It is plopped down in front of you and you poke it and prod it and try to figure out what it is. If it is something very new or very different, it may take more poking than usual, and you may very well come up with the wrong answer. 

This is what it is to be a critic — a real critic, I mean, not one of those Yelp scribblers, or self-certain mandarins with nothing more to offer but thumbs turned skyward or hell-ward. 

I was a critic for 25 years for a major daily newspaper (The Arizona Republic in Phoenix, Ariz.), and I always thought of my job as being a first reader, or first seer, or first listener — a pioneer trying to make sense of something before any sort of consensus has been reached. It is a risky thing to do — to proffer an opinion before you have anyone watching your back. 

When conductor Pierre Boulez first came out with his version of Wagner’s Ring Cycle in 1976, he played as if it were chamber music, it was such a different conception of the music that many critics first assumed it was a failed attempt to make the grand, Romantic mytho-philosophical monster it had always been taken to be. The BBC criticized the conductor’s “ruthless tempi” and “lack of expressiveness.” 

Consensus has now realized it was a brilliant re-thinking of the way the music could make sense. The clarity he brought to the muck (beautiful muck), was transcendent in its own way. Later criticism decided instead that “Wagner’s music doesn’t have to be murky to be metaphysical or massive to be overwhelmingly moving and Boulez gets playing from the too-often turgid Bayreuth Festival Orchestra that makes the music crackle and blaze with musical and dramatic meaning.”

Being a first listener is always risky. You may think something a failure because it doesn’t do what it has done before, failing to hear that it is doing something new brilliantly. 

When the Boulez Ring was new, the critics poked and prodded to see if it was alive. Now, we know not only that it was alive, but that it was a harbinger of a new way of playing classical music that has taken over the business. Out with the Furtwängler, in with the John Eliot Gardiner. Lean and mean drove out lush and weighty. 

I am reminded of this because I have just been confronted by a new recording of Bach’s Goldberg Variations under the fingers of Chinese Wunderklaviermeister Lang Lang. Critical reaction has been all over the place, from deciding it was “the greatest version since Glenn Gould” to complaining that it sounded like a talented conservatory student sight reading. 

And I see what each reaction means: They parallel my own thoughts. Is this a brilliant rethinking, or is this a flaming dumpster fire? 

Mr. Lang, or if I may be so familiar as to call him by his first name, Lang, has always split opinions. Sometimes it is hard to bust through his relentless self-promotion — the kind of commercial huckstering usually identified with charlatans and snake oil salesmen — and then there are his stage antics, eyes closed in thesbian rapture, rolling his head back and forth in a way to make Leonard Bernstein seem like a mechanical clock. How can you take him seriously? 

And yet, there is often magic in his playing. I have heard him live several times, and his performances varied widely, from glib to dazzling to absolutely empty. Yet, at other times, it was profoundly moving. When I heard him play the Chopin E-minor concerto, the way he played the slow movement made time stand absolutely still. It is one of the most soul-satisfying performances I have heard in a half-century of concertgoing. It probably helped that I normally close my eyes when listening to music and therefore was spared his facial contortions. God, was that moment beautiful. 

His recordings are equally all-over-the-place, with some dead-on and concentrated and others distracted and hollow. Lang is capable of so much, but only delivers intermittently. 

So I am now confronted by a Goldberg Variations unlike any I’ve heard before. Is it brilliant or dunderheaded? Is it an aberration or is it a signal that classical music culture is shifting once again? 

I’ve heard a lot of Goldbergs in my time. They were little known or played before 1955, when Glenn Gould launched them on an unsuspecting public, with a blazing performance that redefined Bach playing, clarifying the polyphonic strands and cutting down the pedal, almost mimicking the sound of a harpsichord. Since then, in the same way that no self-respecting art photographer can fail to make a photograph of a green pepper after being shown the way by Edward Weston, so no decent pianist can avoid recording a set of Goldbergs. Most of them are perfectly decent, if anonymous. 

In recent years, a few with real personality have been released. Simone Dinnerstein has her set and more recently Jeremy Denk. Some may remember the brief surfacing of a recording by João Carlos Martins that was almost as idiosyncratic as Gould, although Martins’ piano never seemed to be quite in tune. 

The work has also been transcribed for accordion, marimba, harp, hammer dulcimer, guitar, saxophone quartet, string trio, string orchestra, synthesizer, and brass quintet. I have a recording of parts of them on Japanese koto, and Yo-yo Ma recorded the Aria on his cello. Avant-gardist Uri Caine made his version updating each variations individually for a heterogeneous mixture of voices, instruments and recorded noises. I once put together a CD mixing many of these oddball transcriptions into something I called the “Goldberg Variorum” — each variation played by a different instrument or group. 

Gould recorded the Goldbergs at least four times — with untold bootlegs out there. The initial set has been reissued so many times in different albums, that it is impossible to keep count. Russian pianist Tatiana Nikolayeva recorded them five times and Rosalyn Tureck did it seven times. 

The earliest version I could find was on Welte piano rolls (a kind of player piano) from 1928, by Rudolf Serkin. (His son, Peter, left us three versions). Since then, there have been close to 250 recordings. About 50 of those are on harpsichord. A few are the transcriptions, but almost all are on piano. 

