A linguistic analysis done at Forbes’ request indicates that Newsweek fingered the wrong man when it tried to unmask the creator of the digital currency, Bitcoin.
Bitcoin allows users to conduct transactions with no or low fees and a relative degree of privacy. There is close to $8 billion of the currency on the Internet. But the identity of its creator (or creators), who went by the name Satoshi Nakamoto, has remained shrouded in mystery.
Last week Newsweek said it had discovered that Satoshi Nakamoto was a 64-year-old California engineer born as Satoshi Nakamoto, who legally changed his name to Dorian Prentice Satoshi Nakamoto 40 years ago. But Dorian Nakamoto repeatedly denied any connection to Bitcoin in an interview with the Associated Press. There is legitimate interest in a global phenomenon that could create new markets immune from taxation or regulation. But the story is also now being seen as a test of Newsweek’s credibility, because it was chosen as the cover of the magazine’s first print issue since it became an online-only publication in 2012, when it was owned by IAC. It was sold to IBT Media in 2013.
I asked Juola & Associates, a firm that uses a technique called stylometry to identify the authors of anonymous texts, to compare e-mail, message board posts and letters to the editor by Dorian Nakamoto to writing attributed to the Satoshi Nakamoto who created Bitcoin. The firm’s techniques, developed by Duquesne University associate professor Patrick Juola, seek to identify patterns of language use, like whether a person uses long words or short ones and what the most common words they use are. Juola has over 100 scientific publications and has written two books; he was also one of the forensic analysts who identified J.K. Rowling, the Harry Potter author, as the writer of a mystery novel, ‘The Cuckoo’s Calling,’ that she had published under a pseudonym. After the book’s true authorship was revealed, it went on to become a best seller.
John Noecker Jr., the chief scientific officer at Juola & Associates, had already become intrigued by the idea that his company’s techniques could identify the real Satoshi Nakamoto. In July 2013, after seeing an infographic on the website Chartgirl, he had tried the software on Satoshi candidates identified by journalists like Joshua Davis, in the New Yorker, and Adam Penenberg, in Fast Company. The candidates included Michael Clear, Neal J. King, Shinichi Mochizuki and Vili Lehdonvirta. All of them were researchers in cryptography, mathematics, or related fields who had public communications he could draw upon. All had also reportedly denied being the creator of Bitcoin.
His technique strips out formatting, and then examines: the author’s average word length; the 100 most common words used; groups of consecutive words, because people often use the same phrases unconsciously; and what parts of speech are used – do they use a lot of verbs, or adjectives, for instance. Then those numerical values are analyzed mathematically. First, Noecker looked to see whether the writing samples can really be grouped together. Then he looked to see if any overlapped. If Satoshi Nakamoto were a pen name for another author, their writing styles should match.
Noecker found that King was a better match than any of the other men.. But he says that it was impossible at to be sure King was Satoshi, because it wasn’t possible to say that those men are the only candidates. The problem is akin to a police lineup: If the culprit isn’t there, a witness might pick the best match from the lineup anyway, even though that person is innocent. “If we were trying to identify the Unabomber with a group of writings by President Obama, the works of Shakespeare, and Beowulf, we would expect that President Obama would show as the most likely candidate to be the Unabomber, simply because both authors write in modern American English,” he says. So Noecker did nothing with his finding.
I requested that he re-run the analysis with Dorian Nakamoto, largely based on writing compiled by a poster on the social media site Reddit. These included an e-mail he had sent to his local government, letters sent to model train magazines, and postings on message boards about model trains and firearms. Noecker concluded that King still seemed like the most likely candidate in the lineup; no method ever identified Dorian as being a more likely to be the Bitcoin creator than King, who still can’t be said to be a likely suspect.
“This does let us say fairly confidently that it’s not Dorian,” Noecker says. “If this were a [police] lineup, if everyone picks the same guy that doesn’t mean that guy is the perp. But if they’re picking the same guy and it’s not you, you’re not the perp.” At another point, he said that he can say “with some conviction” that Dorian is not Satoshi Nakamoto.
What about the possibility that Dorian Nakamoto wrote the Bitcoin whitepaper with the help of a co-author or editor who had a significant hand in crafting it? Could it be that these were Dorian’s ideas, but someone else did a lot of the writing?
“It’s possible it could have a significant effect, but normally we can identify both the editor and the author,” says Noecker. “If it was edited really significantly, if Dorian had written up an outline and someone else had followed it, then it is possible that could ‘wash out’ the effects. I would have expected to see a closer correlation than we did even with fairly extensive editing.”
This is only one more piece of data, and it hasn’t been subjected to scrutiny by other forensic experts. But it’s a piece of evidence that seems to exclude Dorian Nakamoto as the Satoshi Nakamoto behind Bitcoin, and to increase the odds that Newsweek got it wrong. At the very least, it makes it hard to say the magazine got it right.
Newsweek and its story’s author, Leah McGrath Goodman, did not return emails requesting comment.
–Andy Greenberg contributed reporting to this story.