Should We Teach Letter Names?

  • Beginning Reading
  • 28 July, 2018

Teacher question: 

Should we teach letter names or letter sounds to beginning readers?

Shanahan responds:

Twice recently teachers have asked this question. In both instances they said they’d been told teaching letter names confused children and that “best practice” was to focus on the sounds rather than the letter names.

As a former first-grade teacher, I vividly remember the kids who when confronted with a word like what would start sounding /d/ (duh). At first I was puzzled, but it quickly caught on that these young’uns were trying to find the sound in the letter name, double-you, and were settling for the first sound in that name.

Obviously, the pronunciation of W was getting in the way (no wonder some teachers tell young children that it is really a “Wubble-you”). W is different than b, d, j, k, p, t, v, and z, in this regard. In each of those cases, the pronunciation of the letter provides a valuable cue as to the most common phoneme represented by that letter. There are also several other letters whose names at least get you close to the right sound (f, l, m, n, s), and still others whose names cue a useful (if not most frequent) phoneme… a, e, i, o, u, c, g.

Beginning reading instruction has included letter name instruction for time immemorial. The very first schoolbooks brought to America from England (The Protestant Tutor) started with the alphabet, as did the first reading books produced here (New England Primer).

Correlational evidence has long supported the practice: beginning readers’ knowledge of the ABCs is a strong predictor of later reading success. The National Early Literacy Panel (NELP) meta-analyzed 52 such studies that had connected ABC knowledge with the later decoding ability of 7,570 kids and found a strong relationship. The more letter names the kids knew, the greater their later success in decoding.

There has been a lot of discussion of all of this over the years (Gibson & Levin, 1975), but with little experimental evidence to go on. The earliest such studies focused on the effects of teaching artificial alphabets, with the conclusion that kids could read words made up of pretend letters even if they didn’t learn their names (e.g., Samuels, 1972). Interesting, but not especially persuasive when it comes to real reading.

Since then there have been several studies that have tried similar things with real letters, typically teaching or not teaching a few to see what happens. The outcomes here have been more mixed, but as with the artificial alphabet studies, the results haven’t been especially convincing because the kids already knew lots of letters which confounds things a bit. No wonder Marilyn Adams (1990) concluded there was little evidence supporting the benefits of teaching the alphabet.

More recent studies have tended to examine the value of the alphabet within the context of phonemic sensitivity training than on its own. The conclusion from these studies? Training in PA and the alphabet together generally has a much higher impact on later reading achievement than PA teaching alone (NELP, 2008). In other words, for some reason, the inclusion of letters in a PA curriculum has a multiplier impact on its outcome.

Jean Foulin (2005) produced one of the most complete considerations of the problem. He reviewed studies that examined the alphabet’s facilitative effects in learning to read (e.g., Roberts, 2003) both to determine whether such instruction made sense and why letter name knowledge might help. His conclusions: (1) beginning reading instruction should include a serious effort to teach letter names and letter recognition and the sounds associated with letters; and (2) we need a lot more research because it isn’t entirely clear why alphabet knowledge exerts the positive effects that have been found for it.

Given that such an erudite and comprehensive analysis failed to determine why alphabet knowledge matters, let me add my opinion to the mix. Here we go:

Letters are concepts. Concepts are abstract ideas that we use to categorize experience. The letter B is not a single thing… it’s a collection of objects that we learn to treat as equivalent. Look at these various renditions of the letter B. They are all b’s. Some of them are upper case and some lower. They are written in different fonts and some are different sizes. Some are printed and some are script… but they are all b’s, and good readers come to treat them all as equivalent.

What we call phonemes (the smallest meaning-varying units of sound in a language) are concepts, too. Each phoneme is a collection of phones that exist along the spectrographic continuum. What you think of as the /p/ sound or the p-sound is actually quite variant depending upon the pronunciation context within which it is produced and heard. Thus the /p/ that you hear at the beginning of the word pin is actually quite distinct from the /p/ sound in the word spin. And, there are obviously pronunciation differences due to pitch and tone (such as the differences between men’s and women’s voices) and there is dialect variation as well. Nevertheless, learning the phonemes means learning to group speech sounds into the categories that we use in English to distinguish meaning.

Letter names are just labels for these visual and auditory categories, and we’ve long known that providing labels for concepts facilitates learning (e.g., Lupyan, Rakison, & McClelland, 2007; Nelson, O’Neil, & Asher, 2008).

Concepts are abstract and providing them with names appears to help children to think of them as real concrete entities. When provided with the names of concepts children were more likely to seek out information about the objects and their functions.

The best evidence seems to support the teaching letter names early on (Ehri, 1983; Foulin, 2005). I think there is good reason to do so.

But if my explanation holds water, then it would be wise to teach letters more conceptually than we often do—getting kids to think more about the variation in the letters than is common. It also suggests why it would make sense to teach the sounds for these letters simultaneously (Piasta, Purpura, and Wagner, 2010), and why teaching kids to write the letters matters, too (Gentry, 2006).

Building letter concepts means teaching kids to group collections of visual and auditory objects together into sets—overlapping sets given the complexity of our spelling system. Instruction should help kids to develop these letter concepts rather than having them memorize simple lists.


See what others have to say about this topic.

Kathy Nall
Jul 28, 2018 10:05 PM

I've been an instructional coach in elementary for more than half of my 23 years in education. I began my career with kindies and firsties. I agree with your thoughts on this and appreciate that you are tackling a primary area! Few do. Perhaps I am too simplistic when I tell mu kindie teachers; Yes DIBELS Next does not count LN, technically and we should teachbthe letter name right along with the sing. I had never thought about why, and your discussion of the abstractnes of letter makes so much sense! May I share this with Colorado Springs School D11? I would share it with teaching learning coaches of elementary and the primary teachers in my building.
Sincerley and
Kathy Nall

kerry hempenstall
Jul 29, 2018 12:38 AM

Thanks Tim.

I looked at some research on this question too.

See at

Tim Rasinski
Jul 29, 2018 12:56 AM

Tim. Thanks. I like to think of letter names as part of the language of early literacy instruction. Without letter name knowledge it becomes more difficult to communicate with children about these important concepts.

