Four-way ablaut in English

[alice]

How many groups of four closely-related English words can you find where the only difference aside from affixes is the vowel? Obviousy, most of them will contain the three principal parts of a strong verb, and they should all continue the same Germanic root. Here are two to start with:

• sing, sang, song, sung
• strike, struck, stricken, stroke

[Pabappa]

hmmm .... i think i saw a different version of this post earlier where you had a stricter qualification ... does that mean that sing/sang/song/sung is literally the only one that differs only by the vowel?

I have step/steep/stop/stoop, but the inclusion of stop is arguable since it connects with the others at a stage much further back and it is not an absolutely certain cognate. There's also archaic stap but Wiktionary doesnt consider it an independent word.

likewise, string/strang/strong/strung works, but i feel even worse about that one, since it again connects much further back than the others and strang is merely a variant of strong.

[KathTheDragon]

So far the only example I've thought up is bind/bound/band/bond.

Edit: apparently bend is a fifth member.

[zompist]

stink/stank/stunk/stench

(I know, got a consonant change in there.)

[Moose-tache]

But technically also from a vowel, yes? It counts!

A fun almost-example is choose, chose, chosen, choice. Astute readers will note that "choice" is from French. But some linguists believe that it comes from a Vulgar Latin borrowing from Gothic that is cognate with English "choose," with the change in vowel being attributed to historical developments in Old French that yielded /oi/. So if we allow the ball to go out of bounds a bit, we could say that this is ablaut, just with another language providing the vowel gradation instead of internal sound change.

[Travis B.]

But technically also from a vowel, yes? It counts!

A fun almost-example is choose, chose, chosen, choice. Astute readers will note that "choice" is from French. But some linguists believe that it comes from a Vulgar Latin borrowing from Gothic that is cognate with English "choose," with the change in vowel being attributed to historical developments in Old French that yielded /oi/. So if we allow the ball to go out of bounds a bit, we could say that this is ablaut, just with another language providing the vowel gradation instead of internal sound change.

The question is this, though - why does choice have /tʃ/ and not /k/ or /s/, considering the usual outcomes of palatalization in Gallo-Romance and OF?

[Richard W]

stink/stank/stunk/stench

That's productive - drink/drank/drunk/drench. Of course, counting umlaut may be a bit naughty.

A variant, courting disqualification: speak/spake/spoken/speech.

In British English at least: see/saw/sight/seer. (Well, I say 'seer' and 'sear' the same.)

Exploring 'closely' - atom/atomic/tome/tmesis.

Natively probing 'closely': wind/wound/wend/wander

Foreign language again (Scots, or maybe Yorkshire): ride/rode/ridden/raid.

Not so obvious: heal/health/whole/hale.

5!: do/does/did/done/deed.

I'm not sure if this counts: man/men/women/postman.

[Curlyjimsam]

bear, bore/born/borne, bairn, bier, birth

[Znex]

The question is this, though - why does choice have /tʃ/ and not /k/ or /s/, considering the usual outcomes of palatalization in Gallo-Romance and OF?

The original vowel was /au/ (/a/ undergoes a different palatalisation in Old French):

PG *kaus(i)janą >> VL *causire > OF choisir

Compare CL causa > OF chose

[Richard W]

Three/third/thrice/thruppence

Twice/twain/twelve/twin (I presume two/tuppence don't count because the /w/ has been dropped).

An/one/only/any/none: An might count for two.

Sit/sat/set/seat. Site probably contains the same morpheme, though I don't know how you'd test. Historically, it's unrelated.

[Travis B.]

The question is this, though - why does choice have /tʃ/ and not /k/ or /s/, considering the usual outcomes of palatalization in Gallo-Romance and OF?

The original vowel was /au/ (/a/ undergoes a different palatalisation in Old French):

PG *kaus(i)janą >> VL *causire > OF choisir

Compare CL causa > OF chose

Ah gotcha. (I knew /a/ caused different palatalization in OF, but I didn't know this word originally had /au/.)

[Pabappa]

Wow this is hard.

Possibly slip/slop/slope/slape, .... again, as with my other sets, there's one member that doesnt quite fit, as Wiktionary claims slape is a loan from Icelandic rather than a native word. If sleep is related .... which neither Wiktionary nor Etymonline claims, but I thought I read it somewhere .... the set could be expanded to six by including slept.

Looking at my first post, i thought i might be able to get strict into this, but no .... straight/straught/stretched are related to each other, but strict is separate from them all the way back to PIE.

The only other thing i can come up with now is feed/fed/food/fodder, and maybe somewhere somehow "fod" has been used in bare form but my guess is that the reflex of the bare form is "food".

edit: found one more .... God/good(bye)/giddy/Godiva

[Risla]

How about rise/rose/risen/raise?

[alice]

How about rise/rose/risen/raise?

"Raise", er, arose by umlaut, as did many other examples in this thread, but otherwise there's no reason why not.

[Richard W]

How about rise/rose/risen/raise?

"Raise", er, arose by umlaut, as did many other examples in this thread, but otherwise there's no reason why not.

Or one could be picky and reject raise as a loan from Old Norse. The native form is rear.

[Pabappa]

lie ~ lay ~ law ~ low.

Possible extended members of the set, in order of best fit to worst, are ...

lair. definitely cognate, but is it really a separate vowel?
lig, a dialectal variant of lie.
ledge and lodge. Etymologies are cloudy. I think lodge may be a merger of a French word with Old English logian, but Wiktionary and EtymOnline both list only the French path. If Im right, I suspect the OE sense was much along the lines of "i lodged the window pane in the groove" and the French provided the sense of lodging for people.
lawyer, if it can be considered to have a vowel distinct from law.

words like lean lien loan lawn lone loon all seem to be unrelated. the common placename ending -ley is not related either. nor are lore leer learn etc.

[KathTheDragon]

lair. definitely cognate, but is it really a separate vowel?

It is IMD, /lɛː/.

[kodé]

Reaaaaaally stretching it here, but: splish, splash, sploosh, splosh. I’ve used and heard the latter two quite a bit before, though since they’re iconic words and non-canonical at that, I’m not gonna die on this hill. OTOH, other vowels don’t sound possible between onset /spl/ and coda /S/ (using X-SAMPA because using tablet): *splesh, *spleesh, ?sploash, ?splush.

[Travis B.]

Reaaaaaally stretching it here, but: splish, splash, sploosh, splosh. I’ve used and heard the latter two quite a bit before, though since they’re iconic words and non-canonical at that, I’m not gonna die on this hill. OTOH, other vowels don’t sound possible between onset /spl/ and coda /S/ (using X-SAMPA because using tablet): *splesh, *spleesh, ?sploash, ?splush.

I am only familiar with splish, splash, and splosh myself...

[Pabappa]

with onomatopoeia you can pretty much do whatever you want, but some forms are more established than others. Ive seen SQUEEEEEEESH in a comic book, so from that we can build a set of squish/squash/squoosh/squeesh. I'd prefer to count only the ones that have at least some use outside of sound effects. So that means for this set only squish and squash count.

Sploosh is definitely a word, .... it's the sound a heavy object makes when dropped in water. Cant just use splash for everything. Likewise splish is for tiny objects. but although some vowels seem to be mostly ignored, i'd still say it's an open process .... if I drew a comic strip where an elephant, a penguin, a cat, a mouse ,and a spider all in sequence leapt into a swimming pool, i would want all of them to have different sound effects and might make a sequence like SPLOOSH! Splosh! Splash! Splish! Spleesh! just because I'd need something funny to finish it up.