Theory behind soloing over chord changes?

Anyone?

Honestly, your question is really vague. A scale works or doesn't work over a chord because of the melody you're playing - there's no magic in the scale itself, and no "use this scale over this progression" will work for all choices of melody.

Harmony is composed of consonance, when the melody note agrees with the chord, and dissonance, when they clash. Use all consonances and you won't sound bad - but you will sound bland. Use too many dissonances and at best you'll sound like you're trying to find the key. So dissonance is like putting salt on your food - a bit improves the taste, but too much makes it inedible.

When you're improvising over a progression (or a single chord), chord tones will be consonant. Make a solo of arpeggios and everything blends - but it feels like it needs some salt. So we open up the spice chest by using a scale, which will have the arpeggio tones plus varying levels of dissonance. This is true of ALL scales. The easiest ones to use, like the pentatonic, have the fewest dissonances. The more dissonances in the scale you use, the more careful you'll have to be with it.

The simple truth is that ANY note can work over ANY chord. How it sounds will depend on context - if it's dissonant, is the dissonance put in a place in your melody where it sounds like you've got an idea? Are you then resolving it in a way that makes a logical musical statement?

Since any note can work over any chord, scales serve two purposes: first, they restrict the dissonances to a limited number. The lower the number, the easier it will be to make something that sounds "good" (or at least, doesn't sound "bad"). The second purpose is to aid in training your ears. As you get to know how a note will sound BEFORE you play it, your solos can become deliberate, and you'll rely less on the scale, and more on the total shape of your melody against the harmony.

Taking your example of A mixolydian over an Am chord, the scale contains A, B, C#, D, E, F#, and G. The chord contains A, C, and E. That gives you these results:

A and E are chord tones - no appreciable dissonance.
B is a second against A, a minor second against C, and a perfect consonance with E. This will give you a slight tension, but won't be objectionable.
C# creates a b9 interval against the C in the chord - that's a harsh dissonace
D is a perfect fourth against A, a ninth against C or E - like B, it'll work, but you won't want to end your solo on it.
F# is a major sixth against A, and a major second against E - but it's a tritone against the C, which is dissonant
G is a major second below A, a perfect fifth above C, and a major third over E - like B and D, it's a tone of interest.

So you've got two notes to watch out for, three that will give some interest, and two that blend perfectly. It'll be slightly harder to use than other choices, like A natural minor, which has three full consonances, two tones of interest, and two dissonances.

The thing that really matters isn't the number of tones that agree, though, it's how you use them. There are a variety of ways to use non-harmonic tones (ones that aren't in the chord). Without going into a full harmony text in this reply, the three most common are:

Passing tone if you play G-F#-E, the F# will be dissonant, but you're moving through from an Am7 harmony (the Am chord with your G note) to a chord tone, with the dissonance in between. As long as the dissonance isn't accented or otherwise emphasized (like holding it longer than the other tones) it'll sound fine.

Neighbor tone if you play E-F#-E, you're moving from chord tone to an adjacent dissonance, and right back to the chord tone, and it'll sound fine.

Appogiatura - the Italian term for "to lean". Hit a dissonance, and immediately resolve it to a chord tone. For example, play the C# against the chord, and immediately drop to C. The dissonance is resolved - you'll hear this in many styles of music. For an additional hint, the appogiatura generally sounds best if you resolve it in the opposite direction from your approach - if you went UP from A to C#, go DOWN to C.

In that example, the C isn't a scale tone - but it is a chord tone. Chord tones will always work over a chord.

Get the hang of how the intervals sound against chords in context and you'll start worrying about the melodies you make, rather than the tool you use to make them - and that's the real key to making a good solo: you know what it sounds like before you play it.

@NoteBoat: Wow, that was very elaborate and thorough! Really great explanation, and not at all hard to understand. Thank you so much. :)

Do you give lessons by any chance? Or perhaps if you could redirect to a good source. Stuff I've come across so far hasn't been this informative and easy to grasp.

I do - I've been teaching guitar since 1978. I also teach piano and music theory, and I own a music school in Plainfield IL (Midwest Music Academy) with teachers for all instruments and voice.

http://mwmusicacademy.com/

I do - I've been teaching guitar since 1978. I also teach piano and music theory, and I own a music school in Plainfield IL (Midwest Music Academy) with teachers for all instruments and voice.

http://mwmusicacademy.com/
Impressive. How about Skype lessons? Can it be possible? I'm not in the US, that's why I'm asking.

I'm afraid we don't do Skype lessons. In my opinion Skype is limiting for effective instruction for a couple of reasons: first, there's the fact that it's 2D. During a lesson, I might need to see what's going on behind the neck (for example, the student's thumb placement) to suggest a solution to a problem. Second is latency - it's really hard to assess timing, which is critical to music.

As technology improves, I might revisit that decision. I do a regular review of technology that we might use, and we implement whatever is practical and effective.

Noteboat, I've been reading your posts for years, and I am sure you are a very good teacher. Your explanations are always spot on and in an easy to understand manner. It's no wonder that you have a successful studio.

I might add a little non-theory to this.

Soloing is also like playing with a cat with a toy on a string. Let the cat catch the toy too often, and the cat gets bored. Don't let the cat catch the toy and it will lose interest.

So you want your audience to predict where you are going next enough so they don't lose interest (catch the toy) but not predict it all the time (get bored).

It's what all good melodies are about.

Notes

I'm afraid we don't do Skype lessons. In my opinion Skype is limiting for effective instruction for a couple of reasons: first, there's the fact that it's 2D. During a lesson, I might need to see what's going on behind the neck (for example, the student's thumb placement) to suggest a solution to a problem. Second is latency - it's really hard to assess timing, which is critical to music.

