A Comparative Study of Fuzzy Aggregation Strategies in Modeling the Syntax Pragmatics Interface

Traditional approaches to grammaticality often rely on binary judgments, overlooking the gradient and context sensitive nature of linguistic acceptability, particularly at the syntax pragmatics interface. Fuzzy grammar models attempt to capture this nuance by assigning degrees of membership; however, most employ a fixed aggregation strategy, typically the minimum operator, to combine syntactic and pragmatic evaluations. This study addresses the research question: How do different fuzzy aggregation operators influence grammatical acceptability judgments at the syntax pragmatics interface? We introduce a comparative framework that systematically evaluates four fuzzy aggregation operators, namely minimum, product, arithmetic mean, and weighted average. Using a pilot dataset of five English sentences rated on a 7-point Likert scale, we normalize the ratings into fuzzy membership values and apply each operator to compute integrated acceptability scores. The results reveal that operator choice significantly influences final judgments, especially in cases where one linguistic dimension is strong and the other weak. While the minimum and product operators enforce strict conjunction, the mean and weighted average provide more flexible and interpretable assessments. These findings highlight aggregation as a critical design factor in fuzzy grammar models, with practical implications for educational NLP systems, grammar assessment tools, and context aware language technologies. Overall, this work contributes a methodological extension to fuzzy grammar modeling and offers guidance for tailoring aggregation strategies to specific linguistic and computational applications.