18 Aug 2016

Groensteen (2.1) The System of Comics, ‘Regarding the Threshold of Narrativity’, summary


by Corry Shores


[Search Blog Here. Index tabs are found at the bottom of the left column.]


[Central Entry Directory]

[Literature, Drama, and Poetry, Entry Directory]

[Graphic Literature, Entry Directory]

[Thierry Groensteen, Entry Directory]

[Groensteen’s The System of Comics, entry directory]


[The following is summary. Boldface and bracketed commentary are my own. Proofreading is incomplete, so please excuse my distracting typos.]




Summary of


Thierry Groensteen


The System of Comics


Chapter 2:

Restrained Arthrology: The Sequence



Regarding the Threshold of Narrativity




Brief summary:

Narrative connections in comics can either be found intrinsically within images (for example, within panels), extrinsically between them, or both. We can first of all say that not every juxtaposition of images necessarily involves a narrative relation. We can also say that the main narrative operations in comics are not found within images. Instead, the contents of images have a sort of prelinguistic significance, in the sense of Deleuze’s “utterables” in his film theory, which provides the basic material for narrative significance of a more linguistically structured sort to be formed secondarily. With this in mind, and by taking notice of the fact that many images present themselves simultaneously on the comics page (rather then appear exclusively to one another like a succession of shots in a film), we can conclude that the proper place to turn our attention when analyzing the narrative connections in comics is between the panels (or simply just between the images).






Groensteen begins by mentioning the questions we will address in this chapter:

Immobile images separated by gutters: how do we tell a story with these things? Is the narration in the images? Is it dispersed between each image, or does it emerge | from being arranged end to end? Does the intericonic gutter have a symbolic function? What part does the text play in the production of meaning? These are just some of the questions posed to those of us who want to theorize the operation of the breakdown, in the same way as they are posed, at least intuitively, to the artist who wishes to translate the story that is in his head into a sequence of images.



Groensteen will now summarize film theorists’ attempts to define “the threshold of narrativity.” While most film theorists may agree that editing plays a decisive role in film narration, they vary on the issue of “whether a single image can itself be considered narrative” (Groensteen 104). [Here by single image with regard to film we seem to be talking about a still image like a frame. I mention this, because we can have one continuous shot, thus not involving any editing, within which a narrative fully unfolds.] Christian Metz thinks that a single photo is unable to tell a story, but two photos juxtaposed [I suppose like comics panels] can tell a story.

[T]he photo is so incapable of narrating that when it wishes to do so, it becomes cinema. The photo-novel is not a derivative of the photo but of cinema. An isolated photo can not narrate anything; that’s for sure. But why must it be by some strange corollary that two juxtaposed photos are forced to narrate something? Moving from one image to two images is to move from image to language.1

(Groensteen 104, qtg. Metz)

[Footnote 1 on p.174 (quoting, except for curly brackets):

1. {Christian Metz.} “Le cinéma: langue ou langage?” Communications, no. 4, Paris, Le Seuil, 1964, p. 63. Text reprised in Essais sur la signification au cinema, t. 1, 1968. This is a position that is similar enough to support, acting this time with regard to painting, Aron Kibedi Varga in Discours, récit, image (Pierre Mardaga, Liège, “Philosophie et langage”, 1989). According to Varga (p. 96 f.), a fixed image, a picture representing “living beings engaged in an action” can evoke a story (particularly if it is already known to the spectator) but could not tell it with strict accuracy. Only “the juxtaposition of images generate stories.”]


Groensteen then mentions Roger Odin’s view, which goes against Metz’. He thinks that even in a still image, we read it (so to speak) from left to right, with eye motions that are thus “vectorized like written discourse,” and thereby a story might emerge (See p.104 for more details).


[I do not follow the next idea of André Gaudreault very well. It might be the following. Narration is a matter of transformation. A single shot, as a moving image, involves transformation and thus narration. But on another level, by means of the editing of shots, there is another layer of narration. I might have misinterpreted, so please consult the quotation to follow.]

André Gaudreault3 takes into his account the two “story principles” announced by Tzvetan Todorov,4 those of succession and transformation: “What could, indeed, be considered as narrative . . . are all utterances that relate actions, gestures or events that have between them a ‘relationship of succession’ and that develop ‘a rapport of transformation.’” He recalls that, for us to be able to talk of transformation between two photographs, for example, their reconciliation must “affirm both resemblance and difference.” Consequently, to observe that “transformation (in the sense of modification) can, to a degree, be considered as the single and unique condition of narrativity since, being by definition a process, it already implicates succession.” Gaudreault concludes that, with respect to cinema, the necessary condition of narrativity resides within movement. Moving images will always be “ready-made narratives,” whatever the “degree of the structurization | of the action that they present.” But they would only enact a narrativity that is “native,” “spontaneous,” or intrinsic (that is, “directly linked to subjects of expression”), to which can be added a “second layer” of narrativity, one that is extrinsic, based on editing, and therefore on the arrangement of “narrative contents.”

