top of page
  • Writer's pictureStephen Parker

The Kilpatrick Conundrum

Updated: Apr 8, 2022

Dr. Nathan Clemens (Univ. of Texas, Austin), along with 8 colleagues from various other U.S. universities, recently penned an article (see here) sharply critical of David Kilpatrick. David responded to Clemens et al here.

Kilpatrick seeks to dismiss the entire Clemens paper on the basis of his (Kilpatrick’s) own understanding of Orthographic Mapping (OM). That's unfortunate. It's precisely Kilpatrick's understanding of OM, in particular, the supposed necessity of "phonemic proficiency" to achieve it, that Clemens et al is questioning. In responding as he does, however, Kilpatrick demonstrates how his own views of OM significantly differ from those of Linnea Ehri. (Ehri is the research scientist who first identified Orthographic Mapping.) This blog will focus on those differences between Kilpatrick and Ehri.

Decoding versus Orthographic Mapping?

The following quotes are from Part 2 of Kilpatrick's response to Clemens. Note how Kilpatrick criticizes Clemens et al for not understanding, as he does, that decoding and orthographic mapping are different processes:

“When discussing phonemic proficiency, which is a cognitive skill needed for orthographic mapping, Clemens et al. confuses… phonetic decoding and orthographic mapping… However, phonemic proficiency is a skill needed for orthographic mapping, not for decoding.”
“Orthographic mapping is the process of remembering written words after they have been correctly identified/decoded (Ehri, 2005; Kilpatrick, 2015)… This confusion by Clemens et al. about the distinction between decoding and orthographic mapping relative to phonemic proficiency nullifies most of what is written in their manuscript.”

And here's Kilpatrick again making a clear distinction between decoding and OM in his book Essentials:

"OM goes in the opposite direction of phonic decoding. Decoding starts with letters and uses letter-sound knowledge and blending to activate a spoken word. There is no need to pull apart the sounds of the words. In decoding, the sounds are already pulled apart. The task is to blend those parts. By contrast, OM requires one to analyze the pronunciation of words in the phonological lexicon. These now separated phonemes will be used as the anchoring points for the individual letters in the written word. Think of decoding as going from text to brain and OM as going from brain to text... It is important that we do not confuse OM and decoding." (pp98-99)

This quote from Essentials raises many questions. What's this distinction between "activating" and "anchoring"? If a word has (literally) just been constructed (via decoding), by blending all its phonemes into a pronunciation, what is now added to this process by "analyzing" that pronunciation and re-separating its phonemes? For Kilpatrick, does OM only occur in "pulling apart" phonemes, but not in combining them? And is this true even if the blending (putting together) of those phonemes was just consciously done by the student?

This requires a closer look...

Orthographic Mapping as an Encoding-Only Process?

To understand this section, you need to be comfortable with this terminology:

  • Decoding and synthesis (putting phonemes together) are used as synonyms. The phonemic awareness skill required to do either one is blending.

  • Encoding and analysis (pulling phonemes apart) are used as synonyms. The phonemic awareness skill required to do either one is segmentation.

With this understanding, look at these quotes from Kilpatrick's books:

  • “The only way a reader can access the phoneme sequence in order to anchor it to the spoken pronunciation is by phoneme awareness-analysis skills.” (Essentials p101)

  • “Phoneme awareness-analysis [as opposed to synthesis-blending-decoding] is essential for storing words in one’s sight vocabulary [OM].” (Essentials p101)

  • “OM is an encoding process responsible for word recognition [OM].” (Essentials p81)

  • “By itself, decoding is not capable of producing a sight word memory [i.e.. not capable of resulting in OM].” (Essentials p95)

  • "By itself, decoding [the word SIT] is not enough to remember that printed word." (Essentials p97)

  • “Theoretically speaking, the only phonemic awareness skill a student needs for OM is segmentation. A student needs to be aware of each phoneme in the word to align those phonemes to the letters. (Essentials p119)

  • “The key is to go from pronunciation to letters (orthographic mapping) rather than letters to pronunciation (decoding).” (Equipped p40).

  • “Decoding uses letter-to-sound relations to activate oral words from an unfamiliar letter string. By contrast, OM uses sound-to-letter relationships to anchor phonemes in a word’s pronunciation to the printed letter strings into long-term memory for future retrieval.” (Equipped p40).

