The study investigates whether sentence planning in writing and speech shares a common scope. Across three eye-tracking experiments using image-description tasks, we found that both modalities obligatorily plan the syntax of the entire initial noun phrase before production onset, but typically do not retrieve non-initial nouns lexically in advance. These results provide the first systematic evidence that phrasal planning scope is modality-independent and primarily syntax-driven, challenging assumptions that extended planning reflects speech-specific demands.
The study examines structural priming in Tagalog, a language with a rare symmetrical voice system, to test whether word order priming is sensitive to verb voice morphology. Across three experiments, we found that priming is strictly voice-specific: word order priming occurs only when prime and target verbs share the same voice, and is stronger in the agent voice than in the patient voice, which shows limited flexibility. These results challenge residual activation accounts and support learning-based models that posit voice-specific syntactic representations, highlighting the importance of typological diversity for psycholinguistic theory.
The study tests whether multisentence text production is best understood as a serial or parallel process by analysing over five million keystroke intervals from six diverse writing datasets. Using Bayesian hierarchical models, we show that a two-distribution mixture model—reflecting parallel planning and execution—consistently outperforms single-distribution models aligned with the serial view. These findings provide robust evidence that writers typically plan upcoming text concurrently with ongoing output, challenging long-standing assumptions in writing theory.
| Paper | Originality | Rigour | Significance |
|---|---|---|---|
| Roeser et al. (2019) | First systematic experimental comparison of sentence planning in writing and speech, demonstrating that planning obligatorily extends to the entire initial noun phrase in both modalities. | Three (appropriately powered, controlled, counterbalanced) eye-tracking experiments with Bayesian mixed-effects modeling to ensure robust inference (code and data is openly available). | Establishes that phrasal planning scope is a modality-independent property of the language production system, challenging assumptions that extended planning reflects speech-specific constraints and informing theories of grammatical encoding. |
| Garcia et al. (2023) | First systematic evidence that structural priming in Tagalog is voice-dependent, revealing that verb morphology constrains syntactic priming in ways not predicted by existing models. | Three (appropriately powered) controlled experiments (one pre-registered), Bayesian mixed-effects modeling, and open sharing of data, code, and materials. | Challenges widely accepted residual activation accounts and supports learning-based theories of grammatical representation, offering crosslinguistic insights that reshape assumptions about universality in sentence production. |
| Roeser et al. (2025) | First empirical evidence to challenge serial models of multi-sentence text production using an implementation of a parallel account in Bayesian mixed-effects mixture models. | Multi-dataset analysis (> 5m keystrokes produced by 967 participants of various demographics), formal (Bayesian) model comparison using leave-one-out cross-validation, simulation checks to rule out overfitting, and transparent reporting of code and data. | Challenges a dominant theoretical assumption in writing research, with implications for cognitive models of language production and interpretation of process data across diverse writing contexts. |
Robust experimental design and analysis exemplifying best practices in experimental psycholinguistics combining
Reshape theoretical understanding of language production: