oui c'est vrai, j'suis vraiment gêné. Je ne dirais jamais "frAnçais", toujours françAIs: c'était qu'une erreur. Et cette erreur bien que j'apprenne le français depuis des années! J'espère que les français me pardonneront. 😞
the stress-accent always on the last syllable means it's a predictable rule the inexistence of minimal pairs is equivalent to the stress being non-phonemic he wrongly put the fran in français in bold, but other than that he described it well note for learners: this applies to 'full' syllables. for example 'fenêtre' will have the stress on the 'ê', as the last 'e' is reduced i can't speak with great confidence but i think some dialects have slightly different stress rules, though it won't matter too much in being understood
@@kklein Oui c'est pas grave, je comprends. Ça me troublait aussi un peu d'abord mais en gros tu l'as bien expliqué. Je voulais tout simplement l'indiquer pour les débutants au français. Moi aussi je suis pas un locuteur natif mais je l'ai appris depuis quelques 7 ans. Alors c'est un petit détail mais je voulais y faire attention. J'espère que c'est plutôt constructif, j'aime trop ton contenu 👍
As a native Chinese speaker, this video is mind-blowing; I'd never given much thought to my language before. Despite the fact that the video is quite complicated, the knowledges presented in it continue to astonish me.
I never realized ðe 'c' in "cold" was different from ðe 'c' in "scold". Ngoldih a Ngand in front of my mouþ really ngelped me get a ngah of ngow aspiration works.
@@halohaalo2583 hah is a native english word but it can also be spelled ha because the second h there is reflecting the fact that we dont have enough vowel letters to represent all english vowel sounds so we use a consonant to modify the vowel like in c versus ch and s versus sh it's not actually an h consonant just a symbol that is written to modify the letter before it to make it more clear what the sound is supposed to be so this doesn't count as a minimal pair
same i took the nz flag as him being half kiwi or something like that, but then when i saw the flags of australia and quebec, i paused for a bit to think about the implications of using national symbols for linguistic separations
Fun fact: another language that has a zero onset in it's syllable structure is Korean and the Korean writing (essentially a letter based script) even has essentially its own character for it (which is the same as the ng sound at the end of syllable). To make it even more interesting there are special pronunciation rules if the previous syllable has a consonant at the end (batchim) to move its sound into the next syllable.
In Middle Korean, it is likely that the ㅇ which now stands for a zero onset was actually pronounced, usually thought to be [ɦ]. After the sound disappeared around the 16th century, ㅇ just came to stand for a zero onset.
@@challalla thanks I didn't know that. I just know that there were a few characters (mainly used in Chinese loan words), like p and z, that were dropped over time as their pronunciation became more "koreanized". This creates some funny transcriptions of now English loan words that would requires these phonemes but do no longer have the "proper" characters for them in Korean.
@@equinoxomega3600 You may be thinking of the letters ㅸ and ㅿ, which were pronounced something like [β] and [z] respectively and were lenited versions of ㅂ and ㅅ. In native Korean words, they were mostly found intervocally, though ㅿ also occurred in initial positions in transcriptions of Chinese characters. Later, [β] changed to [w] while [z] disappeared completely. So you have 고ᄫᅡ corresponding to modern 고와, and 무ᅀᅮ corresponding to modern 무 (earlier 무우). Seen in this light, ㅇ [ɦ] could also be seen as a lenited version of ㄱ which appeared intervocally before disappearing later, perhaps something like [ɡ → ɣ → ɦ → (zero)]. Or you may be thinking of ㆄ which was a weaker version of ㅍ (so maybe [ɸʰ]) but was only used in transcribing Chinese characters. There were attempts in the early 20th century to use archaic letters like ㅸ, ㆄ, ㅿ to represent [v, f, z] in loanwords from English and other languages, but they didn't take off because Korean speakers replaced them with native sounds anyway and Korean spelling is way more phonetic than English. English speakers might write Delhi with d and Gandhi with dh but pronounce them identically as [d] without knowing that they are different sounds in the original language, whereas Korean speakers write 델리 and 간디 using the same ㄷ for both sounds. So there is no need to spell punk and funk differently in Korean as they are both pronounced 펑크. That's a normal part of loanword adaptation in any language. Interestingly, more and more Korean speakers are sometimes pronouncing [f] or at least [ɸ] in loanwords, although not consistently. This isn't considered standard, but maybe if this does become more mainstream then there may be a case for resurrecting ㆄ.
@@challalla Thanks for the detailed information, yes I was referring to ㆄ and ㅿ. I only knew parts of the background to these characters, but as non-Korean living in Korea (but only poorly speaking the language), I always try transcribe my name using the ㆄ (for example when giving lectures), so that it is phonetically as close as possible to its pronunciation. However many Koreans are really confused by it, but it makes for a good conversation starter. Many Koreans are even more confused by another detail in the transcription of my name as they always assume (American) English pronunciation even for non-English names.
When I found your channel it was at 8k subs. After watching it, I clicked on another video and saw you had 9k subs. Now you're at 16k, I don't think I've ever witnessed such a meteoric rise in real time before
In my opinion, the neutral tone constitutes a separate toneme due to the minimal pairs present between words like "卵子" (luǎnzǐ) meaning "ovum/egg" and "卵子" (luǎnzi) meaning "testicles/balls". Although that might also depend on where you are, since Taiwan rarely uses the neutral tone at all.
I feel like you can also interpret the neutral toneme as a kind of phonemic stress as a character with a neutral toneme signifies that the stress lies on the character before
I guess another example would be 孙子（sūn zǐ/zi）meaning grandson, and 孙子（sūn zǐ(stressed)) meaning Sun Tzu the tactician. The stress is used to distinguish the words.
Native Mandarin speaker here and omg thank you so much for putting so much efforts into this! The thing here for us is that we would study all the pinyin stuff including all the vowels and how they pair with consonants etc etc when we were like 5 and by the age of 8 (third grade normally) that chart of what pairings are allowed are essentially second nature already and you wouldn’t even think twice about oh is that a legit mandarin sound or not. So again, thank you for making me relearning my kindergarten and first grade Chinese with these linguistic analysis that no way I would’ve understood or cared back then 😂
dylan wang- wait why does that name look so familiar oh wait, the rubik’s cube guy with a million subscribers has the EXACT SAME NAME! what a coincidence.
I'm curious how native speakers hear some stuff like 一. I've heard people say it ranging from what I heard as "yi" to "yil" and "il" (sorry for no ipa :( hope you get what I mean). Do you hear pronunciations of 他 / 她 that are closer to "ta" or "pa" as distinct? And related, do you hear/say 他 and 她 differently? Started learning 中文 recently and it's fascinating to me how wide pronunciations of some words can vary and how many words seem to be homophonous. But any language where the word fish looks like a drawing of a fish is amazing regardless :D
some regional dialects have the -il added to the end of some words. 他/她 are pronounced the same and even written the same until recently. I can't imagine how someone would mispronounce them as pa🙉
One thing worth noting is that while the palatal sibilants can be considered allophones of the retroflex sibilants (the Wade-Giles system worked like this), they could alternatively be considered allophones of the ALVEOLAR sibilants instead, as these all form minimal pairs with retroflex. This is the case with the Palladius transcription of Chinese (in other words the system used by Russia which is in Cyrillic). The only thing that can really be said for sure is that the palatal sibilants only occur before the palatal glides, so they could certainly be allophonic. It's just hard to know what they're allophonic TO, so pinyin AND bopomofo give them unique representations to sidestep the issue, while also including the triggers (the glides). Edit: According to another commenter the palatals sibilants could also be considered historical allophones of the velar series...
Note that a sound in a language can be an allophone to more than one phoneme, as weird as that sounds. If the sound exists in complimentary distribution to more than one phoneme, like the palatals to the velar stops, retroflex, and dental/alveolar affricatives in standard chinese, and the allophone is distinct from all three, it can be said that it is a single allophone to all three phonemes. Many linguistics dislike this sort of ambiguity though for some reasons, so some try to fix an allophone to a single phoneme. Edit: to help understand, it is best to know how this might come about in languages. Two phonemes/sounds can merge into a single one in certain contexts, but not others, resulting in two phonemes in one position, but one allophone in another. An example of how this might happen would be for example if a language would change /b/ into the fricative [v] between vowels, and then ALSO voice the fricative /f/ into [v] between vowels, making both phonemes have an intervocalic allophone that is the same sound.
@@ThorirPP It does sound pretty weird, but then you'd have to get into the definition of allophones and how they could potentially overlap with other phonemes. It's entirely possible that different palatals are allophonic to different series of phonemes and they've all fused together through the evolution of a language. So unless you wanna introduce historical spelling into something like pinyin, it's probably best to represent the palatal series as its own distinct thing even if it only shows up under specific circumstances
I don't know about the fricatives, but historically the affricates are not allophones of the retroflex series. The palatal affricates come from both velar stops and alveolar stops, the palatized allophones of both merging. The retroflex series originally had a consonant cluster with a medial -r- sound.
This is one of the most accurate summaries of the concepts of "phoneme" and "allophone" that includes some real world applications I've seen. Great job! Analyses aren't really typically picked in order to make it easy to understand a language's phonology, they are generally picked to make it easy to explain it. In this way, I'd be hesitant to include "it makes it easy to read X" as a real advantage of an analysis (don't worry: I know that you're also approaching this a bit more casually and are offering advantages that are beyond just scientific merit). What really matters is "can we predict certain processes we see happen in the language?" Sometimes this means providing an analysis that looks super fucked on the surface but ends up explaining the distribution of sounds extremely accurately. The two-vowel analysis for Mandarin isn't actually nearly as strange as you might be tempted to think. To understand it properly we need to understand what a vowel is in a phonological context. To simplify, all sounds have features that distinguish them from other sounds. The thing that distinguishes /d/ and /t/ in some languages is a feature called [voice], where we say that /d/ has the feature specification [+voice] and /t/ has [-voice]. The feature [voice], then, is used to distinguish voiced from voiceless consonants (/t/ and /d/ actually work a little differently in English but let's simplify their difference to being that of [voice]). The distinction between consonant and vowel also has a feature dedicated to it called [syllabic]. Consonants are [-syllabic] (because them existing doesn't add an extra syllable to a word) and vowels are [+syllabic] (because if there is a vowel, there is an extra syllable). Given that segments may form minimal pairs in which they differ in nothing but one feature, we would expect to have sounds that are identical except for the specification of their [syllabic] feature, and that is certainly what we get: think of the sounds /i/ (like the e in "eon") and /j/ (like the y in "yonder"). How do they differ in terms of quality? Does /j/ sound "harsher" and therefore more consonantal? Not really; the difference is pretty much only in their syllabicity: if you shift the stress in "eon" from the e to the o and then shorten the e a little, you'll see that it sounds identical to the "yon" in "yonder". So here we can say that /i/ and /j/ are identical except that the former is [+syllabic] and the latter is [-syllabic]. Thus: what we see here is that English has essentially pre-decided whether or not a particular sound is syllabic - /i/ is always [+syllabic] and /j/ is always [-syllabic], in the same way that /t/ is always [-voice] and /d/ is always [+voice]. In some other languages, it is completely 100% predictable whether a given sound will end up being voiced or voiceless just by knowing where they are in a sentence and/or the sounds around it. The technical way of putting this is that this language leaves the feature [voice] "underspecified". In English, there are some situations of sounds that are underspecified for the feature [syllabic]. Take the verb "to assemble". This word contains a final /l/, which is essentially pronounced as a vowel (i.e. is syllabic): /əˈsɛmbl̩/. If we were to form the noun meaning "a thing that assembles" by adding -er, then we end up with /əˈsɛmblər/, which has the same amount of syllables as assemble. Suddenly, the /l/ is not syllabic anymore, so we can say /l/ is underspecified for the feature [syllabic], since we can know whether it will end up being syllabic just by knowing what the sounds around it are. In other words: the syllabicity is determined after the word has been assembled. Now that I've rambled on for 9 paragraphs, how does this relate to that analysis of Mandarin? Well now imagine that Mandarin does the same: it leaves the segments /j/, /w/ and /y/ underspecified for syllabicity, therefore underlyingly not strictly vowels, as we can know whether they will end up sounding like vowels just by knowing what the sounds around it are. Fun fact: Proto-Indo-European is said to have had a similar system. The vowels [i] and [u] are said to just be syllabic equivalents of /j/ and /w/.
