Carol Myers-Scotton's reply to Naima Boussofara Omar

[Harold F. Schiffman]

Forwarded from Linguist-List: Carol Myers-Scotton <CarolMS at gwm.sc.edu>
 reply to review of Contact Linguistics in Linguist-List 14.1077


In her review of my book Contact Linguistics: Bilingual Encounters and
Grammatical Outcomes (Linguist 14.1077), Naima Boussofara Omar (hereafter
NBO) gives a reasonable description of the contents of my book, chapter by
chapter, even if it is more of a "trees" overview than a consideration of
the "forest" I try to convey as constituting the grammatical aspects of
contact phenomena.  I am writing only to comment on NBO's critical
evaluation of the "trees".

(1) Overall, NBO seems "stuck in 1993"; that is, her criticisms seem
largely leveled at the my MLF model (of codeswitching) in the 1993 edition
of Duelling Languages (DL).  This edition dealt only with what I now call
"classic codeswitching", switching between two languages in the same
clause, but with only one of the participating languages clearly supplying
the abstract morphosyntactic frame.  The book currently under review
(Contact Linguistics or CL) revises and tries to clarify some claims made
in the 1993 book.  But, more important, CL goes beyond the first edition
of DL in many ways, both in terms of theoretical arguments and data
covered.  NBO largely ignores developments in the model in the last nine
years and new ideas that have been incorporated in the new book.  For
example, she misses entirely how the 4-M model's classification of
morphemes and its hypothesis about differential access of morphemes in
language production strengthens the claims of the MLF model of DL (cf.
Myers-Scotton and Jake 2000).  The 4-M model, and another recent model,
the Abstract Level model, also offer explanations for structures in other
contact phenomena, major goals in CL.

(2) NBO brings up again the old claim that the identification of the
Matrix Language (ML) is circular.  In CL (p. 59) I spell out again how to
identify the ML.  "The MLF model provides the two principles [the Morpheme
Order Principle and the System Morpheme Principle] as tests of the premise
of unequal participation and as a way to identify the Matrix Language.
If the terms of the principles, morpheme order and one type of system
morpheme, both are satisfied, then the Matrix Language can be identified
as that language."  If only one of the two participating languages meets
these criteria, it is the ML.  What is circular about that?

(3) NBO seems to think the MLF model was intended as universal.  It is
based on what I refer to as universally-present principles and processes
in contact phenomena (e.g. unequal participation of the languages
involved, different distributions of content and system morphemes, inter
alia).

But the model itself can hardly be universal.  In DL, the implication is
that the model applies only to classic codeswitching (defined above) and I
make this limit very explicit in CL.  Again, NBO is arguing against any
ambiguity in DL, but not in CL.  The MLF model certainly does not apply
without exception to NBO's examples in her own work from varieties of
Arabic that show structural overlap.  For such varieties, it is no wonder
that one cannot claim that only one of the participating varieties sets
the frame.  Such data are considered under the rubric "composite
codeswitching" in CL, but no claims are made about the details of the
structure.

(4) NBO claims that such examples are "clear violations of the System
Morpheme Principle".  As I just indicated, one would not expect the
principle to apply to such data.  But more important, NBO clearly does not
yet understand this principle.  In DL (1993: 82) the principle clearly
states "...all system morphemes which have grammatical relations external
to their head constituent ...will come from the ML."  This one type of
system morpheme is named the "outsider late system morpheme" in the 4-M
model.  This is clear In CL (2002) and other recent publications.  To my
knowledge, there are no (none) counter-examples to this principle in the
literature dealing with classic codeswitching.  I wish NBO would re-read
DL and read pp. 87-91 in CL.  She seems to have missed something.

(5) NBO also seems to have missed what CL has to say about the relation of
the ML to other designations for participating languages in bilingual data
or settings.  ML, as a label, may refer to the same language as the terms
"dominant" or "unmarked" language do, but it is not identified by the same
criteria.  The ML is a grammatically based construct.  The other terms are
not.  Please read p. 62 in CL.

(6) One more comment about the framework for contact phenomena developed
in CL: NBO complains about my use of "abstract constructs" that "are not
directly testable" to explain the data I consider.  Yes, I am guilty of
proposing such constructs to capture generalizations about the data.  NBO
misses the point that these constructs are part of, or lead to, testable
hypotheses.  This follows the deductive method.  Isn't this what
scientists are supposed to do?  Only abstract constructs that do not lead
to testable predictions are a weakness in a model.

(7) One more point.  No, I do not "invoke extra-linguistic factors when
there are not linguistic 'constraints' available."  True, in CL I mention
such factors as promoting attempts to incorporate elements from two
languages in the case of mixed (split) languages (because the speakers
want to have a code to identify their groups uniquely, with elements from
two languages).  Such factors are influential only in bringing about
conditions for bilingual data to emerge; they do not determine the
grammatical structure of contact phenomena.  Rather, the range of
potential grammatical structures that can emerge in any such phenomenon is
constrained by the cognitively-based principles and processes to which I
refer over and over.  These are universally available and the same ones
show up in contact phenomena over and over.  True, the details of the end
product vary with socio-psychological factors (e.g. the product in creole
formation is different from that in second language learning).  But they
vary only within the limits of the principles and processes just
mentioned..

(8) Finally, let me emphasize that I am disappointed that NBO concentrates
in her evaluation on her view of the "trees" in my models.  I make these
comments only to set the record straight on the true nature of these
"trees" (details of the models).  I had hoped that readers would pay more
attention to the claims about the nature of the overall "forest" in
contact phenomena that I argue for in Contact Linguistics.


References

Boussofara-Omar, Naima. (1999) Arabic Diglossic Switching in Tunisia:
An Application of Myers-Scotton's Martix Language Frame
Model. Unpublished PhD dissertation. Austin: The University of Texas.

Myers-Scotton, Carol. (1993, 1997). Duelling Languages: Grammatical
Structure in Codeswitching. (1997 edition with a new Afterword).
Oxford: Oxford University Press.

Myers-Scotton, Carol. (2002).  Contact Linguistics: Bilingual
Encounters and Grammatical Outcomes.  Oxford: Oxford University Press.

Myers-Scotton, Carol and Jake, Janice L. (2002).  Four types of morpheme:
Evidence from Aphasia, Codeswitching, and Second Language Acquisition.
Linguistics 38,6: 1053-1100.