Stephen Parker
Jul 29, 2018 01:03 AM

In my view, this is a non-issue. I teach reading using Synthetic Phonics. Anyone who does so knows one should teach spelling along with reading. As soon as a child can decode simple words like MOM, DAD, and CAT, they should be asked to spell those words. The child's answer necessarily involves letter names. Thus, letter names and letter sounds should always be taught simultaneously.

Also, the sound of P - symbolically /p/ - is identical in the words PIN and SPIN.

I would even say that the sound of P in PIN, PAN, PEN, and PUN is the same - but a child needs to be taught to decode such words carefully, by taking the vowel following the P into account when decoding.

Tim Shanahan
Jul 29, 2018 01:41 AM


You are wrong about those sounds being identical. They are different spectograohically and English speakers learn to treat them as equivocal. It’s fine that you teach synthetic phonics but you need to know more about the sound system and how kids read.



Stephen Parker
Jul 29, 2018 04:03 AM


Children don't come with spectrographs. Not a child on the planet can hear a difference between the P sound in the words PIN and SPIN.

But you're right: I do "need to know more about the sound system and how kids read." That's why I follow your blog!


Morag Stuart
Jul 29, 2018 06:53 AM

Stephen, the sound of P is not identical in pin and spin. In pin it is unvoiced (its usual articulation); in spin it is voiced and closer to /b/ than /p/. Similarly /t/ in stop is voiced and closer to /d/ than /t/; and /k/ in skip is voiced and closer to /g/ than /k/. See Charles Read’s seminal work “children’s categorisation of speech sounds in English” published in 1975, and much subsequent work by Rebecca Treiman.

Jessica Austin-Burdett
Jul 29, 2018 08:09 AM

The function of drawing the shape of the letter whilst sounding it out and understanding its concept appears to be what is refered to as dual-coding which has been proven to be a very effective learning technique. Great article, as a teacher and a parent I think what you’ve raised here is very pertinent and should form the basis of all literacy teaching.

Tim Shanahan
Jul 29, 2018 01:30 PM

Well answered Morag... and indeed children can perceive the difference, but they have to learn to ignore it.



Jul 29, 2018 03:39 PM

I am afraid I see NO benefit in teaching letter names at all, simultaneously with their sounds or at a later stage. It so happens I posed a question on twitter earlier today asking why we continue to teach letter names.
All English speaking countries have a high incidence of Dyslexia. Could one of the reasons be that just as some of our children, who have some level of phonological deficit, are beginning to make sense of letter sounds we introduce letter names to them. What other country teaches two sets of sounds for their alphabet?
The only letter names we need to know are for the vowels yet we could simply just teach two different sounds for them in the same way as we do for the soft c and g. By teaching the names we are building in an extra step that isn’t needed; it simply adds confusion.
If we encouraged everyone to use and accept the spelling of CAT as cat, using the letter sounds, it would in my opinion help everyone.
I had the privilege to teach a very bright 15 year old dyslexic student using @LexonikST One of the tasks involves listening to and manipulating the sounds heard in polysyllabic words. The group reads the word and then they have to spell one of the syllables. For example, spell the second syllable in the word ‘condensation’ This particular student took far longer than normal to complete this task. At the end of the teaching session I spoke with her, wanting to know why it was so difficult for her to spell ‘den’ .
She explained what she had to do silently in her head. She had to hear the ‘d’ sound then convert it to ‘D’ then listen again to the ‘e’ and convert it to ‘E’ and repeat the process again to be able to say ‘N’.
She then went on at length to explain her difficulties and embarrassment, being laughed at in primary when spelling words using the sounds and that embarrassment still lives with her.
My point is that spelling ‘con den sa tion ’ using sounds is simple. Is sounds like the word but spelling it as CONDENSATION using names sounds nothing like it.
When supporting EAL or ESL students we concentrate our teaching on the sounds of letters, regardless of their age; we chant the alphabet using sounds only and it is set to the tune of ‘Old Macdonald Had a Farm.’ Students quickly learn the letter sounds and learn how to decode words because we focus on their needs and knowing letter names isn’t , in my opinion, one of their needs.

Jul 29, 2018 03:45 PM

Hi Tim,
Wow! I appreciate this post very much. This is a topic that has come up a lot for me as a K-1 teacher. Thanks for the information!

Harriett Janetos
Jul 29, 2018 06:18 PM

Given my many previous posts on the subject it won't surprise you that I totally agree with Katy. Perhaps those of us who primarily work with struggling students see how confused these students are. I'm reminded of the Woody Allen joke: Doctor, my wife thinks she's a chicken. I would do something about it but I really need the eggs. Similarly, my principal might say: Dr. Shanahan, my reading specialist thinks she can teach reading and writing without using letter names. I would do something about it but I really need the results.

Here's what I'm pondering: I'm wondering whether, like math, there is in reading instruction a distinction between conceptual fluency and procedural fluency, and might letter names be best suited for establishing the former and sounds for the latter? I'm going to reread your blog and see if there might be a parallel.

Thanks for a topic that is always front and center for me since I spent an entire year teaching kindergarten without referring to letter names.

Stephen Parker
Jul 29, 2018 08:37 PM

Hello Morag.

I'm not arguing that allophones don't exist - and I understand the difference between aspirated and unaspirated sounds. What I am arguing is that 5, 6, and 7-year olds don't hear that difference and that it would be a silly and unnecessary distraction to the teaching of reading if we were to bring it up with kids. We're not teaching them to be professional linguists!

I don't teach the reading of a word like SPIN, initially, due to the beginning consonant blend, SP. But once a child is competent at reading simple CVC words, I'll place SPIN on an index card with my finger covering the S. When the child reads PIN correctly, I take my finger off the S and ask the child to make an S sound followed closely by the word PIN. Kids catch on immediately. No child has ever said to me: "Hey wait a minute Parker! Now the P sounds different! What're tryin' to pull here?"

My emphasis is on practical teaching - not the fine nuances of linguistics and academicians. I invite you (and you too Tim!) to read my free phonics book for teachers - downloadable from It's the way to teach reading.

