Is syntactic priming semantically independent in first and second language?

Published: 23 December 2022| Version 1 | DOI: 10.17632/mggxhrtc8p.1
Basma Elkhafif,


The study employs syntactic priming to examine whether the prepositional phrase (PP) semantic role contributes to the parsing of PP-attachment syntactic structure in both first and second language. Participants were presented with a list of sentences that all shared both the syntactic attachment and PP semantic role (within-role), or a list where sentences in the last block differed in the PP semantic role (cross-role). Syntactic priming occurred across sentences that had the same and different semantic roles in first language (L1), indicating a semantically independent syntactic processing. Second language (L2) speakers revealed a smaller magnitude of cross-role syntactic priming.


Steps to reproduce

The file “Data.xlsx” contains the data necessary to reproduce the analyses in the paper. The excel file has four taps that include data for within-role condition in first language, within-role condition in second language, cross-role condition in first language and cross-role condition in second language respectively. The data for the within-role conditions include 10 columns that represent subject, structure, block order, answer to the comprehension question (“C” for correct answers and “I” for incorrect answers), item, log transformed stimulus order, first run times, total times, regressions and region (D for disambiguating area and Post for post-disambiguation area). The columns representing cross-modal condition are comparing first run times, total times and regressions either between the fourth blocks of the within-role list vs. the cross role list, or between the first block of the within-role list vs. the fourth block of the cross-role list. In the latter case, the dependent variables were given the names “firstrunf”, “totaltimef” and “regressionsf”. R code #statistical analysis library(lme4) #load the data data= read.csv(file.choose()) #exclude items after which comprehension questions were answered incorrectly list=data[data$CQ=="C", ] #exclude the region that you are not currently analysing list1=list[list$Region=="D", ] #the same is done for each condition before typing the model #the model for first run time as a dependent measure withinrole.model=lmer(firstrun~structure + blockorder + Trstimorder + structure* blockorder + (1+ structure * blockorder * Trstimorder |subject_L)+(1+ structure * blockorder* Trstimorder |item), data = list1, REML = FALSE) summary (withinrole.model) # similar models were used for total times and regressions. # For cross-role analysis, exclude items after which comprehension questions were answered incorrectly. list=data [data$CQ=="C", ] #The model crossrole.model=lmer(firstrun~ structure+group+ structure*group+(1+ structure*group|subject)+(1+ structure*group|item), data= list, REML = FALSE) #similar models were used for total times and regressions.


Helwan University