The previous post argued that the "good services are verbs" heuristic, while valuable as a corrective to bureaucratic naming, obscures as much as it reveals. But that argument assumed a stable boundary between verbs and nouns - one word class for action, another for things. Eve Clark and Herbert Clark's (1979) study of denominal verbs shows that boundary to be far more porous than either the heuristic or its critics typically acknowledge.
A denominal verb is a noun that has come to be used as a verb: to blanket the bed, to bottle the wine, to bicycle to work. English generates these freely and prolifically - about a fifth of all English verbs were originally nouns (Pinker, 1994). What is remarkable is not that this happens but how readily speakers create and understand denominal verbs they have never encountered before. Clark and Clark's (1979) examples range from the everyday ("He kenneled the dog") to the improvised ("We all Wayned and Cagneyed", "She would not try to stiff-upper-lip it through"). In every case, listener and speaker converge on a meaning despite the verb being novel.
The denominal verb convention
How does this convergence work? Clark and Clark (1979) propose that it is governed by a convention in David Lewis's (1969) sense. Their formulation is precise: in using an innovative denominal verb, the speaker means to denote the kind of situation that they have good reason to believe the listener can readily and uniquely compute, on this occasion, on the basis of their mutual knowledge, in such a way that the parent noun denotes one role in the situation and the remaining arguments of the verb denote other roles.
Several features of this convention matter for service design. First, the meaning of a denominal verb is not fixed in the word itself but shifts with context, time, and circumstance. "To porch the newspaper" means something specific on a particular occasion because speaker and listener share knowledge about porches, newspapers, paper boys, and the current topic of conversation. Second, the convention relies on mutual knowledge - what both parties can reasonably be expected to know. Third, the interpretation must be uniquely computable: the speaker must have good grounds for believing the listener will arrive at one meaning, not several competing ones.
This is essentially a theory of pragmatic inference grounded in shared world knowledge. It depends not on the word alone but on what people know about the objects, institutions, and practices the word names.
Classification by case role
Clark and Clark (1979) classify their corpus of over 1,300 denominal verbs by the case role that the parent noun plays in the implied situation - using, explicitly, Fillmore's (1968) case grammar. The classification reveals several major categories.
Locatum verbs are those where the parent noun names the thing that gets placed: to blanket the bed puts blankets on the bed; to butter the bread puts butter on the bread; to wallpaper the room puts wallpaper on the room. The parent noun fills the objective case in a locative relationship.
Location verbs work in reverse: the parent noun names the place. To kennel the dog puts the dog in a kennel; to shelve the books puts books on a shelf; to bottle the wine puts wine in a bottle. The parent noun fills the locative case.
Instrument verbs name the tool or means: to bicycle to work uses a bicycle; to hammer the nail uses a hammer; to sponge the counter uses a sponge. The parent noun fills the instrumental case.
Agent verbs name the performer or entity whose characteristic behaviour is invoked: to ape someone is to behave as an ape does; to nurse a patient is to act as a nurse does; to parrot a phrase is to repeat it as a parrot would.
Goal verbs name the endpoint or product: to powder something is to reduce it to powder; to flour the board is to cover it in flour treated as a goal state.
This classification matters because it shows that the noun-to-verb conversion is not arbitrary. The case role that the parent noun plays constrains what kind of verb it can become - and this constraint is systematic enough to generate predictions about which nouns can be verbed and which cannot. As I explore in the case grammar post, Fillmore's roles reveal the argument structure that any verb implies. Clark and Clark demonstrate that denominal verbs inherit this structure from their parent nouns.
Predominant features and world knowledge
What determines whether a noun can successfully become a verb? Clark and Clark (1979) argue that the answer lies in what they call the predominant features of the objects that nouns name. People have generic theories about concrete objects, specifying three basic aspects: their physical characteristics, their ontogeny (how they come into being), and their potential roles (what they are typically used for).
The predominant feature of a category is the feature that most distinctively characterises it relative to neighbouring categories. For bricks, the predominant features are their box shape and child's-shoebox size. For orphans, it is a fact about ontogeny: their parents died before they were raised. For vehicles, it is a fact about potential role: they are used for transportation.
When a noun becomes a verb, the denominal conversion typically exploits the predominant feature. Instrument verbs exploit the potential role of the named object (what it is typically used for). Location verbs exploit the physical characteristics of the named place (what makes it suitable for containing or holding things). The more salient the predominant feature, the more readily the conversion succeeds.
This has a direct implication for service naming. The reason bureaucratic service names like "RIDDOR" or "Immigration Health Surcharge" fail is not merely that they are nouns rather than verbs. It is that they lack accessible predominant features from which a listener could compute an associated situation. There is no shared world knowledge about what one does with a RIDDOR. The noun cannot be verbed - not because of a grammatical limitation but because the convention has no material to work with. The mutual knowledge is absent.
By contrast, a name like "Blue Badge" succeeds precisely because it names an object with recognisable predominant features: it is a thing you display on a car dashboard, it relates to parking, it signals a particular status. One could readily say "to Blue Badge a car" and be understood. The noun-to-verb conversion is available because the world knowledge is shared.