Stephen Parker

Tim Shanahan
Jul 30, 2018 03:43 AM


The problem is that children can perceive those differences. Pretending that they can’t or not knowing that they can’t will only lead to bad instructional practice. You might benefit from a class in psycho linguistics; they’d take you through children’s language development with regard to these issues. I hope your book reviews the research showing that kids learn phonics more effectively when the instruction includes letter names.

Good luck.

Tim Shanahan
Jul 30, 2018 03:46 AM

Katy and Harriet-
You can’t observe or logic out whether a practice works best...that’s why we do research. Research on children shows that you are wrong—I was just explaining why it works that way. I’ll always go with the research over the opinions. Sorry.


Harriett Janetos
Jul 30, 2018 03:52 AM

Here's the conclusion from the research cited in the link given by Kerry. Note the comments about at-risk students.

"So, what’s the conclusion?

Since each has been shown to be important, it would be wise to include the teaching of both rather than hoping that teaching, say, letter names only will necessarily result in transfer to letter sounds. It appears from research that there is at least some transfer to the other, whichever is taught first. The USA and England differ in which approach is first prioritised, yet the longer term outcomes are similar. Hence, one might infer that it doesn’t matter in the long run which is first taught. Of course it’s possible that the outcomes of the decision may vary given the wide range of student readiness. For some students the decision may not matter, while for others, such as at-risk students, it may be quite important. Additionally, the manner in which the correspondences are taught, their order of presentation, and the way in which they are integrated with phonemic awareness (blending and segmenting, in particular) are also significant influences on reading success. In either letter-sound first or letter-name first approaches, unnecessary confusion is sure to occur if too much information is introduced at the same time, such as teaching simultaneously both upper and lower case letters, and both letter names and sounds. This is more likely with at-risk young readers.

The research so far suggests the need for additional research before the what’s first question can be answered with confidence, if indeed it is crucial. Perhaps the more critical element will be shown to be how well the program is designed and implemented."

shani gill
Jul 30, 2018 09:56 AM

Letters ARE NOT sounds until they are in a word- they are simply letters, the raw material, that children need to know the name of. When we see letters appear in words they can be called graphemes (graph, digraph or trigraph) or maybe letters are in words as etymological markers only (such as the in the word friend) and have NO sound value. Graphemes represent sounds NOT letters. I have worked with struggling, beginning/early readers for over 20 years. I celebrate when my students can articulate the sounds in a word such as by saying '.../w/ like in water, so I'll write a W (double u), /o/ like in swan, so I'll write an A, /z/ like in laser, so I'll write an S. Taught the difference between letters and sounds and the varying ways we can write those sounds from the beginning is sustainable teaching that does not misrepresent the writing system. I teach my students, from the beginning, capital and lower case letter names, the 44 or so 'sounds' we have in English and that we can use graphs, digraphs or trigraphs to represent those sounds. They get it, it makes sense, there are no 'tricky words' or exceptions, there is no need to 'unteach' a simplistic letter/sound correspondence so that they can read/spell words such as 'want' or 'come'. It's really easy just to teach them the writing system as it actually works!!

Tim Shanahan
Jul 30, 2018 01:19 PM

You’re missing the point. Letters are not sounds...they represent sounds. Letters are actually collections of visual phenomena that represent collects of auditory phenomena. Readers have to come to treat these collections as equivalent... having a name for those collections is helpful to learners. Hence the value of letter names.

Stephen Parker
Jul 30, 2018 07:41 PM

Hello Tim,

You've now told me that I "need to know more about the sound system and how kids read." Further, I "might benefit from a class in psycho linguistics." You told Katy and Harriet that "they can't logic out whether a practice works best...that’s why we[?] do research. I’ll always go with the research over the opinions. Sorry. "

Here's what I notice. Whenever someone disagrees with you, you reflexively do two things: 1) you claim your view is backed by evidence and research, a tactic, by the way, which anyone can use, and then 2) you somewhat condescendingly suggest how, if the other party would only educate themselves a little better, they would certainly agree with you on the issue at hand.

Tell me: when you were an avid backer of Whole Language in the 90's, and as a member of the National Reading Panel, where was your concern for evidence and research then? Whole Language never had a shred of evidence behind it. You didn't even know that your own Panel specifically and repeatedly condemned Whole Language when we discussed that in an earlier Shanahan blog.

Above, you also tell me what you "hope" my book does. Again Tim, why not just read it? I'd send you the money myself to purchase it - but it's free! You say, on this blog, that you are "one of the world's premier literacy educators." As such, I'd like to receive some feedback from you (privately, or in a future blog). Maybe there's actually something in my book - a book aimed at "literacy educators" like yourself - with which you'd agree!

Best Wishes.

Tim Shanahan
Jul 30, 2018 08:20 PM

I definitely do present evidence for my claims. It’s funny that you find that to be so offensive. You say anyone can claim evidence, but then you don’t provide any.

I’m not sure what you are referring to about my staunch whole language positions. I’ve taught phonics since the 1960s and have written about it since 1970s.

You’re still misreading NRP.

Given that you are wrong about the role of letter names (according to the research), and you are wrong about the superiority of synthetic phonics (according to the research), why would I read your book on that? Do you have new research? More accurate claims?


Tory Callahan
Aug 02, 2018 03:49 PM

Jeff Bowers recently wrote that new methods are needed. We built one by teaching LN/LS in explicit relationship, a method consistent with Tim's citations. The key is understanding that LN/LS is not dichotomous but overlaps with high frequency. We explicitly teach the overlap and the various LN/LS relationships. Bingo. The kids categorize through structured inquiry. LS can be in "first position" of LN (Ex: b, d, j, k, p, t, v, z) or "second position" (Ex: f, m, n). Or not there at all (ex: h, w, y). We count the phonemes in LN and, wow, there's the sound (or sometimes not but it's explicated..flexible thinking...) Excellent PA training with high utility. We used this Overlap Teaching in preK and K. It developed via backward mapping from typical errors, what confuses children, and capitalizes on lots of research such as desirable difficulty, structured letter inquiry (like SWI), and high-impact phonemic awareness training and phonics. Tim cited importance of labels but also categorization is key because it uses analogy =thinking. (See Surfaces & Essence: Analogy as the Fuel and Fire of Thinking by Hofstadter & Sander.) Our method also exploits the key insight that information is better remembered if it is encoded in an elaborative and organized manner (Bowers & Bowers).
The kids totally get it, especially the strugglers which surprised me.

