Subtlex-pl: subtitle-based word frequency estimates for Polish

We present SUBTLEX-PL, Polish word frequen- cies based on movie subtitles. In two lexical decision exper- iments, we compare the new measures with frequency esti- mates derived from another Polish text corpus that includes predominantly written materials. We show that the frequen- cies derived from the two corpora perform best in predicting human performance in a lexical decision task if used in a complementary way. Our results suggest that the two corpora may have unequal potential for explaining human perfor- mance for words in different frequency ranges and that cor- pora based on written materials severely overestimate fre- quencies for formal words. We discuss some of the implica- tions of these findings for future studies comparing different frequency estimates. In addition to frequencies for word forms, SUBTLEX-PL includes measures of contextual diver- sity, part-of-speech-specific word frequencies, frequencies of associated lemmas, and word bigrams, providing researchers with necessary tools for conducting psycholinguistic research in Polish. The database is freely available for research pur- poses and may be downloaded from the authors’ university Web site at http://crr.ugent.be/subtlex-pl