Cats, catnip, dogs and the absurd

I my previous post, I explained why I thought that Structured Word Inquiry (SWI) and its proponents’ insistence that morphology and etymology should be taught right from the very beginning of reading instruction was at odds with basic cognitive science. I was left with having to speculate on exactly what teaching in this way would look like because I also highlighted the fact that sources that supposedly demonstrate how this is done do not actually feature initial reading instruction.

I was therefore grateful for Jeffrey Bowers, a researcher and advocate of SWI, taking the time to comment and explain what he thinks this would look like:

“This is absurd… let’s consider initial reading instruction… and the case of teaching the word “cat” using SWI. The teacher would talk to children and ask them what does the word “cat” mean? And what about “cats”? and “catnip”? etc., the teacher can talk to completely illiterate children about what is the difference between cat and cats? Do you know what catnip means anyone? No, I’ll tell you. Etc. And then you can break down the word “cat” into its graphemes one at a time. Sound it out. Then add the “s”. Consider how it sounds. Say, OK, we now know that the graphemes “c”, “a”, “t”, and “s” can be pronounced… Next they might learn about the “dog” and “dogs”. And then have a conversation about why there a common letter “s” in “cats” and “dogs” when the sound differently. Etc. The idea that any of this is inconsistent with working memory and cognitive load is absurd.”

I have drawn a map of the items that I think are relevant to the teaching sequence that Bowers describes:

I agree that most, but perhaps not all, children who are about to learn to read would know what a cat and a dog are in the sense that these words would be in their aural vocabulary. That does not mean they would be able to map the written letters of ‘cat’ and ‘dog’ onto these words in aural vocabulary – a process commonly known as ‘decoding’. Similarly, they may have an implicit, or perhaps even explicit, understanding of the phoneme(s) represented by ‘s’ and that this relates to plurality. I doubt many would know about ‘catnip’, as Bowers acknowledges, and so we cannot assume this. So with a little fuzziness around exactly what relevant schemas our children already posses, I reckon there are at least 10-12 distinct items there – well above the four-item estimate of working memory capacity.

As an aside, I have used the word ‘synthesis’ to represent the understanding that the different sounds of the word-parts come together in the whole word. I know that this may remind people of synthetic phonics but I could not think of a more neutral word and the principle of a relationship between the parts and the whole stands whether you are going from the parts to the whole (synthetic phonics) or the whole to the parts (analytic phonics).

To be fair, Bowers does not appear to be suggesting that children attend to all of these items at the same time. We could perhaps imagine working our way through 3-4 of them and then another 3-4 of them and so on. However, there is more to learning than simply ensuring that the flow of information remains under the 4 item limit. Otherwise, we would all remember every detail of most of the programs or films we watch or books we read. No, learning appears to require some form of active retrieval or reconstruction by the learner. It requires practice.

For instance, in Rosenshine’s principles, drawn from the practices of the most effective teachers, we are advised to, “Present new material in small steps with student practice after each step… more effective teachers do not overwhelm their students by presenting too much new material at once. Rather, these teachers only present small amounts of new material at any time, and then assist the students as they practice this material. Only after the students have mastered the first step do teachers proceed to the next step.”

This is mirrored in the experimental work that underpins cognitive load theory. For instance, researchers have found that studying a worked example, immediately solving a similar problem and repeating this process is superior to studying a number of worked examples and then solving a number of similar problems. Many teachers have now drawn on this evidence to build ‘example-problem pairs‘ into our teaching practice.

So perhaps we could add small steps and practice to Bowers’ model. However, it all seems a little incoherent. Do we want students to be fluent with GPCs so that, in the future, they may easily decode unknown words such as ‘gruffalo’, ‘mitosis’, ‘hufflepuff’ and ‘Acidulate’? Do we put this on an equal footing with knowledge of catnip? What about knowing that the ‘s’ in ‘dogs’ represents a (slightly) different sound to the ‘s’ in ‘cats’? How critical is that in reading lesson number 1? Do we want to touch on the GPCs in ‘dog’ and ‘nip’? If not, how are the children meant to be interacting with these words? Do we want them to learn the word ‘dog’ and the ‘nip’ in ‘catnip’ whole? Conceptually, the relationship between cats and catnip is mirrored in the morphology – how important is it for students to learn this word sum? Do they need to know the origin and meaning of ‘nip’ for this to make sense?

We could potentially still go through the same teaching sequence but decide that some items are just not the target of learning. They are there for perhaps motivational reasons and so we don’t need to practice them because we don’t mind if students forget them. This would still come at a cost because we would probably need to use some of the children’s cognitive resources in signalling what is and what is not important. However, even if this worked, the evidence from cognitive load theory suggests that such extraneous information hinders learning. Making progress in learning is likely to be motivational and possibly more of a motivator than these extraneous items. In the longer term, if we use non-optimal approaches to early reading instruction and if the effect of that is more reading failure, we are storing up the issue of disengagement for some future date when the child figures out that he or she cannot read.

I therefore cannot see any advantages in the teaching process described by Bowers over one that focuses on GPCs in order to get students reading words and controlled GPC books as soon as possible. So I won’t be arguing that my school abandon Sounds-Write just yet.

