Timothy Shanahan Points to a Possible Speech-to-Print Advantage

Point taken.

Image by ChatGPT

The Speech-to-Print Advantage

In The Language Literacy Network, Jan Wasowicz (SPELL-Links Learning By Design) has created an infographic that displays the components of four literacy processes: written word recognition, written word production, language comprehension, and language expression. Unlike the strands in Scarborough’s Reading Rope, The Language Literacy Network conceptualizes the functional connectivity of the many reading processes in the brain by depicting the different neural pathways that connect the different language processes.

Important to note is that The Language Literacy Network emphasizes encoding (translating sounds to print) before decoding (translating print to sounds) to reflect what is highlighted as the speech-to-print advantage: a more complete transfer of learning from encoding to decoding with higher quality orthographic representations. In Print-to-Speech or Speech-to-Print? That is the Question Redux Timothy Shanahan leans into this possibility of an advantage to this speech-to-print (S2P) approach and provides an analysis of S2P that is worth exploring and explaining.

First, a clarification. Shanahan writes:

Basically, speech-to-print approaches encourage students initially to perceive their language sounds prior to focusing their attention on the letters and spellings. As children become aware of the phonemic distinctions between words, then these perceived sounds are linked to their visual or print representations.

While this description is true of some speech-to-print approaches, it is not true for those approaches influenced by the Linguistic Phonics Prototype that Diane McGuinness promoted. In Early Reading Instruction: What Science Really Tells Us About How to Teach Reading (2004), she states:

This union of phonemic awareness (PA) with phonics means our students are simultaneously focusing their attention on their language sounds and the letters and spellings representing those sounds. Discussing their 2024 meta-analysis, Florina Erbeli and Marianne Rice emphasize that both PA with and without letters improved phonemic awareness—but it was PA with letters that improved reading and spelling.

Having clarified this point, we can turn to the rest of the article. Shanahan writes:

Comparisons of print-to-speech and speech-to-print programs [is] one especially unfortunate area of neglect. We have some relevant research information about differences between those approaches – just nothing definitive.

A Fresh Start for Phonics?

Shanahan notes that traditional phonics programs have usually emphasized a print-to-speech (P2S) approach with the following steps:

• identify letters

• link sounds to those letters

• sound out words by sounding each letter

He explains that this instructional sequence is identical to the sequence readers must go through during reading: look at the letters and use that information to generate a phonological representation. However, he is quick to clarify an important point:

It seems reasonable to teach students explicitly what we eventually want them to do. That isn’t always the case, however. Learning to read and reading are not the same thing, and one doesn’t necessarily benefit from mirroring the other.

In Teaching reading has its complications, but… Claude Goldenberg reminds us that accessing oral language and accessing print reflect two different processes.

When we see language in written form—print—the signal goes to a different part of the brain than when we hear language in its oral form—speech. When we read, what we had previously only been able to perceive, or access, auditorily now needs to be perceived, accessed, visually. For that to happen, the visual information provided by print must be connected to the oral language system. Such a connection in the brain does not exist at birth.

This distinction between auditory and visual signals may explain Timothy Shanahan’s stated disconnect between learning to read and reading, which seems counterintuitive until we consider that humans as a species began—and humans as individuals begin—speaking long before they began and begin reading. Therefore, reading instruction that recruits what the brain has been primed for might create the optimal conditions that promote efficient recognition of how letters represent speech. Shanahan offers a plausible explanation for this efficiency:

These teaching methods [beginning reading instruction with what the brain has been primed for] encourage much spelling and writing activity – actions that require students to think about what letters represent the language sounds and word pronunciations, rather than solely focusing on sounding out written words.

In addition, before engaging in word-level spelling, the motor act of forming the graphemes (one or more letters) and simultaneously saying the phoneme (sound) as Diane McGuinness advocated, can facilitate letter recognition, automatic retrieval, and the development of this new reading circuit.

The Speech-to-Print Package

This is a good place to stop and consider this distinction between actions that require students to think about what letters represent the language sounds and word pronunciations and those actions that solely focus on sounding out written words. We know that spelling (word recollection) is more difficult than reading (word recognition) as explained by David Share (Spelling as a self-teaching mechanism in orthographic learning, 2008):

Even when spelling highly familiar words, the writer is obliged to retrieve the elements of the visual form of the word whereas reading only requires recognition (Perfetti, 1997). When spelling, furthermore, the writer must process each and every letter. In reading, on the other hand, the orthographic representation may be less than fully specified yet sufficient for word identification, particularly when encountered in meaningful context (Holmes & Carruthers, 1998).

Because reading can occur even if the orthographic representation may be less than fully specified, this may explain the speech-to-print advantage: the brain boost that comes from practicing encoding (writing) in conjunction with decoding (reading). This practice goes beyond written compositions such as the narrative writing found in most kindergarten and first grade classrooms because it emphasizes targeted encoding practice directly linked to reading instruction.

Encoding is the entry point to learning how to decode. While we know that traditional phonics programs do include writing, in S2P writing is an integral part of reading to the extent that dictation often precedes practice with reading decodable text, and this approach teaches students to say the phonemes as they write the graphemes. Spelling is a noisy affair.

