by The elevator in my building has a door-close button that does nothing. I learned this the way everyone learns it, which is to say I pressed it for years under the impression that it was speeding up my departure. The button lights up. It makes a small click when pressed. It provides every sensory signal of function. What it does not do is close the door any faster than the door was going to close on its own. The elevators in most American buildings installed since 1990 have door-close buttons wired to nothing, because the Americans with Disabilities Act requires the door to stay open long enough for a person using a wheelchair or walker to enter, and the button that overrides that requirement is accessible only to the fire department with a key.
The elevator industry has not hidden this. A 2016 New York Times piece by Christopher Mele quoted Karen Penafiel of the National Elevator Industry confirming that the buttons in non-emergency elevators do not function for ordinary passengers. The buttons remain on the panels because removing them would require rewiring, because passengers expect them, and because a button that does nothing costs less to leave in place than a button that does something. The fire department button works; the button passengers press does not. Both are labeled the same way.
The phenomenon extends well beyond elevators. A February 2004 New York Times piece by Michael Luo, titled “For Exercise in New York Futility, Push Button,” reported that of the 3,250 crosswalk buttons in New York City at the time, more than 2,500 functioned as mechanical placebos and only about 750 still worked. The rest were deactivated when the city moved to computerized signal timing in the decades after 1980, and the buttons were kept in place rather than removed because removing them costs more than leaving them attached to nothing. A pedestrian pressing a crosswalk button on Sixth Avenue at Thirty-Fourth Street is performing a gesture. The gesture has no effect on when the light changes. The light changes on a fixed timer that does not know the pedestrian pressed anything.
Office thermostats are the third common case. HVAC contractors working in large commercial buildings have for decades installed decoy thermostats in zones where the actual temperature is controlled from a central building management system. A Wall Street Journal column by Jared Sandberg, who wrote the Journal’s Cubicle Culture column on office life, reported that a significant share of office thermostats were non-functional decoys, installed because building operators had discovered that employees who had a thermostat to adjust reported feeling more comfortable than employees who did not, regardless of whether the thermostat was connected to anything. The decoy thermostat was cheaper than the actual climate complaint.
What unites these three cases is that the placebo button is engineered to produce the sensation of causation without the mechanism of causation. The Harvard psychologist Ellen Langer spent much of her career documenting what she called the illusion of control, the human tendency to overestimate personal influence over outcomes that are actually random or automatic. Her 1975 paper on the subject in the Journal of Personality and Social Psychology remains the source text, and the elevator button has become an unintended memorial to her findings. The button is a compliance device. It gets the passenger through fifteen seconds of waiting without becoming the kind of passenger who kicks the door. The crosswalk button operates the same way. The thermostat operates the same way.
The object lesson is small. A button that does nothing is a minor embarrassment in the built environment. The interesting question is where else this logic operates, and the answer is most of civic life.
Public comment periods on federal rulemaking accept submissions from citizens by the hundreds of thousands. Agencies are required to read those submissions and consider them. In the vast majority of cases, the final rule reflects the draft rule. The comment did not change the outcome. Its filing was registered. Citizens who submit receive an acknowledgment. That acknowledgment was the point.
City council meetings in most American municipalities include a public comment segment during which residents address the council for two or three minutes each on matters of local governance. Council members are not required to respond. A council member is also not required to incorporate the comment into deliberation. Votes that follow get decided in committee or caucus before the comment begins. The resident goes home having pressed the button. Its light comes on. That door closes on the timer it was going to close on anyway.
Surveys distributed by employers after reorganizations, by airlines after delays, by hospitals after treatment, by universities after lectures, arrive with the implicit promise that the responses will influence future decisions. Response data gets aggregated and presented in quarterly reports to executives who have already made the next round of decisions. The survey is a button that lights up. No door closes any faster.
Electoral systems with gerrymandered district lines produce outcomes pre-determined by the shape of the district rather than the preferences of the voters. A voter in a district drawn to favor a party by twenty points is pressing a button connected to a timer set six years before the election. Ballots get counted. The vote does not change the outcome. An illusion of control persists, and the citizen who pressed the button walks out of the polling place with the civic sensation of participation.
The placebo button is not a conspiracy. No central authority installed these systems to deceive. An elevator button represents a cost-benefit decision by building owners. A crosswalk button persists as legacy infrastructure nobody retired. Public comment periods exist as procedural requirements written into administrative law in 1946. Gerrymandered districts result from partisan legislatures drawing maps within a legal framework the Supreme Court has repeatedly declined to revise. Each placebo button got there through a defensible local decision. The pattern they form together is a democracy where the buttons light up reliably and the doors close on the schedule they were always going to close on.
