Thread by @katrinaltrnsl8r

In Japanese 101, you were probably taught that 天才 means "genius".
But "genius" in English doesn't function the same way that 天才 does in Japanese.
Let's talk about meaning, collocation, and the wonderful world of corpora!🧵#l10n

(Yes, this is another short essay that's going to ruin subtitles for you, because once you start seeing it, you won't be able to stop seeing it.
It just can't be helped. You'll never forgive me.)

Let's breakdown 天才 first with a definition from Japanese dictionaries:
天性の才能、生まれつき備わった優れた才能、または生まれつき優れた才能を備えた人物
Natural, innate talent or aptitude; or a person who was born possessing innate talent or aptitude.

Now, let's look at "genius" from an English dictionary:
a single strongly marked capacity or aptitude
extraordinary intellectual power especially as manifested in creative activity
a person endowed with extraordinary mental superiority

Pretty similar at first glance, but take note of the emphasis on intellect in the English definition.
This is the main difference between 天才 and "genius" -- and why you may want to consider other English words when 天才 means "remarkable talent" instead of "very smart".

Genius is also less grammatically versatile than 天才 , especially when it's used to describe natural talent.
A talented painter is more likely to be described as an "artistic genius" than a "genius artist".
This less common construction may sound strange to a native EN speaker.

But how do you know what sounds "natural" to native speakers? Every language is always changing, and one person can't be the authority on what's "natural".
Luckily, we can turn to language data and statistics to help!
Meet: the corpus.

A corpus is a collection of linguistic data, either written text or transcriptions of recorded speech, that can be searched, analyzed, or studied to understand more about how language works.

They have lots of interesting uses, like creating dictionaries, training speech recognition, and developing machine translation and AI.
We can use a corpus to find out when, where, and how certain words or phrases are used in written or spoken language.

There are lots of corpora out there, but for this example, I'm using the Corpus of Contemporary American English.
(fraze.it is a simple free tool! You can also create a free account on english-corpora.org and search a ton of different corpora.)

Let's do a search for the word "genius".
Look at all that information! Take a moment to soak it in before we dive in deeper.

At the top left, we have a chart showing the frequency of the word across text types. Below it, definitions, as well as frequent synonyms.
On the right, we have the most frequent topics the word is used with, as well as its most common collocates.
So... what's a collocate?

A collocate is a word that appears alongside another word too frequently to just be random.
Studying a word's most common collocates can help us understand how a word is used.
Here's the collocates for the word "smoky". How many different meanings can you infer from this list?

Clusters are the most frequent groups of words that come before and after the searched word.
You can expand these results to see even more clusters and their frequency!

From these clusters, we can see that from the 21,000 instances of "genius" in the corpus, the word "genius" isn't often used as an adjective before a noun, and is more likely to be used as a noun after a descriptive adjective.
Musical genius. Comic genius. Artistic genius.

To summarize:
🧠天才 doesn't always mean "genius". Consider the meaning in context and choose an appropriate English equivalent.
📝Corpora can be used to study word order and collocation.
🌟I never want to see the translation "genius singer" ever again.

ASIDE: WHY CAN'T LANGUAGE JUST BE EASY
Meaning is complicated, which is why we have lots of different words that express similar meanings.
But just because the thesaurus says words are synonyms, doesn't mean they're interchangeable.

"That's a lie!"
"That's a falsehood!"

There are two branches of linguistics that study the meaning of words -- semantics, which studies the meanings of words, phrases, and sentences; and pragmatics, which studies how languages is used in different contexts and how meanings are produced through language.

This is why it's so hard for computers to be 100% accurate with machine translation -- there's a lot more going on than just the dictionary meanings of words, and computers aren't (yet) able to understand the contexts that affect the meaning being expressed. One day, maybe!

ASIDE: WHAT DO YOU EVEN DO WITH A CORPUS?
There's a whole lot of cool things you can do, from searching word usage in news articles to seeing how many times a specific one-liner has been used in Hollywood movies.
It can be helpful when writing in your non-native language, too.

I linked a few up in the thread, so if you're linguistically-minded, definitely open one up and play around with it!
There are corpora in hundreds of languages, and some are more specific than others -- check out this corpus of language used in soap operas!

ASIDE: ARE THERE JAPANESE CORPORA?
Absolutely! They're a bit more difficult to access if you don't speak Japanese, as some require registration or payment.
Here I'm using the free Shonagon Contemporary Japanese corpus to find results on 天才的.