Review: The System of Comics by Thierry Groensteen

Groensteen, Thierry. 2007. The System of Comics. Translated by B. Beaty and N. Nguyen: University of Mississippi Press.

Printer-friendly pdf

The best thing about the new translated work The System of Comics by Thierry Groensteen is that it hopefully reflects an increase of English translations of international works on comic theory. There are numerous offerings by European, Japanese, and South American authors that rarely make their way into American scholarship, and more exchange of ideas can only be fruitful to the field. Groensteen is regarded as one of Europe’s leading historians and theorists of comics, and in its original French, the Systeme de la Bande Desinée is heralded by many as a “must-read,” so it would be a logical starting point for such a trend.

Moreover, Groensteen’s book is offered by many as a “serious” alternative Scott McCloud’s Understanding Comics, which has gained massive popularity, especially in America, yet remains dismissed in many scholarly circles since McCloud himself is not an academic. Comparisons between the two works has become commonplace (including the book’s own introduction), and this review will be no exception. For his own part, Groensteen does not engage with McCloud’s theories except in a singular passing endnote, which is a shame since their ideas share many similarities.

While my French is not sufficient to assess the overall quality of translation, it certainly did not read comfortably. At times, the English felt unnatural, and word-choice often seemed clumsy if not uninformed. For instance, calling the psychological device of an “eye tracker” an “eye path follower” betrays a lack of competency and/or desire to find the accurate vocabulary (whether on the part of the translators or author is unknown). Additionally, while the endnotes were often helpful, the absence of a reference section is quite noticeable, and the endnotes do not include all the cited bibliographic information. As an “academic” book, such an oversight is a bit peculiar.

Nevertheless, the writing should only be a surface issue to the actual ideas in the book. Perhaps the most appropriate place to start is where he does: with the definition of “comics.”

Problems with definitions

Groensteen begins with what has become a common exercise in the study of comics: defining “comics” itself (no doubt bande desinée originally). He carefully deconstructs the faults of various definitions that have been proposed by various realms of scholarship. He rightly shows it cannot be guided by a single “essence” like text/image interactions, and decries a definition of comics as episodic narratives, among others.

Rather, at the outset, he identifies “comics as a language” a “system” that arises out of the “combination of a… collection of codes,” most strongly motivated by “iconic solidarity” — a fancy term for the contribution of several images. While somewhat more flexible, this emphasis on the visuals sounds a great deal like McCloud’s more rigid “juxtaposed sequential images,” but lacks a reference or engagement with his ideas despite discussing many other scholars works who Groensteen disagrees with.

Truly, in its similarity, Groensteen’s definition falls into the same formalist trap as McCloud’s in failing to separate the structural notions of creating images/writing from the socio-cultural role of “comics” as objects/artifacts. This habit may have been inherited by them both from Kunzle’s influential The Early Comic Strip, whose recasting of the term “comics” on pre-1800s sequential images engendered many (like McCloud) to seek out a definition to include all possible historical examples. However, this goal is a red-herring, and as Horrock’s points out, there is no reason for sequential images from diverse historical contexts to be bound to the same socio-cultural context of the contemporary notion of “comics.”

Indeed, “comics” as a social artifact refers to numerous qualities, including 1) physical objects (strips and books), 2) a collection of genres, 3) an industry, 4) a culture/community, and others that are all tied to a context of the modern era. On the other hand, sequential images do create a language: a “visual language” that combines with text to be used within those social objects called “comics.” “Comics” are not this visual language. “Comics” are a social object written in a visual language that combines with text. If novels or magazines are written in English, why should “comics” be a language, instead of be written in a language?

While Groensteen strives to strike out an abstract notion for his definition, he frequently reinforces the conflation of “comics as a medium” with “comics as a social artifact” — particularly a physical object. He writes that comics owe their birth to the technological development of lithography (8). Further on, his analysis of both word balloons (69-75), panels as “measurable in square centimeters” (28), and the expressive power of sequential images through “arthrology” rest wholly on the physicality of and across pages. However, physical objects cannot be languages — especially in the abstract sense of them being a “code” — and his reinforcement of it runs contrary to his own definition of “comics” that claims to guide the broader work.