  • “Decoding uses letter-sound skills to identify words while OM uses letter-sound skills to establish a memory of printed words.” (Equipped p40)

  • "Blending puts sounds together while analysis tasks (segmentation, manipulation) pull words apart. Blending is required for decoding while analysis [encoding, segmentation] is required for OM and spelling." (Equipped p73)

So for Kilpatrick, OM is indeed an encoding/segmentation process - not a decoding process. Decoding/blending helps to identify words, but it does not help to remember those words.

Are Basic Encoding/Segmentation Skills Enough for OM?

No, not according to Kilpatrick. Having asserted that decoding is not capable of producing a sight word memory, he now proceeds to say simple encoding ability is not enough either:

  • "Orthographic mapping requires advanced phonemic awareness." (Essentials p96) [This is now called 'phonemic proficiency' by Kilpatrick.] This applies not just to struggling readers, but to "typically developing students" as well. (Essentials p96)

  • "Advanced phonemic awareness appears to be needed for efficient sight vocabulary development [OM]." (Essentials p85)

  • "OM requires advanced phonemic proficiency." (Essentials p97)

  • "Until phonemic proficiency is developed, a student will not have an efficient way to make letter strings familiar." (Equipped p35)

  • "What needs to change? [We must] train phoneme awareness to the level of phonemic proficiency." (Equipped p45)

  • "The One Minute Activities [in Equipped] are designed to develop automatic, effortless phoneme proficiency which is a prerequisite for skilled orthographic mapping." (Equipped p83)

  • "All students" need training in phonological awareness through the "advanced" level (i.e. levels J, K, L, and M in Equipped). (Equipped p84)

Is the Above Compatible with Linnea Ehri?

At this point it would seem prudent to ask a question: Are the above ideas compatible with Ehri's research and writings? Unequivocally, they are not. According to her many writings (not just the 2005 Ehri paper which Kilpatrick usually cites), OM is essentially a connection-making process that makes sight words possible. The connections that must be made are between the individual graphemes (seen in a word’s spelling), and the individual phonemes symbolized by those graphemes (and detected in that word’s pronunciation).

But this connection-making is precisely what’s done in the process of through-the-word decoding. Indeed, the student MUST make these connections for the decoding to be successful. In my conversations with Ehri last summer, she agreed that decoding is the primary way, though not the only way, that OM occurs. Other ways include encoding words by connecting phonemes to graphemes to learn spellings, or viewing a written word, hearing it pronounced, and automatically forming connections between graphemes and phonemes to secure the word’s spelling bonded to its pronunciation in memory. However, the latter path requires someone to identify the word for the student or for the student to predict the word correctly from its context. (My detailed blog on Ehri can be found here.)

Wanting to be sure of my own understanding, I’ve again, just recently, been in contact with Ehri. Here’s what she says:

“David’s concept of OM appears to differ from mine. In my view, OM involves the connection-forming process linking graphemes in spellings to phonemes in pronunciations to secure the words in memory for sight word reading and for spelling. OM covers connections formed when words are decoded and when they are encoded to learn and remember their spellings. The most important phonemic awareness skills contributing to OM are the ability to segment spoken words into phonemes, to blend phonemes to create recognizable spoken words, and to recognize how graphemes in spellings match up to phonemes in pronunciations of words. Perhaps the latter is better described as grapho-phonemic awareness.” [boldface mine]
“Phase theory portrays changes in the types of alphabetic connections that learners form as their sight word reading and spelling skill (memory for the spellings of words bonded to pronunciations) develops, from non-alphabetic visual cues, to partial GP connections, to fully analyzed GP connections, to use of multi-letter, GP-consolidated connections.”

It seems clear Ehri does not think decoding and OM are opposite processes. Likewise, Ehri does not think OM is only an encoding-analysis-segmentation skill. In fact, she agrees that the primary way OM happens is through decoding-synthesis-blending. OM can happen through encoding-analysis-segmentation as well, but only if the unknown word is identified by someone else, or if the child successfully guesses what the word is from its context. Ehri does not think OM requires “advanced phonemic awareness" (now called "phonemic proficiency"), rather, it requires just ordinary blending skills (for reading) and segmenting skills (for spelling).

Additionally, Ehri would never agree that onset-rime instruction occur before phonics even commences. Yet this is precisely what Kilpatrick advises in his book Equipped on pages 49-51, 55-56.

"The linguistic approach focuses on the rime unit... This is often called the word family approach... The linguistic approach is developmentally more appropriate for beginning readers than the phonics approach... The linguistic reading approach is like 'training wheels' for learning to read... If you use this 'linguistics first, phonics second' approach, while systematically training phonological awareness, you will reduce the number of struggling readers."

