Vreleksá Forum Index Vreleksá
The Alurhsa Word for Constructed: Creativity in both scripts and languages
 
 FAQFAQ   SearchSearch   MemberlistMemberlist   UsergroupsUsergroups   RegisterRegister 
 ProfileProfile   Log in to check your private messagesLog in to check your private messages   Log inLog in 

A language without lexical category distinction
Goto page 1, 2  Next
 
Post new topic   Reply to topic    Vreleksá Forum Index -> Conlangs
View previous topic :: View next topic  
Author Message
Aert



Joined: 03 Jul 2008
Posts: 354

PostPosted: Fri Oct 18, 2013 3:59 pm    Post subject: A language without lexical category distinction Reply with quote

So I've been rather busy reading for my linguistics project on Salish lexical suffixes, but I came across some very interesting articles on whether Salish uses the noun/verb distinction, or something else, or not at all.

As a result I'm in the initial planning stages for a language that does not mark any distinction of any lexical category in the root (by which I mean that all non-grammatical roots can take nominal, verbal, or adjectival meaning and corresponding syntactic position etc).

From what I've been reading, it is unlikely that any human language would completely do away with the lexical category distinction, as it would render the language impossible to learn by infants etc.

However, as per my norm, this will be a theoretical exercise to see what such a language a) would look like; b) how it would function; and c) what it could do that 'normal' human languages cannot.


This of course does not mean that no words in this language will have no overt distinction of lexical category, but rather that the roots themselves can become any category. (This is proving to be a challenge, as it can be difficult to consider some of the category shifts implied.)

Again, I will be drawing a lot of my inspiration from Salish languages, including what seems to be a very pervasive (but certainly not homogeneous) system of metaphorical extensions, particularly of body parts.
For example, the lexical suffix =cin (mouth) can be extended to mean 'food (eaten by the mouth)'; 'language (produced by the mouth);' and etc.
Others, eg. =qin (head) can be extended to mean 'top (position of head relative to body);' and 'instance of' (eg. three heads = three people).


So I'm likely going to incorporate a variation of this system into the language, but I haven't done a thorough consideration of what this could be realized as, just yet.

I'll probably post this weekend as I finally have a bit of time off for personal projects, with something more concrete.
Back to top
View user's profile Send private message
eldin raigmore
Admin


Joined: 03 May 2007
Posts: 1621
Location: SouthEast Michigan

PostPosted: Sun Oct 20, 2013 8:37 pm    Post subject: Reply with quote

While every natlang distinguishes between nouns and verbs in the surface form, not all have any lexical category distinctions in the roots. Also, whatever distinction there is needn't be as strong as the ones we're used to.
_________________
"We're the healthiest horse in the glue factory" - Erskine Bowles, Co-Chairman of the deficit reduction commission
Back to top
View user's profile Send private message
Aert



Joined: 03 Jul 2008
Posts: 354

PostPosted: Sun Oct 20, 2013 10:16 pm    Post subject: Reply with quote

Here's a shot of the initial syntactic components, expressed in a generalized (LexXP) form:
The 'n' and 'v' are categorizing morphemes which must attach to a root to form a full lexical item (that includes the category it is, verb/noun/adjective) so it can be used in speech.


Still working on a form of aspect that will cover (generalize to) both nouns (lexical aspect) and verbs (event/verbal aspect), but it's coming along!

Example of the types of semantic variation you can get with the uncategorized root (via morphosyntactic marking, possibly ablaut):

WATER -> water.NOM 'water' (nominal entity); water.STAT 'is water'; water.DYN 'flowing/running water; water flows' (dynamic: eg. water in motion; can be used as a nominal or predicative phrase).

BURN -> burn.STAT: N/A ('burning' is lexically/semantically dynamic only); burn.DYN 'burning, it burns'; burn.RES '(a) burn, (a) burnt thing'; burn.(DYN).NOM 'fire; lit: (the) burning (thing)'

Of course, many possibilities are available - it will take a while to figure out an organized system to be used on roots.
Back to top
View user's profile Send private message
eldin raigmore
Admin


Joined: 03 May 2007
Posts: 1621
Location: SouthEast Michigan

PostPosted: Mon Oct 21, 2013 9:31 pm    Post subject: Reply with quote

"Nominal aspect" is usually the label for the count-noun vs mass-or-measure-noun distinction. Count-nouns are "like" perfective nouns; mass-or-measure nouns are "like" imperfective nouns.

