Image and Narrative
Online Magazine of the Visual Narrative - ISSN 1780-678X



Home archive




Issue 7. History and Theory of the Graphic Novel
special section IAWIS conference, Hamburg 2002

Comic strips and constrained writing

Author: Jan Baetens
Published: October 2003

Abstract (E): This article expands the literary concept of "constrained writing" (i.e. of writing based on a set of preformed formal and semantic rules) to the field of comic strips and graphic novels. First, I examine the way in which one can redefine the notion of constraint in the comics; second, I discuss the way in which the comics may help redefining the meaning and scope of the notion of "obstacle" in literary writing.

Abstract (F): Cet article propose d'étendre le concept litéraire d' "écritures à contraintes" (c'est-à-dire toute forme décriture basée sur la mise en jeu de règles préconstruites, qu'elles soient formelles ou sémantiques) au champ visuel de la bande dessinée. On examine d'abord comment on pourrait envisager la notion de contrainte en bande dessinée, pour analyser ensuite dans quelle mesure la contrainte en bande dessinée peut contribuer à une redéfinition de la contrainte en littérature.

keywords: constrained writing, oubapo, Thierry Groensteen, Jean Ricardou, Jacques Roubaud



In this article, I will expand the literary concept of "constrained writing" to the field of comic strips and graphic novels. Two main issues will be dealt with: first, the way in which one has to define the notion of constraint in the comics; second, the way in which the comics may help redefining the meaning and scope of the notion of "obstacle" in literary writing.

The aim of this article is mainly theoretical, and its scope is not to be reduced to the sole field of comics and graphic novels. In fact, I would like to explore the possibility of a transfer of the literary concept of constrained writing to the field of the visual and graphic novel. Such transfer is both strongly needed in the field, which continues to be undertheorised, and challenging, since it may help to rethink the universalising bias of many formalist and rhetorical concepts in literary theory.

In the first part of the article, I will present the current definitions of the notion of "constraint", which has become familiar to scholars of contemporary literary writing (mostly thanks to the works by the Oulipo group and the on-going experiments in the field of electronic writing and hyperfiction). I will make a distinction between the concepts of "rule" (defined as the rules of either the grammatical structure of a given language or the discursive norms imposed by genres and/or contexts), "device" (defined as a stylistic means of underlining local textual phenomena), and "constraint" (defined as a globally active device whose nature is neither grammatical nor discursive). Also, I will distinguish between various types of constraints.

In the second part, I will discuss the ways in which some "elaborationist" theories of constrained writing, mostly those by Jacques Roubaud ("Each constrained text comments on its own constraint") and Jean Ricardou (who stresses the "staged" nature of constrained writing, bringing to the fore the multiple generic layers between the initial constraint and the final text), may be helpful to analyse the theoretical problems raised by the composition (in the sense of Edgar Allan Poe) of a graphic novel or comic. I will focus especially on the distinction between "integrated" (or complete) and "dissociated" (or incomplete) ways of constrained writing.

In the third part of my article, I will broaden the notion of "constraint" to the field of the so-called "negative constraints" (i.e. the practical and/or institutional obstacles which are the only types of constraints recognised as such - but never analysed - by traditional scholarship). I will try to show how one may establish a fruitful relationship between the "dissociated" (or incomplete) way of constrained writing and the way one deals with "obstacles". In both cases, there should be room for a "global theory" of constrained writing (and reading) in which constrained and unconstrained elements are dialectically linked.

What about constraints in comics?

How should one think of constraints in the field of comics? A simple question, to which one can give at least three basic answers. First, one could try to list all the works containing one or more constraints: examples, astonishing as this may be, are not rare at all. Second, one could try to give a survey of all the existing comments and analyses on the use of constraints in comics: here too, and this is probably even more surprising than the observation made above, one finds numerous examples, often very explicit and detailed. Third and last, one could focus on the works of a very specific group of authors, called the Oubapo group ("Ouvroir de bande dessinée potentielle" = Laboratory of potential comics), whose aim it is to explore in the field of comics what has been realised in the field of literature since several decades by the Oulipo group ("Ouvroir de littérature potentielle") (for a comparison of Oulipo and Oubapo, see Baetens 1998). The group has published several readers, called Oupus (#1, 2, 3, etc.), and these collections offer a good starting-point for a concrete analysis of the role of constraints in comics. To be honest, one should admit that all three of these research programmes have already been outlined in some detail by Thierry Groensteen who is not only the main theoretician in this field but also an important creative collaborator of the Oubapo group. His article "Un premier bouquet de contraintes", released in Oupus 1 (Groensteen 1997), has been a landmark publication and it will be referred to throughout this article.

Nevertheless, it is also possible to think of a different way of analysing the relation between comics and constraints. This way is more theoretical, and it emphasises the notion of constraint rather than that of comics. Indeed, it is one thing to reconsider comics in the light of constraint theory, but another to use the practice of comics reading and writing to critically re-examine the definition, the modalities, and most of all the aims of constraint theory itself.

