Sounding out the word "through"

Earlier this year I was in a meeting in which a reading specialist in a primary school stated that the word through, “couldn’t be sounded out in a million years.”

This statement is very interesting.  The idea the teacher was expressing so colorfully, namely, that attention to letter/sound relationships is of little or no use for learning how to read certain words, is a view shared widely not only by teachers who are skeptical of the use of phonics generally, but also by many enthusiastic proponents of systematic phonics instruction.   Unfortunately, this view, though quite widespread, is demonstrably wrong.  Examining why it is wrong it provides an excellent opportunity to highlight a prevalent misconception about the role of phonics in good reading instruction.

First, to the demonstrably false parts of the statement.  The word through is composed of 3 isolated speech sounds, technically referred to as phonemes. These phonemes are /th/ /r/ /oo/.  The relationship between the sounds /th/ and /r/ and how they are represented in this word, namely with the letters th and r, respectively, is quite regular. In contrast, the way the /oo/ sound is spelled in this word, namely with the letter string ough, is unusual. Indeed, it is the only common word in which the sound /oo/ is spelled this way. 

However, just because a sound in a word is represented in an unusual way does not make the word impossible to sound out.  It is a simple matter to say, “We spell the /oo/ sound in this word ough.”  In ABeCeDarian, in fact, this word is initially presented in Unit 8 of Level B1 in an activity in which students sort words according to how the /oo/ sound is spelled in them.  As you can see the spelling units (often referred to as graphemes) on this list of /oo/ words to sort is presented with space between each grapheme.  For instance, the word food is displayed as f oo d.  The word through, therefore, is presented as th r ough.


The student already knows the regular letter/sound correspondences th and r and there is only one other grapheme in the word.  The student, moreover, has been told that all the words on the list have the /oo/ sound.  With the help of the spacing between between graphemes and the knowledge of the common sound in these words, most students who have advanced to this level of the program have no difficult sounding out all of the word correctly.  If a student does, however, get stuck on what sound to try for the ough, the teacher points to it and says simply, “In this word, that’s /oo/.  You say /oo./“  Armed with this knowledge, the student then sounds out the word.  It’s really that simple, no more difficult, indeed, than sounding out its more regularly spelled homonym, threw.

If it is really so easy to sound out a word with a highly irregular spelling such as through, then why do so many teachers think that it is impossible to do so?  The heart of the error, I think, lies in a view that understands the utility of phonics in an overly narrow way.   According to this view, learning basic letter/sound relationships provides a set of conversion rules that allows a person to translate strings letters into words.  For example, the conversion rules specify that the letter m represents the sound /m/, the letter o represents the sound /o/, and the letter p represents the sound /p/. Thus, the student can figure out that the word mop is /m/ /o/ /p/, /mop/.

These conversion rules, however,  are useful only to the extent that they are relatively simple and straightforward to apply. Or so the conventional wisdom goes.  But as we all know, in English, there is usually not a simple relationship between letters and sounds.  Letters and letter combinations can represent many different sounds, and when there are too many variations, trying out each variation when reading a word makes the conversion too cumbersome.  The string ough, is an extreme example of this variability, representing different sounds in each of these 7 words:  though, through, rough, cough, bought, plough, hiccough.

Because of this variability many people argue that words with highly irregular spellings must be taught in a different way, without recourse to letter/sounds.  Thus, in many reading programs, the words with irregular spellings words grouped together, often often labeled with an slightly accusatory designation, such as “red words,” “outlaw words,” “rule-breakers” or “words that don’t play fair”.  Some teachers don’t analyze these words at all and have students practice reading them as unanalyzed wholes or by embedding them in sentences in which, supposedly, context will allow them to be read properly.  Other teachers do have students analyze the words by having them repeat the letter names and write the words saying the letter names in various exercises.  And of course, some teachers do a little bit of all of these strategies.

