 So, today, we continue with more about onset clusters, and I'm just going to start by, you know, reminding you kind of where we are in the overall presentation. So these are the pre-initials that Baxter and Cigar reconstruct, and they come in loose and tight forms. And last time, I discussed the kind of overall plausibility for reconstructing pre-initials, both in terms of, you know, that there isn't that much evidence, but there's some quite good evidence like these words written with two characters and so on. And then we looked at reconstructing tight pre-initials using Sheshang evidence. In particular, we only looked at the S prefix, which you will be happy to know is the one that, you know, I don't know, that comes up the most. So now, today, we will look at reconstructing tight pre-initials using Sheshang evidence, but for P, for K, and for T, and M, and capital N, and unspecified C, which of course is the same thing as P, K, or T, but when we can't tell which one it is. Okay, so now we're going to start in on that, which is we're looking at Sheshang evidence for tight pre-initial P. Here it goes. So in, you know, pointing to a change of a sort of reduction of P, K, to P in tight, sorry, in tight B syllables, is this Sheshang series? So you see we have a connection between a initial P in one word and an initial K in another word, and actually I would even, you know, go further and say the initial P word is being used as the phonetic in the initial K word. So in order to make the Sheshang hypothesis work, you either need to argue that the first word kind of secretly has an initial K, which is what Baxter and Sagar have done, or you would have to argue that the second word secretly has an initial P. And I just want to emphasize that like the thing that I'm trying to get across right now is that these problems are real, right? Like how to reconstruct the old Chinese of these two characters in order to make the Sheshang hypothesis work is a real problem that you can either ignore or you can try to solve. And if you try and solve it, you could do the way Baxter and Sagar are doing, but you could try other ways, right? So also I would say when I see P and K connections, you can also think about reconstructing labial velars, but, you know, we've already reconstructed labial velars for something else and so on, which is to say I think that, you know, they want to emphasize that it's not like we should take Baxter and Sagar's reconstruction as the only solution, but I think it is a very good reconstruction in terms of telling us which things we need solutions for and offering possibilities for them. Okay, so continuing on here, we have, you know, so PQH leading to aspirated P in type A syllables, and you'll notice that these patterns are much more gappy than we saw for the S prefix, which is to say basically there are specific circumstances in which Baxter and Sagar think there are evidence for reconstructing these things, but there are probably a lot of them, if we agree with their analysis of old Chinese phonotactics, there are probably a lot of cases of these preinitials that have not been discovered. Okay, so yeah, so we have a connection between an aspirated P and voiceless velar fricative, as you know the voiceless velar fricative generally come from uvelars, so this is their solution, yeah, and then here's another one. So this is, you know, P, voiceless R changes to pH, and the sheshen connection is between this aspirated retroflex stop and pH, so there you go, and that was it for the pre-initial P using sheshen evidence, and now on to pre-initial K using sheshen evidence. So we have kla, it goes to kha in type A syllables, kha I should have said without aspiration, and here once again I just include the pre-chin form of a character so that you can more obviously see that there is a sheshen connection between these characters. So we have, yeah, we have a k initial, and then we have a typical lateral series, so that you can see, okay, yeah, we have a voice dental, and we have a voiceless dental, and we have the yuh, so it's a classic kind of lateral series, but then it has this kha initial kind of intruding, so how do we deal with that? Well, why not reconstruct a kla cluster, okay. They appear not to give any examples of the sound change in type B syllables. Okay, now kha goes to kha, and we have an l initial which we reconstruct back to r, and then a kh, so their solution is k before a voiceless r, okay, and they appear not to give any examples of the sound change in type B syllables. Now this one is a little bit, I don't know, finicky, so they think that a k before a glottal stop changed to k. I'm not quite sure, you know, I'm not a funatician, I'm not quite sure how this would end up getting realized, this k glottal stop, and maybe it's, there's no problem there, you tell me, but it seems a little strange to me, and in case they normally reconstruct a series that shows sheshen connections between vealers and glottals, as you have already seen, as deriving from inherited u-vealers, but they also permit this k before glottal as an alternative explanation of such sheshen series. So here it feels like we're getting a little bit sort of not in keeping with Occam's razor, right? If we have one solution for glottal stop veal or interchange in sheshen series, why don't we just use that one solution? What's the motivation for this