Hard vs soft tonal languages

General discussion about learning languages
Sizen
Green Belt
Posts: 259
Joined: Sun Aug 30, 2015 5:53 am
Languages: English (N)
Studying: French, Japanese
Studied: Spanish, Mandarin
Language Log: viewtopic.php?f=15&t=2835&p=35968#p35968
x 706

Re: Hard vs soft tonal languages

Postby Sizen » Thu Jun 10, 2021 7:21 pm

Deinonysus wrote:As I understand it, although there are many patterns that a Japanese word can have, there is a much more limited number of patterns that a word can have given that a certain syllable receives the accent.

Ah, I see what you mean now. Yes, at least in Tokyo-type pitch accent Japanese variants, the placement of the accent determines the pitch of every syllable in a word. There are 5 other variants of pitch accent rules (2 of which are intermediate between Tokyo-type and another variety), and I don't know they're inner workings well enough to comment on them.

Deinonysus wrote:Now, it's hard to find any two papers that agree on how tone works in Xhosa, but based on my current limited understanding, tone isn't determined at the word level but at the morpheme level, so a prefix might have a certain tone pattern, and the root may have another certain tone pattern, and then a suffix may have another tone pattern, so there won't be an overall predictable tone pattern for an entire word based around any particular syllable.

So Xhosa is like Mandarin in this regard: each part of the word has its own inherent pitch contour and the pitch contour of one morpheme doesn't affect the pitch contour of other morphemes in a word (unless there are tone sandhi). I'd still be interested to know if Xhosa's tones can be identified in isolation, because that would also set it apart from Japanese. In Japanese, one-mora words with different pitch accents aren't distinguishable unless they are followed by a grammatical particle, the copula or another inherently accentless part of speech, at which point there would be a change in pitch which differentiates the two words. I've always assumed that this would not be the case with Xhosa and that you could tell the pitches apart regardless of their environment, but I don't know nearly enough about the language.


This is a bit of a tangent, but I think it might be interesting to share. While each morpheme in a Japanese word doesn't carry its own pitch contour, morphemes do actually affect the overall pitch contour of a word. There's an interesting memo out of the University of Osaka (in Japanese) that describers three types of particles (clitics) and auxiliary verbs (verbal morphemes) in regards to pitch accent: 乗っとられ型 ("hijacked" type), 乗っとり型 ("hijack" type), 協力型 ("cooperation" type).

乗っとられ型 ("hijacked" type)
The hijacked morphemes have their own accent that overwrites the accent of an accented verb but that is overwritten in an unaccented verb.

Examples (the bolded syllables are high pitch; they don't indicate the placement of the accent):

取る (to eat: to-ru) is an accented word with the accent on と to. If "H" is high and "L" is low, then it's pattern would be HL.

(さ)せる (sa)seru is a morpheme that attaches to the root of a verb and gives the verb a causal meaning. It doesn't make much sense to talk about it's pitch pattern in isolation because it doesn't occur in isolation, but it is normally accented on せ se.

If we combine the two we get 取らせる (とらせto-ra-se-ru), so LHHL. There is a drop after せ se because (さ)せる (sa)seru's accent is respected.

If we take an unaccented verb like 乗る (to ride: の no-ru), then the accent on (さ)せる isn't respected: 乗らせる (のらせる no-ra-se-ru). There is no drop after せ se because (さ)せる (sa)seru's accent is overwritten by the verb's unaccented pitch pattern.

Summary:
Accented verb:
取る (to eat: to-ru) HL
取らせる (とらせto-ra-se-ru) LHHL

Unaccented verb:
乗る (to ride: の no-ru) LH
乗らせる (のらせる no-ra-se-ru) LHHH

乗っとり型 ("hijack" type)
These morphemes hijack the accent of a word and force its accent on it in all cases.

A common example is the polite morpheme ます masu, which is accented on ま (ma). All words will take this accent no matter what.

