IMAGE 01h2LexDiffFin40.gif

10

01h2LexDiffFin

conspicuous, down to the little frills of language that have little to do with
communicative efficiency, such as gender distinctions. French also have a plural
ending, pronominal cases and relative pronouns inflected for case – all of
relatively direct Latin inheritance. Not only are gender and number relevant
categories in French, but there is also morphological agreement inherited from
Latin. Furthermore, there are suppletive adjectival comparatives and a great number
of Latin-derived prepositions. Also, verbs are conjugated mostly according to
(vulgar) Latin patterns, including such things as subjunctives, and they are even
divided into four conjugational classes – just like in Latin. Needless to say, such
as list could be made longer.

Meanwhile, most of this has been lost in Haitian. [...]

Speaking of grammatical gender, that is a particularly interesting feature. It has
persisted for thousands of years in virtually all non-creolised Indo-European
languages, while disappearing virtually overnight in all languages we would call
creoles (the system persists in some semi-creoles, such as Réunionnais). Could this
be a coincidence? I think not. Acquisition of grammatical gender seems to present
few problems to the L1 learner, but are notoriously difficult for non-native adults.
[...]

I am of course not suggesting that the genderlessness is what defines creoles, but
it is one of several features that are suggestive of a pidgin past.

IMAGE 01h2LexDiffFin41.gif

Emmanuel Faure wrote (07/02/01

[...] what I wanted to say is: before proclaiming anything "not being French at
all", or characterizing something as "typicallly French" (like the alleged
four-classes conjugation -- WHICH DOESN'T EXIST, is it clearer if I express it this
way?), the totality of French variation should be taken into account, not just
school grammar.

IMAGE 01h2LexDiffFin10.gif

Mikael Parkvall replied (08/02/01)

That's what I thought you said the first time, so it's still not much clearer. Are
you saying that the French inflexional paradigms for "manger" and "parler" are no
different from "partir" and "sortir"? Then why don't we get "sorté" and "parté"
rather than "sorti" and "parti"? Even if this exists in in one or two exceptional
varieties which I've never heard of, I can't recall having seen such a dialect
outside Europe.

IMAGE 01h2LexDiffFin21.gif

Salikoko S. Mufwene wrote (12/02/01)

The French gender system is intimately connected to the article system. The little I
have read about French in the 17th century suggests that its article system was not
as regularized as it is today. It was often missing in contexts in which it is
required today. Even in present-day French, I have read documents that would cause
me to flunk my French proficiency tests, as there are several cases where I would
use an article where it should be omitted. I'd be more cautious in developing any
arguments on the disappearance of gender in French creoles based on today's French
alone and on non-native knowledge.

IMAGE 01h2LexDiffFin05.gif

Mikael Parkvall (12/02/01)

From what I have seen, the French article system in the 17th century was pretty much
like it is today (again, overseas Frenches do not function radically different from
European dialects). To get to a stage of drastically different article usage, I
think you'd have to travel further back in time.

Parkvall (2000:15) thanks me heartily for sending him "multitudes of articles and
books", in fact everything I have ever published, alone or with Robert Fournier, on
issues being discussed here. If he had read any of it, he'd know that Magoua and
other koine varieties of French (what I call in my French publications "variétés du
français populaire") have absolutely no "morphological agreement inherited from
Latin", but do have a verb morphology drastically reduced to mere syncopation not
unlike creoles, inflectional affixes drastically reduced in numbers to one or two,
depending on the variety under consideration, and a rule of vowel reduction shifting

IMAGE 01h2LexDiffFin01.gif

10

Printed for Henri Wittmann <hwittman@cgocable.ca>

IMAGE 01h2LexDiffFin01.gif

11

01h2LexDiffFin

/é/ to /i/ in unstressed and word-final open syllables which means that "mangé" and
"sorté" are homophonous in Magoua. Koine French varieties also don't have any plural
endings, pronominal cases (pronouns such as mouén, toué, etc. are invariable), or
relative pronouns inflected for case (the universal complementizer is /k/ and case
must be inferred from any of the following agreement markers such as in /mwén k j/
<moi que je> "I who"). Some varieties such as Magoua have optionally post-nominal
demonstratives (l-ga sti-la "this guy") and post-nominal possessives (l-char à-lui i
màch hén? "his car works (nicely), isn't it?"). All varieties have "bimorphemic"
interrogatives (ki sa k i màlàd-màlàd? "who's being sick like that?") and productive
reduplication (mèy-mèy "very skinny", ti-ti "very tiny", àlé méné-méné "to take a
stroll", fèk manjé-manjé "eating continuously"). Several of these features have been
common knowledge in romanistics at least since Meyer-Lübke and Meillet (cf. Wittmann
1998a for references) and there's no excuse for creolists to ignore them when dealing
with French-derived language varieties, whether creole or koine.

His case for gender seems to be based entirely on the "article system" of historical
French as it survives in modern Standard French. In Parkvall (2000:78), on the other
hand, he bases his conclusion which goes to admit that the post-nominal article /-la/
of creole French cannot be substrate-induced on Fournier (1998; with further
references extrapolated from there) where alternate strategies for definitenes are
said to have developed because the etymological "article system" did NOT survive in
koine French. Though the basically Greenbergian idea that articles grammaticalize to
noun class clitics or affixes as new articles emerge from demonstratives in a cycle
of categorial downgrading alternating with compensatory morphological renewal has
become standard fare in historical linguistics (cf. Wittmann 1983), Parkvall in his
extensive treatment of noun class phenomena (2000:81-83) is either unaware of this or
such input has been weeded out in a cafeteria-principled way.

