|Abstract:||This paper explores the phenomena of ‘persistence’ and ‘emergence’ in the patterns of contrast, distribution, and alternation in the phonology of Malayalee English, a transplanted second language system, and argues for the need to supplement the theory of constraint interaction in Optimality Theory with an explicit theory of constraint generation. The proposal involves two parts. First, when two or more constraints share a common core but vary in the details of their manifestation, their redundancy can be eliminated by deriving them from an underspecified constraint core. Adding further specifications of domain, locus, trigger, or outcome value, would yield the fully specified constraints. The constraints derived from the same constraint core would form a constraint homologue, providing a basis for the study of invariance and probable variations in language typology, language change, and language contact. Second, to capture cross-linguistic probabilities in the ranking of constraints, we propose that each markedness constraint homologue be paired with its faithfulness counterpart, with a universal default ranking relation. Such a constraint pair with default ranking expresses a universal tendency. In a weak tendency, faithfulness outranks markedness; given a language, reversing the default ranking would activate the tendency. In a strong tendency, markedness outranks faithfulness; reversing the ranking would deactivate the tendency. Strong tendencies correspond to the unmarked state of affairs in the SPE sense of ‘unmarked’. Our analysis also reveals that, contrary to expectation in language contact, some patterns in ME that it shares with neither its substrate nor its superstrate sources illustrate the emergence of marked configurations of facts. We suggest that these marked structures can be explained as the resolution of conflicting pulls from the parent languages within the space provided by universal grammar.