EXPRESS: Learning words in diverse narrative contexts disrupts subsequent recall of their meanings

Recent research suggests that encountering words in varied narrative contexts may disrupt the earliest stage of learning (Hulme et al., 2023). The current experiments explore (i) how this diversity disadvantage changes over learning, and (ii) whether diversity primarily impacts encoding or subsequent recall. Adults (Experiment 1: N = 105; Experiment 2: N = 111) encountered novel words with invented meanings within narrative paragraphs across three days. Words were presented within one coherent narrative (low diversity) or across nine different narratives (high diversity). We replicated Hulme et al.'s (2023) results for word-form learning, finding moderate/strong evidence for null effects of diversity on word-form recall and recognition (Experiment 1), and speeded word-form recognition accuracy (Experiment 2), and only anecdotal evidence of a diversity benefit on response times. For word-meaning learning, Experiment 1 replicated the diversity disadvantage on word-meaning recall reported by Hulme et al. (2023; extreme evidence), but in Experiment 2 there was moderate evidence for a null effect on semantic feature recognition. Findings indicate that while narrative diversity does not strongly influence participants' ability to encode or retain semantic features for new words, words learned in low-diversity conditions initially build more coherent contextually-bound representations that facilitate recall of semantic features.