For you, as a skilled speller, writing the words BOAT and NOTE is trivial. You don’t give the task a second thought: BOAT is spelled B-O-A-T and NOTE is spelled N-O-T-E. You don’t have to identify the sounds (phonemes) in these words, and then match those sounds with letters before you write them. Such a process is called segmentation, but you don’t need it.
You don’t even give any thought to the fact that these two rhyming words are spelled differently, not just in their initial letter of course, but in their “rime.” Both spellings – OAT and OTE – are perfectly phonetic, and many other rhyming words (COAT, ROTE, FLOAT, TOTE, GOAT, VOTE, MOAT, QUOTE) share one or the other of these two spellings.
How do you keep all these OAT/OTE spellings straight? And how do you spell the rhyming words GOAL, HOLE, BOWL, TOLL, and SOUL correctly, barely thinking about those disparate spellings? More impressive still, how do you correctly spell COLONEL, LICORICE, TONGUE, and BOUQUET, even though these words are spelled quite oddly given their pronunciations? The answer is that all these words have been orthographically mapped by you at some point in the past. Orthographic mapping (OM) doesn’t simply make reading proficient and accurate; it does the same for spelling. In fact, the word “orthographic” refers to writing words with a correct spelling.
OM was first explicitly described by reading researcher Linnea Ehri [see here]. It’s a process that happens incrementally and automatically when a new word is encountered *if* the reader has been taught the skills to make the necessary connections. The connections that must be made are between the letters (graphemes) seen in the spelling and the individual sounds (phonemes) detected in the word’s pronunciation.
A word that has been orthographically mapped becomes a sight word whose exact spelling is directly linked to the sound and meaning of that word already stored in the brain. Once this happens, any one part of this spelling-sound-meaning triad instantaneously activates the other two.
So a skilled reader, seeing the printed word BOAT, doesn’t need to decode it (sound it out). She simply recognizes the word as a sight word. Its meaning and pronunciation are immediately available to her.
Similarly, a skilled speller, thinking the word, BOAT, or hearing someone else say it, doesn’t need to segment it. Since it’s a mapped sight word, its correct spelling is immediately available.
[If you need quick definitions of phoneme, grapheme, decode, blend, encode, and segment, they can be found at the start of this blog.]
For the beginner, however, someone for whom BOAT is not yet an orthographically mapped sight word, things are more complicated. For him, reading BOAT requires that it be decoded. A phoneme must be attached to each of the 3 graphemes (B, OA, T) and those phonemes must then be blended together to pronounce and identify the word.
And spelling BOAT is more complicated still. He must be able to hear the 3 phonemes in the spoken word, even though those phonemes blend into one another.
Assuming he can do that, and assuming he knows the appropriate letter-sound correspondences, he must still often choose between two (or even more) phonetically plausible spellings: BOTE and BOAT.
Until a word is orthographically mapped as a sight word, spelling it correctly is a fraught enterprise for any student. It requires significant letter-sound knowledge and full segmentation skill. And even then, he may still get it wrong: BOTE.
Given the above, what’s the best way to teach a child to spell? It’s not, and it can’t ever be, to have the child rote-memorize “sight” words. And by “rote” I mean apart from the process of orthographic mapping. Such deliberate memorization is difficult, non-rational grunt work for children, most of whom will never manage to store more than a couple hundred such words in visual memory. (Note: By comparison, most skilled readers and spellers have 50,000 or more sight words stored in their brains by the time they head for college.)
Let’s work backwards. Reading and spelling a word are relatively easy if that word has previously been orthographically mapped. The ability to do such mapping has 3 main prerequisites according to Linnea Ehri. These 3 prerequisites are:
1) knowledge of letter-sound correspondences
2) blending or segmenting ability
3) practice (repeated exposure)
The first one – also called grapheme-phoneme correspondences – refers to knowledge of the Alphabetic Code. The English code is more complex than that of most other alphabetic languages. Teaching it takes time, and it must be done well, starting from the simplest correspondences, and moving, gradually, to the more complex.
I used the word OR in the second prerequisite. That’s because OM can occur either way. The more important route, by far, is via the blending that happens when a child decodes an unknown word for herself. This is fully in line with David Share’s Self-Teaching Hypothesis and with Linnea Ehri’s research into orthographic mapping.
