Skip to main content
Recherche infructueuse

Références de l’article

Stéphane Lojkine, « L’Encyclopédie, édition et subversion », Diderot, une pensée par l’image, cours donné à l’université de Toulouse-Le Mirail, année 2006-2007.


Ressources externes

Encyclopedia frontispiece

"This work will surely in time produce a revolution in minds, and I hope that tyrants, oppressors, fanatics and intolerants will not gain from it. We will have served humanity; but it will be a long time before we are reduced to cold, unfeeling dust, when we will be thanked for it." (Letter to Sophie Volland, September 26, 1762.)

D'Alembert, in the "Avertissement des éditeurs" that opens the third volume of the Encyclopédie, defines the work he is presenting to the public as follows:

"The empire of Sciences & Arts is an irregular, imperfect, & in some ways monstrous palace, where certain pieces are admired for their magnificence, solidity and boldness; where others still resemble shapeless masses; where others finally, which art has not even sketched, await genius or chance. The main parts of this edifice have been built by a small number of great men, while the others either contribute materials or confine themselves to mere description. We shall attempt to bring these last two objects together, to trace the plan of the temple, and at the same time to fill in a few blanks. We will leave many others to be filled in; our descendants will take care of them, & will place the top, if they dare or if they can." (D'Alembert, "Avertissement des éditeurs", tome III, pp. vj-vij; GF I 217.)

The Encyclopédie is an "irregular palace", an unfinished "temple", an "edifice" without a "roof": from the architectural metaphor, traditionally used to designate the discursive object, D'Alembert moves on to the metaphor of the building site, with its never-filled "voids", and from there to the monster: for what opens up to us is "in some way monstrous".

In the same spirit, the Encyclopédie article signed by Diderot envisages the Encyclopédie as an unfinished work on raw materials, delivered formless to future generations to be continued there:

"If our nephews attend to the Encyclopédie without interruption, they will be able to lead the ordering of its materials to some degree of perfection. But, in the absence of a common and constant measure, there is no middle ground; we must first admit without exception all that a science includes, abandon each subject to itself, and prescribe for it no other limits than those of its object. [...] This shortcoming will diminish as the number of editions increases; knowledge will necessarily converge. I examine our work without partiality; I see that there is perhaps no sort of fault that we have not committed, and I am forced to confess that from an Encyclopedia such as ours, barely two-thirds would enter a true Encyclopedia. [...] the first edition of an Encyclopédie can only be a very shapeless and incomplete compilation." (Diderot, art. Encyclopédie, t. V, p. 641c1, 645a; ed. Versini, Bouquins, I 397, 414, 416.)

A work without "common and constant measure", the Encyclopedia abandons "each matter to itself". It is the very work of matter oriented towards "the ordering of its materials", according to a process that has only just begun and will never end. This is a far cry from the illusory orderings of Bacon's inspired tree, deploying his "division of sciences" in a harmonious organigram, in the illusion of an immobile knowledge, capable of being distributed in a logical, homogeneous and stable manner. The Encyclopédie is constituted from the subversion of these limits of taxonomic knowledge. We must not prescribe to the matter of knowledge "any other limits than those of its object", an object in perpetual becoming, as it were absent by the very work of the matter that always exceeds and repels it. The "true Encyclopedia" will first be constituted by this absenteeism, by the elimination of much of the current work2.

Diderot, like D'Alembert, is then led to deconstruct the architectural metaphor. From the Encyclopédie as a statue, he moves on to the immoderation of the colossus and, from there, referring to a famous passage from l'Art poétique, to the Chimera.

"Here we are bloated & of exorbitant bulk; there thin, small, petty, dry & emaciated. In one place we resemble skeletons; in another, we have a hydropic air; we are alternately dwarfs & giants, colossi & pigmies; straight, beneficent & proportionate; hunchbacked, lame & counterfeit. Add to all these oddities that of a discourse sometimes abstract, obscure or elaborate, more often sloppy, dragging & cowardly; & you will compare the whole work to the monster of poetic art, or even to something more hideous. But these defects are inseparable from a first attempt, & it is obviously demonstrated to me that it is up to time & the centuries to come to repair them." (Diderot, art. Encyclopédie, t. V, p. 641c, ed. Versini, Bouquins, I 396.)

As under D'Alembert's pen, the encyclopedic work in progress is here identified with the monstrous abjection of what is exhibited in cabinets of curiosity. Proportion, measure, harmony are now just another category of the world, a world that is no longer ideally ordered, a world inscribed in the becoming of history precisely because it gives itself, endlessly, to repair.

The double materialist spring of the encyclopedic Thing

So, paradoxically, there is no encyclopedic object. The dictionary is there, the alphabetical order is traversed, accomplished, the volumes of plates deploy an impressive array of Enlightenment technical knowledge. Yet there is no object. The Encyclopédie is not the closure of completed knowledge, the circumscription of a limited field of knowledge. It is an undertaking, a process, a movement towards the constitution of an object that always eludes it. While waiting for the unlikely constitution of the object, there are misfires, what Diderot calls bloat. Sometimes it's too long, sometimes it's incomplete, sometimes it's wrong. The Encyclopédie presents itself as that which, below the object, in the formless shape of the abject and fascinating monster, constitutes the Thing. It is "formless", the very work of "matter" abandoned to itself with no other limit than its "object". The object is thus defined as the horizon of encyclopedic matter, as its aim, as if the inverted stakes of this task were to elevate encyclopedic Thing to the dignity of object3.

