Conlangery 135 Using Linguistic Theory for Conlanging

Conlangery 135 Using Linguistic Theory for Conlanging

Published: 18-01-13

Content from the Conlangery Podcast is licenced as Creative Commons Attribution - Non Commercial - Share-Alike. You are free to copy, distribute, remix, and create derivative works from the show, so long as you give attribution, your work is not commercial in nature, and you also use a the same license on your own product. The same licence therefore applies to the following transcript.

Transcript

Speaker 1
Canungra you caught the Lukic get tickin Chicago Allouche . Welcome to Langley the podcast has constructed languages and the people who create them are George Corley .
Speaker 2
With this in the Netherlands we have storefronts here too thats quite a novel . And over in Calgary Calgary That's right .
Speaker 3
Yes in the Alberta Province of Canada we have a new guy on the show Joe Windsor . Yes yes that is .
Speaker 4
That is ed I read your your paper about playing on speakers . By the way it's a very interesting interesting . Excellent . Yeah actually because my dissertation happens to be about L2 stress production . I decided to sort of add that to my notes . I might have . It's not that relevant but I might have a little footnote that that cites you . So I look forward to seeing this . Yeah well we can see I can get my dissertation done . That's always the challenge isn't it . Yes .
Speaker 5
Yeah . And and if it passes with all the .
Speaker 4
Silly errors that I've made and have to work around .
Speaker 6
But anyway going to find errors for the rest of your life every time you open it .
Speaker 4
Trust me yeah . But anyway so today we brought Joey on here because Joe you had an interesting talk at the last LCAC talking specifically about using the Cherian contrastive hierarchy . That's right . When you're when you're constructing the phonology for a language . And I just thought I wanted to bring you on in order to in order to get at the idea of using linguistic theories for conlang it's not like we don't do cirE stuff on this show . We do . We usually stick to like bedrock solid like the most non-controversial theories like most linguists use phonemes . Most linguists like break down the morphology of words in some way . Most most linguists really do have some sort of constituents in syntax but we don't you know go into you know the way that people build up theories of syntaxes go goes way in different directions . And that's part of the reason for doing that is that linguistic theory is like so many different schools that often don't talk to each other enough in my opinion . And you know it's easy not to get that . It's easier to avoid fights but also you can't really make a conlang without at least like intentionally or unintentionally making some sort of theoretical assumptions that like OK phonemes exist or or it makes sense to to put the affixes in this order or something like that .
Speaker 5
So it can be useful .
Speaker 4
I've always thought that looking at real languages and looking at type ology is a little bit more useful for for linguists but theory can can help as well . And by the way we also have Christoph also who's like maybe Joey is maybe more on the side of using theory maybe in the middle . Christoph it seems is going to be disagreeing a little but not so much disagreeing .
Speaker 7
Bringing a counterpoint from somebody who is not formally educated in rustics right and still thinks that he can do coding rather right now . So many others who are not linguists and so bringing by my own point of view on that's which is not necessary that linguistic theory is not useful . That's a we are not linguists we don't intend to be so let's use it as a tool that is useful for us but not to get too obsessed by it .
Speaker 4
And we're going to have a little bit of a discussion about this . And and and Christophe you you bring up a very good point . I've seen your work on your languages and you do very quality work without being a trained professional linguist and most conlang years including a lot of very good con Langer's are not really professional linguists . We're just sort of saying OK what's useful for conlang yours what's easier for conlang yours to like understand and use even if they don't have it based training and we're going to be talking about like like how do you use like a theoretical framework and you know what are the good points and bad points you need to be looking for . So I've talked enough . And Joe I think I want you to sort of talk a little bit about this topic and maybe just talk about starting from the beginning and like why we don't use linguistic theory .
Speaker 8
Sure . So for me to Christophe's point you absolutely do not need to be a theoretical linguist to conlang .
Speaker 9
When I started trying to make a few of my own conlang I didn't necessarily inform them my theory . And with the exception of one of them I don't look back on the phonology I created and hate it . I've gone back and I've tried to apply the theory to some of my Culham pre theoretical conlang and I can't . That doesn't mean that it's necessarily an unnatural phonetic inventory or something like that . But you know I got by without using the theory here .
Speaker 4
I mean we will probably be focusing mostly on phonological theory here because myself and both specialize in phonology . So have you .
Speaker 6
I'm happy to talk about morphology or syntax as well .
Speaker 4
Semantics you might lose me a little bit but you have to draw the line somewhere . Yeah .
Speaker 6
So anyway like you said at the LCAC in Calgary last summer I did a presentation on using dresser's cost of higher Kaid for phonological theory . Specifically segmental or you know individual sounds and's .
