Read and Imagine: Visual Imagery Experience Evoked by Native vs. Foreign Language

The present research examined visual imagery evoked during reading, and how the vividness of imagery depends on language. Previous studies reported that bilinguals’ imagery is less vivid in their second language (L2) than native language (L1). Subsequent criticism suggested that the observed “language effect” may be due to other factors such as L2 proficiency; however, this possibility had not been empirically tested. Our study examined how visual imagery is influenced by the language of text, the nature of the text, and the individual characteristics of the reader. Turkish-English bilingual young adults (N = 382) completed the Read & Imagine Task (R&I) we developed, and a series of other measures testing their English language proficiency and imagery skills. R&I was designed to evoke either object visualization concerning pictorial properties of objects such as color and shape (e.g., a colorful bouquet) or spatial visualization concerning spatial relations, locations, and spatial transformations (e.g., a 3D block tower). The passages used in the task included (1) visual object texts, (2) visual spatial texts, and (3) fiction literature, half of which were presented in L1, and the other half in L2. Each passage consisted of nine segments, and participants rated imagery vividness for each individual segment as well as the entire text. After reading the entire text, they also rated specific imagery characteristics such as color or spatial relations. Regardless of the timing of rating, the vividness of imagery was higher in L1 than in L2 for all types of texts (object, spatial, literature). However, English test scores also predicted vividness in L2 more than vividness in L1, suggesting that L2 proficiency partially explains the language effect on vividness. Further, object and spatial imagery skills selectively predicted imagery vividness in the corresponding object and spatial trials. Our results emphasize the importance of considering individual differences in imagery and testing object and spatial imagery separately when evaluating the effects of language on imagery. Although we found that vividness was lower in L2 than L1, our research suggests that the effect is not extremely robust and may depend on several factors such as the nature and difficulty of the text, when and how vividness is measured, as well as L2 proficiency and other individual differences. The language effect on imagery vividness is far more complex than has been previously proposed.