(Groensteen 104-105)

[Footnotes 3 and 4 on p.174 (quoting, except for curly brackets):

3. I summarize here the essence of chap. III (“A la recherche du premier récit filmique,” pp. 37–51) of Gaudreault’s work Du littéraire au filmique. Système du récit, op. cit. {(Paris: Klincksieck, “Méridiens,” 1988).} All the citations are taken from these pages.

4. Cf. Todorov, Les genres de discourse (Paris, Le Seuil, 1978), p. 66. Underlined in the text.]


Groensteen says he has simplified the terms in this film context so that they apply simply to the concerns we have here for comics. He then summarizes the two sides of the debate. One side thinks there can be narrativity in a single still image. The other side thinks it requires more than a single still image.

I have simplified the terms of these technical debates in order to retain only that which is useful here, that is to say, applicable to comics. The images that interest us are not moving images, but fixed images. According to Gaudreault, we can only talk about their location in terms of extrinsic narrativity: narration is born from the articulation of its contents, but it cannot be found inside each image (even in the “native” state), whereas for Odin, there is narration in the panel itself, which represents a pertinent actantial content, appropriately vectorized by the composition of the image.



Groensteen notes that this vectorization [of eye movements within a single still image by which a narration can unfold] applies more to comics, for two reasons. The first is that “the panel is fixed in the sense that its reading is not impeded by the internal movements of the image” (105). [By this I think Groensteen is saying the following. When we watch a film, what unfolds normally involves motion occurring all over the frame. Also the camera can be moving and the focus can be changing to different objects set a various distances. So here the left-to-right eye motions are distracted by the dynamics of motion in the shot.] The other reason is that unlike in films, the units of a comics work are already spatially arranged so that the eyes must move left-to-right to follow the sequence of panels. Thus, whenever viewing a single panel, our eyes might already have a left-to-right momentum: “the panel participates in a multiframe within which, at the level of each strip, the succession of images is explicitly vectorized from left to right” (105). However, Groensteen adds, in most cases, the contents of the panels are not vectorized, thus we cannot use this notion of vectorized eye motions as the basis for explaining how a single comics panel can by itself present narrative variation: “But if the act of reading obeys an obligatory direction (having already observed that several page layouts render the circulation of the gaze much more complicated, or uncertain), it is sufficient to take any random comic to verify that, in their very composition, the vast majority of images are not vectorized, whether their contents were simply deemed not vectorizable, or whether their narrow and vertical format blocks the vague impulses toward lateral exploration. Consequently, we cannot resolve the question of the panel’s internal narrativity on the basis of this particular criterion” (105). [However, the dialogue balloons seem to operate by the vector principle. (I included a diagram showing this by Scott McCloud from his Understanding Comics, in the summary of section 1.7.2 of this Groensteen book.) However, the story material in the image itself perhaps does not follow this vectorization. But I wonder also, in cases where physical motion is shown, if it is conventional at least in Western comics to position the point of view such that the motion is depicted going from left to right, as by motion lines trailing to the moving thing’s left.]


[Groensteen’s next point seems to be that this film theoretic distinction between moving and fixed images does not really help us entirely for understanding the narrativity in comics’ panels. Instead, we will need to think in terms of narrative drawing, which is something we will learn about shortly. Let me quote as I might have this wrong:]

But the opposition between the two categories of moving images and fixed images is assuredly too crude, even though it is true that each permits an abundance of semiotically and aesthetically differentiated images. Thus, as we will see later, comics lean toward a work of narrative drawing, and its images generally present intrinsic qualities that are not those of the illustration or the picture.



[I am not certain, but Groensteen’s next points seem to be the following. Intrinsic narrativity (that is, narrativity found within a singular image) is more of an issue for film, because the motion of the images does not allow for us to see all the shots at once. We are either in one shot or another. So for film, we might wonder if one shot can by itself have narrative content (although I am confused, because of course it does; the question is if still images do). In comics, however, we see many panels on a page, so extrinsic narrativity is more of an obvious issue. I especially do not grasp Groensteen’s final point on language in the panels. Maybe he is saying that the text in one panel is already narratively referential to text in other panels, but that is just a guess. Let me quote so you can see what it means.]

