The History of the ABC Song

Is it possible that English education is really messed up? Certainly.

Letters as we know them have both names and sounds. A letter's name, like aitch, is not its sound [h], in our spiderwebbed, old, English alphabet. Is it possible to keep name and sound separate in our minds?

I should like us to please de-emphasize the ABC song, because the sounds of BCDGPTVZ have no /iy/, the sounds of FLMNSX have no [ε], etc. "h" contains no [h].

Where did all these funny names come from? They certainly aren't the sounds of the letters. Were they once so?

Meditating upon this question, I remembered my name, Veatch, which seems to tell part of the story. According to a remote cousin's book on the family forebears, one William Le Vache once drove his herd of cattle into the Anglo-Saxon lines during a later battle in the Norman Conquest, helping the Normans win that battle, and justifying his name, forever after, as William The Cow. His descendants being for centuries (at least until my American immigrant ancestor James came over here) the extremely minor nobility of Puddleshire, or somesuch place on the lowest possible low Scots edge next to England.

So now being curious about one's family history, we can look up the word Cow and find, in Latin "vacca"[vakka], Old French "vache"[vače], Middle French "vache"[vače], Modern French "vache"[vaš], Spanish "vaca"[baka], and various spellings in the French region, such as Norman "vaque" or Franc-Comtois "vaitche" (says to Wiktionary.org).

Thus, more or less, into Scottish English comes this Anglo-Norman word, [vačə], which undergoes the Great Vowel Shift and the "Silent-E" deletion event, turning [vače] into [va:č], but maybe it came as [veče] from some Franc-Comtois invader, hence [ve:č], explaining why the Great Vowel Shift which should apply to [va:č] to yield [ve:č] may instead have applied to [ve:č] to yield [vi:č], which is the pronunciation my family uses.

All these were idle thoughts of a plumber that I thought about while carrying pipe or digging a ditch somewhere years ago, but now it comes to mind again considering this most mysterious letter, the horribly named "h" a.k.a. "aitch" or [e:č], which almost rhymes with "Veatch" [vi:č].

Because those same changes would certainly have applied, first from [ače] into [a:č] by Silent e deletion and length shift, then into [e:č] by the Great Vowel Shift. I put it to you: these same long-established sound changes that I've studied and known (with a 300 page dissertation on English Vowels, after all!), when applied to my own name, yield the historical-linguistic path from Veatch to Norman Vache and my great great William the Cow, all that clicks perfectly into place as a reconstruction of the history of this fellow's name. Great.

But now the same reconstruction, the same rules, can hardly less than miraculously be applied to this very monstrosity of a letter name, "aitch", and they lead us to posit no less than a perfectly reasonable attempt at giving the sound of [h], namely [ahe] or may I write it as [əhə], which is just to say that an original attempt to be PHONETIC has been preserved immaculately by centuries of increasingly oppressed and confused schoolchildren and their increasingly oppressive and confused schoolteachers just doing what they themselves once were taught, as the sound changes whirled across town and countryside, making those, originally barely-adjusted, really quite phonetic and useful names, now so remote as to be worse than useless in the education of the young.

This plausible though you might call it a fantastical scenario rests upon a reconstruction of the ABC song's original phonetic form, using known sound changes, and reasonable inference, as follows:

Here:

APA: The APA is what I may call an American Phonetic Alphabet, used by my teacher Labov, and myself in my dissertation; it came from Trager and Smith in the 1940's. You could use [i:,e:,u:,o:] instead of [iy,ey,uw,ow] without signifying a difference, but I'm used to the y's and w's. The APA row in the above table tells us the pronunciations of the names of the letters.

GVS: Let us assume that the song was sung before the Great Vowel Shift (GVS) before the deletion of final e with Length Shift to the previous short vowel (LS). The latter gives us [ahe] for "h", which sounds a lot more like an honest attempt to communicate the sound of [h]. The Great Vowel Shift, a: → e: → i: →ay and ɔ: → o: → u: → aw takes all those long vowels "A":[a:] to [e:], and the E:[e:]'s to E:[i:], and even the I:[i:] and Y:[wi:] to [ay] and [way], which is why we get the sounds of the ABC song, so remote from any reasonable phonetical original.

But if we back out the effects of the Great Vowel Shift, things start to look much more reasonable indeed.

WFL: I should like to propose a word-final short-vowel lengthening rule (WFL), for monosyllabic short-vowel-final words, to precede "Silent-e" deletion with length shift.

