Magic, and the nature of magic, plays a central role in Grant Morrison’s 2005 comic Vimanarama. In this essay I will analyze the way that Morrison cleverly manipulates the conventions of pictorial and textual representation at work in the comics art form in order to provide a unique, extremely effective yet non-realistic pictorial representation of mystical experience. This analysis will depend on Kendall Walton’s examination of fictions as games of make-believe, and on an extension of these ideas to account for the particular way that pictures and text within comics activate different conventions, respectively, with regard to what, exactly, we are implicitly instructed to make-believe when experiencing fiction.
Vimanarama tells the story of Ali and Sofia, two young Muslims living in Bradford, England, whose marriage has been arranged by their parents. At the beginning of the comic these two characters meet for the first time and are immediately flung into an underground world populated by super-powered beings loosely connected to Hindu mythology. After accidentally releasing giant monsters—the Fireborn—who immediately begin trampling through Britain, the pair meet the Untra-Hadeen, a group of magical superheroes whose leader, Prince Ben-Rama, is powered by love. Predictably, Sofia turns out to be the modern-day version of his long-deceased beloved, setting up the romantic triangle around which the plot revolves.
The story is brilliantly illustrated by Philip Bond, whose visual renderings invoke both Bollywood films and Kirby at his Fourth World weirdest. What is of most interest to us here, however, is another aspect of the visual storytelling: the visual depiction of Ali’s mystical near-death experience near the end of the story. At one point in the story, Ali hangs himself in order to petition the angels for help, after Sofia’s admission that she does not love Ben-Rama renders the prince powerless and our world doomed. Ali and his brother Omar, who has passed out from a head injury,1 meet in the afterlife and interact with a number of angelic beings. Notable amongst these is the angel wasp—a shining mystical being shaped like the word “angel wasp” (but with wings) that stings Ali on the nose (82). The angel wasp is a perfect example of Scott McCloud’s notion of a montage word-image interaction.2 In Chapter 6 (“Show and Tell”) of his seminal Understanding Comics (1993), McCloud provides a taxonomy of different ways that words and pictures can interact within a comic panel, isolating seven different types of interaction: word-specific, picture-specific, duo-specific, additive, parallel, montage, and interdependent. We need not delve too deeply into the adequacy and exhaustiveness of this taxonomy, since we are here interested solely in montage word-picture combinations, which McCloud defines as follows: “Still another option is montage, where words are treated as integral parts of the picture” (154). Morrison’s angel wasp, as well as the angel sky and angel clouds that appear on the previous page, but which do not interact directly with Ali, are paradigmatic instances of what McCloud presumably has in mind. Like his three pictorial examples (which involve stressful words such as “cash flow” floating around a sweat-beaded head, happiness expressed by large letters filling the panel, and a human figure made up of rows of type, respectively), the angel wasp is a pictorial (or expressive) non-textual element of the panel that is composed of textual parts. Nevertheless, although McCloud’s identification of montage as one mode of word-image interaction is helpful, his description of the workings of this type of image-text combination is insufficient for our purposes here. The reason is simple: on McCloud’s definition (and quite contrary to what he presumably intends) any text that is included in a panel, for example the text on a street sign or in a shop window, would render the panel an instance of montage. Much more is going on with the angel wasp, however, and much more is going on in McCloud’s own examples, regardless of whether his own discussion of the phenomenon reflects this fact.
Thus, we need to develop a more nuanced account of how images and words work, how they work differently from each other, and how they interact in comics before analyzing the particular case of the angel wasp further. In particular, we need to identify the various conventions that govern how we are to interpret pictures and words within a panel, and how these conventions differ from one another and might produce tensions within a panel as a whole. The basic building blocks of such an account can be given rather straightforwardly in terms of Kendall Walton’s account of fictions as games of make-believe. The theoretical tools provided by Walton’s account of fiction then allow us to explain the power and efficacy of the angel wasp as a non-literal representation of mystical experience in virtue of the fact that the depiction of the angel wasp activates competing, apparently incompatible conventions.
Waltonian Fiction, Text, and Pictures
In his influential Mimesis as Make-believe: On the Foundations of the Representational Arts, Kendall Walton develops a complex and nuanced account of the way that fiction works, and in particular, of the sense in which certain claims are ‘true’ in a fiction and certain other claims are ‘false’ in that fiction (while leaving it open that some claims might, perhaps contrary to how the actual non-fictional world works, remain indeterminate within a fiction). Walton’s account of the nature of fiction can be explicated in terms of three main claims.
The first claim is that fictions, including literary fictions, cinema, comics, and other narrative forms, consist of sets of rules for games of make-believe. Of course, fictions do not at first glance appear to be collections of rules for playing a game—on the contrary, they appear to be descriptions of characters, objects, environments, actions undertaken by those characters, and events that happen to those characters within those environments. Typically, however, the characters, objects, environments, actions, and events described by a fiction do not actually exist—they are fictional.3 Thus, if we intend to explain the sense in which some claims in a fiction (e.g. “Ali lives in Bradford”) are ‘correct’ or ‘true-in-the-fiction,’ and others (e.g. “Ali lives in Boston”) are ‘incorrect,’ or ‘false-in-the-fiction,’ then we need some means other than a traditional semantics that equates the truth of a claim with the existence of the people, objects, environments, actions, and events described in that claim. Walton’s thesis that fictions are collections of rules for games of make-believe is the first ingredient in such an account.
