I recently had the pleasure of reading Kedi Simpson's 2024 PhD thesis The Developmental Trajectory of Second Language Listening Errors. Kedi teaches French, German and Spanish in Oxfordshire and has been researching listening for a few years now, via an MA and doctorate. Since she is at 'at the chalkface' she has an intimate knowledge of what students find easy and hard.
The thesis is almost a bit of a wake-up call, because what struck me the most is how many misunderstandings typical pupils have when they hear quite simple sentences in French. Over a long period Kedi got pupils to write down the meaning of what they heard when listening to short audio snippets. She also got chosen students to tell her about their experience of listening - what goes on in their heads. It's fascinating! They often understand far less than we may think.
A central theme of Kedi’s thesis is Fuzzy Lexical Representation Hypothesis (Gor et al., 2021), which argues that second language learners store words in a "fuzzy" way due to their difficulty in decoding the continuous stream of speech. These difficulties are heightened by the influence of the first language (L1), especially when L2 sounds resemble, but are not identical to, L1 forms. This can cause systematic mishearings, as in one striking example where the phrase bouteille d’eau was misinterpreted by a third of participants as potato, highlighting how learners map new input to familiar sounds or concepts. Another source of misunderstaning is when learners are confused by two similar-sounding L2 words. They may fail to identify individual sounds or the gap between words. A key point she makes is that the L1 (in this case English) is ever present.
Kedi's analysis id also rooted in what is called a Complex Dynamic Systems framework, which is particularly useful in capturing the highly individual, evolving, and interactive nature of listening comprehension. Each listening experience is very personal and is shaped by a combination of factors: vocabulary knowledge, grammar, segmentation ability, personal strategies, and even classroom experiences. No single factor operates in isolation. Her reports from students are an interesting insight into what is going on in their heads when they hear something. For instance, one student heard sa sœur Martine est très travailleuse (“his sister Martine is very hard-working”) and translated it as “Martin has nice eyes.” This misunderstanding arose from a mix of segmentation errors (hearing travailleuse as yeux), misidentifying the name Martine as Martin, and overlooking key grammatical markers like feminine agreement. This is not just a vocabulary issue, but a combination of interacting errors. Field's micro-skills are at work but in a sort of personal complex dynamic interplay of language processing. So much is at play, that predicting where students will go wrong is a bit like predicting the weather.
The young students employed a range of coping strategies to deal with the challenges of decoding. One such strategy is what she terms “self-subtitling,” where learners visualise or mentally transcribe words based on past classroom experiences, such as reading and seeing an image from a PowerPoint slide. This visual reinforcement becomes a kind of scaffold, helping learners link sound to meaning when audio alone isn’t enough.
Interestingly, Kedi found that strategy use develops more effectively in older learners, while younger students often rely more heavily on guesswork or familiar context. One learner, for example, described listening as simply identifying key content words like cycling to answer comprehension questions - evidence of a superficial engagement with meaning, shaped by the limitations of the 'comprehension approach' commonly used in schools. My takeway here is that time is much better spent focuisng on decoding skills (the 'process') than strategies. Strategies have always felt to me like a useful add-on for older students, especially with exams in mind. (The ample research on strategies over the yeras is mixed, but does suggest its utility for some students.) most of all - do lots of listening as this will not only improve students' listening skills, it will also develop their general proficiency through all the comprehensible input it provides.
So, one of the most important pedagogical takeaways from the thesis is a caution against overestimating what beginner and intermediate learners can achieve during listening tasks. Teachers often assume that learners are capable of parsing full sentences and mapping them reliably to meaning, when in reality, many students are barely decoding individual words. As Kedi writes, “a key pedagogical implication must be an overhaul of the expectations teachers have of the purpose of listening in class and what can be achieved from it.”
This aligns closely with Field (2008), who warned that listening development is deeply individual: learners develop at different rates, focus on different cues, and construct their own strategies for coping with uncertainty. Simpson’s research reinforces this point with rich qualitative data. It is the feedback from individual students which I found the most enlightening.
Although she does not go into detail about pedgagical implications, it is clear that she would recommend what is sometimes called the process approach to listening (Field, 2008), an approach Gianfranco and I describe in detail in Breaking the Sound Barrier: Teaching Language Learners How to Listen (2019). Essentially, this involves: (1) breaking down the skill of listening into micro-skills, such as recognisizing phonemes and syllables, retrieving words and chunks, making grammatical sense of input (parsing), as well as bearing in mind top-down knowledge - the background knowledge which helps us make sense of listening input, then (2) designing tasks which focus on each of these micro-skills. At the level of phonology this could include activities or games which get learners to identify different phonemes. At the grammatical and lexical level, this could inlcude spotting where gaps between words occur.
Another key finding is, as others have found, vocabulary knowledge trumps grammatical knowledge. Students need to know lots of words and chunks. Crucially, however, they need to know what they sound like and need to have an accurate (i.e. not fuzzy) mental representation of them. So teaching words through listening is vitally important, as is encouraging accurate pronunciation which helps student develop this accurate mental representatiion of vocabulary.
Ultimately, the thesis reminds us that listening is not passive, and comprehension does not emerge automatically from exposure to input. For beginner learners, the decoding process is messy, unpredictable, and constantly evolving. You can help it aliong by designing tasks which force students to listen carefully, intensively, looking for detail - rather than than doing more superficial tasks which encourage the spotting of occasional words to get meaning. By adjusting our classroom expectations and supporting learners throuhg plenty of scaffolded, low-stakes listening experiences, we can help them develop more stable, less fuzzy representations of the target language.
For more on the problems of listening and how teachers might address them I recently posted here and here.
Kedi's blog with posts about second language listening is here. Her MA dissertation about speech stream segmentation is here.
Read the PhD here.
References
Conti, G. and Smith, S.P. (2019). Breaking the Sound barrier: Teaching Language Learners How to Listen. Independently published.
Field, J. (2008). Listening in the Language Classroom. Cambridge: C.U.P.
Gor, K., Cook, S., & Jackson, S. (2010). Word recognition and lexical representation in L2 phonological processing. Studies in Second Language Acquisition, 32(3), 387–414.
Simpson, K. (2024). The Developmental Trajectory of Second Language Listening Errors. Doctoral Thesis, University of Oxford.
Available at: https://ora.ox.ac.uk/objects/uuid:fd5c080b-850d-4276-8ead-3c6f5dada9cc
Comments
Post a Comment