 Here we go. Voila. This is Nate Simmons, I'm very happy to have you here. He's from France. He is a PhD at the Fuse de Saint-Farbe, right? And then post-doc at CNRS. And what else do you want to say? Oh, can I tell you the good news yet? So, he's starting in the fall at University of Oregon as a penitract professor. Very happy to be here. Thanks to everyone for coming and giving me the opportunity to talk about some of the research I've done on historical linguistics for a language called Arma. So, just what I want to accomplish is to talk about Arma. I've given introduction to talk about empirical materials and to look at some concrete things and the finality of the language. Looking at some of the consonants, the vowels, a little bit of tone and mostly about what we can reconstruct about the consonants of the language. I'd also like to talk a little bit about the morphology of the language and something that I think will be interesting about this directional inner historical perspective. So, Arma is a transomal language, a language complex, a group of dialects spoken along the Upper Mid River in Sichuan, China, by about 70,000 people. Of Chiang and Tibetan nationalities. Here in Sichuan, China, the red counties on this map are places where Arma is spoken and the general prefecture is the yellow. So, just those five counties in northwestern Sichuan. In terms of the other languages, it's part of the Burma-Chiangic branch and this is the structure of the family trees. We see Lolo Burmese on one side including Burmese and Burmese and we're looking at the non-Chiangic part of Burmese-Chiangic and within that, there's a group called Chiangic and there are probably a half dozen languages that they've grouped in with Chiang. So, we can kind of look to see through reconstructing Arma, how can we bring this into the bigger picture of its relationship with other languages in the area. So, here's just another quick map. These areas down here in the south are speaking Chiang, the southern varieties are much more generally diverse than the northern varieties and this is a area here where there's kind of an arbitrary boundary between the line where people become Tibetan at some point, going up the river and that. After a certain county line, everyone here is Tibetan and everyone here is another ethnicity but they're speaking dialects in the same language. So, just to be clear about that, Arma is kind of an amalgam of all the different names that people call themselves. Chiang is the Chinese name for Arma. So, just I'll be talking about some different varieties today and to give an idea. The more important ones I think to remember for this would be Hongxi here, Yonghe, which I've studied the most and then Ronghong here. So, we kind of use three different points of these dialects that we have just to kind of show the diversity of the number of dialects in total and then the varieties that we actually got. So, there's quite a lot of documentation to do for Arma as well. Okay, so then within Arma, how do we characterize the different dialects? I think there are three branches that are non-northern. So, northern is a real branch of Arma and there's northeastern and northwestern and then the southern group is really just the non-northern group. So, we have one southwestern group, the south eastern group and then the Longxi is kind of the out group. So, this structure will help us to determine what to reconstruct as well. If we find something that's just in these two varieties, we think Longming is an innovation. If something is found in Longxi, Manchur, and Yonghe, then there's more evidence for pushing that back further up the tree. So, there's no written tradition for the language. We don't have any texts written in other languages. To determine anything like that, I could use that for other languages. The earliest transcriptions are in Chinese with maybe 200 words in a Chinese word list using Chinese to write the language. There are some transcriptions by Hobson, who's a British naturalist, and an 18. I'm going to do about a 200 word word list, but the variety is not very informative for historical linguistics. Partially because of the time death, but also the variety itself is just not that interesting in that way. Okay, so we can learn about history of the language also through poems from Tibetan and Chinese. Chinese transliterations and local names. Some are my religious texts, which maybe potentially are. So, my goal here is not to give definitive reconstruction of the language, but to kind of give the tools necessary that would be used to do the reading. So, these are kind of the resources we have. But we also have jargonic languages, which are crucial because they're very conservative, including Tangut, which is a test from the 11th century. We have the Burmese side of that family tree as a long written tradition, and some of the other languages were written in Burma County. There's also some reconstructions which we can use as well. Burma is a chronologically progressive subject. So, we lose all the