If fictions (including comics) are collections of rules for games of make-believe, then our next task is to determine what sorts of rules they are (and, relatedly, what sort of game corresponds to a fiction). The answer, according to Walton, is that the rules of the game expressed by the claims in a fiction are rules regarding what we are meant to make-believe to be true when playing the relevant game. Thus, if a novel says “It was a dark and stormy night,” then we are to imagine (when competently playing the corresponding game of make-believe—that is, when experiencing and interpreting the fiction) that it was, at the time described by the fiction, a dark and stormy night, and if Vimanarama depicts Ali as living in Bradford, then we are, when experiencing the fiction, meant to make-believe that Ali is a real person who lives in Bradford (again, when competently playing the relevant game).
Finally, if fictions are sets of rules for games of make-believe, and these rules tell us what we ought to imagine when competently playing the relevant game—that is, when enjoying, evaluating, and interpreting the fiction—then claims are true-in-the-fiction when we are meant to imagine or make-believe that they are true, and claims are false-in-the-fiction when we are meant to imagine that they are false when playing the game. Two things are worth noting here. First, we should not assume that every claim is either true or false in a particular fiction—that is, we should not assume that the principle of bivalence holds in fictions (regardless of what we think regarding the logical status of bivalence in the real world4). Fictions may leave many claims indeterminate, where we are neither required to imagine that the claim is true nor required to imagine that it is false (e.g. the claim “Ali is 20 years old,” which is consistent with what is depicted in Vimanarama but not entailed by it). Second, given a particular fiction, we should not assume that the only claims that are true-in-the-fiction or false-in-the-fiction are those claims that are explicitly asserted (or depicted, in pictorial narratives such as comics). On the contrary, competently playing the make-believe game corresponding to a work of fiction often requires us to fill in the gaps in all sorts of ways beyond what is explicitly stated in the fiction.5
The previous three paragraphs provide little more than a superficial, bare-bones sketch of Walton’s sophisticated account, which must be supplemented with additional theoretical machinery in order to deal with complex or problematic phenomena such as unreliable narrators, contradictions in fiction, and a host of other issues (all of which are discussed in detail in Walton’s 1993 text). Nevertheless, it will suffice for our purposes, as it provides the tools required to fruitfully adapt Walton’s ideas to comics. All that we require here is the idea that fictions encode prescriptions to imagine those people, objects, environments, actions, and events that are described or depicted by the fictions. Given this idea, however, one natural question to ask is whether different sorts of fiction, or even different kinds of parts of a single sort of fiction, might require us to imagine things in different ways. In particular, it is worth asking whether there are any significant differences between the ways that the images in comics require us to make-believe that certain things are true, and the way that the text in comics requires us to make believe. The answer to this question, perhaps unsurprisingly, is “yes.”
The key to our investigation is noting that comics are a multimodal medium—that is, content in comics is conveyed to the reader in more than one communicative mode (e.g. the textual and the visual), and these modes function in distinct ways. The idea that comics are a multimodal medium has become increasingly important in comics studies (and the idea that many ‘new’ media are multimodal has become increasingly important in communication studies and elsewhere). Randy Bomer describes multimodal media as those:
… where print and image do the work of meaning together, where sound and music contribute to the perspectives readers are asked to take, where bodily performance works in tandem with the written word, where print itself is animated and choreographed. (354)
Given their combination of words and pictures, comics are a multimodal medium par excellence.
Comics are not multimodal merely in virtue of the combination of words and pictures, however. In addition, text can appear within a comics panel in at least four different, standard sub-modes. Thus, comics are not only multimodal in virtue of combining images and text, but are also multimodal in virtue of the different roles, or sub-modes, played by text. The first such textual mode is narrative text (or simply narration): the descriptive text that usually appears in rectangular boxes.6 The following is a rough (but adequate) gloss on the convention7 regarding what, exactly, we are prescribed to imagine when engaging with and interpreting narrative text within a comic:
Narration Text Convention: We are to imagine people, objects, environments, actions, and events as they are described in the narrative text. Hence, narration text functions (roughly) the same way that descriptive text does in literature.
Narrative text is actually more complicated than this, since it can come in the form of (apparently omniscient) third-person narration, where the convention regarding what we are meant to imagine is roughly as above, and first-person film-noir style narration (which became prominent in the late 1980s and 1990s due to Frank Miller’s The Dark Knight Returns, amongst other works). With regard to first-person narration, we are not necessarily meant to imagine that things are as the narration describes, but are instead meant to imagine that the narrator (usually the protagonist of the story or some other prominent character) believes that things are as he or she describes them as being, or perhaps that he or she wants us to believe that things are as described (the usual worries about the reliability of first-person narrators of course apply here8).
The second way that text can appear within a comics panel is balloon text (or bubble text): the dialogue or thoughts that appears in smooth or scalloped thought or speech balloons (or bubbles), respectively. The relevant convention regarding what we are meant to make believe is roughly as follows:
Balloon Text Convention: We are to imagine hearing (or telepathically sensing?) the relevant characters speaking the dialogue (or thinking the thoughts) textually represented.
