Over the summer I had the privilege of traveling to Pennsylvania to work with some old colleagues and make some new ones. The travel itself is a slog — three hours in a car, four hours on a plane, several more hours in a rental car for a few days, another four hours on a plane, another three hours in a car. By the time I get home, all I want to do is watch a little Netflix and go to bed, but my brain doesn’t wind down so easily. After my return, I found myself turning over a couple of conversations in my head. I didn’t consider them to be related, but of course, there are no coincidences.
The first thing I kept turning over was something I heard from a colleague while I was on the road. She was summarizing for me some comments she had heard from a bigwig phonics apologist (we’ll call her BW for BigWig) during a professional development event. So first, I want to acknowledge that this is total hearsay; I am now reporting this to you third-hand, which is probably irresponsible, but it’s not like I’m going out on a very far limb: it’s a ubiquitous perspective, and one that confronts me a lot in my work. A lot of people are laboring through the whole erroneous sound-it-out, A-is-for-Apple thing out there. But this story was sticking with me because it offers such clear evidence that people highly trained in multisensory structured language education are getting really bad information, right from the top, as though it were scientific or factual. Again.
It went something like this: after a lively discussion about language and teaching, the training session ended with BW admonishing her teacher trainers: “You’ll hear a lot of talk out there about morphology and etymology,” she said. “But the fact is that children need to master their phonology first.”
Just let that sink in: children need to master their phonology first.
My colleague brought it up because after a few hours of orthographic study with me, she could so clearly see that morphology and etymology both outrank phonology in the hierarchy of the writing system’s concerns, and she was well aware that she was being warned away from my work and the work of my orthographic scholarship community, because it threatens the rabidly phonocentric foundations of the trainer’s life’s work.
I looked at my colleague and asked her, “So at what point does she suggest children have mastered their phonology?”
And now I’m looking at you and asking you to consider at what point you mastered your phonology, and what exactly that means.
* * *
The second thing keeping me up was a question that had popped up on Facebook about spelling and, of course, pronunciation. Actually, questions pop up on Facebook every day. I do some of my best writing in response to those questions. There was a thread on whether the vowel in bang, sang, hang, etc., is ‘long’ or ‘short.’ It was a long conversation about who’s pronouncing what, taking place among people who, like my Pennsylvania colleague, are highly trained and knowledgeable about the English language. What most of them didn’t understand was that regardless of what you feel like you’re pronouncing, the vowels in those words are all phonologically checked, lax, or (shudder) ‘short,’ just like they would be before <tch> or <ck>.
That’s not the question that stuck with me, however; it was a post that took on disyllabic words that end in <et>, like blanket or locket or basket.
The common understanding out there in Phonicland is that these words have an /ɪ/ in the final syllable, but are spelled with an <e>. Because of what is perceived as a mismatch in the orthography, many teachers and tutors try to cue students in to the spelling by articulating the final syllable with a pure [ɛ]: “lock-ET.” The thing is, no one talks like that in real life. That word is never pronounced that way in a normal utterance, only in a phonics lesson. If the kid is writing a sentence sometime later with that word, they’re going to be thinking about the content of their composition, and it won’t feel like “lock-ET” in their head. So if that’s the approach, it’s likely that kids will continue to spell these (and other) words according to the way they pronounce them with little regard for sense and meaning.
Let’s consider these words, or at least a goodly number of them, in terms of our four questions of structured word inquiry. The thing is, we usually take these four questions to the study of a specific word, but this time, I want to take them to the study of words with that have an <et> suffix, and of that <et> in general.
Most of these words are of French origin and have a diminutive denotation, at least historically. The <et> may no longer clearly express smallness in the word (as in musket), and the word’s semantics may have drifted considerably from their etymon, but somewhere in many of these words’ etymologies is diminution (the history of musket includes an association with a ‘little fly’ — think mosquito).
A packet is a small pack of something; a locket is a little locking thing; a rocket is, etymologically speaking, a little cylinder. They are by and large nouns, though many have been zero-derived into verbs, as in snow that blankets the landscape. In the French that they come from, <et> is a masculine diminutive suffix, and <ette> is its feminine counterpart. English uses both: kitchenette, rosette — sometimes with the same base element, even: planchet, planchette, or toilet, toilette.
Well, since <et> isn’t a word, we need to think about this a little differently. It’s a suffix, not a base, but we can still take a look at how its morphology works. It attaches to both free bases (locket) and bound bases (banquet). These words take the typical nominal inflections, as in two caskets or the rocket’s red glare, and they may also compound, as in basketweaving or toiletpaper, or take derivational suffixes, as in toiletries or musketeers. The suffix is never stressed in these word families, so words like magnetic tell us that we’re not dealing with the same diminutive suffix here.
Now, I wouldn’t argue that all of these words are very productive for analyzing morphologically in the present day; you won’t find the <rock> in <rocket> in any other words without the <et>, and the <jack> in <jacket>, while free (a kind of sleeveless tunic or, well, jacket), is archaic and no longer used. But even many of the words that are not morphologically generative are still hugely etymologically productive.