The Encyclopédie is the Thing in more ways than one. Firstly, through this incompleteness of the object that defines the Encyclopédie as a process and, more precisely, as the work of epistemological slippage, as a decentering of knowledge. The Encyclopédie is slippage of the Thing rather than circumscription of the object.

Then there's the disturbing indeterminacy of its contours and, hence, its stakes: should it be seen as the objective sum of knowledge erected to the glory of France, the cultural monument serving the political institution in its own way? Or is it a vast enterprise of ideological subversion, in which knowledge is seized as a lever of cultural revolt before being political? Revolt is at work in the Encyclopédie, not so much in the form of a manifest discourse of revolt, but indirectly, in the reversal that takes place there of the tools of knowledge, not only concepts, notions, which take on new meanings, but the very languages of culture, brutally confronted with a reality that no mimetic distance meditates.

The Encyclopedic Thing is thus defined from the outset as slippage and revolt, i.e. as the work of the symbolic institution, at once the work of undermining and the driving force behind its renovation. There's a double spring in the Thing: in its relationship to the pre-established objects of knowledge, the encyclopedic Thing deconstructs, undoes, negates the gangue of inherited discourse; but in the movement that carries it towards the utopian culmination of an encyclopedic object, it orders knowledge according to a new device that is both semiological and ideological, a device that it doesn't accomplish, that it doesn't theorize, but that its work outlines in hollow.

This device is fundamentally dependent on the movement in which it is inscribed, on this dynamic of the Thing constitutive of the Encyclopédie: the Thing undoes discourse and reorders itself from res, pra`gma, the very materiality of the real. As such, we'll define the double spring of encyclopedic work as the principle of an epistemological materialism.

Assuming the historical legacy of materialism

Our aim today is not essentially to ask whether the Philosophers of the Encylopedia were or thought of themselves as materialists. Clearly, if the question of Diderot's materialism arose and still arises today, it is not in relation to the Materialistic heresy fought by Tertullian in his Treatise against Hermogenes, or even in relation to the materialistic interpretation of Spinozism in the philosophical circles led by Baron d'Holbach or Mme d'Épinay : if the metaphysical question of Enlightenment materialism goes far beyond mere scholarly interest, it's because of a certain historical filiation that we don't really know how to inherit today: dialectical materialism laid claim to the Enlightenment, and precisely to the encyclopedic enterprise. This filiation has not so much to do with a community of theoretical positions, as precisely with a certain relationship to reality, not only of militant involvement in the affairs of the city, but of investment of reality in theoretical reflection. The articulation of labor and value, which forms the problematic core of Book I of Capital, falls within this same double spring of the Thing, the process of which we attempt here to identify in the Encyclopedia. Marx deconstructs the discourses of wealth, denounces the speculations of bourgeois political economy as so many moral and political screens, to reach the Thing itself, fascinating and abject, the use value of things, the incommunicable singularity of worked matter, the infinite difference that inhabits labor; then he remodels wealth in economic terms, constitutes through exchange value things into objects of capitalist circulation, into commodities. This object, the materialist object, no longer has anything to do with the rhetorical object of old-world discourse. Defined as exchange value objectified by labor, wealth in Marx no longer manifests itself as discourse on the mirages of the Thing, in the manner of that tirade from the Timon of Athens quoted in Marx's note4, but as the Thing itself taken in its movement of transformation of matter by the industry of men, the process forbidding definitive distancing, the mimetic cut-off of a closed discourse separated from a closed object.

In this sense, analyzing the materialist decentering of the field of knowledge in the Encyclopédie refers back to the Marxist epistemological revolution, not because the philosophy is the same, but, once again, because a certain relationship to reality, a dynamic of the Thing has been established in the Encyclopédie, a non-discursive, or if we prefer, non-rhetorical rationality, which is visible to us only on condition that we assume this historical legacy by which the unconscious work of the encyclopedic Thing has become a reflexive and conscious process. To refuse this heritage in the name of an archaeological restitution of the authentic truth of texts is not only an illusory undertaking; it is also a condemnation to willful blindness, if Minerva's bird only rises at nightfall.

This decentering cannot be measured simply by the declarations of method and intent in the introductory texts and program articles of the Encyclopédie. The very body of the dictionary, through the interplay of multiple entries, but also through the intertextuality that circulates definitions from one dictionary to another5, is worked by the double spring, by the dynamics of the Encyclopedic Thing. We'll take one example: the article Character.

From taxonomy to technique: the article Caractère

You have to know how to read the Character article for what it doesn't contain. This long, 23-page collective article6, which refers to a series of plates devoted to printing, is constructed as a departure from a known and expected meaning. The notion of character, placed by La Bruyère at the heart of the conservative ideological construction of the moralists, is identified with what Michel Foucault defines as the "taxonomic knowledge" of the classical age. Character, defined as an imprint, a stamped mark, an inflicted hollow, establishes a system of differential signification, a catalog of singularities articulated by the interplay of the Same and the Other. Type separates, divides and compartmentalizes the world; it transforms the continuum of matter into a succession of differentiated objects, capable of being arranged and classified, and thus of signifying. Character thus constitutes the semiotic cut; it is what justifies that strange and sublime axiom of classical semiotics, that in order to signify, one must cut, that meaning is founded on a separation, a gap, the very imprint of the sign wounding the matter where it comes to be inscribed in hollow.


