Nov 21, 2014

The "I Got Rhythm" Bridge - Some Historical Notes

George Gershwin's "I Got Rhythm," published in 1930, has provided the harmonic structure for perhaps hundreds of other jazz tunes. The bridge has a kind of perfect simplicity - a chain of dominant chords that begins on the III dominant, each resolving into the next along the circle of fourths, two bars each, until we land on the V of the original key, setting up the return to the last "A" section.

In the tune's usual key of Bb, the chords to the bridge are:

||  D7  |  D7  |  G7  |  G7  |
|   C7  |  C7  |  F7   |  F7  ||

However, these are not the chords that we see in the "original" sheet music, or in charts that are modeled on the "original" changes. Below is the bridge in what I think was its original form (I've had this sheet lying around for quite a while; I copied it from a Gershwin collection). Click to enlarge.

The chord symbols, taken by themselves, leave a false impression. As shown, they are not functional, but rather just an attempt to represent the notes in the piano arrangement. If a guitar plays the symbols while a piano plays the arrangement, it will sound OK. However, if you follow only the chord symbols, the harmony will sound fragmented and discontinuous - unless you are so extraordinarily clever that you can recreate the piano arrangement from the symbols!

The real gist of the harmony is in the left hand bass notes, which indicate:

|  D7  |  D7  |  G7  |  G7  |
|  C7  |  C7  |  Gb7 |  F7  ||

The many additional chord symbols attempt to represent the mostly-chromatic inner voices in the right hand, particularly the line that starts by ascending from F# in the first bar of the bridge, in the chart above (lyrics: "Old Man Trouble...").

Here's a priceless clip of George Gershwin playing "I Got Rhythm" in 1931. You can clearly hear the chromatic line in the bridge, although he alters it a bit in the third and fourth bars, continuing the line upwards. In this clip, he plays the tune first in Db, goes briefly to D, and finishes in Gb.

I found a few more printed "lead sheet" versions that seem to have been derived from the sheet music, showing similar chord symbols for the bridge:

Tune-Dex fakebook (c. 1949):

|  D7  C  |  Ddim  D7  |  G  D+5  |  Dm  G7  |
|  C7  Bb  |  Cdim  C9  |  Gb7  |  C7 with Gb  ||

Dick Hyman's Professional Chord Changes and Substitutions for 100 Tunes Every Musician Should Know :

|  D7  Am7/E  |  Fdim  F#m7b5   |  G   D7+5  |  Dm7/G   G7  |
|  C7  Gm7/D  |  Ebdim  Em7b5   |   Gb7b5   |    F7   ||

Just Gershwin Real Book :

|  D7  C/E  |  Fdim   D7/F#   |  G7   F/A  |  Bbdim7   G7/B   |
|  C7  Bb/D  |  Ebdim  C/E  |   F7   |    F7   ||

For comparison, here is the progression from the "original" sheet music, above:

|  D7  Am7  |  Fm6  D7  |   G   D+  |  Dm  G7  |
|  C7  Gm7  |  Ebm6   C9   |  C7b5  |   F7    C7 F7  ||

All of these seem to be trying to suggest the chromatic inner voice leading in the original piano arrangement, or some similar line.

Jazz players don't generally pay much attention to this line when playing tunes based on "Rhythm changes." Typically, they'll either just follow the basic chain of dominants, or use that as a basis for an elaborated progression (a couple of examples are shown here).

The pared-down approach seems to have started pretty early. Here's a Red Nichols version from 1930:

There's some interesting material in this Wikipedia article, including the fact that "I Got Rhythm" was actually written in 1928, as a slow song for the musical "Treasure Girl," but was re-used two years later in a faster setting, for "Girl Crazy."

The basic harmonic pattern of the "Rhythm" bridge was used by George Gershwin for the bridges in several more of his songs (I found these in the Just Gershwin Real Book); the first two definitely predate "I Got Rhythm":

Hang On to Me (1924)
Sweet and Lowdown (1925)
Feeling I'm Falling (1928)
Boy! What Love Has Done to Me (1930)

But this bridge idea wasn't necessarily original with Gershwin. The same harmony occurs in the bridge of "Five Foot Two, Eyes of Blue." That song was copyrighted in 1925, but according to Wikipedia, may have been composed in 1914.

Aside from the fact that the "chain of dominants" progression is about as old as European classical harmony, I have to wonder if there isn't an even earlier example in American popular song, operetta, or ragtime. Please let me know if you run across one. To qualify, it has to be an 8-bar bridge in a 32- or 34-bar form, starting on the III dominant, then moving around the circle of fourths, 2 bars per change.


  1. Especially appropriate for this time of year is "Santa Baby," the 1953 composition by Joan Javits, Phil Springer, and Tony Springer. Not older than "I Got Rhythm," but otherwise qualifies.

