I have always been somewhat skeptical of the fact that this number coincides so nicely with the number of digits in a phone number, which the subjects (I presume to be from the US, I know, I have to track down and read the paper) are highly trained to commit to short term memory.

Does anyone know of similar finding were found in experiments with diverse cultures?

This is kind of a chicken-and-egg problem, isn't it? Which came first: the ability to remember 7 digits, or the decision to make phone numbers 7 digits long? It seems more likely that 7-digit phone numbers became common because that was the highest number of digits that most people could remember.

I believe that there is some truth to the claim that the length of a phone number affects our ability to remember numbers of that length; in Norway, we have 8 digits in our phone numbers.

A couple of years ago, I decided to try and remember as much of Pi as I could, and I found that precision to which I could remember it, was in fact 8 decimal digits. Even though I have only been asked to recite as much of Pi as I could remember once during the last year, and I seldom actively try to remember it, whenever I come to think of Pi, 3.14159265 usually pops into my mind.

However, whereas I remember telephone numbers either grouped on the form 12 34 56 78 (twelwe-thirtyfour-fiftysix-seventyeight) or 123 45 678 (hundredandtwentythree-fourtyfive-sixhundredandseventyeight), I remember Pi as a sequence of digits (three-comma-one-four-one-five-nine-two-six-five).

Edit: The two ways of grouping the digits in a phone number that I mention here, is the way that phone numbers are usually written here in Norway. Often, landline phone numbers are written on the form xx xx xx xx and cellphone numbers on the form xxx xx xxx. (Additionally, if I remember a phone number on one form, and someone else says the same number on the other form, I find it hard to tell whether we are speaking of the same number).

phone numbers also play into the concept of chunking

http://en.wikipedia.org/wiki/Chunking_(psychology)

Just as an anecdotal piece of info: most musicians accept that the human brain can keep track of eleven notes. Beyond that, you can change the scale progression and the transition will be (relatively) smooth to the human ear.

It's because eleven notes is the maximum that fit "nicely" into the harmonic series our eardrums pick up: http://en.wikipedia.org/wiki/Harmonic_series_(music)

Perhaps you could clarify or point me to some more information on this? Even as a (somewhat amateur) jazz musician, musical pivots have always eluded me.

I've been searching for an interview with Jimmy Page, where he mentioned this and it's the first I've heard of it, but to no avail. Hence, 'anecdotal'. If I find it, I'll post it here.

That is really interesting.