Learning Spelling by Heart

A teacher in one of my training classes this year sent me a wonderful email this morning, informing me that she and her students had been studying the <ea> digraph. They had studied words with /ɪ(ə)ɹ/ (like ear), words with /ɜɹ/ (like early), and words with /ɛ(ə)ɹ/ (like pear).

But “what about the ea in heart?” she wrote. “I’m reading on etymonline that the ea in the word used to be a long vowel, but then the pronunciation was shifted.  I’m thinking this is the only word like this?”

I appreciated this teacher’s question, and the fact that she had already investigated it herself! I love that she brought me not only a question, but also the evidence she had gathered. She knew to look to the etymology to explain the selection of a grapheme, and she did indeed find a diachronic explanation for the spelling. Here’s how I responded to her inquiry, learning a great deal along the way.

Great question! And it sounds like you’ve already done a thorough investigation. You are right to locate your understanding in the etymology — in the history. And that’s really plenty. But, because I am totally compulsive about spelling, here’s a little more.

If you look at the <ea> card in the LEX deck, you will see that <ea> before an <r> can be pronounced in 3 different ways:

hear     /hɪ(ə)r/
early    /’ɜrli/
bear    /bɛ(ə)r/

[Here’s a picture of the back of that card:]

The word heart, of course, has none of these vowel pronunciations, and instead is pronounced like hart, dart, art, card, etc. So why is it spelled with an <ea>? Well, remember that pronunciation is the fourth and final concern in our questions about orthography:

1. What does it mean?
It’s the cardiac organ, and a lot of figurative meanings (courage, compassion, love, memory, etc.)

2. How is it built?
It’s a free base element, of course — no affixes to peel off.

3. What are it’s relatives?
3a. Morphological relatives?
hearty, heartless, disheartened, hard-hearted, heartfelt, hearts . . .

So no, heart is by no means the “only word like this” — but <heart> is the only base in whose word family the <ear> represents /ɑr/.

3b. Etymological relatives?
cardiac, cardiologist (from Greek), courage, cordial, core, concord, record, discord, accord (Latin/French) — if we go back far enough and look at a wide enough swath of relatives in other languages, we’ll find an <e>, but that may not be helpful. I will say that it’s often the case that an <e> and an <o> (or an <ea> and an <oa>) can mark a relationship — they are both ‘mid vowels’, phonologically speaking: month/menses, broad/breadth — and even moreso, an <ear> often has an <or> relative. Sometimes it’s obvious, like in

Sometimes, a little less obvious:
earth~ore (this one I think is really cool)

So, the fact that <heart> is closely related to all the Latinate forms with an <o> helps us make deeper sense not only of the spelling of heart, but also of a broader pattern in the language.

Synchronically speaking, just as we see similar spelling patterns in heel, feet, knee, we also see heart, head, breast share a spelling pattern as well. These words aren’t historically related, but in the present day, they bear a connection in meaning and in spelling.

4. What aspects of the pronunciation do we have to consider?
Well, in American English, heart sounds like art, as I said. But is some other Englishes, like in Scottish dialects, heart still has a vowel that’s closer to bear. Of course, the info you dug up on Etymonline also offers a diachronic (historical) perspective of the pronunciation. Another reason for keeping the <ea> spelling is to differentiate heart from its homophone, hart, a word that was probably in much more common usage in 1500 than it is today!

Now, in case you didn’t click on the link above, here’s what etymonline actually gives us:


Now, what I didn’t realize until after I clicked send is that the word hearth is also spelled with <ear> but pronounced as /ɑɹ/. Who can find a relative that explains the spelling of <hearth>?

And isn’t word study a heartier endeavor than memorizing a list?


  1. Peter Bowers says:

    Thanks to you and your teacher friend for such an excellent question and response. So true — learning is always so much richer when questions are constructed with the presentation of evidence about the thinking.

    This particular question delves into an area of orthographic phonology that I have been trying to sort out for a long time. I’ve read the post and inspected your LEX cards but I am still left with a question that I would be curious to hear more about.

    My understanding from your grapheme card is that an R immediately following the EA digraph can signal a “rhoticized” phoneme that the EA digraph can represent. (Excuse the caps instead of angle brackets for this comment section.)

    Is that an accurate interpretation?

    What I keep struggling with is sorting out whether it is more coherent to describe EA as a digraph in such a case, or if it would make sense to think of EAR as a trigraph in a word like HEAR? If HEAR is described as having an EA digraph, what is the linguistic term for the R letter in this word? Is it still a grapheme, or perhaps some kind of marker letter?

    One reason for this question is that elsewhere we treat OR, AR, and ER as digraphs.

    For example one of your LEX cards presents digraph OR as representing the phoneme /ɔɹ/ in words such as FOR, CORN and ORDER.

    So here is my question…

    If EA is a digraph that can represent a rhoticized phoneme when followed by an R, (like in HEAR) would it be more coherent to consider the O a single letter grapheme that can represent a rhoticized phoneme when followed by an R?

    Alternatively, would it be more coherent to think of EAR as a trigraph and OR as a digraph? That description helps me explain the status of the R letter in these spellings.

    I am totally open to the possibility that I am missing something very fundamental here about orthographic phonology.

    I’ve asked about this same type of question many times and I’m sure that a key part of my confusion has to do with a need to understand the abstract concept of a phoneme and how to prove the identity of one more clearly.

    I’m not expecting to have a clear understanding to this question I’ve asked many times after this post, but the question of your teacher and your response helped me pose this question in a way that more clearly describes my confusion than I have previously. At least that feels like some progress towards better understanding!