To take the same verbs from before:
Accented verb:
取る (to eat: to-ru) HL
取ります (とりまto-ri-ma-su) LHHL

Unaccented verb:
乗る (to ride: の no-ru) LH
乗ります (のります no-ri-ma-su) LHHL

In both cases, there is a drop after ま ma because ます masu always enforces its accent.

協力型 ("cooperation" type)
These morphemes have their own accent, but never affect the accent of the word beforehand even if that means giving up its own accent.

An example would be そう sou which is used to indicate hearsay. It is accented on そ so. An accented word will keep its accent and cause そう sou to lose its accent, whereas an accented word will keep its accent and let そう sou keep its accent.

Accented verb:
取る (to eat: to-ru) HL
取るそう (るそう to-ru-so-u) HLLL

Unaccented verb:
乗る (to ride: の no-ru) LH
乗るそう (のるそno-ru-so-u) LHHL



Similar phenomena can happen to nouns with morphemes like らしい rashii which can have two different meanings depending on how it interacts with the noun’s accent. When it’s used to describe some form of hearsay, it follows the accent of the noun it follows, but when it’s used to create an adjective describing something that appears like the noun, then it overwrites that noun’s accent and enforces its own accent on し shi.

雨 (rain: あめ ame) is an accented word with the accent on あ a.

雨 (rain: a-me) HL

雨らしい (it looks like rain [i.e. they're announcing rain]: めらしい a-me-ra-shi-i) HLLLL or HLLHL. The accent of 雨 ame is respected.

雨らしい (rain-like: あめらしa-me-ra-shi-i) LHHHL. The accent of 雨 ame is not respected.



There are also suffixes in Japanese that affect the accent of a word. There are two according to the NHK Broadcasting Culture Research Institute:

後部一型 (koubuichi-gata) which indicates that the suffix will overwrite the accent of the word that it attaches to and have its accent on its first mora regardless whether or not the suffix has an accent if it can appear on its own.

前部末型 (zenbumatsu-gata) which indicates that the word to which the suffix is attaching will have its accent overwritten and placed on its final syllable (not mora)

And then there are compound words where each part keeps its respective accent...

Anyways... This is still quite predictable if you know the pitch accent pattern of the word and its morphemes, but it's just to show that knowing the pitch accent of a word on its own doesn't guarantee that you can predict its accent in every situation because morphology can shake things up quite a bit.
4 x

User avatar
leosmith
Blue Belt
Posts: 729
Joined: Thu Sep 29, 2016 10:06 pm
Location: Seattle
Languages: N:English
~C1: Spanish
~B2: French, Japanese, Korean, Mandarin, Russian, Thai, Tagalog, Swahili
Language Log: https://languagecrush.com/forum/t/2093
x 1665
Contact:

Re: Hard vs soft tonal languages

Postby leosmith » Fri Jun 11, 2021 12:11 am

Axon wrote:It's my understanding that the Thai, Lao, and Burmese scripts perfectly indicate tone because of the inherent tone class of a syllable plus a modifier marker if it should be read in a different tone than the tone class would imply. Is that true in practice?

For Thai, this is usually the case, but there are quite a few exceptions. For example, you actually have to memorize the fact that ผลไม้ is pronounced phohnR laH maaiH rather than phohnR maaiH. There are many examples of words for which you have to memorize which consonant belongs to which syllable, and if there are any hidden vowels, to get the correct pronunciation.

The reason I said it happens less often than not in Thai is that I misinterpreted the OP's statement as the script needing to have actual tone marks, which are only seen in about 1/4 of the words.

Deinonysus wrote:That was why I very carefully said a major writing system, not the main writing system.

I didn’t know that you consider pinyin to be a major writing system.
Hanzi don't directly show tone you would not confuse a word with one tone with a word with another tone, because they would use different characters. I have never studied Thai but as I understand it you can generally infer tone based on spelling rules.

Generally, but not always for either language.
In contrast, even pretty serious romanization systems for Japanese (writing Tōkyō instead of Tokyo, for instance), will not use pitch. Pitch would probably only show up in a dictionary.

