Self-segregating morphology again - in simpler terms, with list of methods
|From:||Jim Henry <jimhenry1973@...>|
|Date:||Monday, April 17, 2006, 16:58|
On 4/15/06, Taka Tunu <takatunu@...> wrote:
> Yohemiwi kana kero hipi we kame time hipi wi yapehewa "tara roko hate matu mihi"....
> (Translation: There are several thread on this list about "self-segregating
> morphology" that I cannot understand because these threads use plenty of words
> pertaining to maths or computer sciences that I don't understand. And yet.)
If I recall correctly from the last time I looked at Tunu,
it has a more or less self-segregating morphology.
Self-segregating morphology simply means that the
rules about what shapes a word, or a morpheme, can
have, and how they're put together into compounds
(if the language isn't purely isolating), are such that
you can always tell where one morpheme, or word,
leaves off and the next one begins. Some of the methods
for doing so are complex and require some maths
(e.g., Plan B, and Jörg and Ray's redesigns of it);
others are very simple.
Maybe it would make sense to collect a list of methods
for self-segregation. I'm thinking about this a lot lately
because I want my next conlang to have such a feature.
Tceqli has perhaps the simplest such rule: one set
of phonemes (the fricatives, plosives and affricates)
are the beginning set, and another set (the vowels,
nasals, liquids and semivowels) are the following set.
A morpheme consists of one or more phonemes from
the beginning set, followed by one or more phonemes
from the following set. I'm thinking of stealing this
for my next conlang, but modifying the sets of phonemes
in each set.
In Ilomi a word always begins and ends with a vowel,
and two vowels in a row never occur within a word; a morpheme boundary
in a compound word has an /n/
between two vowels, and /n/ by itself never occurs in root
Plan B and the redesigns of it the consonants grouped
into sets which indicate the number of syllables in the
word. So a word starting with one of a certain set
of consonants will be one syllable, a word starting with
one from another set will be two sylllables, and so forth.
(This is a major simplification.) Yahya recently posted
(probably tongue in cheek) a more verbose self-segregation
scheme with the same basic principle.
Lojban has a complex set of self-segregation rules
that I don't remember well (if indeed I ever studied them
well enough to know them).
gzb started out having a self-segregating morphology,
but in one of the early revisions I decided this wasn't
an important feature for my purposes and in the course
of revising the phonotactics to allow more monosyllabic
words, I made it so there are a few theoretically ambiguous
morpheme boundaries. There, I had certain vowels
that can only occur in the initial or medial syllable of
a polysyllabic root, and others that can occur only in
the final syllable of polysyllabic root or in a monosyllabic
root. This is still the case. But now that a word can
begin with almost any consonant followed by a liquid or
semivowel, and can end in a semivowel followed by
a nasal -- a few boundaries are ambiguous, like
"tĭwmwĭl" ("bed") - which could be tĭwm-wĭl (?-sack)
or tĭw-mwĭl (furniture-sleep) (but in practice the first
parsing is impossible because there is no root "tĭwm").
For this reason I still use hyphens between morphemes
in the online version of the language, though I've dispensed
with them for a couple of years in my journal writing.