Pierre Menard in Logical Space

In accomplishing the formidable task of assessing the truth of something as complex as a story, a reader has to decide between two alternatives. On the one hand, she may decide that the tale she finds herself listening to, reading, or viewing is being related by a teller who has (or purports to have) firsthand knowledge of the characters and situations, and who consequently relates the tale as fact. On the other hand, she may decide that she is experiencing the weaving of a fantasy, in which case the teller, as author of this fiction, would not have had first-hand knowledge of the story's events.[1]

                  To read Jorge Luis Borges' short story “Pierre Menard, Author of the Quixote,” as fiction is, according to David Lewis, to apply an “intensional operator” which one might consider “a restricted universal quantifier over possible worlds” (“Truth in Fiction” 264).[2] Such restricted quantification is necessary for fulfilling the requirement that Pierre Menard's story occupy a world outside that in which the tale's recipient is located because its teller lacks first-hand knowledge. And the events that occur within the narrative frame would also appear to require such a move. In the history related by Borges (via an unnamed narrator), Pierre Menard believes himself capable of performing the incredible act of composing, as though for the first time, the text of Miguel de Cervantes' Don Quixote. In order to actually perform this Herculean task—one that would undoubtedly be impossible in our world—Menard would need to find himself (physically) in some world accessible from ours, but nevertheless located squarely outside its boundaries. However, in order to believe himself capable of doing so, he would need only locate himself in that area of logical and/or ordinary space most capable of supporting his belief state. Neither of these alternatives is a simple task to accomplish or analyze, but deciding whether to take the work as fact or fiction goes a long way in selecting between actual or possible worlds.[3]

            Because fiction concerns events that have not actually occurred in the precise ways represented, it seems plausible to consider reasoning about fiction as similar in method to counterfactual reasoning (TF 269). David Lewis offers his theory of counterfactuals as one particularly suited to the consideration of fiction because it involves worlds external to ours. In such worlds, “causal dependence among events […] may be analyzed simply as counterfactual dependence” (Lewis, “Causation” 165). However, the peculiarities of the story with which we are concerned, that of Pierre Menard, involves his entertaining (strange) beliefs about himself, and its analysis seems therefore to require something beyond his (self) location within a counterfactually determined world or set of worlds. Rather, the de dicto state of affairs that results from mere (self) location in a world requires an additional mechanism for registering Menard's attitudes about himself, specifically, his de se beliefs. By identifying the counterfactual propositions that locate the set of Menard-worlds (M-worlds) and then transforming them (via centered worlds) into a property that he might then self-ascribe, this essay will provide a deeper understanding not only of the relation of our actual world to that inhabited by Menard, but also of the dynamics involved in his adoption of de se beliefs within a counterfactually constructed world.

            The “intensional operator” proposed by Lewis (referenced above) picks out the set of worlds in which Menard's abilities are possible. It would at first appear, from Menard's declaration, “The task I have undertaken is not in essence difficult,” that a necessity (rather than the less restrictive modal, possibility) operator would be required for selecting worlds.  This seems tenable in light of Lewis's assertion that “[w]hat is essential to something is what it has in common with all its counterparts; what it nowhere vicariously lacks” (Counterfactuals 40). Nevertheless, one is simultaneously prodded toward adopting a less restrictive modality based on the narrator's attitude toward his readers. He assumes incredulity on their part in response to his relation of Menard's literary talents. In fact, in response to Menard's dismissal of a particular method of accomplishing the task as “too easy,” the narrator responds, “Too impossible, rather!, the reader will say” (Borges 91), and this anticipated skepticism regarding the veracity of the narrator's claims suggests that Menard's world cannot manifest as vastly different from our world. Despite this, the world would still need to be different enough from ours to support the modest level of success Menard reportedly garnered as a result of his pains. He was able to complete “the ninth and thirty-eighth chapters of Part I of Don Quixote and a fragment of Chapter XXII” (90). This method reasoning about those possible worlds in which Menard's story is likely to find support squares well with Lewis's advice that it is necessary to “depart from actuality as far as we must to reach a possible world where the counterfactual supposition comes true” (TF 269). Hence, M-worlds might be selected based on the Lewis's model of counterfactually derived accessible worlds.