    1. Another such tune is "Across The Breakfast Table," by Irving Berlin, copyright 1929. This song comes from the show "Mammy," which starred Al Jolson. As Berlin and the Gershwins were all active in New York at this time, one must suspect that they knew each other's music. Which begs the question: did the Gershwins borrow, consciously or unconsciously, the idea for the bridge of "I Got Rhythm." I'd guess we'll never know.

      There is a 1930 performance of this song on Youtube. My copy is from The Songs of Irving Berlin: Ragtime and Early Songs, published by Hal Leonard.

    2. It's an interesting question - for a few of us, anyway. My current best candidate is "Five Foot Two," because of its great popularity at the time, and early date (pub. 1925, but perhaps composed 1914 - but I haven't seen a source for that except Wikipedia). I'm waiting for someone who really knows their Scott Joplin to weigh in on this.

  2. Thank you for a very interesting blog. If anyone here can put me onto the original printed sheet for I got rhythm I would be very grateful. It is hard to tell if what we see floating around the net is original or not? I am not so much investigating the harmonic aspects of the song as the rhythmic aspects of the original sheet

  3. An old discussion, but still interesting. I was just reading that some songwriters had control over their sheet music arrangements, and some didn't. Sometimes there's an arranger's credit on the sheet music, sometimes not. And what you get from the original sheet music may not be what was played in the original stage show if it was a show tune. In some cases, songs may have been simply melodies, with staff pianists or arrangers harmonizing them. And on top of that, song pluggers who sold the tunes to singers and show directors may have come up with their own piano parts. Lead sheets didn't start showing up until late in the Tin Pan Alley era, and we shouldn't think of songwriters as necessarily writing tunes as we think the final product should look like. I'm in the process of working out the changes of early 20th century hits now, and the piano parts are often more voice-leading than block-chord styled. In fact, some times there are no real chords, just voice leading, as the above example shows.

    1. Mark, thanks for the comment! You're right about the varying and uncertain reliability of sheet music as a guide to the composer's original intention. But at the distance of so many years, it's often all we have to go on. I'd guess that in most cases, the composer at least approved the piano arrangement, even if he/she didn't actually write it.

      I recently saw a clip of Ethyl Merman singing "I Got Rhythm" in the 1950s, in her inimitable expressive style. Ethyl was the one who sang it in the original 1930 production of "Girl Crazy." Assuming that her style didn't change much between 1930 and the 1950s, it's safe to say that her original Gershwin-approved performance was much looser than the sheet music would lead one to think.

      The block-chord approach, as in modern fakebook versions of standards, is in some ways kind of dumbed-down, but it's an approach that is necessary and appropriate for improvisers. Still, it's good to know as much as possible about the original intention and setting.

  4. I've been reading up on popular music harmony recently. The chain of secondary dominants seems to start with Liszt - at least the earliest example I've seen cited is from him. Ragtime used chains of dominant chords in threes and fours regularly, and they also show up in early Tin Pan Alley tunes from the 1890s, although not in a bridge like this example. The old-time barbershop singers consider this progression their own, calling the dominant chord 'major-minor' - the added b7 being the 'minor.' And I think you'll find them in Sousa's marches as well. The II7-V7 progression seems to be the origin of the jazz/pop ii7-V7 progression - I can't find any minor seventh chords in the pop music of the 1890s-1910s. In fact, the origin of the ii7-V7 was what started me looking.

    1. Mark, thanks for the food for thought. I just put up a post on some of your points:

      About the origin of the ii7-V7 progression - I haven't got concrete evidence to cite, but I think the idea of dom. preparation/dom/tonic goes back pretty far, and includes the IV and ii more than the IIdom. But that's a question in itself, and I didn't address it in the post. Thanks again!

  5. I've just found a transcription online of this recording by Gershwin himself:

    The last two measures of the bridge are played as triads: III-bIII-II-bII.

    Of course, this isn't necessarily the original singer's harmony - it may be a reharmonization for this piano performance.

    1. According to the description on the youtube post, this version is by Leon Bates (not Gershwin). Bates is an excellent modern Gershwin interpreter. His treatment of the bridge uses tritone substitutions. It's a normal jazz approach to the bridge, (though I'm pretty sure that the tritone sub device was known and used in Gershwin's time).

  6. Here's Gershwin's own arrangement in a collection: file:///home/chronos/u-e8a2ce9b75cf7c3160fcbbc9e44631d6607eb066/Downloads/IMSLP10257-Gershwin_-_Own_Transcriptions.pdf

    1. Thanks! I couldn't make that link work, but this one got me a download of the Gershwin arrangement: Very interesting to compare this printed arrangement to the Gershwin video clip above!

      Listening to the Bates clip, it sounds to me like he's playing this arrangement to the end, then going into his own additional material, in the style of Gershwin. That's where the II-bIII-II-bII bridge that you mention occurs.

    2. In the material that (I think) Bates added, there's a cute inside joke at the end of each A section, where he drops in a quote from Gershwin's "Nice Work If You Can Get It." That tune, in turn, had a quote in Ira Gershwin's lyrics referring back to "I Got Rhythm" - "Who could ask for anything more."