The mandate that should have been impossible to pass

The Americans with Disabilities Act is the rare federal statute that imposed a comprehensive civil rights mandate on virtually every employer, state and local government, and place of public accommodation in the country and still cleared both chambers of Congress by margins that read like acclamation. On September 7, 1989, the Senate passed S. 933 by 76 to 8. On May 22, 1990, the House recorded a 403 to 20 vote for the same legislative text. When the two chambers reconciled their differences that July, the conference report passed the House 377 to 28 and the Senate 91 to 6. Thirteen days after the second of those votes, the president signed the bill as Public Law 101-336. Tallies like these invite a simple explanation: the country agreed, and Congress ratified the agreement.

The legislative record tells a harder and more instructive story. The bill that became law was the second attempt, not the first. The first attempt, introduced in 1988, was substantially broader in coverage, remedies, and compliance timelines, and it died without a markup or a floor vote in either chamber. The second attempt was renegotiated from the ground up, and the negotiations that narrowed it are visible in the enacted text: in the disability definition that requires an impairment to substantially limit a major life activity, in the public accommodations title that offers injunctive relief and attorney’s fees instead of damages, in the phased compliance schedules that replaced a flat two year mandate, and in the food handling provision that nearly destroyed the conference agreement weeks before enactment. The supermajorities were real, and they were purchased. The price is written into the statute.

That is the thread this article follows. It opens with the federal advisory council report that put a comprehensive disability rights statute on the congressional agenda in 1986, reconstructs the 1988 bill and explains why it failed, compares the two bills provision by provision to show exactly what the second attempt gave up, introduces the three legislators whose personal investment in disability explains the cross party margins, walks through the Senate and House floor action vote by vote, and then slows down for the two events that define the story as procedure rather than sentiment: the March 1990 demonstration on the Capitol steps that changed the political weather, and the July 1990 conference fight over food handling that came within a single motion of killing the bill. The namable claim is stated up front and defended throughout: the narrowing made it possible. The Americans with Disabilities Act passed not because it asked everything of the country but because, after the 1988 failure, its authors stopped asking everything.

The puzzle has three parts, and each one needs its own explanation. The first part is scope. The statute reaches every employer of sufficient size, every state and local government, and every place of public accommodation in the country. A mandate of that breadth concentrates its costs on identifiable regulated parties, employers who must accommodate, businesses that must remove barriers, governments that must make programs accessible, while spreading its benefits across a population that is politically diffuse. Bills with concentrated costs and diffuse benefits are the hardest kind to pass, because the opponents have more incentive to organize than the supporters do. The 1988 bill is the proof: its costs were maximal, its coalition was unbuilt, and it died without a vote.

The second part is the control case. The 1988 failure shows that the concept alone, even with a federal advisory council report behind it and cosponsors multiplying in both chambers, could not move a bill. The report supplied the idea. The cosponsors supplied the sentiment. Neither supplied the negotiated text that could survive contact with the regulated community’s objections. The second bill’s authors understood the distinction. They did not reintroduce the 1988 text and hope for a better calendar. They rewrote the text to answer the objections, and then they moved it through a Congress that had already seen the unreconstructed version fail.

The third part is the outcome to be explained: four supermajorities in fourteen months, each larger or more decisive than the last in the Senate, and a conference report that survived the defiance of a Senate instruction. Supermajorities of this size are not produced by enthusiasm. They are produced by the removal of reasons to vote no. Every narrowing in the 1990 text removed a reason: the substantially limits test removed the fear of boundless coverage, the limited remedies removed the fear of punitive damages, the phased timelines removed the fear of an impossible deadline, and the HHS list removed the fear that decided the food handling fight. The coalition of Coelho, Harkin, and Dole supplied the yes votes. The narrowing removed the no votes. Both were necessary, and the passage history is the record of both operating together.

The statute’s formal identity

Every passage history in this series begins by fixing the statute’s identity so the page is citable and unmistakable. The bill that became the Americans with Disabilities Act was S. 933 in the 101st Congress, introduced in the Senate on May 9, 1989, by Senator Tom Harkin of Iowa. It was preceded by an unsuccessful bill in the 100th Congress, S. 2345, the Americans with Disabilities Act of 1988, with its identical House companion H.R. 4498. The enacted statute bears the public law number 101-336, was signed on July 26, 1990, and is printed at 104 Stat. 327 under the title “An Act to establish a clear and comprehensive prohibition of discrimination on the basis of disability.”

One detail of the formal identity deserves attention because it marks the bill’s journey through the chambers. As introduced and as passed by the Senate, the short title read “Americans with Disabilities Act of 1989.” In the House-passed version and as enacted, the short title became the “Americans with Disabilities Act of 1990.” The change is more than cosmetic. A bill that keeps its introduced year through enactment usually moved fast; a bill that sheds one year and takes another on usually traveled a longer and harder road. S. 933 was introduced in May 1989, passed the Senate that September, passed the House the following May, survived a conference fight in July, and was signed before the month ended. The two short titles bracket fourteen months of procedure.

The codification anchors the identity in the place lawyers actually look. The statute’s operative provisions sit principally in title 42 of the United States Code, and the food handling compromise that this article treats in full detail is codified at 42 U.S.C. 12113(d), under the heading “Food handling jobs.” The citation matters to the passage history because 12113(d) is the only provision of the statute that was written, in its final form, inside the conference committee in the last weeks before enactment. Every other contested provision was settled before conference. The food handling language was settled inside it.

The Statutes at Large citation, 104 Stat. 327, anchors the identity in the chronological record of enacted laws. The Statutes at Large print every public law in the order of its enactment, so 104 Stat. 327 is the bill’s address in the permanent sequence: volume 104, page 327, the Americans with Disabilities Act as it left Congress and the president’s desk. The United States Code citation, by contrast, organizes the law by subject, which is why 42 U.S.C. 12113(d) is where a researcher looks for the food handling provision as it operates. The passage history moves between these two citation systems, the chronological and the topical, because the story it tells is about how a chronological sequence of votes produced a topical body of law.

For readers who need the enacted text itself rather than the story of how it passed, the series’ guide to the Americans with Disabilities Act covers the statute title by title, and this article will link forward to the provisions article where the narrowed language is explained in operative detail. The task here is narrower and harder: to reconstruct, from the record, how a mandate of this scope got through the building.

The report that put the statute on the agenda

In January 1986, the National Council on the Handicapped, a federal advisory council operating under a statutory mandate in Public Law 98-221, transmitted a report called “Toward Independence” to the president, the president of the Senate, and the speaker of the House. The transmittal is dated February 1, 1986. The report’s first recommendation was the one that matters to this history: Congress should enact a comprehensive law requiring equal opportunity for individuals with disabilities, with broad coverage and clear, consistent, and enforceable standards prohibiting discrimination on the basis of handicap. The Council even proposed a title for the statute it wanted. At page 18 of the report, it proposed that the new law be titled the “Americans with Disabilities Act,” a suggestion that survives, nearly intact, as the short title of Public Law 101-336.

That detail matters more than it looks like trivia. In Congress, naming is the first act of coalition building. A report that proposes both a policy and a title hands the eventual sponsors a finished banner to march under, and it gives every subsequent negotiation a fixed point of reference: the question is never whether there should be an Americans with Disabilities Act but what that act will contain. The Council’s recommendation supplied the broad coverage aspiration, and the eventual statute’s compromises would be measured against it.

Advisory council reports recommending comprehensive legislation are common in Washington, and most of them sink without a trace. “Toward Independence” did not sink, and the reason is worth stating precisely. The report did not merely diagnose a problem; it supplied the legislative concept in finished form. A comprehensive statute, broad coverage, clear and enforceable standards: those three phrases became the working definition of the project for the next four years, and every subsequent draft, including the enacted law, can be measured against them. The 1988 bill took the recommendation at its most literal, and that literalism is part of why the 1988 bill failed. The 1990 statute honored the recommendation’s structure while qualifying its reach, and that qualification is part of why the 1990 statute passed.

The report’s afterlife in the legislative record is unusually well documented. The National Council on Disability, the Council’s successor name, later filed an amicus brief in Olmstead v. L.C. that recounts the report’s history: published in 1986, transmitted to the president and Congress under the Public Law 98-221 mandate, with the comprehensive-law recommendation as its first and most prominent ask. The Council’s own history of the statute, “Equality of Opportunity: The Making of the ADA,” repeats the formulation. The disability community’s timeline of the anniversary years lists 1986 as the year the Council issued “Toward Independence” with legislative recommendations including enactment of a comprehensive equal opportunity law. None of this is contested. The paper trail from the 1986 report to the 1988 bill is direct: the bill’s drafters worked from the Council’s concept, and the Council’s proposed title became the bill’s short title.

For readers tracing the longer arc of federal disability law, the earlier statute that first carried federal disability protections is covered in the series’ guide to Section 504 of the Rehabilitation Act of 1973. The 1986 report’s demand was for something the earlier model had not supplied: a comprehensive prohibition that did not depend on the federal funding hook. Whether and how the enacted statute delivered that demand is the subject of the provisions article. What matters for the passage history is that the demand was stated first, stated officially, and stated with a proposed name attached, and that Congress took up the invitation within two years.

Why the Americans with Disabilities Act failed in 1988

The first bill drafted from the “Toward Independence” concept was introduced in the Senate on April 28, 1988, by Senator Lowell Weicker of Connecticut as S. 2345, the Americans with Disabilities Act of 1988, recorded at 134 Cong. Rec. 9375. The identical House companion, H.R. 4498, was introduced the next day, April 29, 1988, by Representative Tony Coelho of California, recorded at 134 Cong. Rec. 9600. The drafting was done by Robert Burgdorf, a staffer of the National Council on the Handicapped, working from the Council’s report. The bill therefore arrived in Congress as the purest possible expression of the 1986 recommendation: comprehensive, broad, and uncompromised by any negotiation with the parties it would regulate.

Its initial reception looked promising. S. 2345 opened with 13 Senate cosponsors and grew to 41 before the session ended. H.R. 4498 opened with 47 House cosponsors and grew to 125. Cosponsor counts are a rough measure of interest rather than a predictor of passage, but a bill that quadruples its Senate support and nearly triples its House support in a single session is a bill that has found an audience. The audience, however, never got to vote. An unusual joint Senate and House hearing was held on S. 2345 in September 1988, and otherwise there was no legislative activity on the bill at all. No committee marked it up. No floor debate was scheduled. No vote was taken. The bill died with the adjournment of the 100th Congress, as bills do when a session ends without action.

The cosponsor arithmetic deserves a closer look, because it is the evidence most often cited for the claim that the 1988 bill was viable, and it does not support the claim. Thirteen initial Senate cosponsors growing to 41 means the bill at its peak had the declared support of 42 senators counting Weicker himself, eighteen short of the sixty needed to break a filibuster. Forty seven initial House cosponsors growing to 125 means the bill at its peak had 126 supporters counting Coelho, ninety two short of a House majority. Cosponsorship is costless: a member signs on, takes credit with the disability community, and incurs no obligation to move the bill. The growth figures show that the concept was popular. They do not show that the text was passable. The distinction between popular and passable is the entire lesson of 1988, and it is the distinction the 1989 drafters kept in view.

The September 1988 joint hearing was unusual in form and thin in consequence. A joint hearing, with members of both chambers sitting together on a single bill, is a rare procedural choice, and its rarity is part of the record: the bill’s supporters were economizing on scarce hearing time by addressing both chambers at once, which is what a late introduced bill’s supporters must do. But a hearing is not a markup. It takes testimony; it does not amend text. The 1988 bill got the testimony and never got the amendments, which is why its breadth, the boundless definition, the punitive damages, the two year mandate, survived untouched to adjournment. A bill that is never marked up is a bill whose objections are never answered, and unanswered objections are what kill bills in Congress far more often than organized opposition does.

The standard account of why it died, given by Chai Feldblum and her coauthors in their 2008 history of the ADA Amendments Act published in the Texas Journal on Civil Liberties and Civil Rights, is spare and procedural: after the single joint hearing, there was no legislative activity, and the bill died at adjournment. The ADA Legacy Project’s “Moments in Disability History” series is blunter about the substance: Senator Weicker’s S. 2345 had much stronger language and much broader civil rights policy than the act that was eventually passed into law. The two explanations belong together. A bill can die from neglect, and a bill can die from scope, and S. 2345 died from both at once. It was introduced late in a Congress that was already winding down, it was never shepherded through the committee process that turns a proposal into a votable text, and its provisions asked for more than the coalition that would have had to pass it was prepared to give.

Weicker’s role deserves emphasis because it complicates the partisan story. A Republican senator from Connecticut introduced the first ADA bill, and the bill’s House companion came from a California Democrat. The bipartisan character of the disability rights effort predates the 1990 coalition; it was present at the creation in 1988. What was missing in 1988 was not bipartisanship but negotiability. The Congressional Record citations, 134 Cong. Rec. 9375 for the Senate introduction and 134 Cong. Rec. 9600 for the House companion, fix the bill’s birth to two consecutive days in late April, and the Feldblum, Beauchamp, and Taylor article’s footnotes 4 through 7 build the scholarly record of those introductions from the Record itself.

Why was the 1988 version introduced so late in the congressional session?

The bill was introduced on April 28 and 29, 1988, with only months left in the session and a crowded floor calendar. A late bill that needs hearings, amendments, and negotiation runs out of days, and this one needed all three. It received a single joint hearing in September and never reached markup.

