Re: THEORY nouns and cases

On Sunday, April 25, 2004, at 09:21 AM, Joe wrote:

> Ray Brown wrote:
[snip]
>> "1. A distinctive, overtly marked form which can be assumed by an NP to
>
>
> My personal belief, actually, is that it is marked - it's marked by word
> order(Thus, the only language with 1/0 cases is MRLL).
..and my personal belief is different. But I don't see any future in
exchanging personal beliefs - people have been studying linguistics for
some time now. Although I take an empirical view & don't subscribe to any
particular linguistic school, I do take note, as far as I can, of what
others have done and of their (sometimes conflicting) view. Until someone
convinces me otherwise - and personal belief will not do so - I go along
with Larry Task's first definition of 'case'; after all, he was AFAIK a
respected linguist.

> Let me give another example - if a language was to mark tense with word
> order (VSO-preterite, SVO-present, SOV-future) - would you say it has no
> tense?
I would say tense is not comparable to case for a start. Secondly, I find
it difficult to imagine any natlang would behave in the way described.
Indeed, I've encountered a great many natlangs in my 60+ years, but I do
not recall one that used word order to distinguish between time references.

But, again, I will quote Trask: "Tense is a frequent category in the
languages of the world, but is far from universal, Chinese being an
example of a language which lacks tense entirely."

> It clearly has a time-distinction, but no morphologically marked
> tense.
I know of no language which has non-morphologically marked tense. That
does not, of course, mean such a language doesn't exist, but I cannot
comment until I know rather more about the language which is purported to
have tense marked by variation in syntax.
 =========================================================================
On Sunday, April 25, 2004, at 06:58 PM, jcowan@REUTERSHEALTH.COM wrote:

Ray Brown scripsit:

>> If, as you say, you cannot talk without cases, and case is a property of
>> noun phrases, then, surely, it must mean that you cannot talk without
>> nouns. Therefore, if it is agreed that a language has only one class of
>> things, that class must be 'noun' as you cannot talk without nouns.
>
> Non sequitur.  Lojban has noun phrases (sumti), but no nouns; that is, the
> head of a noun phrase is always a verb (selbri), with a determiner that
> means "that which is the (first, second, third ...) argument of this verb"
> .
> Thus the verb "cribe" means "be a bear", and "lo cribe" is "that which
> is a bear".
>
But lojban, like Loglan, was based, I thought, on clausal form logic and,
indeed, in order to test the Sapir-Whorf hypothesis, it was purposely
designed not to be like a natlang. Also, of course, altho lojban describes
'lo cribe' as you say (and you should know  :) it is possible that a
linguist could analyze it as 'lo' being a substantiver, i.e. in 'lo cribe'
, 'cribe' has indeed been substantivized & is a noun.

Is there any natlang where noun phrases occur but no actual nouns?
 ========================================================================
On Sunday, April 25, 2004, at 02:32 PM, Henrik Theiling wrote:

> Hi!
>
> Ray Brown <ray.brown@...> writes:
> ..
>> I apologize if any of what I say seems hostile.
>
> I did not notice you being hostile at all.
Good - I still have this hangover from my few years on a certain other
list where suspicion was rife and flames frequent.

>  I understand that you just
> want to understand what I mean.  That's why we're talking.
>
> Please don't take my postings as offensive.  It's just fun to try to
> find an agreement. :-)
Yes - but as we seem to be dealing with matters upon which there is
disagreement among professional linguists, I think in the end the
agreement might be, as we say in English, "to agree to disagree"   :)

>>> Hmm, I don't mean by case that the nouns are morphologically changed.
>>> E.g. Chinese also has two cases (the one in front of the verb and the
>>> one after) but marks none by morphological processes.
>>
>> Ah - we're talking at cross-purposes. You are using 'case' in the sense
>> it
>> is used in Government-Binding theory;
>
> After reading what you describe, yes, I think I meant that.
I think so.

>> Clearly, if you accept, as you apparently do, meaning (2) above, then
>> Chinese has two cases: nominative and accusative. If, like me, you accept
>> _only_ meaning (1) then clearly Chinese has no cases.
>
> Hmm, but I don't really see a big difference in marking it
> morphologically or marking it by position.
..but the "it" is the cause of the difference. For me it's verbal
arguments such as subject and object, for you it's cases like nominative
and accusative. Even in GB theory it is found necessary at times to
distinguish between 'abstract case' and 'morphological case'.

In practice, few if any natlangs that exhibit 'morphological case' have
their nominative, accusative & dative cases equating exactly to the
categories of subject, direct object and indirect object (I'm less
familiar with ergative languages to know whether those that have
absolutive, ergative and dative are any stricter in their use of these
cases). And do we say that, e.g. 'from the garden' is the surface
representation of an abstract ablative case, 'towards the garden' the
surface representation of of an abstract allative case, 'in the garden' is
a surface representation of the abstract inessive case etc.? (I'm not
familiar enough with GB to know what the theory says on this.)

The (morphological) cases of a language are easy enough to establish. We
say that German has four cases, ancient Greek five cases, modern Greek
three cases etc. We can describe the various different uses of these cases.
  Some will correspond with subject, direct object & indirect object and
some uses will not.

