“Maelzel’s Chess-Player” and
Poe’s Reverse Constraints
Abstract In his essay “The Philosophy of Composition” (1846), Edgar Allan Poe
describes how he composed his lyric poem “The Raven” by following a series of
predetermined steps. My essay shows how Poe’s description of composition as rule
following both has suggestive affinities with and significantly alters the Oulipian
understanding of constraints as axioms that precede composition. Looking closely
at Poe’s earlier essay “Maelzel’s Chess-Player” (1836) and the way it anticipates more
recent debates in artificial intelligence, I show how Poe’s (1984d : 13) constraint,
as stated in “The Philosophy of Composition,” “I prefer commencing with
the consideration of an effect,” is a matter of concealing a decade of experimentation
in previous magazine essays with the effect of a poetry-making algorithm.
I prefer commencing with the consideration of an effect.
—Edgar Allan Poe, “The Philosophy of Composition”
Poe and Rules
In his 1838 tale “How to Write a Blackwood Article,” Edgar Allan Poe
caricatures some formulas of British magazine writing of the time, namely,
the display of erudition and the sensational depiction of death. The
story is both a transcript of Signora Psyche Zenobia’s discussion with a
“Mr. Blackwood” and a facsimile of the story she proceeds to write based
on his advice. This advice includes everything from using “very black ink”
to making sure one’s writing contains “taste, terror, sentiment, metaphysics
and erudition” (Poe 1978b : 339). Zenobia manages to bungle most
of Mr. Blackwood’s rules, notably his suggestion that she make literary
allusions in foreign languages. Thus Cervantes’s (2003: 708)
Ven muerte, tan escondida
Que no te sienta venir
Porque el placer del morir
No me torne a dar la vida
(Come, death, so secret, so still
I do not hear your approach
so that the pleasure of dying
does not bring me back to life)
becomes, in Zenobia’s version:
Vanny Buren, tan escondida
Query no te senty venny
Pork and pleasure, delly morry
Nommy, torny, darry, widdy!
(Poe 1978b : 354)
If the paradox of Death’s approach rousing the dead back to life is lost in
Zenobia’s rendering, the original nevertheless prefigures her own sensational
fate—as she tells it in her story, her head is sliced off by the hands
of a giant clock. That her narration continues, Orpheus-like, even after the
Scythe of Time has had its way with her is part of Poe’s send-up of the kind
of thing he imagined a Blackwood’s audience liking (erudition plus shock).
But in casting his tale in the form of a “how-to” manual, which, if followed,
will lead to the composition of a tale suitable for publication in Blackwood’s,
Poe, as Thomas O. Mabbot puts it, “consciously describes some of his own
methods” (Poe 1978a: 335). In a period when innovators like Poe were constantly
having to come up with ways of filling the columns of American
the kind of literary journalism found in Blackwood’s. Thomas O. Mabbott writes that Poe was
“undoubtedly familiar with Blackwood’s Magazine . . . since his foster father dealt in imported
books and periodicals [and so Blackwood’s] provided a source for ideas made use of, in one
way or another, in many of Poe’s stories” (Poe 1978a: 357).
2. In fact these lines are not even Cervantes’s (2003: 708) but are quoted from Commander
Escrivá, “a fifteenth century poet from Valencia, whose work was greatly admired by many
writers of the Golden Age.” In the prologue to Don Quixote, Cervantes himself spoofs the idea
of rules for literary composition (ibid.: 3–9).
magazines (Poe is of course the inventor of, among other things, the detective
story), “How to Write a Blackwood Article” is not only an example of
Poe’s shrewd imitative ability but an index to his own growing interest in
literary composition as rule following.
Four years after “How to Write a Blackwood Article” appeared in the
American Museum (under the title “The Psyche Zenobia”), Poe published
a review of Nathaniel Hawthorne’s Twice-Told Tales in Graham’s Magazine.
Deepening his interest in the idea of compositional rules, he couched his
praise for Hawthorne in the following terms:
We need only here say, upon this topic [of the superiority of the tale as a literary
form] that, in almost all classes of composition, the unity of effect or impression
is a point of the greatest importance. . . . A skilful literary artist has constructed
a tale [and has] conceived, with deliberate care, a certain unique or single effect
to be wrought out. . . . And here it will be seen how full of prejudice are the
usual animadversions against those tales of effect many fine examples of which
were found in the earlier numbers of Blackwood. (Poe 1984c : 571–73)
In the four years that separate the Hawthorne review from “How to Write
a Blackwood Article,” Poe both has changed his tone from parody to
approval (he may as well be talking about his own earlier piece when he
condemns the “prejudice [of the] usual animadversions against . . . Black-
wood”) and has arrived at a more explicit formulation of his ideas about
compositional rules. Whereas in “How to Write a Blackwood Article” such
rules were a matter of imitating (and caricaturing) the “fine examples” of
Blackwood’s, the Hawthorne review abstracts a general principle from those
earlier examples: the production of “a certain single or unique effect.” In
this way, Poe hints at what will become the premise for his later essay, “The
Philosophy of Composition” (1846), in which he tells us that in all of his
literary compositions he has “prefer[red] commencing with the consideration
of an effect” (Poe 1984d : 13).
