Wiktionary talk:Votes/2016-09/Renaming transliteration

Discussion

I think romanization is also restrictive, because not all the characters used are necessarily actually "Roman". Also, I think transcription is inaccurate even in the cases where the transliteration is phonetic, because to me "transcription" implies that the written word is completely ignored and the sounds of the spoken word are transcribed. People claim that adding unwritten phonetic details to a transliteration somehow makes it not a "transliteration", but in my opinion anything that is meant to be a representation of a written word in another script is a transliteration even if it contains additional information not present in the original script. --WikiTiki89 14:35, 13 September 2016 (UTC)

I think the transliteration systems vary in whether they count as transliterations (one-to-one representation of a native grapheme with a Latin alphabet grapheme) or transcriptions (portrayals of pronunciation).

The automatic transliteration of Ancient Greek is a pretty pure transliteration, aside from removing macrons from the transliteration of eta and omega when they have circumflex: η ω (ē ō) → ῆ ῶ (ê ô), not ē̂ ō̂. And of course smooth breathing is represented by zero at the beginning of a word: ἀ (a). In all (?) cases, you can generate the original spelling from the transliteration. If it were a transcription, the representation would change depending on the stage of the language, since each has a different phonological system: Homeric, Attic, various types of Koine, and Byzantine.

Russian transliteration is very similar: it doesn't show vowel reduction, the contrast between alveolo-palatal and retroflex consonants, or distinguish /j/ from palatalization. Calling it a transcription would be misleading.

On the other hand, the Arabic transliteration is more of a transcription. The letter ā represents at least four different graphemes, as in آسِيَا (ʔāsiyā) and فُصْحَى (fuṣḥā), هٰذَا (hāḏā). There is no way to convert from an Arabic transliteration with ā to the original spelling, because there is no one-to-one correspondence.

Romanization would make sense as a cover term for transliteration and transcription, when the result is a Latin-alphabet-based representation, but transcription would be misleading in the case of Ancient Greek and Russian. — Eru·tuon 20:30, 13 September 2016 (UTC)

To clarify my point, all those systems are different, but they can all still be called "transliterations" even if they aren't 100% "pure" transliterations. They are still fundamentally different from transcriptions, which would be IPA and other things that are purely phonetic and ignore the written word completely. --WikiTiki89 20:47, 13 September 2016 (UTC)

@Wikitiki89 I guess it makes more sense to me to call something that is essentially a system based on graphemes a transliteration, but to call something that is based on phonemes or phones (where these differ from graphemes) a transcription, even if it doesn't use the IPA. Meaning, you ask the question, which is (a particular symbol in) the system representing: graphemes, phonemes, or phones? If graphemes, then transliteration; if phones or phonemes, transcription. If the transliteration or transcription or romanization (or whatever) system essentially corresponds to the IPA symbols that would be used to make a phonetic or phonemic transcription of the language, then it, like the IPA transcription, is also transcription-like. It's just using more accessible or easily understandable symbols in place of IPA symbols. Maybe these definitions of transliteration and transcription are just my own, though. — Eru·tuon 22:23, 9 October 2016 (UTC)

@Erutuon: I guess what I'm saying is that if you are "romanizing" written text, then it counts as a transliteration regardless of what system you use; and if you are "romanizing" live speech or an audio recording, then it's a transcription, again regardless of the system you use. --WikiTiki89 15:43, 10 October 2016 (UTC)

Options that involve renaming the terms without all the parameters

Renaming the tr= parameter is an enormous task that probably won't be done any time soon. I think options should be provided to rename the terms but keep the parameters as-is. Benwing2 (talk) 06:34, 23 September 2016 (UTC)

I'd suggest a two-part vote: (1) decide whether to rename the term, and if so, which term; (2) if renaming the term, decide whether to rename the parameters (and templates such as {{xlit}}). Benwing2 (talk) 06:36, 23 September 2016 (UTC)

@Daniel Carrero: I think we should take this into consideration before starting the vote. --WikiTiki89 15:24, 23 September 2016 (UTC)

 Done --Daniel Carrero (talk) 18:52, 26 September 2016 (UTC)

Support vote removed

My support vote was going to be this.

1. Support --Daniel Carrero (talk) 15:07, 23 September 2016 (UTC)

   I support renaming all transliteration policies (Category:Transliteration policies) to "romanization" and using "r=" or "rom=" as the romanization parameter in templates. "Romanization" is widely used in entries, both in the POS header and in the category name (see suru#Japanese). These entries are categorized in Category:Romanizations by language. We have Category:Transliterations by language also, but their only purpose is serving as middle categories to place the "romanization" categories themselves, with 2 Chinese exceptions I'll mention below. (example: Category:Japanese romanizations is inside Category:Japanese transliterations, which is otherwise empty; the same is true for Egyptian, Sundanese and others)

   Category:Chinese romanizations, Category:Gothic romanizations and Category:Japanese romanizations are the only ones with thousands of entries, if we count the Chinese entries that are placed in subcategories: Category:Mandarin pinyin and Category:Cantonese jyutping.

   Category:Chinese transliterations has two children for transliterations that are not written in Latin script, which can be explained in pages that are not called either WT:Chinese transliteration or WT:Chinese romanization. They are:

   • Category:Dungan xiao'erjing (2 entries) -- the Xiao'erjing system to write Chinese in Arabic script. Dungan is a language.
   • Category:Mandarin zhuyin (41 entries + 1 user page) -- the bopomofo symbols, a phonetic system for Chinese. (the entries are just the symbols, as opposed to words written using them -- like Category:IPA symbols)

   --Daniel Carrero (talk) 15:07, 23 September 2016 (UTC)

When the vote re-starts, I plan to support "romanization" again. --Daniel Carrero (talk) 18:53, 26 September 2016 (UTC)