The latest two volumes of Martin's Song of Ice and Fire have been poorly received compared to the series' earlier books. What's interesting here is that Martin's original plan did not include these books. Rather, these books were a solution to a problem Martin had:

The fourth book, tentatively titled A Dance with Dragons, was to focus on Daenerys Targaryen's return to Westeros and the conflicts that creates. Martin wanted to set this story five years after A Storm of Swords so that the younger characters could grow older and the dragons grow larger. [...] A long prologue was to establish what had happened in the meantime, initially just as one chapter of Aeron Damphair on the Iron Islands at the Kingsmoot. Since the events in Dorne and the Iron Islands were to have an impact on the book, Martin eventually expanded the Kingsmoot events to be told from three new viewpoints since the existing POV characters were not present in Dorne and the Iron Islands.

In 2001, Martin was still optimistic that the fourth installment might be released in the last quarter of 2002. However, the five-year gap did not work for all characters during writing. On one hand, Martin was unsatisfied with covering the events during the gap solely through flashbacks and internal retrospection. On the other hand, it was implausible to have nothing happening for five years. After working on the book for about a year, Martin realized he needed an additional interim book, which he called A Feast for Crows. The book would pick up the story immediately after the third book, and Martin scrapped the idea of a five-year gap. The material of the 250-page prologue for the beginning of A Feast for Crows was mixed in as new viewpoint characters from Dorne and the Iron Islands.

But fan reaction suggests this solution was a poor one, resulting in stretched-out books whose object was to impart key information and to "keep track" of characters who weren't moving as far or doing as much as those of the earlier volumes.

Taking in good faith Martin's descriptions of the difficulties he had with the five-year gap, how else might Martin have solved these difficulties?

I am less interested in how Feast and Dance might have been condensed or abbreviated (that would be an editing question). What I'd like is suggestions for plot/story structures that might have been more appropriate to Martin's purposes. In other words, given that Martin had these difficulties and considered an additional piece necessary - what would have been a better format and structure for such a piece?

  • This question is part of my effort to consider "case-study questions" as a question type for Writers.SE. I'd be happy to hear what people think of this type of question, and what guidelines might be appropriate. If you have any thoughts, please join the meta discussion!
    – Standback
    Mar 5, 2012 at 10:06
  • When Feast was released, it was more of an act of desperation, I felt. GRRM just took an amount of character plotlines, bunched them together and called it a book. He has later said that this was a poor decision. I'm not sure what you are asking here. How to fill the 5-year gap, or how to replace the 5-year gap?
    – TLP
    Mar 5, 2012 at 16:42
  • @TLP : Either is fine. The 5-year gap is the problem, how you suggest solving it - preserving it or eliminating it - is the answer I'd like.
    – Standback
    Mar 5, 2012 at 16:56

4 Answers 4


GRRM's primary goal with Feast was to fill in crucial details, whereas most of the volumes had the goal of portraying epic plot and conflict. I'd suggest that the structure used in the previous volumes - interlacing novellas told from different viewpoints, collectively portraying a larger plot - is a great match for the latter goal, and a poor match for the former.

Interlaced narrative is great for a multi-faceted plot. Individual details, on the other hand, don't need to be forced into a single narrative. There's no inherent reason to try and swirl them all together - while some creative interactions can yield brilliant, unusual plots, it's also possible to simply convey the different points separately, one by one.

In other words, I think Feast would have been better conceived as an anthology of short stories. Short stories are an excellent way to convey focused information. They can zoom in on a particular character, time, place, or society. They're less limited by the narrative needs of a full novel (or a complete volume of a saga). And they can let individual characters quickly sum up a lot of what they're going through, while concentrating on a single, specific exciting event. (Interlaced narrative doesn't let you do that, because you need to either spread the single event over the whole book, or else have lots of exciting events, not one.)

For example, Brienne's oft-maligned arc could have been confined to a short story - for example, beginning with her expecting to find Sansa and discovering the outlaws instead, and then speeding right up to her finale with Catelyn. As a short story, this would have worked very nicely. I think many of the other arcs could be similarly condensed; similarly, using unusual and one-time viewpoints (as in the Dornish and Kingsmoot arcs) would have been easier, more natural, and more flexible - in fact, other more unusual framing POVs could have been employed as well.

As an aside, I'd suggest that considering individual short stories (rather than a full double-volume novel) would make it much easier to identify when a story sags or meanders. This could also be fixed more easily, by shortening the story without rearranging the structure of the entire book.

Lastly, I think an anthology of "tales from the interim" would have better conveyed the nature of the book, its inherent lack of epic-scale developments, its focus on things other than advancing the main characters and plot. It would have stood out as a volume that was fundamentally and intentionally different from the others, and I think it could have served that purpose well.