Last, please remember that LNs are nonsense words until we teach what establishes them as concepts, symbols...Thanks.

Harriett Janetos
Aug 04, 2018 11:56 PM

I've just started reading Gail Gillon's Phonological Awareness: From Research to Practice (2018), and she has a paragraph that supports what you've said on many occasions but can't be overemphasized:

"An interactive relationship between phoneme awareness and letter knowledge is further supported by the results of training studies that have demonstrated stronger effects on enhancing reading when phonological awareness training is combined with letter-sound knowledge training than when phonological awareness training occurs in isolation (Ball & Blachman, 1988; Bradley &
Bryant, 1985; Hatcher et al., 1994).

This is why I really value the word-building activities in Isabel Beck's Making Sense of Phonics.

Pete Bowers
Aug 09, 2018 02:23 AM

Hey all, I wanted to share a couple of responses to Tim's post and some of the comments.

I don't comment a lot, but when I do it is often too long. I'll post this as Part 1 and Part 2.

Part 1

From what I'm reading here, Tim is arguing that there is research evidence finding that we should teach letter names in early reading instead of referring to letters by commonly associated "sounds". However, that research has not shown us why this is the case. Like Tim, I'm happy to share some of my hypotheses about the "why" of this finding especially since that's where we are at when we don't have direct research to guide us.

I think the emphasis on recognizing that letters, graphemes and phonemes are all abstract concepts is a key point as is Tim's point that "we’ve long known that providing labels for concepts facilitates learning" is also a key part of this story.

I was very interested in Shani's post about recognizing that we should be talking about graphemes (single-letter graphemes, digraphs and trigraphs) and phonemes and orthographic markers such as etymological markers in which letters represent no phonology at all.

Tim rightly corrected Shani on one point -- even when they are in words, letters are never sounds. But I'd push back on part of Tim's response to Shani and offer some other observations.

Tim writes to Shani: Letters are not sounds...they represent sounds. Letters are actually collections of visual phenomena that represent collects of auditory phenomena. Readers have to come to treat these collections as equivalent... having a name for those collections is helpful to learners.

Actually, Tim, I think you will agree that letters in-and-of-themsleves do no represent "sounds" (properly referred to as "phonemes") even when they are in the spelling of a word. Letters are abstract raw material of orthography. The letters of our alphabet have many functions -- only one of which is to be the raw material that comprise graphemes for the writing of phonemes. Certainly this is the most common function of letters, but without understanding that letters have other jobs, we end up presenting a confused, unreliable picture of spelling. Consider words like "house" or "two". What is the job of the final "e" in "house" or the "w" in "two"? It turns out that the final "e" here is what Venezky (1999) describes as a plural cancelling marker. We see this same marker in words like "please," "nurse," and countless words that would otherwise have a final single "s" that could look like plurals but are not. (There are other means of plural cancelling as well). The "w" in "two" is an etymological marker that is never associated with any pronunciation in the morphological family of the base "two," but which provides a signal to signal its connection to words that have the sense of "two things" that begin with the letter sequence "tw" in which that "w" is a grapheme (twice, twenty, between, twin etc.). That marker also serves to distinguish this word from its homophones "to" and "too". Similarly, the final "e" in "please" distinguishes this word from its homophone, the plural "pleas".

People may argue that this this is too advanced content for young learners, but that is an empirical claim that I have not seen tested. I have, however, taught such lessons in many kindergarten and Grade 1 classes with very engaged children. When we don't have research, anecdotal evidence is quite useful. Of course, if these ideas are new to teachers, they cannot have tested whether they are too advanced for young children.

On to Part 2...

Pete Bowers
Aug 09, 2018 02:24 AM

Part 2 (see previous comment this builds on)

Since it is the case that letters have jobs other than comprising graphemes to write phonemes, at least educators should be aware that just because (another plural cancelling "e"!) we see a letter in a word, we should not expect there should necessarily be a phoneme associated with it.

But the issue of confusing letters as writing sounds goes far deeper than this issue. By far the greatest confusion I see in classrooms when children are taught to think of letters representing sounds is that they are hindered from understanding when they are looking at a single letter grapheme or a letter in a digraph or trigraph. We have no idea what "sound" is associated with the "t" until we see that letter in a word, and even then, we need to know if it is a grapheme itself, or part of a digraph or trigraph. The words "the" "path" "catch" and "action" all have the letter "t" in them, but none of these words have what would be called the "t sound". Of course in the first three of these words the "t" is not a grapheme, but part of the digraph "th" or "tch". In the word "action" the "t" is writing the phoneme /?/ that we associate with the "sh" digraph we find in "shell" or "bash".

If we are to take on the practice of referring to letters by their supposed "sounds" (even the most common ones) we are deciding to teach young children something that we know we will later have to unteach. How could we make sense of the spelling of "the" or "eat" or "play" if we could not use the letter names?

If we then say we have to choose controlled text that avoids any digraphs or trigraphs we are choosing to work with uninteresting text and we are going to have to teach single-letter graphemes like "j" or "z" that are very rare before we teach incredibly frequent digraphs like "th" or "sh" or any of the vowel digraphs.

So how can we help children become confident with the names of these abstract categories (graphemes) that write the phonemes in English words. My own practice is to find straight forward ways of making these implicit orthographic structures explicit by spelling words out loud in the orthographic structures found in words. So I would spell the base "play" "p - l - ay" and I would spell "the" out loud as "th - e". First I'm just exposing children to these graphemic structures, but very soon, I can ask young to see if they can "spell-out" like me. When they do, I can ask them why they think I announced the "ay" together in "play" but not the "pl" or "la". I find children are very quick to figure out that this "thing" I'm calling "ay" is what is writing that phoneme they hear a the end of that word that sounds like the name of the letter "a". Same with "th" or "igh" when I get to words with those graphemes. Once kids know that there can be digraphs and trigraphs they are quickly on the hunt for more. This is an opportunity for morphological instruction as we can show that the base "graph" is for writing or mark and associate it with things in their world like "computer graphics" the "graphite" in their pencil that leaves a mark on the paper etc. Ironically, "trigraph" is often easier to teach first because kids know what a "triangle" or "tricycle" is. Once that is set, I can show them that "bi-" can be a prefix for "two" just like "tri-" can be a prefix for three.