I see what you mean now. You didn’t specify which writing system, so I assumed you meant the main one. Kanji makes words unique, so if you know the word, you know the pronunciation.
It is fiendishly difficult for a beginner to pick tones or tone patterns out of a sentence.

Not true, imo. I haven’t even studied Japanese pitch accent, other than to watch a couple short videos, and it’s very easy for me to pick out the tones. It took months for Thai and Mandarin.

That is the opposite of my experience.

If you listen to the tts for 箸 I’d be surprised if you couldn’t tell whether it was lo-hi or hi-lo. Regarding Mandarin, well that’s impressive. If you go here, are you able to tell the tones after listening to the audio (to display the answer, just click enter 3 times)? This particular test requires typing out the pinyin, so it tests initials and finals as well as tones. The defunct pinyinpractice used to do something similar, and beginners would rarely do well on it. But that may be because it tests more than tones.
My main point is that pitch accent vs tonal isn't always relevant to the question of whether or not you can get away with not studying tone.

This confused me, because then you said:
no aspect of pronunciation should be ignored. I am a big stickler for getting all the small details right that I am aware of.

So what benefit would you get from reclassifying languages?
Although pitch is a factor in a stress accent, there are not contrasting pitch patterns available on the accented syllable, so I don't agree that it would be appropriate to say English has a pitch accent.

Per Wikipedia, English is not tonal, and doesn’t have a pitch accent. But if we throw out existing classifications, and make up new rules, anything is possible. As I said, it depends where we draw the line.
0 x
https://languagecrush.com/reading - try our free multi-language reading tool

User avatar
Deinonysus
Blue Belt
Posts: 987
Joined: Tue Sep 13, 2016 6:06 pm
Location: Salem, MA, USA
Languages:  
• Native: English
• Advanced: French
• Intermediate: German,
   Spanish
• Beginner: Icelandic,
   Italian, Indonesian,
   Hebrew
x 3386

Re: Hard vs soft tonal languages

Postby Deinonysus » Fri Jun 11, 2021 10:24 am

Axon wrote:Mandarin words are frequently Romanized without tone for and by native speakers in the form of street signs and storefront signs. This happens all over China, whether or not the city is expected to have foreign tourists.

When introducing new words to native speakers (such as rare vocab words for middle school exams or obscure terminology at a museum) the character is written with the Pinyin next to it, always including tone. Chinese dictionaries intended for Chinese people have always had some way to mark tone, whether by dividing the whole book into sections based on tone or adding a small symbol next to the character.

There are some characters (fewer than 20 in common use by my estimate) that are used to represent two words with two different tones, such as 好: hao3 means "good" and hao4 means "to be fond of," appearing in some compounds. These are not marked in any way in texts for natives, though it's not unimaginable that a book using an obscure compound would provide the correct reading.

It's my understanding that the Thai, Lao, and Burmese scripts perfectly indicate tone because of the inherent tone class of a syllable plus a modifier marker if it should be read in a different tone than the tone class would imply. Is that true in practice?
Very interesting! I imagine that's due to lack of access to a proper keyboard layout that can write the tones. I've seen the same thing happen with Yoruba.

I didn't know about the two-way characters, that's good to know!

I've never studied Thai, let alone Lao or Burmese, so I can't comment on the spelling system beyond a couple of interesting videos I've seen about it.
0 x
Corrections welcomed!

Die Liebe liebt das Wandern, Gott hat sie so gemacht.

