are understood in the way outlined above. Once they have been put on this semantic footing, Foucault’s arguments are compatible with acknowledging that it is, of course, possible for a person to write a book, and possible for that person to lay claim to it by writing her proper name on its title page. Those arguments show only that, when an author’s name occurs in such a context, its semantic relation to its bearer ceases to be a matter of direct reference. Foucault accounts for this by taking it to be a consequence of the fact that the speech act performed by the inscription of one’s name on a title page is one of claiming, and not one of denoting, so that the function of authorial names is classificatory, rather than referential. He writes: “These differences [between personal names and authorial names] may result from the fact that an author’s name is not simply an element in a discourse (capable of being either subject or object, of being replaced by a pronoun, and the like); it performs a certain role with regard to narrative discourse, assuring a classificatory function” (Foucault, p. 147).

When an authorial name serves this “classificatory function,” the claim that gets made by its use is most fundamentally a claim about the work with which the author is credited, and about its relation to other works. It is, Foucault thinks, only indirectly about the person who holds the pen (or who answers to the name):

Such a name permits one to group together a certain number of texts, define them, differentiate them from and contrast them to others. In addition, it establishes a relationship among the texts. Hermes Trismegistus did not exist, nor did Hippocrates—in the sense that Balzac existed—but the fact that several texts have been placed under the same name indicates that there has been established among them a relationship of homogeneity, [End Page 394] filiation, authentification of some texts by the use of others, reciprocal explication, or concomitant utilization.

(Foucault, p. 147)

The claims that Foucault is making here are intended to apply only to a restricted set of works. When the word “Gödel” appears in conjunction with Gödel’s Second Incompleteness Theorem, that name makes a contingent claim about who it was that did the work. It may indeed be that the theorem in question can be “reciprocally explicated” with others that have the same authorial attribution, and with which it stands in relations of “homegeneity and filiation,” but the mere appearance of the author’s name can neither establish nor indicate this. A Kripkean name therefore cannot function in the way that Foucault indicates, but this is not a counterexample to Foucault’s theory. To the extent that Gödel’s metamathematics is purely mathematical (but only to that extent; Foucault, p. 152), Foucault would treat it as one of the “scientific discourses,” from which the “author function” has “faded away” (p. 149). His point is quite specifically about the function of authorial attributions in literary contexts (as we saw in section 2 above). It is a point about literary authors qua literary. When we look outside of mathematics, we can find cases of authorial names that do behave in the classificatory way that the above quotation indicates. Such names enjoy a perfectly unparadoxical existence in the domain of genre fiction.

Just as there may be several authorial roles that a single writer, like Kierkegaard, can play, so, in the world of genre fiction, there may be several writers who—either in turn or collectively—occupy one authorial role. The six mystery novels that were published under the name of “Chester K. Steele” between 1911 and 1928 were penned by six different writers, none of who was called by this name elsewhere in their life. The name “Chester K. Steele” therefore operates in exactly the way that Foucault indicates in the quotation above: It “does not pass from the interior of a discourse to the real and exterior individual who produced it.” Instead it indicates a relation of “homogeneity, filiation, . . . reciprocal explication, or concomitant utilization” (Foucault, p. 147).

It would be a mistake to think that such behavior is a special feature of pseudonyms. The proper names “Agatha Christie,” “Ian Fleming,” and “Dick Francis” have all appeared on the covers of books that were penned after the persons who answered to those names had died. If we want to avoid saying that the appearance of those names is a sham, then we shall have to join Foucault in saying that the act of writing one’s name on a title page is something other than an act of direct denotation. [End Page 395] When Ian Fleming wrote “Ian Fleming” on the title page of Diamonds Are Forever, he was not employing a pseudonym. He nonetheless was performing the same classificatory act as that which Sebastian Faulks performed when writing “Ian Fleming” on the title page of Devil May Care. Foucault’s semantic claim is that a name that performs “the author function” always becomes apt for such uses, and so ceases to be an instrument of direct reference. When it is being used to designate a literary author, a proper name is always apt to behave in this filiation-establishing way, so that the names of literary authors always identify “positions that can be occupied by different classes of individuals” (Foucault, p. 153). Rather than making the metaphysically peculiar claim that authors are not persons, Foucault can instead be understood as saying that, when Felix Francis writes as Dick Francis, or Sebastian Faulks writes as Ian Fleming, the significance of these authorial names cannot simply be to tell us which individual held the pen. A person is being spoken about, but he is being identified as the occupant of a certain authorial role.


This role-designating theory of authorial names provides a solution for the problem that was introduced by Foucault’s example of Climacus and Kierkegaard. It does so by treating that case as analogous to the case of an actor and his part. The movie posters can truly say “Daniel Craig is James Bond,” just as they once said “Sean Connery is James Bond.” They can say these things without thereby implying that Sean Connery has ever been Daniel Craig. The same semantic structure enables us to say that Climacus is Kierkegaard, and also that Anticlimacus is Kierkegaard, without thereby implying that Climacus is Anticlimacus.

