Featural Optimality Theory: A Computational Model

Constraint

Meaning

MAX-F(L)

Every instance of [L] in the input has a correspondent feature in the output.

DEP-F(L)

Every instance of [L] in the output has a correspondent feature in the input.

MAX-A(L)

Every association to an instance of [L] in the input has a correspondent association in the output.

DEP-A(L)

Every association to an instance of [L] in the output has a correspondent association in the input.

MAX-F(H)

Every instance of [H] in the input has a correspondent feature in the output.

DEP-F(H)

Every instance of [H] in the output has a correspondent feature in the input.

MAX-A(H)

Every association to an instance of [H] in the input has a correspondent association in the output.

DEP-A(H)

Every association to an instance of [H] in the output has a correspondent association in the input.

ALIGNFEATURE(L, right)

For each token t of feature type [L] such that t is not floating, t is on a path with the right-most anchor in the prosodic word.

ALIGNFEATURE(H, right)

For each token t of feature type [H] such that t is not floating, t is on a path with the right-most anchor in the prosodic word.

ALIGNFEATURE(L, left)

For each token t of feature type [L] such that t is not floating, t is on a path with the left -most anchor in the prosodic word.

ALIGNFEATURE(H, left)

For each token t of feature type [H] such that t is not floating, t is on a path with the left -most anchor in the prosodic word.

ALIGNPRWD(H, right)

The right-most anchor a for feature type [H] in the prosodic word is on a path with a token of feature type [H].

ALIGNPRWD(L, right)

The right-most anchor a for feature type [L] in the prosodic word is on a path with a token of feature type [L].

ALIGNPRWD(H, left)

The left-most anchor a for feature type [H] in the prosodic word is on a path with a token of feature type [H].

ALIGNPRWD(L, left)

The left-most anchor a for feature type [L] in the prosodic word is on a path with a token of feature type [L].

FEATURELINEARITY

The lexical linear order of features is consistent with the linear order of features derived through association to prosodic structure.

GROUNDNEGATIVE(H, L)

if [H] then not [L].

GROUNDNEGATIVE(L, H)

if [L] then not [H].

NOFLOAT(H)

Each token of feature type [H] is associated.

NOFLOAT(L)

Each token of feature type [L] is associated.