Sunday, June 14, 2015

Teaching syllables or morphemes? Why Sandman-Hurley is mistaken

A couple of weeks ago I came across a piece by Kelli Sandman-Hurley in Edutopia entitled 'Teaching Syllables Can Mask Meaningful Morphemes'. In the article, which you can read here, Sandman-Hurley starts by asking how many times you’ve seen the word 'every' spelled as 'evry'.

This is indeed a close approximation of what we hear when we are speaking normally in conversation. On the assumption that you have seen this misspelt hundreds of times (which I have), what then, she asks, would you do to remedy the situation?

Clearly, there is often a mismatch between what we hear and what we spell/write and this is the kind of thing that is happening when writers spell 'every' as 'evry'. When speaking normally we would enunciate the word as two syllables, 'evry', instead of three. Sandman-Hurley states that her practice was to 'over-pronounce' the word, a practice she now believes is 'questionable at best and detrimental at worst'. Unfortunately, the reasoning behind the argument for why she now believes over-enunciated syllabification to be a wrong approach is seriously flawed.

She begins by stating – correctly  that ‘English is not a syllable-timed language. It is a stress-timed language’. But then she goes on to say that ‘this means that syllables bear little to no effect on our writing system’. In actual fact, the opposite is true for reasons I shall sketch out.

What ‘over-pronunciation’ or a 'spelling voice' does is it enables the writer to hear all the sounds in all the syllables in English words. This technique gets over two principal problems: the problem of the schwa in many words and the problem of elision. To take the last first, elision is what happens when we drop out or suppress a sound or syllable in pronunciation. Examples are 'government', in which almost all speakers elide the /n/, and 'chocolate', in which we elide the entire sound/syllable in the middle and produce 'choclut'. 

Then, there is the ubiquitous problem of the schwa, the most common vowel sound in the English language. If you define a schwa in functional terms - as 'a sound that isn’t spelt as it sounds' - once you’ve identified it in a word, the solution to spelling it correctly is solved by saying it precisely in a spelling voice.

So, how we can identify the schwa sounds in words. The answer lies in the fact that schwas are the most frequently occurring vowel sounds in the English language and they are always associated with weak syllables in polysyllabic words. For example, in English, we tend to lay stress (usually) on only one syllable in any polysyllabic word. Thus, in a word like 'chicken' (two syllables), the stress is on the first syllable 'chi'.

So far, so straightforward! The problem is that other syllables in a polysyllabic word may contain a syllable or syllables which are not stressed and it is these that often (though not always) contain a schwa, or weak vowel sound. So, in the word 'chicken', the unstressed syllable is the 'cken'. In the syllable 'cken', the spelling [e] represents the sound /i/ or /uh/, depending on accent, which is a weak vowel sound or schwa.

How does this impinge on writers when they are spelling a word? The answer is that, if they’ve never seen a particular word spelled before and it contains a schwa, they may not know how to spell it, especially if the vowel spelling is one that is less frequently encountered. To compound the problem the weak vowel sound is spoken in different ways, according to accent. For example, a person from the southern states of the United States may say 'chickun'; whereas, people from most parts of the UK will say 'chickin'. In each case, the weak vowel sound will be said slightly differently - as an /uh/ or as an /i/.

So, how we are able to teach pupils to recognise the weak vowel sounds in polysyllabic words. The answer to that is, well, child’s play. Anyone who is a mother-tongue (L1) speaker of English will place stress on one of the syllables in polysyllabic words. Try it with these words: 'chimney', 'magnet', 'cavalry', 'thermometer'. You should find that the underlined syllable in each word is the one you are stressing.

L1 speakers can also break words into syllables very easily. If you remain unconvinced, visit any nursery and watch children clap out syllables in words. It being the case that even young children can syllabify words naturally, by the time they get into Y1, we can make them explicitly aware of the strong syllable. When this is achieved, getting them to recognise the weak vowel sound in sub-dominant syllables is a piece of cake. I know because I’ve been teaching children this for years.

Unfortunately, instead of 'misrepresenting how the written word works', as Sandman-Hurley’s claims, by over-enunciating, we are matching exactly what is written to what we hear. You will also notice if you read her piece that Sanderman-Hurley’s understanding of what a schwa is is incorrect. She describes the second syllable in ‘trusted’ as a schwa, when, in fact, it is only the vowel [e] that is a schwa.

If pupils are properly instructed in how to segment and blend sounds in words into syllables, starting with simple words, such as 'bedbug' and 'desktop', moving on to words with a more complex structure, such as 'groundsheet' and 'earthquake'. From there, it is easy to build on to ever more complex words containing three, four, five and six syllables.

