Nature is cheap, it doesn't allow anything to exist for long unless it's really useful. For that reason we have boolean logic in our minds: it makes the world easier to handle. It's far easier to use statistics to find out what's more likely and prepare for it, instead of preparing for all options regardless of like hood. However, as much as it helps to increase the efficiency of our actions, it distorts our perception of reality. For example, just because about 99% of the trees are 1m or higher is not a reason to state that a tree is a big thing; buy a Bonsai if you don't believe me.
The flaw I see in Socionics is that it promotes the black-and-white-ish world view in which everything is artificially set to belong to a certain arbitrary category, which, in turn, excludes the contrary one. Like with fuzzy sets, the belongship to a category doesn't really imply the absence from other; thus defeating the dichotomic model.
I think I can explain this better through a biological example. You might have heard about DNA code. It's composed by four basic amino acids: adenine, cytosine, guanine and thymine. Now, what makes an organism is not the basic components themselves, but the order in which they are arranged. Trying to state that an organism "is far more cytosine than adenine" is as absurd as Socionics stating that "someone has preference of Se over Ne". It's as simple as that.
The thought process is an extremely complex flow of a few basic operation modes. Just because we are starting to understand the way of operation of each basic mode, it doesn't mean that we understand the whole system.
Individuals are forced to belong to a certain type, but differences between members of a type can be as great as those between members of different types. It's just a matter of what we focus ourselves at.