In what follows, I want to consider Poe’s emphasis on “effects” in light of
the Oulipian notion of “constraint” as “a strict and clearly definable rule,
method or procedure or structure that generates [a] work” (Mathews and
Brotchie 2005: 131). While Oulipian constraints are formal rules for generating
“potential” literary works, this does not mean that texts arising from
the use of constraints are “demonstrations” in the sense that mathematical
proofs are demonstrations. That is, while Oulipian constraints are indeed
“axioms”—“Proposition 14: A constraint is an axiom of a text” (Roubaud
1986: 89)—the relation between a constraint and its potential text is not
analogous to the proof-like relation between axioms and theorems. As
Jacques Roubaud (ibid.), himself a mathematician, puts it in an essay on
Raymond Queneau: “One may think that a text composed according to a
given constraint (or several constraints) will be the equivalent of a theorem.
It is a fairly interesting hypothesis. It is nonetheless true that the foreseeable
passage from the statement of the constraint to its ‘consequence,’ the
text, remains in a profound metaphorical vagueness.” Such “metaphorical
vagueness,” I take it, refers to the fact that the results of the application
of a constraint are not formally derived in the way that theorems are formally
derived from axioms. For example, and to invoke one of the more
celebrated Oulipian works, the move from Georges Perec’s adoption of a
lipogrammatic constraint (the removal of the letter e from the alphabet) to
the final shape and makeup of La disparition (1969) could not be described
as one of necessity.
Emphasizing the way Oulipian constraints are defined as preceding literary
composition, Eve Célia Morisi (2008: 113) points out that in the case
of Poe’s “Philosophy of Composition,” “constraints . . . no longer come
first, prior to the writing, and are no longer preeminent. Their formulation
follows the poem’s composition instead of preceding it, and therefore cannot
be proved to have presided over it . . . [and thus] the a posteriori writing
of ‘The Philosophy of Composition’ symbolically destabilizes the notion of
literary constraint.” Poe’s description of how he composed “The Raven” is
(of course) formulated a posteriori; but rather than “destabilizing the very
notion of constraints” (what Morisi [ibid.] also calls his “capsizing” the
notion of constraint), Poe leads us to rethink how constraints might work.
If “commencing with the consideration of an effect” (Poe 1984d : 13)
can be thought of as a constraint, then in Poe’s case it will remain a consistent
starting point no matter how unforeseeably myriad or “vague” its consequences.
If Poe seems to be saying that “The Raven”—in all of its jumpy,
trochaic detail—arose step-by-step from a series of axiom-like rules, then
this is just the effect he hopes to create in “The Philosophy of Composition”:
the sleight of hand by which a constraint derived a posteriori is
made to look like one which has preceded composition. Thus what Morisi
(2008: 111) calls a “textual inconsistency,” rendering Poe’s account of the
composition of “The Raven” “inauthentic” (ibid.: 112), misses the point,
since “inauthenticity” is just what enables Poe to create his desired effect.
And if Stuart Levine (2009: 58) is no doubt right when he says “no critic,
no literary historian, no poet has ever believed that Poe literally wrote
4. According to the Penguin Dictionary of Mathematics, a theorem is a “statement derived from
premises” (Nelson 2008: 417).
‘The Raven’ as systematically . . . as he says [he had],” this does not change
the fact that Poe was trying to get his antebellum audience to believe that
“The Raven” had been so composed and so remains in accord with his
declared constraint of “commencing with the consideration of effects.” It is
perhaps in this sense that Harry Mathews and Alastair Brotchie (2005: 211)
include Poe in their list of “anticipatory plagiarists” who were “creating
paleo-Oulipian texts without acknowledgment.”
In what follows, I want to uncover a precedent for Poe’s reverse constraints—
the way “effects” are a matter of both enticing and duping an
audience—in looking closely at his essay “Maelzel’s Chess-Player” (1836),
a piece that predates both “How to Write a Blackwood Article” and the
Hawthorne review. When Poe attended an exhibition of Johann Nepomuk
Maelzel’s automaton chess player in Richmond in 1836, he saw not just a
machine that seemed to play (and win at) chess but the effect the performance
had on an audience. Poe treated this performance, I want to argue,
as a model for the kind of effect he wanted the magazine article (whether
short story, poem, or critical essay) to have on a reader. What struck Poe
about the exhibition was the combination of a carefully controlled management
of theatrical artifice with the sensational notion of a thinking
machine. “Maelzel’s Chess-Player” sets up the relation between rule following—
a chess-playing machine would need to follow programmable
rules for making meaningful (and not just random) chess moves—and the
creation of the effect of such a mechanical intelligence. Poe’s earlier essay, I
want to argue, serves as a template for the way his proposed starting point
of the “consideration of an effect,” in “The Philosophy of Composition”
is, as John Tresch (1997: 289) has put it, his “attempt [to] present himself
explicitly as a poetry-automaton.”
Many literary scholars have characterized Poe’s tone of clinical detachment
in “The Philosophy of Composition” as a matter of theatricality and
showmanship. Kenneth Burke (1966: 25), for example, suspects that Poe’s
essay was written “for purposes of showmanship or to compensate for
his own personal shortcomings by representing himself as a paragon of
rational control.” Daniel Hoffman (1972: 83) is careful to mention that Poe
“compar[es] the composing of a poem to the management of theatrical
props and machinery.” Eliza Richards (2004: 53) compares Poe to a “show
man [who] opens the curtains on the theatricality of the lyric.” Lois Davis
Vines (1992: 105) says that in “allowing spectators to witness the process of
creation,” Poe gives them access to an “intellectual drama.” And Levine
(2009: 72) says that Poe “makes obvious use of stage effects of the sort
popular in productions in American cities.”
