If you have even a passing knowledge of Spanish, you know the ns in montaña are different. You can see it. But English isn’t like that. The as in car and care and race look the same, but sound different. But what if we fix that? Several schools in the US are trying a new approach to teaching reading that directly addresses the biggest challenge of learning to read English— English itself.
Many a third-grade English teacher will have a mug on their desk that reads “I before E except after C, unless it’s my feisty neighbor Keith who lifts weights. Weird.” That last word sums it up nicely. English, as a language, is truly weird. The impact is devastating.
English stands alone as the hardest major language to learn to read. The relationship between letters and sounds is just too inconsistent. There are too many exceptions. Every rule can be broken and is broken, often, by the most common words.
This is science. Languages have innate characteristics, and graphed on any scale, English stands alone. Dr. David Share, one of the most cited authors in the science of reading and cross-linguistic research, calls it an “outlier” orthography. It’s so different, and in such wide use, that it wildly distorts the very understanding of how the other 95% of humanity learns to read. English’s complexity drags scholars into an all-out war over how to teach it. It’s an important war, but overall progress against the real enemy is at a stalemate: literacy scores in the US in the last 30 years have not improved. In 2019, before the pandemic and despite all the advances we have made, they declined.
This is a uniquely English problem.
Not a single empirical study has found that English is learned faster and more easily than any other alphabetic language. We spend far more to teach it, with worse results. For English readers in first grade, studies consistently show pseudo-word reading error rates of 40–80%. Corresponding studies in Finnish, Turkish, Italian, Greek, German, Dutch, and Portuguese samples show error rates below 25%.
In the classroom, the conclusion couldn’t be more obvious. A student that spends equal time learning to read English and Spanish will universally achieve faster oral reading proficiency in Spanish. Emma Garcia, a former dual language teacher from Virginia, sums it up: “I had countless hours of instruction and guidance on how to teach my kids to read in English. I had just about none in Spanish, but all of my students could sound out words more easily and accurately in Spanish. It came naturally to them. It’s really easy to teach. It is just a much easier language.”
Even more telling: if you take an L1 English speaker (whose first language is English) and introduce them to L2 Hebrew, L2 reading accuracy will be higher in first grade than L1 reading accuracy in English in fourth grade. To state it simply, an average American kid in a dual language program is more likely to read the foreign language out loud with better accuracy in a single year of instruction than they can English, their native language, after four years of schooling.
This is all the more remarkable because dual language studies control for socioeconomics. The systemic challenges of low-income communities don’t impact the findings above, since the same exact student is learning both their L1 and L2 language.
It is easy to understand that languages like Spanish are easier to learn. But what about languages without alphabets? What about Mandarin Chinese?
English is not alone in having a complex relationship between the written and spoken word. Chinese and the related orthographies of Asia are graphic. Chinese characters are used to represent a word or part of a word. If a student doesn’t know a particular character, they can’t read that word aloud. There is no phonetic information the reader can draw on. They are stuck.
In 1958, this problem was fixed. Zhou Youguang invented Pinyin, a romanized alphabet for “spelling” Chinese words. This has completely transformed how billions of students, native and foreign, have learned Chinese. Pinyin includes six simple vowels, 29 compound vowels, 23 consonants, and the four tones. It’s a comprehensive, alternate way to write words so they can be sounded out and taught to every single language learner, kid or adult. It is structured, systematic, regular, and, as a result, extremely powerful. Armed with this arsenal, a student can attack and successfully sound out every word they see.
Why does this matter? In Spanish (which is regular) or in Chinese languages (which have Pinyin as a support), the language, with a bit of practice blending, enables every single word to be decoded. If a word is already known by the student, the decoded word will match it. If the word is not in their vocabulary, the student can be confident in their pronunciation of the new word and perhaps learn the meaning from the context.
Contrast this to English. Virtually any passage of English will include numerous nondecodable words. Knowing simple letter sounds does not suffice. “The dog was not afraid” would be sounded out “T-h-eh d-ah-g w-ah-ss n-ah-t ae-f-r-aeih-d.” When students attempt to decode in English, the sounds can’t be easily mapped to the correct word, even if this word is known to them aurally. If the word is not in their vocabulary, the likelihood of correct pronunciation on first attempt by the novice reader is minuscule. This is the heart of what makes English so hard to teach and so hard to learn.
The success of Pinyin and these ideas on writing systems sparked similar approaches across the world, including for English. One of the most important efforts here was the Initial Teaching Alphabet (i.t.a.). Similar to Pinyin, the i.t.a. created a simplified, regular teaching alphabet and used this alphabet to write words so that they were easy to read. It accepted the difficulty of English spelling and met the challenge. And the results were extremely promising.
Under timed tests, students were reading 336% more words with the i.t.a. With a few years of study, students were reading books three grade levels higher than the non-i.t.a. control groups. The system was so successful it was deployed for 70,000 students across the US.
Whether looking at the rates of progress for kids learning Spanish, examining the mechanisms used to learn Chinese, or resurfacing early English experiments with alternate orthography, a critical maxim of language learning becomes clear—how a language is written is the primary factor to how easily we can learn to read it.