consonant quotas, and we lose many of their complex onset. So, just to give an example, here's the definitive word for H, which is Burga'ab, I guess, Burga'ab. And versus Taupeng, Burma'che. So, it's a real reduction that we've lost the D is gone, the Burga'ab part has become a cha. There's a lot of change. These are not trivial changes. It's really huge restructuring of the basic syllable shapes, especially for diastronomic forms that we see. So, this is a typical inventory for an amount of variety. I don't think this is representing any one variety, but this is what we typically find in terms of constants. You'll find a three-way distinction between buh and buh, for example, for all the onset except for buh. You'll find a three-way distinction in attributes. We have something like 7-trap and 7-trap. My thinking is that, maximally, you would have three series of attributes at a quarter of a language. There's a large set of reproduces. I think that most of these can be reconstructed. The X here is probably just certain varieties that we can change in certain positions, and so it's not really fun even. Most everything is pretty straightforward for the onset. So, we can look at, for example, we reconstruct a simple, voice-less onset based on nothing to the contrary. There are certain cases where buh will ignite in a metallic position and then buh will ignite as well. So, we kind of see it's really clear for things like the onset, the buh. We also have evidence for this voicing series in Chinese. So, here's some words with voice-onsets in our mind that come from Chinese. Chinese then lost the voicing so we know these are kind of clear by Chinese. So, this gives evidence for this kind of voicing distinction in the earlier studies. Although there are some one-off changes, especially for wong shi. So, we see a loss of distinction between retroflex and non-retroflex. So, wong shi here is really not giving us as much information. Wong shi is generally very conservative and trustworthy, but in this case it's kind of needed additional information to figure out how to read this separately. So, some formal varieties have these waver-less onset, which would form another set of consonants. So, you have this voice set of consonants, and now all the waver-less consonants as well. So, potentially those should be read-instructed based on correspondences with tanghuk here, with other jaronic languages that move and potentially change. I'm not as sure about all of these correspondences, but it seems like there's a big realize that corresponds to other languages that should be read-instructed also. We see a big reduction in the complexity of complex onset in our minds. So, for mao, you have something like shi and yong, heo, piang, shpa, pa, po, re, bu, po, and so on. So, losing the first of two consonants is very common in the southern varieties. Also, we find varieties where there's just a change from a complex onset into an aphrodite. So, from xe to tse, and there are many changes like this. Xe to xe. So, that's another kind of simplification. Not losing the consonant, but changing the complex onset into a simple aphrodite. Yeah, so these are, there's another very wrong one. This is the sound correspondence described in the wrong one. We don't have a complete deletion, but we have a deletion of these. So, we have xe, shi, shi. So, two words like this, but we have a weakening of the consonant. So, in these, basically, we just reconstruct the non-weakening of the consonant. All right. So, we also look outside of Hermann and look at the gyromic languages and see that these consonant clusters are conservative. It's not that they've been innovated in talking to Hermann. So, in Nantru, in talking to Hermann varieties, they have this difference, but the conservative gyromic language looks more like talking. So, we could reconstruct something based on this to say that these are really old consonant clusters in these cases. There's a list of kind of problems in earlier work on protogramma where Jonathan Evans went through and then tried to do a reconstruction. And he did the best he could with the data he had at the time, but there's certain reconstructions that he gives, which lead to multiple correspondences in the daughter language. So, this is what we don't want, because we don't know when is snub becoming nub and when is it becoming d in Nantru, for example, or these three different ones. So, this I just list here as something somebody should follow up on look at these consonants to see what's going on with them. So, yeah, especially the ones with S as an initial I would say are particularly problematic. Although, there are certain, so here's Longshi, for example, in Jaffa we see all of these have primitive plus nub clusters, like snub and snub. And in Longshi we see a d, so that's okay with them. With shnez, the word for a seven, we get a shi. So, there's a real problem with Longshi snub clusters, there's something to be figured out there as a problem that we saw. For the vowels, the vowel systems are pretty similar