The word caractère does appear in le Thresor de la langue françoyse by J. Nicot, in 1606, but in a laconic way, to say the least:

"Caractere, character, characteris, see Charactere."

And in Charactere, we find nothing, without being able to assure, however, that this is an oversight. Considered a pure Hellenism, the word Caractere has not fully entered the language; it means nothing outside its Greek referent.

Character between matter and thought

Furetière's 1690 dictionary is much more prolix. The article begins:

"Caractere, subst. Masc. Certaine figure qu'on trace sur le papier, sur l'airain, sur le marbre ou sur autres matieres avec la plume, le burin, le ciseau, ou autres instrumens, pour signifier ou marquer quelque chose. Letters are characters that serve to mark our thoughts."

Very clearly, typeface is defined as the relationship of the "matter" on which it is traced to the "thoughts" it undertakes to "signify", to "mark7". Of this definition, Trévoux's dictionary retains only the end:

"Character. s. m. Letter of the alphabet. Littera. Letters are characters that serve to mark our thoughts."

Under the guise of plagiarism, the ideological perspective inflects. More concise, narrower too, this introductory definition reduces typeface to the instance of the letter, evacuating the relationship to matter to close the constitution of meaning in a pure combinatorial of signs, an abstract play of letters where neither brass nor marble are worked by chisel or chisel.

The Encyclopédie, on the contrary, initially sets aside reference to the letter, retaining from Furetière only what Trévoux had excluded, the very working of matter:

"This word taken in a general sense, signifies a mark or figure traced on paper, on metal, on stone, or on any other matter, with pen, chisel, chisel, or other instrument, in order to make something known or to designate it." (T. II, p. 645a.)

Furetière's "figure" becomes "a mark or a figure": the word figure, which ambivalently designates an abstraction or a material object, is drawn towards the materiality of the mark, while the poetic designation of noble materials, bronze and marble, is replaced by trivial terms, metal and stone. Finally, character, instead of "signifying or marking", makes it possible "to connote or designate something": it doesn't signify, doesn't mark the cut, the gap of representation; it designates, it makes one see and know; it immediately shows "something"; it makes the Thing work, implements its monstration, makes matter come to knowledge outside the taxonomic game of discursive representation.


From the outset, then, between the Furetière, the Trévoux and the Encyclopédie, the difference in ideological perspectives is marked. What could be read, after all, as no more than a decorative variation between equivalent definitions in fact lays the foundations for divergent trajectories, which become increasingly apparent as the articles unfold.

class="normal" class="normal" class="normal" class="normal" class="normal".

This divergence is not strictly speaking measured in the meaning of the terms. It would be reductive to consider a thematic comparison between the entries in the articles of the different dictionaries, and to conclude that to the moral definition of the classical character, as a typology of persons, are added in the Encyclopédie technical definitions, such as the entry "*Caractères d'imprimerie", written by Diderot if we are to believe the asterisk, and which alone constitutes 16 of the 23 pages of the article. Neither Furetière nor Trévoux forget the printing character, and if they devote a smaller space to them, it's also because space in a dictionary is more counted than in an Encyclopédie.

Printing typefaces: from a catalog of names to the work of history

Once again, it's in the articulation of matter to representation, or in other words of the real to the symbolic, that the materialist decentering of the field of knowledge manifests itself. In Furetière, print is defined as a nomenclature:

"Printers also call caracteres, the letters they use to print, of which voicy the degrees. Gros double canon, gros canon, Trismegiste or canon rapproché, petit canon, gros parangon, gros Romain, St. Augustine, Cicero, Philosophy, little Roman, little text, cute, nompareille, Sedanoise or Parisienne. Almost all these letters have their Italics & capitals."

Unfurling the bigarrangement of their exotic denominations, typeface names grid a closed world of pure language, certainly another language, but still language, conceived as a tool to deliver labels and making elegant economy of the world's matter, which they supplant with the reassuring regularity of their taxonomy.

This stylization of the world, this atemporal purity, is no longer acceptable in the eighteenth century. Trévoux takes up Furetière's paragraph roughly, but places it at the end of a long historical development on the origin and transformations of typefaces, from the ancient to the new Hebrew, which gave rise to Greek, then Latin "which is now commonly used in Europe"; the vicissitudes of Greek and Latin typefaces during the Middle Ages are analyzed from the testimony of medals: from history to medal, the referent returns.

The inscription of the world's taxonomy in a historical process, made sensitive by the examination of medallic material, is treated by Trévoux in an implicitly negative way: from the Hebrew characters, which form the perfect book, the Bible, the history of type is the history of a degradation:

"After the return from captivity [from Babylon], the [Hebrew] people n'wrote more than the Assyrian character"; "After [the emperor] Michael we find Greek characters, which began to alter as well as the language, which n'was more than a mixture of Greek & Latin" ; "Around the time of Decius, the character began to alter, & to lose its roundness & sharpness"; "it fell into the last barbarism"; "it was even worse afterwards"; "when the character is round & well formed, it is a mark of antiquity".

The Trévoux must therefore deconstruct history in order to rescue a taxonomy of the world that can no longer be delivered from the outset, as the a priori form of atemporal knowledge.

