“Trust only your eyes”.
“Reality is never’what we might believe it to be’: it is always what we ought to have thought”.
1L’Illusion comique is as clear as an open book; there is no mystery to it: nothing that cannot be seen if we keep our eyes open. And we can take seriously the change to the title that Corneille made for the 1660 edition of the first volume of his Œuvres, in which L’Illusion comique becomes simply L’Illusion. Far from the theater (to which the adjective “comique” applies) being relativized by this simplification, it is rather generalized as an object of philosophical reflection in the form of a crystallization of thought that only the stage—in its verbal, optical, and temporal aspects—can deploy.
2Pierre Corneille’s L’Illusion not only allows us to experience illusion and to be dazzled by it, but it is also a brilliant object lesson on the possibility, and simultaneously the dissipation, of illusion. By understanding the mechanism of illusion, we discover that everything had always been, from the very beginning, right before our eyes: we therefore understand when we understand why we had not understood, and why we should have understood.
3The device of theatrical presentation does not work well as a didactic vehicle, that is, as a demonstration whose substance is delivered via a distinct explanatory discourse. Its value lies in being, in itself, an object of reflection, inseparable from the two constitutive dimensions of theater. First the stage, both in terms of what separates it from and connects it to the auditorium, and in terms of its depth. Second, the dimension of temporality. For it is, as we shall see, the second step that brings out the moral dimension through its retroactivity: the dimension of what we should have seen. The play is over when, once the depth is deployed, everything experienced up to this point comes to precipitate itself retrospectively in the only instant that matters: the sudden moment of intellection.
The device of illusion and its dissipation
4As we know, the play depicts a father (Pridamant) in search of his missing son (Clindor). Pridamant turns for help to the magician Alcandre, who brings Pridamant into a cave and presents an episode from his son’s past in the form of “phantoms”—characters who are explicitly designated as fictional (including the stock character Matamore) and tell stories that are not themselves depicted. Then Alcandre shifts to a higher gear and “shows” Pridamant what is happening at the present time: Clindor is caught in a tragic sequence of actions that end with his murder. But, just as the events draw to their bloody conclusion, a curtain is raised and we realize that the episode was actually a play performed by Clindor, who has become an actor. We therefore need to add an extra level to the degrees of fictional depth. Freed from the illusion that led him to mistake fiction for reality, Pridamant “returns to his senses” and simultaneously finds himself freed from his previous prejudice against actors.
5The lesson is given by means of the very deployment of the theatrical device. If we are to understand it, we therefore need to present a summary of the play in a different way. For the things that Pridamant experiences on the stage are also experienced at the same time by the spectator in the audience, and therefore with an additional fictional distance, since the viewer knows that he is at the theater. This additional step draws attention to the nesting of fictional levels. Far from weakening the effect of the demonstration, it strengthens it and reveals one of its most profound and unsettling aspects: can this nesting, or series of fictional levels, continue ad infinitum?
6Let us, then, reconsider the schema of the play including this first level.
I am at the theater. I see Pridamant in search of Clindor. Together with Pridamant, I see, by means of Alcandre’s “magic,” an episode from Clindor’s past. Next I see, still together with Pridamant, Clindor being murdered in the present time. Then I discover, with Pridamant, that this murder was fictional: Clindor was performing a dramatic scene. I return to my senses and I understand how the illusion has been able to take place—and not only the theatrical illusion.
8For Pridamant and myself as spectator (since, for the reader, the levels of depth are set out explicitly at the start of Act V Scene 2, where the stage direction is made clear), two fictional levels are flattened into one at the start of Act V. And it is only when the “curtain” is raised that we understand that there was depth where we had seen only flatness. The pleasure offered by the play is therefore found simultaneously in illusion and its dissipation, in a retrospective movement that is characteristic of any correction of error, which involves a temporality of coming to knowledge at the same time as a moral dimension, that of “I should have seen it earlier.”
9Since—this is a crucial point—if I had been capable of hearing what was said and seeing what there was to be seen, I would not have been fooled by the illusion. This theme is introduced by Corneille at two points in the play.
10First, a warning, in Act I Scene 1, when Dorante informs Pridamant about the nature of Alcandre’s magic: it is not a magic of spectacle, of the supernatural, but an effect of immanent reality: “you have no need of such miracles.”  Alcandre does not contradict nature, he limits himself to “inverting” it—a term also used in optics. In other words, we are not witnessing a “spectacular” play such as tragedy with its stage machines, or such as the genre of opera would become. Everything is done naturally and not through artifice.