    • Thanks for your thoughtful comment / question, Pete. You are not alone in wondering about rhotic vowels and their graphemes in English. Many people have asked me if OR or AR is a digraph, and if OUR and EAR are trigraphs. For me, this is part of a desire folks have for definitive answers or definitive labels in language study, and language doesn’t always comply with that desire for definitiveness.

      The best answer I can give to the question “is EAR a trigraph?” is, “It depends.” I’ve also had people write me and say, “There’s no EAR card in the LEX deck, so that must mean it’s not a trigraph, right?” So for starters, I want to make sure that readers understand that the LEX deck is not prescriptive; rather, it’s generative.

      In my estimation, it would not be correct to say definitively “AR is a digraph” or “EAR is a trigraph” because it depends. There are multiple contingencies: it depends on the phonology of these spellings, not only in particular words, but also in particular dialects.
      This question involves an understanding of what linguists call rhoticity — some dialects of English are rhotic; others are non-rhotic. The word rhotic derives from the name of the Greek letter ρ, or ‘rho’, the equivalent of the Latin R. Rhotic vowels are also sometimes called R-colored, or, in phonics, R-controled.

      In rhotic dialects, like mine (a California – New England – Midwest hybrid), non-pre-vocalic /ɹ/ is pronounced; in British Received Pronunciation, it is not. So what in the heck do I mean by non-pre-vocalic? I mean an R that does not precede a vowel — so, it’s either word final, or followed by a consonant. Consider these examples:

      Standard Midwestern
      far /fɑɹ/ or /fɑ˞/
      farm /fɑɹm/ or /fɑ˞m/
      for /fɔɹ/ or /fɔ˞/
      form /fɔɹm/ or /fɔ˞m/
      fir /fɜɹ/ or /fɝ/
      firm /fɜɹm/ or /fɝm/
      fear /fɪɹ/ or /fɪəɹ/ or /fɪɚ/
      hammer /’hæməɹ/ or /’hæmɚ/

      Received Pronunciation
      far /fɑː/
      farm /fɑːm/
      for /fɔː/
      form /fɔːm/
      fir /fɜː/
      firm /fɜːm/
      fear /fɪə/
      hammer /’hæmə/

      Now, what you might notice is that in the Midwestern (rhotic) examples, there are two ways to represent that rhotic vowel. Some linguists use a single symbol, like /ɝ/, while others will use separate symbols, like /ɜɹ/ — one for the vowel, and one for the rhoticity.
      The reason that linguists might use different representations has to do with what rhoticity is. Bear with me, here: you’re probably familiar with the terms onset and rime when talking about vowels. Well, the rime — the vowel and everything that follows it — can be divided into the nucleus (or sonoric peak) and the coda (from the Latin word for ‘tail‘ — a term familiar to musicians as well as linguists).

      In rhotic dialects, some speakers simultaneously articulate the vowel (the nucleus, if you will) and the rhoticity; others pronounce a ‘purer’ vowel and rhoticize only the coda.

      Still with me?

      Even if you answered “no”, you can certainly see that the question of whether AR or OR or EAR spells a single phone or two phones in a given word is it depends. The follow-up question is this: if there are two phones, are they two separate phonemes, or a single phoneme? And herein lies the it depends of your question: if there is only a single phoneme, then the AR or EAR or OR or whatever is a single grapheme; if there are two phonemes, then there are two different graphemes.

      When you “struggle” to sort out whether it’s “more coherent to describe EA as a digraph . . . or . . . a trigraph,” you are not alone. It depends on whether you consider the /ɪɹ/ (or /ɪəɹ/ or /ɪɚ/) to be a single phoneme, or two separate phonemes. And I’m not convinced that anyone has a definitive answer for that.

      If I were to attempt such a dialogue, I might say, “the spelling EAR” rather than insisting on “the trigraph” EAR, because then I don’t have to commit to something definitive. Whether the EAR in hear is one grapheme or two depends on whether the pronunciation is considered one phoneme or two — a question which itself depends on how it’s pronounced, and on how it’s interpreted. Either way, the R is absolutely graphemic, whether it’s its own grapheme, or part of a trigraph.

      You wrote “elsewhere we treat OR, AR, and ER as digraphs” — well, clearly in a non-rhotic dialect, they are. But in rhotic dialects, it’s less definitive. You also wrote, “one of your LEX cards presents digraph OR as representing the phoneme /ɔɹ/” — actually, I never really call OR a digraph in the LEX deck. I hedged my bets by not doing so.

      Finally, you wrote, “I am totally open to the possibility that I am missing something very fundamental here about orthographic phonology. . . . I’ve asked about this same type of question many times and I’m sure that a key part of my confusion has to do with a need to understand the abstract concept of a phoneme and how to prove the identity of one more clearly.”


  2. […] « Learning Spelling by Heart […]

  3. […] finally, year is a bit of a treasure trove. It’s related to yore (as I wrote about here), which derived from the genitive (possessive) plural of Old English gear, meaning […]

  4. Reblogged this on Phonic Books and commented:
    Interesting blog about spelling, word origin and morphology

  5. Ravi Nayak says:

    Hi there

    Could the word ‘Heart’ have its origin in the Proto Indo European language of Sanskrit?


    Ravi Nayak

    • LEX says:

      No, Old English words were not descendants of Sanskrit. More like distant cousins of Sanskrit. The Sanskrit word and the English word both share an ancient Proto-Indo-European root.

      There is no such thing as “the Proto Indo European language of Sanskrit,” by the way. Sanskrit, like English, is an Indo-European language. PIE refers to the ancestor language, not to the whole family.

Leave a Reply

Your email address will not be published. Required fields are marked *