What's more, these categories are not presented as though they are only associated with just one "sound". If we have learned what the "c" grapheme can represent in a word like "cat" this becomes a point of interest when we run into the word "city". Similarly the "ea" is writing a different phoneme in "heal" and "health" -- and that helps us see that these words are related in meaning even though the pronunciation changes.

All of that learning is undermined if they are taught the alphabet as if each letter has a name that is one of the sounds that can be associated with it when that letter is used as a single letter grapheme.

By the way, this is another reason why I introduce the terms grapheme, digraph, and trigraph very early. Having a label for the abstract thing that writes phonemes helps us avoid the single-letter thinking that I find extremely challenging for many children to unlearn.

Here's a way that I like to think about all of this with teachers. If I ask you to write the digit "9" on a piece of paper and then ask you what that digit is worth, you might start by saying "9". But of course, as soon as I put that digit in a NUMBER like 90, we see that this digit just the raw material of math that we can't assign a value to until it is in a number. We have ten digits that can represent any number. We know that there is a difference between "digit" and "number". I would encourage teachers to make sure that they distinguish in their own thinking between "letters" and the orthographic structures for which they can be used.

If anyone is interested, I do have a document at the following link that lays out my own practice for "spelling-out-loud" and "writing-out-loud" word structure. This practice includes conventions for marking morphological boundaries, suffixing changes and orthographic markers.

To be clear, I'm not claiming that these recommendations have been researched experimentally. They are practices that are linked to an accurate linguistic understanding of the morphological, graphemic and orthographic markers that letters are used for in our writing system. A major advantage I find for teachers who take this work on is that they start to deepen their understanding of these structures because they run into situations where they realize that they are not sure what a given letter is doing in a word. When we identify what we don't know, we give ourselves a cue for what we need to learn.

Harriett Janetos
Aug 09, 2018 10:27 PM

There's an interesting discussion of morphology in Ending the Reading Wars: Reading Acquisition from Novice to Expert, Castles, A., Rastle, K., & Nation, K. (2018). Psychological Science in the Public Interest, 19, 5–51. doi:10.1177/1529100618772271. In particular:

"In summary, though we believe that explicit instruction on the nature of morphological relationships in the writing system is likely to benefit the acquisition of literacy, the form of instruction likely to be most effective remains unclear. One important question is this: When should morphological instruction linked to printed words begin? Some researchers have argued that it should be introduced at the earliest stages of learning to read, before alphabetic knowledge is firmly established (e.g., J. S. Bowers & Bowers, 2017; Devonshire et al., 2013). However, this suggestion awaits evidence. Analyses of the Children’s Printed Word Frequency Database (Masterson et al., 2010) suggest that children’s text experience in the first year of reading instruction consists
overwhelmingly of words with a single morpheme (Rastle, 2018). Thus, morphological instruction can play only a limited role and may detract from vital time spent learning spelling-sound relationships. Instead, we would predict that the benefits of explicit morphological instruction are more likely to be observed somewhat later in reading development, promoting learning as
children accumulate the experience necessary to accomplish the direct mapping between spelling and meaning (Rastle, 2018). That is not to say that classroom instruction should not include activities to support the development of rich vocabulary knowledge, which of course will include morphologically complex words. This can be achieved via listening activities, storytelling, and so on (see Section 3.4). When and how explicit instruction regarding orthography-morphology links should be introduced are important questions for future research."

Peter Bowers
Aug 13, 2018 05:03 AM

Hello Harriett,

Your comment about the article above is a bit of a tangent from this thread, but related to discussions we've had in the past.

I encourage you and others to consider the response to this recent article you cite that my brother Jeff Bowers and I have posted titled, "There is no evidence to support the hypothesis that systematic phonics should precede morphological instruction: Response to Rastle and colleagues" at this link:

You may or may not agree with our arguments, but I hope everyone will consider differing positions before drawing conclusions. On that note, I want to highlight that I was most impressed that Ann Castles, the lead author of the article you cite specifically Tweeted her followers to consider our response. That is a wonderful model of scholarship -- pointing people to research challenging one's on articles.

I will just share one observation here for those who don't go on to the full response, or perhaps to encourage you to do so. Consider this statement from the quote you gave:

"Instead, we would predict that the benefits of explicit morphological instruction are more likely to be observed somewhat later in reading development, promoting learning as children accumulate the experience necessary to accomplish the direct mapping between spelling and meaning (Rastle, 2018)."

Notice that they say that they would "predict". This is because they are sharing their hypothesis. This is not a finding based on instructional studies.

Note, like the first claim, the argument that morphological instruction "can only play a limited role" and that it "may detract from vital time spent learning spelling-sound relationships" are NOT based on evidence from instructional studies.

So these are hypotheses awaiting evidence.

I would also highlight that the main argument offered against our claim that morphological instruction should not be introduced before instruction about phonological aspects of spelling is that we do not have enough evidence yet.

It turns out that we do have a fair bit of evidence that is not a direct test of this question, but is definitely extremely relevant, and which they do not address in their paper.

We currently have three meta-analyses of morphological instruction (Bowers, Deacon & Kirby, 2010; Goodwin & Awn, 2010, 2013) and two reviews (Carlisle, 2010, Reed, 2008). Consider this description of the findings from those studies of morphological instruction studies:

Reed (2008), 7 studies K-12, benefits overall, especially less able
Bowers, Kirby & Deacon (2010), 19 studies,K-8, benefits overall, larger for younger and less able
Goodwin & Ahn (2010), 17 studies K-12 (LD), benefits for less able
Carlisle (2010), 16 studies K-12, benefits overall
Goodwin & Ahn (2013), 30 studies K-12, benefits overall, larger for younger

Note the studies that were able to draw conclusions about younger vs. older effects of morphological instruction found younger students gained more. As well less able students gained the most. For the claim in the study you cite to hold up, they need to be able to account for the fact that these findings are in the opposite direction of their prediction. I have seen no such explanation.

