This seems to aligned with the commentary around IBM's Watson approach
that used multiple parallel attacks to decipher the meaning of a
question and derive some possible answers.
The designers seemed to be quite at ease with the fact that formal
systems for NLP would not cut it on their own. (01)
On 08/09/2011 9:33 AM, John F. Sowa wrote:
> I keep saying that the human brain is much more complex than any current
> description. It can deal with anything and everything under the sun
> (and beyond) in an immense number of ways. But as I have said for
> years, what's in the brain is better characterized as knowledge soup
> than a nice neat formal system.
>> A "PutAllSignatures" concept should be able to interpret all
>> the versions of the English "put", all signatures of same, and
>> all semantic meanings as derived from a corpus of "put" sentences.
>> That is what I need to find, and FCA doesn't cover such odd usages.
> You might want that, but there is no magic. Life is tough.
>> Another way to view the question is: how are the FOL expressions
>> for each signature, all using a common set of signature class
>> names for operands, to be processed so that the bindings of each
>> corpus can be automatically switched...
> FOL is great for what it can do. But everybody who tried to develop
> a completely formal system for NLP has failed. Cyc is still plugging
> away at the problem, and I wish them luck.
> Wittgenstein started with his elegant little book in 1922, which was
> strongly influenced by the work of his mentors, Frege and Russell.
> But after teaching school in an Austrian mountain village for
> a few years, he discovered that kids don't think that way.
> He went back to Cambridge in the 1930s, sadder and wiser. He spent
> the rest of his life explaining why that way doesn't work. In 1972,
> Terry Winograd published his PhD dissertation with the title
> _Natural Language Understanding_. But by 1986, he gave up, and
> said that it doesn't work and that nobody can make it work. I believe
> that if Wittgenstein's first book had been right, Winograd could have
> finished the job of NLU in the 1980s.
> Following is an article that discusses those issues:
> Signs, Processes, and Language Games
> See below for a brief summary of my recommended method for
> dealing with the complexity in a way that is manageable.
> For a more recent article, which goes into some of the issues
> in a more detailed way, see
> The Role of Logic and Ontology in Language and Reasoning
> You can do a lot of NLP without getting a system that fully
> "understands" NLs. The following article discusses ways
> of designing systems that relate NLs to more conventional
> kinds of software:
> Future Directions for Semantic Systems
> If you want magic, go to Hogwarts. But there's a lot that
> can be done along the lines I discuss in those articles.
> Source: http://www.jfsowa.com/pubs/signproc.htm
> Multiple Hierarchies. As a way of formalizing a structure of
> microtheories to accommodate an open-ended, possibly inconsistent
> knowledge soup, Sowa (1990c, 2000) proposed to organize them in an
> infinite lattice, which would be rich enough to include any possible
> language game that any finite reasoner (human, computer, or
> extraterrestrial) could ever invent. Central to the framework is a
> partitioning of the knowledge base into four separate hierarchies with
> appropriate mappings to link them. A multilingual system of N languages
> would have 2N+2 hierarchies.
> 1. Words. For each natural language, a hierarchy of words and word
> senses similar to WordNet and EDR or the more automated MindNet
> (Richardson et al. 1998; Dolan et al. 2000). Each word sense would be
> mapped to some type in the type lattice. EDR, for example, has separate
> hierarchies for Japanese words and English words, which are mapped to a
> single type hierarchy that contains the word senses of both languages.
> 2. Types. A lattice of concept and relation types, which are used to
> index the contents of the other three hierarchies. Some types may be
> expressible by word senses in several different languages, but others
> might be not expressible by single words in any of the N languages. A
> type that does not correspond to a single word-sense in some language
> would have to be expressed by a multiword phrase in that language.
> 3. Canonical graphs. A partial ordering of conceptual graphs that
> express the lexical patterns and thematic roles of natural languages.
> Each canonical graph is indexed by the types that occur in it. The
> canonical graphs are generalizations of the more informal frame systems
> that are often used in computational linguistics to represent lexical
> 4. Theories. An open-ended lattice of theories, each of which
> axiomatizes the detailed knowledge about some subject from a certain
> point of view. Each theory is indexed by the types that occur in its
> axioms. The complete lattice of all possible theories is infinite, but
> only a finite subset could ever be implemented in any actual system. Any
> implementation, however, must make provision to accommodate any theory
> in the infinite lattice that might turn out to be useful.
> [Note: Hierarchies #1, #2, and #3 have very few axioms. The
> detailed axioms that can be used for formal deduction are in #4.]
> Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
> Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
> Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
> Shared Files: http://ontolog.cim3.net/file/
> Community Wiki: http://ontolog.cim3.net/wiki/
> To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
Artifact Software Inc
phone: 866-970-2435, ext 102 (04)
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J (01)