That Hypocrite Shanahan -- Cueing Systems vs. Context Analysis

  • context analysis
  • 09 May, 2020

Teacher question:

I attended your recent webinar and you said that students should figure out the meanings of words from context and that they needed to be able to deal with syntax. But I’ve also read that you are against the 3-cueing systems. Isn’t that a contradiction? It seems hypocritical to criticize teachers for teaching 3-cueing and then to turn and around and recommend that they do just that.

Shanahan responds:

Ralph Waldo Emerson wrote that, “Foolish consistency is the hobgoblin of small minds.”

What I said may seem inconsistent, but it would be foolishly so if I had ignored the fact that two distinctly different processes have to be developed in reading —word reading/decoding and reading comprehension. That these two processes have different purposes and operate somewhat differently shouldn’t be beyond the grasp of even the “small minds” among us 

The idea of cueing systems comes from analyses of oral reading errors (or miscues), and a theory of how words are read that simply has not held up to scrutiny. The late Kenneth Goodman examined word reading and found that when words were misread, you could categorize the errors. For example, a student is reading a sentence like:

            The man drove his automobile into the drive.

But instead of saying, automobile, reads he “car.” This error obviously shows no attention to the orthographic/phonological characteristics of the word (its letters and sounds), but car and automobile are both nouns (so they are syntactically similar) and they are synonyms or have similar meanings (which brings in semantics). 

From this, Goodman (1973, p. 9) theorized that a reader collects as little visual information as possible when reading; that he guesses or predicts what is coming based on the semantics and syntax and then “sampling the print to confirm his prediction.” In Goodman’s theory, the best readers minimize the amount of orthographic/phonemic processing that they do and figure out the words as much as possible based on context.

The problem with that theory is that it isn’t right. It turns out to be inconsistent with what we learned about how words are processed during reading. For instance, we know that readers don’t “sample the print” in that way; in fact, studies show that we look at pretty much every letter in a text, including those words that would be highly predictable from context.

Additionally, readers are able to recognize words in about a ¼ second, too fast to allow for the amount of neural processing that would be needed to sample all of these types of information. And, we also know that the best readers are the ones who are proficient with orthographic/ phonological processing, and poor readers are the ones who rely on alternative ways to read the words (Stanovich, 1980).

If the reader could have read “automobile” he would have, but since he couldn’t he used the syntactic and semantic information to make a best guess. (The reader found a work around since he couldn’t really read the word.)

Teaching kids to use these cueing systems to figure out the words is essentially an effort to teach them to read like poor readers. Good readers avoid using anything but the letters and sounds to figure out the words, the poor readers lack this facility so do the best they can otherwise.

Eye movement studies, speed of processing studies, neural processing studies, instructional studies, and so on, all concur. Good readers recognize words by translating letters to phonemes, and poor readers are stuck relying on pictures, and semantic and syntactic contexts to do the best they can under the circumstances.

I do not support the idea of teaching students to read like poor readers, even if this was an interesting and provocative idea in 1965. (And, I’m stunned by people who refuse to change their minds after the accumulation of 55 years of contradictory evidence – talk about “flat-earthers.”)

But reading is not about recognizing words alone. It is also about comprehending and using the information in text.

Reading the words properly enables us to make sense of the message in a text – but that making sense requires additional processing.

That’s why we need to teach phonemic awareness, phonics, and oral reading fluency so thoroughly and so well. We want readers to have automaticity with these; that is, we want them to read the words accurately, but with little conscious attention. This allows readers to devote their cognitive energies to thinking about the ideas in text.

What do we do to comprehend?

One thing comprehenders do is to figure out word meanings. For words we already know, we simply retrieve meanings from long term memory. In other cases, figuring out a word meaning (not the word, but its meaning) may entail the use of a dictionary, guessing based on context, analysis of the morphemes, or asking somebody for help.

Comprehenders also need to make sense of sentence structures and text structures, and to track ideas across a text. They need to bring their prior knowledge about the content to bear on the text, too, and to apply their critical senses to the information (is the information true?).

Word reading needs to be automatic and instantaneous. That’s why you don’t guess words using syntactic and semantic information.

Comprehension, on the other hand is slower and more consciously thoughtful. It requires analysis, reflection, critical thought, and consideration of the language and the content.

My research-based advice is to teach kids both to decode words and to comprehend texts.  Those are different things, they entail different abilities, and therefore sound teaching advice is going to differ for each.