A bill introduced that late in the second session of a Congress faces a structural handicap: committees have set their markup schedules, the floor calendar is crowded with appropriations and expiring authorizations, and anything that needs extended hearings, amendments, and negotiation simply runs out of days. Late introduction did not doom the bill by itself, but it meant the bill had to be noncontroversial to move, and the bill’s breadth made it the opposite of noncontroversial.

What the 1988 bill asked for, in concrete terms, is the subject of the comparison that follows. Three provisions, confirmed in the record, show how much broader the first attempt was than the statute that eventually passed. The disability definition covered any physical or mental impairment, any perceived impairment, and any record of impairment, with no requirement that the impairment substantially limit a major life activity. The remedies allowed people with disabilities to sue businesses for punitive and compensatory damages. The compliance timeline gave the country two years to make everything accessible, a provision the journalist Joe Shapiro, quoted in the ADA Legacy Project’s account, called a radical flat earth bill. Any one of these would have been a hard sell. Together, introduced in April of an election year with no committee groundwork, they were a bill that could attract cosponsors but not a vote.

There is a temptation to read the 1988 failure as proof that Congress did not care, and the temptation should be resisted. The cosponsor growth, from 13 to 41 in the Senate and from 47 to 125 in the House, is the evidence against it. Members were willing to sign their names. What they were not willing to do, in the time available, was to take the political risk of moving a bill that regulated every business in the country, exposed those businesses to punitive damages, and demanded full accessibility within two years. The 1988 bill failed because it asked for the maximum before it had built the coalition that could survive the maximum’s opponents. The 1989 bill succeeded because it built the coalition first and calibrated the ask to what the coalition could carry.

The two drafts side by side: what the second attempt gave up

The enacted Americans with Disabilities Act is often described as the fulfillment of the 1986 report’s vision. Measured against the 1988 bill, it is something more specific: a renegotiated second attempt in which each of the first draft’s most aggressive provisions was narrowed, softened, or deferred. The narrowing happened between the two bills, not during the 1989 to 1990 floor fights, which is why the floor fights could proceed on relatively settled text. By the time S. 933 was introduced in May 1989, the hard concessions were already baked in. Understanding the concessions is understanding why the second bill could attract the votes the first bill could not.

The first and most consequential narrowing concerned who counts as protected. The 1988 bill, in Burgdorf’s drafting, protected any person who experienced discrimination because of a physical or mental impairment, a perceived impairment, or a record of impairment. There was no requirement that the impairment substantially limit anything. The category was as broad as the concept of impairment itself, and it swept in conditions that limited nothing in the person’s daily life. The 1990 statute kept the three prongs, actual impairment, record of impairment, and being regarded as having an impairment, but it added the limiting test that the 1988 draft had omitted: the impairment must substantially limit one or more major life activities. That single qualifying phrase is the most litigated language in the statute’s history, and it was a concession. It narrowed the protected class to people whose impairments have substantial effects, and it did so deliberately, to answer the objection that the 1988 definition would have covered essentially everyone.

Take the coverage standard first, because it is the most technically consequential of the three narrowings. The 1988 bill’s formula, discrimination “because of a physical or mental impairment, perceived impairment, or record of impairment,” contains no limiting principle on severity. Any impairment, however minor or transient, would have sufficed, and so would a perceived impairment that did not exist at all. The enacted statute’s substantially-limits prong inserts the limiting principle: the impairment must substantially limit one or more major life activities. For the business community, that prong converted an unpriceable risk into a bounded one. An employer can assess whether a condition substantially limits a major life activity; an employer cannot assess exposure under a standard with no severity threshold at all. The substantially-limits test was the coverage concession, and it was still being litigated and legislated two decades later, including in the later fight over the definition that produced the ADA Amendments Act of 2008.

The second narrowing concerned remedies, and it is the one the business community cared about most. Under S. 2345, as Shapiro wrote in No Pity, people with disabilities could have sued businesses for punitive damages. Punitive damages are the remedy that concentrates the business mind, because they are uncapped by any compensatory logic and are designed to punish. The enacted statute took that weapon off the table for the public accommodations title: the 1990 law’s remedies there were limited to injunctive relief and attorney’s fees. On the employment side, the statute tied remedies to Title VII of the Civil Rights Act of 1964, which at the time meant injunctive relief and back pay, not compensatory or punitive damages; the series’ passage history of the Civil Rights Act of 1964 reconstructs how that earlier statute reached the president’s desk. Those damages did not become available in intentional employment discrimination cases under the ADA until the Civil Rights Act of 1991, a separate statute enacted the following Congress, as the Equal Employment Opportunity Commission’s history of the law’s development records. The sequence matters. The 1990 compromise was not that damages would never be available; it was that they would not be available in the bill that had to pass in 1990. Deferring the damages fight to a later Congress was one of the prices of the supermajority.

On remedies, the economics are as important as the law. Punitive damages exist to punish and deter, and their availability transforms every employment dispute into a potential lottery ticket for plaintiffs’ lawyers. Compensatory damages for emotional distress do the same on a smaller scale. By exposing businesses to both, the 1988 bill made every covered entity’s worst-case liability effectively unlimited, which is why the business community treated the bill as existential. The 1990 design replaced that exposure with the Title VII model: injunctive relief, meaning court orders to stop discriminating, and back pay, meaning the wages the worker would have earned. Those remedies make the worker whole without bankrupting the employer, and they give the business community a worst case it can insure against and budget for. The public-accommodations title went further toward safety, offering only injunctions and attorney’s fees. The difference between a bill that exposes a business to punitive damages and a bill that exposes it to an injunction is the difference between a bill the business community fights to the death and a bill it can live with. The 1988 bill chose the former. The 1990 bill chose the latter.

The third narrowing concerned time. The 1988 bill’s compliance provision, in Shapiro’s account, required that within two years everything would have to be accessible. A two year deadline for nationwide accessibility is the kind of provision that reads as moral clarity and operates as an enforcement nightmare: no phased schedule, no distinction between new construction and existing facilities, no ramp for the entities that would have to do the work. The enacted statute replaced the flat deadline with negotiated phased timelines that gave covered entities years, not months, to come into compliance across different obligations. The phases are the fingerprints of negotiation. A bill written by advocates alone sets one deadline. A bill written to pass sets several, each one marking a point where someone’s objection was answered with time instead of exemption.

On timelines, the contrast is between a moral statement and an administrative plan. “Within two years, everything would have to be accessible” is a sentence that works as a speech and fails as a statute. The built environment of the United States, every workplace, every place of public accommodation, every government facility, could not be remade in twenty-four months at any price, and everyone in the room knew it. The “flat earth” label the histories attach to the 1988 bill captures the design flaw: it assumed a flat world in which a single deadline could flatten every terrain. The 1990 statute’s phased timelines did the opposite. They broke the obligation into stages, gave covered entities schedules they could plan capital spending against, and, crucially, gave the business community’s negotiators something to negotiate. A deadline can be moved. A flat-earth mandate cannot. The negotiability of the timeline was itself one of the concessions.

What made the 1990 disability definition narrower than the 1988 version?

The 1988 bill covered any physical or mental impairment, perceived impairment, or record of one, with no further test. The 1990 law kept those three prongs but required the impairment to substantially limit one or more major life activities, narrowing the protected class to people with substantial effects.

Why did the 1990 law drop the punitive damages the 1988 bill offered?

The 1988 bill let plaintiffs seek punitive and compensatory damages against businesses, the remedy the business community feared most. The 1990 statute limited the public accommodations title to injunctions and attorney’s fees and tied employment remedies to Title VII’s then available relief. Compensatory and punitive damages arrived only with the Civil Rights Act of 1991, a later and separate fight.

How did the compliance deadlines change between the two bills?

The 1988 bill demanded that everything be accessible within two years, a single flat deadline with no phasing. The enacted statute replaced it with negotiated phased timelines that spread different obligations across different schedules. Each phase marks an objection answered with time rather than exemption, which is why the timelines read as a record of negotiation.

The comparison extends beyond the three headline narrowings to the bill’s overall posture. The 1988 bill was drafted by a council staffer from an advocacy report, introduced late, and never tested against the objections of the regulated community in a markup. The 1990 bill was drafted to survive those objections, and the places where it bends are the places where the objections were strongest: coverage, defenses, remedies, and timelines. The series’ explainer on the ADA’s titles maps the five titles of the enacted statute and shows, title by title, where the narrowed language landed. The artifact table below sets the two bills against each other in the form the record supports, provision by provision, so the trade can be seen at a glance.

The two-bill comparison

Provision	The 1988 bill (S. 2345 / H.R. 4498)	The 1990 statute (S. 933 as enacted)	What the narrowing bought
Bill and sponsors	S. 2345 introduced April 28, 1988, by Senator Lowell Weicker of Connecticut; identical House companion H.R. 4498 introduced April 29, 1988, by Representative Tony Coelho of California	S. 933 introduced May 9, 1989, by Senator Tom Harkin of Iowa; enacted as Public Law 101-336, signed July 26, 1990	A second attempt drafted to be negotiated rather than to make a statement
Who counts as protected	Any physical or mental impairment, perceived impairment, or record of impairment; no requirement that the impairment substantially limit anything	The three prongs retained, but the impairment must substantially limit one or more major life activities	A coverage boundary the business community could price; minor and transient conditions stayed outside the statute
Remedies against businesses	Punitive and compensatory damages available in suits against businesses, as described in Joe Shapiro’s No Pity	Public accommodations title limited to injunctive relief and attorney’s fees; compensatory and punitive damages not available under the ADA until the Civil Rights Act of 1991	Removed the damages exposure that made business opposition existential; the fight moved from killing the bill to negotiating its terms
Employment remedies	Broader damages posture carried over from the bill’s general approach	Tied to Title VII of the Civil Rights Act of 1964, which at enactment meant injunctive relief and back pay	Damages deferred to a later Congress instead of loaded into the bill that had to pass in 1990
Compliance timeline	Within two years, everything accessible; described as a radical flat earth mandate	Negotiated phased timelines spreading obligations across different schedules	A planning horizon covered entities could budget against instead of a single two year cliff
Legislative outcome	One joint House and Senate hearing in September 1988; no markup and no floor vote in either chamber; died with the adjournment of the 100th Congress	Passed the Senate 76 to 8, the House 403 to 20, and the conference reports 377 to 28 and 91 to 6; signed into law	Four supermajorities purchased by the concessions in the rows above

The sources behind the table are worth naming, because the comparison’s authority rests on them. The coverage difference comes from the Feldblum, Beauchamp, and Taylor article in the Texas Journal on Civil Liberties and Civil Rights, which quotes the 1988 bill’s “physical or mental impairment, perceived impairment, or record of impairment” formula and notes the absence of the substantially-limits test. The remedies difference comes from Joe Shapiro’s No Pity, via the ADA Legacy Project’s disability history series, for the 1988 bill’s punitive and compensatory damages, and from the EEOC’s legislative history of the ADA for the Title VII remedy structure and the Civil Rights Act of 1991’s later damages upgrade. The timeline difference comes from Shapiro via the same history series, with the “within two years, everything would have to be accessible” demand and the “flat earth” characterization. The overall judgment that the 1988 bill carried much stronger language and much broader civil rights policy than what eventually passed comes from the ADA Legacy Project’s “Moments in Disability History.”

The table makes the namable claim visible. The enacted statute is meaningfully less sweeping than the bill first drafted, and the concessions on coverage, defenses, and remedies are the reason a mandate of this scope could attract a supermajority. The 1988 bill is the control case: the same concept, without the concessions, could not get a markup. The 1990 statute is the treatment: the same concept, with the concessions, could not be stopped. That is the trade every ambitious civil rights bill in this series has faced, and the ADA’s passage history documents it more cleanly than most, because the two drafts sit only a year apart and the differences between them are a matter of record rather than inference.

The coalition: three legislators with personal stakes

The vote margins that define this history, 76 to 8, 403 to 20, 377 to 28, 91 to 6, cannot be explained by party discipline or presidential pressure alone. They are explained, in the practical sense, by the bill’s champions: legislators from both parties whose investment in disability was personal rather than positional. The brief for this article names three, and the record supports the naming. One member of the House had a disability himself and was the bill’s primary author. One senator’s brother was deaf, and the senator delivered part of his final floor argument in sign language. One senior minority leader was a disabled veteran of the Second World War. Cross party personal investment of that kind does not guarantee passage, but it changes the character of the opposition, because it is difficult to organize a partisan campaign against a bill whose most visible advocates are telling their own stories.

Tony Coelho was the House’s author. Diagnosed with epilepsy at the age of 22, after having been barred from the priesthood under canon law on account of the condition, Coelho served six terms as a Representative from California and was, in the Epilepsy Foundation’s account and his own, the primary author and sponsor of the Americans with Disabilities Act. The biographical details matter to the procedure because they determined who carried the bill and when. Coelho introduced the House companion of the 1988 bill, H.R. 4498, and he was the natural author of the House’s 1989 effort as well. But Coelho resigned from the House on June 15, 1989, about five weeks after S. 933 was introduced in the Senate, and the House’s stewardship of the bill passed to Representative Steny Hoyer, who carried it forward. The resignation is one of the underappreciated contingencies of the passage history. Had Coelho remained, the House bill would have been authored by a member with epilepsy. As it happened, the House bill was managed by a member without a disability, which makes the personal investment story a Senate story in its final phase and a House story in its drafting phase. The bill survived the loss of its primary author, which is evidence that the coalition had grown beyond any single champion by the summer of 1989.

