Idea for a new engelang

[matmcv]

Hi all

I have an idea for a new engelang, I wonder if anyone will be interested in working with me on it or if they have any useful feedback (including telling me it won't work or someone has done it before, that's fine!)

It's basically a language that would have the opposite grammar to English, as directly opposite or distantly different as possible. My current name for it is Shilgngí [ʃɪlgˈŋɪ] [SIlg'NI], although I also think of it as "Antinglish". It began with my answer to a question on Quora: https://www.quora.com/Which-language-is ... ew-McVeagh. Having given a list of what features a language grammatically most opposite to English would have, I was then asked to 'write' something in it (when it doesn't yet exist!) so I 'translated' the Babel text into a form of glossing with abbreviated terms for inflections and shifted word order. Even this doesn't include all required features (e.g. gender) but it gives the gist and made me think a lot about what would be required.

Some notes:

• The purpose is to test whether it's possible to create such a language, whether it can be made to work. Does it 'feel' opposite to English? It could be used to help English L1 speakers get out of their heads a bit, stop taking their native grammar for granted. It's not meant to be used in practice or associated with a fictional people or situation.
• It's not meant to be phonologically and lexically opposite as well, as in lots of clicks and ejectives and whatever. Possibly I could use the phonology of English, only switched around temporally (as in [ʃɪlgˈŋɪ], e.g. /ŋ/ is mostly onset, only coda before vowels), or I could just create a random one, as it's not really what the lang is testing. I could back-engineer English lexemes or generate randomly.
• Some issues do not have an obvious 'direct opposite' to the scenario in English, in which case it's a question of what is most alien, what would be most challenging. E.g. what is the 'opposite' of SVO? Some might say OVS, which is literally the reverse - but that shares the feature of "verb in the middle"! I've already decided the morphosyntactic alignment should be tripartite rather than ergative.

Matthew

[Richard W]

<snip/>
what would be most challenging. E.g. what is the 'opposite' of SVO? Some might say OVS, which is literally the reverse - but that shares the feature of "verb in the middle"

Random sequencing relying on case markings would be sufficiently alien. Perhaps you could just adopt the word order of Latin poetry - except I'm not sure it's been adequately determined. Sticking the object phrase inside the subject phrase can be fun.

[matmcv]

Indeed, someone has suggested non-configurationality would be the way to go - that the opposite of strict word order would be very loose word order, with the role of all arguments clear from inflections. Latin poetry is non-configurational. I think that might be preferable than trying to specify a verbal word order.
Thanks

[bradrn]

Some more ideas, arranged roughly from most impactful to least impactful (formatting: English → Antinglish):

1. Mostly isolating → Mostly polysynthetic
2. Subject prominent → Topic prominent (see Li and Thompson 1976)
3. No classifiers of any sort → Lots of numeral classifiers, noun classifiers, possessive classifiers etc.
4. Lots of determiners → No ‘determiner’ slot at all (see https://wals.info/chapter/88#3._Theoretical_issues)
5. Fully syntactically accusative → fully syntactically ergative
6. No serial verbs → highly productive serial verbs (though this may be incompatible with the previous point)
7. Double-marking, tending towards dependent-marking → exclusively head-marking

[matmcv]

bradrn - very useful substantial suggestions. I am definitely going to make it synthetic, not sure about polysynthetic but I'll look into what difference it will make. Topicality has also been suggested by others and I think it's an excellent idea that goes with non-configurationality. Not sure about having no determiners - I'd want to express some of the meanings of determiners (interrogative, distributive, demonstrative etc.) but I could certainly do it by inflection rather than modifier. My main point with determiners was to deconstruct the English "pre, central, post" complexity as in all the other three adjectival nouns. Reduce that to only one determiner at once.

I've got some questions about some of your suggestions:

• No classifiers of any sort → Lots of numeral classifiers, noun classifiers, possessive classifiers etc.
• Fully syntactically accusative → fully syntactically ergative
• Double-marking, tending towards dependent-marking → exclusively head-marking

(partial quote)

Numeral classifiers I have heard of, what are the other two kinds?
What are syntactically accusative and ergative?
In head and dependent marking in English, do you mean verb inflections like 3.sg. present tense, and pronoun cases? Or number in demonstratives? Ultimately we don't have much marking at all.

[bradrn]

bradrn - very useful substantial suggestions. I am definitely going to make it synthetic, not sure about polysynthetic but I'll look into what difference it will make.

If you want to read some more about polysynthesis, Whimemsz wrote an excellent series on it — it’s inaccessible at the moment, but you can still read an archived copy here.