This would bring much alignment, since the 9-member /iy/ set, BCDEGPTVZ, then can be reconstructed as using the same original short [e] vowel as the /eC/ set FLMNSX.

LS: The bane of English students, the Silent E, is due to the conservation of spelling from a time before certain destructive and confusing sound changes occurred. Today we live after the time of those changes, and we have to make sense of what we face today.

So you know what happened, it is that a final -e after a consonant and a preceding short vowel, as in "tile" or "fine" etc., reduced and reduced, likely becoming a schwa ə since that is the most reduced vowel, and eventually was entirely deleted, while the spellings remained the same and now we have to deal with the Silent E. But it didn't entirely go away, because the sound is not just its sound but its timing, and since it took some little bit of time to pronounce it somewhere in the word or the syllable, and since the timing might with some syncopation or rhythm shift get to be moved around, apparently it did move around, in fact it became length on the preceding formerly short vowel. So I like to call it Length Shift (LS), since if you just move its timing slot over into the previous syllable, then the deletion of any other qualities of the vowel itself automatically occurs, those qualities have to go away, because they don't have any time in which they might exist. That is, Length Shift automatically deletes the thing whose length was shifted. Very cute, I know.

In the ABC song, Length Shift only applies to "aitch"; that's the only one where there is a preceding syllable in the word, containing a short vowel for the length to shift over onto.

(I slipped in a subtle consonant change from [h] to [x], also. A glottal fricative is quieter than a velar fricative, so I'm suggesting the teachers could have done that for audibility's sake.)

-+e: Teachers add a vowel.

In trying to communicate the sound of a letter, it would be better to use no other letters' sounds and just use the actual audible events, the phonetic outputs that may be only parts of sounds, but the ones in particular that are uniquely associated with that phonological unit, such as voiceless bursts alone for /p,t,č,k/, or [h] for /h/.

Then we might more easily teach a phonetic alphabet using examples of sounds which are unambiguous as to what is intended. There are two problems with this seemingly better approach.

First, to say '[h] is the sound of "h"' would be ineffective messaging, unless ear is close to mouth and environment is reasonably quiet. Often such conditions are not available. I've tried it, and it doesn't work.

Second, an adjacent, added vowel is especially unavoidable in trying to communicate the idea of a voiced consonant: any consonant-adjacent voicing (since the consonant is voiced) is actually phonetically identical to a vowel, so it's hard to do without. That's why they are called con-sonant, because they come with a sonant, with a vowel.

Therefore, supposing oneself a teacher in a classroom, in trying to give a example, or perhaps instead of an example, an uptakeable name, for nearly-silent sounds like [h], especially in front of a group of restive children, it helps to add a nice, projecting (loud) adjacent sound, a vowel. My own many attempts to say the sound of a consonant emerge in forms like [Ca], [Cə], [əCə] and [aCə]. These adjacent vowels confuse matters, since the raise the question for listeners, what is Tom communicating, the consonant, or the vowel, or the syllable(s)? But the environment and the need to be heard in that environment overrule the need to be correctly understood. Adding adjacent vowels helps also with voiceless consonants which are inaudible to listeners in any sufficiently noisy environment or remote distance.

So we add a vowel to the consonants to make them teachable in noisy environments. Not because our intention is to give them names, but to make them distinctively audible. That was the crux of the problem.

Because now the student can say back [e:č], and teacher feels duty is done, but the student has not understood [h]. Worse, inserting a vowel beforehand yields an unsatisfactory result for non-continuant consonants. Imagine loudly projecting a syllable [əb] without release or post-release voicing or vowel: the consonant is much less perceptible since it has no release, burst, following-vowel transition or coloration, etc. And if the environment prevents a listener from perceiving such swallowed consonants, then the speaker will not there be able to communicate it without post-consonantal voicing. On the other hand, if you want to put a vowel BEFORE the consonant as name or example of that consonant, it CAN work IF that consonant is a continuant, because you can hold it continuously, as indeed the /eC/ set, FLMNSX can be held continuously, thereby emphasizing and disambiguating the sound even in poor auditory environments.

These considerations control the forms of the ABC song.

To summarize: the key adjustments from the sounds alone here are functional. In other words, it may not be that the author intended names, rather it appears that the author was forced to add adjustments while trying to communicate sounds. It is those adjustments that have ended up confusing us all, and that now come down to us in what we now know as the names of the letters.

Perhaps we might return to the original intention, and teach with letters according to their sounds. Letters that are a phonetic alphabet. That's what IPA First tries to do.

