modified maybe-rule for post-processing

[williampma]

Corresponding rule for opencog/opencog#990

It seems not necessary to pass $r_maybe value to the rule.

[williampma]

Removed the hard-coded varscope stuff for MAYBE, using regular expression instead.

Mentioned briefly at #89

The ScopedVariable implementation might be useful if the matched value needs to be passed to the helper function. But since $r_maybe does not seem necessary, ScopedVariable is not needed in this case.

[williampma]

Ah, I see. That sounds reasonable. However, if the stv for each modal verb would be different, what would determine what is the exact value for each? It seems to be highly susceptible to different people's interpretation.

BTW, the helper function is at the pull request opencog/opencog#990

[ruiting]

In terms of how to determine the stv, that's a question for @bgoertzel ;-P In general, that should vary from different people/characters/agents who say and different people/characters/agents who hear... But we may ignore that complexity at the beginning and just define some initial fixed stv for each modal verb to get things working first.

[amebel]

I have created an issue for discussing the general issue of dealing with stv for each relex relations @ #164

[anitzkin]

A note on that for future consideration:

Some natural languages actually mark their sentences with something like an stv; they are called "epistemic markers" and
they are usually selected on the basis of how the speaker came to believe what they are claiming, such as whether the claim
made in the sentence is something the speaker a) saw for themselves, b) heard from someone, or c) inferred (although I realize
we may want to get more fine-grained than that).

Aaron

---

On Thu, 8/21/14, Ruiting Lian [email protected] wrote:

Subject: Re: [relex] modified maybe-rule for post-processing (#154)
To: "opencog/relex" [email protected]
Date: Thursday, August 21, 2014, 1:15 PM

In terms of how to determine the
stv, that's a question for @bgoertzel ;-P In
general, that should vary from different
people/characters/agents who say and different
people/characters/agents who hear... But we may ignore that
complexity at the beginning and just define some initial
fixed stv for each modal verb to get things working
first.

—
Reply to this email directly or view
it on GitHub.

[anitzkin]

oops, no, now I realize you mean confidence in the parse, not confidence in the claim -- excuse me!

---

On Thu, 8/21/14, [email protected] [email protected] wrote:

Subject: Re: [relex] modified maybe-rule for post-processing (#154)
To: "opencog/relex" [email protected], "opencog/relex" [email protected]
Date: Thursday, August 21, 2014, 3:03 PM

A note on that for future
consideration:

Some natural
languages actually mark their sentences with something like
an stv; they are called "epistemic markers" and
they are usually selected on the basis of how
the speaker came to believe what they are claiming, such as
whether the claim
made in the sentence is
something the speaker a) saw for themselves, b) heard from
someone, or c) inferred (although I realize
we may want to get more fine-grained than
that).

Aaron

---

On Thu, 8/21/14, Ruiting Lian [email protected]
wrote:

Subject: Re: [relex]
modified maybe-rule for post-processing (#154)
To: "opencog/relex" [email protected]
Date: Thursday, August 21, 2014, 1:15 PM

In terms of how to determine
the
stv, that's a question for
@bgoertzel  ;-P  In
general, that should
vary from different

people/characters/agents who say and different
people/characters/agents who hear... But we
may ignore that
complexity at the beginning
and just define some initial
fixed stv for
each modal verb to get things working

first.

—
Reply to this email directly or view
it on GitHub.

[ruiting]

What I meant was the confidence in the claim. What Amen meant was the
confidence in the parse ;-)

Ruiting Lian

On Thu, Aug 21, 2014 at 3:10 PM, anitzkin [email protected] wrote:

oops, no, now I realize you mean confidence in the parse, not confidence
in the claim -- excuse me!

---

On Thu, 8/21/14, [email protected] [email protected] wrote:

Subject: Re: [relex] modified maybe-rule for post-processing (#154)
To: "opencog/relex" [email protected], "opencog/relex" <
[email protected]>
Date: Thursday, August 21, 2014, 3:03 PM

A note on that for future
consideration:

Some natural
languages actually mark their sentences with something like
an stv; they are called "epistemic markers" and
they are usually selected on the basis of how
the speaker came to believe what they are claiming, such as
whether the claim
made in the sentence is
something the speaker a) saw for themselves, b) heard from
someone, or c) inferred (although I realize
we may want to get more fine-grained than
that).

Aaron

---

On Thu, 8/21/14, Ruiting Lian [email protected]
wrote:

Subject: Re: [relex]
modified maybe-rule for post-processing (#154)
To: "opencog/relex" [email protected]
Date: Thursday, August 21, 2014, 1:15 PM

In terms of how to determine
the
stv, that's a question for
@bgoertzel ;-P In
general, that should
vary from different

people/characters/agents who say and different
people/characters/agents who hear... But we
may ignore that
complexity at the beginning
and just define some initial
fixed stv for
each modal verb to get things working

first.

—
Reply to this email directly or view
it on GitHub.

—
Reply to this email directly or view it on GitHub
#154 (comment).