Not sure about having no determiners - I'd want to express some of the meanings of determiners (interrogative, distributive, demonstrative etc.) but I could certainly do it by inflection rather than modifier. My main point with determiners was to deconstruct the English "pre, central, post" complexity as in all the other three adjectival nouns. Reduce that to only one determiner at once.

That isn’t quite what I meant. English has a ‘determiner’ slot in its NP, where you can put an article, a demonstrative, a possessor etc. You can say the car, my car, this car, but you can’t combine two determiners: *the my car, *the this car. But many languages don’t have a ‘determiner’ slot in this way: they certainly have articles, demonstratives and possessors, but they allow them to co-occur with each other. See WALS for more information.

Numeral classifiers I have heard of, what are the other two kinds?

Noun classifiers: they’re morphemes which (optionally or obligatorily) appear before nouns as a classifier. Sort of like a gender system, but not quite. An example from Jacaltec:

xil

saw

naj

CL:MAN

xuwan

John

no7

CL:ANIMAL

labʼa

snake

John saw the snake.

Possessive classifiers: morphemes which classify different types of possession. You may be familiar with alienable/inalienable: that’s one type of possessive classifier, but systems can get a lot more complex. For instance, in Kipeá-Karirí:

dz-upodó

1s-CL:ROASTED

do

POSS

sabuka

fowl

My (roasted) fowl

dz-ukisí

1s-CL:SHARING

do

POSS

sabuka

fowl

My (share of the) fowl

dz-ubá

1s-CL:GIFT

do

POSS

sabuka

fowl

My fown (that was given to me)

And there are other types of classifiers besides those: Aikhenvald’s Classifiers (an excellent book if you want to learn more about these things!) lists genders/noun classes, noun classifiers, numeral classifiers, (three different types of) possessive classifiers, verbal classifiers, locative classifiers and deictic classifiers. And those aren’t mutually exclusive: Palikur, for instance, has all of those except for noun classifiers.

What are syntactically accusative and ergative?

See my ergativity series for details. (You’re in luck: I just finished the section on syntactic ergativity yesterday!)

In head and dependent marking in English, do you mean verb inflections like 3.sg. present tense, and pronoun cases? Or number in demonstratives? Ultimately we don't have much marking at all.

English has head marking through person marking on verbs, and dependent marking through a (minimal) noun case system.

[matmcv]

That isn’t quite what I meant. English has a ‘determiner’ slot in its NP, where you can put an article, a demonstrative, a possessor etc. You can say the car, my car, this car, but you can’t combine two determiners: *the my car, *the this car. But many languages don’t have a ‘determiner’ slot in this way: they certainly have articles, demonstratives and possessors, but they allow them to co-occur with each other. See WALS for more information.

It's an interesting point, but I don't agree with them that the fact that they can have two determiners means they don't have a determiner slot, I think think it means they can have more than one determiner slot. After all does the fact that a noun can have more than one adjective mean that there is no adjective slot? There seems to be an underlying assumption that there can be only one determiner in a NP. I don't see why. 'Determiner' means thing that determines, there can be more than one element to that, such as "definiteness + deixis" as in the example in the WALS section. How should we analyse the French ce truc-ci or cette chose-là? The ce/cette are demonstratives but the ci/là add further demonstrative information. They are not circumfixes because you can do without the ci/là, although the reverse is not true. But in general since Quirk/Greenbaum/Leech/Svartvik created the "pre/central/post"-determiner system for English there's been the sense that English determiners are not simply single words in a NP, that there are three slots all of which can be occupied. No we can't say the this car, but we can say this very car, the same car, that other car, such a car, all the car, half the other car, and other complicated constructions such as all the king's horses (the the has to be taken with king such that the 's applies to that whole phrase; that is then the central determiner, like a personal pronoun possessive, preceded by the pre-determiner all).Thus 'determining' is not automatically a single, simple thing, but can be multi-featured, and it is not surprising if some language separates definiteness and demonstrativeness.

Anyway, this is all by way of saying that I don't think the 'opposite' of English's complicated system is no determiners, or no determiner slot, but an explicitly simple system in place of a complex system. But thanks for pointing the issue out, it's yet another assumption overturned that demonstratives are necessarily definite.

Noun classifiers: they’re morphemes which (optionally or obligatorily) appear before nouns as a classifier. Sort of like a gender system, but not quite.

I will definitely be using noun classes, and more than just in terms of sex-based genders; I was going to prefix the markers as in Bantu languages. Freeing the class morphemes and making them separate words had not occurred to me; but I want to reverse the English tendency of expressing grammatical relations syntactically and lexical relations morphologically, thus complicating a noun by extra inflection. Not only that, but I shall make them irregular, inconsistent, and not entirely rational. Incidentally we do have one noun class distinction in English nouns: count vs. non-count. So I'll have to get rid of that of course, which won't be hard as there won't be plurals or the indefinite article.