This is the way which will be followed in this article. By doing so, I will diverge considerably from Groensteen's more descriptive approach. Indeed, what Groensteen is primarily looking for is a more or less exhaustive taxonomy of constrained techniques in order to provide a fine-tuned analysis of any constrained comics whatsoever (a programme he succeeds very well in accomplishing). In my approach, other questions will come to the fore, more linked as I have said with constraint theory itself than with its many applications in the field of comics.

Let me start from the following comment made by Groensteen, who proposes an elementary distinction between two classes of constraints: "creation and re-creation, i.e. constraints which can be considered generative (they produce new works) and constraints which can be considered transformational (they modify existing works)" (Groensteen 1997: 17). Of course, one could scrutinise here the very distinction between creation and re-creation, but this is not what I intend to do. What interests me is a critical analysis of the notion of "production" (or "generation"), i.e. of the notion which covers the first half of Groensteen's taxonomy (I have no fundamental problem with the second notion of "transformation" or "modification"). One may wonder indeed what happens when one uses a constraint of "production", for instance a constraint based on of the rhetorical figures of "order" (I am following here the examples given by Groensteen). Does the use of such a constraint mean that one suddenly becomes able to "produce", almost ex nihilo, a new work? Can the choice of a "grid-like" page structure suffice to pop up a story? This conclusion sounds a little bizarre since it does not help to consider a crucial point (unless one supposes that each "creation" can only become efficient once it is applied to an already existing set of materials, i.e. once it becomes itself a "re-creation" - but this is clearly not what is meant by Groensteen): the almost metaphysical "leap" from the constraint to the work, that is, from the abstract rule to the concrete text and concrete drawings. Given the fact that a constraint is not just a "supplementary" technique used to reinforce or modify existing structures (this is what a rhetorical devices accomplishes, the theoretical distinction between constraint and device being not a mere detail), this question deserves to be examined very closely.

Two theories on constrained writing

In order to achieve a better comprehension of this leap, it may be useful to bear in mind some well known theoretical views on constrained writing.

The first one has been formulated by Jacques Roubaud, whose "first principle" has become famous: "A text written under constraint often mentions this constraint" (Rouaud 1981). This is a very rigorous principle, and as Roubaud himself admits, it is far from being respected at all times. Nevertheless, it helps to draw a line between authentic constrained texts and texts where similar techniques seem to be at work, but in a less constrained (and we should say here: in a more "device-like") way. But Roubaud's principle should help us at least to understand that a text written under constraint is likely to be thematically autorepresentative (and, of course, vice versa: autorepresentation is often a clue for constrained writing). Moreover, this principle also provides a first explanation for the presumed leap from constraint to work of art.

The second view of the subject we owe to Jean Ricardou, whose "elaborational" theory seeks to establish a more precise framework for the "birth" of a work of fiction (Ricardou 1978, chap. v), i.e. of the gradual transition from a set of predefined but still abstract constraints to a complete and independent work of art. One of the most interesting aspects of this theory - which describes meticulously the successive logical steps followed by the author of a constrained work - is that any elaboration inevitably modifies, sometimes even erases the initial constraint(s). Indeed, constrained writing does not mean passive or mechanical writing at all. The "leap" we are discussing here is never one single step but always a long and complicated process with many and continuous surprises.

Some theoreticians are more given to Roubaud's principle; others feel more confortable with Ricardou's framework. But all who work in one of the two directions have to tackle very carefully the mechanisms leading from the initial constraint to the subsequent work. The reading of concrete works is therefore extremely useful to better understand how things work out in the field of comic strips. For instance, when one takes a closer look at the pieces of fiction collected in Oupus 1, one immediately notices that the interaction between constraint and production follows mainly two roads: dissociation on the one hand, integration on the other hand.

What is meant by the first process of "dissociation"? Very simply put, the fact that during the elaboration of the work, it is possible to separate almost entirely the work the parameter(s) ruled by constraint(s) and the work on the other parameters of the comic. Etienne Lécroart's "acrostics-comics" are an ultimate example of such a dissociation, since one notices that the various readings of the 4 by 4 panels grid allowed for by the imposed constraint are in fact strictly limited to the aspect, or parameter, of the speech balloons. At the level of the drawings all panels are so mutually exchangeable that one immediately notices that they have only been added afterwards, in order to fill in the white spaces of the textual grid. One could amuse oneself and cover the drawings, or colour them completely in black: it would not make any difference as to how the (textual) constraint is imposed here on the comic strip. Inversely, by covering the speech balloons all drawings would be rendered meaningless and dull.

What is meant by the second process - "integration" - may then easily be deduced: it is the permanent interaction, and mutual shaping, of all parameters of the work, subject or not to the initial constraint, during the elaboration of the comic strip. In the famous graphic novel Watchmen by David Gibbons and Alan Moore, for instance, the integrative principle is very clearly at work, since the constrained use of the 3 by 3 panels grid is not mechanically imposed on the entire novel, but rather functions throughout the story as an imaginative tool, linking permanently numeral elements with thematic and iconic clusters.