second solution? Well, here's the series that they use it for, so you see the series, you know, so you have a glottal initial and a k initial, and it appears to be they prefer not to have two sheshen series that write the same type of old Chinese syllable, and this is something that they, I think they're hesitant about, so they don't kind of stayed in a very doctrinaire way, but basically as they've worked over the years with these different hypotheses and their interactions, you basically end up with very few sheshen series indicating the same syllable type, and you can imagine this, just think of something like dentals, right? You might have a series that looks like it represents something like dock, but actually some of those docks will go back to an L because L and type A syllable becomes D, and some of those docks will actually go back to a D, so if you have two sheshen series that look like they are structured round dock, maybe one is actually locked and one is actually docked, so as we kind of apply these various interacting hypotheses to the very sheshen series, you start to get a picture that basically each phonetic determiner serve to index a specific syllable type, so put it in other words that the Chinese script at one point in its history was basically a syllabary, so I think that's one reason why they invoke this hypothesis is that they end up with two sheshen series that seem to indicate the same syllable, so they say, well maybe this one was actually a K prefix to a global stop, but I don't quite know, or let's say what do I know, what do I not know, definitely Baxian cigar have been moving towards this syllabary idea, although they don't state it very categorically in their 2014 book and it is one of the arguments that they invoke in this specific case, but whether or not generally speaking that's what motivates this reconstruction I can't say, by positing a type pre-initial K before a global stop is one of the origins of middle Chinese K, in this case we can understand this series as encoding a shape like god stop UJ as opposed to this other series that is on the slide, which encodes syllables of the type Q UJ, so we're looking at the second series, the Q UJ series, in this series readings with the voiceless velar fricative and with the yuh mean it really has to be reconstructed as a Uvila, right, so yeah, so to just put the argument the other way around, this series has to be Q UJ, so then this series they think well let's look for another solution, how about god stop UJ with a K prefix, so that was it for type pre-initial K using Sheixiang evidence and now we move on to type pre-initial T using Sheixiang evidence, so here is the kra turning into tra and then I generally don't put meanings on any of this stuff because it's I don't know it's not helpful, right, I mean it might be helpful in terms of remembering the particular case but I think it sort of adds a lot of clutter, so I haven't been doing it but I will point out that the first of these means elbow and the second one means nine and both of these have very good cognates in sign or Tibetan languages and you can see that also that the character for nine in the ancient script kind of I think is a picture of an elbow if I remember correctly, so there's clearly they thought there was some kind of you know elbow and nine sound similar, so there you go there's the evidence for a T prefix using Sheixiang series, okay, and then in in this case we have two K changes to tra and here's you know some a velar and a palatal that are linked in the Sheixiang series, so they think wow maybe the velar is the initial and the T they gave rise to the palatal is a pre-initial here's basically the same thing but with an R medial so it becomes a retroflex of an aspirate retroflex stop from the kr and you see that basically it's a pretty clearly velar series but then there's this one retroflex at the end that was it for T pre-initial using Sheixiang series so you see that there's really not nearly as many cases to go through for P, K and T as there was for S but now just a comment about resonance which is that so this one slide has pre-initial type pre-initial P, T and K before resonance and in this case their argument always comes from the fact that character itself has two different readings now generally speaking if a character has two different readings it seems likely that what's going on is some kind of dialect split and then and then emerging of of readings in in let's say the reading tradition so yeah so here are the cases you see we have one character that can be Pim or Lim so that maybe means it would go back to Prim and then Chip and Nep so maybe Tanah yeah and then and then Quet and Mie so well so these differ also in tone yeah let's say you might see a complication in the coda and it and don't worry about it we'll get to it later the codas are fine okay so that's it's for tight stop pre-initials before resonance okay now how about tight nasal pre-initials M and N so so far we've only looked at P, K and T or K, P and T so on the basis of Shesha evidence the tight pre-initials M and N before obstrants are not easily distinguished from each other which is why I'm treating them together and they're also not distinguishable from simplex voice resonance nonetheless they can be isolated before uvulars and before R and M is identifiable on the basis of Shesha evidence before