 Okay, so I actually didn't know the kind of audience I was going to be presenting to, so I didn't know if I wanted to throw OT tableaus. Actually, I ended up deciding regardless of who I wanted to present to, I didn't want to do that. So this is going to be a general descriptive analysis of the evolution of the imperfect. And I'm going to talk about identity constraints and some stuff, but if you're not familiar with, actually who's familiar with OT? Couple people. Okay, everyone else that's totally fine. I deliberately, hopefully this should make sense to everyone who doesn't know anything about OT. And I definitely, okay, so anyway, so Ba is the imperfect marker in Latin. If you look at a Latin paradigm, pretty much no matter what, you have Ba for every imperfect person and number, so amabam, amabas, amabat, et cetera, et cetera, et cetera. Words that Latin students have been repeating, wrote for hundreds of years. So that's pretty expectable. It's a very regular paradigm. It marks pretty much every verb, one exception for essay, the copula. And that's expected, if you're going to have an irregular verb, it's going to be the copula. So something weird happens specifically in Ibero romance, Spanish included, Portuguese, Galician, Catalan, you know, all the ones you typically think of. And that is, so this is pretty normal. So amabam becomes amaba, amabas, you know, all this stuff is regular sound change, you lose the T in the third person, that's totally normal. The U in the second plural, or excuse me, the first person plural turns to, oh, that's normal. This is all expected. But something weird, well, that's just AR stem verbs, right? So this is the low vowel class. That's all normal. This is just like a more apparently normal sound change. But of course, Spanish has two other verb classes, the ER and the IR verbs. And when in Latin, the B is marking every single verb in the language. So you have dikebam or dicebam, that becomes desia in Spanish, and pretty much the same with all other Iberian languages. So you have that, of course, across the entire paradigm. So this is, I guess, interesting because what you have is you have a regular paradigm, and you have some sound changes in some way apply to it, and it becomes irregular. So you have, again, the low class vowel, the AR class, they retain B in all the other classes, you lose it, pretty simple. So if you're a Spanish learner or Spanish teacher, the rule of thumb is, so you use ball with AR verbs, EL with everything else. So it's relatively predictable, not too irregular, but it's a little strange that it's even there in the first place. So the, I guess, possibly surprising thing is the dicebam to desia, that's actually the regular expected sound change. If you apply the sound shifts between Latin and Spanish, this is the thing that's expected. Because you have a general loss of intervocalic voice plosives, so legre becomes leer, cadre becomes caer, et cetera, et cetera, or frididom becomes frio, which is a good one, because you have two voice plosives both disappearing. So this is a general shift, there's a general lanish and basically everything intervocalically in Western romance. So, oh well, I should say, all of these are with G and D. You're not going to find many examples with B outside of the imperfect. That's just because in Indo-European, for different reasons, there are very few B sounds in Latin, and that has to do with glotalic hypothesis or something, but that's another topic for somewhere else. Just know that there are very few Bs, so we don't have many examples of it happening. So the question is, if you have Dikebam to desia, why do we actually maintain the B in the AR class verbs? And if you rephrase that, it might actually make more sense. So if you have Dikebam become desia, why not Amabam to Amaa, or something like that? And when you put it that way, it should be a little obvious. There's something going on. Either Spanish is trying to avoid hiatus, so you don't want to have Amaa with some kind of hiatus or glottal stuff, or it's trying to preserve meter. So Spanish, of course, is a so-called syllable time language, depending on your definition of that. So maybe Spanish is trying to conserve that. Well, in desia, you still have that sort of Latin-made syllable structure with three different things, their vowels. You don't have that in a hypothetical form, Amaa, for the imperfect. But if you look at other romance languages, this is not a universal tendency. It's strong in Spanish and Iberian-Iberia romance languages. It's not universal. So if you look at Romanian, for example, Romanian is sort of the opposite of Spanish in that. Well, first off, Romanian is weird, because it's like the only romance language where you have no good Amare verb, so you have to use adorare. So, which, of course, means exactly what it sounds like. So adorabam becomes adoram, etc., etc., the s to i, that's a normal transformation in Eastern romance, etc., etc. But of course, what you have is a total loss of the b and the a. So what happens is the b is lost and the a coalesces in some way. So Romanian in Eastern romance languages, they actually don't have the same transformations Western romance has, where they lie to everything. But in situations in sort of morphological paradigms and in the post-tonic syllables that you have sort of the same thing happening. And of course, you can look at, you know, the vicere verb as well. You have pretty, I mean, the d to z is expected in this situation, but you see that you don't retain the b in any of these. Now, Italian, of course, is the other leg of the tripod, so to speak, in that, of course, you