Readvertising new paper based on my EPS mid-career prize lecture; sets out my thinking on reading & reading acquisition. Summarised here in a series of tweets journals.sagepub.com/doi/10.1177/17… 1/
Study of reading & reading acquisition has been dominated by concept of primacy. Written language is an expression of spoken language, and reading is parasitic on speech. 2/
This perspective has set much of the research and instructional agendas e.g How do we ‘decode’ words back to spoken language? How should we teach kids to do this? What happens when writing system makes this hard? 3/
To a certain extent this is appropriate for alphabetic writing systems. Visual symbols represent sounds, and kids need to learn this mapping to be successful readers 4/
But there are big differences between written and spoken language. Spoken language comes with coarticulation, intonation, prosody, audio-visual cues, gesture, knowledge of speaker intention. 5/
Even if visual symbols perfectly represent sounds, these sources of information are always absent from text. Thus, decoding writing to spoken language presents many of the challenges of spoken word recognition but without the benefits. 6/
Let’s consider the ancient scriptura continua texts as an example. These provided an incredibly tight link to spoken language. Introduction of spacing and other graphic devices broke this link. Why should this have occurred? 7/
Because simply translating text back to an impoverished spoken language code is not sufficient for rapid, silent reading. Possible that this change in writing system was related to dev’t of modern, silent reading. 8/
I argue that we shouldn’t think of writing just as a route into sound; we need to think about it more broadly, as a learned linguistic environment to be understood in its own right. 9/
Next, we take a deep dive into English writing. Everyone knows that English has relatively poor spelling-sound consistency. But I argue that this is functional: it permits highly reliable morpho-orthographic cues to meaning to emerge. 10/
Though English is not a morphologically rich spoken language, morphology is highly visible in the writing. The availability of multiple possible spellings for particular sound sequences means that some have become reserved to express particular aspects of meaning 11/
Let’s take –ed spelling as an example. In English, this is a highly reliable cue to the past. But use of –ed across all past-tense contexts weakens the spelling-sound relationship (e.g. snored, kicked, dented instead of snord, kict, dentid). 12 /
Interesting that –ed is also virtually never used in words of a single morpheme. In such words, a different spelling is normally used (e.g. instead, horrid). The availability of these other spellings ‘protect’ –ed so that it can be reserved to express the past. 13/
Recent evidence suggests that this is a general principle of English spelling. Disorder of the English spelling–sound mapping is not some nasty feature to be overcome, but a consequence of order of spelling–meaning mapping. 14/
How might this tension between spelling-sound and spelling-meaning information play out in other writing systems? Important question, and I’m not sure. 15/
Interesting that English spelling over past 1000 years has evolved to make these meaningful cues more prominent. Why not become more spelling-sound transparent? Is the evolution tied in some way to the cognitive demands of reading? 16/
Now we move to the mechanisms underpinning reading & reading acquisition. Children start with oral language knowledge, then learn spelling-sound information. We think this is implemented in a particular brain pathway (dorsal). 17/
But ‘decoding’ is not enough for fluent reading; need to learn to relate visual words with meanings directly. This presents a big problem, especially because English vocabulary is vast (ca 70k words by age 20). 18/
Discovery of morphological information might be crucial; those 70k words are built from a more manageable 11k stems. Need to understand how that morphological knowledge is discovered and represented, and how it facilitates orthographic learning 19/
Once that morphological information is discovered & represented, it has a big impact on skilled reading. Skilled readers parse letter strings into morphemes within 200 ms of presentation; even do this for words like ‘corner’ which look complex, but aren’t. Why? 20/
Because the writing system makes sure that these words are rare; usually spelled another way (e.g. martyr). Segmentation is fast because it is superficial; it is allowed to be superficial because of the reliability of the morphemic info in the writing 21/
Emerging evidence that this form of knowledge is represented in a different brain pathway (ventral) associated with dev’t of reading expertise through text experience into adolescence. 22/
Overall, it looks like English readers capitalize on spelling-sound and spelling-meaning information, but at different points in reading acquisition, and underpinned by different brain pathways. 23/
Evidence also that representation of spelling-sound and spelling-meaning information is a mirror of the writing system. Highly consistent information in writing system yields particularly robust representations. 24/
To conclude, study of reading & reading acquisition is about how information transmitted through writing becomes represented in the minds and brains of individuals, through an accumulation of instruction and text experience. 25/
But we know so little about nature of information transmitted through written language; we’ve barely scratched the surface. This is a big challenge for progress in this field 26/
Further, reading acquisition is all about the discovery and representation of case-specific and general knowledge. But how does this arise? We need much more contact with memory theorists to answer properly. 27/
This discussion also raises new questions about writing systems. If meaningful information is such an important part of English spelling and reading, then are other writing systems that prioritise spelling-sound information at a disadvantage? 28/
Tempting to say no, but I’m not so sure. Spelling in most major writing systems is regulated, sometimes in really unhelpful ways for reading (e.g. English braille). English is self-organising. Has it found a balance of information that is successful? 29/
This takes us to a final question about whether particular cognitive constraints drive spelling change (think back to those ancient scripts). Interesting question. 30/
I really liked writing this paper and hope it provides a thought-provoking read. Would love to hear thoughts and links to important references that I might not have included, particularly for the more speculative ideas put forward. End /