The positivization of history in the Encyclopédie: materialization and absenteeism

The history of character is not fundamentally different in the Encyclopédie article, which seems to have taken it over from Trévoux, sometimes almost word for word8. But this time it doesn't lead to Furetière's taxonomy, and is itself embedded in an entirely different device. The origin of character, in the Encyclopédie, is neither Hebraic nor biblical; for obvious ideological reasons, it is in the passage from the state of nature to the state of society, i.e. in a totally profane context, that the new genesis comes to be inscribed. If, in the Trévoux, character was seen as an element of history that was always already there, as a kind of a priori beginning in the face of which no anteriority was thinkable, the Encyclopédie's implicit reference to the state of nature as a state without language and, hence, without character, radically transforms the perspective. The character is invented: history perceived as an alteration of the primitive perfect signs of the divine Word is succeeded by a history of the construction and perfection of signs; on the other hand, a conception of character as an instrument of world representation is succeeded by characters placed at the service of communication in the world. From history as the alteration of representation, we move on to history as the perfecting of communication:

"As soon as men were in society, they felt the need to invent a language to communicate their thoughts. This language probably consisted at first only in designating by certain sounds & by certain figures the sensible & palpable beings they could show each other, & consequently it was still very imperfect: but men were not long without realizing that not only was it necessary for them to represent, so to speak, these beings to the ear by sounds, but also to represent them in some way to the eyes, by agreeing on certain marks which designated them." (P. 645a.)

The beginning is imperfect, and typeface is invented and perfected only to make up for9 this principal imperfection: language materializes the immateriality of thought; typeface, which transforms sound into a visual mark, materializes language and makes it possible to communicate to absent beings. The more immaterial the object of communication (from "sentient & palpable beings they could show themselves" to "these beings present or absent", then to "non-palpable beings" and "abstract terms"), the more distant the interlocutor, the greater the material necessity of the concrete perfection of typeface. From "the very figures of these beings, crudely traced on some bodies", we move on to "gestures", then to "certain arbitrary marks": character is born of the decorporization of the sign, of its separation from the body on which it is first traced, then with which, in the gesture, it is expressed. The arbitrariness of the sign is the product of this work of absence, which separates the abstraction of meaning from the concreteness of the world, which separates the interlocutors, now each absent from the other, which finally separates the technical matter of character from the empathy of the body-symptom making sign, exposing itself in gesture.

The emergence of abstraction, the disembodiment of communication thus introduce absence into the heart of the exchange of thoughts; but at the same time, according to an inverse movement, the invention and refinement of typefaces tend to materialize communication. Type detaches itself from thought, to which it is now linked only by an arbitrary code; it becomes objective; it becomes pure mark; it will soon be nothing more than matter.

This double movement of abstraction and absenteeism on the one hand, and materialization on the other, governs the unfolding of the entire Character article in the Encyclopédie and explains the strange detour it seems to take in approaching the question of perfect language, from the angle of what he calls "a project of universal character10".

The universal character project, or the encyclopedic anti-Babel

Since typeface is no longer envisaged as a stable nomenclature, delivered at most to the alterations of history, but is defined as a process of perfecting communication, typeface functions as the work of the Thing tended towards an inaccessible object. The universal character fulfills the same function for the article Character as the "true"Encyclopédie utopianly to come, for the current encyclopedic construction site. It is the accomplishment of the process, the ideal at once empty and hyper-material towards which typefaces tend.

In fact, after the invention of the first alphabet, the multiplication of mankind having, according to D'Alembert, engendered a splintering of primitive society into different nations possessing different languages, the progress of Science found itself hampered by a new impossibility of communication:

"some authors thinking of freeing the human race from this servitude, have proposed plans for characters that could be universal, & that each nation could read in its language. It is easy to see that, in this case, these sorts of caracteres must be real & not nominal, i.e. express things, & not, like common caracteres, express letters or sounds." (P. 645b.)

The universal character accomplishes both absenteeism and materialization. It is totally cut off from the corporeal expression of language, since any equivalence between a character and a sound is now obsolete. The character refers directly to the "thing", totally dispensing with language. It's no longer a partial supplement, a crutch, but an absolute substitution. Language is thus absent, silence perfect, while the relationship to the thing, the very real, is expressed without mediation. The historical movement of perfecting characters thus tends towards a universalization that simultaneously detaches, evides and materializes it to the extreme.

This materialization of typeface corresponds to a decentering, and even a reversal, of the dominant ideological models. The project of universal character makes up for the splintering of society into several nations and languages in a process that is, of course, totally unrealistic, and which in fact consists of subverting, reversing the story of the Tower of Babel. In the biblical myth, the constructed unity of human society is overturned by God's wrath; the diversity of languages is the mark of divine punishment, the note of infamy that recalls mankind's ceaselessly reiterated u{bri ". In the Encylopedia, on the contrary, it's a purely natural cause that precipitates the splintering of languages, a crisis in humanity's growth. This crisis does not bring mankind back to any fair measure; it creates new needs, which give rise to new technical improvements. The encyclopedic anti-Babel does not destroy the utopia of a perfect humanity and language; on the contrary, it establishes it. In the biblical story, the Adamic language, definitively gone, scattered after the collapse of the Tower, is replaced here by this perfect language of universal characters, no longer lost perfection, but perfection to come.


These characters are not utopian either, since they exist in Japan and China:

"We mustn't imagine that this real character is a pipe dream. The Chinese & the Japanese already have, it is said, something similar: they have a common character which each of these peoples hears in the same way in their different languages, although they pronounce it with sounds or words so different, that they do not hear the slightest syllable from each other when they speak." (P. 646a.)