I also immediately thought of PIE when I saw the 2-vowel analysis for Mandarine. As a native speaker who've been taught the Pinyin system since I could remember, I found this pretty strange at first. But after thinking about it, it made a lot more sense. Together with your comment, it made the fact that PIE is analyzed to have 2 vowels no longer so bizarre as it first seemed to me.
Amazing explanation, those are pretty complex ideas presented in crystal clear language, I thought it would be a bore to read but it's actually so interesting I didn't realise there were so many paragraphs until you mentioned it (and overshooting by two in the process but we'll say it's for the sake of exaggeration and/or the love of the number 9, which I share). Well done for that and thank you, it's a great addition to the video itself.
okay you've been on the platform for 4 months (i mean the first video on this channel was 4 months ago) and already cracked the code -short enough for people to not get bored -jam packed with jokes and witty lines to keep people's attention -educational so people feel like they've learned something which makes them feel better about themselves and gives them the excuse for watching youtube all day for "educational purposes" -a slightly monotone sarcastic voice which is calm enough for people not to get annoyed quickly (as oppose to the OVERLY DRAMATIC INTRO VOICES some youtubers have) all in all amazing videos and channel so riddle me this: How have you only got 18.6k subscribers? (at the date of commenting)
@@shaheenbekk ohhh yeah true! He has some bangers as well hahaha If you like linguistics in general, you’ll love Xidnaf’s GradeAUnderA is more of a satyrical channel lol
im not studying chinese, but i am learning japanese. its so fascinating because this language (chinese) has always seemed like this extremely unapproachable , difficult wall to me. understanding the mechanics behind it a bit more is very helpful.
the 2 vowel system is really bizarre, it's basically loads of allophonic envrionment rules 2:47, also, [k]old will just sound like 'gold' cause it gets lumped together in onset 7:02 historically, a lot of the alveolo-palatals are allophones of velar obstruents, so ki -> qi-, gi-> ji-, hi -> xi. Cantonese, Min, Hakka, Japanese/Korean, etc still retain these. So 京 is Cantonese ging ([ɡ̥ɪŋ] [kɪŋ], same thing just marking aspiration so people don't mispronounce this) becomes Mandarin Personally, I don't think it's all that helpful to count the exact number of phonemes, you just need to know what to do with them. 8:52 well, we say "Mandarin" because it includes more than just the standardized variety (Putonghua, or "Standard Chinese") as most people don't actually speak the way it's taught but still remains mutually intelligible. In terms of language family, it's offhand for the whole Guan branch (官話) and not just Beijing-proximal speech Also, Mandarin refers to "high government bureaucrat of the Chinese Empire" (per. Wiktionary) and probably predates the fruit. You could say the name of the fruit is stupid, not the other way around. The _other_ issue I have with the term "Standard Chinese" is it subsumes all other Chinese topolects under the label "Chinese" and just claims "Standard Chinese" is representative of these varieties. That can be argued for the vocabulary and grammar in formal registers, but it absolutely doesn't apply for Phonology, which are so wildly different that you _have_ to consider them as different languages. If you use "Standard Chinese", then it's unclear to what extent the described processes apply to a speaker's own language production. If you use "Mandarin" you can safely assume it applies to varieties who are intelligible with Standard Chinese and aren't classified into another common topolect like Cantonese or Min.
[k]old being understood as 'gold' (and in general, unaspirated voiceless plosives in onset being perceived as their voiced counterparts by native English speakers) matches my experience of having to pronounce my surname Palacios the English way when I lived in the U.S., otherwise it would get consistently misspelled as having an initial B. Just as an aside, we Spanish speakers do a similar thing: we tend to perceive intervocalic voiced plosives as their voiceless counterparts, so for example a very common adaptation of kibbeh in Spanish is 'quipe'. That’s because voiced plosives in Spanish naturally surface as fricative or approximant allophones when intervocalic.
@@DonPaliPalacios 'and in general, unaspirated voiceless plosives in onset being perceived as their voiced counterparts by native English speakers' Ah yes 'sdop'
@@DonPaliPalacios It's proved that unaspirated voiceless sounds are actually no difference with voiced sounds in English native speakers' ears. Scientists cut off the s(/s/) sound of "spy"(/spaɪ/) and let native speakers listen. Nearly 95% of them said they heard "buy"(/baɪ/), rather than "pie"(/paɪ/).
One analysis that I came up with that I rarely see people use is a three-vowel analysis that conceptualizes three vertical vowels /a ə ɨ/ and redefines all surface vowels except those three as glide-vowel sequences. It's really parsimonious in that it accounts for every single final, collapses many of the consonant phonemes together and eliminates the weird syllabic consonant (by conceptualizing it as the realization of ɨ with no adjacent glides)
I have zero understanding in linguistics and had no idea what was going on for the majority of the video, but you still made it entertaining either way. Good video!
4:39 instead of analysing the "neutral tone" as a 5th toneme, I find it more meaningful to describe it as the lack of tone. There are syllables like 子 and 友 which have tones (both have the 3rd tone) in most words, but lose their tones at the end of certain words, e.g. 兔子, 朋友. So instead, a system of phonemic stress can be described for Standard Chinese (minimal pair: 孫子 sūnzǐ = both syllables stressed: "Sun Tzu" vs 孫子 sūnzi = only first syllable stressed: "grandson"), where only stressed syllables (which correspond to most syllables, though it varies between dialects) have tones while unstressed syllables do not ( = "neutral tone"). Stress-wise, since Chinese words/phrases are generally "initial-heavy", the neutral tone tends to occur at the end of compounds. This also explains why Mandarin dialects/accents which are more syllable-timed (e.g. Singaporean Mandarin, which I speak; also, Taiwanese Mandarin) than stress-timed (Beijing dialect is quite stress-timed) and don't really have any distinct "stress" usually don't feature the neutral tone as much or even at all (in such cases, the "original" tone of the syllable, that is, the tone which the syllable would have if it were stressed, is used instead).
I personally consider it the lack of any vowel, and thus unable to make up any tone. So maybe 兔子 should be written as tùz rather than tùzi, and people likely can still understand how to read it. Also your explaination about syllable-timed and stress-timed probably also explains why as a mainlander, I always find Taiwanese Mandarin sounding quite "soft". To be more specific though, I'm a native Cantonese that speaks "official Cantonese" accent which is even more stressed than most Mandarin accents, so that probably got me into stressing in Mandarin without much issue.
@@FlameRat_YehLon the 友 in 朋友 and the 叭 in 喇叭 also have the neutral tone, despite clearly having vowels. Though it is true that in some northern dialects, vowels get reduced in unstressed syllables (syllables with the neutral tone).
I was about to say the same thing.One of my Chinese teachers said that Mandarin actually has only 4 tones and that so called "neutral tone" is just the absence of any tone.
cool video! i've been into sinitic phonology for a while (mostly by reading wikipedia). you pretty well explained the unique issues that make it hard to define any single best analysis for mandarin. you mostly talked about vowels, but also super interesting is the palatal series which developed only in the last 300 years as a unique feature of mandarin. I will indicate Pinyin in brackets and IPA in slashes. the options, none of which are really satisfactory, are: 1. Analysis by native speakers: are phonemes, and not allophones of anything. This is due to widespread Pinyin education, of course everyone thinks they are separate sounds if they've written them with separate letters since kindergarten. In contrast, nobody in English thinks cold and scold have different /k/ sounds since they have the same letter. Analysis should typically rely on the intuition of natives who have NOT learned about phonology (go to some village and ask "do you think these are the same sound?"). Problem is, due to Pinyin, everyone has learned phonology. This method results in the highest phoneme count which may be considered "not efficient". 2. Analysis by sound resemblance: are allophones of retroflex series . They sound the most alike. Mandarin learners including the speakers of other Chinese languages are commonly mixing up /ʂaŋ/ with /ɕ(j)aŋ/. The two series are in complementary distribution. Problem is it's purely by coincidence that glides were lost after retroflexes. In other words, /ʂjaŋ/ used to exist in the past, where these two series formed minimal pairs and could not be allophones. 3. Analysis by history: are allophones of either the dental series OR the velar series . Pros of this: historically correct because the palatals are a full merger of dental and velars that occurred before glides. In other words, /t͡si, t͡sʰi, si/ and /ki, kʰi, hi/ used to exist, but both turned into /t͡ɕi, t͡ɕʰi, ɕi/. It is the reason for variant spelling of places like Peking/Beijing (velar-->palatal) and Tientsin/Tianjin (dental-->palatal). Cons of this: low sound resemblance. Presumably at some point in history, there were some "intermediate" sounds that were more similar. Another huge problem: to follow the rules of phonemes, you can't merge palatals into both series, nor split them based on individual words. You have to choose either dental or velar, and inevitably the history is wrong for half of the words. If you choose velars (the one with least sound resemblance) you get the slight advantage of being able to reduce the vowel count by merging the empty rime (aka syllabic consonant) /ɨ/ sound with the vowel /i/. Edit: some concrete examples of 尖團合流 (dental-velar merger). the unmerged sounds are reflected in modern Cantonese. 精清星 Middle Chinese corresponding to , Cantonese Mandarin 經傾興 Middle Chinese corresponding to , Cantonese Mandarin Interestingly, the numerous bilingual speakers of Cantonese (or other southern variety) plus Mandarin, have a subconscious awareness of these sound relationships and potential allophony.
Would you mind elaborating on why the syllabic consonant can only be merged with /i/ when the palatals are analysed as allophones of the velars? That's the only part I couldn't follow.
@@franciscoflamenco let's say you merge palatals into retroflex so /ʂ/ = /ɕ/, and then also try to merge /ɨ/ = /i/. Say these are each a pair of allophones, then by definition they cannot cause a change in meaning. This would imply that /ʂɨ/ and /ɕi/ (pinyin shi and xi) are exactly the same word, which is wrong. So either the consonant part or the vowel part must be a phonemic difference. The same thing happens for dentals. It does not happen with velars because velar+/ɨ/ does not exist, so then it's not a problem anymore.
i think 1 might be solved by analysis with the oldest possible generation of native speaker alive - people who received early education before pinyin was widespread should have some intuition not based on pinyin. the problem is though that those who didn’t receive pinyin education probably also didn’t receive mandarin education, so their intuition isnt native…
tbh having you pronounce the Chinese words, even if butchered, would make this a lot easier to understand. I watched this when it first came out, and had no idea what it was about. I watched it two or three times in between and now, after understanding all of the linguistic terminology, had a lightbulb moment after watching this video. After knowing mandarin IPA, I'm only now realizing that at 7:06, you're saying that "x" and "sh", "j" and "zh", and "q" and "ch" are allophones respectively, which I didn't realize without pinyin. With this aside, I do appreciate seeing language content being more objective by analyzing it with IPA, rather than the more common, less objective/analytical content on the web!
I would love for someone to make a video like this about Irish one day. There's actually disagreement amongst linguists about whether and are associated with seperate phonemes. Even weirder, some linguists even think that Irish has a vertical vowel structure with [ɪ] and [ʊ] being allophones of the same vowel, as well as [ɛ] and [ɔ]. They say that which realisation appears depends only on the surrounding consonants. There's also a weird tradition of marking what is clearly [ʌ] with [o] (because of it being spelt ?)
One language family where I've frequently seen the 'underlying glides, no vowels' analysis used is the Wakashan language family. Basically, all glides are called 'resonants', which includes /j, w, h/, which have the (potential) surface realizations of [i, u, a], respectively. When realized "vocalically", they're appropriately called 'vocalic resonants', while when realized "consonantally", they are 'consonantal resonants'-even weirder, glottal stop ia considered an allophone of /h/! The analysis is (as you can imagine) pretty controversial for some folks, but it persists for several Wakashan languages, notably Kwakw'ala, Haisla, and (I think) Nuu-chah-nulth
given that there's phonetic dinstinctions between /ɻ̩/ ([ɚ] in general, [ɑ˞] with the fourth tone (the "leaving tone" as like prefer to say), examples 而, 耳, 二) and /ɻɻ̩/ (just [ɻɻ̩], examples 日), I would highly disrecommend a model that intends to "put the syllablic consonant back into the other consonants".
I personally think it would be more elegant to analyze the syllables as /bʲɛ, kʰʷɑ/. The semivowels never felt like distinct sounds to me, as they tend to be pronounced simultaneously with the "preceding" plosive. Consecutive realization of plosive+semivowel sounds markedly non-native
8:57 About the name of the fruit coming from the name of the language and not the other way around, and how this supposedly makes it not a stupid name for the language: consider that Turkey is trying to convince people to start saying "Türkiye".
