Processing Relative Clause Extractions in Swedish

Area of Science:

• Linguistics
• Psycholinguistics
• Syntax

Background:

• Extraction from relative clauses (RCE) is typically a strong island violation in many languages.
• Swedish allegedly permits RCE, posing a challenge to universal island constraints.
• Existing surface accounts for Swedish RCE have been insufficient.

Purpose of the Study:

• To investigate the processing of Swedish extraction from relative clauses (RCE) using eyetracking.
• To determine if Swedish RCE processing aligns with strong island violations or other syntactic structures.
• To test whether apparent exceptions to island constraints in Swedish reflect deep or surface variation.

Main Methods:

• Eyetracking while reading was employed to measure processing difficulty.
• Experiment 1 compared RCE with licit that-clause extraction (TCE) and illicit non-restrictive relative clause extraction (NRCE).
• Experiment 2 compared RCE with extraction from a subject NP island (SRCE), manipulating plausibility.

Main Results:

• RCE showed early integration patterns similar to TCE but intermediate total durations between TCE and NRCE.
• RCE processing was sensitive to plausibility at early integration points, unlike SRCE.
• Structural facilitation was observed for RCE compared to SRCE.

Conclusions:

• The processing profile of Swedish RCE is compatible with weak island extractions.
• The findings support a surface variation account for the lack of strong island effects in Swedish RCE.
• Swedish RCE may represent a case of weak islandhood rather than a true exception to island constraints.