I will continue to adhere only to the 1st meaning of 'case' as defined by
Trask. To me it seems far less confusing to use 'case', and the
traditional case names, for morphological features, and to use separately
terms like subject, direct object and indirect object verbal arguments.

> Furthermore, IIRC, in the
> case of Chinese, when you front an object for pragmatic reasons,
IIRC topic fronting as in German (the more common reason for fronting; but
some languages, such as Welsh, exhibit focus fronting.

> you
> can mark it with a particle to be the dislocated object.  I think it
> was a suffixed 'ba', but I'm not sure.
>
> You might now say: ok, then *Chinese* does have case,
No I don't. I do not say that when Chinese fronts a noun in accusative
case, it marks the accusative with the particle 'ba'. I say that Chinese
marks a fronted direct object with the particle 'ba';

[snip]

[snip]
>>> Agreed.  But is it necessary to have contrasting forms or would you
>>> say, as I did above, that Chinese has two cases?  (In some verbs even
>>> three (the typical 'give': 'Wo3 gei2 ni3 shu1')).
>>
>> Ooh - so 'gei2 ni3' is dative case? Sorry, I do not agree. I've explained
>> the reasons above.
>
> No, I meant 'ni3' is in the third case, since it is the third argument
> to the verb (which is probably clearer terminology).
Yes, sorry - my typo.  I should have said: "So 'ni3' is dative case?"

I agree that wo3 is the subject, ni3 is the indirect object and that shu1
is the direct object. I disagree strongly that Chinese possesses a
nominative, dative and accusative cases which all happen to have the same
morphological form.

> But ok, maybe the difference between case, prepositional phrases and
> verb+noun phrases is vague.  Or is it not?
Not the way I use case.

> Can Finnish cases like
> allative be analysed as postpositional phrases?  If not, why not?
Because the suffix cannot be separated from the noun and the suffixing of
case endings can change the phonetic structure of the ending and/or the
noun stem, i.e. the two morphemes become inseparable.

>  I
> only have an intuitive idea that they can't.  When is it case, when
> affix?
Case markings are affixes - usually suffixes but infixes, prefixes and
IIRC circumfixes are found also.

> If there was no vowel harmony, is it still an affix?
Yep - Latin is a good example.

> What
> about Japanese?
It uses postpositions.

[snip]
>> Er - but, according to Trask, even in GB theory case is "a
>> putatively universal abstract property of noun phrases", i.e. if you
>> have case, you have 'nouns'.
>
> Ah, ok.  'Putatively.' :-) Does it apply to languages with only one
> class?
I don't know. You'll have to ask someone more familiar with GB theory how
the theory applies to languages with only one class.  My guess is that
there will be considered to be underlying abstract classes even if there
is, apparent;y, only one surface class. But, as I say, someone more
familiar with GB can tell you.

> Which property is a putatively universal abstract property of
> verbs?  If there is only such property, would that forbid languages
> with only one open lexical class?  Or would that suggest that the
> terms apply to languages only that have distinct classes of 'noun' and
> 'verb'?
As I do _not_ hold that notion that case is a putatively universal
abstract property of noun phrases, it's not likely that I'll do anything
similar with verbs.

> Could we agree that the words in AllNouns are 'nouns' and 'verbs' at
> the same time (or, at least, they fulfil one or the other purpose when
> used in a sentence) and that, therefore, 'noun' is inappropriate?
No - I confess I have not examined AllNoun closely enough to form an
opinion on whether Tom Breton was successful in his endeavor or not.

> My current conlang seems to violate the above putative universal
> because it allows phrases to have case.
It's a putative universal of GB theory - not of linguistics generally.
Indeed, many like me do not hold that it is a putative universal at all.
In any case, a conlang is not bound by any so-called universal: it depends
very much on what the aim of the conlang is.

> I'm sure I've seen natural
> languages attaching case markers to phrases.
Like in English: "It's the woman next door's cat" - where "'s" is suffixed
to the whole phrase "the woman next door"  :)
>
[snip]

> Yes.  But although there is case on words, I would not call the
> lexicon entries 'nouns'.
In which case you appear to part company with GB (I think) and seem to be
carving out a theory of your own. No harm in that, otherwise new theories
would never arise.

> What about languages with only one lexical class like Nootka?
I confess I know practically nothing about Nootka. I cannot comment on
your questions without at least some knowledge of the language & it would
be helpful to see different descriptions of it. To my skeptical mind, I
just wonder if the 'one class' will prove to be as elusive as the 'one
vowel' (or no vowel) claimed for some natlangs.

[snip - because without a proper description of Nootka, I cannot usefully
reply]

>> Confusion will arise if one is not aware there are different linguistic
>> theories and different 'schools' of linguistics.
>
> I am aware of the fact, but I am not familiar with these schools.
Nor am I as aware as I'd like to be in some cases.

> I'm a hobby linguist and I hope people will still have fun discussing
> with me and my obviously crude definitions. :-)
I'm a hobby linguist also - just been at the hobby a bit longer. 'Twould
be interesting to see what the 'experts' have to say, especially abut
Nootka.


Ray
===============================================
http://home.freeuk.com/ray.brown
ray.brown@freeuk.com    (home)
raymond.brown@kingston-college.ac.uk   (work)
===============================================
"A mind which thinks at its own expense will always
interfere with language."         J.G. Hamann, 1760

Replies