[Corpora-List] PS:minimal changes in a paragraph (based on a corpus it appeared) ... (2nd attempt (after first one was deleted))

[Bill Louw]

Dear Ramesh, John and all
 
It may help if I give you Russell's precise wording. All I did was automate it, but in fairness, because it gives a quasi-propositional response, it brings us closer to 'truth' than stochastic automata set out to. Logical events will be repeatable. I take the term 'generate' in its mathematical sense and not as mere replication of the kind a stochastic automaton might 'write'
 
The counters, according to Carnap, satisfy Ramesh's worry. Carnap calls them 'truth-constant logical operators'. The main advantage of Russell's method from my point of view is that it allows text to read text AND f it falsifies the Vienna Circle's main position by showing that what they called metaphysics can now be demonstrated to COLLOCATE with the logical operators of argument. Bring science into metaphysics and we will have the long overdue demise of the cognitive with collocation taking the place of the schematic structure of the given (Quine's Third Dogma of Empiricism).
 
But here is the Russell:'A language of that sort will be completely analytic and will show at a glance the logical structure of the facts asserted or denied. The language that is set forth in the Principia Mathematica is intended to be a language of that sort.' 
 
then he continues and applies this reasoning to natural language with one eye on Wittgenstein who argues from the start of the Tractatus for 'natural language philosophy'.
 
Russell continues:'IT IS A LANGUAGE THAT HAS ONLY SYNTAX AND NO VOCABULARY WHATSOEVER.(My emphasis). Barring the omission of a vocabulary I maintain that it is quite a nice language. It aims at being the sort of language that, if you add a vocabulary, would be a logically perfect language.' He adds that actual languages are not logically perfect in this sense... 
 
Hence, I use the method to extract subtext in a quasi-propositional form. The most frequent vocabulary variables found using a reference corpus can be used to grade authorial sincerity in accordance with my 1993 semantic prosody model (see Milojkovic forthcoming). It works better than human intuition ever could. Don't blame me for this. Blame Russell (1956) 'The Philosophy of Logical Atomism.' pp 197-198. It was Freire who wrote:'Read the word in order to read the world.' This method turns Freire on its head: Use a sample of the world in order to read one instance of the word...Hope this helps. Perhaps we need feedback from the person who posted the question. I only intervened because after three days he had no response and thought his question had been deleted. But I am ready for a new Bootcamp Debate if need be :) Bill
 
 
 
 
Here is the Russell, cited by D.F. Pears:'

--- On Tue, 9/8/11, John F. Sowa <sowa at bestweb.net> wrote:


From: John F. Sowa <sowa at bestweb.net>
Subject: Re: [Corpora-List] PS:minimal changes in a paragraph (based on a corpus it appeared) ... (2nd attempt (after first one was deleted))
To: corpora at uib.no
Date: Tuesday, 9 August, 2011, 13:48


On 8/9/2011 12:48 AM, Bill Louw wrote:
> the method I proposed works in other languages.

Which method is that?

> Structural issues of argument are likely to be universal,
> as the article listed below demonstrates.
>
> Milojkovic M. (2011). Semantic Prosody and Subtext as Universal, Collocation-Based Instrumentation for Meaning and Literary Worlds. In Zaharov V.P. (et al.), Труды международной конференции «Корпусная лингвистика – 2011». St Petersburg: St Petersburg State University, Faculty of Philology. 47-52

This doesn't seem to be available on the WWW.  Where can it (or
something similar) be found.

And by the way, the primary requirement for language generation is
to make a meaningful response.  How do you combine the statistical
methods with the semantics?

John

_______________________________________________
UNSUBSCRIBE from this page: http://mailman.uib.no/options/corpora
Corpora mailing list
Corpora at uib.no
http://mailman.uib.no/listinfo/corpora
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/corpora/attachments/20110809/c8f32213/attachment.htm>
-------------- next part --------------
_______________________________________________
UNSUBSCRIBE from this page: http://mailman.uib.no/options/corpora
Corpora mailing list
Corpora at uib.no
http://mailman.uib.no/listinfo/corpora