However, I do think I am starting to understand where all of this comes from. I have recently become aware of a set of posts by Katharine Beals about SWI. In part iii, she notes that SWI proponents seem to favour inquiry learning due to the excitement it is assumed to induce. At one point in his comment on my last post, Bowers asks, “How can someone who agrees that morphology is relevant to instruction not at least be intrigued by the morphological matrices that organize groups of words into morphological families in a way that highlights their spelling, meaning, and phonological consistencies?” The answer to this is, “Morphological matrices don’t strike me as either particularly original or particularly interesting.” However, if you believe that is some deficit on my part and you believe that morphological matrices really are the catnip of reading instruction then you can convince yourself of vast untapped reserves of motivational potential that are going to waste. Hence:

“…there are good reasons to think that SWI can address the main criticism of phonics, namely, the view that an emphasis on grapheme-phoneme correspondences is not engaging for many children. For example, when discussing the disappointing results of some phonology-based intervention studies, Snowling and Hulme (2014) argued that intervention studies need to focus more on pupil motivation with the aim of increasing students’ enjoyment of reading. We would suggest that SWI is a promising approach in this respect given that it aims to give children an understanding of the meaningful organization of the writing system through word investigations. As noted by Dunlosky et al. (2013): Anyone who has spent time around young children knows that one of their most frequent utterances is “Why?” (p. 8) Indeed, nothing motivates like understanding.”

I have no reason to assume that SWI will be superior to phonics in motivating students. All learning requires an element of hard work that is not always fun and anything can be made relatively more or less interesting by how it is presented. I agree that understanding is motivating, but understanding what? Understanding the words on a page because you can decode them and match them to your aural vocabulary or understanding some finer point of a morphological matrix?


5 thoughts on “Cats, catnip, dogs and the absurd

  1. Jenny Chew says:

    Jeff’s account, as quoted by Greg, is useful to have. Would the teacher then check, perhaps the next day, whether the children could read (and possibly spell) ‘cat(s)’ and ‘dog(s)’, or is it not considered important that they should be able to do so?

    It WOULD be considered important in a synthetic phonics approach. Moreover, those few GPCs would have been taught in such a way that the children would know that the letters could also form other words (e.g. as, at, tag(s), cog(s), tot(s), god(s), got, gas – possibly also stag(s) and Scot(s)) and would have practised reading the words by sounding out and blending and spelling them by segmenting the spoken words and choosing letters to represent the sounds. Synthetic phonics stresses the generative power of GPC knowledge from the start. Obviously there comes a time when children have to learn that most graphemes can represent more than one sound and most sounds can be represented by more than one grapheme, but this is not a big problem if the principle mentioned by Greg is followed – ‘Present new material in small steps with student practice after each step’.

  2. katharinepbeals says:

    It’s raining cats and dogs!
    Thanks for bringing up the motivational effects of making progress in learning. One problem with ad hoc inquiry-based approaches (of the sort we see in the various SWI videos) is that there’s very little sense of progression, let alone of progress.
    Thanks also for exploring the cognitive load issues involved in teaching via morphological families. Jenny is right to point out that a much more productive approach would be to teach by what we might call *phonological* families. The most obvious of these are rhyming families: “cat”, “rat,” “bat”, “sat”, “mat”, etc.
    However, as I mentioned in my last post at, I’ve heard one SWI proponent object to grouping things by sound because, she says, it will confuse kids about conceptual categories. Apparently, kids might come away thinking that mats have four legs, or that rodents are felines.
    This SWI prononent then proposed that teaching “cat” along with “cats”, “catnip”, “catty” and, iirc, “she-cat”, both avoids this problem and is more effective. (When I then asked about how novice readers take what is presumably the first step of identifying the word “cat” I got the usual evasive replies.)

    • Jenny Chew says:

      Hi Katharine –

      Just to clarify, synthetic phonics as we know it in England does not stress rhyming word families, but DOES try to ensure that children have early experience of working with the few graphemes that they know in all positions in words. Jeff’s examples used 7 letters (3 in ‘cat’, 3 in ‘dog’, and ‘s’), but he seemed to envisage the children working only with the words ‘cat(s)’ and ‘dog(s)’ at this point. i was trying to point out that if a synthetic phonics approach was being used with children who knew just those 7 letters and 8 sounds (two for ‘s’), the teacher would have the children also working with words in which the same GPCs occurred in other positions in words, but no new letters were introduced.

  3. Harriett Janetos says:

    Katharine Beals refers to a great video where at 8:30 we see the difference between the “spelling out” strategy that SWI uses and the “sounding out” strategy that I use. The student repeatedly reads “healer” as “healthier”. In looking at “heal” I would have my students sound out the three graphemes they see, but this student is directed to spell out the word. Stanislas Dehaene’s book Reading in the Brain explains how the brain attends to phonemes, and one of the problems of spelling out a word is that some letter names contain more than one phoneme: ache-ee-ae-el vs. /h/ /ea/ /l/ which impedes decoding, particularly for struggling readers.

  4. Pingback: Cats, catnip, dogs and the absurd – Filling the pail – The Literacy Echo Chamber

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.