There is a maxim so deeply embedded in the design culture of the twentieth century that it has ceased to feel like a proposition and started to feel like a law of nature: Form follows Function. Louis Sullivan stated it in 1896 for skyscrapers, the Bauhaus turned it into a worldview, and since then designers, architects, software engineers, and HR professionals alike have believed that you first define the purpose and then let the shape follow. The problem with that sentence is not that it is wrong. The problem is that, taken on its own, it faces the wrong direction.
What if the more interesting truth is the reversal? Function follows Form. Not as a refutation of Sullivan but as a complement that the behavioral sciences, cognitive psychology, and UX research of the past few decades have established with growing clarity: the form of an environment configures the behavior we then call function, and it does so before a single conscious decision has been made. The brain reads the shape of the world before it begins to think about the world.
What Gibson discovered in the field
James Gibson was a psychologist with a problem. The prevailing view held that the brain receives raw sensory data and then processes it into meaningful information. Gibson found this implausible. Through years of fieldwork he developed a different model: the brain does not perceive the environment neutrally but as a set of action invitations he called affordances.
A flat stone at the water's edge offers itself as a seat. A narrow passage invites walking through. A protruding handle signals gripping; a flat plate signals pushing. These invitations are not interpreted, they are received, below the threshold of conscious cognition, and the brain calibrates its operating mode accordingly before a person has decided whether to sit, pass through, or push. Affordances are not properties of objects. They are the relation between object and perceiver, and that relation is reliable enough to be engineered.
The environment does not present itself to the brain as neutral scenery. It presents itself as a script. Whoever designs the environment writes the script.
That may sound like an abstract epistemological claim, but its practical consequences are immediately measurable. In UX research the effect is direct: a large, empty input field invites long text. The same field, narrow and single-line, invites short text. Users do not consciously decide to write less when the field is smaller. The field decides for them, long before they have thought about it.
Thaler, Sunstein, and the cafeteria
Richard Thaler and Cass Sunstein formalized this mechanism and brought it into mainstream behavioral economics. Their concept of choice architecture describes the same phenomenon at the level of options, sequences, and defaults: the way choices are arranged determines which choice is made, even when all options are formally equally accessible.
The most cited example is the school cafeteria. When researchers changed the arrangement of food in a cafeteria without altering the offering itself, consumption shifted significantly. What stood at the front and at eye level was chosen more often. What stood at the end or below the line of sight was more often left untouched. The children did not make different choices because they thought differently about nutrition. The form of the display made the choices for them.
This is the behavioral economics version of Gibson's affordances: defaults are the most powerful affordances in human decision environments, because they simultaneously activate cognitive inertia and loss aversion. Whoever sets the default determines the behavior of the majority, not because the majority lacks autonomy, but because the brain allocates the energy cost of deviation only when it sees a clear gain ahead.
The aesthetics-performance paradox
There is an experiment that is almost provocatively simple. In the late 1990s, researchers installed identically programmed ATMs in Japan and Israel, differing in exactly one feature: the visual design of the user interface. One version was aesthetically appealing, the other was functionally plain but identical in operating logic.
Users of the more aesthetically appealing machine made significantly fewer errors and subsequently rated the system as more usable. The systems were identical. The form was not.
The halo effect, usually discussed in the context of person perception, operates on artifacts too. Aesthetically appealing forms are automatically perceived as functionally superior, and that perception changes actual interaction: users approach beautiful systems with less tension and more confidence, which increases available cognitive capacity and reduces error rate. Form produces confidence, and confidence produces better performance.
That is not decoration. That is cognitive architecture.
The open office and the failure of good intentions
Few domains illustrate the mechanics of "function follows form" as precisely as the history of the open-plan office. The idea was compelling: remove physical barriers, encourage chance encounters, overcome silo thinking through spatial proximity. The form was supposed to produce the function of collaboration.
Ethan Bernstein and Stephen Turban at Harvard Business School studied exactly that, using sociometric sensors and server data, in two companies that transitioned from traditional office layouts to open structures. The result was the opposite of the intention: face-to-face interaction fell in both cases by approximately 70 percent. Digital communication via email and messaging rose by 20 to 50 percent.
What the planners had not accounted for is that the brain in an environment of permanent observation displays a very particular protective response, not openness but its opposite: withdrawal, headphones as social barriers, demonstrative disengagement from surroundings as a means of preserving concentration. The form called forth the precise behavior it was intended to prevent, not because of insufficient cooperative willingness among employees, but because of the neurobiology of social behavior under observation conditions.