11Next, a clue, as sumptuous costumes, such as those we see at the theater, are shown: “He waves his wand, and a curtain is raised, behind which the finest actors’ clothes are on display” (I, 2).
12This prepares, and makes understandable in retrospect, the very act of revealing the mechanism of illusion: the curtain raised in Act V reveals that the scene of murder we have just witnessed is a theatrical representation in which Clindor participated as an actor.
13Finally, the last scene of Act V is the scene of disillusionment, both optically and morally, for Pridamant: “trust only your eyes,” Alcandre tells him.
14In order to properly situate the concept of illusion here, we must separate the optical from the spectacular. In fact, the play of optics, both natural and simple, but which allow “inversions,” is necessary for producing and dissipating the illusion. But they are sufficient: everything is right there before our eyes. What happens to Pridamant also happens to us at the same time, and the same thing happens to us every day.
15Everything happens through natural means that are within our grasp: all Alcandre has done is to direct our perception, without changing it or disturbing it. The means that are used are those of the theater, without stage machinery, and natural optics, which are shared, like common sense, by the spectators and characters in the play. Everyone is on the same level: all of us have only the evidence of our own eyes, which we must believe and use well in order to see what there is to see.
Natural, simple, and straightforward means: An ultra-classical play
16Let us return to these means of natural optics and classical dramaturgy. To that end, let us formalize the levels of fiction by representing them in a cross-sectional diagram (see the box below):
R0 designates the position of the spectator (R for reality). L1, L2, and L3 designate the different levels of fiction nested one inside the other. The diagram is a section of the visual apparatus that structures the play, with the eye being situated on the left at R0 (the spectator) then in L1 (Pridamant and Alcandre) and looking toward the right. The names in italics designate the agents who remain attached to a single level of fiction. X and Y are Clindor’s fellow theater troupe members, Isabelle and Lyse, but who are not identified because we only see them in their roles as Éraste and Rosine.
The illusion rests on the temporary flattening of L2 and L3 onto a single level: the spectator and Pridamant “see” the level of L3 flattened onto L2. Only the reader knows from Act 5 Scene 2 that there is actually a third level of fiction, because he reads the explanation of the visual structure and its potentially deceptive nature.
In Act 5, the raising of the curtain allows Pridamant’s eye (placed on the level of L1), at the same time as the spectator’s eye (on level R0), to reinstate the depth that separates L2 and L3, and to understand that L3 is a fiction nested within L2. But, in hindsight, we also realize that we had the means to understand this depth from the beginning, because of the clue given in Act I Scene 2 (a curtain is raised, revealing costumes).
Finally, this discovery leads to one more question, still from a perspective of critical retrospection: if the eye looks behind itself, this begs the disturbing question as to whether there might be a level that we could call R-1, behind level R0 and further to the left, from which an eye would be able to view our own reality as a fiction, and in which we would be like Pridamant. Why, in the final instance, is this not a possibility?
17The distinction between the spectacular and the optical is made explicit in this play, which, in this respect, is classical in the sense that it systematically avoids the baroque spectacular that confuses levels of representation. The fantastical narratives delivered by characters (notably Matamore) are recounted, spoken, and not “hallucinated” on the stage as they could be in a play using stage machinery or in a pastoral, and as they would be later in the opera. Here, the dramatic fiction dispenses with the fantastical or supernatural spectacle.
18The very bodies of the characters, that is, of the actors themselves—their face, voice, movement, and bodily form—are an anchor to reality. It is essential that Clindor, Isabelle, and Lyse remain identifiable throughout the play, of which they inhabit two levels of fiction, and for this identification to also be capable of supporting the ambiguity.  The play is made precisely to be seen by a spectator, since for the reader there is no illusion.
19There are three thresholds, two of which are mobile:
212. The threshold between cave and L2. It must be “sustained” by Alcandre, otherwise the fictions of L2 and L3 (initially conflated with L2) are not possible. This threshold is removed in Act V: “they leave the cave.”
223. The raising of the curtain or backdrop in the theatre (the difference between L2 and L3). This curtain raised in Act V allows us to reinstate, retrospectively, the depth whose flattening had established the illusion. But we should remember that Alcandre revealed it to Pridamant in Act I by revealing the costumes. We therefore had sufficient information to establish this depth, but we did not see it. This is the mechanism of dreams in Freud’s Traumdeutung:  a small detail gives us a “wink” but passes without being noticed, and it is only when we return to it that we see it. Informed vision is always a delayed vision.
