J. Pierpont Morgan was a ruthless financial wizard with superb taste, whose monument--the library designed by McKim, Mead & White in 1906--was a surprisingly restrained product of America's first gilded age. Still more astonishing in that country's latest era of obscene excess, the Morgan has been doubled in size without losing its distinctive personality. The Renzo Piano Building Workshop has wrought its customary magic in weaving together old and new, strengthening the sense of place, and opening up the new central court to views of the street on three sides. Visitors walking into this serene, light-filled atrium, or looking down from two upper-level balconies can savour the sensation of floating within a transparent bubble at the heart of the metropolis.

Nearly all museums have a compulsion to expand, to display more of their holdings and find room for new acquisitions, but also to accommodate ever-greater crowds and boost revenue. A happy few, like the Frick, stay small and are cherished for doing so. In contrast, the Museum of Modern Art abandoned its early role as a tightly focused shrine of the avant garde, and turned itself into an overpoweringly vast emporium with all the appeal of a convention centre. By choosing Piano, who cares as much for the sacred (contemplating art) as the profane (socialising, shopping and eating) and manages to keep the two kinds of space distinct, the Morgan avoided that fate.

The institution badly needed more gallery and storage space for its 350 000-item collection of rare books, master drawings, and manuscripts that range from priceless medieval miniatures to musical scores and correspondence from Ernest Hemingway, plus a better performance space for its renowned concerts. It also wanted to appear less intimidating (Morgan's library was a hermetic strong-box, designed to exclude the hoi polloi and natural light) and to develop its role as an art museum.

For the architects, the challenge was to find a footprint on which to build. The library, the 1850s Morgan family brownstone to the north, and the Classical-style annex that J. P.'s son added in 1928 were all listed properties, and the spaces between were cluttered with later additions. The Landmarks Commission would have opposed a tower. The solution was to clear the additions and to go down, blasting out the Manhattan schist to a depth of 18 metres to accommodate three levels of storage vaults, and a steeply raked auditorium. More than 50 per cent of the 13 800sqm complex is now located below ground. Three new pavilions have been inserted between the existing buildings: offices on 37th Street to the north, a 6m cube called the Thaw Gallery to the south, and a three-storey entry pavilion on Madison Avenue that, in its transparency, offers a symbolic welcome mat. New and old structures frame the 15m, glass-roofed courtyard, evoking an Italian piazzetta.

The Americans With Disabilities Act (ADA) of 1990 is "An Act to establish a clear and comprehensive prohibition of discrimination on the basis of disability" (Preamble, ADA, 1990). One definition of disability discrimination given in the law is the failure to provide accommodation (ADA, 1990). The Equal Employment Opportunity Commission (EEOC; 1999) called the accommodation request process a fundamental feature of the ADA. The ADA does not mandate employment, it prohibits discrimination.

Research into the employment goals of the law makes the assumption that the ADA accommodation request process is viable (DeLeire, 2000; Moon, Chung, & Yang, 2003; Stapleton & Burkhauser, 2003, Wells, 2001). These studies, some using the same data sources, generate opposing answers on the ADA's impact on employment all the while ignoring the fundamental request process of the ADA. Other research looks into the behaviors of entities covered by the law, but these may not be relevant to the ADA's goals. For example, Bruyere (1999) surveyed 1,402 human resource departments on their preparedness to accommodate and reported that most said they were making changes. However, there was no way of knowing from that study whether the changes were initiated by, or even affected employees and/or job seekers with disabilities. Hernandez, Keys, and Balcazar's (2000; 2004) indicated that employers and representatives of the private and public sector express positive attitudes about the employment and access rights of workers with disabilities, but their behaviors may not match their attitudes. Relevant ADA behaviors have not been researched.

The meaning of the attitudes and opinions of people with disabilities in terms of their actual behaviors and the effect of ADA is also unclear. People with disabilities know about the law and are in favor of it (NOD/Harris, 2000; 2002), and some think it has not accomplished much (Hinton, 2003; NOD/Harris, 2004). The National Council On Disability's (NCD) (1995) collection of testimonies from people with disabilities affirmed the well known benefit of receiving accommodations. However, the NCD study did not describe the ADA request process in light of disability discrimination, that is, the refusal to accommodate. The functional impact of the law--the behaviors of the people who could make ADA requests and those who receive ADA requests is not known.