Shanahan speculates:

Perhaps the opposite – starting with phonemes and pronunciations and then connecting those to letters and printed words – might be a good idea. It’s possible that trying to spell and write words does more to enhance phonemic awareness and it may somehow make the phonology more prominent or easy to perceive (Wasowicz, 2021).

This good idea to connect phonemes and pronunciations to letters and printed words is, in fact, the overarching principle behind Richard Gentry and Gene Ouellette’s book Brain Words: How the Science of Reading Informs Teaching (2019). They write:

New information from cognition psychology demonstrates how having a deep level of knowledge of words in the brain—including how to hear them, say them, read them, and spell them correctly—turns out to be a very big deal. Yet for too many reading programs, establishing a dictionary of brain words is either missing or attempted through teaching methods that aren’t scientifically supported. Instruction is all too often piecemeal—failing to integrate this brain connection and neglecting the aural component of written language. For decades schools have given instruction in deep word knowledge—including explicit spelling instruction—very low priority while touting phonological awareness alone, phonics only for isolated decoding lessons, fluency and comprehension as separate entities without capitalizing on their essential deeper connection to word reading and spelling.

Tim’s Timeline

Timothy Shanahan provides a quick rundown of S2P approaches over the last century which helps situate where the approach has been and where it is going.

• The possibility that kids might benefit more from speech-to-print than from print-to-speech first started to appear in the 1910s – in the work of Ernest Horn (1919). He explored the idea that spelling instruction and practice might benefit reading.

• Later, this notion gained purchase when Helen Murphy (1943) discovered that the ability to auditorily discriminate the phonemes in words was a powerful predictor of later reading progress, and she and her advisor, Donald Durrell, built this concept into their beginning reading program (Durrell, Sullivan & Murphy, 1945), a program that eventually morphed into Speech to Print Phonics (Durrell & Murphy, 1964).

• When Jeanne Chall (1967) published her landmark review of phonics research, she concluded that those programs that included spelling, writing, and/dictation did better than those that did not.

• I followed up on that in my reading-writing relationship research in the 1980s. I found spelling and decoding to be closely related, even when a lot of other variables were available to suck up the variance (Shanahan, 1984).

• Marilyn Adams (1990) and Linnea Ehri (1997) both theorized on that possibility.

• Later, Ginger Berninger and her colleagues improved on that with even more ambitious efforts with the same results (Berninger, Abbott, Abbott, Graham, & Richards, 2002).

• Steve Graham reported a meta-analysis that found spelling instruction improved reading – most likely because of its contribution to decoding (Graham & Santangelo, 2014).

Shanahan concludes:

But that evidence is all indirect. It suggests the value of the approach. It doesn’t prove it.

This point bears repeating and emphasizing: It suggests the value of the approach.

Given current concerns about the ineffectiveness and inefficiency of some phonics approaches, a suggestion is a promising start—even more promising when we realize that left out of this timeline are the books by Diane McGuinness, beginning with Why Our Children Can’t Read and What We Can Do About It: A Scientific Revolution in Reading (1997). Those of us who practice her S2P approach called Linguistic Phonics have been using it with success for many years.

According to McGuinness, there are a few basic elements necessary for a beginning reading/spelling program, which she labels a Prototype for Teaching the English Alphabet Code. I’m listing the elements of this prototype below. In the parentheses following each one, I’ve noted one or more researchers/educators whose publications over the past two decades have aligned in various ways (if not completely) with these recommendations.

Note in the first bullet point that what Diane McGuinness is referring to as sight words are high-frequency words that are memorized as a whole—not words that become recognized by sight after they have been orthographically mapped by connecting phonology, orthography, and semantics (sounds, spellings, and meaning)—which include all words, not just those that are high-frequency. And in the second bullet point, clarification is also needed. It is important to separate teaching the sounds of letters as part of reading instruction and teaching the names of letters for use as labels. (Bolded words were originally italicized in the book.)

• No sight words (except high-frequency words with rare spellings). (Linnea Ehri)

• No letter names. (Stanislas Dehaene)

• Sound-to-print orientation. Phonemes, not letters, are the basis for the code. (Richard Gentry and Gene Ouellette)

• Teach phonemes only and no other sound units. (Susan Brady)

• Begin with an artificial transparent alphabet or basic code: a one-to-one correspondence between 40 phonemes and their most common spelling. (David Share)

• Teach children to identify and sequence sounds in real words by segmenting and blending, using letters. (Susan Brady)

• Teach children how to write each letter. Integrate writing into every lesson. (Steve Graham)

• Link writing, spelling and reading to ensure children learn that the alphabet is a code, and that the code works in both directions: encoding/decoding. (Linnea Ehri)

• Spelling should be accurate or, at a minimum, phonetically accurate (all things within reason). (Gene Ouellette and Monique Sénéchal)

• Lessons should move on to include the advanced spelling code. (Louisa Moats)

Shanahan’s cautious optimism:

There is a preponderance of evidence on that side of the argument, but I’m not satisfied that it is yet proven.