Hence, balloon text functions roughly the same way that dialogue does in literature. There is an important dis-analogy between what exactly we are meant to imagine when confronted with thought balloon text, or analogous descriptions of a character’s thoughts within prose literature on the one hand; and what we are meant to imagine when confronted with speech balloon text, or analogous descriptions of speech within prose literature, on the other. On the Waltonian picture, when confronted with speech balloons we are meant to imagine not only that the depicted words are spoken, but we are in addition meant to imagine hearing the dialogue (analogous comments apply to dialogue in prose literature). Clearly we are not meant to imagine that we somehow hear the thinker’s thoughts (it is not at all clear what thoughts ‘sound like’), nor are we meant to imagine that we are thinking those thoughts—the thoughts in question belong to the characters, not to ourselves. Thus, unlike speech balloon text, which on a Waltonian view we are meant to imagine that we hear, it seems that when engaging with thought balloon text we are meant to imagine that the thoughts in question are thought by the relevant character, but we are not meant to imagine that we somehow experience the thoughts ourselves.9
The third way that text can standardly appear within a comics panel is pictorial text: the text that appears as part of the image contained in the panel. Pictorial text includes writing, street signs, and other text that is a part of the picture. Thus, the relevant convention regarding what we are directed to imagine when confronted by pictorial text can be summed up as follows:
Pictorial Text Convention: We are to imagine seeing the text in question, and to imagine that it appears to us as it appears in the image.
In short, pictorial text is a picture of text. As noted before, it is here where McCloud’s discussion of montage text-picture interaction goes awry, since pictorial text, governed by the Pictorial Text Convention, is clearly a case where “words are treated as integral parts of the picture.” In the next section we shall mobilize the resources developed here to provide a better gloss on what McCloud likely intended.
The second clause in the Pictorial Text Convention deserves a bit more attention. Note that it does not state that the text appearing within the panel appears within the fiction (that is, fictionally appears to the characters in the fiction) exactly as it appears on the page. Instead, the presence of pictorial text (or, in fact, any pictorial content within a panel) cues an inferential process by which we deduce various fictional truths regarding the (fictional) appearance of the text (or, more generally, any people, objects, environments, actions, and events pictured) indirectly by considering how we are meant to imagine that it appears to us. The importance of such indirect, inferential fictional truth is emphasized by Patrick Maynard in Drawing Distinctions, his masterful application and extension of Walton’s ideas to pictures:
First, it is clear that what is depictively important about depictions is not only what they depict—that is, what things, situations, events and properties they get us to imagine seeing when we look at them—but how they depict them … [Stylistic differences] literally affect our activities of seeing: first, seeing the drawings, since they affect the activities required to work up a relevant imagined seeing from the marks on the surface before us; second, our ways of conceiving the subjects depicted. (220, emphasis added)
Maynard’s point, simply put, is that there is no sharp distinction between content (that part of the pictorial fiction meant to indicate fictional truths) and style (roughly, that part not directly and immediately relevant to fictional truth). On the contrary, since understanding the pictorial aspects of a pictorial fiction involves imagining that we see what is depicted in the picture and imagining, further, that things appear to us as they appear in the picture (but not necessarily imagining that these same things appear within the fiction exactly as they appear to us), we cannot immediately conclude that people, objects, environments, actions, and events appear in the fiction exactly as they appear in the panels.10 Thus, the Pictorial Text Convention given above is nothing more than a special case of the following, more general principle:
Pictorial Content Convention: When confronted with pictorial content within a panel, we are to imagine seeing the people, objects, environments, actions, and events in question, and to imagine that those people, objects, environments, actions, and events appear to us as they appear in the image.
Of course, how things are depicted in the picture will affect our understanding of, and interpretation of, what is fictionally true in the story being told. In particular, it will affect what we imagine regarding the appearance of the relevant people, objects, environments, actions, and events. But this relationship is complex: differences in the appearance of elements from one panel to the next, or one story to the next, often do indicate some interpretationally important change, but such differences need not indicate that the elements in question have themselves changed in appearance.
There are, of course, exceptions to the Pictorial Content Convention. For example, emanata—pictorial elements within a panel that conventionally represent non-literal content, such as motion lines, or little birds circling a character who has been struck on the head—are not meant to be imagined to be seen.11 Similarly, some comics incorporate one-off non-literal, metaphorical pictorial elements (such as Gary Trudeau’s representation of Bill Clinton as a waffle) that are not instances of well-established conventions but which nevertheless are not meant to be imagined to appear as they appear in the comic art.12 Nevertheless, the vast majority of imagery appearing within a comics panel is governed by something like the convention codified above, and in particular the example we will be most interested in here—the angel wasp—involves a subtle manipulation of the Pictorial Content Convention.
The fourth and final way that text can standardly appear within a comics panel is SFX text (or sound effects text): the BAMFs, THWIPs, and SNIKTs that appear within the panel, often in bold, brightly colored block letters, which indicate loud or important sounds occurring within the fiction. The convention regarding what sound effects text prescribes us to imagine is relatively straightforward:
SFX Text Convention: We are to imagine hearing a sound appropriately related to the events depicted and onomatopoetically suggested by the SFX.
