These "gendered styles" they found are "what men and women are paid to write about". We already know this, for the breakdown of jobs by sex: View attachment 1182901
You clearly didn't bother to read the study before chimping out, because if you did you'd know that they controlled for that:
For each genre we used precisely the same number of male- and female-authored documents (Fiction: 123 male documents, 123 female documents; Nonfiction: 179 each, including Nat Science: 2 documents each; Appl. Science: 13; Soc. Science: 60; World Affairs: 34 Commerce: 4; Arts: 31; Belief/Thought: 18; Leisure: 17). Documents were chosen in each genre by using all available documents in the smaller (male or female) set and randomly discarding the surplus in the larger set. No single author wrote more than 6 documents in this corpus.
I don't know what you're so upset about anyway. The tendencies they found don't favor either gender, they aren't immutable or absolute, and they don't prevent men or women from being effective writers in either genre. It's not like they're claiming that women are peabrains who can't do science or that men are robots who can't relate to people. They're just observing general tendencies in writing.
I used to play Final Fantasy XI, a rather grindy MMORPG, which had two single-gender races: mithra (catgirls) and galka (big bearded muscular men with tails who undergo some vague supernatural process of reincarnation where young galka just kind of mysteriously show up one day after an older galka dies). It was a well-known fact that the overwhelming majority of mithra were actually played by men. (The galka were played by men too, with the main difference between the mithra and galka player populations being that the galka players were all cripplingly retarded.) So I would agree with this.
That "Gender Analyzer" is fucking exceptional. Here's a GIF that demonstrates why:
Here's what you're seeing in those three frames:
The tool rating its own description from the site.
The tool rating that same description, except every word that doesn't affect the score is replaced with the word "nigger"
The tool rating the four words from the description that actually affect the score.
While the actual algorithm is hidden behind an API, I was able to determine that it uses a 0-100 scale for its ratings where 0 is most feminine and 100 is most masculine. It defaults to 0 so any text that doesn't contain a "masculine" word will be marked as most feminine (0). Any text that contains a single "masculine" word without any "feminine" words will be marked as most masculine (100).
After determining that, I hooked into the API and fed it a list of words from the study I mentioned in my last post. Of the 457 words in that dataset, Readable only acknowledged 12 "masculine" words and 13 "feminine" words:
As
Ever
Good
If
In
Is
Now
Some
Something
The
This
Well
Actually
Am
Because
But
Everything
Has
Him
Like
More
Out
Since
So
Too
Each word has its own particular value, so rating one word from each set won't necessarily result in a neutral score of 50. However, there are three pairs of words that do balance perfectly (e.g. now and has). I used these to run the following tests:
now now now has: 75 now now has: 67 now has: 50 now has has: 33 now has has has: 25
Do the numbers in these results remind you of anything? They should, because they're percentages of basic fractions: 3/4 = 75%, 2/3 = 67% (66.66% rounded up), 1/2 = 50%, etc. This means that Readable's "Gender Analyzer" algorithm is just (Masculine Points)/(Masculine Points + Feminine Points).
In case it isn't obvious, that algorithm is a joke. It doesn't account for the position of words within sentences, the length of the text, or anything else beyond its tiny pool of "masculine" and "feminine" words. Whoever made that shit should be embarrassed.
That "Gender Analyzer" is fucking exceptional. Here's a GIF that demonstrates why: View attachment 1186694
Here's what you're seeing in those three frames:
The tool rating its own description from the site.
The tool rating that same description, except every word that doesn't affect the score is replaced with the word "nigger"
The tool rating the four words from the description that actually affect the score.
While the actual algorithm is hidden behind an API, I was able to determine that it uses a 0-100 scale for its ratings where 0 is most feminine and 100 is most masculine. It defaults to 0 so any text that doesn't contain a "masculine" word will be marked as most feminine (0). Any text that contains a single "masculine" word without any "feminine" words will be marked as most masculine (100).
After determining that, I hooked into the API and fed it a list of words from the study I mentioned in my last post. Of the 457 words in that dataset, Readable only acknowledged 12 "masculine" words and 13 "feminine" words:
As
Ever
Good
If
In
Is
Now
Some
Something
The
This
Well
Actually
Am
Because
But
Everything
Has
Him
Like
More
Out
Since
So
Too
Each word has its own particular value, so rating one word from each set won't necessarily result in a neutral score of 50. However, there are three pairs of words that do balance perfectly (e.g. now and has). I used these to run the following tests:
now now now has: 75 now now has: 67 now has: 50 now has has: 33 now has has has: 25
Do the numbers in these results remind you of anything? They should, because they're percentages of basic fractions: 3/4 = 75%, 2/3 = 67% (66.66% rounded up), 1/2 = 50%, etc. This means that Readable's "Gender Analyzer" algorithm is just (Masculine Points)/(Masculine Points + Feminine Points).