We know that, in reality, Japanese and Chinese ideograms, though related, are far from interchangeable and do not achieve that perfect communication between people of different languages that universal characters aim for. In fact, the Chinese and Japanese provide the ideal pretext for a geographical decentering of the ideological model. Let's look no further11. The Far Eastern model disengages from the biblical model, following a strategy widely used by Voltaire, and favors the shift from a taxonomic logic of culture, both logo- and theo-centric, to a materialist logic of meaning production.

Materialist semiology of character

Finally, what essentially stumbles in the notion of "Character", what enables and favors decentering, is the materiality it introduces into the sign: character is neither the sign of discourse, nor the phoneme of language; it is the materiality of the sign; even more material than the signifier, it identifies the work of the signifier with the very work of matter. This materiality of character leads to its demonization, as evidenced by this final paragraph of Furetière's Caractère article:

"Caractere, se dit aussi de certains billets que donnent des Charlatans, ou Sorciers, qui sont marqués de quelques figures talismaniques, ou de simples cachets. They make foolish people believe that they have the virtue of making people do marvelous & incredible things, such as travelling a hundred places in three hours, being invulnerable to the army, &c. When someone tells of one of these alleged effects, they say that this man must have a character, that he has made a pact with the Devil."

Furetière's paragraph is repeated word for word in the Trévoux, but no longer closes the article, which adds to it the notions of "generic character", through which Linné founded botanical science, and "character in painting", which makes it possible to account for non-discursive representation. Both botany and painting go beyond the strictly taxonomic definition of character: plant characters shift the taxonomy of the world towards an organicist model12; painting characters divert the categories of discourse towards the image. In both cases, the character refers to the materiality of the referent, no longer the mere satanic materiality of the non-discursive sign, of the disquieting and dubious talisman, but a materiality that grounds the sciences of nature and reveals the rationality of the image, which invests science and culture.

This reversal, which remains in outline in the Trévoux, conditions the unfolding of the entire article in the Encyclopédie, from which the paragraph on talismans, incidentally, disappears altogether. In the Encyclopédie, the identity between the work of the signifier and the work of matter orders an overall device, in which much more is at stake than the mere materiality of the mark printed on paper, metal or stone. It is within this framework that D'Alembert's question of perfect language is followed by the entry *Caractères d'imprimerie, which deals with the technique of engraving punches and casting type. This succession should not be seen as a purely random occurrence in the dictionary. The organization of entries within an article does not obey the arbitrary constraints of alphabetical order. Together with the play of cross-references and the supplement of plates, it constitutes the very mechanism of the Encyclopédie.

The material production of characters accomplishes the materialist decentering of the question of perfect language: from language, we have moved on to signs; from signs, to the production of their marks. This production, too, has its own history: to the perfection of communication, marked by the invention of language, then signs, then the alphabet, corresponds the perfection13 of printing, marked by the invention of movable type (Guttemberg), by the modeling of character sets and their shapes (Garamond), by the distinction of i and j, u and v.

The differential system of signs, which organized classical semiotics, the nomenclature of printing typefaces, presented in Furetière and Trévoux as a stable, regulated world, become here not only something material, metal to be engraved, melted and struck, but by this very work, constitute an economic stake14. At first glance, one might think that the semiological stakes are lost in technical explanations. By focusing too much on the art of engraving punches and casting typefaces, it seems that the Encyclopédie article forgets its original aim, that typefaces were invented to communicate meaning.

In fact, if we consider the succession of entries in the article Character as a signifying device, we find that the entry *Printing typeface functions as the key moment in a reversal of meaning, enabling the passage from the proper to the figurative, from type as a material mark to type as an abstract principle for constructing an identity. This reversal is first indicated in the most material way possible:

"It is conceivable that the character which is to leave its imprint on the paper, must be turned in the opposite direction to the imprint. Example, for the character B to give the print B, this character must be arranged as follows ; for if we assume a paper applied to this , so that it receives the imprint, it is obvious that when we turn the paper over to see the imprint left, the parts of what was on the left, being on the right, & those that were on the right being on the left, we will no longer see the figure, but the figure B. It's precisely as if the paper were transparent, and we were looking at the character from behind. This is what makes reading a form difficult for those who are not used to it." (P. 650b.)

The printing process involves turning the characters upside down, thereby confronting us with illegibility. From the outset, the technical dimension of the thing is presented as a passage in reverse, as a test and obstacle of strangeness, and finally, symbolically, as an elision of the paper support: "It is precisely as if the paper being transparent, one looked at the character from behind". What falls away here, as the world of technology opens up, is the support of the graphosphere: the engraver's art elides the book, offers itself to be viewed outside the categories of the legible.


For it's the engraver and not the printer we're talking about. Here, materialist decentering is coupled with social and technological decentering:

This is the engraver, not the printer.

"Engravers of characters are little known in the Republic of Letters. By an injustice of which there are more important examples, we have attributed to the Printers who have made the most beautiful editions, a reputation & praise which should at least be shared with them by the skilled workmen who had engraved the punches on which the caracteres had been cast [...] Much has been said about the Plantins, Elzevirs, Etiennes, & other Printers, whom the beauty & neatness of their characters have made famous, without observing that they were not the authors [...]. But isn't it enough for the Printer to be given the loüange that belongs to him for the meanness of the composition, the cleanliness of the printing, the purity of the correction, &c. without carrying over to him the loüange that belongs to men who have been left in oblivion, even though we were obliged to them for the most beautiful thing in printing? [...] although the Printer, or rather the Typographer, is to the Engraver no more than a skilled singer is to a composer of Music." (P. 651a and b)

The technical knowledge of type makes it possible to decentralize the representation of the work of printing, to shift the scene from the traditional gallery of humanist printers with famous names to the obscure, nameless world of "skilled workers", the Engravers. Using a musical metaphor, Diderot goes so far as to turn the social hierarchy on its head, turning the worker from a performer into a composer, and the printer from a master craftsman into a mere "skilled singer". The adjective "skillful" passes from Engravers to Printers, while the art of composition passes from Printers (who compose the typefaces) to Engravers.