Importantly, and like balloon text and narrative text (but unlike pictorial text), SFX text does not require that we imagine that we see the text in question, despite the fact that SFX text is incorporated directly within the panel rather than being separated from the pictorial content by a narration box or speech or thought balloon.
This completes our survey of the four standard ways that text can occur within a comics panel, and the conventions governing how we are to interpret—that is, what we are meant to make-believe with regard to—such text in each case. Of course, the ‘completeness’ of this taxonomy is not meant to imply that it is impossible that text could appear in a comic panel but not play one of the roles described above, since avant-garde comics artists are free to do all sorts of other things with text. The point is merely that these are the four standard, conventional ways that such text can and does so appear in mainstream Western comics, and any other usage of text within a comics panel is non-standard and non-conventional.13
Vimanarama‘s Visual Magic
During Ali’s near death experience, a number of panels represent the mystical, ineffable aspects of this experience by blending or merging two of the roles that text can play within a comics panel. In particular, the text in this passage functions as both narrative text and pictorial text. Even before the appearance of the angel wasp, Ali’s eyes leave his head and float around outside his head for a page or so (76 – 77). When his eyes eventually return to his head, one of them no longer appears to be a normal human eye—instead, the eye has been replaced by a three-dimensional block-letter version of the word “eye” (and his mouth has likewise been replaced by the word “mouth”). In the next panel, his entire upper body, including his eye, nose, face, head, ear, fingers, thumb, neck, shoulder, arm, teeth, and collar have been replaced by similar block-letter instances of the corresponding words. Ali’s dialogue in the last three panels of this page further emphasizes the blurring of representational images and descriptive text:
I came on a mission. My name is … My name is Ali. I can even spell it. (Morrison and Bond 77)
A few panels later, the angel wasp appears: a yellow winged creature made up of the words “angel wasp” written, as before, in three-dimensional block letters (82). The angel wasp (as well as the McCloudian montage version of Ali’s face, and the angel sky and angel clouds that fill the background of the panels on the page immediately prior to the appearance of the angel wasp) represents in the way that it does due to a non-standard combining of two of the standard roles that text plays within a comic.
On the one hand, the block letters that make up the wasp itself function as narrative text: although the angel wasp has wings and the bottom angle of the “w” in “wasp” functions as a stinger, pricking Ali on the nose, it would be impossible to know that this creature is a type of angel (literally, an angel wasp) without applying something like the Narrative Text Convention to the angel wasp, imagining some sort of angel-wasp hybrid as described by the text.
On the other hand, as is already implicit in the previous paragraph, the same text functions as pictorial text: we are clearly meant, at some level, to interpret this creature as looking like the word “angel wasp.” There are numerous conventional cues within the panel indicating that both Ali and his brother Omar see the wasp, and that it is meant to be interpreted as an integral part of the image itself, rather than functioning merely as descriptive narrative text: Omar reaches for the angel wasp in the penultimate panel of page 82, Ali is stung on the nose by the angel wasp in the next, and both panels include motion lines indicating the ‘spatial’14 movement of the angel wasp.
Thus, we are instructed to imagine two quite different, and in fact incompatible, things when reading this panel. On the one hand, application of the Narration Text Convention indicates that we are meant to make-believe that a creature that is both an angel and a wasp (and thus presumably resembles both an angel and a wasp, or perhaps resembles the sort of wasp-like creature that would torment angels) flies up to Ali and Omar and stings Ali on the nose.15 Application of the Pictorial Text Convention, however, indicates that we are meant to make-believe that a creature that resembles the words “angel wasp” flies up to Ali and Omar and stings Ali on the nose. It is hard to make sense of this at first glance, however (pun very much intended!). Nothing—not even an angelic creature—can simultaneously look like just a wasp, or look like an angel, and also look just like the words “angel wasp.” More to the point, it is just impossible to imagine or make-believe that an object of this sort exists! Thus, the panel seems to be prescribing incompatible imaginative acts—nothing can both look like a sequence of letters and look like what those letters describe.16
Some of the interpretational tension can be dispelled by attending to Maynard’s point discussed in the previous section, and noting that the two prescribed imaginative acts function at different interpretational levels, via different communicative modes. The Pictorial Content Thesis, and its special case, the Pictorial Text Thesis, do not require that we imagine that things—including words—appear (within the fictional world in question) exactly as they are depicted in the panel. Rather, we are merely required to make-believe that we see the people, objects (including words), environments, actions, and events depicted in the panel, and that we imagine that these people, objects, environments, actions, and events appear to us as they are depicted in the panel. Thus, when reading and interpreting the panels involving the angel wasp, we are not required to make-believe that the angel wasp genuinely appears (to Ali and Omar, for example) to be constituted out of letters. Rather, we are merely meant to imagine that we see the angel wasp, and that it appears, to us, when we see it, as if it were constituted out of letters. Thus, there is no contradiction: we are meant to imagine that an angel wasp appears to Ali and Omar, and we are meant to imagine that we see it doing so, and we are meant to imagine that, when seeing this, that the angel wasp appears to us (but not necessarily to Ali and Omar) as if it were shaped like the word “angel wasp.”17
If we are merely meant to imagine that the angel wasp appears to us, when we imagine it, as if it were shaped like the word “angel wasp,” but that it does not actually appear this way, then the next obvious question is: what does the angel wasp look like (when seen by Ali and Omar)? In attempting to answer this question, it is important to remember that the angel wasp appears to Ali and Omar during a mystical experience. A careful interpretation of these panels along the lines sketched above suggests that we cannot know what, exactly, the angel wasp (or angel sky, or angel clouds, etc.) actually looks like. On the contrary, these panels are Morrison and Bond’s way of non-literally representing the brothers’ mystical near-death experience in a manner that communicates the important narrative content of the episode to us, the readers, while retaining and emphasizing the idea that the experience, being magical, mystical, and ineffable, cannot be translated faithfully to the comics page in terms of images or text. Thus, we know that Ali and Omar have a mystical experience, and that they encounter an angel wasp, but we are also made quite aware that we cannot know exactly what the angel wasp is like: the relevant bits of information we are given are in tension with one another.18
Of course, this is far from the first time that a comic book has attempted to capture the nature of an ineffable, incommunicable mystical experience on the comics page. After all, Doctor Strange, the Silver Surfer, and a host of others have been achieving this with Kirby krackle and other depictive strategies for decades. Thus, we can (and should) ask: why did Morrison and Bond adopt the particular strategy they did adopt—manipulating the conventions of textual representation in comics—in order to represent the mystical in Vimanarama?
