For these Ngrams, I used wildcard search for the first, and comparing the occurance of a phrase in two languages in the second.
Unsurprisingly, the most frequent word is the name of the main character, Raskolnikov.
The next function I explored was creating a line graph in the "trends" section of the interface. I used the graph to plot not only the frequency of
4 characters' names, but when in the novel that those names are most and least significant. Sonia, the second most mentioned character in the book,
doesn't hit her stride in popularity until the eighth segmented section of the book.
Lastly, I was curious of how the book overall contextualized the two most significant characters, Raskolnikov and Sonia.
Russian literature is not always known for its fleshed out and open-minded portrayal of female characters, and as much as
Dostoyevsky may have tried, his writing still has biases. As you can see, the words that most often surround Sonia are one-dimensional.
“The reality of a world without bees is a scary one indeed. Without them, where are our flowers? Our sweet honey? Our beautiful nature as we know it?
Honeybees are kind, gentle, smart, and so hardworking. We owe it to them to clean up our environmental messes.”
Both sentimood and text2data rated this as strongly positive, but only sentimood acknowledged the word “scary” as negative.
“The little room into which the young man walked, with yellow paper on the walls, geraniums and muslin curtains in the windows,
was brightly lighted up at that moment by the setting sun.”
Both sentimood and text2data identified this sentece as completely neutral
“With a sinking heart and a nervous tremor, he went up to a huge house which on one side looked on to the canal, and on the other into the street.
This house was let out in tiny tenements and was inhabited by working people of all kinds—tailors, locksmiths, cooks, Germans of sorts,
girls picking up a living as best they could, petty clerks, etc. There was a continual coming and going through the two gates and in the two courtyards of the house.
Three or four door-keepers were employed on the building. The young man was very glad to meet none of them, and at once slipped unnoticed through the door on the right, and up the staircase.
It was a back staircase, dark and narrow, but he was familiar with it already, and knew his way, and he liked all these surroundings: in such darkness even the most inquisitive eyes were not to be dreaded.”
Sentimood analyzed this as pretty positive, while text2data gave it the most negative score that it can- text2data also flagged the word “Germans” as negative, interestingly.
“And here... I am again on the same errand,” Raskolnikov continued, a little disconcerted and surprised at the old woman’s mistrust.
“Perhaps she is always like that though, only I did not notice it the other time,” he thought with an uneasy feeling.”
Text2data considered this text semi-negative, and Sentimood considered it to be semi-positive.
"it's impossible to be unhappy in this weather! There is not a single drop of rain, no storms, nothing but a breeze coming
through and just the perfect amount of heat: neither burning nor freezing."
Both sentiment analyzers rated this text as negative.
"This honey is unbelievable. In fact, it is so strong and sweet that it drowns the rest of the cake in a permanent, sugary coat.
Now the cinnamon undertones combine with the honey in this grandiose flavor puddle!
What will the bride think of this, on what is supposed to be her happy wedding day?
she should be in love, dancing and smiling with her new husband in front of all her friends and loved ones.
Now, she'll be firing the cake baker on the spot. Forgive me for my confidence in myself, that I had hoped would be properly placed!"
Both programs identified this text as strongly positive, rather than the slightly negative I perceive it to be.
This original text is the first line of the Part II of Master and Margarita. It's widely regarded as the absolute best Soviet novel.
Initial text:
“Follow me, reader! Who told you that there is no such thing as real, true, eternal love? Cut out his lying tongue!”
Became this via Google:
Follow me reader! Who told you that there is no such thing as the present,
really eternal love? Cut out his lying tongue!
Not terrible, but mistranlated the "real" and "true", while throwing "present" in there as well.
translated by Bing:
"Follow me, reader! Who told you that there is no such thing as real,
true, eternal love? Cut out his tongue!"
Bing did a surprisingly good job with this!
This next text was intended to be a bit more of a challenge. The poem's english title is "Incantation by Laughter"
by V. Khlebnikov. In Russian, every single word in the poem is derived from the root of the word "to laugh",
turning it into made-up adjectives, nouns, participles, and so on. The key to translating this is that every
word in the poem MUST have the same root as whatever "laugh" is in that language. Both translators failed in this regard.
This text:
O, laugh, laughers!
O, laugh out, laughers!
You who laugh with laughs, you who laugh it up laughishly
O, laugh out laugheringly
O, belaughable laughterhood - the laughter of laughering laughers!
O, unlaugh it outlaughingly, belaughering laughists!
Laughily, laughily,
Uplaugh, enlaugh, laughlings, laughlings
Laughlets, laughlets.
O, laugh, laughers!
O, laugh out, laughers!
Became this via Google:
Oh, laugh, funny!
Oh, laugh, funny!
You laugh with laughter, laugh with laughter
Oh laugh laughing
Ah, laughable laughter, laughter of laughing mockers!
Oh, make him laugh with laughter over laughter!
Laugh, laugh
Uplaugh, laugh, laugh, laugh
Laugh, laugh.
Oh, laugh, funny!
Oh, laugh, funny!
and this via Bing:
Oh, laugh, laugh!
Oh, laugh, laugh!
You who laugh with laughter, you who laugh ridiculously
Oh, laugh with laughter
Oh, laughter - laughter laughing laughter!
Oh, I didn't rant it, mourning the laughter!
Laugh, laugh,
Uplaugh, enlaugh, laughter, laughter
Laughter, laughter.
Oh, laugh, laugh!
Oh, laugh, laugh!
I was curious if google or bing would be better or worse at translating colloquial language, rather than literary.
I tried to trip the translators up by using a ton of negatives in a row (which is what you're supposed to do in Russian!)
This sentence:
I ain't never seen nothing like that, never.
Became this via google:
I have never seen anything like it, never.
and this via bing:
I've never seen anything like it, ever.
Again, neither of them got it quite right.
Lastly, I threw this sentence in as a hopeful trick- and it didn't work! The word for butter
in Russian, масло, is also the same word for olive oil and oil paint. I was hoping that the ambiguity would
give pause, but I suppose "butter" is more often used than the other two definitons of the word.
This sentence:
I can't believe I paid fifty dollars for butter last week.
stayed the exact same in both translating programs.
My second experiment was much simple: can I train this network to recognize stripes and patterns as distinct from solidly colored
surfaces? This one turned out great; the network had a 100% success rate and even categorized patterns beyond just stripes.
