In my last post, we reviewed a round up of some research on phonology, with clarifications around what we know and what we do not yet know regarding the relation of advanced phonemic awareness training and phonemic proficiency to outcomes for struggling readers.
One piece I briefly mentioned in that post and which I’d like to dig further into is from David Share, “Is the Science of Reading Just the Science of Reading English?“
This is an important question to ask, because while research into how children who speak English learn to read in English has become quite substantive (even if still mostly unknown in too many classrooms), there is still quite a bit we don’t know about learning to read in English if you don’t speak English as your first language, and there’s even more we don’t yet know about learning to read in languages other than English.
Share’s main critique against the “science of reading” is that it is “Anglocentric,” such that research presented on reading in English doesn’t even take the time to mention that it is about English and seems to expect findings to bear universal implications, whereas any research in another language explicitly notes that specificity, with an expected delimitation of applicability mostly to that language itself.
I have seen such vanilla bias in my own writing — when reviewing my updated post for Nomanis, “I think I was wrong about phonemic awareness,” for example, I realized that I hadn’t specified I was referring specifically to phonemic awareness as it relates to English — and updated my post accordingly. And I think the post is much clearer and accurate as a result of this.
The fact is that when it comes to something like phonemic awareness, this linguistic skill is most critical for learning to read in an alphabetic language such as English — whereas “in reading syllabaries and morphosyllabaries phonemic awareness is not a uniformly important factor” (Verhoeven & Perfetti).
What is universal across languages and writing systems is the importance of phonology: “The predictive power of syllable-level awareness across different writing systems and languages – including alphabets, which do not encode syllables – highlights the general importance of attention to the sound structure of a language for learning to read” (Verhoeven & Perfetti). When we discuss learning to read in English, phonemic awareness surfaces as a critical subskill of phonological awareness that develops reciprocally with learning the alphabet, and that furthermore, higher level phonological awareness is developed via phonemic awareness training with letters. But if we were teaching our kids to read in Mandarin Chinese or Japanese, we would be far less concerned with stressing phonemic awareness in pairing symbols to speech, because that wouldn’t match the encoded language forms we would be trying to teach to our students.
While Share acknowledges the importance of phonology in all written languages, he challenges how Anglocentrism has biased our viewpoint of the concept, such that we tend to think of phonological awareness linearly in a written language. Yet he argues that “many writing systems…exhibit substantial nonlinearity, or multilinearity with multiple axes.” In English, we think of phonological awareness in relation to reading as sounds “strung along a single (acoustic) axis.” But what does “nonlinearity” even mean in phonological awareness?
It took me a while to begin to understand what Share is saying about this, as he doesn’t provide a more concrete example until later in the paper. It did take me on a fruitful rabbit hole excursion on investigating “extralineal diacritics” (searching for this term brings up more papers by Share!), but I didn’t fully get what he meant until I reread this example under the section “Multilinearity and Nonlinearity”:
In Devanagari, for example, a noninitial /i/ is written before the consonant after which it is pronounced, and this nonlinearity appears to create problems for the learner (Kandhadai & Sproat, 2010). In Malayalam, the official (scheduled) language of the southern Indian state of Kerala, the syllable pronounced /ktro:/ is written with the symbol for /r/ preceding the symbol for /k/ and the two-part vocalic circumfix for long /o:/ surrounding the ktr consonant cluster, so the reader decodes the five-symbol string ക്ത്രോ in the following order: 3, 4, 2, 1, 5! In some scripts, characters are nested in vertically aligned syllabic units, such as Indic aksharas, Arabic mashkul script, and Korean syllable blocks, which may facilitate reading acquisition (at least initially) by obviating the need to access phonemes. Chinese semantic-phonetic character compounding typically positions the semantic component to the left of the phonetic, but this component can also appear above, below, after, or surrounding the phonetic.
This is an important aspect to consider when learning to read other languages, as “Learning the many positional regularities (and exceptions) of a script appears to tax visuospatial skills (McBride, 2016; Yang & Meng, 2020), a factor that the science of reading has concluded is unimportant in English reading (Vellutino, Fletcher, Snowling & Scanlon, 2004).”
Share also claims that Anglocentrism originally resulted on a focus primarily on word reading accuracy, rather than rate and fluency. This may have been true in the past, though I certainly have not seen this in my own time in the field, so thankfully that has been rectified. But part of Share’s point is also that for reading in other languages, speed and fluency are often the more important factors in reading development, especially in relation to dyslexia.
Here he makes an interesting statement: “These observations suggest that reading rate may be the most intractable component of the reading of students with reading disability/dyslexia and perhaps the core deficit.”
Another casualty of an Anglocentric focus has been that the “study of meaning (i.e. morphology, morphological awareness) . . . has only begun receiving the research attention it deserves over the past decade or so. . . it is clear that morphology has replaced fluency as the neglected stepchild of the science of reading.” He has a point here — I’ve become more and more aware of the importance of morphology in my own practice, most especially for students learning the English as a new language, and it does seem to be growing as an area for research.
According to Share, these and other biases based on focusing solely on English and other European languages have meant that our views of orthographic complexity are limited and do not yet have universal applicability. The few theoretical frameworks that can describe cross-linguistic orthographies, “orthographic depth” and “psycholinguistic grain size theory,” “promote the one-dimensional view of script variation.”
To combat this, Share proposes 10 dimensions of orthographic variety that can do more to be inclusive of global diversity:
- Spoken–Written Linguistic Distance
- Multilinearity and Nonlinearity
- Visual Confusability and Visual Complexity
- Historical Change: Retention of Historical Spellings Despite Pronunciation Change (“This category has understandably preoccupied English spelling reformers for centuries, as well as generations of Anglophone reading researchers, but may come at a price, namely, at the expense of research into other importance factors of reading (e.g. morphology, meaning)”
- Spelling Uniformity Despite Morphophonemic Alternation
- Omission of Phonological Elements
- Dual-Purpose Letters
- Inventory Size
All in all, an interesting read that got me exploring many other orthographic related factors more deeply. I learned terms like “extralineal diacritics” and “ductus,” and am continuing on a quest to dig further into a statement Share makes at one point in summarizing the research on the dimension of “Spoken-Written Linguistic Distance”:
The evidence is overwhelming that when students learn to read written forms that diverge from their spoken vernacular, this has a profoundly detrimental impact on learning to read (August, Shanahan, & Escamilla, 2009; Gatlin & Wanzek, 2015; Myhill, 2014; Saiegh-Haddad & Schiff, 2016).
While there is most definitely evidence on this point, I would hesitate to state “overwhelming.” I’ll write more about this in another post!
If you’re interested in digging further into the idea of “universals” in writing systems, here’s a few recommendations to continue beyond this paper:
- Frost, R. (2012). Towards a universal model of reading. Behavioral and brain sciences , 35 (5), 310-329.