The limits of a memorable heuristic
Duncan Stephen's recent post, "Services are verbs combined with nouns", addresses the nouns, verbs best way to describe service ontologies problem, and which I also wrote about in states and services back in November. Lou Downe's influential framing - "good services are verbs, bad services are nouns" - has shaped a generation of government digital work. But as Duncan observes, the original problematic service names ("RIDDOR", "Immigration Health Surcharge") failed because they were jargon, not because they were nouns.
Every task, Duncan notes, is actually a verb combined with a noun: "Apply for a Blue Badge", "Check your National Insurance record". The insight is deceptively simple, but it opens onto deeper questions about what verbs and nouns actually reveal - and what they systematically obscure.
What verbs capture: language as action
The "verbs" framing draws, whether consciously or not, on a rich intellectual tradition. J.L. Austin's How to Do Things with Words (1962) established what linguists call speech act theory: the insight that language doesn't merely describe the world but actively shapes it. As Gond and Cabantous (2015) explain, "a performative utterance is one 'in which to say something is to do something; or in which by saying something we are doing something.'" When a registrar says "I now pronounce you married", or when a service declares "Your application is approved", these aren't descriptions of pre-existing facts - they are declarations that constitute the institutional reality they name.
Austin distinguished three dimensions of any speech act: the locutionary (what is literally said), the illocutionary (what is done in saying it - requesting, promising, declaring), and the perlocutionary (the effects produced by saying it). This matters for service design because services are saturated with illocutionary acts: users make requests, providers make commitments, systems issue declarations that change status. Iqbal (2018), whose promise framework I explored in the first post of this series, makes this explicit: "Services are a set of promises... Agreements are based on promises. Services require agreements, therefore services are also based on promises".
Winograd and Flores (1986) applied speech act theory directly to systems design through their "Conversation for Action" framework. As McCarthy and Wright (2007) summarise, they "describe organizational life as comprised of speech acts combined into 'recurrent patterns of communication.'" Work, in this view, consists of networks of requests and commitments - someone asks for something, someone else promises to deliver, and the cycle continues until the commitment is discharged or broken. This anticipates service design's focus on what users are trying to accomplish and how services respond.
Linguistic theory adds another lens. Fillmore's frame semantics (1968) showed that verbs encode structured relationships between participants. As Gärdenfors (2017) explains, extending his conceptual spaces framework into linguistic semantics, "the agent and the patient of an event model are the two most central examples of thematic roles". When we say "book an appointment", the verb implies an agent (the user doing the booking), a patient or affected object (the appointment being created), and potentially an instrument (the booking system mediating the action). Understanding these argument structures helps designers identify what roles a service must support and what entities it must represent.
So verbs aren't arbitrary - they capture something real about how services create value through action. The question is what they leave out.
The hierarchy problem
Activity theory offers a more structured critique. Leontiev's hierarchy distinguishes three levels: activities (oriented to motives), actions (oriented to goals), and operations (oriented to conditions). A "verb" in service design typically captures the action level - "apply", "check", "book" - but understanding requires connecting to the motivating activity above and the automatised operations below.
Kaptelinin and Nardi (2012) explain that "activities are composed of a sequence of steps, each of which may not be immediately related to the motive". The classic example is Leontiev's hunting party: the beater who scares game away performs an action seemingly opposed to the goal, but it only makes sense within the collective activity of hunting where beaters drive prey toward ambushers.
For service design, this means "book an appointment" (action) only becomes meaningful when connected to "get access to healthcare" (activity) and mediated by "select date from calendar" (operation). Verb-focused design tends to flatten this hierarchy, treating all actions as equivalent when they occupy different structural positions in human activity. A ward nurse "updating patient status" and a consultant "declaring medical optimisation" are both verbs, but they sit at different levels of the activity system and carry different institutional weight.
The abstraction hierarchy alternative
Cognitive Work Analysis offers an even more developed alternative. The abstraction hierarchy represents a work domain at five levels - from functional purpose at the top through abstract function, generalised function, and physical function, down to physical form at the bottom. Moving down the hierarchy answers HOW; moving up answers WHY.
Burns and Hajdukiewicz (2017), whose ecological interface design approach I explored in an earlier research post, emphasise that "EID has adopted the Abstraction Hierarchy as a fundamental way to analyze the environment, the work domain". Crucially, this is not a task analysis - it represents invariant constraints, not sequences of action. The work domain analysis captures what remains true regardless of specific tasks, events, or operator strategies. As Vicente put it: the "stage" rather than the "play".
