Standards for Chinese Pronunciation in Quizbowl: An Essay (Part One)

[wadegilespromoter]

Dear mods: Would it be possible for this post to be moved to Collegiate Discussions? I'm so sorry about this.
Standards for Chinese Pronunciation in Quizbowl: An Essay (Part One)

Hello everyone. I’ve been playing Quizbowl for about two years now, and as an international student from Taiwan, and also a native speaker of both Mandarin Chinese and Taiwanese (also known as Hokkien), much of my knowledge of Chinese history and literature stems from reading original language sources and/or general knowledge as a Taiwanese - which has (obviously) been to my advantage but has at times been to my detriment due to some inconsistencies regarding answerlines in original languages, which, although most of these situations have been protested and resolved, still leads to unnecessary frustration at times. This, as well as a general desire to promote better Quizbowl with regards to East Asian content, has motivated me to write this essay to hopefully push Quizbowl in a better direction.

I will be using database questions I’ve found on Qbreader as examples - but they are meant in no way to criticize the great people behind these packets and questions, but rather to illustrate the need for such a standard to be established.

I also anticipate for this essay to be a prequel to an East Asian-focused tournament/side event in the far, far future, so any feedback on this essay would be greatly appreciated. With that, let’s get started…


Languages - Transliteration and Pronunciation

Chinese is a very different language from English. This can result in problems when we translate original language terms into English. Unfortunately, this is where the “consonant rule” can fail us: different romanization systems can use different consonants to ascribe to the romanization. As you can probably tell, I am a big proponent of the Wade-Giles romanization system (for a variety of reasons), but I recognize that Pinyin is the dominant system currently to Romanize Chinese. I do think that, where it matters, one must ensure that both Pinyin and Wade-Giles should be acceptable. For example:

This emperor unusually succeeded to the throne as the younger brother of his predecessor, thus becoming the second emperor of the Sòng dynasty.

ANSWER: Tàizōng (“tye-DZONG”) [or Zhào Kuāngyì; or Zhào Guāngyì; or Zhào Jiǒng; prompt on Zhào; reject “Taizu”]

While “Tai-tsong” (Wade-Giles) and “Tai-zong” (Pinyin) would be virtually indistinguishable to the reader, if someone says “Chao” (Wade-Giles) rather than “Zhao”, the player may be penalized if the moderator happens to be too stringent. This should be something that editors should be aware of, in my opinion.

Moreover, since English is a non-tonal language, and it would be completely unreasonable to expect Westerners with no Chinese education to speak tones, I believe that we should do away with tonal markings altogether - most of the time readers won’t read them correctly anyway.

Moving on, oftentimes readers can be confused by certain ways that Pinyin spells things in questions. While it may be foolish to replace all Pinyin with a better romanization system (as it prevents people from searching up questions easily), we could improve this by having a better pronunciation guide. Yes, we currently do have a pronunciation guide library, but I find it inconsistent, inaccurate, and sometimes have two pronunciations for the same term, which can lead to inconsistent pronunciation guides across packets.

It’s important to note that we are not linguistic experts here, just quizbowlers (unless you're a linguistics expert playing Quizbowl, which, props to you). All we’re trying to do here is to ensure that Western readers can read past these words without trouble, not to establish new linguistic standards for people to learn Chinese. I would also note that this is probably the part where I am least qualified to speak, as I grew up learning Chuyin/Zhuyin (bopomofo) as a native speaker from Taiwan, and not Pinyin, so I am not necessarily familiar with how romanization necessarily works, and
I also have no idea how Westerners would perceive Romanization, so any input by Westerners (especially those who are learning Mandarin!) would be greatly appreciated.
Let’s take a look at some examples:

At a banquet, this figure's spirit possessed and killed the (*) general Lǚ Méng (LOO mung), who led the forces responsible for his death. Prior to his deification, this wielder of the Green Dragon Crescent Blade spared Cáo Cāo (TSAO tsao) at Huáróng (HWAH-rong) Pass and swore the Peach Garden Oath with Zhāng Fēi (jahng fay) and Liú Bèi (l'yoo BAY). For 10 points, name this Chinese martial deity who was once a Three Kingdoms-era general.
ANSWER: Guān Yǔ [accept Guān Dì or Guān Gōng; accept Guān Yúncháng; accept Kuan for “Guan” in all answers; prompt on Guan or Kuan; prompt on martial god or wǔdì]

First of all, I strongly commend the editor for allowing “Kuan” in place of “Guan”.
Second of all, we can see many common pronunciation missteps here. Inevitably, “Cao Cao” will be pronounced as “Kow Kow” rather than “Tsao Tsao”, and so we should always include a pronunciation guide for “Tsao”. However, using the letter Y would be a bad idea, since some readers may “l’yoo” and pronounce it as “lie oo” instead, which would throw off the reader.
So in general:

1. “C” (ㄘ) in pinyin should have a “Ts” in the pronunciation guide: “Cao”(曹) -> "Tsao"
2. “Q” (ㄑ) in pinyin should have a “Ch” in the pronunciation guide: “Qin” (秦) -> “Chin”
3. “X” (ㄒ) in pinyin should have a “Hs” in the pronunciation guide: “Xi” (習)-> “Hsi” - although I’m interested in whether or not Westerners may struggle with this as well (I’m not sure), so we may need to find another way for this.
4. Avoid using apostrophe marks in pronunciation guides. Liu should be adequate.
5. Differentiating between “u” (ㄨ) and “ü” (ㄩ): “u” (no umlauts) should be pronounced as “oo”, and “ü” should technically be pronounced like in German (with the Umlaut), but putting “oo but with your mouth pursed” in pronunciation guides is unreasonable, so I think having “Uh” may better (?), or perhaps we can use “yu”, as it is used in German but then we run the risk of having readers pronounce “Lyu” (呂) as “Lie yu”, so I don’t know, man.
6. Not using “Y” except in “You”. This prevents readers from pronouncing any “y”s as a vowel rather than a consonant.
7. In cases where the pinyin uses an apostrophe, a space should be used in the pronunciation guide instead. For example “Chang’e” may be mispronounced as “Change”, which would be really bad, so we should use “Chang Eh” - since the original Chinese is two words as well (嫦娥), we can save some headaches. In another example, Xi’an (西安) should have the pronunciation guide of “Hsi An”.

Otherwise, I believe that the remainder of Pinyin pronunciations should be adequate for Quizbowl purposes. In the first example above, for example, Zong (宗) should be adequate to pronounce for most Westerners, and have "dzong" as a pronunciation guide may have readers read as "dezong" which would be disastrous since it would refer to another emperor (Dezong of Tang 唐德宗).

Accepting regional language answers

I am a big proponent of regional languages, and I therefore believe that, wherever reasonable and possible, we should accept both the Mandarin Chinese pronunciation and the pronunciation of the local language. For example:

[10] Name this politician who claimed “We are victims of policy!” after Cathay Pacific aircrew members were allowed to attend a birthday celebration maskless at the Reserva Iberia restaurant. This man who was the victim of a 2019 stabbing has claimed that Nobel Peace Prize nominee Alexandra Wong was “a member of Al-Qaeda.”

ANSWER: Junius Ho Kwan-yiu

Here, the answer line is formatted where, if we give the English first name (Junius) or the Cantonese (Kwan-yiu) first name, it would be acceptable. However, we should probably allow for “He Jun-yao” so that the Mandarin would also be acceptable.
This should also extend to geographical names as well. For example:

For 10 points, give this city in Southern Taiwan, which contains the 17th largest port in the world.
Answer: Kaohsiung City (accept Gaoxiong, Ko-hiong, prompt on Takao)

Here, we have a Taiwanese city. The primary answer line is the most common English name, but Gaoxiong is also acceptable due to the Pinyin romanization. The Taiwanese term Ko-hiong is also acceptable. The historical term of Takao is prompted on a case-by-case basis.


Requiring full names

I believe, at least at the collegiate level, that we should start requiring full names for East Asian names, rather than just surnames. While it is understandable when referring to Western figures to only give the last name, surnames in East Asia are much more commonly shared with one another. While “Smith”, the most common surname in America (and also a very common Quizbowl guess), Smiths only composed 0.9% of the American population in 2000 (source). On the other hand, “Wang” compose 7.2% of the population, and the top 20 surnames alone compose more than half of the current population in China. Requiring full names can lead to less fraud.
Obviously, this may be an unpopular opinion, so I recognize how this would make things way too complicated for something that should be simple and consistent.


Any input, critiques, or comments are welcomed. I will put out a part two specifically on Chinese Emperor names in the future (most likely after my MCAT in April), so please stay tuned for that too.

[dni]

Most of this seems sensible to me. I even somewhat agree with "Requiring full names" in theory, though I would never advocate for it in practice for obvious reasons. The argument you lay out is also far more applicable to Korean surnames and may also be relevant to other countries/languages as well.

I believe that we should do away with tonal markings altogether - most of the time readers won’t read them correctly anyway.

I disagree with this point. The tones are definitely helpful for people who can read them and are also useful when reading these packets later. Moderators who don't know how to read tones (or similar phonetic notation, in any language) simply ignore them. It is very easy to do this given the original letters are all still there. If the letters alone are insufficient, there should be a separate pronunciation guide anyways the moderator will just skip to, instead of attempting to "read [the tones] correctly," which I don't see how they could even attempt, given they don't know what the tones mean.

I did also wanted to also add one other thing I don't believe you covered. I've seen a few pronunciation guides adding erhua for seemingly no reason? Here's one example:

... For 10 points, name these now-defunct objects that used keys to create hanzi ("HAN-zur") symbols.
ANSWER: Chinese typewriters [or The Chinese Typewriter: A History; or mingkuai until it is read; prompt on typewriter with "what language is it designed to write?"]

Ignoring how terrible the content of this tossup is, this kind of pronunciation guide seems like something that should basically never be done, as the vast majority of players are unlikely to have encountered this pronunciation before.

[wadegilespromoter]

Thanks for your reply!

Regarding erhua, this is something that didn't really enter my mind as I was writing, since such a feature is extremely rare in Taiwanese Mandarin, and for many Taiwanese people (like me!), the pronunciation of the word "er" (兒) is sometimes closer to "eh" anyways.

