 Okay. Okay. So we're going to talk about the initials today. So, you know, broadly speaking, the methodology is the same as we did with the vowels. So first internally reconstruct middle Chinese, and then look at Shesong evidence. And we can't look at rhyme evidence in this case, right? Because the initials are not a, you know, are not relevant for the study. Or yeah, when you rhyme things, you're using the vowel and the coda of a syllable, you're not using the initial. So the poetic device that would be useful for studying initials would be alliteration. But as far as I know, Chinese hasn't used alliteration in a systematic way like Old English did. So it's not, you know, so, so it's only really internal reconstruction and Shesong evidence in the first instance that helps us reconstruct the initials. Okay. So first then the internal reconstruction. And we will remove some things. So we'll remove H and the palatal advocates, and the retroflex stops. And we will add labial velars. So here goes. I think I mentioned this in passing, just in terms of, you know, to exemplify how we use internal reconstruction to go from middle Chinese to old Chinese. But we notice that voiceless velar occurs in both type A and type B syllables, that voiceless aspirate velar occurs in both type A and type B syllables. But the voiced velar only occurs in type B. And the H, which is probably a gamma, I mean, this is one of the confusing things about Baxter's transcription, but the H means something like a voiced velar frigate, that occurs only in type A syllables. So we see a complementary distribution. And then we think, okay, we can do internal reconstruction and propose that G changes to gamma in type A syllables. So now we've gotten rid of gamma for old Chinese. Now we look at the palatal advocates. And we notice that dental stops initials only occur in type A rhymes. And the palatal advocates only occur in type B rhymes. So then we can suggest that the, you know, that type B syllables are a conditioning environment for this palatalization. And we don't have to reconstruct palatal initials to old Chinese. And now just a notational thing. So you're aware, Division III syllables in Baxter's system have some kind of Y or J in them. So you can always notice Division III because of that. And then, and Division I and IV are unmarked in his system. Whereas in old Chinese, Baxter and Sagar write type A with this little pharyngealized, you know, this little raised, you know, you see it in the T, right? And then type B they don't mark. So I think for our purposes now, we don't need to feel commitment to this phonological interpretation of the type A, type B distinction, but you can just treat it as a handy way of indexing the conditioning environment. So you see that, you know, dentals come from type A dentals and palatal advocates come from type B dentals. Okay, now a look at the origin of L and the retroflex consonants. And we hit on a little bit of this before, but there are two vowels, the A vowel and the E vowel, which both occur in Division II, that do not co-occur with, no, I'll just read that. Yeah, so dental initials don't occur with these vowels and only very rarely with L initial. And these vowels, which are Division I and IV, don't occur with retroflex stops. So Yachondorf proposed that the dental and retroflex initial classes were originally the same, because this is again a sort of complementary distribution. So we can write it like this, that L comes from R and that T-R, well, comes from T-R, but this is a place where Baxter's transcription is a little unhelpful, because the T-R in middle Chinese means a retroflex stop, whereas the T-R in old Chinese means a T followed by an R. So that's why it puts the retroflex stop in brackets. So, and then also I'll just say that like as an aside, so Yachondorf actually, I think he originally looked at this and I'm actually, I don't want to say that, it might have been Puliglik, because Puliglik and Yachondorf kind of proposed this around the same time. That what I'm trying to get at is why do we reconstruct the L as an R? Why not just leave it in L? Well, at this point, we can say, because it seems sort of more normal to take retroflex initials back to a cluster with an R, then to a cluster with an L. But we will also want to use L in old Chinese for other things. So, and I mean, there's also some direct evidence, but yeah, just to say that, you know, I don't think that this change of R to L is a big deal. Yeah. And it doesn't need to bother you. Okay. So, and then similarly, the retroflex africates originate from africates followed by a medial R. So, that was, that's kind of the R hypothesis, right? The R, the medial R explains both division two rhymes and retroflex initials. Now, on to labial velars. If we look at the distribution of medial W in middle Chinese, it falls neatly into two categories. The first is checked rhymes that occur only after velars or glottals. And these are the rhymes in question. And I'm giving that the name of the rhyming pinion followed by the characters, whoops, followed by what the, how the rhyme is represented in Baxter's transcription. Okay. And there are no middle Chinese syllables, you know, of the types you see on the screen, which is, which is to make the same point, basically, that acute initials don't occur with these rhymes. Now, the second class is all other rhymes, which appear freely with all initials. So, these two distributional classes imply two different origins for W. So, I will just call them kind of mechanically W one, which only occurs after velars and glottals, but with any rhyme. And W two, which only occurs in certain rhymes, but with any initial. So, and, and, and you might notice that these, these statements of W one, W two are a little bit different than the, they're the mirror image of the distributions I was talking about between initials and, and rhymes. And that's because I'm presuming that in some contexts, we can't distinguish the two, right. So, after velars and glottals, W one and W two merged. So, Audrey core was the first person who then said, well, you know, W one seems to have this special relationship with, with velar initials. So, maybe it's from labial velar consonants. So, just, just to give you some examples, yeah, so we, you know, I mean, you can almost see this as a kind of rewriting rule mechanically that going from middle Chinese to all Chinese, we, we take the W and we put it upstairs as a superscript. Yeah. And this also shows you, I mean, it is a nice, I think the case just in terms of simple examples. So, how, how we sort of notate middle Chinese and all Chinese differently, where the, the type A syllables, we add the pharyngealization and then prefer using IPA characters to, to sort of standard Roman characters. Okay. And then for the moment, we will presume that there were also labial glottals because you, you know, if you, if you, if you just follow the way I've been doing things, then you have to, right? So, the same, these