Senator Tom Harkin of Iowa was the Senate’s author and the bill’s most visible advocate. Harkin’s older brother Frank had lost his hearing at the age of five, and Harkin had grown up in a household where disability was not an abstraction. When Harkin introduced S. 933 on May 9, 1989, he was not merely sponsoring a bill; he was carrying a family history onto the floor. The moment that fixed his authorship in the public memory came more than a year later, on July 13, 1990, during the Senate’s final debate on the conference report. Harkin became the first person to deliver a Senate floor speech in American Sign Language, opening with a tribute to his brother and then translating his remarks into spoken English: “Congress opens the doors to all Americans with disabilities, today we say no to ignorance, no to fear, no to prejudice.” Asked later why he had used sign language, Harkin gave the simplest possible answer: he wanted Frank, who was deaf, to understand him. The drafting note that matters here is the date. The sign language speech belongs to the July 1990 conference report debate, not to the September 1989 passage debate, and confusing the two collapses the story’s timeline. The speech was not an opening argument for a bill seeking votes; it was a closing argument for a bill that had already survived conference, delivered in the language of the brother whose experience had helped put the bill on the agenda.

Senator Bob Dole of Kansas was the minority leader and the bill’s most important Republican. Dole had been seriously wounded in Italy on April 14, 1945, shattering his right shoulder and arm, an injury that left him permanently disabled. He received two Purple Hearts and the Bronze Star. By the time of the ADA, he was the Senate Republican leader, the senior minority leader of the brief, and the Congressional Record’s findings on his later Congressional Gold Medal describe him as a lifelong advocate for the disabled and a key figure in the passing of the Americans with Disabilities Act in 1990. Finding 9 of that Record entry is as close as Congress comes to an official history, and it names Dole as indispensable. Dole’s role in the procedure is easy to understate because he did not author the bill, but it is difficult to overstate in political terms. A minority leader with a visible disability, from the president’s own party, standing for the bill, deprives the opposition of its most natural organizational base. Republican senators who might have been willing to vote against a Democratic civil rights bill were being asked to vote against a bill their own leader had made his cause. The 76 to 8 and 91 to 6 Senate margins are, in significant part, Dole margins.

The three biographies also explain something about the bill’s timing. Coelho’s resignation in June 1989 could have stalled the House effort; instead Hoyer picked it up without a visible break. Hoyer is the least storied of the bill’s managers, and his very lack of a personal disability narrative is part of the point: by the summer of 1989 the bill no longer needed a champion with epilepsy to carry it in the House, because the coalition and the narrowed text were doing the carrying. The transition from Coelho to Hoyer, about five weeks after S. 933’s introduction, tested whether the bill was a personal project or an institutional one. It was institutional, and it survived. Harkin’s persistence through fourteen months of procedure, from introduction to conference report, gave the Senate effort a single continuous author. Dole’s presence as minority leader gave the bill bipartisan cover at every stage, including the conference fight where it mattered most. None of this was sentiment. It was the practical infrastructure of a supermajority: authors who could not be dismissed as dilettantes, a manager who survived a resignation, and a minority leader whose biography made opposition personally costly.

The mechanism by which personal investment becomes votes is worth stating plainly, because it is the least sentimental part of the story. In a legislature, members take cues from members they trust on issues they have not mastered. Disability rights was not a mastered issue for most senators in 1989. What they had mastered was their colleagues. A Republican senator uncertain about the bill’s business costs could look at Dole, a disabled veteran and the minority leader, and conclude that the bill had been vetted by someone who understood both the policy and the politics. A Democratic senator uncertain about the bill’s scope could look at Harkin and Coelho and conclude the same. Personal investment functions as a heuristic for seriousness, and seriousness is what moves undecided votes. The 76 to 8 and 91 to 6 margins are what that heuristic produces.

The demonstration that changed the weather

On March 12, 1990, with the bill between its Senate passage and its House action, disability rights activists staged the demonstration that became the most reproduced image of the legislative fight. Organized by ADAPT, a grassroots disability rights organization, the protest began with a march from the White House to the Capitol. At the Capitol, more than sixty activists, by the Museum of Disability History’s account, left their wheelchairs and mobility aids at the base of the west steps and crawled and dragged themselves up the stairs. Other accounts put the number of marchers above one thousand, with about a thousand onlookers watching the crawl. The event was photographed and broadcast widely, and it entered the political vocabulary under the name the Capitol Crawl.

The procedure of the demonstration is worth reconstructing, because its power came from its staging. The activists did not simply protest. They removed the very devices that made their mobility possible and then performed, on the steps of the building where the bill was stalled, the physical cost of inaccessibility. Every step crawled was an argument that no lobbyist could make with words: this is what a world without the ADA requires of us. The march from the White House added a second layer of meaning, connecting the executive branch’s inaction to the legislative branch’s delay in a single line of bodies moving across the capital.

The procedure-minded reader should ask what a demonstration can actually do to a bill’s prospects, and the answer is specific. The Capitol Crawl did not change any vote count directly; the House vote two months later was 403 to 20, a margin that suggests the outcome was not in doubt. What the demonstration changed was the cost structure of opposition. Before March 12, a member could oppose or stall the bill as a routine regulatory matter, one more mandate with a price tag. After the crawl, with the images of activists dragging themselves up the Capitol steps in national circulation, opposition carried a different and much higher political price. The demonstration converted the bill from a regulatory question into a civil rights question in the public mind, and civil rights questions are ones members answer on the record, in daylight, with their names attached. The 403 to 20 margin is consistent with a chamber in which the price of a no vote had become prohibitive.

The political weather metaphor is exact. Weather is not argument; it is the condition in which argument happens. After March 12, every House member who considered weakening the bill had to do so against the background of those photographs. The business community’s negotiations continued, and the concessions the business community won were real, but the price of killing the bill outright rose sharply. The Capitol Crawl did not write a single line of the statute. It changed what was politically possible to do to the statute, and that is the more important effect.

The timing also matters. March 12, 1990, sits in the window between the Senate’s September 1989 passage and the House’s May 1990 action, which is to say it sits in the window when the House was deciding how far to go. The Chapman amendment, which the House adopted on May 17, 1990, about two months after the crawl, shows that the demonstration did not eliminate controversy; the food handling fight was still to come. But the crawl set the terms on which the remaining fights would be conducted. After March 1990, no one in the House debated whether disability discrimination was a legitimate subject of federal civil rights law. They debated, instead, the scope of the defenses, the shape of the remedies, and the mechanics of compliance. That is a narrower and more technical debate, and it is the debate the bill’s authors wanted, because it is the debate their narrowed text was designed to win.

The numbers attached to the demonstration vary by source, and the variation is worth recording rather than smoothing over. The Museum of Disability History counts more than sixty activists leaving their wheelchairs and mobility aids to make the climb. The Zinn Education Project counts more than a thousand marchers on the day, with about a thousand onlookers watching the crawl itself. The two figures are not contradictory: a march of a thousand can contain a climbing party of sixty. What matters for the passage history is not the exact headcount but the image, which was photographed and broadcast widely, and the organization, ADAPT, a grassroots disability rights group that understood that the Capitol steps were the most legible possible stage. The steps are where Congress enters its workplace. Forcing the country to watch disabled Americans crawl up them made the abstract question of accessibility concrete in a way no hearing testimony could.

ADAPT’s identity matters to the story. Then called Americans Disabled for Accessible Public Transit, it was a grassroots direct-action organization, not a Washington lobbying shop. The distinction is the point. The disability rights coalition had both kinds of organizations, the inside negotiators who worked the bill’s text and the outside agitators who worked the political weather, and March 12 showed the division of labor at its most effective. The negotiators could not have staged the crawl, and the crawlers could not have drafted the compromise, and the bill needed both.

The staging deserves one more layer of analysis. The activists began at the White House, the seat of the executive, and marched to the Capitol, the seat of the legislature, tracing in physical space the path a bill travels from signature back to passage. They left their mobility aids at the base of the steps, which meant the climb was performed without the devices the bill was meant to make unnecessary, a literal enactment of the world the ADA promised to change. And they climbed the west steps, the ceremonial front of the Capitol, where the cameras were already positioned for the iconography of American government. Every choice in the staging maximized the image’s circulation, and the circulation is what changed the weather. A demonstration that is not photographed is a private act. A demonstration that is broadcast is a political fact.

Why did activists crawl up the Capitol steps in March 1990?

The March 12, 1990 demonstration, organized by ADAPT, was staged to pressure the House into acting on the ADA. More than sixty activists left wheelchairs and mobility aids at the base of the Capitol’s west steps and crawled up them after a march from the White House. The broadcast images made inaccessibility visible nationwide.

The Senate acts first: September 1989

S. 933 was introduced on May 9, 1989. Four months later, on September 7, 1989, the Senate passed it with an amendment by a yea-nay vote of 76 to 8, recorded as Record Vote No. 173. The four month interval between introduction and passage is short by the standards of major civil rights legislation, and it is worth pausing over what the speed signifies. The bill had the advantage of the 1988 groundwork: the concept was familiar, the cosponsor base was already warm, and the narrowing that made the bill passable had been completed before introduction rather than negotiated on the floor. A bill that arrives pre narrowed can move quickly because the fights that would have slowed it have already been fought in the drafting room.

The congress.gov action log’s phrasing for September 7, 1989, “Passed Senate with an amendment by Yea-Nay Vote. 76-8. Record Vote No: 173,” compresses a day of floor action into a single line, and each element of the line carries information. “With an amendment” means the Senate did not pass Harkin’s introduced text untouched; the floor process modified it, as floors do. The yea-and-nay vote means every senator’s position was recorded by name, which matters because a recorded vote is a commitment device: members who vote yes on the record cannot easily vote no later without explanation. Record Vote No. 173 fixes the vote in the Senate’s numbered sequence for the session, the citation a researcher uses to pull the roll call.

The 76 to 8 margin deserves the same careful reading the House margins will get later. Eighty four senators recorded votes, and seventy six of them voted yes. In a chamber where sixty votes is the threshold for breaking a filibuster and sixty seven the threshold for overriding a veto, seventy six is a supermajority with room to spare. It is also a margin that crosses every plausible partisan and regional line. No purely Democratic bill passes 76 to 8 in the Senate of the 101st Congress, and no purely Republican bill does either. The margin is the arithmetic signature of the coalition described above: a Democratic author with a personal stake, a Republican leader with a personal stake, and a text that had been negotiated down to the point where the remaining objections were not worth a no vote.

The speed deserves equal attention. Four months from introduction to passage is fast for a bill of this scope, and the speed is itself evidence about the bill’s preparation. A bill that needs extended committee hearings, hostile amendments, and floor negotiation does not move in four months. S. 933 moved in four months because the narrowing was complete before introduction, the coalition was assembled before introduction, and the Senate’s floor action was therefore a ratification of work already done rather than a negotiation conducted in public. The contrast with the 1988 bill is instructive: the 1988 bill had about six months and produced a single hearing, because its text was not ready for the objections it would face. The 1989 bill had four months and produced a 76 to 8 passage vote, because its text had already answered them.

What did the 76 to 8 Senate vote actually show?

Seventy six senators voted yes and eight voted no on September 7, 1989, in Record Vote No. 173. With 76 votes the bill cleared the 60 vote filibuster threshold and the 67 vote veto override threshold with room to spare, proving the narrowed text had assembled a coalition far beyond either party.

It was the arithmetic proof that the 1988 failure had been a failure of calibration, not of concept.

The Senate’s version, as passed, contained no food handling provision of the kind the House would later adopt. The Senate relied instead on the statute’s general direct threat defense, the provision that allows employers to act where an individual poses a direct threat to the health or safety of others. That reliance would become the fault line in conference, because the House, eight months later, would decide that the general defense was not specific enough for food handling jobs. The Senate’s confidence in the general defense and the House’s insistence on a specific one were both reasonable readings of the same text, and the conference committee would have to produce a third reading that satisfied both chambers. The September 1989 vote, for all its decisiveness, settled nothing about food handling. It settled that the Senate wanted the bill.

The House: the Chapman amendment and the two votes of May 22

The House took up the bill in the spring of 1990, and the House’s proceedings supply the two most procedurally intricate episodes of the passage history: the adoption of the Chapman amendment on May 17, and the double passage mechanics of May 22. Both need to be reconstructed carefully, because the second is frequently misdescribed and the first nearly killed the bill.

On May 17, 1990, the House adopted H.AMDT.450, offered by Representative Jim Chapman of Texas, by a recorded vote of 199 to 187. The amendment added subsection 103(d) to the House bill, H.R. 2273, and its content must be stated exactly, because the entire conference fight turns on what it said. The Chapman amendment provided that an employer could refuse to assign, or continue to assign, an employee with an infectious or communicable disease of public health significance to a food handling job, provided that the employer offered the employee an alternative employment opportunity for which the employee was qualified and as to which the employee would sustain no economic damage. The vote was close, 199 to 187, a twelve vote margin in a chamber that would, five days later, pass the underlying bill 403 to 20. The closeness of the Chapman vote against the lopsidedness of the passage vote is the first clue to the amendment’s political character: it was the one provision on which the House’s consensus broke down.

The Chapman vote of May 17 deserves its own arithmetic. One hundred ninety-nine to 187, with 386 members voting, is a margin of twelve in a chamber of 435. That is not a consensus. It is a near-even split with a slight lean, and it tells us that the House was genuinely torn on the food handling question. Compare it to the 403 to 20 vote on the bill five days later, a margin of 383. The same chamber, the same week, the same bill, and the difference between a twelve-vote margin on the amendment and a 383-vote margin on the bill is the difference between a contested detail and an agreed whole. The ADA’s managers learned from the Chapman vote that the food handling provision was the bill’s most flammable material, and the conference’s subsequent decision to strip it reads, in retrospect, as an attempt to remove the flammable material before it could ignite the whole structure.