23This great simplicity in the sole use of natural means makes L’Illusion a play about certainty, inasmuch as this certainty is constituted by uncertainty. The moment when I understand the mechanism capable of blurring the various levels is the very moment when I can see and understand their distinction: one who has never confused levels will never know what distinct levels are. One who has never been fooled will remain foolish all his life. Uncertainty and falsity do not impede certainty and truth, and are not even their regrettable companions: they are instead the necessary operators for certainty and truth. Knowing is not a state of tranquility nor the result of peaceful accumulation, but an uneasy operation in which it is necessary to take the wrong path.
24Vision is the operation of an eye that is first “surprised.” This surprised eye is, in the first instance, “curious,” astonished: the moment of curiosity is that of juxtaposition, of the confusion of levels, of the flatness that presents itself as wondrous. We must progress to the moment of the “methodical eye”  for the confusion to take shape, and pass from the order of matters to the order of reasons. Vision is the paradigm for this journey: to see is not sufficient, to see we must see that which makes vision possible. The regressive moment of retroaction is the condition of the deployment of this device. Therefore, the path of coming to knowledge is inseparable from a story that we must constantly relive. In order to understand, we must begin by being deceived: error not only comes first chronologically, it must also come first epistemologically. Here we find some of the key ideas from the psychology of knowledge developed by Bachelard.
25The aesthetic consequence of this redeployment in cognitive time and in a hierarchically structured space is not negligible. The “baroque” aesthetic is mobilized here, but held in contempt, shown for its flatness, its profusion, and its illogical regime of juxtaposition, and then reordered by the economy of vision according to other, simpler principles of intelligibility. Corneille places this baroque moment in his play as if in a jewelry box, which reveals its deceptiveness and shows its worthlessness as well as its superficial sparkle: once the box is opened, we understand that the baroque pearl is artificial. Considered in terms of its contents, the play is baroque (narratives, excessive events, changes of location, stretching out time). But considered in terms of the viewpoint that it establishes, and if we take its title seriously (L’Illusion and not, as in 1639, L’Illusion comique), it is regular and classical: unity of place, and perfect unity of time (the duration of represented events and the actual duration of the play completely coincide).  The play contains baroque illusion while dismantling it: there was really no reason to be dazzled. Curious wonder is followed by the disillusioned amazement appropriate to classical reason: “that’s all there was to it!”
The philosophical moments of illusion and disillusionment
26We need merely analyze the mechanism of illusion to see that it is also the mechanism of disillusionment. However, this mechanism functions every day, every time we correct an error, and as is the case here, it does so in a temporality of retrospection and a mode of “should have been.” This is the first lesson that we can draw from it: the correlate of coming to knowledge is doubt. Certainty—unlike being sure—is subject to uncertainty. Someone who is sure of something will never think to establish its certainty. In order to demonstrate something, you must first be in doubt.
27But another proposition, a more disturbing one, can be drawn from this theatrical journey: this is the immanence of our knowledge. We have only our eyes with which to see, only our thoughts with which to understand, and no magician, no external interpreter can give us the truth, which can only be constructed with great effort, without being able to avoid the stage of error, of the false path. No god presides over our coming to knowledge, nor any evil genius: I must do it myself. We could also conclude, more strongly again, that the very idea of infallible knowledge is absurd, and therefore that God, who cannot be mistaken, has no knowledge in the strict sense and cannot understand what knowledge is.  Someone who has only ever seen light and never shadows would not even know what light is.
28In Corneille’s play, Alcandre enjoys no privilege: all he does is to reveal to sight its own conditions, the mechanism that makes it see, and that it can see poorly. But this mechanism was itself visible. Furthermore, it can only be deployed in time—eternity would be of no help to it, on the contrary, since it would abolish the moment of being mistaken and therefore also the moment of understanding. Only retrospection gives it its validity. And just because one has been mistaken once, does not mean that one is henceforth protected from error: one can (one must) be mistaken each time in order to be able to understand anything at all.
29Let us consider more closely how this uncomfortable self-revelation functions. We can identify three moments in it.
1 – The moment of illusion and its revelation (or the moment of coming to knowledge)
30This moment can be analyzed in two steps. The illusion is produced by the canceling of depth, which is presented as a flatness. What Pridamant lacks—as do we spectators—in this moment, is a space that situates the forms of knowledge in relation to one another in a relation of the condition of possibility: an absence of principle, absence of order of reasons, a “curious” moment. Its revelation establishes a critical experience that reconsiders the very possibility of vision and of “false” vision: we understand that we have seen poorly, and we also understand why it was possible to see poorly. Our gaze is therefore not only “ahead,” but also “askew.” Not only do I see what I see, but I see how it is possible to see. When we correct an error, the depth is reinstated: we do not see A then B while saying “B is truer than A.” No: in seeing B, we say “I now understand why I believed A to be true.” This critical experience can be schematized as a circulation in a hierarchically structured space (this is the schema of depth), but it also presupposes a circulation in time that informs the attitude of “now, I see,” resulting from previous errors.