You could also have locational aspect.
Perfective locations would be, in English, "at" and suchlike prepositions.
Imperfective locations would be, in English, "in" and suchlike prepositions.

Cities could be either perfective or imperfective in English.
Named areas bigger than cities would always be imperfective in English.
Certain locations smaller than cities -- streets and buildings, for instance -- might always be perfective in some dialects of English, and sometimes perfective but sometimes imperfective in other dialects of English. For instance, in my 'lect of English, streets are usually perfective "on", but there are other 'lects for which streets are usually imperfective "in".

_____________________________________________________________

Not all languages have adjectives (modifiers that can modify noun-phrases) as a separate category from nouns and verbs. For some languages, anything we'd use an adjective for, can be expressed via a stative verb; for some languages, anything we'd use an adjective for, can be expressed via a noun (e.g. "has hunger" rather than "is hungry"); and some would "do" some of our adjectives by means of their verbs, and others by means of their nouns.

Even if a language has modifiers, it may not split them between adjectives (which can modify noun-phrases) and adverbs (which can modify any kind of phrase besides a noun-phrase; verbs, adjectives, other adverbs, and adpositional phrases, among others).

OTOH some languages that do distinguish adjectives from adverbs, go further and have two different categories of adverbs; one kind can modify verbs, the other kind can modify modifiers (adjectives and adverbs).

IME nearly any ad-adjectival or ad-adverbal adjective is a degree-word like "very" or "somewhat" or "totally" or "negatively". In English some of those can also be used as ad-verbal adverbs, while some -- "very" and "too", for instance -- can't modify finite verbs (though they can modify participles, which are a kind of adjective).


_____________________________________________________________


Will your language have adpositions, conjunctions, and pronouns, as separate lexical categories?
Or will you derive conjunctions and adpositions such as "and" and "with" from verb-forms of numerals such as "two"?
Or use all cases instead of any adpositions?
But then what will you do for conjunctions and pronouns?
Will you use "head" or "body" instead of reflexive pronouns? Will you use "servant" instead of first-person, and/or "your grace" instead of second-person? In other words, will you use nouns instead of pronouns? You might inflect all verbs and adpositions for person and/or number and/or gender and/or definiteness and/or etc., thus avoiding the need to mention a pronoun; and/or use proper nouns instead of pronouns, like "Eldin says Aert should give the cake to Eldin".

Another possibility is to somehow derive the following seven pronouns;
one that means "the first thing mentioned",
one that means "the second thing mentioned",
one that means "the third thing mentioned",
one that means "the last thing mentioned",
one that means "the next-to-last thing mentioned",
one that means "the third-from-last thing mentioned",
and one that means "anything else I've already mentioned (later than the third thing but before the third-from-last thing)".

In reality you could probably do with just five:
one that means "the first thing mentioned",
one that means "the second thing mentioned",
one that means "the last thing mentioned",
one that means "the next-to-last thing mentioned",
and one that means "anything else I've or you've or we've already mentioned (later than the second thing but before the next-to-last thing)".

Somebody once posted a "nounless" (actually, "almost nounless") conlang that had just four pronouns;
the last thing mentioned;
the next-to-last thing mentioned;
the main topic (probably the first thing mentioned);
and anything else.
As near as I can remember, verbs could only have pronouns as participants, adpositions could only have pronouns as objects, and adjectives could only modify pronouns; a non-pronominal full noun-phrase was never part of a clause, though of course it could be part of a sentence.
_________________
"We're the healthiest horse in the glue factory" - Erskine Bowles, Co-Chairman of the deficit reduction commission
Back to top
View user's profile Send private message
Aert



Joined: 03 Jul 2008
Posts: 354

PostPosted: Wed Oct 23, 2013 4:12 pm    Post subject: Reply with quote

Thanks for the info on perfective/imperfective for non-verbal aspects - very interesting!


As for the adjective/verb distinction - I'm not sure what I want to do about that. I'll still want to be able to have relative clauses and other phrases with modified nouns, but I'm not sure how I want to do that yet - I'll have to see what kind of structures should be used based on the design principles.

Right now I think that the only form of agreement that should exist is argument agreement, which would mean that phrases that, eg. in English would require a certain syntactic category could be rendered in any form.

Eg: possession in English requires a nominal possessee; if there is no overt marking of syntactic category, then I would assume that the possessive structure would 'force' a nominal reading. However, if there is overt marking, then you could potentially have things like 3SG.POSS boy.STAT 'his boyishness; his acting like a boy; he/his behaviour is boyish; etc.'