The Senate had never adopted any such provision. The Senate’s bill relied on the general direct threat defense, and the Senate’s managers evidently considered that defense sufficient for food handling jobs as for every other workplace question. The difference between the chambers was therefore not a disagreement about whether infectious disease could ever justify removing a worker from food handling. It was a disagreement about who decides, and by what standard. The Chapman amendment gave the employer a specific, self executing defense for food handling assignments, conditioned on offering alternative work without economic harm to the employee. The Senate’s approach left the question to the general defense and, ultimately, to case by case adjudication. Disability advocates read the Chapman language as a broad employer license to exclude workers with conditions like HIV from food handling on the basis of fear rather than evidence. The amendment’s supporters read it as a necessary public health protection that the general defense did not clearly supply. The passage history does not need to resolve which reading was right. It needs to record that both readings were sincerely held, that the House adopted the amendment 199 to 187, and that the Senate’s rejection of it set up the conference collision.

Five days later, on May 22, 1990, the House produced the vote most often cited in summaries of the ADA’s passage, and the one most often described imprecisely. At 3:00 PM, the House passed H.R. 2273, the House companion bill, by a recorded vote of 403 to 20, with 9 members not voting, recorded as Roll No. 123. The Clerk’s vote question named the Americans with Disabilities Act of 1990, and the tally is the House’s official recorded judgment on the bill’s text. Then, at 3:04 PM, the House took up S. 933, the Senate-passed bill, struck everything after the enacting clause, and substituted the text of H.R. 2273, all without objection. At 3:05 PM, the House passed S. 933 itself without objection, which is to say by voice vote, and laid H.R. 2273 on the table.

Why did the House pass the ADA twice in one afternoon?

The 403 to 20 recorded vote at 3:00 PM was taken on H.R. 2273, the House companion bill. At 3:04 PM the House substituted that text into S. 933, and at 3:05 PM passed S. 933 by voice vote, tabling H.R. 2273.

The procedure is worth unpacking because it is the standard House method for getting to conference and because it is routinely misreported. When the House and Senate each pass their own bill number containing similar text, the chambers cannot go to conference until one chamber has acted on the other’s bill. The House’s solution is the enacting clause substitution: take the Senate’s bill number, empty it of the Senate’s text, pour in the House’s text, and pass the resulting vehicle. The bill that goes to conference is then S. 933 in number but H.R. 2273’s successor in text, and the Senate’s conferees are appointed to reconcile the Senate’s original text with the House’s substituted text. The 403 to 20 tally belongs, strictly speaking, to H.R. 2273’s passage. S. 933 passed the House by voice vote the same afternoon. To describe S. 933’s House passage as a 403 to 20 recorded vote is therefore imprecise, though the Clerk’s own phrasing, “the House passed the ADA by a recorded vote of 403 to 20,” captures the substance correctly: the recorded judgment of the House on the bill’s text was 403 to 20.

The 403 to 20 margin, with 9 not voting, means 423 members recorded a position and 403 of them voted yes. In a House of 435 seats, that is as close to unanimity as a contested regulatory bill ever gets. The Chapman amendment’s 199 to 187 adoption, five days earlier, is the contrast that gives the passage margin its meaning. The House was nearly unanimous on the bill and nearly evenly divided on the food handling defense. The bill’s managers had built a consensus that held for everything except the one provision the conference would have to resolve.

The nine members not voting on Roll No. 123 are a footnote worth one sentence. In a 435-member House, 403 plus 20 plus 9 equals 432, leaving three seats vacant or unrecorded in the Clerk’s detail, the ordinary churn of resignations, deaths, and absences that the raw tally does not explain. The margin among those who voted was better than twenty to one.

Why the Senate vehicle had to survive is a question of legislative mechanics. The Senate had passed S. 933 first, in September 1989. If the House had simply passed H.R. 2273 and sent it to the Senate, the Senate would have had to pass the House bill, and the bill’s number, history, and committee lineage would have restarted. By substituting the House text into S. 933, the House preserved the Senate’s action, kept a single bill number for the conference, and ensured that the conference committee would reconcile the Senate-passed text with the House-passed text rather than starting over. The voice vote was not a shortcut around deliberation. The deliberation had happened on H.R. 2273, on the record, at 3:00 PM. The voice vote was the chamber’s acknowledgment that there was nothing left to deliberate.

The doctrinal disagreement: who decides the risk

Beneath the procedural drama of the food handling fight lay a genuine doctrinal disagreement about who should decide whether a worker’s condition makes food handling unsafe, and the disagreement deserves to be stated on its own terms before the conference’s resolution is assessed.

The Senate’s position was that the statute’s general direct threat defense was sufficient. The direct threat concept, as the Senate’s bill embodied it, allows an employer to act where an individual poses a direct threat to the health or safety of others, with the threat assessed on the basis of reasonable medical judgment and objective evidence rather than stereotype. Applied to food handling, the defense would work the way it worked everywhere else in the statute: the employer would have to show, if challenged, that the worker’s continued presence in the food handling job posed a genuine threat that could not be eliminated by reasonable accommodation. The Senate’s managers evidently considered that framework adequate for kitchens as for every other workplace, and the Senate-passed bill contained no food handling specific language at all.

The House’s position, as expressed in the Chapman amendment, was that food handling was different enough to need its own rule. Chapman’s text gave the employer a specific, self executing defense: where the employee had an infectious or communicable disease of public health significance, the employer could remove the worker from food handling, provided alternative qualified employment was offered without economic harm to the worker. The amendment’s logic was preventive rather than adjudicative. Instead of requiring the employer to prove a direct threat after the fact, it gave the employer a clear rule to apply in advance, with the alternative employment condition as the worker’s protection. The 199 to 187 adoption vote shows that nearly half the House found this logic persuasive.

The amendment’s three-part structure is a small lesson in legislative drafting. First, the trigger: “an infectious or communicable disease of public health significance.” Note what the phrase does not do. It does not list diseases. It does not define public health significance. It leaves the category’s boundaries to the employer’s judgment, bounded only by the requirement that the disease be infectious or communicable and of public health significance. Second, the power: “refuse to assign or continue to assign” the employee to a food-handling job. The “continue to assign” language means the amendment reached current employees, not just new hires; an employer could remove a worker already in the job. Third, the condition: the employer must offer “an alternative employment opportunity for which the employee was qualified and would sustain no economic damage.” The condition is the amendment’s conscience. It says the employer’s power to remove is paired with a duty to reassign without pay loss, which means the amendment’s drafters anticipated the objection that removal equals termination and tried to answer it in the text.

The advocates’ objection to Chapman was that the amendment’s key terms, infectious or communicable disease of public health significance, left the decisive judgment with the employer, and that employers applying that judgment would exclude workers on the basis of fear rather than evidence. The objection was not abstract. The conditions most likely to trigger the defense were the ones most surrounded by public misunderstanding, and a defense that turned on the employer’s assessment of public health significance would, in practice, license exclusions that the statute’s general framework would not have permitted. The Senate’s reliance on the direct threat defense and the advocates’ opposition to Chapman were the same position stated from different sides: the question of risk should be answered by evidence and medical judgment, not by the employer’s discretion.

The Senate’s alternative, the general direct-threat defense, worked on an entirely different theory. A direct-threat defense asks whether the individual employee poses a direct threat to the health or safety of others that cannot be eliminated by reasonable accommodation. It is individualized by design: the question is always about this worker, this job, this risk. The Chapman amendment asked a categorical question instead: does this worker fall into the disease category, and is the job food handling. The union’s May 31, 1990 letter to the Senate, from the United Food and Commercial Workers, made the case against categories: workers with conditions like HIV, which is not transmitted through food handling, could be swept out of food jobs by employer fear rather than by transmission science. The categorical rule’s virtue, speed and clarity for the employer, was also its vice, overbreadth for the worker. The individualized standard’s virtue, precision, was also its vice, uncertainty for the employer who had to guess right about each case.

The conference substitute answered the doctrinal question with a third allocator of decision authority. Neither the employer alone, as Chapman would have had it, nor the general adjudicative framework alone, as the Senate would have had it, but the Secretary of Health and Human Services, acting through a published, annually updated, scientifically grounded list. The list determines which diseases count. The reasonable accommodation qualifier determines whether the risk can be managed without removal. The employer retains only the conditional right to act within those boundaries. The doctrinal disagreement was thus resolved not by choosing between the chambers’ positions but by moving the decision to a forum neither chamber had proposed: a federal scientific determination, transparent, revisable, and bounded. That is why the provision reads as the most carefully engineered language in the statute. It had to settle a three way disagreement about who decides, and it settled it by inventing a decider.

The near death in conference: the food handling fight

The conference committee is the least visible and most consequential stage of the legislative process for a bill that has passed both chambers in different forms. Its members, drawn from both chambers and both parties, are charged with producing a single text that both chambers can accept, and their product, the conference report, is voted up or down in each chamber without amendment. The procedure gives the conference enormous power and no margin for error: if either chamber rejects the report, the bill returns to negotiation or dies. The series’ explainer on conference committees sets out the institution’s rules in full. That all-or-nothing structure is what made the Chapman fight existential. What follows is how those rules nearly destroyed the Americans with Disabilities Act in the last weeks before enactment.

The collision was set up by the chambers’ different answers to the food handling question. The House had adopted the Chapman amendment, giving employers a specific defense for removing workers with infectious or communicable diseases of public health significance from food handling jobs, conditioned on alternative employment without economic harm. The Senate had no such provision and relied on the general direct threat defense. The conference committee had to choose: keep the Chapman language, drop it, or write a third provision that both chambers could accept.

Before the conferees met, the Senate took a step that raised the stakes. Senator Helms moved to instruct the Senate conferees to support the Chapman language, and the full Senate voted to instruct its conferees to accept the House’s food handling provision. A motion to instruct is, in Senate procedure, a formal direction from the chamber to its conferees about the position to take in conference. It is not absolutely binding as a matter of parliamentary law, but it is a public instruction adopted by the full Senate, and conferees defy it at their political peril. The June 1990 fight over the Helms instruction, including the tabling battle documented in the disability community’s legislative files, showed that the food handling question had organized a real constituency in the Senate. The United Food and Commercial Workers union wrote to the Senate on May 31, 1990, urging senators to reject any motion to instruct the conferees to recede to the House approved Chapman food handler amendment, which is evidence that organized labor was watching the instruction fight and wanted the Chapman language kept out.

What the conference committee did next is the most dramatic procedural fact in the passage history. Less than three weeks after the full Senate voted to instruct its conferees to accept the Chapman language, the conference committee stripped the Chapman amendment from its report. The conferees defied the instruction. The National Restaurant Association, writing to senators on July 9, 1990, in a letter preserved in the Dole Archives, made the timeline explicit: the committee had removed the amendment less than three weeks after the Senate’s instruction vote. A conference committee that defies a Senate instruction is betting that the full Senate will accept the defiance when the report comes back for a vote. Conferees are appointed to produce an agreement both chambers can pass, and their loyalty is to the package, not to any single instruction. It is a bet the conferees can only win if they have replaced the stricken provision with something the Senate’s Chapman supporters can live with, or if the politics of the conference report as a whole make a no vote unthinkable.

The conferees’ replacement was the provision codified at 42 U.S.C. 12113(d). Instead of Chapman’s employer operated defense, the conference wrote a science driven federal listing mechanism. The Secretary of Health and Human Services, not later than six months after enactment, must publish a list of infectious and communicable diseases that are transmitted through handling the food supply, with the methods of transmission stated, disseminate the list to the general public, and update it annually. Where an individual has a listed disease that is transmitted through food handling, and the risk cannot be eliminated by reasonable accommodation, a covered entity may refuse to assign or continue to assign that individual to a food handling job. And nothing in the act preempts state or local food handling laws. The full mechanics of the provision are walked through in the next section. What matters for the conference story is the character of the trade. The Chapman amendment had given employers a judgment to make. The conference substitute gave the Secretary of Health and Human Services a list to publish. The discretion moved from the employer’s office to a federal scientific determination, updated annually and disseminated publicly. The House’s public health concern was answered with a federal list. The Senate’s and the advocates’ anti discrimination concern was answered by taking the judgment out of employers’ hands.

The difference between the Chapman amendment and the compromise is the difference between an employer-defined category and a science-defined list. Chapman let the employer decide which diseases of public health significance justified removal from food handling, subject to the alternative-job condition. The compromise took the disease list out of the employer’s hands and gave it to the Secretary of Health and Human Services, bounded it by actual transmissibility through food handling, required publication and annual updating, and preserved the reasonable-accommodation backstop: removal is permitted only where accommodation cannot eliminate the risk. The supporters of the categorical rule lost the category. The supporters of the individualized standard lost the pure individualized standard. What both sides got was a federal list, compiled by public health authorities, that would tell employers, workers, and courts which diseases actually travel through food. It is a characteristically congressional solution: when the two chambers cannot agree on who decides, create a list and let the list decide.

What did the Chapman amendment actually say?

The Chapman amendment, H.AMDT.450, adopted by the House on May 17, 1990, by 199 to 187, added subsection 103(d) to the House bill. It let an employer remove a worker with an infectious or communicable disease of public health significance from food handling, provided the employer offered alternative qualified work with no economic harm.

How did the conference committee resolve the food handling fight?

