“For I am a bear of very little brain, and long words bother me.”
A. A. Milne

Over the past few months I’ve been working with Leila. I met Leila when she was in Year 3.

I am teaching her class their very first Sounds & Syllables spelling lesson. We are practising counting the number of syllables in a variety of words. It is going well, so as a fun aside, I say, ‘I’ll be really impressed if next week anyone can tell me how many syllables are in supercalifragilisticexpialidocious.’

Five seconds later a hand is up.


‘Fourteen.’ This is Leila.

Feeling slightly ambushed, I frantically do a mental count. “Ummm, yes. Well done.”

Leila’s mother works at the school. I’m regularly there before school starts as is Leila. So the next time I’m there, I catch her and teach her to spell supercalifragilisticexpialidocious using the Sounds & Syllables method. Here are her first three attempts. [fig. 1]

[fig. 1]

The first time through we go slowly. I say each syllable clearly in a spelling voice and she spells it. We work through the whole word in this way. She doesn’t make a single error. The second time through, I say the word normally and off she goes. She makes one spelling error, using <a> instead of <i> for the schwa towards the end of the word. You can see she has circled it to target the tricky spelling. The third attempt is again on her own; she knows her tricky bit so she nails it. There are also attempts four, five, six and so on days and weeks later to ensure it has stuck. It has.

And that led me to thinking, ‘Just how long and difficult a word could I teach Leila to spell?’ We started off with something fairly simple (compared to where we are at the moment, anyway) with floccinaucinihilipilification. [fig. 2] She nailed it in a few attempts.

[fig. 2}

And on we went. Honorificabilitudinitatibus next. [fig. 3] She breezed through it.

[fig. 3]

As she did with hippopotomonstrosesquipedaliophobia. [fig. 4]

[fig. 4]

And this left me with somewhat of a problem. Where do we go from here? There aren’t many words in the English language that can top these. There are a few ridiculously long medical words, but almost nothing else. But there was one long and difficult word that came to mind, the infamous Welsh place name: Llanfairpwllgwyngyllgogerychwyrndrobwllllantysiliogogoggoch.

The problem was I didn’t know how to spell it. Worse than that, I didn’t know how to say it either. And I knew next to nothing about the Welsh language. So I was going to have to learn the word myself before I could teach it to Leila. I did manage to learn how to say and spell it, as did Leila, remarkable as she is. [fig.5]

[fig. 5]

In one sense all of this is nothing more than a little fun. What better way to spend a few minutes than playing around with ridiculously long words. And let’s face it, most teachers are not going to be teaching these words with their classes tomorrow. But because I had to teach myself how to both pronounce and say these words before I could teach them, it sharpened my thinking about how to teach spelling.

And this brings me to the point of this post – what I know about learning to spell by teaching first myself, and then Leila, to spell Llanfairpwllgywngyllgogerychwyrndrobwllllantysiliogogogoch (herafter referred to as Llanfair… for brevity’s sake) and other ridiculously long words.

1. Pronunciation is the key that unlocks spelling

Before I could learn to spell the word, being able to pronounce it correctly was essential. At first I found versions of the word that had been written phonetically. But it wasn’t enough; I couldn’t be sure I was saying the word correctly. I needed it modelling. Not having access to a Welsh speaker, I went to Youtube [fig. 6] and listened to the correct pronunciation repeatedly, after each listen trying to replicate what I’d heard. Good pronunciation is the key that unlocks spelling.

[fig. 6]

This was highlighted while quizzing Leila a few weeks after teaching her hippopotomonstrosequipedaliophobia, I noticed an error had sneaked in: she was spelling dalio as dala. I asked to say the word back to me and she was pronouncing the this part of the word as day|luh instead of day|lee|uh. The cause of the misspelling was the mispronunciation.

Of course, this makes complete sense. In essence, written English is a code for replicating spoken English, and to spell a word correctly we need to move from sound to print. If we are not representing sounds clearly (or even at all) in spoken language, that can impact on our attempt to capture those sounds as spellings. Establishing clear pronunciation is an essential first step in the spelling process.


  1. Listening to carefully to clear pronunciation from an expert (the teacher) is vital. Ensure that teacher pronunciations are clear and crisp in a silent environment with the unswerving attention of every child. If they don’t hear the word correctly, whatever the reason, they are at an immediate disadvantage.
  2. Ensure that all children pronounce the word – no opt outs – both chorally as a whole class, but also selecting individuals to be sure that pronunciation is correct, particularly for those who might struggle
  3. Repetition is important – locking in the pronunciation doesn’t happen after a couple of attempts. Practice. Practice. Practice. Have children say the word a dozen times. And again tomorrow and next week to be sure.