The dangerous version of this pattern is the one where citizens learn the buttons do nothing and continue pressing them. Compliance does not require belief. The door-close button still gets pressed by passengers who know it does nothing, because pressing it is what one does in an elevator, and because the alternative is standing silent in a small box with a stranger for fifteen seconds. Voting in a gerrymandered district still gets done by voters who know the district will produce the predetermined outcome, because voting is what one does as a citizen, and because the alternative is admitting the system is closed.
The recognition that matters is the one that separates the buttons that work from the buttons that do not, and the organizing that follows is the organizing that rewires the boxes rather than the gesture of pressing. Ballot initiatives that establish independent redistricting commissions are rewiring. Lawsuits that force agencies to respond substantively to public comments are rewiring. Municipal charters that require council members to respond to comment on the record are rewiring. The door-close button in the elevator is a lost cause and not worth the fight. A crosswalk button at Thirty-Fourth Street is not worth the fight either. Ballots and comment periods and council meetings are worth every fight we can bring, because the apparatus behind those buttons is still capable of connection, and the difference between a democracy and an elevator panel is whether the wires on the other side of the panel still reach the doors.
@boles
This is excellent. Well done and said. This kind of zombie function applies in other countries often too. Why? Because US companies design things for use under US law, so having everything they sell be compliant with their biggest market's rules means buttons here in Canada follow the same logic, even where our laws are different.
Very interesting and thought-provoking piece. Thank you.
Thank you, and the Canadian angle is exactly right. Otis and Schindler and KONE and Mitsubishi design for their largest single market and ship the same panel to Toronto and Berlin and Singapore. The button on the wall does not know which country it is in. Regulatory scholars call this the Brussels Effect when EU privacy law exports into US sites that serve European users, and the California Effect when CARB emissions rules become national because no automaker builds California-only cars. The American ADA elevator panel rides the same supply chain.
The political consequence is that the country writing the rule for the biggest market is, in practice, writing the rule for everyone in that market’s supply chain. The US wrote the rule that disabled the door-close button. Canadians inherit a nonfunctional button under a different legal framework, the same way Europeans inherit American car cupholder dimensions even though no one in Berlin commutes with a Big Gulp. The button does not close any faster on either side of the border, and the explanations behind it diverge in interesting ways depending on which country’s regulatory text you read first.
Thanks for the read!
@boles
You've found a reader. I'll be subscribing. Thanks for the great response.
So kind! Thank you! And, welcome!
@boles They, in fact, do not ship the same model to Berlin and Singapore. They do ship them to Canada. The US/Canada market (and probably Mexico?) for elevators has different standards that make elevators more expensive and the market more consolidated than the rest of the world.
Uytae Lee has a great video about the North America-specific problem.
You are correct and I owe a correction. My earlier comment overstated the geographic reach by claiming manufacturers ship the same panel to Berlin and Singapore. That was wrong. The North American elevator market runs on ASME A17.1 in the US and CSA B44 in Canada, which are harmonized to each other but diverge significantly from EN 81 in Europe and from the standards used across most of Asia. Manufacturers build region-specific products for those code zones. The US to Canada spillover is real, since Canadian code references the ASME standards directly and Canadian buildings receive what amounts to the same product the US gets. The transatlantic and trans-Pacific spillover I implied does not exist in the way I described.
Uytae Lee’s video is excellent on the North America-specific problem. He documents how the code differences combine with consolidated manufacturer relationships, restrictive permitting, and limited installer competition to produce elevators that cost three to four times what comparable European installations cost, with longer install times and fewer model options. The Canadian inheritance of the US-cost structure is one of the cleaner natural experiments in the field, since Canada could in principle have adopted EN 81 instead.
Thank you for the correction. The Brussels Effect and California Effect frames I cited are real, but the elevator example does not extend across the Atlantic the way I implied. It extends across the Canadian border, which is the precise pattern the original commenter was describing.
@boles For certain, and I agree the problem exists, but the distinction is useful in figuring out what can be done to improve things.
In elevators, as with so many things, the solution is to accept international standards so that we're not stuck in a siloed market dominated by one foreign country's eccentricities.
Agreed, and the framing of the United States as the foreign country whose eccentricities Canada is stuck with is the right way around. Most American writing on this topic treats ASME A17.1 as the neutral baseline and EN 81 as the international variant, which gets the geography of the problem exactly backward. The international standard is the international standard. The North American carve-out is the parochial one. Canada inherited it through proximity and code reference rather than through any deliberate choice to align with the smaller, more expensive, less competitive market, and the cost of that inheritance shows up in every elevator quote across the country.