I'll probably use a variation of the lexical structure for grammatical marking and modifier usage from my earlier (yes, still unfinished but not forgotten) project.


The language will definitely have pronouns, at least as pronominal clitics (but probably not full stand-alone forms); conjunctions yes; and adpositions I'm not sure yet. I'm planning to use some sort of case-like system for at least some adpositions, but this is still very preliminary.

There will definitely be grammatical inflection: person/number, maybe gender; probably also human/non-human/inanimate as well (the 'standard' human pronouns, and then an animate/inanimate distinction used elsewhere).

I suppose then I'm not going for full 'nounlessness' or 'categorylessness' but rather, that roots (only) can take multiple syntactic categories. So the words that are not 'full' lexemes (content words, as opposed to grammatical/functional words) are restricted to their category, if they have one. If pronouns are going to be clitics, then that would suggest to me that they are not the same as content words (at least, in that they are bound morphemes that must occur on a verbal root), and so should not be able to be interpreted as non-nominal.


So then, the following sentences would be allowed:

happy-NOM-3SG.M 'he [has] happiness'
happy-STAT-3SG.M 'he is happy'
happy-DYN-3SG.M 'he is happy'

But each of these would have to have different interpretation, due to the different category marking. My initial thoughts are that a nominal predicate (of a non-concrete sensation/etc) would be the least 'permanent,' compared to the stative being more permanent, and most commonly used in terms of expression; and that the dynamic marking would indicate that there are different types of happiness that 'he' is experiencing, and as a metaphor to the dynamic nature of reality, that things marked as dynamic would be the most 'tangible/real/concrete' version of an abstract concept. So in effect, the experiencer is moving through different types/stages/varieties of happiness, and is closer to being 'truly happy' than with the stative-marked predicate.

I think this could also be used on non-emotion 'adjectives' as well, eg. 'cold' or 'hungry' (physical sensations) and maybe even 'descriptive' adjectives eg. 'grassy' or 'green' (physical appearance/properties), with the NOM marker being the least permanent (most fleeting), and dynamic being most real/true (eg. to a prototype concept or the real-world - a hill that is grassy with bright green, healthy grass and dew etc would be grassy.DYN) or 'most/superlatively X' (eg. cold: cold.DYN would probably be freezing, whereas cold.NOM would be a fleeting, insubstantial cold not really to be considered/worried about/etc).

So this sounds like a kind of cultural philosophy about the world - that things/objects are (while still real), fleeting in existence and so there would be more of an 'awareness' if you would call it that of the shifting/changing nature of reality over time, due to natural forces.

Interesting, thanks for your questions!
Back to top
View user's profile Send private message
Aert



Joined: 03 Jul 2008
Posts: 354

PostPosted: Fri Oct 25, 2013 3:42 am    Post subject: Reply with quote

So the overall syntactic structure and the noun-incorporation structure have changed already...

The order of a typical sentence looks like it will be:

Tense/Aspect; verb, agreement marking; subject; object.

Agreement marking will be done by pronominal clitics, while tense and aspect may be inflectional clitics (but I'm not sure yet, they may instead be particles separate from the verb).

Incorporation of an object requires the object (of nominal form or interpretation, see below) to attach to the root verb, resulting in an overall complex/phrasal verb that has the same valency regarding transitivity as the root verb does. So eg. incorporated 'dish-washing' is intransitive, but phrasal 'washing dishes' is transitive.

I'm not yet sure about whether different lexical categories will be overtly marked yet, but if they are, I think that verbs will be marked with a null morpheme, so that they appear to be base roots (and hence, that functional nouns and adjectives are in a sense derived from the verb, though they're all derived from the root).


Initial thoughts have been floating around about the phonology, and I'm going to make this one much closer to a standard polysynthetic language in that it will allow for complex syllables:

max. C(C)V(C)(C) for roots
max. C(C)V(C)(C)(C) for syllables

I'm thinking about doing some sort of vowel deletion system when roots compound, but while I have a few conditions under which it could occur, I don't know Why it would occur (what 'problem' it is avoiding by doing the deleting).

The sounds will include:

a plain/glottalized nasal and approximant consonant distinction
a voiced/voiceless/ejective contrast for most stops and some affricates
the point vowel system /a i u/ as well as the e in 'bed' and schwa

And I'm trying to figure out what the phonotactic rules should be - they will definitely allow different clusters than English, but not quite as crazy as some of the Salish and Wakashan languages out there!

