Hacker News Reader

Comment by acchow | Hacker News Reader

> “Linux Kernel Networking Subsystem”

How do they say this in Japanese? In Chinese (Mandarin/Cantonese)?

In Chinese, "Linux内核网络系统", literally "Linux Kernel Network System". Unlike Japanese and Korean, the Chinese language generally uses full meaning-to-meaning translations of technical terms as much as possible, and only occasionally uses phonetic nouns (e.g. 安卓 "Anzhuo" for Android) or English words (e.g. "Linux").

Chinese does, however, use phonetic transcriptions for proper names of people and places. This leads to some very unfortunate contortions, as the Chinese language is much less suited to phonetic transcription as compared to syllabary-based languages like Korean or Japanese. There's no "one obvious way" to map a foreign name onto the Chinese sound-set, and even when you do, there's likewise no obvious way to choose the corresponding characters (and you must be cautious to avoid characters with unwanted meanings or connotations).

Sounds like this won't be news for you, but for others in the thread, the Chinese sometimes play around a lot with puns and rhymes in transliterated names.

Here's a terrific twitter thread about Chinese nicknames for NBA players: https://twitter.com/i/moments/993587263283048449?lang=en

An example from the thread[1]:

Stephon Marbury - 马政委 "Marburyist Political Commissar"

This is a pun on "Marxist Political Commissar," since in Chinese both "Marbury" and "Marx" are spelled with the same first character "ma" ("horse").

[1] https://twitter.com/nick_kapur/status/993571262327910400

Chinese in many respects is a more explicit language. Variables are generally required to "wear their types on their names". All the disease names contain the word disease (imagine if we must speak of "rabies disease" and "AIDS disease" as opposed to "rabies" and "AIDS"). Country names in general contain the word country, so you can tell from the name that France (lawful country) is a country whereas Chicago is not - something left implicit in English. (There are exceptions of course)

Although that is somewhat of a simplification, it is certainly more true of Chinese than of English.

The written language is also typed to a certain degree. Almost all disease words use a common radical (e.g. 癌症 for cancer), for example. As a more extreme case, every chemical element name in Chinese incorporates a radical that reflects the chemical's basic state - liquid (water), gaseous (air), metallic (gold), or non-metallic (earth) - symbolically connecting the old elemental system with modern chemistry.

The vast majority of country names in Chinese do not contain 国 but are rather purely phonetic transcriptions. Only a handful of countries have 国 in the common name.

That's true, but I think the full name usually has, just as the full names of Beijing Shanghai Tianjin etc contain 市 even though the short names doesn't.

And the short names of England, France, America, Germany, Korea, Thai, and China itself contain 国. This seems to suggest that there was a naming convention initially being followed.

Better examples than country names for illustrating this difference between Chinese and English include disease names (mentioned above), fish names, bird names, tree names, flower names, mountain names, river names …

I recently learnt the Chinese name for San Francisco: Old Gold Mountain. From the gold rush days.

Similarly, New Gold Mountain is Melbourne Australia near where the Australian gold rush started soon after.

With phonetic systems like Japanese and Korean, terms like this regularly get represented as-is in the script. Sometimes they don't get translated, as you can see plenty of jargon in this Korean wiki page on C++. [0]

One example is 'computer' -> 컴퓨터 in Hangul. It's still 'computer', just pronounced with the Korean pronunciation rules.

[0]: https://namu.wiki/w/C%2B%2B

Something like

リニキス　カーネル　ネットワークイング　サブシステム

It's basically English pumped though a lossy algo. I hate katakana and hope it dies soon.

I think that the proper term is "Linuxカーネルネットワークサブシステム" meaning "Linux Kernel Network Subsystem".

What is katakana?

Japanese uses three distinct character sets in writing: Hiragana (a syllabary), Katakana (another syllabary) and Kanji (Chinese characters). Hiragana and Kanji are usually used together to write native Japanese words, whereas Katakana is used almost exclusively for loanwords.

Katakana syllables are either bare vowels, or consonant+vowel (with few exceptions), meaning that transliterated words must be contorted to fit the syllabary (the "lossy algo" referred to by the parent comment). So, something like "kernel" in English becomes kaa-ne-ru, and "subsystem" becomes "sa-bu-shi-su-te-mu".

Katakana [1] is one of the two syllabary writing systems in Japanese, with Hiragana [2] being the other. IIRC katakana is the one where every syllable starts with a consonant and ends with a vowel, so transliterated loan words with consecutive consonants end up growing vowels in the middle. They also have kanji (characters), and various romanization systems (https://en.wikipedia.org/wiki/Romanization_of_Japanese). It's really amazingly inefficient.

[1] https://en.wikipedia.org/wiki/Katakana [2] https://en.wikipedia.org/wiki/Hiragana

Why is everyone posting squares?

Unicode characters probably aren't all rendering on your end due to your font.

The Japanese first adapted writing from Chinese; this adapted system is called kanji and is highly similar to, and often homographic with, Chinese hanzi. However, Japanese and Chinese have very different phonological structure: Chinese is mostly comprised of monosyllabic words, whereas Japanese has lots of polysyllabic words. This makes it more difficult to do phonetic transcription in Japanese kanji than in Chinese hanzi.

The Japanese got around this by simplifying the script into a syllabary (every character represents roughly a syllable, or more often, a consonant-vowel pair). They did this twice: one of these syllabaries is hiragana, and the other is katakana. In modern usage, katakana is used largely for phonetic transcription, such as transliteration of foreign words and names, or onomatopoeia, whereas hiragana is used for writing out Japanese words.

As others have pointed out, Japanese has a rather constrained phonological system, so a word like strengths cannot be represented directly but rather more like "su-to-re-n-ge-tsu." Of course, this feature isn't limited to Japanese; it's how an island called "Christmas" gets transliterated to "Kiritimati", just like its parent "Kiribati" is the local pronunciation of "Gilbert." While people think it's annoying, it's largely because they haven't faced languages with challenging transciptions into Latin script. There's a reason why there's a plethora of transcriptions of "مُحمّد‎" after all.

Thank you for explaining this to me! (And thank you to everyone else's comments too of course). This is fascinating.

> There's a reason why there's a plethora of transcriptions of "مُحمّد‎" after all.

There's a plethora of transcriptions into (more) phonetic alphabets because there are a plethora of regionalized pronunciations [1]. And there are a plethora of pronunciations because Arabic uses an impure abjad [2]. Since the vowels are not always, exactly or uniformly specified in writing, unspecified behavior leads to varying results in each compiler.

[1] https://en.wikipedia.org/wiki/Muhammad_(name)#Transliteratio... [2] https://en.wikipedia.org/wiki/Abjad#Impure_abjads

Reply on news.ycombinator.com