Let’s take a look.
Well, I’ve already given several examples of the kinds of morphological relatives you’ll find with these words. Some compound prolifically (basketball, breadbasket, wastebasket), others compound a little (straightjacket), while still others don’t compound at all (blanket). Where we really hit paydirt is with the etymological relatives. When we peel off that <et> and identify the base, even if it’s not morphologically generative in present-day English, it often points to relatives that help us deepen our understanding not only of the meanings of these word families, but also of the kinds of relationships we mark in writing. Consider the following relatives and their shared historical denotations:
banquet ~ bank ~ bench ‘table’
blanket ~ blank ~ blanch ~ blanquette ‘white’
bracket ~ breeches ~ britches ‘support or armor’ (influenced by, but unrelated to, brace)
crocket ~ crook ~ crochet ~ croquet ‘hook’
latchet ~ lace ~ lasso ‘lace’ (historically, ‘rope, noose, snare;’ influenced by, but unrelated to, latch)
market ~ merchant ~ merchandise ~ mercantile ‘to buy’
picket ~ pike ~ peak ~ pique ~ pitch ‘to prick or point’
pocket ~ pouch ~poke (as in pig in a poke) ‘bag’
planchet ~ planchette ~ plank ‘board’
placket ~ placard ~ plaque ‘plate or tablet’
ratchet ~ rocket ‘spindle or cylinder’
ticket ~ etiquette ~ sticker ~ stitch ‘to stick, to fasten by sticking’
As we study these shared denotations and the stories they tell, we also begin to see another dimension of the writing system come to life: the etymological relationships between phonemes and graphemes. In the examples above, we can see certain relationships repeated in the graphemes: <tch> or <ch> and <ck> or <k>; between <c> and <ck> and <k>, and <qu>; sometimes multiples of these. We see these same relationships across English orthography, even in words without the <et> including dike~ditch, break~breach, wreak~wretch, mystic~mystique, and so many others.
Some of the morphological relationships that emerge in this study are built around free bases and are fairly intuitive, like <face> + <et> or <cab> + <in> + <et>. Others have bound bases whose denotations make perfect sense, but may not be guessable, like the <buck> shared by <bucket>, <buckle>, and <buckboard>, which denotes ‘bulge.’ The <cors> in <corset> is related to corps and corpse — they all denote ‘body.’ I know, right? Wow.
We also see in many of these words all these medial spellings that we associate with being final to monosyllables, like <tch> and <ck> and <nk> and <dge>. Once we understand that, historically at least, those spellings are final to a base element, we understand those words and their spellings better. If we look at relatives like toilet and toilette or blanket and blanquette, floret and flowerette, ticket and etiquette, we ought to be able to surmise that the <et> and <ette> suffixes are related somehow.
Basically, the more relatives we can gather, the more data we have to support our understanding. I could go on, but I want to save some of the study for other people. So let’s wrap up with the fourth question.
This is really the question that started it all, as it so often does. Even though it should be our final question, this is frequently where people start when studying spelling: with the pronunciation. And that’s a mistake. It’s a mistake, as it was in the Facebook discussion, to zero in immediately on “why are these words pronounced like this but spelled like that?” without considering the other questions, the context of the meaning and structure and relationships that govern the spellings.
Just as with the <a> in rang, sang, and hang, people get hung up on what the physical pronunciation of a word is — its phonetics — rather than considering how we organize a word’s possible pronunciations within the whole system of our language — its phonology. All writing systems represent some aspect(s) of the language’s phonology, its psychological organization of the pronunciation of meaning, but no orthography represents phonetics, the physical properties of speech.
With the frequent phonics practice of stressing that second syllable in words like locket or packet — a with many frequent phonics practices — we not only fail to study how these words’ pronunciations actually work; we also obscure other facts about the pronunciation of this suffix. Sometimes it is stressed, especially in musical terms like quintet or quartet or duet or clarinet. Some of these can be spelled either with <et> or with <ette>: quintette, quartette. We also find the same diminutive suffix stressed but pronounced only as /eɪ/; these are later French loanwords like ballet, filet, bouquet, sachet, croquet. In French, which is a syllable-timed language, the tonic stress falls at the end; the closer a word is to its French origin in time, generally the Frencher it is in English.
Rather than fixating on whether a child spells the second syllable of these words correctly immediately and consistently, what if we left that alone for a bit and studied instead what is actually happening in these words, with this suffix? So far, we know that the diminutive suffix <et> can be pronounced as [ət] (as in musket), [‘ɛt] (as in clarinet), or [‘eɪ] (as in ballet).