in cross-variety. You see a high and this high rounded, high-varated vowel, a bar die, and two mid vowels, ar shua and a, it's really common for Longshi. For Ronghong we see a difference here where we added an additional vowel and we lost the bar die, so the vowel responses don't cause as much problems, although there's this chain shift that happens in the history of the language, such that you have the same but the correspondences have shifted. So in long shi, you get e, but in long hong, you get e. And in long shi, you would get o, and in long hong, you would get u. So there's this huge change and shift that happens, that we have to kind of unwind as we do these comparisons. Looking at tone, another important feature of the language, most strategies that are two-way distinctions. Not like Chinese, we have multiple tones, but just two simple tones, high versus low. Most varieties have this distinction. Some don't have this distinction. So I would really expect this. I'd say, okay, maybe we don't call it high and low, but it's type 1, type 2, or 80 or something. And here, there's some non-tomal varieties that do have specialized vowels. So they have uvularized and non-uvularized vowels. I think that there's a correlation between the vowels and the tones. They really come from the same source, or the distinction comes from the same source. So just looking at these examples here, we have a high tone in long shi. And we have a plain vowel in hong, which is a variety when it comes to the vowels. So the high tones correspond to plain vowels, and the low tones correspond to these uvularized vowels. So it's kind of a dark vowel quality, something like ho, or something like that, I suppose. So there's a correlation between the vocalic type and the tone type. And we only need to re-instruct one. So there's some sort of super-segmental distinction at the earlier stage. So here's a minimal error. We have so, and so versus so, and so, something like that, or there's a couple minimal errors for this. The long shi varieties with the tone are actually more conservative because they preserve this distinction in all cases where in hong yin period you have two concepts with uvulars, and they can't not have uvularized vowels. They have to have uvularized vowels. So this shows that long shi is actually giving us more value here. There's more information on the side of long shi. And there's some varieties which don't have either the vowel distinction that I talked about or the tone of distinctions. They just have the same vowels here. Su, Laird, and cannabis. Su are the same. They're just homophones, like, yes, with the same vowel and like. All right, so now we'll look at a little bit of the disolator form. I'm just going to show you how the bigger picture of how words are structured and how this has changed over the language. I think this is something people have not written about before, and I find it interesting. So long shi, again, is our conservative kind of shiny noise, and it's more tonal than in other varieties. So in long shi you can have a low tone and a high tone. So low high, high low, high high, or low low. So all four combinations for the two tonal categories. So when we compare this with another variety, we find that the tonal correspondence are regular. So we have moussi and moussi, low high, low high, and then khutah, khutah, high low, and high low. But in Yonghe, the high highs have become high low. So we get khutah for kick, and then the bobo's have become low high. So wasa for monkey. So if you take a four-way distinction, you boil it down to a two-way distinction. For Yonghe, there's really only high low versus low high difference, and moussi has this more conservative four-way contrast. Another dialect to be ensured has a three-way contrast. And we can look at this and see that it has the same patterns for Yonghe and in the low highs, low high, high low, and low high. But kick patterns with moussi have the low low forms, and it adds separately from either of these in changing high high to low high versus high low high. So we have a split here, where we would have a four-way distinction, preserved in moussi, you have a merger here in moussi, and you have a merger in Yonghe. So if we took a Yonghe word, we wouldn't really know to reconstruct this as high low or high high because it could be either. We would really need to know either a combination of the value of the word in Yonghe and the ensured or the value in moussi to reconstruct the re-instruction. There's other varieties, so lobojai, I'll go through each of these individuals, but you can see how these are variations. So lobojai is the same as moushe, except here it subsequently changed the high low into a high high. There's additional twists and helping is more complex. So there's various ways where you can go from this four-way distinction to more simpler terms, and it works that way rather than to try to go the other way around. Okay, so moussi is a H1L variety, it's spoken in the southwestern region of the county, and it lacks all sorts