The Contribution of Learning and Memory Processes to Verb-Specific Syntactic Processing

TY - JOUR

T1 - The Contribution of Learning and Memory Processes to Verb-Specific Syntactic Processing

AU - Ball, Lewis Vincent

AU - Mak, Matthew

AU - Ryskin, Rachel

AU - Curtis, Adam

AU - Rodd, Jennifer

AU - Gaskell, Gareth

N1 - © 2024 The Author(s).

PY - 2025/2/1

Y1 - 2025/2/1

N2 - Certain aspects of lexical knowledge can be primed by recent usage, with effects observed up to 24 hours later in some circumstances. Here, we used syntactically ambiguous sentences (“The man hit/chose the dog with the stick”) to explore the longevity of priming of syntactic structure. Some verbs provide a bias towards an instrument interpretation (the stick was used to hit the dog), whilst others are biased towards the modifier interpretation (the man chose the dog that possessed the stick). Experiment 1 revealed an effect of pre-existing verb bias on resolving syntactic ambiguities. In Experiment 2, we primed specific verbs towards their dispreferred interpretation in an exposure phase (e.g., hit was primed to the modifier interpretation). ~20 minutes later, the same verbs, along with unprimed verbs, were encountered in syntactically ambiguous contexts in a test phase. Exposure to the dispreferred interpretation in the exposure phase increased the preference for the same interpretation in the test phase, particularly for instrument-biased verbs. In Experiment 3, the exposure and test phases were separated by a ~12-hour interval that included sleep. No overall effect of exposure was found, but again a simple effect of priming was found for instrument-biased verbs. Finally, in Experiment 4 using a sentence completion task, we found that instrument-biased verbs had significantly stronger pre-existing biases, which we discuss as a possible explanation for the imbalance in priming between verb bias conditions. Our results suggest verb-bias priming is maintained over relatively long periods such as 20 minutes, and possibly as long as 12 hours, consistent with a contribution of episodic memory to maintenance of verb-specific syntactic biases.

AB - Certain aspects of lexical knowledge can be primed by recent usage, with effects observed up to 24 hours later in some circumstances. Here, we used syntactically ambiguous sentences (“The man hit/chose the dog with the stick”) to explore the longevity of priming of syntactic structure. Some verbs provide a bias towards an instrument interpretation (the stick was used to hit the dog), whilst others are biased towards the modifier interpretation (the man chose the dog that possessed the stick). Experiment 1 revealed an effect of pre-existing verb bias on resolving syntactic ambiguities. In Experiment 2, we primed specific verbs towards their dispreferred interpretation in an exposure phase (e.g., hit was primed to the modifier interpretation). ~20 minutes later, the same verbs, along with unprimed verbs, were encountered in syntactically ambiguous contexts in a test phase. Exposure to the dispreferred interpretation in the exposure phase increased the preference for the same interpretation in the test phase, particularly for instrument-biased verbs. In Experiment 3, the exposure and test phases were separated by a ~12-hour interval that included sleep. No overall effect of exposure was found, but again a simple effect of priming was found for instrument-biased verbs. Finally, in Experiment 4 using a sentence completion task, we found that instrument-biased verbs had significantly stronger pre-existing biases, which we discuss as a possible explanation for the imbalance in priming between verb bias conditions. Our results suggest verb-bias priming is maintained over relatively long periods such as 20 minutes, and possibly as long as 12 hours, consistent with a contribution of episodic memory to maintenance of verb-specific syntactic biases.

U2 - 10.1016/j.jml.2024.104595

DO - 10.1016/j.jml.2024.104595

M3 - Article

SN - 0749-596X

VL - 141

JO - Journal of Memory and Language

JF - Journal of Memory and Language

M1 - 104595

ER -