It all started with a scared little boy who spoke no English, crying on his first day of school.
Norbert Rennert was six years old. His parents, who were refugees from World War II in Eastern Europe, had just moved to Canada by way of Paraguay. “I remember that experience quite vividly,” Norbert says, “So I know what it’s like going to school and being confronted with a language you don’t understand.”
It was a key moment that would lead him, eventually, to create a language program called “SynPhony,” which is currently alleviating the fears of children all over the world.
The next key moment came with his daughter.
In 1994, Norbert moved to Suriname with his wife and four children, serving SIL International by working on a New Testament revision in the main Creole language of that country. While there, they discovered that their third daughter had some hearing issues that contributed to her delayed language development.
When they moved back to Canada after ten years they were able to get her a set of hearing aids. They hoped that the hearing aids (and access to the education system of Canada) would bring her up to speed. But despite their daughter’s consistent effort, her reading level lagged several years behind where it should have been.
The Rennerts found a program claiming to be able to teach their daughter to read in a few months, which seemed to them to be an outlandish claim. But they were desperate, so they went for it.
Three days of the week, Norbert did exercises with his daughter. With his background as a linguist, he was able to look at the lessons and see the logic behind them.
The program started her with nonsense words.
Norbert told his daughter, “You see these words? They’re nonsense. They don’t mean anything. So all you have to do is sound out the letters.”
Norbert points out that this strategy “takes a load off the mind of a poor reader, because they don’t have to figure out the meaning. No one will judge them for saying nonsense, because the words are nonsense. So all they have to do is sound out the letters.”
He and his daughter sounded out the words, which were designed so that they would use one sound (the most common pronunciation) consistently for each letter. When she was able to sound out the nonsense words, the lesson switched to real English words—but with the key criteria that all the pronunciations of those letters were the same as the pronunciations used with the nonsense words.
His daughter’s accuracy went up to 90 percent and then, with practice, to 100 percent. Just a week and a little bit more of practicing nonsense words, and a reader was born before his eyes. After that, it was just a matter of introducing new spelling patterns and practicing, always building on what had come before. So it never used spelling patterns that were outside of her reading ability.
"We covered over 150 spelling patterns,” Norbert said, “And the academic in me started to wonder how many spelling patterns there actually were."
That was another key moment.
When he had finished tutoring his daughter, Norbert had a vision: “What if we could make a computer program that could handle the complexity of English, but would make this educational program even more flexible? A paper based product has one sequence,” he realized. “A computer program could do this dynamically, and you could come up with almost any sequence.”
It occurred to him that before a computer program could happen, he needed to analyze a bunch of words according to their phonemic and orthographic form, and then do an inventory. “I analyzed over forty-four thousand words exhaustively into their grapheme/phoneme correspondences,” Norbert says. “Many dictionaries include a pronunciation (or phonetic) form of the word for each entry, but what I did was merge the two.”
Norbert points out that English has 26 letters, but 44 sounds. It solves that disparity in a lot of ways, but in doing so creates a lot of complexity. He ended up finding over 450 spelling patterns in the English language. For example, “ea” can be used in the words “leaf,” “great,” and “head” to make three completely different sounds. In his system, those would be three different spelling patterns, because they map onto three different phonemes
Norbert thought, “I work in an organization that works with languages. What if I were to make a computer program that could work with any language?”
There was only one problem: Norbert Rennert did not know how to program computers.
He took the idea to his supervisor and his supervisor said, “Nah. It’s too big a project.”
Determined, Rennert spent the next two years learning to program, and created a proof of concept that he could use on a language as complex as English.
From the beginning, though, he made sure the program was language-agnostic, so that none of the language-specific features were hard-coded. By maintaining the language data as a separate component, he was free to create other language databases and apply the software to them as well. Now his program can handle any language in the world, with the exception of Egyptian and Mayan hieroglyphics, and Chinese characters (which are symbolic languages that can’t be broken into their component sounds).
For all other languages, SynPhony (a portmanteau of “synthetic” and “phonics”) can search a database in any language, and with his continued development has become a sort of “Swiss Army knife” of a language program that can be used to analyze languages, to create word lists and decodable sentences (sentences that children can understand based on the letters and words they’ve already learned), and ultimately to provide quality education materials in the mother tongues of students around the world.
Norbert recently traveled with SIL LEAD to Maputo in the south of Mozambique, where he met with over thirty participants working in fourteen different languages, supporting the goal of creating educational materials that follow one, key principle: “That we don’t ask children to read something they’ve never been taught how to read—which requires paying a lot of attention to the letters they’ve learned, and finding words that consist only of those letters.”
This approach to education uses language in a dynamic and adaptive way, allowing educators to meet children exactly where they are. No longer do they have to attempt to force a square peg into a round hole. SynPhony is a search engine for literacy that can match the words and texts of any language to the reading ability of any student.
It’s is a computer program, but it’s also the culmination of the story of a little boy crying at school over an incomprehensible new language.
Norbert Rennert created SynPhony because he saw a need and wanted to help.
One day you, Norbert Rennert, and the rest of us will all be gone from this world. We can’t know what our greatest impact or legacy will be—whether it will be a project we focused on for years, or a passing comment to a friend—but we can hope that whatever it is will be something that helps… that serves the greater good.
For children all over the world, SynPhony is Norbert Rennert’s legacy of service.