The conferees stripped the Chapman amendment, defying the Senate’s instruction to accept it, and substituted the provision codified at 42 U.S.C. 12113(d). The substitute directs the Secretary of Health and Human Services to publish and annually update a list of diseases transmitted through food handling.

Only listed diseases, unfixable by reasonable accommodation, would justify removing a worker from food handling jobs under the substitute. The conference report then went to the floors, and the food handling fight got its last hearing in the House. On July 12, 1990, Representative Dannemeyer moved to recommit the conference report with instructions directing the conferees to insist on including the Chapman food handling language. A motion to recommit with instructions is the minority’s last procedural weapon against a conference report: it sends the report back to the conference with directions, and if it passes, the report must be renegotiated. The Dannemeyer motion failed, 180 to 224, recorded as Roll No. 227. The conference report, containing the HHS list compromise instead of the Chapman amendment, was then agreed to by a recorded vote of 377 to 28, Roll No. 228.

Consider the arithmetic of the firewall. The Dannemeyer recommit failed 180 to 224, with 404 members voting. A majority of those voting was 203. The motion’s supporters had 180, which means they were twenty-three votes short. Twenty-three members switching sides would have sent the conference report back, reopened the Chapman fight, and put the whole ADA back into negotiation with the legislative clock running toward adjournment. The margin that saved the bill was not the 377 to 28 vote on the report itself, impressive as that was. It was the 224 votes that refused to reopen the deal. Conference agreements survive on the discipline of members who vote against their preferences on individual provisions to protect the package, and July 12, 1990, is a clean example.

The sequence of the two July 12 votes is the whole story in miniature. The House first voted, 180 to 224, against restoring the Chapman language, and then voted, 377 to 28, for the conference report without it. At least 196 members voted against the Chapman restoration and for the report, or in any case voted in a pattern that produced those two tallies. The food handling defense that had divided the House 199 to 187 in May could not command even 181 votes in July when the question was posed as a choice between the Chapman language and the bill itself. That is what a conference report does to a contested provision. It bundles the provision’s fate with the bill’s fate, and members who cared about the provision more than the bill turned out to be a minority of a minority.

The conference committee’s defiance of the Senate instruction deserves a final procedural note, because it is the kind of move that looks reckless until the vote proves it was calculated. The conferees who stripped the Chapman language were betting that the Senate, faced with a choice between the conference report and no bill, would choose the report. They were also betting that the HHS list compromise gave the Senate’s Chapman supporters enough to accept. Both bets were about to be tested on the Senate floor, where the conference report needed not just a majority but the political cover to make the defiance stick.

The Senate concurs and the president signs

On July 13, 1990, the day after the House agreed to the conference report, the Senate took up the report. The debate was the bill’s last legislative hurdle, and it was here that Senator Harkin delivered the sign language speech that the verification record dates precisely to this debate. Harkin became the first person to address the Senate floor in American Sign Language, opening with a tribute to his brother Frank, who had lost his hearing at the age of five, and then translating his remarks into spoken English: “Congress opens the doors to all Americans with disabilities, today we say no to ignorance, no to fear, no to prejudice.” His stated reason, given later, was disarmingly direct: he wanted Frank, who was deaf, to understand him. The speech is sometimes misdated to the September 1989 passage debate. It belongs to July 13, 1990, to the conference report debate, to the day the Senate decided whether to accept a conference product that had defied its own instruction on food handling.

Why did Senator Harkin deliver part of his floor speech in sign language?

On July 13, 1990, during the Senate’s conference report debate, Harkin became the first person to deliver a Senate floor speech in American Sign Language, opening with a tribute to his deaf brother Frank before translating into English. His reason was personal: he wanted Frank to understand him. The speech marked the bill’s final debate, not its 1989 passage.

The Senate agreed to the conference report by a yea-nay vote of 91 to 6, recorded as Record Vote No. 152. The margin is the largest of the four roll calls in the passage history, and it is the vote that vindicated the conferees’ gamble. The Senate that had instructed its conferees to accept the Chapman language accepted, 91 to 6, a conference report that had stripped it. Fifteen more senators voted yes than had voted yes for the bill’s original Senate passage, and the six no votes were a smaller opposition than the eight who had opposed the bill in September 1989. The HHS list compromise had done its work. Whatever the Senate’s Chapman supporters had wanted from the instruction fight, the conference substitute gave them enough, or the prospect of killing the most significant disability legislation in the country’s history over a food handling provision gave them pause, or both. The 91 to 6 tally does not distinguish among those explanations. It records only the outcome: the conference report passed with votes to spare.

Thirteen days later, on July 26, 1990, the president signed the bill. S. 933 became Public Law 101-336, printed at 104 Stat. 327. The public law number encodes the bill’s place in the 101st Congress’s output: the 336th public law enacted. The public law text’s bracketed “[S. 933]” preserves the bill number in the law itself. The signing statement, preserved by the American Presidency Project, is dated July 26, 1990, from the White House, and notes the public law number. The short title as enacted was the Americans with Disabilities Act of 1990, the House’s styling, not the Senate’s 1989 version. From introduction on May 9, 1989, to signature on July 26, 1990, the bill had traveled fourteen months, four roll calls, one demonstration, one amendment fight, one defied instruction, and one recommit motion. Every stage had left its mark on the text.

The thirteen days between the Senate’s agreement and the signature were the bill’s quietest procedural interval, and they followed the standard final steps. Once both chambers have agreed to the conference report, the bill is enrolled, which means it is prepared in its final form, certified, and signed by the presiding officers of the two chambers. The enrolled bill is then presented to the president, who may sign it into law or return it with objections. No drama attached to these steps for S. 933: the supermajorities in both chambers had removed any doubt about the outcome, and the only remaining question was the date of the ceremony. July 26, 1990, is the date the enrolled bill became Public Law 101-336, and the date from which the statute’s own clocks, including the six month deadline for the HHS food handling list, began to run.

The compromise that saved the bill: 42 U.S.C. 12113(d)

The food handling provision that the conference committee wrote in place of the Chapman amendment is worth walking through clause by clause, because it is the passage history’s most intricate piece of legislative drafting and the clearest evidence of the series’ thesis that procedure determines text. The provision is codified at 42 U.S.C. 12113(d), under the heading “Food handling jobs,” and its mechanics are as follows.

First, the duty. The Secretary of Health and Human Services, not later than six months after the date of enactment, shall publish a list of infectious and communicable diseases which are transmitted through handling the food supply. The six month deadline is the provision’s urgency mechanism: the list cannot be deferred indefinitely, because the statute commands its publication within half a year of July 26, 1990. The diseases listed are not any infectious diseases, and not any diseases of public health significance in the abstract. They are the diseases transmitted through handling the food supply, a category defined by a mode of transmission, not by a level of fear. The list must state the methods by which such diseases are transmitted, which turns the publication into a scientific document rather than a bare roster of names. And the Secretary must disseminate the list to the general public and update it annually, which makes the determination a living one, revisable as medical knowledge changes, rather than a fixed statutory catalog that would freeze the science of 1990 into law.

Second, the employer right. Where an individual has an infectious or communicable disease that appears on the published list, that is transmitted to others through the handling of food, and that cannot be eliminated by reasonable accommodation, a covered entity may refuse to assign or continue to assign the individual to a job involving food handling. Each qualifier does work. The disease must be on the list, which means the employer’s judgment is constrained by the Secretary’s scientific determination; a condition the Secretary has not listed cannot trigger the provision no matter what the employer fears. The disease must be transmitted through food handling, which ties the exclusion to the actual mechanism of the job rather than to the diagnosis in the abstract. And the risk must not be eliminable by reasonable accommodation, which imports the statute’s central balancing concept into the food handling context: if gloves, reassignment of duties within the job, or other adjustments eliminate the transmission risk, the employer must make them rather than removing the worker.

Third, the savings clause. Nothing in the act preempts, or is intended to preempt, state or local laws concerning food handling. The federal provision sets a floor for the employer’s rights and the worker’s protections, but it does not displace stricter or more specific state and local food safety regimes. The savings clause is the kind of provision that gets added when two sovereignties both claim the field, and its presence here reflects the reality that food safety had long been a state and local regulatory domain before Congress entered it through the disability statute. It is the federalism receipt: states and localities with their own food-handler rules kept them, and the federal list set a floor for the ADA’s purposes, not a ceiling for public health regulation.

Set against the Chapman amendment, the conference substitute is a study in how a conference committee narrows a discretion. Chapman had given the employer a defense to assert: the employer determined that the worker had an infectious or communicable disease of public health significance, and the employer removed the worker from food handling, subject to the alternative employment condition. The conference provision gives the Secretary a list to publish and the employer a conditional right that activates only for listed diseases, only for food transmitted conditions, and only where reasonable accommodation cannot eliminate the risk. The employer’s discretion shrinks. The scientific determination grows. The annual update keeps the list honest. And the reasonable accommodation qualifier keeps the statute’s core promise intact even in the food handling context. The provision is longer than Chapman’s, more qualified than Chapman’s, and harder for an employer to invoke than Chapman’s. That is why the advocates could accept it and why the conference could strip Chapman’s language without losing the bill.

Three design choices in the substitute deserve emphasis because they show the conferees thinking about enforcement, not just principle. The six month publication deadline prevents the listing duty from becoming a dead letter. A Secretary who could publish the list at leisure might never publish it, and the conferees, writing under the pressure of a Senate instruction fight, were not inclined to trust leisure. The annual update requirement prevents the list from fossilizing. Medical knowledge about transmission changes, and a list fixed in 1990 would eventually misdescribe the science it was meant to encode. The update duty makes the provision self correcting. And the reasonable accommodation qualifier does the heaviest lifting of all: it ensures that even where a listed disease is transmitted through food handling, the employer must first ask whether an adjustment eliminates the risk before removing the worker. The qualifier imports the statute’s central mechanism into the one context where Congress was most tempted to abandon it.

The compromise the conferees wrote in place of the stripped language is a masterclass in bounded delegation. The drafting also shows the conference committee doing what conference committees are for. The House wanted a specific food handling defense. The Senate wanted the general direct threat defense to suffice. The conference produced neither the House’s employer discretion nor the Senate’s generality, but a third thing: a bounded, science driven, annually updated federal list with a reasonable accommodation qualifier and a state law savings clause. Neither chamber got exactly what it wanted. Both chambers got something they could vote for, 377 to 28 in the House and 91 to 6 in the Senate. That is the conference function performed correctly, and it is why the provision bears the marks of its authorship in every clause.

What the narrowing bought: the arithmetic of the supermajority

The four roll calls of the passage history, set in sequence, tell the story of a coalition that was built before the first vote and held through the last. On September 7, 1989, the Senate passed S. 933 by 76 to 8. On May 22, 1990, the House recorded 403 to 20 for the bill’s text. On July 12, 1990, the House agreed to the conference report 377 to 28, after rejecting the Dannemeyer recommit motion 180 to 224. On July 13, 1990, the Senate agreed to the conference report 91 to 6. The yes votes rise across the sequence in the Senate, from 76 to 91, and the no votes fall, from 8 to 6. In the House, the conference report’s 377 yes votes are fewer than the 403 recorded for the bill’s text in May, which is the expected pattern: a conference report that has stripped a provision one chamber adopted will lose some votes on each side, the Chapman supporters who wanted the amendment and the advocates’ allies who disliked the compromise. The net result, in both chambers, was still a supermajority with room to spare.

The namable claim of this article is that the narrowing made it possible, and the vote sequence is the evidence. The 1988 bill, without the narrowing, could not get a markup. The 1990 bill, with the narrowing, could not be stopped. The concessions are visible in the text: the substantially limits qualifier in the disability definition, the injunctions and fees ceiling in the public accommodations remedies, the phased timelines in place of the two year mandate, and the HHS list mechanism in place of the Chapman employer defense. Each concession answered a specific objection from a specific quarter, and each one converted potential no votes into yes votes or at least into abstentions. The business community negotiated hard on the defenses, the remedies, and the compliance timelines, and the negotiations are visible in the statutory text, which is why the consensus was real and why it was purchased rather than spontaneous.

The counter reading that this article was commissioned to address holds that moral consensus alone carried the bill. The record supports a more qualified version. The moral case for the bill was genuine, the Capitol Crawl changed the political weather, and the personal stories of Coelho, Harkin, and Dole gave the coalition a human face that no whip operation could manufacture. But the moral case had been available in 1988, when the broader bill died without a vote. What changed between 1988 and 1990 was not the country’s moral sentiment. It was the text. The advocates traded coverage breadth for the substantially limits test, traded punitive damages for injunctions and fees, traded the two year mandate for phased timelines, and traded the Chapman fight for the HHS list. Those trades are the reason the bill’s opponents could not assemble a blocking coalition, and they are the reason the bill’s supporters could assemble a supermajority.

The documentary proof that the consensus was negotiated rather than spontaneous is the paper trail of the spring and summer of 1990. Consider what the two interest letters show. On May 31, 1990, the United Food and Commercial Workers, a labor union representing food workers, wrote to the Senate urging rejection of any instruction to accept the Chapman language. On July 9, 1990, the National Restaurant Association, an industry group representing food employers, wrote to senators protesting that the conferees had stripped the Chapman language less than three weeks after the Senate voted to keep it. These are not the letters of bystanders. They are the letters of organized interests fighting over statutory text in real time, and they bracket the conference’s decision from both sides. The union feared the categorical rule would sweep workers with non-transmissible conditions out of food jobs. The industry feared the individualized standard would leave employers guessing. The conference heard both fears and wrote the list. A consensus that produces this kind of contemporaneous lobbying is not a mood. It is a settlement.

