More on teaching phonemes

This post is part of an exchange about phonemic awareness that originated on the SPELLtalk listserve. It's awkward having it in two places, but the posts may be hard to find on SPELLtalk. Here I'm responding to a reply from the original poster, Dr. Herron. Her post is here.

I don’t know if we agree about phonics, but let’s get back to phonemic awareness. Thanks for clarifying that you don’t advocate teaching 44 phonemes as a prerequisite skill. I had in mind the many people who are doing this. It’s a practice that is based on a deep misunderstanding of phonemes, phonemic awareness, and reading, which I’m trying to explain. As I ask myself periodically, why am I doing this? Because practices based on a misunderstanding are likely to be ineffective. Because I think many teachers would want to know if what they’ve been told to teach has a valid foundation or not. Finally, because sharing this knowledge may encourage people–including publishers–to pursue instructional practices that have better empirical and theoretical foundations. 

You ask if you (and other educators) are using the term “phoneme” wrong, and maybe this sets the linguists off, so perhaps we should use another term? You are using the term incorrectly. It would indeed astound my phonologist colleagues to learn that in the “science of reading,” people teach up to 44 phonemes as though they are units of speech with “correct” pronunciations. 

And now, having said this, many readers will conclude, “Oh. This is just a quibble about terms. We can use another term, sure. End of problem.”

The name doesn’t matter but what’s being taught does. The rationale for teaching children the phonemes of a language is based on a mistaken assumption: that phonemes are discrete units with correct pronunciations. If the assumption is false, is there any reason to engage in this kind of instruction? Perhaps the answer is yes, but the reasons will need to be specified.

What’s being taught in such activities is a version of what I know as “letter sounds”: the conventional sounds that are associated with letters in teaching the alphabet. Such as “bǝ” (the syllable with the “b” sound in the context of a schwa vowel) as the sound for B. As I noted in the previous post, some authorities recommend saying the syllable without the schwa to pronounce the initial “phoneme” correctly, but that’s phonetically disastrous: it mangles the “b”, leaving a very unnatural sound behind. I don’t see any rationale for teaching these artificial sounds, or shaping one’s own mouth or a child’s in order to produce them. Perhaps others can articulate (so to  speak!) the logic behind these activities.* 

Whereas I do not see any practical, empirical, or theoretical basis for teaching these weird “phonemes,” the evidence is very strong that letter sounds are important. They are essential for getting sounding out words–reading aloud–off the ground. The standard sounding out activity plants the idea of treating words as if they consist of spoken parts, which correspond to graphemes. The activity has a lot of other benefits, of course. The main point is that the letter sounds are only a tool, not the object of extensive instruction. They are there to bootstrap reading aloud. I’m not going to lecture people on SPELLtalk about how to teach phonics. But, there are ways to transition smoothly from knowing the letter sounds to using some of them to sound out words letter by letter to sounding out words using bigger chunks that have natural pronunciations. 

This treatment of the letter sounds in early reading instruction is fundamentally different from the treatment of “phonemes” in a lot of PA instruction. The letter sounds are only important for how they are used in a real reading task, reading aloud. They aren’t the building blocks of spoken words. The child doesn’t have to perform at a high level on “letter sound awareness” tasks. The focus is entirely on reading: sounding words out, blending sounds, and so on. The letter sounds are like a booster rocket that drops away once the child’s reading aloud gets off the ground. At that point they can learn from pairings of spellings and the actual pronunciations of words and parts of words (e.g., rimes). The ability to treat spoken words as if they consist of segments develops out of gaining skill at these realistic tasks. Gaining this type of phonemic awareness then facilitates getting even better at reading words. It’s a feedback loop.

These observations may seem pretty opaque but could be spelled out in a future post. I think that quite a lot changes when teachers shift from teaching phonemes to teaching reading, using the letter sounds as a tool.

Obviously I’m questioning some pretty widely held beliefs here. I understand that people may not be ready to drop doing these activities because some guy on the Internet said they were problematic. I am trying to seed future advances. There will eventually be people who develop new curricula and other materials that aren’t subject to the same concerns, one hopes. 

Your other concerns:

> I think this sentence doesn't make sense to teachers: "Spoken words do not consist of spoken phonemes.  (They are not, as you said, "individual sounds you make with your mouth when you pronounce a word.")" 

An example may help. We don’t say the word “car” by speaking a sequence of three sounds (speech isn't beads on a string). However, it’s easy to treat the word as if it consists of three sounds. It’s hard not to because we read an alphabet. Using a term other than “phoneme” doesn’t change these basic facts. These concepts may be hard to accept if you were taught a different story, which you’ve incorporated in your work. It’s disheartening, for sure. As a scientist I’ve sometimes had to modify beliefs based on new evidence. It’s not easy, but what’s the alternative? Continue with mistaken ideas because they are familiar? 

If I haven’t addressed your concern, you could spell out what exactly doesn’t “make sense” about this stuff?

Finally, 

> Why do we have to say "as if they consist of segments"? 

The issue gets back to what we should be teaching, the pairing of letter sounds with printed words. In the sounding out of “bat”, the teacher and child are treating the sounded out word as if they consist of three discrete segments/phonemes/speech sounds. That’s an illusion but essential for binding speech and print. If instead we say that there really are phonemes in spoken language, this view leads to the unnecessary PA training of phonemes in the absence of print in pre-readers, the arguments over exactly what’s the right way to train PA, and other activities that are delaying the actual goal, which is getting the child to pair sound and print early and often. 

>Do we have to be so vague and circuitous with teachers? 

The fact that phonemes are abstract doesn’t make this account “vague and circuitous”. I think that finding out that phonemes are an illusion/abstraction is upsetting because many teachers have been taught that they really do exist in speech, and have been encouraged to teach them.  As I’ve said, this approach would have been entirely sensible if spoken language consisted of discrete phonemes. It’s important to correct this misperception because it leads to undesirable practices.  

> Do spoken words consist of spoken segments (whatever you call them) or not?  

No. The feeling that we can hear or produce individual segments/phonemes/discrete speech sounds comes from experience with an alphabetic language. It’s one of many illusions in our perceptual world. If you’re interested, this website from the American Museum of Natural History shows visual illusions. Especially relevant are the illusions where our brain fills in something that is not there. As the first page of the website says, “your brain does not simply receive this information [from your senses]. It creates your perception of the world.” Our illusion of phonemes in speech is the brain filling in something that is not there.

>If they don't consist of segments, then what do letters stand for? 

I don’t think this is the right question. The goal isn’t teaching the child “what letters stand for”. It’s teaching them how to read and spell words containing letters–A, for example. The child doesn’t need to be armed with all the ways that A’s are used in written English. They need to learn how the As function in combination with other letters in actual words. That is phonics: learning about patterns in print that correspond to patterns in pronunciation. 

The short answer to your question, though, is that graphemes represent phonemes. Phonemes represent the sounds of words at an abstract level that is like a mashup of print and sound.  This phonemic level omits a lot of information about how words are actually pronounced, and it is more segmental than real speech (because it has been shaped by spelling). This phonemic representation is used to generate an actual pronunciation, in reading aloud. In silent reading, it is also involved in gaining access to other information associated with a word. 

* Interventions focused on articulation may be relevant to individuals with speech impairments, whose needs are quite different than typically developing folk