Creolization, Sound Change and OT: toward an explanatory framework

        Recent years have seen increased application of constraint-based analyses to historical
sound change (see e.g. Holt 2003), including phonological restructuring in Creoles (e.g.
Singler 2000, Uffmann 2003). In most Optimality Theoretic (OT) analyses, phonological
transformations are accounted for by constraint re-ranking, where surface change is seen as
the end product of innovative ranking of violable, universal constraints. While providing
rather straightforward description of diachronic language data, re-ranking analyses fall short
of true explanation (see e.g. McMahon 2000, Gess 2003). Application of OT to analyses of
creolization also fails at explanatory adequacy, although this has not been widely discussed or
even acknowledged. Constraint re-rankings such as (1) provide a grammatical snapshot prior
to and after Creole emergence; however, these neither refer to the mechanisms of change, nor
explain the trajectory of change, i.e. the events and causalities which gave rise to grammatical
restructuring. More importantly, re-ranking ignores fundamental questions about faithfulness
and input representations; learners of the emerging Creole could not be assumed to have
access to and subsequently categorize underlying representations of either lexifier or substrate
languages, rendering traditional faithfulness-markedness interaction models insufficient for
the analytical task at hand.
        This paper has two aims: the adaptation of Gess' (2003) explanatory model of sound
change, providing for lexical and productive evaluation in a sequence of idealized temporal
stages; and its application to Creole phonological restructuring, where distinction is made
between pidginization, emergence and stabilization. In contrast to `normal' sound change,
where transformation is incremental and occurs nearly imperceptibly over several generations,
creolization involves the relatively rapid emergence of an independent grammar, most often
following a period of pidginization during which the grammar and lexicon are reduced. For
discussion of issues particular to creolization and illustration of the model, I analyze the loss
of front rounded vowels, a common outcome in French Lexifier Creoles, as in (2). The loss
of distinction (e.g. [± round]) is explained as a product of interactions between effort
minimization and perceptual cue maximization during pidginization and Creole emergence.
Lexical instability is reflected at these stages in variable input representations (here [±
round] and [± front]), reflecting assumptions about the acquisition of meaningful
distinctions and their categorization as distinctive features by pidgin speakers and,
subsequently, the first generation of Creole speakers. Each stage of the model incorporates
lexical (declarative) and post-lexical or productive (procedural) evaluation, where
grammatical transformation is described by constraint re-ranking and explained by reference
to the causes of constraint re-ranking, e.g. phonetic principles, loss of featural distinction.
        The approach to phonological change and creolization outlined in (3) presents several
advantages over antecedent approaches. Firstly, it does not assume that all speakers share or
have access to the same representations at any diachronic period and describes change both
lexical and grammatical terms. Secondly, it provides a means by which to motivate constraint
re-ranking and representational change by reference to the causalities leading to these, here
the interaction of articulatory and perceptual forces during different stages, as well as
language-external factors. Thirdly, this analysis addresses the role of faithfulness in
creolization, considered to be the purview of both lexical and productive phonological
interactions, rather than solely the latter, and acknowledges the expanded scope and multiple
directionality of faithfulness during different stages. Finally, the model formalized here
provides a means by which to both describe and explain decreolization, as a quasi-reversal of
previous restructuring due to, among other causes, renewed or increased contact between the
speakers of the Creole and its historical lexifier.

(1)    Constraint Re-Ranking: Description of Creolization
       a. Lexifier
/y/             PRES(y)       *V[front, rd]
       y                            *
       i           *!
       b. Creole
/y/                             PRES(y)
            *V[front, rd]
       y           *!
       i                            *

(2)     Front Rounded Vowels in French Lexifier Creoles: Examples from Haitian
                              [syk] `sugar'                                 [sik]
        French        /y/                            Haitian         /i/
                              [jø] `eyes'                                   [je]
                      /ø/                                            /e/
                              [soe] `nun'                                   [s]
                      /oe/                                           //

(3)     Three-Stage Constraint Re-Ranking: Explanation of Sound Change (adapted from Gess 2003)
        FS = French speakers (also pidgin speakers)
        PS = Pidgin speakers (native speakers of various languages)
        CS = Creole speakers (first/native speakers)
        *RV = no round vowels, *FV = no front vowels, *FRV = no front round vowels
        Pres(lowF-cues) = preserve cues relevant to lowered formant frequencies (round vowels)
        Pres(F2-cues) = preserve cues relevant to tongue position (front ­ back)
        CAE = conserve articulatory effort (assume [y] is more effortful than [i] or [u])

        Stage 1 ­ Pidginization
                             PS: V [+hi, ± frt, ± rd]
Lexical representation                                          FS: /y/
Corresponding ranking        IdentV, *RV, *FRV, *FV             IdentV >> *RV, *FV, *FRV
Production ranking           CAE, Pres(cues)                    Pres(cues) >> CAE
                             [y], [i], [u] (variation)          [y]
Pronunciation
        Stage 2 ­ Early Creolization
Lexical representation     CS (majority):                       CS (minority):
                           V [+high, +front, ± round]           V [+high, ± front, + round]
Corresponding ranking *FRV >> IdentV >> *RV                     *FRV >> IdentV >> *FV
Production ranking         CAE >> Pres(F2-cues)                 CAE >> Pres(lowF-cues)
                           [i] >> [u]                           [u] >> [i]
Pronunciation
        Stage 3 ­ Stabilization
Lexical representation       /i/
Corresponding ranking        *FRV >> IdentV >> *RV
Production ranking           None with respect to Stage 1 /y/
Pronunciation                /i/

                                             Works Cited
Gess, R. 2003. On re-ranking and explanatory adequacy in a constraint-based theory of phonological
         change. In Holt (ed.), 67-90.
Holt, D.E. (ed.) 2003. Optimality Theory and Language Change. Dordrecht: Kluwer.
McMahon, A. 2000. Change, Chance and Optimality. Oxford: Oxford UP.
Singler, J.V. 2000. Optimality Theory, the Minimal-Word Constraint, and the Historical Sequencing
         of Substrate Influence in Pidgin/Creole Genesis. In Language Change and Language Contact
         in Pidgins and Creoles, J. McWhorter (ed.). Amsterdam: Benjamins, 335-351.
Uffmann, C. 2003. Markedness, faithfulness and creolization: the retention of the unmarked. In
         Phonology and Morphology of Creole Languages, I. Plag (ed.). Tübingen: Niemeyer, 3-23.