Innovations versus idioms

Innovations that are contextuals are common in other areas of the language too. Among noun compounds we find egg plate (said of a plate decorated with pictures of eggs), mystery woman (said of a woman who wrote mystery stories), and umbrella head (said of a man who wore an umbrella on his head). Downing argues that these cannot be handled by derivation. Among denominal adjectives we find Stonehengey (said of a man with conservative views), dinnery (said of food more appropriate for dinner than for other meals), and Beethoveny (said of music that sounded as if Beethoven could have composed it). Among possessives we find Justin’s bus (the bus Justin watched yesterday), my tree (the tree I always point out as we pass), and Erie Stanley Gardner’s lawyer (Perry Mason, the lawyer Gardner created). And among shorthand expressions we find two Picassos (two people who paint much like Picasso), my street (the majority of adults on my street), and San Francisco (the basketball team based in San Francisco). These four categories are just a sample of the places in English where innovations are common.

Like denominal verbs, these categories span the continuum from innovation to idiom (or well-established construction). Among noun compounds, for instance, contrasting with the innovations are virtually opaque idioms like bulldog, thunder-mug, and tinderbox; while birdcage, dogsled, and bookshelf probably lie in the middle. But just what is this innovation-idiom continuum? We will consider this question briefly from four different points of view: the history of English, the acquisition of English by the child, the processing of English by the language user, and the synchronic description of English. Each viewpoint yields a somewhat different answer. In our discussion we will use only denominal verbs as illustrations, but the points we make are equally applicable to other types of innovation as well.

5.1. historical change. Nearly every denominal verb in the lists above, we assume, was introduced into English as an innovation. This is generally confirmed by the OED — though there may be exceptions. Each verb, then, arrived at its present form by a complex historical process we will call idiomatization. Because the process is gradual, each verb passes through several stages on its way from innovation to idiom; and because fresh verbs are being introduced into this process all the time, at any moment there are verbs at each stage. We tentatively propose six stages in this process, and illustrate each with examples from present-day-English.

(a) complete innovations. Denominal verbs begin their lives as complete innovations. To our ears, Wayne, Cagney, pie, erratum, and bargain-counter are complete innovations. Some may remain nonce forms, while others may proceed to the next stage.

(b) near-innovations. When a speaker or group of speakers uses an innovation more than once, and it is recognized as the same form, then we have a near-innovation. Thus Herb Caen has used houseguest and chopstick more than once, and readers have begun to recognize these as ‘his’ words. Once again, these may or may not proceed to the next stage.

(c) half-assimilated transparent idioms. Some verbs become transparent idioms for one group of speakers, but remain innovations for everyone else. Key in the data, for example, appears to be idiomatic among computerniks, as noted earlier, but it is still perceived as an innovation by the rest of us. Satellite the broadcast was probably idiomatic within CBS before it was used on television, where it was perceived by most viewers as innovative. For verbs at this stage to move on to the next, they must generally be transparent to the outgroup, as both key in and satellite are.

(d) assimilated transparent idioms. Verbs like bicycle, truck, crowbar, and paperclip appear to be fully assimilated into English as verbs, even though their interpretations are fully transparent. We would know what they meant even if we hadn’t heard them before.

(e) partly specialized idioms. A great many verbs are no longer entirely transparent because they have become partly specialized. Four we have already mentioned are smoke a pipe, park the car, ground the plane (‘keep down’), and land the plane (‘bring down’). Land, for instance, was originally used in navigation to mean ‘disembark’. Earlier in this century, it was transferred (along with many other navigational terms) to aeronautics — for airplanes putting down on land from the air. When airplanes were designed to put down on water, the idiomatization was apparently so complete that it didn’t seem odd to ‘land’ on water.

(f) opaque idioms. Many verbs have come to have such obscure denominal origins that, for most people, they are completely opaque. Boycott, dun, diddle, and fudge are based on the names of long forgotten individuals, and most people now think of them as verbs pure and simple. Charleston and shanghai are based on familiar place names, but few people know the connection. Riddle and ferret are based on common nouns unfamiliar or unknown to most people. Slate, badger, and beef are based on familiar nouns, but few people are aware of the noun-verb relations.

