Different past tense for various *et words

Question

The past tense of Bet, Let, and Set are Bet, Let, and Set; but the past tense of Jet, Net, Pet, Vet, and Wet are Jetted, Netted, Petted, Vetted, and Wetted. Is there a reason for this difference?

Answer 1

There is a set of English irregular verbs that have four unusual characteristics:

1. they consist of only one syllable
2. they end in a dental stop, /d/ or /t/
3. they have a lax or low vowel - /ɛ ɪ æ ɔ ʊ ə ɚ/
4. they are not inflected for past tense or past participle

Examples are the verbs bet, let, set, as noted, but also others, like cut, spit, and cast.
There are around 25 verbs in the list:

• beat, bet, bid, burst, cast, cost, cut, fit, hit, hurt, knit, let, put, quit,
  rid, set, shed, shut, slit, spit, split, spread, thrust, upset, wed.

As can be seen, the verbs jet, net, pet, wet, and vet are not in the list,
so those are regular verbs with regular past tense forms in -ed.

Why those verbs and not others? Because that's the way irregularities work.
Irregular verbs do not happen by rule -- irregular means 'not by rule'.

Answer 2

Many of the words with irregular past forms are very old (derived from Old or Middle English), while most of those with regular forms have been introduced more recently. For example, "vet" is fairly new, originating from the late 19th century and surprisingly seems to be related to "veterinarian".

Irregular

let: from Old English lǣtan (Wiktionary)

bet: from 16th-century criminal slang, likely from abet or Old English bætan (Wiktionary)

set: from Old English settan (Wiktionary)

hit: from late Old English hyttan, hittan

Regular

vet: late 1800s

jet (as a verb): 1690s, from French jeter "to throw, thrust"

net (as a verb): 1400s

pet: Sense of "to stroke" is first found 1818

wet: from Old English wætan (*This may be an exception, but as noted in a comment, the irregular past form "wet" is used in some senses.)

Answer 3

WaterMolecule's answer seems to be right in attributing the difference to historical causes. There seem to be complicated reasons for the development of the two kinds of inflection for verbs ending in -t or -d.

Verbs that originate from "class I weak verbs" tend to contain the vowel E and have irregular past-tense forms ending in T or D

It seems that the verb set was a class I weak verb in Old English. Verbs of this class originally had a -j-/-i- suffix that caused umlaut on the vowel (which explains why many of them have the vowel e). But not all of the irregular verbs like this end in -et: there are also ones ending in -it (such as knit, although it also has an alternative form knitted) and -ut (shut).

The presence of this -j-/-i- suffix was relevant to the development of the past tense form because the i vowel in the past tense form eventually ended up being lost, causing the past-tense suffix to come directly after the consonant at the end of the verb root. When the verb root ended in t or d, the final consonant of the verb root merged with the past-tense suffix (or the past-tense suffix was dropped), and shortened the preceding vowel if it wasn't already short.

Meet, bleed, breed, feed all have long vowels in the present tense, and set has a short vowel in the present tense, but historically, these verbs all belonged to the same conjugation class.

The development of other classes of verbs is also relevant

There was another class of weak verbs where it appears to have been usual for the past tense to end in -ed.

And there were also several classes of strong verbs that had members ending in -d or -t that developed irregular past-tense forms similar to those seen in class I weak verbs. Let is in this category, as is shed.

An account of the history of these verbs is given in "The Regularization-Through-Derivation Effect and the Historical Development of Verbs in the West Germanic Languages", by David Fertig:

The distinction between regular and irregular weak verbs in modern English is a partial continuation of the distinction between Class-2 and Class-1 weak verbs in OE, and thus ultimately a partial continuation of the distinction between verbs formed with the -ja-derivational suffix and those formed with -ō-. [...] In early Middle English, after reduction of unstressed vowels, the descendents of the class-2 (-ō-) verbs had a connecting schwa before the dental suffix in the past tense, whereas in the descendents of Class-1 (-ja-) verbs, the dental suffix was generally attached directly to the end of the verb stem, with no connecting vowel. [...]

Many of the T/D+Ø weak verbs in Modern English were Class-1 verbs in Old English, including spit, knit, shut, set, sweat, wet, spread, probably slit, perhaps cut, and partially put. Others were from Old Norse, borrowed into OE as Class-1 verbs and first attested (with no connecting schwa) in early ME, including hit, thrust, cast, and rid. Still others were strong verbs in OE and apparently owe their T/D+Ø past forms to reanalysis of the suffixless strong past. These include shit, burst, let, bid, and shed. Finally, quit, cost, and hurt were borrowed from French into early Middle English at a time when the general inflectional-class distinction between verbs with and without a connecting schwa was still very much alive. Most French borrowings were conjugated with a connecting vowel, including most of those ending in -d or -t, such as butt ('hit'), mend, tend, fend, plead, treat, cheat, pleat, waste, taste, as well as others that rhymed with weak verbs of the no-connecting-vowel class, such as glean. A considerable minority entered the no-connecting-vowel class, however, for reasons that are not entirely clear. In any case, 22 of the 25 verbs in the modern T/D+Ø class belonged to the no-connecting-vowel class in early Middle English. The modern irregular forms of these 22 verbs therefore have nothing to do with any analogical attraction of the T/D+Ø class. Only three verbs appear to have genuinely been attracted into that class: split was borrowed from Dutch in the late 16th c.; bet, of controversial origin, is also first attested in the late 16th c.; and fit, originally derived from the adjective in the late 16th c., has, in the 20th c., developed T/D+Ø forms in certain meanings.

Similarly, breed, bleed, feed, lead, read, speed and meet are all from OE weak class 1, making plead the only verb that has (in some dialects) been attracted into this subclass. Among the other T/D-with-laxing verbs, hide, light, and alight were in weak class 1 in OE, while the remaining three, slide, bite, and shoot, were strong.

(pp. 20-21)

It seems that verbs ending in -t or -d that were formed after the Middle English era have mostly followed the -ed pattern.