Learning to read is not something that ever finishes. I'm still learning new words and improving my understanding of morphology, etymology, syntax, style, and so forth. But Helen can read now, in the sense that the problems she faces reading are mostly the same ones an adult faces, albeit at higher frequency and in a different mix, rather than the basic decoding she was struggling with a year ago.
Talking to the professor of phonetics at work, I discovered that recent studies involving brain scans prove I was basically right: people reading don't activate many of the areas used when listening to speech and the processes are clearly different. Reading is a heuristic process involving word prediction based on features such as first letters, length and shape, which have no aural parallels, as well as the pragmatic and syntactic context.
I don't think this is inconsistent with phonics being the most effective way to teach reading for most children. But that is a scaffold that gets thrown away as reading fluency is established, only deployed by fluent readers in extremis. (School's "alien words" phonic check preparation notwithstanding, Helen's reluctance to revert to phonics is notable with unfamiliar names, where she hesitates before trying to find a plausible sounding-out, almost as if she has to switch from one "drive chain" to another.)
One of the big constraints is pragmatic. Anything scary can put her off proceeding further. And we're not talking about age-inappropriate material here: even a character being turned into a frog can be a problem, as can social awkwardness. Sometimes reading the last page, to confirm that the protagonist is alive and happy — and no longer a frog — can reassure her. (In a way I share this concern: I don't like real horror fiction or the kind of crime fiction that verges on it.) Often starting a story is the hard part — series are easier — so perhaps she finds the emotional engagement fiction demands off-putting (she seems more confident tackling non-fiction).
Lexical complexity is still a problem. There are plenty of words Helen doesn't know or has only a partial grasp of, and she has a lot to learn about using morphology and etymology to guess meanings. A fair bit of this is domain-specific, of course, and reflects her lack of knowledge of whole areas. But the extent of her general vocabulary continues to astound me. The difficult words in the first two pages of The Lord of the Rings, read cold, were "tongues" (a decoding problem) and "reminiscences" (most likely an unfamiliar word).
It's hard to get a grasp for her ability to tackle syntactical complexity. The only really gnarly sentences she's encountered are in passages I've read to her, and much academic and literary prose is designed to be read and works poorly when listened to. And syntactical complexity usually accompanies conceptual complexity, either in technical and academic writing or in literary prose and poetry, making it hard to distinguish them. Helen didn't take to (my reading) the opening sentence of the English translation of Jose Saramago's The Stone Raft —
When Joana Carda scratched the ground with the elm branch all the dogs of Cerbere began to bark, throwing the inhabitants into panic and terror, because from time immemorial it was believed that, when these canine creatures that had always been silent started to bark, the entire universe was nearing its end.
— but was that because of the syntax or just conceptual overload?
We've read a fair bit of poetry, so she's had experience with non-standard word order, and with similes and analogies, but it's mostly written for children or derived from an oral tradition (Homer, Gilgamesh). I think The Wasteland and Paradise Lost will have to wait quite some time. She copes fine with simple narrative complexity, as in The Odyssey (nested narratives) and Marta Morrazoni's The Invention of Truth (unrelated alternating plot strands). And her memory is better than mine, so I expect she'll cope better with the sheer number of characters in the Mahabharata or War and Peace than I do. But she has not yet encountered stream of consciousness, unreliable narrators, or any number of other complexities.
For non-fiction, the major constraint is (as with adults) the assumed background knowledge. This is very much domain-specific: Helen is remarkably competent, and confident, with history or biology aimed at older children or even adults, presumably because Camilla I read a broad range of history and science with her. On a good day she can tackle a British Museum exhibition. Levels of abstraction are obviously a constraint. She was noticeably happier — "more! more!" — with my reading bits of narrative history from Jonathan Israel's The Dutch Republic than with more analytical passages from James Scott's Against the Grain — "enough of that" — and I have to carefully select and drastically reformulate bits of mathematics texts to make them accessible.