OM can occur via segmentation as well. For this to happen, however, someone other than the child must first pronounce the word, or the child must correctly guess the word’s identity. Assuming that’s done (and assuming the child has full segmentation ability and appropriate letter-sound knowledge), the word can be mapped to memory via this route as well.
Full segmentation skill also gives the child a reasonable chance of correctly spelling a word which has not yet been orthographically mapped. While this is the case, either spelling, B-O-T-E or B-O-A-T, would demonstrate considerable skill. Spelling a word (and getting corrective feedback in the case of BOTE) also serves to reinforce the letter-sound connections needed for orthographic mapping if the word has already been decoded a few times.
The third prerequisite (above) is the common-sense requirement that rehearsal is needed to make a newfound skill “stick.” What may be surprising, however, is that only 2-5 successful decoding episodes for a new word are required (for a typically developing child) in order to add that word to the child’s ever-growing collection of sight words in memory.
Continuing to work backwards, what’s the quickest, easiest, and most efficient way to teach a child letter-sound correspondences and blending/segmenting skills? It’s to do so directly, from the very start of reading instruction. This will allow OM to commence as soon as possible. The wrong decision here, at the beginning of reading tuition, will serve only to waste valuable time and to make both reading and spelling more difficult for children.
So what’s the wrong decision? It’s to start instruction with any type of top-down orientation, that is, with an initial emphasis on rote-memorized “sight” words and various word-guessing strategies, rather than on letter-sound relationships and blending/segmenting. It’s to stress reading comprehension and meaning at the start of instruction, rather than decoding and automatic sight word creation.
Part of the reason for this can be understood from the perspective of the Simple View of Reading. (A discussion of the Simple View can be found here for those interested.) Just as important, however, is this: any instructional method that starts with spoken whole words makes phoneme identification harder than it need be. That’s because, in spoken words, phonemes are coarticulated, that is, they blend seamlessly into one another, making their individual identification a difficult task for any beginner.
A better decision is to start reading instruction using a bottom-up, synthetic phonics approach. Here, phoneme identification is easy because children are taught, directly, what 5 or 6 of them are. They practice saying these single phonemes apart from whole words. Once they’re pronouncing the phonemes correctly, the teacher can associate a letter with each one, and begin teaching the skill of blending those phonemes into whole words.
Here’s a brief, sketched example of what I mean for getting started:
Example: Teach your children 6 phonemes /m/, /n/, /s/, /a/, /o/, /e/ and the letters that represent these sounds: M, N, S, A, O, and E. (Note: the 3 vowel sounds are the “short” sounds you can hear at the beginning of the words APPLE, OX, and ED.) Now show your children how to blend these sounds into 30 words: MAN, MOM, MEN, SAM, NAN, MESS, SASS, MASS, ASS, AM, AN, SEN, SAN, MAM, SOM, NES, NAS, MON, EM, EN, ES, NOS, SEM, MEM, NAM, NON, NOM, SESS, NEM, and SON (as in SONIC).
Maybe you’re thinking: “Hold on Parker, those last 19 aren’t even words.”
My response:
True enough. But not too far down the line your children will be faced with reading words like SENt, SANd, MAMmal, SOMber, NESt, NASty, MONsoon, EMber, ENvy, ESsay, NOStril, SEMinar, MEMber, dyNAMic, NONstop, NOMinate, obSESS, NEMesis, and SONic. So, it’s hardly a bad thing to have them learn to “read” these 19 word parts now.
The primary goal, here in the first months of kindergarten, is learning some phonemes, the letters that represent them, and the skill of blending. That some of the blends are only word parts is okay.
As soon as another 3-4 letter-sound correspondences are added to the above six, there will be more than enough actual words to go around.
Maybe you’re wondering how to get started teaching the skill of blending. Well that’s easy. I chose these 6 sounds to start because they’re sustainable, that is, they can be made as long as the child’s breath holds out.
So write a word (or word part), spreading out its letters, like this:
M A N
Tell a child he or she is to make the sound of each letter for as long as you’re pointing to it.
Now point to each letter in turn, for about 2 seconds each, with no break in between: MMMMMMMMAAAAAAAANNNNNNNN.
Do it again for one second each: MMMMAAAANNNN.
Now for a half-second: MMAANN.
Finally, sweep your finger across all the letters quickly: MAN.