I guess that's it for now Smile
Back to top
View user's profile Send private message
eldin raigmore
Admin


Joined: 03 May 2007
Posts: 1621
Location: SouthEast Michigan

PostPosted: Sat Oct 26, 2013 7:47 pm    Post subject: Reply with quote

One of the things about non-finite verb-forms, if they are action-nominals (is it also true of other verbal nouns?), is that frequently the Subject and primary or direct Object, or the Actor and Undergoer, are often required to be in the Genitive case (or one of the Genitive cases) or one of the similar cases (e.g. Possessive or Attributive or Partitive); or, must be treated grammatically as if it were a possessor or an attribute or a part, in languages that don't have cases.

One thing worth noting is that, in English, the "Saxon genitive" and the "Norman genitive", which are pretty much synonymous in nearly any other use, are different when used to denote such a participant of a non-finite verb-form which is a verbal noun; or at least of a gerund.

We seem to use "noun's gerund" when "noun" is the gerund's Agent or Actor, but use "gerund of noun" when "noun" is the gerund's Patient or Undergoer.

If we have "N1's gerund of N2" we know that N1 is the Actor and N2 is the Undergoer.
_________________
"We're the healthiest horse in the glue factory" - Erskine Bowles, Co-Chairman of the deficit reduction commission
Back to top
View user's profile Send private message
Aert



Joined: 03 Jul 2008
Posts: 354

PostPosted: Sat Nov 23, 2013 8:37 am    Post subject: Reply with quote

Quote:
While every natlang distinguishes between nouns and verbs in the surface form, not all have any lexical category distinctions in the roots. Also, whatever distinction there is needn't be as strong as the ones we're used to.


I finally see what you mean - Salish languages don't seem to have lexical category distinctions at the root level, though they do at the lexical level. I was actually thinking about this project in terms of Salish linguistics, but trying to take it a step further. I'm finding some very interesting hypotheses/structural models that could be quite useful to this endeavour!
Back to top
View user's profile Send private message
Aert



Joined: 03 Jul 2008
Posts: 354

PostPosted: Wed Nov 27, 2013 2:19 am    Post subject: Reply with quote

So I've decided to adopt a syntactic root structure as advanced by Wiltschko (whose framework I'm using for my project), with a bit of an extension.



So in these there's a categorizing head (n/v/a) that marks a structure as noun, verb, or adjective (and there are lots of morphemes that will be able to do so). Some of these morphemes can produce either noun/verb or noun/adjective items, as in the object of the 1st example (phrasally 'at the flowing-place'; nominally 'the river').

Then there's the modifiers, of which there are four so far; these apply before categorization and so can apply to any lexical category, producing different interpretations.

Eg. AUG (augmentative) on a noun produces the plural; on a verb it produces the iterative interpretation (do many times); on an adjective makes for a 'very adjective.'

For the first example, there's both a categorizing morpheme (required) and a modifier (optional). The dynamic marker indicates some sort of active process, and the DIM (diminutive) marker produces the reading 'a little bit.'

The root itself is pretty straight-forward, the important thing is that it can be interpreted as any lexical category depending on morphosyntactic markings, and position in the syntactic (sentence) structure.

The lexical suffix (LexSuf) is a class of bound (usually nominal) entities which are roots (and so can also take modifiers, and further lexical suffixes). They are category-neutral in one type of incorporation (as simple root complements), but are full nouns in another type (basically equivalent to the example phrase structure).


Some examples (where the LexSuf is a full noun, indicated by '=' ):

'we built a/the house'

1PL PERF-√build=house

There are many ungrammatical ways this could potentially be produced:

*1PL DYN-RES-√build=house
LIT: we an (active) house(-building)

*1PL PERF-RES-√build
LIT: we (finished) (a) built (thing) [can't have VERBAL cat.head on a nominalized Root]

*1PL RES-√build
LIT: we a built (thing)/building (person marking on a non-predicative noun)

*1PL DYN-√build=house RES-√house
LIT: we were (actively) house-building a built house
(object not in the complex verb)

Other potential (acceptable) phrasings:

1PL DYN-√build=house and PERF-√build
LIT: we were (actively) house-building and (we) built (it) [3Obj = null]

1PL PERF-√build=house RES-√build
LIT: we house-built the building

However, the same item indicated by the lexical suffix cannot be an object of the verb ('no syntactic doubling' - a property of noun incorporation)

*1PL PERF-√build=house RES-√house
LIT: we house-built the house


Last edited by Aert on Thu Nov 28, 2013 8:08 am; edited 2 times in total
Back to top
View user's profile Send private message
Aert



Joined: 03 Jul 2008
Posts: 354

PostPosted: Wed Nov 27, 2013 8:12 am    Post subject: Reply with quote

Update:



Not completely sure yet, but it produces the order I want, at least.