Speaker 9
I've started religiously using this when I'm conlang . I just finished two phonetic inventories this morning where I worked out before contrastive hierarchy and everything . And the reason I do that is because by strictly following theory I know I'm going to end up with at least a semi natural phonetic inventory and the second point to that is some of the work for me is already done . Once I start with this theory so I know if some of my consonants have this lateral plus minus lateral feature then I know I can use lateral in phonological rules to add some flavor to my conlang later on . And I know that not every single consonant is going to alternate with plus or minus lateral because you know maybe a majority of them don't care about lateral ness . So using the theory can can help me later on by already making a few choices for me . One of the things you talked about is how accessible are some of these theories to not formally trained in linguistics con langurs . I think the contrast of hierarchy is accessible . I think threshers book is fantastic . He's got some video lectures on the Internet that are searchable and you can really get into it . On the other hand one of the more advanced syntactic theories I use in my own research is called phase theory . And I don't think I'd recommend it for any not formally trained Kung Langer again because I use face theory some of my decisions are already made for me because I start by drawing the syntactic tree and if something's in a lower phase then it can't move to the top of the tree so I don't get topical ization of indirect objects and some of my languages are or things like that . But it's a very inaccessible theory . If you don't have the formal training so I wouldn't expect anyone to give it a second thought but for me there's there's two real reasons to use theory . One if you do the work up front some of your decisions later on can be easier or made for you because of your earlier choices too . If you want to make a natural language following theory will probably lead to that . If you want to make an unnatural language violating theoretical constraints will probably help you in that direction .
Speaker 4
Right . And I just want to say I do agree that the the the contrastive hierarchy is definitely something that a lot of people will understand and be able to use the only sort of prerequisite is you need to have some knowledge of some kind of distinctive feature theory . It doesn't necessarily have to be one feature set or another . It can be used with any of them . But yeah it's . Definitely it's just a way of once you have the idea of distinctive features you're choosing what distinctive features are active in the phonology . That's exactly it which is absolutely something useful for you to decide ahead of time . On the other hand what you were just talking about there is the theory . You're using the theory and your . This is making some things easier for you . It's also imposing your limitations on you . Yeah absolutely correct . Yeah . Once he once you figured out your contrastive hierarchy which obviously you could revise it later then this is like making this . This is the way the language is at least at some particular stage and that's constraining what you do later . Now that's not necessarily a bad thing but you know it's something you have to be aware of and you have to be aware of . With all these theories of like what is it intending to describe and what can you make from what you make from it . And you know what's what's acceptable for you in terms of limitations .
Speaker 6
You're absolutely right . I think one of the topics I ended up getting into at the LCAC last year was historical change .
Speaker 9
And I said I had this proto language that had certain features and then I decided to invert some of the positions of those features in the tree . And that made some daughter languages for me or if I use some as under specification if I didn't use every single feature that you'll find in just about any introductory linguistics textbook then when I ended up with phenology they really didn't like I said wow I don't need by layby all trills the combination of the features I have in this language don't necessarily have to be by label tril they could just be ejective stop still by label still non Sonor and still non syllabic way I go . So that's that helped me revise things too . You know after I got through it I really don't like what I did . I can go back I can revise things and still have the same basic structure of the phonology that I had created originally .
Speaker 7
Right . Basically that sounds like something that I would have a problem with . Although I'm concerned with the constraints that you create yourself if not in English like me you try to apply a theory that you don't completely understand the the consequences off like just say this so that's always the issue I have with especially with linguistics which is still a rather young science .
Speaker 10
And though when you read about Devizes disagreements in the field it's getting quite difficult to know what you can use and what you can use what is currently considered accepted might . In five years beat's or that it was on or completed garbage . So it's you as a non linguist . I'm always concerned about . What am I going to use and am I really understanding what I'm doing . Yeah and .
Speaker 8
I think that's a really good point Christoph . What I would say to that is if if your use of theory least there's nothing hanging on it you're not publishing a paper that says this is the way that the mind works of a Somali speaker or something like that it doesn't really matter if you use an outdated theory theories or self-contained units .
Speaker 9
And as long as you follow some theory that you're comfortable with it's going to lead to a certain amount of structure in your conlang and I'll preface everything I'm going to say with .
Speaker 4
If that's your goal in your online right and all all linguistic theories are intended to describe real linguistic data even the ones where you know a lot of people a lot of people express doubts about the those theories that are based a lot on introspection and native intuition because the data can be a little bit less lower quality on those . Even that it's still real data . You can question about whether there might be some some intrinsic biases in that and and some . And whether that's low quality or or some of the conclusions but it's a serious describes a real language so most theories applied with a little bit of care will get you a realistic language they might bias you in certain directions but they will . However you do have to watch out and when you're reading about these theories pay attention to places where you know they're trying to describe real languages . But there's two two possible ways where a theory can go wrong and where most of them do go wrong at least a little bit is either they under generate which means that they can't describe certain things that actually do occur in real languages which is not necessarily Fatum fatal but it's it's biasing you away from certain structures that you might otherwise have gone to . And then there are ones that over generate and create weird sayings that never never actually happen . Now we're all familiar with it too . And you know you can find a most almost any linguistic feature you can think of but there are sometimes things that we don't find . And you know one theory that ten get into a problem with over generation if you're doing it to naively is optimality theory . I'm going to tell a little story I went to a conference once where some people were trying to apply optimality theory to stress systems . And they came up with a bunch of universal constraints if you don't know optimality theory the basic idea is you're generating a bunch of different possible forms at once and then you're checking about them against ranked constraints and basically the one that that is the least bad . Basically that that violates the lower ranking constraints and has the least number of violations is going to win . Right . So you're ranking your constraints . What they did is they had all their constraints and they just ran it through a computer to do all all possible constraint rankings on their constraints and what they ended up with is they started getting odd things and then they started to add orchestrates to try to deal with things that it made it worse than you got weird things like where how the stress pattern works is dependent on the weather . The syllables have an odd or an even number which is not a thing that happens in natural languages . So it was . So . Sometimes you have to be a little bit careful and kind of think about like is what I'm getting from this theory something that could actually happen . And . Again like if you go into OT like they know about these problems and there are people working on OT who are trying to address those problems so when you start reading about it you will find people saying OK we've had this criticism here and we've had this problem with our theory here and talk about like you know what things that it makes that they don't find in natural languages and ways that they're trying to to get rid of the problem right .
