Translation. Immanuel Kant claimed that we understand new things in relation to what we already know (this is a rather extreme reduction of trascendental idealism). In more recent times, a similar argument was used as part of the underlying structure of Blending Theory. This process of integrating the new into the old is essentially a method of translation. Or so it would seem that translation is the basic mechanism for understanding the new when one looks at the various "translations" provided for learning new things.

An example of this process of learning via translation was recently shown to me by a coworker. The translation at hand here is the article Python for Lisp Programmers by Peter Norvig. In fact, according to Norvig, this really isn't translation as much as understanding that Lisp and Python are (arguably) essentially the same. To illustrate his point, he provides examples of data types, etc. between the two languages. One might play Devil's Advocate with his article by claiming that two spoken languages are the same because they have the same number of adjectives to describe cat hair. Yet, I think that he has a good point, if for no other reason than newer languages tend to borrow ideas (either deliberately or not) from older ones, and Lisp has been influential for about a half-century.

Another "translation" guide that I recently encountered was a Unix to Plan 9 command translation cheat-sheet. Granted, many of the same people worked on both operating systems, but, aside from that, it seems that here we see translation at play not only in learning but design. But how much difference is there between learning and building. According to Marvin Minsky, not much.

Perhaps, if we acknowleged that we learn best by actively relating new information to old, then we could accelerate our
learning. We could move post the ten year learning curve. We could find relevance for our knowledge, tie it together. Or, perhaps we could simply avoid feeling like we are "starting over" when we embark upon something new.