            In Counterfactuals, Lewis provides a model involving sets of worlds accessible from the actual world and represented as concentric “spheres of accessibility”[4] centered on or surrounding[5] the actual world and arranged in order of accessibility therefrom. Each sphere comprises a set of worlds that stand in a certain counterfactual relation to the world of the reader—the actual world. The counterfactual proposition (if X were true, Y would be true) encapsulates the condition that determines the degree of accessibility that can be granted to each sphere. The resulting schematic places  the worlds that are most similar to the actual (in physical laws, attributes, content, etc.) in closer ontological relation to the actual world upon which they are all centered. Each concentric sphere in effect encapsulates our world, and each successively less accessible set is a superset  that contains all previous ones, so that the intersection of all the counterfactual sets of worlds is the actual world. The spheres, in short, form a nested set (Counterfactuals 14). For reasons cited above, it seems desirable that the Menard fiction should belong to a set of worlds that is located ontologically close to ours, and it turns out that according to Lewis's schematic, this might actually be indispensable. As we have realized, it is the standards of similarity that determine the accessibility of the spheres. Since counterfactuals are similar to strict conditionals,[6] the compound that would determine the location of the set of worlds that support the claims made in Pierre Menard's history would amount to a strict counterfactual.[7] The Lewisian model proves its consistency in that the strictness of the counterfactual varies inversely with the standards that determine the worlds' degree of similarity, and less stringent standards have the effect of “bring[ing] more worlds into accessibility” while also making it “more difficult for anything to hold at all these worlds” (14). Since the sets that are farther away from our world are in that position based on their levels of accessibility, it follows also that the antecedents and consequents occur together far less frequently as one proceeds “outwards.” Then, Lewis's limit assumption would guide us toward the closest as the most suitable set of worlds for Menard's history, for “[i]f the consequent of a counterfactual holds at all antecedent worlds within some antecedent-permitting sphere around [the actual world], then also the consequent holds at all antecedent-worlds in the smallest antecedent-permitting sphere” (20). What follows from this is that in order to determine the set of worlds in which the fiction of Pierre Menard is tenable (if such determination is at all possible), we need to construct a counterfactual conditional whose truth value,[8] can be computed based on the attributes of various possible worlds accessible from our own.

       An M-world  accessible from the actual world of the reader can be located based on the following counterfactual: If Menard were able to configure his frame of mind in such a way that it exactly matched that of Cervantes, then he would be able to compose verbatim, but as if for the first time, the text of Don Quixote. The framing of the antecedent derives from the specificity of Menard's desires concerning his project. In proceeding toward his goal of composing the Quixote, the narrator makes it clear that Menard eschews the puerile method of being (or becoming) Cervantes; he opts rather for a more difficult method of achieving his objective. Borges writes, “Being, somehow, Cervantes, and arriving thereby at Quixote—that looked to Menard less challenging (and therefore less interesting) than continuing to be Pierre Menard and coming to the Quixote through the experiences of Pierre Menard” (Borges 91; emphasis in original). The phrase “frame of mind,” used in the counterfactual antecedent, is suggestive therefore because it introduces the concept of attitudes and their objects, a category that subsumes that of the beliefs Menard entertains about himself. It is not surprising then that it is in his essay on “Attitudes De Dicto and De Se,”[9] that Lewis provides insight into the implications of achieving a mental state very similar to that sought by Menard. Lewis explores the relationship between two belief states: one held by a person, Heimson, who erroneously believes he is David Hume, and one held by Hume himself who is correct in his belief that he is Hume. Lewis writes, “Heimson may have got his head into perfect match with Hume's in every way that is at all relevant to what he believes. If nevertheless Heimson and Hume do not believe alike, then beliefs ain't in the head!” (ADDS 142). The difference, of course, is that Menard does not actually believe that he is Cervantes—but the analysis by Lewis nevertheless applies because it confirms that what Menard wants is possible by allowing him to adopt the configuration of Cervantes' psyche while circumventing the logical fallacy of his becoming Cervantes. At this point, the idea of identity across possible worlds, or the Lewisian alternative—counterpart theory—seems relevant to the discussion. Menard's narrative faces the challenge of having an empty existence if the counterfactual argument that determines its location in logical space turns out to be vacuously true: that is, if no worlds exist in which he is able to configure his mind to match that of Cervantes and yet remain Pierre Menard. The possible worlds conceptualized as comprising logical space is populated by beings considered by some to have at least the capacity of being identical to each other. Lewis, who dissents to this trans-world-identity paradigm, offers an alternative. He writes, “I shall assume that each subject of attitudes inhabits only one world. He may have counterparts to stand in for him at other  worlds, related to him by bonds of similarity, but he himself is not there” (ADDS 517). From this perspective, beings in possible worlds may occupy roles akin to others in other worlds; yet, rather than having one person existing simultaneously in multiple worlds (trans-world identity), these beings exist as each other's counterparts.[10]