As to the second claim that morphological instruction may detract from phonological learning -- there is no instructional evidence cited for this either. It is also worth noting that in both of Goodwin and Ahn's studies, the greatest effect of the morphological instruction were for phonological measures.

In their 2010 study (focusing on students with literacy difficulties) effect sizes for phonological awareness outcomes were d=.40 and morphological awareness was d=.40.

In their 2013 study that included all students the effects for phonological awareness was d=.48 and morphological awareness was d = .44

Again these are effects from instructional studies that are not in the direction that would be predicted by the hypothesis that including morphological instruction should hinder learning about phonological aspects of the writing system.

It turns out the research evidence we have that is relevant to this question is in the exact opposite direction that most researchers have assumed for decades and which is asserted in the quotes you identify.

For there to be research evidence that we should avoid morphological instruction at the beginning of instruction, we would need instructional evidence that showed weaker outcomes when morphology is included from the start with phonologically based instruction compared to when grapheme-phoneme correspondences are taught from the start BEFORE morphology is included. I have yet to see anyone point to any such evidence. If you or anyone can point to such evidence, I would be very eager to study it.

While we don't have much direct studies of this question, one study that does do this directly is by Devonshire, Morris and Fluck (2013). They compared a phonics condition to an SWI type condition with 4-7 year olds. The SWI condition taught the interrelation of morphology, etymology and phonology. They found significant effects for standardized measures of reading and spelling. This study has been challenged by Rastle and others for not really being a "phonics condition". In our response you will see that we find no justification for rejecting this study for not using a phonics condition.

You can judge for yourselves by reading the Devonshire paper at this link:

These are just a few of the issues that readers/researchers need to keep in mind when reading the paper you have identified Harriett. I do hope you and others will read our more detailed response I've linked.

I hope you, Tim and his readers find all of this productive food for thought. I'm very open to any arguments showing flaws in this analysis. I do hope, however, that if you do, you first read our paper linked at the top with much more detail.



Tory Callahan
Aug 13, 2018 01:37 PM

Thanks to the many links provided which I look forward to reviewing and also the interesting comments.

The issue of how single grapheme to phoneme teaching potentially hinders "...understanding when they are looking at a single-letter grapheme or a letter in a digraph or trigraph..." is important. This is conceptual teaching that we do early. Explicitly teaching the relationship between LS and LN and how they overlap in simultaneous LN/LS teaching does not preclude teaching "system behaviors" like di/trigraphs. We teach early how T and H and then TH behave. It's a very common behavior and key concept: one language sound can be represented by more than one letter. TH is the earliest example we teach since it is such high frequency in early words. (I'm talking kindergarten here.) It is absolutely the truth, as stated already, that timing is key. If /t/ and /h/ are too firmly established/routinized before /th/, it is much harder to teach and you do have to "undo." The kids start to trust you a little bit less!

We have tried to build a researched-informed "systematic" sequence that exploits both morphology and phonology. Ex: We use the "tw" to teach concept of twoness first through twin--it's rigorous PA training and also a concrete introduction to how orthography marks concepts, meaning. Starting with "twin" where one can discern /w/ (and they can as Tim noted earlier) helps anchor concept and spelling first . Also, "twin" is better known for children who do not yet know numbers and start school with less language/book experience. Details matter a lot we find.

Harriett Janetos
Aug 13, 2018 03:16 PM

I, too, look forward to pursuing the links. I think teaching "twin" first, which is perfectly decodable, before asking children to recognize and remember "two" makes perfect sense. That's how I see phonics preceding morphology in beginning reading.

Like Tory, I also introduce digraphs to kindergarteners early on as they appear in high frequency words like "THem" "SHe" or "plAY" so there is not an exclusive emphasis on one letter one sound but on however many letters represent ONE sound.

See the article "Rethinking Sight Words" by Katharine Pace Miles in the May/June Reading Teacher for a good review of Linnea Ehri's research on beginning reading.

Harriett Janetos
Aug 13, 2018 09:32 PM

The article "There is no evidence to support the hypothesis that systematic phonics should precede morphological instruction: Response to Rastle and colleagues" is definitely worth reading because it really frames the debate. The following statement about the types of words taught in beginning reading stood out because it is certainly not in line with all the beginning reading texts I've worked with.

"Most words that children need to learn at the start of instruction are multisyllabic and/or multimorphemic, and these words are far less regular according to phonics."

And the later reference to words like "walk" and "sign" are not examples of words that are taught to beginning readers at our school. Those of us recommending that morphology instruction begin after beginning readers are grounded in decoding through PA and phonics instruction (the "phonology first" hypothesis) might be making this recommendation because for us monosyllabic words are the majority of words taught to our beginning readers.

Peter Bowers
Aug 14, 2018 05:36 PM

Thanks Harriett for your comments and for taking the time to read the article.

You write:

" Those of us recommending that morphology instruction begin after beginning readers are grounded in decoding through PA and phonics instruction (the "phonology first" hypothesis) might be making this recommendation because for us monosyllabic words are the majority of words taught to our beginning readers."

I'd encourage you to consider why "monosyllabic words are the majority of words taught to our beginning readers."

I would argue that this is not because the majority of the words beginning readers encounter in their world, and in the books parents and teachers read to them. If a child's beginning reading instruction is limited to mostly monosyllabic words, that is because a conscious choice has been made to restrict the words that are presented to them.

One can make the argument that this is a wise choice, but once again, I know of no instructional evidence on which to base that choice compared to the way I would read a story that kids love and then use words from that text as a launching pad for investigating how the written word works.

This is not to say that reading instruction cannot include texts with very basic language, but I know of no principled reason or research evidence that shows it should be restricted to such texts.

For example, I often point to this video of a preschool class that is investigating the morphological family of the word "rain" because they had just had some major rain after a period of severe drought.

