"Regarding Twain's complaints, the biggest lesson I learned from learning german at an older age was: languages are nonsensical, illogical, and the sooner you stop trying to make sense of it all, the faster you will advance. Children do not learn languages by being analytical and by comparing to what they already know, they learn by just accepting it as it is."Oh how absolutely true this statement is. As an adult, I find myself deconstructing and atomizing every aspect of the language I'm learning and then reconstructing it bit by bit into English. Not only does this approach slow one's progress down to a crawl but also it limits one's fluency even after one has some basic understanding of the language. [Incidentally, I found this to be more of a problem with French than with German even though I was taught French at a much younger age.]
From my observations, I have noticed that I'm not alone in having adopted this approach, especially so for some of us with a technical bent. It seems to me that we adopt the same analytical approach to learning a language as we do to when learning a technical skill or analyzing some physical phenomenon. Clearly this is the wrong when it comes to learning languages but it's a very difficult habit break. In a similar vein, I recall getting somewhat annoyed when trying to learn Morse code and making no progress. Whilst I was told never to analyze the individual dots and dashes of characters but nevertheless I still did and it was not until I broke this habit (with some difficulty mind you) and that I listened only to the rhythm of the sound of each character that I started to make progress.
Obviously, learning a language is a much more difficult exercise than learning Morse code but the same rules apply in both instances. I recall several decades ago attending a lecture by Noam Chomsky and he made a comment that I have never forgotten as it resonated with me then and ever since. That was that he—a world renown linguist—had difficulty in learning languages—and he cited words to the effect 'that a new language can waft down the corridors of MIT and that most of his colleagues catch it with ease as it passes by—except for him'.
Even at the time of the lecture, I could not help but think that he like me dissects and atomizes a language in ways his colleagues have learned not to do (but perhaps when you're a linguist that approach may be useful—after all, Chomsky is very famous and his colleagues are seemingly very much less so).