allnoun, Linear P

>Anyway - what exactly _is_ a postposition or, for that matter, a
>preposition?  I don't think I've found a convincing definition that covers
>them all.  The term 'adposition' doesn't help at all and offends my
>classical sensibilities - but the "correct" form 'apposition' has got
>another meaning  ;)
True.  Besides, even Romans didn't *always* assimilate the final consonant of
their prepositions.  Or if they did, they didn't always spell it that way!

I like Cognitive Grammar a la Ronald Langacker and George Lakoff and company.
I suspect they would say that an adposition construction has these elements:

(phonological)
1 2 3, in either order.
1 could be any of a number of things (noun, verb, adjective, phrase of several
kinds)
2 is the adposition in question
3 is a noun.

(semantic)
1' is the semantic content of 1.
2' is the semantic content of the adposition: it has two elaboration sites*,
one for 1', which is profiled*, and one for 3', which is not.  2' includes
some specific relationship between the two, prototypically but not necessarily
a spatial one.  2' does not call for sequential scanning* on the part of the
interpreter.
3' is the semantic content of 3.

*elaboration site: a part of the semantics of a construction which is
unspecified -- a semantic blank to fill in.
*profiled: the "profile" is that content which is "designated" -- pointed out
-- by the construction.  Its complement is the "base."  For example, the word
"nose" profiles the nose, but equally a part of its meaning is its base, which
is the face as a whole.  Modifiers, in Ronald Langacker's system, profile a
part of the content which is an elaboration site -- a blank to be filled in.
Heads profile explicit content.
*sequential scanning: an unusual feature of Ronald Langacker's cognitive
grammar that sets verbs apart from other relation-expressing constructions.
Langacker theorizes that verbs are actually processed differently by the mind
when they are interpreted: the "scene" profiled by the verb is scanned
sequentially in time rather than considered simultaneously in time.
Adpositions do not call for sequential scanning in and of themselves.

>Ah well.
>But the point I was trying to make - I think - is that it is not going to
>obvious to everyone that 'on' in 'table:on' is a noun, other
>interpretations are possible.
Well, yes, depending on how you define nouns. :)  But "on" functions in all
ways like a noun, grammar-wise, in "allnoun."

>-----------------------------------------------------------------------------
>At 3:02 pm -0400 20/4/99, Ed Heil wrote:
>[snip]
>>
>>I'm not sure how far you can push this argument, though.  I mean, surely
>>"perch" is a noun, and yet it means "thing-something-is-on."  "Location"
means
>>"place-something-is-at."
>
>table:perch  ??
>
>I think 'table:topside' might be better.
No, remember, we're talking about value:role pairs.  If a pen is on the table,
the table is its "perch" -- the place on top of which it is resting.  I know
"perch" is a little infelicitious cause it's really designed for animate nouns
and has more specific content than simply a place something is resting.  But
if a pen is on the table, the table is not the topside of the pen.

>>In natlangs, semantic content which would
>>prototypically appear in one part of speech sometimes shows up in another.
>
>I agree - all I'm saying is thatsome of Tom's nouns can be interpreted
>differently.
>[snip - I think, in fact, we don't differ too much - it's rather, it seems
>to me, a matter of emphasis]
Agreed. :)  I don't think I disagree with anything but small details and
points of emphasis.

>>Ray,  I think that we have to accept that his nouns are really nouns.  If we
>>don't, to be consistent, we have to deny that natural language gerunds and
>>certain other nouns are nouns, and surely we can't do that.
>
>Yes, but the whole point of things like gerunds and participles is that
>they combine both _verbal_ functions and that of nouns or adjectives
>respectively.  They stride both camps.  Once we try to eliminate finite
>verbs altogether it seems to me very difficult to maintain that the gerunds
>and/or participles use aren't in fact fulfilling the function of the finite
>verbs we've bannished.
Of course they're fulfilling the function of verbs.  I can say, "there was a
robbery, masterminded by Bob and Dave, at the lighthouse," and the noun
"robbery" is fulfilling the function of the verb "rob" in the sentence, "Bob
and Dave robbed the lighthouse."  But clearly "robbery" is still a noun,
despite the fact that it is carrying semantic information customarily carried
by verbs, and has its agents and patient expressed in the sentence.

>>This is how AllNoun works: in natlangs, the content of other parts of speech
>>can be squished into nouns (i.e. gerunds, nouns like "perch" and
"location"),
>>and AllNoun simply pushes this tendency to its ultimate limit.
>
>Basically I go along with that.
>
>>The problem of course is that nouns, unlike other parts of speech which
>>"modify" or "govern" other words, have no built-in way to relate to each
other
>>in any coherent way,
>
>Cases  ;)
Good point!!  I had not thought of that.  Hmm...

>>so Tom had to come up with his own, artificial way to
>>relate nouns to each other -- part-whole and role-value relations, expressed
>>with punctuator morphemes.
>>
>>I think this is extremely extremely clever but unnatural and awkward to most
>>human minds.
>
>I certainly don't disagree on that - I said in my original mail that I
>thought the experiment was well worth making.
>
>>If there were a way to implement it less awkwardly, I'd love to
>>use it in a conlang.
>
>Well, why not try  ;)
HMMMMmmmmmmmm... :)  Not a bad idea.  Not a bad idea at all.

How about a case-ending that expresses "is the role filled by"?  That would
eliminate those silly colons.  Only problem is when you have to add that case
ending to a long (parenthesized) string of words.  Hmmm...

>You might even persuade my to forget Mathias' paroxystic poncifs and have
>another go at 'All-Verb'   :-)
No thanks.  The nouns work for me just fine. :)

>PS.
>
>>     "Koy tse tl'an tse tum gen nekom payaw;
>>           ts'enra me hlay man yatam."
>>"The noble nation of Atlantis is greatest among men;
>>    And its reign shall extend unto eternity."
>>          (from a Linear P inscription.)
>>****************************************************
>
>Linear P?  Tell me more ;)
P as in "palaeo-", you see.  It was deciphered on the grounds of being a very
distant ancestor of linear A and B (in the process, linear A was cracked.
Turns out it's a dialect of Luwian with heavy non-Indo-European, presumably
'Minoan', influence...).  Atlantean, or "tl'an-tse" (properly Mal-Tl'an-Tse,
the Language of Tl'an the Noble) turns out to be a simple, isolating, SVO
language.

Somehow this didn't make the academic journals.  Not sure why... ;)

--
****************************************************
Ed Heil ..................... edheil@mailandnews.com
****************************************************
     "Koy tse tl'an tse tum gen nekom payaw;
           ts'enra me hlay man yatam."
"The noble nation of Atlantis is greatest among men;
    And its reign shall extend unto eternity."
          (from a Linear P inscription.)
****************************************************