 I've entitled this thing a preliminary mix of messy Tibetan verb paradigms. And I mean quite literally a mixology because I think that the messy ones, that the messy Tibetan paradigms really are mixed. And I have a number of sources to show you as to where or how that mixing happens. At least, you know, in my view, I'll have to say that not all of the ideas I will present today are equally nicely put together. Some of them are, others really aren't and you will be able to see it. So this is my introduction. The verbs in Tibetan have four stems. They're traditionally called present, past, future and imperative. Each of these stems is marked by a semi-regular combination of prefixes, suffixes and in the case of the imperative and of our alternation, of our outlaw from out to all, if, or from yet from out to all, if applicable. So it's not, so it happens, it doesn't always happen. And this is what the paradigms can look like. Now, what I mean by semi-regular is that, A, it's not entirely regular. You can see that this is by Nathan's reckoning in 2010 that there are at least 12 patterns. So they're not entirely regular, but they're semi-regular because if you look at, if you look at what these, what these different paradigm variants look like, they all look similar enough, right? So in all cases with the present stem, you have a prefix that is essentially some kind of velar. This Archeong letter is variously reconstructed, but Nathan's view is that it is a velar fricative. But you have a velar prefix for the present stem. You have a labial prefix for the past stem. You have a variously a labial or a dental for the future stem. And then nothing by means of prefixes, but an S or historically a D for the imperative. So it's regular, but it's regular in that sense, but you can't really clearly predict, given a verb root, which paradigm it's going to take. Tibetan allows a lot of consonant segments before the initial, so before the main onset. So these are called pre-initials. Tibetan is not the only language that has pre-initials, but it has a lot, it allows a lot, but it doesn't allow an infinite number. So sometimes you get these paradigm, a paradigmatic prefixes attached to the root and sometimes it doesn't work. Sometimes phone attack constraints limit the number of pre-initials that surface. And sometimes historically they, they condition certain sound changes. So what you see is not always what you get, but this is the idea, this is the general semi-regularness of Tibetan verb paradigms. Today, these paradigmatic forms are really only preserved in written Tibetan. Most of today's Tibetan anyway, have basically lost all of the pre-initial segmental material. A lot of them develop tones, some of them became pre-nasalization, pre-aspiration, that sort of thing, but by and large you don't get this rich segmental material anymore in modern Tibetan languages. So really we are relying on dictionaries. And obviously with lexical graphical data, you always have issues of prescriptive bias and unclear sources and so on and so forth. So there are views that say, well, it's just futile, working with dictionary data on verbs. I don't think that's the case. There are principled ways to deal with the messy, the slight messiness that you see across different lexical graphical sources. Here are a couple of examples, right? So today's talk basically speaks to this idea that you can actually find some regularity in the messiness in dictionary data on Tibetan verbs. So let me just show you the patterns, deviant pattern number one. This is a vowel alternation in the present stem. This is a set of ablot patterns. One of them is that if you have an a in the root of the verb, like to kill, sat, it's not sad, even the spell that way, you get ge-sod in the present stem. The vast majority cases of this ablot is associated with a particular present stem prefix, the ge prefix, although there are exceptions. That's one pattern. The other pattern is a fronting pattern. So ab becomes e, u becomes e. Now I'm gonna show you examples below because this one co-occurs with deviant pattern number two. So this is the one that is a little tricky to explain. In this pattern, you have a single verb with a paradigm in which different stems may have different voicing. These are examples. You have verbs like cost of fall, right, bub, bub, right? So you see that in the present stem, you have a voiced initial for the root of the verb and that's the case for the future stem as well. But in the past stem, you get a voiceless initial and also similarly for the imperative. We can kind of schematically think of it as voiced in the present, voiceless in the past, voiced in the future, voiceless in the imperative. Same thing here, right? So bing, biong, biong, biong. One more thing to note is that aspiration is no longer contrastive in older and classical Tibetan. So if you have a voiceless initial that does not have a pre-initial before it, right, then it becomes aspirated by a phonological rule. So basically Tibetan only has two laryngeal categories, right, basically voiced and voiceless. That's all that Tibetan has. So in these sets of examples, you also see this vowel alternation from the even pattern number one. So you get a, which is the root vowel, but that uplots