Speaker 11
Yeah I used to teach optimality theory and when I was a graduate assistants for the phonology classes and I'd start by getting my students to make coffee we have constraints like Star better and don't depend the size cream and don't depend the size sugar and star black things like that and we end up getting the coffee that they liked by reorganizing the constraints and that example was to show them how powerful optimality theory is because you literally can't just make up new constraints .
Speaker 12
But that doesn't make them good constraints . And when I used to say was a good constraint is something that can be found to be violated in some language can be found to not be violated in some other language or be a higher ranked constraint like you . You've raised it .
Speaker 9
And something that is phonologically plausible . When we have a wealth of history even as a young science you know we've been doing phonology for 50 60 years at least quite seriously and the observations from S P E or or element theory or contrastive theory of phonology or or .
Speaker 8
Feature geometry or whatever you want to do there are still valid observations and optimality theories shouldn't be reinventing the wheel . It should build on some of those observations in that tradition that we've always had .
Speaker 10
So I mean you're going to address Crystal . I completely agree with that . And I've heard about the Optimality theory and also complains about how it can be too powerful if it's applied to natively .
Speaker 7
On the other hand that's also one thing that's interesting that's as I've always been fascinated by the limits of naturalism and the how far you can go and a year and a half in my list of books that I still have to read I have two books about court and the court rethinking universe and another called Herrera and Rory Seima about the Raiva stuff that is found on the languages that we know of which is by the way just a small sample compared to all the languages that have possibly existed in the history of mankind or humankind that you want to call it .
Speaker 10
So what I want to get there that's that's a lot of stuff that we don't know . We don't know that there may be some things that we think are impossible but are actually happening to language that's unfortunately what the last in the midst of time all just happened in the language that we have not discovered yet . So I would be especially if you're like me interesting in making language that taught .
Speaker 7
No just run them run of the mill not too artistic but try to push the envelope a bit . Try to not to be careful to let yourself be . How could I say that . Let's just have go away from a promising structure or something just because the majority of people say that it's impossible .
Speaker 9
Yeah and that's a really good point thanks .
Speaker 8
So what I was going to say was this actually came up in my dissertation and actually in my dissertation defense I had made a statement that one of the predicted structures of wisdom to be found in a natural language and my Examiner said why not and I stepped back and said well I guess there's absolutely no linguistic principle that prohibits this structure .
Speaker 13
I don't think it's a tested and I would be very surprised to find it but there's no principle that prohibits it .
Speaker 14
So maybe Christophe is right and in a language that became extinct before we started cataloguing things are one we haven't encountered yet .
Speaker 9
Maybe that one particular structure would exist and it would surprise me . But just you know their theory has its limitations which as George was saying may or may not align with the tested structures . Yeah and I mean you you don't need to be too like overly careful about because I mean if you're not creating something that's not totally attested then then you're not even making a conlang you're going to be making like .
Speaker 5
Like you're just going to be using some other language . Right . There's For a start the key there's you know an interaction of morphological features with the direct inverse I have inflecting prepositions I have Dravidians style gender .
Speaker 4
I don't know of a language that has all of those features but I can make them work together . And I don't really see any reason why a language couldn't arise that has that particular collection of features working together the way that I have them so obviously just because like some particular combination is not attested doesn't mean it's not possible . Still you mean you have to be aware of like what things are not attested but plausible . And then what things are not attested and much less likely to happen . Things like the the you know like why I was talking about . Like even or odd syllables affecting the stress pattern that really you know we don't find any language that does that and the languages we do know they just really seed was signs stress by counting from one end of the word or the other . So it doesn't seem like the like the number of syllables shouldn't matter right .
Speaker 8
No phonology counted it famously can't can't .
Speaker 4
Yeah . Or . Well people say that it can count too it can count to two or maybe count to three but it doesn't count . Like all the syllables of the word that way definitely .
Speaker 7
No I haven't had to count so far . Oh okay . I not know you at least in the in the stress patterns it counts to fall and that and I did that on purpose . Yeah just said that counting to three . Okay . I come to four and I'll see if I can handle it or not .
Speaker 15
No . So is it . So where's the stress put .