Minor point, but in French, stress goes on the last syllable of the intonation phrase, so in just "français," it's on /se/, but in the compound noun "français standard," it would be on the last part, and if that were at the beginning of the phrase, there wouldn't be any stress on those two words, since it'd come at the end of the whole phrase, so... The nitpicky thing is to say stress placement on that one French word is wrong, sorry. The more interesting takeaway is that French probably doesn't have lexical stress at all, and that it's important to distinguish between stress on the word level and on a phrase level. English has both, and they work independently, as in: "I didn't say HE loved her" vs "I didn't say he loved HER" vs "I didn't SAY he loved her" or phrasal/focus/sentential stress put on any other word in that sentence. In French, even that kind of focus stress shifting isn't possible, so you use other things like clefts instead. French really just doesn't have stress. Pick Spanish, haha.
This video reminds me of a convo I had with my parents on tones and how shi and xi, and chu and qu are pronounced differently which can confuse non native putonghua speakers. But as much as i find this interesting, there’s a lot I don’t understand as a layperson ;; eg tonemes, minimal pairs, segmental, suprasegmental?? It gets a little hard to follow as the video goes on... I want to learn about linguistics but I’m not sure where to start ;; is there any resources you can recommend? Or maybe make a video on the basics/terms?
a few people have asked me to make a video about more of the basics now, so I might do that actually. But fyi a toneme is a tonal phoneme. A minimal pair is a set of two words (eg. "cold" and "gold" used in the video), which are only separated bis y one sound (in the example, the only difference between "cold" and "gold" is that one has a /k/ sound and the other has a /g/ sound). "Segmental" and "suprasegmental" are a bit harder aha - segmental phonemes are like the individual sounds involved in making a word, which suprasegmental phonemes carry on over several segments - like the tones in Chinese, which carry through the entire word.
you can check out the crashcourse linguistics series for intro to linguistics! the youtube channel trevtutor is another great resource if you would like to dive a bit deeper into the linguistics rabbit hole ;)
great video! i am a chinese native speaker from Taiwan, studying how to teach chinese as second language.your video provided me a new perspective of how you learn chinese!
Good job! Your channel has some high quality stuff. Kinda remind me of Xidnaf [I just typed a 10-line comment about sibilants and dental-alveolar merger in Standard Chinese, and I discovered that someone has covered it in another comment way better than I could ever do.]
You could argue, somewhat in jest, that there is a minimal pair between /ŋ/ and /h/. The last name Nguyen, and a more conservative pronunciation of when. /ŋwɛn/ vs /hwɛn/.
But I'd argue (also somewhat in jest) that /hw/ is its own phoneme, since there's clearly minimal pairs between it and /h/, and between it and /w/, and posit the phoneme /ʍ/ (as some analyses do)
You are right, [i] and [ji] sound pretty much the same to me, only difference being the 2nd one is slightly louder, or [i] being the first tone, while [ji] being the 4th.
@1:33 The vertical vowel system is so true, as my professor told me before: so basically Chinese is one of the weirdest 30ish languages in the world whose "consonants (including liquids and glides) bears more weight than vowels in their syllables". So many languages that we know, like English, are "vowel dominant": meaning that their whole phoneme/phone system is evolved from vowels. Then we were used to thinking this way for the rest of the languages. However, we then observed that there are a handful of languages like Circacusan, Abaza, that have only two (or maybe only ONE @Karbadian) vowels but 134ish consonants and consonantal features. AND yes, these features are "PHONEMEIC" unlike the ones in English where they are only stylish (think about vocal fry in North American English but can actually word meaning in Circacusan). Anyhow, weird-aXX languages like Chinese and Circacusan keep us thinking: maybe vowels are not that important in our mental grammar.
I just recently stumbled upon that other video and was really happy to find another place in the rabbit hole of linguistics, and now there's already a follow up, just lovely! Keep it up
7:04 You should rather merge ‘j, q, x’ with 'g, k, h' or 'z, c, s', but not the retroflex consonants. The reason for the former one is that, 'j, q, x' are palatalised 'g, k, h', they share the same ancestry. For the latter one, in modern standard Chinese, 'j, q, x' absorbed all the syllables formed by 'z, c, s' + 'i', e.g. 'si' became 'xi', 'sia' became 'xia'. (N.B. real 'si' /si/ here, not the pinyin 'si') That's why they got the space to scribe 'sr' as 'si'.
I feel like when deciding how many phonemes a language has, it important to take historical linguistics into consideration as well. For example, the zero onset in Mandarin mostly came from the glottal stop initials and velar nasal initials in Middle Chinese.
In your last video, I noticed you didn’t count erhua syllables such as “ber” (贝儿). Any reasons for that? Also how are the unusual syllables used in interjections counted? For example, n (嗯), m (呣), hm (噷), and 哼 (hng).
As for interjections, I think they're considered paralinguistic, not linguistic. And as for erhua syllables, I don't think they're considered a feature of proper Modern Standard Chinese, they're just a feature of northern Mandarin dialects (which are the dialects that most resemble MSC, because it's based off of the Beijing dialect)
@@archkde students of Chinese do learn Chinese with at least some erhua. Usually there are alternative non-erhua versions like 这里 'here' instead of 这儿 'here' and adding -儿 'diminutive suffix' to the end of words is generally a very Northern Chinese thing to do (why say 小孩儿 'child' when 小孩/小孩子 does the trick?) but a handful of words have no non-erhua alternative like 二 'two', 儿子 'son', 耳朵 'ear' etc.
@@niku.. Correct me if i’m wrong, but afaik, erhua exclusively refers to 儿 being used as a suffix in a non-syllabic fashion, so words like 儿子 and other characters that have the pronunciation of “er” like 耳朵 don’t count.
Great video. My only real nit pick is with the section where you discussed how we know how some phonemes are different, like /h/ and /ŋ/. You mentioned that the two ways to do this is by looking at minimal pairs and by looking at how different the sounds are phonetically. But that latter method is actually very rarely used, if ever. Instead, most linguists when building phonology's and deciding phonemes have a few different approaches, for finding what phonemes a language has, and what allophones said phonemes themselves possess. The first method is obviously minimal pairs, which work great for many simple morphemes, like pin/bin, tip/dip, etc... But as you mentioned before, sometimes, some phonemes don't have minimal pairs to work with. In English, this would be /h/ and /ŋ/. However, we do know that they are different phonemes, not because they are phonetically different, but because they can exist in identical environments. /h/ can only exist in the onset, and /ŋ/ in the coda, yes, but they can also exist intervocalically. reheat /rihit/, dehiscent /dıhısǝnt/, singing /sıŋıŋ/, hangout /hæŋaʊt/, etc... Looking at words that place the phones in near identical or identical environments, we can then determine if said phones are independent phonemes, or allophones of one phoneme.
Minimal pairs follow-up video idea: /ʃ/ and /ʒ/ in English. The only sort-of minimal pairs I can think of for the two sounds are fission/fishin' (dialectal), shush/zhuzh (very dialectal), confusion/Confucian (which I would call valid, but some pedants might point out violates Scrabble's rule against proper nouns being words I guess), and leash/liege (also valid IMO but when's the last time someone used "liege" beyond a historical context?). Even if, sure, there do seemingly exist some tentative examples for ʃ/ʒ, it raises an interesting question: what if there weren't any? I'm sure there's some contextual rule about phonemes occupying the same "phonemic space" or whatever, but it seems like it could get confusing, controversial, and tantalizingly elusive to define. EDIT: as far I've been able to tell since, yes, linguists are fine with distinguishing phonemes based on being perceived differently when occurring in similar context, if not the exact same. Many English speakers would consider it a mistake to pronounce rouge "roosh" or potion "pozhun," so we could determine that they're perceived as different sounds. More examples of "questionable" pairs include: Aleutian/allusion(/illusion); measure/mesher; cash(/cache)/cazh (a clipping of "casual"); Haitian/Asian; hash/hajj
So minimal pairs are a sign of phonemic contrast, but they're not actually necessary for two sounds to be analyzed as phonemes. A language might coincidentally not even have such a pair of similar-sounding words. What's really important is that their distribution is unpredictable by rule - that is, you have no complimentary distribution. That's why the pleasure / pressure example is great even though the words aren't true minimal pairs. The sounds still exist in similar enough environments to suggest that environment is not a determining factor of which sound you get.
Your videos are really intriguing for me as a native Mandarin speaker! And no, I have no idea why the language is called Mandarin in English. Maybe the reason is lost among translations between European travelers centuries back. In primary school, I was taught that Chinese has six vowels (a o e i u yu) but your explanations make so much more sense in linguistic studies. Language barriers have created and will continue to create distrust and distorted views among people. I am learning my third language now, and looking forward to more quality videos from you!
A "mandarin" is a bureaucrat in the Imperial Chinese government (according to wikipedia, the term is probably a corruption of some portuguese word), so "mandarin chinese" was the variety spoken in the court. Britain mostly did business in Guangdong, so the variety they were most familiar with was cantonese and they just knew mandarin chinese as the language that the bureaucrats in Beijing spoke in.
Teachers of English to foreign students always seem to go overboard on the aspirated unvoiced plosives thing. The only people who I've noticed aspirate cluster-final plosives are non-native speakers, yanks who say fiftʰeen and sixtʰeen, and RealLifeLore all the bloody time.
8:50. Don't worry K Klein, I got you on why Standard Chinese is called Mandarin. Mandarin, as we know today was called 官話 or the 'language of the officials' when it was used as the adminstrative language of the Ming and Qing dyansties. In the 16th century, the Portugese, having established logistical colonies in India (Goa) and the Malay Peninsula (Melaka) were among the first Europeans to establish proper contact with the then, Ming dynasty China. Thus, they had the honour of translating and transcribing many Chinese words into the Latin script. The Portugese translated 官話 into 'mandarim', which was derived from the Malay word 'menteri' meaning minister. The Malay word itself has its origins in the Sanskrit word 'mantri', which again has the same meaning of minister. Both the Merriam-Webster and Oxford Dictionary’s acknowledge the origins of the word 'mandarin' in the Portugese 'mandarim' and was incorporated into the English language in the 16th century. So, it seems to check out. According to an article I read: Mandarin is what linguists call an exonym, a name given to a place, people or language externally from the place of origin.' This article talks about it in more detail: www.theatlantic.com/science/archive/2019/01/where-does-mandarin-come/579073/
As for why the fruit is also called mandarin, I don't have a credible source for that; but I have heard that since the fruit originated in China and shares the same colours of the yellow robes worn by Chinese officials and emperors, Europeans named the fruit as such to highlight the colour similarities.
In Korean language, P,T,K sounds are distinguished by three differences in aspiration. For example, /tal/(moon), /t͈al/(daughter), /tʰal/(mask) and /ki/(energy),/k͈i/(talent),/kʰi/(height).
/ŋ/ vs /h/ does matter when it comes to morphological alternation where /h/ may be yeetted, /ŋ/ goes to /ŋg/ (with a few exceptions) as such /ŋ/ is argued by some to be an allophone of /ng/ with the stop being unpronounced word finally (an other special word-morpheme)
Chinese has a set of consonants and a set of the 5 vowels you can use, but there's a whole list of the possible final segment of a word in pinyin that we learn and memorize in elementary school. I can't list all of them off the top of my head right now, but it starts like this: a, o, e, i, u, v(for notation reasons I'm using v, since chinese pinyin does not use v), ai, ei, ui, ao ou, iu etc. These can contain non vowel letters, for example ang, eng, ing, ong are all in this list. There's also another form of chinese word, which in pinyin form would be consonant - connecting vowel - vowel. Usually the connecting vowel is "i", but this list is pretty limited, so it only works with a small combination of connecting vowels and the list of final segments. One example would be i-ao, and a word with that could be xiao, or 小. There's a few final segments that can be used without a consonant, namely a, ai and er. I'm pretty sure these are the only ones. There's also a list of words that sound the same as either the consonant or the vowel. For example, zhi sounds like the consonant zh, and yi sounds like the vowel i.
I'm amazed by you. Found you on the start page and caught my attention immediately. I was looking for youtube channels about linguistics because either the few left upload videos very irregular or don't at all. I'm glad that I found you. Your content is superb. But one question: Why is your overlay language in word and youtube German but your location England, Uk? Are you German living in the UK? But if so why don't you sound German at all? How is this possible? Keep up with that content. I see that your subs are rising extremely fast :)
6:39 The z and the inverted r are not pronounced as z and r, both are actually vowels. In fact z and r are used to represent the vowels because other letters we usually use to represent vowels (a, e, i, o, u) have been used up. How z and the inverted r (which you use r to represent) sounds like is somewhere in between i and u.