to a in the present stem, right? If you, and then if you get o, then it that uplots to e in the present stem. People have said that this is conditioned by the or the d suffix in the present stem. I'm not gonna engage with that idea. I'll take that as pretty reasonable, even though it's not watertight. Okay, so let's look a little deeper at this even pattern number two. When you get a voice alternating paradigm, when you have one paradigm that involves this pattern of voiced in the present in the future, voiceless in the past and imperative, you get this cluster of properties. A, they're always transitive. B, they display this uplot in the present stem if applicable. And third, and this is the important thing that will come back today later. These voice alternating paradigms have associated or related paradigms which are built on one or the other or both of the voice alternating roots and taken together these related paradigms co-alternate in transitivity. So this is what I mean. So look at this second line here, right? So bub, bub, cost of all, this is the one that you saw. This is the one that's familiar. It's transitive, it's voice alternating across the paradigm. Now, it's what happens that there is one other verb, paradigm with a root that is only voiced, that is only bub, right? And this one means fall in transitive. So to fall to fall down, right? So you get a bub, bubs, a bub, bubs, bubs. There's a third one, which in this case happens to mean to come in an honorific sense. It's the honorific verb to come. Historically, I reconstructed anyway to a meaning that means to descend. Now, notice that this is intransitive again, right? To descend is intransitive. So you get voice intransitive, alternating transitive and voiceless intransitive. So this is kind of your triplet, a triplet of related Tibetan verb roots slash paradigms that alternate in transitivity. It's intransitive, transitive, transitive, generally. This is a pattern that is messy because why, right? Why does it happen to take this form? Why do these things come in triplets? We will deal with different aspects of this for a lot of today. But let me just also observe that it's not in all cases that you actually have all three attested. So this case, right? So the remove drive out case, right? Is one where you have the voiced intransitive, right? So you have the biung, that means to come out to emerge intransitive. You have the biung, biung, biung, which is alternating transitive. That's your second part. You don't have the third one. It's not attested, but it doesn't mean that it cannot exist. It could have existed, right? So this is one of Nathan's papers from 2014 where he basically collects a whole bunch of these triplets or defective triplets. So you have these examples, which are not defective, which are full. They have all three. You have these ones, which are defective in the last one. You have this set, which are defective in the first one. There's a lot of questions we can ask about this, but let's just say that the current impression is that this root alternation, this one, this BP alternation, there's an impression that it correlates somehow with transitivity in transitive-transitive. There's also an impression, which is I think correct that it's old. It's very old, older than old Tibetan. But that's kind of about as much as we know explicitly about it. Alternations that correlate transitivity and initial laryngeal features is widespread in Sino-Tibetans, but we have to say that not all of them are historically cognate. And it's also not always the same laryngeal features involved, right? So in Tibetan, it is voicing, right? So voiceless in Burmese, it's aspiration. So even though this is a common feature in the family, we don't know enough to say that all of them are historically connected in this way or that. So let's focus just on Tibetan and Buddhist languages. So that's kind of the status quo of deviant pattern number two, right? Good, okay, so deviant pattern number three, this one is easier to explain. So this is a set of verbs, which have immutable G prefixesion across stems. So if you remember our introduction, right, we would expect a verb in a Tibetan paradigm to display a velar prefix for the present, a labial prefix for the past, labial or velar or dental for the future, variously, and nothing, usually nothing for the imperative. Now, that's not true for a certain set of verbs. These are some examples. So kaduk, right? You get kaduk, kaduks, not paduks, right? You get kaduk and kaduks. It's g all around. Kaduk is the same, kaduk, kaduks, kaduk, kaduks. Now, kadal is interesting because in at least some sources, you get kadal, badal. So this one is expected, badal, which is also expected. And then kadal, which is not expected. So what is going on, right? Up to my 2022 paper, the usual treatment is kind of pretty tongue-in-cheek. So basically people say, well, G is part of the verb root, not a prefix, that's correct, right? Because it doesn't alternate. You can, you treat it as part of the verb root, right? So that's correct. But we don't know where the prefixes went, right? So let's just not talk about them. So if you look at earlier studies of Tibetan verbal paradigm morphology, like this important one by Hablin, 76, this