before the ·ing suffix

Question

Page 1579 of the CambridgeGEL reads

For die the ie is the default spelling, so that the replacement works in the opposite direction: ie is replaced by y before the ·ing suffix.

Why was a replacement in either direction diachronically necessary? (I infer its necessity from the wording so that)

To avoid having three vowel letters in a row: dieing looks bad, or something. There's no accounting for English spelling, however -- most of it is arbitrary. — John Lawler, Jul 04 '20 at 16:42
@FumbleFingers just wanted to point out the overall preferance for -ie vs -y, where both are possible — GJC, Jul 04 '20 at 17:28
I don't know if there are any better known or more clear-cut examples than *stymieing. The "ugly" spelling also dominates with birdieing* and *sortieing, but they're both pretty unusual as verb* forms, compared to *stymie. And with more common verbs like to boogie* we again get the "opposite", as per *dying*. — FumbleFingers, Jul 04 '20 at 17:47
@FumbleFingers If you’re Scottish, as a bonus you also get ghillieing and the ineffabling tuilyieing and spulyieing to boot. :) — tchrist, Jul 04 '20 at 17:55
@FumbleFingers BTW, how can I get the NGram total relative freq? E.g. 3:1, 2:1, etc. — GJC, Jul 04 '20 at 18:04
@FumbleFingers die is one syllable long (therefore it's stressed) and has a diphthong, while boogie has the stress on the first syllable and the last syllable ends in a monophthong. The last vowel of both words is represented by what CambridgeGEL calles a two letter symbol — GJC, Jul 04 '20 at 18:25
Nobody has yet taken into account the usual syllabi(fi)cations. According to Encarta, fre·er, fre·est, freed, free·ing, frees. — GJC, Jul 04 '20 at 19:59
You can do limited "math" in NGrams. I'm not great at it, but I think it amounts to enclosing your search string in (brackets), then treating that as a numeric value (= number of hits) which can be added / subtracted / multiplied / divided by some other "value" represented by another bracketed search string - or just by a constant, as in (search string) / 5 — FumbleFingers, Jul 05 '20 at 11:10

score 1 · Answer 1 · answered Jul 05 '20 at 18:47

It's not a matter of "necessity" per se: the spelling conventions that happened to become standardized for English are largely arbitrary.

I think the following tendencies lie behind the spelling alternations for words like die and try (none of them are very strongly synchronically active):

Word-final "i" is strongly dispreferred. Therefore, in most words where it would occur, "y" is used in its place (e.g. cry, rely, copy can be thought of as "standing for" *cri, *reli, *copi).
Non-grammatical words tend to have spellings that are more than two letters long, so words that end in an "i" sound and would otherwise be spelled with only two letters are spelled with a "silent e" at the end: e.g. die, lie, tie, vie, pie. A few words end in -ye instead of -ie: rye, dye, lye.
Silent e is usually present or added between an i (word-finally, -y) and the suffix -(e)s.

This seems similar also to the use of -es in the verb forms does, goes, and various noun plurals in -oes.
Silent e is usually absent or "dropped" before the suffix -ing.

This is more of a default, rather than a strong dispreference for words ending in eing, and there are multiple exceptions that I have already tried to cover in an answer to a previous question: Is "ageing" the only exception? Other suffixes that show similarly variable behavior are (e)y, (e)able, (e)ish.
the sequence "ii" is strongly dispreferred and so "y-ing" is used in place of "i-ing"

Both the dispreference for word-final i and the dispreference for ii might be related to the concept of "minim letters" and the idea that in certain types of handwritting these letters are less visually distinctive, especially when they occur in sequence, than other letters.

As I said, these are obviously not hard constraints on contemporary spelling so there are plenty of words that don't follow these tendencies; e.g. ski, skiing, Latinate plurals like radii, abbreviations like bi.

It seems that by and large, the spelling of verbs ending in /ai/ seemed to settle down from about 1600. From the OED, the chronology of the spelling of Dying,: 1435, deyng; 1450, dieng; 1556, dying; 1563, deand; 1590 and onwards, dying. Spying, 1338, spiyng; 1398, spying; c1430, spiyng; 1495, spienge; 1523, spyeng; a1568 and thereafter, spying. Lying 1382, liggynge; 1455, liggyng; c1470, liggand; 1488 lyand; 1495, lying; 1564, lieng; 1583, lieng;1619 and thereafter lying. — Greybeard, Aug 04 '20 at 23:01

before the ·ing suffix

1 Answers1