When it comes to word reading, I'm going to teach students to decode. When it comes to figuring out word meanings, i'm going to teach students to use context to make sense of the words (and morphology and references). Just like the research says.

That’s wisdom, not inconsistency!


See what others have to say about this topic.

Sam Bommarito May 09, 2020 05:10 PM

First- you know I deeply respect your opinions. You also know I sometime disagree with them. This is one of those times. Here is a brief excerpt from a blog entry I made on this topic.

D--r r.-s--rch-rs,

Th- r--d-rs o- my bl-g c-n r--d th-s w-th--t th- v-w-ls m-ss-ng. I th-nk th-y -r- us-ng th- thr-- c—s t- h-lp th-m d-c-d- th-s m-ss-g-. Cl-y r-c-mm-nds w- t—ch cr-ssch-ck-ing. C-nt-xt pl-s th- f-rst s--nd.=  cr-ssch-ck-ng.

Most folks can decode this just fine without all the letters. Other examples have given the letters being out of order and folks still able to decode it. On the one hand- teaching the three cues directly to students did not prove to be a worthwhile activity. On the other hand- informing your instruction about what cues the reader does use does pay off handsomely- just had direct experience with that with three students I tutored last year. I have to wonder outloud about how really good readers are able to make sense of things examples like I've just given. Fairly obvious they are using more than the letter information. BTW there are MANY things you have said about the teaching of phonics I agree with. First and formost among them is that not enough time is spent on the direct teaching of phonics. I don't actually fancy myself a flat earther. I rather think of myself as someone who discovered that everything the universe of literacy does not circle around the star we call phonics, Rather it circles around the star we call comprehension. Thanks for considering these remarks (@doctorsam7), Sam from St. Louis

Scott Geisler May 09, 2020 05:37 PM

Oh you hypocrite, always making sense of what some folks struggle to grasp. I appreciate that you don’t spout opinions, just evidence-backed reasoning, when you break the complex into such simple logic. While the universe doesn’t circle around anything, the literacy is centered upon the twin concepts known as decoding and comprehension. And these twin “stars” are interconnected by phonology, fluency and vocabulary because without proficiency in any one of these, one cannot be a highly skilled reader of anything that matters in the building of knowledge. These are truths, not opinions. Well done, Tim.

Kirsten Snook May 09, 2020 05:48 PM

Thank you Tim. A great summary.

There's just one thing that my pedantic mind picked up on in here: "In other cases, figuring out a word meaning (not the word, but its meaning) may entail the use of a dictionary, guessing based on context, analysis of the morphemes, or asking somebody for help."

Re the ref to 'guessing based on context'. Would you say it's more systematic than that? So, deducing or inferring meaning from context? I just worry when I hear practitioners say the word 'guess' to children, as it does sometimes lead to haphazard guesswork. I know that has been much more prevalent when referring to decoding than comprehension, but I do still hear it. Like I say, just my pedantic mind at work here! Would welcome your thoughts.

Loving your blogs.

Donna M Corrigan May 09, 2020 06:56 PM

AMEN!!! Usng context clues when reading should not be for the sake of decoding....but to help with the meaning of unknown words!!
Thank you!!

Sam Bommarito May 09, 2020 07:08 PM

I'll stand corrected by Scott Geisler. I like the idea of the twin stars. Thanks Scott!

Julie May 09, 2020 07:21 PM

What are your thoughts on Share (1995) idea that "partial" decoding will bring a child close to correct pronunciations which can then be refined?

Kim May 09, 2020 10:00 PM

What about early readers who are able to make meaning from predictable texts? Should we not encourage them to use whatever they have (background knowledge, letter/sound relationships, words parts, memory...any source!) so they understand that reading is about gaining meaning? As these early positive experiences compound they build self-efficacy. When we work with these readers we don't blindly accept their approximations - we use it for teaching. We use the error within text to teach more about orthography. *I really think we'd all benefit from a more open way of thinking and adding respect for the work of others into our discussion. Goodman made contributions to our understanding of the complexity of reading, as you are, but we don't have to all subscribe to one, black and white theory. I'm so very tired of the either/or debate. You can think me simple minded if you like but there are strengths to more than one approach - I feel it makes me more responsive to student needs during the reading and writing process. I continue to seek out the varied opinions of various teachers/researchers to enrich my evolving theory of literacy learning.

Julie Dinan May 09, 2020 10:30 PM

The "sin" is this respondent does not understand how to teach reading. There's a balance of using explicit phonics strategies and using context cues. The bottom line is our 3 cueing fan base still needs a lot of work. It's time to check your ego at the door. Our kids are tired of chunky monkey, eagle eye, etc. Yes, I still see these posters in use.