special rhymes that have medial W and only occur with velars and glottals. Well, if we're going to reconstruct it as labial velars, then we also need to reconstruct labial glottals. And then some of you will maybe feel uncomfortable with that. I do at least from phonetic reasons, but we'll, we'll come back to that. So, so don't let it bother you. Okay. So, if we just take the initials of middle Chinese and we do these moves that, that I've already explained, which is we, we get rid of the palatal afrogates and we get rid of the retroflex consonants and we add in labial velars and labial glottals. This is what you get as an inventory. So, this is a kind of, yeah, a good place to sort of pause and take stock of our progress, which is to say, this is what you get from internal reconstruction before looking at Shesham series. And the, I mean, these are the type B and these are the type A, they're the same, right? It's just that the type A we write with the little pharyngealization marker. Okay. I also forgot to mention that we, we get rid of the, the, the voice feel of fricative too, right? So, you see, it's, that's what we have. Okay. So, I mean, I feel like maybe it's a little early and there won't be any questions, but might as well give a try because it's a kind of logical break in the presentation. So, any questions? No questions. I guess I will go on then. So, now we're going to, you know, look at Shesham series kind of with this progress we've made in mind and see what happens and come up with some other proposals. So, according to the, the Shesham hypothesis, and this is not in, well, yeah, I should say last time I talked about the Shesham hypothesis as articulated by Duan Yuzai, which is that if two things are in the same Shesham series, they will, they would have rhymed in all Chinese poetry. Well, there's a sort of corollary to that, which you will have already sort of seen evidence for and, and we'll continue to see evidence for in a minute, which is that Shesham series tend to have a whole more organic initials. So, you know, initials from the same place of articulation. And then that observation, that tendency has been sort of, you know, you know, paraphrased back as a methodological principle where, where we say that a Shesham series should have initials only from the same point of articulation. So, Shesham series that mix middle Chinese pronunciations with non-homer organic initials provide an opportunity to explain the divergent middle Chinese pronunciations as phonetically conditioned developments of all Chinese readings with homo-organic initials. So, it's a kind of internal reconstruction, again, if you like, but now we're not going from the distributional patterns of middle Chinese, you know, itself, but we're working from trying to fix those places where the work we've done so far does not yet bring the initials into conformity with the Shesham hypothesis. Okay, so some examples. Here we have a Shesham series where you have a yuh initial and a druh initial and a duh initial. Okay, so we're looking for something that can explain this. And this is what, you know, Backstrom-Sugarre proposed building on an idea of pulley blanks, which is that initial L develops into yuh in type B syllables and develops into D in type A syllables. And this kind of funky cluster, LRA, develops into the retroflex, you know, voice retroflexed initial stop in type, both type A and type B, which is why the little pharyngealization thing is in brackets. So that's their hypothesis. And look, it cleans up this Shesham series nicely, yeah. So now the initials are all, you know, home organic. Okay, we can also propose a voiceless residence because there are some Shesham series with predominantly voiced resident initial readings that have the occasional character with an obstruent, a voiceless obstrant. So let's look at some examples. Yeah. Just to tell you a little bit about my conventions, no one has been bothered by it, but the number after a character, so here like 2415a refers to its number in Shusler's 2009 book, which lists all the Shesham series. So it's a kind of a, you know, unique identifier. And it, and the first number tells you, actually, it's sort of rhyme class in his analysis. So it's in sort of rhyme class 24 in his analysis. And then the 15 tells you the specific Shesham series under that class, and then the A is which character in that series. Yeah, okay. So here we have, you know, contact between the voiceless velar fricative X and the voice velar stop. So the proposal is that we had, you know, a voiceless velar resident that in type A syllables changed to a voiceless velar fricative. And now, again, we have managed to make the initials in this Shesham series home organic. And then I just give one example from type A and one example from type B. And I'm trying to be sort of systematic in that way. In some cases, you can get both type A and type B examples within one Shesham series. But in other cases, you need to look at different Shesham series to get both types, which is what I'm doing here. So there we have our voiceless velar resonant. Yeah. And here's an example for M. So voiceless M in both type A and type B syllables, again, becomes a voiceless velar fricative. And you see this, you know, this evidence of it from the Shesham series. And I hope this works. But when I'm talking about Shesham series, I use a bigger font for the Chinese characters, hopefully in order to draw your eye into the actual structure of the character so that you can see like, Oh, yeah, I have the second character wholly contains the first character. And they have kind of similar pronunciation. So probably the first character is the phonetic determiner of the second character. And then we try to make them fit even better with this hypothesized sound change. Okay. And now voiceless N. So here we get some more complication. So it changes to a th in type A syllables and to shuh in type B syllables and before an R to this aspirated retroflex voiceless stop. And I don't know, I feel torn between sort of talking you through each one and it being boring or just saying, Well, you can see it immediately. And it being maybe too fast. So let's just look at the second group of three. So we have here the word for woman as the as the, you know, origin point of the Shesham series. And it has an N R. And then the this, the second character has N, both of those in type B. And then the third character starts with shuh, right? So how does it, how is it possible that why would anyone link you use an N sound as the phonetic determiner in a character that's read with shuh? Well, maybe it's because it starts from, you know, in old Chinese of voiceless nasal. Okay, so then we can also have our voiceless lateral here. And in type A syllables, it turns into ta. And in type B syllables, it turns into shuh. And here you see I'm relying on the earlier discussion of laterals. So we have, I mean, and, and this is a nice Shesham series in, in terms of how messy things can look from the