I didn't understand most of it but i Guess i could start with learning about phoneme. Anyhow, I like your manner of speech and how you explain things. Keep it up.
One thing to note: French does not have any stress at all. It makes use of intonation patterns which lay over entire phrases and clauses, meaning that individual words may alter in their accentuation depending on the surrounding sentence and the pragmatic context around it
But the stress is always on the last syllable of any such segment. Which means that when french speakers try learning other languages, they'll first learn individual words and will put the stress on the last syllable of those individual words
@@bacicinvatteneaca I don't think you're right. All syllables are (supposed) to be given equal weight. This is now seen as a very, very "formal" or "old-school" way to speak, associated with, say, journalists or old teachers, but it is still the standard. Or at least it is in my dialect. However, in a question (for example) the intonation will be rising all throughout the sentence. I think Parisians also have a rising tone, but I am not sure. A good word for it would be in my opinion "civilisation". You're expected to say "la cI-vI-lI-sA-tION".
@@bacicinvatteneaca No. There may be a phrase final high intonation tendency but it’s certainly not a rule. Personally when it comes to isolated words with no conversational context i tend to keep the intonation flat
i'd say at least in southern france, we tend to put the stress on the last syllable, or second-to-last when we pronounce the -e (personally, i do about 50% of the time inside sentences, but some people do it much more often), and secondary stress before any word-internal e (as in développement, 2 or three stressed syllables)
Im not a native english speaker and thank you for providing me with the ability to just check how perfect my english accent is. I do indeed say the k of kold and skold very differently when checked by air ammount as the first has much the other has none. And I dont even know how to asperate something, I so struggle with this trying to say latin or PIE or ... words. Knowing that everything is unasperated in my language I also checked if I speak english with a latvian accent what then happens and yes neither skold nor kold have any air reaching my hand.
It's cool to see another way minimal pairs are used. As an English teacher, I use minimal pairs in games to help force my students to hear English phonemes (say, by having students try to guess the differences between baa and bar so they can hear the r final better, or having them find either fan and van in a scavenger hunt). I've always been kind of sad that Chinese phonotactics is too restrictive to allow people to even make up plausibly fake minimal pairs like xao/shao (as x + ao must have a glide, xiao). It would certainly help non-native speakers develop phonemic awareness of the sh/zh/ch xi/ji/qi distinction.
As someone currently in the process of learning chinese this is hitting pretty close to home. ji and qi sound very similar to me, to the point where I think I can hear the difference between 人 and 认 better than 俏 and 叫 despite my native language not being tonal. What really fascinates me is that there are words which seem to have a lot of variance in pronunciation like 一 or 他 / 她. Also if I'm hearing it correctly and 他 and 她 are really homophonous, that's crazy. Not having to specify the gender of someone when referring to them in 3rd person while speaking, but kinda having to do so while writing is pretty funny. (better than my native tongue, there we always have to specify -.-)
@@chalkchalkson5639 this is because 她 was a recent invention originally used to translate "she" in English textbooks- originally 他 was all-inclusive and had the same meaning as 她/它/牠
6:42 it’s funny because it makes sense. In Miyako (a Japonic language) i and u when between consonants (and other places in some dialects) become syllabic s and f like pstu “person” (cognate to Japanese hito) or tabz “journey” (cognate to Japanese tabi) so it makes complete sense. The r syllabic consonant probably happened as z merged in place to the retroflex consonants it was after. The theory from Miyako is that it became ɨ before breaking out of the vowel space to z (then in the case of Miyako dialects, assimilating in voice)
The analysis with three glides and two vowels is based on historical Chinese phonology. There's an entire system established though the likes of 切韻 (600 CE), 韻鏡 (c. 10th Century), and 廣韻 (c. 1008). The system aims to help poets find words that rhyme in Middle Chinese, not to minimize the number of phonemes in Modern Standard Chinese. The linguists who made bopomofo were trained in Chinese Classics so they had an affinity for the system. Furthermore, the linguists were envisioning a system that can transliterate all Chinese languages, and some of those languages have inherited these phonotactics from Middle Chinese.
Actually the vowels in 安an and 昂ang are different but written as if they were the same. As a native speaker, I can say most of Chinese people don't really get the difference between [ŋ] and [n], we use the vowels to differentiate them. Vowels in Chinese may be a lot more than you expected.
Chinese Zero to Hero just made a video about this. Actually, it makes sense. N is articulated more in the front of the mouth than NG, so naturally the vowel preceding it would be more in the front, as well. It makes it easier to pronounce.
A question that troubles me from time to time: How are vowels different from (certain) syllabic consonants? How many vowels are there that cannot be regarded as syllabic consonants? BTW: According to my experience, the zero-onset is often realized as either a glottal stop, or as ng before /a/ and schwa, but not when a glide is present, so I would consider it as an initial in its own right, or a glide equivalent at the very least. (The ng variant is not relevant to the 疑母 initial ng from Middle Chinese, as it is also witnessed in e.g. 愛., which does not have a ng initial in MC.)
Vowels are vocalic, so no air is being obstructed when you say the phoneme. Syllabic consonants may be sonorant, meaning that air is not entirely stopped at any point, but the tongue and/or lips are used to obstuct sounds so they are not considered vocalic.
At 5:40 I don't really understand why it's said that "o" and "e" both represent the vowel ə, which is not really accurate. They both represent different sounds: o is /o/ while e is normally /ɤ/ but can also be /ɛ/ if preceeded by /j/ or /ɥ/... So, pinyin uses 6 symbols to represent 7 vowel sounds, note that zhuyin does in fact use 7 symbols. On the other hand, some linguists consider the syllables "zi, ci, si, zhi, chi, shi, ri" to have a /ɨ/ vowel, while it can also be considered as being a syllabic consonants as explained in this video.
I would be interested to see historical linguistics used more in these phonemic analyses, since complementary distribution often comes from sound changes to a single phoneme- we can see for example that scold and cold's /k/ also come from the same historical origin, or that the chinese palato-alveolar series [tɕ], [tɕʰ], [ɕ] stems from the velar series /k/, /kʰ/, /x/ and thus makes sense to analyse as allophones of the same velar series. Calling them allophones of the retroflex series seems a little weird- its functional, but ahistorical, and both moves reduce the phonemic inventory the same amount. This way analyses have more historical grounding and give some insight into dialectal differences.
For example, this makes Cantonese, which retains the velar series where Mandarin has the palato-alveolar, make more sense. We can simply say they are the same phoneme, but Mandarin has a different allophonic pronunciation.
This also adds some more context to why certain phonemic pairings exist. If we include Japanese [h] and [f] under the same phoneme /h/ since they exist in perfect complementary distribution, it makes more sense if we also consider the historical sound change that turned [f] to [h] except before /u/. They exist in complementary distribution because they come from the same phoneme that simply gets pronounced differently in a certain context. [h] -> [f] is a rather far-fetched sound change, and so knowing that [f] -> [h] happened makes grouping them under a single /f/ or /h/ more reasonable.
i would categorize syllabic consonant as 0 final imo like in bopomofo there just isn't any and i would merge the alveolar-palatal series as either velar of alveolar bc of historical sound changes
I don't really like using historical sound change as the sole reason for classification. It's valid of course but not necessary and we are trying to analyze the modern language and how speakers percieve things right now. The alveolo-palatals are indeed in complementary distribution with the velars, the alveolar affricates and the retroflex consonants so it's difficult to assess where to put them.
The advantage of saying the alveolar-palatals are allophones of the velars specifically is that you don't need to include the syllabic consonant as a separate phoneme.
I'd heard the idea floated that Bopomofo would've likely been more accepted over Pinyin had the "Initial + Final + Tone" symbols been arranged into syllabic blocks (like what Hanguel does). What are your thoughts?
That would definitely fit closer with the general shape of the Hanzi. However, Zhuyin (or pinyin) isn't used in the language except for marking the sound, unlike how Hangul is used for Korean or Kana for Japanese. So it shouldn't matter that much.
So" It's "i" before "e", expect after "c", unless sounded as "a" as in neighbor and weigh? Your explanation is wonderful. (not my crude joke, above) Do you know of Simon Roper on UA-cam? You two should do some videos together. I'm excited for it.
To be more accurate the vowel in "keng" is more like /ɤ/, you can hear that when an average Chinese speaks English: sometimes they pronounce the schwa vowel like /ɤ/.
2:59 Fun fact, the Korean alphabet DOES consider them as allophones of the same phoneme, as Hangul uses one letter to represent BOTH the glottal stop initial consonant AND the /ng/ final consonant...
4:26 No, French accent is always on the last syllable of the word, that's why there are no minimal pairs
ikr? found that strange, never said frAncais
oui c'est vrai, j'suis vraiment gêné. Je ne dirais jamais "frAnçais", toujours françAIs: c'était qu'une erreur. Et cette erreur bien que j'apprenne le français depuis des années! J'espère que les français me pardonneront. 😞
the stress-accent always on the last syllable means it's a predictable rule
the inexistence of minimal pairs is equivalent to the stress being non-phonemic
he wrongly put the fran in français in bold, but other than that he described it well
note for learners: this applies to 'full' syllables. for example 'fenêtre' will have the stress on the 'ê', as the last 'e' is reduced
i can't speak with great confidence but i think some dialects have slightly different stress rules, though it won't matter too much in being understood
@@aiocafea Je crois que le québécois est un peu différent, mais je peux pas le confirmer
@@kklein Oui c'est pas grave, je comprends. Ça me troublait aussi un peu d'abord mais en gros tu l'as bien expliqué. Je voulais tout simplement l'indiquer pour les débutants au français. Moi aussi je suis pas un locuteur natif mais je l'ai appris depuis quelques 7 ans. Alors c'est un petit détail mais je voulais y faire attention. J'espère que c'est plutôt constructif, j'aime trop ton contenu 👍
As a native Chinese speaker, this video is mind-blowing; I'd never given much thought to my language before. Despite the fact that the video is quite complicated, the knowledges presented in it continue to astonish me.
As someone who's learning Chinese it amazes me that you guys have a fully working language whilst using so few sounds
very few people actually think about their native language a lot.
@@flubadubdubthegreat1272 Extremely simple grammar and a small number of sounds. We are all about simplicity 😂
You have just 12 videos but the quality of your channel is comparable with UA-camrs that have 4+ years of experience. Keep it up!
Damn! I was about to embark on a binge of their channel only to realise that I watched them all already.
@@jan_franzke Ikr? i wish he'd started 10 years ago
Not really, but he does remind me of Xidnaf. I miss him :'(
@@jan_franzke
Their*
The creator of this video uses they/them pronouns.
I can't believe that omg, I thought they would've made hundreds of videos 🥺 they're all such high quality
I never realized ðe 'c' in "cold" was different from ðe 'c' in "scold". Ngoldih a Ngand in front of my mouþ really ngelped me get a ngah of ngow aspiration works.
minimal pair hah vs hang
hah should be considered a native english word
This feels like I’m having a stroke.
@@halohaalo2583 hah is a native english word but it can also be spelled ha because the second h there is reflecting the fact that we dont have enough vowel letters to represent all english vowel sounds so we use a consonant to modify the vowel like in c versus ch and s versus sh it's not actually an h consonant just a symbol that is written to modify the letter before it to make it more clear what the sound is supposed to be so this doesn't count as a minimal pair
@@yesid17 that was super interesting to read, and a good explanation, but I think it was even harder to read than the original comment 😭 why no commas
ʍɒt ðə..
4:07 Love the flag usage, really shows how idiotic it is to represent languages with national flags
I didn't even notice it was unusual. Union Jack and fleur-de-lis, yep makes sense.
same
i took the nz flag as him being half kiwi or something like that, but then when i saw the flags of australia and quebec, i paused for a bit to think about the implications of using national symbols for linguistic separations
you should have used 🇮🇳 for maximal confusion
Those very some *really* english and french icons though.
I assumed he used those because it was the accent he was basing the analysis on, which would be a smart use of national flags for that sort of thing
Fun fact: another language that has a zero onset in it's syllable structure is Korean and the Korean writing (essentially a letter based script) even has essentially its own character for it (which is the same as the ng sound at the end of syllable). To make it even more interesting there are special pronunciation rules if the previous syllable has a consonant at the end (batchim) to move its sound into the next syllable.