of distinctions with so-called, so how did it get to be that way? We can look at moussi as our four-way distinction, helping as an intermediate stage, and moussi as the final product, where we have lojai becomes lojai, that's fine, and it becomes stretched on the second syllable, so this is the accent here, so we must sub, must sub, I guess, bashe, so we have lojai becoming a weak strong stress pattern, we have the high low becoming a weak strong stress pattern as well, which is bizarre, but in helping, we can low low, and then the low lows differentiated the low bits, so the first low is lower than the second low, so you have 3, 1, 3, 3, and this got reinterpreted as weak and then strong, so then you have weak strong syllable structure here. The high highs became high low in helping, and they continued to evolve just into a collapsed monosyllable in moussi, where you get this nice crunchy word, this quote for September from two syllables collapsed, and then the low low is this week, so in moussi you really only have a two way difference between this kind of collapsed syllable type and the non-collapsed syllable type, so there's not as much information in moussi to try to figure out what the origin system is, so it looks like this, we think we have just a two way. Okay, so then in the north west, the mall varieties they've done a similar thing where they've collapsed monosyllables, but that's all into monosyllables, and here's the example of that, so we have this high low distinction, it becomes just a monosyllable. So all of these went from high high to high low to then just one syllable. That's how you lose tones to change the contrast on to that of the syllable structure. For the forms which have low initial, so low high and low low, in mall they just get a disyllable with no real pitch pattern to speak of, so it's not that these are in total, but they just don't contrast with these anymore because they're separate syllable structures. The next one. Okay, so we can schematize this and say we have high low, it becomes monosyllable, low high becomes kind of a disyllable, and you can think of this one change where this change kind of results in these being not contrasted with anything else. We have these list of changes, I won't go through each of these, but this is just to kind of get improved each one of those notes on the tree that I showed earlier is validated by one of these innovations. So, except for long chi, which I can't find any total changes for, I based that tree on these correspondence, so management of wu wu jai, they share innovation, wu wu jai has an additional change, wu shi has an additional change on top of helping. Here, for example, so we can kind of see the divergence from long chi. Alright, looking at codas then from the tone it's not clear that we can really reconstruct any codas for codular amount. Maybe it needs a lot more work, but here we can see that conservative line is like Tibetan has hung codas here rung, for example, or oversell, wu wu jai says that these are cognates, these hung forms with pole in japu and this would be cognate with pole in long chi in these forms and wu wu rung, so we have that change from pole to wu but I think these are parallel innovations for japu wu wu jai, I don't think this is a common change that they've lost these codas to all of them together, because for japu wu wu wu jai still had the codas, so probably parallel innovations for which makes it better. But yeah, we can look to more conservative line that is defined where the codas drop and what the zodiac has been. All right, so now I have the rest of this to talk a little bit about the non-phenology portion of which can be more exciting. There's these prefixes in our how to go with verbs and they have to you have to say them. So this is a typical inventory of these prefixes where there's and you have up and down towards and away from a center, in and out of something and then up river and down river. So we'll kind of talk a little bit more about this and how they work. Here's an example from the Yonka Variety of Vermont, just a spoken narrative that shows these direction prefixes and how they're used. So we have this first sentence So here it's down river. So I went downstream, there's a lot of this is a literal direction because he could have been going from somewhere else or find another way. Fuku tatsuna I met a ghost, so I bumped into, I met up with a ghost where this is acting as a protective marker, not really any literal direction. And then we have do lundas so because I got scared so this would be away from center prefixes, which is functioning as a protective but has its semantics from this away, he got scared away or he scared something out of me. There's a lot of kind of English flow fictions, they're like similar things we can do with racial verbs that look like he's listening up, quiet down and so on. But in this language it's a way to correct this in the protective. So I'll look at one of these pairs direction prefixes and see what we can say about it up to proto or