Challenges in the assumptions of using a flemma-based word counting unit

Abstract

The choice of word counting units (i.e. word family, flemma, or lemma) is of great importance in vocabulary list and test creation, as there are assumptions underpinning the use of each. Flemma-based counting assumes that if a learner can understand the meaning of a word in one part of speech (POS), they can also understand its meaning when the same word form is used in another POS. A previous quantitative study showed that such an assumption is not always valid, but it did not provide reasons as to why. Therefore, this article presents an interview study probing the challenges learners face when they fail to understand the meaning of a known word form used in a new POS. The data were collected through one-on-one interviews with 16 university students in Japan, where they were asked to demonstrate comprehension of target words embedded in short sentences, as well as to explain how they approached the task. The interviews revealed that factors related to both the words and the learners play a role: some words have more complex meaning relationships across different POS, and some learners did not have advanced enough parsing skills to overcome such complexity. Based on these findings, this article concludes that the flemma might not be a suitable word counting unit for Japanese university students with relatively low levels of proficiency.