Repeat with other children. Done. Your children have learned to read their first word.
Repeat this procedure for the words MOM, MEN, SAM, and NAN.
Write these 5 words out, normally spaced (MAN, MOM, MEN, SAM, NAN), and have various children read them as you point to the words in random order. Continue with this procedure until all 30 “words” can be read. Note: double SS words can be introduced this way:
M E SS
Once the above is accomplished, in the first few months of instruction, your children will be reading – not memorizing sight words, not guessing – but reading. They’ll already be starting to realize how the whole enterprise works: letters in words stand for sounds, and those sounds can be blended to figure out what the word actually says. This is called the alphabetic principle. Orthographic mapping will have already commenced. Your children will be starting to conclude: “Hey, I can do this!” and “This is fun!”
If, as soon as possible, you reverse the above, and ask your children to spell a word (or word part) they have just recently built by blending, your children will likely be able to do so. Spelling is harder than reading. It involves segmentation: hearing a word pronounced, detecting all of its phonemes, and then attaching the correct letter to each phoneme heard. This task becomes much easier if the child just recently built the word, phoneme by phoneme, by blending.
[See my FREE books for how to get your children from this point to full, independent reading.]
Now contrast this example with how Phonemic Awareness (PA) must be approached using any top-down method of instruction. Since top-down methods start with whole words, children must become aware of phonemes by hearing them, not individually spoken, but mixed in with other phonemes. In full words, phonemes are in their coarticulated form, that is, they bleed into one another – a fact that makes hearing individual phonemes difficult for beginners.
Consider that for a moment. A bottom-up approach, like Synthetic Phonics, makes PA as easy as possible by teaching isolated phonemes directly and then blending those phonemes into whole words. All top-down approaches make PA as hard as possible by requiring beginners to hear and distinguish phonemes in their coarticulated, whole-word form.
Starting reading instruction, as in the above example, not only allows orthographic mapping to commence as soon as possible, but it also shortens, or even eliminates, Ehri’s partial alphabetic phase. This is due to the fact that children are being taught through-the-word decoding from the beginning of instruction, giving them rapid access to Ehri’s full alphabetic phase, where fully connected sight words are permanently and accurately mapped to memory.
An orthographically mapped word is one that has become a sight word for an individual. It’s a word that can easily be read (without sounding it out) and spelled (without the need to consciously memorize the spelling or to segment the word). The point here, is that to make spelling as easy as possible for students, teach them, properly, how to read. “Properly,” in this context, means: teach them to decode right off the bat, so that orthographic mapping starts early and proceeds continuously, rather than starting in a year or two, or not starting at all.
How do we know AFRICUH, MOTER, DOLLER, MOUNTIN, and COTTIN are misspelled even though, phonetically, these words are spelled in a perfectly reasonable manner? Without thought, skilled readers spell and write these words correctly: AFRICA, MOTOR, DOLLAR, MOUNTAIN, and COTTON. Orthographic mapping is the only way words containing these notorious “schwa” sounds can be spelled accurately. [For more on schwa sounds, which commonly occur in the unstressed syllables of multi-syllable words, see here.]
What about those irregularly-spelled, yet common one-syllable words in English – words that children need to read and spell in the early stages of instruction? I’m referring to words like DONE, HAVE, LOVE, COME, and SAID. Can such words be orthographically mapped?
Researchers David Share and Linnea Ehri both say yes. As Share points out, most so-called “irregular” words are irregular only in the vowel; the consonants are perfectly fine. That’s true of all 5 words just listed. Ehri suggests having children pronounce such words exactly as they’re spelled, thus drawing the needed extra attention necessary to map these particular words to memory.
Accordingly, DONE would be pronounced as rhyming with BONE, HAVE as rhyming with CAVE, LOVE as rhyming with COVE, COME as rhyming with HOME, and SAID as rhyming with PAID. Reading simple sentences, with these common words mispronounced in this manner, quickly demonstrates to students that the expected pronunciation simply doesn’t work. Teachers can also discuss with their children the way these irregular words would be spelled in a perfectly phonetic world: DUN, HAV, LUV, CUM, and SED.
That still leaves a handful of common words whose spellings are so at odds with their pronunciations, there’s no recourse but rote-memorization. Happily, there are only a few such words relevant for beginners – words such as OF (UV), ONE (WUN), ONCE (WUNS), TWO (TEW), and EYE (I).