User avatar
Deinonysus
Blue Belt
Posts: 987
Joined: Tue Sep 13, 2016 6:06 pm
Location: Salem, MA, USA
Languages:  
• Native: English
• Advanced: French
• Intermediate: German,
   Spanish
• Beginner: Icelandic,
   Italian, Indonesian,
   Hebrew
x 3386

Re: Hard vs soft tonal languages

Postby Deinonysus » Fri Jun 11, 2021 10:44 am

Sizen wrote:So Xhosa is like Mandarin in this regard: each part of the word has its own inherent pitch contour and the pitch contour of one morpheme doesn't affect the pitch contour of other morphemes in a word (unless there are tone sandhi). I'd still be interested to know if Xhosa's tones can be identified in isolation, because that would also set it apart from Japanese. In Japanese, one-mora words with different pitch accents aren't distinguishable unless they are followed by a grammatical particle, the copula or another inherently accentless part of speech, at which point there would be a change in pitch which differentiates the two words. I've always assumed that this would not be the case with Xhosa and that you could tell the pitches apart regardless of their environment, but I don't know nearly enough about the language.

Wow, that post was packed with some really interesting nuances!

My main problem with identifying Xhosa tones is that I have not run into any learning materials that give more than a couple examples before abandoning the topic entirely, and there is not a reliable online source for tones. Tone information has recently been added to Wiktionary but it cites no sources and I don't trust it. I'm trying to get my hands on a dictionary set with tone markings for headwords but it can only be purchased from South Africa and it's slow going. Anyway, because of these difficulties I haven't had enough data on what tones should be to train my ear to know what to listen for. It's hard to identify even in isolated words because the tones follow a sweeping melody rather than going straight into a high or low tone directly, so it's particularly hard to tell a low tone from a falling tone.

A few words whose tones I've seen identified are:
  • mólò: Hello (loanword from Afrikaans "môre" and/or English "morrow", but it can be used any time of day).
  • èwé: yes
  • íthàngà: pumpkin
  • íthângà: colony or cattle outpost
  • íthàngá: thigh

In the examples beginning with í, the í is a noun class prefix and I believe it always takes a high tone.
0 x
Corrections welcomed!

Die Liebe liebt das Wandern, Gott hat sie so gemacht.

User avatar
Deinonysus
Blue Belt
Posts: 987
Joined: Tue Sep 13, 2016 6:06 pm
Location: Salem, MA, USA
Languages:  
• Native: English
• Advanced: French
• Intermediate: German,
   Spanish
• Beginner: Icelandic,
   Italian, Indonesian,
   Hebrew
x 3386

Re: Hard vs soft tonal languages

Postby Deinonysus » Fri Jun 11, 2021 11:07 am

leosmith wrote:For Thai, this is usually the case, but there are quite a few exceptions. For example, you actually have to memorize the fact that ผลไม้ is pronounced phohnR laH maaiH rather than phohnR maaiH. There are many examples of words for which you have to memorize which consonant belongs to which syllable, and if there are any hidden vowels, to get the correct pronunciation.

The reason I said it happens less often than not in Thai is that I misinterpreted the OP's statement as the script needing to have actual tone marks, which are only seen in about 1/4 of the words.

Generally, but not always for either language.

I see what you mean now. You didn’t specify which writing system, so I assumed you meant the main one. Kanji makes words unique, so if you know the word, you know the pronunciation.

If you listen to the tts for 箸 I’d be surprised if you couldn’t tell whether it was lo-hi or hi-lo. Regarding Mandarin, well that’s impressive. If you go here, are you able to tell the tones after listening to the audio (to display the answer, just click enter 3 times)? This particular test requires typing out the pinyin, so it tests initials and finals as well as tones. The defunct pinyinpractice used to do something similar, and beginners would rarely do well on it. But that may be because it tests more than tones.

This confused me, because then you said:
no aspect of pronunciation should be ignored. I am a big stickler for getting all the small details right that I am aware of.

So what benefit would you get from reclassifying languages?

Per Wikipedia, English is not tonal, and doesn’t have a pitch accent. But if we throw out existing classifications, and make up new rules, anything is possible. As I said, it depends where we draw the line.

Pinyin is major in my experience because I consistently see it in all learning materials for Mandarin whereas many learning materials for Norwegian and Xhosa do not mark tone at all. But I suppose it could be that it's never used by native speakers. I only have very passing familiarity with Mandarin.

I have never systematically studied Japanese pitch accent. My main experience with "soft" tonal languages is with Norwegian and Xhosa.

