in Writing

The staged problem with closure

Reading Tamryn Bennett’s Comics poetry: Beyond sequential narrative, which argues convincingly for an alternate theory of comics that  may embrace “non-narrative, multi-linear, simultaneous, experimental, abstract or poetic” comics, I was struck by an odd argument against Scott McCloud and his concept of closure:

According to McCloud, the navigational process known as ‘closure’ requires negotiation of ‘gutters’ by the audience. In opposition, [Neil] Cohn claims ‘closure’ is a cloudy term incapable of describing the complex cognitive processes of making ‘meaning’

Bennett goes on to quote Cohn as saying:

If closure occurs ‘in the gaps between panels’ then how does it work if a reader cannot make such a connection until the second panel is reached? That is, the gap cannot be filled unless it has not already been passed over, making closure an additive infereence that occurs at panels, not between them.

I honestly don’t see what they’re arguing about; Scott McCloud is a populariser of the mechanisms that make comics tick, but I missed the part where he claimed to give hard scientific definitions? As for the nitpicking of where exactly closure takes place, don’t underestimate McCloud’s rhetorical slight of hand: Saying that closure occurs in the gutter is simply his visualisation of the language idiom of “filling in the gap”. A reader would necessarily have to read both panels to make any connection between them, linear or otherwise. Besides, as David Mazzuchelli illustrates it, there can be no gap between one panel; there can’t even be a “between”:

Asterios Polyp - Transitions

For me, though, the point of confusion about Bennett’s argument is that whereas McCloud’s gaps don’t apply, the concept of segmentivity (adopted from poetry studies to augment the narrative/cultural studies approach in comics scholarship) emphasises gaps as an integral part of the analysis:

In poetry, as in comics, there is a greater emphasis on segmented ‘matter’, spatial arrangement and ‘ gaps’ than in prose. In both poetry and comics formal analysis of material components is encouraged, as ‘segments’ can be examined independent of narrative ‘closure’.

I hope this doesn’t come off as snarky, I just struggle a bit with why this exactly becomes a point of contest. Bennett’s aversion toward closure may be due to its lack of academic refinement, or to the sequential baggage that McCloud loads it with. She writes of his panel-to-panel transition types, essentialy closure subcategories:

Five of the six examples given by McCloud illustrate a linear interpretation of panels that restrict sequences to ‘moment to moment’, ‘action to action’, ‘subject to subject’, ‘scene to scene’, and ‘aspect to aspect’. By focusing on the immediate panel relations of sequential images critics like McCloud have overlooked the possibility for comics to operate outside of linear grids and conventional narrative structures…

– which I happily co-sign to. My own problem with McCloud’s list of transitions is that last category which Bennett doesn’t mention, dismissively named the Non Sequitur. Here it is presented in Making Comics:

McCloud - making comics_trans 6

“Look, he drew himself as a metronome! You can tell those non sequiturs are only for whacked-out artistes, better steer clear of them!” It’s obvious that McCloud doesn’t find much use for this transition form, tucking it away at the end of his list and writing it off as nonsense gags (though it isn’t quite clear to me from the phrasing if that’s the role he sees it playing, in “otherwise rational” experimental work? Perhaps not even McCloud himself knows). His examples do read as the stereotype of a Tourette’s patient shouting incoherent profanities, and might discourage readers from lending any credence to this type of transition, but again we shouldn’t overlook McCloud’s rhetorical deftness. The category doesn’t fit his theory of comics as a basically narrative art form, and must therefore be painted as a cul-de-sac.

In one last salute, from Understanding Comics, McCloud concedes that there may be something to the juxtaposition of more disparate images, and that even though “such transitions may not make sense in any traditional way, […] a relationship of some sort will inevitably develop.” He leaves the category name as an indication of how much more the reader can expect to hear about it. This omission, in turn, stands out in his work as much as the dismissal of closure does in Bennett’s study, and unfortunately closes the door on a motherlode of potential for expression in comics.

McCloud - understanding comics_trans 6

What McCloud barely scratches the surface of above – and still in that dismissive (BANG!) tone revealing his eagerness to go on and tell stories – can in fact be seen as a parallel structure to the concept of sequence. In Bennett’s segmentivity model, derived from poetics, it would amount to seriality, but in cinema it is established as a form of montage, specifically the Kuleshov effect. It’s a classic example of early experimentation into the possibilities of visual expression:

Essentially, Soviet film director Lev Kuleshov found that by cutting between two unrelated images he presented them in a context that an audience would interpret as a whole larger (or at least different from) its parts. Representing his temporal cuts instead as spatial arrangements, we get comics strips that look like examples of the “aspect to aspect” transition introduced in Understanding Comics, but in their collected context reveal the first and last panels of all the strips to be the same, neutral facial expression. Only the central panel, or cut, is swapped for different contextual interpretations. The “alchemy” McCloud talked about is at work, “endowing the images with a single, overriding identity” – in the below examples hunger, sorrow, and lust, respectively:


Once this effect is established, one can’t help but wonder just how dissimilar images might be juxtaposed to produce more free or ambiguous reader association, or instilling sensation rather than logical context (“otherwise rational stories”), much as Bennett argues for in Comics Poetry.

…segmentivity doesn’t discount narrative possibilities or the value of sequence, instead it enables visual and veral components to be understood as a series of pieces that can contain both narrative and non-narrative elements depending on cognitive inclinations.

It appears stranger still, then, that she criticises the five sequentially oriented transition types but opts not to delve into the potential for comics as a form of (poetic) expression represented by Kuleshov’s montage. As with McCloud dispensing with the wider potential of seemingly discontinuous juxtapositions, I suspect that acknowledging that near-sequential function might blur Bennett’s rationale for multidirectional and spatially arranged comics readings? Not logically blurring the soundness of her point, I mean, but in terms of the rhetorical construction of the study’s argumentation.

I’ve nitpicked a bit myself at Bennett’s article, but as I wrote the other day, I’m quite convinced by the brunt of it. Mostly I took the opportunity here to air some of my own thoughts on comics, and will continue in another upcoming blog post.

Also on:

Write a Comment


  1. I’m enjoying reading your recent posts delving into this stuff. Related to your question about my view of closure… I know McCloud isn’t trying to formulate scientific mechanisms, but you’d be surprised how many full on scientists I’ve met who blindly parrot his claims about closure.

    While I definitely believe that inference happens in reading sequential images (and have done some studies on it), I don’t think its the singular comprehension process it is made out to be. Plus, there are places where “closure” can’t be the mechanism. This is either because panels need to be connected across distances, because they “transition” multiple dimensions at once (time AND characters AND location). Also, we have evidence of prediction occurring *prior* to a reader reaching a subsequent image to be able to make that relation, as described in this video:

    What I find weirder about the post you mention is that Bennett seems to collapse page layout and panel-to-panel meaning. These are certainly connected, but they are not the same thing.

  2. Hi Neil, thanks for replying to my posts, and apologies for not approving the comments earlier.
    It’s interesting to hear your thoughts on closure, especially in this case where your words are being used to argue someone else’s points.
    I’m not really getting Bennett’s aversion towards (layout) grids, either. No reason for this is spelled out that I recall – except perhaps implied in the segmentivity emphasis on “gaps and pauses”, those would be restrained in a rigid grid.


  • This Article was mentioned on

  • This Article was mentioned on