            Counterpart theory works to Menard's advantage, for even if the set of worlds in which his story, as fiction, is located did not already contain its own version of Cervantes (which it does), the controversy surrounding the possibility of trans-world identity raises doubts about the  conceivability of any world in which Menard's counterfactual antecedent is tenable. Lewis's alternative counterpart theory provides us with more conceptual options, allowing us to hypothesize a world in which Menard exists as other than himself (that is, as a counterpart in possession of those abilities that are impossible at our world) while allowing Cervantes to exist in a relationship of self-identity across worlds.[11] This hybrid model also supports another important aspect of Menard's history: the fact that he does not claim identity with Cervantes. In fact, Menard himself advocates a kind of strange hybridity when he charges his postmodernist rivals with writing fiction that is “good for nothing but […] captivating us with the elementary notion that all times and places are the same, or are different” (Borges 90-91). The text advocates, rather, the cohabitation of (degrees of) temporal difference and similarity, exemplified by his desire as Pierre Menard to enter into a state commensurate with Miguel de Cervantes, and the conflation of counterpart and trans-world identity fortuitously allows this desire to remain feasible. This analysis leads to a situation in which at least one set of worlds is configurable that supports the counterfactual antecedent formulated above. It remains, now, to be shown whether the consequent has any hope of holding in the antecedent worlds, and this requires, further, an analysis of the belief state in which Pierre Menard finds himself.

            Representationalist philosophies and psychologies of belief formation suggest that this attitude might result from a reconfiguration of one's brain to accommodate the particular fact or propositional content with respect to which one adopts the attitude of belief. William G. Lycan gestures toward this paradigm that entertains a quasi-physicalization of beliefs (as states of the subject's head) in a way that reflects their content (as propositions) by describing belief as a dyadic relation formed between a subject and a nominalized complement ascribed to that subject (Lycan 6–7).[12] Berating the lack of attention paid by possible-world theorists to the psychological components of the truth conditions of belief sentences, he notes that “[i]t would be strange if the ontology of belief and the logical form of belief-ascriptions had nothing to do with each other”[13] (7). In our story, Menard's belief ascription takes precisely the logical form of the antecedent of the counterfactual (cited again below) whose truth value would determine the existence of the set of worlds that includes his. This antecedent (if Menard were able to configure his frame of mind in such a way that it exactly matched that of Cervantes) expresses a counterfactual proposition, and it is propositions that generally fill the object slot of an attitude (Feit 7).[14] The challenge, then, is whether it is possible to nominalize the situation described by Menard's counterfactual antecedent and then manipulate it in a certain way so that, as he self-ascribes the particular belief it contains, he also self-locates to that world.[15]