But none of these critics traces such “stage effects” and “showmanship”
back to Poe’s witnessing, and writing about, the sensational spectacle of
the mechanical chess player. Nor do they consider that part of the performance
of “The Philosophy of Composition” is precisely the way it creates
the illusion of a literary composition which appears to arise by necessity
from the following out of a set of axiom-like rules. If Poe (1984d : 15)
says he composed “The Raven,” “step by step to its completion with the
rigid consequence of a mathematical problem,” he simultaneously conceals
a series of prior experiments like “How to Write a Blackwood Article,”
the Hawthorne review, and the essay on the chess player itself. “Maelzel’s
Chess-Player” lays out in detail Poe’s interest in the way the careful creation
of surface effects—what appears to be going on—informed his understanding
of how literary compositions could produce similar effects on a
reader. Such illusionism is analogous, I would argue, to the reversal at
work when “effects” are generated from a hoax-like transposition of an a
posteriori explanation into an axiomatic starting point. “The Philosophy
of Composition” makes a “constraint” not out of the determinate algorithmic
series by which Poe claims to have composed the poem but from the
effect of having composed the poem in this way.
“Perhaps no exhibition of the kind has ever elicited so general attention
as the Chess-Player of Maelzel” (Poe 1984b : 1253). So begins the
unsigned editorial Poe contributed in 1836 to a new Virginia magazine
called the Southern Literary Messenger. In it he attempts to debunk the illusion
produced by the chess-playing automaton of the Bavarian inventor
Maelzel, a traveling exhibition he saw on a number of occasions during one
of its U.S. tours. Closely observing the way Maelzel’s machine appeared
6. Maelzel purchased the chess automaton (sometimes called “the Turk” because of its
sultan’s attire) from the German inventor Baron Wolfgang von Kempelen. After touring
Europe a number of times and being sold to Maelzel, the chess player made a series of
appearances (often accompanied by musical automatons) along the eastern seaboard of the
United States in the 1820s and 1830s (Standage 2002). Before seeing Maelzel’s chess player,
Poe relied heavily on pamphlet copies of David Brewster’s Letters on Natural Magic for his
knowledge of the machine. Poe’s essay appeared in the April 1836 issue of the Southern Literary
Messenger. For a detailed publication history of how Brewster’s writings about the chess
machine reached North American periodicals, see Wimsatt 1939: 144–46.
to win chess games against volunteers picked from the audience, Poe was
fascinated by the effect it had on a paying spectator and was convinced that
the machine was a fake.
Poe (1984b : 1253) begins his essay by pointing out the distinction
between a piece of mechanism and a human mind. “Everywhere
men of mechanical genius, of great general acuteness, and discriminative
understanding . . . make no scruple in pronouncing the Automaton a pure
machine, unconnected with human agency in its movements.” Poe (ibid.:
1255) then compares the chess-playing automaton to Charles Babbage’s
difference engine, saying that a machine able to win at chess must be a far
more complex mechanism than a machine that merely churns out sums:
It will perhaps be said in reply that a machine such as [Babbage’s calculating
machine] is altogether above comparison with the Chess-Player of Maelzel.
By no means—it is altogether beneath it [since] arithmetical or algebraical
calculations are, from their very nature, fixed and determinate. Certain data
being given, certain results necessarily and inevitably follow. These results have
dependence upon nothing, and are influenced by nothing but the data originally
given. And the question to be solved proceeds, or should proceed, to its final
determination, by a succession of unerring steps liable to no change, and subject
to no modification. This being the case we can without difficulty conceive the
possibility of so arranging a piece of mechanism, that upon starting it in accordance
with the data of the question to be solved, it should continue its movements
regularly, progressively and undeviatingly towards the required solution,
since these movements, however complex, are never imagined to be otherwise
than finite and determinate.
Poe is concerned here with making the distinction between Babbage’s difference
engine and Maelzel’s chess player as that between a “determinate”
calculation and making inferences about data that cannot be known in
advance. A machine like Babbage’s calculator “proceeds . . . to its final
determination, by a succession of unerring steps liable to no change, and
7. Interestingly, von Kempelen himself alludes to the importance of the “effect” of the
machine upon a spectator in a quotation Poe includes in his essay: “[The chess player] is
a very ordinary piece of mechanism—a bagatelle whose effects appeared so marvelous only
from the boldness of the conception, and the fortunate choice of the methods adopted for
promoting the illusion” (Poe 1984b : 1256–57).