2. Breaking words down into manageable chunks is critical

I didn’t just listen to Llanfair… in one go and then try to repeat it. As a novice Welsh speaker, such a long word was far beyond the limitations of my working memory. There were simply too many parts to hold onto at once.

Breaking the word into syllables was necessary to avoid cognitive overload. I would practice pronouncing the first three syllables Llan|fair|pwll several times to lock them in and then practice the next two gwyn|gwll, again repeating them several times over before combining them into a single chunk Llan|fair|pwll|gwyn|gyll. I’d then repeat this until I had three larger chunks that I could pronounce reasonably well. [fig. 7]

[fig. 7]

Now, with three chunks I could load from long-term memory. I was able to practice pronouncing the whole word.

I took the same approach when teaching Leila. Individual syllables followed by clusters of syllables followed by the whole word.


  1. Never overestimate the limitations of working memory in the novice learner. Their environment or cemetery is my Llanfair…. Just as I need to take time to break the word down into manageable chunks, so will they: en|vi|ron|ment and ce|me|te|ry.
  2. Break words into syllables. Master each syllable, both the teacher and the children pronouncing each crisply and clearly.
  3. Listen carefully and drill down into the pronunciation of individual syllables if necessary.

3. Sweat the small stuff.

There were some sounds I really struggled with, particularly those that are not used in English. The pronunciation of <ll> was particularly challenging. It’s an aspirated /l/ sound. Place the tip of your tongue against the back of your top teeth as if about to say /l/ and then push the air over it. It took a while to first practice the pronunciation of this sound individually and then to integrate into the word I was attempting to pronounce. Another good example was <ch> pronounced as in the Scottish loch. I found it fairly easy to pronounce the sound at the end of the word in goch but really struggled when moving from ych into wyrn.


  1. Ensure that all children can pronounce sounds correctly (wherever possible – impediments and missing teeth allowing). Now, because for many children they are saying and spelling words in their first language, most sounds in most words this shouldn’t be too problematic. But some children might find things a little trickier, particularly those with hearing difficulties (and it is estimated that as many 8 in 10 children will have had glue ear at least once before they are 10 years old, for the most part undiagnosed) and English Language Learners who might find the pronunciation of some sounds challenging if they are not routinely used in their first language.
  2. Because a child can say a sound in isolation, does not necessarily mean they can integrate it into a complex word, so pay attention and respond to how children are pronouncing sounds in context and provide clear modelling and support where necessary.

4. Know the code.

I have stored away in my long term memory a fairly comprehensive schema of the complex English alphabetic code. I know that the sound /p/ can be spelled as <p> (pig) and <pp> (happy), and that /oo/ can be spelled as <o> (to), <oo> (too), <ew> (chew), <oe>, (shoe), <ou> (soup), <ui> (suit), <u> (ruin) and <u-e> (chute).

But it isn’t just sound-spelling matches that I have stored away. I’ve also stored away a lot of statistical knowledge that helps me to select spelling choices. This has been refined over many years has by speaking, reading and writing millions of words. It helps me me to determine the probability that one spelling choice might be selected over another.

For example, when spelling floccinaucinihilipilification, I noticed that /ee/ is typically represented by <i> (flocci, nauci, nihili, pili). This tends to be true for longer multisyllabic words (particularly those that have their origins in Latin). Having taught this to Leila, she was able to then apply this knowledge to honorificabilitudinitatibus (abili, tundini, tati). Leila has now refined her statistical model to become that little bit more sophisticated, and this will help her to spell many other multisyllabic words.

There are four things to focus time and attention on when refining the statistical model and developing their English spelling schema.

Frequency. Some spelling choices are much more common than others. If I hear a new word containing the /ee/ sound, I’m more likely to guess at <e> (me), <ee> (bee) or <ea> (team) than I am <eo> (people). I know the former are fairly common spelling choices for /ee/ and the latter is rare. I’d be foolish to bet on the longshot.

Position. If I hear a new word that begins with the sound /v/, I am likely to plump for the spelling <v>. If the same sound is at the end of the word, I’ll select <ve>. Even though <v> is the more common choice overall, my statistical model tells me that <ve> is a much more likely spelling at the end of a word (love, move, gave, have, live, etc.)

Relationships. If you asked me to spell the pseudoword /w/ /ur/ /ch/, I’d spell it as worch. There are other plausible spellings, but my statistical model tells me that following /w/, /ur/ is often spelled <or> (worth, world, word, worm, worse) and that <ch>, and not <tch>, is the likely choice following a long vowel sound. such as /ur/.

Morphology. When spelling hippopotomonstrosesquipedaliophpbia, I have a lot to draw on. I already know <hippo> (horse) from hippopotamus, <monstro> (unnatural) from monstrosity and monstrous, < ped> (foot) from pedestal and pedestrian, and <phobia> (a fear of), from arachnophobia and others.

