The Corpus of Bidirectional Reading of Traditional Chinese Text: An Extension of the Multilingual Eye-Movement Corpus

We present a corpus of eye-movement data collected during the reading of passages in traditional Chinese and developed in the framework of the Multilingual Eye Movement Corpus (MECO) project. A distinctive feature of traditional Chinese writing is that texts in this written language can be formatted and read either horizontally or vertically. Leveraging this characteristic bidirectionality, our corpus, based on 60 Hong Kong participants' eye movement data, is the first to offer a within-subject and within-item comparison of eye movements during horizontal and vertical text reading. In addition to the corpus itself, we provide reliability estimates of the eye-tracking data and linear mixed-effects model analyses for its technical validation. The results demonstrate that horizontal and vertical reading of traditional Chinese passages yield comparable spatial and temporal eye movement patterns, contributing to our understanding of both universal and language-specific cognitive, visual, and oculomotor processes in reading. Ultimately, this corpus supports the development of more comprehensive and ecologically valid models of eye movement control in reading.