This is where Western and Japanese linguistics show their biggest difference. The former has a tendency to try making heavy generalizations (finding the universal properties of language/s) and prove it by crafting a few examples fitted for the task (peak armchair linguistics). In the latter, a huge amount of data is collected. I've seen at library a book with hundred maps, relatively precise, about the fine variation in Japanese dialect. It's however sometimes lacking explanation. So both approaches have their strengths and weaknesses and I think that more ground breaking word can be done if relying on works from the two traditions.