Wait — what? So there’s no /ɪ/ in locket, bucket, basket, musket, and the like? Nope. Some people might feel like they’re pronouncing [ɪ] in these words, but there’s no /ɪ/. If you don’t know the difference between [ɪ] and /ɪ/, you need to take my IPA LEXinar post-haste. There is an /ɪ/ in fidget and in ticket, but it’s in the first syllable, not the second, and it’s spelled with an <i> as it should be. In fact, /ɪ/ is one of the simplest vowel phonemes to spell in English because it only has two spellings: <i> and <y>. The grapheme <e> is not associated with the phoneme /ɪ/ in English. It could be spelling the phoneme /ɛ/ in these words, arguably, but unless it’s stressed, it’s a schwa.
Now, not every di- or trisyllable that ends in <et> has a demonstrable or identifiable diminutive suffix, even diachronically. The word that started it all, basket, is of obscure origin. Hornet is Old English, not French, and isn’t actually related to a horn or anything. But as words orbited around each other in Middle English and beyond, surely their spellings have been influenced by that ubiquitous <et> suffix.
Whether the schwa in these words is phonemic on its own or just an unstressed allophone of /ɛ/ in these words is anyone’s argument to make. But it’s not an /ɪ/. There are words, also schwaed at the end, that arguably do have an /ɪ/ phoneme in the unstressed syllable, because it’s spelled with an <i>. Latinate words like habit, limit, merit, and implicit all have a Latinate <it> suffix (compare rehabilitate, subliminal, meretricious, and implicate). Since its pronunciation is similar if not identical to an unstressed <et>, it bears examining what differentiates these kinds of words. First, the <et> suffix is French; these words reliably made their way into English via French, some from Latinate roots (like facet) and others from Germanic roots (like bucket). The <et> in French could be word-final, as it can in English. The <it>, on the other hand, is a Latin stem suffix, meaning it’s followed (in Latin, and often in French) by another suffix: habitare, limitem, meritare, and implicitus). While these words may have traveled via French, they retain a Latinness in their morphological families that the <et> ones don’t: compare habit, habitual, rehabilitate, inhabit to bucket, buckle, buckboard. The relatively concrete homeliness of the <et> words is what makes a lot of Phonics Pholks misidentify them as “Anglo-Saxon.” The Latin <it> almost always fixes to bound bases (inherit,circuit, audit), while the <et> is less selective.
Here’s what’s up, phonologically speaking: words that end in an unstressed <et> — whether or not it’s a suffix — do not have an /ɪ/ in the final syllable; if you want to name a pure vowel phoneme, it has to be an /ɛ/ or an /eɪ/; that’s all the orthography allows for. The unstressed ones have a schwa, because they’re unstressed. They may have an [ɪ], but when people are trying to identify the phoneme by saying “Well I pronounce it this way,” they’re missing the point. They’re looking for love in all the wrong places. Phonology isn’t in your mouth; it’s in your brain. Not in your physics, but in your psychology.
Orthography is human thought made visible as text. Because phonemes are in our heads, we can argue for a long time about their identity. Because phonemes are sets or categories in our linguistic psychology, we can debate their contents in perpetuity. If we want to identify them, literally, to figure out their identity, we need only look to the orthography. The orthography takes our phonology — our human thought — and makes it visible as text. Many orthographic scholars understand that the orthography doesn’t represent phonemes; rather, it pinpoints and identifies them with graphemes.
Here’s my hypothesis: [ɛ] isn’t typically realized in unstressed syllables in English. In compounds like redhead or daybed, the second element carries secondary stress; it’s not neutralized. In other words, you have to stress it to fully pronounce it. If you pronounce basket as *[‘bæsˌkʰɛt] you end up giving the second syllable some stress — it’s louder, longer, and maybe higher than in [‘bæskʰət]. It ain’t natural. So, instead of emphasizing or overemphasizing pronunciation (indeed stop doing that — good for the teacher who posted for identifying that this practice is both ineffective and misleading), zero in on meaning. Use the four questions as a guide. Talk about the phonology only after after you’ve made sense of the meaning, the structure, the history and the relatives of these words.
In these Facebook conversations, which are brave and honest and sometimes even moving, people reshape their understanding of English orthography. People are vulnerable and willing to be wrong. Everyone learns from each other that way. It struck me that in this friendly, brilliant, experienced, well-educated group of parents and educators and activists and innovators, conversations about phonology were still often woefully wandering. Participants remained fixated on the physics of their pronunciation, rather than looking at the grapheme that make human thought visible.
“You people have a phonological awareness problem and I am not kidding,” I posted. “I mean, I am, but I’m not. I mean this in the friendliest of ways.” Then I put a smiley face for good measure. I don’t want to insult anyone, you know.
I do want to suggest, however, that it’s possible they haven’t yet mastered their phonology. So rather than waiting until everyone totally agrees on the pronunciation of bang and sang and hatchet and blanket to start studying the structure and history, how about instead we use the morphology and etymology to make sense of the phonology? Go ahead, BW, keep all your eggs in one phonocentric basket and keep insisting that children master the phonology, whatever that means. I’ll be over here helping people understand it instead.