Consider also the Chapman vote itself as evidence against the spontaneous-consensus story. A 199 to 187 adoption margin is not the vote of a chamber that has made up its mind. It is the vote of a chamber split down the middle on a provision its own leadership had not fully vetted, and the twelve-vote margin meant the provision’s supporters could plausibly have believed they would win the conference fight too. They had reason for confidence: the Senate then voted to instruct its conferees to accept the language. Two recorded expressions of legislative will, one in each chamber, both favoring the Chapman approach, and the conference overrode both. That override is the strongest single piece of evidence that the ADA’s passage was managed, not merely felt. The conferees decided that the package mattered more than the provision, and they were willing to defy a Senate instruction to prove it.

It is worth pausing on how substantive the business community’s defensive victories were, because there is a version of this story in which the concessions were cosmetic and the supermajority was inevitable. The record does not support that version. The substantially limits qualifier is a real filter: it excludes from the protected class people whose impairments do not substantially limit major life activities, which is a large category under the 1988 bill’s boundless definition. The remedies ceiling is a real economic difference: injunctive relief and attorney’s fees expose a business to the cost of compliance and the cost of the plaintiff’s lawyer, but not to the open ended punishment that punitive damages represent, and the difference between those two exposures is the difference that concentrates the business mind. The phased timelines are a real planning horizon: years to come into compliance across different obligations, rather than a single two year cliff, changes what compliance costs and when. And the HHS listing mechanism is a real constraint on employer discretion: an employer who wants to remove a worker from food handling must point to a published federal list, not to a private judgment about public health significance. None of these are cosmetic. Each one reduced the bill’s cost to its opponents by a measurable amount, and the supermajority is the sum of those reductions.

And consider the shape of the final compromise as evidence of who sat at the table. The 12113(d) list is not a disability rights provision in its origins. It is a public health provision grafted onto a civil rights statute, and its presence in the ADA is the fingerprint of the food industry’s bargaining power. A pure disability rights bill would have relied on the direct-threat defense alone, as the Senate version did. A pure industry bill would have kept Chapman’s categorical employer defense. The enacted provision is neither. It is a hybrid that gives employers a clear rule, gives workers a science-based boundary, gives the health secretary a deadline, and gives the courts a list. Hybrids like that do not emerge from moral consensus. They emerge from conference rooms where both sides have leverage and neither side has enough to win outright. The ADA’s conference room had both.

The ADA’s passage history is therefore the cleanest documented case in this series of the thesis that procedure and negotiation determine text. The two drafts sit a year apart. The differences between them are a matter of record: the definition, the remedies, the timelines. The conference fight that produced the final disputed provision is documented to the level of dated letters and briefing papers. The roll calls show the coalition growing as the text narrowed. Few statutes in the series offer all four of those evidentiary elements at once. The ADA offers them because its first attempt failed in public and its second attempt was negotiated in the open, with the concessions written into the text where any reader can find them.

Two cautions close the analysis, both required by the article’s neutrality flags. First, the concessions should be reported as what they were: negotiated settlements between organized interests, with winners and losers on each point. The disability community won the comprehensive statute and lost the broader 1988 definition, the punitive damages, and the flat timeline. The business community won the narrowed definition, the limited remedies, and the phased compliance, and lost the fight to keep the statute from existing at all. The food handling dispute should be described by its content, as this article has done, rather than by characterization: the House adopted an employer defense conditioned on alternative employment, the Senate relied on the general direct threat provision, and the conference substituted a federal scientific listing with a reasonable accommodation qualifier. Second, the 1990 coalition belongs to 1990. The bipartisan margins of the ADA’s passage describe a specific Congress, a specific bill, and a specific set of negotiations. They are not evidence about any later alignment, and this article draws no conclusions from them about any.

The fourteen month calendar: the bill in sequence

A passage history told as procedure benefits from a straight chronological pass, because the order of events is itself an argument about causation. The Americans with Disabilities Act did not move in a steady line from introduction to signature. It moved in bursts separated by long pauses, and the bursts are where the procedure happened.

The story opens in January 1986, when the National Council on the Handicapped published “Toward Independence” and transmitted it on February 1 to the president, the president of the Senate, and the speaker of the House. For more than two years the report’s recommendation sat as a proposal without a bill. Then, on April 28, 1988, Senator Lowell Weicker introduced S. 2345 in the Senate, and on April 29 Representative Tony Coelho introduced the identical H.R. 4498 in the House. The cosponsor counts grew through the spring and summer, 13 to 41 in the Senate and 47 to 125 in the House, and in September 1988 the two chambers held their unusual joint hearing on the Senate bill. After the hearing, nothing happened. No committee marked up the bill, no floor time was scheduled, and when the 100th Congress adjourned, the bill died with it. The first attempt had lasted barely six months from introduction to adjournment, and its only procedural footprint was a single hearing.

The second attempt began on May 9, 1989, when Senator Tom Harkin introduced S. 933. About five weeks later, on June 15, 1989, Coelho resigned from the House, and Representative Steny Hoyer took over the House’s stewardship of the bill. The Senate moved first and moved fast: on September 7, 1989, four months after introduction, the chamber passed S. 933 by 76 to 8. Then came the long pause. For six months the bill sat while the House did its work, and the pause is where the political weather changed. On March 12, 1990, the Capitol Crawl put the bill’s moral case on the national stage. On May 17, 1990, the House adopted the Chapman amendment 199 to 187, opening the food handling fight. On May 22, 1990, the House passed the bill’s text 403 to 20 at 3:00 PM, substituted it into S. 933 at 3:04 PM, passed S. 933 by voice vote at 3:05 PM, and laid H.R. 2273 on the table.

Read the intervals, not just the dates. Four months from introduction to Senate passage: a pre-negotiated bill moving through a willing chamber. Eight months from Senate passage to House action: the chamber where the real bargaining happened, including the demonstration that changed the weather and the amendment that nearly killed the deal. Five days from the Chapman adoption to the House passage vote: the amendment and the bill decided in the same week by the same members at wildly different margins. About seven weeks from House passage to the conference report: the period when the instruction was adopted, defied, and replaced with the compromise. Thirteen days from final passage to signature: enrollment and presentment. Twenty-seven months from the first bill’s introduction to the second bill’s signature: the full arc from maximalist failure to negotiated success. The chronology is the argument. Bills that move fast in one chamber and slow in the other are telling you where the negotiation lives, and the ADA’s negotiation lived in the House in the spring of 1990 and in the conference room in the summer.

June 1990 belonged to the instruction fight. Senator Helms moved to instruct the Senate conferees to accept the Chapman language, the full Senate voted to adopt the instruction, and the tabling battle over the instruction consumed the month’s procedural energy. The United Food and Commercial Workers union weighed in on May 31, 1990, with a letter urging senators to reject any instruction to recede to the Chapman amendment. The conference committee, undeterred, stripped the Chapman language from its report, replacing it with the HHS listing mechanism. On July 9, 1990, the National Restaurant Association wrote to senators documenting the timeline with precision: the committee had removed the amendment less than three weeks after the Senate voted to instruct its conferees to accept it.

July 1990 was the endgame. On July 12, the House rejected Representative Dannemeyer’s motion to recommit the conference report with instructions to restore the Chapman language, 180 to 224, and then agreed to the conference report 377 to 28. On July 13, Harkin delivered his sign language tribute to his brother Frank on the Senate floor, and the Senate agreed to the conference report 91 to 6. On July 26, 1990, the president signed S. 933 as Public Law 101-336. Fourteen months from introduction to signature, four roll calls, and a conference fight that filled the final three weeks. The calendar shows a bill that was never in danger of dying after September 1989 and never safe from amendment until July 1990.

The pacing also reveals where power sat in each phase. In the first phase, power sat with the drafters: Burgdorf’s text defined the debate, and its maximalism defined the failure. In the second phase, from May to September 1989, power sat with the Senate coalition, Harkin and Dole, who moved a pre-negotiated bill through a willing chamber in four months. In the third phase, from September 1989 to May 1990, power sat with the House, which held the bill for eight months while the demonstration changed the weather and the members fought over the details. In the fourth phase, June and July 1990, power sat with the conferees, who defied an instruction, stripped a provision, and wrote a compromise. And in the final thirteen days, power sat with the enrolling clerks. A bill’s history can be read as the migration of power from phase to phase, and the ADA’s migration runs from the council staff to the Senate to the House to the conference room to the president’s desk.

Reading the four roll calls as one story

The four recorded tallies of the passage history repay being read together, because each one measures a different stage of the coalition and the differences between them are as informative as the margins themselves.

The Senate’s September 7, 1989, vote, 76 to 8, Record Vote No. 173, came on the bill as the Senate wrote it, before the House had added anything. Eighty four senators recorded votes. Seventy six voted yes. In Senate arithmetic, 60 is the number that breaks a filibuster and 67 the number that overrides a veto, so 76 is a supermajority with sixteen votes of cushion above the filibuster line and nine above the veto line. The eight no votes were a small enough minority that the bill’s managers never had to bargain with them. The vote’s message was that the narrowed text had cleared the Senate’s institutional hurdles before the House had even begun.

The House’s May 22, 1990, vote, 403 to 20 on H.R. 2273, Roll No. 123, with 9 members not voting, is the most lopsided tally in the history. Four hundred twenty three members recorded a position, and 403 of them voted yes. Set against the 199 to 187 Chapman vote five days earlier, the passage margin shows a chamber that was nearly unanimous on the bill and nearly evenly divided on a single amendment to it. The 9 not voting are worth noting only for what they do not change: even if all nine had voted no, the margin would have been 403 to 29. The House wanted the bill. The House disagreed about food handling. Both facts are true at once, and the conference existed to reconcile them.

The House’s July 12, 1990, pair of votes is the most revealing. The Dannemeyer motion to recommit with instructions to restore the Chapman language failed 180 to 224, Roll No. 227, with 404 members voting. The conference report was then agreed to 377 to 28, Roll No. 228, with 405 voting. The 152 vote gap between the recommit motion’s 180 supporters and the report’s 28 opponents is the measure of the conference effect. Dozens of members who were willing to vote for restoring the Chapman language were not willing to vote against the conference report that lacked it. That is the structural power of the conference report as a legislative vehicle: it forces a single up or down choice on a package, and members who lose on one provision must decide whether the provision is worth killing the package. On July 12, 1990, the answer was no by a margin of 377 to 28.

The Senate’s July 13, 1990, vote, 91 to 6, Record Vote No. 152, is the vindication of the conferees’ gamble. Ninety seven senators voted. Ninety one voted yes. The Senate that had instructed its conferees to accept the Chapman language accepted, by 91 to 6, a report that had stripped it. The yes votes rose by fifteen from the September 1989 passage vote, and the no votes fell from eight to six. A conference report that defies a Senate instruction is supposed to be in trouble on the floor. This one passed with the largest margin of the four roll calls. The HHS list compromise had given the instruction’s supporters enough to come along, or the cost of killing the bill over food handling had become too high, or both. The tally does not apportion the motives. It records the outcome.

One procedural footnote belongs with the four tallies. S. 933 itself passed the House by voice vote on May 22, 1990, at 3:05 PM, immediately after the 403 to 20 recorded vote on H.R. 2273 and the text substitution. A voice vote, in which the presiding officer asks for ayes and noes and judges the louder side, is used when the outcome is not in doubt and no member demands a recorded vote. That no member demanded a recorded vote on S. 933, five minutes after 403 members had gone on the record for the same text under a different bill number, is itself a datum. The House had said what it thought. Repeating the exercise under the Senate’s bill number would have added nothing but time.

The instruction fight: how the conferees defied the Senate

The June 1990 fight over the Helms motion to instruct deserves a closer procedural reading, because it is the hinge on which the conference turned and because the conferees’ defiance of the instruction is the boldest move in the passage history.

A motion to instruct conferees is one of the few tools a chamber has for controlling its negotiators after they have been appointed. In Senate practice, the motion directs the conferees to take a specified position in the conference, in this case to support the Chapman food handling language. The instruction is adopted by the full Senate, which gives it the chamber’s formal imprimatur, but it is not an absolutely binding order as a matter of parliamentary law. Conferees can defy it. The price of defiance is paid on the floor when the conference report comes back: the chamber that was defied gets to vote the report up or down, and the defiance becomes part of the debate. Senator Helms’s motion, and the Senate’s vote to adopt it, put the conferees on notice that the chamber wanted the Chapman language kept. The tabling fight that surrounded the motion in June 1990, documented in the legislative files preserved from the period, showed that the instruction had organized support and that its opponents had to fight it procedurally rather than simply outvoting it. The June tabling fight over the instruction, documented in the Dole Archives’ Chapman file, shows that the instruction itself was contested; there was a fight about whether to even have the fight.

The organized interests lined up as the instruction fight made the stakes clear. The United Food and Commercial Workers union wrote to the Senate on May 31, 1990, urging senators to reject any motion to instruct the Senate conferees to recede to the House approved Chapman food handler amendment. The union’s position is a reminder that the food handling fight was not only about disability rights doctrine. It was about the workers who handle food, their unions, and the conditions under which they could be removed from their jobs. The National Restaurant Association, writing to senators on July 9, 1990, took the other side of the ledger, documenting with some heat that the conference committee had stripped the Chapman amendment less than three weeks after the full Senate voted to instruct its conferees to accept the language. The letter is preserved in the Dole Archives’ Chapman file, along with the Senate briefing paper on the Helms motion and the scanned text of the conference substitute, which is why the mechanics of this fight are recoverable in unusual detail.