perspective of middle Chinese, because you have in one phonetic series, you have ya, you have da, you have shuh, you have ta. It's quite a mess. It's quite far from the Shesham hypothesis. But we can say, okay, we have this theory of sound change where they all were initially laterals, the voiceless laterals in type B and type A, and then the voiceless laterals in type B and type A. Okay, so that's it for the, for the laterals. And here we go with the rotix. So type A voiceless r, changes to th, and type B voiceless r to trh. Yeah. And you see the, the evidence here. And, and this is where, you know, so jumping back a little bit, you know, you see when we initially proposed reconstructing a medial r to explain the division two vowels and the retroflex initials, it seems a little bit kind of arbitrary, like, well, you know, why r, why not, why not l? Since the argument is based in part on the fact that, you know, that l initial doesn't occur with the division two, for example. Well, now you see how I think it becomes plausible as part of an overall system, which is we have work we want the r's to do and we have work we want the l's to do. And that it all gets a little bit simpler, if you assume that, so we started with this, with voiceless r's and voiceless l's, had all these sound changes that I'm describing, and then kind of late in the process, those remaining initial r's just changed to l. Yeah, okay. So now just to kind of give you a bird's-eye view of the voiceless resonance, no, first actually, I'm going to discuss the intellectual history. So, Don Ponghe first proposed this voiceless resonance idea only for the labial nasal, so only for m, yeah. And then Polly Blank extended it to velar nasal, to labial velars, to dental nasals, and so on. And then Baxter added the voiceless rotic, okay. So, and then, you know, Baxter and Sagar keep all those ideas. So here's the bird's-eye view, which hopefully makes the proposals seem a little less ad hoc, right? So, yeah, I mean, let's even try and paraphrase what's going on here. So in type A syllables, the grave initials, and these terms grave and acute, they come from Roman Jakobsen and are based on the observation that in many patterns in language, labials and velars have something to do with each other, and then kind of stuff involving the roof of the mouth has to do with each other. So we can broadly speaking divide constants into the graves, which are the ones at the front or the back of the mouth, or the acutes, which are the ones kind of in the middle of the mouth. This terminology seems to be quite out of fashion now in kind of phonological circles, but we use it in all Chinese. Of course, you know, Roman Jakobsen didn't have Chinese in mind at all, but Baxter uses it a lot in his 1992 book, and I do think it's quite helpful. So here, you know, we can use those terms to see why the divisions are, where they are in terms of outcomes, right? So the voiceless resonance in type A that are grave turn into voiceless veal-affricative, and those that are acute turn into dental, voiceless dental aspirin stop. And then in type B, it's quite similar. It's the same for the graves, and then in the acutes, we get this further split with voices N and L going to shah, and voices all are going to trah. Okay, so that's overview of the development of the voiceless resonance, and now we move on to uvulars. Maybe before this, I'll give you just a tiny sense of how controversial different bits of this are. So the uvulars, which we're going to get to just now are quite controversial. So Baxter and Cigar are kind of a little bit isolated for believing in uvulars, whereas the voiceless resonance are much more widespread. They're in, for instance, Li Feng Gui's system, but still, I don't think they're in Wang Li's system, so there are big parts of China where they won't be believed. But the takeaway message, voiceless resonance, not particularly controversial, but uvulars are more controversial. So let's look at the uvulars. We noticed that there are shesheng series that mix initial velars and glottal stops or initial ya. So the point is that k, g, r and ch are all in the same place of articulation. So they're so they're homo-organic, so they're in keeping with the shesheng hypothesis, but the glottal stop and the initial ya are in different places of articulation. So they violate the shesheng hypothesis. So we want to propose something to get rid of them. Okay, so Pan Wuyun is the first person who said, well, why don't we have uvulars to get rid of these? And he had a particular formulation, which I won't go into. Also, Baxter and Sagar had a formulation kind of different than his that was their own in around 2012. They wrote an article about it, but then they changed their mind and I'm only going to present it as their new version, yes, which is to say, even if you believe that using uvulars is a good way of solving this problem in the shesheng hypothesis, there's clearly more than one way to do it. But we're only going to look at the way they do it in 2014. But first, some positive evidence that these uvulars are a good idea. And this comes from work of Sasha Vovan's, Vovin, yeah. So there's a, these are both sort of titles of foreigners in Dahon, yeah. So we have this Huyu, and there's it like, so this is, this is Vovan's argument. Great ruler is, I don't, I don't feel terribly comfortable pronouncing uvulars, but it's Gah, yeah, in Proto-Yanisayan. And it, you know, if you write those characters, if you reconstruct those characters as uvulars, oh, look, it matches uvulars in Proto-Yanisayan. And similarly, there is a title, you know, Zhenyu or something, which would reconstruct to Dar-Kwa, and he compares it to old Turkic Tar-Khan, yeah. So they're not, I mean, they're not perfect comparisons, like where's the n in the second one. But I think they do help build a case for plausibility of uvulars in, you know, Chinese. Okay. So here are the proposals. A, I don't know what to say, a Q, a voiceless, unaspirated uvular stop becomes a glottal stop in Middle Chinese, whereas a voiceless, aspirated uvular stop becomes a voiceless velar fricative. And that's, so that's, yeah, those are in some ways the more simple case. And then the voiced uvular stop, which is this capital G, becomes the voiced velar fricative, which we confusingly write as an H. And then here is some evidence, yeah. So if you just look at this series from Middle Chinese, it's a bit of a mess. Yeah, you have a K initial, you have a glottal initial, and then you have these these two velar fricative initials. And if you reconstruct it using the uvular hypothesis, well, it almost works. Because you'll say, well, Nathan, what about the velar? You know, you, okay, fine, we reconstruct the glottal as a uvular, and we reconstruct the velar fricatives as uvulars. But then what about this velar? Well, we'll get to that. But it involves this theory of pre initials, which is not today's topic. But but basically, to make a long story short, if there was a