labial nasals so let me just say kind of what we're not going to see which is there might be things like an M before a T yes but an M before a T would have led to a D in Middle Chinese and a D is allowed to interchange with a T in a Shesha series so kind of by the very nature of a Shesha series we are not going to be able to see prefix M and N with kind of most stops in on the basis of Shesha evidence that's the the point that I'm trying to make here which is to say I'm only going to show certain kinds of examples but those are the only examples we would expect to see yeah okay so M and capital N before some uvulars okay so we have a classic uvular series because we've got a velar and we've got a yaw and we've got a glottal oh but look we've also got a velar nasal how do we explain the velar nasal well it's the first thing on the slide and you see that basically it could be a lot of things but all of them would be some kind of nasal prefix followed by some kind of uvular and I've written out all the options to just make this point that like Baxter and Cigar will reconstruct this word a particular way because they will have other ideas about etymological relationships and and loans into other languages but from Shesha series alone in their system you can say that it's some kind of nasal prefix followed by a uvular we can also see evidence of the nasal prefix in their system before R actually in old Chinese leading to a yaw in middle Chinese and I am going to give you this example which is two readings of the same character so one with a yaw and one with a luh yeah in middle Chinese and then just remember that middle Chinese L goes back to R and middle Chinese yaw well can go back to various things including L and uvulars but in this case they say well we want some kind of explanation for maybe how an R changed into a yaw and so their explanation is the nasal prefix before an R changed into a lateral changed into yaw connections between yaw and la in middle Chinese they see as evidence for a no or a ma prefix before a R in old Chinese yeah and you know it's another case where you have two readings of the same character which they take as evidence for dialect developments and here's what they propose is that in one dialect MR develops similarly to L which is what you already saw in the example I discussed so here's an example of this ying from and this mring from and then in another dialect MR merges with M but it seems to me that reconstructing let's say MR with the medial R and M as a pre-initial before R already indexes these different outcomes so I'm not really sure you know why they need to propose different dialect developments although it kind of this gets back to the question of how do we think MR and M dot R were pronounced differently so let's say you know if that's not different right if the difference between a pre-initial and an initial is neutralized before R because R conserves the glide well then maybe you do need the dialect difference so I just want to sort of flag that problem and think it through so now an M pre-initial in a predominantly nasal series from another point of articulation so just talking you through this example the the sheshing series actually sees a connection between nia nga and ma but nia is a very normal type B development from nga so we don't worry about it so basically it seems like the series overall has a nga initial but then there's this ma example and they reconstruct that with an M prefix now here's a good case for me to to just give you a sense of you know a possible alternative approaches so Chris Beckwith who unfortunately hasn't you know there's no sort of Beckwith system he just sort of comes by and says a few ideas here and there he thinks that series like this should all be reconstructed with an M initial and that under certain phonetic conditions M changed to to to another nasal in this case nga or in other cases nga and he you know tries to formulate some kind of sound change that that would account for that each of these surprising sheshing connections could in principle be argued the other way okay and then here's here's an example with the nga initial so we have a clear nga series but then there's a ma initial that's sort of interfering in it so they reconstruct that's Baxuan cigar a tight M pre-initial before the nga so now turning to tight pre-initials before you viewers as a source for middle chinese viewers i may have misstated this in another lecture and taking them back to loose pre-initials which is an idea that Baxuan cigar toyed with before the 2014 book i think in 2012 but in a case here goes here we have a classic uvular series right there's a glottal initial and a yaw and a velar well now when i've previously presented these things i've sort of said oh don't pay attention to the velars the velars are a reason that we're reconstructing a uvular initial in this series but let's not worry about where they come from well what Baxuan cigar do is they propose that there's some kind of tight pre-initial stop right because we've already seen that a ma and nga would have led to a voiced voiced velar nasal onset right uh so instead we're getting a velar stop so either this would be a top prefix apocry prefix or a cup prefix in their formulation uh and that's how they reconstruct velars that interrupt uh uvular series and then similarly if there's a medial r and similarly