There are a few general ways to play this (or just about any) tune. You can generally play in G major (E minor) for the whole tune- it doesn't get too far from there. You can look for ii-V7 progressions and get a little more detail in your playing, or you can approach each chord individually.


Personally, I use combinations of each of those ideas- they specific way in which you approach playing over a tune helps define your style.

JC

I like posts like these, because I'm also constantly curious as to what goes on in people's minds when they improvise (especially over jazz-type progresions). Myself, I tend to be a vertical player (treating each chord separately vs horizontal-finding a scale that works over alot of chords). I also actively look for and accentuate the ii-V-I's in a progression, because it's at the resolving V chord where the biggest tension and release happens. But alot of the time I'm treating each chord separately. I tend to think chord shapes and arpegios, I think arpegios rule for improvising. Playing off the arpegios gives you the sound of "playing the changes", which all the good players do and it is a skill that many "scale" players lack. Having said this, I'm finding that my favorite sections of alot of my favorite players solos, as well as my own, are the repeating-type phrases that don't seem to follow the changes, or incorporate common tones between the chords. Like an idea carried "through" the changes. What I feel I might "think" when playing at my best would be: carrying an idea or theme through the changes while keeping an eye on the chord tones as they pass. Trust me, I don't allways succeed.

I admit, the verticle approach can break down when the tempo is fast and the chords fly by at high speed. That is when horizontal type thinking can be handy. But no one says you have to outline "every" chord. And that's when "grouping" certain chord together can be a benefit. Alot of bebop tunes have many ii-V's that fly by at high speed. A way to deal with this is to basically think of the ii-V as one thing. Alot of people "minorize" the ii-V, and think of it only as the ii (play minor ideas over the ii and the V).

Anyway, I think I'm talking too much. I just like the subject you brought up, so I went off rambling for awhile.

"Autumn Leaves" is really an alternation between the major key and its relative minor:

ii-V-I-IV-vii in G major

 _______________________

/                       Am7-D7-Gmaj7-Cmaj7-F#m7b5-B7-Em-Em7

                   \___________/

                 ii-V-i in E minor

The Em7 becomes the vi in G major again, leading back to Am7.

In the bridge, the sequence is reversed. It starts with the ii-V-i in E minor (F#m7b5-B7-Em) and goes on to a ii-V-I in G major.
Finally (C section) it works its way back to the minor key tonic (Em).

In the melody, the E melodic minor scale is used on the B7 (bar 6) - which is a clue for what you can do as a soloist.

Otherwise, the G major scale will cover just about everything - as long as you're aware of how it shifts to E minor all the time.
The D minor pent - is that what you really meant? - won't work in the bridge (IMO) - it has an F natural, which is definitely out of key - except on the B7 (oddly, see below).

I'd suggest using E minor pent (or even B minor pent might sound cool), but also thinking about the chord arpeggios. Go for the interesting chord tones that aren't in the scale: the D# on the B7, the C on the F#m7b5 (the B7 can have a b9, C, too).

A more advanced jazz choice would to superimpose other minor pents or arpeggios: Am on the F#m7b5, and Cm (yep Cm!) or Dm on the B7.
D minor pent on the B7 chord gives you the #9 (D), b5 (F), #5 (G), b7 (A), and b9 (C) - a highly sophisticated jazz choice!:-)
(IOW, it helps to know why these strange choices work - because you are introducing certain valid chord extensions.)
D minor pent will NOT work anywhere else on this tune.

JonR

BTW, I should add that when using these "altered" scales (D minor pent over B7), it's very important you come out of them well. It's best to resolve by a semitone up or down to a chord tone on the Em (which could be a 9th, F#, or 6th, C# as well as E, G or B).

This means you need to plan your lines. Have a "target note" you're going to land on on the Em chord, and construct a little phrase over the B7 that will lead to it:

  B7 (alt)                Em

  (b7  #5   b7  #9  b9.... 5)

e|--------------10---8--|--7----------

B|-10---8---10----------|-------------

G|----------------------|-------------

D|----------------------|-------------

A|----------------------|-------------

E|----------------------|--------------

  (D minor pent)

Hi Jon,

This idea of chord tone relationships has always interested me. I always thought of modes as chords in motion with passing tones since they strongly help to frame the sound of a particular chord.