Bernstein compares this to actors performing for an audience who must construct an imaginary fourth wall between themselves and the spectators in order to forget themselves and inhabit the character. In the open office, employees collectively built such a wall because the form of the space left them no other option.
94 percent belongs to the system
The economist and quality pioneer W. Edwards Deming spent decades of industrial fieldwork developing a number so uncomfortable that it rarely appears seriously in HR discourse: approximately 94 percent of all performance deviations in organizations are attributable to the system, meaning processes, tools, information flows, and physical conditions. About 6 percent trace back to individual factors.
This is not a romantic excuse for individuals. It is an empirical distribution of causal responsibility, and it carries far-reaching practical implications. When an organization has a performance problem and responds with a motivation workshop, a coaching program, or a new bonus structure, it is statistically addressing the 6 percent share and ignoring the 94 percent. That is not merely inefficient; it is actively destructive, because it confronts people in a system they cannot change and frames as personal failure what is a system failure.
Deming's classic Red Bead Experiment makes this transparent. A task is designed so that the outcome is fully determined by the system: whoever must draw white beads from a container in which red beads predominate cannot physically influence the result. Yet praise, criticism, bonuses, and warnings are dispensed as if the workers were responsible for the outcome. The joke is mordant, but it is not a joke. It is standard practice.
UX as behavioral architecture: what this means for design
When we bring these insights together, a very clear design-theoretic consequence emerges that extends well beyond conventional usability discussions. UX design is not the design of a surface on which people make decisions. UX design is the configuration of the conditions under which certain behaviors become more or less probable, before conscious decision-making begins at all.
That is not the same thing. The first framing implies that designers provide possibilities and users choose. The second implies that every design decision is already a behavior-relevant choice that the user in the next step merely confirms or expends effort to revise. To build an interface where purchase is the next logical step is not to have provided a neutral action option. It is to have configured the behavior.
This becomes legible in the debate around dark patterns, which extends well beyond illegal manipulation tricks. A dark pattern in its purest form is not a lie in the user interface but an information architecture that shapes form so that a specific function arises from sheer inertia. A "cancel subscription" button positioned three clicks deeper than "renew subscription" does not manipulate the user's beliefs. It manipulates their energy budget. The form produces the function the provider wants, at the cost of the function the user wants.
The difference between an ethical nudge and a dark pattern is not the mechanism. The mechanism is identical. The difference is whose interests the form serves.
Applying the principle: five diagnostic questions
For organizations and design teams that want to take "function follows form" seriously, a concrete diagnostic path emerges. Before any measure targets individual motivation, training, or culture work, five questions should be posed and honestly answered.
First: what defaults are currently active, and what behavior do they produce automatically? Every system has default states that become norm through cognitive inertia. The double-sided printing standard that cuts paper use by 40 percent. The default meeting slot of 60 minutes that stretches meetings to 60 minutes even when 30 would have sufficed. The pre-filled option in the checkout flow. These defaults were rarely consciously designed, but they are the most powerful behavioral drivers in the system.
Second: where is a performance problem being attributed to the person when the environment has not been examined? Deming's 94/6 rule should be the first counter-check before a coaching budget is released. What does the system structurally prevent, and do the people working in it know?
Third: what affordances does the physical or digital environment emit, and are these the signals that produce the desired behavior? A call center whose acoustic environment generates permanent stress will not produce calm, empathetic customer conversations regardless of how many active listening courses are completed. An interface that hides complexity will not produce informed decisions.
Fourth: in what sequence is information presented, and what anchor does that sequence set? If a budget discussion begins with cost items, investment decisions are made differently than if it begins with impact potential. The sequence is behavior before the content is.
Fifth: what would change if the next employee review began with an audit of the work environment rather than a conversation about the person? That question alone, applied consistently, would invert the structure of many performance dialogues.
The responsibility that knowledge creates
There is a consequence rarely discussed honestly in design conversations. If form produces behavior, then everyone who designs form also bears moral responsibility for the behavior that form evokes. That applies to interface designers as much as to office planners, process owners, and organizational developers.
Lawrence Lessig articulated for digital space what holds for all designed environments: architecture is a form of regulation that operates preventively rather than punitively. A system that makes a certain behavior technically impossible or structurally very costly regulates behavior more reliably than any norm, conviction, or sanction that kicks in afterwards. That is an enormous power, and like every enormous power it demands a serious engagement with whose interests it serves.
The question, then, is not only whether a design "works" in the sense of conversion rates, click paths, or productivity metrics. The question is what function the form produces, for whom, at whose cost, and whether the people whose behavior is being configured know it. Sullivan was right that form and function belong together. What he did not state explicitly is that causality runs in both directions, and that the direction we pay less attention to is the more powerful one.