don't lose anything. So Italian, basically, b's become v and everything is retained. So you retain the meter here and you retain, you don't actually lose anything in any verb class. I'll just skip through those. So to generalize what we've talked about so far, so Spanish, you have, Spanish, another Iberian romance language is you have partial loss of the b. Romanian total loss Iberian. Well, you change it to a v, but it's totally conserved. Okay. So traditionally, in a sort of rule-based account of a phonological change, what you would have is each language, when you, you know, each language basically performed a different sound change on its, you know, funnel, its morphological inventory. So each one, you know, the speakers of Spanish would have a morphological rule that said, okay, well, we lose b in this situation, not in this situation, or Romanian would have one that universally lose it and coalesce as vowels, Italian would have another one. But that's a little strange because what's going on is phenomenologically the same thing. So it's a little weird to say that there are all of these similar sound changes that are occurring all over the Romance-Sprockblund, which are sort of the same and sort of different, but we have to model them formally as being totally different animals. So it'd be nice if there were just one way of unifying all of this change. So how do we do that? Now, of course, we, one can do that with constraints, OT, something like that. So in case you're not familiar, traditionally, in linguistic sense, you know, the classic Sanskrit grammarians, everything has been rule-based. So you take an underlying representation of a sound, you perform operations on it, intervocalic flapping, inner, excuse me, inner vocal, yeah, or inner vocalic voicing or something like that. A lot, a lot of different operations occur on an underlying form until you get to whatever we actually say, which may be totally different. While, of course, during, you know, the past couple of decades in formal phonology, there's been a shift to the constraint-based view where there's an underlying representation. And what determines how you pronounce that isn't necessarily a sequence of rules, but it's different constraints like this particular sequence is difficult to pronounce. You know, I want to conserve loyalty to an underlying form. You know, I want to avoid this or I like this. There are different constraints, and the differences between languages are ultimately the differences between what constraints they like more than other ones. So we can model the difference in romance languages like this. So we can formalize these, these names aren't very religious or shouldn't be taken religiously. So we can think of the Leninian tendency as just a constraint. And I'll talk more about this later. This isn't the way you should formally formalize it, but I'll just call it Len for now. That is, at some point in time, there's a phonological shift that sort of doesn't like the intervocalic voiced sound that wants to get rid of them. Aside from that, there are countervailing tendencies. So there's the tendency to conserve each sound as it's represented. So you don't want to lose, there's, there is a part of the language faculty that says, I don't want you to get rid of sounds. You know, that's a constraint as well. And also there's a tendency to preserve meter like what we saw in Spanish. So I'll call this max syllable, but you could formulate this in different ways. What you call it isn't super important. So if you put it this way, you actually see the differences between romance languages are really just the differences in what constraints they prioritize. So in Spanish, you have a very strong preference for preserving syllable structure. So you want to have, that's the most important thing. Even if you have this really seductive sound change, you don't want to lose that syllable structure. So max, or excuse me, Spanish prioritizes max syllable. That's number one. Number two is LaNitian. They'll lanide if it doesn't violate the first thing. That's fine. That's fine. And last constraint is max. And this, the last max is in reference to B. So they don't really, they care about these theoretically, but not as much as they would like to do the LaNitian, the sound change. No, Romanian just has a different order. So their priority is LaNitian. They don't care if they have to bulldoze over different, you know, metrical feet or something like that. They don't care about that. They just want to get the sound change done. In Italian, I only have two constraints here because the max syllable constraint actually doesn't even matter in this situation because Italian just wants to keep that B or V, wants to keep some kind of consonant there. And we can't actually tell where max syllable is in this situation. So the difference here is, so well, I should just say, what's the essence of the sound shift? The sound shift is really just a higher prioritization of the LaN constraint. So what happened diachronically, you can simply say, instead of different language communities learning different sound changes, what happened formally is just Lin became higher prioritized in all languages. Now, what you did with that depends on what other constraints, whatever language community, whatever sprock one actually happened to like. So as generation, it should be said, as generations go by, people lose this underlying representation of B. So no Romanian speaker today thinks that there is being a B in the imperfect. There's no such thing. So since that's the case, you no longer have to think about this constraint that's theoretically