Gärdenfors and the force patterns of denominal verbs
Peter Gärdenfors (2017), building on Kiparsky (1997), extends this analysis into the framework of conceptual spaces. Gärdenfors argues that denominal verbs refer to generically intentional activities: the verb "paint" involves the intentional covering of a surface with paint, which is why "The explosion painted the workers red" sounds odd. An explosion cannot have an intention.
Kiparsky (1997) draws a further distinction between "true" denominal verbs and "apparent" ones. True denominal verbs like "box" require the literal object: you cannot box a present in a paper bag. Apparent denominal verbs like "shelve" are "bleached" - you can shelve a book on a windowsill, not just on a shelf. Gärdenfors's analysis is that the apparent denominal verbs refer to the force patterns associated with the named object rather than to the object itself. "Shelve" means to put something on a thin, flat, horizontal, elevated surface that counterbalances the weight of the object - a description of the characteristic force relationship, not of the shelf per se. Similarly, "to hammer" refers to the force pattern used with hammers, which is why "He hammered the desk with his shoe" makes sense even though no hammer is involved (Kiparsky, 1997).
This connects to Gärdenfors's (2017) single-domain thesis: a verb root refers to a convex region of vectors in a single semantic domain. Manner verbs refer to force vectors (how something is done); result verbs refer to result vectors (what changes). Denominal verbs, on this account, typically lexicalise the force pattern associated with an object's canonical use. The object's predominant features - the same features Clark and Clark identified as enabling the denominal convention - determine which domain the verb enters.
What this means for the verb heuristic
The Clark and Clark analysis reframes the "good services are verbs, bad services are nouns" discussion in several ways.
First, it dissolves the sharp boundary the heuristic presupposes. If nouns routinely become verbs through a productive and systematic process, then asking whether a service should be named as a verb or a noun is the wrong question. The question is whether the noun chosen has the kind of shared world knowledge - the predominant features, the accessible force patterns - that would allow users to compute the associated activity. "Apply for a Blue Badge" works not because "apply" is a verb but because both "apply" and "Blue Badge" evoke recognisable situations through mutual knowledge.
Second, it explains why some bureaucratic names resist conversion while others do not. "Google" became a verb almost immediately because the object's predominant feature - its role as a search tool - was universally accessible and the associated activity was singular and distinctive. "Universal Credit" resists verbing not because it is a noun but because the system it names has no single predominant feature from which a canonical activity could be computed. The institutional complexity defeats the convention.
Third, it connects naming to the argument structure of action. When "to Blue Badge" succeeds as a verb, it inherits a case grammar: there is an agent (the person), a patient (the car or parking space), an instrument (the badge itself), and a set of institutional preconditions (eligibility, application, assessment). The noun does not just label the service; it implies the entire frame of roles and relationships that the service instantiates. This is the connection to frame semantics - the denominal verb convention is, in effect, a mechanism for evoking frames through single words.
Fourth, and perhaps most importantly for design, it suggests that the material and institutional infrastructure discussed in the previous post is not just context for verbs but the very substance from which verbs are made. Denominal verbs work because nouns carry world knowledge - about objects, their physical characteristics, their typical roles, their institutional settings. The richer and more shared that knowledge, the more readily the noun-to-verb conversion succeeds. Service names that work are not just verb-shaped; they are anchored in objects and institutions that users already understand.
The limits of denominal conversion
There are limits to how far this analysis extends. Clark and Clark (1979) focused on concrete nouns - things with palpable physical characteristics, ontogenies, and roles. Many service nouns are abstract ("benefit", "assessment", "referral") and their predominant features are institutional rather than physical. Whether the denominal convention extends smoothly from concrete objects to institutional abstractions is an open question.
There is also the problem of pre-emption. Clark and Clark (1979) note that an innovative denominal verb will normally be blocked if it is homonymous with a well-established verb. You cannot use "fly" as a denominal verb (from the noun "fly", meaning the insect) because the existing verb "fly" pre-empts it. In service naming, similar pre-emption may occur when institutional terms already have strong competing meanings.
Nominalisation - the reverse process, turning verbs into nouns - creates its own problems. As Peace (1999) argues, nominalisation is a kind of naming that turns an action into a thing, and in doing so can obscure the agency, the temporality, and the relationships that the verb made visible. "Assessment" hides who assesses, what is assessed, by what criteria, and to what end. The noun strips the verb of its argument structure. Service names built from nominalisations - and government services are full of them - may suffer from exactly this loss.
References
Clark, E.V. and Clark, H.H. (1979). When nouns surface as verbs. Language, 55(4), 767-811.
Fillmore, C.J. (1968). The case for case. In Bach, E. and Harms, R. (eds.) Universals in Linguistic Theory. Holt, Rinehart and Winston.
Gärdenfors, P. (2017). The Geometry of Meaning: Semantics Based on Conceptual Spaces. MIT Press.
Kiparsky, P. (1997). Remarks on denominal verbs. In Alsina, A., Bresnan, J. and Sells, P. (eds.) Complex Predicates. CSLI Publications.
Lewis, D. (1969). Convention. Harvard University Press.
Peace, R. (1999). Surface tension: Place/poverty/policy - from "poverty" to social exclusion: Implications of discursive shifts in European Union poverty policy 1975-1999. PhD thesis, University of Waikato.
Pinker, S. (1994). The Language Instinct. William Morrow.