I think,
8. Erhuas should never be written into pronunciations guides if the original Chinese doesn't have the corresponding character in it. Now, if we are indeed talking about a specific character, work, or object that does have "er" at the end in the original language, (example below), we should include it. In such a case, we should treat "Er" as a separate word/character for pronunciation purposes.

The serving girl Pang Chunmei is featured in the final parts of this novel, which earlier sees the sordid marriages of Li Ping’er and Pan Jinlian. For 10 points, name this novel about Ximen Qing, loosely based off an incident in The Water Margin and known for its explicit material.
ANSWER: The Plum in the Golden Vase [or Jin Ping Mei; or The Golden Lotus]

I would revise this question to add the following pronunciation guides:

The serving girl Pang Chunmei is featured in the final parts of this novel, which earlier sees the sordid marriages of Li Ping’er (Lee Ping Er) and Pan Jinlian. For 10 points, name this novel about Ximen Qing (Hsimen Ching), loosely based off an incident in The Water Margin and known for its explicit material.
ANSWER: The Plum in the Golden Vase [or Jin Ping Mei; or The Golden Lotus; Accept Chin Ping Mei]

9. Although, if we have a Beijing player who is very used to adding the "er" sound left and right, we should also possibly accommodate for that in the answerline? I'm not entirely sure what the rule of erhua is supposed to be since I never do it myself, to be honest.

I disagree with this point. The tones are definitely helpful for people who can read them and are also useful when reading these packets later. Moderators who don't know how to read tones (or similar phonetic notation, in any language) simply ignore them. It is very easy to do this given the original letters are all still there. If the letters alone are insufficient, there should be a separate pronunciation guide anyways the moderator will just skip to, instead of attempting to "read [the tones] correctly," which I don't see how they could even attempt, given they don't know what the tones mean.

That's an interesting take. My concern is that it may be possible for readers to be confused by tonal markings as emphasis or as a diacritic, or for some readers to do tones poorly (for example, some may not realize the 3-3 turns to a 2-3) and cause a unnecessary distraction.
Furthermore, whenever I read during practices, some people (non Chinese speakers) have complained that my tonal pronunciation of Mandarin words makes it harder for them to understand what I'm saying (ex: pronouncing it as Wang2 An1 Shi2 rather than Wang Anshi). I am concerned for native speaker readers to add tones in their pronunciation may lead to less comprehension by players, which is obviously not a good thing.

I have also heard arguments that reading Chinese terms atonal is more fair to Westerners, but I am not sure how valid that argument is.

[1.82]

I appreciate this post (and Alan's recognition that Wade-Giles is a superior transcription system to Hanyu pinyin, which most people fail to ever acknowledge because of their mental subservience to the idea that new things are better than old things) and I agree with Dan's addenda. Removing diacritics from all foreign names and terms is a position that Bruce Arthur long espoused on this forum. My feeling is that tone markers are useful at best and harmless at worst since they provide no information at all for someone who doesn't understand them, but on the other hand it is inconsistent to use tone markers for Mandarin transcriptions but not for those in other Chinese languages.

I do disagree specifically with point 3 in the pronunciation guide: pinyin "x" should simply be transcribed in pronunciation guides as "sh" because anglophones do not distinguish between /ɕ/ (Wade-Giles hs) and /ʂ/ (Wade-Giles sh). For a similar reason, the distinction suggested in point 5 would be inadvisable, since the distinction between ü and u does not exist in English; asking readers to attempt such a distinction would be counterproductive, which is why pronunciation guides for terms in German and Turkish do not. Whatever ambiguity exists as a result of failing to make these distinctions is an inevitable product of asking about concepts from a foreign language.

I think that existing naming policy is sufficient for Chinese names. To provide the most absurd example, it would be unreasonably pedantic to demand "Mao Tse-tung" for a question on Mao, who is immediately recognizable and generally known by a single name. Chinese naming conventions are not identical to Western naming conventions, but we have to allow that quizbowl is a game being played among anglophones who are used to Western naming conventions by which it's customary to refer to people simply by family name: an answer of "Wang" could refer to Wang Hung-wen or Wang Ching-wei, it's true, but we would only ask for a full name if the two individuals are close enough that there could be confusion, just as with Andrew Johnson and Lyndon Johnson.

[theMoMA]

This is an interesting topic. As the possessor of about 200 words of Chinese (and shrinking), I feel comfortable expounding on a certain "westerner" perspective. I cannot read Chinese and probably never will be able to, even if I resume my interest in learning to speak and listen. I understand that tonality is important and the basics of how it works but would find it unhelpful in any respect in a quizbowl question. I know the basics about how to pronounce Chinese as it is romanized in either Pinyin or Wade-Giles. These are my biases and background.

For quizbowl, my preference (which I understand is somewhat in tension with the preferences of people who actually read, speak, and understand the language) is for there to be no tonal markers, no tonality in pronunciation, and well-written but "basic phonetic" pronunciation guides for most terms.

I think the overriding consideration for pronunciation should be that most quizbowlers are not Chinese speakers and will have encountered these names and terms by reading them on paper or hearing them pronounced by a non-Chinese quizbowl reader, and not by comprehending them as they were spoken in authentic Chinese. In broad terms, I am in favor of applying basic rules of pronunciation so that Qing is like "ching" and Cao Cao is like "tsao tsao" but avoiding overly "authentic" pronunciations that may help Chinese speakers but are likely to confuse me and the large majority of quizbowlers who don't speak Chinese and would probably not be able to make the phonetic connection between the authentic pronunciation and a name or term they've seen.

I can't tell if this means I entirely agree with Alan or just mostly agree with him, but I find his basic precepts (removing tonality and tonal markers, pronouncing terms broadly correctly but not aggressively so) to be in line with what I would prefer. (For the answer line questions, I obviously think both romanizations and appropriate regional-language alternatives should be included in answer lines; and I think there may be some cases when it is appropriate to ask for both family and given names in answer lines but this should be on a case-by-case basis.)

[Stinkweed Imp]

I was going to post my own anglophone perspective but I basically agree with Andrew, so I'll just add that quizbowl is an unusual context where it is both normal to encounter a word only in writing and have no opportunity to clarify what the meaning is, so it may be necessary for writers to sometimes sacrifice correct pronunciation for intelligibility to English-speaking players and moderators. Trying to include tones or consonants that aren't used in standard English in a pronunciation guide is liable to confuse players that aren't familiar with Chinese pronunciation and could easily result in the moderator saying a totally mangled pronunciation that confuses English and Chinese speakers alike. To counter my own bias, I'll note that this is by no means a problem unique to Chinese. There are many common quizbowl names that either have a counterintuitive pronunciation ("Bronisław"), use sounds that don't exist in English ("Llewellyn"), or have a standard English pronunciation that is different than the original language ("Rheims"), and quick search of qbreader shows that sets are inconsistent on what pronunciations to use (or if they include a pronunciation guide at all).

[Fado Alexandrino]

I think question text should be pinyin only, while answer text can have whatever various romanizations exist. I soft prefer having no tones in question text, because the questions flow more smoothly and I'm not speaking Chinese for various words, and that there is no variation in how different readers across different rooms pronounce these words.

[wadegilespromoter]

I want to clarify by saying that with this proposal, I have primarily three aims:

1. Have a pronunciation guide system that, when read by a Western reader, can somewhat resemble, rather than be a 100% replication, of the original term.

2. Have a system that prevents readers from saying something that is very very wrong, or refers to another different term entirely.

3. Be relatively simple and concise.

The thing about Quizbowl is that, while readers are only encountering something only in writing, players are encountering something only by listening. It is completely unreasonable to expect readers to master every single language on Earth (I myself struggle with the pronunciation of certain French and Italian words, the former coming up quite a lot (more than Chinese at the one dot level, I believe), so the goal is only to ensure that Western readers (those whose native language is English, and do not know a lick of Chinese), with relative minimal effort can still pronounce a term that is understandable by both Western players and those with a greater understanding of Mandarin. In summary, "good enough" for Quizbowl.

I do disagree specifically with point 3 in the pronunciation guide: pinyin "x" should simply be transcribed in pronunciation guides as "sh" because anglophones do not distinguish between /ɕ/ (Wade-Giles hs) and /ʂ/ (Wade-Giles sh).

Thanks for your input! I guess I'm too used to seeing "Hsu" and "Hsi" and "Hsinchu" (Taiwanese city) to be bothered by the "Hs" spelling. "Sh" should probably be sufficient at Quizbowl.

To provide the most absurd example, it would be unreasonably pedantic to demand "Mao Tse-tung" for a question on Mao, who is immediately recognizable and generally known by a single name.

That is also an interesting point. I guess in the west, "Mao Tse-tung" is the first thing that comes to mind when the surname "Mao" is mentioned, while in East Asia it would require a disambiguation (Chiang Kai-shek's first wife, for example, is Mao Fu-mei). I do recall someone complaining that, at a tournament, there was a question asking "the son of Chiang Kai-shek" and the answerline only asking for Chiang, and not the full name Chiang Ching-kuo. In that particular case, the full name should have been required.

This was partially inspired by my own experience at this year's ACF Regionals, where there was a bonus (which I loved) on KMT figures, but only asking for last names (Chen, Sung, and Chiang, respectively). I felt that it was too easy only asking asking for last names (especially Chen, as I figured there would be teams guessing a random common Chinese surname, Chen being one of them), but looking at the conversion rates it was probably at the right difficulty for Regionals.

I will say, in the East Asian-content focused tournament/side event that I have in the works, we will probably still ask for full names to promote deeper knowledge.

[eygotem]

Frankly, it seems unfair if some moderators are going to use the tone markings and other moderators don't. I speak Mandarin relatively fluently, but usually I make an effort to not pronounce tones during my reading because it simply would be inconsistent with what every non-Chinese reader would say, and I very much wouldn't want to cost someone points just because they heard something one way but are hearing it again a different way. I can definitely see where tones create additional clutter and have the potential to trip up readers who don't know what they mean. In my view, they should be removed at least from the question text so that different rooms can hear the same question regardless of who's reading.

[Skepticism and Animal Feed]

Diacritics are great for expanding the capability of the Latin alphabet and I would heartily recommend diacritical marks to anyone who is developing a romanization for a language. I even use diacritical marks in my own languages that I make up.

