Abstract
This paper examines the fifth major vector of fragmentation in A Song of Ice and Fire: the proliferation of television adaptations and licensed products that now generate canon rather than merely reflect it. Where the first four papers traced fragmentation within the published textual property, this paper traces fragmentation across media boundaries. Game of Thrones, broadcast on HBO from 2011 through 2019, established the pattern by producing narrative material that overtook its source and by arriving at an ending the novels have not yet reached. House of the Dragon, ongoing since 2022, has continued and intensified the pattern by dramatizing a pseudo-history whose source text deliberately preserved ambiguities that television cannot. Further adaptations are in production or development, and a substantial body of licensed material exists across games, comics, and other products. Each of these constitutes a version of the property with some claim to authority, and the collective effect is that the reader and viewer now face a landscape of authoritative sources whose relative standing is contested and, in several cases, genuinely undecidable.
1. Game of Thrones (2011–2019) as the First Major Divergence
The HBO adaptation of the main novel sequence, developed by David Benioff and D. B. Weiss, aired in eight seasons from April 2011 through May 2019. For the first four seasons the series drew closely on the published novels, with A Game of Thrones and A Clash of Kings adapted in single seasons and A Storm of Swords split across the third and fourth. From the fifth season onward the series began to outpace the published source material, drawing increasingly on Martin’s notes and outlines rather than on finished prose, and by the sixth season the series had effectively moved beyond the published record and was constructing its own continuation of the story.
Several features of the adaptation deserve attention for the purposes of this paper.
The first is the compression of material. The novels’ viewpoint structure, with its many narrators and its parallel geographies, does not translate directly to television, and the showrunners made substantial consolidations from the beginning. Characters were combined, subplots were excised, and timelines were adjusted to permit the parallel tracks of the novels to converge more quickly than the novels themselves permit. These decisions were necessary for the television form, but they were also interpretive: they selected among the novels’ possibilities and foreclosed some that the novels leave open.
The second is the production of material not present in the novels. Certain scenes and sequences in the adaptation have no counterpart in the published books, either because they dramatize events the novels report in passing, or because they develop characters whom the novels treat more briefly, or because they supply connective tissue that the novels accomplish through different means. This material is partly derived from authorial consultation, partly the invention of the showrunners, and partly the work of individual writers and directors. Its canonical status has never been fully adjudicated, and readers of the novels disagree about how much of it, if any, should be treated as reflecting the novels’ own world.
The third, and the most consequential, is the ending. The final two seasons of the adaptation, covering material the novels have not yet reached, arrived at resolutions for the principal storylines that are now, for much of the global audience, the effective ending of the story. The novels, when they are completed, will arrive at their own ending, and the two endings will stand alongside each other. Whether they will be substantially the same, substantially different, or different in respects that are more significant than the respects in which they are similar, is a matter on which reasonable judgments can be made but not settled.
This is a new situation in the history of fantasy publishing. Earlier novelistic properties that were adapted for film or television produced adaptations after the source novels were complete, and the adaptations therefore drew on finished material. Game of Thrones drew on unfinished material and produced its own ending in the absence of an authorial ending, and the result is that the property now has two endings planned: the adaptation’s, already broadcast, and the novels’, still to come. The existence of two endings is itself a form of fragmentation, and it is a form that cannot be resolved by any future authorial decision. Even if the novels’ ending is eventually received with universal enthusiasm, the adaptation’s ending will remain in existence, and viewers who encountered the property through the adaptation will have received that ending as the story’s resolution.
The reception of the adaptation’s ending was, as a matter of public record, substantially hostile. Critics and viewers objected to compression, to character decisions, and to thematic resolutions that many regarded as unearned. The strength of the hostile reception has produced a particular feedback effect on the novels: the authorial ending, when it arrives, will be received not merely as an ending but as a response to an ending already rejected, and the terms of its reception will include comparison with what the adaptation produced. This is an unusual pressure to place on a concluding novel, and its effect on the finished work, if it can be assessed at all, will be assessable only when the work appears.
2. House of the Dragon (2022– ) and the Live Revision of Fire & Blood
House of the Dragon, developed by Ryan Condal and Martin, premiered on HBO in August 2022 and has continued across subsequent seasons. It adapts the Dance of the Dragons as presented in Fire & Blood Volume I and the predecessor novellas. As noted in the third paper, the Gyldayn frame in the source text presents competing versions of contested events and declines to adjudicate among them, and the adaptation must therefore choose versions in order to dramatize them.
