My recent posting on two-part back-formed verbs has elicited queries about how and why these verbs arise. The first of these questions is easier to answer than the second, but I’ll take a stab at both.
(Much of this material has been discussed, on several occasions, by Neal Whitman on his blog Literal-Minded.)
Some background: there are two principal types of back-formed verbs: simple and two-part. These back-formings are then a kind of verbing — by subtraction, rather than by zero conversion (as in the verb verb itself) or suffixation (as in sissify and funeralize).
In simple back-formings, an affix (or apparent affix) is eliminated, to yield a new “base verb”: burgle from burglar, incent from incentive. The story is more complex for two-part back-formings (like the verb back-form itself), as I’ll argue below.
But some things generalize across the two types: (1) back-formed verbs seem to have begun to appear in any number early in the 19th century; (2) many (though not all) of them seem to have been originally playful (“humorous”, as the OED puts it) in tone, though a number of them (like burgle and enthuse) seem to have lost this tone, though some continue to be colloquial; (3) many of them are marked in the OED as “originally” or “chiefly” U.S. (or North American); (4) a huge number of them are in fact (relatively) recent; (5) like verbings in general, and in fact like innovated vocabulary in general, many of them quickly developed extended senses not immediately predicable from their sources (see my discussion of back-formed gay marry here).
The two-part back-formations themselves have (at least) two types of sources:
(1) noun-noun compounds, in particular ”synthetic compounds”, in which the first noun represents a non-subject argument of the base verb in the head noun, this head noun most often being a form in -ing (as in child-bearing and absentee voting) or -er (as in timekeeper and bartender) (often it’s hard to know which form is the source); and
(2) modifier-noun combinations where the modifier is “non-predicating” (see here) and the head noun is an abstract noun based on a verb (as in gay marriage and free association).
Synthetic compounding — in which a non-subject argument is, in effect, “incorporated” into a verb — is quite productive, though the considerable literature on them (which was spurred by the phenomenon lying in some sense on the borderline between morphology and syntax) has suggested some possible constraints on it. Synthetic compounds start life usable only as nouns (though the uses of the -ing compounds are more complex than that), but they are easily subject to reanalysis as derived from compound verbs — giving things like child-bear, absentee vote, timekeep, and bartend.
A similar reanalysis gives rise to new two-part verbs in type (2): gay marry, free-associate. Non-predicating modification is very open, though some combinations are hard to interpret out of context — but when the combinations become conventionalized in certain meanings and reasonably frequent, again reanalysis is easy.
Both Neal Whitman in a blog posting and Ian Preston in a comment on my last posting in this blog have wondered why some back-formings occur and others do not. Whitman is even willing to asterisk some back-formed verbs.
As I have more and more experience with two-part back-formed verbs, I am less and less willing to label candidates as unacceptable or impossible. I keep finding surprises. I’m inclined to think that so long as the back-formed verb would be useful to speakers, sooner or later someone will come up with it.
Ian Preston asks:
Why is it that “nail clipper” or “goal scorer”, say, don’t seem to yield back-formed verb phrases when “hair dryer” and “goal keeper” do?
Ah, but the verb nail clip is attested:
How often to nail clip and bathe a housepet Beagle? She’s 2yo and I have rescued her. (link)
The examples of goal score I’ve found are not very clear. But wait a while and a good one will probably turn up.