But the i.t.a. ultimately failed. There was a problem. It didn’t transfer.
Kids who learned to read with the i.t.a. could read in that system extremely well, but when it came to transition back to “regular” reading, the results didn’t stick. The i.t.a. was too different from standard English. The approach of using a simplified, regular alphabet was correct, but the particular implementation failed because it didn’t prepare students for success with English as it is currently written.
While the creators of the i.t.a. consciously avoided labeling it as respelling, the idea that words were respelled was a key challenge the approach failed to overcome. It was not the first to stumble on this rocky ground. Luminaries throughout the centuries have, rather unsuccessfully, fought for change. Benjamin Franklin, Theodore Roosevelt, Brigham Young, Mark Twain, and Andrew Carnegie all championed various forms of spelling reform in their time. Carnegie went as far as to say that standardizing and simplifying English would lead to world peace, donating a portion of his fortune to the cause.
Ultimately, when pressed, English readers have chosen to keep things as they are. Among all the systemic challenges that face our country, none seems more obvious than this. The English-reading world, dominated by those with the education and resources to master such a challenging language, has chosen to preserve etymology and the status quo at the direct expense of making the language easier to learn.
The failure of the i.t.a. and spelling reform both add evidence to support another maxim of language learning— spelling is set in stone. This is not to say it cannot change. Stones erode over time. Great storms cause entire cliff faces to detach (as happened when Pinyin was introduced). But, typically, change is slow. Very, very slow. And the sort of wholesale spelling change needed to make English regular failed in times far less contentious than the one we live in. It isn’t happening.
While grand change is off the table, a more modest intervention can prove just as effective, and thankfully, history provides a highly relevant case study. Centuries ago, the languages of the Middle East encountered the same exact challenge.
English and Hebrew have very little in common linguistically. They evolved on divergent paths. They are not even read in the same direction. And yet one trait co-evolved in both—letters in both languages evolved to support various sounds. Reading the newspaper in Israel today, one would come across a letter that, as written, is phonetically ambiguous to a novice reader.
If the same word appeared in a text for children, however, it would be “coded.” Hebrew, when presented in religious texts or for language learners, uses diacritical marks above and below the letters to indicate specific sounds in specific words. These marks are learned quickly and easily applied, as any colleague who has had a Bar or Bat Mitzvah will attest.
Like Pinyin in Chinese languages, these markings provide novice readers a simple, consistent way to sound out unfamiliar words. They reinforce correct pronunciation, a critical feature when transcribing a religious text (and quite important for any speaker of a language). As with Pinyin, this supplemental orthography enables independent reading. Words within the reader’s vocabulary are understood. Words outside the reader’s vocabulary are at least confidently, accurately pronounced, and may be learned through context.
Where the i.t.a. transformed spellings to improve reading speed, languages like Hebrew augment spelling with visual cues for pronunciation. This enables rapid onset of oral reading fluency, without the drawbacks inherent to respelling. Students come to learn the base orthography of the word and then, when presented with the same word without the diacritic support, effortlessly recall the word. It is no longer unfamiliar and can be read on sight. As this sight-reading ability is achieved, the scaffolding is removed. Every single reader of Hebrew has learned to read this way.
It is critical proof for a third, and final, maxim of learning a language—words learned to be read on sight with scaffolding don’t need scaffolding to be read on sight.
Taken together, these three maxims point toward a compelling, simple solution for dramatically improving the ability to teach reading in English. Since the rate of learning of the language is a direct function of the complexity of the orthography, and since spelling as a whole can’t change, a diacritic system is needed. Such a system allows students to learn to read in English as fast and accurately as they can in simpler languages. And, following the Hebrew example, that progress would be transferable to reading without the supports. In practice, that’s exactly what the early data is showing.
In 2021, several schools across the US broke new ground to adopt a literacy program that had been developed with these challenges in mind. The program, called TIPS, uses diacritical marks to help beginning readers sound out words. Crucially, their technology includes these TIPS in the material only for the first few instances of the word in a particular book, after which the TIPS are removed and the student reads the word on sight. Teachers also receive access to digital teaching tools, lesson plans, and a library of content that includes the diacritics.
The results of the intervention were astonishing. At one school, a group of third graders gained a full year of reading skill in ten weeks, five times more progress than they had made in prior periods. At another, a group of special needs students increased word-reading skills by 58% over the same time frame, despite years of alternate interventions with limited gains. At another, the school improved overall reading scores by six percentile ranks vs. a comparison group without the intervention that declined by one point. Yet another achieved over 300% faster progress than the normal expectation for ELL students. At this school, every single student in the intervention group was categorized in the high-risk zone for DIBELS, and within ten weeks not a single student was in this category.
While the work above has yet to be published and does not yet include randomized controls, it lines up with the most obvious conclusion of the science of reading—a language that is easy to read will be easier to learn.
Zachary Silverzweig is a mission-driven entrepreneur, a technology leader, and the innovator behind TinyIvy, which develops literacy tools for schools based on the TIPS system.
He holds master’s and bachelor’s degrees from Columbia University and lives in New York with his wife and two children.
