Saturday, April 3, 2010

Phonics vs. "Balanced Literacy": the Hebrew experiment

Learning how to break the code of written language is one area in which all the evidence shows that the left-brained/analytic/phonics approach trumps the right-brain/holistic/Balanced Literacy approach--for all types of students.

The most recent example of how ineffective the Balanced Literacy approach is can be seen on the following Youtube video, which was recently posted on kitchentablemath:

In the ensuing discussion thread, I had the following idea for an experiment.

The idea is to take American adults and turn them back into novice readers via a not-too-complicated phonetic writing system consisting of sound-to-symbol correspondences with which they are totally unfamiliar. The Arabic and Hindu writing systems strike me as too complicated; the Greek and Cyrillic systems involve too many familiar symbols and sound-symbol correspondences.

But you'd have to do this without introducing the added challenge of learning another language.

So what you do is you transliterate a bunch of simple English words--of the sort you'd start beginning readers on ("cat", "mat", "can", "pan", "and", "in", "on", etc.)--into Hebrew letters.

Then you teach American adults who don't already know how to read written Hebrew how to read this list of transliterated words. Group A learns via X hours of phonics-based lessons; Group B via X hours of balanced literacy. Then everyone gets a post-test in which they have to read a list of Hebrew transliterations of the words in question.

Of course, these adult subjects aren't completely comparable to American children learning to read for the first time. On the one hand, they already know how to connect phonemes together into syllables (p-e-n -> pen). On the other hand, they haven't had the kind of prior exposure to the Hebrew alphabet that beginning readers usually have had to their native alphabets by the time they start to learn to read. On the third hand they are adults, not children.

However, I still think that this experiment would be quite revealing, both in the likely results it would produce, and in how it would remind those who need reminding of (1) what it's like to learn how to read, and (2) how impractical it is to memorize written words as graphical wholes.


Liz Ditz said...

Katherine, Dehaene reports on just such an experiment in Reading in the Brain, p. 225-228. Yoncheva, Blau, Maurer & McCandliss (2006) Strategic focus during learning impacts the neural basis of expertise in reading, Poster presented at the Association for Psychological Science Convention, New Your, May 25-28.

The task involved memorizing words written in a novel orthography. Whole language did better the first day, but direct instruction pulled ahead by the second day and trumped whole language in the end.

Seth said...

The Arabic writing system isn't really that complicated. For both Hebrew and Arabic though you'd probably want to include the short vowels, typically left out of writing.

On a different topic, have you seen the ACM paper on teaching computer science in the math curriculum?

Note in particular the "level 1" teaching. Mentions Logo, etc.

LexAequitas said...

The kana in Japanese would be terrific for this -- they're phonetic, and each is probably pretty comparable in complexity to English letters.

A=あ (hiragana),ア (katakana)
I=い (hiragana),イ (katakana)
U=う (hiragana),ウ (katakana)
E=え (hiragana),エ (katakana)
O=お (hiragana),オ (katakana)

Katharine Beals said...

Thanks for the suggestions! Thanks for summarizing Dehaene's study, Liz; not sure how easily I can read it without actually buying the book.

Haven't seen the ACM paper, Seth, can you send a link? (I gather you've been learning Arabic!)

LexAequitas, thanks for the Kana suggestion. I did think about the Japanese system, but wondered about a syllabary like this one introducing confounding variables.

Seth said...

The link was in the above:

or more directly, here:

Let me see if this works so that it will come out as an actual link:


re Arabic, yes a few years now.

Liz Ditz said...

Katherine, you could contact McCandliss -- a copy of the poster presentation wasn't on either of his faculty web pages.

And there's this:

Maurer, U., Blau, V.C., Yoncheva, Y.N., McCandliss, B.D. (in press).
Development of visual expertise for reading: rapid emergence of visual
familiarity for an artificial script. Developmental Neuropsychology

It looks like they did some EEG research as part of the research, too.

A specific component of the EEG waveform is called the N170 (since it is negative, and occurs at 170 ms). McCandliss and colleagues have several papers in press with N170 in the title:

Maurer, U., Blau, V. C., Yoncheva, Y. N., & McCandliss, B. D. (in
press). N170 in learning to read a novel script: early visual familiarity enhances right lateralization. Developmental Neuropsychology.

