After a lengthy period of sharply reduced activity due to heart problems (he underwent a three-way coronary bypass operation last year), George Balanchine at age 76 is back doing what he does best -- creating ballets for the New York City Ballet.

Among the recent products of his restored energies is one that has provoked more discussion than any of his works of the past half-decade. This ballet, with the jaw-crushing title of "Robert Schumann's 'Davidsbundlertanze'," premiered last June in New York and can now be seen for the first time in Washington during the company's program Wednesday evening at the Kennedy Center Opera House.

"Davidsbundlertanze," which means "Dances of the League of David," is a ballet for four couples (at the Kennedy Center they'll be Adam Luders and Karin von Aroldingen, Jacques d'Amboise and Suzanne Farrell, Peter Martins and Heather Watts, and Ib Andersen and Sara Leland) with decor by Rouben Ter-Arutunian, set to the collection of 18 "character pieces" for solo piano by Robert Schumann named in the title.

The League of David was an imaginary literary creation of Schumann, who was an assiduous writer as well as a composer. In a musical journal he edited and helped found, Schumann wrote prolifically about the artistic developments of his time. The League of David was his fancied coterie of "good guys," the defenders of high principle and progressive esthetics who warred unceasingly against "the Philistines," the benighted, reactionary enemies of true art.

Two members of the league, among others, embodied antipodal sides of Schumann's own emotional nature. Named Florestan and Eusebius, they represented the complementary dynamic and contemplative poles of Schumann's introspective self, and he signed many of his essays and reviews with their names or initials. Similarly, the enchanting piano vignettes of "Davidsbundlertanze" -- which was composed during the trying period of Schumann's courtship of his future wife, Clara Wieck (whose father was opposed to the match) -- were each inscribed with an "F" or an "E" in the first edition of the score.

Given this literary, autobiographical and philosophical underpinning from Schumann's side, one might expect a ballet, if not exactly narrative in form, at least strongly dramatic in outline, with clearly defined characters. the production supports such expectations in its period costumes and decor suggesting an imaginary ballroom, with a mist-shrouded cathedral against watery depths in the background. Nevertheless, Balanchine had no dramatic specifics in mind, and in characteristic fashion resists adamantly all attempts at verbal exegesis of his ballet. "If you keep reading things into it, you miss the dancing," he says.

As he made clear in talking about the new work this past week, the music as always, was both his springboard and his destination -- the music as an abstract interplay of emotional curents, not as any kind of map of events, real or imagined.

"I knew and loved this Schumann music for many, many years, way back in Russia," Balanchine said. "When I was studying music as a youngster, I passed an examination one time playing Schumann's 'Novellettes,' also passed an examination one time playing Schumann's 'Novellettes,' also wonderful litle pieces. I even used some Schumann pieces to make little concert dances for myself in Russia, and later on, I danced in Fokine's 'Carnaval' [the best known of Schumann's piano cycles, choreographed by Fokine in 1910] with the Diaghilev company."

He insists too that the "subject" of the ballet is not Robert Schumann. "It's not about Schumann," he says, "it's about the music. Whatever I may have been thinking or feeling while I was making the ballet, that's nobody's business, that's only my business. It's like people will say about some music: 'Oh, that's 12-tone music.' So what? It has nothing to do with what one hears. After the first statement of the tone row, you get totally lost trying to follow it, it would take the most fantastic ear and harmony training and even then it would be nearly impossible. Or people talk about a fugue -- 'This is a four-part, or a six-part fugue' -- as if this had something to do with understanding the music. These are matters for the composer, not the listener. And so it is with my work -- I think as a dancer, I think in steps, not about characters or stories or philosophies."

Some concrete connotations of the ballet seem inescapable. Adam Luders has a solo that appears to depict beyond question the mental derangement which Schumann suffered, and his final shadowed exit inevitably suggests the composer's suicide attempt and asylum death. At one point, moreover, a number of black-clad figures who threaten Luders demand to be regarded as the hostile Philistines. Even Balanchine concedces this last point: "Yes, I put them in to be the element against Schumann. That's what the 'Davidsbundlertanze' is all about."

But the other dancers, he maintains, have no fixed identities. "Luders isn't Schumann -- he doesn't look like him, with that small head of his, and Schumann didn't dress that way. And none of the dancers is Florestan or Eusebius. Nobody knows what they looked like. Schumann made them up, not me. The dancers are just four couples who dance in a variety of ways, sometimes vigorously, sometimes poetically. The explanation is in the music. The dancers don't have to 'act' a part -- it's all in the music. These dancers are intelligent people, and they're more musical than musicians. So many musicians just play their instruments, and are only interested in how to make more money, how to play for television. But our dancers love the music and understand it."

In the last analysis, Balanchine construes the world of expressive movement and the world of verbal discourse as universes apart. "I don't have the brain of a writer, I don't think like a writer, I don't live in the world of words," he says. "I also don't have the time to put lots of things on paper and throw away all that isn't good. We dancers can't do that -- we make something as fast as possible, put it in the memories of dancers and show it immediately to the public. And much of it comes by chance, by accident -- remember Dr. Fleming, who discovered Pencillin. One day mold formed on his experiment, and that was it. That's what we do -- we start something and mold comes over it, and we say, 'That's great, that's it, let's keep it.' And we do."