  • I think you are missing the point, like GRRM did. He introduced in the first three books a "battle" between two families providing several POVs on both side (three with Targaryen). The reader bought that story. But Feast does not bring that story any further. Even worse, it introduces new POVs nobody is interested in (who cares about a Kingsmoot if Others are in the north and dragons oversea?). Brienne's arc at least hold a connection to the previous characters, so it was for me one of the more interesting nevertheless disappointing ones. Jun 5, 2012 at 10:48
  • @JohnSmithers: I'm not sure what point you think I've missed. - are you saying that since Feast doesn't advance the primary story, it should have been excised entirely? Dance focused more on familiar characters - did that solve the issue for you?
    – Standback
    Jun 5, 2012 at 11:48
  • Excised sounds good, Standback. If it does not add to the story, kick it. I really do not understand the purpose of Feast. That book is only bridging weeks, maybe months, with mostly irrelevant information (not showing development of the main characters). If he keeps that pace, he needs a ton more books for the five-year-gap. I haven't read Dance yet (and refuse to do till the next volume is published), because the critiques say it's just like Feast. Maybe the right characters this time, but the story does not go any further. The suspicion arouses that GRRM doesn't know how the story ends. Jun 5, 2012 at 14:41

This won't be an essay, as I'm short on time, but my personal answer is this:

Martin's world expanded exponentially with each additional book in the series, and along with all the new information came many new characters, a lot of them to carry exposition and information, yet initially served no real purpose of their own in the sense of the grand epic that was being told. None of the readers was emotionally invested in Aeron Greyjoy, for example, but through him we gained a better understanding of the background that shaped his nephew Theon's character.

This became a problem when the expository characters became secondary characters in their own rights – now each of them had to have his or her own arc, background, characteristics, etc. Martin felt, whether rightly or not, that he couldn't just leave these new characters in the lurch when he executed his originally-planned five-year time skip.

However, what he failed to realize was that there were only so many fictional characters readers could care about. They tended to bond with the ones introduced in the first novel and see those subsequently introduced as no more than tools of exposition, information, or adversity for the primary characters already established. No one cares about Aeron's problems, and I have personally long suspected Brienne to be nothing more than a second attempt on Martin's part to create a more sympathetic and interesting Ser Loras Tyrell, who's been hanging around for five novels now and has yet to serve any real purpose. In that he only partially succeeded, as Brienne serves a classic example of a second-tier character in which no one is really invested, despite Martin's efforts to the contrary.

This crucial misjudgment on Martin's part was what led him to scrap the time skip to begin with, and it was an error from which the series has yet to recover, and I seriously doubt it ever will. Martin should have realized his primary characters was where the focus of the story always lay. He should have either abandoned the tertiary ones or relegated their fates briefly after the fact, no matter how unfair to them and their fans it seemed, and stick to his original plan. The fact he didn't do so served to seriously water down the focus and intensity present in the initial first three books as well as eliminate the powerful dramatic effect the skip would have possessed – we would have seen our beloved characters beaten down only to rise again, observe the changes that whould have occured in them for ourselves without having to agonize over every tiny step of their way. Some of the changes the loss of the time skip necessiated - such as suddenly turning Cersei from a flawed, conniving, yet ultimately brilliant player into a drunken moron complete with a previously never-heard-of mystical prophecy tailor made to explain her obvious overreaching and mistakes - could have been avoided, and our respect for the character would have remained in tact.


It's interesting to see @Standback propose dealing with the problem by creating an anthology of short stories. Clearly there's a dissonance of execution between the first installments and the last two.

But perhaps the real resolution of fitting format and structure with story would require stepping outside the print notions of (one-volume) books and the rejected two-volume novel.

In a fully electronic age, the 2000 pages of Feast and Dance wouldn't have had to be shoehorned into two arcs (as two books) but could have been released as multiple segments of one continuum. In print, ending one character's chapter requires the pre-release agents (author, publisher) to decide which character's story will be brought up next. Imagine what it would be like if instead the post-release agents (individual readers, book groups) were presented with multiple choices.

A future Martin-like author might tell such a saga with such different focal points as epic novel and short-story as suggested but also with more, running the gamut from internet meme to forum thread to graphic novel.

Well, I don't know what the epic story of the future will look like, but it very likely will be more like an enormous interconnected website than a boxed set of paperbacks, to be read in no prescribed order and not seriatim.


It's not an easy question you pose. For one thing, this problem stumped George R R Martin, who should be considered more experienced than most users on this website. For another thing, we do not know all the variables involved that depended on the 5-year gap. I would say that this question is far too complex to handle in this format, and would be better off handled one piece at the time, e.g. "How to solve the dragon's growth", "How to solve the Stark kids' growth", etc.

