<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
</head>
<body>
<p>Glen -</p>
<p>In the spirit of some of our other woven threads here, I'm
wondering how practical it would be to build a natural language
parser which could apply rules such as those found in:</p>
<blockquote>
<p><a class="moz-txt-link-freetext" href="https://en.wikipedia.org/wiki/Straight_and_Crooked_Thinking">https://en.wikipedia.org/wiki/Straight_and_Crooked_Thinking</a><br>
</p>
<p><a class="moz-txt-link-freetext" href="https://en.wikipedia.org/wiki/List_of_fallacies#Informal_fallacies">https://en.wikipedia.org/wiki/List_of_fallacies#Informal_fallacies</a></p>
<p><a class="moz-txt-link-freetext" href="https://en.wikipedia.org/wiki/Cognitive_bias#List_of_biases">https://en.wikipedia.org/wiki/Cognitive_bias#List_of_biases</a></p>
</blockquote>
<p>Seems like a next obvious step after spell and then grammar
checkers?</p>
<p>Maybe something to be applied broadly to social media?<br>
</p>
<p>Is it a first step toward an AI Overlord that is more like an
overzealous English Major than HAL (2001 space odyssey) or the
Lawnmower Man or SID 6.7 (Virtuosity) maybe? I am reminded of
the cautionary tale offered in Jack Williamson's "<a
moz-do-not-send="true"
href="https://en.wikipedia.org/wiki/With_Folded_Hands">With
Folded Hands"</a> or "the Humanoid Touch". Even though I knew
Jack and heard anecdotes about these stories, I never heard the
one described in this Wikipedia article... a root source of
those who fear "the Nanny State" (scare quotes intended) perhaps?<br>
</p>
<p>Maybe easier to implement with Machine Learning and ubiquitous
training (details left to the reader) than something closer to a
rule-based system? My most recent brush with Natural Language
Processing is over 10 years ago and I sense things have progressed
in the interim.<br>
</p>
<p>- Steve<br>
</p>
<p><br>
</p>
<p><br>
</p>
<p><a class="moz-txt-link-freetext" href="https://en.wikipedia.org/wiki/List_of_cognitive_biases">https://en.wikipedia.org/wiki/List_of_cognitive_biases</a><br>
</p>
<blockquote type="cite"
cite="mid:02283207-2fdb-aba8-693e-27911921809a@gmail.com">
<pre class="moz-quote-pre" wrap="">Interesting. Thanks. I used Ada for 5 years contracting (mostly for the Army) and never once heard any of these terms used that way. And I was part of the group that pitched new projects to our clients. I'm wondering if it simply fell out of use or if I was too holed up in my own little world. We did commonly use the phrases "skeleton" and "fleshing out".
Also FWIW, the straw man fallacy is not solely an adversarial concept. You can straw man yourself. You can accidentally straw man someone. There are 3-way attempts at constructive ... what? ... trialog (?) where each party straw mans the others position on the way to a common ground. Etc.
It's unfortunate that we focus on competitive, zero-sum, and adversarial senses of such things. But that need not be the case.
On 1/28/21 12:30 PM, Steve Smith wrote:
</pre>
<blockquote type="cite">
<pre class="moz-quote-pre" wrap="">Perhaps no-one cares or shares my confusion with the use
Strawman/Steelman championed by Glen and adopted by others, however:
consensus development:
<a class="moz-txt-link-freetext" href="https://en.wikipedia.org/wiki/Straw_man_proposal">https://en.wikipedia.org/wiki/Straw_man_proposal</a>
vs polemical debate:
<a class="moz-txt-link-freetext" href="https://en.wikipedia.org/wiki/Straw_man">https://en.wikipedia.org/wiki/Straw_man</a>
</pre>
</blockquote>
<pre class="moz-quote-pre" wrap="">
</pre>
</blockquote>
</body>
</html>