moth. So there's this sort of prefix which has been called a lot of different things this is located towards speaker, clockwise right word for the reverse of these leaf labels are different, but I believe they all have a similar core semantics. The dot one is out from the center like this and the other one is in towards the center. So let's look at some examples of these prefixes and there are fixed coefficients with verbs that we can get a sense of these semantics. So these are just verbs in most of your mouth which take the dot away prefix. So we have make noise to unfold release animals, dissolve land, so many things that happen in this direction generally. And we have things which are fixed coefficients but don't really have to do with space. So to have a bitter taste in your mouth you can imagine a spatial meaning for that other than it's just a perfect form of the verb. To be crazy maybe you went away crazy for something, there's something there here, okay. So for each of these it's kind of tricky to look and see if you can convince yourself if there's an etymological reason for these prefixes. Here's another variety. We have shoots catch on fire, flee, blend expand to die something an earthquake to scatter so all of this a way kind of stands I think pretty clearly. Some cases where there's not really so clear semantics to shrink I think is the opposite of expansively to get something to buy something. I don't see those as having the same semantics and we'll see a little bit more about that in a minute but there's some kind of semantics to do with crossing as well, traversing going across some sort of boundary. Okay, just another one of these varieties, just to show that the vowels of this dialect have changed a lot depending on its environment. So we have we look at longxi it's all odds and in this one it's a bit of a mess. So long long saving kind of a way outwards semantics to die, to spread, to release an exit and the opposite preface the fourth one, to cleanse up your fists, to fold an umbrella put on clothes to buy something, all things that have action towards the center. So it's the z, z, preface which we talked about the z in longxi comes from both z and z, z and zh, so we would have to see what's the reason for it is based on other evidence. In rong homo we have zu and z so zu su to sail someone from old directions zh, zhi to get hit now from any direction I guess to the buy towards the center. So there is less of these towards ones in the narrative it's interesting also probably more conducive to a grammaticalized, perfected meaning of some chemical way then something like this or it could be just an exit So here is the schema for that one on the right should have arrows pointing inwards. But you can imagine a right handed person kind of sewing seeds, and this would be away from self, clockwise, and rightward. And if somebody was collecting grain, maybe they would be doing something in the opposite direction. So this is what I think are the four semantics to the extent that that's possible for these two preaches. And in opposite order, it's this one to the z and this one to the da, away. And I think that these form an opposition in a language that's early to grow over, because you find these constructions like to do something back and forth, or do something this way and that, right? That can change. The word for go, for example, might be da, da, da, da, to go this way and that. So I think that these form kind of a nice pair. In different subgroups of Ramah. So in one, one, two, not one. Subgroup. To compare this with orientational preaches in another language, there is this neutral prefix in trustus, which I think is cocky with this da. So I'll try to show that. And this here in trustus has a neutral semantics, which kind of fits with the idea that we have some negations. Away. And that semantics which we should become the general vector form. So this says neutral. It doesn't have any real spatial input in trustus. But we need to look at how long the prefix is that the prefix, the specific prefix verb parents. So if we look at the verb to appear in all the varieties of Ramah, except Pusy we have the away form. So Pusy uses the upwards prefix. Got scared up by something. Something that he got shook up. But I think this is innovative. And then this is insert. So I would say here should be a thoughtful kind of reconstruction. Okay, same for the word for flea. So flea takes the translocative in all these varieties, except for Pusy where it's cocky outwards. So I'm going to have to think that Pusy is kind of on its own here. We have the verb like hits, which has all that evidence pointing towards the dot. So I'm going to just reconstruct the dot a lot, I guess. This one is interesting. The word for forget has the translocative in most. But then in some varieties, it's been replaced with the inwards marker. And Yoko has been replaced with the downwards marker. So I would reconstruct the translocative as basic and same. These are innovations for these. These got rematched with another