Optimality Theory

Optimality Theory models phonology as the selection of the best output form from a set of candidates, evaluated against a hierarchy of ranked, violable constraints.

Definition

A constraint-based framework in which surface forms are selected by evaluating candidate outputs against a language-specific ranking of universal, violable constraints.

Scope

This topic presents the constraint-based framework introduced by Prince and Smolensky. It treats the architecture of a generator that produces candidate outputs and an evaluator that selects the optimal one according to a language-specific ranking of universal constraints. It distinguishes markedness constraints, which penalize marked structures, from faithfulness constraints, which penalize differences between input and output, and explains how reranking the same constraints accounts for cross-linguistic variation. It also notes extensions and the treatment of opacity. The treatment is descriptive and analytic.

Core questions

• How does Optimality Theory select a surface form from candidates?
• What distinguishes markedness from faithfulness constraints?
• How does constraint reranking account for cross-linguistic variation?
• What challenges, such as opacity, does the theory face?

Key theories

Ranked violable constraints

The core proposal that grammars consist of a universal set of constraints ranked in a language-specific order, with the optimal output being the candidate that best satisfies higher-ranked constraints even at the cost of violating lower-ranked ones.

Markedness and faithfulness interaction

The account of phonological patterns as the resolution of tension between markedness constraints favoring unmarked structures and faithfulness constraints preserving the input, as elaborated in McCarthy's guide.

History

Optimality Theory was introduced by Prince and Smolensky in a manuscript circulated in 1993 and published in 2004, and was quickly extended by McCarthy and Prince and others. It became a dominant framework in phonology in the 1990s and 2000s, shifting the field from derivational rules to constraint interaction.

Debates

Opacity in a parallel model

Because the standard theory evaluates outputs in parallel, opaque interactions that were straightforward under rule ordering are difficult to capture, prompting extensions such as stratal and serial versions of the theory.

Key figures

• Alan Prince
• Paul Smolensky
• John McCarthy
• René Kager

Seminal works

• prince2004
• mccarthy2002

Frequently asked questions

What does it mean for a constraint to be violable?

In Optimality Theory constraints can be violated; the winning candidate is not one that satisfies every constraint but the one that best satisfies the higher-ranked constraints, even if it violates lower-ranked ones.

How does the same set of constraints produce different languages?

All languages are assumed to share the same universal constraints; languages differ in how those constraints are ranked, so reranking the same constraints yields different optimal outputs and thus different sound patterns.