Last year, I tried to apply this thinking to the TCH journey map, layering physical evidence, view layers, model layers, and abstract conceptual layers alongside the action layer. The innovation - if it deserves that name - was showing these layers simultaneously, revealing means-ends relationships between them. A physical artefact (printed list) supports a view (discharge dashboard) which represents a model (patient list filtered by status) which embodies a concept (medically optimised patients) which enables an action (identifying patients ready for discharge).
This is what verbs alone cannot capture: the systemic architecture that makes particular actions possible and meaningful.
What nouns reveal
Duncan's argument for nouns is sound, and supported in lots of other places. In OOUX (Object-Oriented User Experience), which he references through Sophia Prater's orca method, "calls to action" - the verbs - are described as "the affordances of your objects". A social media post calls a user to like it, comment on it, bookmark it, share it. The object comes first; the actions are what the object affords. This is affordance theory in action, and connects to the triadic model I explored in the opening post of this series.
This inverts the verb-first logic. Instead of asking "what do users need to do?" and working backwards, object-oriented approaches ask "what are the things in this system?" and then enumerate what can be done to them and how to emphasise these capabilities. Kim Goodwin's (2011) design framework makes this explicit, listing object definition, relationships, states, actions, and attributes as the components of a conceptual model.
The connection to my earlier in this view is more explicit states and services: objects have states. An application object might have a status attribute that takes values like "draft", "submitted", "approved". Understanding the state space of objects - what conditions they can be in, what events trigger transitions - reveals things that verb-focused analysis misses: error states, edge cases, loops, dead ends. The statechart perspective asks not just "what can users do?" but "what condition is this entity in, and what can happen to it from here?"
Practice theory's integration
Practice theory offers perhaps the most sophisticated integration of verbs and nouns. Shove, Pantzar, and Watson's (2012) The Dynamics of Social Practice argues that practices - not individuals or behaviours - are the proper unit of analysis. Practices are constituted by three interconnected elements: materials (objects, infrastructures, tools), competences (know-how, skills), and meanings (social significance, aspirations). This integrates some other things I've written about before about the primacy of meaning in design.
Katrini (2018) summarises: "practices in general consist of three main elements: materials, competencies, and meanings". This maps onto our grammatical categories: nouns for materials, verbs for the competent performance of actions, and something like adjectives and adverbs for the meanings and manner of practice.
The insight is that these elements change together. Showering practices changed not because attitudes changed (meanings) or because people decided to shower differently (actions), but because piped hot water, bathroom design, and cleanliness norms co-evolved. For service design, this suggests attending to the material infrastructure and the competences alongside the actions - and understanding that services recruit practitioners into careers of increasing commitment, not just discrete interactions.
What this means for some of my dashboard work
Returning to the dashboard work that has occupied these weeknotes, the implications become concrete.
The digital dashboard operates in the register of status and metrics - patient flow, discharge readiness, bed occupancy. It represents nouns: patients, beds, tasks, statuses. These are the objects that persist between interactions, the entities whose states the system tracks. The doctor's notebook that I wrote about in December operates as a record of the actions - it captures verbs: who to call, what conversations to have, which results to chase. These are the speech acts still to be performed, the commitments not yet discharged.
Neither alone is sufficient. The dashboard without the notebook loses the action orientation that makes information useful - knowing a patient is "medically optimised" means nothing without knowing what to do about it. The notebook without the dashboard loses the systemic view that makes action coordinated - knowing you need to "chase bloods" doesn't help if you can't see which patients are waiting on results. The seam between them is where the actual work happens - translating between registers, connecting nouns to verbs, objects to actions.
Duncan's conclusion resonates: "by codifying our verbs as service patterns and nouns as information models, we have the potential to unleash great efficiencies".
But the hierarchy matters too, and it exposes a deeper problem with the verb-plus-noun formula. When Duncan says a service is "Apply for a Blue Badge" or "Check your National Insurance record", at what level of abstraction does that verb-noun pair sit? "Apply" is an action in Leontiev's sense - a conscious, goal-directed step. But the activity that motivates it (securing mobility support, planning for retirement) and the operations that constitute it (filling in form fields, uploading a photo, clicking submit) are invisible in the name. The verb-noun pair captures the middle of the hierarchy and flattens everything above and below it.
Not all verbs are equivalent; some are activities, some are actions, some are operations. Not all nouns are equivalent either; some name purposes, some name functions, some name forms. The abstraction hierarchy reveals which level each element occupies - and therefore how it relates to others.
Consider the board round through Leontiev's hierarchy. Ensuring safe and timely patient flow through the hospital is an activity - oriented to a motive, organising everything the board round does. Within that, declaring a patient medically optimised, requesting a social care assessment, escalating a delayed transfer of care are actions - conscious, goal-directed steps with institutional weight. And scanning the dashboard for patients, filtering the list by discharge status, ticking a checkbox on the discharge checklist are operations - automatised routines shaped by the tools at hand. The same nurse "updating patient status" could be performing an action (consciously reclassifying a patient's readiness) or an operation (routinely entering data at the end of a shift) depending on context. The level isn't fixed by the verb - it depends on where attention and intention sit.