When Is ‘Close’ Close Enough?

Speech-to-Print advocates like me have been satisfied for a long time despite lack of definitive proof because we see how much more efficient this approach is, even if it is only equally effective as the best P2S approaches. Efficiency really matters because of the time constraints in the classroom as discussed in The Science of Reading Meets the Science of Learning. For example, the speech-to-print approach teaches flexible pronunciations rather than syllable types to help students decode multisyllabic words. Note the conclusion about efficiency and effectiveness in this study.

It was more important to teach students to form complete graphosyllabic connections between spellings and pronunciations within words than to teach them to apply rules for locating syllable junctures in words. The finding that our form of syllable instruction proved effective, whereas the rigid approach did not, supports this view. Other experiments examining the effectiveness of syllabication instruction have yielded positive outcomes on students’ word-reading skill . . . However, the time spent teaching students was more extensive in these studies than in the present study (Bhattacharya & Ehri, 2004).

Steve Dykstra cautions that a lack of bullseye evidence for a particular instructional method shouldn’t automatically rule out that method if it can be part of making reasonable decisions that we believe will benefit our students. He says:

We are not required to be right; we are required to have ‘right’ reasons that we can explain to other professionals . . . Know where the science ends and your best judgment begins.

In the comments section of Shanahan’s blog, S2P practitioner Jennifer Newman offers the following support for this approach:

• McGuinness, McGuinness, & McGuinness (1996) reported early findings from the development of Phono-Graphix.

• Denton et al. (2006) evaluated an intensive intervention that included Phono-Graphix and found meaningful gains.

• Endress et al. (2007) published a program evaluation of Phono-Graphix with students with disabilities, again showing promising results.

• Simos et al. (2007) used brain imaging to show that phoneme-based instruction like Phono-Graphix shifted activation to typical language regions in struggling readers.

• In Essentials of Assessing, Preventing, and Overcoming Reading Difficulties (2015), David Kilpatrick reviewed intervention research and noted that traditional phonics programs tended to yield only moderate standard score gains, while programs like Phono-Graphix showed some of the largest—suggesting they deserve further investigation.

• Case, Philpot, & Walker (2023) conducted a longitudinal study on the Sounds-Write program, showing strong outcomes across multiple years.

Proof? Perhaps not. But Timothy Shanahan has conceded that there is a preponderance of evidence. He notes:

Perhaps the closest thing we have to a direct test of the proposition is a meta-analysis of 11 studies (Weisler & Mathes, 2011). It concluded that instruction that integrated encoding into decoding instruction led to significantly higher reading achievement. Still not the strongest evidence – because it combined investigations that compared encoding with decoding (Christensen & Bowey, 2005) along with those that compared encoding instruction with things like extra math lessons (Graham, Harris, & Chorzempa, 2002).

Until there is conclusive evidence, he says it’s fair to surmise that there are real benefits to be derived from activities like:

• explicit early instruction in phonemic awareness [I would add “with letters”]

• invented spelling encouragement

• explicit spelling instruction

• word construction from sounds

  These kinds of speech-to-print activities increase learning . . . Get a phonics program that includes such activities or layer them into a traditional print-to-speech program.

Noted. Speech-to-print is showing promise. Let’s lean into these components as a first step. Then, going forward, we can drill down into other possible low-investment/high-yield practices:

• introducing the ‘advanced’ spelling patterns as soon as possible

• making the best use of decodable books while also integrating a range of text types

• facilitating multisyllabic word construction early on

• integrating phonemic awareness and phonics with writing

Finally—and most important—let’s cultivate curiosity and show humility as we examine the best reading instruction methods available. Let’s be highly selective and choose methods that align with best practices regardless of their origin, freeing decision-making from trends and traditions that may not be serving our students in the best possible ways. Linguistic Phonics points to a possible path forward.

Spoken language really matters, as Trina Spencer (Story Champs) reminds us:

The way to teach writing is to teach speaking; children do not write what they cannot say.

Those of us who are speech-to-print proponents would also say that the way to teach reading is to teach speaking. Children need to unite phonology, orthography, and meaning to facilitate the orthographic mapping that is necessary for automatic word recognition. And speech brings built-in phonology: speak to read. Is that part of Timothy Shanahan’s point?

In a recent post in the SPELL-Talk listserv, research neuropsychologist Jeannine Herron (Making Speech Visible) is emphatic when it comes to the importance of speech:

My point about barking up the wrong phonological tree is simply this—letters stand for sounds we SAY. So let's help children pay attention to SAYING the sounds, not just hearing them. If children learn that letters stand for any sound they say, they soon realize that they can write any word they want to write—whether those words are single or multi-syllable, whether they are monomorphic or polymorphic.

Whatever it takes—including turning to that tired term, paradigm shift—let’s embrace the best methods to teach reading both effectively and efficiently. There is too much tribalism and not enough trial and error in considering new and/or repurposed approaches to reading instruction.

Let’s make changes that really make a difference. That’s the point.

Always striving to make sense—please let me know when I fail.