I was talking to another physics graduate student the other day which turned out to be fairly interesting. He told me that he had joined the Stanford summer chorus and was wondering what the best way to learn his part was, since he knew I was in the chamber chorale. I thought about it for a bit, and noted that, while learning the notes of your part in isolation can be useful, it is not sufficient for learning your part, since the way that an individual part sounds is highly dependent on the other parts. That is, although the notes are the same in both instances, they sound completely different if they are sung by themselves, and if they are sung within the context of the piece. This is part of what makes music so amazing - listening to how the composer has woven each of the parts together to create a fabric that is more than the sum of the individual parts. In my case, I realized that it's often easier to learn the part with the rest of the choir, since then many aspects of the individual part make more sense. For instance, Bach is the absolute master of 4 part harmony and counterpoint. In fact, one music theory prof I had at MIT claimed that music teachers used to give their students Bach chorales with notes missing, instructing the students to fill in the missing notes. If the student did not do what Bach had done, it was wrong - since Bach had set it up so that there was only one way out within the "laws" of counterpoint. When singing Bach, while the individual part sometimes looks odd, everything makes perfect sense once the other parts are added - Bach needed notes to fill in this chord in this way, and it's easy to hear what note you need to sing. Of course, sometimes I get it wrong and end up singing the wrong note (even though it sounds right to me) because Bach wanted a different chord than I thought, but let's not get into that. This led us off into a discussion about gestalt. Being the free-associating kind of person I am, this connected in my mind to how I apparently tend to learn things. At MIT, I consistently did very poorly in my physics courses for the first 4 weeks or so. However, as the term wore on, and the "big picture" of whatever subject was being covered became clear, I got a much better idea of what I was doing and finished the term strong. In other words, I was mostly incapable of learning a subject piece by piece - I had to get the big structure into my head which then made it obvious to me how the pieces fit together and how to use the individual techniques. Most classes work in the opposite direction - where the individual pieces are introduced one by one, and slowly they are put together to form the big picture. I hadn't thought about it this way at the time, but it's sort of the conflict between inductive and deductive reasoning. I guess I tend to be a deductive person - from a general principle (the big picture) I deduce the specific application (individual pieces/techniques). Whereas most classes tend to be taught inductively since that's a more intuitive form for most people - where we learn from the individual pieces how to put together the big picture.