As technology improves, I might revisit that decision. I do a regular review of technology that we might use, and we implement whatever is practical and effective.
That's too bad... Technique-wise, I'd say I'm at an advanced level, so I'm not much worried about it. It's the theory that I'm focusing on now. I want my songwriting and soloing abilities to improve, and I've been trying to achieve this by myself, but it's proving to be difficult. I haven't been able to find a good local teacher, since in my town and around I'm one of the best guitarists (not boasting, just the truth), and there might be someone who could help me but he lives really far which makes it very hard for me, and makes Skype lessons cheaper (because of travel expenses) and more convenient.

If you're willing to give me at least one or maybe a few Skype lessons regardless of the drawbacks, if nothing but to have some of my questions answered, I'd be more than glad. I mean after seeing your explanations, I think it would be totally worth it, since I have many questions.

Noteboat, I've been reading your posts for years, and I am sure you are a very good teacher. Your explanations are always spot on and in an easy to understand manner. It's no wonder that you have a successful studio.

I might add a little non-theory to this.

Soloing is also like playing with a cat with a toy on a string. Let the cat catch the toy too often, and the cat gets bored. Don't let the cat catch the toy and it will lose interest.

So you want your audience to predict where you are going next enough so they don't lose interest (catch the toy) but not predict it all the time (get bored).

It's what all good melodies are about.

Notes
Thank you for your response. Great advice. :)

In fact that's exactly what I'm trying to achieve. I'd even like to learn jazz at some point, but this is too advanced for me at the moment. I guess with jazz you'll lose most of your listeners since most of them won't understand anything. :lol:

I'll try to elaborate my original question further. Some guitarists play a different scale over each chord. Not just focus on the chord tones and stick to the main scale, I mean change the entire tonality over that chord. Say if they're playing natural minor scale over an Am chord for example, if after that there's a G chord let's say, what I mean is that some don't just focus on the G chord tones and stay in the same scale, but change everything. Let's say they might play G Phrygian Dominant for example, or just G Ionian, or even G blues (I've heard Tim Pierce doing this a few times). Then when it's Am again they may play A Dorian over it this time instead of natural minor. So instead of sticking to one scale, the melody or the solo keeps changing according to the chord. I guess this is very common in jazz, in addition to the other scales used in jazz. So my original question was how does this work? I've tried it myself, and it sounds good and more interesting, though sometimes a bit strange (but still nice), but I guess it's just because my ear is still not used to this sort of thing.

It's true that many guitarists think of different scales over different chords. Some even make it work - but more end up doing what one of my teachers called "mental masturbation" - the thought process doesn't result in anything musically rewarding.

For example, it's common to see things on the internet like "over C major, play C Ionian, and when the chord changes to F, switch to F Lydian". That's musically useless. Both chords have exactly the same tonality - C major - so while you might think you're using a different scale, what's coming out is all C major.

It's also true that jazz artists will think in terms of "chord scales". But they're driven by different reasons. The chord/scale thinking sprang from bebop, which is characterized by a couple of things: fast tempos (many bop tunes are counted in at 280-300bpm), and frequent chord and key changes. A tune like Coletrane's "Giant Steps" at 300bpm will change keys once every 1.2 seconds, which isn't a lot of time to think about what you're going to do. Individual chords fly by in less than half a second. As a result, many bebop solos will simply rip through segments of arpeggios and scales, because unless you've spent considerable time working with the changes that's all a player can do.

Another use of changing scales in jazz solos is seen in modal jazz - tunes that have a simple two or three chord vamp with no dominant chords, like Miles Davis' "So What". The A section is an Em7-Dm7 vamp; the B section moves up a half step to Fm7-Ebm7. Solos are usually done using a Dorian scale, so after 16 bars you switch from D Dorian to Eb Dorian.

This practice can be extended to individual chords: you could use A Dorian over an Am or Am7 chord and switch to D Dorian over a Dm or Dm7 chord. Notice that this is quite different from the internet example of C Ionian over C and F Lydian over F, because the scales have DIFFERENT tonalities: A Dorian is A-B-C-D-E-F#-G and D Dorian is D-E-F-G-A-B-C-D. Using F sharp over one chord and F natural over the other will sound different - the key here is to use the same type of scale over each chord root.

The strategy you describe for Tim Pierce is another, called "parallel scales" (Steve Vai has called this "chord axis theory", but I avoid this term - there's a composition technique developed by Bela Bartok that goes by the same name). Instead of keeping the same scale type, you keep the same root: over a G chord, use a G root scale that agrees with the chord tones - you wouldn't use a D half diminished scale over D major, because the third and fifth don't match, but you could use a D mixolydian or D Lydian, because they contain the chord tones. Over another G chord (or over another section in G) you use a different G-root scale.

My personal view is that focusing on a scale doesn't say "I can use these notes" so much as it does "I'm going to avoid these notes" (the ones not in the scale). It's what you use at each moment, and how they relate to each other that makes the solo good or not. Think of the scale you choose as an artist's palette: you've mixed the colors you can use to make a painting.

Imagine your palette has all greens and blues. There's no red - so it's a color you'll be avoiding. If you think of the palette as "I can use any of these colors" you paint like Jackson Pollock: anything can go anywhere. Soloing like this is what I call "poke and pray" - you hope the result is pleasing.

If you want to be a Rembrandt, you're thinking "I'm working on the sky now, so I'll only use the blues and ignore the green.... now I'm doing the grassy meadow, so it's time to put down the blues". Same palette, different approach.