8. Babbage was a British mathematician who takes his place in history as the inventor of
the first programmable calculating machine. For accounts of Babbage’s place in the history
of computing, see Morrison and Morrison 1961; Haugeland 1986; Breton 1990; Swade
9. The difference between intelligence understood as rational calculation and intelligence
understood as skillful coping in the world (and so involving embodied experience) is the
main animating issue in research and debate on artificial intelligence. See especially Haugeland
1986; Boden 1990; Dennett 1998; and Dreyfus 1999.
subject to no modification,” since it treats the data fed into it in a “fixed and
determinate” manner, with the conclusion implicit in the premises from
the beginning, just as a deductive inference moves from a major premise,
through a particular case, and to the conclusion by necessity. 0 Maelzel’s
chess player, on the other hand, would have to respond in real time to an
opponent’s chess moves and convert those data into meaningful moves of
its own. Thus with the chess player,
there is no determinate progression. No one move in chess necessarily follows
upon any one other. From no particular disposition of the men at one period
of a game can we predicate their disposition at a different period. Let us place
the first move in a game of chess in juxta-position with the data of an algebraical
question, and their great difference will be immediately perceived. From
the latter . . . the second step of the question, dependent thereupon, inevitably
follows. It is modeled by the data. It must be thus and not otherwise.
But from the first move in a game of chess no especial second move flows of
necessity. (Ibid.: 1256)
If algebraic data, according to Poe, advance inexorably from premise
to conclusion, moves made by the mechanical chess player would occur
without such predeterminations. Poe’s contrasting of Babbage’s difference
engine and Maelzel’s chess-playing machine is thus a matter of spelling out
the difference between a priori deductions and a posteriori inductions—
a distinction that will serve as the basis for his argument attempting to
debunk the machine as a fake.
For Poe’s readership, the most captivating question concerns whether or
not somebody is concealed within the machine—a question that amounts, for
the antebellum audience, to “Can a mechanical contraption actually think?”
or, to update the wording, “Is there such a thing as artificial intelligence?”
Poe (ibid.: 1264) takes up this question with a lesson in inductive reasoning:
Some person is concealed in the box during the whole time of exhibiting the
interior. We object, however, to the whole verbose description of the manner in
which the partitions are shifted, to accommodate the movement of the person
concealed. We object to it as a mere theory assumed in the first place, and to
which circumstances are afterwards made to adapt themselves. It was not and
could not have been arrived at by any inductive reasoning. In whatever way the
shifting is managed, it is of course concealed at every step from observation.
As with the distinction between the “determinate calculations” of Babbage’s
calculating engine and the more improvisatory making of chess
moves, Poe will not accept the inference based on a “mere theory assumed
in the first place.” One should not speculate about hidden partitions,
because one should not make inferences about what cannot be observed.
As a matter of fact, Poe was only partly correct in his assessment of the
chess player’s mechanism—yes, there was somebody concealed within the
cabinet guiding the movements of the chess player’s mechanical arm, and
his or her remaining concealed during Maelzel’s exposing of the cabinet’s
interior did depend on movable partitions. But what concerns us here is
less the final accuracy of his inferences than the way the staged effect of
Maelzel’s performance serves as a kind of template for Poe’s later devising
the constraint of “commencing with a consideration of effects.” In making
explicit the link between the mechanical chess player enchanting spectators
with the effect of artificial intelligence and a kind of magazine writing
that would entice potential readers of the Southern Literary Messenger, Poe
translates the effect of the exhibition into a writing designed to seize a
reader’s attention. This dynamic, which James Berkley (2004: 369) calls
Poe’s “tak[ing] up and usurp[ing] the sublime theatricality that had previously
belonged to [the exhibition of ] Maelzel’s Chess Player,” is, as I will
show below, the germ for Poe’s reverse constraints.
It is just this dynamic—the way the difference between Babbage-like
necessity and real chess moves can be manipulated for the creation of a
theatrical effect—that animates some of the key debates in a more recent
discussion about artificial intelligence. Describing the chess match in 1996
between IBM’s Deep Blue computer and Russian chess champion Gary
Kasparov, John Searle (1999) writes:
When it was first announced that Deep Blue had beaten Gary Kasparov, . . . I
suspect that the attitude of the general public was that what was going on inside
Deep Blue was much the same sort of thing as what was going on inside Kasparov.
. . . [But, unlike Deep Blue,] Kasparov was consciously looking at a chessboard,
studying the position and trying to figure out his next move. He was also
planning overall strategy and no doubt having peripheral thoughts about earlier
matches, the significance of victory or defeat, etc. . . . [he] was, quite literally,
playing chess. None of this whatever happened inside Deep Blue.
In contrast with the description of what Kasparov was doing when he was
playing chess, this is how Searle (ibid.) describes Deep Blue:
11. For a detailed description of the truth behind Maelzel’s illusion, see Standage 2002:
12. Searle’s example is a slightly altered version of the “Chinese Room” argument from
Imagine that a man who does not know how to play chess is locked inside a
room, and there he is given a set of, to him, meaningless symbols. Unknown to
him, these represent positions on a chessboard. He looks up in a book what he is
supposed to do, and he passes back more meaningless symbols. We can suppose
that if the rule book, i.e., the program, is skillfully written, he will win chess
games. People outside the room will say, “This man understands chess, and in
fact he is a good chess player because he wins.” They will be totally mistaken.
The man understands nothing of chess, he is just a computer. The point of the
parable is this: if the man does not understand chess on the basis of running the
chess-playing program, neither does any other computer solely on that basis.