This hierarchical reversal, like the reversal of the typeface, participates in the same deconstruction of the taxonomic and logocentric universe bequeathed by conservative classical culture. At the very moment when type appears as an illegible thing where matter works, the figure of its master-builder becomes a figure without a name, a figure excluded from the republic of letters, destituted by the text:


"For one thing that must astonish, is that the Writers who have made in different times the history of Printing, who have followed its progresses, & who have shown themselves to be the most learned on this object, have greatly extended themselves on the merit of Printers, without almost saying a word about Engravers in characters" (p. 651b).

The inversion and decentering are reiterated when we move from the engraving of punches to the foundry of type. In fact, from counterpunch to punch, punch to die, die to typeface, typeface to printed letter, the letter will be turned over four times. As for the hierarchical shift, it was accentuated by the appearance of this new socio-professional category of foundry workers. While the conception, invention and design of typeface were the sole responsibility of the engravers, for the world, for the republic of letters, typeface only materialized from the "imprint" of the punch, from which the foundrymen drew the typeface, and from the imprint of the typefaces, with which the printers published the books. What of the typeface is visible in the world and represented in discourse is therefore only the imprint of an imprint, the doubly, and even quadruply inverted simulacrum of what was decided in the engraver's gesture alone, in his creative implementation of matter.

But as the text sinks into this play of reversals and these effects of epistemological decentering, thereby confronting the force of de-emiotization exerted by the work of the Thing, a new instance appears that supplements the deconstructive dynamic of the text: these are the plates.

The text-image device

The representation of engraving and foundry processes is part of a seemingly simple device: the text of the article refers to a series of plates supposed to illustrate it. In reality, this text-image arrangement obeys a changing economy. We might even go so far as to say that it's the changing relationship between text and image within the device that conditions the unfolding and semiological logic of the article.


In the first instance, the reference to the plates shows objects, whose manufacture the text recounts. The text then functions as the work of matter, as the implementation of the Thing. The images to which it refers constitute the aim of the text, the constituted objects towards which it tends. The plates are the horizon of the text, as if the work of illegibility in writing (the description of technical procedures becoming incomprehensible) were tending towards the surpassing of writing in the image, according to a movement identical to that described in the process of historical perfection of the typeface, whose achievement, the universal character, can only be represented in the iconicity of the ideogram.


This first economy of the text-image device is reflected in the omnipresence, in the text, of the verb faire: "cette gravûre se fait", "On fait les poinçons", "Le contre-poinçon fait, il s'agit de faire le poinçon", "On fait une marque de repaire", "Cette opération faite, on retire le contre-poinçon". The use of the material is at the same time the use of a device, as indicated by the verbs indicating the arrangement of the objects: "on dresse un morceau", "on le serre dans un tas", "on l'affermit par deux vis", "on présente à cette face", "on le dresse sur la pierre". The text implements a device for transforming matter, tending towards the constitution of an object, i.e., a form, a structure, as these characteristic phrases indicate:

"To form the hollow parts [of the punch], we work a steel counterpunch in the shape of the white parts." (P. 652a.)
"That's all about the structure of the mold" (p. 654b).

But as we progress through the manufacturing process, the text-image relationship becomes more complicated. The surface of the plates is subdivided into a so-called vignette space, where the place in which the machines are implemented is represented, and an out-of-place, out-of-space entity, where the machine parts are arranged. Between the textual support of the implementation and the iconic support where the object is accomplished, there is the dimension of place, the space of the vignette where the device is imaged. This tripartition is echoed by Diderot's lexical remark in the entry De la Fonderie en caractère:

"The term Fonderie en caracteres has several meanings: it is taken either for a complete assortment of punches & dies of all the caracteres, signs, figures, &c. used in Printing, with the molds, furnaces, & other utensils necessary for melting the characters; or for the place where the characters are made; or for the place where the metal of which they are formed is prepared; or finally for the very art of melting them: it is in this last sense that we shall deal with it particularly." (P. 653a.)

Between the collection of objects, the place where the objects are made and the making of the objects themselves, we find here the same three levels of the text-image device, first the objects figured on the plates, then the vignettes, finally the text which, in the process it implements, refers to both.

But foundry is a more complicated process than engraving. It's no longer a matter of making objects directly, but of describing and breaking down the machine that will make it possible, from the few punches supplied by the engraver, to produce the thousands of characters ordered by the printer. From artisanal to industrial implementation.


The text no longer refers to objects, but to parts, or even parts of parts. Referring to the boards, i.e. moving from the legible to the visible, separates, divides, breaks down the object:

"That done, we practice the entailles, a, b, c that we see fig. 12 & 13."
"Les deux premieres parties qu'on peut considérer dans le moule, sont celles qu'on voit Planche II [...] fig. 20 et 21." (P. 653b.)

The word "piece" then returns obsessively: "garnie de toutes ses pieces", "toutes les autres pieces", "La première piece", "Cette longue piece". All these pieces assembled by the text and scattered, shattered in the image, accomplish a device that is completed on the following page:

"Voilà bien des pièces assemblées: cependant le moule n'est pas encore formé; il y manque la pièce principale, celle pour laquelle toutes les autres ont été inventées & disposées, la matrice." (P. 654b.)