One way to understand what is going on here, and to further develop the analysis of the angel wasp begun above, is to note that the angel wasp is (unlike more traditional spirals-and-krackle representations of the magical) an exercise in metafiction, which Patricia Waugh defines as follows:
Metafiction is a term given to fictional writing which self-consciously and systematically draws attention to its status as an artefact in order to pose questions about the relationship between fiction and reality. In providing a critique of their own methods of construction, such writings … examine the fundamental structures of narrative fiction … (2)
Waugh clearly has metafictional (prose) literature in mind here, but the basic idea is this: Metafictional works of art are works that are in some sense or another about the process, status, or mechanics of art in general, or that artwork in particular.19 Morrison and Bond’s angel wasp can be understood quite naturally as achieving its indirect representation of an unrepresentable mystical experience in terms of exactly the sort of metafictional strategies that Waugh has in mind.
In attempting to imagine the people, objects, environments, actions, and events depicted and described within the relevant passage of Vimanarama, the reader finds himself at an imaginative impasse: when confronted by the angel wasp, the reader is instructed (on the narrative text reading of the angel wasp) to imagine a creature that (presumably) looks something like a wasp (and, in addition, looks something like an angel, or perhaps looks something like a combination of the two). On the pictorial text reading of the image, however, the reader is instructed to imagine that this same creature looks like the words “angel wasp” (or, at the very least, appears that way to the reader, given the Maynardian epicycle discussed above). Upon detecting the irresolvable tension in these competing prescriptions, the reader is forced to reflect on the manner in which the structure and content of the panels generate these prima facie incompatible imaginative requirements, to reflect on the manner that comics narratives represent their fictional worlds more generally, and to consider the different levels at which, and communicative modes within which, these apparently incompatible prescriptions function. In short, the angel wasp (and similar phenomena throughout this portion of Vimanarama) “systematically draws attention to its status as an artefact in order to pose questions about the relationship between fiction and reality” (Waugh 2)20 We shall examine the metafictional structure of the angel wasp in more detail in the final section of this essay.
Before doing so, however, two additional observations are in order. First, this sort of phenomenon seems to be exactly what McCloud has in mind when introducing the notion of montage text-picture interactions. It is not just that the linguistic expressions in a montage panel “are treated as integral parts of the picture.” In addition, in a montage text-picture combination, the words are treated as pictorial—that is, as activating the Pictorial Text Convention—in addition to activating one of the other three standard conventions for interpreting text within comics panels. In other words, a McCloudian montage combination occurs when a piece of text within a panel prescribes multiple (and sometimes apparently incompatible, as in the case of the angel wasp) imaginative acts due to the invocation of multiple conventions, functioning at distinct levels, for interpreting text and image.
Second, it is important to note that there is something more going on in these panels that helps us to understand why Morrison and Bond used this particular metafictional technique to draw the reader out of the fiction, simultaneously highlighting the nature and mechanics of fictional narrative and indirectly expressing the unrepresentable nature of Omar and Ali’s mystical experience. This particular strategy functions via the construction of an image that forces a tension between the conventions that govern our interpretation of (narrative) text and the conventions that govern our interpretation of in-panel images (including pictorial text). The key to fully understanding this choice of strategy lies in the close connections that are typically drawn between a thing (be it a person, object, environment, action, or event) and the proper, or true, name of that thing in many magical practices and traditions. For example, in his influential text on magic, The Golden Bough, James Frazier describes the connections between knowledge of the true name of a person, and power over that person, in Hindu tradition as follows:
A Brahman child receives two names, one for common use, the other a secret name which none but his father and mother should know. The latter is only used at ceremonies such as marriage. The custom is intended to protect the person against magic, since a charm only becomes effectual in combination with the real name. (245)
Of course, Hindu mythology of exactly the sort described by Frazier plays a central role in Vimanarama, and Morrison is familiar with Frazier’s work: in the 1989 Arkham Asylum: A Serious House on Serious Earth, Dr. Amadeus Arkham is shown reading the text during his descent into madness. Morrison’s own views on magic, and his adherence to a version of chaos magic as expressed in his 2008 essay Pop Mag!c,21 can be understood as a modern version of this idea, with the more abstract sigils and hypersigils (which are constructed from linguistic descriptions—that is, names—of the object of desire in question) replacing the literal use of names.22 Thus, there can be little doubt that traditional ideas regarding the power that names have over the object named, and regarding the intimate connection between the world and the language we use to describe it more generally, are central to Morrison’s understanding of magic. Morrison and Bond’s depiction of the angel wasp as simultaneously an object and a description of that object plays on this traditional blurring of name and thing named within magical traditions, further emphasizing the mystical, non-literal nature of the depiction of Ali and Omar’s near-death experience.
