Re: A ravening of ravens

On 3/29/06, Yahya Abdal-Aziz <yahya@...> wrote:
> Jim Henry wrote:
> A pair of antonyms, as we all know, are words
> opposite in meaning to each other.   A contronym
> is a word which is its own antonym, eg (the verbs)
> "rent" or "cleave".
>
> Does your conlang have any contronyms or (to use
> its synonym) antagonyms?
I'm not sure if this counts, but:

The gzb root {ðân} can mean either victory or defeat.
After all, what is a victory for one person is defeat for
another, in many instances. But when inflected as a
verb (active or stative) the ambiguity disappears:

ðân-zô - to defeat, to conquer, to win a victory over

ðân-van -  to be defeated, to be in a state of defeat

As a noun, though, it could be ambiguous in some contexts.

The root {kâj} means "exchange" and the active verb
{kâj-zô} can mean "exchange", "buy" or "sell" depending
on (hopefully) context.

> And could one design a language to make the
> likelihood of contronyms small?
I suppose any well-designed engelang or auxlang
would start out with no such antonymous polysemy.
But once it's released into the wild, people will use
words in new senses and maybe some contronymy
is inevitable sooner or later -- though I can't think
of any examples yet in Esperanto, I shouldn't be
surprised if some turn up later.

> A related question: how to design a conlang to
> minimise - or at least delay - the occurrence of
> homophones - words of similar sound but different
> meanings.  English abounds in homophones, eg "pier"
> and "peer"; is this due to its almost indiscriminate
> borrowing, or just bad luck (aka "the law of
> averages")?
Some time ago there was a discussion here about
designing a language to maximize the phonetic
distance between actual words.  In the extreme case,
no two realized words would be minimal pairs; all
would differ by at least two phonemes.  If such
a language started to be used more natively than
learnedly and underwent sound change, it would not
develop so many homophones so soon as would a
language like Esperanto which starts out with many
words only distinguished by one phoneme or
even by one distinctive feature - but if the sound
changes went on long enough, such homophony
would eventually appear.

A less extreme method would be to make sure
all the words _within a given semantic domain_
differ by at minimum two phonemes.  Or all words
within a given contextual or distributional category
should differ by at least two phonemes.  (Vorlin
applied this principle in some areas, e.g. its pronouns;
I'm not sure how systematic it was elsewhere in the
language.)  For instance, all personal pronouns and
monosyllabic nouns that might be agents of a verb
could be min. 2 phonemes distinct from each other,
but some might be only one phoneme distinct from
some inanimate nouns or physical-quality adjectives or
temporal adverbs...  In such a case the homophony
resulting from sound change would leave the homophones
very likely to be distinguished by context.

Such a language would probably have to be a priori..

> I'm still interested in knowing whether any of
> your conlangs has some collective nouns beyond
> the ordinary.
None right now, but I'll keep the possibility in mind next time I'm
working on a naturalistic conlang (or one sketch
I've got in progress, which is an auxlang/engelang
that escapes into the wild and mutates following
a crash of civilization).

Collective nouns of the English sort are unsuitable to
an engelang or auxlang, but some such languages have
a variety of collective derivations (beyond Esperanto's
generic -ar suffix).  A while ago there was a
discussion about such.  gzb has four, Ithkuil has
(IIRC) nine.  Some possibilities for such engelang
collectives are:

- unordered set
- pair, trio, etc.
- ordered sequence, series
- ordered matrix, grid, table in 2+ dimensions
- network of interconnected parts
   - decentralized
   - hierarchical
- the whole set of of ~ everywhere

with further distinctions for the homogeneousness or
heterogeneousness of the elements of the collective
(identical, same kind but not identical, different kinds
but common purpose, different kinds but common
location...)

--
Jim Henry
http://www.pobox.com/~jimhenry

Reply