group of verbs is actually very vastly undersided. They don't cite a lot of them, even though they existed. So I don't know what went on there, but it felt like this was being swept under the rug as somehow not very regular. Let's just also observe that with this group of verbs with immutable G prefixation, right? They are also all transitive. They do not display the R to O up-laut in the present stem, right? So if this G were the usual present stem G, then oftentimes you would expect this R to up-laut to O, right? Like the Gesalt example, but this is an example where it doesn't. And also you can find counterpart roots to these roots as well. And these roots have the following property. They do not have the G and they're in transitive for the most part, for the most part in transitive. So you would have a root that is duk without the G. You have a root that's dal without the G and you would have a root that's just duk without the G. And these are also for the most part in transitive. So we have another potential scenario of sort of transitivity related alternation, right? So this is deviant pattern number three. This is an example, right? So duk, this is to meet, to reach, to arrive at. This is in transitive, this lacks G, right? As opposed to cost to touch, cost to meet, compete, sue. I'm showing you all of the examples I just talked about. Okay, so these are the three deviant patterns. Now I'm not gonna talk about the first one because I don't have an explanation for it that's different from the ones already given, but I will talk about pattern two and pattern three. And I'm gonna do it in reverse order. So I'm gonna talk first about pattern three and then about two different aspects of pattern two. So pattern three in 2022, I said that the G, even though it in Tibetan is already fossilized as part of the verb root, historically is a derivational prefix. I identify the semantic value or function of this derivational prefix as agentive transitivization. I'll talk about exactly what it means, but it has a certain kind of transitivizing function, but it's not your typical causative. I also show that the agentive transitive G is probably at least as old as Bodish. So this is what the Bodish subgroup of Sino-Tibetan looks like according to the current working hypothesis. You have Tibetic, which includes all the Tibetan quote unquote dialects. These and some of these East Bodish languages, mostly in Bhutan and North Eastern India form a subgroup. And then this subgroup together with this group called Tamang, it's spoken in central and Northern Nepal, which is the one that I'll focus on a lot today. These together maybe plus a couple more stragglers will form your Bodish subgroup of Sino-Tibetan. So my idea is that this agentive transitive prefix is at least as old as Bodish. So after I introduce this historical point made in the paper, I'll come back and show you that identifying this G historical prefix can help explain some of the messy transitive paradigms with immutable G. Let's just go through what we find about this G in Tamangic. So Tamangic is similar typologically to languages in this region. Modern Tamangic is tonal. It has four tones. Mazodong's work, which is very convincing in this regard shows that these four tones arise from two prototones. She designates those as A and B, so I'll follow that. And this comes about due to your typical tonogenesis you lose initial voicing, you gain tones. So the idea is that if you have ba A tone in proto Tamangic, you get first tone ba in modern Tamangic. If you have ba B tone, you get second tone ba. If you have ba A tone voiced, you lose that voicing in modern Tamangic and you get a third tone. And then if you have ba with B in proto Tamangic, you get fourth tone ba. So you get one, two, three, four all voiceless in modern Tamangic. Over here with aspirates, right, you get basically A translating to one and B translating to two. So in general, you get this pattern of in total four tones and two aspiration categories in modern Tamangic. And then you have three voicing aspiration categories across two tones. So six laryngeal categories in total in proto Tamangic. I'm counting tones into laryngeal categories. Okay, so Bodish pre-initial. So Tibetan like pre-initials, right? So the ones that you see stacked before the main initial in Tibetan, you don't see that anymore in even in proto Tamangic. You don't see that in the proto Tamangic syllable system. It only allows a main initial with nothing before it. The syllable structure is just tone plus an initial plus one or two medials plus a rhyme that is either a vowel or a diphthong or a vowel plus a consonant that is but the g. So pretty typical of this region. Modern Tamangic morphology, verbal morphology are all innovative. So there's not much to see there. So what do we know even as early as 1953? So we had an inkling back then that some pre-initials in Bodish may have been lexically preserved in Tamangic. And the idea is that those pre-initials would have been preserved just because they have excluded the main initials. So back then, the examples that Schaefer gave us were the Tibetan word eight, number eight, Brgyad. In Tamangic, you get something like Brat. Okay, so you get lexical preservation of the Br which are both