Some of this, however, went out of the window when learning to spell Llanfair… My statistical model is lacking. Apologies to any Welsh speakers here, because my statistical model of Welsh spelling is based upon a handful of words (all place names) so is severely lacking, but I had to learn a few important things about the code and how it appears to differ from English.

I had to learn that <ll> doesn’t represent /l/ as it does in English. It represents an aspirated /l/ sound. I had to learn that <ch> here doesn’t represent /ch/ or /k/ as it might in English but /kh/ (loch). I had to learn that <yr> can represent /eer/. And I had to learn that /i/ is more commonly spelled as <y> and less commonly as <i> whereas as in English the opposite is true.

And I have no morphology to draw upon. I now know a little: llan (church), drobwll (whirlpool) and goch (red). I half know other bits. I know, for example that somewhere in pwllgyngyll there’s a meaning of ‘white’ and ‘hazel’ and the somewhere in gogo is the meaning of ‘cave’ but as yet as I still hazy and need to check to be sure.


  1. Understanding and applying the alphabetic code is the beating heart of good spelling instruction. If you do not know that /g/ can be represented by <gu> you will always struggle with words such as guard, guess, and beguile. If you don’t know that /o/ is often spelled <a> after /w/ then spelling watch, what, quality and squalid will be that much harder.
  2. Mastering the code is not a job done by the end of Year 1 or 2. By the end of key stage 1, children should be a long way towards mastery of the code, but there is a long journey ahead and wonderful things still to discover. To borrow a phrase from my mother, any teacher in any year group who does not understand the relationship between letters and sounds is neither use nor ornament to the teaching of spelling.
  3. Morphology matters. Knowing the base <pict> (from the Latin ‘pictus’ meaning to make pictures, paint, or embroider) will help us to spell both picture and depict even though the pronunciation of <t> is different: /ch/ in picture and /t/ in depict.

5. Focus time and attention on tricky within word misspellings.

The average person isn’t going to spell a difficult word correctly the first time. So focusing on those parts that you find tricky is important.

Once I had the pronunciation more or less down (and for Llanfair… that took a few days of practice), and I’d taught myself the sound-spelling matches that were unfamiliar to me) I tried to spell it using my Sounds & Syllables approach one syllable at a time, segmenting each syllable into sounds before moving onto the next. I’d then look at the correct spelling, circle the spellings I’d got wrong. I’d pay attention to these within word misspellings, making links to known words (which is tricky when you’re spelling a word in a language in which you are a novice learner. But I could find help by linking to what I knew: for <eer> spelled /yr/ I recalled the old Anglo-Saxon word wyrd (thanks Terry Pratchett) ; for <ll> I knew Llandudno and Llewellyn. I’d then cover everything up and try again, thinking hard about those tricky spellings. I’d do that until I’d spelled the word correctly a few times.

When teaching Leila, for her first attempt I’d let her spell the word without me showing her the spelling . I’d say the first syllable in a spelling voice (see my Sounds & Syllables approach for more guidance), let her spell it and move onto the next. This allowed me to see how she was approaching each syllable and each sound, providing useful assessment data. If I’d shown her the word before her first attempt, how would I know if she was just remembering from sight?

This allowed me to see what she knew and what she didn’t: the double <c> representing /k/ and /s/ (comparing it to words like success, accept and accident), <au> representing /or/ (so we talked about naughty) and the <ti> spelling of /sh/ (which she could spell but didn’t know how to represent it as a two-letter spelling). It also drew her attention to what she knew and what she didn’t. And now she knew what she didn’t know, she could fix it.

We’d then attempt the word again. Leila being Leila, she’d invariably nail it on the second attempt. For Llanfair… it was the third.


  1. Don’t show the word before children’s first attempt. See what they do, the kind of mistakes they make. It’s a simple but powerful assessment tool.
  2. Do something with you what you find! Focus on common within-word spelling errors.
  3. Cluster, compare and contrast tricky spellings with words they already know. If they misspell the /ai/ in neighbour, cluster and compare with eight and weight and contrast with straight.
  4. Spell words several times. Make sure that everyone walks out of the lesson having spelled it correctly at least once.

6. Practice. Revisit. Practice. Revisit. Practice. Revisit.

When learning to spell Llanfair… I’d come back to the word after several hours, perhaps for a day or two and try again. Unsurprisingly, mistakes had crept in. Although I’d spelled the word correctly in my previous spelling session, I hadn’t learned the spelling. I’d just performed well. I spaced out my attempts in this way until I could spell the word first time every time.

