I've never found a definitive answer, but it seems pretty obvious that there was a practical reason for doing so.
It's not likely to be that much of an advantage for single-note playing, since patterns would be easier to learn, and movable across the fingerboard. It's not like the scales can't be played on that system - ask a bass player (a bass player who knows scales, anyway

).
It's possible that it was an engineering problem meant to reduce the tension on the top two strings. But I doubt that's why - the lute had a bazillion courses, and that held up in a smaller form.
That pretty much leaves the answer as being that it's easier to finger (some) chords.
It's not difficult to figure out the "open" fingering for a guitar using pure fourths (there's probably some errors, but whatever):
A: 0 0 2 2 1 4
B: 2 2 1 4 3 1
C: 0 3 2 0 0 2
D: 2 0 0 2 2 1
E: 0 2 2 1 2 3
F: 1 0 3 2 0 0
G: 3 2 0 0 2 2
Here are the minor chords:
A: 0 0 2 2 0 4
B: 2 2 1 4 2 1
C: 0 3 1 0 0 2
D: 1 0 0 2 2 0
E: 0 2 2 0 2 2
F: 1 3 3 2 0 0
G: 3 2 0 0 2 2
One thing that I notice is that there are a lot of four-finger chords.
Edit: This is an issue because if you need four fingers to make a chord, you won't have any fingers left to use a barred version of the chord. That means that you can't "share" common forms, and have
lots of shapes to memorize.
In "standard" tuning, with the top two strings a B and E, you've biased the guitar to playing in the key of E. The main chords using this system are E, A and D. They can be seen as variants of each other - the E is the "main" chord, with the scale degrees (from low to high) are Root, Fifth, Root, Third. It's easy to create variants of the chord by shifting the root on the fourth string to get a maj7 or dom7, and the third on the third string to get a minor third or a suspended fourth. All these variants (E, E7, Esus4, Emaj7, Em, Em7) fit nicely in the hand, since they only require three adjacent fingers to span no more than three frets.
E : 0 2 2 1 0 0
One particularly nice feature is that there's only one instance of the 3rd scale degree, so you only have to change one string to convert the chord to a minor (or suspended).
The A is basically the E form, shifted one string up (a fourth):
You can think of the high E "wrapping around" to the low E. Note that the only thing that changed was the note on the second string (because it's a third, not a fourth):
A : 0 0 2 2
2 0
It still has all the easy-to-finger convenience of the E chord (and variants). Shift it up the strings once again and you've got a D chord:
D : 2 0 0 2
3 2
Again, you've got all the convenient variants of the E chord in easy reach.
Shift again and you get an open G chord, which isn't quite as nice, because you've doubled the third:
G : 3 2 0 0
3 3
With the third doubled (on the second and fifth string), it's much less convenient to convert this "natural" form into it's minor variant. Shift again and you've got a C, with the same issues when it comes to the third scale degree:
C : 3 3 2 0
1 3
So the G and C are much less "friendly" than the E, A and D and their variants.
Shift up another string and you're on a F major, which is a barred E, and the circle is complete.
So... Having a B on the second string gets a finger-friendly E chord with lots of easy-to-finger variants. It also gives the A and D forms, which are almost as nice. Because they are three-finger chords with an easy reach, they lend themselves to barring if you need to create a different chord.
All in all, it's a pretty good compromise.