Speaker 7
Any of the . It can be any of the last four syllables of the word . Okay . Yeah . And if it's partially it's a it's how to say that it's partially semantic so it depends on the word . This one constraint it can never appear on the first syllable the word okay unless the word has a single syllable then then that's easy but it can get to Pyon the first syllable and it can be on the end of the last four . Okay and if you add suffixes and clitics you get all kind of interesting phenomena happening . Yeah . Which . But it's it's rather simple it's just that instead of having up to this to the last I went to the fourth . So far it's works fine .
Speaker 4
And there there what you did is you took basically the theoretical assumption that we always have in phonology of . And it depends on like how you're constructing your theory whether you're saying you can count to two or we can count to three . But you're just taking that and taking one assumption and you're knowingly violating that . I don't see a problem with that if that's what you want to do because I mean I mean that's how we've got going on that . Yeah . Joey Well you just wrote a paper on how really freaking weird the cling on stress system is and it is really freaking weird . I will put that paper in the show notes because it's if you know anything about like weight sensitive stress it doesn't make any sense .
Speaker 8
Yeah I got to chat with Mark Okrand about it and I said you know I have this theory of how clean water works and you know now I have a chance to ask you I'm going to ask you and I explain what it was and he said so what you're saying is the system I came up with breaks your theory . Yes . He's a good . Yeah I mean it was intentionally a non-human unnatural by a human centric point of view conlang and he achieved it by breaking rules .
Speaker 4
But what you proved is that at least English speakers can acquire it .
Speaker 16
It is a learnable system it's just so I it barely .
Speaker 4
So it's it's it's very like it's totally on a test and it's very weird and you know we can we can we can go about like I have a hypothesis about it that we could test possibly but that's an interesting thing that that and it was deliberately weird . Everything about hanging on is deliberately weird . The the the funny men Vittori doesn't make sense to have an ologist the syllable structure doesn't make any sense . So yeah it's it's intended to be that way and if you're intending to do it that way then that's another way that you can even be useful if you want to make something that's intentionally not human then you have to figure out like what do we know human languages do OK . So any notes before I kind of want to get into like examples of things that I've done we've talked about Joe you've talked about ways that you've applied theory . Anybody have other things you want to talk about before I get into that . I'm happy to move topics . Yeah . OK . So I just wanted to talk about a couple of examples . First of all I did . I kind of did it post facto just to like confirm that it works . First start to key is to do the contrastive hierarchy just for files to see if I could make my vowel harmony work in there . Right . And it's very hard to go post hoc with that system . Yeah it is not impossible but very hard .
Speaker 15
I did actually end up modifying things a little bit and like the key thing is that I made . I was making the off oul transparent to back Harmonie but it does trigger lowering the height Harmonie so it makes things it lowers things so you have a low specification higher in your tree in the back specification .
Speaker 4
And yeah branches quarantined it worked off and it worked out that it made sense . And I could make . Basically it's not back yet . And in terms of advancement it's nothing but it is low so it can lower things . Or maybe it's not even specified for height but if it's not specified for height then the next thing is automatically lower right because the lower vowels are not actually low . They're their mid vowels . But anyway this gets complicated .
Speaker 6
I think I was making some notes for this show and one of the things I put in there was where is it here . If you're if you're not a trained linguist a sentence like debacle ization of the Kornel spirant results in a glottal consonant which may be subsequently deleted if it does not carry a secondary palatal place feature is an unapproachable sentence and we just demonstrated exactly that .
Speaker 8
Yeah you get lost in her own vocabulary here .
Speaker 7
Yeah let's let's just say that I read linguistic linguistic books and even that call . Yeah no it's racist . Like some of the stuff that I had to read when I was doing through Dynamix .
Speaker 4
Well but I can say but like I could translate that a little bit . So like DBI equalization . That just means loss of oral articulation . You know that's that's like t going to h h or to . Yeah .
Speaker 17
So the coroner experienced that could be any of swarth or not . She can be Kornel . Yeah . So it's some it's some fricative . That's that's that's articulated with the front of your tongue . And yet so DBI equalization and that becomes a global consonant which is probably it's probably an H .
Speaker 14
This actually describes something in Irish where you get the cunning . It has a secondary place articulation in which case it's a powerful .
Speaker 4
Yeah . Yeah . Which you know once you know about all those words mean that makes total sense that that's what it means . So and that's just like terminology that has been used for a long time . So as you know you just have to get sort of used to different things . But .
Speaker 15
But I use that resharing tree and I made made you know the the the the vowel harmony I had worked with that now . And you know that's useful .
Speaker 4
I was just showing what features can be active and it worked out . The other thing I did was I wanted weight sensitive stress and this is a thing about like the question of whether you need a theory if you have completely regular stress on a particular syllable like you're putting always final stress like French or always penultimate or always initial always pen . That's rare but it can happen . Then you don't need to worry about any metrical theory because it's you just you just say you're assigning stress to whatever syllable right .
Speaker 8
Well then you have to ask yourself what is a word like . Are you assigning pen initial air or penultimate stress based on your Stemm or do you assign stress .
Speaker 18
After all of your affixes have been added or before all of your affixes have been added to prefixes and suffixes pattern separately . We see that in Algonquian languages for example where prefixes are part of a different word but suffixes are part of the same phonological words right . All kinds of stuff you can still use theory for there .
Speaker 19
Yeah definitely .
