Search Abstracts | Symposia | Slide Sessions | Poster Sessions
An MEG Investigation of Adjective Order Preferences as a Syntactic Constraint in English
Poster Session A, Friday, September 12, 11:00 am - 12:30 pm, Field House
This poster is part of the Sandbox Series.
Boxuan Li1, Amilleah Rodriguez1, Liina Pylkkänen1; 1New York University
When multiple adjectives are used in a noun phrase, as in beautiful blue dress or large brown bag, their relative order tends to follow a consistent pattern across speakers. This pattern, known as adjective ordering preferences, reflects a shared sense of which adjectives should come first. Linguists have argued that these preferences arise not from surface-level habits but from an underlying syntactic structure. According to this view, different adjective types, such as size, color, and material, occupy distinct syntactic positions within the noun phrase, forming a fixed hierarchical order (Cinque, 1994; Scott, 2002). For example, we prefer big red ball over red big ball because "big," as a size adjective, appears earlier in the syntactic structure than "red," a color adjective. This syntactic account contrasts with alternative explanations based on semantics (Scontras et al., 2017). If adjective order preferences are syntactic in nature, then violating them should produce behavioral and neural effects like those elicited by uncontroversial syntactic violations, such as violations of phrase structure. In this study, we tested this prediction using behavioral and MEG measures, asking whether the brain treats adjective order violations as genuine syntactic violations. Our data collection is ongoing and will be completed before the annual meeting. We designed three combinatory test conditions and two non-combinatory control conditions. Each test condition consists of a noun phrase composed of two adjectives and a noun, while in the control conditions, a list of three plural nouns is presented. The test conditions were: (1) Right order: adjective phrases that followed canonical English ordering preferences; (2) Wrong order: adjective phrases that violated the preferences by reversing the two adjectives (e.g., red big ball) and (3) Phrase structure violation: phrases where the noun appeared ungrammatically between the adjectives (e.g., big ball red). The control conditions were: (1) Related list: three semantically related plural nouns in a list (e.g., cats dogs rabbits); (2) Unrelated list: three unrelated plural nouns in a list (e.g., desks clouds apples). In a Rapid Parallel Visual Presentation (RPVP) paradigm, noun phrases and word lists are presented for 300 ms and followed by a 500 ms blank screen, creating a total of 800 ms of processing time. This was then followed by a task stimulus, which either matched or mismatched the first stimulus, with the participant’s task being to indicate whether the two stimuli were identical. We length matched the test stimuli and the following task stimuli. In mismatch trials, one word was replaced in the task stimuli with a same-length distractor. We measured reaction time and accuracy as behavioral indices and track neural responses with MEG, with the first test stimulus being the target of MEG analysis. If adjective order violations are processed similarly to phrase structure violations, we expect similar behavioral and neural responses for the wrong-order and phrase-structure violation conditions. However, if adjective order violations are perceived as grammatically correct, their responses should resemble those of the right-order condition in the earliest stages of structure perception (Fallon & Pylkkänen, 2024; Flower & Pylkkänen, 2024).
Topic Areas: Syntax and Combinatorial Semantics,