An Experimental Investigation of HAM, a Novel Mnemonic Technique for Learning L2 Homonyms and Homophones

Over the past 40 years, extensive research has examined the effectiveness of mnemonics for vocabulary learning. Much of this research has investigated the keyword method (Atkinson & Raugh, 1975), which involves linking an image related to a to-be-learned L2 word with an image related to a similar-sounding L1 word. Whereas most research has shown the keyword method to be effective (Webb & Nation, 2017) with impressive long-term retention rates (Beaton, Gruneberg, & Ellis, 1995), some have questioned its usefulness, particularly due to the quality of the resulting lexical representations and extended latencies associated with recall (Barcroft, Sommers, & Sunderman, 2011; Van Hell & Candia Mahn, 1997). Other drawbacks of the keyword technique are the equating of dissimilar L1 and L2 phonemes and the difficulty in creating associations for languages with markedly different phoneme inventories. The current study presents a novel approach called the Homonym/Homophone Association Method (HAM). It overcomes some of the drawbacks of the keyword method by associating meanings of L2 homonyms or homophones, one known by the learner and one unknown. Because the pronunciations of the L2 target words are identical (or nearly identical), learners only need to associate two distinct meanings. A quasi-experiment (N = 71) employing a within-subjects design compared the effectiveness of (1) HAM using researcher-generated associations and images, (2) HAM using self-generated associations, and (3) production practice that involved writing target words in sentences. Results on an unannounced posttest given 3 weeks after instruction suggest an advantage for HAM using researcher-generated associations.