The conferees’ decision to defy the instruction was a calculated risk, and the calculation is worth reconstructing. The conferees knew the Senate had voted to keep the Chapman language. They also knew that the disability community and its allies regarded the Chapman language as an unacceptable employer license to exclude workers on the basis of fear, and that keeping it would cost votes on the other side. A conference report must pass both chambers, and the conferees’ job is to find the text that maximizes the combined yes votes, not to obey one chamber’s instruction at the expense of the other’s acceptance. By stripping Chapman and substituting the HHS list mechanism, the conferees bet that the Senate’s instruction supporters would accept the substitute and that the House, which had adopted Chapman 199 to 187, would accept its removal. The July 12 and 13 votes proved the bet correct in both chambers. The instruction fight is therefore the passage history’s clearest demonstration of the conference committee’s independence: instructed by the Senate to keep a provision, the conferees removed it, replaced it with a compromise of their own drafting, and were sustained by both floors.

The drafting chain: from a council staffer to the statute book

The Americans with Disabilities Act had an unusually visible drafting chain, and following it from the advisory council to the enrolled bill shows how the narrowing entered the text.

The chain begins with Robert Burgdorf, a staffer of the National Council on the Handicapped, who drafted the 1988 bill from the Council’s “Toward Independence” report. Burgdorf’s draft is the purest expression of the report’s concept: comprehensive coverage, no substantially limits qualifier, punitive and compensatory damages against businesses, and a two year deadline for full accessibility. The draft was introduced without the softening that a committee markup would normally apply, because it never got a markup. It went from the drafter’s desk to introduction on April 28 and 29, 1988, to a single joint hearing in September, to death at adjournment. The 1988 bill is therefore a clean specimen of advocacy drafting, and its failure is a clean test of what advocacy drafting can and cannot achieve in Congress.

The chain continues with the redrafting between the two bills. The memo and the brief agree that the enacted statute is the renegotiated second attempt, which means the narrowing happened in the drafting room between the 100th and 101st Congresses rather than on the floor of either chamber. The substantially limits qualifier, the injunctions and fees ceiling, and the phased timelines were all in the bill when Harkin introduced S. 933 on May 9, 1989. No floor amendment added them. No conference produced them. They were the price of reintroduction, paid before the second bill ever saw a vote. This is why the Senate could pass the bill four months after introduction: the fights that normally consume a bill’s committee stage had been fought in advance, in the drafting, and the bill that emerged was already a compromise.

The chain passes through the authors. Coelho, the primary author and sponsor in the House, introduced the 1988 companion and would have authored the 1989 House effort had he not resigned on June 15, 1989. Hoyer, who carried the bill forward after the resignation, inherited a text that was already narrowed and a coalition that was already built. Harkin, who introduced S. 933 in the Senate, carried the bill through all fourteen months from introduction to conference report, providing the continuity of authorship that the House side lost with Coelho’s departure. The drafting chain thus has a break in it, the resignation, and the fact that the bill survived the break is evidence that the narrowing and the coalition were properties of the text and its support rather than of any single author.

The chain ends in the conference committee, which wrote the only provision of the final statute that was drafted in the last weeks: the food handling substitute codified at 42 U.S.C. 12113(d). Every other contested provision was settled before May 1989. The food handling language was settled in late June and early July 1990, by conferees working under the pressure of a Senate instruction they had decided to defy. The provision’s intricacy, the six month deadline, the methods of transmission, the annual update, the reasonable accommodation qualifier, the savings clause, bears the marks of that pressure. It is the most carefully conditional language in the statute, because it had to satisfy the most divided set of negotiators.

The documentary record: where this history comes from

A passage history is only as reliable as its sources, and the ADA’s passage is unusually well documented, which is why the narrowing story can be told with confidence rather than reconstructed from inference. The sources fall into four groups, and each group anchors a different part of the narrative.

The first group is the congressional record itself. The bill’s action log on congress.gov records S. 933’s introduction on May 9, 1989, its Senate passage on September 7, 1989, by 76 to 8, the House’s July 12, 1990, agreement to the conference report by 377 to 28, and the Senate’s July 13, 1990, agreement by 91 to 6, with the roll call numbers attached to each. The House Clerk’s vote pages supply the detail the action log summarizes: the May 22, 1990, 403 to 20 vote on H.R. 2273 with 9 not voting, the 199 to 187 Chapman amendment vote on May 17, 1990, and the 180 to 224 Dannemeyer recommit vote on July 12, 1990. The Congressional Record citations pin the 1988 introductions to the page: 134 Cong. Rec. 9375 for Weicker’s April 28 introduction of S. 2345, 134 Cong. Rec. 9600 for Coelho’s April 29 introduction of H.R. 4498, and 136 Cong. Rec. H10911 for the Chapman amendment’s May 17, 1990, adoption. These are the load bearing facts of the passage history, and they are all matters of public record rather than recollection.

The second group is the archival record of the food handling fight. The Dole Archives at the University of Kansas holds the Chapman file, including the Senate briefing paper on the Helms motion to instruct conferees, which records that the Chapman amendment was agreed to in the House on May 17, 1990, by 199 to 187 and added subsection 103(d) to the House bill. The same file holds the National Restaurant Association’s July 9, 1990, letter to senators documenting that the conference stripped the amendment less than three weeks after the Senate’s instruction vote, and the scanned text of the conference substitute that became 42 U.S.C. 12113(d). The United Food and Commercial Workers union’s May 31, 1990, letter to the Senate, preserved in the legislative files at Carnegie Mellon, shows organized labor urging senators to reject the instruction to recede to the Chapman language. Without these archives, the conference fight would be a matter of conflicting memories. With them, it is a matter of dated documents.

The third group is the secondary literature that the passage history draws on for the 1988 bill’s content. Chai Feldblum, Katie Beauchamp, and Lauren Taylor’s 2008 history of the ADA Amendments Act, published in the Texas Journal on Civil Liberties and Civil Rights, supplies the introduction dates, the cosponsor counts, and the critical procedural fact that the only legislative activity on S. 2345 was the single September 1988 joint hearing. The ADA Legacy Project’s “Moments in Disability History” series supplies the cosponsor growth figures, 13 to 41 in the Senate and 47 to 125 in the House, and the judgment that Weicker’s bill had much stronger language and broader civil rights policy than the enacted act. Joe Shapiro’s No Pity, as quoted in the Legacy Project’s account, supplies the two most vivid specifics of the 1988 bill’s breadth: the punitive damages remedy against businesses and the two year flat accessibility mandate. The Equal Employment Opportunity Commission’s history of the law’s development confirms that compensatory and punitive damages in intentional employment discrimination cases did not arrive under the ADA until the Civil Rights Act of 1991.

The fourth group is the biographical and commemorative record. The National Council on Disability’s amicus brief in Olmstead v. L.C. and its “Equality of Opportunity” history document the 1986 “Toward Independence” report, its first recommendation, and its proposed title. The Epilepsy Foundation’s account of Coelho and Coelho’s own site record the epilepsy diagnosis at 22, the six terms in the House, and the primary authorship. The Des Moines Register’s interview with Harkin and the New Mobility profile record the brother Frank, the July 13, 1990, sign language speech, and Harkin’s stated reason for it. The House Report 107-474 and the Congressional Record’s Dole findings record the 1945 wound, the decorations, the minority leadership, and the lifelong advocacy. The Museum of Disability History and the Zinn Education Project record the Capitol Crawl’s date, its organization by ADAPT, and its numbers. The American Presidency Project preserves the July 26, 1990, signing statement. A passage history built on these sources is built on documents that any reader can check, which is the standard this series sets for itself.

Who won what: the concessions by name

The neutrality flags for this article require reporting the concessions and the parties who won them by name, describing the food handling dispute by its content, and drawing no conclusions about any present day alignment from the 1990 coalition. This section does that accounting.

The disability community won the comprehensive statute it had sought since the 1986 report: a clear prohibition of discrimination on the basis of disability, enacted as Public Law 101-336. It lost the broader 1988 definition, accepting the substantially limits qualifier. It lost punitive and compensatory damages against businesses in the 1990 bill, accepting injunctions and fees in the public accommodations title and Title VII remedies in employment, with damages deferred to the Civil Rights Act of 1991. It lost the two year flat mandate, accepting phased timelines. And in the conference, it lost the outright defeat of any food handling defense, accepting instead the HHS listing mechanism, which constrains employer discretion but preserves a federal role in the question. The community’s negotiators traded breadth for enactment at every point, and the trades are visible in the text.

The business community won the narrowed definition, the limited remedies, and the phased compliance schedules, the three concessions the brief identifies as the ones it negotiated hardest. It lost the larger fight: the statute exists, it is comprehensive, and it binds every employer, government, and place of public accommodation in the country. The business community’s victories were all defensive, limitations on a bill it could not stop, and the limitations are real. A substantially limits test that excludes marginal impairments, a remedies scheme without punitive damages, and timelines measured in years rather than months are substantive protections for regulated entities, not cosmetic ones.

On the food handling question specifically, the winners and losers can be named from the roll calls. Representative Chapman won the House vote on May 17, 1990, 199 to 187, and lost everything after: the conference stripped his amendment, and the Dannemeyer motion to restore it failed 180 to 224. Senator Helms won the instruction vote in June 1990 and lost the outcome, as the conferees defied the instruction and the Senate sustained the defiance 91 to 6. Representative Dannemeyer lost the recommit motion that was the last chance to save the Chapman language. The advocates who opposed the Chapman amendment lost the May 17 vote and won the conference, the recommit vote, and the final tallies. The HHS listing mechanism was the conferees’ invention, and its authorship belongs to the conference committee rather than to either side of the floor fight.

The coalition managers won the largest prize, which was the bill itself. Harkin carried the Senate from introduction through the conference report and delivered the sign language speech that closed the debate. Dole, as minority leader, gave the bill the Republican cover that made the supermajorities possible. Hoyer carried the House effort across Coelho’s resignation without a visible break. Coelho, though gone from the House before the bill passed, remained the primary author and sponsor in the record, the member whose own epilepsy had made the bill personal before it was political. The supermajorities were purchased with the concessions catalogued above, and the purchasers can be named. That is the accounting the record supports, and it is the antidote to the version of the story in which moral consensus alone carried the bill.

Closing: procedure as the author of the text

The Americans with Disabilities Act passed because its authors learned the lesson of 1988 and applied it with unusual discipline. The first bill asked for everything and got nothing. The second bill asked for slightly less than everything, in the places where slightly less was the price of passage, and got a statute. The report that started it, the demonstration that changed its weather, the three legislators who carried it, the four roll calls that recorded its margins, and the conference fight that nearly killed it are all part of the standard telling. The procedure is the part that explains the text. Every narrowed provision in the enacted statute is a fossil of a negotiation, and the negotiations are the reason the mandate could be as broad as it was.

There is a final procedural moral, and it is the reason this article belongs in a series about how Congress works rather than in a celebration of a single statute. The ADA’s passage shows that in the American legislature, breadth is the enemy of enactment and negotiability is its precondition. The 1988 bill was broad and died. The 1990 bill was negotiable and lived. The negotiability was not a betrayal of the cause. It was the cause’s translation into the only language Congress speaks, which is the language of provisions that can be amended, defenses that can be bounded, timelines that can be phased, and lists that can be compiled. The members who learned that language between 1988 and 1990, Weicker’s and Coelho’s successors, Harkin, Dole, Hoyer, Chapman, Helms, Dannemeyer, and the conferees who defied an instruction to save a statute, wrote one of the most consequential civil rights laws in American history by treating it as procedure. The feel-good version remembers the crawl up the steps. The procedural version remembers everything, and it is the better memory.

Readers who want to work through the roll calls, the two bill texts, and the conference substitute in detail can keep their notes organized in a legislation study notebook. The passage history ends at the president’s signature on July 26, 1990. The statute’s life after enactment, in the agencies, the courts, and the workplaces it governs, belongs to the other articles in the cluster.

The One Test for this article asked whether a reader could explain how a sweeping mandate passed by overwhelming bipartisan margins, name the act of civil disobedience that changed the political weather, describe the earlier and more sweeping bill that failed, and identify the amendment that nearly destroyed the conference agreement. The answers are the 1990 narrowing, the Capitol Crawl of March 12, 1990, S. 2345 of 1988, and the Chapman amendment. They are procedure, not sentiment, and they are the reason the Americans with Disabilities Act is Public Law 101-336 rather than a forgotten bill from the 100th Congress.

Frequently Asked Questions

Q: How did the Americans with Disabilities Act pass Congress?

The bill, S. 933, was introduced in the Senate on May 9, 1989, by Senator Tom Harkin, passed the Senate 76 to 8 on September 7, 1989, and then moved to the House, where a food handling amendment adopted 199 to 187 on May 17, 1990, set up the conference fight. The House passed the bill’s text 403 to 20 on May 22, 1990, substituted it into S. 933, and passed S. 933 by voice vote the same afternoon. The conference committee stripped the food handling amendment over the Senate’s instruction to keep it, substituting a federal disease listing mechanism. The House agreed to the conference report 377 to 28 on July 12, 1990, after rejecting a recommit motion 180 to 224, the Senate agreed 91 to 6 on July 13, and the president signed it as Public Law 101-336 on July 26, 1990.

Q: What was the Capitol Crawl before the Americans with Disabilities Act?

