Recognition of spoken pseudowords

• Author(s) / Creator(s)

  • Matthew C. Kelley
  • Benjamin V. Tucker

• bBoiA sizable number of phonetic and psycholinguistic experiments have been conducted to investigate the recognition of real words. From this work, researchers have found that various characteristics of lexical items affect the recognition process, such as lexical frequency, phonotactic probability, phonological neighborhood density, and uniqueness point. Such studies have also described processes that occur during word recognition, such as activation and deciding which among a number of candidates is the item being heard. The present study continues in this line of research and examines the recognition of phonotactically legal pseudowords.

  Many psycholinguistic studies use pseudowords in their stimuli set for experiments so that the task asked of the participants requires linguistic processing to occur. Yet, these studies tend to disregard the responses to pseudoword stimuli because the studies are focused on characteristics of the words. As such, the process of recognizing pseudowords themselves has gone under-described for some time, making it difficult to say, for example, what effect the pseudowords may be having on participants during their participation. Additionally, understanding how pseudowords are processed may have implications for understanding how new vocabulary items are acquired in first- and second-language contexts, because a new word is effectively the same as a pseudoword when it is first heard.

  In the present study, we investigate how listeners respond to pseudoword stimuli in an auditory lexical decision task. The data are from the Massive Auditory Lexical Decision data set (Tucker et al. 2017), which consists of auditory lexical decision responses to 26,800 words and 9,600 pseudowords, with a mean of 11.88 responses per pseudoword from 232 monolingual native Canadian English speaking participants. We use linear mixed-effects regression to model the effects of phonotactic probability, phonological neighborhood density, and uniqueness point on participants’ response times to the pseudowords in the auditory lexical decision task. Significant effects were found for each variable of interest, and all showed a trend of higher values predicting longer response times, suggesting overall that more remarkable sequences (that is, less likely sequences) are easier to recognize as pseudowords than more common sequences. We conclude our analysis by discussing similarities between pseudoword processing and real word processing and examining the implications of these findings for experimental design and models of word recognition.

  Presented at WECOL 2017 in Boise, ID.

• Date created

  2017-10-20

• Subjects / Keywords

  • speech perception
  • phonetics
  • psycholinguistics
  • spoken word recognition

• Type of Item

  Conference/Workshop Presentation

• DOI

  https://doi.org/10.7939/R3HM53109

• License

  Attribution-ShareAlike 4.0 International