Is Morphology Training Better Than Phonics Instruction?

Man, sometimes when you publish a blog entry you’d wish you stayed in bed.

You hope to write something that someone will find useful. But the responses might make you feel more like you’ve been dropped onto the set of Fox News or MSNBC.

Recently, I’ve experienced some interesting responses.

For example:

Studies show that phonemic awareness (PA) training helps young kids learn to read (PreK through Grade 1). From those studies, I claimed we should teach PA to the point where kids can fully segment words into their individual phonemes. This conclusion was based both on the experimental impacts of training studies and on large-scale analyses of children’s development.

Before long I heard from David A. Kilpatrick, author of Essentials of Assessing, Preventing, and Overcoming Reading Difficulties. He took exception with my claims about the importance of segmentation. He rightly pointed out that studies show that phoneme manipulation tasks are harder than segmentation tasks; that phoneme manipulation ability continues to develop through Grade 4; and, that PA continues to correlate with reading through Grade 6. Based on this, he believes we should continue to teach PA long beyond full segmentation.         

How can one disagree with such a thoughtful and reasonable argument?

My position, ultimately, is based on the fact that there are more than 100 studies showing the learning benefits of PA instruction—in PreK, K, and Grade 1.

And, how many studies show its benefits in Grades 2 and up?

None.

David Kilpatrick may be right about the value of continuing the PA instructional regime into these upper grades. Perhaps there are achievement points to be found in continuing PA instruction to the point where kids can easily do mental manipulations of phonemes (e.g., adding, deleting, reversing). But you won’t see me touting that until someone actually proves a learning benefit for kids.

Hypothesizing a benefit and proving a benefit isn’t the same thing.

More recently, I’ve been smacked upside the head by several readers upset with me for not proposing more, and more thorough, spelling instruction and morphology instruction focused on spelling aimed at advancing students’ reading ability.

Some of those arguments have been enthusiastic.

Traditional phonics instruction emphasizes letters and sounds but ignores the morphological and etymological reasons for spelling, my critics have pointed out. Reading experts have long recognized the importance of the morphological aspects of word meanings, but there has been little pedagogy aimed at the morphological aspects of spelling.

I’ve been sent lots of linguistic evidence to convince me of the morphological nature of our spelling system — and most of that work cites Dick Venezky’s seminal work.

In the 1960s, when computers first allowed for the ambitious quantitative study of language, Dick revealed the surprising consistency inherent in the English spelling system. Contrary to what was long believed — that our spelling system was a confusing mess — Venezky argued that whatever was lost in ease of pronunciation, was more than regained in the consistency of meaning inherent in our spellings. Hence, the endings of dogs and cats may be pronounced differently: /z/ and /s/, but their identical spelling consistently and helpfully signaled plurality.

I’m happy to see that Dick’s work continues to bear fruit in linguistics (he was one of my teachers — he even helped me to design morphology-oriented spelling measures for my doctoral dissertation). But I think he’d be surprised to hear his work used as an argument against phonics instruction — he was a big phonics proponent (though I’ve seen him offer the same kinds of linguistic critiques of phonics programs that have been sent to me recently).

Dick not only had expertise in linguistics but extensive knowledge of psychology and computer science. He knew that teaching kids to read was different than inputting a linguistic system to a computer. Despite the flaws and shallowness of many (most) phonics programs when it comes to features like morphological sophistication, such teaching still gives students a clear learning benefit.

What Dick Venezky came to believe was that phonics instruction gave students “clues” to the English spelling system. Students then use those clues to figure out how the system works. Phonics instruction does not teach everything one would need to “decode” text, but it provides useful pointers and puts kids into a mindset of trying to understand the system.

That doesn’t mean he would — or that we should — reject the idea of introducing morphological explanations and “clues” earlier, only that we shouldn’t be so sure that it would improve things as much as some morphology proponents assume.

For example, one colleague pointed out that in some phonics programs, kids are taught to divide the syllables of “action” in the following manner: ac/tion. He argued that this was a bad choice because it obscures that the root word is “act.” That’s correct linguistically, but does it matter when you’re 7?

Initially, we hope to teach kids enough to allow them to come up with an approximate pronunciation of a word that is in their mental lexicons (primary grade kids know 5,000-10,000 words). It is more likely they’ll come up with “action” by saying “ak/shun” then by saying “act/ion.”

Of course, if they don’t know the word action — don’t know what the word action means — then, breaking the word the second way (emphasizing “act”) may just get them to the meaning no matter what the pronunciation.

The issue here turns on what would be best for beginning readers … is it best to help them to figure out the meanings of unknown words or to help them to translate print to pronunciations of words already in the child’s oral language? I think it is the latter, so I don’t mind delaying most morphological work until phonics is mastered (e.g., Words Their Way).

However, there are arguments that we should teach morphology earlier and even in place of phonics instruction (one critic wrote that the National Reading Panel findings were out of date since we now know morphological training to be more beneficial than phonics). Eeks!

I looked at these critics’ evidence (Bowers & Bowers, 2017 provides a nice summary of this work). Specifically, they point to two studies of morphological training for young children. One especially weak study — impossible to tell if the outcomes were due to the training or to existing ability differences in the participants — claimed long-term benefits to preschool morphology training.

And, an experimental study that examined the impact of 10 hours of morphology teaching: This one claimed to enhance reading performance by more than a grade level! Not surprisingly, the outcome measures used were tightly aligned to the training and there were other design problems, too.

That’s the entire body of instructional research one could use to prescribe instruction for preschool and primary grade kids (and in both studies, everyone got lots of phonics instruction, too—not exactly proof of the inadequacy of phonics).

Again, I can’t really say these folks are wrong — we might be able to affect clear reading improvement by teaching the morphological aspects of spelling earlier and more thoroughly, instead of what we currently provide with phonics.

But I definitely won’t be prescribing reading instruction based on a single 10-hour study.

The reason why I insist that we teach phonemic awareness, phonics, vocabulary (word meanings including morphology), oral reading fluency, reading comprehension strategies, and writing is because there are dozens, even hundreds, of studies done by different researchers, with different kinds of kids, with different variations on the instructional routines, but with a consistent and substantial learning payoff. Why trust 100 such studies on phonics — some carried out for as long as 3-years — over a single small study of 10 hours of morphology instruction? I think you can probably answer that one for yourself.

I hope researchers will continue to propose provocative hypotheses about learning, and that they’ll continue to evaluate these ideas rigorously under a broad array of instructional conditions. And, if they find something that consistently helps kids, then I hope we’ll adopt their ideas. Until then, I won’t be recommending morphology over phonics or other terrific but unproven ideas — no matter how intelligently, reasonably, or vociferously those opinions may be stated.