I’m in the process of proofreading Shannivar, the second volume of a fantasy trilogy (The Seven-Petaled Shield). As is typical, I swing between elation at what I’ve accomplished (“This is brilliant!” “I nailed it!”) and wishing I could take the whole 
thing apart and put it back together right. I’m also reflecting on the challenges and joys of “middle books.”
Middle books present particular challenges that reflect whether they are truly the second of three parts or whether they are “the continuing adventures” of a successful-but-complete first book. A trilogy is like a three-act structure, only on steroids. The whole work gets fractal, if I’m using that term correctly. Overall, you have three books, but each book has a three-act or four-act architecture within it. And each scene has its own buildup and partial resolution of tension, etc.
In a successful trilogy, the second book soars. It takes off like a rocket from the firm foundation that has been established in the first book, using the unresolved or partially-resolved tension to get a running start. There’s a great freedom in middle books because the “problem” – the threat or goal – has already been established. It may be clarified or elaborated or modified, but we’re not starting from scratch. Now we have the freedom to ratchet up the tension, increase the stakes, have a gazillion things go wrong wrong OMG DISASTER. I wonder if many middle books have a soggy quality because they limit themselves to “getting from here to there” instead of “swamp-malaria-alligators-sinkhole-hurricane-ALIENS FROM SPACE-PLANET GO SPLODY!” Middle books work when every turn makes the situation exponentially worse and our characters have to work that much harder and suffer that much more.
Like the first book and stand-alone novels, middle books need a three-part structure, too, with plot twists and rising action. Because they begin at an unstable place, the initial establishment of character and problem (something the protagonist wants or must prevent, etc.) can be accelerated. That is, we’re not introducing the problem or what leads up to it; we’re beginning with what makes that problem even more pressing. I love it when an author is skillful enough to do this at the same time as reminding me of the crucial points from the last book, particularly when it may have been a year or more since I’ve read that. Backstory infodump, whether in narration, flashback, prolog, or inane “as you know, Bob” dialog, squanders the momentum that is the parting gift of the first volume to the second. These things are counterproductive because they are static, backward-looking. The result is a flabby middle book, one that is stuck with having to “come up to speed” from a dead-in-the-water halt.
In much the same way the ending of the first book gives the second a running start by incomplete resolution of tension, the second book ends in an even less stable, more unsettled place. I don’t mean that a cliff-hanger is necessary, only that the reader is left with the certain knowledge that the story is not yet over. We are at a temporary resting-place, with even greater perils to come. A partial resolution is necessary for the dynamic shape of the book, but because the book is not complete in itself – it is part of a larger overarcing tale – all the ends should not be neatly tied up. (Well, unless we’re talking about our hero being bound in chains while a dragon approaches…)
At the beginning, I mentioned “middle” books that are really sequels to stand-alone novels. The author may leave specific elements unresolved, “open doors” if you will, when submitting a book in today’s market. Publishers are increasingly reluctant to offer multi-book contracts to unknown or lesser-known authors, or if the genre is one in which an author does not already have an established following. Established midlist authors face the same situation. On the one hand, only a single book contract may be offered; on the other hand, if the book does well, the publisher is likely to ask for more in that world with those characters. That’s where the “open doors” come in, or an author might develop secondary characters or introduce problems not present in the initial volume. I would say these are not true “middle books” but sequels in a series. Characters and their conflicts may carry over from one volume to the next, but the progression structure is essentially episodic, not one long story told over three acts. This is neither good nor bad, apples nor oranges; it presents its own set of challenges and rewards.
