 Right so Yeah, so it's just I mentioned this already, but in some cases back from scar have reason to Suspect there's a pre-initial, but they can't tell what its place of articulation is so they write it with a capital C Okay, so we start with pre-initial s Basically all researchers who think there are constant clusters at all Think there is a pre-initial s, but there is kind of quite I mean I would even say vicious Disagreement about in which circumstances you should reconstruct This type pre-initial s and Zev Handel has a nice article. They kind of where he's trying to kind of make the camps kiss and make up So so so I think that's a good place to go for that, but you know, I'm gonna stick with back from cigars presentation so Type pre-initial as one of the origins of Middle Chinese type re-initial s as one of the origins of Middle Chinese s So here's a shei sheng series where you see contact between M initial words and an s initial word. Yeah so Here and in other places you will see that I give a Pre-chin form of the script And I'm not doing that in any kind of rigorously paleographic way It's just to help you see that the analysis of these characters as belonging to the same shei sheng series is Plausible so if you look at the fourth character on the slide It's not really obvious that it belongs in this shei sheng series But if you look at its pre-chin form it it seems pretty plausible, right? so in this series we have you know, let's say two two words that start with M one that starts with a With an X which actually it's probably some kind of font problem in my presentation There should be a little circle under that M because as we saw yesterday one of the sources of velar fricatives is voiceless labial Resonance so, you know in if you look at those first three you're like, oh, okay We we did that yesterday But now today we add in look what is an S initial word doing in this series? well Baxter and cigar say Actually, it also was an M initial, but it had an S pre-initial. So that's their solution to that and And yeah, and then I just am trying to go through kind of systematically in type a syllables So here we have an SM reconstructed in a type a syllable and here we have SM Reconstructed in a type B syllables. Okay, so now Yeah, I don't want to go through this too fast, but it's kind of Much of a muchness is what they would say here in the UK so SN develops into S in type a syllables and Here's the evidence. We have a sheshring series. That's mostly an initial, but it has an S initial word in it and In type B syllable so again mostly N initial but one S initial word And now with no so we have in type a Here I'm mostly no initial series, but with some words that start with S and in type B Again, a mostly no initial series, but with one word that begins with S in middle Chinese Okay, now With the lateral so I mean and this is actually a nice series in terms of showing you just how messy a Middle a middle Chinese, you know in middle Chinese garb a sheshring series can look but you we have this Contact between ya and da and ta so we saw yesterday. We reconstruct that as a lateral series And then We We have some S initial words one type a one type B which it's convenient because then we can just use the one series to prove both sound changes so we reconstruct SL as one of the sources of S and And You know a same style of argument here. We have a uvular series That suddenly has and it's actually a labial uvular series where we have a character with an S initial so So why not reconstruct a type pre-initial S? I will state it slightly differently, which is when a person has a syllable-based Orthography and they want to write a word that they don't know how to write they turn to a similarly pronounced word and And then press it into service, right? So I'm actually this is I haven't written it yet But this is what I'm planning to talk about in my evening lecture next week. So a very simple You know example, let's say let's say we speak modern English and We've just invented writing and so we have we draw a picture of a deer to mean the animal deer And then we find ourselves wanting to write a letter back home to our mother And we want to start it with your mother, but we have no no writing system So we don't know how to write the word deer So we draw a picture of the word deer and and now that works well because it deer and deer have segmentally identical Pronunciation at least in my dialect. Yeah, okay and then and then later, you know in the letter I find myself wanting to say I bought a new steering wheel for my car and I think gosh, how am I gonna write the word steer and Then I think well, it's awfully similar to deer, but it's not the same So maybe I'll like write a picture of a deer with a little car next to it. And then my mom will Will be able to say to herself, okay What word sounds like deer but has to do with cars and then she'll say oh, he probably means steer here Yeah, so that is the That is called the rebus principle Which is basically I mean, I don't know it. Let's stick with bolts here He thinks basically whenever you invent a script, you know ex nihilo you you use the you start with very iconic You know things and then you move to more abstract things through pronunciation using the rebus principle so we imagine That's what we're looking at when we see a shesham series is like this is a series