pre-initials to the exclusion of the main initial g and also in this case, the mediodia. Similarly, the Tibetan word hundred, Brgyad, is actually pretty regularly Bram, right? So this was Schaefer's impression and Schaefer had very little to work on beyond a couple of words, but this was his impression. Now, it turns out that these are the wrong examples because later on, we know that this g in Tibetan is actually inserted in even just an earlier stage of Tibetan, you would have something like Brgyad. The g is actually not there. So in this case, Br are actually the initials, not pre-initials, same thing with, yeah. So this is called Yifangui's law. So Schaefer happens to have gotten the wrong examples, but I think that he's got the right conclusion. Some pre-initials are indeed lexically preserved in Tamangic. So here's one. Now, I was looking at one of the Tamangic languages, Rishyam Kutama. There's a really nice dictionary done by Mazodong and I found this. So you have a verb, rep, second tone, which means to break off. So it's more like this kind of snapping, kind of breaking off, right? So break off small branches or the ends of a bundle to even out a bundle of small branches. There is another verb, grep, also second tone. That means to cut wood by hand or to prune. So if you think about it, it's pretty much the same notion, it's the same action, right? Now Mazodong really hopefully gives this note under this entry on grep, right? So she says that grep involves wood that is larger than does this rep and it implies an effort. So that's actually very helpful. So apparently, right? You have this verb doublet within Tamangic that is apparently semantically related and the relation seems to be that you have a difference in the intensity of the action and also the amount of effort. So the question is, well, can you get more, right? Well, it turns out that really with Tamangic, it's difficult because there's only about a thousand or so words for any Tamangic language recorded. Now, if we expand our search space to include Tibetan cognates, we can actually find more doublets like this, differing in a velar pre-initial, a ge, and also semantically related in the following ways. We can actually see three slightly different kinds of semantic relations, which I will eventually unify. So first, you get verb doublets that are just transitivity alternating. So you get an intransitive that has no ge and transitive that has ge. Here are some examples. So in Proto Tamangic, you have this verb klak, which means to spill, to abandon, to pass by a place, to let a place remain, to divorce, which means to let your wife remain, right? To spill, to let something remain behind, to abandon, to let something remain behind. This is basically to cause to remain, to be leftover. In Tibetan, you have plak. This doesn't have the ge, right? And this one is intransitive. The meaning is the intransitive version of this, basically to be left to remain behind, to remain as a surplus. You get other examples like these as well. So to grill, to make something hard or stiff, to slurp up or drink soup, noisily, like, right? That's transitive. And then lep, lep is making noises, like. So this is intransitive. There are cases like glon, to swallow, to call something, to go down. I had to do a little bit of reconstruction, but I think it works. So in Tibetan, lung is a noun, which means valley, or it could mean scriptural transmission. What connects these two noun meanings? Well, it seems to be a verb meaning that means to be down or to go down, right? This is not a tested, but it can be reconstructed this way. So this is pattern one, right? So pattern one, transitivity alternation. Pattern two, you get noun verb doublets. So in these doublets, the noun is the one without the g and the verb contain the g. These are denominals verbs. Specifically, these verbs usually mean calls to be whatever this noun is. So example, leba or leb, a piece or fragment, like a fragment of paper. Sle, in tamang means to rip something, right? To cause something to become fragments. This one is really good. So ruma is a yogurt culture, leavening dough. So basically things that are fermentation material. In tamang, you actually have a verb, right? That looks like groom. That means to cause to ferment, right? To cause something to become fermentation material. That's actually quite nice in that way. So you get this other type of pattern, right? A noun versus a denominal verb. Third, this is the one that's really interesting. You get verb, verb doublets. Both of them are verbs. They're not transitivity alternating, but the g version has an augmented agentivity. You can see best from this example, right? This is the tamang internal example, right? So the idea is that you break off wood versus you break off bigger pieces of wood with more effort, right? This is what Mazanong tells us. Here you have a couple more. So in Tibetan, you can have to quiver slightly and then you get the tamang version that has a g kliam kliam. That means to wriggle or slither on one's stomach, right? So you have an infant who's not even able to crawl. I imagine like, you know, this infant trying to move around, slither, wriggle on their stomach. So the root rad in Tibetan to scratch to scrape. In tamang, you