line-eyeding it. It becomes unimportant. So the whole game sort of becomes a reset. So you can actually have, yeah, so nowadays Lin could be ambiguous. But you can sort of have the same shift happening again, actually synchronically in Spanish. Some speakers exhibit an alternation that looks like this. So you have the word hablado, which some people will seamlessly pronounce hablado. That's no big deal for them. But on the other side, with respect to hablada, very few, I don't know, I don't, there might be someone who does this, but I don't think anyone says hablada or totally line-eyeding the D. But this is exactly the same thing. So you have to continue a new linician process, but in the same way that Spanish prioritized that syllable maintenance, it's still doing it today. So in some sense, you can make predictions of how sound changes permeate through a language based on the general constraints that they still might be prioritizing. So I should say what is, how much time do I have? Ten minutes. What's that? Eight minutes. Okay, great. So I should say what exactly is lin? So I've just said there's a linician process that happens, and I've said that's a constraint. So that's terrible to do an optimality theory, even though people do stuff like that all the time. But that's theoretically not something you want to do. And I'll put it this way. So traditional OT had these things called ident constraints. So an ident constraint is basically a constraint that says don't change phonemes. So if it's underlyingly a B, you want it to be a B. Or you could have it referred to specifically voice. So don't change voice or don't change place of articulation, any of those things. That's generally what an ident constraint is. And so they tie how you pronounce things to the underlying form and vice versa. So for example, if you have an underlying B, it is bad to pronounce that as B. That's just, you don't like that. Now this ident constraints are a classic part of optimality theories used everywhere. But I think it's interesting to think about what are ident constraints really for? So one thing I think that's hard to get through your head is that the phoneme B is not a voiced bilabial plosive. The phoneme B is an abstraction. It is something that theoretically exists in your head. There are some words that have B in it, and they are networked with all the other words that have B in it. And we use those phonemes to distinguish the words with B from the words without B. So bat and vat are different words. And that's not, we don't necessarily need bat to have a bilabial plosive. There's nothing essentially about that. It's just conventionally, we use this particular pronunciation to differentiate classes of words. So the voiced bilabial plosive is something that exists on the surface representation. So ident constraints, they've sort of been cheating all this time and saying that there's some way in which all of these features are sort of part of your, all these articulatory features are part of what's actually going on in your brain. But I think if you really think about ident constraints are formally doing, is they're keep, as I said, they're sort of keeping different words to different categories. So the reason we want an ident constraint is so we can distinguish bat from vat, or other words like that. You want to keep different words sort of in the pile they belong in. And so B can, the phoneme B conventionally rules the sound ba and everything around there in each phoneme, you know, has its own sort of territory. And the important thing is not necessarily that they keep preserving that exact sound, but that they keep distinguishing words. And that's sort of the important information theoretic part of it. So, but the thing about it is that goal of ident falls apart. Like there's no reason to conserve a historical pronunciation if you have something like a chain shift, or if you have no other words that you're distinguishing. So for example, and of course, in Spanish at the time, or in Western romance at the time, and in some portions of Eastern romance, you have a general chain shift. So you have gotta becoming gotta, you know, toto becoming todo, you have the general shift. So, you know, reduplicant or excuse me, long consonants become just plain voiceless stops, voiceless stops become voice stops, voice stops become, you know, they either are lanide or historically they go through fricative phases or something like that. And I guess in Spanish, todo is the fricative depending on your definition. So you have this general shift of everything in the whole language in a particular direction, in a particular direction. So if you think of it this way, when a chain shift gets in motion, if you want to distinguish words, it's good to follow the chain shift. It's good to actually keep that sound distinct by changing how it's been historically pronounced. So if you really think about it, this is ident doing what it's supposed to do, but it's sort of I'm turning it upside down. So what I'm trying to say is what I modeled as being a tendency to linician, you can just actually model it as being an identity constraint, if you understand that differently. It's really just a tendency for languages to differentiate different words. It's just sort of in motion. So it's actually doing the opposite of what it historically did, but with the same goal, if that makes sense. So in conclusion, so just a recap, I guess. So descriptive status of the imperfect, how it looks in different romance languages, we can treat sound changes generally as sort of shifts in constraint orderings. And because different languages have