But in quizbowl they should be universally stripped out of the vast majority of words, perhaps left only in words like fiancee where we usually spell it in English with a diacritical marks.

The problem is that diacritical marks are used differently in different languages and this sows the seeds of confusion. The same diacritical marks used as Chinese tone markers are also used to indicate stress, vowel length, lowering/laxing the quality of a vowel, and various other things in languages that quizbowlers and moderators might be familiar with. A person familiar with Spanish, notably the second most spoken language in the USA and probably the only foreign language that most Americans ever have any meaningful interaction with, might stress the vowel that has a rising tone marker on it, a person familiar with Latin or Maori might pronounce a vowel with a high tone marker as a long vowel, etc. This literally makes it worse for everyone involved.

Then you get the other population, moderators who instead of misunderstanding a diacritical mark are just confused by it. They do what confused moderators do and they pause. Avoidable delays are the biggest evil in quizbowl. They lead to longer tournaments, to teams wasting their time standing awkwardly in front of room where tossup 16 is still being heard, and to quizbowl being less enjoyable. I literally once played a tournament that was shut down by police because it went past the time of the building reservation. You can't just think about ACF Nationals or Chicago Open where the moderators are all worldly people familiar with what diacritical marks mean in different world languages. You have to think about the marginal moderator, the 10th moderator being recruited for some obscure tournament out west because a new team joined the field.

If this were 20 years ago I'd add another paragraph about how not having to deal with diacritical marks makes things easier for editors but I think editors today are better and can handle it. Still it's a lot of work, especially if you take it upon yourself to add diacritical marks into submitted questions that lack them. There's also room for false positives (say, people whose families have been in the USA for generations and no longer put the accent mark over that letter) though that's probably a situation that will come up less than once per tournament.

I think it's a good point to consider the ethics of a moderator who knows 100% how to correctly pronounce a foreign word pronouncing it that way during a game. I am haunted by something I once did: I was reading a tossup about something Argentinian and while reading it I pronounced a Spanish word with an Argentinian accent. Shortly thereafter, noted native Spanish speaker and Chilean-American Seth Teitler buzzed in with the right answer. Maybe he just knew it (he did know a lot of things!), maybe he was being justly rewarded for his knowledge of how Argentinians talk. Or maybe I unnecessarily gave him an unfair advantage by showboating as a reader. Maybe I disadvantaged players on the other team who did not grow up with Chilean parents who made fun of how Argentinians talk and were just confused by my pronunciation.

[dni]

I make an effort to not pronounce tones during my reading because it simply would be inconsistent with what every non-Chinese reader would say, and I very much wouldn't want to cost someone points just because they heard something one way but are hearing it again a different way

This is a sentiment I have heard expressed a few times in the past as well and I don't think it makes sense. At the very least, empirically, it is not something everyone does, and I have had several moderators at ACF Regionals, ACF Nationals, SCT and ICT (which I would argue are the only tournaments where uniformity really matters) who, using their knowledge of Chinese, pronounced Chinese words with the tones wherever they could. A lot of these sets didn't have tone markings.

I'm not a linguist so I'm sure there's some subtle differences here, but words getting pronounced with their original language pronunciation (which I personally would consider to be more "correct" in most cases) and causing some player confusion is something that regularly happens in other languages as well (e.g. happens a lot to me for Fr*nch). It's a little annoying when it happens but it has never felt truly unfair -- I just didn't know how that word was actually pronounced -- and I have never heard anyone bring up this argument in relation to any language other than Chinese. The reverse of this also happens all the time: moderators mispronounce words (often due to a bad pronunciation guide) and cost players games. At the end of the day, moderators will never be truly uniform and removing information from question text that facilitates more correct pronunciations doesn't make sense to me and also doesn't seem like an improvement to that.

Whether it's in English or another language, if a moderator knows how to pronounce a word correctly, I think they should just ignore whatever pronunciation guide there is and pronounce it correctly. Uniformity across rooms is a nice ideal to strive for, but to make it happen in this context means instructing every moderator at every key tournament to actively make their reading of certain words more incorrect and defer to (often imperfect or absent) pronunciation guides. I recognize some players will get the short end of the stick here, but this feels more fair to me than the alternative, which would also be to the detriment of another set of players. I do, however, recognize it would be impossible to prove or disprove this, so in practice I would say sets can keep doing whatever they want :^)

With regards to the tones confusing non-speakers, I agree with Naveed that "tone markers are useful at best and harmless at worst since they provide no information at all for someone who doesn't understand them, but on the other hand it is inconsistent to use tone markers for Mandarin transcriptions but not for those in other Chinese languages."

As some others have touched on, this is not an issue unique to Chinese. Other languages also have diacritics, most of which I (and I assume most readers) do not understand. Here's some random examples from 2025 ACF Regionals:

A novel set in this city that retells the killing of Trotsky in multiple styles also includes the wordplay-heavy dialogue of Bustrófedon. The gay poet José Cemí (“say-MEE”) grows up in this city in the 1977 novel Paradiso.

This concept and “light” provide an alternate title for the “Light Prop for an Electric Stage,” a rotating kinetic “modulator” by László Moholy-Nagy (“LOSS-low MOH-ho-lee-NAHJ”).

Bach’s D minor concerto for this pair of instruments was performed in 1905 by Fritz Kreisler and Eugène Ysaÿe (“oo-ZHEN ee-ZYE”).

As a reader, when I see symbols like í or ÿ, which I do not understand and never encounter outside of quizbowl, I just go off the pronunciation guide where there is one or do the best I can while ignoring any diacritics if there isn't. I don't really see how this procedure would differ by language. Moderators are always going to pause at words they don't understand, I really don't believe there's a meaningful difference the presence or absence of diacritics make here.

[ahuff]

I do think we should probably be a bit more consistent with the goal and use of pronunciation guides.

Recently, a lot of quizbowl tournaments have moved to pronounce Mandarin words like a native Chinese speaker would say them.

At the same time, Spanish pronunciation guides are heavily Anglicized, and have a tendency to add new syllables, change the "e" vowel to an "ei" diphthong, and use English "v" instead of the Spanish one. Sets also have a tendency to change their mind about how to pronounce certain things question to question.

Brazilian Portuguese pronunciation guides seem to fluctuate between pronouncing words like English or Spanish, with only very rare instances of using Portuguese pronunciations.

I think all of these policies are fine in a vacuum, but I don't like the overall effect of seemingly only striving for accuracy for one language.

[wadegilespromoter]

I think all of these policies are fine in a vacuum, but I don't like the overall effect of seemingly only striving for accuracy for one language.

To be fair, we should ensure that your average moderator can pronounce foreign words reasonably with minimal effort for all languages. I recall reading a packet at practice one time where "!Kung" was in a question, which I believe most readers would likely just say "Kung!" or "exclamation mark, Kung" like I did. The only reason why I focused on Chinese is because it's a language that I know.

This is a sentiment I have heard expressed a few times in the past as well and I don't think it makes sense. At the very least, empirically, it is not something everyone does, and I have had several moderators at ACF Regionals, ACF Nationals, SCT and ICT (which I would argue are the only tournaments where uniformity really matters) who, using their knowledge of Chinese, pronounced Chinese words with the tones wherever they could. A lot of these sets didn't have tone markings

In my personal experience, I do try to say Mandarin as I would normally say in normal conversation, because it feels really weird to an extent if I say Mandarin terms I am very familiar with in an atonal manner. This also includes when I give answers as a player (I obviously will give answers that I am the most familiar with).

The same diacritical marks used as Chinese tone markers are also used to indicate stress, vowel length, lowering/laxing the quality of a vowel, and various other things in languages that quizbowlers and moderators might be familiar with.

This is the reason that, when I need to use romanization, I prefer having Wade-Giles system of numbers for tones (ex: Wei1 To3 Ma3), which is also the predominant system in Cantonese (jyut6 ping3) as well. However, the dominant system for Taiwanese is with diacritics (Tâi-lô).

When we have a Cantonese/Taiwanese/other non-Mandarin language term in the question (which I'm not sure how often that would happen), we could simply add the tonal system in as well, which wouldn't take that much more effort if people are already putting in tonal markings for Mandarin. However, as I said originally, I believe tonal markings should be done away altogether, as it unnecessarily makes things more complex than they should be.

[andyyu]

You bring up a lot of good points here. Most of your points seem reasonable, especially mentioning the need for accepting all relevant romanizations of a word in answerlines, and, if necessary, the inclusion of regional languages in answerlines. I also agree with Dan's reply regarding erhua and other dialectal variations.

I disagree with

Avoid using apostrophe marks in pronunciation guides. Liu should be adequate.

and

Not using “Y” except in “You”. This prevents readers from pronouncing any “y”s as a vowel rather than a consonant.

I've heard a surprising number of people unfamiliar with Mandarin pronunciation pronounce "Liu" as "loo", which can further add to the confusion between Lu/Lü. "y" is the best way we have to transcribe the consonant [j], and apostrophes are used to help clarify whether "y" is pronounced as a consonant or vowel. Admittedly, readers may not be familiar with apostrophes, and it isn't the prettiest-looking, but it does help instinctually nudge moderators in the right direction. I'm not saying we should go all out on the apostrophes and y's, but rather both can and should be used tactfully on a case-by-case basis.

From Ophir's style guide on writing pronunciation guides:

Use an apostrophe to break up non-English consonant clusters within the same syllable, especially if there is no good alternative.
for /kn/: ✔︎ Knuth (“k’NOOTH”)
for /ʃj/: ✔︎ Xia (“sh’YAH”), not ✘ Xia (“shyah”) — because shy suggests a beginning rhyme with “shy,” such as shy-ah /ʃaɪ.ɑː/ instead of /ʃjɑː/
for /jʃ/: ✔︎ Reich (“RYE’sh”), not ✘ Reich (“ryesh”) — because esh suggests a rhyme with “mesh,” such as ry-esh /ɹaɪ.ɛʃ/ instead of /ɹaɪʃ/

Taking "Xia" as an example, there's no good way of transcribing this without using an apostrophe or a "y". Removing the apostrophe suggests /ʃaɪ.ɑː/ as mentioned, not including the "y" can lead to /ʃɑː/, which can be confused with "Sha", etc. Maybe "shee-AH" could work, but it's still not perfect: this suggests two syllables rather than one. In the most extreme case, I could see people pronouncing it as /ʃi.ˈʔɑː/, with a glottal stop in the middle, which is far from ideal. Going back to the "Liu" example, I think "l'YOH" and "lee-YOH" are the "least bad" versions I could come up with, but the first involves an apostrophe and both involve a "y", and the second suffers from the two-syllable problem I mentioned earlier.

