Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Ché Biko]

I pasted the entirety of stuff I wrote (Fanfiction. [..]

If I get this right, the Guesser does not work as well when you use stuff like that:

Lyrics, lists, poems, and prose are special writing styles. This tool is unlikely to classify these texts correctly.

[Shiori]

This is amusing. If I know programming I would love to crack that wide open to see how it works.

It's not terribly sophisticated. Basically it has two word lists giving a number of relatively subject-neutral words a score proportionally to how often they occur in 'male' versus 'female' writing, adds up their scores for every time they occur, and compares the final scores. For the curious, the entire dictionary it uses is this:

[spoiler]  var DictionaryInformal = new Array();
  DictionaryInformal['actually']= -49;
  DictionaryInformal['am']= -42;
  DictionaryInformal['as']= 37;
  DictionaryInformal['because']= -55;
  DictionaryInformal['but']= -43;
  DictionaryInformal['ever']= 21;
  DictionaryInformal['everything']= -44;
  DictionaryInformal['good']= 31;
  DictionaryInformal['has']= -33;
  DictionaryInformal['him']= -73;
  DictionaryInformal['if']= 25;
  DictionaryInformal['in']= 10;
  DictionaryInformal['is']= 19;
  DictionaryInformal['like']= -43;
  DictionaryInformal['more']= -41;
  DictionaryInformal['now']= 33;
  DictionaryInformal['out']= -39;
  DictionaryInformal['since']= -25;
  DictionaryInformal['so']= -64;
  DictionaryInformal['some']= 58;
  DictionaryInformal['something']= 26;
  DictionaryInformal['the']= 17;
  DictionaryInformal['this']= 44;
  DictionaryInformal['too']= -38;
  DictionaryInformal['well']= 15;

  var DictionaryFormal = new Array();
  DictionaryFormal['a']= 6;
  DictionaryFormal['above']= 4;
  DictionaryFormal['and']= -4;
  DictionaryFormal['are']= 28;
  DictionaryFormal['around']= 42;
  DictionaryFormal['as']= 23;
  DictionaryFormal['at']= 6;
  DictionaryFormal['be']= -17;
  DictionaryFormal['below']= 8;
  DictionaryFormal['her']= -9;
  DictionaryFormal['hers']= -3;
  DictionaryFormal['if']= -47;
  DictionaryFormal['is']= 8;
  DictionaryFormal['it']= 6;
  DictionaryFormal['many']= 6;
  DictionaryFormal['me']= -4;
  DictionaryFormal['more']= 34;
  DictionaryFormal['myself']= -4;
  DictionaryFormal['not']= -27;
  DictionaryFormal['said']= 5;
  DictionaryFormal['she']= -6;
  DictionaryFormal['should']= -7;
  DictionaryFormal['the']= 7;
  DictionaryFormal['these']= 8;
  DictionaryFormal['to']= 2;
  DictionaryFormal['was']= -1;
  DictionaryFormal['we']= -8;
  DictionaryFormal['what']= 35;
  DictionaryFormal['when']= -17;
  DictionaryFormal['where']= -18;
  DictionaryFormal['who']= 19;
  DictionaryFormal['with']= -52;
  DictionaryFormal['your']= -17;
[/spoiler]
(The dictionary is split between formal/informal words as well. Negative values are more 'female.')

[Elmund Egivand]

How exactly does one determine if a word is male or female?

[Jace]

How exactly does one determine if a word is male or female?

The post directly above yours explained this...

[Louella Dougans]

i think he means, why is "him" a -73 highly "female" word, while "her" is a -9 slightly "female" word ?

[Jace]

i think he means, why is "him" a -73 highly "female" word, while "her" is a -9 slightly "female" word ?

Ah, that would depend on which linguists the makers of the program listened to. There is a tendency to use different words by different groups of people, but how it is quantified depends upon the person researching it.

[Shiori]

i think he means, why is "him" a -73 highly "female" word, while "her" is a -9 slightly "female" word ?

If they've done it with any kind of rigor, they've taken a large corpus of writing by men and women authors, counted up how many times every word occurred in each, and noted the relative frequency. If a word occurs a lot more in the text of one gender than another, it's a fairly decent predictor that if it occurs in a text, that text was written by someone of that gender, and gets a high score; if the difference is less pronounced, it'll get a lower score. How exactly that's scaled is probably arbitrary.

[Lunarisse Aspenstar]

This sounds like a stereotypical piece of clap trap because the writings it probably used wouldn't pass the "Bechdel test" .  Because it sounds like it used emo romance novels to find women writers if you look at the list and think of how the words would be used.

Genre: Informal
  Female = 1820
  Male   = 421
  Difference = -1399; 18.78%
  Verdict: FEMALE

Genre: Formal
  Female = 959
  Male   = 277
  Difference = -682; 22.41%
  Verdict: FEMALE

to follow che and arista 

She returned barefoot to the kitchen to bake cookies and to wonder why he wouldn't date her because she is a commoner.  Who knew his mind?  If only she had noble birth! If only!  She knew she was everything he sought, elegant and beautiful and smart.  Why did it matter to him that she was a commoner?