Possessive classifiers: morphemes which classify different types of possession. You may be familiar with alienable/inalienable: that’s one type of possessive classifier, but systems can get a lot more complex. For instance, in Kipeá-Karirí:

dz-upodó

1s-CL:ROASTED

do

POSS

sabuka

fowl

My (roasted) fowl

dz-ukisí

1s-CL:SHARING

do

POSS

sabuka

fowl

My (share of the) fowl

dz-ubá

1s-CL:GIFT

do

POSS

sabuka

fowl

My fown (that was given to me)

It's interesting; do we have to analyse this as possessive classifiers though? To me the possessive is split between "do" which indicates general possessiveness and "dz" which specifies the person and number of the possessor. The morphemes following "dz" indicate a circumstance of possession, but it's like a distinct factor, like an adjective after "my", as in the English glosses. I guess the quality is wrapped up with the person-indicator, and in the case of 'share of the' that actually affects the nature of the possession, whereas 'roasted' is surely just a quality of the fowl. It would be different also if the range and applicability of the classes was different for the different persons/numbers, and thus depended on them: that would indicate more that the classes were factors associated with the possession rather than just factors associated with the possessed thing and morphologically mixed up with the possession.

See my ergativity series for details. (You’re in luck: I just finished the section on syntactic ergativity yesterday!)

I also Googled this (Polinsky 2017) and yes, it's yet another twist in the tale of morphosyntactic alignment! I've already decided to go with tripartite alignment rather than ergative, so I don't think the language would have the same issues with ellipsis in a conjoined clause. "Mother(A) saw father(O) and ∅(S) returned" wouldn't work; unless I allowed it with ambiguity, perhaps with topic-indication, salience, obviation or simply contextual clueing.

Incidentally, why can't some ergative languages allow relativising the A? English and many accusative languages can relativise the O. That restriction seems to be on top of an ergative alignment per se. Also, in what sense is English "fully syntactically accusative", rather than just syntactically accusative?

[bradrn]

That isn’t quite what I meant. English has a ‘determiner’ slot in its NP, where you can put an article, a demonstrative, a possessor etc. You can say the car, my car, this car, but you can’t combine two determiners: *the my car, *the this car. But many languages don’t have a ‘determiner’ slot in this way: they certainly have articles, demonstratives and possessors, but they allow them to co-occur with each other. See WALS for more information.

It's an interesting point, but I don't agree with them that the fact that they can have two determiners means they don't have a determiner slot, I think think it means they can have more than one determiner slot. After all does the fact that a noun can have more than one adjective mean that there is no adjective slot? There seems to be an underlying assumption that there can be only one determiner in a NP. I don't see why. 'Determiner' means thing that determines, there can be more than one element to that, such as "definiteness + deixis" as in the example in the WALS section. How should we analyse the French ce truc-ci or cette chose-là? The ce/cette are demonstratives but the ci/là add further demonstrative information. They are not circumfixes because you can do without the ci/là, although the reverse is not true. But in general since Quirk/Greenbaum/Leech/Svartvik created the "pre/central/post"-determiner system for English there's been the sense that English determiners are not simply single words in a NP, that there are three slots all of which can be occupied. No we can't say the this car, but we can say this very car, the same car, that other car, such a car, all the car, half the other car, and other complicated constructions such as all the king's horses (the the has to be taken with king such that the 's applies to that whole phrase; that is then the central determiner, like a personal pronoun possessive, preceded by the pre-determiner all).Thus 'determining' is not automatically a single, simple thing, but can be multi-featured, and it is not surprising if some language separates definiteness and demonstrativeness.

Interesting perspective… I think the issue here is that you define ‘determiner’ semantically, whereas I define it syntactically (as ‘a modifier that can go in the determiner slot and can’t co-occur with another determiner’). Thus from your perspective, it’s perfectly natural for a language to have more than one determiner, whereas from my perspective (and WALS’s) that means that the language in question must not have determiners!

Anyway, this is all by way of saying that I don't think the 'opposite' of English's complicated system is no determiners, or no determiner slot, but an explicitly simple system in place of a complex system. But thanks for pointing the issue out, it's yet another assumption overturned that demonstratives are necessarily definite.

I wouldn’t go that far! By their semantics, demonstratives are always definite; it just so happens that some languages allow definite markers and demonstratives to co-occur. In some of them, the article is obligatory, while in others it’s optional; in those languages, it appears from my limited reading that the presence/absence of the article makes no semantic difference, although I would be very hesitant to say that conclusively without any corroboration.

(Though interestingly, what you say here does apply to possessives: in some languages, possessed NPs are not necessarily definite, and need an article to make them definite.)

