Fuzzy Voice

The "voice" modals, hiah (Agent) and fiuh (patient) are binary two
valued true/false modals as is traditional in western logic. Anything is
either and agent or not an agent in this mind-set. Thus in the legal
system, if

Bill hiah am gape Monica.
Bill was kissing (as active agent) Monica. is true,
He is a criminal, but if
Bill fiuh am gape Monica.
Bill was being kissed as (passive recipient) by Monica,
he's home free.
[hiah] is the agent tag on [Bill, it also affects the verb].
[am] is a tensor.

This post will explore the implications of applying fuzzy logic to the
grammatical categories "agent" and "patient". There is a background
example in the Vector tense series explaining fuzzy membership.
Suppose:
Bill rozi hiah am gape Monica.
Bill rozih am gape Monica.
Bill was kissing (low fuzzy membership value=rozi) (actively=hiah)
Monica. is true.

(A contraction for [rozi hiah] is [rozih].)
This leaves Monica's role undefined.
(Below are listed three main fuzzy membership categories defined
in NGL. )
Suppose she was more active:
Bill rozih am gape Monica roze hiah. or:
Bill rozih am gape Monica rozeh.
This claims they were both agents, with Monica more active.
Suppose as some lawers asserted that Bill was completely the patient
and Monica was completly  the agent:
Bill fiuh am gape Monica hiah.
Bill (entirely patient) was kissed by Monica (entirely agent).
This form is the classical dichotomy. Eve gets the blame.
Egalitarian form:
Bill rozga fiuh & hiah am gape Monica rozga fiuh & hiah.
Bill equally patient and agent was kissing Monica equally patient and
agent.
Another form:
Bill rozof am gape Monica rozoh.
Bill (mostly patient) was kissed by Monica (mostly agent).
Bill rozif am gape Monica rozoh.
Bill (slightly patient) was kissed by Monica (mostly agent).  etc,etc.
&etc.
----------------------------------------------------------
                TABLE OF VOICE CONTRACTIONS

rozi fiuh--> rozif   low membership patient category
roze fiuh--> rozef   medium membeship patient categoy
rozo fiuh--> rozof   high membership patient categoy
rozga fiuh-> rozgaf  equal membership patients, requires two categories.

rozi hiah--> rozih    low membership agent category.
roze hiah--> rozeh    medium membership agent category.
rozo hiah--> rozoh    high membership agent category.
rozga hiah-> rozgah   equal membership agents, requires two categories.
#roz::- quantitative fuzzy membership, # is 0-1.
roz#::- same as above.
---------------------------------------------------------

Ken caused the trial.
Ken hiah pa fe ku pilixi.

Bill hiah pa fe ku pilixi.
Bill caused the trial.
These statements are mutually exclusive when the hiah, agent, is used
for emphasis on the causal person. This is the classical binary logic
metaphor. In the minds of classical logicians, this logic is primal and
takes precedence over any modal logic.

Bill rozga hiah pa fe ku pilixi.
Bill (fuzzy value .5) caused the trial, leaves the other .5 membership
in the trial-causer category open.

Bill rozi hiah pa fe ku pilixi. OR:
Bill rozih pa fe ku pilixi.
Bill (fuzzy value lowest category) was the agent that caused the trial.

Bill rozih pa fe ku pilixi.
Richard rozeh pa fe ku pilixi.
Linda rozih pa fe ku pilixi.
Ken rozih pa fe ku pilixi.
Monica rozih pa fe ku pilixi.
Henry rozih pa fe ku pilixi.
.
.
.
All 6 statements are consistent as modals, with the greatest agency
assigned to Richard (rozEh).

Suppose the polls say that 62% of the people believe that Richard caused
the trial.  We can verbalize this by saying:

Richard .63roz hiah pa fe ku pilixi.
Richard, with a fuzzy membership value of .63 in the causal agent
category, caused the trial.
------------------------------------------------------------------
FUZZY MODAL DEFINITIONS  3/6/99

All nilenga modals are second order logic claims about propositions,
implicit or explicit, but they can be used in a usual way; thus Ka can
be used as "can".
RO type of modals. [ was PO] [potentials]
..............(deletions)

roz#::-fuzzily      x says "p" has fuzzy #=[0-1] membership in
                    '"P" is true.' (2)
#roz::- as above.
rozI::-low  membership value. ?-.33
rozE::-medium membership value. .33-.66
rozO::-high membership value. .66-.99..  [fuzzy modal predictions]
rozOL::-certain, fuzzy value 1.0. Means same as "p".
rozGA::- equal membership value. 0.50. Implies two categories.
----------------------------------------------------------------------




\1itm fe
\2tp V
\4st acc JD
\5def make, do, cause to be
HOM

\1itm pili
\2tp V
\4st acc JD
\5def test, prove

For the NGl project,

Gerald Koenig