Harriett May 09, 2020 10:32 PM

Thank you so much, Tim, for adding your voice (again) to such an important topic. As you know, this subject has received a lot of attention recently, including Emily Hanford's audio documentary At a Loss for Words: How a flawed idea is teaching millions of kids to be poor readers, but I think your piece concisely captures the take-away message for me when you say: "The problem with that theory is that it isn't right".

Emma May 09, 2020 10:34 PM

Thanks Dr Shanahan,
Sharing with my teacher friends; as usual you clarify and reinforce the evidence.
B/r/i/ll/i/a/n/t :-)
Miss Emma

Tarjinder Gill May 09, 2020 10:49 PM

In terms of 'guessing' based on context - do you mean the difference between row (argue) and row or do you mean a completely unknown word? In which case could you give an example of what you think is reasonable. I ask because guessing as a reading comprehension strategy was something I was expected to teach pupils. While I can understand this as a strategy of less resort in a test, my concern is that it teaches pupils who are learning to comprehend to guess randomly and not really read.

Jackie Kerlalyn May 10, 2020 03:43 AM

Ugh. I’d like to know what you’re selling. Seems to me like we’ve gone back to the Reading First era with this one. While one can argue that strong foundation skills truly lays the foundation for vocabulary development, fluency and ultimately comprehension.

I am truly frustrated with “leading experts” in this field convincing educators that there is only one area of literacy development that lends itself to growing a reader. We all know how that ends up. Why not educate your followers as to how ALL areas of literacy development build upon and work together in literacy acquisition? I saw how you took the whole “complex text” movement and it slaughtered reading for many children and teachers; probably because you were working for some publisher trying to sell their program. I won’t say who but let’s be honest here. Now, because teachers pushed so hard with shoving complex text down the throats of children, we’ve now realized we’ve forgotten to teach phonics and now you’re on that bandwagon. Not surprising.
I challenge you, to use your platform to do a much better job on helping educators bring it all together for readers. But...that doesn’t sell programs now, does it?

Do better my friend.

Jeff Bowers May 10, 2020 10:08 AM

Except not clear that the science agrees with these claims. For example see:

And how is the fast rate of word identification inconsistent with prediction? Prediction is supposed to speed up processing.

And skilled readers also good with morphological processing.

When you write: "Eye movement studies, speed of processing studies, neural processing studies, instructional studies, and so on, all concur. Good readers recognize words by translating letters to phonemes". I thought the evidence that good readers learned to identify words lexically. Isn't that even the claim of strong proponents of phonics who talk about phonology as a self-teacher?

Maria May 10, 2020 01:58 PM

Would love to hear your suggestions on Morphology strategies for secondary students. I have shared ideas on this other colleagues and they are resistant to it. Their biggest issue is that there is not enough time.

Thank you

Susan Board May 10, 2020 06:53 PM

It's clear that Tim's opinions are based on research, not hypocritical at all. Nearly 40 years of reading instruction and observations...I have to say, science nailed it. Thanks for clarifying in the article!

Linda May 10, 2020 08:01 PM

... and the beat goes on.

Timothy E Shanahan May 11, 2020 02:05 PM


That is a silly example and that phenomenon has been adequately addressed (repeatedly), most recently by Mark Seidenburg in Language at the Speed of Sight. Of course, your example isn't an example of reading, and doesn't summarize what readers do when they are actually reading. If it did, then we would find readers skipping letters (and even whole words) when reading... they don't.

This is the problem with the anti-phonics position: it tries to reason from incidents when people are either not reading (how often do you read texts that leave out letters?) or when their reading efforts have failed (focusing on mistakes rather than success).

Finally, haven't you noticed that in 55 years no one has managed to conduct a study in which teaching the cueing systems gave students a learning advantage? Essentially, you aren't teaching reading, you are teaching students to use a workaround instead--it can work (you can get the word by some other means), but it isn't reading. There is an interesting literature on workarounds in different domains; you might want to take a look.


Timothy Shanahan May 11, 2020 02:10 PM


Share (and Patricia Cunningham) are correct as far as I can tell (from experience and the research record) that decoding is an approximating process (at least when it is being learned). One doesn't come up with perfectly accurate pronunciations, but pronunciations close enough to trigger success. However, instruction and reading practice refines this process (Ehri's work) and speed and accuracy improve. Share's work also suggests how readers can develop phonologically without explicit instruction in phonics, too.