The question of an intrinsic narrativity to the image beckons us less directly than it preoccupies the theorists of the seventh art. The co-occurrence of panels within the multiframe, their simultaneous presence under the eye of the reader, and also the visibility of the intervals between these panels, that is to say, the locations where their symbolic articulation is carried out, function so that we are naturally inclined to credit narration to the sequence—whereas with cinema, when it is a matter of a clear cut, the moment of passage between two shots is | not visible: I am in one shot, then suddenly I am in another. This tendency is reinforced by the fact that the inscription in the text (caption or dialogue) in the midst of the panel itself imposes the level of language on the image, thus obscuring the speculations on its eventual intrinsic narrativity.



[In the next paragraph, Groensteen makes the point that “between two images, transformation does not automatically ensure a relationship of narrative order”. But his account for why this is depends on his summarizing a demonstration he made in a talk he once gave. The summary is a bit too abbreviated and the terminology too underdefined for me to follow. (He does cite an article which perhaps gives the same demonstration, but I do not know yet.) If I had to guess, he may have in a scientifically demonstrative sort of way shown that with certainty two certain images involve a transformation but the link between them cannot be seen as involving some logical (narrative) sort of ordering. He also distinguishes a visual series from a narrative sequence, and it reminds me of the same terminological distinction Hatfield made in section 3 of “The Art of Tensions”. Perhaps Groensteen is saying that an image series may not evoke a narrative procession when viewed, or in other words, the series may not constitute a narrative sequence. I wish Groensteen had made his demonstration here in this book, because a very important conclusion of his depends on it. For, this goes against McCloud’s demonstration where he suggested that a series of non-sequitur transitions will still come to obtain some kind of coherence in the viewer’s mind. Whether or not this would be a narrative coherence, I am not sure.]

In privileging the sequence, we accomplish nothing more than displacing the problem, since it is not true, contrary to what Christian Metz has postulated, that two juxtaposed drawings (like two photos) are forced to tell us something (or three, or n drawings that are assembled within the same page). I demonstrated this point at the Cerisy colloquium on comics in 1987, identifying five intra-narrative modes wherein panels can be regrouped within a multiframe, namely amalgam, survey, variation, declension, and decomposition. But, within these fundamental modes of organization (qualified, at the time, by “primary distributive functions”), two at least —variation (where the images define the same thematic paradigm) and declension (where an identical motif is submitted to different stylistic treatments) — verify the dual conditions of resemblance and difference between images and can therefore be placed under the regime of transformation. For as such, the linking of panels is not determined by any logical inference nor by any causal-deductive order. It follows that between two images, transformation does not automatically ensure a relationship of narrative order. In fact, once images present a rapport of transformation between them, they constitute at the very minimum a series (the minimum being only two images) but not necessarily a narrative sequence.5

(Groensteen 106)

[Footnotes 5 on p.174 (quoting, except for curly brackets):

Cf. “La narration comme supplément,” Bande dessinée, récit et modernité, op. cit. {ed. Thierry Groensteen (Colloque de Cerisy, Paris: Futuropolis-CNBDI, 1988)}, pp. 45–69. In this text, I define a series as “a continuous or discontinuous succession of images linked by a system of iconic, plastic or semantic correspondences.” Note that the survey and decomposition can also produce series, without being based on a rapport of transformation.]


Groensteen then address the issue of the threshold of narrativity. [The ‘threshold of narrativity’ seems to be what is the least necessary for narration to occur.] Groensteen says that we can at least be certain that “the juxtaposition of two images, taken in a rapport of transformation, does not necessarily produce narration” (106). Regarding whether or not a single image can contain a narrative is a matter we examine later (106).


[Groensteen’s next point is very important and fascinating, but I might not get it exactly right. I think he might be saying that the comics image is to be understood in terms of Deleuze’s notion of the prelinguistic sign, which Deleuze also calls an ‘utterable’. The idea here is that there is a visual given with significance, but that significance is not raised to a linguistic signification in its immediate reception. So it is an utterable rather than an enunciation. Groensteen might be arguing that the comics image begins as an utterable in the sense that it has some sort of prereflective narrative significance even before that significance is brought explicitly to light by means of placing it within a discernible narrative sequence where it plays a role in advancing that narrative procession. I follow the part on François Wahl even less. But Groensteen might be saying that the way we perceive something visually is already structured such that it lends itself to linguisticized (and also narrativized) articulation.]