Etc.: There remains a so-far-unexplained residue of monosyllabic letter names:

• AIOUWY the names of vowels and glides, which are exceptions to the rule of having an adjacent [e] added for audibility. These have good reason to be exceptional, namely that according to phonetic naming, naming the sound by using the sound itself, those letter names should use those very vowels that they are names of, [a, i, ɔ, o|u, W, wi], (including E:[e] is also consistent), and

• JKQR: [ja, ka, kyo, ar] fall out of this reconstruction, applying the general rules backwards. What makes them exceptions to the rule of adding [e]?

K has reason to be specially differentiated from C and S since it is the remainder of C when S is removed, by that old Latin palatalization rule c->s/_{i,e}. It couldn't be preceded by [e] since it is not a continuant and we only let continuants FLMNSX be preceded by [e] in this game; and it can hardly be followed by [e] without the same palatalization rule applying again and losing the [k] of it. Hence K:[ka].

And Q, that doubly-redundant letter (since it always precedes U, so it's redundant, and since it could have instead been written with K, so that's redundant), is confused by its near-universally-adjacent letter U. U is named [yo] not [o] or [u], so the [k] sound in Q and the [yo] name in U are combined as this reconstructed name for Q, [kyo], which is a confusing mess, but at least a natural one.

Finally, [ar] gets a note below.

In this way, we could reconstruct a day in history, centuries or millenia ago, in which the ABC song was actually an honest attempt to give the actual sounds of the letters, adjusted by:

• 1) the projection requirement that adds a minimal (about to be deleted, after all), short [e] vowel adjacent to each non-vocalic letter (that is, excluding AIOUWY), with its position..:

• 2) before the consonant for some continuants FLMNSX (counting X, [ks] as continuant due to the included [s]), but not other possible continuants CHVZ,

• 3) otherwise after the consonant.

• 4) JKR in this treatment would previously receive an [a] to distinguish them from G, C, and S/H respectively, which bleeds the rule (1). In particular:
  • The ambiguity of J:je vs G:je would motivate the ABC-song-writer to change the vowel in one of them.

  • Similarly the potential confusion of C:[se|ke] with K:ke would motivate a name change like K:[ka].
  • Finally note that final [&flap;] r may be phonetically devoiced and indeed ended with a fricative, as in Mexico City Spanish "ir" (to go): [i:š]. If such final devoicing (and frication) were part of the language (which may not have been English!) at that time (long before /r/ vocalization as discussed in Veatch 1991), then R:[er] and S:[es] would be potentially confuseable thus motivating the Author to use a different vowel for R:[ar] than for S:[es], to maintain the distinction teachable.

    Incidentally the claims regarding JKR of sound modifications for functional reasons are attributed to the author of the ABC song, the namer of the letters, carrying out his or her task of making distinctive names for a song to be used in teaching. Exceptionally giving names with [a] to JKR in the reconstructed ABC song, is the author's work -- not due to historical sound changes which, much research has shown, mysteriously ignore such functional reasons for keeping sounds apart. Historical sound changes, such as mergers, which create homophony, that is, mass confusion, between formerly distinct pairs of words, are innumerable.

I can say it more strongly: if this is NOT true then there sure are a ton of remote coincidences you must explain. The Great Vowel Shift is not up for debate, nor length shift and final e deletion. When you apply those backwards to do a historical reconstruction of what we know as letter names, what you get are forms which clearly and obviously are attempts to say the sounds themselves, with the minimum adjustments required by the acoustic environment of classroom communication. I don't see any reasonable alternative. So:

The ABC Song was originally intended and designed to communicate and teach a phonetic alphabet phonetically: a list of sounds by their sounds! -- Not, as we have come to know it, a list of letters by their names.

Clearly, the random cross currents of many centuries of linguistic history has obscured this basic intention.

Those who believe in and insist upon using and teaching the ABC Song as some kind of a foundation for literacy teaching, should reconsider their view given the reconstructed and evident intention of its author, namely, that the path to teaching literacy is by teaching a phonetic alphabet phonetically, by teaching a list of sound symbols by their actual sounds.

Songs are fun, and I am not so close-minded as to deny the utility of names for things including names for letters and sounds. We sometimes do have to talk ABOUT things, after all. But the task of reading goes from the letters we read to the sounds they signify, and inserting the wide detour of letter naming into that path makes it much longer and more difficult, and, according to my belief, causes most English illiteracy.

So let's teach a phonetic alphabet first.