Of course, teaching pupils about morphemes is useful to them – if teachers are able to explain what exactly they are and what they mean. Unfortunately, this isn't always the case. In a recent study, Herrington and Macken-Horarik (2015) found that in a group of twelve teachers, not a single one could ‘identify the correct definition of a morpheme’.

So, going back to Sandman-Hurley’s example, how would I teach ‘every’? The division is best taught as /e/ | /v/ /e/ | /r/ /ee/. Here, the schwa is the [e] in the middle syllable. A spelling voice, enunciating the spelling of the schwa [e] as /e/ overcomes the problem. If this is followed up by asking pupils to write the word as three separate syllables and to say the sounds as they write, the practice greatly helps them to remember it.

As a teacher, one would also make absolutely clear to pupils that we are not trying to get them to speak in this way in everyday conversation, and that we would only employ these techniques when faced with a word that contains schwas or in which we elide a sound or syllable. 

Ignoring syllables to focus on morphemes is putting the cart before the horse. My advice is to teach pupils to read and spell efficiently before you embark on teaching the first steps in morphology.


Holly Shapiro, Ph.D., CCC/SLP said...

John, I'm excited to have an opportunity to engage in a healthy debate with someone I truly respect. And good catch on the schwa error. Yes, only the vowel sound is unstressed in the second syllable of . On to the meaty stuff. Believe you me, I have taught many a student to use their "spelling voice" in my day. And it did get the job done for the word in question. However, my thinking has since changed. Take a word -- the first one that comes to my mind is as in "Please separate these beads for me." We can teach our student to use a spelling voice, overpronounce the second syllable, and say it as "sep-air-ait". On the other hand we can explore the base which has the sense "to make ready." In one brushstroke, we can build an understanding of a larger family of words (think pare, apparatus, compare, disparate, preparatory) that share spelling structure and meaning but not necessarily pronunciation. Lately, I've been going In this new direction (well, new for me) not because it's "right" but because it's more productive. I also must quibble with you about teachers. The fact that teachers may not know a morpheme from a hole in the wall cannot be used as justification for teaching one way over another. However, what a great conversation this could be about how teachers, myself included, are trained.

Holly A. Shapiro, Ph.D., CCC-SLP
Ravinia Reading Center

John said...

Hi Holly and thank you for posting a comment.
My response is going to be nuanced and I'm going to agree and disagree. I disagree with you when it comes to teaching beginning readers. This is because I stand by what I said in this post When children have acquired a degree of fluency, then it is possible to start bringing in aspects of grammar and lexis on a more controlled and systematic way.
I'll clarify more exactly what I think can be taught from a morphemic perspective in another post I hope to find the time to write in the next two weeks.
I wasn't suggesting, by the way, that the whole project of teaching children what morphemes are should be constrained by the fact that many teachers aren't taught anything about them in training. Obviously, we do what we can to bring this knowledge to the attention of teachers through the various means at our disposal but what I was saying is that it's difficult when the training institutions seem to show little or no concern about this area.
I would also bring back the teaching of Latin in schools on the grounds that it provides a superb vehicle for preparing students for learning a Romance foreign language and because (circa) 60% of English words have a Latinate base. Take your example 'pare', for instance. As I'm sure you know, it is derived from 'paro, parare, paravit, paratum', meaning 'prepare, get ready, provide, obtain'.
How I would teach 'separate'? I'd split it as /s/ /e/ | /p/ /a/ | /r/ /ae/ /t/, 'se', 'pa' 'rate'. There may be some differences on the way we in UK pronounce the word; however, there is only one /r/ sound in the word and that belongs to the syllable 'rate'. What precedes the /r/ is a schwa. Thus, if the learner is told that when writing 'separate', they need to say 'se' 'pa' 'rate', emphasising the /a/ sound, they will probably never spell it incorrectly again. Actually, I'm living proof in the effectiveness of this approach: many years ago, I spelled 'kangaroo' with the spelling [e] rather than [a]. My Latin master told me to say 'kan' 'ga' 'roo' and said that if I did, I would never again spell it incorrectly. And I never have!

kelli said...

Hi John:

Thank you for taking the time to read my article and comment on it. But, I am perplexed by the title of your blog post. It states that I am mistaken, but I only presented linguistic facts with evidence. The point of my article was that written English is not based on syllables, but morphemes, which is a linguistic fact. The point of the article was to start thinking about teaching how English is truly structured to help students understand all the 'exceptions' that syllables produce.

So, by explaining to the student that is just ever + y is helping them think about language differently. It is helping them understand that English spelling is not a pure sound/symbol relationship and that by teaching that way we are causing more strife. Why not have the student (yes, even the 5-year-old) think about the meaning of the word to help them spell it? This is not to say that phonology is not important, because it is, but meaning is first and only when we know how a base surfaces in a word can we know how to pronounce and/or spell it.

