The Adaptive Threshold Model: Communicative Competence as Conditional Performance

This paper advances the Adaptive Threshold Model (ATM) as a theoretical framework for understanding communicative competence as a conditionally accessible phenomenon rather than a static inventory of abilities. Building on long-standing work in communicative competence, pragmatics, and second language acquisition, the paper identifies persistent problems in the field, including construct fragmentation, intra-speaker variability, and the limited predictive validity of global proficiency measures. ATM addresses these issues by proposing four interacting performance thresholds—linguistic density, cognitive load, contextual support, and strategic control—that govern when underlying communicative competence can be successfully realized in real time. The model is articulated with explicit theoretical restraint: it does not posit new competence components, nor does it function as a theory of learning, acquisition, or long-term change. The exposition proceeds by defining the proposed thresholds and their governing principles, illustrating their interaction through analytically isolated violations, and clarifying how optimal performance emerges through task-appropriate alignment rather than maximization. The paper situates ATM in relation to existing research traditions, including Gricean pragmatics, task-based performance theory, and work on language attrition, in order to delineate its explanatory scope and prevent conceptual overextension. Finally, the model’s implications for understanding fluency–accuracy trade-offs, fossilization, and language assessment are explored, with particular emphasis on threshold-sensitive performance profiling as an alternative to global proficiency scores. Taken together, the paper presents ATM as a unifying, performance-oriented framework capable of explaining systematic variability in communicative success across contexts while remaining compatible with established theories in linguistics and applied linguistics.