One of the things about writing is that if you want to improve, you have to work at it yourself. Nobody is going to make you practice; nobody is going to force you to get better. Even taking writing classes is a choice – I’ve known people who took how-to-write classes simply to have a deadline to work to (but that’s a whole ‘nother rant).
Some writers are perfectly happy letting things come naturally. Writing is a skill, and like most skills, it improves with practice. You can get quite a long way just by writing one novel after another, without paying much conscious attention to any particular area that may need improvement.
Eventually, though, one reaches a point where the pace of improvement slows. One then has a choice: one can accept that one has reached the plateau section of the learning curve, and just continue to take whatever progress shows up in the course of one’s regular writing; or one can deliberately push oneself.
There seem to be two schools of thought as regards pushing. The first set of writers treat improving their skills as something separate from their daily word count. They take classes and write exercises that are targeted on whatever they perceive their writing weaknesses to be, then integrate their new skills with their actual writing. The second sort treat pushing themselves as part of their normal writing process; they set themselves challenges and take on stretchy projects that will force them to improve while they produce their daily word count.
I’ve always been the second sort of writer, which is a bit odd because I’m normally not much for risk-taking, and deciding to write a stretchy, different kind of book is definitely a risk when you are making your living this way. The editors may not like it; the readers may not like it; people may be so put off by whatever-it-is that they never buy any of my books ever again. On the other hand, what most readers notice first is the content; playing with things like viewpoint and structure and word choice are not so much of a risk as far as the readers go, unless I fail utterly and disastrously.
I never saw the use of exercises; all the ones I’d ever seen wanted you to write a page describing a girl in blue, or two people watching a convertible at a stop sign, or something similar, and if I wanted to do that, I’d rather do it writing pay copy.
Then in 1998 Ursula le Guin’s Steering the Craft came out. It was the advanced writing manual I’d been craving, and it was full of exercises that I would never, ever have actually written in a novel. I like to experiment, yes, but I’d never try for an entire page with no punctuation whatever, or a scene written in sentences of less than seven words each, or a 300-word grammatically correct sentence.
In other words, you can get quite a long way by just writing and by setting yourself challenges, but there are some things that are much easier to get at in the artificial setting of an exercise.
Regardless of the way one chooses to push oneself, though, diagnosis is important. It’s less important, I think, if one is pushing by writing stretchy books; “stretchy” is subjective, and as long as it feels stretchy, it’s probably working on something, even if it’s not quite the thing one thought it would stretch when one chose the project. Exercises tend to be more pointed at one specific area, and if it isn’t an area you have a problem with, the exercise probably isn’t going to be much benefit.
The two easiest things to push on are probably viewpoint and structure, because they are the two aspects of fiction that are clearest and most obvious. First-person and third-person viewpoint are clearly different and easily definable, in a way that differences in description or narrative style or backstory revelations or even plot are not. A lot of structural techniques, like flashbacks or parallel scenes or multiple viewpoint, are likewise extremely easy to define. One can set oneself a task: write this book in first person; write the next book with two viewpoint characters in strict alternation, chapter by chapter. When the book is done, it’s obvious whether one succeeded or failed.
And sometimes one discovers something unexpected along the way. The book with the strictly-alternating-viewpoints has a character who enters the Elf Hill and is out of the story for ten years; does the writer skip ahead, forcibly following the set pattern, or break the pattern in mid-book? The first-person narrator is unexpectedly possessed by a second character; how does first-person work for that?
With an exercise, one rarely, if ever, runs into these unanticipated events. Exercises are short and targeted; there isn’t time or room for one’s subconscious to take off in a totally new direction. Novels and short stories are different. They have their own agendas, which take precedence over whatever challenge the writer set herself to begin with. In the end, the question isn’t really “should I break the alternating viewpoint pattern when the character enters the Elf Hill, or not?” It’s “which way is this story going to work better?” The important thing is to end up with an interesting story; exactly how one gets there is irrelevant.