I try to help people shape their prose faculty, their facility with prose. About a week ago Greg Ashman tweeted an NPR interview about the “science of reading,” which occasioned mixed feelings in me. It’s always nice to hear science confirm one’s teaching philosophy and, though I don’t teach reading to grade schoolers, but writing to university students and scholars, I find “phonics” to be both a compelling theory and a useful practice. When students want to know whether or not they are “doing it right”, i.e., whether or not they are writing well, I tell them to read their paragraphs out loud. Even better, I tell them to get a classmate to read it out loud to them. The way a paragraph sounds, the ease with which it comes off the page, tells you a great deal about how well it is written.
But I’ve also long been skeptical about the scientific study of ordinary cognitive abilities like reading and writing. Claudio Sanchez introduces his interview with Mark Seidenberg with this observation:
Mark Seidenberg is not the first researcher to reach the stunning conclusion that only a third of the nation’s schoolchildren read at grade level. The reasons are numerous, but one that Seidenberg cites over and over again is this: The way kids are taught to read in school is disconnected from the latest research, namely how language and speech actually develop in a child’s brain.
At first pass, this seems like a reasonable point. But suppose I said that only one third of the nation’s school children eat a healthy diet. And suppose I explained this by way of a “disconnect” between what kids are fed and what the latest research shows us about how foods and beverages actually affect the brains of children. The research may be perfectly sound (or it may not) but did we really need brain research to understand what children should eat? This becomes still more clear when we hear what the science actually shows.
Success in reading depends on linking print to speech. There’s a massive amount of behavioral research, neuroimaging research, on brain organization and brain development, which conclusively shows that skilled reading is associated with children’s spoken language, grammar and the vocabulary they already know. It’s about teaching kids the correspondence between the letters on a page and the sounds of words.
This sounds very “old school” to me and (as with all things old-school) immediately sensible. What should puzzle us is that grade-school teaching was ever disconnected from this insight. And this is where things get tricky for me. I want to celebrate Seidenberg for speaking the truth to teachers, but I fear that the problem itself arises because the teaching profession is, increasingly, guided by research. If teachers had been able to maintain autonomy over their own teaching methods, they would never have abandoned the close connection between learning to read and reading out loud. And then I wouldn’t have to teach students to read out loud when learning how to write clear, scholarly prose. It would just be natural.
I don’t know much about the scientific literature on reading at the grade school level, so I don’t know exactly when exactly what went wrong. But I do suspect that the distance between the spoken and the written word grew substantially under the so-called “post-modern” conditions that were inspired by Derrida’s “deconstruction” of “logocentrism”. At one level, after all, it was precisely an attempt to free the written word from its servitude to speech. I don’t think it’s entirely coincidental that complaints about the “turgidity” of contemporary academic prose are often traced back to Derrida’s influence. And it can certainly be demonstrated that composition studies has been profoundly affected by this influence. Indeed, I’m increasingly confident that literacy studies has been deconstructed as well, so it would not surprise me to find that grade-school literacy practices have been deliberately freed from the shackles of logocentrism. It would not surprise me if this can be shown to have had a detrimental effect on the reading level of school children.
My mixed feelings about Seidenberg’s suggestion, then, stem, not from any disagreement I have with him, but from the authority that science increasingly has over teachers. I don’t think teachers should adopt phonics on the advice of science, but on the counsel of common sense. (Indeed, it was also common sense that should have pushed back against the “ideological turn” in literacy studies and the “process turn” in composition studies.) It has never really made sense to separate writing from speech entirely–to let writing live a life of its own, independently of the sound that our words make. It only made sense after we taught ourselves to trust “research” more than the evidence of our own senses. Or rather, at that point we had begun to happily believe things we didn’t understand, to adopt practices that didn’t really makes sense to us, because one or another “study” had “shown” that some new pedagogy was needed to get us “beyond” traditional teaching methods. I don’t think it made things easier.
“… on the counsel of common sense.” I really like that phrase.
And you’re right about the argumentative role of “brain research” in that article, i.e. it serves no substantive argumentative role in that article (and its potential evidential role is left completely unspecified). It is mere rhetorical flourish. “Mere ornament” or “an idle cog” in Wittgensteinian parlance.
I’d be very cautious advocating ‘common sense’ because beating children when they get a word wrong an assignment was also once considered ‘common sense’. But what changed it was not more science but a renegotiation of what made ‘common sense’ in this context. The ‘research guidance’ can ultimately only be one of the interlocutors in this negotiation.
