Chicken and egg; sound and form

Hi all,

Apologies in advance to anyone to whom all this theory
stuff is already familiar - just skip to (***) for my own
comments.

I googled for "morphome", wishing to understand
David's point a little better, and tripped over this
fascinating document on MORPHOLOGICAL TYP-
OLOGY:
http://www.mml.cam.ac.uk/ling/courses/ugrad/morph_4.doc

The author, Paul Heggarty, has an attractive analysis
of morphological typology - rather than seeing discrete
classes of types, he extends the reasoning of Comrie,
to consider continuous variation along a number of
dimensions.  These include, most significantly, the
lexical-grammatical dimension, that underlies the
notion of "grammaticalisation", which is posited to
reduce the lexical (content) meaning of a word or
morpheme while increasing its grammatical (function-
al) meaning, along with a host of concomitant changes.

Under the general heading
7.3 CHANGE IN MORPHOLOGICAL TYPE,
Heggarty concludes his section
7.3.1 REPERCUSSIONS OF GRAMMATICALISATION ON MORPHOLOGICAL TYPE
thus:
"Note that these changes seem to be an exception to the neogrammarian principle that
sound change is exceptionless. It seems like that principle has to be qualified
to say that the context conditioning a sound change may have to include not
only the phonological and morphological context, but also even the lexical or
grammatical status, in that morphs with grammatical status undergo changes that
morphs with lexical status resist. Hence the noun will has not reduced to [l],
and the numeral one is not reduced to [«]. "

In section
7.3.2 THE COLLAPSE OF THE LATIN CASE SYSTEM:  CAUSE AND EFFECT?,
Heggarty continues:
"Let’s take an illustration of the link between grammaticalisation and sound change.

It has long been noted that the two fundamental changes on the way from Latin to
modern Romance happened at around the same time:
• the Latin case system ‘collapsed’, and Romance switched to marking the
same meanings by word order and prepositions instead, increasingly
grammaticalising particles that earlier had a more lexical meaning such as de
down from.
• phonetic distinctions between the sounds at the end of nouns, part (or all) of
the case suffixes, were lost.

In most Romance languages the basic singular form of nouns derives from the
accusative case form: note the [t] present in Italian notte, Spanish noche, and
pronunciations of nuit at least in old French (now visible only in the
spelling, and surviving in derivations like nuitée). These come from Latin
accusative noctem, not nominative nox.

The final -m has been lost entirely in all Romance, however, just as has the final
s of the nominative, so on most nouns there was no longer any distinction
between subject and object.

One big question, still unresolved is, which ‘caused’ which?
• Did the sound changes come first, and force the language into recruiting
prepositions to make the crucial difference clear between subject and object?
• Or did the grammatical structure change happen anyway, so that the distinct
case endings were no longer necessary, and were free to ‘erode’
phonetically?
• Or did the two process simply go hand in hand and feed off each other (perhaps
the best analysis)?"

(***) My thoughts:
When applying sound changes, whether using a
programmed "sound change applier" or doing it
by hand, I think it would be instructive to test
analytic daughter languages of a fusional parent
for usability, given two extreme paths and one
middle path (all sound changes first, all grammar
changes first, both sound and grammar gradually
changing together).

Of course, in the case of Latin or Germanic, we
do have considerable historical evidence as to
the actual sequence of many of the changes of
the last two or so millennia.  But starting from
a completely new fusional conlang, there's no
certainty that the evolution of its daughters
need parallel those two trees.  And it might
provide a test of how reasonable Heggarty's
alternative evolutionary paths are.

(Still haven't gained a clear notion of the
linguistic meaning of "morphome"!  Heggarty gives
the example of "l...d" as a morphome that subsumes
both "lead" and "led".  Google does point up other
meanings for "morphome" in computational biology
and in environmental engineering.)

Regards,
Yahya

--
No virus found in this outgoing message.
Checked by AVG Free Edition.
Version: 7.1.392 / Virus Database: 268.6.0/341 - Release Date: 16/5/06

Reply