OPTIMALITY THEORY: ISSUES AND CONSTRAINTS
Introduction
The last nine decades (1920-2010) has witnessed a substantial change in the nature of research in phonological theory. This change has been marked by the development of several theoretical frameworks both linearly and nonlinearly. In the 1990s, phonological investigation reached one of the most fascinating and challenging stages as phonologists began to chart a new direction for phonological theory. Following this search, Prince and Smolensky (1993) introduced the phonological framework called Optimality Theory (OT), which encourages a deflection from rule-based to constraint-based analysis. This theory, thus, tried to account for how the natural intelligence processes utterances as it relies on well-formedness constraints, the domain of universal grammar.
Optimality theory, to an extent, appears to be one of the relatively new theories of phonology. Before now, what existed as the most viable theoretical model was the generative phonology (with its subsequent modifications), which relies a great deal on phonological rules. These rules specify the relationship between the underlying forms and the surface representations. The major function of these phonological rules, according to Oyebade (1998), is to account for changes in the value of segments. This is achieved by showing the derivational sequence or path of an item in its journey from the underlying level to the surface level.
The standard phonological rule aimed at capturing grammatical generalizations in this model appears in this format: A B/C – D where a structural description (SD) delimits a class of inputs and a structural change (SC) specifies the operations that alter the input (Omachonu, 2011:91) Besides, one crucial assumption about this format is that any of the elements or variables (A,B,C,D) could be null or empty. And this makes it possible for the generative phonologists to capture different phonological processes within this format. It means therefore, that the central thrust of linguistic investigation, would be to focus on a system that can be used to explain or analyse inputs, according to Prince and Smolensky, in terms of ‘the possible structural descriptions of rules – and to define the operations available for transforming inputs – the possible structural changes of rules’ (cited in Omachonu, 91). This conception, as they argued, ‘has been jolted repeatedly by the discovery that the significant regularities were to be found neither in input configurations nor in the formal details of structure deforming operations, but rather in the character of the output structures…’
Prince and Smolensky went further to say that for the standard phonological rule format
A B/C – D to be worth pursuing, a theory which defines the class of possible predicates CAD (structural description) must be there, and another interesting theory which defines the class of possible operations A B (Structural Changes) must also be put in place. However, if these theories are not vigorous as indeed they might have proved to be in reality, Prince and Smolensky submitted that one must entertain or consider one of these two conclusions below:
- Phonology itself simply does not have much content, is mostly “periphery” rather than “core”, is just a technique for data compression, by the inevitable idiosyncrasies of history and lexicon; or
- The focus of explanatory action is elsewhere.
The latter appears a more viable conclusion judging from expert opinion (see McCarthy and Prince 1993; Oyebade 1997; 1998)
Optimality theory came as a child of necessity due to the flaws of the previous phonological models. Two related strands of evidence, according to Omachonu, inform the change in phonological approach that brought in OT. First, in the early days of generative phonology, it was found that the burden of the explanatory adequacy of most phonological events does not reside in the feature changing rules but with the phonotactic constraints that languages onbserve. The second strand of evidence emanates from the fact that since the late 1980s, evidence from syntactic analysis had shown that rule operated derivational grammars were already on their way out because constraint based grammars seem to have greater explanatory force. This, in the opinion of experts, is because constraint based grammars have brought the focus of the linguists and researchers alike back to the issue of identifying the properties of Universal Grammar. On the basis of this conviction, scholars like Prince and Smolensky argue that any serious theory of phonology must be committed to Universal Grammar which means that it must rely heavily on the well-formedness constraints. Thus, the goal of optimality theory is to develop and or examine thoroughly the way that representational well-formedness determines the assignment of grammatical structure.
In looking at the theoretical assumptions and principles of optimality theory, MaCarthy and Prince (1993b) assume that:
…the role of grammar is to select the right output form from among a very wide range of candidates, including at least all of the output that would be possible in any language whatsoever…language – particular rules, or procedures for creating representations have no role at all in the theory and the …burden of accounting for the specific patterns of individual languages falls on the well-formedness constraints. (4)
Besides, optimality theory, as observed by Oyebade, abandons two key presuppositions of generative phonology. First, it does away with the rule-formation of generative phonology and enthroned the function Gen which generates for any input a large space of candidate analysis. Here, phonotactic constraints become the present in any grammar. Optimality also abandons the widely held view that constraints are language – specific statements of phonotactic truth. Therefore, what marks individual grammars, in this model, is the order in which these constraints apply.
According to Omachonu, Optimality operates on five fundamental principles. These are:
Universality: Universal grammar provides a set CON of constraints that are universally present in all grammars.
Violability: Constraints are violable; but violation is minimal.
Ranking: The constraints of Con are ranked on a language – particular basis; the notion of minimal violation is defined in terms of this ranking. A grammar is a ranking of the constraint set.
Inclusiveness: The constraint hierarchy evaluates a set of candidate analyses that are admitted by very general considerations of structural well-formedness.
Parallelism: Best-satisfaction of the constraints hierarchy is computed over the whole hierarchy and the whole candidate set. There is no serial derivation.