My take on the 5-year gap is that what GRRM was envisioning was something like the time that passed between episode 3 and 4 in the Star Wars trilogy, when we "begin" by watching Luke grow into his hero role. Well, something like that, anyway. His problem was that his original plans were thwarted by the story growing too much on the way there, to the point where he felt it was no longer feasible with flashbacks and other means of telling a story retrospectively.

I don't think there is a solution. You either have gaps, or you don't. You can allow time to pass by simply not mentioning it. The problems begin when you start thinking, like GRRM did, that you can allow time to pass by adding fluff and new storylines (Brienne, Ser Arys Oakheart, Arianne, Asha Greyjoy, Reek, etc). By fluff I mean things like trying to add complexity and richness by enumerating long lists. For example, trying to illustrate a forest like so:

There were all kinds of trees: tall ones, thick ones, short ones, wide ones, spruce and pine, deciduous trees and coniferous trees, trees with faces and trees without faces. Even mighty stumps that had once been trees. There were a lot of trees.

Heck, I can't even do it well, but you get the gist I hope. I am particularly un-fond of this writing style, which I consider to be just a lazy brute-force attempt at adding spam in order to simulate some kind of complexity. The human mind does not easily adapt a set of varied information into the concept of complexity. Rather, the mind simulates complexity by being given examples of it:

From atop the tower, the forest seemed endless, disappearing in the morning fog, stretching out into the white nothingness.

Fluff can also be extending pointless scenes, such as Tyrion signing a bunch of papers, and delineating time by writing "Tyrion signed another paper" from time to time. Or by people shoving through crowds and mentioning people along the way -- for 20 pages straight.

The point is that fluff does not bring the story forward. While reading 200 pages of fluff, time may not have passed at all in another story line. This is also explicitly mentioned as being the case in the prologue to the recent books. So what did he really gain by doing this? In my opinion, nothing at all.

What he has done is add a sort of buffer in the reader's mind, a long stretch of pages that may be interpreted as a long time passing. But in doing so, he has sacrificed a lot of the reader's goodwill and diminished the effective style of writing that permeated his first three books. Go back and reread A Game of Thrones, if you haven't done so in a while, and you will be astonished by how quickly so much happens there, compared to later books.

While important information was given along the ~3000 pages that make up Dance and Feast, much of it very interesting, actually, the critical failure is that most of the story lines we read are uninteresting. I think GRRM had the right idea from the start, and that a 5-year gap would have been preferable to 5-years worth of fluff.

Also, while a 5-year gap might be hard to bridge, it could/should have been abridged in shorter pieces. He could have made note of things that were to happen in this time span, and strung them together, while making obvious references to time passing. Such as showing the Dragon's growing in size. I think it is important not to become a hostage of your own story line, but always remember to keep things interesting.

For example, while Brienne's story might be interesting on its own merits (her being a possible descendant of Dunk's), writing hundreds of pages of her making a journey were know from the start to be pointless, during which nothing really happens, except at the end, is to annoy the reader, I feel. I feel that the story could have been left with only Jaime's perspective, and been all the better for it: We would follow Jaime's journey towards redemption with some interest, as he is a main character, which Brienne is not.

  • I appreciate the lengthy response, but most of your comments refer to criticism of Martin's writing - too long, too slow-moving, too much padding. I asked what different structure he might have used, but you didn't really offer one :-/
    – Standback
    Mar 6, 2012 at 14:26
  • @Standback Writing is structure. If you ask a vague question, you will get a vague answer. I cannot go into each character arc and specifically give advice for each of the characters. Well, I could, but it certainly would not be a concise answer. =) What kind of answer were you looking for?
    – TLP
    Mar 6, 2012 at 16:04
  • @Ido has kindly provided me a positive example for comparison: he suggests that the structural flaw with Feast and Dance is in including too many secondary characters, and trying to give them plots while the reader considers them secondary. You criticize the same arcs, but your criticism is "they were poorly written," which is content/style/pacing - not structure.
    – Standback
    Mar 6, 2012 at 17:42
  • "The problems begin when you start thinking, like GRRM did, that you can allow time to pass by adding fluff and new storylines (Brienne, Ser Arys Oakheart, Arianne, Asha Greyjoy, Reek, etc).", "...the critical failure is that most of the story lines we read are uninteresting." As well as the entire last paragraph. "Fluff", my main topic, to me include uninteresting story lines. So, no, that's not an example: Ido's answer says pretty much the same as mine, but in a different way.
    – TLP
    Mar 6, 2012 at 17:59
  • Ido explains why the problem is a structural one, not a quality-of-writing one. IMHO.
    – Standback
    Mar 6, 2012 at 19:25

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.