In Middle Korean, it is likely that the ㅇ which now stands for a zero onset was actually pronounced, usually thought to be [ɦ]. After the sound disappeared around the 16th century, ㅇ just came to stand for a zero onset.
@@challalla thanks I didn't know that.
I just know that there were a few characters (mainly used in Chinese loan words), like p and z, that were dropped over time as their pronunciation became more "koreanized". This creates some funny transcriptions of now English loan words that would requires these phonemes but do no longer have the "proper" characters for them in Korean.
@@equinoxomega3600 You may be thinking of the letters ㅸ and ㅿ, which were pronounced something like [β] and [z] respectively and were lenited versions of ㅂ and ㅅ. In native Korean words, they were mostly found intervocally, though ㅿ also occurred in initial positions in transcriptions of Chinese characters. Later, [β] changed to [w] while [z] disappeared completely. So you have 고ᄫᅡ corresponding to modern 고와, and 무ᅀᅮ corresponding to modern 무 (earlier 무우). Seen in this light, ㅇ [ɦ] could also be seen as a lenited version of ㄱ which appeared intervocally before disappearing later, perhaps something like [ɡ → ɣ → ɦ → (zero)].
Or you may be thinking of ㆄ which was a weaker version of ㅍ (so maybe [ɸʰ]) but was only used in transcribing Chinese characters.
There were attempts in the early 20th century to use archaic letters like ㅸ, ㆄ, ㅿ to represent [v, f, z] in loanwords from English and other languages, but they didn't take off because Korean speakers replaced them with native sounds anyway and Korean spelling is way more phonetic than English. English speakers might write Delhi with d and Gandhi with dh but pronounce them identically as [d] without knowing that they are different sounds in the original language, whereas Korean speakers write 델리 and 간디 using the same ㄷ for both sounds. So there is no need to spell punk and funk differently in Korean as they are both pronounced 펑크. That's a normal part of loanword adaptation in any language. Interestingly, more and more Korean speakers are sometimes pronouncing [f] or at least [ɸ] in loanwords, although not consistently. This isn't considered standard, but maybe if this does become more mainstream then there may be a case for resurrecting ㆄ.
@@challalla Thanks for the detailed information, yes I was referring to ㆄ and ㅿ. I only knew parts of the background to these characters, but as non-Korean living in Korea (but only poorly speaking the language), I always try transcribe my name using the ㆄ (for example when giving lectures), so that it is phonetically as close as possible to its pronunciation. However many Koreans are really confused by it, but it makes for a good conversation starter. Many Koreans are even more confused by another detail in the transcription of my name as they always assume (American) English pronunciation even for non-English names.
@@challalla but isn't ㅇthe NG sound?
When I found your channel it was at 8k subs. After watching it, I clicked on another video and saw you had 9k subs. Now you're at 16k, I don't think I've ever witnessed such a meteoric rise in real time before
18.5k 20 minutes later. Crazy!
minutes later at 18.5k
And this is totally deserved too, these are amazing
19k!
In my opinion, the neutral tone constitutes a separate toneme due to the minimal pairs present between words like "卵子" (luǎnzǐ) meaning "ovum/egg" and "卵子" (luǎnzi) meaning "testicles/balls". Although that might also depend on where you are, since Taiwan rarely uses the neutral tone at all.
I feel like you can also interpret the neutral toneme as a kind of phonemic stress as a character with a neutral toneme signifies that the stress lies on the character before
Not really though. The tone for the word "卵" is different when spoken due to the tone sandhi
I guess another example would be 孙子（sūn zǐ/zi）meaning grandson, and 孙子（sūn zǐ(stressed)) meaning Sun Tzu the tactician. The stress is used to distinguish the words.
that pair has got to make sexual education in taiwan really difficult.
@@spcxplrr Dunno about Taiwan but for Standard Mandarin, the use of that word is very infrequent at best, there’s far more common alternatives instead
Native Mandarin speaker here and omg thank you so much for putting so much efforts into this! The thing here for us is that we would study all the pinyin stuff including all the vowels and how they pair with consonants etc etc when we were like 5 and by the age of 8 (third grade normally) that chart of what pairings are allowed are essentially second nature already and you wouldn’t even think twice about oh is that a legit mandarin sound or not. So again, thank you for making me relearning my kindergarten and first grade Chinese with these linguistic analysis that no way I would’ve understood or cared back then 😂
dylan wang- wait why does that name look so familiar
oh wait, the rubik’s cube guy with a million subscribers has the EXACT SAME NAME!
what a coincidence.
@@txtp Wang and Dylan are both incredibly common names, plus your odds go up considering you're watching a video on the Chinese language in English
I'm curious how native speakers hear some stuff like 一. I've heard people say it ranging from what I heard as "yi" to "yil" and "il" (sorry for no ipa :( hope you get what I mean). Do you hear pronunciations of 他 / 她 that are closer to "ta" or "pa" as distinct? And related, do you hear/say 他 and 她 differently? Started learning 中文 recently and it's fascinating to me how wide pronunciations of some words can vary and how many words seem to be homophonous. But any language where the word fish looks like a drawing of a fish is amazing regardless :D
some regional dialects have the -il added to the end of some words. 他/她 are pronounced the same and even written the same until recently. I can't imagine how someone would mispronounce them as pa🙉
@@shuyangjiang9466 maybe not full pa, but one of the people recorded for duolingo definitely has a fair bit of labial plosive in there
One thing worth noting is that while the palatal sibilants can be considered allophones of the retroflex sibilants (the Wade-Giles system worked like this), they could alternatively be considered allophones of the ALVEOLAR sibilants instead, as these all form minimal pairs with retroflex. This is the case with the Palladius transcription of Chinese (in other words the system used by Russia which is in Cyrillic). The only thing that can really be said for sure is that the palatal sibilants only occur before the palatal glides, so they could certainly be allophonic. It's just hard to know what they're allophonic TO, so pinyin AND bopomofo give them unique representations to sidestep the issue, while also including the triggers (the glides).
Edit: According to another commenter the palatals sibilants could also be considered historical allophones of the velar series...
Note that a sound in a language can be an allophone to more than one phoneme, as weird as that sounds.
If the sound exists in complimentary distribution to more than one phoneme, like the palatals to the velar stops, retroflex, and dental/alveolar affricatives in standard chinese, and the allophone is distinct from all three, it can be said that it is a single allophone to all three phonemes.
Many linguistics dislike this sort of ambiguity though for some reasons, so some try to fix an allophone to a single phoneme.
Edit: to help understand, it is best to know how this might come about in languages. Two phonemes/sounds can merge into a single one in certain contexts, but not others, resulting in two phonemes in one position, but one allophone in another.
An example of how this might happen would be for example if a language would change /b/ into the fricative [v] between vowels, and then ALSO voice the fricative /f/ into [v] between vowels, making both phonemes have an intervocalic allophone that is the same sound.
@@ThorirPP It does sound pretty weird, but then you'd have to get into the definition of allophones and how they could potentially overlap with other phonemes. It's entirely possible that different palatals are allophonic to different series of phonemes and they've all fused together through the evolution of a language. So unless you wanna introduce historical spelling into something like pinyin, it's probably best to represent the palatal series as its own distinct thing even if it only shows up under specific circumstances
I don't know about the fricatives, but historically the affricates are not allophones of the retroflex series. The palatal affricates come from both velar stops and alveolar stops, the palatized allophones of both merging. The retroflex series originally had a consonant cluster with a medial -r- sound.
@@ThorirPP So like the alveolar tap in English dialects that have that
This is one of the most accurate summaries of the concepts of "phoneme" and "allophone" that includes some real world applications I've seen. Great job!
Analyses aren't really typically picked in order to make it easy to understand a language's phonology, they are generally picked to make it easy to explain it. In this way, I'd be hesitant to include "it makes it easy to read X" as a real advantage of an analysis (don't worry: I know that you're also approaching this a bit more casually and are offering advantages that are beyond just scientific merit). What really matters is "can we predict certain processes we see happen in the language?" Sometimes this means providing an analysis that looks super fucked on the surface but ends up explaining the distribution of sounds extremely accurately.
The two-vowel analysis for Mandarin isn't actually nearly as strange as you might be tempted to think. To understand it properly we need to understand what a vowel is in a phonological context.
To simplify, all sounds have features that distinguish them from other sounds. The thing that distinguishes /d/ and /t/ in some languages is a feature called [voice], where we say that /d/ has the feature specification [+voice] and /t/ has [-voice]. The feature [voice], then, is used to distinguish voiced from voiceless consonants (/t/ and /d/ actually work a little differently in English but let's simplify their difference to being that of [voice]). The distinction between consonant and vowel also has a feature dedicated to it called [syllabic]. Consonants are [-syllabic] (because them existing doesn't add an extra syllable to a word) and vowels are [+syllabic] (because if there is a vowel, there is an extra syllable).
Given that segments may form minimal pairs in which they differ in nothing but one feature, we would expect to have sounds that are identical except for the specification of their [syllabic] feature, and that is certainly what we get: think of the sounds /i/ (like the e in "eon") and /j/ (like the y in "yonder"). How do they differ in terms of quality? Does /j/ sound "harsher" and therefore more consonantal?
Not really; the difference is pretty much only in their syllabicity: if you shift the stress in "eon" from the e to the o and then shorten the e a little, you'll see that it sounds identical to the "yon" in "yonder". So here we can say that /i/ and /j/ are identical except that the former is [+syllabic] and the latter is [-syllabic].
Thus: what we see here is that English has essentially pre-decided whether or not a particular sound is syllabic - /i/ is always [+syllabic] and /j/ is always [-syllabic], in the same way that /t/ is always [-voice] and /d/ is always [+voice]. In some other languages, it is completely 100% predictable whether a given sound will end up being voiced or voiceless just by knowing where they are in a sentence and/or the sounds around it. The technical way of putting this is that this language leaves the feature [voice] "underspecified". In English, there are some situations of sounds that are underspecified for the feature [syllabic]. Take the verb "to assemble". This word contains a final /l/, which is essentially pronounced as a vowel (i.e. is syllabic): /əˈsɛmbl̩/. If we were to form the noun meaning "a thing that assembles" by adding -er, then we end up with /əˈsɛmblər/, which has the same amount of syllables as assemble. Suddenly, the /l/ is not syllabic anymore, so we can say /l/ is underspecified for the feature [syllabic], since we can know whether it will end up being syllabic just by knowing what the sounds around it are. In other words: the syllabicity is determined after the word has been assembled.
Now that I've rambled on for 9 paragraphs, how does this relate to that analysis of Mandarin? Well now imagine that Mandarin does the same: it leaves the segments /j/, /w/ and /y/ underspecified for syllabicity, therefore underlyingly not strictly vowels, as we can know whether they will end up sounding like vowels just by knowing what the sounds around it are.
Fun fact: Proto-Indo-European is said to have had a similar system. The vowels [i] and [u] are said to just be syllabic equivalents of /j/ and /w/.
I also immediately thought of PIE when I saw the 2-vowel analysis for Mandarine. As a native speaker who've been taught the Pinyin system since I could remember, I found this pretty strange at first. But after thinking about it, it made a lot more sense. Together with your comment, it made the fact that PIE is analyzed to have 2 vowels no longer so bizarre as it first seemed to me.
Great comment!
Um I might read this later. The replies will remind me.
Amazing explanation, those are pretty complex ideas presented in crystal clear language, I thought it would be a bore to read but it's actually so interesting I didn't realise there were so many paragraphs until you mentioned it (and overshooting by two in the process but we'll say it's for the sake of exaggeration and/or the love of the number 9, which I share). Well done for that and thank you, it's a great addition to the video itself.
@@gautiermilewski4322 Thank you for reminding me to read the comment!
You are filling the hole Xidnaf left for me. Keep up the videos dude!!
who the hell is Xindfg?
Damn, this guy is pumping out some great contents!
he has a really promising future
okay you've been on the platform for 4 months (i mean the first video on this channel was 4 months ago) and already cracked the code
-short enough for people to not get bored
-jam packed with jokes and witty lines to keep people's attention
-educational so people feel like they've learned something which makes them feel better about themselves and gives them the excuse for watching youtube all day for "educational purposes"
-a slightly monotone sarcastic voice which is calm enough for people not to get annoyed quickly (as oppose to the OVERLY DRAMATIC INTRO VOICES some youtubers have)
all in all amazing videos and channel
so riddle me this: How have you only got 18.6k subscribers? (at the date of commenting)
This is the exact same formula that people like Xidnaf and GradeAUnderA used and honestly it’s fckin awesome I ADORE these kinda videos
@@deithlan oh idk those channels i was kinda thinking more cassualy explained but i'll check out those channels for sure
@@shaheenbekk ohhh yeah true! He has some bangers as well hahaha
If you like linguistics in general, you’ll love Xidnaf’s
GradeAUnderA is more of a satyrical channel lol
The channel's first video is only 5 months old. That's pretty new for UA-cam, all things considered
im not studying chinese, but i am learning japanese. its so fascinating because this language (chinese) has always seemed like this extremely unapproachable , difficult wall to me. understanding the mechanics behind it a bit more is very helpful.
