Language distance shapes fluency
The ‘language distance’ between English and a language learner’s first language affects reading fluency, but not comprehension according to a study by Victor Kuperman at McMaster University, Canada.
It is clearly easier for someone to learn a language that is more like their native language than one that is more distant: learning English is generally easier for Italian than for Japanese speakers. What is less clear is which processes in language learning are affected and how to quantify those effects.
Kuperman set out to clarify the effect of language distance on reading fluency and comprehension. Past studies have given mixed results, possibly because the focus has been on comprehension, which is relatively easy to measure.
Data for this study was taken from the Multilingual Eye-movement Corpus. This database has measures of reading fluency and comprehension from 1,105 students and staff from 24 universities having 18 distinct first languages.
Measuring language skills
Reading fluency was measured in three parts: length of the first look at a word to recognise and decode it; total gaze duration assessing the plausibility of the word in context; and number of words read per minute. Participants also answered comprehension questions on the 12 short texts, as well as several measures of specific skills such as decoding and vocabulary as well as overall language proficiency. All participants were judged to have advanced level English proficiency.
Language distance was taken as the ‘genetic distance’ as calculated using 18 lexical items in the open access source: eLinguistics.net. Differences of written script were not explicitly incorporated into this measure which may seem surprising, but it is a lot more difficult to quantify distances between scripts than between spoken languages. How can we meaningfully measure the ‘distance’ between an abugida script such as Hindi, the logographs of Mandarin, and the Roman or Cyrillic alphabets? Fortunately, to some extent, these script differences are reflected in the phylogenetic trees used to calculate language distances.
While the runners up for most distant language are Turkish and Hebrew, first place goes to the geographically much closer Basque (or Euskara). Basque is the only surviving pre-Indo-European language still spoken in western Europe and is unrelated to any known language family. Dutch was the closest language, followed by Danish and Norwegian.
A series of regression analyses sought to identify whether language distance was influencing reading fluency and/or comprehension and if so, whether that influence was direct or indirect.
To illustrate what this means, imagine two students, one a Dutch speaker and the other Turkish. It may be that the Turkish student has more difficulty acquiring the necessary skills such as decoding and vocabulary and that this influences reading. But if the influence of language distance is only indirect, i.e. via these underpinning skills, then once the two students have reached the same level of skills, their reading fluency should be the same.
Alternatively, if language distance influences reading fluency directly, then the Turkish student will always be less fluent even though the two students measure equally on the underpinning skills of English proficiency.
The results found that while language proficiency predicted reading comprehension, language distance had no effect – direct or indirect – on reading comprehension.
Reading between the lines
The results for reading fluency, however, were quite different. Here, language distance had consequences for both the underpinning skills such as decoding and word identification, and directly on reading rate. So even if our Dutch and Turkish students have the same level of comprehension, the Turkish student will tend to be less fluent.
These results begin to get under the bonnet to investigate how language distances incur challenges for language learners but also raise more questions. These participants all had advanced English proficiency – are the challenges of language distance better or worse over time? Can targeted training early in language learning help to close the gap?
It seems intuitive that reading comprehension must be influenced by differences in script which are not fully measured here. Teaching English to learners who are learning the Latin alphabet for the first time has very particular challenges and speaking and listening tends to race ahead of reading and writing. How do learners respond when the language is largely the same – but the script is completely different, such as for Hindi and Urdu speakers? That could tell us something about the specific role of differences in script.
Fluency is very often taken as a proxy for language proficiency, especially in everyday settings. But the results of this study suggest that where language distance is relatively large, comprehension is likely greater than fluency suggests.
[Reference]
[Source: EL Gazette]
