Pattern to Old English verbs-of-making-adjective?

Question

The other night (after hearing someone on TV say "smoothen"), I noticed that a fair number of Anglo-Saxon-derived adjectives tend to come in pairs where the more "distinguished" or "active" adjective becomes a verb by adding -en while its opposite adjective is the same as the verb form.

Positive	Negative
roughen	smooth (*smoothen)
sharpen	dull (*dullen)
widen, broaden	narrow
brighten	dim (*dimmen)
thicken	thin
fasten, tighten	loose (but also "loosen")
sweeten	sour

My question is: Is there a known or theorized linguistic reason for this pattern? I have at least three hypotheses: (1) The -en marker connotes human activity: one must act positively to brighten and sharpen things, but they will dim and dull on their own. (2) The marked words are more frequent and/or more important; it's useful to distinguish [please] brighten from [it is] bright, but the distinction between [please] dull and [it is] dull is less practically relevant and therefore verbally unmarked. (3) This is entirely a case of confirmation bias.

---

Of course the pattern isn't universal; I can think of more exceptions to the rule than I can followers. I'll list them here, just in case it's useful to someone and/or someone thinks of a reason to move one up into the table above.

Positive	Negative
heat (*hotten)	warm, cool
speed (but cf. modern "quicken")	slow
blacken, redden	whiten
darken	lighten
gladden	sadden
quicken, liven	deaden
stiffen	—
sicken	—
straighten	crook, bend, curve, curl
moisten, dampen	wet (*wetten), dry (*dryen)
heighten (*tallen), lengthen (*longen)	shorten
strengthen (*strongen)	weaken
— (*heavyen, *weighten)	lighten
—	deafen
— (louden?)	soften
flatten	—
ripen	—
fatten	— (*lean)

Answer 1

This is a partial answer due to the wide-ranging examples in your question. Based on the evidence I've seen, this is an accident that depends on two things: (a) when the verb in question was formed and (b) whether an -en form emerged from the time when that suffix was most generative in English. Beyond that, when both forms exist, it's a matter of luck or guessing whether the -en form is more successful.

The -en suffix is primarily generative in Early Modern English (EME)

Something you'll notice if you look up words like widen and sweeten is that they are relatively recent formations from the 16th century, and not themselves Old English formations. The Oxford English Dictionary elaborates in its definition of the suffix -en (5):

Etymology: Most of the words in sense 1 [darken, deepen, harden, madden, moisten, widen] seem to have been formed in late Middle English or early modern English, on the analogy of a few verbs which came down from Old English or were adopted from Old Norse; e.g. fasten v. < Old English fæstnian ; brighten v. < Old Northumbrian berhtnia ; harden v. < Old Norse harðna.

There are other ways these verbs are formed too, such as from nouns or directly from Germanic strong verbs, but for the adjective-verb pathway, most of your examples are early modern. I'll put the first use of each verb in brackets to illustrate a pattern:

Positive	Negative
roughen [1582]	smooth [1340] (*smoothen)
sharpen [1530]	dull [c1374] (*dullen)
widen [1566], broaden [1727]	narrow [eOE]
brighten [intrans. OE, trans. 1567]	dim [a1300] (*dimmen)
thicken [1425]	thin [c900]
fasten [OE], tighten [1727]	loose [a1225] (but also "loosen")
sweeten [1552]	sour [1340]

I see a pattern. The left column almost always comes from late Middle English or Early Modern English, with two from Old English; the right column comes largely from Middle English with two from Old English. The -en suffix we are looking at wasn't generative in the period of Middle English the verbs from the right column come from, but it was generative in the period after. If a word like sour had formed as a verb in the 1500s, we may have indeed seen souren.

Actually, that's not a bad idea. Sometimes those verbs also formed.

A few of the verbs in the right hand column did develop -en forms, even if they didn't persist

If you go looking, you can actually find some -en forms of those latter verbs in the Early Modern period:

• smoothen, v., from 1655. Several subdefinitions.
• dullen, v., from 1832. Marked as rare.
• dimmen, v., undated. Marked as rare.
• loosen, v., in first definition rare/obsolete and from 1382 (Wycliffe Bible), in subsequent definition from 17th century from 1645 (Milton)
• souren, v., marked as northern dialect: "intransitive. To become sour." First use in 1570.

In most cases these remained rare or dialectical. Smoothen and loosen are somewhat common, and they seem to really catch on in the Early Modern period. Aside from some possible phonological reasoning or the fact the -en forms had prior competition from simpler forms without -en, it's an accident or guessing game to explain why souren isn't at least like loosen. Similarly, it's a matter of guessing why some words in the left hand column do have common verbs without -en (rough, v., e.g., "rough the passer") but others don't (sharp, v., whose uses are almost all obsolete).