AgrO and AgrS are agreement for subject and object.

IP is for tense/inflection; I'm not sure what I want to do with this yet as I'm considering trying a system that marks temporal reference solely on the root structure (which would mean you could get a single phrase, no embedded clauses etc, which have different temporal references).

AspP is aspect, standing in here for vP (not really sure why, just aspect seemed more suited to where transitivity markers would live.

[k] is the lexical category marker, n/v/a.

and I think that's it!

So the maximum surface form for a single complex 'word' would be:

[[v-Mod-√Root=LexSuf]-TR=[n-Mod-√LexSuf]]-(Tense?)-AgrO-AgrS

possible example: [[DYN-INC-build=house]-TR=[N-AUG-child]]-3SG-3PL 2SG.POSS
"your children did a little bit of house-building"/"your children started house-building"/"your children have been house-building (but haven't finished)."
Back to top
View user's profile Send private message
eldin raigmore
Admin


Joined: 03 May 2007
Posts: 1621
Location: SouthEast Michigan

PostPosted: Tue Dec 03, 2013 12:19 am    Post subject: Reply with quote

Aert wrote:
Quote:
While every natlang distinguishes between nouns and verbs in the surface form, not all have any lexical category distinctions in the roots. Also, whatever distinction there is needn't be as strong as the ones we're used to.
I finally see what you mean - Salish languages don't seem to have lexical category distinctions at the root level, though they do at the lexical level. I was actually thinking about this project in terms of Salish linguistics, but trying to take it a step further. I'm finding some very interesting hypotheses/structural models that could be quite useful to this endeavour!


Re-reading this I'm not sure I made it clear that; some conlangers (and maybe some linguisticians, since these conlangers think they got their ideas from academic linguisticians) think that in some languages even a fully-inflected finite word, just sitting isolated in the lexicon, doesn't have a lexical category: --- instead, when it gets used in a clause, one can tell whether it's a noun or a verb (or maybe something else) as used in that clause.

(I'm not sure that's true, and I'm not sure who said it or where.)

That means that if a lexicographer were to compile a dictionary of such a language they'd have to leave out the n or vi or vt or what-have-you that we're used to seeing in dictionaries of English.

_____________________________________________________________

I need to read the two posts you've made since the one I just quoted.

I did read the first one, but I can tell I need to study it, not just read it.
I have not tried to peruse the second one; I didn't have to try to read it to tell it would require study.

_____________________________________________________________

Anyway, congratulations on doing such good work!


_____________________________________________________________



Aert wrote:
So I've decided to adopt a syntactic root structure as advanced by Wiltschko (whose framework I'm using for my project), with a bit of an extension.

So, who's Wiltschko? And where can I see the framework s/he advanced? Is there a URL I can look at, or do I need to look in a book, or in a journal?

Your "syntactic root structure" looks an awful lot like and X-bar theory syntax tree.
_________________
"We're the healthiest horse in the glue factory" - Erskine Bowles, Co-Chairman of the deficit reduction commission
Back to top
View user's profile Send private message
Aert



Joined: 03 Jul 2008
Posts: 354

PostPosted: Fri Dec 06, 2013 10:23 pm    Post subject: Reply with quote

Quote:
when it gets used in a clause, one can tell whether it's a noun or a verb (or maybe something else) as used in that clause.


I think here you're talking about the Functional Determinism Hypothesis, which states that the category of an item is dependent on the category of the functional head above it. Eg. there's a category-less root 'run' which, when it merges into a nominal functional head like NP is interpreted as nominal, whereas when it merges with a verbal functional head like VP, then it's interpreted as verbal.

This was disproved by Davis & Matthewson, 1999 "On the Universality of Syntactic Categories" (Revue québecoise de linguistique 27, 29-69) in their analysis of English and Lillooet (an Interior Salish language). Since the two languages are so different, it is at least possible that this is universal.

The alternative, of course, is that roots are specified for category before they merge with the nominal/verbal functional heads, which is the Lexical Determinism Hypothesis. The 'syntactic root structure' I'm using is intended to work under this hypothesis, where a root is categorially specified before it merges with the functional head (that is, it merges with a categorizing head before the functional head).

