When I talk with students of any age about how English spelling works, I introduce them to the term <grapheme>. I also talk with them about letters, but letters are not the same as graphemes. English words are spelled with graphemes, not letters.
A grapheme is a unit consisting of one, two or three letters. A grapheme signals or represents a phoneme. Here are some examples of graphemes in English:
<sh, th, a, qu, r, tch, oi, ugh, ph, or, w, ck>.
And here are the graphemes in several words:
In <mop> there are three graphemes: <m, o, p>.
In <shop>, although there are four letters, there are three graphemes: <sh, o, p>.
In <match> there are five letters but three graphemes: <m, a, tch>
The <sh> in <shop> is one single unit, representing that first phoneme in the pronunciation of <shop>. The <tch> in <match> represents the final phoneme in that word when it is spoken. You can read more about phonemes here.
You might be thinking, “Wait a minute — you said that graphemes are units of one, two or three units, but there are graphemes that consist of four letters, like <eigh> or <ough>.” In fact, these are not graphemes.These strings of four letters are often referred to as phonograms, which is an educational rather than a linguistic term. The term phonogram is used in education to refer to several different categories of items including graphemes and rimes. But by definition, a grapheme represents a single phoneme; the evidence is clear that there are no graphemes that consist of four letters. I’ll write more about that soon.
Most graphemes in English can signal a number of different speech segments in spoken words. Whether or not those different pronunciations are actually different phonemes is an interesting question; there can be arguments for and against, depending on context. By orthographic linguistic definition, one grapheme represents one phoneme, but the pronunciation of a phoneme may vary in different situations. Those varying pronunciations are called allophones. In fact, we need to think carefully about differing speech segments in related words that are represented by a single grapheme; are they allophones or separate phonemes? An example would be the phoneme signaled by the first <i> in <finite> and the same <i> in <infinity> (<infinity ➞ in + finite/ + y>). Are those separate phonemes or are they allophones of the same phoneme? This can become extremely complex and thought-provoking but orthographic phonology is not simple and can only be understood in the context of morphology and etymology.
When I’m helping students learn terms, we often start by looking at the morphological structure of the term. I didn’t always do this, but over time I realized that this practice was helping me to sort out terms that I was trying to learn, and I’m fully literate. Why wouldn’t I show these same meaningful connections to every student?
So we can see that the morphological structure of the word <grapheme> is <graph + eme>. The base <graph> has an orthographic denotation of “write, something written.” We see this base in words like <graphics, autobiography>. The <-eme> suffix is used for terms in linguistics that represent the smallest distinctive unit that carries meaning. The morphemes in the word <grapheme> make it easier to understand and remember that a grapheme is the smallest unit of writing that is distinctive for meaning.
A grapheme is an abstract concept that can have various physical realizations. Those various realizations are called allographs. These are all allographs of the grapheme <a>:
Notice that we can physically represent this concept of a grapheme <a> in many ways. These representations have different forms, but represent the same meaningful unit of writing. This is the concept that we also see in allographs, allophones and allomorphs, which are the various ways that the concepts of graphemes, phonemes, and morphemes physically appear in the English language.
So what are the roles of letters in spelling, and what should students know about letters? Letters are the building blocks for graphemes, much as digits are the building blocks for numbers. When we look at the digit 9, we have no idea of its value unless we know what number it appears in. The 9 in 29 and the 9 in 945 have very different values. In much the same way, the letter <w> can be a component of several different graphemes: <w, ow, aw>, for example. The letter <t> can be used to build a number of different graphemes: <t, th, tch> and others. And some letters in English words are not part of a grapheme at all, but serve a different function. They can be a component of a marker of some kind. Markers can also have several different functions in words.
So when you think about graphemes and letters, these points are key:
- Graphemes are units that are made up of 1, 2, or 3 letters.
- We spell words with graphemes, not letters.
- A grapheme represents a phoneme, and both graphemes and phonemes are abstract concepts that can have various physical realizations.
- Varying physical realizations of a given grapheme are called allographs.
- Not every letter in a word is a grapheme; there are also markers in words that do not represent phonemes. Markers have other functions.
One of the reasons that I call this site Learning About Spelling is because that’s what I’m doing as I write this. Every single day I learn something new about our writing system and replace pieces of my old thinking with new. This is my current understanding and I welcome questions and responses. I’ll edit this as my understanding grows and changes.