It's true that you probably won't learn social skills solely by reading a text in an empty room and then leave it, having fully assimilated the described social skills.

Beyond that -- and not to single you out, of course -- your comments present what I have found to be a widespread, counterproductive misconception. What I have found, instead, is that:

1) Believing that learning a skill you possess requires extensive experience, is the quickest way to "compartmentalize" and weaken your own understanding of the skill, and dull your ability to pass it on to others. If you start from the premise that it's all an inarticulable black block, you will completely miss out on the parts that can be verbally communicated. I have seen this all the time in instructors who lament that they can't just tell me how to do something, and at the end I find that "er, you could have just told me all along that ...".

2) Spending time around others who exhibit a high level of X will do very little for your skill at X. Everything that was a mystery will remain so, because you won't see the "model" that led the expert to act one way rather than another, and people woefully overestimate their ability to "infer" the "algorithm" that the expert is using. (This is very much related to the underdetermination problem.) Time and again, I've apprenticed with others and learned precisely nothing, while I've trained others up to my level in a fraction of the time I required to learn it, or of the time expected for the person to learn the skill.

3) Much experiential learning -- not all, of course -- can be obviated by a relatively small amount of verbal instruction, because it singles out the non-obvious, hard-to-experientially-infer part of the problemspace. Again, with the instructors I've been involved with (and contrasted with my role as an instructor), I could make no progress learning alongside them until I could verbalize the skill -- which typically reveals holes in the instructor's own understanding!

This has all led me to strongly suspect that people who fall back on an inarticulability defense, typically lack understanding in crucial ways themselves.

Recently, I wrote an introduction to asymmetric cryptography that was lauded as far more helpful than anything else available on the matter, even despite covering less material. My trick? Actually have explicit understanding of the matter, not just the "learn by watching" kind.