pre initial, you and you had a sort of vowel uvular vowel, they think that that merged with velars. So and then the pre initial was lost. So if you didn't follow that, that's fine. But let's say there we will see that there is machinery for changing the velars in uvular series into back into uvular. Okay, so now here's, yeah, just the point that the the capital the small capital G changes to a voiceless velar fricative in type A syllables, but in type B syllables, it changes into yeah. So this this is part of the, if you like, the overall pattern of palatalization that we see in type B syllables. So now, um, yeah, the question is, did old Chinese have an initial yeah. And a backstern cigar would say no. So so far, we have seen two origins for yeah. And then I'm just going to mention briefly a third that involves pre initials. So yeah, can go back to a lateral where in shesheng series where it has context with dentals, right? So if you have a yeah, and a duh in a shesheng series, then you reconstruct them as laterals. If it has connections with velar fricatives or with velar stops, then we reconstruct it as a voiced uvular stop. And then there's also this option of are preceded by a pre initial turning into yeah. And actually, that's kind of the same as the L because what they actually think happened is the C dot R changed to L and then the L changed to yeah, right in type B syllables. So they have these three origins for yeah. And then they say, well, look, you know, if if I have three ways explaining yeah, in middle Chinese, I don't need to it's Occam's razor would say, you know, decide among these three based on the shesheng connections. And we don't need to have a yeah in old Chinese. But Axel Schussler and Guillaume Jacques disagree. And they point to this comparative evidence in particular. So the old Chinese word for young, sorry, the old Chinese word for sheep, which is young, would reconstruct in Baxter and scar system to having a voiced velar, sorry, voiced uvular stop initial. But if we look at other sign of Tibetan languages, the word for sheep seems to have a yeah in it. Yeah. So we have this young car, which would be like white sheep in Tibetan and this job of form where the Joe part descends regularly from young. So yeah. And then similarly, there's a comparison is made to this word for itch, although we have to admit that the Ryan correspondence is not particularly convincing for itch, because Tibetan and just have something like yeah. But all Chinese has young. In any case, looking at this evidence, Axel Schussler and Guillaume Jacques think, no, don't don't just blindly reconstruct all of your yas to something other than yeah, you know, in some cases, there's reason to think that all Chinese itself had a yeah. And now I think Schussler would say all Chinese had no uvulars. You should reconstruct all of these cases as yeah. Whereas, I think what Guillaume Jacques would say is, yeah, he's perfectly happy with the three origins of the yas that Baxter and cigar propose, but he wants a fourth origin of yas, which is just yeah. Yeah. Okay. Now, we also have evidence for a labial uvular. When you see a series like this, yeah, I mean, it's what to say. It's the kind of series where we reconstruct uvulars because you have a glottal initial, and it has contacts with velar initials, right? But it's hookah, so which is to say there are there are medial W's. And it's one of those hookah rhymes that we would normally reconstruct labial velars for. But we're going to reconstruct these as uvulars, not as velars. So we reconstruct them as labial uvulars, labial uvulars, yeah. And yeah, and if you look at these examples, they just happen to be, I mean, at least as far as we can tell, this is the claim made by Baxter and cigar, and I've checked it to some extent, they just happen to all be before front vowels. And then that means that they can still use the reconstructed labial uvular initial for something else, which they do, and they say, okay, it becomes this, I don't even know how to talk about this initial in Middle Chinese, but it's also like a voiced velar fricative, but in type B or division three syllables. There's a kind of a phonemic issue in Middle Chinese, whether this is seen as a separate initial or not, but we don't need to worry about that. And Baxter and cigar and basically mechanically project this particular Middle Chinese initial back to a voiced labial uvular stop. And that, yes, let's say, I mean, I don't know if it's a positive reason for doing so, but is convenient in terms of this initial in Middle Chinese basically exclusively occurs in type B syllables with medial w. And you see that that's a natural, if it began life as what happens to a labial uvular in type B syllables, then that makes sense why it would only occur in those environments. Okay, and then just some evidence. We have this series where you see there is this initial we're talking about, the HJ initial, then two velar initials, then two glottal initials, and this is good evidence or let's say it's the kind of series that we reconstruct as uvulars and in this case as labial uvulars. So I'll just sort of say, okay, take a good hard look at that series. And then we bounce back to this one and you see, okay, so this change only before front vowels and this change before all other vowels as a kind of prediction that you could go through all the shesheng series and series with these kinds of contacts should only be not with front vowels and series with the other kinds of contacts should be only with front vowels. So now to give you the bird's eye view of uvular developments, first we'll look at type A, the voiceless uvular stop becomes a glottal stop. The voiceless aspirated uvular stop becomes a voiceless velar fricative. The voiced uvular stop becomes voiced velar fricative and then we get basically the same stuff but with the labialization. So the labial velar, the voiceless labial velar stop becomes a glottal stop in a hookah syllable. The aspirated voiceless uvular stop becomes a voiceless velar fricative in a hookah syllable and the voiced uvular, sorry the voiced labial uvular stop becomes a voiceless, sorry a voiced velar fricative in hookah syllables. Okay, I mean you can see it on the page but I hope that you know me reading it out is a little helpful to kind of have you think through it in your own mind and then I don't you know misstate it too often. But now we look at the type B syllables which are a little bit more complicated. So q becomes glottal stop, qh becomes a voiceless velar fricative, so far so good the same as the other case but the voiced uh uvular stop becomes ya, then the the the next two are the same as in the type A syllable so qw is becomes a glottal stop, qwh is the the x but then the as we as was discussed before but you know just as a reminder the labial uv, the voiced labial uvular has a split where it usually becomes uh this middle chinese initial that we call hj but it also becomes a y before the vowels a and