'Imitation' in this context does not mean simple mimicry but is a shorthand for a complex process of exposure to as vast an array of stimuli as possible. We leave the subconscious brain to work out all the most useful connections, and the more we expose it to, the more useful its networks become. We can guide that process a little by rewards and comments, so that the end result is that the child or pupil ends up doing more or less what we do. It has ended up in that sense as imitation, but imitation of actions along the way has not necessarily played a huge role in how we got to that situation.We don't learn to speak by mere imitation and we certainly don't subject our toddlers to Winston Churchill speeches alone.
In go, if you do, say, 100 tsumego problems a day, that sounds impressive, but you are not actually giving your subconscious brain much to work on. You'll end up as a one-trick pony. By looking at lots of pro games, with ideally other aspects of go also in the mix, you are giving your subconscious plenty of compost to turn into bio-fuel.
During that process (which doesn't apply in the same way to every activity, of course), trying to understand something is an often delusional ego-trip. You isolate A and B, have a think and come up with C. You may think you have discovered something, but the odds are that your subconscious brain has already discovered that and is just allowing you a cheap thrill - your own "discovery" is no more than a surface manifestation of that. In language, grammar rules are usually little more than such post-facto descriptions of innate processes.
To improve at go it seems better to feed your subconscious than to feed your ego.