I don't really have a strong personal preference on tone marks within sets, as long as they are consistent throughout a set: i.e. if one question uses tones, all questions should. Generally, I don't think moderators think too much of them while reading. I've tried moderating with and without tones, and while moderating with tones is definitely more natural for me, moderating without tones ensures consistency with non-Sinophone moderators. Dan also brings up a good point that inclusion of tones can be useful in later reading or packet study.

That being said, one should note that some Romanization systems are more "invasive" than others for untrained moderators. Pinyin and Cantonese Yale's diacritical marks are much easier to ignore than the numbers of Wade-Giles or Jyutping, for example. It would be incredibly impractical to write pronunciation guides every single time such a system appears to prevent the moderator from saying things like "jyut-six ping-three" or "chung-one kuo-two".

As people have mentioned, all these problems are not unique to Chinese. I am of the opinion that pronunciation guides should lead Anglophone moderators as close as possible to an accurate pronunciation of the word in question using the phonemes of English (so, for example, Portuguese pronunciation guides should not have Spanish pronunciations) while still being understandable to players, and I also agree with Dan that someone who knows how to pronounce a word correctly should not be forced to "English-ify" their pronunciation. I do think some sort of cross-linguistic standard should be established for pronunciation guides in quizbowl, and that this standard should be adhered to whenever possible.

[Tejas]

I broadly agree with Andrew that the reader's primary goal should be to pronounce words in a manner easily intelligible for a fluent English speaker, with PGs meant to aid in that goal. Like others have said I don't think that "authentic" pronunciation is really a primary consideration if it impacts how a non-speaker would be able to interpret the word. When I'm adding PGs for Indian/Sanskrit-derived words there are sounds that a non-native speaker would have no way of distinguishing, so I try to just make sure that the vowels and the emphasis are correct since those are usually what trip me up when I'm playing and a reader can't pronounce a word properly.

I'll also add another point, which is that the necessity and format of PGs differs based on the reader. If I'm writing or editing something for ICT or ACF Nationals, I can safely assume that the readers are all experienced quizbowl players who don't need to be told pronunciations of common words that come up. If I'm editing for an IS set, then I assume the baseline reader is a coach or parent who may not have experience with a lot of these words/languages and I add more common and detailed PGs accordingly.

[Owen Mimno]

Maybe I’m not the most qualified person to be responding (having spent all of one hour on the continent of Asia) but I will offer my input.

I’m really glad that Alan has brought up the issue of answerlines. The existing rules regarding pronunciation are definitely adequate in most cases, but fall short in scenarios where there are many possible ways of transliterating consonants, which is often the case in Chinese languages. This problem is exacerbated by the fact that we are generally raised to perceive the sounds we are familiar with as individual units in an alphabet with no sense of the similarity between them. This is not just a “haha learn IPA suckers” post — as far as I know, linguistics has never developed any kind of phonological Levenshtein distance taking into account the similarity of speech sounds. The best we can do for now is making sure that a player doesn’t get negged for saying something that sounds to the mod like “Kuan Yu”.

I agree that using “hs” in pronunciation guides to represent the <x> in pinyin makes a lot of sense abstractly, as the best possible way to approximate the sound of /ɕ/ (especially the way I hear some Chinese speakers realize it — sort of a “hissy s”). That said, I don’t think it is the optimal transliteration when you are trying to elicit the correct pronunciation from people who do not know that it represents /ɕ/, or how to pronounce /ɕ/, or what that funny squiggle even is. To be fair, I think it is certainly better at this purpose than “x”, which is a point in favor of Wade-Giles, but at the same time non-Chinese-speaking quizbowlers seem to have internalized the “x on paper → say sh” rule, which hopefully trips readers up less than “hs on paper → say whatever hs is” would.

In any case, it is my understanding that in Mandarin the alveolo-palatals are in complementary distribution with the retroflex consonants, and it is a quirk of Wade-Giles that /ɕ/ is transliterated differently from /ʂ/ when both /tɕ/ and /ʈʂ/ are written as “ch” and both /tɕʰ/ and /ʈʂʰ/ are written as “ch’”. And furthermore, /ɕ/ and /ʂ/ are contrastive in other languages yet would both be written in pronunciation guides as “sh” — for instance, Polish, where they are represented by <ś> (or <si>) and <sz>, respectively (2016 NASAT uses “k-SHISH-toff kyeh-SHLOFF-ski” for Krzysztof Kieślowski). Anyway, all of this is to say that regardless of whether editors use pinyin or Wade-Giles in transliterations, I believe “sh” is better for PGs.

The debate on tones seems to me to be more philosophical — what is the purpose of pronunciation guides? Ideally, to be a correct written approximation of the pronunciation of all non-English words, but if that were all then we would just stick the IPA in questions. It might then seem like the purpose is for every reader’s pronunciation to be equal, such that no player’s ability to buzz is affected by who is reading. In this case, we would demand that all readers strictly obey the pronunciation guides, and not add tones to Chinese words even if they know them. This, too, seems problematic — should someone who learned French in school be making an effort to bring the quality of their pronunciation down to the same level as someone who didn't?

The further problem that results from this is that accurate pronunciations of foreign languages can disadvantage players who know the clues but perhaps have never heard the correct pronunciation, or are unable to distinguish certain sounds. Maybe it’s reasonable to expect that players will put in some effort to improve their pronunciation (learning that “x” isn’t /ks/ in Chinese), but not that players should all feel obligated to add Hungarian IPA transcriptions to their Moholy-Nagy cards. Quizbowl is meant to reward knowledge, not language skills, but we can't deny that quizbowl inherently does occasionally reward language skills (how would you stop players from lingfrauding? ban foreign words?), and thus I don't think it's a problem that having rudimentary sound perception ability should be a prerequisite for getting points in certain rare instances.

The compromise that seems to present itself is that diacritic marks present in other languages can be classified into three categories — firstly, those which can be represented in English without causing confusion (e.g. accents to represent stress, like how they are sometimes used in English poetry); secondly, those which represent phonological features with close or exact English analogues that can be clarified in the pronunciation guides (e.g. accents to represent vowel length, sound change markers like umlauts and tildes); and lastly, those which do not represent phonological features that can be explained in English pronunciation guides (e.g. tones in Chinese, underdots in IAST to represent retroflex sounds). My take is that the former two classes should be included in question texts, and the latter removed. Hopefully this minimizes confusion on the part of moderators, while permitting those who can actually pronounce a language correctly to do so. But of course, this leaves as many questions unanswered as those it solves, and I expect there will hardly be agreement.

Beyond diacritical marks, I think that pronunciation guides should generally follow this idea — sounds that can be somewhat accurately reproduced in a PG should be, and those that can’t shouldn’t be. In my amateur phonologist perspective, editors should perform their due diligence in ensuring that Joinville, Brazil is not pronounced like Jean de Joinville and neither are pronounced the way most English speakers inevitably will. However, I don’t think it’s possible to figure out a better PG for the Spanish word “nevada” /neˈβaða/ than what English speakers already say, however incorrect it may be. And unrelatedly, I think we should replace the Eszett "ß" with “ss” across the board, but this may be getting too far off track…

[ahuff]

To be clear, I like this post quite a bit.

I would welcome a publicly available document that summarizes a few recommendations for how to write pronunciation guides in various common languages, and would happily contribute to a section on Spanish

[forrestw]

As someone who doesn't speak any tonal language, I don't find that moderators giving words the appropriate tones impedes my comprehension. In my experience, I'm basically just "tone-deaf" and don't process the tones as meaningful.

However, as Tejas brings up, there are sounds in many languages that have no equivalent in English. To use an example from earlier in this thread, Mandarin has the /ɕ/ phoneme that most closely resembles English /ʃ/. If a moderator does use /ɕ/, I think that probably does slow down my comprehension but not to the point where I feel a particular need for accommodation, since there's only one English sound it resembles. But what about Cao Cao? Someone who doesn't know how pinyin works and has only read about him can't buzz on a correct pronunciation of his name, but pronouncing it as "kow kow" is not only wrong, it punishes players who do know how to say his name. Ultimately, I think moderators should pronounce words as accurately as possible--this means that speakers of a language should not attempt to "anglicize" their pronunciation and pronunciation guides should tell non-speakers to use the closest reasonable approximation in English phonology. I think that failing to buzz from not knowing how to pronounce a name is a skill issue, I at least would feel like I was the one at fault in such a situation.

[wadegilespromoter]

To be clear, I like this post quite a bit.

I would welcome a publicly available document that summarizes a few recommendations for how to write pronunciation guides in various common languages, and would happily contribute to a section on Spanish

There is this, but currently there is only French. In my opinion, based on some of the feedback in this post, I would have the following recommendations:

Pronounciation Guides For Readers

1. “C” (ㄘ) in pinyin should have a “Ts” in the pronunciation guide: “Cao”(曹) -> "Tsao"

2. “Q” (ㄑ) in pinyin should have a “Ch” in the pronunciation guide: “Qin” (秦) -> “Chin”

3. “X” (ㄒ) in pinyin should have a “Sh” in the pronunciation guide: “Xi” (習)-> “Shi”

4. Avoid using apostrophe marks in pronunciation guides. "Liu", and not "L'yoo".

5. Differentiating between “u” (ㄨ) and “ü” (ㄩ): “u” (no umlauts) should be in the pronounciation guide as “oo”, and “ü” should be "uh".**
"Lu" (陸) -> "Loo", "Lü" (呂) -> "Luh"

6. Not using “Y” except in “You”, the consonant. This prevents confusings in whether or not to pronounce "Y"s as a consonant or a vowel.

7. In cases where the pinyin uses an apostrophe, a space should be used in the pronunciation guide instead. For example “Chang’e” (嫦娥)-> “Chang Eh”, Xi’an (西安) “Shi An”.