The “System”

As invoked by the title, Groensteen’s work seeks to sketch a “system” by which “comics” operate. To his credit, he carefully moves in analysis through varying levels of interactions in comics’ structure, accounting for most of its forms. He eschews the method of previous structuralist approaches to treating the form as a “language,” which aim to dissect the medium into its minimal units, instead aiming for its broader “articulation” — larger levels of structure. However, Groensteen’s grand system is little more than an extensive taxonomy with terms that essentially mean “the principle by which (pick your taxonomic portion) operates.”

Groensteen begins by discussing the “spatio-topical system,” the various spatial elements at play in comic pages. He first notes the “hyperframe” as the delineated space of the page, in contrast to the “multiframe” — meaning the relation of all frames that constitute a comic piece, including “the sum of the hyperframes.” This chapter also includes discussions of the functions of a panel, usefulness of margins, positioning of balloons, and various facets of page layout, such as inset panels, the “strip” of a tier of panels, and possible functional roles for types of layouts.

Now, my harshest criticism is not that I think Groensteen’s theory of comics is invalid or wrong, but that it is uninteresting.

While at times insightful, he labors through discussing details of relatively lackluster observations, such as the various ways in which a balloon visually interacts with a panel border. It is understandable why a taxonomic distinction can be made between balloons that touch the panel border and those that overlap it, but does it really add substantially to knowledge about the function and understanding of balloons as a graphic/narrative element on its own?

If this discussion led to a substantial observation about the constraints that this interaction creates on such a relationship, this breakdown would seem more significant (for instance, that two balloons cannot cross tails such that their speakers are in opposite panels than the balloons). However, no revelation of this sort is reached, and most of his analysis focuses solely on the physical aspects of the relations of balloons to panels. Such is the features of most of his discussions. All of this leaves the question of what this analysis offers a theory of comics understanding except for surface descriptions of a (fairly banal) phenomenon?

The next two sections are devoted to the principle of “arthrology” to describe either the linear relations of panels to each other (restrained arthrology) or the relations that one panel might have to others in a non-linear sense (general arthrology). Unlike McCloud’s panel transitions, Groensteen does well to recognize that panels make connections beyond their immediately juxtaposed neighbors, yet does not give any hint as to how. McCloud’s transitions at least attempted to characterize relationships between panels, which is largely why the approach is so appealing, but Groensteen leaves such detail aside, preferring instead for gross scale abstract principles. He describes “braiding” as the principle guiding arthrology — essentially the function of making connections across the multiframe.

On the surface braiding and arthrology appear to be about the creation of meaning, like McCloud’s “closure,” or perhaps a comic version of Halliday and Hasan’s (1976) classic notions of local and global discourse coherence, which Saraceni (2000) also attempted to map to comics’ analysis. However, Groensteen does not even come close to talking about semantic concerns in this way — despite purporting to (the back cover describes the book as “An authoritative exploration of how the comics achieve meaning, form, and function.”).

Really, braiding and arthrology are a theory of compositional relationships. What Groensteen focuses on is not in any way a system of how the content of panels lends towards making meaning — it is primarily about the relationship of the visual composition within a panel to others on a page and other panels scattered throughout the broader work. In particular, his theory observes the recurrence of compositional or thematic similarities (such as motifs) across varying distances of space, be it a page or a whole work, and whether the position in layout of those panels has significance (such as in the first or last position of a page).

While these observations provide a very different perspective on comics pages, they do not even scratch the surface of the “holy grail” of questions about comic sequences that they appear to address: how do sequences of images create meaning? Indeed, Groensteen dismisses such aspects as merely about “storytelling.”

Rather, arthrology, like the other aspects of Groensteen’s system, once again highlights a physical feature of comic pages in what is proposed as an abstract code of understanding. The theory treats the medium as an “art object” to be analyzed like a multi-canvas painting, as opposed to a communicative medium in the McCloudian sense. Indeed, such an approach would be akin in verbal language to recognizing that various patterns of sound appear throughout different words in a story — perhaps telling you something about phonetics, but nothing about meaning. All told, Groensteen’s system limits itself to only describing the surface aspects of the medium, without pushing towards any deeper constraining principles.