I mean China do be known for its huge, (once) extremely unapproachable, difficult wall
Japanese is a whole different ballgame in terms of phonology so...
the 2 vowel system is really bizarre, it's basically loads of allophonic envrionment rules
2:47, also, [k]old will just sound like 'gold' cause it gets lumped together in onset
7:02 historically, a lot of the alveolo-palatals are allophones of velar obstruents, so ki -> qi-, gi-> ji-, hi -> xi. Cantonese, Min, Hakka, Japanese/Korean, etc still retain these. So 京 is Cantonese ging ([ɡ̥ɪŋ] [kɪŋ], same thing just marking aspiration so people don't mispronounce this) becomes Mandarin
Personally, I don't think it's all that helpful to count the exact number of phonemes, you just need to know what to do with them.
8:52 well, we say "Mandarin" because it includes more than just the standardized variety (Putonghua, or "Standard Chinese") as most people don't actually speak the way it's taught but still remains mutually intelligible. In terms of language family, it's offhand for the whole Guan branch (官話) and not just Beijing-proximal speech
Also, Mandarin refers to "high government bureaucrat of the Chinese Empire" (per. Wiktionary) and probably predates the fruit. You could say the name of the fruit is stupid, not the other way around.
The _other_ issue I have with the term "Standard Chinese" is it subsumes all other Chinese topolects under the label "Chinese" and just claims "Standard Chinese" is representative of these varieties. That can be argued for the vocabulary and grammar in formal registers, but it absolutely doesn't apply for Phonology, which are so wildly different that you _have_ to consider them as different languages. If you use "Standard Chinese", then it's unclear to what extent the described processes apply to a speaker's own language production. If you use "Mandarin" you can safely assume it applies to varieties who are intelligible with Standard Chinese and aren't classified into another common topolect like Cantonese or Min.
[k]old being understood as 'gold' (and in general, unaspirated voiceless plosives in onset being perceived as their voiced counterparts by native English speakers) matches my experience of having to pronounce my surname Palacios the English way when I lived in the U.S., otherwise it would get consistently misspelled as having an initial B.
Just as an aside, we Spanish speakers do a similar thing: we tend to perceive intervocalic voiced plosives as their voiceless counterparts, so for example a very common adaptation of kibbeh in Spanish is 'quipe'. That’s because voiced plosives in Spanish naturally surface as fricative or approximant allophones when intervocalic.
@@DonPaliPalacios 'and in general, unaspirated voiceless plosives in onset being perceived as their voiced counterparts by native English speakers' Ah yes 'sdop'
Yeah I thought it was weird lol. That’s why I asked
@@DonPaliPalacios It's proved that unaspirated voiceless sounds are actually no difference with voiced sounds in English native speakers' ears. Scientists cut off the s(/s/) sound of "spy"(/spaɪ/) and let native speakers listen. Nearly 95% of them said they heard "buy"(/baɪ/), rather than "pie"(/paɪ/).
@@Anonymous-df8it watch geoff lindsey's video to disgover how english is really sboken
One analysis that I came up with that I rarely see people use is a three-vowel analysis that conceptualizes three vertical vowels /a ə ɨ/ and redefines all surface vowels except those three as glide-vowel sequences. It's really parsimonious in that it accounts for every single final, collapses many of the consonant phonemes together and eliminates the weird syllabic consonant (by conceptualizing it as the realization of ɨ with no adjacent glides)
Never in my life have I been confused by my native language more than I am at this very moment. Cheers, I guess?
Welcome to General Linguistics
same here.
I have zero understanding in linguistics and had no idea what was going on for the majority of the video, but you still made it entertaining either way. Good video!
Same
Yeah same with me I don’t really understand what you were saying with allophones and segments and suprasegements and stuff
Ughh these videos and comments are simultanionsly very intresting and also make me feel like I know nothing.
I found you just yesterday, watched the chinese video just yesterday, and here you are. life's good. and you deserve 100 times the subs and the views
4:39 instead of analysing the "neutral tone" as a 5th toneme, I find it more meaningful to describe it as the lack of tone. There are syllables like 子 and 友 which have tones (both have the 3rd tone) in most words, but lose their tones at the end of certain words, e.g. 兔子, 朋友.
So instead, a system of phonemic stress can be described for Standard Chinese (minimal pair: 孫子 sūnzǐ = both syllables stressed: "Sun Tzu" vs 孫子 sūnzi = only first syllable stressed: "grandson"), where only stressed syllables (which correspond to most syllables, though it varies between dialects) have tones while unstressed syllables do not ( = "neutral tone"). Stress-wise, since Chinese words/phrases are generally "initial-heavy", the neutral tone tends to occur at the end of compounds.
This also explains why Mandarin dialects/accents which are more syllable-timed (e.g. Singaporean Mandarin, which I speak; also, Taiwanese Mandarin) than stress-timed (Beijing dialect is quite stress-timed) and don't really have any distinct "stress" usually don't feature the neutral tone as much or even at all (in such cases, the "original" tone of the syllable, that is, the tone which the syllable would have if it were stressed, is used instead).
I personally consider it the lack of any vowel, and thus unable to make up any tone. So maybe 兔子 should be written as tùz rather than tùzi, and people likely can still understand how to read it.
Also your explaination about syllable-timed and stress-timed probably also explains why as a mainlander, I always find Taiwanese Mandarin sounding quite "soft". To be more specific though, I'm a native Cantonese that speaks "official Cantonese" accent which is even more stressed than most Mandarin accents, so that probably got me into stressing in Mandarin without much issue.
@@FlameRat_YehLon the 友 in 朋友 and the 叭 in 喇叭 also have the neutral tone, despite clearly having vowels. Though it is true that in some northern dialects, vowels get reduced in unstressed syllables (syllables with the neutral tone).
I was about to say the same thing.One of my Chinese teachers said that Mandarin actually has only 4 tones and that so called "neutral tone" is just the absence of any tone.
@@FlameRat_YehLon I doubt that since in 吗 the vowel "a" is still heard, even if it lacks tone.
@@FlameRat_YehLon also, the lack of true vowels does not prohibit tones, as syllabic consonants in words like 四and 纸 are capable of carrying tone
cool video! i've been into sinitic phonology for a while (mostly by reading wikipedia). you pretty well explained the unique issues that make it hard to define any single best analysis for mandarin. you mostly talked about vowels, but also super interesting is the palatal series which developed only in the last 300 years as a unique feature of mandarin. I will indicate Pinyin in brackets and IPA in slashes. the options, none of which are really satisfactory, are:
1. Analysis by native speakers: are phonemes, and not allophones of anything. This is due to widespread Pinyin education, of course everyone thinks they are separate sounds if they've written them with separate letters since kindergarten. In contrast, nobody in English thinks cold and scold have different /k/ sounds since they have the same letter. Analysis should typically rely on the intuition of natives who have NOT learned about phonology (go to some village and ask "do you think these are the same sound?"). Problem is, due to Pinyin, everyone has learned phonology. This method results in the highest phoneme count which may be considered "not efficient".
2. Analysis by sound resemblance: are allophones of retroflex series . They sound the most alike. Mandarin learners including the speakers of other Chinese languages are commonly mixing up /ʂaŋ/ with /ɕ(j)aŋ/. The two series are in complementary distribution. Problem is it's purely by coincidence that glides were lost after retroflexes. In other words, /ʂjaŋ/ used to exist in the past, where these two series formed minimal pairs and could not be allophones.
3. Analysis by history: are allophones of either the dental series OR the velar series . Pros of this: historically correct because the palatals are a full merger of dental and velars that occurred before glides. In other words, /t͡si, t͡sʰi, si/ and /ki, kʰi, hi/ used to exist, but both turned into /t͡ɕi, t͡ɕʰi, ɕi/. It is the reason for variant spelling of places like Peking/Beijing (velar-->palatal) and Tientsin/Tianjin (dental-->palatal). Cons of this: low sound resemblance. Presumably at some point in history, there were some "intermediate" sounds that were more similar. Another huge problem: to follow the rules of phonemes, you can't merge palatals into both series, nor split them based on individual words. You have to choose either dental or velar, and inevitably the history is wrong for half of the words. If you choose velars (the one with least sound resemblance) you get the slight advantage of being able to reduce the vowel count by merging the empty rime (aka syllabic consonant) /ɨ/ sound with the vowel /i/.
Edit: some concrete examples of 尖團合流 (dental-velar merger). the unmerged sounds are reflected in modern Cantonese.
精清星 Middle Chinese corresponding to , Cantonese Mandarin
經傾興 Middle Chinese corresponding to , Cantonese Mandarin
Interestingly, the numerous bilingual speakers of Cantonese (or other southern variety) plus Mandarin, have a subconscious awareness of these sound relationships and potential allophony.
Would you mind elaborating on why the syllabic consonant can only be merged with /i/ when the palatals are analysed as allophones of the velars? That's the only part I couldn't follow.
@@franciscoflamenco let's say you merge palatals into retroflex so /ʂ/ = /ɕ/, and then also try to merge /ɨ/ = /i/. Say these are each a pair of allophones, then by definition they cannot cause a change in meaning. This would imply that /ʂɨ/ and /ɕi/ (pinyin shi and xi) are exactly the same word, which is wrong. So either the consonant part or the vowel part must be a phonemic difference. The same thing happens for dentals. It does not happen with velars because velar+/ɨ/ does not exist, so then it's not a problem anymore.
@@jasonfang9903 Thank you, that's a great explanation.
i think 1 might be solved by analysis with the oldest possible generation of native speaker alive - people who received early education before pinyin was widespread should have some intuition not based on pinyin. the problem is though that those who didn’t receive pinyin education probably also didn’t receive mandarin education, so their intuition isnt native…
tbh having you pronounce the Chinese words, even if butchered, would make this a lot easier to understand. I watched this when it first came out, and had no idea what it was about. I watched it two or three times in between and now, after understanding all of the linguistic terminology, had a lightbulb moment after watching this video. After knowing mandarin IPA, I'm only now realizing that at 7:06, you're saying that "x" and "sh", "j" and "zh", and "q" and "ch" are allophones respectively, which I didn't realize without pinyin. With this aside, I do appreciate seeing language content being more objective by analyzing it with IPA, rather than the more common, less objective/analytical content on the web!
I would love for someone to make a video like this about Irish one day. There's actually disagreement amongst linguists about whether and are associated with seperate phonemes. Even weirder, some linguists even think that Irish has a vertical vowel structure with [ɪ] and [ʊ] being allophones of the same vowel, as well as [ɛ] and [ɔ]. They say that which realisation appears depends only on the surrounding consonants. There's also a weird tradition of marking what is clearly [ʌ] with [o] (because of it being spelt ?)
One language family where I've frequently seen the 'underlying glides, no vowels' analysis used is the Wakashan language family. Basically, all glides are called 'resonants', which includes /j, w, h/, which have the (potential) surface realizations of [i, u, a], respectively. When realized "vocalically", they're appropriately called 'vocalic resonants', while when realized "consonantally", they are 'consonantal resonants'-even weirder, glottal stop ia considered an allophone of /h/!
The analysis is (as you can imagine) pretty controversial for some folks, but it persists for several Wakashan languages, notably Kwakw'ala, Haisla, and (I think) Nuu-chah-nulth
given that there's phonetic dinstinctions between /ɻ̩/ ([ɚ] in general, [ɑ˞] with the fourth tone (the "leaving tone" as like prefer to say), examples 而, 耳, 二) and /ɻɻ̩/ (just [ɻɻ̩], examples 日), I would highly disrecommend a model that intends to "put the syllablic consonant back into the other consonants".
I personally think it would be more elegant to analyze the syllables as /bʲɛ, kʰʷɑ/. The semivowels never felt like distinct sounds to me, as they tend to be pronounced simultaneously with the "preceding" plosive. Consecutive realization of plosive+semivowel sounds markedly non-native
8:57 About the name of the fruit coming from the name of the language and not the other way around, and how this supposedly makes it not a stupid name for the language: consider that Turkey is trying to convince people to start saying "Türkiye".