Possessive classifiers: morphemes which classify different types of possession. You may be familiar with alienable/inalienable: that’s one type of possessive classifier, but systems can get a lot more complex. For instance, in Kipeá-Karirí:

dz-upodó

1s-CL:ROASTED

do

POSS

sabuka

fowl

My (roasted) fowl

dz-ukisí

1s-CL:SHARING

do

POSS

sabuka

fowl

My (share of the) fowl

dz-ubá

1s-CL:GIFT

do

POSS

sabuka

fowl

My fown (that was given to me)

It's interesting; do we have to analyse this as possessive classifiers though? To me the possessive is split between "do" which indicates general possessiveness and "dz" which specifies the person and number of the possessor. The morphemes following "dz" indicate a circumstance of possession, but it's like a distinct factor, like an adjective after "my", as in the English glosses. I guess the quality is wrapped up with the person-indicator, and in the case of 'share of the' that actually affects the nature of the possession, whereas 'roasted' is surely just a quality of the fowl. It would be different also if the range and applicability of the classes was different for the different persons/numbers, and thus depended on them: that would indicate more that the classes were factors associated with the possession rather than just factors associated with the possessed thing and morphologically mixed up with the possession.

I think you may be misinterpreting this. The possessive classifiers here are upodó, ukisí, ubá; the rest is just the remainder of the possessive construction. And we analyse these as possessive classifiers (and not anything else) because they form a closed set of only twelve items, which — as far as I can tell — are used only in possessive constructions. If you’re interested in the complete paradigm, here’s what Aikhenvald says:

Kipeá-Karirí (Rodrigues 1995; 1997) had a system of twelve classifiers characterizing the relationship between the possessor and the possessed in terms of how the possessed could be handled by the possessor. Classifiers used with food items categorize them in terms of (a) acquisition: (i) gathering of wild plants: classifier uapru; (ii) raising of animals: enki; (iii) cultivation of manioc: uanhi; (iv) cultivation of other plants: udje; and (b) preparation: (i) boiling: ude; (ii) roasting: upodo; (iii) maturation at home: ubo. Other items are also categorized in terms of (a) acquisition: (i) finding: uito; (ii) sharing: ukisi; (iii) gift from outsiders: uba; (iv) booty: boronunu, or (b) transportation—carried goods: e.

I should also add that there are at least three types of classifiers attested with possessive constructions. Some languages have classifiers which categorise the possessed noun; some languages have classifiers which categorise the relation between the possessor and the possessed; and (very rarely) a few languages have classifiers which categorise the possessor. As Aikhenvald says, the classifiers of Kipeá-Karirí characterise ‘the relationship between the possessor and the possessed in terms of how the possessed could be handled by the possessor’; thus they are an instance of a relational classifier.

See my ergativity series for details. (You’re in luck: I just finished the section on syntactic ergativity yesterday!)

I also Googled this (Polinsky 2017) and yes, it's yet another twist in the tale of morphosyntactic alignment! I've already decided to go with tripartite alignment rather than ergative, so I don't think the language would have the same issues with ellipsis in a conjoined clause. "Mother(A) saw father(O) and ∅(S) returned" wouldn't work; unless I allowed it with ambiguity, perhaps with topic-indication, salience, obviation or simply contextual clueing.

That Polinsky article is excellent — I used it as one of my main sources! (Though it doesn’t tell the whole story: Polinsky places more emphasis on Ā-movement than I would like.) But if you’re using a tripartite alignment, then I wouldn’t worry too much about syntactic ergativity.

Incidentally, why can't some ergative languages allow relativising the A? English and many accusative languages can relativise the O. That restriction seems to be on top of an ergative alignment per se.

‘Why’ questions are always tricky with this sort of thing — we just know that a surprising number of morphologically ergative languages disallow relativisation of the ergative argument. As I said in the ergativity thread I linked:

Finally, it is remarkable that the majority of morphologically ergative languages show some sort of syntactic ergativity (again discounting active-stative languages like Tibetan and Basque, which never show syntactic ergativity). For instance, using the WALS language sample, Polinsky finds only seven ergative languages which do not display syntactic ergativity. … But the WALS sample is admittedly small … Polinsky notes that ‘we do not have sufficient information … so any judgment calls about the relative frequency of certain patterns over others are necessarily preliminary’. Still, it is remarkable that — at least when it comes to relativisation — it is far easier to find morphologically ergative languages which are syntactically ergative than it is to find morphologically accusative languages which are syntactically accusative.

Also, in what sense is English "fully syntactically accusative", rather than just syntactically accusative?

You’re right — I shouldn’t have said that. As you noted above, English is syntactically neutral with respect to relativisation (and Ā-movement more generally).