Searle couches his concern over the confusion of “understanding” with
the running of a program (in this case, a man in a room converting symbols
from a book into chess moves) as something an audience presumes
to be going on “inside” Deep Blue. For Searle, it is a mistake to imagine
the complexity of human neurochemistry as being synonymous with the
manipulation of “meaningless” tokens. Poe’s analysis of Maelzel’s chess
player is also a matter of making the distinction between a “pure machine”
that would win at chess due entirely to a formal system of symbol manipulation
and the theatrical effect of such a machine. This effect is something
close to what those standing outside the room in Searle’s example perceive
when they see chess moves made upon a chessboard. Deep Blue, perpetually
translating “meaningless [to it] symbols” into moves in the game,
creates the illusory effect of somebody (or something) that “knows” how to
play chess, just as Maelzel’s exhibition—as the word exhibition literally indicates—“
showed mechanism without itself being mechanical, and provoked
evaluation of the secret workings of the machine, beyond the spectacle of
its effect” (Sussman 1999: 83). Whatever the relative abilities of Maelzel’s
chess player and Deep Blue, then, the link between Poe and Searle is their
having in common the way theatrical artifice works to generate the effect
of a machine that thinks.
This resemblance between their descriptions of Maelzel’s chess player
and of IBM’s Deep Blue nevertheless arises from a very important differ-
his essay “Minds, Brains, and Programs” (1990). The essay is a critique of what Searle calls
“strong AI,” the claim that intelligence is formal symbol manipulation. Searle thinks that
no system of formal symbol manipulation can amount to intelligence understood as belief
and intentionality. For some counterarguments to Searle’s critique of “strong AI,” see Boden
1990; Copeland 1993.
13. An important practical difference here between Poe and Searle is that Searle is talking
about an algorithm that really can play good chess, whereas Poe had no concept of something
as sophisticated as Deep Blue (with processors able to make about 200 billion calculations
per second). But in 1836 Poe was justified in his skepticism about such a mechanical
For Poe, winning at chess serves as a valid example of thinking (as
something Babbage’s machine cannot do), whereas for Searle, winning at
chess might be merely the result of a program translating symbols into
positions on a grid. Unlike Deep Blue, Kasparov was “consciously looking
at a chessboard, studying the position and trying to figure out his next
move . . . planning overall strategy . . . having peripheral thoughts about
earlier matches, the significance of victory or defeat,” all things that the
IBM machine does not do, because it is not (and cannot be) programmed to
do them. Thus winning at chess is not, for Searle, an example of thinking,
whereas simply playing chess is, with all of the errant, peripheral thoughts
that would accompany such an activity (like “What time were we supposed
to meet?” or “I wonder what this chessboard is made of ?” or “I smell
smoke; is the building on fire?”—all cognitive operations of which Deep
Blue would be incapable). While Poe was, as Shaun Rosenheim (1997: 100–
101) points out, “striving toward a notion of artificial intelligence” in his
essay on the chess player, by the time we get to Searle the effect of thinking
shifts from a surface illusion of intelligence to machines that actually play
chess (and win, even against human world champions) but for all that do
not “understand” chess.
At issue in the convergence, and divergence, between Poe and Searle
is the relation between artificial intelligence and what we might call artificed
intelligence. Whether you are standing outside the sealed room saying
“this man understands chess” or are convinced by the performance of a
chess-playing automaton, in both cases the question of what is to count as
evidence of intelligence is bound up with the observable effects of intelligence.
This is exactly what Poe takes up from the performance of Maelzel’s
chess player and will later use as the basis for the constraint he claims
has guided all of his compositions. If audiences remained captivated by
the illusion of a piece of mechanism able to do things that they felt only a
human being could do (an issue that continues into the present, as we’ve
seen with Searle’s argument), then Poe sees this as an opportunity to dazzle
his readers by presenting himself as a poetry-making machine.
Just prior to the publication of his poem “The Raven,” Poe published an
essay, “A Chapter of Suggestions,” in which he proposed an idea for a new
kind of magazine essay:
An excellent Magazine paper might be written upon the subject of the progressive
steps by which any great work of art—especially literary art—attained
completion. How vast a dissimilarity always exists between the germ and the
fruit—between the work and its original conception! Sometimes the original
conception is abandoned, or left out of sight altogether [but] pen should never
touch paper, until at least a well-digested general purpose be established. In
fiction, the dénouement—in all other composition the intended effect should be
definitely considered and arranged, before writing the first word: and no word
should be then written which does not tend, or form a part of a sentence which
tends, to the development of . . . the strengthening of the effect. (Quoted in
Levine 2009: 56)
When “The Philosophyof Composition” appeared in Graham’s in April 1846,
much of the material of this essay (which appeared in the Opal in 1845)
was reworked for the larger audience Poe had gained with the intervening
publication of “The Raven,” a poem that brought him wide fame. Levine
(ibid.: 55, 59) writes that it was a “crowd-pleaser [and its] fame [was] probably
critical [to Poe’s] decision to write an essay on how a poem is made,”
and Richards (2004: 53) says that “The Philosophy of Composition” “rode
the wave of [‘The Raven’s’] success.” Such descriptions make clear Poe’s
keen sense of an expanded audience for the just previously published idea
about an essay that would explain the compositional process.
