Yes, the behaviour classifications of Socionics (and MBTI) are very subject to (mis)interpretation and controversy but so are many aspects of science. I agree that Socionics is largely based on anecdotal data but there are germs of consistency in this data suggesting perhaps a finite number of ways that information can be processed. The current models seem downright silly to me. It will take the right person to bring in a true scientific approach to prove or disprove the existence of predictable information control system structures (firmware?) because logic would suggest that they must exist. Wasn't Darwin's theories considered poppycock for many years before and after he published?

a.k.a. I/O