Not all denominal verbs will pass through the stages in this order, or even complete the series. A verb like key in could lose its transparency within the computer community even before it is assimilated into English; and verbs like ground and eye will probably never become opaque. The majority of denominal verbs, it seems, have become assimilated just because they are virtually transparent. This makes them readily understood by people who have never heard them before, especially children, and they are therefore readily maintained with a stable meaning. Yet when there is a lexical gap that could usefully be filled, opaque verbs like lynch, boycott, and pander are also readily maintained — but as verbs unconnected to nouns.

Because one of the main functions of idiomatization is the creation of special-purpose verbs, dictionaries are strewn with partly specialized idioms. The verbs formed from shell, as listed in the OED, are quite typical: ‘remove (a seed) from its shell’, ‘expel (a growth)’, ‘shed (milkteeth)’, ‘drop out of a shell’, ‘remove the shell of, ‘bring forth as from a shell’, ‘scale off’, ‘enclose in a shell’, ‘furnish with shells for collecting oyster spawn’, ‘spread shells on’, ‘bombard with shells’, and ‘drive out by shelling’. Many of these senses are utterly unfamiliar to modern ears, as we would expect. Such specialized senses should be abandoned when the object is no longer is use, as in to archie or to roneo (Partridge); when the object no longer has the particular use, as in Zeppelin the fleet, used in World War I to mean ‘bomb the fleet from Zeppelins’ (Jespersen); or when the special allusion is no longer recognizable, as in Copenhagen the fleet ‘sink without warning’, or Burgoyne a general ‘capture’ (Partridge). Dictionaries probably underestimate the number of specialized uses that have arisen by this process.

5.2. language acquisition. Children learning their first language almost certainly do not distinguish innovations and idioms the way adults do. They appear to treat many adult idioms as if they were innovations, and many adult innovations as if they were idioms. These two ‘errors’ have important consequences.

Children produce innovations from a very early age, and some of these conform to the adult constraints on innovations. [...] From our observations, many (perhaps most) children produce innovative denominal verbs, though they vary greatly in how often they do so.

Because of this early facility, children may produce and understand particular denominal verbs very differently from adults. Consider these four possibilities. First, they may learn the noun hammer, and then create and use the verb hammer, even though it is idiomatic for adults. Here their innovation corresponds to the adult’s transparent idiom. Second, they may hear the verb truck, and from their prior knowledge of the noun truck interpret it as an innovative denominal verb. Again, their innovation corresponds to the adult’s transparent idiom. Third, they may learn the noun dial as applied to clocks, bathroom scales, and gas meters, and separately learn the verb dial the number for push-button (and dialless) telephones, never realizing that the two are related. In this case their opaque idiom corresponds to the adult’s partly specialized idiom. Fourth, they may hear a near-innovation like Let’s chopstick for dinner again — in the absence of chopsticks — and interpret it as an idiom meaning ‘have Chinese food’. In this instance, their opaque idiom corresponds to the adult’s near-innovation. There are other possibilities, too, including those in which the child and adult agree in their treatments. The point is that, as children create their own system of language, they may alter the status of verbs as innovations or idioms.

Children, then, may play a role both in keeping language stable and in speeding language change. They probably contribute to language stability when they treat adult idioms as innovations. For example, in producing or understanding bicycle, truck, and jeep as innovations, they may prevent those verbs from deviating too far from the paradigm ‘go by [vehicle]’, from becoming partly specialized idioms like land, ground, and smoke. On the other hand, children probably spur on language change when they treat near-innovations and partly specialized idioms as opaque. Thus, treating the noun and verb dial as unrelated, they may contribute to the acceptance of dial as an opaque idiom; in treating chopstick as an opaque idiom, they may effectively be introducing it into English as just such an idiom. Here as elsewhere, children are probably instrumental in both maintaining and changing language.

5.3. language processing. In speaking and listening, people must certainly process innovations and idioms very differently. Consider comprehension. For innovations, at one extreme, people must create completely new meanings: confronted with Wayned, they cannot retrieve a ready-made meaning from their mental lexicon, since they have none for verbs they have never heard before. If the line we have taken is correct, they must construct the meaning of Wayned in conformity with the innovative denominal-verb convention. For idioms, at the other extreme, listeners must retrieve ready-made senses: they must look for boycott as a verb in their mental lexicon, since they don’t have the parent noun Boycott available. Parallel arguments hold for innovations and idioms in production.