Minor point, but in French, stress goes on the last syllable of the intonation phrase, so in just "français," it's on /se/, but in the compound noun "français standard," it would be on the last part, and if that were at the beginning of the phrase, there wouldn't be any stress on those two words, since it'd come at the end of the whole phrase, so...
The nitpicky thing is to say stress placement on that one French word is wrong, sorry.
The more interesting takeaway is that French probably doesn't have lexical stress at all, and that it's important to distinguish between stress on the word level and on a phrase level. English has both, and they work independently, as in: "I didn't say HE loved her" vs "I didn't say he loved HER" vs "I didn't SAY he loved her" or phrasal/focus/sentential stress put on any other word in that sentence. In French, even that kind of focus stress shifting isn't possible, so you use other things like clefts instead. French really just doesn't have stress. Pick Spanish, haha.
This video reminds me of a convo I had with my parents on tones and how shi and xi, and chu and qu are pronounced differently which can confuse non native putonghua speakers.
But as much as i find this interesting, there’s a lot I don’t understand as a layperson ;; eg tonemes, minimal pairs, segmental, suprasegmental?? It gets a little hard to follow as the video goes on... I want to learn about linguistics but I’m not sure where to start ;; is there any resources you can recommend? Or maybe make a video on the basics/terms?
a few people have asked me to make a video about more of the basics now, so I might do that actually. But fyi a toneme is a tonal phoneme. A minimal pair is a set of two words (eg. "cold" and "gold" used in the video), which are only separated bis y one sound (in the example, the only difference between "cold" and "gold" is that one has a /k/ sound and the other has a /g/ sound). "Segmental" and "suprasegmental" are a bit harder aha - segmental phonemes are like the individual sounds involved in making a word, which suprasegmental phonemes carry on over several segments - like the tones in Chinese, which carry through the entire word.
you can check out the crashcourse linguistics series for intro to linguistics! the youtube channel trevtutor is another great resource if you would like to dive a bit deeper into the linguistics rabbit hole ;)
Please never stop making videos!!!
Babe wake up new k klein video
great video! i am a chinese native speaker from Taiwan, studying how to teach chinese as second language.your video provided me a new perspective of how you learn chinese!
Pronouncing /p/, /t/, /tʃ/ and /k/ in positions where they are meant to be aspirated may sound like [b], [d], [dʒ] and [ɡ] to English speakers.
i literally just found your channel through that video and now there’s another one!! this is amazing i’m obsessed w all these videos now
Good job! Your channel has some high quality stuff. Kinda remind me of Xidnaf
[I just typed a 10-line comment about sibilants and dental-alveolar merger in Standard Chinese, and I discovered that someone has covered it in another comment way better than I could ever do.]
who the hell is xdiif
Glad to be part of the wave of new subscribers to this channel!
You could argue, somewhat in jest, that there is a minimal pair between /ŋ/ and /h/. The last name Nguyen, and a more conservative pronunciation of when. /ŋwɛn/ vs /hwɛn/.
Nguyen is not an english word though so it does not apply.
@@halohaalo2583 but /ŋwɛn/ is the English way if pronouncing it
But I'd argue (also somewhat in jest) that /hw/ is its own phoneme, since there's clearly minimal pairs between it and /h/, and between it and /w/, and posit the phoneme /ʍ/ (as some analyses do)
You are right, [i] and [ji] sound pretty much the same to me, only difference being the 2nd one is slightly louder, or [i] being the first tone, while [ji] being the 4th.
@1:33 The vertical vowel system is so true, as my professor told me before: so basically Chinese is one of the weirdest 30ish languages in the world whose "consonants (including liquids and glides) bears more weight than vowels in their syllables". So many languages that we know, like English, are "vowel dominant": meaning that their whole phoneme/phone system is evolved from vowels. Then we were used to thinking this way for the rest of the languages. However, we then observed that there are a handful of languages like Circacusan, Abaza, that have only two (or maybe only ONE @Karbadian) vowels but 134ish consonants and consonantal features. AND yes, these features are "PHONEMEIC" unlike the ones in English where they are only stylish (think about vocal fry in North American English but can actually word meaning in Circacusan).
Anyhow, weird-aXX languages like Chinese and Circacusan keep us thinking: maybe vowels are not that important in our mental grammar.
@8:09 don't worry @K Klein, Chinese is not the only one that has only 2 vowels in the nuclei: en.wikipedia.org/wiki/Northwest_Caucasian_languages
Ah, another awesome and interesting video. Great job as always! :D
Your channel is gold. Completely underrated. Great job!!!
Another great video as always, are you getting monetised now or is UA-cam still being hell?
am actually getting money now so that's good!!!
@@kklein Nice! Good to hear! :))
I just recently stumbled upon that other video and was really happy to find another place in the rabbit hole of linguistics, and now there's already a follow up, just lovely! Keep it up
7:04 You should rather merge ‘j, q, x’ with 'g, k, h' or 'z, c, s', but not the retroflex consonants.
The reason for the former one is that, 'j, q, x' are palatalised 'g, k, h', they share the same ancestry.
For the latter one, in modern standard Chinese, 'j, q, x' absorbed all the syllables formed by 'z, c, s' + 'i',
e.g. 'si' became 'xi', 'sia' became 'xia'. (N.B. real 'si' /si/ here, not the pinyin 'si')
That's why they got the space to scribe 'sr' as 'si'.
at first glance i thought the title said "the pheromones of cheese"
i thought it was gonna be how cheese can control behavior using it's smell
As a native Chinese, I actually do not consider and never considered “toneless” 4:54 to be a toneme, and that’s how we are taught growing up
I feel like when deciding how many phonemes a language has, it important to take historical linguistics into consideration as well. For example, the zero onset in Mandarin mostly came from the glottal stop initials and velar nasal initials in Middle Chinese.
In your last video, I noticed you didn’t count erhua syllables such as “ber” (贝儿). Any reasons for that? Also how are the unusual syllables used in interjections counted? For example, n (嗯), m (呣), hm (噷), and 哼 (hng).
As for interjections, I think they're considered paralinguistic, not linguistic. And as for erhua syllables, I don't think they're considered a feature of proper Modern Standard Chinese, they're just a feature of northern Mandarin dialects (which are the dialects that most resemble MSC, because it's based off of the Beijing dialect)
How the fuck did you type m and hm?? 😳
erhua seems northern-centric to southerners like me, it's rather dialectal
@@archkde students of Chinese do learn Chinese with at least some erhua. Usually there are alternative non-erhua versions like 这里 'here' instead of 这儿 'here' and adding -儿 'diminutive suffix' to the end of words is generally a very Northern Chinese thing to do (why say 小孩儿 'child' when 小孩/小孩子 does the trick?) but a handful of words have no non-erhua alternative like 二 'two', 儿子 'son', 耳朵 'ear' etc.
@@niku.. Correct me if i’m wrong, but afaik, erhua exclusively refers to 儿 being used as a suffix in a non-syllabic fashion, so words like 儿子 and other characters that have the pronunciation of “er” like 耳朵 don’t count.
Great video. My only real nit pick is with the section where you discussed how we know how some phonemes are different, like /h/ and /ŋ/. You mentioned that the two ways to do this is by looking at minimal pairs and by looking at how different the sounds are phonetically. But that latter method is actually very rarely used, if ever. Instead, most linguists when building phonology's and deciding phonemes have a few different approaches, for finding what phonemes a language has, and what allophones said phonemes themselves possess.
The first method is obviously minimal pairs, which work great for many simple morphemes, like pin/bin, tip/dip, etc... But as you mentioned before, sometimes, some phonemes don't have minimal pairs to work with. In English, this would be /h/ and /ŋ/. However, we do know that they are different phonemes, not because they are phonetically different, but because they can exist in identical environments. /h/ can only exist in the onset, and /ŋ/ in the coda, yes, but they can also exist intervocalically. reheat /rihit/, dehiscent /dıhısǝnt/, singing /sıŋıŋ/, hangout /hæŋaʊt/, etc... Looking at words that place the phones in near identical or identical environments, we can then determine if said phones are independent phonemes, or allophones of one phoneme.
You had me at the difference between the scold and cold
I'm sorry the editing that works to highlight the different noises is incredibly choppy and distracts a lot.
Minimal pairs follow-up video idea: /ʃ/ and /ʒ/ in English. The only sort-of minimal pairs I can think of for the two sounds are fission/fishin' (dialectal), shush/zhuzh (very dialectal), confusion/Confucian (which I would call valid, but some pedants might point out violates Scrabble's rule against proper nouns being words I guess), and leash/liege (also valid IMO but when's the last time someone used "liege" beyond a historical context?). Even if, sure, there do seemingly exist some tentative examples for ʃ/ʒ, it raises an interesting question: what if there weren't any? I'm sure there's some contextual rule about phonemes occupying the same "phonemic space" or whatever, but it seems like it could get confusing, controversial, and tantalizingly elusive to define.
EDIT: as far I've been able to tell since, yes, linguists are fine with distinguishing phonemes based on being perceived differently when occurring in similar context, if not the exact same. Many English speakers would consider it a mistake to pronounce rouge "roosh" or potion "pozhun," so we could determine that they're perceived as different sounds.
More examples of "questionable" pairs include: Aleutian/allusion(/illusion); measure/mesher; cash(/cache)/cazh (a clipping of "casual"); Haitian/Asian; hash/hajj
Wikipedia uses near minimal pairs pleasure/pressure
So minimal pairs are a sign of phonemic contrast, but they're not actually necessary for two sounds to be analyzed as phonemes. A language might coincidentally not even have such a pair of similar-sounding words. What's really important is that their distribution is unpredictable by rule - that is, you have no complimentary distribution. That's why the pleasure / pressure example is great even though the words aren't true minimal pairs. The sounds still exist in similar enough environments to suggest that environment is not a determining factor of which sound you get.
@@cubing7276 treasure/tressure is a true minimal pair.
delusion, dilution
"rage"/"rash" is a minimal pair (fricative), as well as the "age"/"hache" pair (affricated).
Good use of humor throughout your video. Very professionally done.
I'm officially subscribed to your channel.i believe it was a glorious choice .
Your videos are really intriguing for me as a native Mandarin speaker! And no, I have no idea why the language is called Mandarin in English. Maybe the reason is lost among translations between European travelers centuries back. In primary school, I was taught that Chinese has six vowels (a o e i u yu) but your explanations make so much more sense in linguistic studies. Language barriers have created and will continue to create distrust and distorted views among people. I am learning my third language now, and looking forward to more quality videos from you!
A "mandarin" is a bureaucrat in the Imperial Chinese government (according to wikipedia, the term is probably a corruption of some portuguese word), so "mandarin chinese" was the variety spoken in the court. Britain mostly did business in Guangdong, so the variety they were most familiar with was cantonese and they just knew mandarin chinese as the language that the bureaucrats in Beijing spoke in.
Teachers of English to foreign students always seem to go overboard on the aspirated unvoiced plosives thing. The only people who I've noticed aspirate cluster-final plosives are non-native speakers, yanks who say fiftʰeen and sixtʰeen, and RealLifeLore all the bloody time.
8:50. Don't worry K Klein, I got you on why Standard Chinese is called Mandarin.
Mandarin, as we know today was called 官話 or the 'language of the officials' when it was used as the adminstrative language of the Ming and Qing dyansties.
In the 16th century, the Portugese, having established logistical colonies in India (Goa) and the Malay Peninsula (Melaka) were among the first Europeans to establish proper contact with the then, Ming dynasty China. Thus, they had the honour of translating and transcribing many Chinese words into the Latin script.
The Portugese translated 官話 into 'mandarim', which was derived from the Malay word 'menteri' meaning minister. The Malay word itself has its origins in the Sanskrit word 'mantri', which again has the same meaning of minister.
Both the Merriam-Webster and Oxford Dictionary’s acknowledge the origins of the word 'mandarin' in the Portugese 'mandarim' and was incorporated into the English language in the 16th century. So, it seems to check out.
According to an article I read: Mandarin is what linguists call an exonym, a name given to a place, people or language externally from the place of origin.'
This article talks about it in more detail: www.theatlantic.com/science/archive/2019/01/where-does-mandarin-come/579073/
As for why the fruit is also called mandarin, I don't have a credible source for that; but I have heard that since the fruit originated in China and shares the same colours of the yellow robes worn by Chinese officials and emperors, Europeans named the fruit as such to highlight the colour similarities.