It's based off Martina Wiltschko's (2005) Root Structure, which is itself based on X-bar/Minimalist syntactic theory, so there's going to be a lot of similarity. I've taken her Root Structure model to its logical extension for this; unfortunately, it doesn't appear to work with the Salish language data, but that's okay because I'm making a language that does Razz

Thanks for the comments!
Back to top
View user's profile Send private message
eldin raigmore
Admin


Joined: 03 May 2007
Posts: 1621
Location: SouthEast Michigan

PostPosted: Fri Dec 06, 2013 10:51 pm    Post subject: Reply with quote

Aert wrote:
.... Davis & Matthewson, 1999 "On the Universality of Syntactic Categories" (Revue québecoise de linguistique 27, 29-69) in their analysis of English and Lillooet (an Interior Salish language). ....

Thanks, I'll try to find that.
_________________
"We're the healthiest horse in the glue factory" - Erskine Bowles, Co-Chairman of the deficit reduction commission
Back to top
View user's profile Send private message
Aert



Joined: 03 Jul 2008
Posts: 354

PostPosted: Fri Dec 06, 2013 11:16 pm    Post subject: Reply with quote

Send me a PM with your email, I've got a copy (and notes)
Back to top
View user's profile Send private message
eldin raigmore
Admin


Joined: 03 May 2007
Posts: 1621
Location: SouthEast Michigan

PostPosted: Sat Dec 07, 2013 8:54 pm    Post subject: Reply with quote

Aert wrote:
I think here you're talking about the Functional Determinism Hypothesis, which states that the category of an item is dependent on the category of the functional head above it. Eg. there's a category-less root 'run' which, when it merges into a nominal functional head like NP is interpreted as nominal, whereas when it merges with a verbal functional head like VP, then it's interpreted as verbal.

Could be. If you could restate that in theory-neutral (meta-)language, I might understand it better.

Was that Hypothesis intended to be a universal? Because AFAIK the languages in which it was claimed that such a thing was true for most words and/or most clauses, were few. And the possible candidates in well-studied and presumably well-understood languages like English are equivocal and a minority.


Aert wrote:
This was disproved by Davis & Matthewson, 1999 "On the Universality of Syntactic Categories" (Revue québecoise de linguistique 27, 29-69) in their analysis of English and Lillooet (an Interior Salish language). Since the two languages are so different, it is at least possible that this is universal.

What did they disprove? That all words already have a part-of-speech before use in a clause universally for all languages, or that it happened ever for any language that a word's part(s)-of-speech weren't settled until it was actually used in a clause? Or is there any difference in the Genrativist program(s)?

It seems, based on what you've written so far, that they probably proved it couldn't be universal that in every language all (most?) words in all (most?) clauses aren't assigned to a part-of-speech while just occupying the lexicon, but rather until they are actually used in a clause. They did this by proving that in two highly dissimilar languages, viz. English and Lillooet, many (most? all?) words already have their word-category(ies) assigned while in the lexicon, before actually being used in a clause. The fact that those two languages are so dissimilar makes it seem likely that for all languages all (most?) words are categorized before using in the clause -- their category is part of their "definition" in the lexicon.

Have I understood you, and them, aright?

(and btw are these the Davis and Matthewson relevant here?)

Aert wrote:
The alternative, of course, is that roots are specified for category before they merge with the nominal/verbal functional heads, which is the Lexical Determinism Hypothesis. The 'syntactic root structure' I'm using is intended to work under this hypothesis, where a root is categorially specified before it merges with the functional head (that is, it merges with a categorizing head before the functional head).

Once again, this is not theory-neutral language. I think I understand and so I think I agree (as if my agreement were necessary, the point is if I understand then your "of course" above is appropriate). But it would be much easier for me to comprehend if it were stated in terms someone who hadn't put much study into Generativism could understand.


Aert wrote:
It's based off Martina Wiltschko's (2005) Root Structure, which is itself based on X-bar/Minimalist syntactic theory, so there's going to be a lot of similarity. I've taken her Root Structure model to its logical extension for this; unfortunately, it doesn't appear to work with the Salish language data, but that's okay because I'm making a language that does [Razz]

I found several links concerning Martina Wiltschko. Are these her? Or rather, which, if any, of the following links refer to the one you're talking about?

http://www.linguistics.ubc.ca/people/martina
http://www.academia.edu/3118781/Decomposing_the_Mass_Count_Distinction_Towards_a_Formal_Typology
http://www.academia.edu/3118784/The_composition_of_INFL_An_exploration_of_tense_tenseless_languages_and_tenseless_constructions
http://www.zoominfo.com/p/Martina-Wiltschko/303665301
http://link.springer.com/article/10.1007%2Fs11049-008-9046-0
http://linguistlist.org/issues/17/17-511.html

