 This talk is in some way complementary with La Infance talk of yesterday. La Infance presented something about prefixes and I'm going to talk about suffixes. How does evidence from languages with conservative morphology in a scientific family, to what extent are these relevant to studying Chinese? Before I start, let me say a few words about methodology. I think that in a scientific family, the two groups among the 30 uncontroversable subgroups of the family that have the most complex morphology of gyarungik on the one hand that Infance has been talking about yesterday and gyarungik on the other hand. There is evidence that I cannot go into too much today that these languages present archaic features that some of these morphology practices are, to some extent, ancient and cognate between these two families. And although the number of cognates between gyarungik, gyarungik and Chinese is very limited, I mean, in the Infance presentation he tried to put together all the cognates and we have less than 200 cognates between gyarungik and Chinese. So these languages are not so useful to reconstruct actual clusters, for instance, or confirm many hypothesis about clusters in all Chinese because common vocabulary is so limited, but on the other hand the rich morphology of these languages can be a guide to at least create hypothesis to interpret the morphology of all Chinese. I believe that when studying a language whose morphology has been eroded it becomes impossible to really recover the meaning of the lost affixes only on a language internal basis. I think it's when we look at Indo-European, for instance, if you study a language like Albanian or Armenian you can know the entire literature by heart. You will never be able to etymology any word if you don't have knowledge of Sanskrit and Greek. You always need to understand the basic morphological structure of total language and try to recover the traces of it you can find in languages that have less conservative morphology even if they are not closely related in Shambha. So I'm going to talk about one topic, the S suffix because it's the most conspicuous suffix in all Chinese. I think there are several distinct suffixes, actually, that correspond to Chuchon. I hope I can be forgiven not to present exhaustive references on this topic. For instance, I didn't quote Sun Yuban. I was aware of his existence thanks to Chuchon, but I will just use Downer's study of the Greek. More than 50 years ago, Downer collected examples of the derivation by Chuchon. He found eight different parts with widely distinct semantic properties as you can see in section 2. His interpretation of this diversity in the meaning of this suffix is that the Chuchon tone derivation, as he calls it, was not really the trace of a new fractional system but simply a way to create new words, as you can say in the quote taken from his article. And this view has been recreated by several other authors. I think that this view, at the time when Downer wrote, was perfectly legitimate, that with 50 years later, as you can see, of course, to criticize the previous authors, but I think that such a view, if you hold a view like that, then that's it, you can't go any further. Basically, you say that any study of trace morphology is pointless. I mean, I think that's a view that in some way leads nowhere. At least we should try to interpret the remnants of morphology in Chinese using our knowledge of more conservative languages, and that's what I'm going to try to do. Of course, the comparison I'm going to present in this paper I have made no claim that they are all true. I'm just presenting a series of possibilities, and I leave it up to the time specialist to determine to what extent they can accept them a lot. So, Downer found eight times. I'm not aware of... Since I haven't read in detail Suni Wen's book, I don't know if more recent work on this topic has discovered new types of new meanings for this derivation, but in any case, the types that Downer has proposed are all quite well-attested. Derivation from verb to noun. So it's not explicit that there are various types of nouns, actually. Like you have to be high and high, so it's an abstract noun. Be at the place, but in some cases it's the object of a transitive or it may correspond to the only argument of an intransitive. So it seems to be used to derive various types of nouns, it doesn't have a very specific meaning. You can derive a verb from a noun. There are also many examples, please. So the exact opposite of previous one. And the meaning of the verb is also quite varied. Sometimes you use an implement as you have in this pillage. You have many examples of causative. In the case of bi and cell, it's not really a causative, but in any case, many languages in the world use causative markers to derive words like cell or oral from bi. That's a well-known situation. And you have a few examples, like how and how, to like to be good, which are not really causative. They are more like what I have called using the Arabic tradition, tropative, so to consider something to be good. And there is evidence, for instance, in general language, that causative can also be used in some cases with a tropative meaning. There are examples of applicative. So applicative, I mean, you take an intransitive verb and you derive a transitive verb from it, unlike in the case of a causative. In a causative, you should take an intransitive verb and you causativeize it. The object of the transitive verb, the cosi, corresponds to the subject of the transitive verb. I said to make someone sit, the person who sits is the object of the second, of the causative verb. In the case of an applicative, the argument that you add with the derivation is the object. It can refer to the subject of the derived verb and correspond to the subject of the original verb. I will show you more concrete cases using living languages to explain what I mean by applicative for those of you who are not familiar with it. Let's give the example of taking a run in the out-run. You have a few lexicalized applicative verbs in English, German, you have a few examples like that. It's not productive in any European language but in some language family it's