Timothy Shanahan May 11, 2020 02:14 PM


Studies of predictable texts show them to be somewhat damaging as they apparently encourage kids to NOT look at the letters in the words. When I was a first-grade teacher, I did use such books (Brown Bear, Brown Bear), but I limited its use severely -- using much more vocabulary controlled text and providing explicit phonics instruction. Reading is a meaning getting/making activity, but it requires you to do so via an author's words (not just the ideas).



Timothy Shanahan May 11, 2020 02:19 PM


Teaching students to guess the word as a "last resort" is what is known as a workaround -- a method for overcoming a problem or limitation by skipping the usual/required steps. If you observe much in classrooms, you'll see teachers trying all kinds of methods to try to get kids to call the right word without actually reading it... "rhymes with..., if you go up in the air, you are in a ______?, pointing at the picture, etc." These efforts may or may not result in the youngster getting the right word, but they aren't reading and they aren't showing students how to make meaning.



Harriett May 11, 2020 10:38 PM

From "How the Science of Reading Informs 21st Century Education", Florida Center for Reading Research, April, 2020:

The three-cueing approach to support early word recognition (i.e., relying on a combination of semantic, syntactic, and graphophonic cues simultaneously to formulate an intelligent hypothesis about a word’s identity) ignores 40 years of overwhelming evidence that orthographic mapping involves the formation of letter-sound connections to bond spelling, pronunciation, and meaning of specific words in memory (see Ehri, 2014). Moreover, relying on alternative cuing systems impedes the building of automatic word-recognition skill that is the hallmark of skilled word reading (Stanovich, 1990; 1991). The English orthography, being both alphabetic-phonemic and morpho-phonemic, clearly privileges the use of various levels of grapheme-phoneme correspondences to read words (Frost, 2012), with rapid context-free word recognition being the process that most clearly distinguishes good from poor readers (Perfetti, 1992; Stanovich, 1980). Guessing at a word amounts to a lost learning trial to help children learn the orthography of the word and thus reduce the need to guess the word in the future (Castles et al., 2018; Share, 1995).

Deb Dunn May 12, 2020 12:57 AM

Having used both phonics and the three cueing system over the years, I'd like to add that there can sometimes be confusion about the three cueing system. I have found that with some students an overabundance of explicit, focused instruction in phonics can leave them comprehension deprived, UNLESS we are ALSO giving them opportunities to real real text and "think about what makes sense", use context clues, etc. to gain meaning. I have had a number of students over rely on phonics (with challenging texts) and under rely on making meaning, leaving them dumbfounded as to what a passage was all about. It is critical to balance both at the same time so there isn't an over reliance on sounding every word/syllable out at the expense of comprehension (otherwise what would be the point of reading?), and and so there isn't an under reliance on phonics which is the tool used for decoding. Each has an explicit purpose, but neither lives in a vacuum.

John Walker May 12, 2020 02:47 PM

How I have enjoyed reading this post, Tim! It's a splendid summary of what is going on when we're reading.
Thank you!

Matt May 12, 2020 05:50 PM

Hi Tim,

Can you respond to Jeff Bower's comment? I'm interested to hear what you have to say in response.

Thank you for your blog!

Matt S.

Timothy Shanahan May 13, 2020 01:59 PM

Jeff and Matt--
Jeff is absolutely right about the study that he cites, but he ignores the other 350+ studies on the topic. I'm not saying that his point is ridiculous -- or even wrong -- but scholars who do such work have lots of concerns about the design of that study and haven't drawn the same conclusions Jeff has (there are reasons to think the undergraduate students may not have really been skipping the words (Pickering & Gambi, 2018)). It is also true that most studies (there is an exception) have not found such word skipping in children to be related to reading ability. But as should be clear, there are a large number of studies of this and they vary greatly in their methods, designs, and outcomes. I think the current consensus on that data would agree with Jeff's view, at least with regard to reading-proficient adults reading easy texts (usually single sentences). They do appear to skip some of the words (though it is also possible that they might not be skipping these but reading them parafoveally or through something called previewing effects).

In any event, my point remains the same. There are no studies showing that teaching students to try to figure out words from syntax and semantics improves their ability to recognize words (though as the blog entry makes clear, there are benefits for such instruction when it comes to things like figuring out the meaning of words).