e. Okay, kind of I shouldn't put them in brackets because they're not optional it's they have to be there but the reason I'm putting them in brackets is because I don't quite know how to formulate but because I mean you see the chinese characters right so if you look at the the the chinese character I think it's called yin that means shadow it does not mean glottal stop plus w it just means glottal stop so that's kind of why I put the w in in brackets so you know probably what I should have done is written like yin he and then had the apostrophe w with with no brackets okay um but I but it felt wrong to kind of put the middle the actual middle chinese term there next to something that was not in fact yeah so that's why I put the w is in in uh in brackets but no it's these sound changes you know these sound changes on this slide are presented to be about hook oh syllables only and now we kind of uh take a step back in terms of you know we started with initials of middle chinese and we did some internal reconstruction on them and we came up with something and then we added more proposals based on um based on um shesham series so uh so where are we now yeah well let's look at uh origins for the voiceless vealer fricative uh we have three that we've proposed you know in this presentation one is from uh voiceless uh one is from voiceless ma and one is from an aspirated u-vealer stop so um baxter and cigar again like they did with um with initial yeah they think that's enough you know we have enough origins for um for the voiceless vealer fricative so if a series has only vealer readings but some of those are x then they reconstructed to a u-vealer now uh let me kind of put that another way in terms of why is it worth mentioning this if we have such a series it's already conforming to the shesham hypothesis right it has only vealers so kind of from a naive methodological perspective you could just leave it alone yeah you there's nothing there's no problem to fix but if you had that approach it would mean reconstructing x into old chinese so uh we look at uh one such series yeah so here's an example we have this series and in this series you know you have the the the gamma initial and the x initial and it's not a problem in terms of the shesham hypothesis right they are already homo-organic but um uh baxter and cigar's feeling is they don't want to reconstruct a a an x and a gamma into old chinese when there's already the machinery available to explain where they come from so it's just a sort of simpler explanation to take series like this back to u-vealers and then the result is there are there is no x in old chinese right that's that's their that's their conclusion that's we've managed to get rid of x so similarly we can look at uh at z so baxter projects initial z in 1992 baxter projects initial z and z are all the way back into old chinese uh and schussler also allows for old chinese z although it uh it it seems to me like it only comes up for him as some as a pre-initial before w so that seems like a really weird um distribution so uh you know i think he has some explaining to do there but um baxter and cigar in 2014 uh suggest uh origins of uh of old chinese uh or of of z in old chinese that involve uh clusters with s so things like sb and you know sd and whatnot become uh become z in their um in their system and i haven't gone into that because this presentation is about simplex initials and it's not about constant clusters but it's just to tell you they have reason to to remove z as a simplex initial so um kind of summing this all up we you know what what are the moves that we've we've done we've proposed laterals uh in in old chinese uh that explain these uh the connections and situation series we've proposed various kinds of voiceless resonance that um clean up uh various situation series uh we've proposed u-vulers and then uh we've removed x and we've removed z so um now i'm just gonna whiz back to the summary uh chart yeah so this is the inventory of old chinese initials that comes purely from internal reconstruction and then you see that it's got the hj initial still and it's got the y initial and it's got the z initial and it's got the x and it's got the labialized x and it's got the labialized glob stop and the glob stop so that's what we got when we just internally reconstructed middle chinese and then left it alone but now we've managed to you know add the laterals add the voiceless resonance get rid of the x's get rid of the glob scops get rid of the z and what does that give us we just whiz ahead ah here we get it so this is the uh the type b initials of old chinese as reconstructed by baxter and sagar uh i only feel inspired to make one observation which is they still have a glob stop and uh i think this is something that they felt kind of of two minds about because you could take uh the glottal stop back to um uh a uvular across the board but they think that they're there's there they're that you can kind of see there's some shesham series that have glob stops mixing with dealers of the of the type that we reconstruct um uvulars for but they did but there are also a few uh shesham series that are just glottal stops all the way through so then they think well in order to kind of index that difference why don't we keep the glob stop as an initial so they so they do that and uh and then you get the same stuff for uh type b syllables and those are all of the initials of of old chinese the the the the pharyngealization is the um is the way baxter and sagar index uh type a syllables right so they think type type a syllables were pharyngealized and type b syllables were plain and uh as you see uh from these you know kind of from all the stuff i'm talking about this a b distinction has a huge effect on sound changes uh and you know in terms of um if something's type a it does this if it's on type b it does this so the the the and and the predominant effect there is um is palatalization in type b so there's a question you can ask yourself which is like like why did everything palatalize in type b so the the the uh the original proposal that baxter had in 1992 was everything's palatalized in type b because there was a palatal medial in type b and that kind of follows on from charlton but it ends up being uh not a very nice um um proposal for one reason because it means like half the syllables in the whole language had a medial yeah in it and that's just typologically bizarre right whereas the pharyngealized pharyngealized non pharyngealized is a kind of contrast that you do get in languages at a kind of half half ratio so like particularly in somatic languages you have these emphatic and non-emphatic syllables so um that's maybe the closest point of comparison where type a would be emphatic and type b would be non-emphatic and um and and then the the question is well why would all non-emphatic consonants palatalize generally and um this gets into this kind of the areas of linguistics that i am not super comfortable with in terms of you