8. Erhuas should never be written into pronunciations guides if the original Chinese doesn't have the corresponding character in it. Now, if we are indeed talking about a specific character, work, or object that does have "er" at the end in the original language, we should include it. In such a case, we should treat "Er" as a separate word/character for pronunciation purposes. "Li Ping’er" (李瓶兒) -> "Lee Ping Er"

9. Concision should be the goal, and so tonal markings are optional.

Answer Lines:

1. Answer lines should accomodate for both Wade-giles and Pinyin. In most cases one answer should be sufficient despite the spelling difference, but the following should be considered

a) b (Pinyin) <-> p (Wade-Giles)
b) d (Pinyin) <-> t (Wade-Giles)
c) g (Pinyin) <-> k (Wade-Giles)
d) j (Pinyin) <-> ch (Wade-Giles)
e) zh (Pinyin) <-> ch (Wade-Giles)
f) r (Pinyin) <-> j (Wade-Giles)

2. Whereever reasonable, answerlines should accomodate for both the Mandarin and the local regional language(s) pronounciation. We should also accomodate for possible erhua in pronounciations as well.

3. Primary answer lines, as well as the term being read in a question, should be the spelling that a work or figure is primarily known in the west for, whether that be Pinyin (Xi Jinping), Wade-giles (Lai Ching-te), or name in a language other than Mandarin (Sun Yat sen), or maybe even in English (Andy Lau), despite the above accomodations.


**Pending more input on this - would "Lyu" or "Luh" be a better approximation for 呂?

[andyyu]

**Pending more input on this - would "Lyu" or "Luh" be a better approximation for 呂?

I tend to agree with Naveed here:

the distinction suggested in point 5 would be inadvisable, since the distinction between ü and u does not exist in English; asking readers to attempt such a distinction would be counterproductive, which is why pronunciation guides for terms in German and Turkish do not.

Many people have pointed out how it is impossible to distinguish between all the sounds in a language using only the phonemes of English. English speakers won't be able to distinguish between /u/ and /y/, no matter what phonetic respelling you throw at them. "Lyu" is problematic for previously mentioned reasons. "Luh" to me suggests /lʌː/, which I would interpret as Le/乐 /lɤ/, a different word entirely.

In this case, I would say to just remove pronunciation guides and mark the umlaut if needed. People who know what the umlaut does can adjust their pronunciation accordingly, while people who don't will just skip over it. Maybe put "loo" if you really want a pronunciation guide. It isn't perfect, but it's the best solution I can think of.

[meebles127]

This thread is extremely interesting.

I've been studying Chinese for around 12 years now and spent 4 months living there last year. God willing, I'll be going back after I graduate to temporarily escape this fascist hellhole. Anyways......

I've been involved in proofreading/packetizing/adding PGs to a significant number of sets in the last several years. In this time, I have never felt comfortable writing pronunciation guides for Mandarin because I simply can no longer understand how English speakers "think about" Chinese.

I distinctly remember one instance where I was helping John Lawrence on a set and he asked me if I could PG a certain Chinese phrase. Instead of doing something that would be exceedingly unhelpful to English speakers, I instead voice recorded me saying the phrase and sent it to John. Apparently this was helpful.

[wadegilespromoter]

This thread is extremely interesting.

I've been studying Chinese for around 12 years now and spent 4 months living there last year. God willing, I'll be going back after I graduate to temporarily escape this fascist hellhole. Anyways......

I've been involved in proofreading/packetizing/adding PGs to a significant number of sets in the last several years. In this time, I have never felt comfortable writing pronunciation guides for Mandarin because I simply can no longer understand how English speakers "think about" Chinese.

I distinctly remember one instance where I was helping John Lawrence on a set and he asked me if I could PG a certain Chinese phrase. Instead of doing something that would be exceedingly unhelpful to English speakers, I instead voice recorded me saying the phrase and sent it to John. Apparently this was helpful.

Voice recording is certainly an interesting approach to do pronunciation guides, although I don't believe it's scalable.

This also poses another question on whether native speakers should be exclusively responsible in developing pronunciation guidelines, and to what involvement do people of varying abilities with the language should contribute to developing such a guideline. As a native speaker, I certainly believe the "u" and "u-umlaut" (ㄨandㄩ) sounds are very different, but based on some of the feedback in this thread perhaps they should have the same pronunciation guide marker. Same thing with the "X" (ㄒ) sound - I certainly thought it was very different from a "sh" (ㄕ) sound, but it seems that having "sh" for both in the pronunciation would be the most straightforward option.

Again, probably good enough for quizbowl.

[meebles127]

This thread is extremely interesting.

I've been studying Chinese for around 12 years now and spent 4 months living there last year. God willing, I'll be going back after I graduate to temporarily escape this fascist hellhole. Anyways......

I've been involved in proofreading/packetizing/adding PGs to a significant number of sets in the last several years. In this time, I have never felt comfortable writing pronunciation guides for Mandarin because I simply can no longer understand how English speakers "think about" Chinese.

I distinctly remember one instance where I was helping John Lawrence on a set and he asked me if I could PG a certain Chinese phrase. Instead of doing something that would be exceedingly unhelpful to English speakers, I instead voice recorded me saying the phrase and sent it to John. Apparently this was helpful.

Voice recording is certainly an interesting approach to do pronunciation guides, although I don't believe it's scalable.

This also poses another question on whether native speakers should be exclusively responsible in developing pronunciation guidelines, and to what involvement do people of varying abilities with the language should contribute to developing such a guideline. As a native speaker, I certainly believe the "u" and "u-umlaut" (ㄨandㄩ) sounds are very different, but based on some of the feedback in this thread perhaps they should have the same pronunciation guide marker. Same thing with the "X" (ㄒ) sound - I certainly thought it was very different from a "sh" (ㄕ) sound, but it seems that having "sh" for both in the pronunciation would be the most straightforward option.

Again, probably good enough for quizbowl.

The Chinese u and ü sound completely different to me.

Quizbowl's insistence that xi and shi are similar is one that continues to completely baffle me. As a non-native learner they couldn't be more different.

[1.82]

This also poses another question on whether native speakers should be exclusively responsible in developing pronunciation guidelines, and to what involvement do people of varying abilities with the language should contribute to developing such a guideline. As a native speaker, I certainly believe the "u" and "u-umlaut" (ㄨandㄩ) sounds are very different, but based on some of the feedback in this thread perhaps they should have the same pronunciation guide marker. Same thing with the "X" (ㄒ) sound - I certainly thought it was very different from a "sh" (ㄕ) sound, but it seems that having "sh" for both in the pronunciation would be the most straightforward option.

Again, probably good enough for quizbowl.

The important factor here is not whether these phonemes sound different but whether it's reasonable to ask a reader to produce them differently, since quizbowl sets are written for other people to read aloud. To use an extreme example, the letters "ngq" in "Nongqawuse" are pronounced /ᵑǃʱ/, but no anglophone reader can reasonably be expected to pronounce this, so pronunciation guides should render it as "ng" instead. It doesn't matter that even someone who speaks no Xhosa can tell that that's not the same sound. (If you are interested in how best to anglicize click consonants, here is an interesting page that I have posted on this forum before.) The distinctions in Mandarin between /u/ and /y/ and between /ʂ/ and /ɕ/ are not replicable by anglophone readers and so they should not be represented in a pronunciation guide. This applies to all languages.

[joshxu]

...
I'm not a linguist so I'm sure there's some subtle differences here, but words getting pronounced with their original language pronunciation (which I personally would consider to be more "correct" in most cases) and causing some player confusion is something that regularly happens in other languages as well (e.g. happens a lot to me for Fr*nch). It's a little annoying when it happens but it has never felt truly unfair -- I just didn't know how that word was actually pronounced -- and I have never heard anyone bring up this argument in relation to any language other than Chinese. The reverse of this also happens all the time: moderators mispronounce words (often due to a bad pronunciation guide) and cost players games. At the end of the day, moderators will never be truly uniform and removing information from question text that facilitates more correct pronunciations doesn't make sense to me and also doesn't seem like an improvement to that.

Whether it's in English or another language, if a moderator knows how to pronounce a word correctly, I think they should just ignore whatever pronunciation guide there is and pronounce it correctly. Uniformity across rooms is a nice ideal to strive for, but to make it happen in this context means instructing every moderator at every key tournament to actively make their reading of certain words more incorrect and defer to (often imperfect or absent) pronunciation guides. I recognize some players will get the short end of the stick here, but this feels more fair to me than the alternative, which would also be to the detriment of another set of players. I do, however, recognize it would be impossible to prove or disprove this, so in practice I would say sets can keep doing whatever they want :^)
...

I pretty much agree entirely with this post, especially these highlighted portions. As someone who's lost buzzes due to mods Anglicizing/mispronouncing Chinese words, I find this situation more unfair than mods pronouncing words correctly and confusing non-speakers. I don't fault the mods in question here, and I concede there's nothing really that could solve this problem, but I'd very much not like it to be a norm where players whose interaction with the term lies exclusively in the original language and its pronunciation are systematically put at a disadvantage.

As a mod, I ignore pronunciation guides if I know the correct pronunciation regardless of the language. I don't feel too strongly about this, but I generally appreciate tone markings for Chinese words if only to prevent me/mods from using the wrong tones. There's only one way for me to pronounce Wang Anshi with or without tones or a pronunciation guide. There's no way any attempt for me to pronounce that atonally wouldn't be more confusing to players.

[Eddie]

Hello, everyone.

Given that my main involvement in quizbowl these days is in proofreading and providing pronunciation guides, I'd like to offer my support for various positions that have been advanced herewithin, and offer a new proposal as to the transcription of high front rounded vowels [y ʏ].

First, I propose that we generally* transcribe the high front rounded vowels [y ʏ] with the digraph <ew>, as in 呂 Lǚ ("lew"). This has been my choice for several years, and I would like to argue for it as follows:

• First, it is easier to parse and read than <'yoo>, as in 呂 Lǚ ("l'yoo"), which I have found many American readers unable to pronounce anyways, instead resorting to either "lee-YOO" or "loo".
• Next, it avoids the ambiguous use of <y> as a consonant or vowel.
• Third, it maintains a formal orthographic distinction with <oo> [u ʊ], which provides peace of mind.
• Fourth, if the reader has yod-dropping† or is unable to pronounce the given /Cj/ cluster, then it is simply pronounced as <oo> [​u], providing a free fallback pronunciation.