of morphemes that That could that could be related to each other via the rebus principle Which means that their pronunciation has to have been Pretty similar. Yeah, and then the question is which which is partly just a question of your a priori assumptions is How similar can something does does something have to be to be permissible to use this rebus principle in the let's say in leaf on quays version of the shesham hypothesis you say Syllables must rhyme and have home or organic initials in order to qualify for this kind of rebus relationship and And and then I would say that you can really Set up different Chinese historical phonologists on some kind of spectrum In terms of how doctrinaire they are about this principle Where Baxter and cigar are very doctrinaire like they really think, you know, you cannot leave that as alone You cannot say that, you know, people are like, oh, yeah, you know guns son, whatever. Yeah Whereas other schools feel like, you know, come on like you The each act of inventing a new character happened in a specific time and place By a specific person and there's no way we can get in their heads and know exactly what they were thinking and Similarly, we can be pretty sure that they didn't hold some giant committee meeting where they all agreed like we will only invent new characters That have home or organic initials and rhyme. Yeah, so so so those people feel like you shouldn't push the shesham hypothesis too strongly And this is I mean kind of maybe you'll get used to this But I feel like, you know, the truth doesn't matter what matters is what's good methodology and clearly good methodology is to be as doctrinaire as possible, right and then and then Let yourself propose ridiculous things for ideological reasons and That is how you will discover Why they are ridiculous Yeah So that's you know, my feeling on on on this and I think that actually in the preface to his his dictionary Schuessler does this Schuessler is Thinks back strings guard go way way too far And so he tries to tell little stories about like why in this and this situation There was a counter prevailing consideration where basically, you know as someone found himself in in a tricky situation and So he found a pragmatic solution in terms of how to write something that that over read the shesham Hypothesis and I think that's you know, that's you know, how how that that's right, you know, which is to say That's exactly the conversation we need that it's yeah, exactly. Yeah, we need a conversation between the the the the ultra ideological shesham adherents and And and their critics who can come up with very good stories about why Being so ideological doesn't work in particular cases another labor uvular now in type B a Series that has an S initial. So let's reconstruct another as cluster But in this case, I'll just say it's not clear to me how Baxter and cigar decide whether or not This particular word is aspirated And I will also point out that there are no type B syllables that Baxter and cigar reconstruct with an S prefix to uvulars, but without Labio Labelization so we don't get any, you know as far as I can tell in the 2014 book We don't get any sq's without a w or a medial R and I just think that's a The pattern I noticed maybe it's totally meaningless, but that's I think that's the sort of we need to be on our toes for these sort of Phonotactic holes and and wonder whether maybe that they point to You know, maybe problems in their analysis. Okay, so Now Here's a series where we actually have an S connected to To a velar But and actually this is this is this is one where Morphological speculation comes in they think script notches and wedge are etymologically related And I think that's actually how they presented in their book But I I just sort of point out to you that this is also a sheikh and relationship So you can you can you can pause it at least some kind of Of S cluster in the word wedge purely based on the sheikh and relationship But the mismatch of the u-vehler with a velar is a problem And you you could of course, you know reconstruct the the velar as a u-vehler But the series overall does not have any other words That are incompatible with a velar initial reconstruction. Okay, so now, you know, I'm moving on again another example where we have a u-vehler series but with a With an S word in it And in this case, I would say it's not again. It's not clear to me why they choose aspiration rather than without aspiration Or why they don't reconstruct this series with with velar nasals because you can have a velar nasal origin of both the velar fricative and the And of course the velar nasal, right? So and and maybe this is just a case of showing you that that there's a lot of flexibility in their system, right? If you just look at these three examples, you could reconstruct it as they have s s q h m q h and and q h or you could reconstruct it s ng ng and and And you know ng the sort of voiceless ng and I think somehow that's this, you know I don't know this kind maybe just a question case I think that would have been a slightly simpler analysis because it would have left the the the view the Second character the velar nasal it would have just left it alone. Yeah, okay but But in case they do what they do and just to point out that they think there's actually a dialect split where s q h develops into in a