Modifying slightly, but tellingly, the language of “A Chapter of Suggestions”
for his now more confident performance in front of the audience he
had amassed with “The Raven,” Poe (1984d : 14) begins his account
in “The Philosophy of Composition” as follows:
I have often thought how interesting a magazine paper might be written by any
author who would . . . detail step by step, the processes by which any one of his
compositions attained its ultimate point of completion. . . . Most writers—poets
in especial—prefer having it understood that they compose in a fine frenzy—an
ecstatic intuition—and would positively shudder at letting the public take a
peep behind the scenes, at the elaborate and vacillating crudities of thought . . .
in a word, at the wheels and pinions—the tackle for scene shifting—the step ladders
and demontraps—the cock’s feathers, the red paint and the black patches,
which, in ninety nine cases out of the hundred, constitute the properties of the
Poe’s rhetoric here, in which literary production is propelled by the “wheels
and pinions” of hidden contraptions and the “tackle for scene shifting” of
the theater, recasts Maelzel’s mechanical spectacle into the backdrop for
his constraint of effects. But something curious occurs in this comparison.
What Poe (ibid.: 13) had derided as mere “calculation” in the essay on the
chess player has now been elevated to the process by which he claims to
have made “The Raven”:
I prefer commencing with the consideration of an effect. Keeping originality
always in view . . . I say to myself, in the first place, “Of the innumerable effects,
or impressions, of which the heart, the intellect, or (more generally) the soul is
susceptible, what one shall I, on the present occasion, select?” Having chosen
a novel, first, and secondly a vivid effect, I consider whether it can best be
wrought by incident or tone—whether by ordinary incidents and peculiar tone,
or the converse, or by peculiarity of both incident and tone—afterward looking
about me (or rather within) for such combinations of event, or tone, as shall best
aid me in the construction of the effect.
The “combination of event and tone” that Poe “selects” here, I would
argue, is the one he remembers being impressed by in Maelzel’s exhibition,
but meanwhile he has inverted the very distinction that had earlier
enabled him to debunk the machine as a fake. If Poe, as Tresch (1997: 288)
says, “claims his own mode of poetic production is just as mechanical as
that by which Babbages’s calculator produced his table,” then here Poe
both redeploys the idea of determinate necessity and uses it to conceal the
chain of tentative experiments that have led to the essay of compositional
theory. And when he says that he “look[ed] about [him] (or rather within)”
for a place to begin, he further compounds observational sense perception
with the “within” of a rational starting point, again fusing a Babbage-like
rationalism with the act of “selecting” from the amassed devices of his
own earlier essayistic experiments. Even more strikingly, in reusing, with
only slight modification, an entire paragraph from “A Chapter of Suggestions”
for “The Philosophy of Composition,” and at just the moment when
he reintroduces the idea for a magazine piece that would detail the steps
by which a literary work is made, Poe replaces an all-too-human process
with the effect of a completely determined process. Indeed, Poe’s modifications
of the earlier version are precisely the figures of mechanism (“wheels
and pinions”) and theater (“tackles for scene shifting”) that he picked up
from Maelzel’s performing automaton (and which anticipate Searle’s man-
computer transmitting “meaningless symbols” that look like chess moves
to the spectators outside the sealed room). At the same time, the earlier
version’s admission of the “dissimilarity [which] always exists between the
germ and the fruit” of a compositional idea is replaced in “The Philosophy
of Composition” with a desire to show in detail how a work “attained its
ultimate point of completion” through rational control rather than through
the “fine frenzy” of “inspiration.”
In finding new use for earlier published material to be presented before
a larger audience, Poe’s showman-like instincts tend toward the concealment
of the “vacillating crudities” of earlier trials through the claim that
he has composed “The Raven” with the “precision and rigid consequence
of a mathematical problem.” Rather than dismissing this last claim as a
piece of strident hyperbole, we should read it against Poe’s (1984b :
1255) earlier assessment of Babbage’s calculating engine: “The question
to be solved proceeds, or should proceed, to its final determination, by a
succession of unerring steps liable to no change, and subject to no modification.”
So if the lesson in logic Poe dramatized in the Maelzel essay was
that one ought to make inferences based on observation alone (rather than
beginning, like Mr. Blackwood and Babbage, with axiomatic premises),
then in “The Philosophy of Composition” that insight is reversed so that
Poe can turn his own process of composition into something like the working
out of a “mathematical problem.” All of this puts Poe “in a position
directly analogous to that of the exhibitor of the chess-playing automaton”
(Tresch 1997: 289), only here he’s reversed the logic earlier used to debunk
the machine as a hoax—that predetermined calculations could never be
enough to get a machine to win at chess—in order to create his desired
effect of having composed “The Raven” in accordance with rules postulated
in the first place.
Following the precedent of Maelzel to the end, Poe thus perpetrates a
hoax. Like Maelzel narrowly masking the sound of a hidden chess master’s
sneeze with a crank of ersatz clockwork, Poe (1984d : 17) divines his
starting point a posteriori:
I betook myself to ordinary induction, with the view of obtaining some artistic
piquancy which might serve me as a key-note in the construction of the
poem. . . . I did not fail to perceive immediately that no one [artistic effect]
had been so universally employed as that of the refrain. The universality of
its employment sufficed to assure me of its intrinsic value, and spared me the
necessity of submitting it to analysis.
