Tom, Pat C, and John B, (01)
TK
> I was wondering if this is a relatively stable number of concepts that
> should be in any ontology/repository: the concepts of the English
> language that cannot be described through a semantic sub-net. (02)
General points: (03)
1. Philosophers and scientists since Plato and Aristotle have observed
that only the terms (or concepts) agreed by convention can be fully
defined by necessary and sufficient conditions -- the most common
examples are in mathematics and in stipulations by lawyers. (04)
2. Empirical concepts (in science and everyday life) change with every
novel experience, and no two people have the same experience. (05)
3. The distinction between a word and a phrase is vague. Is 'ice
cream' one word or two? How would you count "words" such as the
German 'Lebensversicherungsgesellschaftsangestellter' or its
English equivalent 'life insurance company employee'? (06)
4. What do you do about the many homonyms of 'bank'? Dictionaries
don't agree on any exact number: river bank, bench or table for
money changers, financial institution, to bank a fire, to bank
an airplane in a turn... All those meanings evolved from the
same Germanic root -- from which we also get the word 'bench'. (07)
PC
> several dictionaries use a "defining vocabulary" as the list
> of words to define all the other words in the dictionary...
> "Longman Dictionary of Contemporary English" ... has about
> 2150 words... Some of the words are used in more than one sense. (08)
*All* of those so-called defining words are used in many more than
one sense. And dictionaries don't even attempt to state necessary
and sufficient conditions for most words. For plant and animal
species, their definitions are hopelessly vague. Some include
a picture; others cite the Latin name, which they don't define. (09)
See below for a brief summary of Adam Kilgarriff's paper, whose title
is a quotation by Sue Atkins. Sue A is a professional lexicographer,
who worked with computational linguists. Adam K is a computational
linguist who has a company that processes large volumes of NL data.
For further discussion, see the other slides in goal3.pdf. (010)
JB
> it is hard to define a concept. (011)
That is the understatement of the year. For conceptual graphs, I used
a clear, simple, and precise way to *avoid* a definition: a concept
is a node in a conceptual graph. I wrote a book and many articles to
explain how those nodes map to the many ways that people (and their
computers) use words like 'word', 'phrase', 'concept', etc. (012)
JB
> The general consensus in The Brain (the journal) community is that
> columns serve as points that function to draw together data from a
> larger area around the column which might be interpreted as a concept... (013)
Yes. That may clarify some issues, but it raises many, many more.
For a summary, see slide 34 of http://www.jfsowa.com/talks/goal2.pdf (014)
John
____________________________________________________________________ (015)
Source: Slide 14 of http://www.jfsowa.com/talks/goal3.pdf (016)
“I don’t believe in word senses.” (017)
The title is a quotation by the lexicographer Sue Atkins, who
devoted her career to writing and analyzing word definitions.
In an article with that title, Adam Kilgarriff observed that (018)
● “A task-independent set of word senses for a language is not
a coherent concept.” (019)
● The basic units of meaning are not the word senses, but the
actual “occurrences of a word in context.” (020)
● “There is no reason to expect the same set of word senses to be
relevant for different tasks.” (021)
● “The set of senses defined by a dictionary may or may not match
the set that is relevant for an NLP application.” (022)
● Professional lexicographers are well aware of these issues. (023)
● The senses they select for a dictionary entry are based on editorial
policy and assumptions about the readers’ expectations. (024)
* See http://www.kilgarriff.co.uk/Publications/1997-K-CHum-believe.pdf (025)
_________________________________________________________________
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 (026)
|