Form "obstacle" to constraint

For constraint theory, comic strips can serve as an immensely helpful example to establish a clear distinction between mechanisms of dissociated and integrated writing. (Until now, this distinction has not been fully acknowledged: almost everybody seems to take for granted that constrained writing is necessarily integrated; dissociated or mechanic procedures generally are excluded from critical attention).

But the significance of comic strips for literary theory goes far beyond this. A second and probably much more important debate to which the analysis of comic strips can bring much clarity concernc the status of one specific type of constraints which literary theory often considers as "false" or "lacking of significance", namely the many material and institutional "obstacles" that are mostly defined as mere "handicaps" and almost never as creative tools. In the low-art field of comics strips (ruled less by Creative Geniuses than by the culture industry) obstacles of this kind are much more ubiquitous and drastic than in the high-art field of literature. Their example, however, could compel literary theorists to seriously reconsider what it means for creative production that a book is commissioned, edited, published, distributed, etc.

Let me give some examples of such obstacles: censorship (or self-censorship) as an example of ideological constraint; the necessity (in European comic strips) to tell a story over either 48 or 62 pages as an example of an editorial constraint bearing no relation at all to the "content" of the book; the time schedule imposed by the publisher as another example of institutional and commercial constraints; etc. All these constraints, and many, many others, are often neglected as being mere obstacles or handicaps, instead of getting recognised as what they are to no lesser extent, viz. just as important features of the creative process as their more prestigious brothers and sisters, the "literary" constraints. And although everybody knows that most authors manage to make very astute and clever use of them (Hergé for instance uses them as a real substitute for other "authentic" constraints; see Baetens 2002), most theoreticians continue to equate these obstacles with creative castration. An excellent proof of their neglect and indifference is the fact that literary scholars only accept to discuss the handicaps in question in the field of popular literature ( or paralittérature, Trivialliteratur). There, a handicap may happen to assume a creative function. In "real" literature, however, this is never the case (except in a completely anecdotal way). Comic strips should compel literary scholars to free themselves from their stereotypical rejection of these handicaps which are as complete and stimulating examples of constrained writing as any other.

Moreover, the distinction between dissociated and integrated constraints can also be related to the concept of "negative constraint" (or obstacle, handicap, etc.). If one accepts that, in most cases, it is not possible to maintain a clear-cut distinction between parameter(s) ruled by initial constraint(s) and other elements of a work written under constraint, one must also accept that the elements produced during the elaboration of the text (and as we noticed, most of these elements are not preprogrammed, but "pop up", often very surprisingly, during the creative work of the author) cannot be considered as the "remainder" or "leftovers". Instead, they ought to be seen in the same way as we have examined the "negative constraints" or obstacles.

This way of analysing obstacles has the advantage of making us understand to what extent two mechanisms at work here should be theorised simultaneously: on the one hand the fact that a constraint is modified by the work on which it is imposed; on the other hand the fact that a constraint also modifies the form, the status, and the position of the so-called "remainder" of the work, i.e. those elements which are not directly linked with the operation of the constraint (but which the constraint inevitably produces during the elaboration of the work). One could even say, and this will be my conclusion, that an expanded vision of constrained writing (and comic strips make it more easy for us to envision such an expansion) not only establishes a less schematic definition of what a constraint really is and means, but also, and maybe even more so, changes the status of the non-constrained elements of a work which can now be analysed as negative constraints in their own right. Constrained theory thus has no reason any longer to separate constrained and non-constrained elements, as is still done by most scholars, and instead should think of the non-constrained parts or aspects of the work as participating "negatively" in the overall action of the writing under constraint. Such a reorientation of the theory implies that reading under constraint no longer means trying to isolate what is constrained from what is not, but trying to see what is made possible by a rule or constraint, and what is not.


Jan BAETENS (1998). "Une déclaration d'indépendance". In: 9e Art, pp. 124 s.

- (2002). "Hergé, écrivain à contraintes". In: Philippe Marion (éd.). Hergé. Colloque de Chaudfontaine. Liège: Ed. de la Bibliothèque des Paralittératures.

Thierry GROENSTEEN (1997). "Un premier bouquet de contraintes". In: Oupus 1 (in: OUBAPO 1997). pp. 13-59.

OUBAPO (1997). Oupus 1. Paris: L'Association.

Jean RICARDOU (1978). Nouveaux problèmes du roman. Paris: Seuil. 1978.

Jacques ROUBAUD (1981). "Deux principes parfois respectés par les travaux oulipiens". In: Oulipo. Atlas de littérature potentielle. Paris: Gallimard. p. 90.



Maerlant Center Institute for Cultural Studies

This site is optimized for Netscape 6 and higher

site design: Sara Roegiers @ Maerlantcentrum