get kret, which is to scrape or to erase with a knife. So here, right? It's very subtle, but the idea is that the g version has a heightened agentivity. Both of them can actually be transitive. Both of them are agentive, but this one, this g version, gives you an augmented agentivity. So how do you unify these three, right? So the unifying theme I suggested is just augmented agentivity. In all three cases, what you get is that you add agentive meaning to the word that you are deriving with this g prefix. Now, specifically, when you have a predicate, when the word that you're trying to derive does not have an agent argument already, you can actually add it, right? You add an agent argument and you also add the sort of cause-to-be, cause-to-become meaning. In these cases, and in these cases only, you get this apparently causative meaning. So this accounts for both these transitive alternating doublets and also the noun denominal verb doublet meanings, right? So the idea is that predicates can be verbs and nouns, right? So a noun can be thought of as a predicate that means be now, right? So the other pattern, this is the interesting one. If you have a verb that already has an agent argument or already has agentive meaning, you augment that agentive meaning, you heighten that agentive meaning by increasing the intensity of the action by making the verb mean something like do this with more vigor and so on, right? But you don't actually add an agent argument. So this is a case where it deviates from your typical causative, but you don't get a dye transitive out of a verb that's already agentive transitive. So this is what we find when we compare Tamangic and Tibetan. If you come back to Tibetan, it turns out that you can actually find a lot of these verb doublets that show a fossilized G. I found 39 in total, but let me just give you some examples. The important thing is that all of these show exactly the same characteristic pattern of augmented agentivity. Here's how, right? Remember, we have three different sub-patterns, right? Of this augmented agentivity, right? So if you have an intransitive verb that's not agentive that does not have an agent argument, you add an agent argument, you causativeize it, right? So this is what you see here, right? So to come together, to collect, to assemble all of these intransitive, right? To cause, to come together, this is pattern one. Pattern two, right? You have this verb, zen, which itself can be either a noun or verb, but actually this verb is itself denominally verbalized from a noun. The noun itself is zen, that means food. So you have a noun, right? And you add a g to it, you get a verb that means to devour, right? To consume, to cause something to become food for you, right? Essentially. So you get the second pattern where you get a noun that is denominally verbalized into an agentive transitive verb, right? Pattern three. So here you get a verb, roam, wonder about some place, steal through some place, linger at some place. Well, it is not necessarily intransitive. And if you add good to it, you get a verb that means to search or to spy on a place, right? So this is not exactly a transitivity alternation. It is more like an augmented agentivity interpretation, right? Because to wander about, to linger at one place, that's to be around a place without too much intention, without too much effort, right? To spy on a place or to search a place, right? That actually means to be around a place with an intention and with effort. There you go, you have a third pattern here as well. So I'll just say that Tibetan also has these roots that are not really verbs but adjectives, right? So that's accurate, detailed. Sheik boh, broken in pieces. Tam, which you see in tam ba, which is complete or full. Tam jia, which is an older word, means all and entire. These are fine because we can always reconstruct a stative verb root. Adjectives in these languages are generally secondary developments. So these are not problems, right? So these are just like verb transitivity alternation. Okay, so you find it in Tibetan, you find it in Tamangik, you can reconstruct it proto-bodish, right? So my reconstruction is that there is a G prefix in proto-bodish, I call the function, the augmented agentivity function, I call it agentive transitivization, right? So you have this in proto-bodish, okay? So what does that mean from a Tibetan verb mixology standpoint, right? If you have a messy transitive paradigm in Tibetan, which mixes up these two things, A, G pre-initial forms in paradigmatically unexpected places, and B, non-G pre-initial forms, we know that it's probably due to incomplete leveling or completion. So here's some examples, right? So Dao, Ba Dao, Ba Dao, Ba Dao is expected, Ba Dao is unexpected because this would have been, would have involved a root that's Dao, this was due to analogical leveling, but this leveling has not extended here yet, that's the idea. Example two, right? So you have these, you have these verbs, dem, ba-dams, ba-dams, across different dictionaries, you get different forms. Here in this dictionary, you get ba-dams for the past, you would expect ba-dams, right? Here in this other dictionary, you get ba-dams in the present, right? You would expect a dem, in fact, given that