I tried a few examples on that site and it was the consonants that gave me trouble. I got the tones right every time, and I thought it would be hard to mistake them. But maybe my results are atypical and Mandarin tones just happened to click with me, and other people may not have the same experience that Mandarin is one of the easiest tonal languages to identify apparent tone in (I don't know any of the sandhi rules so I can't speak to actual underlying tone, just apparent tone).

My reason for proposing a new categorization system is that in my experience in dabbling in a large number of languages, there are two general types of learning experiences presented by tonal languages, and existing classifications such as "pitch accent vs tonal" or "simple vs complex tone" are inadequate to identify these two types of experiences. I think it's very useful to know what sorts of challenges you will face when learning a given language.

I merely identified that for "soft" tonal languages, people will commonly recommend against beginners learning tone. This is not an attitude I would ever take myself, but this attitude is an identifier of "soft" languages and is comorbid with a lack of proper tone drills in resources for "soft" tonal languages.
2 x
Corrections welcomed!

Die Liebe liebt das Wandern, Gott hat sie so gemacht.

User avatar
leosmith
Blue Belt
Posts: 729
Joined: Thu Sep 29, 2016 10:06 pm
Location: Seattle
Languages: N:English
~C1: Spanish
~B2: French, Japanese, Korean, Mandarin, Russian, Thai, Tagalog, Swahili
Language Log: https://languagecrush.com/forum/t/2093
x 1665
Contact:

Re: Hard vs soft tonal languages

Postby leosmith » Fri Jun 11, 2021 8:28 pm

Deinonysus wrote:I suppose it could be that it's never used by native speakers. I only have very passing familiarity with Mandarin.

Natives mostly use it for typing - they input pinyin (without tone marks) which gets converted to characters. There is also pinyin (with tone marks) in dictionaries to clarify pronunciation. Pinyin and other phonetic systems are important, but natives don't read texts in those systems.

I have never systematically studied Japanese pitch accent. My main experience with "soft" tonal languages is with Norwegian and Xhosa.

There are only 2 possibilities in Japanese, which makes distinguishing tones very easy.

maybe my results are atypical and Mandarin tones just happened to click with me, and other people may not have the same experience that Mandarin is one of the easiest tonal languages to identify apparent tone in

I don't know if your results are atypical; maybe nobody checks tones only.

My reason for proposing a new categorization system is that in my experience in dabbling in a large number of languages, there are two general types of learning experiences presented by tonal languages, and existing classifications such as "pitch accent vs tonal" or "simple vs complex tone" are inadequate to identify these two types of experiences. I think it's very useful to know what sorts of challenges you will face when learning a given language.

I merely identified that for "soft" tonal languages, people will commonly recommend against beginners learning tone. This is not an attitude I would ever take myself, but this attitude is an identifier of "soft" languages and is comorbid with a lack of proper tone drills in resources for "soft" tonal languages.

I realize you are talking about some languages I've never studied, but to me, the issue is not that the languages are miscategorized. I believe the main problem is lack of information about pronunciation in beginner material. Not just about tones/pitch accent, but pronunciation in general. For example, in my studies, there has been very little instruction regarding sentence level pronunciation, that is, how it differs from just the sum of the words. That's probably the biggest, most obvious, thing that is missing from pronunciation information for most languages. But individual languages often lack word level info too. As another example, in Tagalog there are simple stress patterns that exist in verb conjugations, and these are almost never covered even though most materials will tell you that stress is very important in Tagalog. Regarding pitch accent languages - I think they should provide the information required to learn it, along with a paragraph or two explaining why some choose not to.
5 x
https://languagecrush.com/reading - try our free multi-language reading tool

User avatar
devilyoudont
Blue Belt
Posts: 550
Joined: Tue Jun 26, 2018 1:34 am
Location: Philadelphia
Languages: EN (N), EO (C), JA (B)
Language Log: https://forum.language-learners.org/vie ... 15&t=16424
x 1682

Re: Hard vs soft tonal languages

Postby devilyoudont » Sun Jun 13, 2021 7:59 pm

I don't think this division makes sense. If you want to have a good accent, you need to study pitch if your target language has pitch, or tone if your target language has tone. If you are mainly concerned with learning languages primarily for reading, you can choose to neglect your pronunciation regardless of your target language.

