Vocabulary for listening: Emerging evidence for high and mid-frequency vocabulary knowledge Joshua Matthews University of New England, Armidale, NSW, 2351, Australia article info Article history: Received 18 January 2017 Received in revised form 13 October 2017 Accepted 17 October 2017 Keywords: Second language listening Listening comprehension Vocabulary Aural vocabulary knowledge abstract This article presents empirical evidence aimed at informing approaches to vocabulary development for the purpose of supporting L2 listening comprehension. Inferential sta- tistics were used to analyze the relationship between second language (L2) aural vocab- ulary knowledge (AVK), L2 listening comprehension and the overall L2 proficiency among 247 tertiary level L2 learners of English. Three frequency based levels of AVK were measured. Measures of level 1 (0e2000 frequency range) and level 2 words (2001e3000 frequency range) tapped AVK of high frequency words, and level 3 words (3001e5000 frequency range) tapped AVK of words just beyond the high frequency range. Listening comprehension was measured with a version of the International English Language Testing System (IELTS). Regression modeling showed that AVK at each of the three levels contributed uniquely to the prediction of L2 listening for the entire cohort. Only measures of level 2 and 3 AVK were uniquely predictive of L2 listening for a relatively high profi- ciency subgroup, whereas only level 1 AVK offered a unique contribution to the prediction of L2 listening scores for a relatively low proficiency subgroup. Results are interpreted in order to provide a range of pedagogical recommendations. © 2017 Elsevier Ltd. All rights reserved. 1. Introduction Successful listening comprehension depends on a language user's knowledge of L2 words (Vandergrift & Baker, 2015). Words represent the lowest level of representation at which a stable connection between the form and the meaning of a word can be reliably established (Hulstijn, 2002); therefore, once words are recognized and associated with their literal meanings, larger semantic units can be built in the mind of the listener. Key to the listening comprehension process is the rapid and appropriate association of such semantic units (linguistic knowledge) with the listener's pre-existing schemata (non-lin- guistic knowledge). The ability to recognize, understand and interpret strings of words in connected speech typically poses little difficulty for native speakers, as throughout their lifetime native speakers have been exposed to an immense quantity of contextualized spoken input (Ellis, 2002). This is not the case for many L2 listeners (Field, 2008a). Without having encountered the equivalent degree of exposure to the target language characteristic of native speakers, L2 learners typically have sub-optimal aural vocabulary knowledge; that is learners have difficulty recognizing words in the spoken form. Learners may even have dif- ficulty in recognizing words in speech which they know when presented in the written form (Goh, 2000). E-mail address: joshua.matthews@une.edu.au. Contents lists available at ScienceDirect System journal homepage: www.elsevier.com/locate/system https://doi.org/10.1016/j.system.2017.10.005 0346-251X/© 2017 Elsevier Ltd. All rights reserved. System 72 (2018) 23e36