different prosodic constraints, they can respond to sound changes differently across different languages, but they can also respond to sound changes systematically the same way as they occur. So when you have chain shifts or when you have something that we mentioned, the Ablau thing in Spanish, when you have a shift like that occurring, it's understandable that if what's going on synchronically in the language looks like what it was doing 500 years ago, because a lot of the constraints of the language, a lot of the prosodic desires are going to remain the same. And I didn't, as I said, had many faces. And I think, yeah, that's it. So any questions? Analyze and count it. Sound deletion in words like ablau, ablau, things like this, where you would have two hands encountering it. Sound was too polite. Nothing, nothing synchronically. Yeah, because I think you would find as they do. Yeah, it depends on, I guess it's an issue of how far gone the sound shift is going right now. Yeah, I mean, what I'm trying to say is it's going to be formally similar to what has happened before. I don't know how it is now, I just use it as an example. Yeah, yeah, yeah, yeah. Any QP that is fewer cases than the ones that have ablau and QP. It's just interesting to see if synchronically you find the same pattern. Yeah, would you, do you have the intuition that people say ablau or something like that? Oh, yeah. I think there's a thing that I would say that would do, yeah, especially in Spanish. Okay, yeah, I'll check into that. And even that is true too. Yeah, because the present D and then you have ablau. Especially in the parts, yeah. Chileans, Spanish, they say like weah, weah, which even like, well, nothing, it's relevant, but like even orthographically, they don't even write the second A, they just weah, sometimes they write a WEA with an accent. Yeah, I mean, my general framework here wouldn't necessarily not predict something like that, but I do think there's a, my intuition is there's definitely a tendency for it to happen more often than you have ablau or something like that. So the deep optimality, if you're ready for questions. I'm watching. Okay, is there still time? Oh, is there still time? Yeah, no worries. So you're essentially like Lynn and I get on the scene. Is that like the last technical way of like merge or over the OT way of trying to preserve a finite contrast? Yeah, exactly. So I, you know, I was saying in a different ways of doing that in OT, it's sort of hard other than me just sort of saying, I think that's the goal of what I've been supposed to be. But the problem with it, the problem in implementing it is because I'm, you basically have to say every single phoning is sort of connected to every other phoning. That causes really big OT tableaus. That's what you need OT help for. But yeah, I'm not entirely sure that how to implement this, like I said, general, I mean, there are people, there are people in my department who actually do this stuff, Andy Guido, for example, does this stuff where sound change is motivated by contrasts. So like if you have a lot of, you know, two sounds that are very highly contrasted in a language, you're not going to use the distinction between them. But if you have, you know, two sounds that, you know, the distinction isn't very important for a couple of sounds, you have to use this all the time. And that's sort of the same intuition I have here. You know, but then that kind of applies to the purposeful situation that we're talking about. Like, you're, the two vowels we delete in Ablau is you have two very distinct vowels that are making something different. Whereas when you delete that D in the part simple in Ablau, then you have the two same vowels, so they're the same kind of asymmetrical relationship. If I have a distinction, big distinction, I'm not going to lose it. But if I have a smaller distinction, then maybe it will. Yeah, well, it's in the opposite direction. Yeah, well, it's, yeah, no, yeah, you're right. It's not exactly what I was getting at, though, like what I was getting at is like, so we have a distinction in English, dilution versus delusion. It's one of the only minimal pairs of show and show. So that's a good place for us to lose that distinction. So the idea is comparing different words and like, how many, where did they actually contrast to something like that? But yeah, I mean, with respect to what you said, yeah, there is a tendency because the vowels are different, you can dip them on it, dip them on it or something like that. And again, most languages really despise hiatus and they really despise, like, if Spanish were to implement like a super long vowel, that would be a huge shift in the finality of everything. Like speakers would have to learn this entire category. But, you know, and there are languages that are different as well. So ancient Greek, for example, happens to love vowel line. They're just like, ancient Greek is the weirdest language that actually is like, there's a semi-vowel between those vowels. I'm going to get rid of it just because this is how it is. I'm thinking, of course, it is you have the a position because the a are to go together and then does it. So you have a, a, but you don't do a, I just simplify to one vowel. So it's a, and I'm thinking that would be a much better solution than to do both of them. Yeah, yeah. In general. Yeah, I think if they have, if they have no, no kind of epithetic thing that they can do between them. And they emerge. Yeah, in the same time, of course, happens in Romanian. They don't want to have adoram, they say, adoram. Yeah, in the same, so. Okay, thank you. Great, thank you.