Re: Most developed conlang

On 4/26/07, Henrik Theiling <theiling@...> wrote:

> Jim Henry writes:
> > On 4/25/07, Henrik Theiling <theiling@...> wrote:
> >...
> > But what if you are adding a word that could be equally plausibly
> > derived from more than one word already in the lexicon -- by more
> > than one route?
....
> >Ah, here's one, perhaps a bit contrived: mal-mangx-em-a.  Is it
> >derived from "mangx-em-a" or from "mal-mangxi-i"?  Does it mean "not
> >tending to eat = not hungry" or "tending to vomit = nauseous"?
> >("Nemangxema" would not be ambiguous like this.)
>
> This seems an easy case then: just use the one that reflects the
> meaning of the new word best.
I think I like Alex Fink's analysis better in this case -- we have
two homophonous words that might need separate lexicon entries
(at least the less transparent one, "((malmangx)em)a" deserves
a lexicon entry, and the more transparent one needs another
entry to note its homophonousness or polysemy with the other),
even though they're composed of all the same morphemes
in the same order.  Of course if E-o developed so that one
of these meanings was conventionally selected as *the*
meaning of the word and the other was never used, we
could drop the unused one.  Or say that the use of the
word in one sense blocks its use in the other sense,
like the root "postul-o" blocks "post-ul-o" requiring the
longer "post-e-ul-o" for the sense of "afterward-person
= successor".

\> There are also quite interesting cases where compounds are compounded
> from parts that don't exist in isolation.  E.g German 'Himbeere'
> ('raspberry') and 'Brombeere' ('blackberry'). *'Him' and *'Brom' do
> not exist.  Quite frequent this is actually, and one would of course
> assume full opaqueness, but it's funny in compounds anyway where one
> would expect both parts to exist.  English 'raspberry' may be similar
> (or does it indeed derive from 'to rasp'??).  And there was a
> discussion about 'strawberry' on this list once.
"cranberry" is the more canonical example; I think I've heard
this phenomenon called "cranberry morphs".

"huckleberry" is another English example.  And, in modern
English, "ruthless", though "ruth" exists as an archaic word
for mercy or pity.  There are scads of words which were
analyzable in Latin but incompletely analyzable in English,
as you point out, thus jokes about "feeling ept, ane and sipid".

In the history of Esperanto there are instances where
root words borrowed from another language were
reanalyzed as a recognized morpheme + a cranberry morph,
which was then used on its own and ceased to be a
cranbery morph; e.g. "frauxlo" ("bachelor") back-formed from
"frauxlino" < Fraulein.  A little bit as though "cran" were
extracted and used as a stand-alone adjective to
signify some aspect of the taste or appearance of
cranberries.

> The longer one thinks about this, the easier it becomes to do
> something less abysmal... :-)
?

--
Jim Henry
http://www.pobox.com/~jimhenry/review/log.htm
New book reviews 2007/4/25

Reply