Anyway, I heard on another board that Wes once told Benson that everything is a V or a I. I did not know what this meant so I looked into the matter and came up with a possible explanation. In a I chord you could get at least five tones in say C major9, CEGBD. For a V chord, G9 you would get the following tones GBDFA.

Put the two together you have CDEFGAB or all the notes in the scale. The 5 chord creates the following extensions for the I chord beginning on the 3rd: 7, 9, 11, 13. So it stands to reason that pivoting between a I and V tonality would pretty much give you lots to work with.

I have read that Charlie Parker was most interested in the upper extensions in soloing, those being the color tones. So he probably played the V over lots of changes to get the upper extensions.

Looking at ways to derive altered tones is what you spoke about here. Using a framework to come up with these is quite useful. One would think you could come up with some handy concepts here to generate altered upper and lower extensions based upon a more familiar tonality or superimposition of sound that creates tension but also has continuity.

BTW, Pentatonic scales might as well be considered arpeggios in my way if thinking. You are only adding one extra tone to an arppegio and therefore can think of arpeggios and their related chord tones as a basis for some interesting concepts.

I'm ramblin a bit but find the topic to be thought provoking, especially as it relates to looking at chord tones for analysisand how to use substitution principals to create dissonance or color.

All the best,

Jim

Good stuff, Jim. The V chord (G7) creating the extensions you suggest for the I chord (CMaj7) is a nice example of diatonic chord superimposition. If you were to superimpose a non-diatonic chord, such as D7 over the CMaj7, you'd get the additional upper extension color tones (9,11,13, etc) with the D F# A C# arpeggio. Not sure if this has anything to do with what Wes was talking about, but non-diatonic seventh chord superimposition also seems to create the color tones and tensions that Charlie Parker, Wes and others used alot.

Tom

The D7 arpeggio should be D F# A & C, rather than C#.

The idea that everything is either V or I is the basis of Schenkerian analysis. (Interesting the uneducated Wes came to the same conclusion as some advanced academic!)

What Charlie Parker was getting at (AFAIK) was something a little different: the upper extensions of either maj7s or dom7s.
That would include 9, #11 and 13 on a maj7 (ie lydian mode), and probably the same on a dom7 (= lydian dominant).
But no doubt he was thinking about alterations as well on dom7s: b9, #9, b13.

As Tom says, you can use superimposed arpeggios to get these upper extensions.
E.g., a Bm7 chord on a Cmaj7 (for lydian);
Daug or Gm(maj7) on C7 (for lydian dominant);
Dbm6 or Ebm7 on C7 (for C altered).

JonR

Great post Jon,

I liked the examples you gave and examples of other altered tone possibilities.

The idea that everything is either V or I is the basis of Schenkerian analysis. (Interesting the uneducated Wes came to the same conclusion as some advanced academic!)

I could not help commenting about this statement... Schenkerian analysis is all about the theory of how we hear music, if I'm correct. I would bet that Wes heard music in a skillset that was very advanced considering his remarkable talent. So it does not surprise me that he would make these observations without the perscribed theory of the day to refer to.

But I have too heard that he was not too well versed in theory. However, uneducated is kind of a harsh word since education can occur on so many levels along with experience. I'm not intending to argue the point but rather shed a different light on what I beleive to to a genius of the guitar. Some might question the education history of Einstien as well, but not the intellect. In this regard, I beleive that Wes was a brilliant musician. He obviously had some valuable information he passed along to not only Benson but also to legions of guitar players who followed him.

Pretty cool stuff Jon. I did a bit of research into Schenkerian analysis and found it to be a very good system for categorizing melodic concepts.

Thanks for bringing this info to my attention.

Jim

I didn't intend "uneducated" to be a pejorative term! I meant "educated" in the adademic "been to college" sense.
He was of course highly "learned", having educated himself to a high degree by ear (learning Charlie Christian solos note for note from records). All he lacked was (probably) most of the terminology the rest of us use. And no doubt he picked up a lot of that later on just from talking to more "educated" musicians.

I'm sure you're right he was a highly intelligent man. If there's one thing that's struck me about ALL the pro jazz musicians I've met, it's their high intelligence - not just good ears (which comes from practice), but quick thinking and (usually) a dry wit too.

Interesting to note that Django Reinhardt was also "uneducated" (as a gypsy, I think he was actually barely literate). And Django and Wes would probably fight it out to claim the title of "greatest guitarist of the 20th century" (you wouldn't find many pro guitarists who wouldn't agree).

Dude - what about Charlie Christian?