“Dumbledore will get out from behind a cream cake” isn’t an original sentence after all
Jonathan Basile has written this sentence 3,153 times
A week ago, I published a new Harry Potter chapter written by a machine learning algorithm.
In an interview about the piece, I commented on the computer’s creativity:
“…you can argue that the Harry Potter neural network was very creative. In fact, this might be the first time anyone, or anything, has constructed the sentence: ‘Dumbledore will get out from behind a cream cake.’”
It turns out though that I was wrong. The sentence “Dumbledore will get out from behind a cream cake” has been constructed at least another 3,153 times. And all of the instances can be attributed to the same guy: Jonathan Basile.
The Library of Babel
Jonathan Basile runs a website called The Library of Babel, which attempts to contain “every book that ever has been written, and every book that ever could be — including every play, every song, every scientific paper, every legal decision, every constitution, every piece of scripture, and so on.”
To do this, Jonathan has used a computer to generate every possible combination of 3,200 characters (with the aspiration of reaching every possible combination of 1,312,000 characters).
For the most part, Jonathan’s website looks like this:
“vw.kvxhclctvbny,zjfrq,,ym,uaggibseekhmkltfjdmuzamsysnrqvomirfataytthudkqtq,ezvlflmynbdqrssliqjbkpapratod,qaijshhjunqkuhidlspgjvbbyoadufbkrgeowy,i.pkft,dptiuovqbgx.lpsyikpevksnztkqyhnsfzst.gmicw.j,.tpirr”
Yet, occasionally, a “lucky” streak of characters will connect and spell something a little bit more meaningful.
“psianpvwhswixmaysobfzhpel.,wlzdumbledore will get out from behind a cream cakeiheupslf pttbzpz.vetygeklgap.ivrkruu,pxgcmcxagnnxafanehmtjq”
And since this sentence is only 48 characters long, it appears on Jonathan’s website 3,153 different times* (since the sequence can start on all but the last 47 characters of a 3,200-character chunk).
The Unbelievable Nature of Random
When Carson Kahn first showed me the “Dumbledore Sentence” amongst a random assortment of characters, my instinct was to think it was fake. Even though I know it’s inevitable that this sequence must be generated (based on the premise of the site), my brain still has trouble dealing with this reality. Since I only see a very tiny portion of the entire text, my probabilistic intuition is completely confused.
Nevertheless, here it is. And it is very real.
I’ve searched for a bunch of other sentences on Jonathan’s site, and sure enough, they too were there. Of course they were. They have to be. And yet, seeing them there still created this feeling of wonderment.
There is something really pleasurable about knowing that enough randomness can produce a little bit of order.
This entire post has already been written
I encourage you to play around with The Library of Babel, and search for your own obscure, impossible-to-believe-they-exist sentences. Even though you know they are there, you’ll still be surprised when you find them.
You’ll even find this entire Medium post already written.
*The number 3,153 disregards all possible combinations of random letters, and assumes that the sequence only shows up once on each page. This is a massive underestimation, but aids storytelling. The true number is closer to 9 x 10⁴⁶¹².