“Ordinary induction” is here illusorily raised to the level of a “universal”
rule, but the fact that such a posteriori inferences are always open to
revision, and so cannot serve as immutable universals, in no way makes
Poe’s constraint of “commencing with the consideration of an effect”
inconsistent with his overall compositional practice. For the sleight of hand
by which an ordinary induction is elevated to the status of a timeless axiom
is precisely the effect Poe hopes to create for the audience of “The Philosophy
of Composition.” Just as Poe invokes a Babbage-like determinism to
conceal the chain of writing experiments that have led up to this essay, so
here he consolidates and makes explicit that effect in turning an “ordinary
induction” into a “universal” rule. And just as Maelzel opens the front of
the cabinet box and rotates it 360° on its iron castors to assure his audience
that no person is concealed within it, only to allow the chess master
crouched inside to shift in sync with mechanically contrived partitions, so
Poe will show us the “wheels and pinions” that churned out “The Raven”
while dramatically unveiling its “commencement” after the fact.
Poe’s description of the individual steps by which he composed “The
Raven” shows this reversal of constraint at work so as to foreground the
precedent of Maelzel’s machine (and of Poe’s prescient prefiguring of more
recent debates in artificial intelligence) in yet another way. If an algorithm
is “a mechanical procedure for solving a problem in a finite number of
steps” (Nelson 2008: 7), then Poe (1984d : 14) echoes this definition
in describing how he wrote his poem “step by step,” in accordance with
an algorithm that might look something like this:
1. Intend to compose a poem that should suit at once the critical and the
2. Conceive the proper length of the poem.
3. Design the poem in accordance with the universal rule that Beauty is
the sole legitimate province of the poem.
4. Select the tone of the poem in accordance with the rule that melancholy
is the most legitimate of poetic tones.
5. Use the device of the refrain in the poem.
6. Apply the refrain throughout the poem.
7. Select the phoneme ‘or’ as a sound for the refrain.
8. Select ‘Nevermore’ as a word embodying this sound.
9. Select a raven as a non-reasoning creature capable of uttering the
10. Select the death of a beautiful woman as the most melancholy topic
according to the universal understanding of mankind.
11. Combine the idea of a lover lamenting his deceased mistress and a
raven continuously repeating the word “Nevermore.”
12. Begin to compose “The Raven.” (Ibid.: 14–20)
There could be no more aggressive overturning of the romantic shibboleths
of organic expressivity than this recipe for cranking out poetic effects.
As a “healthy corrective to over-romantic portrayals of the poetic process”
(Levine 2009: 57), such a procedure is a matter of moving mechanically
through a series of steps. While we are told we are reading an essay about
the assembly of “The Raven,” in fact we are—if we pay close enough
attention to Poe’s literary illusionism—witnessing a literary performance in
which what appears to be a poetry-making algorithm in fact conceals the
tentative experiments of a working magazinist—the author of “Maelzel’s
Chess-Player,” “How to Write a Blackwood Article,” the review of Hawthorne’s
Twice-Told Tales, “The Raven,” “A Chapter of Suggestions,” and
“The Philosophy of Composition.” Far from an airtight sequence of determinate,
algorithmic steps, these steps amount, as Levine (ibid.) has put it, to
“a lot of bumbling efforts, false starts [and] failed experiments.”
If the success of “The Raven” led to a (relatively) large audience eager to
catch a glimpse of the workshop of a famous poet, it accordingly provoked
Poe into wanting to stage himself as an uncanny poetry-making machine.
And this, then, is just the “effect” he wanted to create in “The Philosophy of
Composition.” That the essay purports to attribute such “effects” to “The
Raven” and not explicitly to the essay explaining its composition (as I am
claiming it does) does not make it any less accordant with Poe’s proposed
constraint. Indeed, the essay becomes all the more striking, since Poe adds
to the effect of presenting himself as a poetry-making machine the further
effect of claiming to tell us how he made “The Raven,” when in fact he is
using the poem as a stage for a striking essay of literary theory in which he
presents himself as a kind of literary computer. Thus the whole move from
“The Raven” to “The Philosophy of Composition” creates a discontinuity
between the mournful lyricism of the poem and the machinelike detachment
of the essay, such that, as Richards (2004: 53) puts it, the “dramatic
difference between his critical and poetic voices enhances his authorial
mystique.” We might say that this difference heightens the “effect” of his
I agree, then, that this operation “reverses constraint” (Morisi 2008: 113).
But to say that because of this reversal Poe’s essay “does not seem to call
for a comparison with an Oulipian enterprise that meticulously defines its
literary constraints before applying them with rigor” (ibid.) is to deny that
illusions like these could count as “effects” in the sense Poe uses the word—
as a guiding rule for composition.
The specific question I have tried to raise here is whether Poe’s line—“I
prefer commencing with the consideration of an effect”—can be thought of
as an Oulipian constraint, and I think it can. To the argument that Poe does
not use constraints in the conventionally Oulipian way—since he declares
such constraints only after the work is finished—I have responded that part
of what Poe means by “effect” is the way the reader is led to believe that
“The Raven” had been so composed; that we suspect the explanation of
“The Philosophy of Composition” to be hoax only heightens the effect.