different dialect and actually I think one of the most interesting and kind of productive areas of their book is pointing to Hypotheses of dialect variation that can then be more systematically explored Okay, so now, you know sts cluster. Well, here's this one's a little well. Yeah, this one's a little oh Yeah, yeah, okay So you have a ts in the first one and you have an s in the second one So they say well, maybe that's it comes from an sts. Yeah, and here's the same in type b So we have a ts h and s. So you think ah, maybe it's the sts Okay, and now those were all Examples of s clusters that became s in middle Chinese So now we will look at s clusters that become SR in middle Chinese and a you know is basically Not going to shock you that they're the same kinds of situations where there's a medial are so in the most Simple case They think SR becomes SR. Yeah, and Here's the the the evidence for it. We have a kind of our series And and there's an s in it. So Or there's an SR in it. So maybe take it back to SR and I Will just point out that They appear not to reconstruct any words with type a SR Although they do reconstruct s Voiceless are in type a so that's just I don't know. I just feel It may seem kind of random from you say, oh, look, here's a hole. Here's a hole But I think it's good to to you know, draw attention to these things. Okay, so then we also have Soon becomes SR like in this case where you know, you have Preclear, right? You have Vila nasal series And then you have an SR in it. So you take it back to s Vila nasal are They appear not to reconstruct any words in type a with this cluster. Okay Now that was it for SR. Now we move on to s y in middle Chinese So they propose these changes in Type a syllables so stuff stuff They all become sure But the SK to show change only occurred before front vowels and this is part of the the so-called first palatalization and Which was which we came which I hit on a little bit in the presentation about the six valve theory Which is that you got palatization of Vila's before front vowels as one, you know moment of Palatelization before you get the overall palatization of type B initials. Yeah Okay, yeah, that's what I just explained so so the restriction is result of the first palatization of old Chinese and And Yeah, so and and and this is basically, you know How you get the the the relative chronology, right? Yeah, so we imagined that you had something like ski So let's say key changes to chi. Well, so then you think ski probably changes to ski and then Ski changes to she as as an exact parallel of the change that we also saw saw of ste changing to s Yeah, and I'll just actually I feel obliged to say because I'm a tibetologist these sound changes of shuh to shuh and star to Or sir to sir happened in old to them like or are like Philologically attested to have happened in early old to bed and so on so their sound changes that feel, you know, very dear to my heart okay now aspiration after Initial s shih sheng series mix home or organic initials of different manner types. So it is not possible to distinguish Aspirates on the base, you know aft like it's Let's put it this way sp and sph, you know are neutralized in middle Chinese because both of them turn into s. Yeah, if you like, yeah so So back strings to guard do reconstruct these differences, but on the basis of aspiration in min or other dialects So that's so that's just basically an answer to if if you were wondering to yourself How can you know whether these these things were were aspirated or not because it's neutralized in middle Chinese After a pre-initial s that's the answer is is you look at things other than shih sheng series So here we have as tea goes to shuh. So you see, you know a nice dental You know dentals and pals actually this this answers This is an example. It came up very end of yesterday. Are there shih sheng series that mix? Palatels and dentals that argue for the sound change of dentals to palatals in Tai B. Well, here's an example, but it's also an example of Shuh initial words intruding into a dental shih sheng series. So is evidence for sth goes to shuh change And here's You know, I just I've discussed I presented these changes already now. I'm giving you the evidence, right? That's what's happening here. So here's a velar series and you notice one with high front vowel and evidence of for scuff because there's a Shuh that is intruding into it Yeah, and then I Will just mention that back string cigar think that in type a syllables scuh changes to cuh rather than to shuh But they don't use shih sheng evidence for that. So I'm not going to mention it now and by implication They would say that scuh develops into cuh also in Tai B syllables You know when not in front of high front vowels, but they don't explicitly say that as far as I can tell Okay, so now with aspirated k And they propose the aspiration Because of aspirate pronunciations such as actually including in mandarin in in guanzhou Which point to a middle chinese reading With an with an aspirate palatal But the but the actual middle chinese reading is With an s, right? Okay Now tight pre-initials as an origin of middle chinese z and you'll remember yesterday when I presented the The simplex initials I said we can get rid of the z because it has so many origins Well now is where we're going to see those origins of z They all