Here’s what everyone knows about Proust: he was languishing in misery and ennui until one day he ate a madeleine and suddenly remembered the past he had so long forgotten.  Immediately he felt happy again, “the vicissitudes of life indifferent to [him], its disasters innocuous, its brevity illusory.”

Here’s the problem with what everyone knows: this never happened.  Not in real life.  In his novel, Proust makes this happen to his character.  But his character remembers things that are not part of Proust’s life (such as “Aunt Léonie”); his character is Catholic (Proust was Jewish); his character grows up to be a heterosexual (Proust himself preferred men).

Can’t we at least say that Proust had a special, epiphanic experience while eating a madeleine one day, even if its content was different?  Unlikely.  This scene actually went through more than one draft, and an earlier version has the character eating not a madeleine but a biscotte (somewhere between a cracker and a piece of toast).  If the madeleine scene were simply a transcription of Proust’s diary, these different versions would not make any sense.

This didn’t trouble people like George Painter.  Painter accepted that there couldn’t have been a real epiphany with a madeleine... so it must have been a real epiphany with a biscotte!  Proust was languishing in misery and ennui until a fateful moment—“on or about 1 January 1909” (I love this precision!)—when he ate a biscotte and suddenly remembered his past; involuntary memory changed his life.  That became the standard view for a while, and Roger Shattuck, Ronald Hayman, Derwent May and others all went along with it.

But hang on: for this to be true, it would have to have been Proust’s first experience of involuntary memory.  Surely no-one’s life is going to be changed by the nineteenth time something or other is revealed to them; it has to be the first.

Now Proust did not have his first experience of involuntary memory in 1909.  Not even close.  Jean Santeuil—a manuscript Proust was writing in the 1890s, over a decade earlier—is packed with involuntary memory scenes.  (I’ve counted about a dozen, including—interestingly enough—one involving a cup of tea.)  And it offers lavish accounts of why they are so important.  (Yes, they reveal the existence of a consistent self, an atemporal true self that feeds on the eternal; that’s why they fill you with joy; they are a spur to writing; they offer a past purified of desire...)

So a second generation of Proust biographers patiently explained that in fact the madeleine scene is made up.  Proust, they pointed out, already knew about involuntary memory in the 1890s; he already had theories about it; he already planned to include it in his writing.  Nothing further needed to happen along these lines “on or about 1 January 1909.”  And it almost certainly didn’t.

But here’s the thing: all this patient explanation doesn’t seem to have done much good.  Let’s look at Jonah Lehrer, whose 2007 Proust Was a Neuroscientist made such a splash.  Incredibly, Lehrer goes even further than Painter and company: he thinks it really was a madeleine, not a biscotte!  “Proust’s solution,” he says, “arrived in the unexpected form of a buttery cookie flavored with lemon zest and shaped like a seashell.”  (He's clearly been eaten one too many products from the Illiers gift shop. http://www.tourisme28.com/uk-proust-s-madeleine/uk-proust-s-madeleine.php)  He even dates the purported event two years later, in “1911, the year of the madeleine.”  1911?  The biscotte episode was written in 1908-9!

This fallacy is like a zombie: you shoot it twenty times, and it still keeps on coming.

So here’s my question: how can we kill it once and for all?