Say L is a human language (e.g. German, Chinese, etc.) and w is a string in L of length n>1. Is it known for different languages what the probability is that w is a word in L? And if S is an ordered set of strings, the probability that S is grammatically correct in L? I mean, I know or have a good idea how to answer this _if_ I had access to the right database. If being the key word here. Mayb this has to see with entropy?

Thanks.