On March 12, 1990, disability rights activists organized by ADAPT marched from the White House to the Capitol, left their wheelchairs and mobility aids at the base of the west Capitol steps, and crawled and dragged themselves up the stairs. More than sixty activists made the climb while about a thousand supporters watched, and the event was photographed and broadcast widely. The Capitol Crawl did not change any vote directly, since the House passed the bill 403 to 20 two months later, but it changed the political cost of opposition by converting the bill from a regulatory question into a civil rights question in the public mind. After the crawl, the remaining fights were about the scope of defenses and remedies rather than whether disability discrimination belonged in federal civil rights law.

Q: Who wrote the Americans with Disabilities Act?

The primary author and sponsor in the House was Representative Tony Coelho of California, who had been diagnosed with epilepsy at age 22 and served six terms. Coelho introduced the House companion of the failed 1988 bill and was the natural author of the 1989 effort, but he resigned from the House on June 15, 1989, and Representative Steny Hoyer carried the bill forward. In the Senate, the author was Senator Tom Harkin of Iowa, who introduced S. 933 on May 9, 1989, and whose older brother Frank was deaf. The 1988 bill had been drafted by Robert Burgdorf, a staffer of the National Council on the Handicapped, working from the Council’s 1986 “Toward Independence” report. Senator Bob Dole, the Republican minority leader, was the bill’s most important Republican champion though not its author.

Q: Was the Americans with Disabilities Act bipartisan?

Yes, by every available measure. The Senate passed S. 933 by 76 to 8 in September 1989, the House recorded 403 to 20 for the bill’s text in May 1990, and the conference reports passed 377 to 28 and 91 to 6 that July. Margins of that size cannot be produced by one party alone. The bipartisanship was personal as well as numerical: the Democratic author in the Senate, Tom Harkin, had a deaf brother; the Republican minority leader, Bob Dole, was a disabled veteran; and the House’s primary author, Tony Coelho, had epilepsy himself. The 1988 bill’s sponsors were also cross party, with Republican Senator Lowell Weicker introducing S. 2345 and Democratic Representative Coelho introducing its House companion. The coalition was built into the bill from the first draft.

Q: What nearly killed the Americans with Disabilities Act in conference?

A House adopted amendment on food handling, offered by Representative Jim Chapman of Texas and adopted 199 to 187 on May 17, 1990, let employers remove workers with infectious or communicable diseases of public health significance from food handling jobs. The Senate had no such provision and relied on the general direct threat defense. Senator Helms moved to instruct the Senate conferees to accept the Chapman language, and the Senate voted to do so. The conference committee then stripped the amendment anyway, less than three weeks after the instruction vote, substituting a compromise that directed the Secretary of Health and Human Services to publish a list of diseases transmitted through food handling. When the conference report reached the House floor on July 12, 1990, a motion to recommit it with instructions to restore the Chapman language failed 180 to 224, and the report passed 377 to 28.

Q: What was the final tally on the Americans with Disabilities Act?

The final votes were the two conference report votes in July 1990. The House agreed to the conference report on July 12, 1990, by a recorded vote of 377 to 28, after first rejecting Representative Dannemeyer’s motion to recommit with instructions to restore the Chapman food handling amendment by 180 to 224. The Senate agreed to the conference report on July 13, 1990, by a yea-nay vote of 91 to 6. The earlier passage votes were also lopsided: the Senate passed S. 933 by 76 to 8 on September 7, 1989, and the House recorded 403 to 20 for the bill’s text on May 22, 1990, before passing S. 933 itself by voice vote the same afternoon. The president signed the bill as Public Law 101-336 on July 26, 1990.

Q: Why did an earlier Americans with Disabilities Act bill fail in 1988?

S. 2345, introduced April 28, 1988, by Senator Lowell Weicker, with the identical House companion H.R. 4498 introduced the next day by Representative Tony Coelho, was substantially broader than the enacted statute. It protected any impairment without a substantially limits test, allowed punitive and compensatory damages against businesses, and demanded full accessibility within two years. It was also introduced late in the session: the only legislative activity was a single joint House and Senate hearing in September 1988, with no markup and no floor vote in either chamber, and the bill died with the adjournment of the 100th Congress. It attracted cosponsors, growing from 13 to 41 in the Senate and 47 to 125 in the House, but it never built the negotiated coalition its breadth required. The 1989 bill succeeded by narrowing each of the aggressive provisions before introduction.

Q: Which members of Congress had personal reasons to back the Americans with Disabilities Act?

Three champions carried personal connections to disability. Representative Tony Coelho of California, the bill’s primary House author and sponsor, had been diagnosed with epilepsy at age 22 after being barred from the priesthood under canon law. Senator Tom Harkin of Iowa, who introduced S. 933, had an older brother, Frank, who lost his hearing at age five; Harkin later delivered part of his final floor speech in American Sign Language so his brother could understand him. Senator Bob Dole of Kansas, the Republican minority leader, had been seriously wounded in Italy in 1945, leaving him permanently disabled, and the Congressional Record describes him as a lifelong advocate for the disabled and a key figure in the act’s passage. Their cross party personal investment is the practical explanation for the supermajority margins.

Q: What did the Toward Independence report recommend?

“Toward Independence,” published in January 1986 by the National Council on the Handicapped under a statutory mandate in Public Law 98-221, was the origin document of the ADA. Its first recommendation was that Congress enact a comprehensive law requiring equal opportunity for individuals with disabilities, with broad coverage and with clear, consistent, and enforceable standards prohibiting discrimination on the basis of handicap. The report also proposed the statute’s name: at page 18 it suggested the new law be titled the “Americans with Disabilities Act.” Transmitted to the president and congressional leaders on February 1, 1986, the report gave the eventual sponsors both a policy blueprint and a finished banner, and the 1988 bill was drafted directly from it by the Council’s staffer Robert Burgdorf. The 1990 statute honored the recommendation’s structure while qualifying its reach.

Q: Who introduced the 1988 version of the Americans with Disabilities Act?

Senator Lowell Weicker, Republican of Connecticut, introduced S. 2345, the Americans with Disabilities Act of 1988, in the Senate on April 28, 1988. Representative Tony Coelho, Democrat of California, introduced the identical House companion, H.R. 4498, the next day, April 29, 1988. The bill opened with 13 Senate cosponsors and 47 House cosponsors, growing to 41 and 125 respectively before the session ended. The drafting had been done by Robert Burgdorf of the National Council on the Handicapped, working from the Council’s 1986 “Toward Independence” report. The bill received a single joint House and Senate hearing in September 1988 and no further legislative activity; it died with the adjournment of the 100th Congress. Weicker and Coelho are therefore the sponsors of the failed first attempt, distinct from Harkin, the sponsor of the enacted second attempt.

Q: How did the 1990 law narrow the 1988 bill’s definition of disability?

The 1988 bill protected any person discriminated against because of a physical or mental impairment, a perceived impairment, or a record of impairment, with no further test. The category was as broad as the concept of impairment itself. The 1990 statute kept the three prongs but added the limiting test the 1988 draft had omitted: the impairment must substantially limit one or more major life activities. That qualifying phrase narrowed the protected class to people whose impairments have substantial effects and answered the objection that the broader language would have covered nearly everyone. It is the most consequential single narrowing in the passage history and the most litigated language in the statute. The concession was deliberate, made during the redrafting between the two bills, and it is one of the prices the advocates paid for the supermajority.

Q: Why did the 1990 law drop the 1988 bill’s punitive damages against businesses?

The 1988 bill allowed people with disabilities to sue businesses for punitive and compensatory damages, the remedy the business community feared most because punitive awards are uncapped by compensatory logic and designed to punish. The 1990 statute took that weapon off the table: its public accommodations title was limited to injunctive relief and attorney’s fees, and its employment remedies were tied to Title VII of the Civil Rights Act of 1964, which at the time meant injunctive relief and back pay. Compensatory and punitive damages in intentional employment discrimination cases did not become available under the ADA until the Civil Rights Act of 1991, a separate statute from the following Congress. Deferring the damages fight was one of the negotiated prices of the 1990 supermajority.

Q: How did compliance deadlines change between the 1988 bill and the 1990 law?

The 1988 bill required that within two years everything would have to be accessible, a single flat deadline with no phasing that the journalist Joe Shapiro called a radical flat earth mandate. The enacted 1990 statute replaced it with negotiated phased timelines that spread different obligations across different schedules, giving covered entities years rather than months to comply. A bill written by advocates alone sets one deadline; a bill written to pass sets several, each one marking a point where an objection was answered with time instead of exemption. The phases are the fingerprints of negotiation between the disability community, which wanted prompt accessibility, and the business community, which needed workable schedules. The two year mandate was one of the three confirmed ways the first bill was broader than the second.

Q: What did the Chapman amendment to the Americans with Disabilities Act actually say?

The Chapman amendment, H.AMDT.450, offered by Representative Jim Chapman of Texas and adopted by the House on May 17, 1990, by 199 to 187, added subsection 103(d) to the House bill. It provided that an employer could refuse to assign, or continue to assign, an employee with an infectious or communicable disease of public health significance to a food handling job, provided the employer offered the employee an alternative employment opportunity for which the employee was qualified and as to which the employee would sustain no economic damage. The Senate never adopted the provision and relied instead on the statute’s general direct threat defense. The amendment was stripped in conference and replaced by the federal disease listing mechanism codified at 42 U.S.C. 12113(d).

Q: How did the conference committee resolve the food handling dispute?

The conference committee stripped the Chapman amendment from its report, less than three weeks after the full Senate had voted to instruct its conferees to accept the language, and substituted a new provision codified at 42 U.S.C. 12113(d). The substitute directs the Secretary of Health and Human Services, within six months of enactment, to publish a list of infectious and communicable diseases transmitted through handling the food supply, with methods of transmission stated, to disseminate it publicly, and to update it annually. An employer may remove a worker from food handling only for a listed disease transmitted through food handling where the risk cannot be eliminated by reasonable accommodation. State and local food handling laws are expressly not preempted. The discretion moved from the employer’s judgment to a federal scientific determination.

Q: Why did the House pass the ADA twice in one afternoon in May 1990?

At 3:00 PM on May 22, 1990, the House passed H.R. 2273, its own companion bill, by a recorded vote of 403 to 20, with 9 not voting. At 3:04 PM it took up S. 933, the Senate passed bill, struck everything after the enacting clause, and substituted the H.R. 2273 text, all without objection. At 3:05 PM it passed S. 933 itself by voice vote and laid H.R. 2273 on the table. The substitution is the standard House route to conference: by putting the House text inside the Senate’s bill number, the House created a single vehicle the Senate could act on and that conferees could reconcile. Strictly speaking, the 403 to 20 recorded tally belongs to H.R. 2273’s passage, while S. 933 passed the House by voice vote minutes later.

Q: What was the Dannemeyer motion to recommit during the July 1990 debate?

On July 12, 1990, when the conference report reached the House floor, Representative Dannemeyer moved to recommit the report with instructions directing the conferees to insist on including the Chapman food handling language that the conference had stripped. A motion to recommit with instructions is the minority’s last procedural weapon against a conference report: had it passed, the report would have gone back for renegotiation. The motion failed 180 to 224, recorded as Roll No. 227. The conference report, containing the HHS disease listing compromise instead of the Chapman amendment, was then agreed to 377 to 28, Roll No. 228. The two votes together show that the food handling defense, adopted 199 to 187 in May, could not command even 181 votes in July when posed as a choice between the Chapman language and the bill itself.

Q: Why did Senator Harkin deliver part of his floor speech in sign language?

On July 13, 1990, during the Senate’s final debate on the conference report, Senator Tom Harkin became the first person to deliver a Senate floor speech in American Sign Language. He opened with a tribute to his older brother Frank, who had lost his hearing at age five, and then translated his remarks into spoken English: “Congress opens the doors to all Americans with disabilities, today we say no to ignorance, no to fear, no to prejudice.” Asked later why he used sign language, Harkin said he wanted Frank, who was deaf, to understand him. The speech is sometimes misdated to the September 1989 passage debate, but it belongs to the July 1990 conference report debate, the bill’s last legislative hurdle, which the Senate cleared 91 to 6 the same day.

Q: How did the Senate’s 1989 passage vote compare with its 1990 conference vote?

On September 7, 1989, the Senate passed S. 933 by 76 to 8, Record Vote No. 173. On July 13, 1990, it agreed to the conference report by 91 to 6, Record Vote No. 152. The yes votes rose by fifteen and the no votes fell by two across the ten month interval. The increase is the measure of what the conference process accomplished: the report that came back to the Senate had stripped the Chapman food handling amendment over the Senate’s own instruction to keep it, substituting the HHS disease listing compromise, and the Senate accepted the defiance overwhelmingly. The 91 to 6 margin is the largest of the four roll calls in the passage history and the vote that vindicated the conferees’ gamble that the chamber would choose the report over no bill.

Q: Why did the Senate vote to instruct its conferees to accept the Chapman language?

After the House adopted the Chapman food handling amendment in May 1990, Senator Helms moved to instruct the Senate conferees to support the Chapman language in conference, and the full Senate voted to adopt the instruction. A motion to instruct is the chamber’s formal direction to its conferees about the position to take, and the vote showed that the food handling question had organized a real constituency in the Senate, with the National Restaurant Association and others pressing for the employer defense. The instruction was not absolutely binding as a matter of parliamentary law, and the conference committee defied it, stripping the Chapman amendment less than three weeks later and substituting the HHS listing mechanism. The Senate then accepted the defiance 91 to 6. The instruction vote is the high water mark of the Chapman amendment’s support.