The choices the adaptation has made have been extensively documented in interviews and in critical commentary. In several prominent cases the showrunners have selected versions attributed to Mushroom, the dwarf jester whose Testimony the source text treats with particular skepticism. In other cases they have selected versions attributed to the more respectable sources. In still other cases they have constructed composite dramatizations that draw on multiple sources or that introduce material with no direct source in the book.
The effect of these choices is a live revision of Fire & Blood. The adaptation does not simply present one reading among several; it presents, for the audience of the adaptation, the reading, and the competing versions that the book preserved become, for that audience, non-existent. A viewer of House of the Dragon who has not read Fire & Blood will typically believe that events happened as the adaptation depicts them, and this belief is not, in any strong sense, mistaken. The adaptation has authorial imprimatur through Martin’s involvement, and the depicted versions are among those the source text contemplates. The viewer is receiving a legitimate version of the history, but is not receiving the preserved ambiguity that the book offered.
The pressure this places on Fire & Blood Volume II, discussed in the third paper, is substantial. The portion of the Targaryen history that precedes the Dance has been adapted; the portion that follows the Dance has not. Volume II, when it appears, will therefore be received by an audience that has seen the Dance dramatized and that holds particular versions of its events as canonical. The written continuation must be coordinated with the dramatized precedent, or must diverge from it at acknowledged cost, and in either case the authorial freedom to write the continuation as it would have been written in the adaptation’s absence has been constrained.
A further feature of the House of the Dragon situation is the ongoing nature of the adaptation. Game of Thrones was, by the time the present paper is written, a concluded artifact; the discussion of its relationship to the novels can take the adaptation as fixed. House of the Dragon is unfinished, and its future seasons will continue to produce canonical material. The adaptation’s choices about later portions of the Dance, about the post-Dance regency, and about any further material that falls within its eventual scope will continue to constrain the written chronicles as those chronicles attempt to reach readers. The two projects are, in this sense, in real-time competition for canonical authority, and neither the author nor the showrunners have offered a framework for adjudicating between them.
3. Future Adaptations in Development
Several further adaptations are in production or development as of this writing.
The Dunk and Egg adaptation, developed for HBO under the title A Knight of the Seven Kingdoms: The Hedge Knight, is in production and has been discussed in the second paper. The pressures it places on the novella cycle are the pressures Game of Thrones placed on the main novels at an earlier stage of that project’s development: adaptation of a significantly unfinished source, with the attendant risk that the adaptation will produce its own continuation of material the novellas have not yet written. Because the Dunk and Egg cycle is episodic rather than serial, the adaptation may be able to handle the incompleteness differently from Game of Thrones, but the structural pressure is real and will need to be managed.
Nine Voyages, an animated project reportedly covering the adventures of Corlys Velaryon during his travels before the events of House of the Dragon, has been discussed publicly and has moved through stages of development. Its status at any given moment is difficult to verify from outside the production, and it may or may not reach finished form. If it does, it will add another authoritative version of material that the supplementary texts currently treat in reference form, and it will produce further canonical decisions about events the source material treats only briefly.
Additional animated projects have been discussed, including concepts focused on Aegon’s Conquest, on the Golden Empire of Yi Ti, and on other regions and periods that the main property has mentioned only in passing. The status of these concepts varies widely, and several have been discussed publicly only to be shelved, revived, or reconfigured in subsequent announcements. The pattern is consistent with the broader history of HBO’s development of the property: a large number of concepts are considered, a smaller number reach production, and a smaller number still reach broadcast.
The general point is that the adaptation pipeline is open-ended. Further projects will appear, each will produce canonical material, and each will interact with the written property in ways that the preceding adaptations have modeled. The written property is no longer the only generator of canon, and the expectation that it will regain a privileged position is not, at this point, reasonable.
4. Licensed Materials
A substantial body of licensed material exists alongside the televised adaptations. It includes video games, tabletop roleplaying games, collectible card games, board games, comics, graphic novels, and various ancillary products. The canonical status of this material varies widely, and in most cases has been neither formally adjudicated nor formally disclaimed.
Several categories deserve mention.
The graphic novel adaptations of the main novels, published by Dynamite and Bantam across the 2010s, dramatize the novels’ material in comics form without substantial deviation. Their status is that of straightforward adaptation: they present the novels’ events in a different medium, and they do not generate canon independently.
The video game Game of Thrones: A Telltale Series, released in 2014 and 2015, places original characters within the televised adaptation’s setting and advances a side narrative that neither the books nor the show dramatize. Its status is more complicated, because it introduces named characters, events, and houses that have no counterpart in the primary property, and the question of whether these additions should be treated as part of the world or as parallel invention has not been settled.