know articulatory phonetics and whatnot there are parallels in arabic dialects but there is a easy uh or easy ish thing to imagine which is um if you have a pharyngeal non pharyngealized uh distinction uh it has a quite a big you know semantic load if you like uh so over time you are trying to keep like pharyngealized ba and um and non pharyngealized ba distinct yeah and that will generally mean that the non pharyngealized ba is kind of further forward in the mouth and the pharyngealized ba is further back in the mouth and things that are further forward in the mouth might have then start to be indexed by palatalization so that there's a kind of let's think of it this way at one point the phonemic contrast was uh pharyngealized non-pharyngealized then the um phonemic contrast or then there's a sub-phonemic contrast which is which is in order to keep them at the front of the face uh you start to get palalization in the non pharyngealized syllables and then at some point you know the the next generation thinks the contrast is between palatalized and non-pharyngealized and then they reinterpret the pharyngealized non-pharyngealized distinction as a palatalized non-pharyngealized contrast I think something like that is the is the idea um and then something but something to flag is um this reconstruction of old chinese is typologically bizarre because it has velar uvular distinction and pharyngealized non-pharyngealized distinction so you have you know so you have pharyngealized velars and non-pharyngealized velars and you have pharyngealized uvillars and non-pharyngealized uvillars and as far as I know which is to say as far as is discussed in the sinological literature no real languages are likely okay so you know this is not an objection that backstab and cigar are naive to what they point to is that the evidence for these two things uh a velar uvular contrast and an ab contrast come from radically different time periods so basically the ab contrast comes from the I mean in a sense from the rhyme uh tables even but but probably I mean you can push it back to the writing books yeah so it's kind of late whereas the uh the uvular velar distinction comes only from shesham zu so it's very early so so so their feeling is look we don't really know what's going on uh which is to say at the moment there was a uvular velar contrast probably the ab distinction was something different than pharyngealization but pharyngealization explains the palatialization so so probably it's something like there was you know there was this kind of primeval ab distinction that we really have no idea what it is at the time you had the uvular velar distinction then the uvulars were lost then the loss of the uvulars kind of allowed whatever the ab distinction was at that time to change into a pharyngealization distinction and then you had the pharyngealization distinction until you uh until that led to the sort of vowel warping and palatiation of the sort of early home period so that's let's say i i think that's the my best attempt to say what backstreet cigar things actually happen and i uh have done a little bit of work on this uh trying to kind of really pin down when the different sound changes happen and i think it sounds about right to me because you can imagine it's something like uh like this as you saw that the there's kind of simplifying a lot of issues the uvulars generally turn into frickets yeah so you get um uh you get um well yeah the uvulars tend to turn into this so i think that's can also be understood as a kind of re phonology re sorry re phonologization right which is that you had a you had you and at one point you had this nice uvular velar distinction and then you know to speak a little tele teleologically and then they decided we don't want that in me we're going to only have velars and then the uh the uvular stops were reinterpreted as velar frickets and then you have that moment uh and then at that moment it's perfectly fine to have a uh uh uh pharyngealized non pharyngealized distinction i mean the the way the the the way things are usually presented it's kind of the reverse which is so i don't like to talk in terms of marketness i think it's kind of unhelpful generally but i'll allow myself the indulgence at the moment which is um that the ab distinction uh before a certain moment uh uh you had type b as unmarked and type a as marked with pharyngealization but then at a later moment that flipped and you get type a is unmarked and type b is marked with paloalization and so um if you want to kind of think of it in terms of you know features i don't know i don't know how to do stuff but actually like you know you have uh you have kind of uh plus uh the the plus pharyngealized becomes the the minus paloalized and and the minus pharyngealized becomes the plus uh paloalized um but uh but another way of putting it is like it you know what's a good phonemic analysis of middle trends and i just have to admit that that's not a question that interests me yeah so uh in in this presentation and in my book for example i just treat the rhyme categories of middle trends as as rhyme categories you know i i don't even try and figure out how many nuclear valves there are in middle trends you know because i mean and and i would say this sort of i don't know in the same way that like there are several different ways to phonologize uh put an flaw and that's kind of interesting i mean actually it's interesting methodologically speaking was very interesting and it was very important in the history of uh phonemics as a discipline right that they're like oh my god there's more than one way to analyze the language um i think that uh middle chinese looks like that to me where it's kind of well you know it's clearly chinese what to what to say you know and there's that if you if you really want to think about well what's in the medial and what's in the initial and how many vowels are there and when i i think it starts to get a little bit um i don't know drawing but not not to say you know my work isn't dry but uh the analysis of middle chinese in terms of features or in terms of phonies is just not something that excites me so but there's a huge literature on this kind of stuff let's try it again for sort of you know uh once upon a time there was a language that had a uvular velar distinction and half of its syllables were something and half of its syllables were something else maybe that was long and short vowels maybe that was two syllable and and one syllable you know that these are these are existing proposals that some people believe um who knows more let me just talk it through in terms of stages you have sort of sort of the real or stage we'll call it sort of stage zero we have a uvular velar distinction and some other distinction yeah then at stage oh and incidentally we have no fricatives yeah i mean no except s yeah uh and then we have stage b uh or sorry i switch from number it's the letters okay we have stage one version of stage zero so now we have stage one