Speaker 15
I'm just can't is just saying that the basic part of it . You don't necessarily need to . There are still some choices to be made once you've decided that . I'm just getting into it too . If you want to do weight sensitive stress though then it can help to use a metrical grip which a metrical grid . Basically all you're doing is you're saying based on criteria you have several stages in your Métro grid and you're saying OK at this point I'm going to add a mark to these particular syllables and it can be based on position that it can be based based on civil weight which is how you get though the weight sensitivity and as you build up then then you get to the the the syllable that's going to be stressed . I'm just working at the word level for that but it can even build up into other bigger prosodic domains .
Speaker 14
So I was just going to just say I think you're right the metrical grid is probably the most accessible theory for people with no formal linguistic training . If people looked up how to do metrical grid theory it's probably going to make a lot more sense than say optimality theory or the other Zadock .
Speaker 4
Optimality theory you have to actually know a lot of other linguistics in order to actually apply it . Yeah Métro good theory you can use just on its own . Now there's different Métro theories around . Some of them do like domain boundaries and some of them don't . So it's you can sort of pick the one that works the best for you . But basically what I ended up doing I ended up with a pretty simple material grid that basically ends up assigning . If nothing if it automatically assigns stress to the penultimate syllable unless the final syllable is heavy and then the final syllable gets stress and then it also computes where secondary stress will appear . It's very useful if you need secondary stress . So that worked out and I was also able to implement that in phonics using the scalar features and phonics . I was able to basically make the computer do the metrical grid calculation itself and then I could automatically assign stress to my words and do stress based phonological rules so you can even use a theory and another tool in order to like make things work out right . Did that make sense it did to me .
Speaker 10
It make sense to cross off things so it's . I'm just concerned about it there in this kind of theory and the how to say that and the assumption that people actually make these kinds of calculations when they're speaking because I get the idea that it's a whole complicated theory to describe something that's well I'm not that people speaking a language actually do these kind of definitions when they speak . That's not how things happen .
Speaker 8
Yeah we're talking about whether or not you actually have this kind of online processing . It's what we call it . And the theory is you have this present hierarchy so you have your actual sound segments and then get developed or divided into possibly timing units and then syllables and then metrical feed and then words and phrases and international phrases and however many levels you happen to have in your particular theory . And what tends to happen is one syllable within a metrical foot boat will gain prominence and one foot within a phonological word will gain prominence and one word within a phonological phrase will gain prominence .
Speaker 13
And what this grid does it just tracks the prominences from top to bottom or bottom to top depending on your language .
Speaker 9
And so whichever one happens to wind up where you have five levels of prominence all over the exact same syllable that one is your primary stress . And some of them might not gain prominence from every single level but they might have one or two levels of prominence and so these are secondary or tertiary stresses . And we we know that this matters to people because if you put emphasis on the wrong level things can get confusing Yeah .
Speaker 4
Now you do raise a good point and though in that is this internally what people are doing on every single word at the phrase level we probably are processing these things so that we end up like when we are accenting one word in a sentence we end up accenting the stressed syllable of that word . The primary stress on that word . At the word level so metrical grid theories .
Speaker 17
Have to contend with the fact that they have to make exceptions for languages like English and like Russian that have lexically determined stress that has to be remembered .
Speaker 4
So and that's getting into like you know what does a theory do and what can the theory not do . That's that's actually attested . So if you want lexical stress you can still use a grid theory but you have to read into it and understand like how are they handling these cases where you have . You have to remember the stress pattern on an individual word or divide words into classes and remember what stress class the word belongs to . So there is definitely .
Speaker 20
Something you have to deal with . Are you hearing that . Are you hearing the baby . No
Speaker 4
I'm not . No I'm not . OK . You should be fine . Yeah . So and that's that's that's a thing you have to think about with any theory is is sometimes they run into places where they don't necessarily describe all languages and usually authors will be aware of that and will tell you that especially you know try to find like more recent explanations . So but it's it's a useful thing to get to that stress assignment .
Speaker 17
I actually do have doubts about it because of other stuff from language acquisition that seems to indicate that the bar for the bar that ends up requiring a like an infant acquiring language to start memorizing the words is before you really get out of completely regular stressed languages even a completely regular stressed language . It seems like if there's something that causes them so like Polish has regular penultimate stress but you can have at the end of the word a content word that that is one syllable and the idea is that maybe those children end up having to actually memorize stress just because at the end of an endurance they can't always have a regular rule .
Speaker 4
So there's the theory the Metro grids are useful they're not necessarily 100 percent the only way that stress gets assigned because we have these counterexamples to work with and it doesn't really .
Speaker 17
It's not you need to be aware of that but it's not necessarily a fatal thing knowing that but even knowing also this this other theory about acquisition can also tell you what languages are likely to have exceptions because one thing about Polish is there are lexical exceptions from loanwords that have a different stress pattern . And that's one of the clues that people that Polish people acquire the language and memory start memorizing stress on words . Whereas French you don't get exceptions . It's always final stress .
Speaker 10
And actually you notice that because if you talk about stress much French people who are not linguistically minded won't even know what you're talking about you know they don't hear it right .