            The representationalist conceptualization of belief is roughly translated as the fact of having stored in the brain or psyche a particular token that is the nominalization of the content of a certain proposition. It accomplishes (in theory) the unity of ontology and logical form toward which Lycan gestures above, and in doing so represents the act of belief ascription as a sort of etching of logical form into the neuronal configuration of the brain. As Greg Janzen notes, representationalism “is tied to a certain naturalizing project” involving the explanation of consciousness attitudes such as beliefs as “complex information-providing entities housed in the brain” (138; emphasis added). Therefore, when our narrator quotes Menard's letter in which he avers, “I can premeditate committing it to writing, as it were—I can write it,” that verb “premeditate” refers to the sort of consideration or contemplation that, in informing Menard of his ability to perform the given task, precedes (or coincides with—perhaps defines) the adoption of a belief. Further, the verb also suggest that a psychological process involving the construction of Menard's intent might be taking place, accompanied by a reconfiguration of his brain or psyche that precipitates his final assertion, “I can write it” (92).
            The text contains certain other psychological hints that the representational method of adopting propositional attitudes (as impressing logical form upon the brain) is applicable to Menard—but in ways that introduce problems related to de dicto versus de se attributions. As a denizen of the twentieth century, Menard's achievement of Cervantes' seventeenth century style is interpreted as “an admirable (typical) subordination of the author to the psychology of the hero” (93). First of all, the looseness of the term “psychology” makes it less than obvious that this “subordination” occurs as a result of alteration to the author's brain itself, or merely to some epiphenomenological mind or psyche.[16] But a more troubling problem is that the use of the adjective “typical” generalizes the sentence's evaluative content beyond its application to Menard—a desirable trait of any theory of belief, but one that underscores the de dicto status of his counterfactually derived set of accessible worlds. And a similar problem arises when referents for the terms “author” and “hero” are sought, for their non-specific nature points again toward the de dicto.
            Let us return to the counterfactual antecedent (along with its consequent in those worlds in which they occur together), which forms an axis via which the M-world(s)' accessibility from the actual world is gauged, and attempt a connection with the representationalist view of belief.  It might be useful to think of this antecedent as expressing a representational belief and then to ask whether such an entity can be operationalized; that is, whether the purely psychological can be shown to manifest physically. This question is motivated by Perry's observation that “change in beliefs seem to explain […] change in behavior” (3), and the text appears to affirm this, for Menard writes, “My general recollection of the Quixote, simplified by forgetfulness and indifference, might well be the equivalent of the vague foreshadowing of a yet unwritten book” (Borges 92). What he describes is the glimmer of psychological substance that is both the impetus and the intention to write, and impetus as action or proto-action gestures toward a connection between psychological belief and those behaviors that support or exemplify it.
                Enter the interpretationist approach as a rival of representationalism that offers a behavioral method of determining belief by founding its ascription upon those actions that demonstrate (or imply) the existence of that belief (Burwood et al. 98). The emphasis placed by each on action (interpretationism) and inscription of logical form (representationalism) forges an analogy between the two purported manifestations of belief—with, however, the interpretationist view boasting the advantage of an empirical component that allows—even demands[17]—direct observation and falsifiability of the belief claim. Though advocates of each would likely eschew a fusion of these perspectives, their alliance is favorable at least for our progress toward locating the M-world. By combining these two approaches, we seek to make a judgment about Menard's belief state that also performs an oblique analysis of the truth conditions of our counterfactual. For if behavior can be shown to confirm belief, then the consequent of Menard's counterfactual (then he would be able to compose [...] Don Quixote) might be verified. Therefore, when the text says “he undertook a task of infinite complexity [,] dedicated his scruples and his nights 'lit by midnight oil' to repeating in a foreign tongue a book that already existed; when it goes on to say, “[h]is drafts were endless” and Menard, in his own voice, describes the components of the task as a game that “first allows me to try out psychological variants” on the text and then “forces me to sacrifice them to the 'original' text,” this is a demonstration or rendering of his belief into action (Borges 95, 93). He represents himself as truly performing the task: composing the Quixote without copying or transcribing Cervantes' version, for he makes it clear that his task involves sacrifices of his “psychological variants” to the “original text,” arrived at “by irrefutable arguments” he presumably has with himself during the process of composing his Quixote (93). He represents himself performing all the rigors of the act, and at this point we might truly have located a world in which the truth conditions of Menard's counterfactual are realized.

            But not so fast; there is trouble in paradise. These beliefs that we have attributed to Menard are beliefs about himself—de se beliefs, and his adoption of them are a self-ascription of their content. Yet, according to John Perry, the problem of the essential indexical threatens the view that “belief is a relation between subjects and propositions conceived as bearers of truth and falsity” (3-4). The reflexivity of the belief is of consequence to the truth value of a proposition: its truth or falsity depends upon whether and how a person (Menard) comes to believe particular things about himself. Menard's actions have indeed shown him to self-ascribe, but some doubt exists as to whether self-ascription is in itself enough to instantiate the de se relation. The mechanics of self-ascription, which have yet to be explored, may shed some light on this problem, and (again) Lewis provides a clue on how to proceed. Not surprisingly, a similarity exists between the Lewisian account of counterfactuals as picking out accessible worlds, and propositions as sets of possible worlds. Lewis proposes a definition of properties that subsumes that of the proposition, expanding the idea of a property to one that has the concept of location built into it (ADDS 521). He writes, “to any proposition there corresponds the property of inhabiting some world where that proposition holds” (516), but goes further, however, to note that when one can “self-ascribe the property of being in a certain perceptual situation,” that property “does not correspond to any proposition, since there are worlds where some have it and others do not” (520). His belief is that the property is a more fine-grained analytic tool than the proposition because it is able to perform in a wider range of situations.