I am only presenting the evidence, not a theory. I am also not selling any programs, so I have no reason to anything other than present English the way it actually is and I think our kids deserve those truths.

As for the word , perhaps I should have stated that the has a schwa in the nucleus of the syllable.

Holly's question about all the words that are related to is a very astute observation, I would love to see your response to it. Look at all the words the student can not only spell, but all the words that they understand...

Please take some time to really think about what I am presenting before telling the world that I am mistaken, because I have more evidence that I am not mistaken than any syllable program can present.

Lastly, if we are measuring success based on what has worked for us, then I would have to say that you are mistaken, because my way has had a profound impact on my students. In fact, just yesterday when I showed a severely dyslexic fifth grader that the word is really not a sight word (sight words don't exist) by showing him that we need the for related words like and he looked at me and said, "mind blown. Can I take this home and put it on my wall." for the first time in his life someone showed him the underlying structure of English.

Dick Schutz said...

The thing is, the Alphabetic Code of written English IS a "pure sound/symbol relationship" but the relationship is comprised of 180ish grapheme/phoneme correspondences, plus conventions relating to abbreviations, contractions, and the pronunciation of many person/place names. So the purity is complex. The written language is a means of representing spoken language, not for conveying meaning directly, and one of the beautiful features of the Code is that it permits wide latitude in the pronunciation of words. That is, English CAN be pronounced as "syllable-stressed" but when it is so pronounced it comes across as "barking at text," "spelling voice," "robotic"--or otherwise unconventional.

Some of the confusion in the exchanges above is that several of the technical linguistic terms have different implications for written and spoken English and for reading and spelling instruction. For example, the pronunciation of any phoneme outside the context of a syllable is inherently distorted. But no matter. Children can be reliably taught how to handle the Code using approximations. Another example, children "speak prose," so matters of intra-word characteristics such as morphemes, stress, and syllables are acquired before complexities of reading and spelling arise. The terms are important considerations in architecting reading and spelling instructional programmes, but children can be taught to "read linguistic text" without any reference to the technical linguistic terms.

The proof of the instruction is in the programme, not in the linguistic terminology. The "proof" can't be sorted out using examples of individual words and individual children--there are just too many individual words and individual children, so that the discourse morphs to tis-taint disagreement.

When it comes to spelling, in today's world teaching children how to intelligently use SpellCheck (it's getting better all the time) and to txt competently makes more sense to me than fretting about anachronistic communication, but that's an opinion, not a "fact."

John said...

Hi Kelli and thank you for your reply.
(Part 1) To clarify: as you will see if you have time to look further into the Literacy Blog, I start from the position that spoken language is primary, that is to say that wherever one goes in the world, everyone learns to speak, to listen and to understand their own spoken language(s) naturally.
What no infant does is to absorb the script of a language naturally: written language must be studied. So, specifically, where, unless I am misinterpreting what you say, I think you are mistaken is in suggesting that the teaching of morphemic understanding comes before teaching pupils how to decode and encode.
Of course, I wouldn’t want to be too mechanistic about this. As children, by the age of around five years, already have a considerable ‘knowledge’ of the grammar of their own L1, pointing out that we add the inflection -ed to many simple past tenses can only be helpful. But then they would be trying to represent the sounds /d/, /t/ or /Id/ already if they have been taught to represent what they hear and we, as teachers, would be teaching them how to spell these inflected endings. However, they don’t need to be taught that the simple past tense of ‘laugh’ is ‘laughed’. They know this and use the form in spoken language naturally. The same is true in the case of other inflected endings.
Enhancing meta-understanding as we teach the complexities of the writing system is what knowledgeable teachers will do. What’s more kids love to learn stuff like this. Teaching inflectional and simple derivational morphology is not difficult largely because many forms are comprised of simple one-to-to-one sound-spellings. These simple forms can therefore be taught alongside teaching children to read and spell from the beginning.
What is more problematical is teaching longer, less frequently encountered, more abstract words, which pupils/students of any age will probably be learning from their reading. The question is: how to teach these words. If the student is reading the word ‘demonstrate’ and they read it as ‘dee’ ‘mon’ strate’, ‘deemonstrate’, the chances are that, as long as the word is in their lexical repertoire and it is being read in context, they will normalise the word to ‘demonstrate’. If they don’t, we need to point to the (first) letter [e] and say that this can be /ee/ in some words but that in this word it is /e/. But, as Dick points out, you can’t teach every word in the English language.
Regardless of whether the word is longer or not, we need to teach pupils to break words into syllables, starting, I would suggest, with simple compound words; otherwise they often try and read all the way through the word dropping out sounds and syllables. [I have a reading test in which children have to read (in context) the word ‘convenient’. They very often don’t even try and, if they do, they read it as ‘convent’.]
The other problem is that many students have no idea where morphemic boundaries begin and end. I taught the word ‘paucibacillary’ to some Year 2 children (ages around seven years) a few years ago. The word came up in the context of collecting money for children suffering from leprosy, and paucibacillary leprosy is a particular kind of leprosy. I wrote the word on the board and put in some syllable boundaries pau | ci | ba |ci | lla | ry and a good three-quarters of the class were able to read it successfully straight away. What we have here is a six-syllable word that is comprised of two free morphemes and a bound morpheme. Thus: free + free + bound: pauci + bacill + ary. I did teach them that ‘paucus’ is Latin for ‘a few’ or ‘few’ and that ‘bacillus’, a Latinate adaptation of ‘baculus’ meaning ‘rod’ or ‘stick’, means ‘germ’. The bound morpheme ‘ary’ is also derived from Latin to mean ‘a thing or person that is connected with or belongs to’. Which gives us a type (‘paucibacillary’ - an adjective meaning a thing connected with few germs) of leprosy...