Here the text no longer implements the material; it arranges the constituent parts of the machine. We're moving further and further away from a discursive logic. Diderot no longer narrates a manufacturing process; he spatially arranges the elements of a machine. This marks a shift from a logocentric to an iconocentric model: the image no longer complements a text in which the action takes place; the text is reduced to describing the image. At the same time, what is at stake is no longer the making of the punch (with all that this implies of the phallic imaginary), but the setting up of the matrix: in the world of technique, the process of symbolization is ordered according to the symbolic feminine.

But the economy of the text-image device does not stabilize in this new balance of power. The shift towards the image continues in the third part of the entry *Caractère d'imprimerie, devoted to the fonte d'caractères d'Imprimerie, the font and not the foundry, i.e. the very assortment of typefaces, which Furetière and Trévoux summarized by laconically enumerating the names of the fonts. In this third part, the image bursts into the very text of the article, first in the form of tables, then with "examples of all the Caracteres in use" (p. 663). Here, the text becomes an image as a typographic device, giving itself to view as a form of the signifier, independent of its signified. The writing of the logocentric world here completes its desemiotization.

The symbolic reconstruction of typefaces

What is written there, however, is not indifferent: these are moral maxims, veritable pastiches of the Maximes or the Characters, incidentally never explicitly cited in this article. The characters of the moralists are integrated into the work of the encyclopedic Thing, as the ultimate moment of textual deconstruction and the work of illegibility on a surface won over by the logic of the visible. Moral literature arrives in the text before the moral meaning of the word; it precedes it as a kind of negative avant dire.

The figurative and abstract meaning of the character is then the occasion for a vast parody, in which we are told that nothing is signified in it other than what was signified in the proper. The reference to La Bruyère is carefully avoided, in favor of Duclos, who draws character towards mores and mores towards the character of nations: the materialistic decentering of the field of knowledge produces the representation of another world, multicultural and polycentric. But that's another question.

Analysis of the article Caractère in the Encyclopédie has enabled us to identify what constitutes the very movement of a materialist culture in the Philosophes' work of thought. A collective article, bringing together in a rare balance the extremes of the Encylopédie, from the most abstract metaphysical considerations on the origin of typefaces to the most technical accounts of typefaces, the Caractère article constitutes in a way a condensation of what we have defined as the work of the Encyclopedic Thing.

It is no longer the taxonomic efficiency of the classical typeface, with what it implies of metaphysical idealism and moral conservatism, that in the notion of Caractère interests the Encyclopedists. From the outset, Character is defined not as a state, a habitus, but as a process of both materializing and perfecting communication. From then on, the whole thrust of the article will be to explore in greater depth this materialization of character, which detaches itself from the body, becomes an arbitrary, cut-off signifier, an object therefore, whose manufacture we can from then on envisage, very materially.


This materialization deconstructs all biblical references, from the origin of the typeface, which is no longer Hebrew but natural, to its ultimate end, which rebuilds a blissful Babel in reverse.

But the semiological stakes of this materialist decentering only become fully apparent in the technical part of the article, when the relationship between text and image gradually shifts in favor of the latter, to turn around, negate the moralists' discourse on type. The materialist decentering of the field of knowledge then reveals the passage from a logocentric to an iconocentric world, and from a semiotics of cut (where typefaces compartmentalize, isolate singularities) to a semiotics of continuum (where typeface is a matter of foundry and inversion).

This movement is found in many other articles: the Essai article combines the rise of a new literary genre with the test furnace where metals are melted and analyzed, where the value of money is tested; the Matière article focuses on the question of the materiality of light, shifting the obsolete question of metaphysical materialism to the realm of the visible, which becomes the field of all semiotic stakes.


"We have warned that Trévoux's Dictionary is largely copied from Basnage's Furetiere. So when we quote Trévoux's Dictionary in the following, it is only because the name of this one is better known, & without pretending to do wrong to the other one which was its model. Many of the articles in the Encyclopédie that have been claimed to be imitated or copied from Trévoux, are themselves imitated or copied from Basnage." (T. III, p. xvj.)

In the article Langue, Chevalier de Jaucourt explicitly builds his grammatical definition of language on criticism of Frain du Tremblay's, which the latter moreover delivered in his Traité des langues only "after censuring the definition of the word langue, given by Furetiere" (De Jaucourt). The work of definition is a work of intertextuality.

"Écriture, sub. f. (Hist. anc. Gramm. & Arts) We'll define it with Brebeuf:Cet art ingénieux
De peindre la parole & de parler aux yeux,
Et par des traits divers de figures tracées,
To give color & body to thoughts.

The method of giving color, body, or to put it more simply, a kind of existence to thoughts, says Zilia (that witty Peruvian, so well known by her works), is done by tracing with a pen, small figures that we call letters, on a white & thin material that we call paper. [...] (Article by M. le chevalier De Jaucourt.)" (T. V, p. 358a.)

"Chinese script: [...] the characters of Cochinchina, Tongking, & Japan, by the admission of P. du Halde, are the same as those of China, & signify the same things, without however these peoples in speaking express themselves in the same way. Thus, although the languages of these countries are very different, and the inhabitants cannot hear each other when speaking, they get along very well when writing, and all their books are common, as are our arithmetic numbers; several nations use them, and give them different names: but they all mean the same thing. There are as many as eighty thousand of these characters.
However disguised these characters may be today, M. Warburton believes they still retain features which show that they derive their origin from painting & images" (t. V, p. 360a et b).

