Bass player play 6th over power chords?

Sure it's possible - they did it :)

Seriously, theory always follows practice - not the other way around. Whenever theory comes up with a 'rule', it's based on what's been done by composers before that time. And when composers do something outside of accepted theory, the rules get rewritten.

But that's not what's happening here. There are three basic principles at work. One is that 'chord' doesn't describe a fingering - it describes the ENTIRE harmony. The arrangement then distributes the notes of that harmony among the various instruments. Giving it a quick listen, the full harmony becomes apparent when the lead guitar comes in... those notes are held a little longer, and contribute to the overall sound. So the chords are:

E-G-B-C# = Em6
E-(G)-B-F# = Em add9
G-(B)-D-E = G6
G-(B)-D-F# = Gmaj7
C-(E)-G-A = C6
(D)-F#-A-C-(E)-G = D11
G-B-D-A = G6

So when the guitar plays a 'power chord', that isn't the full picture. Guitar players often confuse 'chord' (harmony) with 'fingering' (their small part of it) when it isn't a solo guitar piece.

Second, you've got memory. The chord tones I've got in parenthesis aren't being played - but in each case they were JUST played. There's no way we'll hear the G-D-E as anything other than G6, because our ear is filling in the blank with B natural, rather than Bb. Short term memory plays a big part in how we hear music. That's why I know the second to last chord is a D11 - the dominant chord of G, rather than some sort of F#m7b5: we've already heard enough to center our ears on the key of G, and we expect a dominant chord type as the final dissonance.

And third, you've got the arrangement. Looking at just the tab, you might say "yuck, those notes will sound awful - they're too close together!". And you'd be right if they were in the same register. But arranging recognizes a difference between a closed voicing and an open voicing. This is easier to explain in terms of a piano rather than a fretted instrument. You can have that Em6 played as a closed voicing (which basically means putting all the notes in the same octave) like this - I'm approximating the keyboard spacing:
B C# E G

Or you can spread them out in an 'open' voicing, as they do. Here, the lead guitar is more than an octave above the highest rhythm guitar tone - it doesn't eve fit on the same line in the code :) :

C# E B E G

Although dissonance in a chord is part of its quality, open voicings don't have the tension sound as harsh as closed ones. And that definitely plays a role here.

Make sense?

I was hoping to get any answer for my question, but this is a great gift Tom! I just read your analysis several times. It does makes sense and you just changed (for the better) my way of thinking about songs in general.

Man, I feel like I just had a breakthrough. :D This has made my day!

Happy to help, Dennis. Give me a call and we'll jam :)

Everything Noteboat says is absolutely true. However, I'll add an even simpler way of looking at it.

I'll use the first chord (G power chord over E) as an example -- If the guitar is playing a G power chord (GDG) and the bass plays an E you have an E minor chord minus its fifth. The notes are EGD in total (EGBD would be the complete minor seventh). Since the E is in the bass it gives the chord it's root and the guitar provides the all-important 3rd and 7th tones. So the progression is a string of minor seventh chords -- C#m7, Em7, Am7, Bm7 -- with the notes divided between guitar and bass. Kind of a cool sound.