What should I say to someone who claims “phonics doesn’t work”?

Sometimes people try to tell me that “phonics doesn’t work”.

Written words are not exact matches for spoken words, but this doesn't mean that "phonics doesn't work".

The process of reading is about taking an imperfect written representation of a word and "fuzzy matching" it to an actual word that you know in your head.

Some of these imperfect written representations are so close to the spoken word that we don't even notice they're different.

For example, "cats" and "dogs" have two different "s" sounds. (The "s" in dogs switches to a "z".)

"Stop" and "top" have two different "t" sounds. (Put your hand in front of your mouth to feel the puff of air from the aspirated "t" in "top")

"Cat" and "can" have two different "a" sounds. (It's subtle, but you can hear the nasalized difference if you compare "plan it" vs "planet")

The fact that written words don't precisely capture our spoken language doesn't mean that "phonics doesn't work."

But being aware (or unaware) of these nuances does have a large impact on teaching effectiveness. For example, the popular app Reading Eggs introduces the sounds “a” and “s”, then has the student combine them in “as”. But that “s” sound just changed to a “z”!

Kids will need to learn the word “as” eventually, but if you’re trying to teach the concept of blending two sounds, it’s a bad idea to use an example where one of those sounds changes!

In Mentava, you’ll notice us being extremely precise with our pronunciations when introducing the concept of blending. A very common mistake that everyone else makes is using “a” and “m” to teach blending, but nasaling the “a” to say “am” instead of “aaaaaaamm” when blending them together.

So when teaching phonics, we begin with a simple model where there’s a 1-1 correspondence between sounds and letters, and as students become more advanced they learn to handle increasingly imprecise word representations.

This doesn’t mean that phonics doesn’t work, it just means letting go of the false assumption that every written word must be an exact match for a spoken word.

Written words *can* be used to represent spoken words precisely - if you're willing to learn the 107 letters, 31 diacritics, and 17 additional signs of the international phonetic alphabet.

But even in English, it's much easier to learn the 26 letters of our alphabet and accept that the written forms of words are actually just really good hints for spoken words.

In English, most of these hints are good. Decoding simple written words like "cat", "sat", "hop", etc is straightforward.

If you learn a few more unusual patterns, you can easily decode slightly less good hints like "high", "sight", "other", "brother".

And even some of the most irregular written representations (like "said", "colonel", or "rendezvous") aren't completely terrible hints.

It's not like we're out here pronouncing "would" as "gadumbal". The phonetic pronunciation of the word still gets you 90% of the way to realizing what spoken word it represents.

Despite what critics say, non-standard words like "rendezvous" or "colonel" or even "said" are not failures of phonics.

It's just the way our writing system has evolved, and understanding the phonetic basis (with software like Mentava) gives kids the best possible tool to decode it.