Questions: Constraint Interaction and Ranking in Optimality Theory

Question 1

Language A always deletes word-final consonants; Language B always preserves them. An OT analyst claims both languages possess the same constraints NOCODA and MAX-IO. How can opposite outputs arise from the same two constraints?

• A. Language A has NOCODA but lacks MAX-IO; Language B has MAX-IO but lacks NOCODA
• B. In Language A, NOCODA outranks MAX-IO, so coda deletion is preferred; in Language B, MAX-IO outranks NOCODA, so consonants are preserved even at the cost of a coda violation
• C. Language A applies NOCODA only in word-final position as a phonological rule; Language B's rule is different
• D. The constraints are the same but Language B applies them only to stressed syllables

Answer: (B)

This is the central explanatory mechanism of OT: the same universal constraint inventory produces different phonological systems through different rankings. NOCODA (prefer no coda consonants) and MAX-IO (preserve input segments) conflict when an input has a coda. Which constraint wins depends on ranking. Language A prioritizes surface syllable well-formedness over faithfulness to the input; Language B does the opposite. No language-specific rules are needed — only a language-specific ranking order.

Question 2

In an OT tableau, how is the winning candidate selected from a set of competitors?

• A. The candidate with the fewest total constraint violations across all constraints wins
• B. The candidate with no higher-ranked constraint violated more than any competing candidate wins — a single fatal violation of a top-ranked constraint eliminates a candidate regardless of how few lower-ranked violations it has
• C. Candidates are scored on a weighted sum of violations; the highest score wins
• D. All constraints must be satisfied; if no candidate satisfies all constraints, the input is ungrammatical

Answer: (B)

OT uses winner-take-all evaluation at each constraint level, not a sum of violations. Higher-ranked constraints take absolute priority. If candidate A violates a constraint ranked above any constraint that candidate B violates, B wins — even if A has fewer total violations overall. This is why the tableau procedure works column by column from left (highest ranked) to right: a candidate can be eliminated by a single fatal violation at the top, regardless of its performance on lower-ranked constraints.

Question 3

Two languages can exhibit entirely different phonological processes while sharing an identical inventory of universal constraints, if their constraint rankings differ.

• T. True
• F. False

Answer: True

This is OT's core prediction about cross-linguistic variation. English and Japanese have dramatically different syllable structures — English tolerates complex onsets and codas; Japanese strongly prefers open CV syllables and inserts epenthetic vowels. In OT terms, they share the same constraints (ONSET, NOCODA, MAX-IO, DEP-IO, etc.) but rank them differently. Japanese ranks syllable-structure markedness constraints very high; English ranks faithfulness constraints higher. The factorial typology claim holds that all possible human phonological systems are predicted by the set of possible rankings of a universal constraint set.

Question 4

In Optimality Theory, a language that deletes consonants has a 'delete' rule that a language preserving those consonants lacks.

• T. True
• F. False

Answer: False

OT replaces language-specific rules with language-specific constraint rankings. Both languages possess the same constraints — including MAX-IO (don't delete) and whatever markedness constraint motivates deletion. The language that deletes has simply ranked the markedness constraint above MAX-IO; the language that preserves has the reverse ranking. No language-specific deletion rule is required or posited. This is what makes OT a theory of linguistic typology: variation comes from ranking differences, not from different grammars being built from different primitive operations.

Question 5

Why does OT predict that the set of possible human phonological systems corresponds to the set of possible rankings of a universal constraint inventory, rather than requiring separate rule inventories for each language?

Model answer: Because the same finite set of constraints, ranked in different orders, generates different outputs from the same inputs without language-specific rules. Each ranking is a grammar; each possible ranking of n constraints is a possible human language. This means OT's typological predictions are generated mathematically — by permuting rankings — rather than by stipulating different rules for each language. The theory explains why cross-linguistic variation is bounded (only outputs achievable by some ranking of the universal set are possible) and why languages differ systematically rather than arbitrarily.

This is OT's boldest theoretical claim, called factorial typology. If there are n constraints, there are n! possible total rankings, each predicting a different phonological system. Researchers test the theory by checking whether observed cross-linguistic patterns are a subset of the predicted typological space. Languages outside the predicted set would falsify OT. This makes the theory empirically testable in a way that language-specific rule systems are not.