Speaker 4
Yeah that's called stress deafness . They don't they they they cannot even they don't actually hear it because it's completely regular . They don't store it they just automatically apply their final stress rule to every word for precisely .
Speaker 7
And if you speak French with a different special trespass and they will hear that it's we get what they want . We'll never be able to explain you know what's what's wrong with the way you speak . And yeah and then you have french people trying to learn other languages that do have stress especially English and then you have these very specific way of speaking that you just heard again a few days ago a French person trying to speak English and it sounds quite like that . And I can't even do it right where everything is and stress at the end of of nouns of phrases . It's a very weird way of it sounds very weird right .
Speaker 17
So I think the bottom line on the discussion we just had the Metro reds are useful and I use them and they gave me what I wanted to do on .
Speaker 4
So it's not that a particular theory is not useful if it doesn't do everything that that you might wish for it . On the other hand like if there's something else you want to do that's a little bit outside the theory or falls onto her on on into an exception to the theory . Well if there's other natural languages that do that then you're going to know that like totally faithfully applying the Metropol grid is not necessarily going to give you all of the information . So when you're working with linguistic theory and conlang based on linguistic theory it's all about finding the tool that works for you to do what you want to do and also knowing that none of these frameworks is perfect . So . So you have to you have to be able to like hold those two ideas at the same time . I didn't need to worry about any like lexical stress or anything for what I was doing . I just needed to have something that would do weight sensitive stress the way I wanted it . So I used the metrical route .
Speaker 10
And completely agree that it's actually what I wrote in the notes about my my own opinion and that's linguistic theories they are there . They're fine for linguists for Condong as they are on their own they are useless tools that they provide that are useful . Excellent point . Yeah and if the tools are useful . By all means use them but do not focus too much on the linguistic theory except to know what's in the Rancho funny that the author tour you're using right . So I would love for you to have something like a linguistic primer for cloning as of all the main tools in every area of combining with its phonology morphology syntax program everything tools that are accessible and use for how to apply them where they apply when they don't apply and just search like a basically I could do books for calling as off of stuff that can be useful some day when I am a little bit less swamped .
Speaker 4
I have . I have had the idea of of actually doing some something like that for Finau just for phonology . Joey maybe you and me could work together on this . Maybe just even do a series of Lengua articles doing different we could do the contrastive hierarchy we could do that the metrical grid we could do some feature geometry and it's an awesome idea . Yeah I'd like to lay it out as . Okay this is what this theory is good at . These are cases where where it falls short . These are cases where it makes weird things and then sort of like just lay out so that con lawyers can look at this and say OK whether this particular tool is useful for me or not for the kind of language I want to do .
Speaker 8
Another thing is you actually have to do every step of the theory . So I'm like David Peterson kind of famously doesn't like morphemes and this traces back to morphology classy head and I think probably his master's degree . And if I had to guess based on who that class was with the was probably in the theory of word and paradigm morphology with what he personally what I use as well . Yeah it's the theory of morphology I ascribe to . But I don't go through and lay my lexicon and my Radikal on it look at the interface between lexicon and Radda con which hopefully means nothing to most people . But I know that I don't need a separate morpheme for plural and the separate morpheme for definite and a separate morpheme for animate and a separate morpheme for stationary and a separate morpheme for whatever else I want to do case or something like that I can have a morpheme that is definite plural nominative right and not something that's word and paradigm morphology can get used to and you can also fill in gaps like or you can purposely leave gaps like plural second person in English . You know some of us have yours are you all but for most people it's you and it's indistinguishable from the first person and then a new paradigm where college can get you that natural morphology without having to go off the deep end and have a noun that suddenly has 16 affixes on it .
Speaker 19
And that's what you want .
Speaker 4
And it's all about what you want . And this also gets into like series will bias you in a certain direction . So you kind of want to have an idea of what kind of language do I want before I like that theoretical tool to cause some people some people want like a gloomy loop mating language that has a lot of different things and they can just use morphemes just without without worrying about that . The objection that I saw from David I mean he . My my my daughter just opened the door for a second and you guest on the program . But that talking to people OK . Yeah she's she's she's she's she's learning English and hopefully a little bit Chinese now . OK . What I was saying OK . What I was saying is David's objection from what I understand yes he does believe that morphemes don't exist and if he's arguing the like the actual merits of the theory . But but it seems like . His main theory idea is that . Adhering too much to morphemes is going to bias conlang yours into making the kind of language that you were just talking about where it's highly agglutinating and like each morpheme has a very like particular strict meaning and each affix has a particular meaning . And you get all these very long words . If you want to do that then you don't have to worry about it but if you would prefer to have some you know multiple exponents and stuff you can do that in morpheme theory it might be easier to do it in word and paradigm . So . So at the same time word and paradigm is all about inflection and it seems like I'm not sure what it really says about derivation .
Speaker 21
And I don't remember it drawing an actual distinction between inflection and derivation which has been an ongoing theoretical battle for at least 40 years . Now whether or not inflection and derivation as separate categories and morphology is something we impose as linguists or if it's actually inherent in the system . And I don't have a good answer for you I have friends that are working on that problem but .
Speaker 4
Not one .