To continue to borrow from film theory, I find my strongest support in the analysis of the “movement-image” (the shots) proposed by Gilles Deleuze. And most particularly in the following passage: “On the one hand, the movement-image expresses a total that changes and establishes itself between objects: It is a process of differentiation. . . . On the other hand, the movement-image includes intervals. . . . It is a process of specification.”6 This dual characteristic of the movement-image presents a subject that is “semiotically, aesthetically, [and] pragmatically” formed, but non-linguistically:

It is not an enunciation, and these are not utterances. It is an utterable. We mean that, when language gets hold of this material (and it necessarily does so), then it gives rise to utterances which come to | dominate or even replace the images and signs, and which refere [sic]  in turn to pertinent features of the language system.7

Wouldn’t the fixed image of comics, which by definition does not change, be semiotically, aesthetically, and pragmatically formed (structured) as well, but in a different manner? The question can now receive a partial response, since the first half of this book has already brought to light the structuring power of spatiotopical parameters: the forms, dimensions, and contours of the frame, the site of the panel, the methods of text inclusion, etc.


To my mind, the status of utterable can and must be extended to all forms of images, but it is not enough to take note of its semantic potential. Indeed, the image, as we will see shortly, is not only an utterable, it can also be a descriptable and an interpretable. The meaning that the reader (of a comic) or the spectator (of a film) constructs the reading that is executed from the image has as conditions a selective description and a personal interpretation. This appropriation can ultimately be converted into an utterance; it can also steer us toward an aesthetic judgment, one that would consider the image for its appreciable qualities.


Deleuze, however, did not extend to all images the enunciable quality: He reserved it solely for movement-images. In chapter 7 of What Is Philosophy? (“Percept, Affect, and Concept”), he assigns the meaning of all works of art, notably the tableau (painting), toward the single register of sensations (percepts and affects),8 that is to say that he places it “outside all mediation of language,” a fact that ought to stun François Wahl. Indeed, one can consider what would make moving images a subject that language necessarily seizes, as “the visible, such that painting expresses it” — or better: if it is the painting that expresses it — will not be “torn from the register of sensation.” For Wahl, Deleuze strangely strikes an impasse on the “discursive being of the tableau and its constitution by the phrase,” which is precisely the subject that he consecrates in his work titled Introduction au discours du tableau.9 François Wahl’s thesis is that “it is in the structure of language that perception is constructed,”10 the conversion of the picture into linguistic propositions makes it all the more natural that the picture (more generally: the visible) already obeys a specific discursive organization, its configuration consisting of a game ruled by contextual rapports.

(Groensteen 106-107. The ‘[sic]’ in the Deleuze block quotation is mine. Also for that quotation, the italicization in my edition is different than as it is given here.)

[Footnotes 6-10 on p.175 (quoting, except for curly brackets):

6. {Gilles Deleuze.} Cinema 2: The Time-Image, trans. Hugh Tomlinson and Robert Galeta (University of Minnesota Press, 1989), p. 29.

7. Ibid., p. 44.

8. Cf. Gilles Deleuze and Félix Guattari, What is Philosophy? trans. Hugh Tomlinson and Graham Burchell (New York: Columbia University Press, 1994), pp. 154–188. Most notably this sentence: “The work of art is a being of sensation and nothing else: it exists in itself ” (p. 164).

9. {François Wahl. Introduction au discours du tableau.} Seuil edition, “L’ordre philosophique,” 1996. What preceded was a summary of no. 87, p. 199.

10. Id., p. 17, emphasized in the text.]


[As a note, I am not certain that “Deleuze, however, did not extend to all images the enunciable quality: He reserved it solely for movement-images.” This might be right, but the reason I am not sure is because I would think that at least Deleuze would also consider time-images in film as having prelinguistic signs (and thus as being utterables). Perhaps Groensteen means that Deleuze only reserves it for cinematic moving images (within which a movement-image or a time-image can appear). And while it is true that Deleuze studies the non-narrative, non-figurative elements in Francis Bacon’s paintings in the Francis Bacon book, Deleuze also considers the non-narrative, non-figurative elements to present “brute facts” in some way. As such, I wonder if they would not also function as prelinguistic signs or utterables. We will explore this question later when summarizing parts of that book.]


Groensteen then claims that we should not look within images but rather between them to study the narrative elements or functions, because that is where “the pertinent contextual rapports establish themselves with respect to narration” (107).

Again, demonstrating that meaning is inherent to the image is not something that directly speaks to comics, since it is between the panels that the pertinent contextual rapports establish themselves with respect to narration. It is moreover at this level that we will quickly verify that the linkage of images constructs articulations that are similar to those of language.







Groensteen, Thierry. The System of Comics. Translated from French to English by Bart Beaty and Nick Nguyen. Jackson, Mississippi: University Press of Mississippi, 2007. Originally published as Systém de la bande desinée. Paris: Presses Universitaires de France, 1999.




No comments:

Post a Comment