To summarise the modus operandi of OT, it may be plausible to argue that these principles together with the basic assumptions have given birth to the three fundamental functions in OT. These are GEN, which generates all candidate sets to be evaluated, CON, which stands for constraints and EVAL, which does the evaluation of candidates (cited in Oyebade 1998:175)
One of the most famous illustrations of this framework is presented by McCarthy and Prince (1993) in their technical report on generalized alignment, which sought to show how OT can provide a principled treatment of the ways in which the edges of prosodic or morphological constituents align with one another. The authors give what is now considered a classic analysis of Tagalog language using Um-infixation. If we take the root /aral/ and add the /um/ affix, we get the inflected form /umaral/ meaning ‘teach’, but with the root /gradwet/, we get the inflected form /grumadwet/. McCarthy and Prince showed that the optimal output can be achieved for each variant by invoking two constraints, first a syllabl structure constraint, and the then an ‘edge-most’ or ‘align’ constraint, which deals with how far a linguistic entity like the Um-infixation can be from the edge of some unit (like the prosodic word or stem)
One can use this theory to describe the English syllable structure. The Un-prefix and the full-suffix have provided the optimal output in ‘Unfaithful’ with /faith/ as its stem or root. The relevant constraints are shown below:
- NO-CODA (i.e. syllables must be open)
- ALIGN-un, L, Stem, L (align left edge of affix with left edge of stem)
There is a violation of the open syllable rule in (a) above in the word /unfaithful/ as we find each syllable having a closed ended feature rather than being open-ended. When such a violation of constraints occur, the* is used to show the number of times of any violation. When candidates are rule out as optimal, an exclamation mark ‘!’ is placed in the cell showing the number of violations. In Unfaithful, there are three violations of the Open ended rule as each syllable of the word carries closed ended elements of /n/, /ɵ/ and /l/ respectively. Thus, in ‘unfaithful’, we can have the violations represented as ***!
- For a language learner, especially in English, there are certain phonemic problems that form a tollgate to the successful learning of a second language. A lot of factors account for these problems especially in the case of the Nigerian child trying to learn English.
One major problem is the phonological interference. As Spencer cited in Ogbuehi points out, “the phonology of a second language will almost always receive some imprint from the phonology of the mother tongue” (30) One can easily identify people from different ethnic groups when they speak English in Nigeria. This is not only peculiar to Nigerians alone. It is common problems of second language learners of the language from other areas of the globe, and even among those who speak the language as a first language. The Indians, for instance, find it difficult to pronounce words beginning with /v/, so ‘Vice-chancellor’ is pronounced as ‘wice-chancellor’.
David Abercrombie records the effect of phonological interference in the Old Testament story of Ephramites and the Gileadites over River Jordan. The fleeing Ephramites were asked to pronounce the word ‘shibboleth’ as a test word for distinguishing them from the Gileadites. The fugitive Ephramites pronounced ‘sibboleth’ because they could not pronounce ‘sh’. Consequently, “the Hebrew word ‘shibboleth’, meaning ‘ears of corn’, is now used to denote any particularity in speech of a geographical area.” (Ogbuehi 31)
Today there are many ‘shibboleths’ for identifying people from different areas of Nigeria. A few examples from the three major indigenous Nigerian languages will suffice to illustrate this point.
In the Igbo-speaking areas of Nigeria, the liquid /r/ is used interchangeably with the lateral /l/, thus producing such pronunciations as ‘rook’ for ‘look’; ‘bred’ for ‘bled’; ‘road’ for ‘load’, ‘Lun’ for ‘run’ etc. This problem is found in other indigenous Nigerian language speakers of English due to the fact that that there are the absence of certain phonemes from the indigenous languages that are present in English. The best the Nigerian speaker of English could do is to substitute one phoneme for another from his own language into the English language. Thus, phonemic substitution is a major challenge or problem area for the second language learner.
Another problem area is that of stress and intonation. While English is a stressed timed language, our indigenous languages are syllable timed which makes use of tonality to show voice quality. As such, Nigerian second language speakers of English give equal voice stress to English words which brings about pronunciation anomalies. For instance, Nigerians and Britons will pronounce the following words with unequal stress and with different stress shift:
| Nigeria | British |
| Coll’eage | ‘colleague |
| Jo’seph | ‘Joseph |
| Accommo’date | ‘accommodate |
| Demon’strate | ‘demonstrate |
Another problem area among some Nigerian speakers of English is that of the infusion of phonemic sound segments where they are not required. For instance, the use of the Anaptictic vowel is common among the three major tribes in Nigeria. Ball becomes ballu; table is tabulu; admit is admit; love is lovu; leg is legi, etc. This is somehow due to the fact that most indigenous languages have open-ended syllable structure unlike English that is both open and closed ended.
REFERENCES
Chomsky, N. The Sound Patterns of English. New York: Harper and Row, 1968.
McCarthy, J. and Prince, A. ‘Generalised Agreement’, Yearbook in Morphology, 1993: 79-154.
McCarthy, J. and Prince, A. Prosodic Morphology 1: Constraint Interaction and Satisfaction. Ms. University of Massachusetts, Amherst and Rutgers University, New Brunswick, New Jersey, 1993a.
Ogbuehi, C. English as a Second Language in Nigeria: An Introductory Text.
Nsukka: Magnet Business Enterprises, 2001.
Omachonu, G.S. Nsuk Texts in General Linguistics Vol 1. Enugu: Rossen Publications Ltd., 2011.
Oyebade, F. ‘Phonology II’. In Ore Yusuf (ed.) Introduction to Linguistics. Ilorin: University of Ilorin Press, 1992.
Contents