Is she your instructor or major advisor or doktorvaterin or whatever the right word is?
_________________
"We're the healthiest horse in the glue factory" - Erskine Bowles, Co-Chairman of the deficit reduction commission
Back to top
View user's profile Send private message
Aert



Joined: 03 Jul 2008
Posts: 354

PostPosted: Sat Dec 07, 2013 9:23 pm    Post subject: Reply with quote

Well, I was studying for a phonology exam, and accidentally made up a pseudo-syntactic account of syllabification and stress assignment, so I'm going to use it Smile

The syllabification and stress systems will be pretty simple, I think, but the phonotactics will be a little more open to complex clusters.

The root structure maps onto the syllable/stress structure pretty simply, with the end result being that the first syllable of the root takes primary stress.
This is because the language underlyingly has an iambic stress-pattern (second syllable of a foot gets stressed) - if there is only one syllable in the foot, then it gets stressed. Since prefixes (eg. categorizing morphology or modifiers) occur before the root, they are positioned to take the foot that is unstressed, so overall the root always takes stress.

(Except if the prefixes end up being more than one syllable - I haven't figured that bit out yet, unless I simply make all modifiers (C)VC, which is quite likely.)

The order of the complex word (outlined earlier) is [categorizing morphology, (modifier), root, (lexical suffix), etc], so if we have:
categorizing morpheme s- (nominalizer)
modifier -an (~plural)
category-less root kint (person)

Then s-an gets grouped as a full syllable, and kint remains a syllable on its own. The foot (two syllables) receives stress assignment on the last syllable: (san)(KINT) 'people'

If the modifier is not included, then the prefix and root must be parsed into syllables - if they can act as a single syllable (dictated by phonotactics), then they will. Otherwise some repair strategy will occur, eg. vowel epenthesis. Since <sk> is an acceptable onset, the two morphemes are parsed into one syllable (SKINT) '(a) person.'

If other items are added, then more interactions can occur. If we have:
st(i)- (perfective)
-an (~plural)
p'at (root, 'build')
su (LexSuf, 'house')

then we're talking about building multiple houses (e.g. as a distributed event: house-building, multiple (iteration) times), as being completed (past tense).
The linearization is st(i)-an-p'at-su. (i) indicates it only surfaces if no other nucleus is found.
The syllabification gives (stan)(p'a)(tsu), where (stan)(p'a) is a single foot.
[the <t> moves due to a preference to maximize the onset and reduce the coda, if possible]

Because stress goes on the last syllable of a foot, the word is pronounced (stan)(P'A)(tsu) [<p'> is the ejective <p>]

I'll try to put up a diagram of the phonology-(morpho)syntax interface I'm talking about soon, but I should resume studying.

(EDIT: just saw your new post, one second!)
Back to top
View user's profile Send private message
Aert



Joined: 03 Jul 2008
Posts: 354

PostPosted: Sat Dec 07, 2013 9:41 pm    Post subject: Reply with quote

Alright I emailed you the files.

Quote:
What did they disprove? That all words already have a part-of-speech before use in a clause universally for all languages, or that it happened ever for any language that a word's part(s)-of-speech weren't settled until it was actually used in a clause? Or is there any difference in the Genrativist program(s)?


Yes that all words have a part-of-speech before use in a clause (i.e. before merging into a clausal structure). They talk about this being a potential universal for all languages, because English and Lillooet are so different, but obviously some more research is needed. It is interesting that these two very different languages (one of which is in the language family where lexical category distinction has been debated) satisfy the same criteria for underlying lexical category, at least.

Quote:
Have I understood you, and them, aright?

(and btw are these the Davis and Matthewson relevant here?)


Yes, from what I understand, and yes.


Quote:
Once again, this is not theory-neutral language


Right, I'm basing my linguistic and conlang work in specific theories.

The two hypotheses differ in one key (binary) aspect: either the root is specified for category underlyingly (and so can only occur in structures which have the same category, such as noun or verb); OR the root is unspecified for category, and receives specification from the syntactic structure above (that is, the root 'takes' the category of the structure it appears in).

Quote:
I found several links concerning Martina Wiltschko. Are these her?

Yes on all accounts; unfortunately her 2005 article on Root Structure isn't listed on her UBC website, but many of her other papers are.

