Last year I had a mildly ranting blog-thing (What – Weaver, rant? No way!) on a peer review site, wherein I attempted to explain to several of my fellow writers that the “I Write Like” analyzer is for amusement only and not to be taken seriously. Some of these kids were so very, very proud that their analysis came back telling them they could write just as well as Stephen King or J. R. R. Tolkien… A simple computer program cannot possibly determine this, I said; all it does is look for certain keywords (such as if the protagonist is named Bella, the result will almost certainly be “Stephenie Meyer” regardless of actual writing style), and the list of possible writers for comparison is too small to mean much anyway.
To make my point, I ran several of my own stories, and some written by my twin, through the analyzer and posted the results. Alas, I lost that post when I quit the site it was on. So I just had to do it again.
This blog post consists mainly of those results.
The Excalibur Mission
This is a science fiction novel (or science fantasy, if you’re the type who insists that psionics cannot exist in “real science fiction” – but if you are that kind of reader, don’t tell me, or I’ll have to write a ranting blog-thing all about you) that I and my clone-sibling have been working on together.
Prologue: Stephenie Meyer (This result is based ENTIRELY on a character name, which seriously pisses me off. I changed that character’s name to his most recent alias, and the analysis came back with “Cory Doctorow.”
Here are the results for the other chapters, by count:
Arthur C. Clarke: 12
David Foster Wallace: 4
Dan Brown: 3
Raymond Chandler: 1
Cory Doctorow: 1
Oscar Wilde: 1
William Gibson: 1
H. G. Wells: 1
How likely is it that two dozen chapters of the same novel are going to be this different? Even sticking with the sci-fi authors on the list, it’s all over the place. No one actually writes with that much variation in style unless they do it deliberately or they’re totally incompetent at keeping narrative voice consistent. [Idea for a test: give the analyzer a fragment of autobiography of CC/CB and see if it squawks at the “inconsistent” voice. A few critics certainly did — the idiots.]
Now, I’ve read novels by Arthur C. Clarke and H. G. Wells, and last year I finally read Neuromancer by William Gibson. Earlier this summer, I read a single short story by Cory Doctorow. But I’ve never read anything by Oscar Wilde, David Foster Wallace or Raymond Chandler (I can only extrapolate what Chandler’s style may be like from one critic’s comment many years ago that Roger Zelazny’s Amber novels read like ‘The Lord of the Rings as told by Raymond Chandler’), and I wouldn’t read anything by Dan Brown unless paid a lot of money to do so.
Then there were my shorter works.
version 1: Gertrude Stein
version 2: Gertrude Stein
This is the other story that has come up in my posts about peer critiques. Version 2 is my clone-sibling’s rewrite to see if we could change readers’ perceptions of the protagonist. (Some readers kept insisting that my female protagonist is unbelievable because ‘men can’t write female characters.’ Funny how they liked the revision even though it, too, was written by a man. We just didn’t tell them that, and the default assumption seems to be “female protagonist = female author = realistic writing of female protagonist.” Idiots.) Granted, we also didn’t want to drastically change the narrative voice from one version to the other.
To test a hypothesis that had shown positive results in the past, I temporarily changed the protagonist’s first name to Bella, and got the same results. Conclusion: the “Gertrude Stein” keywords (whatever the hell those are) must be a lot stronger than the four times the protagonist’s name appears in my story.
“Nonlinear” (poem): Chuck Palahniuk
I don’t often write poetry. Not my thing. I occasionally read it, as when I read everything by Robert Frost as research for my fiction (after JG informed me that he really, really likes “Stopping by the Woods on a Snowy Evening,” although he’s not sure how he feels about “Mending Wall”), and I nearly did a happy dance when I found “To Spin Is Miracle Cat” online, but my own poetry is generally too idiosyncratic — or perhaps idiopathic — for audiences. I’d never heard of Mr. Palahniuk before doing this test; I had to look him up on Wikipedia.
[Edit: You can find this poem here: Nonlinear ]
“Ice Is Also Great”: Arthur C. Clarke
“A Dangerous Thing to Bear”: Douglas Adams
These short stories feature the same protagonist and are written in the same style, no matter what the “I Write Like” analyzer says. I won’t say that “Ice” was not strongly influenced by other writers, but Clarke isn’t on the list.
“Redoak”: J. D. Salinger
Maybe because the story – what there is of it so far — appears to be mostly about normal people (university students) doing normal things (a freshman meeting his roommate for the first time, various characters hanging out with friends and talking about classes, etc.). As I said, I don’t know Salinger’s writing, so I can only guess, and my guess is that this analysis is inaccurate.
“Raven’s story”: James Joyce
A fragment that has been absorbed by one of my twin’s novels. This one, oddly enough, makes sense, given the limited options. Lots of sentence fragments in this narrative. Lots of strange imagery, streams of consciousness, etc. From what little I know about Joyce, this fits a little bit. However, that does not mean that I actually write like Joyce. I simply had, in this fragment of story, enough of the right keywords and sentence-structure elements for a computer program to find a “match” and say that it’s just like something James Joyce would write.
Sometimes it’s good for a laugh, but don’t take it seriously.