To see the flaw in these approaches, it is necessary to tease apart some related but distinct goals of reading instruction.  The act of translating print into speech is usually referred to, quite aptly, as decoding.  When a person looks at the word “mop,” and says /mop/, for example, she has decoded the word.

There are two kinds of decoding.  One kind is used for translating into speech a word that a reader is unfamiliar with.  This is the type of decoding a person uses when she tries to read a word she has never seen before, such as the nonsense word fluntercumicality.  Even experienced and highly adept readers will read this word relatively slowly and have to approach it by translating individual letters, or, in some cases, perhaps small groups of letters, into sound and then combining those sounds in the proper order.  This type of decoding is often referred to as sounding out a word.

But as readers acquire increasing skill, they also acquire the ability to translate printed words into speech almost instantly, without overtly converting individual letters or sub-groups of letters into sounds first.  The acquisition of this automatic word recognition for a large number of words is, of course, essential to the development of the ability to read well. Reading would be of little utility if a person had to sound out most of the words on a page.

Even for students who learn to read rather quickly and easily, however, the ability to read a particular word automatically generally requires that the student be exposed to the word at least several times.

As I’ve noted, many teachers maintain that teaching letter/sound correspondences to students might very well help them decode the word hand if they had never seen it before, but it will not help them read words such as through if they had never seen it before.  This observation is probably true, especially in the case of very beginning readers who have very little experience reading.

But that is hardly the end of the story.  In addition to being able to sound out unfamiliar words, as I just noted, students also need to be able to instantly recognize many common words.  The word through, for instance, is  a word we would want second graders to be able to read easily in text. Indeed, a large number of the most common words in English have somewhat irregular spellings.

Given that most students need repeated exposures to words, we are now confronted with an important pedagogical question: What sort of presentation and practice allow the student to instantly recognize irregularly spelled words like through, with the least number of repetitions.

The answer, perhaps surprising to many,  involves showing the student quite clearly and explicitly how the letters in the word represent the sounds of the word. In our example of the word through this kind of presentation means showing students quite explicitly and clearly that the graphemes th,  r and ough represent the sounds /th/, /r/ and /oo/.

While it might seem counterintuitive at first that a “phonological” presentation is key to developing rapid, automatic word recognition, It’s not difficult to show why this approach is more efficient than others.

As I mentioned earlier, one common “non-phonoogical” option is to read the word frequently in context.   Reading the word frequently and in many different contexts will certainly help a student learn to read the word, but if the student doesn’t get help analyzing the letter patterns in the word, she will make many mistakes before learning the word correctly.  The semantic information in a sentence, such as, “The children walked through the park” doesn’t provide sufficient information to narrow down what the word is.  The words in, into, around, by, toward, from, and others could replace the word through in this sentence and preserve a grammatical sentence.  Moreover, such practice leaves the student guessing at what the connection between the letters and the word are, and many students will not be able to analyze it without some explicit and clear assistance.

One can even better appreciate the confusing diffuseness of the information presented to the student with this approach when one asks, “How should the teacher correct a student who makes an error reading the word through in this sentence?”   As I just noted, there is not enough semantic information in the sentence to determine that the word is indeed through and not some other similar preposition.  If the teacher starts pointing out letters within the word to help the student understand why this word is through, and not into, or around, then one has to ask why not make sure the student is clear about all of the letter/sounds in the word.

This criticism of using context to help develop decoding skill is not to say that semantic information is irrelevant to decoding.  Students should be told constantly that what they read should make sense.  Good readers certainly do and should use this guideline to monitor whether they have decoded words accurately.  But the role of the semantics of the sentence and other information related to the meaning of the text is properly information that a reader uses to monitor the accuracy of her decoding, but it does not serve well as the principal information used to practice, develop, and improve a student’s decoding.

Another option is to present the word in isolation and drill it along with other irregularly spelled so-called “sight words.”  Because the student has to decode the word without any information from context, this approach probably does help focus the students attention a bit more on the letters in the word. But again, why should the student be left to analyze the word herself, when the letter/sound patterns can be shown to her and then rehearsed explicitly. Correcting any errors is very cumbersome without explaining clearly how the letters in a particular word represent its sounds.