involve pre-initial s. Okay, so here we go before voiced uvular, so this is a A uvular series Why is it a uvular series just to kind of you know, I don't repeat things so you get used to them? Well, because you see contact between yaw initial and a velar So it has to be a uvular series and we reconstruct the yaw to uh to a voiced uvular And then we have a z that you know that intrudes in this Series that we think is a uvular So why not reconstruct it as an s before a voiced uvular? And then you can understand the uh the voicing assimilation As the reason that it changes into a z Bactionary cigar also reconstruct sd and smt as sources of z But not on the basis of sheshen evidence. Okay, and then we also have you know, uh s Little g r changes to z r Uh in this series And uh now, you know, we go from origins of z to origins of zh And they're very parallel to the origins of shah that we saw so Here is a series that is clearly velar Because it it mixes uh You know jh in in in type b and dh in type a And then it has a zh initial. So let's you know, let's uh propose that it comes from sg And uh in general, uh, they believe that sg developed into g But uh their evidence for this Does not come from sheshen series. So again very parallel to the sk situation Which is that this sg goes to zh would only have happened before high front valves in the because of the first parallelization Okay, and now Tight pre initial s as one of the origins of sra So bactionary cigar, uh maintained that a tight pre initial s before q r gave rise to sra And here's uh evidence for that. So we have a glottal Uh initial which might come from a uh From a uvular And we have contact with sra So they they let's say this is the kind of problem That motivates them, um Reconstructing, uh This sound change and actually maybe this is a good one in terms of uh, you know shusler I mean, I don't know about what shusler would say about this particular case But he would look at something like this and say like like really, um, it maybe I don't know maybe maybe these aren't a sheshen series or maybe there's you know, this is the kind of place Why I think he would be a little bit, uh Suspicious in terms of well, it's not Prima fascia totally obvious that that a sra would come from a scura uh But if you want to make The sheshen hypothesis work on this sheshen series. Well backs from cigar have managed to do that. Okay And then a similar one from type b So you have sra, uh Connected to a glottal So, you know, there you go And now the aspirate equivalent. So this is, um Pre-initial s As one of the origins of tsah. So backs from cigar Take tsah as a reflex of a voiceless resonant Prefixed with s if there is any evidence of connections to words that are otherwise known to have resonant initials Whether voiced or voiceless So here's an example S before voiceless n in type A syllables they think changes into tsah And here's the reason why It's uh because of uh, these two words, uh, and you see why I've given the, um The, um Pre uh chin orthographic forms because the the the modern forms of characters doesn't make it obvious that these are related So the second one means person and the first one means a thousand. I think if I if I'm wrong about that, uh, someone told me um So, you know, they were written very very similar in in, uh, right back to oracle bone inscriptions So they think well, they must have been similar then, right? And they don't cite uh station evidence for the same change in type b syllables But they do reconstruct, uh, the change in some words. So in particular this, um, this word, uh twice And I don't they don't quite spell it out But I think it's pretty clear that what they're thinking is that uh, the word twice is related to the word two Yeah, so it's a etymological speculation. Uh, and, um, and that this sound change Would give you the machinery to link those two words. Yeah, okay So now, uh before a voiceless resident, uh, sorry voice of lateral So we have a nice lateral series no connection between a da and a ya And then, uh, there's a suh initial that intrudes. So why not use, uh, you know, the machinery we've already developed to solve that problem They appear not to reconstruct any words, uh in type a with this cluster Now the aspirate equivalent, uh, so here, uh, we have, uh, a connection between, uh, A tsuh initial and a thuh initial With the with the chub, you know, it's the usual dental stabilizing type b So they say, okay, maybe, uh, you know, we've we've already seen this kind of argument before Uh, yeah, so so that's another sound change. Uh, and then onto this one Uh, uh Which I think I've mentioned already that s q h changes to suh or tsuh Depending on dialect I include this here. It's not exactly xie sheng evidence, but it's pretty similar because two readings of the same character Right. So this character means slipper or shu and, uh, it has two readings one with, uh, uh, uh, a tsuh initial and one with a Tsuh initial. So I would say generally speaking, I don't think they articulate it as a principle, but it makes sense When there are two readings of the same character in the same meaning, that's kind of, uh, pretty good indication of of, uh, let's say a dialect Split and then, uh, borrowing into the same, um let's say standard, uh That gets represented in the guanyun okay So it's not clear to me on what grounds they, uh, reconstruct