Metafictionality and Multimodality
With much now said about Vimanarama, it would be remiss to ignore the consequences all of this has for comics, and for metafiction, more generally.23 Thus, I shall wrap up this examination of Morrison and Bond’s Vimanarama with two questions of exactly this sort. First: if the metafictional effects of the angel wasp in Vimanarama arise due to a conflict between two or more conventions governing what we are meant to make-believe when experiencing the fiction, is it the case that all metafiction within comics (or within narrative more generally) is of this sort? Second: if the answer to the first question is negative (and I will suggest it is), then is there something specific or special about comics that makes them particularly well-suited for this sort of metafictional exploration? I will conclude the essay by arguing that the answer to this second question is a resounding “yes”!
To review: the metafictional effect of the angel wasp was generated by depicting it in such a way as to activate two distinct conventions regarding what we are meant to imagine when interpreting the episode in which the angel wasp occurs. The Narration Text Convention instructs us to imagine a creature described by the phrase “angel wasp” (that is, something that is both like an angel and like a wasp), while the Pictorial Text Convention instructs us to imagine a creature that appears to us as resembling the words “angel wasp” (and hence something that appears to us as visually quite unlike either an angel or a wasp). There is no doubt that much metafiction in comics, and elsewhere, is the result of a conflict between incompatible conventions regarding what we are meant to make-believe of exactly this sort. Is it possible that all metafiction, or at the very least, all metafiction in comics, arises via this sort of conflict between conventions, temporarily ‘bumping’ our attention from the narrative to a meta-narrative awareness of and reflection on the conflict?
An affirmative answer to this question seems difficult to defend, at least in any form that is substantial enough to be interesting. Of course, we could just stipulate that one of the conventions regarding fiction (and hence comics) is that the narrative should not “systematically [draw] attention to its status as an artefact in order to pose questions about the relationship between fiction and reality” (Waugh 2), and thus that any narrative that does so is violating this convention. But such a move is rather unilluminating. What would be more impressive would be an argument that any intentional narrative effect that forced the reader to abandon the narrative and contemplate the nature of that narrative, or the nature of narrative itself, would of necessity need to involve violating some specific convention of the sort discussed in the second section of this essay. But such an argument does not seem forthcoming: while much metafiction does proceed in this manner, including the angel wasp episode, there are no doubt other ways to systematically draw attention to a work as a created work. For example, a narrative that included two characters discussing the nature of fiction, or even the nature of metafiction, would have the requisite metafictional effect without violating any of the conventions of standard storytelling.
Nevertheless, there does seem to be something special about the sort of metafiction at work in the angel wasp. Thus, it is worth asking whether this is a special sort of metafiction, and whether the nature of this sort of metafiction is intimately related to comics in some manner. The first thing to note about the angel wasp is that the metafictional effects depend on the mobilization of two distinct conventions regarding what we are to make-believe when interpreting the content of comics panels. The second thing to notice is that the two conventions in question—the Narrative Text Convention and the Pictorial Text Convention—normally govern different sorts of text within a comics panel: the former governs text that occurs narratively, usually (but not always) separated from the image in a narration box, and the latter governs text that occurs within the panel, as a proper part of the image itself. Thus, the conflicting conventions involve different roles that text can play within a comics panel—that is, different modes.
An important reminder: comics are not merely a bi-modal medium, combining two pre-existing modes of communication (e.g. textual and pictorial). In addition, as already emphasized above, the role of text within comics can be separated into four distinct, standard roles or sub-modes, and there are distinct conventions regarding what we are meant to imagine or make-believe when confronted with text in each of these sub-modes. Normally, a particular piece of text within a comic will function in only one of these modes (i.e. it will be exactly one of narration text, balloon text, pictorial text, or SFX text). The metafictional effects of the angel wasp, however, are a result of the presence of text that functions in more than one mode at the same time, invoking multiple, conflicting (or apparently conflicting) conventions. This sort of phenomenon is not limited to comics, but can occur within any multimodal work where distinct conventions govern how we are to interpret elements from the different modes involved in the medium—we need only include an element in the story that activates two or more conventions in virtue of the fact that it instantiates two or more modes (just as the angel wasp is simultaneously a description and an image). Let us call such metafictional effects:
Multimodal metafiction: Metafictional effects generated by the presence of a single narrative element that instantiates two or more communicative modes at work within the fiction, thereby activating two or more conflicting conventions regarding proper interpretation of the fiction (i.e. regarding what we are meant to make-believe).24
Although not all metafiction needs to involve a conflict between two or more conventions governing how we are to interpret various elements of the fiction, multimodal metafiction by definition does. More importantly, however, multimodal metafictions like the angel wasp require that the fiction is told within a multimodal discourse. Thus, with the help of Morrison and Bond’s Vimanarama, we have identified an extremely interesting sub-category of metafiction that proceeds by taking advantage of the particular structural affordances of multimodal media. Further in-depth investigation of this special category of metafiction, however, will have to wait for another time.
