Re: Another Sketch: Palno

On Mon, 25 Aug 2008 19:19:25 -0400, Logan Kearsley <chronosurfer@...>
wrote:

>Palno is essentially an experiment in weird grammar, inspired by
>postfix mathematical notation. It started out also being an attempt to
>create the simplest unambiguous grammar possible, but that quickly
>gave way to aesthetic considerations. It's still simple enough that
>the complete grammar is only 3 pages, including inflection tables and
>example sentences.
>Palno has only three parts of speech (atoms, predicates, and
>conjunctions), and only two basic rules:
>1. Predicates follow their arguments.
>2. Conjunctions are infixed between arguments.
>Conjunctions and the 2nd rule are included to aid human
>comprehensibility; the language is complete and unambiguous without
>them.
Given that clauses can be zero-converted to arguments just as atoms can, is
there really a distinction between atoms and predicates?  Why not analyse
atoms as 0-ary predicates?

One change that doing this would make to your grammar as presented is that
an atom or a conjunction of atoms standing alone would become a complete
clause, where it's not now.  But I don't see much else.  Maybe you mean
there to be other distinctions.

>Atoms are approximately equivalent to nouns, predicates subsume all of
>the functions of verbs, adjectives, adverbs, and adpositions and some
>extra weird stuff, and conjunctions are as usual.
>Predicates act like postfix mathematical operators to describe actions
>on or relations between their arguments. Predicates are marked for
>valency, and can take 1-3 arguments. Any collection of a predicate
>preceded by the correct number of arguments forms a complete clause.
How do you deal with nominal sorts of notions that "want" to have an
argument, or verbal sorts of notions that don't "want" any?  (The former IMO
is a reasonable interpretation of "obligatorily possessed nouns" in langs
that have such.  Weather verbs are in many langs a good prototype of
argumentless verbs.)

>There are three types of arguments- atomic arguments, formed from
>single atoms, and predicate arguments, formed from complete clauses,
>and compound arguments, formed from groups of atomic or predicate
>arguments bound by conjunctions.
>Arguments are marked for number (singular or plural) and case
>(nominative, accusative, or dative). The case marking allows the
>argument of a predicate to appear in any order; this is extremely
>useful for avoiding deep nesting of clauses, which can quickly get
>confusing. Compound arguments mark each sub-argument separately with
>the same case. Predicate arguments are marked on the final predicate.
I had an engelang whose grammar was of a similar sort, and eventually I got
really tired of the inelegance and inflexibility of different fixed argument
sets and return values, as it were, of various types of word.  Just as Palno
seems to be doing, {nouns} (including those with arguments, like 'mother
(of)') returned references to the object, but {verbs} returned references to
an event nominalisation (when subordinated).  I had {prepositions} set up to
take two arguments, just as yours do, but they returned the located noun
instead of some kind of event nominalisation, to make them maximally useful
when subordinated: so _pencil table on_ would be 'the pencil on the table',
not 'that there is a pencil on the table'.  So klugey.
(In this paragraph {foo} means a prototypical lexeme whose English
translation is a foo.)

(FWIW that lang of mine had other ungainly features, too, such as a
lexically fixed tuple of argument roles for each verb and no recognition
whatsoever of theta-roles, a flaw shared (and likely inspired) by
Lo(gl|jb)an.  No case etc. marking, anyway.)

So.  Do you have a way of rendering relative clauses, or anything like them?

What happens if you want to leave components unspecified?

If
  'I did the job well' == ['I did the job'].NOM 'good',
then what is 'it is good that I did the job'?

And, it looks like your preposition-like words are showing the same
schizophrenia wrt interpretation of subordinated predicates that mine were.
 With reference to

>The sentence "I hit the cat on the table with a ball" is structurally
>ambiguous over whether the cat had the ball or whether I used the ball
>to hit the cat. Palno renders each case differently:
>"I cat table-ACC on ball-ACC with-ACC hit." vs. "I cat table-ACC
>on-ACC hit ball-ACC with."
check my work, if you would: are these
  I [[cat table-ACC on] ball-ACC with]-ACC hit
resp.
  [I [cat table-ACC on]-ACC hit] ball-ACC with  ?

Anyway, it seems that here you're failing to recognise a polysemy of English
"with", that coincidentally has just come up on the list: check the thread
"Help with grammatical term".  In the first "with" is 'accompanied by', the
comitative; in the second it's 'using', the instrumental.  The syntactic
ambiguity you mean to get at here is a valid one, but the example is kinda
tarnished by using a semantically different prepositiony relation in each.

Alex

Reply