*The exception is that if [ty dy] are PGed as ("tew dew"), then a reader with yod-coalescence, as in most British dialects, will read them as ("chew jew"). I do not know if this has ever happened, and if it has, if it has affected gameplay.
†Most Americans have yod-dropping. If you say "tune" and "toon" the same, then so do you.

As to the issues specific to Chinese languages, my thoughts largely align with those that have come before me, and I would simply like to show my support for these positions:

• I agree that answer lines should always provide Chinese romanizations in Pinyin, Wade–Giles, and the individual- or language-specific romanization, if distinct from the first two.
• I agree that pronunciation guides should be given in lay English eye spellings that conform to English phonotactics, and in particular, that [ʃ ʂ ɕ] should all be transcribed as "sh", and likewise [ʒ ʐ ʑ] as "zh", [ʧ ʨ ꭧ] as "ch", and so on.
• In conjunction with the previous point, the Chinese name or term itself should be given in whatever romanization is more appropriate or used by the person themself (e.g. Xi Jinping but Lai Ching-te).
• I agree that features specific to a dialect should not be transcribed, whatever be its prestige or standardness. Hence, erhua should not be transcribed, and conversely, neither should, for example, nonrhoticity in German German.
• I agree that it is an intractable problem to find an ideal transcription for the Mandarin palatal–glide clusters (as in 姐 jiě, 喬 qiáo, 夏 xià). I personally prefer the two-syllable solution ("jee-YAY", "chee-YAO", "shee-YAH"), but do not believe it to have an objective advantage over the apostrophe solution ("j'YAY", "ch'YAO", "sh'YAH").
• I have no strong feelings on the transcription of tone, other than that it be consistent throughout the set.
• I agree that diacritics in general should be kept as in their original language, and I agree in particular with Dan's point that moderator slowdown is caused by unfamiliar words in general, and not by diacritics specifically. By the same logic, for example, we might propose that quizbowl avoid foreign words with the letter <c>, since the letter is liable to be pronounced as any of [c ʤ s ʦ ʧ θ k |], among others.

Lastly, in regards to Andy's point about the Anglicization of foreign sounds, I propose that foreign sounds should be transcribed with the English sound that is both (a) the most perceptually similar and (b) the most harmonious with English phonotactics. In particular, I propose that we generally avoid transcribing the lax vowels [ɪ ɛ] ("-ih" "-eh") in open syllables. So for example, we should PG Spanish de [de] as ("day") rather than ("deh"), but keep des [des] ("dess"). To this end, I have earlier likewise transcribed Mandarin 姐 jiě as ("jee-YAY") rather than as ("jee-YEH"), and in hindsight, I suppose this also explains my preference for the two-syllable solution over the apostrophe solution.

[Eddie]

Hello again, everyone.

I have made this post in a new post, as it is about a different issue in the PGing of Sinitic languages and sinoxenic vocabulary. How should PGs mark stress when transcribing from a language without phonemic stress, such as Korean, French, and most Chinese languages? I see two basic options:

1. Do not mark stress at all, as in Mandarin Taizu ("tye-tsoo"), Korean Taejo ("tay-joh"), and French Paris ("pah-ree").
2. For each language, pick a distinguished syllable position to always stress (ultimas in Mandarin and French; initials in Korean).

My personal preference is for the second option. English has phonemic stress, and if stress is not marked, then the reader will unconsciously choose a particular syllable to stress anyways, so it is better to avoid the risk and explicitly notate which syllable to stress.

[wadegilespromoter]

Thanks for your reply!
Some thoughts:

First, I propose that we generally* transcribe the high front rounded vowels [y ʏ] with the digraph <ew>, as in 呂 Lǚ ("lew"). This has been my choice for several years, and I would like to argue for it as follows:
First, it is easier to parse and read than <'yoo>, as in 呂 Lǚ ("l'yoo"), which I have found many American readers unable to pronounce anyways, instead resorting to either "lee-YOO" or "loo".
Next, it avoids the ambiguous use of <y> as a consonant or vowel.
Third, it maintains a formal orthographic distinction with <oo> [u ʊ], which provides peace of mind.
Fourth, if the reader has yod-dropping† or is unable to pronounce the given /Cj/ cluster, then it is simply pronounced as <oo> [​u], providing a free fallback pronunciation.
*The exception is that if [ty dy] are PGed as ("tew dew"), then a reader with yod-coalescence, as in most British dialects, will read them as ("chew jew"). I do not know if this has ever happened, and if it has, if it has affected gameplay.
†Most Americans have yod-dropping. If you say "tune" and "toon" the same, then so do you.

This is an intriguing proposal! I didn't think of "ew" ("new", "few") would be a good approximation of u-umlaut (呂) but I do think this is a great idea. I have a concern though: if we mark "lew" in the PG, this may sound similar to what some western readers pronounce "Liu" (劉) (even though that's is not how it's pronounced), which would replace one ambiguity (Lu vs Loo 呂 陸) with another (Lu vs Liu 呂 劉).

I agree that features specific to a dialect should not be transcribed, whatever be its prestige or standardness. Hence, erhua should not be transcribed, and conversely, neither should, for example, nonrhoticity in German German.

As state previously, I strongly agree with this idea, as erhuas is not present in conversational Taiwanese mandarin at all. However, a quick caveat: there are instances where the character "er" (兒) is used in a form of erhua, explicitly written in text, such as in character names (ex: 李瓶兒/Lee Ping'er). As I pointed out before, this would be the only case where writing out the "er" would be expected and required, which you then designate "er" as a separate character (Lee Ping Er) in the PG, rather than treating the erhua sound as a modification of the previous character (which is what some Northern Chinese dialects do, if I am correct).
If "Lee Ping'er" is the answerline (assume we want the full name), "Lee Ping" alone should not be acceptable, because that's not the character's name.

I agree that it is an intractable problem to find an ideal transcription for the Mandarin palatal–glide clusters (as in 姐 jiě, 喬 qiáo, 夏 xià). I personally prefer the two-syllable solution ("jee-YAY", "chee-YAO", "shee-YAH"), but do not believe it to have an objective advantage over the apostrophe solution ("j'YAY", "ch'YAO", "sh'YAH").

I disagree, since for 姐 喬 夏, I feel like the vowels parts do not warrant a PG for them (Q and X do, as stated previously). Since English does have diphthongs (a combination of two vowels), we may need to ensure no one pronounces "ie" as in "Tie", but we should be able to have western readers pronounce "iao" and "ia" correctly and would not warrant a PG.
Perhaps we can use "eh" as a vowel ending for the ㄝ (Pinyin: e, Wade-giles: Eh). So 姐 would be "Jie (PG: Jieh)" in a question text.

The concern with having multi-syllabic PGs (ex: jee-YAY) for one character is that it may be easy for a zealous but uninformed western reader to pronounce as "Jee Yay", which may be misinterpreted by players to be two characters (ex: 吉野 instead of 姊), which would not be good.

I have made this post in a new post, as it is about a different issue in the PGing of Sinitic languages and sinoxenic vocabulary. How should PGs mark stress when transcribing from a language without phonemic stress, such as Korean, French, and most Chinese languages? I see two basic options:

In my opinion, I think stress marks should only be present where in normal English contexts it is used (ex: fiancée), because stress marks aren't really a thing in Mandarin. If a reader stresses it one way or another, it doesn't affect the meaning, it just sounds weird (which, when factoring that most westerners don't speak tones properly, means that they already do "sound weird" in the eyes of a native speaker). Having readers stress the ultima just sounds even weirder.


--

我們都應該學中文

沒辦法，畢竟要真的學會一門新的語言對很多人來說並不簡單，所需的時間和精力也是許多人沒有的。
Translation:
Em: We should all learn Chinese.
Me: Unfortunately, it is not easy for many people to learn an entirely new language, and many people do not have the time and effort required to do so either.

[Eddie]

This is an intriguing proposal! I didn't think of "ew" ("new", "few") would be a good approximation of u-umlaut (呂) but I do think this is a great idea. I have a concern though: if we mark "lew" in the PG, this may sound similar to what some western readers pronounce "Liu" (劉) (even though that's is not how it's pronounced), which would replace one ambiguity (Lu vs Loo 呂 陸) with another (Lu vs Liu 呂 劉).

I think that there's a categorical difference here in the ambiguity: for lü 呂, we're approximating a sound [y] that isn't present in English, whereas for for liu 劉, we're transcribing a sound [ow] that is (basically) present in English, but just incorrectly identified. Here I think the onus is on the player to recognize that there is a distinction between lü 呂 ~ liu 劉, much as the onus is on the player to hear the French PG ("feen") and know that it denotes fine and not fin ("fan"). Hence, I propose the threefold distinction lu 陸 ("loo"), lü 呂 ("lew"), liu 劉 ("l'YOH" or "lee-YOH"), and have it be incumbent on the player to know that the latter two are distinct.

As state previously, I strongly agree with this idea, as erhuas is not present in conversational Taiwanese mandarin at all. However, a quick caveat: there are instances where the character "er" (兒) is used in a form of erhua, explicitly written in text, such as in character names (ex: 李瓶兒/Lee Ping'er). As I pointed out before, this would be the only case where writing out the "er" would be expected and required, which you then designate "er" as a separate character (Lee Ping Er) in the PG, rather than treating the erhua sound as a modification of the previous character (which is what some Northern Chinese dialects do, if I am correct).
If "Lee Ping'er" is the answerline (assume we want the full name), "Lee Ping" alone should not be acceptable, because that's not the character's name.

I am not a Sinologist, but my understanding is that erhua, as its name suggests, specifically denotes the phonological process by which various northern dialects rhotacize syllable codas (ping 瓶 /pʰiŋ/ > pingr 瓶兒 [pʰjɤ̃ʵ]), and not the general occurrence of rhotacized vowels in, say, èr 二 and various personal names. Hence, personal names like Lee Ping'er 李瓶兒 [pʰiŋ.ˀɚ ~ pʰjɤ̃ʵ.ˀɚ] would be outside the domain of any "no erhua" policy, and would continue to require the rhotic syllable as usual.