One quick observation can be made, however: While multimodal metafiction must occur within a multimodal media (as a matter of definition), it is doubtful whether multimodal metafiction can be achieved in any multimodal medium whatsoever. The reason is simple: multimodal metafiction requires a single narrative element that is simultaneously an instance of two or more modes or submodes mobilized within the medium. This, in turn, requires that at least two of the modes or submodes in question be of the same ‘sort’ (i.e., loosely speaking, be processed via the same sense or senses). For example, the metafictional effects of the angel wasp, and its effectiveness in communicating the incommunicable nature of Ali and Omar’s mystical experience, depends on the fact that text and image are both visual modes of communication, and thus that a manifold of lines and color on the page can be interpreted simultaneously as text and picture. Similar effects would be more difficult to achieve in opera,25 where the two primary modes are of different sorts: the music of opera is auditory, while the theatrical aspects are primarily visual. It is difficult, if not impossible, to imagine how a single element of any work of art could function simultaneously as auditory and visual, and as a result it is hard to imagine multimodal metafiction in opera. This observation is not a drawback, however, but instead illustrates the special nature, and structural complexity, of multimodal metafiction.
 Of course, it is worth noting that comics need not involve this sort of interaction. Most readers are familiar with the existence of ‘silent’ or ‘mute’ comics (e.g. the comics collected in Marvel Comics’ 2002 Nuff Said trade paperback anthology), which tell a story solely through the use of pictures. Additionally, the possibility of comics that involve no pictures (but might still involve text of various sorts) is explored in detail in Cook (“Comics Without Pictures”).
 Of course, non-fictional characters, objects, events, and environments often appear within fictions. For a discussion of how we are to interpret such real-world entities within the Waltonian account, see Walton.
 The principle of bivalence is the claim that every sentence is either (determinately) true or (determinately) false. Bivalence should be distinguished from the law of excluded middle (or tertium non datur): the claim that every sentence of the form “P, or it is not the case that P” is a logical truth. The two principles are independent of one another. In my alternate role as a philosopher of logic, however, I am in fact doubtful that either the principle of bivalence or the law of excluded middle holds generally as a matter of logic, even in precise scientific or mathematical contexts! For discussion, see Cook (“Should Antirealists be Antirealists about Antirealism?”).
 Scott McCloud’s notion of closure—filling in those actions and events that occur between the panels (i.e., in the gutters, in a metaphorical sense)—is a special case of exactly this sort of gap-filling (see McCloud Chapter 3).
 It is unclear whether narration really appears ‘inside’ the panel (at least standardly), rather than ‘outside’ it (i.e. it is unclear whether the panel border proper includes the line drawn outside the narration in standard cases, or whether it includes merely the line separating narration from image). Similar worries apply to text occurring in speech or thought balloons, which might be thought to occur ‘above’ or ‘outside’ panels in some relevant sense. This worry, while important, is orthogonal to the issues examined here, since the narrative, or semi-narrative, text we shall concern ourselves with here unambiguously occurs within the image.
 It is worth emphasizing that these are clearly conventions—that is, tacit agreements regarding how agents are meant to act. There is nothing in human nature or the natural world that requires that we interpret the various modes of narrative storytelling in the manner in which we do. Rather, these rules are the result of regularities in interpretational behavior that have evolved over long stretches of time, but which nevertheless govern how interpretation (i.e. make-believe) should be carried out in the future. For an in-depth discussion of both the contingency and normativity of conventions, see Lewis.
 Of course, concerns regarding the reliability of third-person narration also arise, in roughly the manner that they arise in prose narrative. Again, we need not concern ourselves further with these complications here.
 Inferring that people, objects, environments, actions, and events (always) appear exactly as they do within the panels of a comic—that is, accepting the Panel Transparency Thesis—commits us to very weird things, such as the Joker’s physical appearance varying drastically (and in anatomically impossible ways) from one story to the next. For an extended discussion of this phenomenon the reader is encouraged to consult Cook (“Does the Joker have Six-inch Teeth?”).
 In addition, there are numerous other ways that text can appear within or ‘on’ a comics proper, including title, credits, issue number, advertisements, etc. I take these modes of text within comics to be cases where the text, strictly speaking, does not appear within the panels of the comics (even in those cases where the text might appear within panels in some sense, such as in advertisements taking the form of single-page comic strips).
 The term “spatial” is scare-quoted since this near-death experience takes place in the afterworld, and not in the ordinary three-dimensional physical world that the reader inhabits (actually) and that Ali and Omar normally inhabit (fictionally).
 There is another possibility here: perhaps the narrative reading of the angel wasp text is meant to function much like the actual term “dragonfly,” which is not meant to evoke a creature that looks both like a dragon and like a fly, but instead functions more metaphorically. This will not get us out of the present puzzle, however, since even if the angel wasp is not meant to resemble either a wasp or an angel, the description clearly suggests a creature of some sort, and presumably nothing can look both like a creature (of any sort) and like the words “angel wasp.”
