You’re right, @gman, I don’t know if there's hard data about what’s easier to learn, gestures or keyboard shortcuts.
I was basing my assumption on gestures being more aligned with some of the metaphorical structures we use in natural language (e.g. “stock market is up”, “up” is metaphorically structured to mean “more” or “positive”; that's hopelessly oversimplified, see Lakoff's work for details).
Gestures can have direction and physicality (also talked about in the video I linked to above). This I believe to encourage intuitive understanding, because it is tapping into some very deeply embedded metaphors that we use all the time in language and which are based on how we interact with the world (think of how much in language is structured with e.g. spatial direction, “stock market up”, “feeling down”, “going back in time”, “moving forward with a decision”, etc.).
However, how well that works, might also be dependent on how we acquire that skill, much like language itself. I have spend almost seven years officially (and many more before and after that) in that world of Apple's user interface design and experience principles, and for me a lot of these gestures just make sense on a very intuitive level with not much cognitive effort involved. I find it much harder to memorize the right chords of keys to press for more advanced keyboard shortcuts. The ones I do use all the time, are “muscle memory”, as in I wouldn’t be able to tell you the keys without watching my fingers — and isn’t that also a gesture?
So from that I was following that gestures are on a somewhat lower level, where there is less cognitive effort involved. Of course, assuming that the metaphors used for gestures are aligned with metaphors we use in natural language, for which pinch-to-zoom seems to be an excellent example for instance.