The tabletop roleplaying game material, published across several editions by different companies over the years, includes substantial worldbuilding content that draws on authorial consultation without being formally published under Martin’s name. Some of this material has been subsequently confirmed or contradicted by later authorial publications, and some remains in the semi-canonical middle category discussed in the fourth paper.
The collectible card game A Game of Thrones: The Card Game, published in two editions by Fantasy Flight, produced quotations and flavor text attributed to characters and sources from the novels. Some of this flavor text was authorially supplied; some was written by the game’s designers with Martin’s approval; some was written within editorial parameters that did not involve direct authorial review. The resulting corpus of quoted and attributed material is substantial, and parts of it have entered fan discussion as though they were drawn from the novels themselves.
The general pattern in the licensed material is that canonical status is a matter of degree rather than of category. A licensed product with heavy authorial involvement is closer to canon than one with light involvement. A product that dramatizes material the novels have already produced is closer to canon than one that generates new material. A product that has been subsequently referenced by the author is closer to canon than one that has been ignored. The gradations are fine, and no published standard exists for applying them.
For the reader or viewer attempting to navigate the property, the practical effect is that the licensed material operates as a kind of penumbra around the canonical center. Most readers ignore it. Some readers engage with selected portions of it. A small number of specialized readers engage with it systematically, and these readers have constructed frameworks for incorporating it that other readers have not endorsed. The property has, in this sense, multiple canons operating at once, each internally consistent but none fully reconciled with the others.
5. The Unresolved Authorial Position on Cross-Media Canon
A final feature of cross-media fragmentation deserves separate treatment. Martin has not, across the long history of the property, issued a consistent or comprehensive statement of how the various media versions should be related to one another. His public remarks have been, in general, respectful of the adaptations and the licensed products, and he has declined to issue the kind of authorial rulings that would establish a clear canonical hierarchy. The consequence is that the question of cross-media canon remains unresolved, and the readership and viewership must adjudicate it as best they can.
Several considerations bear on the authorial silence. The commercial relationships with HBO and the licensees are complex, and authorial rulings that disfavored adaptation choices might affect those relationships. The dramatic choices made by adaptations are sometimes improvements on the source material, or are at least defensible as alternative readings, and the author may reasonably decline to overrule them. The novels themselves remain unfinished, and authorial statements about cross-media canon would commit the novels to positions the novels have not yet taken. And the sheer volume of adaptation and licensed material now exceeds what any author could comprehensively rule on.
The result, however, is a property without a canonical framework. Readers disagree about what counts, and the disagreements cannot be settled by appeal to authorial ruling, because no such ruling exists. This is a genuinely unusual situation for a property of this scale, and it distinguishes the Ice and Fire landscape from several other large fantasy properties in which authorial or estate-level rulings have established the canonical status of various derivative works.
One effect of the unresolved position is that the practical canon is negotiated through reception. A reading is canonical, in this de facto sense, if substantial portions of the readership and viewership treat it as such, and it is non-canonical if they do not. This is not a stable basis for canonicity, but it is the basis that currently operates, and the fragmentation traced throughout this paper is in large part a consequence of the absence of a more stable basis.
6. Conclusion
Cross-media fragmentation is the most public form of fragmentation in the property and, in many respects, the most consequential for the general audience. It is the form in which fragmentation is experienced by viewers who may never have opened a novel, and it is the form through which most of the global audience now encounters the world. The adaptations have produced their own endings, their own resolutions of contested histories, and their own authoritative versions of material the books present in different form. The licensed products have further multiplied the canonical sources. The authorial position has remained, for reasons both commercial and artistic, carefully non-committal.
The audience that results is fragmented in ways the earlier papers have anticipated and that the sixth paper will attempt to synthesize. Novel-first readers, show-first viewers, and House of the Dragon-first audiences hold different versions of the world and read or watch each new installment against different backgrounds. The property is the same property, in the sense that a single creator has overseen its development and that the various versions share substantial material. It is not the same property, in the sense that its canonical boundaries are drawn differently by different members of its audience.
The sixth and final paper will take up the task of synthesis. It will propose a taxonomy of contradictions across the corpus, examine the limits of the unreliable-narrator defense, consider the cohorts of the divided readership, weigh the commercial incentives against completion, and offer recommendations for the editorial stewardship of a property whose coherence has become a matter of active negotiation rather than of settled fact.
