Wednesday, April 7, 2010

Optimality Theory

A relatively new linguistic theory is Optimality Theory. Though it can be applied to many branches of linguistics, it is most common in phonology. Optimality Theory claims that the forms of a particular language are the result of an interaction among competing constraints.

The three basic components of the theory are Gen, Con and Eval. The first component, Gen, generates the list of possible outputs that the language can have. The second component, Con, provides the constraints which are used to pick the optimal candidate. The third component, Eval, picks the optimal candidate.

Constraints can be divided into two types. They are faithfulness and markedness. Faithfulness ensures that the input does not deviate too much from the output. Markedness ensures that the output does not violate the phonotactics of the language. For example, a markedness constraint against complex codas ensures that the language does not have a form with more than one consonant in the syllable coda.

Optimality Theory claims that the constraints are universal. This means that they must apply to every language. Another claim is that all infants possess every possible speech sound of every language but later filter out the speech sounds that their language does not need. Though the click sounds of the Bantu languages of southern Africa are rare in the languages of the world, Optimality Theory claims that people who never use these particular sounds could produce them in their infancy. That no evidence can be provided to support this claim is of little concern to the proponents of this theory.

The word "constraint" strikes me as a negative term. Optimality Theory dislikes the word "rule" because it wishes to portray itself as superior to rule-based phonology. Instead of rules, it uses tableaux with a series of constraints used to pick the optimal candidate.

Other words which could be used instead of "constraint" are "value", "preference", "condition", "obligation" and "requirement". Not every constraint is really a constraint. For example, the constraint which states that the input must equal the output is not truly a constraint but rather a necessity or obligation. Of course it can be reformulated as a constraint by stating that that output cannot differ from the input, but it is simpler to state that the input and output must be identical.

A big problem with Optimality Theory is that it provides little insight into how languages work. The constraints used to choose an optimal candidate and the ranking of the constraints are solely for the purpose of obtaining the correct result. It is in essence manipulation. As a Finnish professor once told me, it is similar to a chess game. In chess one manipulates the pieces on the board to mate one's opponents. In Optimality Theory one manipulates the constraints to choose the correct candidate. To me this seems unsatisfactory.

Optimality Theory is a modern linguistic theory which was proposed by American linguists in the 1990s. It became rather popular in Canada and the United States but never became very popular in Europe. This is probably because it reveals nothing new about language but is more like a game. It merely manipulates constraints to achieve a desired result. This theory is not likely to remain popular for long.

1 comment:

于名于倫 said...
This comment has been removed by a blog administrator.