[Arista Shahni]

*Arista silently made a cup of tea for Lunarisse, knowing her mind.  She had a deep suspicion as to why Che was in the kitchen, however, though quickly dismissed the random staring into the refridgerator with no particular goal in mind as STEREOTYPICAL MALE BEHAVIOR.  >

[Mizhara]

Oi. No mocking one of our most sacred rituals of manhood, misandrist shitlady!

Hmm. Shitlady doesn't work as well as shitlord, does it?

[Morwen Lagann]

*Arista silently made a cup of tea for Lunarisse, knowing her mind.  She had a deep suspicion as to why Che was in the kitchen, however, though quickly dismissed the random staring into the refridgerator with no particular goal in mind as STEREOTYPICAL MALE BEHAVIOR.  >

Confirming it's not a successful trip to the kitchen until all the cold air has been let out of the fridge and/or freezer.

[Kala]

...I do that 
In my defence, the freezer sometimes doesn't close properly >.>

[Nicoletta Mithra]

i think he means, why is "him" a -73 highly "female" word, while "her" is a -9 slightly "female" word ?

If they've done it with any kind of rigor, they've taken a large corpus of writing by men and women authors, counted up how many times every word occurred in each, and noted the relative frequency. If a word occurs a lot more in the text of one gender than another, it's a fairly decent predictor that if it occurs in a text, that text was written by someone of that gender, and gets a high score; if the difference is less pronounced, it'll get a lower score. How exactly that's scaled is probably arbitrary.

The problem here is in first establishing what a 'male' and a 'female' author/writing style are: After all, it's supposedly about gender and not sex. And the former is an even more supple concept than the latter (and knowing my biology the latter isn't as clear cut as one usually thinks already).

In the end it turns out to be a bit circular: You first define what is a 'male' and a 'female' writing style, then you count word frequencies, and as a last step use correlations in word frequencies to writing style to 'predict' writing style. I don't think that much is won with this, safe maybe that you don't need to read the text anymore to assign a 'male' or 'female' style.

 It's more a descriptive statistics than genuine insight that this 'test' gives you - which might be useful for research, though.

[Vieve]

Apparently I'm most often a guy.

1)  Test #1:  EVE Fan Fiction (specifically, "Flotsam in Everyshore (2)")

Genre: Informal
  Female = 2456
  Male   = 3386
  Difference = 930; 57.95%
  Verdict: Weak MALE
Weak emphasis could indicate European.

Genre: Formal
  Female = 2707
  Male   = 2060
  Difference = -647; 43.21%
  Verdict: Weak FEMALE
Weak emphasis could indicate European.

2) Test #2, Excerpt from Current Fiction Project.

[spoiler]

"Oh, good.  You are all awake." Ambru said. He'd brought a lantern. Rather, he and the four men in the identical wine-colored coats who were walking across the deck behind him all had lanterns. Through a new, rather strikingly symmetrical rent in the fog, Gev could see the side of the ship they'd come from, and the weighted rope ladder they'd dropped over it. A ship. Not the ship they'd been steering for, and certainly not any of the men they'd left behind on that ship, because for one, they didn't have any men left on their ship, and for two, their cutter was most certainly not a heavy Tannaliri cruiser.

It had a name -- Relentless-- that said that its captain had already earned his or her own surname, and the odds were unlikely the honor'd been awarded by the Mariner Guild's esteemed Tannalir-based schools of drydock etiquette and long afternoon naps. Gev took a moment to marvel at how it was nothing short of miraculous that their boat had been found in the vast southern ocean beneath the trackless stars right when they needed help.

But just a moment.

"You've been drugging me for a while…" he guessed.

"For the past month, yes," Ambru agreed.

"….because you knew I wouldn't have taken the fucking job if I'd smelled the slightest thing off about…" The sudden 'ow' that escaped Gev's lips surprised him by how soft it was, because he'd been on the verge of screaming.

One of the anonymous men with Ambru was now holding the bulk of Gev's voice in his pinched fingers. It occurred to Gev as he rubbed his throat and stared groggily at him, that the man wasn't quite as anonymous as the others: the effort of catching and holding a voice seemed to be making him appear more distinct, more memorable … and the other men less obviously flesh and blood. The illusory men re-stabilized when their master opened his fingers.

Gev cleared his throat. "I'm going to kill you both."

"Certainly," Ambru said. "Come."

[/spoiler]

Genre: Informal
  Female = 582
  Male   = 682
  Difference = 100; 53.95%
  Verdict: Weak MALE
Weak emphasis could indicate European.

Genre: Formal
  Female = 288
  Male   = 485
  Difference = 197; 62.74%
  Verdict: MALE

3)  Test #3 - A sample of work writing.   (Trust me, you don't want to read it: it's pretty tedious.)

Genre: Informal
  Female = 385
  Male   = 1507
  Difference = 1122; 79.65%
  Verdict: MALE

Genre: Formal
  Female = 830
  Male   = 981
  Difference = 151; 54.16%
  Verdict: Weak MALE
Weak emphasis could indicate European.

[Jace]

It's more a descriptive statistics than genuine insight that this 'test' gives you - which might be useful for research, though.

This was the original intent of this sort of research. It was done for the sake of information the provide an interesting stepping stone to further research or possible ideas, not meant to be something terribly innovative in and of itself.