OPEN APP
Home / Opinion / Columns /  A tenth of Ludwig Van Beethoven
Listen to this article

When I was younger, so much younger than today, I learned the piano. Western classical music: Mozart, Strauss, Bach, Beethoven, Chopin, Debussy ... I grew to love Chopin’s heartfelt C# minor waltz, Beethoven’s playful Für Elise, Mozart’s elegant C Major Sonata and Strauss’s lyrical Artist’s Life. There were other charmers too: Joplin’s peppy ragtime classic The Entertainer; the yearning in the tango El Choclo; the spirited El Jarabe Tapatio or Mexican Hat Dance, that was adapted into Jeevan ke Safar Mein Rahi from the film Munimji.

Seems I’m showing off? Maybe I am. So much gorgeous music that I played on my piano—and then it was lost. That story, another time.

Of these composers, I nursed a slight preference for Beethoven. In my mind, I often compared him to Mozart. My likely superficial impression remains: Mozart’s music is forever soaring and elegant; Beethoven’s is deep, intense and introspective. Both challenged and thrilled me, but in a way I never quite put my finger on, Beethoven resonated a little more.

And that, coupled with a brush with artificial intelligence (AI) in my computer science days, is why some recent Beethoven news grabbed me. This news: Not long ago, an orchestra performed his 10th Symphony.

What’s newsworthy about that, I’m sure you’re asking. The man has been dead nearly 200 years, and his compositions have all been performed innumerable times the world over. Why did this performance of his 10th make the news?

Because Beethoven never composed a 10th Symphony.

There’s his famous 5th (Victory) with its stirring four-note riff to start. There’s his majestic 3rd, originally called Bonaparte, but renamed Eroica when Beethoven became disillusioned with Napoleon. There’s his 9th (Choral).

There isn’t a 10th.

Though Beethoven did start composing it. He wrote “10th Symphony" at the top of blank musical sheets and scribbled some cryptic musical notes below. But these were really just musical ideas. Where Beethoven intended to go with them, we don’t know. He died in 1827, before he was able to take the ideas any further, let alone complete his 10th Symphony.

In 2019, a team of computer scientists and musicians—led by a professor in AI at Rutgers University, Ahmed Elgammal, and the Austrian composer Walter Werzowa—decided to try to go where Beethoven wasn’t able to. They set out to complete his 10th, calling their project Beethoven X. They put AI to work to expand the few notes from 1827 into a full-fledged symphony. Beethoven’s 10th, come to life.

That was the goal.

Let’s try to spell out some of what went into this. A symphony is typically 20-30 minutes long, divided into four sections (“movements"). The first is usually lively and crisp; the second is slower and expressive; the third is often a dance (called a “minuet") or a playful form known as a scherzo (which means “joke"); and the fourth is an exuberant, up-tempo end to the composition.

Beethoven’s symphonies follow this general schema, but saying so speaks of only the broadest possible outline of his works. The job this team set for themselves was to teach their computer something approximating the great man’s creative process. Where did his musical ideas spring from? How did he transform a few scribbled notes into his eloquent, lyrical compositions?

To answer that, the team used all nine of Beethoven’s symphonies, trying to decide what a given skein of notes from his sketches for the 10th meant. Did it belong in a minuet? Or the slower movement? But perhaps harder, how would the simple melody in that skein grow into the intricate musical structures Beethoven built? All of the 5th Symphony, for example, is founded on those four famous notes, the first three of them even identical.

More questions: How do you find harmonies? How do you end a passage of music, then feed into the next? What about ending an entire movement, and the symphony itself? How do you return to an original idea with enough variation to make it interesting, but still recognizable? What happens with an entire orchestra—how do you tease out different strands of music for different instruments, making sure they are appropriate for those instruments and still combine well?

Much to teach the AI programme, so that it could then teach itself still more and eventually start producing music. After all, that’s how creativity works with us. We are taught the basics—writing out words, playing notes on an instrument, hitting a tennis ball across a net, whatever. Those are the tools that let us learn even more and discover other tools and experiment with what we know and push our boundaries and learn still more ... and eventually it’s hello Sally Rooney, or Meghan Trainor, or Emma Raducanu.

Or Ludwig van Beethoven.

In late 2019, several months after they started, the team recruited a pianist to perform the snatches of music their programme had so far generated. In the audience were several people who knew Beethoven’s music well. When the pianist performed, nobody listening was able to distinguish Beethoven’s notes from what AI had rendered. That seemed like something of a triumph, for sure. Still, these were just snatches, only a few minutes long at best. But even so, one of the musicians on the team said that the programme “reminded him of an eager music student who practises every day, learns, and becomes better and better."

How else, after all, do we become better at our respective crafts?

The team stuck to their task. So did the AI programme. The hope was that they could premiere the constructed (or reconstructed, whatever you prefer) symphony last year, 250 years since Beethoven’s birth. The pandemic interfered with those plans. But on 9 October this year, the Beethoven Orchestra in Bonn performed the new symphony for the first time.

I’ve listened to it. In many ways it is very Beethoven-like—the patterns of the notes, the crescendos, the finales of the movements. There are echoes and reminders of his other symphonies—the four-note motif of the 5th, for example. Taking a break from writing this paragraph, I played the third movement, the scherzo, for a fan of classical music, without telling her what it was. A couple of minutes into it, she asked: “Is this Beethoven?" And I replied: “Yes and no."

So at some level, you might say the team has succeeded.

But as you can imagine, the project and its musical result has its critics. Jan Swafford, author of a biography of Beethoven, wrote a critique in the online classical music magazine Van (The Intelligence of Bodies, bit.ly/3p5E88a). Listening to the symphony, he wrote, made him “more philosophical than annoyed."

Computers “can’t account for the heat or alarming lack of it in the room, sensations in the groin ... the woman or man you just met who excites you and whom you hope to excite ... the bastards who write better than you ... hairs falling out of your head onto the page" and more. In other words, AI can never understand the creative expression that floods out of “living, breathing, loving, lusting, aspiring human beings."

It’s true: Swafford is a Beethoven expert and knew from the start that this wasn’t a Beethoven symphony. Still, for him the music “doesn’t sound like Beethoven ... in the end it’s aimless and uninspired and sometimes tries too hard."

Decide for yourself. Beethoven X: bit.ly/3DRbSfj.

Once a computer scientist, Dilip D’Souza now lives in Mumbai and writes for his dinners. His Twitter handle is @DeathEndsFun

Subscribe to Mint Newsletters
* Enter a valid email
* Thank you for subscribing to our newsletter.

Never miss a story! Stay connected and informed with Mint. Download our App Now!!

Close
×
Edit Profile
My ReadsRedeem a Gift CardLogout