As we have already seen, Magoua and related dialects have three noun class proclitics
/l-, (l)à-, (l)é-/ pairing into two genders, Z ("masculine") and L ("feminine").
Noun class gender controls subject agreement which is enclitique to the preceding
subject head DET (when DET is filled, the clitic fuses with the preceding /la/),
proclitique to what follows (TMA, AGRO or whatever predicate head there may be), and
controls object agreement (direct, indirect or locative) which is proclitique to the
verb. Definiteness is marked by subject-object agreement or by filling the DET head
position of the Determiner Phrase (the noun phrase + DET) with /la/, optionally with
agreement, obligatorily whenever clitic agreement strategies are unavailable.

Comparatively, Swahili (a Bantu language) has noun class gender controlling agreement
(subject, direct object, indirect object, locative) in a way very much like Magoua:

(9)aan-a-*(u-)onam-ti(gender 2, class 3 SG
"I-PRESENT see the tree" )
(9)abn-a-*(ki-)onaki-tabu(gender 4, class 7 SG)
"I-PRESENT see the book"

(9)ban-a-*(i-)onami-ti(gender 2, class 4 PL)
"I-PRESENT see the trees"
(9)bbn-a-*(vi-)onavi-tabu(gender 4, class 8 PL)
"I-PRESENT see the books"

(10)a ma-tunda*(zi-)*(Ø-)chunga(gender 6, class 10b PL)
CL-fruits*(AGRS-)(*CL-)bitter
"The fruits are bitter" (stative predicate)

(10)b ma-tunda*(ni-)*(ma-)chunga(gender 6, class 10b PL)
CL-fruitsAGRS-*(CL-)bitter
"The fruits are bitter ones" (non-stative DP predicate)

However, Swahili (and Bantu in general) has quirks and frills which indeed have
"little to do with communicative efficiency":

(11)a.Swahili has 16 noun class prefixes (surviving from the original 18 of Bantu)
pairing into 7 genders. If we were to add nouns with non-third-person reference
(nouns referring to "I", "we", "you") paired into genders, we would have 9 genders
controlling each one its own set of gender agreement prefixes whereas we'd have in
Magoua 4 genders controlling sets of clitics marking gender incidentally.

IMAGE 01h2LexDiffFin01.gif

11

Printed for Henri Wittmann <hwittman@cgocable.ca>

IMAGE 01h2LexDiffFin01.gif

12

01h2LexDiffFin

(11)b.In Swahili, agreement marks definiteness only in object positions and with
inanimate nouns that do not belong to gender 1; in all other cases, agreement is
obligatory and based on gender only.

(11)c.Magoua agreement is not sensitive to nominal class membership with
non-definite subject agreement, indirect object agreement, locative agreement and
reflexives.

(11)d.Agreement markers are "bound" affixes in Swahili implying fusional inflective
morphology, clitics in Magoua. [For Magoua, conventional /-/ is to be interpreted as
syntactic clitic if it antecedes the item it cliticizes on (phonologically, the
clitic might enclitic), affixal when it follows;in Swahili (and in Bantu in
general), /-/ has always affixal interpretation.]Absence of fusional inflection was
McWhorter's yardstick number one out of three to measure the "creole prototype" (cf.
Wittmann 1999). [McWhorter, in the publication I'm referring to, just talks about
inflection implying fusional inflection via bound affixes. Any language, even the
most isolating one, would have to have inflection, fusional, clitic or isolating, in
order to qualify as a natural language. Any non-lexical "functional" item is
inflective by definition. Isolation-fusion is a matter of degrees along a continuum
which doesn't contribute significantly to typological differences. Movement rules or
strictly fixed word order are conceivable anywhere along that continuum.]

(11)e.

Gender 7 of Swahili has only one intrinsic member (-hali "place").

(11)f.Noun class prefixes are the only prenominal material there can be in Swahili
which gives the illusion they're "agglutinated" whereas in most other languages
PRODUCTIVE noun class markers have to compete with other material in the same
syntactic position.

(11)g.In present tense contexts, the zi/ni alternation of (10a)/(10b) corresponds
to the i/sé alternation of Magoua in (7a)/(7b). [In (8b), AGRS agrees superficially
with the class clitic of the subject but is coindexed semantically with the object
DET;in (10b), the same effects are achieved almost identically, except that object
"DET-ness" surfaces as a class prefix which, though agreeing superfically with the
subject, retains in this context a "functional load" inherited from its past as an
article.]Though this seemingly reflects subject agreement marking historically
derived from an earstwhile copula, Swahili has overt copula phenomena emerging in
typical copula positions with non-present tense reference just as they would in
Hebrew and Arabic (I can't go into details here on the difficult subject of the
copula in Swahili). This is not so for Magoua.

Presenting the Magoua paradigm of the old "copula" in pre-Magoua garb would look
something like this:

(12)aje_suis tait
(12)bt_es tait
(12)cail_est tait
(12)cbelle_est tait
(12)ccc_est tait
(12)daon_est tait
(12)dbc_est tait
(12)eaon_est tait
(12)ebson tait
(12)ecc_est tait
(12)fason tait
(12)fbc_est tait

where <_> represents synchronically opaque fusion. Note here that only verbs,
including auxiliaries, can support object agreement clitics. Since Magoua does not
have an auxiliary TO BE or a syntactically active position supporting copula
phenomena, non-verbal predicate heads support object agreement only to the extent
this can be achieved through suppletive object agreement as discussed with (7b).
Non-koine non-creole varieties of French (such as Standard French, Acadian French or
any other pre-koine dialect of Gallo-Romance) retain a fully conjugated auxiliary TO
BE supporting productive object clitic attachment as in (13):

IMAGE 01h2LexDiffFin01.gif

12

Printed for Henri Wittmann <hwittman@cgocable.ca>

[made with GoClick]