Yoncheva, Y. N., Blau, V. C., Maurer, U., & McCandliss, B. D. (in
press). N170 in learning to read a novel script: the impact of attending to phonology on lateralization. Developmental Neuropsychology.

Katharine Beals said...

Seth, thanks for the links on the ACM standards. It's exciting to see their interest in pushing programming skills in k12. I agree with you that we are missing many opportunities to teach these skills--the truest problem solving skills out there, imho!--to young children. Logo seems OK to me, but when we tried it with J., we found it somewhat limiting in terms of real basics. Why not Basic? (not VB, but plain old B)?

Anonymous said...

Interesting idea, except that Hebrew orthography is not all that suitable for English vowels.

For examples, I;ll use your list:

"cat": Hebrew can note "cut" and "ket" but not "cat."

"mat": there's "met" and "mutt", but not "mat."

"pan": "pen", "pun", but no "pan."

Armenian or Georgian might work better for this experiment.

Katharine Beals said...

Thanks for the suggestion, anon. Armenian looks perfect! An alphabetic writing system with symbols that are truly novel to most Americans.

14lisawho said...

Theory and practice are the theory. Of course you can teach reading with phonics when there is direct letter-sound correspondence. Too bad the majority of English words don't have that correspondence. And how would you measure "reading ability?" Decoding words is not reading. Reading is about making meaning from print. "Phonicators" are not readers.

FedUpMom said...

The difference is that phonics-first means you read the word, then figure out the meaning. Whole language means you try to predict the meaning, then try to guess what the word is based on your prediction of the meaning.

I just blogged about this problem:

A Pox on "Progressive" Whole Language

Katharine Beals said...

"Theory and practice are the theory... Too bad the majority of English words don't have that correspondence."

The majority of English words have enough of a letter sound corresponence that sounding them out via the standard rules of English phonics consistently provides more clues about their identity than any other reading strategy. Consider, for example, the words
"theory", "and", "practice", "are" "the" "same" "in" "theory".

Or consider how you, as an adult, would deduce the pronunciation of an entirely new word--say a new medical term like "cysticercosis".

"Decoding words is not reading. Reading is about making meaning from print. "

Decoding is the first step. Making meaning from print is easiest when decoding happens automatically. For this, for nearly all kids, phonics instruction is a necessary boostrap.

Jenny said...

The video clip in this example is really more of a Whole Language activity. A Balanced Literacy classroom would have lots of phonics activities in addition to patterned book. In this example, of course I would have the child sound out "cat" and other words she might use in addition to pointing out the picture cues.

When I taught in a Balanced Literacy classroom, I had Wright Group/benchmark books to use, but I also taught all of the kids with a program called Systematic Sequential Phonics They Use. The only two kids who couldn't read on or above grade level by the end of the year ended up being diagnosed with LDs.

A final point is that your experiment with Balanced Literacy and adults really isn't a fair comparison at all, because you and I don't have young brains. I can't tell you what we listened to in SOTW #3 in the car two days ago, but my son can tell you every last detail of what happened and compare it to SOTW #2 which we listed to a month ago.

Young brains learn language and other things faster, which is why early education is so important. It's also why Balanced Literacy, which includes the good elements of Whole Language plus a ton of phonics, works so well.

Katharine Beals said...

I'm not aware of any reasearch showing that young brains are better at memorizing strings of graphemes than older brains are. There are certainly differences between old and young brains, but so far as I know, this isn't one of them. Do you know of any studies on young brains and grapheme-string memorization that you can link to here?

Regardless, it seems prudent to use teaching strategies that, per a given teaching goal, place a minimum strain on memory, and make maximum use of powerful mnemonics (such as phonics).

Balanced Literacy clearly means different things to different people. If it means intensive phonics instruction in the very early grades, I don't object to it on those grounds.

My source on Balanced Literacy is Diane Ravitch's "Death and Life of the Great American School System," pp. 36-37. She writes that while BL uses some phonics (and whole language) "it focuses mainly on teaching reading strategies" like prediction, visualization, and inference, and has the teacher functioning as facilitator while students work in groups. If applied to the early grades, this focus, and this pedagogy, does not optimize the mastery of phonics.