• Member Since 11th Apr, 2012
  • offline last seen 1 hour ago

Bad Horse


Beneath the microscope, you contain galaxies.

More Blog Posts758

Jul
24th
2015

Did Harper Lee write "Go Set a Watchman"? · 2:50am Jul 24th, 2015

I haven't read Go Set a Watchman, the new old book by Harper Lee, who wrote To Kill a Mockingbird. But Stylo has.

At rem-dog's suggestion, I tested whether Harper Lee really wrote Watchman using Stylo. Here's Stylo's cluster analysis of Go Set a Watchman, To Kill a Mockingbird, a set of essays by Harper Lee, and a bunch of other stories. As I explained earlier, Truman Capote might have written To Kill a Mockingbird, so I included his work. I also included a bunch of other American English writers. (I had some British writers at first, but they all clustered together far across the water, so I took them out.)

I took up to 2 samples of 9000 words from each story. We'll look at 2 presentations of Stylo's data, cluster analysis and PCA into two dimensions. The PCA approach is less deceptive, because it plots things in a metric space, meaning that things that look close together really are close together, while things that look far away really are far away. However, the cost is that you're using less information--only 2 dimensions of data that has many dimensions--so the picture is inaccurate. The clustering approach has more power, because it can look at all the dimensions and use them to recursively split the stories into 2 groups that are similar to each other. But it's more deceptive, because stories that appear close are close, but stories that appear far away can also be close.

Here's the cluster analysis. (I used the 90 most-common words, instead of the 100 most-common as I usually do, only because the latter split Ernest Hemingway across the top and bottom in a way that made them look very far apart when they were really pretty close. Other than that they were nearly identical.)

We see that it does a great job of putting the different pieces of stories by the same author, and different stories by the same author, together. We also see that 1) Watchman and Mockingbird are right next to each other, 2) both are far away from Harper Lee's essays, 3) neither are near Truman Capote's books, but 4) Truman Capote's fiction isn't close together, either.

Here's another cluster analysis, using 6500 words per story, taking up to 3 samples per story, and including the English authors. Pretty much the same results, but note that here Truman Capote's nonfiction is right next to Harper Lee's nonfiction. About half of them were written during the years she hung out with Capote.

You might say, "So what? Watchman and Mockingbird are about the same people, in the same town. They'll have the same words just because of their names." But I checked the frequent word set used, and there are no names in it.

Here's Stylo's 2D PCA analysis of the same set of stories:

We see the same thing again. Note the PCA puts Harper Lee's fiction and Truman Capote's fiction closer together. (In Cold Blood isn't fiction.)

I still don't know whether Harper Lee or Truman Capote chose the words used in To Kill a Mockingbird. I wouldn't say that Truman Capote "wrote" Harper Lee's books even if the stylometrics said so with certainty, because even if she told him the stories and he rewrote them in his own words, I'd still consider her the principal author. But it looks like Watchman and Mockingbird were penned by the same person.

Comments ( 29 )

Well, that's a relief at least. Won't have to turn the whole literary world on its head.

It would be terribly awkward to out Go Set A Watchman as a fake on a My Little Pony fanfiction site.

Or alternatively, I suppose, out To Kill a Mockingbird as having been written by Capote.

3264713

It would be terribly awkward to out Go Set A Watchman as a fake on a My Little Pony fanfiction site.

Just wait. I didn't show you the stylometrics versus Argembarger yet.

You said "I tested whether Harper Lee really wrote Watchman using Stylo" which is a very interesting way to put it. And probably because I don't know better, I have to ask: how effective (or difficult) would it be for someone to actually use Stylo as an aid to copy the style of other authors, such as for this novel, to give the result that much more authenticity?

It would be terribly awkward to out Go Set A Watchman as a fake on a My Little Pony fanfiction site.

Is it bad that I find that thought to be absolutely hilarious?

3264831

how effective (or difficult) would it be for someone to actually use Stylo as an aid to copy the style of other authors, such as for this novel, to give the result that much more authenticity?

Stylo looks at word choice, not style. It's easy for a writer to change his style; it's hard for her to use a different vocabulary. Stylo would only be helpful to prevent being outed as a fake by people using tools like Stylo. You could write something that everybody thought sounded like Hemingway, but Stylo wouldn't say it did. And if Hemingway had written an H.P. Lovecraft parody, Stylo would have a pretty good chance of still saying it was Hemingway.

3264784

* BREAKING *

Harper Lee's Editor: "Main Character Was Originally Scoutaloo"; Name Changed For Publication


Publisher Hints At 2016 Sequel: "The Killing Birdseses"

3264849

Stylo looks at word choice, not style.

Of course! I was just thinking that copying something as complex as an author's actual style of writing - especially someone whose works are well-read and thoroughly analyzed - would require some serious skill as well as an intimate knowledge of the writer's works, and that an analysis of those works using tools like Stylo would help to craft something that could add authenticity to pass a work off as another writer's efforts. It would be a tremendous amount of effort, like attempting to perfectly copy the style of a painter to create "newly discovered" works. And yet there seems to be so much controversy about this particular novel.