Speaker 7
Yeah . And that's going back to my first comment on that linguistics even though it's been going on for quite a while . He's a very young science and there's still a lot of stuff that we simply don't know and we need to be aware of that and especially people who want to look to use a theoretical framework needs to need to know .
Speaker 8
That the notation we have is is limited then that those frameworks have limitations because that's yeah I mean so one of the things I try to drill into to my students heads constantly is in linguistics we don't prove things we offer evidence in favor or against things because until the technology exists where you can actually stick an electrode into someone's brain and say oh look there's a phoneme .
Speaker 9
Oh I see what you're doing to that phoneme . We don't know what the phoneme is doing or what the actual phoneme is .
Speaker 8
We find cases in phonology where you have this really really regular pattern and you go OK this must be the phoneme but it's actually the exception that's the phoneme in the regular pattern is computed by majority rule . And all we can do is we can we can offer evidence in support of a viewpoint but we can't actually prove things the way that biology can slice something open and say hey this thing actually does have an internal skeleton or this thing does have a spleen or you know gravitational waves are real because we can capture them and we can measure them by seeing how light bends over four kilometer stretch we can do that nearly as much in linguistics as you can in some of the physical sciences so our theories are always going to be revised . There's always going to be debate where people are in favour of one theoretical commitment to another and that's okay . If you want to use linguistic theory for your conlang would you want go with it . And if your next time line you want to use a different theory pick a different theory and go with it . There's nothing wrong with doing that because it really is an art form and you're the creator of it . So have fun .
Speaker 4
And again all theories are trying to describe natual data . Know whatever theory that you're doing you are working your fate . You will you'll be seeing the examples of the data that people are trying to to work on and that itself gives you inspiration . I have I had a note in here like there's infrasonic phonology there are arguments about the domains and there's just arguments even about like what to call cert domains . There's something called a clitic group that's like a fun a logical word and any clitics that are attached to it and people some people want to call it a composite group because they're like oh well in a lot of languages compounds and clitics act like they are the same domain like the two words in the compound have the same phonological behaviour as a word and it's clever and I'm like Okay well that's the debate . But look at what the data they're pointing to is is that hey you can phonologically treat a clitic the same as the two components of a compound . While I use that in a to key the the the the the the vowel harmony is confined to the final logical word . So . So the two words in the compound do not have to harmonize and a a clitic doesn't have to harmonize with the though whatever it's being attached to . And you know I even that is like a historical layer of historical compounds that also that did harmonize and stuff but that's where I got the idea was looking at this debate and say oh I can use like what they're arguing about . The last thing I kind of want to say I'm want to talk about your idea about proving things in science really you don't absolutely prove things Matthew failed to reject the hypothesis . Yeah yeah precisely the proving is a math thing in science you do evidence . But some of the more real and physical things that we do find and me since I do phonetics you know I know like the most physical thing a lot of the narrow phonetic stuff isn't that useful but knowing some of that . Will help you and maybe I could do a whole episode on this but one tidbit I'm going to say there are a lot of con ligers I see who want to be really fiddly about how they label their vowels with IPA and use like the raised and lowers and loader and stuff on the phoneme or try to do like 60 different vowels and . I understand the phonetics of vowels and from the phonetics of veils that doesn't really make a whole lot of sense if you're going for naturalism because a vowel is not a point in the vowel space the vowel is . It's like a probabilistic cloud . People people will pronounce a within this party . This general area in the vowel space with you know within this you know set of F1 and F2 values but it's not it's not like they're not going to consistently always pronounce it the same way twice . It's it's it's a statistical thing . So you can't have tons of different Vout though that distinctions because soon they're going to overlap and it's not really that important if there's no distinction between like tents and Laxon and it doesn't really matter which one you choose that much . Maybe maybe you want to say the center of that space is closer to one or the other but really you don't have to be too picky about it .
Speaker 8
So this is actually why use that contrastive hierarchy . So you one of the languages I was just working on as rounding distinctions in the consonant series but no rounding distinctions in the vowel series so I chose the phoneme would be if it was a friend . Now all the unrounded and if it was a back value be rounded as a phoneme . But when I got to do is say you know it doesn't really matter if I have . Aw . Aw . You know they're both backload vowels once round ones not round and there's no specification for roundness on that vowel so just because of this language it's not going to matter .
Speaker 18
But what I can do is say it agrees with ever whenever consonant follows if it's around the consonant use rounded valves and rounded consonant USAn unrounded vowel and nobody's going to know the difference .
Speaker 4
Are native speakers that combine right and that's one way you can . You can use a contrastive hierarchy . You could also make those different Sosi phonetic variables too like some people pronounce it as an R . Most the time some people pronounce it as an all . Most of the time there's there's a lot of things you could do with that and I like that example and contrastive Harkey does a lot of good stuff for vowels . I just want to say you're not being too picky about the vowels . You can you can not worry about being too picky with some kinds of consonants too but there there is a difference in that some consonant sacred distinctions really are category that distinction between plain and aspirated . That like there is like a line where people just like somewhere between 20 and 40 milliseconds it depends on the place of articulation and where it is in the word but somewhere between 24 milliseconds there will be a line where okay . This side it's played on the side it's it's it's aspirated so some things are pretty discrete but vowels are not vowels or not . Anyway that's all I've got to say about that because I could I could end up going all wrong and my daughter is awake so I probably should play with her . Any last things that you guys want to share . I think this has been a pretty good discussion .
