Phonemic Awareness and Phonics Instruction Are NOT the Way We Learn to Read

    In Language Magazine's April 2021 issue, Stephen Krashen presents an essay in honor of Kenneth Goodman*

    An appropriate way to honor Kenneth Goodman is to describe a presentation I gave recently as a guest lecturer in a class at the University of Southern California. I began by telling the students (mostly first-year grad students) that I wanted to teach them a new word, one that I use in conversation all the time because it has so much snob appeal: zeitgeist. Zeitgeist is German for spirit (Geist) of the times (Zeit). It refers to an idea that is widely accepted and representative of a large number of people in a profession. The zeitgeist in reading education is, of course, phonics, and the current belief that children don’t read well because they don’t know phonics. Phonics dominates the zeitgeist, but the zeitgeist also includes phonics’ partner in crime, phonemic awareness.

    I presented the case AGAINST phonemic awareness and the case AGAINST systematic phonics instruction, then discussed how language acquisition and literacy development happen and why prediction is the core of both. And yes, why they use the same process, quoting Ken Goodman’s phrase “learning to read is natural” and why his term the “psycholinguistic guessing game” is so accurate.

    Supporters of phonics claim that learning to read is not natural and it must be taught. They assume that in order for children to learn to read we must first teach them that words are made up of separate sounds: we need to teach them “phonemic awareness.” Phonemic awareness instruction is of two kinds: segmentation, knowing how to break up a word into its component parts, and blending, combining sounds in words.

    I presented the case against phonemic awareness training by reviewing research that claims to show that phonemic awareness instruction works—that is, it results in better reading. It doesn’t. In my review, I found that children who had phonemic awareness training did not clearly do better on reading tests than those who didn’t have it. In fact, I found only six studies of the impact of phonemic awareness instruction on reading. The only two publications that claimed victory for phonemic awareness had a small number of subjects, and in one case the language involved was Hebrew, not English. I published my results (Krashen, 2001a). The response of the National Reading Panel (Ehri, Shanahan, and Nunes, 2002), champions of phonemic awareness instruction, was that if we had more studies the results might be different.

    But there is more: studies show that children improve in phonemic awareness without instruction (Krashen, 2003) and that children who have low phonemic awareness often learn to read and write very well, especially when they find texts that really interest them (Krashen, 2001b).

    The current zeitgeist is that we understand texts by decoding print—that is, by using the rules of phonics to pronounce the words we see. We do this all in our minds. We then “listen” to our mental pronunciations of the words and understand what we “hear” in our mind’s ear. To do this, children need to study and learn the rules of phonics. The favored approach is phonemic awareness training followed by “systematic intensive phonics,” teaching ALL the rules of phonics in a strict order to all students.

    A major problem with systematic intensive phonics is that the rules are often very complex. Expert researchers and even teachers don’t know all the rules, students don’t learn all the rules, and different programs teach different rules.

    I found a good example of a complex rule in Goodman’s Phonic Facts, p. 41: the b at the end of comb is not pronounced, it is silent. In my presentation I asked the students, all graduate students and excellent readers, if they knew the rule. Only a few did (b is silent after m at the ends of words). I then asked them how to pronounce combing and combination and asked them why b is silent in combing but is pronounced in combination. Nobody had any idea. But they could pronounce both words. Here are the rules (again, I learned them from reading Phonics Facts): b stays silent when the ending has a grammatical function; otherwise, we pronounce it.

    Also, many rules don’t work very well; they have numerous exceptions. Clymer (1963) investigated 45 phonic generalizations that applied to words included in four basal series. The well-known rule “when two vowels go walking, the first does the talking” (when two vowels appear side by side, the long sound of the first is heard and the second is silent, as in bead) worked in only 45% of cases in which words had vowels back to back. Other rules did just as poorly.

    Frank Smith, in Understanding Reading, pointed out that a considerable number of phonics rules are “unreliable… there are too many alternatives and exceptions… 300 ways in which letters and sounds can be related” (p. 41). He also pointed out another problem with systematic phonics teaching: different phonics programs teach different rules. Clearly, straightforward rules of phonics can be consciously learned by most students, but most of our knowledge of phonics is acquired via reading.

    Even when rules are consciously learned, do they do us any good? Not much. Elaine Garan has concluded that phonics study pays off when reading tests only ask children to read lists of words out of context. Phonics knowledge does not contribute to comprehension. Other studies confirm this (Krashen, 2009).

    If Not through Phonics, Then How?
    The best hypothesis is what Goodman (1967) has called the psycholinguistics guessing game. We learn to read by understanding what is on the page. This is the way we acquire language in general. We make predictions based on (1) our knowledge of the world, (2) our knowledge of the language, and (3) what we have heard/read already.

    We don’t have to teach students to predict or urge them to predict. Frank Smith (2004) points out that we are predicting all the time, and our predictions are usually correct, which is why we are so rarely surprised. Successful prediction helps us get through the day.

    If our prediction is correct (confirmed) according to what we hear and read, we understand. If what we understand through hearing or reading contains unacquired aspects of language, we may start to acquire the form and meaning of these items, a little at a time. As Goodman has pointed out, we learn to read as we are making sense of print.

    Of course, if Goodman and Smith are right, it is a multitrillion-dollar disaster for the textbook industry. I’m afraid intensive phonics phans are going to have to find another way of making a living.

    Clymer, T. (1963). “The Utility of Phonic Generalizations in the Primary Grades.” Reading Teacher, 29, 333–342.

    Ehri, L., Shanahan, T. and Nunes, S. (2002). “Response to Krashen.” Reading Research Quarterly 37(2), 128–129.

    Goodman, K. (1967). “Reading: A psycholinguistic guessing game.” Journal of the Reading Specialist 6, 126–135.

    Goodman, K. (1993). Phonics Phacts. Portsmouth: Heinemann.

    Krashen, S. (2001a). “Does ‘Pure’ Phonemic Awareness Training Affect Reading Comprehension?” Perceptual and Motor Skills 93, 356–358.

    Krashen, S. (2001b). “Low PA Can Read OK.” Practically Primary, 6(3), 17–20.

    Krashen, S. (2003). “The Unbearable Coolness of Phonemic Awareness.” Language Magazine, 2(8), 13–18.

    Krashen, S. (2009). “Does Intensive Decoding Instruction Contribute to Reading Comprehension?” Knowledge Quest, 37(4), 72–74.

    Smith, F. (2004). Understanding Reading. Mahwah, NJ: Erlbaum.

    *Kenneth Goodman passed away on March 12, 2020. This paper was written as part of a memorial to Kenneth Goodman, December 12, 2020.

    Stephen Krashen is professor emeritus, Rossier School of Education, University of Southern California.


    Please enter your comment!
    Please enter your name here