I must be getting old, but this morning I learned about an entirely new concept called “inauthentic text.”
An old post was pinged by a blog which linked what I said, and when I went there to read the post and possibly thank the blogger, I saw that the link had absolutely nothing to do with the point in my post. Nor could I make any sense out of what the “blogger” was saying. I thought she might be schizophrenic, because she would use certain stock phrases, but also, the grammar and syntax were in other places so twisted that the text appeared to be computer generated translations.
The blog is nothing but a spam blog promoting a book I will not mention lest I help the S.O.B. who wrote it.
It consists of what is called “inauthentic text.” I copied and pasted a paragraph to this “Inauthentic Paper Detector”
This text had been classified as
INAUTHENTIC
with a 32.0% chance of being authentic text
OK, so for control purposes, I copied and pasted every word from this recent post, and to my horror, I got this:
This text had been classified as
INAUTHENTIC
with a 40.3% chance of being authentic text
So what’s up? If my writing is inauthentic, what is to become of me?
Not wanting to take this personally, I entered M. Simon’s latest post, and it, too, was labeled “INAUTHENTIC.”
Next, I copied and pasted Sarah’s great post about the Republican “war on women.”
This text had been classified as
INAUTHENTIC
with a 34.0% chance of being authentic text
What could be going on? How about Dave? According to the miserable excuse for an authenticity detector, Dave’s writing was said to be even less authentic:
This text had been classified as
INAUTHENTIC
with a 18.1% chance of being authentic text
Could all four bloggers here be writing inauthentic text? I tried entering randomly selected news articles, and even an editorial piece by Glenn Reynolds. (All were called inauthentic!)
Is it possible that the “Inauthentic Paper Detector” (which I found linked at the Wiki entry on the subject) is itself less than authentic? It is quite clear to me that it simply labels whatever text is entered as inauthentic. Why would Wiki link such fraud? What’s more shocking than that is that scientists seem to have taken it seriously.
For control purposes, I thought I would try entering the text of the Gettysburg Address.
This text had been classified as
INAUTHENTIC
with a 17.4% chance of being authentic text
Well, at least the detector reveals what we write here is more authentic than the most famous speech by Abraham Lincoln, written by him at a crucial period in our history.
Small comfort that.
UPDATE: In my vain attempt to find any redemption in the “Inauthentic Paper Detector” I noticed that the site’s emphasis seemed to be on scientific papers. So I copied and pasted the first page of Einstein’s famed paper, Relativity: The Special and General Theory:
Einstein, it turns out, was inauthentic too!
This text had been classified as
INAUTHENTIC
with a 32.2% chance of being authentic text
If Einstein’s relativity is inauthentic, the world is in bigger trouble than I realized.
Comments
8 responses to “Is nothing authentic anymore?”
If the detector always returns “TRUE”, and is usually tested with suspect test from a spam-blog, then it is easy to fool most people into viewing the detector as accurate.
That’s one explanation.
Another explanation is that the detector has some very shoddy algorithms.
Dammit! You caught me!
[…] See Eric’s latest post for details on why I will not mention the book. Print PDF Categories: Uncategorized 1 […]
Well, it claims to be checking for authentic scientific papers, not simply authentic text.
I tried feeding it selections of some technical reports (written in casual style) and got scores in the 20s.
I’m guessing it wouldn’t like the Chicken Paper either: right form, wrong content.
I don’t have any bona-fide scientific papers, written in the approved style, handy for electrical cut’n’paste.
Well, Einstein’s relativity paper is one of the most famous of all time and it failed.
Hey, how about some Global Warming science? Maybe that would work!
I submitted chapters 5 and 10 of Darwin’s Voyage of the Beagle and got scores of about 75 and 82 percent.
“The main purpose of this software is to detect whether a technical document conforms to the statistical standards of an expository text.”
I will LMHO if their statistical base is skewed toward long and archaic, or puffed up B.S
[…] anyone demand, say, a burger with artisanal mayo; as Cheryl Lynn says, it’s got to be real. Eric Scheie reports that his text apparently lacks authenticity: Could all four bloggers here be writing inauthentic text? I tried entering randomly selected news […]