In case it isn't obvious, that algorithm is a joke. It doesn't account for the position of words within sentences, the length of the text, or anything else beyond its tiny pool of "masculine" and "feminine" words. Whoever made that shit should be embarrassed.
Wow, thanks. That is fascinating. Your autism (I mean that in the kindest way possible) could be utilized to bring you to wonderful places in life, if you want to.
I mean, probably doesn't help I'm surrounded by boys 24/7 and they're not good boys for the most part. Also haven't been around girl friends since high school, yet they weren't all that feminine in mannerisms so I never got into make-up and frills and laces (but I like plushies, do they count as accessories?). Been accused of not being feminine in my speech/mannerisms but I'm a hopeless romantic cooing over romance and babies.
Guess tomboys can't be girls after all in CURRENT YEAR.
EDIT: Oh, so now there's some estrogen in this post. This analyzer must be relying on buzzwords that're "gender specific".
It does, but probably not the words that you'd suspect. Here's some Chad-tier autism that explains how it works, and here are the words it's picking up in your particular example:
I mean, probably doesn't help I'm surrounded by boys 24/7 and they're not {GOOD} boys for {THE} most part. Also haven't been around girl friends {SINCE} high school, yet they weren't all that feminine {IN} mannerisms {SO} I never got into make-up and frills and laces ({BUT} I {LIKE} plushies, do they count {AS} accessories?). Been accused of not being feminine {IN} my speech/mannerisms {BUT} I'm a hopeless romantic cooing over romance and babies.
Guess tomboys can't be girls after all {IN} CURRENT YEAR
It does, but probably not the words that you'd suspect. Here's some Chad-tier autism that explains how it works, and here are the words it's picking up in your particular example:
That "Gender Analyzer" is fucking exceptional. Here's a GIF that demonstrates why: View attachment 1186694
Here's what you're seeing in those three frames:
The tool rating its own description from the site.
The tool rating that same description, except every word that doesn't affect the score is replaced with the word "nigger"
The tool rating the four words from the description that actually affect the score.
While the actual algorithm is hidden behind an API, I was able to determine that it uses a 0-100 scale for its ratings where 0 is most feminine and 100 is most masculine. It defaults to 0 so any text that doesn't contain a "masculine" word will be marked as most feminine (0). Any text that contains a single "masculine" word without any "feminine" words will be marked as most masculine (100).
After determining that, I hooked into the API and fed it a list of words from the study I mentioned in my last post. Of the 457 words in that dataset, Readable only acknowledged 12 "masculine" words and 13 "feminine" words:
As
Ever
Good
If
In
Is
Now
Some
Something
The
This
Well
Actually
Am
Because
But
Everything
Has
Him
Like
More
Out
Since
So
Too
Each word has its own particular value, so rating one word from each set won't necessarily result in a neutral score of 50. However, there are three pairs of words that do balance perfectly (e.g. now and has). I used these to run the following tests:
now now now has: 75 now now has: 67 now has: 50 now has has: 33 now has has has: 25
Do the numbers in these results remind you of anything? They should, because they're percentages of basic fractions: 3/4 = 75%, 2/3 = 67% (66.66% rounded up), 1/2 = 50%, etc. This means that Readable's "Gender Analyzer" algorithm is just (Masculine Points)/(Masculine Points + Feminine Points).
In case it isn't obvious, that algorithm is a joke. It doesn't account for the position of words within sentences, the length of the text, or anything else beyond its tiny pool of "masculine" and "feminine" words. Whoever made that shit should be embarrassed.
Even amongst weeb fanfic culture you can find the difference between MAB & FAB.
Lady weebs tend to write very expressively with delicate words and small details. They'll sound "anime", but like a shoujo anime rather than a shonen. Fanfics tend to focus on feelings and relationships, but ofc can be anything. Usually interested in deep character development.
Male weebs tend to go for blunter language and copy the style of light novels more. More beige prose and focus on actions, along with favoring porn or action oriented genres. Occassionally throw in philosophical musings foe either actual discussion or most likely lip service like shonen sometimes does. Ive seen male writers use purple prose too, and it tends to come out when said action or sex scenes occur, or when an object of their weeb crush appears. They tend to focus on moving the plot along and action, and do extra worldbuilding (basically what the study and other users have said).
This is all on a spectrum of course. Most weebs write somewhere in between and go back and forth. Things like Mary Sues and other bad writing topics are universal.
The thread where Jamias played (read) Ana Valens' awful awful game gives a good example of this. I could tell Ana was trans because no women ever thought like the main character did, and the entire thing focused on fetish/jerking off more than it did character. To be fair, Ana wrote all the characters like awkward robots who never met another human, which is a mistake anyone could make.
When someone starts using lots of emojis and or laughs, that tells me female more than anything. Actual writing style, without all that, I find it can be quite difficult to tell.
