DrewPeterson7
Sage of the Order
The real problem is thinking of theory as a series of rules - its not. Its a language, primarily useful in telling other people WTF you are hearing in your head or what you are playing. My two least favorite musician behaviors are saying either, "I don't want to learn theory because I don't want to be limited by the rules," or "You have to know the rules before you break them." Both groups miss the point completely.
I've always thought of musical theory as sort of a "toolbox" for a player. I'll try to explain what I mean by that, and I think it'll probably put me dangerously close to your second group, but I don't think that's quite where I fall so bear with me.
I would maybe accept the argument that straight up diatonic harmony ("this is a major scale, and these are the chords you get if you harmonize it into triads) is a "rule," but that seems a little strict for me. Rather, it's just sort of a definition of what works - that if you group the notes of a scale together into chords, these are the chords you get, and they'll sound good together. It's sort of like a musical safe zone.
Of course, things that are safe are never any fun. So, then you get into what I really consider the meat and potatoes part of music theory - an understanding of WHY those notes sound good together when grouped like that. Tension and resolution, basically. As you begin to understand that concept, then you also begin to see how you can use that concept to step outside of purely diatonic chords - say, that in a blues in A minor, quick change chords, after that Am-Dm-Am opening couple bars, you can add some pretty cool tension before you go up to the Dm (the fourth) by briefly raising that Am to an A7, because there's a whole shitload of half-step resolutions between A7 and D minor. There's a word for this in music theory, it's called a secondary dominant, but you can infer it pretty easily from just a pretty basic understanding of how resolution works. Similarly, you can use these same concepts to modulate in and out of keys, or write progressions that don't fall purely in any "key" to speak of - one of my favorite sets of chord changes is Stone Temple Pilots' "Interstate Love Song," where you've got this incredibly tense chromatically descending half step bassline that goes down until basically the bottom falls out of it and it collapses upon itself down into the tonic. It's brilliant, but a good understanding of theory helps you understand WHY it works, and furthermore how you can apply the same concepts yourself.
Anyway, I'm kind of rambling, but the gist is you can write some pretty badass progressions entirely by ear, but an understanding of theory really helps you understand why they work (which in turn can make it easier to finish a progression if you get a couple chords that sound good together, but can't finish it and make it come back together into something that sounds good repeated), as well as gives you an idea what you can play if you have to take a solo over it - what scales will probably sound consonant, what ones will probably sound dissonant, and what situations a dissonant sounding scale may sound more appropriate or interesting than a consonant sounding one.
I remember one of the first times I ever had this conversation on the net, some kid was talking about how he didn't want to learn theory because he didn't want to "limit" himself, and then when pushed gave as an example a chord progression from one of his band's songs. I don't remember the exact changes, but it was basically A minor chords - Am, C, Dm, and Em, except at one point it resolved Em-Am, and at another it was E-Am. He claimed that this sort of radical progression, one that used both an Em AND an E, was the kind of thing no one who studied theory would ever write, because it was "against the rules." I started cracking up at the softball I'd just been lobbed, for reasons that I'm sure will be readily apparent to anyone who recalls the history of the harmonic minor scale.