John said...

Kelly, (Part 2)
... As a teacher, I would always be looking to extend understanding of the pupils/students I teach and, as you can see, in this instance, my approach was talk + spelling + analysis of meaning, with the option, if the pupils already understand what is meant by different types of morphemes, of couching it in more technical language. The thing is that this is a trade-off: how much time do you have, how knowledgeable is the teacher, how well-integrated is the approach in the school, and so on.
On our courses many teachers express excitement and intense interest in the aspects of etymology and philology we get to touch on. The reality is though that since classics (almost) disappeared from the curriculum and MFL was so poorly taught for a couple of decades and at the mention of grammar everyone ran a mile, for many teachers, getting up to speed is a long, hard road.
Best wishes,

John said...

Hi Dick,
Thanks once again for writing in!
I must say that I agree with you about linguistic terminology. We keep such stuff as simple as possible, not even referring to vowels and consonants as vowels and consonants when we begin teaching young children to read and spell.
The crucial thing, as you rightly point out, is the proof of the pudding. I've set this out before but here it is again anyway.
At the end of the first three years of schooling at St Thomas Aquinas school in Bletchley (UK), we tested fifty children in Y2 [average age seven years and three months] who had been taught using Sounds-Write from the beginning of YR - i.e. as soon as they entered school. On the spelling test we used (Dennis Young’s Parallel Spelling Test, 2nd edition, which has a ceiling of eleven years), 44% (22 out of 50) of the pupils hit the ceiling on the test, 64% (32 out of 50) of all pupils had a spelling age above nine years and six months, and 90% (45 out of 50) had a spelling age of eight years or above. Only one pupil had a spelling age below his chronological age and he had some speech and language problems.
Not bad, eh Dick, considering there are always difficulties in getting teachers to teach a programme with fidelity?
Now, when I'm appointed dictator of Britain... ;)

LEX said...

John, you contradict yourself when you espouse here "keep[ing] such stuff [linguistic terminology] as simple as possible, not even referring to vowels and consonants as vowels and consonants when we begin teaching young children to read and spell" but in another post ( you recognize how many young children are interested in the big words of dinosaurs, like, well, dinosaur and archaeopteryx. Little kids love big words. How curious.

John said...

Hi Lex,
I don't believe I am contradicting myself at all here.
Why don't I advocate teaching young (four and five years) children words like 'consonants' and 'vowels'? Because the terminology is irrelevant in the early stages of learning to read. Why, for instance, might it be useful to teach the word 'mat', say, and then tell the children that the /m/ and the /t/, spelled respectively [m] and [t], are consonants and that the /a/, spelled, [a], is a vowel. At age four years or in the very beginning stages of learning to read, this is not only irrelevant, it is also too much information/cognitive overload. Later, it might well be useful to teach this knowledge.
Similarly, when children have learned the one-to-ones and have gone on to learn that sounds can be spelled in more than one way, they can easily progress to polysyllabic words. These would start with simple words, such as 'desktop' and 'Batman', and move on to words like 'dinosaur'.
So not really very curious at all.
And, by the way, I won't be responding to your lack of civility on Twitter. I pointed out that letters don't 'have' or 'make' sounds, not because I'm being excessively literal, but because young children are highly literal and get the wrong idea if we tell them that letters 'make' sounds. I explained my reasoning here:
If you should happen to read this reply to your comment, I would suggest that you might adopt a more polite tone. Disagreement is good and helps us all to reshape our ideas.
Best wishes,