Standards harmonization is the right remedy and probably the only remedy at scale. The economic case for adopting EN 81 in North America is strong on its own terms: a larger competitive pool of manufacturers, faster install times, lower per-unit costs, more model variety, and the regulatory savings of not maintaining a parallel code apparatus for a market segment that has no good reason to be parallel. The political obstacles are domestic and predictable. The ASME and CSA standards bodies have institutional interests in continuing to exist. The consolidated manufacturers benefit from the consolidation and will resist any change that opens the market. The trade associations have their constituencies. None of those obstacles is a good reason for Canadian buildings to keep paying three to four times the European price for the same vertical transportation, but each of them is a real political force that any harmonization effort would have to overcome.
The same code-island pattern applies to electrical standards, plumbing fittings, and a long list of other infrastructure categories where North America runs its own apparatus without measurable safety benefit. The elevator case is the cleanest because the consolidation and cost penalty are easiest to document, but the wider pattern is the more useful argument for policy purposes, since a coalition that wins on elevators can use the same playbook on the next category and the one after that.
Thank you for the correction and for staying in the thread to push the analysis forward. The distinction between domestic and international causes was exactly what was missing from my earlier framing, and the harmonization remedy you named is the right answer to the better-stated question.
@boles
My downtown has those crosswalk buttons with a red LED that says "Wait!" when you press it.
It's delivering instructions to me, not the traffic signals.
I still press it 3-4 times rapidly because it's fun.
Welcome, and the Wait LED is the perfect example of the second-generation placebo button. The first generation pretended to give you control over the signal. Your downtown’s version dropped the pretense and now openly issues instructions. The button used to lie about the direction of communication. The new one tells you where the actual authority sits, which is in the signal controller upstream, not in your finger. Honesty of a kind, even if the honesty did not come with a refund.
Your three-or-four-rapid-press habit is the sound of a citizen who has read the system correctly. The pressing is not for the traffic light. The pressing is for you, a small kinetic ritual at a corner where you are otherwise just standing. Behavioral researchers have a name for this. The button becomes a fidget device once the user knows it does nothing, and the fidget function turns out to be real even when the signaling function is fake. The wood pencil tapping on a desk during a meeting works the same way. The hand needs something to do while the situation decides itself.
Thanks for reading and for the LED detail. The “Wait!” sign is going in my notebook for the followup.
@boles
We also have several mid-block crosswalks where pressing the button immediately activates yellow flashing lights above the crosswalk sign.
For the most part, drivers do tend to give way, but there's definitely not the same obedience given to a "real" traffic signal, e.g. drivers tend to keep rolling if you're not actually in the street when the lights are flashing. If you stand on the curb waiting for them to stop, they don't.
I've developed a habit of staring at the driver as I cross the street; not sure about the psychology behind it, but it does tend to get them to slow down or stop further away.
Welcome back, and the mid-block flashing-yellow signal is a fourth category worth thinking about. The placebo button does nothing, the honest button announces what you are supposed to do, and the dishonest button does covert work while looking innocent. The conditional button you are describing actually works, but only partially, and the compliance is tied to a calculation the driver runs every time the lights go on.
The driver’s calculation has three inputs. First, is the pedestrian committed enough to actually step into the street? Second, is there a camera that will record a yield failure? Third, is there any enforcement infrastructure attached to the camera? The mid-block flashing yellow without a camera is a request that drivers respect on average and ignore individually. The same flashing yellow tied to a camera tied to an automated citation tied to a fine is a command. Here in Jersey City the cameras at the mid-block signals get respect because drivers know yield failures get recorded. Some of those recordings turn into mailed citations, and the fear of those citations does the compliance work the flashing yellow alone cannot do. The same signal, fifty miles away in a town without camera enforcement, gets the rolling-yield behavior you describe.
Your eye-contact strategy has actual research behind it, which is why it works for you. Traffic psychologists have been studying pedestrian-driver eye contact for decades, and the consistent finding is that pedestrians who make eye contact with approaching drivers receive higher yield rates than pedestrians who do not. The mechanism is straightforward. A driver who has been seen by a pedestrian has been identified as a specific actor making a specific choice, and the social cost of choosing not to yield increases. A driver who has not been seen can tell themselves later that they did not register the pedestrian was waiting. The eye contact eliminates that deniability. You have converted yourself from generic obstacle into specific witness, and the witness role pulls the driver out of the anonymous-flow behavior that produces rolling yields.
Thank you for the second observation. I am putting both crosswalk variations and the eye-contact research into the notebook for the next piece. The four buttons together are starting to look like a real civic taxonomy: placebo, honest, dishonest, and conditional.