The concept-space of rhetoric
Jorn Barger
http://groups.google.com/groups?th=786339f80f1046fe
--Silke-Maria Weineck challenged me:
> > From an artificial-intelligence/natural-language-processing
> > perspective, ['metonymy' is] a really stupid, poorly defined
> > rhetorical form, as well:
> > http://www.robotwisdom.com/jaj/ulysses/eolustropes.html
>
> Why should terms of literary or discourse analysis answer to the
> standards of an AI perspective?
It's not yet commonly understood that AI is a general advance on
the scientific method-- applicable to every domain of knowledge--
rather than a specialised subdomain of computer science. Its
goal is to use the discipline of programming to force every
domain to spell out its knowledge precisely, and to test the
result for hidden flaws.
But the AI of _rhetoric_ is an especially rich and apt domain
that natural-language specialists in particular should be
exploring.
Joyce used ch7 of Ulysses (Eolus) to inventory the classical
rhetorical forms, which have ridiculously fancy names that date
back to the ancients... but which, surprisingly, are often very
poorly defined.
And Joyce innovated new forms so generously that the analysis
of Eolus is an ideal startingpoint for rigorously exploring
the 'concept space' of classical rhetoric. (The already-named
forms are so arbitrary that they make an extremely
impoverished startingpoint by comparison.)
Starting simple:
I don't know if there's any classical term for a Joycean trope
like 'dullthudding', where he simply deletes the space
between two words to achieve various subtle effects.
[more: http://www.robotwisdom.com/jaj/ulysses/compounds.html ]
There are plenty of examples of this happening naturally as
language evolves, and you can generalise the concept to
cover, by analogy, fusing two _sentences_ into one, or two
paragraphs into one, or two (eg) book-reviews, etc.
(The Silva website lists Ellipsis, Asyndeton, and Brachylogia
as other tropes involving 'Omission':
http://humanities.byu.edu/rhetoric/Figures/SCHEMES.HTM )
This 'dimensional analysis' has a rich payoff for programmers
when it succeeds-- a word-processor might have one menu that
allowed selection of 'granularity' (letter-syllable-word-
phrase-sentence-paragraph-chapter-work) and another that
allowed selection of functions like 'fuse', 'swap', 'sort',
'insert', 'delete', etc.
On my Eolus page I also wrestle with the difference between
'chiasmus' and 'antimetabole' which hinges on whether words,
phrases, *or ideas* are repeated in reverse order. (Adding
ideas to the mix is a brilliant extension of the challenge.)
Joyce uses a wonderfully effective trope to evoke a fat neck:
"Fat folds of neck, fat, neck, fat, neck."
This isn't even literally correct-- between the stripes of
fat-neck would be black lines of nothing-- but the concept
of a metaphor between the-words-themselves on the page and
the-sensory-impression in the world (or imagination) is
extremely rich and well-plumbed by JAJ. His:
"...trams slowed, shunted, changed trolley, started..."
is a marvel of kinesthetic onomatopoeia, for which there
needs to be a more-disciplined rhetorical vocabulary.
I could go on and on with this, and will at the drop of a
hat-- but will anyone drop the hat?
Lionel Bonnetier
> But the AI of _rhetoric_ is an especially rich and apt domain
> that natural-language specialists in particular should be
> exploring.
It seems to me that what we call figures of speech
is actually our more intuitive way of speaking, and
"grammatically pure" language is a subset of these
figures that has been frozen into a standard.
But the study of language generation is in infancy,
even though more advanced in some ways than the
moldy study of literary style. The situation is
worse in language understanding: it relies on
dictated formalism and has no chance of feeling what
is behind figures of speech. Even a catalog of such
figures wouldn't capture the essence of it, and the
catalog would never be complete.
I don't know whether the guy who pricked you on
rec.arts.book pranced on ignorance, or actually
knows anything about current AI, in either case he
illustrates the gap between the deep, but obscure,
human feeling for language, and the precise, but
shallow, approach of AI.
smw
...
> Joyce uses a wonderfully effective trope to evoke a fat neck:
You have the nerve of attacking people for "topos" while using "trope"?
pish-tash, as my best girl friend would say.
Jorn Barger
news:<8281d0b6.02101...@posting.google.com>:
> > Joyce uses a wonderfully effective trope to evoke a fat neck:
>
> "trope"?
I'd love to substitute a less-pretentious synonym if I knew one.
(SSC: I have no idea what 'trope' actually means, it's just an
innocuous gap-filler for 'figure of speech'.)
smw
Jorn Barger wrote:
It means figure of speech. And you use trope because it's shorter. Just
as other people prefer to say "a topos in 18th century thought" rather
than use a cumbersome paraphrase of that perfectly useful little word.
s
Siri Atma Oaklander De Licori
> hat-- but will anyone drop the hat?
_
_|_|_