I disagree, since for 姐 喬 夏, I feel like the vowels parts do not warrant a PG for them (Q and X do, as stated previously). Since English does have diphthongs (a combination of two vowels), we may need to ensure no one pronounces "ie" as in "Tie", but we should be able to have western readers pronounce "iao" and "ia" correctly and would not warrant a PG.
Perhaps we can use "eh" as a vowel ending for the ㄝ (Pinyin: e, Wade-giles: Eh). So 姐 would be "Jie (PG: Jieh)" in a question text.

The concern with having multi-syllabic PGs (ex: jee-YAY) for one character is that it may be easy for a zealous but uninformed western reader to pronounce as "Jee Yay", which may be misinterpreted by players to be two characters (ex: 吉野 instead of 姊), which would not be good.

I agree with the risk that you describe for the two-syllable solution; I suppose the issue is whether it outweighs the risk of "j'YAY" or "j'YEH" being misread. I think that this is an empirical question: if readers have no problem with "j'YAY" / "j'YEH", then I will happily adopt this solution henceforth.

In my opinion, I think stress marks should only be present where in normal English contexts it is used (ex: fiancée), because stress marks aren't really a thing in Mandarin. If a reader stresses it one way or another, it doesn't affect the meaning, it just sounds weird (which, when factoring that most westerners don't speak tones properly, means that they already do "sound weird" in the eyes of a native speaker). Having readers stress the ultima just sounds even weirder.

My apologies; by the marking of stress in PGs I was referring to the use of capitalization to denote a stressed syllable (as in Xi Jinping "shee jeen-PEENG" rather than "shee jeen-peeng"). I agree that to a native Chinese ear, stressed ultimas sound unnatural, but this does happen to be how native English speakers usually pronounce Chinese names. Hence, I think that we should continue to PG Chinese names with the last syllable capitalized, since (a) most English speakers will unconsciously do this anyways, and (b) those that don't will be dissuaded from unconsciously stressing a different syllable.

[L.H.O.O.Q.]

**Pending more input on this - would "Lyu" or "Luh" be a better approximation for 呂?

Coming at this from someone who does not speak a lick of any Chinese language but recognizes the [\u] vs [y] from German and Turkish, the most recognizable Anglophone analogue to /呂/ [ly] would probably be "Lee." This has obvious problems if there is also a [li] syllable that forms a minimal pair with [ly]; however, this seems to me like a case where there's not going to be any one satisfactory mapping of all the phonemic vowels of Chinese to all the phonemic vowels of English, which means that concessions would have to be made.

I, as a Midwesterner, would pronounce a guide of "Luh" as [lʌ]. I think in Pinyin this sound would more closely resemble "Le," and this character /乐/ sounds close when I make the Google Translate TTS say it. I think [\i] is closer to [y] than [ʌ].

EDIT: I wonder if there's some way to get an English-speaker to produce [ʊ] (as in /look/ [lʊk]) through a pronunciation guide if it appears in an open syllable. That'd most likely be the way forward, but I have no idea how you'd get there - my best attempt at transcription is "leuh" but that'll obviously parse as a diphthong to people who're reading it blind.

[wadegilespromoter]

**Pending more input on this - would "Lyu" or "Luh" be a better approximation for 呂?

Coming at this from someone who does not speak a lick of any Chinese language but recognizes the [\u] vs [y] from German and Turkish, the most recognizable Anglophone analogue to /呂/ [ly] would probably be "Lee." This has obvious problems if there is also a [li] syllable that forms a minimal pair with [ly]; however, this seems to me like a case where there's not going to be any one satisfactory mapping of all the phonemic vowels of Chinese to all the phonemic vowels of English, which means that concessions would have to be made.

I, as a Midwesterner, would pronounce a guide of "Luh" as [lʌ]. I think in Pinyin this sound would more closely resemble "Le," and this character /乐/ sounds close when I make the Google Translate TTS say it. I think [\i] is closer to [y] than [ʌ].

EDIT: I wonder if there's some way to get an English-speaker to produce [ʊ] (as in /look/ [lʊk]) through a pronunciation guide if it appears in an open syllable. That'd most likely be the way forward, but I have no idea how you'd get there - my best attempt at transcription is "leuh" but that'll obviously parse as a diphthong to people who're reading it blind.

I would argue that "Lee" would be much worse than any exisiting system - because not only is it farther away from Lu or Loo, it raises confusion with Lee as well.
What about "ue"? I think even if it's parsed as a diphdong it would be relatively close as well. The concern is if someone parses as two separate vowels.

[23nl1us]

Thanks for raising this issue! I think you bring up a lot of good issues. I agree wholeheartedly that pronouncing Wade-Giles as if it were English should be accepted---as is the case for mispronunciations of Pinyin, HK Govt Romanization, SG-MY 'dialectical' romanization, etc. I also think that pronunciation guides should be standardized. (Actually as someone who does have some linguistics background I think there should be some attempt at creating a simplified form of IPA.)

However, there are some concerns I have. It is late where I am right now so I will have to read it more closely in the morning, and give it a more well-thought response it deserves. I just wanted to get some initial thoughts down while they are fresh. Apologies if some of what I mention is not accurate to what you said, and if my current message is a bit unpolished or unstructured.

Just to the point about allowing regional pronunciation. I do agree in spirit--I have sometimes felt the urge to answer Chinese tossups in something other than Mandarin, and when writing my own tossups on Chinese history I have thought hard about what to include in the answerlines. However, I worry that allowing too many may overwhelm readers (I think that in general Quizbowl answerlines need to be shorter not longer), while giving too few will create the same problem as the current Mandarin-centrism. Except, for better or worse, Mandarin has become the lingua franca of Sinophone world and is thus more "neutral" despite its Northern origins.

The problem is that many answerlines are quite broad. For example, dynasties. If you accept Cantonese and (Taiwanese) Hokkien, what about Teochew, Shanghainese, Hakka? If you're accepting these relatively mainstream varieties, what about less common varieties that still have millions of speakers in some cases, such as my grandmother's native variety of Min Chinese [which is not Southern or Eastern]. Actually, there are many non-Han ethnic groups which had prolonged contact with Han Chinese and developed reading systems for Chinese characters---in some cases these are people who were part of the dynasties in question, so it is part of their history as well. What do we do here?

Now, in practice the likelihood of someone both speaking such a language and choosing to answer in it is not particularly high. Generally, I think that if someone has enough knowledge of their (or their family's) mother tongue(s) to be able to use it to answer questions about history, then they probably have enough knowledge to describe in a more mainstream variety. However, Mandarin is already such a variety, and as I already mentioned has now become a lingua franca for the Sinophone world. If someone can say Hu Shih's name in Changshahua they probably can say it in Mandarin. It is a problem I see with a lot of answerlines that include alternative names or terms that would only be known to someone who is familiar enough with the answer to also know the more common name. I understand that sometimes the less common name is the first to come to mind---I've been negged before because a correct but obscure answer was not on the answerline---and that's especially for a mother tongue, but still worry about how to manage the scale.

I do think that in the case of answers that do pertain to a specific city or region, it should be acceptable to provide, but even that has issues. Let's say the answer is Fujian. Are we accepting answers in all the many mutually unintelligible varieties. This is a region (in)famous for its linguistic diversity where every 10 li has a different tongue. What is a "fair" amount of diversity in this case? (Also if we allow Hok-kian how do we know if the player meant to say the English term "Hokkien" which refers to the language; should it matter?).

I don't know if this is really relevant. I have a tendency to be pedantic sometimes, but I thought it would be better to get my thoughts out first. Like I said, I will re-read it in the morning and provide further comments if appropriate. Thank you again for bringing up the topic. The Chinese-speaking world and non-Japanese East and Southeast Asia in general is quite under- and at times misrepresented in Quizbowl, so it is very important to have these conversations.

Also: Lyu I think is better than Lu and Li for the /y/ vowel.

[23nl1us]

I believe my previous reply is still under review and I'm not feeling super well right now, but there is another point that came up after I read through the whole thread, which is the apparent (unless I missed it) conversation about the <zhi> <chi> <shi> (<chih> <ch'ih> <shih> if my WG is correct) and <zi> <ci> <si> (<tsu> <ts'u> <su>?) series. Since /i/ does not occur after retroflexes, nor after dentals (except in certain groups that merge the alveopalatals with the dentals, such as some younger speakers, especially girls and women, or speakers in Malaysia and Singapore). The actual vowels in <zhi> <chi> <shi> and <zi> <ci> <si> are very different from English and some linguists analyze them not as vowels but syllabic consonants /ʐ/ and /z/. I am not familiar with Bopomofo but I believe that there in fact is no 韵母 (rime, analgous to the nucleus and coda in Western syllabl analysis) symbol used when rendering these sequences using Zhuyin characters; rather the initial/consnant is used on its own. This poses as much of a challenge for non-Mandarin speakers and I'm not sure there is a clear solution. For the retroflexes, the vowel can have somewhat of rhotic quality, especially in Northern speakers, but for many speakers (including myself) this is less pronounced, even assuming that they split retroflexes at all. I have nonetheless experimented with the idea of using -r to mark certain vowels (cf. -ir in Tai-lo), although this might be more effective if the Anglophones in question were non-rhotic speakers of English and interpreted it (correctly) as more a vowel change than actual rhoticity (cf. "Shernberg" for Schoenberg and other errors introduced when pronunciation guides on umlauts meant for British speakers are adopted by North Americans). It would also, similar to what you pointed out, lead to potential ambiguity as to whether there is supposed to be erhua (although the Hanzi example has z- so it should not be rhotic in any case). As for the dentals, I sometimes find myself wondering if older Romanizations such as tsz or even tsu (the latter seemingly used in WG as well, unless I'm mistaken) would be more accurate.

Also, RE: stress marks, I actually find that marking "stress" is a possible way to get English speakers to produce some approximation of tone, although it can be hard to actually implement.