In Korean language, P,T,K sounds are distinguished by three differences in aspiration.
For example, /tal/(moon), /t͈al/(daughter), /tʰal/(mask) and /ki/(energy),/k͈i/(talent),/kʰi/(height).
/ŋ/ vs /h/ does matter when it comes to morphological alternation where /h/ may be yeetted, /ŋ/ goes to /ŋg/ (with a few exceptions) as such /ŋ/ is argued by some to be an allophone of /ng/ with the stop being unpronounced word finally (an other special word-morpheme)
What about words like singer? With an ŋ not followed by a word-final g?
@@LunizIsGlacey that why it's debated
Chinese has a set of consonants and a set of the 5 vowels you can use, but there's a whole list of the possible final segment of a word in pinyin that we learn and memorize in elementary school. I can't list all of them off the top of my head right now, but it starts like this: a, o, e, i, u, v(for notation reasons I'm using v, since chinese pinyin does not use v), ai, ei, ui, ao ou, iu etc. These can contain non vowel letters, for example ang, eng, ing, ong are all in this list.
There's also another form of chinese word, which in pinyin form would be consonant - connecting vowel - vowel. Usually the connecting vowel is "i", but this list is pretty limited, so it only works with a small combination of connecting vowels and the list of final segments. One example would be i-ao, and a word with that could be xiao, or 小.
There's a few final segments that can be used without a consonant, namely a, ai and er. I'm pretty sure these are the only ones.
There's also a list of words that sound the same as either the consonant or the vowel. For example, zhi sounds like the consonant zh, and yi sounds like the vowel i.
earned a subscriber, hope you pump out more vids!
I'm amazed by you. Found you on the start page and caught my attention immediately. I was looking for youtube channels about linguistics because either the few left upload videos very irregular or don't at all. I'm glad that I found you. Your content is superb.
But one question: Why is your overlay language in word and youtube German but your location England, Uk? Are you German living in the UK? But if so why don't you sound German at all? How is this possible?
Keep up with that content. I see that your subs are rising extremely fast :)
6:39 The z and the inverted r are not pronounced as z and r, both are actually vowels. In fact z and r are used to represent the vowels because other letters we usually use to represent vowels (a, e, i, o, u) have been used up.
How z and the inverted r (which you use r to represent) sounds like is somewhere in between i and u.
I didn't understand most of it but i Guess i could start with learning about phoneme.
Anyhow, I like your manner of speech and how you explain things. Keep it up.
Interesting analysis video! Thanks for uploading!
Wow you gained a few thousand subscribers in just a few days!!! Good job
Damn this is well explained. Learned something new about a language I have been using for 18 years.
3:16 omg, I really want someone to make a terrible spelling reform based on the idea that ŋ and h are alophones now!
One thing to note: French does not have any stress at all. It makes use of intonation patterns which lay over entire phrases and clauses, meaning that individual words may alter in their accentuation depending on the surrounding sentence and the pragmatic context around it
But the stress is always on the last syllable of any such segment. Which means that when french speakers try learning other languages, they'll first learn individual words and will put the stress on the last syllable of those individual words
@@bacicinvatteneaca I don't think you're right. All syllables are (supposed) to be given equal weight. This is now seen as a very, very "formal" or "old-school" way to speak, associated with, say, journalists or old teachers, but it is still the standard. Or at least it is in my dialect.
However, in a question (for example) the intonation will be rising all throughout the sentence. I think Parisians also have a rising tone, but I am not sure.
A good word for it would be in my opinion "civilisation". You're expected to say "la cI-vI-lI-sA-tION".
@@bacicinvatteneaca No. There may be a phrase final high intonation tendency but it’s certainly not a rule. Personally when it comes to isolated words with no conversational context i tend to keep the intonation flat
i'd say at least in southern france, we tend to put the stress on the last syllable, or second-to-last when we pronounce the -e (personally, i do about 50% of the time inside sentences, but some people do it much more often), and secondary stress before any word-internal e (as in développement, 2 or three stressed syllables)
I recently found your channel and have seen all of your videos dozens of times already and I hope you keep making your amazing content :)
Omg that screenshot of my comment was UNEXPECTED IM FAMOUS
“Disadvantages: Stupid and Dumb”
Haha very scientific
Im not a native english speaker and thank you for providing me with the ability to just check how perfect my english accent is.
I do indeed say the k of kold and skold very differently when checked by air ammount as the first has much the other has none. And I dont even know how to asperate something, I so struggle with this trying to say latin or PIE or ... words.
Knowing that everything is unasperated in my language I also checked if I speak english with a latvian accent what then happens and yes neither skold nor kold have any air reaching my hand.
It's cool to see another way minimal pairs are used. As an English teacher, I use minimal pairs in games to help force my students to hear English phonemes (say, by having students try to guess the differences between baa and bar so they can hear the r final better, or having them find either fan and van in a scavenger hunt). I've always been kind of sad that Chinese phonotactics is too restrictive to allow people to even make up plausibly fake minimal pairs like xao/shao (as x + ao must have a glide, xiao). It would certainly help non-native speakers develop phonemic awareness of the sh/zh/ch xi/ji/qi distinction.
As someone currently in the process of learning chinese this is hitting pretty close to home. ji and qi sound very similar to me, to the point where I think I can hear the difference between 人 and 认 better than 俏 and 叫 despite my native language not being tonal. What really fascinates me is that there are words which seem to have a lot of variance in pronunciation like 一 or 他 / 她. Also if I'm hearing it correctly and 他 and 她 are really homophonous, that's crazy. Not having to specify the gender of someone when referring to them in 3rd person while speaking, but kinda having to do so while writing is pretty funny. (better than my native tongue, there we always have to specify -.-)
@@chalkchalkson5639 this is because 她 was a recent invention originally used to translate "she" in English textbooks- originally 他 was all-inclusive and had the same meaning as 她/它/牠
The "scold vs cold" thing really broke my brain
Is this part of the test? Do we get a printout?
Some have suggested that /ɥ/ is actually an allophone of /jw/ or /wj/, so you could lower the count by one again
oh, that's very interesting 🤨
Isn't that how Pinyin does it?
Huh, I never noticed that insight and incite have the same pronunciation but different stresses.
6:42 it’s funny because it makes sense. In Miyako (a Japonic language) i and u when between consonants (and other places in some dialects) become syllabic s and f like pstu “person” (cognate to Japanese hito) or tabz “journey” (cognate to Japanese tabi) so it makes complete sense. The r syllabic consonant probably happened as z merged in place to the retroflex consonants it was after. The theory from Miyako is that it became ɨ before breaking out of the vowel space to z (then in the case of Miyako dialects, assimilating in voice)
The analysis with three glides and two vowels is based on historical Chinese phonology. There's an entire system established though the likes of 切韻 (600 CE), 韻鏡 (c. 10th Century), and 廣韻 (c. 1008). The system aims to help poets find words that rhyme in Middle Chinese, not to minimize the number of phonemes in Modern Standard Chinese. The linguists who made bopomofo were trained in Chinese Classics so they had an affinity for the system. Furthermore, the linguists were envisioning a system that can transliterate all Chinese languages, and some of those languages have inherited these phonotactics from Middle Chinese.
Actually the vowels in 安an and 昂ang are different but written as if they were the same. As a native speaker, I can say most of Chinese people don't really get the difference between [ŋ] and [n], we use the vowels to differentiate them.
Vowels in Chinese may be a lot more than you expected.
Chinese Zero to Hero just made a video about this. Actually, it makes sense. N is articulated more in the front of the mouth than NG, so naturally the vowel preceding it would be more in the front, as well. It makes it easier to pronounce.
I absolutely love your videos
A question that troubles me from time to time: How are vowels different from (certain) syllabic consonants? How many vowels are there that cannot be regarded as syllabic consonants?
BTW: According to my experience, the zero-onset is often realized as either a glottal stop, or as ng before /a/ and schwa, but not when a glide is present, so I would consider it as an initial in its own right, or a glide equivalent at the very least.
(The ng variant is not relevant to the 疑母 initial ng from Middle Chinese, as it is also witnessed in e.g. 愛., which does not have a ng initial in MC.)
Vowels are vocalic, so no air is being obstructed when you say the phoneme. Syllabic consonants may be sonorant, meaning that air is not entirely stopped at any point, but the tongue and/or lips are used to obstuct sounds so they are not considered vocalic.
Just discovered your channel. I love it. More linguistic channels!
You're suddenly getting popular very quickly
2:41 There are also no minimal pairs between unaspirated k and g. Why can't they be the same phoneme? Sgold and scold are both understand as scold.
i think this is a better analysis than what i put forward in the video, i have changed my mind to support this
PLEASE DONT STOP MAKING VIDEOS
At 5:40 I don't really understand why it's said that "o" and "e" both represent the vowel ə, which is not really accurate. They both represent different sounds: o is /o/ while e is normally /ɤ/ but can also be /ɛ/ if preceeded by /j/ or /ɥ/...
So, pinyin uses 6 symbols to represent 7 vowel sounds, note that zhuyin does in fact use 7 symbols.
On the other hand, some linguists consider the syllables "zi, ci, si, zhi, chi, shi, ri" to have a /ɨ/ vowel, while it can also be considered as being a syllabic consonants as explained in this video.
I would be interested to see historical linguistics used more in these phonemic analyses, since complementary distribution often comes from sound changes to a single phoneme- we can see for example that scold and cold's /k/ also come from the same historical origin, or that the chinese palato-alveolar series [tɕ], [tɕʰ], [ɕ] stems from the velar series /k/, /kʰ/, /x/ and thus makes sense to analyse as allophones of the same velar series. Calling them allophones of the retroflex series seems a little weird- its functional, but ahistorical, and both moves reduce the phonemic inventory the same amount. This way analyses have more historical grounding and give some insight into dialectal differences.
For example, this makes Cantonese, which retains the velar series where Mandarin has the palato-alveolar, make more sense. We can simply say they are the same phoneme, but Mandarin has a different allophonic pronunciation.
This also adds some more context to why certain phonemic pairings exist. If we include Japanese [h] and [f] under the same phoneme /h/ since they exist in perfect complementary distribution, it makes more sense if we also consider the historical sound change that turned [f] to [h] except before /u/. They exist in complementary distribution because they come from the same phoneme that simply gets pronounced differently in a certain context. [h] -> [f] is a rather far-fetched sound change, and so knowing that [f] -> [h] happened makes grouping them under a single /f/ or /h/ more reasonable.
i would categorize syllabic consonant as 0 final imo
like in bopomofo there just isn't any
and i would merge the alveolar-palatal series as either velar of alveolar bc of historical sound changes
I like that - a completely valid analysis.
I don't really like using historical sound change as the sole reason for classification. It's valid of course but not necessary and we are trying to analyze the modern language and how speakers percieve things right now. The alveolo-palatals are indeed in complementary distribution with the velars, the alveolar affricates and the retroflex consonants so it's difficult to assess where to put them.
@@niku.. lol you could also categorize as 0 coda too
The advantage of saying the alveolar-palatals are allophones of the velars specifically is that you don't need to include the syllabic consonant as a separate phoneme.
@@antoniozavaldski yeah i was thinking about that too lol
I'd heard the idea floated that Bopomofo would've likely been more accepted over Pinyin had the "Initial + Final + Tone" symbols been arranged into syllabic blocks (like what Hanguel does). What are your thoughts?
That would definitely fit closer with the general shape of the Hanzi. However, Zhuyin (or pinyin) isn't used in the language except for marking the sound, unlike how Hangul is used for Korean or Kana for Japanese. So it shouldn't matter that much.
5:42 pinyin and phonemes 1
7:01 simplification for complementary distributed allophones
In the word "français" the stress is actually in the last word. Cool vid btw ;)
So" It's "i" before "e", expect after "c", unless sounded as "a" as in neighbor and weigh? Your explanation is wonderful. (not my crude joke, above) Do you know of Simon Roper on UA-cam? You two should do some videos together. I'm excited for it.
To be more accurate the vowel in "keng" is more like /ɤ/, you can hear that when an average Chinese speaks English: sometimes they pronounce the schwa vowel like /ɤ/.
Thank you for this video. I always intuitively has this hunch about the Chinese language but did not no how to approach it. Thank you.
2:59 Fun fact, the Korean alphabet DOES consider them as allophones of the same phoneme, as Hangul uses one letter to represent BOTH the glottal stop initial consonant AND the /ng/ final consonant...
I understand almost nothing, but i was very amused. Love it dud