Speaker 8
Kristoff anything else from from your point of view .
Speaker 10
Just an example that I always gave you that theory for none theorists is only as useful . Like I said the tools it brings and I always give the example especially in the eyes of artists like we all I think we are . Painters don't need to know or the Indians out of color theory to be able to paint . You don't need to know that it can be useful to know what to do because the color .
Speaker 7
Sorry that's some . There were some I think a police car driving by knowing about the kowtows circle and the what are complimentary and supplementary columns can be useful but knowing the ins and outs of rods and cones in the high end of the V . The entire reason why the color gamut is as it does for human beings is not going to be that's useful unless you want to do some special things for colorblind people . So as a painter it's good to have some idea of color theory but you don't need to do to tonight everything just know you need to know the tools that color theory can bring you or not . You can just project can be a great painter without knowing any color theory you mostly paint us with experience what works and what doesn't . And that in the end with enough experience and knowing nearly as much that's actually studied theory .
Speaker 22
I completely agree with you . I think the biggest thing that linguistics can contribute to conlang is George societ a number of times describing naturalistic data so we know what popular trends in human languages are . You can probably gain just as much if not more information to inform your conlang from picking up a couple of popular grammars rather than a linguistics textbook but some of the tools that linguistics can give us I think is the language of linguistics will allow us to communicate our ideas to one another better .
Speaker 14
We have the Conlang listserv and people are sending out questions and giving examples from their own current langues and things like that . And one of the posts I recently made on the list serve was I would encourage people to look up they'll be glossing rules and so you can do interlinear or glossing .
Speaker 8
If you can show me your sentence and you can show me whether you believe in more things or not the morphine by morpheme break down and what the morphemes mean I will be better able to give you feedback on your conlang . I will be better able to say I think one of the discussion points was is this really a dative or is it an allative case or does it even matter who knows .
Speaker 19
It all depends on the conflict .
Speaker 18
If if we're all speaking the same analytical language we can communicate her ideas to one another better and we can support each other a little bit better and I think that's one of the fantastic tools that linguistics irrespective of theory provides us as artists and as complainers .
Speaker 7
Yeah I completely agree on that . Learning linguistic recovery is basically as it is a given . It's not just as you said about communication but it will actually streamline your thought process and it will remove you from some of the assumptions that you may have due to the languages you already know and especially the descriptions that you've that you've got those languages that you probably didn't get from actual linguists but from from school which are quite which I of which I have a lot of unspoken assumptions . Knowing the language of linguistics helps you get rid of all that baggage and look at languages in a more word say unbiased news or .
Speaker 4
A point although not totally neutral I think David Peterson even suggests just like when you're doing morphological categories like cases just like First going with like what is it used for and figuring out like different examples of what it's used for and then he actually tells people to let like linguists later analyze and figure out what I mean . I think there's some value in learning how to describe it yourself but it makes sense to say like at first OK . I want a case that does this or I want my cases to do this this and this and this and figure out like what they do and then sort of sort of like working out like what to call them based on what's seems to be the primary function . Right . But where were this episode is all about starting with theories and we didn't do that . The thing that people leave with in this episode and then I'm going to end it is don't get hung up on what the theory is true or whether this is really what's going on psychologically . Just think about how is it useful for your conlang . OK . Like you know . So don't let linguists figure out what's true and what's not . And I can guarantee you all of the theories that we currently have are false in some way . But you know if you want to be a linguist then you can you can get into that work . But if you just because you're you don't need to worry about whether the theory is true . Just think about how it can be useful to you and how it can do it and whether it can do what you want to do with your language . And with that I'm just going to say it first of all Joey has provided some some textbooks that can get you started on mostly some some fun logical theories some different things that that will help you out . I might try to think of some war I probably linked to . Joe your talk at LCAC and to Williams talk as he talked about another using Siri using the construction grammar which is another theory that he applied to a conlang and it gave him the results he wanted .
Speaker 8
And Doug Balz Oh yeah .
Speaker 3
We can lick the ball too . Well like those and we'd be happy to have people's feedback too . Absolutely . What will what theories to use how to use them what are your goals when doing that . But with that . I am going to end the podcast so that I can I can go and be a good dad .
Speaker 23
And I'm going to say happy Conneally thank you for listening to conlang or you can find our archives and show notes that conlang or com can also follow us on Facebook Twitter and Tumblr our web spaces provided by the language creation society . Our site was designed by Bianca Richards and our theme music is by no Novis Nigeria's under the Creative Commons Attribution Non-Commercial share alike license . You are free to use our show for any non-commercial work as long as credit is provided to con re podcast and you use a similar creative commons license . Online or is supported by our listeners . Please visit patriot on dot com Logicon langar . Give me your support . Thank you .

Conlangery Podcast/Conlangery 135 Using Linguistic Theory for Conlanging (last edited 2018-02-06 07:52:40 by TranscriBot)