LING 301: Constraints

Markedness constraints

Penalise “marked” or complex structures in the output

Faithfulness constraint

Penalise differences between the input (underlying form) and output (surface form)

Positional faithfulness

Constraints relativised to certain contexts that have greater prominence

• e.g., word-initial, onset, root or lexical morphemes, etc.

Faithful candidate

A possible candidate that has no change from the input

• all candidate sets must have this

Optimal candidate

winning candidate

The emergence of the unmarked (TETU)

occurs when faithfulness to the input isn’t crucial, allowing markedness constraints to take precedence and causing the unmarked form to emerge as the winner

• e.g., when the optimal candidate incurs the same number of violation marks as a losing/marked candidate, so a markedness constraint becomes active to favour the winner

*COMPLEX

Assign one violation mark for every syllable margin (onset or coda) that contains more than one consonant

M

NOCODA

Assign one violation mark for every syllable ending in a consonant

M

ONSET

Assign one violation mark for every syllable that starts with a vowel

M

*VOICEDCODA

Assign one violation mark for every voiced obstruent that occurs in coda position

M

ALIGN-LEFT(ROOT, σ)

Assign one violation mark for every segment between the left edge of the root and the left edge of the syllable

M

CULMINATIVITY

Assign one violation mark for every prosodic word that has no stress

M

EXHAUSTIVITY

Assign one violation mark for every prosodic node that immediately dominates a prosodic unit more than one level below it in the prosodic hierarchy 

• every segment must be immediately dominated by something

M

*V[phrase

Assign one violation mark for every short vowel at the end of the phrase

M

*NC̥

Assign one violation mark for every sequence of a nasal consonant followed by a voiceless obstruent

M

IDENT()

Assign one violation mark for every x in the output that does not have the same y feature as its input correspondent

F

MAX-IO

Assign one violation mark for every segment in the input that has no output correspondent 

• no deletion

F

DEP-IO

Assign one violation mark for every segment in the output that has no input correspondent

• no epenthesis

F

CONTINGUITY

Assign one violation mark for every instance where contiguous segments in the input are not contiguous in the ouput

• no metathesis

• Segments that are adjacent in the input must also be adjacent in the output. (no skipping or breaking sequences apart).

F

LINEARITY

Assign one violation mark for every pair of input segments whose order is reversed in the output

• dont make segments switch places

F

UNIFORMITY

Assign one violation mark for every output segment that corresponds to more than one input segment

• no coalescence

F

ROOTUNIFORMITY

Assign one violation mark for every pair of distinct input segments belonging to the same root that correspond to a single output segment (i.e., fusion within the root)

• All segments that are part of the same root morpheme should have the same features across all its occurences in the output

F

OBSPLACEIDENT 

Assign one violation mark for every obstruent in the output whose place of articulation differs from its input correspondent

F

NASPLACEIDENT

Assign one violation mark for every nasal in the output whose place of articulation differs from its input correspondent

F

OBSMAX

Assign one violation mark for every obstruent in the input that has no correspondent in the output 

• i.e., every deleted obstruent

F

AGREE[voice]

assign one violation mark for every consonant cluster that does not have the same voice feature value

M

MAX[F]

Assign one violation mark for every input segment whose feature F is absent in its output correspondent

F

IDENT(OBSVCE)

Assign one violation mark for every obstruent in the output whose voicing differs from its input correspondent

F

MAX[CONT]

Assign one violation mark for every output segment whose [cont] feature has been deleted or altered compared to its input correspondent 

F

*V#

Assign one violation mark for every vowel that occurs at the end of a word 

M

*VTV

Assign one violation mark for every occurence of a voiceless obstruent between vowels 

M

ALIGN-R

Assign one violation mark for every morpheme or grammatical word (input) whose right edge does not align with the right edge of the prosodic word correspondent (output)

epenthesised sound is not part of the prosodic word —> dirsupts the alignment

ALIGN-um-L example

ALIGN-LEFT(il, stem) example

*NUC/PLOSIVE

Assign one violation mark for every plosive that occurs in the syllable nucleus 

Linguistic scale - hierarchy of constraints

NUC/PLOSIVE » NUC/FRICATIVE » NUC/NASAL » NUC/LIQUID

MAX-BR

Assign one violation mark for every segment in the base that is absent in the reduplicant

• enforces full reduplication

RED=σμμ

Assign one violation mark for every reduplicant that is not a heavy syllable

RED=σμ

Assign one violation mark for every reduplicant that is not a light syllable

RED = Ft

Assign one violation mark for every reduplicant that does not consist of exactly one prosodic foot

Coda-Cond

Assign one violation mark for every word-final consonant other than [x] and [y]

TROCHEE

Assign one violation mark for every foot that is not left-headed 

• not trochaic

IAMB

Assign one violation mark for every foot that is not right-headed

• not iambic

PARSE-Syl

Assign one violation mark for each syllable which is not

parsed to a prosodic word

19

FTBIN (Foot Binarity)

Assign one violation mark for every foot that does not contain two syllables

ALIGN-STRESS-RIGHT

Assign one violation mark for every syllable intervening between the main-stressed syllable and the right edge of the prosodic word

*CLASH

Assign one violation mark for every pair of adjacent stressed syllable in the output

*LAPSE

Assign one violation mark for every pair of adjacent unstressed syllables in the output

ALL-FT-X

Assign one violation mark for each mora that imtervenes between a foot and the X edge of the prosodic word

• requires every foot in a word stand at a particular edge (X) of the prosodic word

NONFINALITY

Assign one violation mark for every stressed syllable that occurs at the right edge (final syllable) of a prosodic word

• prevents final stresses (can leave some final syllables unfooted)

GrWd=PrWd

Assign one violation mark for every grammatical word that does not correspond exactly to one prosodic word 

LEFTMOST

Assign one violation mark if the head foot of a prosodic word is not the leftmost foot

RIGHTMOST

Assign one violation mark if the head foot of a prosodic word is not the rightmost foot

ALIGN-WD-L

Assign one violation mark for each mora or syllable intervening between the left edge of the PrWd and the first foot

• every prosodic word begins with a foot 

ALIGN-WD-R

Assign one violation mark for each mora or syllable intervening between the last foot and the right edge of the PrWd.

• every prosodic word ends with a foot

WEIGHT-TO-STRESS-PRINCIPLE (WSP)

Assign one violation mark to every heavy syllable that is unstressed in the output

ALIGN-L

Assign one violation mark for every segment that intervenes between the alignment of the left edge of the grammatical word and the left edge of the prosodic word

• prevents epenthesis outside the grammatical word

CONTINGUITY-BR

Assign one violation mark for every reduplicant in which segments that were contiguous in the base are no longer contiguous.