            This brings us to a problem associated with Menard's self-location in the M-world. Since the counterfactual propositions we chose correspond to a set of accessible worlds, the problem arises how Menard is able to select his own M-world from the potentially infinite set returned. Perry sheds light on this problem when, in analyzing his own messy-shopper example, he notes that his actions to prevent the mess were prompted not simply by the realization that John Perry was making a mess, but by the concomitant belief that he was himself John Perry (Perry 4-5). How, then, does our Menard know that the flaccid designator “Menard” of the counterfactual antecedent (and the worlds it returns) refers to him? Lewis's 'perceptual belief' that allows a person (Lingens it was) to locate himself in a given de dicto (or even de re) situation conflates with Perry's identification of the need to invoke the essential indexical (Perry 5). In Lewis's observation that some worlds “have it and others do not,” it refers to an indexical relation, a self-ascription of “the property of being in a certain perceptual situation.” This is what reduces the set of worlds—the set of de dicto situations—to one de se world in which the subject comes to identify himself with a given player, as a particular individual “whose boundaries don't follow the borders of the worlds” (ADDS 519).[18] Subjecting counterfactuals to a process similar to that performed by Lewis upon propositions, therefore, results in a similar (counterfactual) property—one that would select from among the set of accessible worlds the single one to which our Menard belongs, and beyond that (or, perhaps, concomitant with it) the single person—“Menard”—to whom our Menard can form a relationship of identity. For Lewis this is indispensable, for “there is a kind of ignorance that cannot be remedied by any amount of self-location in logical space” since self-location is involved in the adoption of both de dicto and de se attitudes (ADDS 523). However, with the de se, this self-location is achieved via the ascription of a property that has a built-in locating component.  At this point, it becomes useful to explore the concept of properties as objects formed in conjunction with centered worlds, thereby facilitating their (self) ascription as the means of self-location in possible or accessible worlds.

            A centered world is a conceptual mechanism for spatio-temporally locating an entity that possesses an intentional state (Liao 11). It is defined in the Lewisian sense as an instantaneous person slice—a person along with his/her temporal location—often represented as an ordered pair: (person, time).[19]  According to this model, a proposition as set of worlds becomes centered upon a given person at a given time. The point at which the proposition (as set of worlds) is identified  is one at which the de dicto self location has already been performed [20]; self-ascription is now on the verge of being performed, and de dicto is now in the process of becoming de se. This is accomplished by the Lewisian maneuver that transforms the proposition (qua set of worlds) into the “property of inhabiting some world where that proposition holds” (ADDS 516). In this situation, the persisting person (corresponding to the temporally constrained person upon which the world is centered) has the capacity to self-ascribe the property into which the centered world qua proposition has been transformed. The property now consists of a time-slice of some objectified version of the subject performing the ascription—and this ascription of the self to the self is what transforms the de dicto into the de se (ADDS 521).

            With Menard, the strongest evidence of self-ascription is his declaration, “I have assumed the mysterious obligation to reconstruct, word for word, the novel that for him [Cervantes] was spontaneous” (Borges 92-93; emphasis added). While this sentence does not explicitly express his affirmation of himself as the Pierre Menard of the counterfactual accessible world, he does ascribe to himself the task whose performance defines that “Menard.” Furthermore, he uses the indexical 'I,' which Perry adamantly insists is indispensable to the adoption of a de se belief (5). In addition to this, a weaker bit of evidence suggests that Menard also self-ascribes a time and thereby a definite location, for the narrator writes that he “dedicated his scruples and his nights 'lit by midnight oil'” to the task. The time, “midnight,” is quoted, and all the narrator's quotes have so far been attributed to Menard via his letters, unless otherwise indicated. Since the narrator attributes this quote to no one in particular, it is reasonably safe to consider it as uttered by Pierre Menard. Upon this evidence, we may indeed say that he self-ascribes a property in the form of a world centered upon “Menard,” thereby identifying himself with a temporally located individual, and consequently adopting a de se belief toward the set of accessible worlds generated by our counterfactual proposition.

            Despite all this, the joke of the text cannot be ignored: Menard repudiates everything that he affirms. He describes the Quixote as “contingent,” “inevitable” and “impossible” inside the same paragraph,[21] and while he indeed writes the text he sets out to write, he destroys it as it is written—apparently for the purpose of ensuring that it does not survive him. In the end, all that does survive is his intent to write the text, which is all the narrator can truly convey—but according to the interpretationalists, this is not enough to underwrite belief. Rather, mere intent is belief cauterized, action stifled in the moment of its execution. So the final world might be that all we are truly left with in the actual world is evidence of the existence of antecedent M-worlds at which none of the consequents hold.