The nouns stratify similarly, but something interesting happens as you move up. At the form level, the nouns are unambiguously object-like: the digital dashboard, the printed patient list, the doctor's notebook, the ward whiteboard. These are the objects in the OOUX sense - things you can point to, pick up, interact with; things that have states, afford actions, and persist between interactions. A patient list is an object whether it's printed or digital; a whiteboard is an object whether anyone is looking at it or not.
At the function level, the nouns become more abstract but still recognisable as entities: discharge readiness, patient status, bed occupancy. These are trackable and representable - the kind of thing a dashboard column or a status badge can display - but they're no longer objects in the same sense. "Bed occupancy" isn't a thing you can hand to someone; it's a property of a system, a function that multiple forms can serve. The dashboard tracks bed occupancy; so does the whiteboard; so does the ward clerk's mental model. The function persists across its material realisations.
At the purpose level, the grammatical categories start to blur entirely. Patient safety is nominally a noun, but it's really a value; timely discharge is a nominalised verb; continuity of care names an ongoing process rather than a thing. These aren't objects at all - they're the reasons the objects exist. The higher you climb the abstraction hierarchy, the less cleanly verbs and nouns separate; actions become reified into purposes, and objects dissolve into abstract constraints. The companion posts on the theoretical foundations and case grammar explore this blurring in more depth.
A fuller grammar
The strongest critique of "good services are verbs" isn't that verbs are wrong, but that they're incomplete. Duncan gets us to "verb + noun". Activity theory gets us to "motive + goal + conditions". Practice theory gets us to "materials + competences + meanings". The abstraction hierarchy gets us to "purpose + function + form".
The full picture is something like: verb + noun + adjective + preposition + institution + practice + network + means-ends hierarchy. The elegance of "good services are verbs" was always a simplification - useful for cutting through bureaucratic jargon, powerful as a reorienting human and action-centred heuristic, but incomplete as an analytical framework. The companion research posts on the theoretical foundations and case grammar in service semantics develop this critique in more depth.
What would it mean to take this fuller grammar seriously in service design? At minimum, it suggests enriching our representations to show not just actions but the objects they operate on, the states those objects can be in, the means-ends relationships connecting actions to purposes, and the material infrastructure that makes action possible.
The journey maps, abstraction hierarchies, and stakeholder mappings I've developed through this TCH work are attempts in that direction - imperfect, but reaching toward a richer representation than verbs alone allow.
References
- Austin, J. L. (1962). How to Do Things with Words. Oxford University Press.
- Burns, C. M. & Hajdukiewicz, J. R. (2017). Ecological Interface Design. CRC Press.
- Downe, L. (2020). Good Services: How to Design Services That Work. BIS Publishers.
- Fillmore, C. J. (1968). The case for case. In E. Bach & R. Harms (Eds.), Universals in Linguistic Theory. Holt, Rinehart & Winston.
- Gärdenfors, P. (2017). The Geometry of Meaning: Semantics Based on Conceptual Spaces. MIT Press.
- Gond, J.-P. & Cabantous, L. (2015). What do we mean by performativity in organizational and management theory? The uses and abuses of performativity. International Journal of Management Reviews, 18(4), 440-463.
- Goodwin, K. (2011). Designing for the Digital Age: How to Create Human-Centered Products and Services. Wiley.
- Iqbal, M. (2018). Thinking in Services: Encoding and Expressing Strategy Through Design. BIS Publishers.
- Kaptelinin, V. & Nardi, B. (2012). Activity theory in HCI: Fundamentals and reflections. Synthesis Lectures on Human-Centered Informatics, 5(1), 1-105.
- Katrini, E. (2018). Sharing culture: On definitions, values, and emergence. The Sociological Review, 66(2), 425-446.
- Leontiev, A. N. (1978). Activity, Consciousness, and Personality. Prentice Hall.
- McCarthy, J. & Wright, P. (2007). Technology as Experience. MIT Press.
- Shove, E., Pantzar, M. & Watson, M. (2012). The Dynamics of Social Practice: Everyday Life and How It Changes. SAGE.
- Stephen, D. (2025, June 26). Services are verbs combined with nouns. Duncan Stephen. https://duncanstephen.net/services-are-verbs-combined-with-nouns/
- Vicente, K. J. (1999). Cognitive Work Analysis: Toward Safe, Productive, and Healthy Computer-Based Work. Lawrence Erlbaum.
- Winograd, T. & Flores, F. (1986). Understanding Computers and Cognition. Addison-Wesley.