The remaining option for teaching the word without reference to letter/sounds is to have the student spell the word out loud and then say the whole word: tee, aitch, ar, o, yu, gee, aitch, /throo/.  This is the approach of some explicit phonics programs such as Orton-Gillingham, which adds motor gestures to the performance in the belief that this aids memory.

Of all the alternate methods of teaching irregularly spelled words, this is the best because it focuses the student on letters.  But the focus is still too broad.  It calls the student’s attention to individual letters rather than functional groups, and as a result, it doesn’t help to clarify how our spelling system works, nor does it help students make connections among words with similar patterns. The complexity of decoding through, resides primarily in the facts that ough, not only represents /oo/ exclusively in this word, but also is used in other words to represent so many different combinatins of sounds in other words. Nevertheless, this spelling does bear some connection to the spelling of other words. Specifically, the ou in through represents a similar sound to the ou in several other words, such as soup, group, route, you, and youth, and having a “mute” gh is a feature of many common words, such as sight, eight, and dough).

In short, words such as through do not have to be learned as discrete and isolated wholes, but can still be integrated with the patterns that appear in other words. Only when a student is helped to analyze explicitly how the letters in words represent sounds do these patterns and relationships become clear. And it is this understanding of the patterns and relationships which will generalize and help students read not only the single word through correctly, but make it easier to learn other words with ou and gh. The approaches using context only, or isolated, whole word practice or oral spelling don’t readily promote this kind of generalization.

These connections exist among the thousand or so most common words in English, that is to say, the words we want students to learn to recognize automatically in primary school. But what about less frequent words with very unusual spellings, such as colonel and yacht. Is it really possible to help people learn to read these words by pointing out letter/sound relationships?

I would argue with a resounding, “Yes!” Here, the description will often involve more than just the explication of letter/sounds, but delve into even deeper relations, regarding the etymology of the word and how the word was pronounced at various times in various different languages. For example, the word colonel, comes from the Italian colonnello, meaning a person who commanded a column of soldiers. The pronunciation changed in Spanish and Portuguese, with the l replaced with an r (coronel in both languages). The modified pronunciation was retained in English along with the older spelling. (A similar sort of change in which a changed pronunciation is matched with an earlier spelling, explains why the pronunciation of the word one doesn’t match its spelling.) This approach, which strives to explain letter/sound relationships clearly, not only helps students learnn to read the words easily, but exposes them to the richness of language and how words evolve and are connected across different languages..

I should add that it is very useful to pronounce words such as this to adopt a “spelling pronunciation” in order to help people learn to spell them. In other words, for most people, spelling performance is improved if the letters can be connected to individual sounds, thus ensuring that the person does not have to rely solely on visual memory to spell the word correctly.

Cognitive scientists who have developed neural networks to simulate reading have observed a similar pattern in these networks, namely, that information about letter/sound relationships, is discovered and utilized even for words with irregular spellings. For instance, the word give, seems to be an exception to the pattern of pronunciation found in dive, drive, jive, hive, thrive, alive, strive, but there are nonetheless words that have similar spellings yielding similar pronunciations, such as gift and river, as well as the dual pronunciations of live. So there is some good evidence from cognitive science that letter/sound knowledge can be utilized even for the decoding of words with irregular spellings.

Furthermore, we have ample evidence that good decoders differ from poor decoders primarily with regard to their letter/sound knowledge and ability to sound out unfamilar words. Therefore, it is important to make sure that all of students understand how letters represent sounds, and not rely on trying to read words by treating them as unanalyzed or poorly-analyzed wholes. This approach reduces guessing, and, generally reduces student confusion and stress.

So, it is indeed possible to sound out the word through and others like it that have irregular spellings. Moreover, if teachers fail to make the letter/sound relationships explicit in these words, they cheat their students of a valuable tool to accelerate their reading growth.