References

Balderston, Daniel. 1993. Out of Context: Historical Reference and the Representation of Reality in Borges. Durham: Duke UP.

Borges, Jorge Luis. 1939. “Pierre Menard, Author of the Quixote.” Collected Fictions. Andrew Hurley (Trans.) New York. Penguin, 1999.

Burwood, Stephen, Paul Gilbert & Kathleen Lennon. Philosophy of Mind. Fundamentals of Philosophy Series. John Shand (Ed.). London: UCL Press.

“Desire.” 1988. Routledge Encyclopedia of Philosophy. New York: Routledge.

Feit, Neil. 2008. Belief about the Self: A Defense of the Property of Theory Content. Oxford: Oxford UP.

Janzen, Greg. 2008. The Reflexive Nature of Consciousness. Philadelphia: John Benjamins Publishing.

Lewis, David. 1979. “Attitudes De Dicto and De Se.” The Philosophical Review. 88(4): 513-543.

—. 1986. “Causation.” Philosophical Papers. Vol. 2. New York: Oxford UP. 159-213.

—. 1973. Counterfactuals. Oxford: Blackwell.

—. 1983. “Truth in Fiction.” Philosophical Papers. Vol. 1. New York: Oxford UP. 261-280.

Liao, Shen-Yi. “What Good are Centered Worlds?” University of Michigan. http://www-personal.umich.edu/~samliao/liao-cw.pdf [Link now broken, unfortunately]

Lycan, William J. 1988. Judgement and Justification. Cambridge: Cambridge UP.

Miller, Alexander. 1998. Philosophy of Language. Fundamentals of Philosophy Series. John Shand (Ed.). London: UCL Press.

Perry, John. 1979. “The Problem of the Essential Indexical.” Nous. 13(1): 3-21.

---

[1]    The use being made here of the term “first-hand knowledge” restricts its application to those events which have occurred in a world shared by both the teller and recipient of a tale. See Lewis' “Truth in Fiction,” p. 266.

[2]    Hereafter abbreviated as TF.

[3]    This importance of this decision is not to be belittled. Pierre Menard was a real person who did, in fact, compose all the titles listed in the first pages of the story Borges writes. The only discrepancy between his actual corpus and that cited by Borges (via his narrator) is “the subterranean, the interminably heroic production […] that must remain—for such are our human limitations!—unfinished” (Borges 90). Daniel Balderston offers evidence to this effect in the chapter “Menard and his Contemporaries: the Arms and Letters Debate” in his text on representations of reality in Borges' fiction.

[4]    Or balls. Lewis notes that, mathematically speaking, the objects described here as spheres are actually balls, since each concentric sphere is meant to be a solid comprising a set of worlds whose common bond is the possession of a certain degree of accessibility from the actual world (Counterfactuals 7).

[5]    Not to be confused with centered worlds, which are distinct philosophical entities—though an argument for the similarity of the uses to which both might be put will be made subsequently.

[6]    That is, strict conditionals whose strictness cannot be predetermined because of its variable nature. See Lewis in Counterfactuals (4).

[7]    One also wonders about the extent to which the strictness of the counterfactual also interacts with its components, to increase or decrease the accessibility of the worlds. This is based on the fact that the counterfactual components must represent the extreme nature of the claims made in Pierre Menard's tale—his self-ascribed properties or beliefs about himself.  Lewis allows for similarity (or even identity) of pre-antecedent worlds to the actual world “at the cost of a small miracle” (77), but it is not yet clear to me what happens when the components that determine the accessibility of the counterfactual (set of) worlds are so vastly dissimilar from what could occur in the actual world that they would be deemed impossible.

[8]    Truth value might here be relatively safely replaced with necessity, for according to Lewis, “[n]ecessity is truth at all accessible worlds” (5).

[9]    Hereafter referred to as ADDS.

[10]  See also Counterfactuals, p. 39.