orientation. And you can kind of see these. The so-called structure thing we talked about earlier. It's a double, two nice so-called here. And a dire, or just done. So there's quite a difference from the so-called. And then this word for tired. This is less clear. I think we have conflicting evidence, upwards, downwards, and translocative. I'm not as sure about. But kind of needs more work. Each word would have to be done this way. Look at the cross for rights. Here's the verb to buy, which in most varieties takes the forwards. So I think that is original. And then it got replaced with upwards, or with away in the entry. Where away in the entry is more of a general perspective than having a real translocative. Okay, last. Then we have this verb to die, which I think is really complicated. Here in Ochi, it's got a downwards prefix. In this Northwestern subgroup, it's got the away prefix. Here it has inwards. Inwards again. I'm not really sure what to do with this one. I'll say for now that it's translocative. That's not. Yes, it's translocative. Let's see. There's cases where we have long words from Chinese also. So this is a long word Chinese verb to be long. Has the translocative prefix in most varieties but it's been replaced in Ochi and in Ochi. Ochi did it with the downstream prefix talking with the upwards prefix. So this kind of confirms to me that Ochi is really kind of innovative. It's innovative with multiple forms and with long words as well. So I would really check this with the dot organizational prefix. Looking beyond for our minds to see if they have these prefixes with other languages, Evans has suggested that this veranda would possibly cognate the Tongans. Da. And with kumi ta away. So I like the first comparison. I don't think the second one works. The Tongan prefix da has been noted as having no directional semantics. It's a perfect marker. Or it's been called a case of a neutral directional prefix. There is a closely related language called Disha, which has a prefix da with the exact same onset and with a similar semantics of being orientational in this way. And the crucial thing I think is that there's forms with the dot prefixes that are cognates in all three groups. So if we look at these words here to forget, to die and to move. I'm not saying these are my final answers to reconstructing the program up, but if we reconstruct these verbs, having the da prefix fits really nicely with the Disha dot prefix where you get dabble, dasa, dabble, and here we have something like dabble, dasa, dabble. Yeah, I don't want to commit to these. I'm committing only to reconstructing the prefix in the slot. And then we have these verbs to become, which fits nicely with Tongvit as well. I'm not a Tongvit expert, so I'm sure there's people who need to find more cognates or maybe correct some of these words. So this means that there's validates this place. If you have a close relationship between dharma and dharma, I would say it's a way to treat this. We don't see an persuade, for example. We don't see an uya. So just to give a quick look at the inventory of things from other language groups, here's this tree structure. We have persuade from Chang'e and the ones inside it. And our sue has no real overlap with dharma. So that makes sense. It has an innovative form for dharma. The forms used are dharma, away prefix, which is Tongvit. They're cognates of Tongvit, yes, and dharma. We see dharma as a formative and I think that there's an interesting thing going on with this dharma, away prefix in our store, so that could mean to say well maybe persu, and wumi belong to some sort of way, or it's through contact. If we look at the geographical relations, this is a terrible map, really sorry to have to show it. It's got a lot of great accuracies, but it does have this nice portion of showing persu here, two different wumi varieties, and wumi, they're all in the same area. These are all persuic varieties, so I'm more tempted to think about a way prefix as being an innovation or a parallel development across those, but I think that the contribution here would just be to say that the dharma way prefix fits nicely with this other way prefix. For more examples of that. I was kind of half-stepping about Changi for a long time about those in Ratbag, but now I think there's some kind of an angle, especially for these western-studded vanguards that don't really help that much, so I think it's kind of representative of where we're at now for, yes, so north-western varieties, if you were to reconstruct Kodahari, you could use the north-western variety for their complex concepts. You would need wonshi for the vowels and the toms. You would want to look at the southern varieties for the Kodahari, and yes, some of these southern changes. And also to look at cross variety, I think it's important to look at cross varieties, looking at one variety and seeing which period it takes to the cross variety, rather than just looking at the vanguards by themselves. So, that's I think all I have to say is thanks very much.