While "rain" is a monosyllable, of course the kids suggest all sorts of multisyllabic words to study at the same time. And in the process they encounter interesting grapheme-phoneme correspondences such as the fact that when you add the "-s" suffix to this base to write something like "it rains a lot," that suffix written with "s" is representing the pronunciation /z/. There is no big deal made about it, but this is an early introduction to the reality of our writing system. Most graphemes can write more than one phoneme, and that is a good thing. When that "-s" suffix is marking a plural, we can make sense of the spelling structure and meaning and phonology of words like "cats" and "dogs". This is pre-school SWI instruction that is planting seeds of understanding that never has to be undone. It may be sparked by a monosyllable word in this case, but it does not restrict the words they encounter -- and as you'll see in the video, the children themselves are proposing the morphologically complex and multisyllabic words. Also, this is not the only type of "spelling and reading instruction" but when Carolee sits down and reads with a child in this class one-on-one, she can draw on these common learning experiences when encountering countless words.

And this example returns this string in part to the topic of the post - using letter names or pronunciations in beginning literacy instruction. If we call the letter "s" by one of its pronunciations /s/ and the letter "z" by it's main pronunciation /z/, how would a child make sense of monosyllabic words where that "s" grapheme is writing the phoneme /z/ that has been learned as the name of the letter "z"? And its not just a problem for words with an "-s" suffix (e.g. dogs, bags, days and countless others), it is also instruction that would have to be ignored in words like "has," "use" or "is".

If readers are interested in another video of SWI instruction with high frequency words that cannot be understood with phonics alone, the first video in the series of videos of lessons I taught at a school in Switzerland some time ago gets into the spelling of homophones and function and content (lexical) words in a Grade 1 class.

From the "phonology first" perspective, the instruction in both of these videos would be in appropriate instruction. Both videos are integrating other linguistic features of English with grapheme-phoneme correspondence from the beginning.

One of the challenges in communicating the logic of teaching the way grapheme-phoneme correspondences are influenced by other linguistic factors including morphological and etymological constraints (the homophone principle is part of etymology) is that few teachers or researchers have any frame for what this kind of linguistic instruction could look like in young classrooms. But if that statement is fair, it means the "phonology first" perspective is making an recommendations that teachers avoid instruction that they themselves to not understand. Not only is there no instructional research to support the argument that phonological aspects of spelling should be taught before other linguistic aspects are introduced -- most making that claim are doing so without an understanding of the instruction they are encouraging others to avoid.

I'd make one last point that grows from your highlighting of the Castles, A., Rastle, K., & Nation, K. (2018) paper and the response from Jeff and I that I shared.

Obviously I am engaged with literacy research. However, too often research is downloaded on teachers who are not in a position to critically analyze that research. Talk to any researcher ask them if they see problematic research cited all the time as though it were fully resolved. We need to read research knowing that it is supposed to be a presentation of ideas to a community that is EXPECTED to challenge that research. That's one reason I was so impressed that Ann Castles, the first author on the paper you cited points her Twitter followers to the challenge to her paper by Jeff and I.

I'm not saying that people should agree or disagree with the Castle's et al paper or our response. Instead, I'm suggesting that reading these kinds of articles and responses provides a useful reminder to the reader that all research should be critically analyzed by readers.

In my philosophy of science 101 course in my grad work, I was taught that scientists should not accept any hypothesis as being confirmed until all plausible alternative hypotheses have been falsified. So far I've seen no falsification of the plausible hypothesis that from the beginning, children should receive instruction that accurately represents how their writing system works. The "phonology first hypothesis" demonstrably misrepresents how English orthography works. We should therefore be very cautious about how tightly we hold onto that hypothesis. And if we are doing the best scientific work, we should be actively trying to disprove the hypothesis we think may be best. On that note -- I hope any readers will point to logical or research errors in what I'm posting here, or on Jeff's blog where the linked paper is. Jeff put the paper up on that blog specifically to encourage a scholarly discussion.

If any readers have challenges to our paper, please articulate them at Jeff's blog!

I always appreciate your openness to engage with these debates Harriett.


Tory Callahan
Aug 14, 2018 08:07 PM

I respond here on the point Pete relates to the original post and will follow up on the other issues on Jeff's blog. Really appreciate the comments, Harriet and Pete; they are important. For now, related to original topic of LS/LN post, I am commenting as someone who is not "phonology first" but understands the difficulties children are trying to sort out. I respond to this:
"And this example returns this string in part to the topic of the post - using letter names or pronunciations in beginning literacy instruction. If we call the letter "s" by one of its pronunciations /s/ and the letter "z" by it's main pronunciation /z/, how would a child make sense of monosyllabic words where that "s" grapheme is writing the phoneme /z/ that has been learned as the name of the letter "z"? And its not just a problem for words with an "-s" suffix (e.g. dogs, bags, days and countless others), it is also instruction that would have to be ignored in words like "has," "use" or "is". "

Teaching children to be flexible is the very point--it illustrates how the system works!-- but it doesn't mean that the explicit relationship between primary LS and LN has no initial utility. It clears up a lot of difficulty. Nothing is always. We do anchor s as writing /s/ and z as writing /z/. But, quickly, early, we teach is/his/as/has (we resequenced the Fry list..) and what is going on in these words and the related conceptual teaching (one letter can represent more than one phoneme, depending on position, etc.) We also expose the children to the sounds of plural s but emphasizing, in this context, the meaning rather than pronunciation or phonology...and remember that Spanish and other "non-standard American" dialect speakers pronounce /s/ for standard American pronunciations of words where s takes /z/.

You just have to teach the interplay within the structure. As we've already established, phonology changes based on context (and dialect); it's not absolute. We also explicitly teach the difference between "use" /yoos/ as in "what's the use of..." vs /yooz/ " I will use this pen..."

I just don't see things as "first" anymore. That "phasing" served a role but now, I think, takes us off track. Explicit overlap and interleaving better describes how it unfolds. I think meaning based teaching--across unit sizes-- is as important as phonemes; they are both "binding agents." I could argue the meaning links are more important--certainly equally so-- for children who come to school with less language background. That's what likely brings them down eventually. With the previous example, "tw" becomes a binding agent for twin, two, twice, etc. It's not a morpheme or a phoneme but it's meaningful in the concept of twoness, as is "both" which we also connect...nothing is always, including "tw" to signal twoness. And "tweet" and "twitter" are the disrupters here, right? Another example of why we have to be systematic and yet nimble whichever unit size we are stressing in whatever context.