a uvular in this word kind of, uh per se So it appears to be because, uh The the the Because of this loan into pro-mian, which has a vela fricative initial Uh, yeah, so it so I think the presumption is that this word shares the same root Uh Yeah, I won't I this I don't want to get sort of drawn into these details But they're in the presentation if you want to look at them later Um, but they do show kind of that, um The level of sophistication, I think and I mean that in a good way, uh in the 2014 book, which is, you know, you look at, uh this intricate interrelationship between, uh, borrowings into unrelated languages, uh, and, um Sheshaeng series and etymological speculation okay, so now just, uh Sticking with uh, well, yeah, I'll try to get let's try to get through the the s initials so They propose, um Both sra and sra as two sources of sra in middle chinese Here's, uh, some sheshaeng evidence Uh, a character that starts with sra having contact with the character that starts with sra um in type a syllables, uh Instead as we saw earlier, um, they have sra developed into tsa And then here's the other source of tsa, uh, which you get from, uh, this case of a character with two different readings But I would say that I'm a little skeptical here because if the s prefix which you see in the first of the two readings in their system Is morphological is a morphological aspect one would expect it to somehow contribute to the meaning And And uh, and it doesn't the two readings, you know referred to the same meaning. Okay Uh, oh, yes, so we're almost done with the s is so so we get s as a source of the so So so in the first of these bullet points, you see sheshaeng evidence. So you see a contact between piang and Uh Zhang ziang So they think well, uh, you know, the the the second character must somehow have a labial in it because uh, Otherwise they wouldn't be the same sheshaeng series. So they reconstruct that back to uh, sb And then uh, this is you know, kind of, uh Not technically a sheshaeng series, but exactly that sort of rebus principle I was talking about which is this one character Uh In early sources Splits into two characters to distinguish two different meanings. It meant nose and self So the reading meaning nose is with a b initial and the reading meaning Self is with the z initial. So it's it's very similar and they take it back to an sb cluster So now that's basically it for uh, a pre-initial s is based on, you know, mostly sheshaeng evidence uh, and I'll just sort of, uh You know wrap up that section and I have a concern because, um I don't see much sort of elegance in this Uh in these proposals. Yeah, so so tsa tsa tsa you might expect them to have similar sources in in pre-initial s, but they don't Uh, and I mean, maybe this isn't a big consideration But I think it points to the fact that what's missing from the 2014 book which was in the tooth in the 1999 sorry the 1992 book is really ordered sound changes like Uh, we let's move away. This is sort of, you know, I don't know a proposal if you like as a discipline We should move away from this sort of ad hoc Uh proposing sound changes to fix particular sheshaeng series Uh and say no, we need we need to come up with a coherent story of how did the one language development the other one through ordered sound changes And I don't see that that is very easy and actually, you know at one point I sort of intended to do that in my book and it what it turned out to be it would have delayed the book by You know, uh a few more years. So I decided not to uh, and yeah, so that's a uh, that's one sort of reservation and then also, uh You know, I've been pointing out gaps. So let's look at the distribution of pre-initial s. Well, we don't have, you know spa spa ska But we do have zba and and ska in type b syllables. So it just seems very like You know Weird distribution. Now, I think what back to Mr. Garwood say is well these clusters probably existed and we just haven't figured out what uh Evidence to look for to reconstruct them Or uh, they were lost without a trace things can be lost without a trace But I think both of those stories of where they lost without a trace or what evidence should we be looking for to reconstruct them Would be would sort of be natural consequences of Of an ordered list of sound changes as part of our theory of olcanny's uh reconstruction okay, um Oh, and then yeah, and then as mentioned this is just just there's more, um kind of Objections I have to the distribution of the s So it's clear that much more work remains to be done on old chinese phonotactics I think that there are these things like also, you know, having a first person pronoun that starts with a velar nasal happens in sino-tibetan and in, um in, uh In uh, what language family is kasi and um in uh, uh, austro asiatic, uh and Yeah, um, and and I don't know I My inclination is to say these are these are sort of, um, that there are materialist explanations for these things somehow in terms of, uh, of, um, you know, um Like I don't know and human biology and communication needs. Yeah, uh that are also things like, um Actually one I've been thinking a lot about recently is is it a coincidence or is it a borrowing that? That uh, uh, indoeuropean has a word something like quen for dog And that old chinese has a word something like quen for dog. Um But actually baxter and