But since Gould, most pianists have hewn to the stricture that, since they were composed for the harpsichord, they should be performed as drily as possible, and with little or — preferably — no pedal. And since the arrival of “historically informed performance practice” (fie on the miscreants, I say, fie) boatloads of pianists have done their best to erase any notion that a performer should “interpret” the score. Just the notes, ma’am. 

This has led to quite able, but faceless performances by such as Angela Hewitt, András Schiff and Murray Parahia. I don’t mean to poo-poo these recordings, They are all excellent of their kind, but they are chaste. 

And so, we come to Lang Lang’s two-disc set, taken at a rather leisurely pace, but with lots of spark and crackle in the details. He likes to thump hard on stray notes and he adds many ornaments, especially in the repeats — and not all the trills and mordents are stylistically appropriate. Extra passing tones and tons of rubato. Worse: Pedal. In modern terms, this is Bach done in “bad taste.” 

In the old days (pre-HIPP), pianists tended to play Bach on piano as if he had written for piano. They brought out tunes and backed them with accompaniment. Now, we revel in the polyphonic strands, each brought out cleanly. If you listen to pre-World War II recordings of Bach, you will hear pianists such as Edward Fischer or Wilhelm Kempff play their Bach as if he were the godfather of Chopin. 

You can hear the echoes of this Bach in the Well-Tempered Clavier of Daniel Barenboim. They are magnificently played, but purists cover their ears and bray “Nyah-nyah” to block out the sound. Yet, there is a long tradition, now largely buried, of approaching Bach’s music as a pretext for piano playing, showing off the performer’s skills and sensibilities. After all, do you go to La Boheme for the story, or to hear Pavarotti? The present orthodoxy considers this a kind of blasphemy. 

Yet, the music no longer belongs to Bach; it is ours and we can express our ownership of it any way we wish (as someone once said about modernized performances of Shakespeare, set on the moon or done with an all-female cast, “It’s OK. They haven’t destroyed the text. It is still there, unharmed.”) And no matter what you may think of Lang Lang’s performance, the text is still there. It belongs to you, too. But Bach himself is no longer here; he has no say in the matter and we are presumptuous if we claim to speak for him. 

So, maybe, after 40 years of increasing musical priggishness and the cult of “composer’s intentions” we are beginning to loosen up. After all, it is Postmodern doctrine that it is all just “text” to be worked on by each of us. 

The audience that actually cares about classical music seems divided into two unequal groups. The larger posits a Platonic ideal performance and judges each concert by how close to this ideal it reaches. Of course, each listener has his own vote for what that ideal is. But the goal is always the “perfect” realization of the score. 

But the other group seeks constantly to be surprised, to see the notes through a newer lens and have the music refreshed. “To hear it again for the first time.” They expect each concert to give them a different version even of old chestnuts. The standard issue performance bores them. 

I can give a great demonstration of the difference. When Anne-Sophie Mutter first recorded the Beethoven Violin Concerto with the Berlin Philharmonic and Herbert von Karajan, she gave what must be the closest to the perfect ideal. Nothing out of place, everything beautiful and expressive. As if played by angels, not humans. It has remained in print since it was released in 1979. She recorded it again in 2002 with Kurt Masur and the New York Philharmonic, in a performance much more personalized, with very different phrasings and dynamics. No longer a Platonic ideal, but much more a here-and-now. You would never confuse it with any other performance or violinist. I love it; many fans hated it. Hate, hate, hated it. 

The problem with the Platonic performance is that the ideal changes over time. Once, the perfect Beethoven was Furtwängler, then it became Szell, and after that, it became John Eliot Gardiner. Tastes change over time. 

We seem to be in another shift, giving up the impersonal historically strait-jacketed version for a reintroduction of the more individuated performance. We hear it in the recordings not just of Lang Lang, but of Mikhail Pletnev, Valery Gergiev or the late Nikolaus Harnoncourt (operating under the deceitful guise as an “original instrument” guy — but really just sui generis.)

Really, the Platonic template has been in place mostly from World War II on. Before that, performance idiosyncrasy was the norm, from Vladimir de Pachman to Willem Mengelberg. Leopold Stokowski was famous for tinkering with scores and glamming up what he was conducting. Now, having gone through that, into the Post-war standardization and then the HIPP diminution, we seem to be re-entering an era of increased personal interpretation. 

And that brings us back to Lang’s Goldbergs. If we are in the cusp of a change, we cannot really be sure where the gamepiece comes down. This new recording may indicate a reshaping of the way we play Bach, the way Gould reshaped it from 1955. Or maybe it’s just a garish one-off. 

I poke it; I prod it. I place my bet. So many of the hundreds of recordings of the Goldberg Variations are magnificently well-played and satisfying in their own way, but how many are memorable? You could replace one with another and be equally pleased, indeed not even to notice the difference. Gould was memorable: You can spot it in a crowd of hundreds. Lang Lang’s recording is the most memorable I’ve heard since then, and I’ve heard a boxload of them. It is memorable, but is it good? Will its novelty wear thin, or become the new norm? 

I am going back for a fourth dive into the new recording. Then, maybe, a fifth. Perhaps after that, I’ll have an answer.