31The critical moment is illuminating but frightening: because when we see that, we also see that we cannot escape from vision in order to attain once and for all an absolute, definitive perspective. We can always be caught in a false depth that has the appearance of flatness, because this is constitutive of all vision. This leads to the following moment.
2 – The hyperbolic moment
32I discover that coming to knowledge is a succession of acts of self-sequencing, operations of self-reflection, and I see that everything can be reordered without end. From this comes a doubt, a confusion, that is crystallized here in the character of Alcandre. Alcandre folds and unfolds perspectives before Pridamant’s eyes. But what if someone else were to do that for me too? The trick of the curtain that is raised up could also spread to the threshold between the auditorium and the stage, and why should this not continue to a “hyperbolic” vanishing point where what I take to be reality (life) would be nothing but a fiction: and what if I were someone else’s dream? We then find ourselves imagining, as a continuation of the optical schema, a space behind the level of R0 that would perceive R0 as a fiction, and so on ad infinitum. This is the Cartesian hypothesis of the evil genius in Meditations on First Philosophy, which imagines a level behind the subject from which the instrumentalization of his thoughts might be controlled, reducing him to the state of a manipulated object.
33However, what Descartes shows us is precisely that this is impossible. Nothing could remove me from my position as subject once the operation of my own reflection has considered itself as reflection, recognized itself while producing itself, with the result that I cannot be mistaken when I say “I think,” not even when I think an erroneous proposition: the reflective position absorbs all possible reflexivity. The optical paradigm offers the same device. Once I have understood what it is to see, I understand that there is no vision except that which I can put into practice: I can be mistaken, I can see poorly, but never with regard to the critical status of vision—thus the “I think” (Cogito) is the affirmation of this critical status. No matter how many levels I posit, they will all always presuppose an eye that sees them, that often confuses them, but is also capable of situating them, and nothing can spare me from the task of having to order them in relation to one another. The example of the reabsorption of reflection ad infinitum in the very operation of reflection is also considered by Spinoza in On the Improvement of the Understanding: to know that there is a mirror, I must know what a reflected image is, and from this point the regression of reflections is itself only thinkable around this end point that is itself the source of all reflection—knowing presupposes an instance that knows that it knows. 
34In other words, the immanence of vision to itself, the immanence of coming to knowledge and of thought to themselves are the guarantees of our discernment, at the same time as being the seat of error: no external power conceives of me as a fiction, since the critical mechanism consists precisely in knowing how a fiction is possible, but it is due to this very mechanism that error is possible, or even inevitable. There is no hyperbolic point behind the operation of coming to knowledge.
3 – The moral moment
35We then reach a third moment, which is essentially critical and moral. The dimension of morality establishes that only the work of self-reflection makes it possible, not to escape definitively from error, but to treat the question of error as constitutive of knowledge and as a dimension of action. This is a moral moment because it concerns a return to oneself, which the philosophical tradition terms a praxis, a reform. In the play, this moment coincides with that in which Pridamant frees himself from his prejudices against the theater and the condition of actors. As we say commonly, and with good reason: he has returned to his senses. Yes, we get over our errors and our prejudices by returning to them, returning to ourselves, and also turning back to a past that is then redefined—a little like a game of checkers, where we see the effects of our actions after each move. This return is moral in nature: it makes me see what I should have seen. This is what Bachelard calls a movement of “intellectual repentance.” The mind, he adds, “must be formed by being reformed.” 
36That is why we leave the performance of this play in a different state from the one in which we entered it.
37According to Christian Biet, L’Illusion can be considered as belonging to the category of epic theater in Goethe’s sense, and the sense that Brecht would give this term later on: this theater is made to be seen from several viewpoints, and when we leave it we have been “edified.”  It requires us to walk around it and receive it like a boomerang. What is interesting is not the plot, nor even the nesting of plots, but the reflective device that allows them to be nested. Furthermore, Corneille demarcates very clearly the firm boundary between fiction and reality, which also allows Pridamant to identify his own prejudices. Not, in my view, as Christian Biet argues, because Corneille was fearful of the possibility of madness,  but because of his insight into the nature of knowing, of vision, and into their conditions of possibility: it is therefore an epic theater because it is also a “transcendental” theater, a theater on the conditions of possibility of the theater, as well as of all vision and all knowledge of what is true and false.
