Personally I find Jonathan’s formulation unappealing. Of course the reading and writing of poetry takes place mainly in the mind, but to call it a kind of ‘thinking’ seems to unhelpfully foreground intellectual and analytical qualities – rather than, say, linguistic or sensuous ones.
Kasey’s approach is more sympathetic to me. To excerpt a bit that is, if nothing else, easily excerptable:
“There are contexts, obviously, when it is perfectly sensible to ask what a poem means. For a student reading Frost’s “Stopping by Woods on a Snowy Evening,” it is quite reasonable to ask what Frost means by looking into the lovely, dark woods and then saying “But I have promises to keep / And miles to go before I sleep.” It is also fairly easy to answer: he means that oblivion is tempting, but there are reponsibilities the living must honor. It is less easy to answer the question “What does the repetition of the phrase ‘miles to go before I sleep’ mean?” In fact, it may be a meaningless question.”
I’ve never felt that ‘meaning’ was a good focus when trying to understand what literature is and how it works. The very fact that paraphrase kills poetry seems enough reason to approach it from some other angle than ‘meaning’. The old MacLeish chestnut – ‘a poem should not mean, but be’ – doesn’t really get you any further. Not only does it offer no way of distinguishing a good poem from a bad one, it offers no way of distinguishing a poem from a lamp-post. More tempting is the formulation ‘a poem should not mean, but do’: a poem is a way of doing things to people with words. But then, so are novels and political rhetoric. Perhaps there is no functional difference between rhetoric and poetry – just a formal one.