Poe’s essay “Maelzel’s Chess-Player” is the precedent for such “reverse”
constraints, I have argued, because he sees in the exhibition the way a surface
artifice of a priori axioms (artificial intelligence) conceals a series of a
posteriori inferences (a person crouched inside the machine making chess
moves that do not necessarily follow). For Poe, this means turning a decade
of toil in the columns of American magazines into the cool announcement
that his compositions are the result of a step-by-step algorithmic procedure.
Finally, if it has been throughout a logic of the hoax that has led Poe to
his reverse constraints, then he is the “anticipatory plagiarist” of a further
feature of Oulipian practice, as François Le Lionnais (1973: 18) described it
in the first Oulipo manifesto: “When they are the product of poets, amusement,
farce and hoax [supercheries] still belong to poetry. Potential literature
therefore remains the most serious thing in the world.”
Allen, Michael L.
1969 Poe and the British Magazine Tradition (New York: Oxford University Press).
2004 “Post-Human Mimesis and the Debunked Machine: Reading Environmental Appropriation
in ‘Maelzel’s Chess Player’ and ‘The Man That Was Used Up’,” Comparative
Literature Studies 41 (3): 356–76.
1990 “Escaping from the Chinese Room,” in The Philosophy of Artificial Intelligence, edited by
Margaret Boden, 89–104 (Oxford: Oxford University Press).
1990 Une histoire de l’informatique (Paris: Seuil).
1966 Language as Symbolic Action (Berkeley: University of California Press).
2003 Don Quixote, translated by Edith Grossman (New York: Ecco).
1993 “The Curious Case of the Chinese Room,” in Artificial Intelligence: A Philosophical Introduction,
121–32 (London: Blackwell).
1998 Brainchildren (Cambridge, MA: MIT Press).
1999 What Computers Still Can’t Do (Cambridge, MA: MIT Press).
1986 Artificial Intelligence: The Very Idea (Cambridge, MA: MIT Press).
1972 Poe Poe Poe Poe Poe Poe Poe (New York: Avon).
Le Lionnais, François
1973 “La Lipo (Le premier manifeste),” in Oulipo 1973: 15–18.
1 Poetics Today 31:1
2009 “Notes on ‘The Philosophy of Composition,’” in Edgar Allan Poe, Critical Theory: The
Major Documents, edited by Stuart Levine and Susan F. Levine, 55–60, 71–76 (Urbana:
Illinois University Press).
Mathews, Harry, and Alastair Brotchie, eds.
2005 Oulipo Compendium (London: Atlas).
McKeon, Richard, ed.
1941 The Basic Works of Aristotle (New York: Random House).
Morisi, Eve Célia
2008 “The OuliPoe; or, Constraint and (Contre-)Performance: ‘The Philosophy of Composition’
and the Oulipian Manifestos,” Comparative Literature 60 (2): 107–24.
Morrison, Philip, and Emily Morrison, eds.
1961 Charles Babbage and His Calculating Engines (New York: Dover).
Motte, Warren, ed.
1986 Oulipo: A Primer of Potential Literature (Lincoln: University of Nebraska Press).
Nelson, David, ed.
2008 Penguin Dictionary of Mathematics (London: Penguin).
1973 La littérature potentielle: Créations, recréations, recreations (Paris: Gallimard).
1969 La disparition (Paris: Gallimard).
Poe, Edgar Allan
1978a Tales and Sketches. Vol. 1, 1831–1842, edited by Thomas O. Mabbott (Urbana: Univer
sity of Illinois Press).
1978b  “How to Write a Blackwood Article,” in Poe 1978a: 336–57.
1984a Essays and Reviews, edited by G. R. Thompson (New York: Library of America).
1984b  “Maelzel’s Chess-Player,” in Poe 1984a: 1253–76.
1984c  “Twice-Told Tales (review of Nathaniel Hawthorne),” in Poe 1984a: 568–88.
1984d  “The Philosophy of Composition,” in Poe 1984a: 13–25.
2004 Gender and the Poetics of Reception in Poe’s Circle (Cambridge: Cambridge University
1997 The Cryptographic Imagination (Baltimore: Johns Hopkins University Press).
1986 “Mathematics in the Method of Raymond Queneau,” in Motte 1986: 79–96.
1990 “Minds, Brains, and Programs,” in The Philosophy of Artificial Intelligence, edited by
Margaret Boden, 67–88 (Oxford: Oxford University Press).
1999 “I Married a Computer,” New York Review of Books, April 8, http://www.nybooks.com/
1991 Edgar A. Poe: A Biography (New York: Harper Perennial).
2002 The Turk: The Life and Times of the Famous Eighteenth-Century Chess-Playing Machine (New
1999 “Performing the Intelligent Machine: Deception and Enchantment in the Life of the
Automaton Chess Player,” Drama Review 43 (3): 81–96.
2000 The Cogwheel Brain: Charles Babbage and the Quest to Build the First Computer (London:
Grimstad • Antebellum AI 1
1997 “‘The Potent Magic of Verisimilitude’: Edgar Allan Poe within the Mechanical Age,”
British Journal for the History of Science 30: 275–90.
Vines, Lois Davis
1992 Valéry and Poe: A Literary Legacy (New York: New York University Press).
Wimsatt, William K.
1939 “Poe and the Chess Automaton,” American Literature 11: 138–51.
(Courtesy of Paul Grimstad Yale University)