The problem with both ‘common sense’ and research in complex issues like this that both can be marshalled for both sides of the dispute. Reading is so different from speaking/listening that it makes just as much sense to assume that children should just form a direct link between the shapes on the page and the words they represent. It’s more important that they read things for pleasure and use develop reading skills by frequent use.
And the fact of the matter is that, the latter is more important for true literacy. And it works for a huge part of the bell curve of reading skills distribution. Common sense tells us that non-phonics focused approaches did not transform our cultures into wastelands of literacy. Common sense also tells us that everybody reads in different ways and has different reading needs. So we should be very cautious of all these measures of illiteracy. We are also moving the goalposts on literacy, so it is possible, things got a lot better in this area since the 1960s (when the new approaches were introduced), but we just expect more.
Now, this is where research comes in. It is there to help us adjudicate these common sense impressions (kinds of stylised facts or hypotheses). But both the correlations and effects while consistent are incredibly small by common sense standards. The time frames are too short and the research guided advice is contradictory.
Now, the research is unequivocal that for children with reading difficulties, synthetic phonics is best. But this needs to be combined with very structured, methodical instruction, ideally in small groups.
But synthetic phonics has nothing to do with what you imagine linking sound to writing to be. It is purely about developing and building on a phonological awareness. It will let children learn things like s and i and t, pronounced together give you ‘sit’. It is never about reading whole sentences for meaning (that is the not research approved method). All of this method will not make children better writers – just spellers. Or even better readers – just better decoders (decoding letters into sounds).
And, this works despite common sense. Synthetic phonics makes sense for regular languages like Italian because most sounds match their letters more or less regularly. You would expect it to be much less effective in English where almost no regularity is present. All three letters in ‘sit’ can be pronounced differently in different contexts ‘sugar’, ‘bite’, ‘tuition’. So it would make sense that it’s better to learn the bigger patterns. But no, it is better to start with the regularity, even if some of those mappings will then have to be unlearned in some contexts. Without this research evidence, synthetic phonics would be even more controversial now than it is.
Another example of research going against common sense in a definitive way is bilingual education (or the promotion of bilingualism in general). You would think that teaching Mexican students only in English would be better for them but it turns out (quite definitively) that teaching some subjects in Spanish is much better. You would also think that bringing up children speaking only one language would be better for them and that speaking two languages at once could be harmful to their abilities – but research shows conclusively that it’s not true. But then research goes an extra step to say that bilingualism is good for the brain and here we can say – hm – if that were the case wouldn’t we expect bilingual countries to have significantly higher IQs, lower dementia or higher GDP? Common sense tells us that this is not the case in any noticeable way. Research may find some slight correlation but it’s very noisy and for any individual deciding to invest in becoming bilingual to increase their IQ – using this research is little better than flipping a coin.
So we need both common sense and research in conversation here. The research is not as determinative as we often make it out to be – basing decisions on research results in some areas of education is no better than flipping a coin. But common sense needs to be interrogated because it is too often steeped in prejudice and unexamined assumptions. Research is one way in which this interrogation can take place.
Thanks for the thoughtful comment, Dominik. I agree with most of what you say. I definitely think common sense needs to be challenged in conversation. The trouble with research, however, is that it’s often implemented by people who are not qualified to challenge it. That is, the research is taken to be “true”. It’s not the beginning of a conversation but a call to action.
The key is to make sure that the people who are implementing pedagogy have the authority to critique it. If somebody claims their pedagaogy is “common sense”, you can challenge them with your own reason and experience. But if someone says, “Studies show that…” you have to go out and do research yourself to show how it’s wrong. To me, the important thing is the conversation about our reasons and the experiences that we derive them from and test them against. Research too often disenfranchises would-be critics.
Hi Thomas, I couldn’t agree more. Research is often bandied about too easily. And it is used as a deus ex machina (btw: Research Ex Machina would be a great blog post title, hint); the end of a conversation, not a part of it. Neuroscience is the worst offender these days (as the story you reference illustrates).
Some years ago I co-wrote a paper called Epistemology as Ethics as part of a project on how philosophy can help policy makers evaluate research (Excerpts here: http://metaphorhacker.net/2011/03/epistemology-as-ethics-decisions-and-judgments-not-methods-and-solutions-for-evidence-based-practice). Our conclusion was that it can only do that as part of a conversation. Now, I would think about it in terms of Rorty’s work on the limits of epistemology.
Another way (less offputting to many), is to think about it in terms of Bayesian updating. I’ve been thinking about a modern take on Dewey’s pedagogic creed as a list of priors about education against which to assess how much we should worry about jumping on the latest fad definitely, absolutely confirmed by research results.