Despite the criticism of the workings of his system, Groensteen is nevertheless an astute observer of various components of the comics form, and his numerous insights on comic elements betray a deep commitment to dissecting the medium. The System of Comics does reveal gems of this intuition at times when discussing various components, though the overall theoretical architecture in which they are embedded is not commiserate with the value of their insight.


The most troubling thing I found about The System of Comics is the overall orientation to scholarship that this work represents. Throughout, Groensteen’s writing conveys an attitude as if the theories are entitled to be significant, echoed in the introduction where Beaty and Nguyen hail it for emerging from the rich semiotics tradition, as if that inherently legitimates its ideas. This is a direct knock to McCloud, who they note has “been criticized for…lack of theoretical sophistication” (1). Though, this criticism has only come from academics unfortunately perpetuating the stereotype of ivory tower snobbery, as if scholarship must be done in the academy. It is hypocritical at best for any scholar to deride McCloud for not engaging a broader literature while lauding a book with very similar ideas to McCloud’s without any citation or discussion of them. So much for the value of “engaging the literature.”

Truly, Groensteen is the anti-McCloud, keeping the keys to theory locked away in the ivory tower, reachable by only those willing to slog through the exposition to reach it. While McCloud’s work about comics is presented in its visual language, there is some irony that The System of Comics takes “iconic solidarity” of images as its crux, yet is almost devoid of graphic examples. Perhaps this is why McCloud is ridiculed so much by Groensteen-minded theorists: he willingly gives away the goods to the rabble without a fight.

If McCloud lies on one extreme of being too accessible (as if that’s a bad thing), Groensteen’s work is the inverse, reflecting the worst of academic jargon and inapplicability of “theoretical sophistication” — to the extent that the terminology obfuscates the actual theories (to academics and layfolk alike). Groensteen offers complicated names and lengthy descriptions to what are otherwise fairly facile observations about surface phenomena.

Moreover, numerous questions are left unanswered, for instance how this theory is useful or applicable to 1) describing how this medium of sequential images communicates, 2) contrasting various comics’ structure with each other, or 3) describing the relationship of the visual language in comics to other modes of human expression?

Groensteen’s “system” accomplishes none of these. These are all questions that are important to semiotics as a discipline, so why is Groensteen’s approach unable to even broach them? It only states vague principles for nearly obvious observations, with fancy names and a semiotic tradition to validate its claims. It is scholarly hand-waving at its best, and a reflection of why the invocation of “semiotics” garners more eye-rolls than awe in many circles.

Paradigm Shifts

In Kuhn’s renowned discussion of paradigm shifts, he describes that prior to a major shift many similar theories will emerge in competition with each other, yet all pointing towards similar intuitions. In this case, Groensteen taps into the common intuition that the system found in comics is somehow similar to the system of language. Indeed, he begins with a strong statement about how comics are a language, yet his analysis paints a picture of the comic medium that is decidedly un-language-like.

While his stated focus is not on “minimal units” — thereby avoiding topics such as how the system is understood as a graphic domain and meaning-making for signs and symbols — he states nothing resembling a grammar for how the sequence creates meaning (though claiming that is his intent). If his aim is to describe a “code” that shows this is a language, one would think it would involve structures akin to language. All this leaves doubt as to whether Groensteen would actually know what being “a language” would entail in the first place (aside from metaphoric extension).

To this extent, Groensteen’s total work certainly marks a valiant attempt that can take its place next to other approaches that reflect the growing pains of a discipline, sharing the intuition that the comic form should be compared to language (including: Eisner 1985; McCloud 1993; Saraceni 2000; Cohn 2003, etc.). However, this piece and its theories are not the revolutionary new paradigm for comics that will sweep away all others, and will likely follow the path of the branch of semiotics from which it hails — being considered largely passé in the study of language.