Edit: One final comment. I think that another area that could be improved is better accomodation of pronunciation variations within Mandarin (that is to say, Standard Chinese, Putonghua, Kuoyu, Huayu, etc.). One obvious example is <r>. As the existI believe my previous reply is still under review and I'm not feeling super well right now, but there is another point that came up after I read through the whole thread, which is the apparent (unless I missed it) conversation about the <zhi> <chi> <shi> (<chih> <ch'ih> <shih> if my WG is correct) and <zi> <ci> <si> (<tsu> <ts'u> <su>?) series. Since /i/ does not occur after retroflexes, nor after dentals (except in certain groups that merge the alveopalatals with the dentals, such as some younger speakers, especially girls and women, or speakers in Malaysia and Singapore). The actual vowels in <zhi> <chi> <shi> and <zi> <ci> <si> are very different from English and some linguists analyze them not as vowels but syllabic consonants /ʐ/ and /z/. I am not familiar with Bopomofo but I believe that there in fact is no 韵母 (rime, analgous to the nucleus and coda in Western syllabl analysis) symbol used when rendering these sequences using Zhuyin characters; rather the initial/consnant is used on its own. This poses as much of a challenge for non-Mandarin speakers and I'm not sure there is a clear solution. For the retroflexes, the vowel can have somewhat of rhotic quality, especially in Northern speakers, but for many speakers (including myself) this is less pronounced, even assuming that they split retroflexes at all. I have nonetheless experimented with the idea of using -r to mark certain vowels (cf. -ir in Tai-lo), although this might be more effective if the Anglophones in question were non-rhotic speakers of English and interpreted it (correctly) as more a vowel change than actual rhoticity (cf. "Shernberg" for Schoenberg and other errors introduced when pronunciation guides on umlauts meant for British speakers are adopted by North Americans). It would also, similar to what you pointed out, lead to potential ambiguity as to whether there is supposed to be erhua (although the Hanzi example has z- so it should not be rhotic in any case). As for the dentals, I sometimes find myself wondering if older Romanizations such as tsz or even tsu (the latter seemingly used in WG as well, unless I'm mistaken) would be more accurate.

Also, RE: stress marks, I actually find that marking "stress" is a possible way to get English speakers to produce some approximation of tone, although it can be hard to actually implement.

Edit:

One final comment (block of comments). I think that another area that could be improved is better accomodation of pronunciation variation within Mandarin (that is to say, Standard Chinese, Putonghua, Kuoyu, Huayu, etc.). One extremely common example is <r> which is alternatively realized by some speakers as an approximant similar (though rarely identical) to the English r, or a "zh" sound, like the g in certain pronunciations of "beige". Other speakers pronounce it like an English z, or do not differentiate between <r>, <l>, and/or <n>. The latter are proscribed and generally the result of influence from other varieties of Chinese, but extremely common. These are sound correspondances which would not be familiar to. Another example is merger of -n and -ng finals, particularly -in/-ing, -en/-eng, but also -an/-ang in some speakers. As someone with near-complete merger of -in/-ing in ordinary speech (in most cases I know the "correct" pronunciation but because I do not regularly produce or percieve the contrast it's not always easy to recall) and partial merger of -en/-eng, I am sometimes scared of making these mistakes that seem "obvious" in the context of Romanization but not as obvious in non-Romanized (whether written or oral) contexts.

Then there are retroflexes. Retroflexes are commonly cited as something that Western European speakers struggle to pronounce (presumably it is easier for Slavic speakers who have retroflexes as well), and in fact many Chinese do as well, realizations range from partial to full mergers, various intermediate pronunciation, differing extents of tongue curling (I tilt my tongue slightly, many more Northern speakers curl it more dramatically), etc. There are varieties of Chinese that do have retroflexes but have them in different words than in Mandarin, something their speakers carry over when speaking Mandarin. Given the above, to what extent should "Shu" vs "Su" actually be penalized? In my experience the affricates zh/z ch/c are already judged leniently (because English speakers also do not distinguish them well). I have not encountered ambiguous scenarios with sh/s, but since they do (loosely) correspond to a pair of contrasting phonemes in English, I worry that this may be judged differently.

This probably applies to some other languages as well. For example I have noticed that <r>, <g>, and <ch> in German is pronounced differently by different people, although I am not sure of the precise regional correlates in this case. Generally there is a lot of leeway for pronunciations of European languages but readers may be unaware of the kinds of variation that happen in non-European languages.

[Eddie]

I think that once an answer line threatens to have more than three or four answers, it's time to let the protest system take over. On the issue of Chinese dynasties, for example, I think an answer line in just English and Mandarin will be enough for the vast majority of cases. As John has noted, Mandarin is currently the prestige language in mainland China, and in the interest of keeping an answer line to a reasonable length, Mandarin should be the go-to choice of non-English answer, for better or for worse. Moreover, I think that an answer line in a language other than English and Mandarin should be given only if it is directly relevant to the content of the question. For example, a question on the Yuan dynasty could have an answer line in Mongolian, a question on Qing–Taiwan relations in Hokkien, and so on. The hypothetical answer line on Fujian, then, would be a rare situation in which an abundance of alternate answer lines could be justified. This is analogous to our existing treatment of European-language answer lines; a question on Spain gives an answer line in English and Castilian only; an additional answer in, say, Catalan, would be justified only if the question is on, say, Barcelona.

As to the issue of "Hok-kiàn" versus Hokkien, I think that the distinction is too minute for English speakers to realistically try to police.

As for the pronunciation of the syllabic sibilants, I think the best practice is to just transcribe in the way that an English speaker would expect (e.g. 是 shì ("shee")), despite the inaccuracy. Attempting to mimic a sound not present in English will, at best, yield a flawed rendition on the reader's part, and at worst, render the pronunciation incomprehensible to the listener. I feel the same way about using stress to mimic tone, and believe that we should stick to the usual English pronunciation in which stress falls on the final syllable of a compound word. My thoughts on this issue are informed by my position that the purpose of a pronunciation guide in quizbowl is to facilitate the reader's ability to verbally transmit a written phrase to the listener. In particular, I don't think that the purpose of a pronunciation guide in quizbowl is to accurately reproduce pronunciations internal to the language; at best, this is a welcome side effect. You may disagree with this position, in which case you will likely disagree with the rest of this paragraph, and we must simply agree to disagree.

[23nl1us]

As for the pronunciation of the syllabic sibilants, I think the best practice is to just transcribe in the way that an English speaker would expect (e.g. 是 shì ("shee")), despite the inaccuracy.

Yeah I think this is alright too. In my edit (pending review) to one of my previous replies I talk about variations in the pronunciaiton of Mandarin itself, and actually this would coincidentally be an example of that: there are definitely Mandarin speakers (mostly older speakers in, for instance, Cantonese- and Hokkien-speaking regions) who would pronounce 是 as something similar to "shee" when speaking what would sociolinguistically be considered "Mandarin" (or "regional Standard Mandarin" in the academic literature).

[wadegilespromoter]

The problem is that many answerlines are quite broad. For example, dynasties. If you accept Cantonese and (Taiwanese) Hokkien, what about Teochew, Shanghainese, Hakka? If you're accepting these relatively mainstream varieties, what about less common varieties that still have millions of speakers in some cases, such as my grandmother's native variety of Min Chinese [which is not Southern or Eastern]. Actually, there are many non-Han ethnic groups which had prolonged contact with Han Chinese and developed reading systems for Chinese characters---in some cases these are people who were part of the dynasties in question, so it is part of their history as well. What do we do here?

I think with regards to languages other than Mandarin - I generally agree (and it is implied) in my original post that the particular local language + Mandarin should be explicitly allowed, but other languages should not. For example, if the answerline talks about a geographical feature in Guandong, both the Mandarin and Cantonese answer should be explicitly acceptable (as well as the English, obviously), but the Taiwanese answer should not, or at the very least be directed through the protest system. This would also be the case with people of the early-modern/modern periods as well, as well as any cultural practices specifically studied in a particular region (sin-pu-a 媳婦仔/tongyangsi 童養媳 in Taiwanese and Mandarin respectively), for example.

Hopefully, this would serve the dual goals of preserving local languages and preventing answerline bloat.

--

For "general" terms, such as dynasty names, historical events, etc - I would argue that only the Mandarin (and English, of course) should be acceptable. However, consider the following question....

[10h] This two-month long battle starting in November 1948 targeting KMT headquarters in Xuzhou led to the end of the Chinese Civil War. This battle gave the communists control over central China and resulted in the PLA crossing the Yangtze and conquering Nanjing.
ANSWER: Battle of HuaiHai [or HuaiHai Campaign; or Battle of Hsupeng]

A bit of a backstory about this battle - the PLA (Communists) wanted to take the Lianghuai (兩淮) and Haizhou (海州) - which is why on their side, Huaihai (淮海) is what the battle is known as. On the other hand, the main ROC (KMT nationalists) were based in Hsuchou (徐州) and there were additional forces based in 蚌埠. This is why on the ROC side, this battle is known as Hsubang (徐蚌) (not Hsupeng).

Now, 蚌埠 has two pronounciations, the ROC pronounciation for this city is Bangfu (source), while the modern PRC pronounciation is Bengbu/Pengbu (source). Obviously the ROC would use the ROC pronounciations of the city, so this battle, in Taiwan, is known as Hsubang. (source

The important thing to note is that even in Mandarin, for a variety of reasons, there are pronounciation/terminology differences between that of China and Taiwan (due to the fact that they are two different countries).
As a pertinent example of what I believe will come up in one way or another in the next few ACF tournaments, the mythological figure Nezha (哪吒) has a different pronounciation in Taiwan's Mandarin, Nuozha (source - as the source notes, the Nezha pronounciation is of the Beiping [Beijing] pronounciation). Also, unlike what a recent Chinese movie may suggest, Nuozha (also known as Third Lotus Prince - 蓮花三太子) is seen as an important religious figure in Taiwan that adherents of Taoism (including yours truly) pray to. (whereas in China it's primarily only a mythological figure).

I do believe that when there are difference of Mandarin pronounciations between China and Taiwan, both pronounciations should be outright acceptable on the answerline, without need for protest (obviously, I, as a Taiwanese native speaker of Taiwanese and Taiwanese Mandarin, would stand to benefit from this - I would know what a thing is called in Taiwan, but I would not have any clue as to how the pronounciation in the PRC Mandarin would be).

[23nl1us]

I agree that differing standards should be acknowledged when appropriate (and actually, I go a little further in the edit that is still pending review, in terms of addressing the fact that "plausible mispronunciations" as are generally accepted for European terms, should encompass a different range of alternative pronunciations for Mandarin including r-l/n-l/f-h mergers)