if there's aspiration i'll just say that it for this one don't let it confuse you that there's this uh capital n before a q that's based on evidence other than the shesheng evidence for our purposes at the moment you could imagine that as a capital g okay and then uh you know just another example yeah so so uh the velar uh let's say velar stop velar fricative connections they reconstruct as uvulars and we've already discussed that that the aspirate uvular is the source of the voiceless velar fricative well here uh we're we're pointing out that the that the tight pre-initial before a uvular is the origin of the velar reading in the series okay and now with uh uh with labio uvulars the voiceless velar fricative with a rounded vowel in middle chinese is going back to a uh an aspirate labio uvular stop in all chinese in order to explain the velar stop in the middle chinese uh we reconstruct this uh tight pre-initial c i should have stated it more precisely it's clear that generally speaking uh manner doesn't matter right so so there's not going to be a separate phonetic for dak and a separate phonetic for tak the shesheng hypothesis in the form i'm using it which is the form that baxter and cigar are using it comes from li fang kuei and says four two things to be in the same shesheng series they have to have the let's say the same nuclear vowel the same kota consonant or no kota consonant uh and a homo organic initial so that means a medial r can come and go who cares it also means that that let's say the tone if you like but for us that means a final s or a final global stop can come and go so so you could have you know bak and prak written the same way if you like and you could also have you know paks uh written the same way now it tends to be that this doesn't happen for nasals which is to say that mak would be written with a different uh phonetic uh although uh that's not kind of categorically the case uh and then i would also say in bigger series there are subseries that seem to be aiming at finer phonetic differences so um so so one thing that happens in a big series is you'll get a subseries for type a and a subseries for type b um and uh this is so modus listen i wrote a paper about kind of looking for this kind of phonetics uh like what what what are the phonetic details you get in subseries and that was the only one that really popped out but i i think there's more to be done there the more we work on this then then the less we will be able to hold fast to the sheshaing hypothesis as uh leafon quay formulated because what you'll end up getting is like in in in many cases fine phonetic detail will be in the in the sheshaing series like i mean you also get for example you also get um some sheshaing series where every character has the same tone you know where it's all uh pox and and uh box and uh prax with the s on it but i think that everything works according to plan in a sense right you you come up with a very doctrinaire sheshaing series hypothesis it it proves very productive in terms of figuring out what's actually going on and then you can kind of let it go when when it's clear what is happening which in some cases is even more strict than the sheshaing hypothesis and in some cases is even more loose and that seems to have to do with how many syllables of that type are in the language right which is to say like um gosh i'm i'm sort of tempted to make an analogy with poetry right which is like um uh if you're writing a poem in english and uh then you will you know you'll find it very easy to rhyme love with dove and above and whatnot but then you'll find it very hard to rhyme things with orange um so so similarly like when they had a lot of sort of uh words or morphemes let's call them that had a similar phonetic structure broadly similar phonetic structure then the sheshaing uh the the use of the the phonetic principle in the script allowed for more refined pronunciations whereas if it were a more rare syllable type they they had to kind of it was economical in a sense to use one uh phonetic for you know even for different nuclear vowels for example if we want to believe this syllabary story it would go something like this in the very early days you have some other story to tell and it's the kind of the story of how people invent scripts you you kind of haven't quite worked out any principles yet and and you start with pictures of concrete stuff you know you're like ah it's a dog you know and and it's a cat uh and and then now you have the phonetic repertoire that you can start to build other things with anyhow that's the story in the early period and at that stage every uh grapheme let's call it has a phonetic and a semantic meaning and and i don't want to say that necessarily graphically each character has uh oh this is the phonetic and this is the semantic no no no the regardless of the graphic structure in the early days all signs were to use kind of boltz's uh terminology plus p that's had a phonetic meaning and plus s had a semantic meaning uh well eventually as you get to the kind of let's say late warring states period it seems to be the case that um the phonetic principle is predominating which is to say that the that there are fewer cases of phonetic ambiguity and more cases of semantic ambiguity and that's where you're sort of heading towards a real syllabary yeah uh but then it seems like uh i mean this is