The case
From James Boyle, The Public Domain, ch. 6 (LIT-1301, read in NOTE-1021). Page numbers are the Yale edition’s.
- The received origin. Most sources, and the biography on Charles’s web site, say Ray Charles made “I Got a Woman” (1955) by rewording the 1904 hymn “Jesus Is All the World to Me” (p. 127). Boyle hears “little resemblance” (p. 133).
- The actual source. Claims pointed to a song by the Bailey Gospel Singers, “I’ve Got a Savior”, that no catalogue or library could find. Boyle’s research assistant found a 78 rpm copy in the Rodgers and Hammerstein Archives of the New York Public Library. “It sounds like the same song”: the introduction and backing differ, but “the central melody is almost exactly the same”, down to the intonation (pp. 133–134). The record names its composer only as “Ward”, perhaps Clara Ward. Charles turned another Ward song, “This Little Light of Mine”, into “This Little Girl of Mine” in the same way (p. 134). “I’ve Got a Savior” itself takes lyrical patterns from a 1940 spiritual, “Ain’t That Good News” (p. 133).
- Charles to West. For “Gold Digger” (2005) Kanye West had Jamie Foxx, who had played Charles in Ray, imitate Charles’s style and the melody of “I Got a Woman”. An eight-bar loop of Foxx singing it runs under the whole song. The fragment was cleared, with payment to the Charles estate (pp. 138–140). The words invert the original: “She gimme money / when I’m in need” becomes “She takes my money, when I’m in need”.
- West to The Legendary K.O. In September 2005 two Houston volunteers recorded “George Bush Doesn’t Care About Black People” within days of Katrina (pp. 125–126). Its background is “a looped, or infinitely repeated, version of the hook that Kanye West and Jamie Foxx had in turn taken from Ray Charles” (p. 144).
- The phrase pattern. Boyle sets it out: “I’ve got a savior, Oh what a savior” becomes “I got a woman, way over town” becomes “There’s a Gold Digger, way over town” (p. 142). The Legendary K.O. “found themselves sampling Jamie Foxx, who was copying Ray Charles, who was copying the Bailey Gospel Singers, who themselves may have borrowed their theme from an older spiritual.”
Documented: Boyle names his sources and the archive holding the record. That the Bailey record precedes Charles’s song is his dating, “a mere three or four years earlier”. Its copyright status is unknown. The Copyright Office has no registration for it (pp. 136–137).
What it can show
CLAIM-002. The chain is a worked instance of a retelling recognizably descended from an original it no longer resembles in function. The tune and the phrase pattern survive four steps, while what the song praises or condemns changes at each one. The links are not one relation repeated:
- Bailey to Charles is a rewording, with the melody and lyrical pattern kept.
- Charles to West is a licensed re-performance by an imitator, with no part of Charles’s recording used.
- West to The Legendary K.O. is a sample of the recording itself.
So identity of the artifact holds only on the last link, and continuity of transmission holds on all three. Similarity of function holds on none.
The relations come apart in the record of the case, too. The transmission story that sources repeated, from a 1904 hymn, was wrong, and it was the melody that located the real source. That bears on CLAIM-002’s defeater, which asks whether descent judgements are fully predicted by present similarity. Here similarity corrected a false story about transmission. That does not defeat the claim: it shows similarity used as evidence of transmission, not standing in for it.
CLAIM-210. The Legendary K.O. step is a rendering in the strict sense, a transport from a given source (the recording). The Foxx step shares the melody’s kind with Charles’s song and was produced from it, so it is a rendering too, though of the composition, not the recording. The law tracks the same line. West’s use of the composition was cleared (p. 140). The Legendary K.O. took the sound recording itself, which is the use Bridgeport governs. Boyle sums up its rule as “Get a license or do not sample” (p. 151).
What it cannot show. It is one chain, picked out because it is striking. Boyle’s reading of the musical similarity is his own, with no musicological analysis. Nothing here measures how much of the “message” a listener hears in the borrowed melody. The record’s other cases of change under serial transmission, such as CASE-068 (risk messages in diffusion chains), are experiments, and this is not.