cigar reconstruct that that not with an n but with an r So that you get something like core as um as dog and then I sort of think well it could be a coincidence because Uh, because you know having a word for dog like curl or you know Is is not so surprising I wouldn't I wouldn't go that far Because the the use of a second character to represent loose pre initials is is not at all systematic in old Chinese and I would again, uh sort of compare to um sumerian where sumerian had uh Like actually it's a very interesting thing, which I think is is is a very good comparison Maybe some of you know more more about sumerian than I do but the earliest Stages of sumerian have the worst evidence for prefixes Uh, and then the as the prefixes actually became Unproductive or as no one knew sumerian and had to learn it as a second language The the writing of the the prefixes becomes more systematic And I think that has a natural explanation, which is actually, um That when the prefixes were totally productive You didn't have to write them, you know, you could just say the kind of uh overall Syllables and people would know the the necessary grammatical interpretation and apply the necessary prefixes for that grammatical interpretation, but the the the less you could rely on people to um To actually Productively use this morphology Or or to have command of it the more you had to write it explicitly And I think actually in in a sense the the problem for you know a whole old Chinese Is that we have no akkadian, you know, there was no there was never a moment where a huge body of of foreign speakers Needed to learn old Chinese when it still had pre-initials and if that had been the case I think we would have much stronger evidence, but I I think kind of I mean you can almost think like as a thought experiment What would have happened if the sumerians lived until today? Well, they might well have Ended up with a kind of highly analytic, uh, you know tonal language And the philological evidence for those prefixes would be extremely scant Because the occasion never would have arisen naturally to write them down. They just You know, uh, yeah, anyhow, that's my sense Which is to say the preservation of those two character writings is really specific to this For the reasons that I uh, I said we will get to is there shesheng evidence for loose Pre-initials, uh, I think so, uh, but it's later in this presentation This really doesn't bother me. I mean, maybe it's because I'm a tipatologist, but uh, I think that having you know, uh You know sp Like like I don't know s m t are Yeah, like having loads of initial clusters, but only one glide it doesn't bother me. I mean, I maybe should Oh, yeah, so so I have, uh Self-consciously not said anything about Morphological structure, you know, because I think shesheng series as evidence don't point to that So I think that yeah, you can you can understand that that actually all pre-initials. That's what backsters Sorry, that's what cigars 99 the nine theory was that all pre-initials are morphological affix all roots have a kind of c optional r vowel c But it's clear that in Certain cases actually I would point to elbow as an example, which is now they reconstruct elbow is something like the crew and I think there's very good reasons for Reconstructing elbow that way, but they don't have any theory about the morphological function of the toe Yeah, so so I think that yeah, this is what this is exactly the sort of thing. I was explaining which is that like Uh, the the current version of the pre-initial theory passed through a phase Where where morphological structure of the root was the primary consideration And and I think that was very helpful for them, but it's not clear how much they're sticking to that as an idea at the moment Let's say this is actually gets to the the controversy about the the pre-initial s that I that I referred to sort of at the beginning where Me Too Lin in particular reconstructs an s Where Baxter and cigar reconstruct voiceless resonance so, um Baxter and cigar Sort of it's not worth rehearsing that whole discussion But I would say the problem that me Too Lin has is Explaining the evidence that they explain with Pre-initial s. I don't remember whether he has an alternative theory for that But what your point is is that like, you know, uh, they're clearly our cases and we saw some with like person and thousand Where there's variation Initiation series or you know word family so-called Uh between voiced resonance and voiceless resonance So Baxter and cigar recognize that problem and say Someone else can deal with that later maybe Probably these voiceless Resonance did start as some sort of constant cluster, but you know, they think at those at the level of at the synchronic level They're interested in reconstructing Uh, all chinese had You know, let's say n voiceless n s n and s voiceless n Yeah, I yeah, exactly. Yeah, if you like at at at You know at whatever old chinese was yeah Which is let's say In their approach as early as possible based on entirely chinese data And I think they're very good about that much more so I think kind of they go a little too far Much more so than let's say pang pang wu yun or jang jang shang fun baxter and cigar really feel like you cannot use comparative data When you're working out old chinese Um, uh comparative data to say to ben and vermice