quite productive. Five, you have this restricted meaning, kind of a category which appears to be a casual for various things and I won't attempt to discuss it more. You have passive. So passive would be the object of the causative. And you have quite a few examples. And you also have, that one did not use this verb because it wasn't invented when you wrote this article, anti-passive. What is anti-passive? Anti-passive is to derive an intransitive verb from a transitive verb. But unlike the passive where the only argument of the intransitive verb corresponds to the patient of the transitive verb. In the case of anti-passive, the only argument of the intransitive verb corresponds to the subject. So to put it bluntly, simply like in French, in French we have use of the reflexive in an anti-passive like fashion with some verb like bat, bat to beat someone, so bat to fight, which is intransitive, which works intransitively, let's say. In French, you only have very few examples like that with the reflexive used to create this anti-passive construction but again, some languages have productive ways of making anti-passive. In European languages, generally we don't have anti-passive for one reason because nearly all European languages have widespread agent preserving ability. So basically most transitive verbs, whether in English, French, you can omit the object and it's interpreted as meaning an unspecified object. So you can say, I eat, I eat an apple, I eat means I have a meal. But in languages that, in most languages of the world actually, if you omit the object it will mean, it's with a specified object and you need to have an anti-passive marker to indicate that you have an unspecified object. So you have a few examples of anti-passive in Chinese. Then you have adverb derivation with very few examples, just the three or four. And a forming compound. So you have in some cases the first or the second element of a compound with an attution added. Okay, so in the first glance it looks like what Donald says, it's just a big mess and you'll never be able to solve anything. And it's just, let's do something else. Well, I'm just going to try to propose a series of etymologies in other languages. Some previous colleagues have already pointed out some of them. The first of them following Audrey Kouin was a discovery of the tushon comes from an S-sophics, comes from an S, was a forest and of course there are many other scholars have built on that idea. Before I go any further, I will propose, I will start by proposing a sound law, a potential sound and you are free to criticize and to comment that. In Tibetan, I first start with a typological example. In Tibetan, it's well known that in what we call the present tense, whether it's proper or not to call it this way, I would argue. We have some transitive verbs, have a conjugation by which we add either an S or a final D. And all people who have worked on Tibetan internal reconstruction agree that you have a sound law which tells you that something to write on the blood there. Oh, can you just... Yeah, yeah. So if you look at these two examples, you have in this a tap to plant or this is a verb that has many meanings. So, when you have a verb ending whose root ends with pa-ma-ga and various others, you end up with S. So, let's say this was from something like en-da-de. Let's reconstruct it for the... whatever. And the vowel R changes to A. When you have a verb ending in a vowel, you still get the fronting of R to A, but it remains the same. So you have a dit, pa-tas, to chase, so it comes from en-da, and it stays that has a fronting and something like that. So these would be the rough protocols for the present tense forms that I put here. And thus, we have a sound law in Tibetan which tells you that p-tah, p-tah, k-tah, n-tah, etc. change to p-tah, k-tah, m-tah. It's a fact that you don't have clusters anywhere in Tibetan. Are you the only possible final clusters we have in Tibetan? Well, I won't talk about the tatang, but these consonants are with S. So, let's imagine for a moment that since in Chinese, the only final cluster we can reconstruct involves consonants plus S, and there are many of them. I mean, it's striking that in any reconstruction of a Chinese, you get S's everywhere. You get also R's everywhere. I think this is a problem, I think, for all reconstruction of a Chinese, not specifically Laurent's or Schuessler's or anyone else's. The communist opinion of the present reconstruction reconstruct, I think, too many R's and too many S's. My view is that maybe the present reconstruction don't make enough distinctions. One of the reasons is that consonants different from S and R have merged with the original R's and S's and from the present data it's not always possible to distinguish between them. Let's accept for the time being that T could change to S when it occurs as the element of final cluster. At least maybe not after all consonants, but maybe after some consonants. We have evidence in some languages of the family, especially curanty for the existence of cluster like that. So it's not legitimate to suppose that it may have any case. If we do that, then we are entitled to compare the final S in Chinese not only with S suffixes but also with T suffixes in other languages. And that's what I'm going to try to do. Besides for yourself whether you believe it's reliable or not. So let's go one by one. I won't discuss all the downers categories. I won't discuss the restricted minimum but most of the other ones. The first most obvious that was discovered is the nominalization issue. Because it's not productive anymore but we have quite a big series of nouns derived from the S suffixes. I think this is completely un-controversial and there are even a few downers that are probably connected to Chinese. Actually there is a typo on the middle Chinese of cloth. It should be, there's no IJ. It's just simply I. Oh yeah, I didn't include the old Chinese. It's just out of laziness. I have added it in the eventual version. I will add them. But I think you can try to figure out for yourself. I use a slightly modified version of middle Chinese of back says middle Chinese. So I won't dwell too much on normalization