Quote:
Is she your instructor or major advisor or doktorvaterin or whatever the right word is?


Nope, I'm at the University of Victoria; I just came across her work in class last year and wanted to pursue it in the Interior Salish language Nxa'amxcin. I just missed meeting her a few months back though, but I'm planning to send a copy of my essay (based on her model) to her once it's done.
Back to top
View user's profile Send private message
eldin raigmore
Admin


Joined: 03 May 2007
Posts: 1621
Location: SouthEast Michigan

PostPosted: Sun Dec 08, 2013 8:11 pm    Post subject: Reply with quote

Aert wrote:
Well, I was studying for a phonology exam, and accidentally made up a pseudo-syntactic account of syllabification and stress assignment, so I'm going to use it Smile
....
I'll try to put up a diagram of the phonology-(morpho)syntax interface I'm talking about soon, but I should resume studying.
....


Looks great! Cool


_____________________________________________________________


Aert wrote:
Alright I emailed you the files.

Thanks!

Apparently they're claiming that in these two languages every root has its lexical category assigned as part of its "lexicon entry".
If so, I seriously doubt that's a universal; consider Hebrew and Arabic and other Semitic and Afro-Asiatic 3Cons (triconsonantal-verb-root languages). I am pretty sure that these, and also several other languages, are considered to have roots that are simultaneously noun-roots and verb-roots and adjective-roots.

But I don't really know of any natlangs in which the finite forms ("surface" forms) of any finite words fail to have a part-of-speech as part of their "lexical entry". I've seen it claimed for some language(s), but I don't remember who made the claim or when or where or about what language(s). The above languages aren't them, because once a transfix or "binyan" (vowel-pattern and possibly gemination) is applied to a root, the result has a part-of-speech, even before any prefix and/or suffix is applied.

BTW Do you know how to turn off autocorrect in creating a post for this phpBBoard? I just got <b> autocorrected to <b>. ANd then when I tried to space it out it removed the space between the i and the n and then capitalized the i.
_________________
"We're the healthiest horse in the glue factory" - Erskine Bowles, Co-Chairman of the deficit reduction commission


Last edited by eldin raigmore on Fri Jan 29, 2016 8:00 pm; edited 1 time in total
Back to top
View user's profile Send private message
Aert



Joined: 03 Jul 2008
Posts: 354

PostPosted: Thu Dec 12, 2013 7:11 pm    Post subject: Reply with quote

I'm not sure why the syllable/foot structure diagram isn't showing up anymore, but I'll put up a revised version later on anyway (the current updated one has so many notes on it that it would be impossible to read!)

The basic gist is that I've changed the foot structure to trochaic (stress on the first syllable of each foot), but that prefixes/etc are considered to be outside of the foot (and so don't influence where the stress goes). I've made a basic formalism of this, I just have to make sure it works for all cases.

The important thing is that the first syllable of the root gets stressed, and every second syllable after that gets secondary stress.


The root structure section is going quite well, I'm currently working on a way to use semantic extensions on verbal prefixes to be true tense markers (i.e. not marked above the clause in the tense/inflection node of the syntax).

But I should resume studying again...
Back to top
View user's profile Send private message
eldin raigmore
Admin


Joined: 03 May 2007
Posts: 1621
Location: SouthEast Michigan

PostPosted: Thu Dec 12, 2013 10:50 pm    Post subject: Reply with quote

If there are two or three syllables before the root does the syllable two away from the root get secondary stress?
If there are four or five syllables before the root does the syllable four away from the root get secondary stress?
How many prefixes can there be on a single word? And how many syllables can those prefixes add up to?
Similarly, how many suffixes can there be on a single word? And how many syllables can those suffixes add up to?
What about compound words?
_________________
"We're the healthiest horse in the glue factory" - Erskine Bowles, Co-Chairman of the deficit reduction commission
Back to top
View user's profile Send private message
Display posts from previous:   
Post new topic   Reply to topic    Vreleksá Forum Index -> Conlangs
Goto page 1, 2  Next
All times are GMT
Goto page 1, 2  Next
Page 1 of 2

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum


Powered by phpBB © 2001, 2002 phpBB Group
Theme ACID © 2003 par HEDONISM Web Hosting Directory


Start Your Own Video Sharing Site

Free Web Hosting | Free Forum Hosting | FlashWebHost.com | Image Hosting | Photo Gallery | FreeMarriage.com

Powered by PhpBBweb.com, setup your forum now!
For Support, visit Forums.BizHat.com