Kim Sande May 13, 2020 03:47 PM

Kim here again - could you please share some references with me so I can read about the negative impact of predictable text? I think it's such a logical way to support readers who do not yet have enough decoding knowledge. We aren't teaching them to avoid looking at letters - they just need time and experience to develop HOW to look at letters. Using their errors as teaching points gives them experience in looking at letters while cross-checking with meaning. Once I read the studies using predictable text my question will be HOW did they USE it -- I think that good be a big variable - and a key to success or failure. I think we make the most of predictable text and gradually increase the text complexity as they integrate their decoding abilities into their problem solving while reading/writing. I look forward to citations on these studies! Thank you!

Harriett May 13, 2020 04:32 PM

Kim, here's an 11-minute video by a speech pathologist about the problems with predictable text. If you just watch for four minutes from 4:30, you'll see exactly why these texts confirm Tim's statement: "Studies of predictable texts show them to be somewhat damaging as they apparently encourage kids to NOT look at the letters in the words." I've worked with hundreds of struggling readers, and they all come to me frantically searching the pictures for clues as to how to read the words because they have not developed and practiced their decoding skills.

Greg May 14, 2020 03:21 PM

I'm not so convinced that Scott's interesting example detracts from decoding at all. When reading his altered text (i.e., the systematic deletion of vowels), my decoding simply becomes much more laborious and annoying, but I still decode what is there (I have to and then use whatever else in my reading arsenal to make sense of the text ). However, delete the onsets of his words, then my decoding becomes more challenging: _e _ea_ers _f _y _og _an _ead _is (and all the context in the world will only help me marginally). Everything written goes through the lens of the code: rapidly, automatically and remarkably accurately (for good, skilled readers). Degrade the code, you degrade the message. There really is no way around it.

Greg May 14, 2020 03:41 PM

My apologies, Scott. The example of degraded text I referenced was from Sam, not you.

Lynne Raiser May 23, 2020 03:43 PM

Context clues are used for word MEANING not decoding the word. It's that simple.

Mat Jun 17, 2020 03:04 AM


Although guessing words from context is not as efficient as phonic decoding, Kilpatrick (2015, p.38) notes Share's research findings (1995, Phonological recoding and self-teaching) that phonic decoding combined with contextual guessing is more accurate than either strategy alone.

From an instruction point of view, this has left me confused!

Linda Giles Sep 20, 2020 08:42 PM

And I thought something was wrong with me when I thought there was something wrong with the way my grandchild was being taught to read. Then I read this in your article, "Good readers recognize words by translating letters to phonemes, and poor readers are stuck relying on pictures," and realized that I was right. My grandchild had stopped using her decoding skills which I had taught her and started relying on memory and pictures. My daughter had to make her read the sentences backward just to force her to read the words. I know that might sound silly, but she refused to look at the words and was just reciting what she had remembered from class. I decided to watch her eyes and realized she was reading the pictures. She was scoring poorly in reading and had been placed in remedial class. When Covid-19 caused school closures and we were forced to homeschool, her reading level improved to the shock of her teacher and principal. The proof was in the online assessment scores they had received. I am enjoying your articles, Mr. Shanahan. They have been useful to me and on point.

E Morse-Gagne Dec 05, 2021 02:04 PM

@Sam Bommarito--It looks to me as if you're saying that you can tell how advanced readers interpret text by looking at how they solve the problems posed by corrupted text. The one doesn't map onto the other. Yes, it's possible to figure out that "th- r--d-rs -f m- bl-g" means "the readers of my blog." But what you're doing is taking away the main method fluent readers use (the actual letter sequences) and throwing them back upon the resources available to novice readers. And the result is that your "readers" are now operating as puzzle-solvers, not as readers. They are decoding more slowly than usual and their comprehension is altered, because you've shifted them into struggling-reader state.

averbisnotadoingword Mar 12, 2021 11:56 AM

'But reading is not about recognizing words alone. It is also about comprehending and using the information in text'

The mystifying point for the phonics-wary like myself is how a word could even BE recognized if its orthography is irregular ( eg 'put') or if it contains unstressed vowels. Without morpho-syntactic context, how could one even start to distinguish between PROduce and proDUCE?

Either we don't need to process the phonology of the word, and skip straight to meaning, which does not look good for the more extreme phonics models. Or, we do process the phonology of the word, and do so using morpho-syntactic cues, which does not look good for the more extreme phonics models.

What Are your thoughts?

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

Comment *

That Hypocrite Shanahan -- Cueing Systems vs. Context Analysis


One of the world’s premier literacy educators.

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