 Well, it is not quite true that nothing can look both like a sequence of letters and look like what those letters describe. For example, self-descriptive phrases such as: “appended to its quotation” appended to its quotation do function this way. The angel wasp, however, is clearly not meant to function in this manner.
 One way of putting Maynard’s point: pictorial content in comics panels activates two kinds of conventions regarding how we are to conclude that the subject matter of that content appears. On the one hand, we are given (direct) information regarding how we are to imagine that the subject matter in question appears to us. On the other hand, we are given (indirect, inferentially obtained) information regarding how we are to imagine that the subject matter appears to the characters.
 Thus, by the same token, when Ali ‘s various body parts transform into words describing those parts, we are not meant to imagine that exactly this transformation takes place, but are meant to imagine that the transformation appears this way to us, and appears to Ali in a way that, perhaps, cannot be expressed in words or images.
 Morrison also plays metafictional games with Max Thunderstone’s visible thought balloons (which turn out to be visible to other characters) in The Filth. In this case, the standard conventions for interpreting balloon text (i.e. the Balloon Text Convention) is subverted in a manner similar to how the angel wasp subverts other conventions for interpreting text, but in The Filth these strategies are not mobilized in order to represent a magical experience.
 Relevant to this, but a bit tangential to the primary focus here, is the fact that Morrison views many of his comics as magical sigils themselves. For example, he contrasts his earlier work, which is often about magic, with later works such as The Invisibles as follows:
Kid Eternity was about Magic, and Zenith mentions chaos magic. But they weren’t a magical thing in themselves. All it was, I was interested in magic and I was throwing references in. Then I started to think about the potential of comics to actually do magic. I thought I’d do a comic that’s not just about magic and anarchy, but will actually create them and make them happen. (Salisbury 210)
 This sentence could be misinterpreted as indicating that I believe all comics are metafictional (a disturbingly prevalent view within comics studies). While I do not ascribe to such a thesis, there is, as I shall show in this section, something special about comics (and other multimodal art forms) that makes them especially ripe for certain kinds of metafictional exploitation. This might go some ways towards explaining the intuition that all comics are metafictional, without actually endorsing that stronger claim. I leave full exploration of this issue to another time, however.
Bomer, Randy. “Literacy Classrooms: Making Minds out of Multimodal Material.” Handbook on Teaching Literacy Through the Communicative and Visual Arts. Ed. James Food et al. New York: Lawrence Erlbaum Associates. 2008. Print.
Cook, Roy. “Should Antirealists be Antirealists about Antirealism?” Erkenntnis 79.2 (2014): 233-258. Print.
—. “Comics Without Pictures, or Why Batman #663 is a Comic.” Journal of Aesthetics and Art Criticism 69 (2011): 285-296. Print.
—. “Does the Joker have Six-inch Teeth?” The Joker: Critical Essays on the Clown Prince of Crime, Eds. Robert Weiner & Robert Peaslee. Jackson, MS: University Press of Mississippi (forthcoming). Print.
Frazier, James. The Golden Bough: A Study in Magic and Religion. London: Wordworth, 1890. Print.
Inge, M. Thomas. Anything Can Happen in a Comic Book: Centennial Reflections on an American Art Form. Jackson, MS: University Press of Mississippi, 1995. Print.
Lewis, David. Convention: A Philosophical Study. Oxford: Blackwell, 1969. Print.
Maynard, Patrick. Drawing Distinctions: The Varieties of Graphic Expression. Ithaca, NY: Cornell University Press, 2005. Print.
McCloud, Scott. Understanding Comics: The Invisible Art. New York: William Morrow, 1993. Print.
Meany, Patrick. Our Sentence is Up: Seeing Grant Morrison’s The Invisibles. Maro, IL: Sequart, 2012. Print.
Miller, Frank. The Dark Knight Returns. New York: DC Comics, 1986. Print.
Morrison, Grant. Supergods: What Masked Vigilantes, Miraculous Mutants, and a Sun God from Smallville Can Teach Us About Being Human. New York: Random House, 2011. Print.
—. Pop Mag!c in Book of Lies: The Disinformation Guide to Magick and the Occult. Ed. Richard Metzger. New York: Disinformation Books, 2008. Print.
Morrison, Grant and Philip Bond. Vimanarama. New York: DC Comics, 2006. Print.
Morrison, Grant and Chris Weston. The Filth. New York: DC Comics, 2004. Print.
Morrison, Grant and Dave McKean. Arkham Asylum: A Serious House on Serious Earth. New York: DC Comics, 1989. Print.
Morrison, Grant et al. Nuff Said. New York: Marvel Comics, 2002. Print.
Salisbury, Mark. Writers on Comic Scriptwriting. London: Titan Books, 1999. Print.
Walker, Mort. The Lexicon of Comicana. Lincoln NE: iUniverse Press, 1980. Print.
Walton, Kendall. Mimesis as Make-believe: On the Foundations of the Representational Arts. Cambridge MA: Harvard University Press, 1993. Print.
Waugh, Patricia. Metafiction: The Theory and Practice of Self-Conscious Fiction. London: Routledge, 1984. Print.