Cohn, Neil. 2003. Early Writings on Visual Language. Carlsbad, CA: Emaki Productions.
Eisner, Will. 1985. Comics & Sequential Art. Florida: Poorhouse Press.
Halliday, M. A. K., and Ruqaiya Hasan. 1976. Cohesion in English. London: Longman.
McCloud, Scott. 1993. Understanding Comics: The Invisible Art. New York, NY: Harper Collins.
Saraceni, Mario. 2000. Language Beyond Language: Comics as Verbo-Visual Texts. Dissertation, Applied Linguistics, University of Nottingham.


  • Very interesting review. I haven’t read all of Groensteen’s book, but I was struck — even in the first tens of pages — by some of the same things you’re mentioning. It struck me as unnecessarily obtuse and jargon-y at times, which probably is why I put it down and never got around to picking it up again. I think that might be the cardinal sin of theory: worshiping at the altar of jargon and thinking that it will make one’s work seem sophisticated. Hmm. Thanks for the review.

  • Neil,

    I agree with many of your criticisms, but do think that Groensteen can be good, especially when he focuses on a more local level. Some of his discussions of actual comics pages, those towards the end of the book for example, are careful and interesting. I wish the book featured more of these.

  • Groensteen’s work is the inverse, reflecting the worst of academic jargon and inapplicability of “theoretical sophistication” — to the extent that the terminology obfuscates the actual theories (to academics and layfolk alike). Groensteen offers complicated names and lengthy descriptions to what are otherwise fairly facile observations about surface phenomena.

    I disagree strongly with this on two levels: one, that Groensteen’s language serves merely to disguise the obviousness of his insights (first off, his language is appropriate to his project, and, second, his insights are IMO not so obvious); and, two, that academic jargon (terminology) is somehow inappropriate to an academic project. There’s an implicit charge of snobbery here that is false. Fact is, Groensteen wrote System to engage a long-standing European tradition of semiological work on comics, and jousting with that tradition in language appropriate to that tradition is only fitting. What’s more, Groensteen (and his translators) have engaged comics on multiple levels, varying their language in ways appropriate to each project.

    The old charge v. academics, of snobbery or willful obscurity, is a tired rhetorical trope that serves only to discredit academic work on the basis of some pseudo-populism. It is routinely used in comics fandom to fend off academic attention, or deflect work that is genuinely challenging.

    I see two red herrings in your argument: one, the idea that Groensteen in locked in an “ivory tower”; and, two, that Mcloud “is ridiculed … by Groensteen-minded theorists.” The former is an anti-intellectual cliche; the latter is false. The reasons McCloud has been held up for criticism (even by those who, like me, admire Understanding Comics and learned a lot from it) are many and varied, including, for example, his idiosyncratic (and to students potentially confusing) use or misuse of established terms and the haziness and subjectivity of some of his arguments, e.g., those about “identification.”

    And, BTW, not only academics have raised serious questions about McCloudian formalism; look at some of the spirited rejoinders from working cartoonists as well.

    I grant you that System is a difficult read; at least it was to me. And I do not agree with all of it, by any means. I refer readers to reviews by Craig Fischer and myself, over at Thought Balloonists, for further commentary:

  • Thank you for the comments.

    Charles, as I said on the comixschl list, I have heard several people critique McCloud’s work for being “non-academic” and praising Groensteen’s for the opposite. I am targeting those claims, not other critiques of McCloud.

    Truly, I have no problems with criticism of McCloud and his formalism: at this point I disagree with most of his ideas! But I don’t think people can dismiss him outright as “not written for academics” when clearly his work has had an impact in academia.

    My critique here is that Groensteen just doesn’t engage with McCloud enough though being aware of him and having similar concepts, and that his work is so inaccessible that it alienates people from the ideas. Whether he’s “engaging in tradition” or not, this is a problem (perhaps with him AND the tradition). This doesn’t have to be intentional, it just is what it is and I find that disappointing.

