These data are for the experiment reported in [http://journals.plos.org/plosone/article?id=10.1371%2Fjournal.pone.0175199][1] The .mat file contains objects giving the channel location information (a structure), time in milliseconds corresponding to each sample (a vector of integers), the identifiers for participants (a cell array of strings), and two structures called 'subject_averages' and 'item_averages' each containing the ERPs averaged by subjects or by items, respectively. Each 'averages' structure has a separate field for each condition, and within that a field for each participant; the ERP for a given condition/participant combination is a Channels × Samples matrix. Note that the item averages for the clause-initial before condition contain missing values for item 10, which contributed no data to this condition (for each participant who saw this item in this condition, there was artifact in the corresponding trial). These files are provided for researchers interested in verifying the claims of this paper through reanalysis of the data; researchers interested in using the data for other purposes (such as to test new research questions) should obtain prior agreement from the authors (per section 8.14 of the American Psychological Association's Ethical Principles of Psychologists and Code of Conduct http://www.apa.org/ethics/code/index.aspx?item=11#814). The folder 'raw data' contains the raw data in Brain Vision Recorder format. [1]: http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0175199