[11]  A complication arises when one notes that Menard's determination to remain himself while conceiving the work of Cervantes forces him to alter a part of that work, at least in intention. The narrator assures us that this determination to write as himself has obliged Menard to “leave out the autobiographical foreword to Part II of the novel. Including the prologue,” the narrator continues, “would have meant creating another character—'Cervantes'—and also presenting Quixote through that character's eyes, not Pierre Menard's” (Borges 91). This decision both respects and flouts the trans-world identity theory, for while it seeks not to create another Cervantes, respecting that one already exists (the historical Cervantes that obtains also in the M-world), it ignores the identity relation that does need also to subsist in both worlds between the (two) “Cervantes” of the prologue in order to maintain the overall identity of the real Cervantes. Of course, one may simply appeal to counterpart theory to avoid this difficulty. Menard's assessment, “The Quixote is a contingent work; the Quixote is not necessary,” supports this (92). However, this omission might yet be made consistent with the foregoing identity-based analysis by noting that much of the rest of the Quixote was also “left out,” since the world Menard inhabits is one in which he is, before his death, able to complete only two full chapters of the novel using his method. We could, therefore, simply count the abandoned portion of the prologue as simply part of the section(s) Menard was unable to complete.

[12]  A revamping of the definition of belief as a relationship between an agent and a proposition that emphasizes his particular ontological view of the proposition.

[13]  Perry problematizes this relationship in his essay “The Problem of the Essential Indexical,” contending that “its solution requires us to make a sharp distinction between objects of belief and belief states” (4). It is worth noting, however, that this analysis might be guilty of mingling otherwise distinct conceptions of the propositional object. As Lewis notes, “Not everyone means the same thing by the word 'proposition.' I mean a set of possible worlds, a region of logical space. Others mean something more like a sentence, something with indexicality and syntactic structure, but taken in abstraction from any particular language” (515). It is with the latter description that this paragraph deals, although propositions as sets of possible worlds will certainly be the focus of a significant portion of the rest of the paper.

[14]  Lewis's problematization of this generalization will be dealt with presently.

[15]  This is feasible. It might be necessary to alter the form of the counterfactual, changing it from a subjunctive into an indicative mood, but since this proposition is meant to hold outside of the actual world, its content would, in fact, exist there in precisely the form necessary for Menard to adopt it as a belief upon which he could then act This is in contradistinction to a propositional or intentional attitude that exists as the object of a desire, which might innocuously remain in the subjunctive (Routledge 30).

[16]  The connection between [the reconfiguration of Menard's brain to fit that of Cervantes] and [the reconfiguration of his brain to reflect the adoption of the belief that (the above-mentioned reconfiguration is both dispositional and occurrent)], is interesting and difficult. This paragraph is more concerned with the second; it assumes the first—perhaps too hastily.

[17]  In the tradition of logical positivism, interpretationists demand that belief be “truth-apt” (Miller 106).

[18]  Presumably Menard does achieve the required perceptual situation, since he proceeds to perform actions that reflect this. The question arises, however, whether the “he” of the consequent takes on the value assumed by the “Menard” of the antecedent once Menard (our Menard) comes to associate himself with the name. What is its status as a variable: bound or free, and to what degree? [Counterfactual: If Menard were able to configure his frame of mind in such a way that it exactly matched that of Cervantes, then he would be able to compose verbatim, but as if for the first time, the text of Don Quixote.] Is the “he” also de dicto prior to Menard's identification with the name “Menard”? It would appear so, and in that case the assumption made earlier (that Menard achieves the required perceptual situation) would be erroneous. And this would make self-ascription of the belief all the more crucial for him, for without self-ascription, he would actually not (know or believe himself to) be performing the task of re-composing the Quixote. In that case, any self location he would have performed so far would have been de dicto self location. As Lewis quips: “I reject Perry's terminology: I say that all belief is 'self-locating belief.' Belief de dicto is self-locating belief with respect to logical space; belief irreducibly de se is self-locating belief at least partly with respect to ordinary time and space, or with respect to the population” (ADDS 522).

[19]  The Quinean view of centered worlds couples the time coordinate with a three-dimensional space coordinate, but this runs into the problem of how to account for more than one thing located at exactly the same place and time—such as Schrodinger's cat and its ghost, for instance (Liao 9).

[20]  See note 18 above, or Lewis's “Attitudes De Dicto and De Se,” p. 522.

[21]  Menard writes, “Composing the Quixote in the early seventeenth century was a reasonable, necessary, perhaps even inevitable undertaking; in the early twentieth century, it is virtually impossible” (Borges 93). See note 11 above for his earlier reference, in this same letter, to the contingency of the Quixote.