Part of writing a play is letting it go. It's both exhilarating and a bit frightening when you turn your script over to the director and actors who will try to make it live. It's a risk--you hope you'll get lucky. With "Proof," I did. But when I let this play go I had no idea how far it would travel.

The play has been done in London, Tokyo, Manila, Stockholm, Tel Aviv and many other cities; the definitive New York production, directed by Daniel Sullivan, opens in Beverly Hills this week at the Wilshire Theatre.

"Proof" started with two ideas. One was about a pair of sisters: What if, after their father's death, they discovered something valuable left behind in his papers?

The other, more of a visual image than anything else, was about a young woman: I saw her sitting up alone, late at night, worried she might inherit her father's mental illness.

While trying to see if these ideas fit together, I happened to be reading "A Mathematician's Apology," the memoir by the great Cambridge mathematician G.H. Hardy. It's probably the most famous attempt to explain the pleasures of doing math to a non-mathematical audience. One passage particularly startled me.

"In a good proof," Hardy wrote, "there is a very high degree of unexpectedness, combined with inevitability and economy. The argument takes so odd and surprising a form; the weapons used seem so childishly simple when compared with the far-reaching consequences; but there is no escape from the conclusions."

That sounded like a definition of a good play, too. Math was alien territory to me--I had barely made it through freshman calculus in college--but I decided to set my story in Hardy's world. A mathematical proof became the "thing" the sisters find; my protagonist, Catherine, became convinced that she may have inherited her father's talent--he was a legendary mathematician--as well as his illness. With these elements in place, and feeling inspired by the meetings with the mathematicians I'd begun to have, I was able to finish a draft of the play quickly, in about six months.

My first play, "Skyscraper," had been commercially produced off-Broadway in 1997. Its run was short, but long enough for the literary staff at Manhattan Theater Club to catch a performance. They had invited me to submit my next play--a good break for me, since MTC is the best venue for new work in the city. I sent "Proof" to them. A few weeks later, it had a star, Mary Louise Parker, a director, Daniel Sullivan, and an opening date for what I assumed would be a six-week run.

"Proof" has now been running for two years. In that time, I've often been surprised at the responses it has generated. At a New York University conference on the play, a panel of women mathematicians used it to discuss questions of sexism and bias in their profession. After a performance on Broadway I got a note from an audience member backstage: "My daughter is just like Catherine," it said. "I can't communicate with her. Can you help me?"

In Chicago, a woman confronted me after a book signing. She told me her father had been a mathematician who'd lost his mind and she'd spent her whole life caring for him. "This is the story of my life," she said. "How did you know?"

The answer, of course, is that I didn't, any more than I intended the play to speak directly to the concerns of female academics, or could tell a stranger how to break through to his daughter. When you let a play go, you also take the risk that it will take on associations for people that you didn't intend and can't account for.

That risk is the prerogative of art, however, and of the theater in particular. The theater affects us more directly, and unpredictably, than any of the other arts, because the actors are right there in front of us, creating something new every night. Something, as Hardy might put it, "Unexpected and inevitable." Which makes it all worth the risk.

*

David Auburn won the Pulitzer Prize for drama in 2001. "Proof" also won the 2001 Tony Award for best play; Daniel Sullivan won for best director; and Mary-Louise Parker won best actress (play).