and in stage one we we no longer have uvulars but we've got more fricatives so we've got uh we've got uh velar fricatives in particular but we've also got the glide yeah which we might not have had before or we might have had before we don't know about that uh and we and the second distinction whatever it was has now changed into a fringelized non-fringelized distinction so so this is kind of stage one then at stage two you have uh you you still don't have uvulars you still have velar fricatives uh and you've lost pharyngealization and you've gotten powers and then what you're saying is at that moment half of your syllables have a medial yeah so so this whole pharyngealization theory has gotten us exactly nowhere okay that's a reasonable point it's not necessarily the case that at a segmental level at that moment half of your syllables actually had a medial yeah it could be that that what has happened is you have now dentals and palatal initials that it could be that you go in a sense straight from a pharyngealized non-fringelized distinction to whatever happens immediately after this you know imaginary moment where half the syllables have a yeah without actually passing through the stage where half the syllables have a yeah yeah that's a good question actually um so the question in a way is where did this pharyngealization idea come from uh and um it comes from an article of Jerry Norman's and so I would have to actually really look at what Jerry Norman proposed but I will tell you something that's highly indicative and I left it I didn't mention earlier I mean because you can't mention everything but also because I didn't want to confuse it with the uvular question so now we have to say forget that we ever knew anything about uvulars and the uvular hypothesis so we're talking about something totally different which is just the abyss okay so in early loans from chinese to mong mian type b vealers in chinese are borrowed as mong mian vealers well that's not a shocker okay but type a vealers are borrowed as mong mian uvulars so that I think is I mean let's say I think there are other arguments but I think that the best one is stuck in my mind as that means that that that the aliphonic distribution of vealers in that moment in chinese history was such that type a vealers were were perceived of by mong mian speakers as uvular and uh it's it's completely normal in languages that have this kind of emphatic non-emphatic distinction to have the emphatic vealers be phonetically realized as uvulars I think that would be fun uh so to put that another way uh type a was uh pharyngealized and type b was paralyzed at the same time yeah I think that's fine and then actually it's quite sort of elegant because then you you just sort of dropped the marking of the pharyngealization and keep the palalization and then the palalization moves from a feature associated with the syllable to a feature associated with the initial that I think that's a beautiful explanation yeah I think this is this very much uh also seems to be the case to me an extra stage um of some other previous contrast like pharyngealization uh it seems like a necessary complication because uh from my experience uh both from you know how have managed to analyze uh the phonetics of pharyngealization and the palalization myself and from how um what I've heard um of other people uh say how they perceive both phenomena uh it would seem that um and this ties into what you said that indeed palalization and pharyngealization are sort of opposites on the same spectrum and it's really about the phonetic contrast and you know speakers who are for instance more uh used to um the default rather uh palatalized uh production of sounds which doesn't mean that they use some kind of a yoglide no that that that is uh um that you know that that is a part of it but you know they perceive sounds that are less palatalized as being uvularized for instance you know in russian um we have the the so-called palatals and the the the normal constants now some foreigners experience our normal unmarked uh stops as actually being uh sort of uvular so I think it's it's it's as much about the perception of the people who um have a different uh who position their uh unmarkedness a different point in that spectrum that that has as much to do um with the the you know palalization or uvularization or pharyngealization of sounds as as the actual sound production does and in this case uh I don't think you really need to assume this second later stage of um pharyngealization in an earlier period because I mean ultimately it's about it's about a phonetic contrast and if you look at um I mean I mean you don't you don't really find pharyngealization much I think um in um in northern or eastern Asia whereas palalization does occur a lot and and then again maybe something else to think about is um sort of develop the vowel harmonies which actually evolve vowels more than consonants so yeah uh basically yeah the main point is that exactly I wanted to reinforce your suspicion that which is also my suspicion as of already a few years that so-called palalization and pharyngealization are just opposites and you know in very basic terms phonetically palalization what people call palalization involves uh somewhat a constriction of the air passage with in most cases a yeah promotion of of the tongue um although in the case of the the velars for instance um yeah the tongue is actually pulled kind of up and backwards a bit but in the end it still achieves um a hollower um mouth cavity whereas for instance if you look at um maybe the Arabic emphatics they actually produce a much greater volume of the cavity um you know you could do you know you could also kind of translate uh this contrast figuratively onto sort of you know and this will maybe sound a bit crude but you know the the more feminine and the more masculine sounds and and and you can see easily how that becomes a continuum yeah so can I jump in here so just a few observations uh what uh so if I can concretely paraphrase what you're saying there's one particular um claim that I think is is useful to highlight which is that uh you could have had the old Chinese uh in let's say the post-yuvular era but the pre-palatal palatalization as a sound change era um you could have had them speaking Russian and have the Hmong Miens still borrow type A velars as u-velars because of because of how they perceive the contrast in in old Chinese I think that's a very important point it has not been sufficiently understood in the literature which is same yeah I think there has been a sense that well if the Hmong Miens borrowed these things as uh as uh as a u-velar velar then they must have been somehow u-velar velar whereas you're saying no that's perfectly consistent with them being sort of velar and palatalized so if their velars were mostly sort of palatal velars then a normal velar to them would sound like something that's really yeah exactly yeah so I think that sounds fine um then I do have a slight