If the ideogram is given as a model of universal character, it is in a certain way from its iconicity that its universality proceeds: the absolute abstraction of universal character thus coincides with the primitive materiality of the image, according to the contradictory double movement of materialization and absence that defines the work of the Thing in the historical process of perfecting character.

"Marque, (Comm.) dans le commerce & dans les manufactures, c'est un certain caractere qu'on frappe ou qu'on imprime sur différentes sortes de marchandises, soit pour montrer le lieu où elles ont été fabriquées, & pour désigner les fabriquans qui les ont faites, soit pour témoigner qu'elles ont été vûes par les officiers ou magistrats chargés de l'inspection de la manufacture, soit enfin pour faire voir que les droits auxquels elles sont sujettes ont été acquittés, conformément à l'ordonnance. [...] Marque is also a particular sign or character used by merchants, which is known only to them, & by which they remember the price they paid for the merchandise to which it is attached." (Unsigned article; t. X, p. 136b.)

The typeface is therefore part of a dynamic of commerce that is not only linguistic, but also economic. It regulates exchange, it stamps merchandise, it fixes value: through typeface, the thing traded is elevated to the dignity of an object, the manufactured material is integrated into the taxonomic universe of the symbolic institution, which assigns it an origin, a label, a tax, a price. The character transforms the thing into an object, reduces, minimizes the disquieting strangeness of what is exchanged, integrates the incessant labor of matter, the fluctuation of value, into the illusion of a stabilized nomenclature.


In this article, exceptionally, the Encyclopédie is not paginated double-sided, hence the four columns, noted a, b, c, d.


On the thwarting of the idealist perspective in the Encyclopédie article of the Encyclopédie, see G. Benrekassa, Le Langage des Lumières. Concepts et savoir de la langue, PUFécriture, 1995, chapter 7.


For a psychoanalytic definition of the Thing (Das Ding), see J. Lacan, Seminar VII, "The Ethics of Psychoanalysis. The Lacanian relationship of the Thing to the object is, of course, strictly the opposite of the one we are attempting to analyze here, from a materialist perspective.


Karl Marx, Le Capital, book I, ed. J-P Lefebvre, Messidor/éditions sociales, 1983; chapter III, "Monnaie ou circulation des marchandises", note 91, p. 149.


This circulation is very well summarized at the end of the Errata for the first two volumes, at the head of volume III of the Encyclopédie:


The beginning is signed (O), i.e. D'Alembert; then (F), in other words Du Marsais (for hieroglyphics); then again (O); then (E), which indicates the Abbé de La Chapelle (for characters in geometry and perhaps for all mathematical characters) ; then again (O); the entry *Caractère d'imprimerie is marked with an asterisk, indicating Diderot; the entry "Caractère des sociétés ou corps particuliers" is signed again (O); "Caractère d'un auteur" is signed (G), i.e. abbé Mallet; "Caractère, terme moderne de botanique" is signed in full M. le Chevalier de Jaucourt; "Caractère en peinture" is signed (R), M. Landois.


The beginning of the Écriture article, to which Caractère refers, insists on the same articulation between thought and the "white matter" of paper:


See p. 646b.


This logic of the supplement is explicitly inscribed in D'Alembert's text: "Men who had the facility of speaking to each other by designating palpable beings by sounds, could supplant by other signs, as by gestures, to what might otherwise be lacking in this language; it is thus that a mute makes his thought heard by showing the objects of which he speaks, & supplant by gestures to the things he cannot show". The sign supplements the gaps, the voids of language. Gesture makes up for the handicap of the mute, for the void of sound. In a word, character supplements the absence of object.


As we shall see below, universal character and perfect language constitute two different constructions, though the aim is common. On the construction of "un nouveau système de Grammaire, pour former une langue universelle & abrégée, qui pût faciliter la correspondance & le commerce entre les nations de l'Europe", see the entry Langue nouvelle in the article Langue, by M. Faiguet, trésorier de France.


The same discourse on ideograms is found in the article Écriture:


Michel Foucault, Les Mots et les choses, chapter V, Gallimard, Bibliothèque des sciences humaines, 1966.


The word comes up several times: "We can regard the Engravers des poinçons as the first authors of all the movable characters, with which we have printed since the origin of Printing : they are the ones who invented, corrected & perfected them by a series of long & painful progressions, & who brought them to the state we see them in"; "This discovery was made in Germany around the year 1440; the general usefulness it was found to have, made its success very rapid. Several people occupied themselves at the same time with its perfection" (p. 651, 1st column); "Claude Garamond, natif de Paris, parut en 1510, & porta ce travail au plus haut point de perfection qu'il ait jamais acquis"; "Vers le commencement de ce siecle on a perfectionné quelques lettres" (p. 651, 2nd column); "for it is from the perfection of the punch that will depend the perfection of the characters which will emanate from it" (p. 652, 1st column).


The economic dimension appears only indirectly in the body of the Character article itself. But if we refer to the first reference, the Brand article is much more explicit:

Référence de l'article

Stéphane Lojkine, « L’Encyclopédie, édition et subversion », Diderot, une pensée par l’image, cours donné à l’université de Toulouse-Le Mirail, année 2006-2007.



Les Salons

Diderot philosophe

Théâtre, roman, contes