the imperative is a dem, you would expect a dem in the present as well, or you would expect doms if the present was ba-dams. So how do you explain this? Well, it's pretty simple, right? So the idea is that you have two roots, dam and ba-dams, one with the agent of transitivization prefix, one without. In this case, note that because both of these are agent of transitive, dam is already agent of transitive, adding the G to it is not going to do too much to the semantics, right? So you get two verbs which are basically synonymous. So then you get two regular paradigms, right? So if these were regular, and then you can say, well, these dictionary forms are just differently subpleted paradigms made up of different parts of these two paradigms that happen to be essentially identical in meaning. That's the mixology, right? The mixology is that you have subpletion. Now, this is a more elaborate example. Basically you have two transitivity related pieces of protomorphology. There's initial voicing alternation. And there's also this fossilized G. Turns out that you can get dictionary forms of transitive paradigms that mix up forms from multiple different transitive paradigm strategies. So we have three here. And in fact, all three can be kind of mixed up in one paradigm. Okay, that's mixology point eight. Okay, so we've done that. Mixology point two, point two is actually simpler. So we're gonna go back to our initial voicing alternation examples. So we have the bub, pub, cost of fall, right? Recall that it has a related intransitive which is voiced, bub, and an intransitive which is pimp. So here's the example that we would have just seen, right? So you have another of one of these triplet, right? So brawl, intransitive, brawl, brawl, transitive, but brawl, that's also transitive. How does that happen? Didn't we just say that this was supposed to be intransitive? In general, in these triplets, this voiceless third paradigm, this voiceless third root is kind of complicated. It's messy. It's messy in two ways. First, it's messy in transitivity. Most of these voiceless third paradigms are intransitive but a few are transitive, right? So here's one. Second, the root vowel is also messy because you would expect bub to be here, also bub, but this one actually displays the r to a ablot which is characteristic of the present stem of the transitive, right? So it's kind of complicated. It's like, why did you mix those up? And this is actually quite regular across all these, all these instances of a voiceless third paradigm. So questions to figure out, why is there this variability in transitivity? And also why is there this r to a ablot? How does it get it? The ultimate question is, is there any correlation at all between voicing and transitivity? So I'll answer that towards the end of today. First answer, these two questions. I think that it's a pretty simple answer. The third voiceless root is not primary. It is secondary. It developed from a transitive into an intransitive. So I call this secondary intransitivization. Now, in Tibetan, the semantic connection between the two intransitives is not always obvious. So you have fo, you have to honorific to come. You have pour out, you have to float to swim. It's not always clear how the two intransitives are connected semantically. So we just kind of leave them as is in Tibetan. Now, if you look at tamagic, it's actually more clear. In tamagic, you get these intransitive doublets. So bub to fall down. Pub is a certain kind of going down. It's for a human to come down. This is the example, right? I come down pub. Pure for spring water to gush out from the ground, for rainwater to be gushing along on the street. Pure also intransitive. It's for water to boil in like the pot. So you have these roots, first and third, that denote the same general event concept, which are both intransitive. But the third one is narrower than the first, right? And the event denoted by the third root is narrow in this following way. It involves an agent causation at the background versus these first roots, which denote events that are kind of uncaused or naturally occurring. There's no sort of causer or agent in the background, right? So I come down, well, I have to kind of, of my own volition cause myself to come down, right? This fall down can be trees, snow, earth, whatever. Same thing here, right? So spring rainwater versus boiling water, right? So you actually are boiling water. You're causing water to spill out in a certain way. Occasionally, if we look back at Tibetan, you also get this semantic connection. So we don't have to go through the detailed examples here. So my proposal here is that the third roots that are intransitive are secondary. The intransitivity developed out of a reanalysis of transitive meaning that is syntactically encoded. So you start with a transitive second root, a transitive voiceless root, pure. So this would have been something like to cause to gush forth at a certain point. At this stage, you could say a sentence like, John caused water to gush forth by boiling it. Now, these languages often drop arguments, right? So you can drop the agent argument very frequently, right? So you could just say things like, somebody caused water to gush forth by boiling it, presumably. Now