My basic understanding about pitch vs tone... pitch creates a prominent syllable in a word similar to stress in English, but based primarily on tone rather than volume. Speakers of tonal languages do not perceive any particular syllable as being prominent based on tone.

Stuff like how badly your pronunciation will impact the ability for others to understand you is going to vary based on the specifics of your native language and your target language, as well as things you can't control like how often people who speak your target language interact with L2 speakers from your linguistic background.

In terms of why a generic English to Japanese course does not teach pitch, I believe it's only tangentially related to the utility of pitch in Japanese. These courses are mostly designed for American college students. A college course needs to give you some measurable progress towards something to justify itself to the administration and to students who want to take the class. This means that practically every aspect of the course will be incomplete in some way or other so that at the end of the semester it will look like you have accomplished a lot. Here's another example of the incompleteness of these courses: they tend to ignore aspect when teaching verb conjugations (to the later confusion of those students). Best case scenario, they are more or less hoping to give you a baseline which you can build off of in your time outside of class. When it comes to the unit on pronunciation, the teacher has much more important issues to combat: their American students' tendency to turn every vowel into a diphthong and obliterate any distinction between long and short vowels. Textbook writers know that this will be the case and tailor the unit to those ends. Pitch gets left out not because it is useless or unimportant, but because the diphthongs will create more immediate issues for an enterprising student who attempts to say hello to a random person. At least this is my theory.
3 x

User avatar
Saim
Blue Belt
Posts: 541
Joined: Sun Nov 15, 2015 12:14 pm
Location: Novi Sad
Languages: Native: English (AU)
Advanced fluency: Catalan, Serbian (+heritage), Spanish, Polish
Basic fluency: Hungarian, French, Galician, Asturian
?? (depends on register): Urdu
Intermediate (mostly passive): Hebrew, Punjabi, Russian, Portuguese, Italian, Occitan, Dutch, Turkish, German
Basic/dabbled: lots of Slavic languages, Romanian, Esperanto, Basque, Arabic, Mandarin
x 1661

Re: Hard vs soft tonal languages

Postby Saim » Mon Jun 14, 2021 12:37 am

devilyoudont wrote:My basic understanding about pitch vs tone... pitch creates a prominent syllable in a word similar to stress in English, but based primarily on tone rather than volume. Speakers of tonal languages do not perceive any particular syllable as being prominent based on tone.


I think that's a good description of Japanese pitch accent but doesn't really apply to everything that is referred to as pitch accent.

Serbian for example has both normal European-style stress and tonal patterns placed over it. There is a prominent syllable where there is also a drop in pitch. If after the drop the pitch rises back up, we call it a "rising" accent (if the stressed syllable is short, this rise occurs after the stressed syllable; if it's long, it occurs during it). If it doesn't rise back up, it's a "falling" accent. But I think that stress is still primarily distinguished by loudness, and so sounds more like, say, English stress than Japanese pitch accent; that's probably part of the reason why pitch accent is much less distinctive in Serbian than in Japanese.

Then there are languages like Turkish where stress is primarily marked by tone, but it's not generally analysed as having pitch accent since there's only really one pattern. Then again, you could reanalyse "final stress" (which occurs in the vast majority of words, so it seems to be the default form of "stress") in Turkish as a sort of heiban or acentless pattern (L-H) and forms of non-final stress would be words with an accent (H-L or L-H-L; i.e. H on the "stressed" syllable).

All this kind of points to the thesis that pitch accent isn't much of a coherent category. There are just different restrictions languages can place on their use of lexical tone.
7 x


Return to “General Language Discussion”

Who is online

Users browsing this forum: No registered users and 2 guests