Leila knows that before I teach her a new word, she has to demonstrate that she can still spell the previous words. She is quizzed regularly on what she has previously been taught. Learning the new word is the reward for knowing the previous ones. And like me, mistakes occasionally creep in with previously learned words. These mistakes are addressed before moving onto new learning. It’s a constant cycle of revisit – practice – revisit – practice – revisit – practice.

This has implications for weekly spelling tests. Personally, I’m not against them. I think tests are a powerful tool not for proving learning but for improving learning: just not in the way they commonly used in most schools, which has two flaws. The first, and worst, is that the test often replaces rather than supports the teaching of spelling. If you do that, my advice – stop. If you only have time for either teaching or testing spelling, teach it. The second flaw is testing words only once. Children spell them correctly, promptly forget them and spell the same words incorrectly in their writing a few weeks later. Rinse and repeat. The one thing that is almost certain after something is learned, if there is no further attempt to retain the information, is that it will be lost over time. This is known as the Ebbinghaus Forgetting Curve [fig, 8], named after Professor Hermann Ebbinghaus who ran a self-study (1885), trying to memorise nonsense syllables (wid, zof, etc.), testing himself over time to see how much information he had retained and recording the results. The curve shows that people tend to forget much of what they have learned quite quickly.

[fig. 9]

So, if you learn once and test once it should be no surprise to find that children will forget much of what they have learned. What we ought to do is to periodically test the same content multiple times [fig. 10], allowing for some forgetting between each test (this makes the recall more effortful and, therefore, more likely to stick). After each test information is likely to be forgotten more slowly, so the gaps between each test can lengthen until we are sure learning is secure.

[fig. 10]

So the problem with spelling tests isn’t the testing, it’s mistaking the result of a single test as learning. If you use spelling tests to support the teaching of spelling, my advice – draw half of your words on your current spelling test from pervious ones, including words that children spelled correctly, constantly revisiting them to lock them in for the long term.


  1. Just because children are spelling words correctly in the spelling lesson and/or in the weekly test, do not mistake that for learning. Constantly revisit to lock them in.
  2. The best way to revisit might be through low-stakes quizzing. Make this a part of your regular practice. Schools I work with often run paired quizzing sessions during registration time.

7. The spelling voice should be a temporary scaffold.

When learning Llanfair… I initially couldn’t make head nor tail of the pronunciation when spoken naturally, The word was too complex, the sounds and syllables flew by before my brain could fully register them. Thankfully the speaker [fig. 6] slowed the word down, exaggerating sounds and syllables – in essence, using a spelling voice, making the word easier for me to follow. And initially it was easier for me to think of fair pronounced /f air/ rather than /v ire/. The latter was too far removed from the spelling; thinking of the word in this way provided the early support I needed.

But after a while I found I no longer needed the spelling voice. I could hear and say the word naturally (or as naturally as I am able) and still spell the word correctly. The spelling voice was a temporary scaffold.

And it ought be that way. Pronouncing the <ai> in mountain as /ay/ provides children with some guidance as to the correct spelling, but this should only ever be a temporary support. If children can only spell mountain when I say it in a spelling voice, we have a problem.

I’d do this for Leila too. When spelling hippopotomonstrosequipedaliophobia, I’d initially pronounce the <o> at the ends of hippo, poto and dalio as /oa/, and the <e> in ped as /e/ when in a natural speaking voice they are typically pronounced as schwas. Once she was comfortably spelling the word correctly with the support of the spelling voice, we’d take it away. I’d pronounce the word normally from then on.

There were occasions, however, where I would revert back to the spelling voice. Leila’s bugbear in honorificabilitudinitatibus was spelling the final <i> as <a>. So, I’d temporarily revert back to a spelling voice, pronouncing what is typically a schwa as /i/.


  1. A spelling voice is a powerful scaffold in the early stages of learning to spell a word, but children do need to be able to spell words without it.
  2. Once children are spelling a word correctly with the support of a spelling voice, remove it.
  3. If they start to misspell the word again, temporarily reintroduce the spelling voice.


With the right instruction, almost all children can be good spellers. The principles I outline here can be applied to simpler words and to spellers of varying ages and abilities. And they can spell words you perhaps think they can’t.

At Leila’s school most of the year 6 children can now spell hippopotomonstrosesquipedaliophobia. And in Leila’s Year 4 class, the words she is learning have spread around the classroom. Having success with these long words creates a buzz around spelling, and this success feeds into attitudes about spelling; if they can spell these words, they can spell anything.

You might not be spelling these kinds of words in your classrooms tomorrow, but the principles I’ve outlined above can certainly help your children to be stronger spellers.

And more children in your class could spell honorificabilitudinitatibus then you think.

To subscribe to this blog please fill out the form below.

About the Author

Leave a Reply