S normalization because I think it's quite un-controversial. I will talk a little on the causative applicative. Okay, so for the causative applicative there is good evidence who are the scientific family of T suffixes that derive applicative or causative verbs in various languages. It's mainly in Quarante languages that you have a lot of them but you get a few examples in many unrelated branches including German languages. In Tibetan there are many good examples but... So, libu is probably the language that has the best cases because it distinguishes very clearly an S causative and an S applicative. So here taking data from Boyd Mikhailovsky so these are the root forms. It's not the actual way these forms are pronounced because they always occur with suffixes and I won't go into the morphophonology of it it's 111 to represent a talk. You have an intranstive verb harp to weep. You can derive from it a causative verb haps with an S suffix to cause to weep and an applicative verb like here you see the difference. Weep mom for. In the case of cause to weep the person who weeps is the patient corresponds to the... the subject of the intranstive verb. In the second case the applicative derivation the person who weeps is the agent corresponding to the... to the subject of the intranstive verb. So you see the different relations between causative and applicative derivations that it's difficult to claim that it's productive in any Kiranti languages in the sense that you cannot as far as I know apply to long words but in some way the Kiranti languages you have to understand that in Kiranti verbs are a closed class large closed class but you only have 700 or 800 verbs and that's all and you cannot create new ones you barely can borrow verbs but you can borrow them but it's really... it's completely different from anything can be turned into a verb. So... the fact that we don't have example of for instance, Blippali verbs to which these are added is not really... does not really indicate that these derivations are... are dyed. So outside of Kiranti we don't have a nice example, so I'm going to... in order to make you understand the meaning of these suffixes I'm going to show you a few more from Kali, the language on Blippali so here are examples of applicative in Kali so for instance you have to be angry to scold to be angry at someone a rid laugh a rid double T laugh at that's again a root for it's not pronounced like that it's going to dry if you use it in a closed class then the road to code actually it may be a cognate with the fey to bark I don't know whatever the R is who are you? So basically these verbs either the recipient of something or the stimulus of a feeling to be frustrated from something so either it's a feeling and you have the experience which corresponds to this subject and you add it with these suffixes you add the stimulus the causative T can also be used in Kali to express causative so you have examples like word to run word to run P calm this one is interesting because you have the only verb with a T applicative in gyarnic is the g from we and it's completely this one so this is the evidence that this derivation has some degree of antiquity now in Chinese the only proposal for a T applicative was by Laurent I won't boil too much on it I think that one possibility to explain some of the causative applicatives if you accept the sound law I proposed would be the following at some stage you have this sound law of estimating my time sorry so you have this sound law and you had quite a lot of pairs between verbs with S suffix expressing applicative and in transitive verbs and then since you also had a causative suffix S the distinction between applicative and causative is already quite tiny even in language that had where this derivation was quite productive became blurred and the S causative was general the meaning the causative applicative meaning of what was extended to a context outside of the application of this sound law and that's why you also have the S suffix expressing applicative or causative with verbs in open syrup for the passive and anti-passive I think now we can propose a different comparison in Qienti languages and also in various languages with relatively conservative morphology like Toulon and like Maga we have a C suffix or X suffix this reflexive has it's the basic meaning it's the most common meaning but it has many additional meanings it can be used for impersonal subject in some of the cases it's used for anti-passive you have example 7 and 8 L'Opée Grande and discussed by which is Gramsina I feel disgust so here you see that the subject of the of the reflexive corresponds to the subject of the transitive it also has so so I think that one possibility for the for the use of Qienti to express both the passive and anti-passive would be comparison with this with this reflexive suffix for the remainder for the the denominal use of S I think that this is somehow of a problem for scientific comparison the art to the extent of manual is no good example of S suffixes deriving verb from nouns so whatever the explanation must be for Chinese I think that Giaoling or the Qienti evidence of little use for the duration of adverbs one possibility is the the use of an S suffix which appears to be all because it's shared by Giaoling and Qienti languages to which whose original meaning must have been locative it's used in Tibetan it appears as in case markers various case markers including the the adjective and the additive in the general languages it's used mainly for for locative with motion and I think that this can potentially be compared with the use of S suffix of the S suffix in Chinese with the as an adverb derivation so yeah for the other ones I don't think I have good comparisons but I think that the comparison I propose today I have to stress that there is some degree of speculation here I am not claiming that they are necessarily related to Chinese but at least I think it's better to explore how the meaning of the derivation we observe in Chinese might be linked to living languages rather than speculate from a Chinese internal perspective and I hope that can comment on the proposal I made here today thank you