objection to your sort of I don't know in the Middle East you u-velar eyes in in Asia you parallelize that's that's a you know that's a caricature but my understanding is that actually a lot of Mongolian uh vowel harmony has to do with you know plus minus ATR which is I don't I don't I've never even I've never understood what what it means for me to advance my tongue word I just don't relate to my tongue in that way but um uh but uh the I don't even know it's uh but the the back vowel yeah the back the the velars in Mongolian uh back vowel uh words are pronounced phonetically as u-velars so um so I think let's put it this way that I I object to your sense of you know palatalization somehow more likely than u-velarization but I completely agree with you that it's kind of six and one half a dozen of another and is in all of these things u-velarization palalization and vowel harmony are sort of mixed up together in a bucket yeah all I'm saying is I mean you would have to actually I think find some uh examples of East Asian languages or at least you know languages sort of to the east maybe of the Zagros mountains that uh actually engage in something like you know the examples that you draw you and uh Mark whose presentation I also watched they draw on examples from Arabic but I mean look let's be honest I mean the distance and oh sure I but no I don't think I that's just that that's just the typological argument of this can happen it's not an argument to contact yeah um but I would say that um actually you know what I would point to you know which I think is the best thing to point to is the Guarani languages have uh let's say depending on the Guarani language u-velar-velar contrast and a uh a systematic sort of uh contrast on the vowel between uh what they call velarized and unvelarized now I I'm not a phonetician so I don't have a good sense of how pharyngealization is different than velarization Gongshun actually doesn't work on this uh both in terms of Tongan and in terms of Chinese that is unpublished uh and I don't think it's particularly likes being published soon it would be nice you know to ask him about it um but but but but here's I think the crux of the issue is is at some point in Chinese history you start getting mergers right which you you start getting like uh the the you start getting the let's say the type a voiceless nasal and the type a voiceless lateral merging as and you start and you get the type b voiceless lateral and the type b voiceless nasal merging as right so so I think there's just a question you know kind of I don't know to those of you who are phonetically sophisticated about like what kind of thing would go on it would make those mergers happen and and and we can call those parallelization not in the sense of you know synchronic parallelization but in terms of diet so that's you know the ab distinction triggered those mergers and then the question is um let's say really from the Sino-Tibetan perspective like really really old what should we do with this ab distinction which is to say you know I'm I feel like look at well at one point it's parallelization at one point it's very unlikely to be uh pharyngealization because we reconstructed u-wheeler contrast so I basically agree with you and with Sid then maybe it's a little unhelpful to reconstruct it as uh pharyngealization you should just if like just keep it as parallelization or figure out what it actually comes from uh but we do need to figure out what it actually comes from and then again I mean why should we be uncomfortable with just the system of six vowels as opposed to three vowels that are governed by these two types I mean if you look at no we with that no that's that's wrong we have six vowel with with with both types okay so let's say which is to say um you you I would prefer to understand for pharyngealization so we'll just say the type A type B distinction as uh linked to the whole system but Baxter and Cigar link it to the initial and then they come up with a very gorgeous list of initial that's fine with me but if you linked it to the vowels then we would have a 12 vowel hypothesis super super easy um so this is roaming this goes back to Roman Jacobson where basically he says peas and k's I call grave you know peas peas and the lake I call recue it's I mean you can really think of it as a purely terminological issue and it's just that many many sound changes are conditioned by this difference right that like the vealers and labials behave one way and that's sort of dentals and advocates and red reflects behave another so it just simplifies the discussion of historical phonology immensely if you can just say grave grave grave um I mean and and and then there is a there is a question that uh which is is there a reality behind this and I believe there is and and it's it's not articulatory it's acoustic right which is why you get things like uh you know labial vealers turning into labials in greek and whatnot so there's some kind of acoustic things having to do with formants that you know the people kind of people who care about that stuff you can ask about that you know which is I I think it's a real difference in terms of acoustic phonetics between grave and and uh and acute but we don't have to treat that way we can treat it as just a kind of term watcher thing you know like uh let's say Baxter says there aren't very many and forget about it yeah um so it's a it's a philological question and actually if anyone right wants to write a paper you know for this class I know that's an option uh or you know in life it would be uh interesting to look at all of these types of syllables that we're choosing to ignore you know the the the laimu division two initials the the kaiko uh yunmu uh initials there's there's various kind of small classes of phenomena in in the guanyun that that historical journey technology tends to ignore and then it's it's a it's a philological question of like can those examples actually be explained because if they can't be they shouldn't be ignored basically they think that a loose pre-initial before a uvular changes the the uvular into a velar and then the the loose pre-initial is locked and this is something we'll get to not at the next you know maybe in the next lecture maybe in the next um and and I'll tell you that that to me this really feels like they have too many tools in the toolbox it's like oh well you know if it's like whack-a-mole or something it's like well if this comes up you hit it with this and then if this comes up you hit it with this and um and uh you know actually one way of thinking about this is like if you if you so we've tried to formalize or we I don't know what I mean but I and some people who helped me do things that I'm unable to do myself um I've tried a little bit to formalize the sound changes with uh with finite state transducers and the and the nice thing about a finite state transducers is that you can run them backwards right so you can say okay rather than always thinking like okay