  • You say Groensteen does not engage sufficiently with McCloud. This is because there is little to engage with in McCloud’s commentary. His analysis is basic, his theories flimsy and conclusions little more than unsubstantiated overgeneralisations. What you (and most of the detractors on the comments board) seem to be finding ‘difficult’ in Groensteen’s book is more properly termed intellectual rigor. Yes, it’s not easy. That’s because it’s thorough, and more theoretically sophisticated than McCloud’s weak formulations. Hopefully, if more works from the Francophone critical tradition make it into translation, it will do something to address the dearth of decent Anglophone criticism, of which this is an example (though, having already risked sounding utterly derisory, I’d say you had a future in a Cultural Studies department, given the unfortunate shortcomings of much of the latter).

  • First, let me say that while I welcome responses on my blog, I respond much better to opinions that are not made anonymously, especially those insulting me.

    While you may think Groensteen is rigorous — and I would agree that he is at least in scope — the problem is that his ideas and observations are NOT overly complex but his jargon and description of them are. It’s one thing to be rigorous, but at least make the rigor count for something of use. I found little of use in Groensteen’s actual theories that weren’t already obvious, facile, or uninteresting.

    And, I’d say that my own work fills that void quite well, and does so far better than both McCloud and the Francophone tradition. Perhaps you should read some of it. I’m quite happy in psychology and cognitive science — disciplines which aren’t outdated in research on language like the “semiological” tradition that Groensteen hails from — thank you very much.

  • I wish to make just a few comments about your very critical review.
    1. I don’t understand how you manage to discuss my “definition of comics” since I don’t give any, and carefully explain why. You build up a definition by using bits of my texts that are not supposed to be used as a definition.
    2. I agree with you about “comics” refering at the same time to physical objects, a collection of genres, an industry, etc. But I believe I make it clear from the very beginning that this particular book deals only with the question of comics as a language. The original French version of this essay was published in a Semiotic series, and this is just not the place to discuss economical or sosiological issues.
    Moreover, this is the only one of my books that is currently translated in English, but I wrote a lot more books and ‘The System of Comics’ is not supposed to concentrate all my views about the comics phenomenon.
    3. From various comments I understand that there are some real problems with the English translation, and I suppose that it doesn’t read comfortably. This is a shame, because I do not think that the vocabulary used in the French version is very sophisticated : my basic notions, “decoupage”, “mise en page” and “tressage” – translated as breakdown, page layout and braiding (please note that the assonance is lost) – are very simple terms that make immediate sense for a French reader. The only new technical terms that I introduce are “spatiotopie” and “arthrologie” and etymology makes them easy to understand too.
    4. I do not consider myself at all as an “anti-McCloud” and I have never criticized his work. I believe that the main opposition between “Understanding Comics” and my book is that McCloud examines panel-to-panel relations without any consideration for the complete page, whereas I consider the page to be a basic unit of the comics language.
    Thank you for your interest.
    Thierry Groensteen

  • Thierry, thanks very much for replying to my. I greatly appreciate your willingness to engage and clarify your positions, especially in such an open manner given how critical my review is. Some responses:

    1/2. My main concern with definitions is the idea that “comics are a language” — as opposed to the idea that “comics are written in a language”, which is my own position. I feel this is an important distinction, because so long as all those socio-cultural things exist, “comics” cannot “be” a language in any real non-metaphorical sense.

    3. I’m glad to hear that the French edition has less intended jargon. It is very unfortunate that the translation makes things so difficult, because it feels a hinderance to being able to get at the ideas — which is what colored most of my problems with the book. Unfortunately, my French is too poor to appreciate it as intended (I’ve tried…). Hopefully the (inevitable) next translations of your books won’t have these issues.

    4. I don’t mean to imply that you are against McCloud (I was more replying to critics who I’ve read that compare your approaches). Rather, I would have just liked to see you engage with McCloud’s ideas more in the text, since I’m sure you would have a lot of interesting things to say about them.

    Again, thank you for your willingness to make comments. I would love to hear your thoughts on anything else on my site if you have the opportunity or desire.

  • Write a Reply or Comment