a kind of a just so story i'm sort of telling you a parable right how close it is to reality is something that we need to do a lot more investigation on but then it seems like that process uh of let's say de-semanticization of the script was kind of uh became a political question yeah it became seen as some kind of decadence and uh and you know ah the kids today they don't know how to spell anything yeah so um so around the chin period which is very short right when there was this effort to standardize things although that's overstated it's pretty clear now that even the standardization of weights and measures was not as successful as he's sometimes talked about but there clearly was some sort of political project of standardization uh um that let's say project from the chin even whether or not was successful in the chin sort of said no no we need to try like a chinese character should have a phonetic and a semantic component and then um eventually the kind of variation that you see in the warring states period and this tendency towards uh uh uh uh uh syllabary is sort of reversed and you and you get kind of certainly by the the early han period and probably yeah well let's say certain about the early han period you you get this you know the script as we know it today where like you are not allowed to play around with what goes on inside a chinese character boltz is the person to look for on this stuff and what i would suggest actually which i guess i'll send out is his article in the brillian cyclopedia about the origin of the chinese script i'm not able to answer that which is say like i've just been trying to draw attention to that as a question but i would say that my impression is that baxter and cigar distinguish them so uh which is to say i mean we would really have to look at the book very carefully but my impression is that like uh i mean i'm not let i don't know about a given case but let's say p dot r and pr my impression is that that those would have different outcomes in middle chinese where you know what pr would be you know just p followed by a um second division yeah and p dot r would be something like uh l initial i've never had occasion to ask baxter or cigar exactly what they have in mind phonetically with this pr versus p dot r i would think it would be very nice if the pre-initial initial distinction was was uh merged or let's let's you know what trebet squire would say is you have some kind of let's say archie phoneme in that spot um or i guess archie phonetic position uh but but anyhow that that then they should they should you know if there was not a phonetic distinction then they need to develop the same way and vice versa if they if they don't develop the same way then there must have benefited distinction yeah um and i think that this is a kind of um uh like i don't want to put it too strongly but let's say a rhetorical weak point in baxter and cigar system which is not to say that they've done anything wrong or that they're wrong about it but i just feel like it could have been uh you know that that that if you're a neogramarian like me and you read the book you find yourself worrying about this problem but yeah uh although i will point out that you know uh none of the reviewers worried about it um and maybe that's because very few people operating in chinese historical linguistics are neogramarian yeah i don't have a strong feeling about that but what i would say is it points to something that i think needs uh work that baxter and cigar have only kind of scratched the surface of which they've done in another context which is final r where whenever you have one of these dialect very you know you know dialect split proposals you you would really like to figure out um um you know which dialect right like like like um let put another way uh dialect mixture is the kind of uh an easy uh rabbit to pull out of your hat like yeah you know every time your historical phonology gets hard you say ah it was dialect mixture now i think that multiple readings of the same character with the same reading is a priority very good evidence of dialect mixture but still you you want to say like where and when did the dialect split where and when did the dialects merge and it it may well be that in a given circumstance you can't figure that out there just isn't that evidence but uh you would want to see someone go through the effort right and as i said in the case which which i haven't talked about at all but in the case of final r where there's a dialect difference where in the west it changes to n and in the east it changes to yeah uh baxter and cigar have have have great evidence of oh it happened you know in this place at this time so i would just really like to see in all of these other cases including this one that some effort is put in to isolate you know where and when this dialect split happened and and then it will feel more like a genuine explanation and not a um and not pulling a rabbit out of a hat now that's just sort of slightly different there's answering a slightly different question than the one you asked which is you know can like if we have to invoke dialect variation maybe dialect variation is enough and we don't have to distinguish between uh these different phonotactic types of pre-initials and i think that's an excellent point yeah we we we should only have as much machinery as we need and not anymore