The above image is from p56 in Equipped. Here's what Kilpatrick says about it:

"Rime units can play a key role in early reading... As students learn rime units, they can be put on a 'word wall.' With this word wall there is a quick reinforcement activity that can be used at various times during the day. With a pointer, the teacher can randomly point to the rime units on the word wall and have the children call them out."

[Note: reading by using larger units such as rimes, prefixes and suffixes, occurs, according to Ehri, in her final (4th) consolidated phase.]


Here's one final idea from Kilpatrick (not directly related to OM) that I believe merits close scrutiny. In my view, it's seriously misguided. Judge for yourself whether or not you think this is reasonable thing to say to teachers, or to expect of children:

“If a student cannot do the more basic phoneme-level processing (levels H, I, and J in this program), he or she is not ready for a phonics approach. Such a student should continue with linguistic materials [onset-rime materials] until the basic phoneme-level skills emerge." [David Kilpatrick, Equipped for Reading Success, p51]

Here are some examples from Levels G, H, I, and J in Equipped. Level G (onset-rime) is assumed already-mastered in the above quote:

Level G:

six instead of /ix/ say /ack/ sack

lamb instead of /am/ say /aust/ lost

Level H:

flame instead of /f/ say /b/ blame

shriek instead of /sh/ say /k/ creek

Level I:

master don’t say /s/ matter

reception don’t say /p/ recession

Level J:

knit instead of /i/ say /E/ neat

ball instead of /aw/ say /e/ bell

robot instead of /o/ say /O/ rowboat

wonder instead of /uh/ say /ah/ wander

commit instead of /i/ say /U/ commute

Let’s suppose you can do this sort of thing. But consider: you’re a skilled reader, and you have this print right in front of you. And even if the print was not right in front of you, as soon as someone pronounces the above words, you can picture their spellings – and that greatly assists you.

Yet Kilpatrick says six-year-olds, with none of your advantages, must be able to do this type of oral-only phoneme manipulation BEFORE they’re ready for phonics, that is, BEFORE they’ve begun to study which sounds go with which letters, BEFORE decoding/blending instruction even begins.

How could one devise a more effective plan than this one for keeping children, for entire semesters, at the level of phonemic awareness training without letters? This is a level at which children have yet to begin the process of acquiring the actual skills necessary for competent reading and spelling! I don’t understand how anyone could agree with such an idea.


It's not as if the changes Kilpatrick has made to Ehri's research and writings are inconsequential. On the contrary, they represent a thoroughgoing revision and correction of some of the most important aspects of Ehri's work. Kilpatrick has made OM the opposite of decoding when Ehri says decoding is the primary way OM occurs. He's made OM exclusively an encoding/segmenting activity when Ehri says OM occurs via both decoding/blending AND encoding/segmenting. He's made his "phonemic proficiency" a requirement for OM, when Ehri says the primary phonemic awareness skills required for OM are simple segmentation (for spelling) and simple blending (for reading). He's placed onset-rime at the start of reading instruction, when Ehri has consistently argued against doing so. Last, Kilpatrick's insistence on an astonishingly sophisticated level of phonemic awareness (without letters), even before phonics commences, virtually assures that precious and irreplaceable classroom time will be wasted everywhere Kilpatrick-style phonemic awareness programs are implemented.


Reading reformers and educators have a choice to make. Should they design instructional materials and spend classroom time based on OM principles elaborated by Ehri or by Kilpatrick? The positions of these two individuals, on many important and practical matters, are simply incompatible.

In trying to decide, all stakeholders should keep the following in mind. It was Linnea Ehri who first explicitly identified OM and provided it with its current name. She spent her professional lifetime doing the research, conducting the controlled trials, and publishing her results in peer-reviewed journals for all to examine – 160 such publications to date.

Kilpatrick has not done original research. He’s conducted no randomized controlled trials to test the validity of his conjectures. And the total number of peer-reviewed journal articles he has published is one. (I couldn't find any others. I will happily update this number if anyone points me to other relevant publications.) I encourage David to conduct the research needed to test his claims, and to get that research published. This is especially important in those many areas where he differs with Ehri.


I look forward to the final version of the Clemens et al study. But even this first “working paper” is worth your time. Don’t allow Kilpatrick’s attempt to “nullify” the entire paper (see his quote above) dissuade you from reading it. Along with Mark Seidenberg, I agree it’s a “terrific article, a classic.” I, too, “hope it's widely read and digested.”

Stephen Parker

Boston, February, 2022


Find me on Twitter: @ParkerPhonics

I offer free books and a reading science blog for teachers and interested parents:



bottom of page