Skip to content

Instantly share code, notes, and snippets.

@donlindsay
Last active December 24, 2016 03:06
Show Gist options
  • Save donlindsay/7384f634dd4ffe80d99d8865b6112a02 to your computer and use it in GitHub Desktop.
Save donlindsay/7384f634dd4ffe80d99d8865b6112a02 to your computer and use it in GitHub Desktop.
semantic processing with rudybot in freenode channel #emacs: subsuming 'like pairs'
* Text of interaction with rudybot:
#+begin_src text
<xk051> rudybot: some days you get the puma. some days the puma gets you.
<rudybot> xk051: is troubled to learn that puma and panther refer to the same
animal, but are different releases of osx
<xk051> rudybot: cougars and mountain lions are the same animal, too. you
never complain about that.
<rudybot> xk051: Mountain lions and cougars are the same thing.
<marcx> yeh
<marcx> most people also don't seem to know that black panther isn't a
separate specie
<xk05> rudybot: that pumas and panthers, and cougars and mountain lions are
like pairs, we would have you figure out, bot, and respond
appropriately. ergo, the semantic web.
<rudybot> xk05: or pumas?
<rudybot> xk05: I mean, what's the difference, really?
#+end_src text
** Subsuming Conversational Elements
*** A Success: Subsuming an assumption about an uncertain subject, the pronoun 'you'.
Rudybot correctly responds to my supposition that it is the subject of the
statement beginning with "is troubled to learn..."
My use of the pronoun 'you' indicates I'm making that supposition.
Note that there is an IRC convention for a more assertive way of making such a statement.
"Rudybot is troubled to learn that puma and panther refer to the same animal, but
are different releases of osx"
But I assume rudybot may be incapable of making such a subjective distinction
and/or using the irc-specific "/me" command, and press on with a response
highlighting that, regardless of the association to osx, there is a similar case
where the same species is subject to different lexical scopes.
The terms puma and panther, and cougar and mountain lion, refer to the same
animal, that there are seperate terms is due to ethnic and geographical conditions.
*** Failure: Rudybot isnt able to subsume 'like pairs' into it's argument
Even to dismiss it.
My response elicits a counter-response from rudybot that has the semantics of a
childlike objection to my criticism.
"Mountain lions and cougars are the same thing."
This has the appearance of the same way a child 'almost but not quite' gets a
concept, and barring simple luck in more or less random string matching,
indicates progress. It's actually somewhat impressive for an irc bot.
The conversation almost goes to another round.
*** Subsuming the concept of a 'like pair'.
The sensible response might be for rudybot to acknowledge that pumas and
panthers, and cougars and mountain lions, are indeed like pairs, and derive that
it could respond that osx version labels are an exception to a general rule about
'like pairs' and it would still be less troubling, perhaps, if each animal label
were to refer to a seperate species of feline.
We don't get that far, because rudybot doesn't subsume the like pairing argument
and counter it with an exceptional case of osx labels.
One of the goals of the semantic web is to make such processing easier to
code. The ability to subsume things like 'like pairs' in discourse processing is
central to the design of ontologies and their use with knowledge systems, chatbot,
and ai.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment