adding explanation of lexical entries

This commit is contained in:
Noah Diewald 2021-02-25 19:54:31 -05:00
parent 7c78b560da
commit bdf1b2608b
No known key found for this signature in database
GPG Key ID: EC2BAE1E100A5509
2 changed files with 55 additions and 4 deletions

View File

@ -4,21 +4,26 @@ The files in this directory are mostly more fleshed out examples from
talks. They still consitute incomplete fragments but provide details
that are otherwise missing in overviews.
** Potawatomi possession PotawatomiPoss.v
Some of the explanations in these files assume a knowledge of LCG. For
a brief introduction, see the files:
- introlcgho.pdf
- raisingho.pdf
** Potawatomi possession: PotawatomiPoss.v
The phenomenon is described in:
- [[https://noah.diewald.me/files/aimm4poster.pdf][The structure of Potawatomi hybrid-class overabundance]]
** Quechua plurals QuechuaPlural.v
** Quechua plurals: QuechuaPlural.v
The phenomenon is described in:
- [[https://noah.diewald.me/files/diewald2018wp.pdf][WP for sign-based CGs: a focus on morphotactics]]
** Wao Tededo future tense WaoFuture.v
** Wao Tededo future tense: WaoFuture.v
I have not uploaded the slides yet.

View File

@ -291,3 +291,49 @@ Axiom pa_bo_ke_to_v_fut_1 : forall
(mt : mtrip), (Fentry mt) ->
(lem (fst (fst mt)) [ pa ; bo ; ke ; base ]) ->
Sentry V_fut_1.
(** ** An explanation of the syntactic categories
In reality, a verbal syntactic category is generally of the form NP -o
S for an intransitive and NP -o NP -o S for a transitive. Since the
transitivity of the periphrase should be determined by what I am
calling the participal, the actual scheme looks like the following:
|- \lambda t s.s(t)·kebopa;(NP -o V.fut.part) -o NP.1 -o S;\lambda x
P.fut(P(x))
This lexical entry for kebopa says that it takes an intransitive
future participle and then feeds it a first person argument NP.1.
|- boto;NP.1;first
This is the lexical entry of a first person pronoun.
|- epe;NP;water
The above is the lexical entry for water.
|- \lambda u v.u·v·beke;NP -o NP -o V.fut.part;\lambda w z.drink(z w)
The above is the lexical entry for beke.
First beke is combined with epe using modus ponens resulting in the
following:
|- \lambda u.u·epe·beke;NP -o V.fut.part;\lambda w.drink(water, w)
This is now of the right type to combine with kebopa resulting in:
|- \lambda t.t·epe·beke·kebopa;NP.1 -o S;\lambda x.fut(drink(water,
x))
The expression can now be combined with the first person:
|- boto·epe·beke·kebopa;S;fut(drink(water first))
The basic notion is that any participal can be made intransitive by
applying one arguement. Agreement can be handled completely by the
auxiliary, which absorbs the subject argument of the participal when
it combines with it.
*)