3264885
To Kill a Mockingbird really came under suspicion mostly because A) it was really good and B) Harper Lee never wrote anything else. This is a very unusual combination of attributes for a work to hold; generally speaking, producing a high-quality work of literature requires a great deal of practice, and you are unlikely to have your very first work to be one of the best works of the 20th century, and then ever produce anything else, ever.

The fact that she was closely associated with Capote - who was a very good writer - made for a very easy and obvious conspiracy theory - that Capote had written the book and attributed its provenance to Harper Lee for whatever reason.

Go Set a Watchmen really does get rid of a lot of the need for a conspiracy theory there - she did write other things, but they simply were never published, and she ended up taking good ideas from her lesser, discarded works and doing something else (which ended up being much better) with them. It explains why the book came out of nowhere and was good, and might also indicate why she never wrote anything else again - she had basically written about as good of a book as she might expect to write, and nothing else was nearly as good, so she'd rather be "That person who wrote that one really awesome book" than "that mediocre writer who got lucky once". A lot of public attention might have also scared her into not wanting to publish anything else, because she thought if she failed she'd be upsetting a lot of people. It also explains why she never published it.

Who knows for sure, though. It is not like other writers haven't gone crazy after writing some great book and then refused to write anything else because of the attention.

Heck, the creator of Fez kind of did the same thing; he was a person with a lot of problems, produced a surprisingly decent video game for a one-person effort (though it isn't actually as good as people say it is, and I'd say Cave Story is much better as a one-person effort), then started lashing out angrily at the world and has to date never finished anything else, cracking under the pressure of public attention and scrutiny and various hangers on and whatnot.

But the fact that Go Set a Watchman falls close to To Kill a Mockingbird stylistically does make it much less likely that Capote wrote them both, though it doesn't rule out his involvement in other ways.

3264901

Phil Fesh was cracking under pressure way before Fez ever came out. If you watch Indie Game you can see that, well, he's basically on the verge of a breakdown the entire time.

3264901 You know if I had been told 4 years ago I would be reading indepth literary analysis in a community gathered around My little pony, I would have laughed. And probably believed you; four years ago I was into the Power Puff Girls

Go Set A Watchman is a 1st draft of To Kill A Mocking Bird. It shows the evolution of the story. It is literally half-baked (To Kill A Mocking Bird is fully baked). On its own merits, it is not very good, but is a good companion-piece To Kill A Mocking Bird, showing how the finished book came to be.

The ethics of publishing Go Set A Watchman is interesting:

Harper Lee did not want this mediocre early draft published, until she went into cognitive decline. ¿Was she legally competent for consenting to publication of the story? ¿How ethical is this?

Mister Atticus Finch, who became a paragon in to Kill A mocking Bird, is a racist segregationist in Go Set A Watchman. The Atticus Finch in Go Set A Watchman is not my Atticus Finch:

I am on record of this site as writing that colts wanting to know how to be stallions should emulate Big MacIntosh. In the Human World, I tell boys who ask how to be a man to emulate Atticus Finch. Now, I must qualify this statement as the Atticus Finch in To Kill A Mocking Bird —— ¡not the Atticus Finch of Go Set A Watchman!

Shouldn't this post have been titled "Six Degrees of Francis Bacon?" :unsuresweetie:

This is awesome. Can you go beyond single-word-level analysis?
3264849 Author style may be too high-level to capture, but how about particular turns of phrase? For example, "to do so" is uncommon among writers but may be common within a particular author's corpus. Some writers hate ellipses, and others use them liberally. Searching for periods inside of quotation marks should turn up American writers.

TL;DR: Can Stylo use character combinations other than words as units for analysis?

3264861 Go Set A Chikun

:pinkiecrazy:

Interesting. Hard to tell where influence ends and ghostwriting begins, at least from this point in time, but certainly interesting.

3264861
Atticus Finch is Scootaloo's father? I like this idea, not least because it means she has a parent.

3265312 No. I checked a Scootaloo plushie at Trotcon. She has no belly-button. She must be an immortal goddess.

3265221 IIRC Stylo can check groups of up to 3 words, but if that's what you ask it to do then that's all it does. Further fancy stuff you need to program in. Punctuation is a clue, but it ignores punctuation.

You forgot to compare to Francis Bacon. I think Francis Bacon ghost-wrote TKaM and GSaW.

That's not really the controversy (at least in literary circles): everypony believes she wrote Watchman. In fact, it was initially a previous draft of Mockingbird.

The controversy is that she probably never wanted it released, particularly given the degree of contradiction in character and inability to focus on details of the editing in her advanced age, and supposedly she privately told her split-class lawyer/caretaker she wanted to publish the old draft as a sequel, and made this decision immediately after her sister died (who was the other caretaker). It seems highly unlikely her lawyer's story is true.

But at least this is hilarious.