Notice—contrary to one strand of poststructuralist thinking that takes Foucault as its inspiration—that this account does not depend on James Bond (or Anticlimacus) having the metaphysical status of a fictional entity. Just as Daniel Craig was James Bond in Skyfall, so Charles Dance was Ian Fleming in a 1989 biopic of that writer. Fleming’s proper name can be used to denote a role, even while being the name of a real individual. It is even possible to play the role of oneself. To say that Kierkegaard is Climacus, or that Kierkegaard is the author of Sickness Unto Death, is not therefore to confine Kierkegaard to a shadowy, quasi-fictional existence. The role that the authorial name designates need not be a role within the fiction of the authored work. Instead, as Foucault says, it “has no legal status, nor is it located in the fiction of the work; rather, [End Page 396] it is located in the break that founds a certain discursive construct and its very particular mode of being” (Foucault, pp. 147–48).

By taking authorial names to denote individuals only insofar as those individuals are the occupants of a role, the Foucauldian theory enables us to avoid saying that the appearance of Fleming’s name on Faulks’s book is a sham. I take that to be an advantage of the theory, but this point is one that needs to be handled with care. The theory does not immediately clear Faulks of all charges. Instead it tells us that, if there is a sham in a case like that of Faulks writing as Fleming, it is the sham of imposture, rather than of impersonation. By treating it in that way we are able to recognize that Faulks’s claim strikes the same false note as that which is struck when Gore Vidal’s name is placed on the cover of the posthumously reissued When Thieves Fall Out. Vidal did indeed write that book. He wrote it rather quickly, at a time when he needed the money. It was published under the pseudonym of Cameron Kay. While he was alive, Vidal prevented the book’s reissue. It is only after his death that the book has been issued with his name on its cover.

Foucault’s theory gives a unified treatment of these cases. If we want to say that something false is perpetrated when the name “Gore Vidal” is placed on the cover of his pseudonymous thriller, despite the fact that that name correctly identifies the book’s writer, then it cannot be that the significance of that name is exhausted by its reference. Instead, as Foucault says, “The author’s name manifests the appearance of a certain discursive set and indicates the status of this discourse within a society and a culture” (Foucault, p. 147). The name of Gore Vidal misleads when it appears on the cover of When Thieves Fall Out because that novel has a different “status within a society and a culture” from the novels that form the canonical output of Gore Vidal. When the name of Ian Fleming appears on the cover of Devil May Care, that book thereby purports to have the status of a book by Fleming, and not of a book by Faulks. One’s answer to the question of whether Faulks’s use of Fleming’s name is appropriate will be derived from one’s estimation of whether the book makes good on this purport (and not from one’s theory as to Faulks’s true identity).

Foucault’s theory does not imply that anything goes when we are in the business of making authorial attributions in literary contexts. Contrary to a fear that poststructuralist interpretations of it sometimes provoke, it provides us with grounds for determining what it is that goes, and what it is that does not. Some roles place extraordinary demands on their occupants. There is no reason why these should not include [End Page 397] constraints as to their personal identity. The theory therefore allows that some attributions can never be appropriate. It would be a sham if Sebastian Faulks wrote “Shakespeare” on a volume of sonnets (much as it was a sham when Pascal Pia wrote “Rimbaud” on a volume claiming to be La Chasse spirituelle).

Foucault’s theory gives a satisfactory account of all these matters, and does so while avoiding any metaphysically peculiar claims, but it does this only after being grafted to Kripkean stock. By developing Foucault’s ideas from these Kripkean roots we have also been able to show that there is a cogent argument to be given in support of a Foucauldian theory.

Christopher Mole
University of British Columbia


1. Saul A. Kripke, “Naming and Necessity,” in Semantics of Natural Language, ed. Gilbert Harman and Donald Davidson (Dordrecht: Reidel Publishing Company, 1972), pp. 253–355, reissued as Naming and Necessity (Cambridge: Harvard University Press, 1980); hereafter abbreviated Kripke. All page references are to the 1980 edition.

2. Michel Foucault, “Qu’est-ce qu’un auteur?” Bulletin de la Société française de philosophie 63, no. 3 (1969): 73–104, reprinted as “What Is an Author?” in Textual Strategies: Perspectives in Post-structuralist Criticism, ed. and trans. Josué V. Harari (Ithaca: Cornell University Press, 1979), pp. 141–60; hereafter abbreviated Foucault. All page references are to Harari’s translation.

3. Saul A. Kripke, “A Completeness Theorem in Modal Logic,” Journal of Symbolic Logic 24, no. 1 (1959): 1–14.

4. Michel Foucault, Naissance de la clinique (Paris: Presses universitaires de France, 1963).

5. Saul A. Kripke, Reference and Existence (New York: Oxford University Press, 2013), p. 4.

6. John Stuart Mill, A System of Logic, Ratiocinative and Inductive, vol. 7 of The Collected Works of John Stuart Mill, ed. J. M. Robson (Toronto: University of Toronto Press, 1974).

7. John R. Searle, Speech Acts: An Essay in the Philosophy of Language (Cambridge: Cambridge University Press, 1969).

8. Roland Barthes, “The Death of the Author,” in Image-Music-Text, ed. and trans. Stephen Heath (New York: Hill and Wang, 1977), pp. 142–48.

9. Such uses are emphasized by Tyler Burge, “Reference and Proper Names,” The Journal of Philosophy 70, no. 14 (1973): 425–39. Burge takes them to be problematic for the Kripkean. The Kripkean might insist that these uses are figurative, but—given the remarks noted above, in which Kripke allows that information about naming can be conveyed when a name is used—he need not deny that such uses of a name exist. [End Page 398]

Additional Information

Print ISSN
Launched on MUSE
Open Access
Back To Top

This website uses cookies to ensure you get the best experience on our website. Without cookies your experience may not be seamless.