Reading and spelling an orthographically mapped word are fully automated processes for a given child because the exact spelling of that word has now become part of what Ehri calls an “amalgamated unit” in that child’s brain. These units start forming in the child’s earliest years, as soon as speech gets underway. But they initially contain only the pronunciation and meaning of each word in her speaking or listening vocabulary. With orthographic mapping, exact spellings become a part of these already-existing amalgamations, creating the triads I mentioned earlier.
But prior to orthographic mapping ability, reading and spelling are far from being automated. They’re conscious, and often painstaking processes – and they’re the reverse of one other. The former involves decoding and blending; the latter, encoding and segmenting. So it should come as no surprise that children are poor spellers because they’re poor readers. They have limited ability to identify unknown words on their own via decoding. This means they have limited ability (and opportunity) to do the conscious analysis that makes orthographic mapping possible.
To identify an unknown word, only two paths exist for children who can’t decode – and both paths are poor alternatives. They can ask the teacher what the word is, or they can guess its identity based on pictures, context, or first letter.
A wrong guess – increasingly likely as texts become more complex – is a dead end. But let’s assume he guesses correctly, or that the teacher supplies the word’s identity. Will orthographic mapping occur now? Not likely. Having identified the word in this manner, the child will want to quickly move on, without doing the connection-making analysis that makes orthographic mapping possible.
Even the rare child who stops at this point in an attempt to align the letters she sees, with the phonemes she now hears, would need full segmentation ability, that is, the ability to hear all the spoken word’s phonemes. How likely is it that a child who can’t decode, nonetheless has full segmentation ability, and will use that ability at this juncture to consciously make the necessary connections?
The central importance of decoding lies in this: it forces a child to do the analysis that makes orthographic mapping possible. She must consciously assign a phoneme to each grapheme (letter or letter combination) in the unknown word and then blend those sounds into a full pronunciation. Thus, she sees and hears exactly how all the graphemes and phonemes line up. Decoding the word a few more times maps the word to her permanent memory as a sight word.
The problem with segmentation skill is that, even if it’s present, it doesn’t require a child (who has guessed the word’s identity) to do the necessary analysis to make it a sight word. And most children who can’t decode don’t have full segmentation skill anyway. Such children have only limited ability to map words to memory.
Above, I said children are poor spellers because they’re poor readers. We can be even more specific: children are poor spellers because they can’t decode. The lack of decoding ability makes orthographic mapping difficult-to-impossible. And with relatively few words mapped to permanent memory, spelling will forever be a problem. Is it B-O-T-E, or B-O-A-T, or B-O-T? Spelling every unmapped word is a struggle for such children because their only recourse is to try to hear the phonemes in the spoken word, and then match the correct grapheme to each phoneme heard. (It’s this, or it’s rote-memorization.) Under such conditions, we should not be surprised when children get frustrated and give up.
For the child who can’t decode unknown words on her own, improving spelling will be a frustrating, difficult, and time-consuming task for everyone involved. I’m reminded of a quote attributed to Desmond Tutu:
“There comes a point where we need to stop just pulling people out of the river. We need to go upstream and find out why they’re falling in.”
Children are drowning in a river of misspelled words because they’ve not been taught how to properly decode. It’s far better to teach reading in a competent manner, and then watch, in awe, as orthographic mapping kicks into place – a mapping that makes correct spelling nearly inevitable.
The bottom line is this. Fluent word recognition and skilled spelling both depend on orthographic mapping. Orthographic mapping, in turn, depends primarily on through-the-word decoding. And a child's decoding ability, for better or worse, depends on choices made by adults – adults who mean well, but who are far too often influenced by fad and ideology. “Constructivism,” “discovery learning,” and “three-cueing,” imposed on six-year-olds, are three examples. So is Whole Language, the root and foundation of “Balanced” Literacy.
Emphasize decoding from the first day of instruction, practice it with decodable text, and practice spelling of already-decoded words. If you do this, your children will become skilled readers and skilled spellers, in tandem, in the least possible amount of time.
Stephen Parker
Boston, January 2022
To download any of my 300-page how-to-teach-reading guides, click here and decide what best fits your needs.
Thanks to my mentor, Dr. Pamela Snow, for her many helpful suggestions.