Modelling Response Selection in Task Switching: Testing the Contingent Encoding Assumption

Schneider, Darryl W.; Logan, Gordon D. · 2013 · OpenAlex-citations

DOI: 10.1080/17470218.2013.843009

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This study investigates the "contingent encoding assumption" in task-switching models, which posits that response selection cannot begin until both the task cue and the target stimulus have been encoded. Schneider and Logan (2013) argue that this assumption is prevalent in computational models but has never been empirically tested. To evaluate its validity, the authors designed two experiments manipulating response congruency, stimulus order, and stimulus onset asynchrony (SOA). The core hypothesis was that if response selection can occur prior to cue encoding, it would be observable specifically for congruent targets (where the response is identical regardless of the task) presented before the cue. The experiments utilized a task-switching paradigm where participants judged target words as either living/nonliving (origin) or small/large (size). The researchers manipulated whether the cue appeared before the target (cue–target order) or after the target (target–cue order) at SOAs of 0, 400, or 800 ms. Crucially, participants were allowed to respond immediately after the first stimulus appeared, enabling the detection of "early responses" executed before the second stimulus was even presented. This design allowed the authors to test if response selection could proceed based on target information alone when the target was congruent across tasks. The results provided strong evidence against the contingent encoding assumption. In the target–cue condition, the response congruency effect increased significantly with longer SOAs, indicating that more time allowed for greater advance processing of congruent targets. Most critically, participants frequently produced early responses for congruent targets at long SOAs (800 ms), executing responses before the cue was displayed. This demonstrates that response selection can indeed begin and complete prior to cue encoding. Conversely, in the cue–target condition, the congruency effect decreased with SOA, consistent with standard task-switching dynamics. The authors modified an existing computational model by introducing "baseline evidence"—task-neutral evidence that drives response selection before full stimulus encoding. Simulations showed this modified model successfully reproduced the empirical data, including the occurrence of pre-cue responses. The findings imply that the contingent encoding assumption is invalid and must be removed from current task-switching models. The study suggests that cognitive systems can utilize task-neutral information from targets to initiate response selection, challenging the view that cues are strictly necessary prerequisites for any response processing. This insight refines the understanding of cognitive flexibility and control, highlighting that response selection is not strictly contingent on the joint encoding of cue and target, but can proceed based on partial information when response mappings align.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-25; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• task switching