Between the two extremes, it isn’t always clear what should happen. Transparent idioms, for example, could be processed either as innovations or as opaque idioms; both processes would lead to the right interpretation. But these verbs are so frequent, and so well assimilated as verbs, that they are presumably processed most of the time like opaque idioms. In comprehension, it would be inefficient for their meanings to be recreated each time when they could be retrieved from the lexicon ready-made — like most other word meanings. Indeed, this almost has to be true if we are to account for pre-emption. When a verb has a common idiomatic sense, that normally takes precedence over certain innovative senses. Thus, although on reflection the noun bottle may be recognized in bottle the beer, this information isn’t normally used in the process of saying or interpreting it.

Yet, in the right circumstances, transparent idioms may be processed as innovations. Imagine hearing We used everything — we snowmobiled, snowshoed, and skated, as opposed to We did everything — we hiked, drank beer, and skated. The first sentence contrasts the three instruments, and invites skated to be processed as an innovation on a par with snowmobiled and snowshoed. But the second contrasts three activities, and invites skated to be treated as an opaque idiom on a par with hiked and drank. With contrastive stress, the noun origins of a verb are readily brought to the fore. In We didn’t use our car — we taxied to the airport, the instruments are contrasted, a car vs. a taxi, while the rest of the meaning of taxied, ‘went by X’, is backgrounded (Watt). How transparent idioms like skate and taxi are processed, therefore, may depend on the context. This may also be true of partly specialized idioms.

The presence of innovations, near-innovations, and idioms sometimes processed as innovations offers a distinct challenge to most theories of comprehension and production. These theories implicitly assume that all word meanings are available ready-made in the mental lexicon. That assumption is clearly wrong. If innovations of all types are as common and as readily understood as we suppose, then no theory of comprehension or production can be complete unless it handles them in the natural course of the relevant processes. Right now this goal seems far off.

5.4. synchronic description. Contemporary English has denominal verbs at each stage of idiomatization — from full innovations, like bargain-counter, to opaque idioms like boycott. How much of this information belongs in the synchronic description of English? If such a description is supposed to characterize the ideal speaker / listener’s ‘knowledge’ of English, we have a problem — because, as applied to denominal verbs, ‘knowledge’ has at least four interpretations. First, it could mean ‘always-used information’: in comprehending bargain-counter, listeners probably always use the fact that it comes from the noun. Second, it could mean ‘usable information’: for taxi, listeners may not normally use the fact it comes from a noun, but in contrastive contexts they can. Third, it could mean merely ‘awareness on reflection’: many people are surprised when they are shown that the noun and verb land are related — but, on reflection, they could probably figure this out for themselves. Fourth, it could mean simply ‘intellectualizable information’: most people could not figure out for themselves the relation between boycott and Captain Boycott; but when informed by a dictionary or a specialist, they would in some sense ‘know’ the denominal character of boycott. These successively more inclusive criteria for ‘knowledge’, of course, lead to different synchronic descriptions of denominal verbs.

Forced to make a choice, we would probably opt for a synchronic description that included only ‘usable information’, knowledge that is or can be accessed in normal language use. But it may be more defensible to include all information about denominal verbs, yet distinguish which parts are known at which level of knowledge. The synchronic description, in any event, will have to do more than just dichotomize denominal verbs into innovations and idioms.

There is a further complication: note that idioms and innovations can co-exist with the same parent noun. The idioms shelve the books and shelve the closet coexist with innovative uses of shelve, as in While maneuvering through the door, the carpenter shelved his assistant in the back (‘poked with a shelf’). The complication is that these idioms often shade off into innovations, with no clear boundary. In Alex forked the peas into his mouth, for example, fork has an idiomatic sense ‘convey in the normal manner by means of a fork’. We all know, of course, what the normal manner is — which, if this sense is to be idiomatic, must be defined independently of any context. But if it is mutual knowledge that Alex is a child who uses a fork two-fisted, or backwards, or only as a means to catapult food into his mouth, the speaker would intend fork to mean ‘convey by a fork in a two-fisted manner’, or ‘convey by a backward fork’, or ‘catapult by means of a fork’. At what point has Alex strayed too far from the ‘normal manner’? At what point have we moved from the idiomatic sense of fork to an innovative one? There is no obvious answer, and this adds still another complication to the synchronic description of denominal verbs.