in this case, you realize that the boiling is actually not so clear anymore, right? Because you don't have the person there anymore. So if you have this frequent enough, what I predict can happen is that the being boiled, the cause part of the event will just be reanalyzed as the lexical meaning of the verb, right? So originally this verb just meant to cause to gush forth. Now you drop the agent argument, but you incorporate this being boiled meaning into it. So water was caused to gush forth, water became boiled. So this is how you get this sort of agent caused intransitive meaning that is secondary. So what does that tell us about messy paradigms in Tibetan? Well, it tells us that if you accept the conclusion that voiceless third roots, what voiceless third paradigms are generated from secondary intransitivization, then it can be variable, right? So some of the voiceless roots may not go through intransitivization. It may still be transitive. So pep has gone through intransitivization, prow has not, right? That's variable. And second, how do you get the ablot? Well, it's through simple analogical extension. So here's the analogical formula, right? So you can model it up against other abloting paradigms and then you can get it through the entire paradigm. Okay, so this is the mixology point, right? Okay, so we've done both of our kind of more put together points. So let's accept the idea that voiceless third roots develop secondarily, right? They're secondarily intransitivized. But what have I been assuming in giving that analysis? Well, if we go back to our triplet, your first root is always voiceless. It's always intransitive. That's not a problem. Your third root is voiceless. We've said that if this third root is intransitive, it's due to a secondary intransitivization. I was assuming that this thing was transitive then, right? So does that mean that voicelessness, right? Just correlates with transitivity. But if that is the case, how come that in all of these second paradigms, these are actually always transitive paradigms, right? You always get a part of it that involves the voiced root and a part of it that involves the voiceless root. And this is regular as you get it all across these examples. I don't have a put together idea for this. And I don't think we have very good analysis for this just yet, but here is a thought. I think that it might actually be kind of a historical phonological reason. It's a possibility. If you look at unpre-initialed syllables in Tibetan, they generally correspond to tamagic tone A. And tone A is supposed to be this just general modal tone. So that's expected. So gap, gap, gong, gong, gong, gong, gong, gong, zin, zin, za, za, di, di, do, do. It's pretty good. There is a non-trivial number of exceptions where you get B tones. I can explain some of them already. So it's actually true across laryngeal categories. So you get voiced, voiced correspondences, aspirate, aspirate correspondences, za, zor, chop, tang, top, pa, po, same thing. You also get plain voiceless, plain voiceless correlations, which is not very easy to find because Tibetan doesn't like plain unpre-initialed voiceless consonants, but you have one here. So it's true. So the correlation between unpre-initialed syllables and tone A seems to be pretty general between Tibetan and Tamangic. Now, there is another correspondence class. So you get in Tamangic A tones, but voiced. So voiced, za, A tone, voiced, zang, A tone, zi, do, and bon. The expected correlation would be also za, zang, zi, do, bon, right, in Tibetan because of, you know, the pattern we just showed, but you get actually aspirates. So za, zang, zi, do, bon. You know, why am I taking this so seriously, right? Because there's a lot of nouns. There are even very basic sign-all Tibetan nouns, so like za, right, sun, that you get in Chinese too. Nouns, unlike verbs, are morphology poor, generally speaking, in Tibetan. So then if you get these funky correspondences across different laryngeal classes, I think it's kind of telling me something. Could Proto-Bodish have a laryngeal class, which always surfaces as voiced in Tamangic, so you get these, right? But surfaces as voiceless in Tibetan when you don't pre-initial it, which is the case in these nouns, but when you pre-initial it with, you know, certain pre-initials, I think, you get a b instead. What that would buy us is that we could say that what you have in your second class, your voice alternating class of transitive verbs in Tibetan, they actually don't reconstruct two different roots, differing in voicing. It actually reconstructs back to the same root, which belongs to a funky laryngeal classes, neither voiced nor voiceless. That is your transitive root, right? So the root-initial of your intransitive root is always voiced, but the root-initial of your transitive root is this funky laryngeal class, which surfaces as P, this is a typo, right? When not prefixed or, you know, prefixed with B or something like that. That would be how this can be explained. I suppose you could further hypothesize that this transitivity alternation between roots was in an even earlier stage of the language due to some kind of segmental prefixation that transitivizes. And that's all, that's the presentation about mixology.