Thanks Pete and Harriet.

Harriett Janetos
Aug 14, 2018 10:24 PM

Since our dueling anecdotes won't get us anywhere, until more research is done I highly recommend these books featured in "Ending the reading wars: Reading acquisition from novice to expert". Until next time . . .

Dehaene, S. (2009). Reading in the brain. New York, NY: Penguin Viking
Kilpatrick, D.A. (2015). Essentials of assessing, preventing and overcoming reading difficulties. Hoboken, NJ: Wiley
Seidenberg, M. (2017). Language at the speed of sight: How we read, why so many can’t, and what can be done about it. New York, NY: Basic Books
Stuart, M., & Stainthorp, R. (2015). Reading development and teaching. Thousand Oaks, CA: SAGE
Willingham, D. (2017). The reading mind: A cognitive approach to understanding how the mind reads. San Francisco, CA: Jossey Bass
Wolf, M. (2007). Proust and the squid: The story and science of the reading brain. New York, NY: Harper Collins

Bev Sace
Aug 15, 2018 01:33 AM

As a kindergarten teacher and an early years language and literacy teacher and consultant in Hong Kong, where children are trilingual (or quadringual even), I think letter names should be taught only after they have learnt the most common sound values of the graphemes (most common spellings) of the 44 English sounds.

I speak through experience. Many times, a child would read the word cat either by sight (very prevalent in HK) or some attempt at synthesizing. However, those with both letter name and letter sound knowledge use both names and sounds! For example, and this is a real one, a child once read cat as /see-ah-tt/ and blended it as /sat/. They oscillate between their knowledge of letter sounds and letter names, and come out frustrated, and with lower reading confidence.

If children are taught just the sounds, there will be no delay in blending time (as sounds are recalled with automaticity), no confusion between names and sounds, at least at the very early stages of reading instruction.

Granted, many children come to school already knowing their letter names. That is why 1) I explain to them (and their teachers!) that same with the cat animal, we call the animal “cat” but its animal sound is “meow”. And for letters - this is the letter but its sound is /a/. 2) I say that when they are in Ms Beverly’s lesson, we will only use letter sounds (until they are ready to move on to alternative spellings where we need letter names to talk about the correct spellings of the same sound, eg etc.)

I suppose it’s a bit like knowing a few languages and code switching. I mix up English, French, Tagalog and Chinese when I speak at home sometimes because I know these languages and they are available to me.

Also the alphabet and song is quite limited in its power to provide children a mnemonic when they feel lost. However if they sing the alphabet with sounds instead of names, it would be a much better help for reading. I have personally come up with a vowel song too, with all the vowels of the English language. This helps my students. If there is one to learn the consonant digraphs as well, children would have a better way of consolidating all 44 English sounds (or 40, depending on regional English).

This is my experience and it’s worth exploring research to give me better information on this - or start one! Especially in an early childhood second language context.

Bev Sace
Aug 15, 2018 02:06 AM

I didn’t see the long discussion that’s happened prior to my comment. Very interesting to learn about what teachers and psychologists think!

I have attended a seminar in Perth on literacy and language in 2017 and I dont remember now who said it - teachers like to keep doing what they think works, without being informed by research.

On the other hand, sometimes research gives us non-conclusions or confusing ones! If I would argue for analytic phonics (I do not!) I would find research that would seem to back it up.

We have first language settings and L2 settings. We have different points of views from SLTs, linguists, psychologists and teachers.

I can tell you what works best in Hong Kong, but it may not be the same as what works best in another country due to different first languages (for example in Hong Kong, children learn to read Chinese through sight reading, so many teachers think that English reading should be by sight reading too, whereas in India, they use Hindi sounds to try to read English words).

So I try to keep abreast of research on reading in both L1 and L2, in ESL and EFL contexts, and on child dev and reading science. It’s a complicated undertaking.

My point is, while it is superb for us professionals to keep arguing our points, and be informed, we still need to translate all of what we know to make our teaching fit for purpose for children at the right times. For me, no teaching of letter names until they master all the basic sounds of the basic code.

Tory Callahan
Aug 16, 2018 11:37 AM

Bev, interesting on many counts. The issue with spelling "cat" is a great example of why explicitly teaching the difference between LN and LS, in terms of how they relate, can help a child understand how this LN/LS system works and clear up confusion. And the concept of language sounds vs. meow. LN for C doesn't contain its primary sound /k/ (as so many letters do such as B, V, M, etc.) Since C writes /s/ cued by following letter, "place value" or position in word comes into the picture. At that point, you are into word level, or how letters behave in word contexts. (For us, this is later teaching unless a specific child asks.) Words are the point of course; putting letter knowledge in word context is important (vs. extended, isolated phoneme-grapheme teaching). (Bev, do you use the pinyin system? I have wondered about this for American students: a subset of letters, say two vowels and 5 consonants and a consonant digraph. Teach how the system works from letter knowledge into word level. Just an overview sort of thing. Then teach the rest of the letters, once this conceptual teaching with limited examples about how the system works has been seeded or introduced.

As I think about this thread, I wonder about "phonology first" as a possibility for first letter knowledge teaching. It's interesting how little research exists on this topic. The recent Castles et al piece barely mentions letter knowledge. The research typically takes off at word level.

Harriet, I've read and learned from those books many times (and the Gillon). Thank you. When studies don't clearly guide us, it's especially useful to consider, for example, what Tim writes about LN/LS teaching. And even with studies, what gets studied and what does not, impacts what we see and also important things we might then miss just because they aren't yet studied. I can think a key issue is resolved. Then I search for the evidence and it's not there. Tim tried to keep us to the evidence trail on this LN/LS issue (even thought evidence about specific teaching strands is limited.) And, yes, as Bev notes, if one doesn't like what the evidence says, it tends to be ignored! Castles et al repeatedly call for further research even, for example, to determine what implementations are effective in terms of "systematic" phonics, what's the sequence, how much to teach, how deep it goes, etc.

What Are your thoughts?

Leave me a comment and I would like to have a discussion with you!

Comment *

Should We Teach Letter Names?


One of the world’s premier literacy educators.

He studies reading and writing across all ages and abilities. Feel free to contact him.