3265409
Hay now. You should ask the next time you go checking the anatomical details of my plushies.

3265520 I know better than to check the anatomical details of your plushies. :rainbowderp:

I haven't read it, but I'm intrigued by reports that Atticus Finch goes back in time to set off an atomic bomb in Manhattan.

You might say, "So what? Watchman and Mockingbird are about the same people, in the same town. They'll have the same words just because of their names." But I checked the frequent word set used, and there are no names in it.

Actually, I was going to say that hanging out together, they probably influenced each other's writing styles the same way that the manga art produced by Seishi Kishimoto and Masashi Kishimoto (two brothers) looks virtually identical.

3265737 Given how much writers cluster together based on what country they're from and what years they wrote in, just coming from the same town at the same time is problematic.

3265977
Right but actually knowing one another and probably discussing writing together is going to exacerbate it.

Okay seriously then:

As a control, compare the works of two OTHER well-known writers who were known to hang out and write about the same stuff, but definitely DID NOT write each others' books. See if their works show the same convergence-yet-not-coincidence as Capote's and Lee's.

Who am I thinking of? Tolkien and Lewis, that's who.

I suppose I should let people know why I have some problems with this new old work by Harper Lee. Therefore, I will list out some of the reasons why I thought things might not be kosher between these two works. Allow me to warn the reader beforehand, this is liable to have many spoilers, so reader be warned.

There are two main statements made by the publisher regarding Watchman that I question. One is that this was an early draft of Mockingbird, written several years before it, and the other is that it was published without edits. There are many elements of this story that, in my mind, calls into question whether or not these statements are true.

-The books dovetail too well together. Now, on its own, this could be written off as its being first draft, but in my opinion, too much of the story was left open to character development. However, If someone were to read this book without reading Mockingbird, I think they'd be completely and utterly lost. The central pivotal point in the book, Atticus's face/heel turn has little to no impact if we do not know the particulars of the first book, nor how devastating Scout's visit to Calpurnia's home is. Another big problem is the character of Henry. He is just to convenient, being a particular friend of Jem's, almost a Dill surrogate for during the school months, to not be thought of after the fact, and seems to me to be a plug-in character for when Dill is not there. It is also very convenient that Uncle Jack Finch has retired from medical practice in Tennessee and moved back home to Maycomb.

-The characters are too underdeveloped. In addition to Atticus and Calpurnia mentioned above, Alexandra's and Jack Finch's characters are not well established, making us rely upon the characterization in the other book to understand them. Scout's reaction to Atticus's turn makes no sense to us, since within the context of Watchman we have no idea what type of pedestal Scout had put her father on. There are three brief mentions of Atticus's goodness that I recall in Watchman: a brief mention of the rape trial, and the fact that he would wait his turn in line at the Jitney Jungle no matter who was in front of him, and that the "n-word" was never uttered in Scout's hearing while growing up. If this was supposed to be a work that was independent of the other, why was there not more time allotted to the establishment of his righteousness?

There are several other things that I could mention, but will not spoil for those who wish to read this. The book comes up short in many things, that make me question whether or not they are playing this as they say, that it is an early draft of what became Mockingbird, and whether or not some other hand edited it. The flashbacks are scanty, so I don't know how an editor could tell Ms. Lee to write a book based on them, there are anachronisms that I question, and aspects regarding Capote's work Other Voices, Other Rooms that raise further questions. All told, the work in my mind is highly questionable.

Of course, I could be, and probably am totally wrong. The stylometric verdict is quite damning, indeed. And, I suppose I do have my biases, but still, this book irks me in a way that I have rarely experienced before. In the grand scheme of things, I suppose it does not matter, but it still is an itch that is quite vexing. Thank you, Bad Horse, for humoring me in this little experiment.

3266281

Who am I thinking of? Tolkien and Lewis, that's who.

Ooh. Nice idea. Though there are several CS Lewises! He's one of those rare writers who wrote different stories with different styles, as I recall. I wonder if that will have an effect.

3272188

I would say to compare The Lord of the Rings to the Narnia stories, as they're the closest in subject matter and produced during their period of closest association. Anything else would be apples-to-oranges relative to the comparison in question here (Mockingbird and Watchman to Capote's novels).

But if you want some grown-up Lewis fantasy fiction--no, skip the Perlandra series: what you want is 'Til We Have Faces.

3272202 I agree. I expect the Narnia Chronicles will have a simpler vocabulary. Still, Tolkien's very distinct, particularly in The Silmarillion. I'd say Till We Have Faces vs. The Hobbit. Or maybe "The Dark Tower". (The short story.)

Sherwood Anderson vs. William Faulkner would be another case. Or Philip K. Dick vs. Tim Powers.

Look up some of the Greenville, Mississippi authors, perhaps? Walker Percy and Shelby Foote come to mind.
And if you do Tolkien and Lewis, it might be interesting to include something like Mere Christianity and some of Tolkien's Christian apologist correspondence, perhaps. And to add to it, Charles Williams, as well.

Login or register to comment