On the different versions
38In the first version of L’Illusion comique (1639), Act V Scene 3 ends with the announcement by Lise-Clarine of the arrival of Princess Rosine. Clindor-Théagène then asks Isabelle-Hippolyte to secretly observe his conversation with Rosine. Scene 4 is a dialogue between Clindor-Théagène and his lover Rosine, during which Clindor-Théagène renounces his love and endures Rosine’s reproaches. They are surprised (Scene 5) by the arrival of Éraste and the servants of Florilame, who kill Clindor-Théagène and Rosine and take away Isabelle-Hippolyte, with whom the prince has fallen in love. Scene 6 involves Alcandre and Pridamant.
39In the version of 1660, Corneille has removed this Scene 4 and the character of Rosine. Scene 3 ends with the arrival of Éraste, who (Scene 4) murders Clindor-Théagène, Isabelle-Hippolyte dies (or faints?) on the body of Clindor-Théagène, and a curtain is raised at the end of Scene 4 (“Here we raise a curtain that covers the garden and the bodies of Clindor and Isabelle, and the magician leaves the cave with Pridamant”); we rejoin Alcandre and Pridamant in Scene 5 (the final scene), during which the curtain is raised (“Here we raise the curtain, and all the actors appear with their doorman, counting out money on a table and each taking their share”). The original version is therefore one scene longer than the version of 1660, and includes one additional plot detour. The shorter version of 1660 is also simpler and removes one character (Rosine) who belonged only to this level of the fiction.
40In the edition of 1639, in Act 5, the murder of Clindor-Théagène takes place in Scene 5. There is no direction for a curtain to be lowered between the end of Scene 5 and Scene 6, which involves Alcandre and Pridamant, and consequently there is no direction to raise this curtain in Scene 6 (the final scene). The explicit directions, which make the device completely clear, appear in the edition of 1644.
This article began as a lecture given at the Théâtre du Nord (Lille) in 2005. The first draft of the lecture is available online at: http://www.mezetulle.fr/lillusion-comique-de-corneille/.
Translator’s note: Unless otherwise stated, all translations of cited foreign language material in this article are our own.
The first line was changed by Corneille for the 1660 edition: “This magician whose art commands nature” becomes “This magician who inverts nature with a word.” In an excellent, unpublished article entitled “Certitudes dans l’illusion,” François Regnault writes: “In place of the complicated practice of his predecessors, Alcandre uses only his magic wand.” I thank the author for sharing this text with me.
Some characters do however remain attached to a single level of fiction: Matamore and Adraste in L2, Éraste in L3 (and also Rosine in the edition of 1639–44).
See Christian Biet, “L’avenir des illusion ou le théâtre et l’illusion perdue,” Littérature classiques 44 (2002): 175–214.
Sigmund Freud, Traumdeutung, II, Die Methode, in Sigmund Freud, Gesammelte Werke (Berlin: Fischer Verlag, 1973), vols. 2–3, 108.
These expressions are borrowed from Françoise Siguret, L’Œil surpris: perception et représentation dans la première moitié du XVIIe (Paris: Kincksieck, 1993), and from Victor Stoichita, L’Instauration du tableau: métapeinture à l’aube des temps modernes, 2nd ed. (Geneva: Droz, 1999 ).
See, on this point, Marc Fumaroli, “Rhétorique et dramaturgie dans L’Illusion comique de Corneille,” Dix-Septième Siècle 60–61 (1968): 107–32. Let us add to this the scene in which Clindor and Isabelle declare their love, which is distinctly different from the baroque style, and the various excesses that are parodied by Corneille.
We can note that this relation to time (“I have already been mistaken and that could happen again, but this very uncertainty is the mark of the possibility of my being certain”) is the same as that of the “generous soul” in Descartes’ Passions of the Soul; he maintains both self-esteem and self-distrust, and establishes the perfectibility of man by excluding the possibility of perfection.
Spinoza, On the Improvement of the Understanding; The Ethics; Correspondence, trans. R.H.M. Elwes (New York: Dover Publications, 1955).
Gaston Bachelard, The Formation of the Scientific Mind, trans. Mary McAllester Jones (Manchester, UK: Clinamen Press, 2002), 24 and 33.
For example, as in Peter Handke’s play Offending the Audience (Publikumsbeschimpfung), which leads the spectator to the point of madness.