When the ADA complaint processes are used they are not effectual means of acquiring accommodation. In a comprehensive study of all ADA EEOC employment discrimination charges (N = 149,143) between July, 1992 and September, 2000, Moss, Burris, Ullman, Johnsen, and Swanson (2001) found that most complaints were rejected and that when a complaint was accepted, the person with the disability lost most of the time. Furthermore, Moss et al. found that the result of a fully processed EEOC complaint, win or lose, was most often simply a letter to the complainant. The authors concluded that the EEOC complaint process is ineffectual, and noted, "Aside from a chance to tell their stories, most claimants will not benefit from filing a claim." Colker (2000) reviewed ADA litigation and concluded that covered entities knew it was highly unlikely that they would ever face any enforcement action for noncompliance.

Sullivan (2001) found the major reason people charging disability discrimination lost their litigation was because of a breakdown in the negotiation process whereby an employee and an employer discuss what accommodation is needed. He noted the ways employees were blamed for contributing to that breakdown but wrote that the ways employers affected that breakdown were too varied to list. In contrast, Harlan, and Robert (1998) listed some of the ways employers effect the request for accommodation process. They labeled these tactics "employer resistance strategies." These included generating fear of reprisal, giving misinformation, pejorative labeling of requesters as "lazy" or "trouble makers," and telling requesters accommodations would take a long time to arrive and would be of poor quality. Informants in the Harlan and Robert study said they did not request accommodation because they feared reprisals such as being fired, or passed over for promotion. Intimidation, rather than outright refusal followed by a redress process is a means some employers use to avoid fulfilling ADA requests.

Harlan and Robert (1998) found that professionals and managers with disabilities did not negotiate for accommodations. At that level, accommodations were a perquisite of the position. At the entrance level, Hauser, Maxwell-McCaw, Leigh, and Gutman (2000) found a blanket refusal to accommodate or hire Ph.D. clinical psychology interns who are deaf or hard of hearing. Some of McNeal, Somerville, and Wilson's (1999) informants with post polio syndrome did not make requests out of fear of being fired.

Another impairment that is specifically mentioned in the law is visual impairment. However, instead of examining disability discrimination against people who are blind in the accommodation request process, the literature focuses on their assumed deficits. The Director of the National Eye Institute, Dr. Carl Kupfer, launched an education campaign saying "The major problem is that people with low vision do not seek help." (National Eye Institute, 2000). They were the problem. Rumrill (2001) wrote that the major problems with the ADA were that people who are blind lacked knowledge of what to request, and the limits of their rights, and skills for effective communication. Such opinions, in effect, blame people with disabilities for the discrimination they experience.

If at first you don't succeed, try, try again" didn't spring from the Americans with Disabilities Act (ADA), but it may as well have, as the law creates an ongoing obligation to reasonably accommodate qualified individuals with disabilities.

"Many managers are confused about how much they have to do and how far they have to go" when workers request accommodation, noted Sharon Rennert, senior attorney adviser at the Equal Employment Opportunity Commission (EEOC).

[ILLUSTRATION OMITTED]

"It is imperative for employers and managers to understand that the reasonable accommodation obligation is an ongoing obligation," she said, citing Humphrey v. Memorial Hospitals Association (239 F.3d 1128 (9th Cir. 2001)). That ruling states that "the duty to accommodate is a continuing duty that is not exhausted by one effort."

The fact that you've provided one accommodation and it works or doesn't work "does not necessarily mean you're done," she observed. "Employers need to stay on top of" accommodation requests, according to Rennert, who said that employers should:

* Research possible accommodations thoroughly, checking first with the employee requesting accommodation, but also checking, for example, with the Department of Labor's Job Accommodation Network ((800) ADA-WORK), the EEOC, organizations representing people who have the same disability as the employee and the vocational rehabilitation agency.

* Continue checking with the employee for input and ideas.

Second Choice Redux

After providing an accommodation, the employer should periodically check on whether the accommodation is working, Rennert recommended.

When an employer discovers that an accommodation is not working, it should reconsider the information it had beforehand and continue to engage in the interactive process with the employee. The employer may have narrowed the accommodation choices down to two and selected the one it preferred, which it has the right to do. But there may be no reason to think the other choice will not work, and the employer may need to give it a try.

Often that choice is one that managers are not thrilled with, such as telework or a modified schedule. "Managers' discomfort or dislike of an accommodation is not a valid reason for turning it down," Rennert cautioned.

Encourage managers to try the second option on a trial basis. "If it works for two weeks, there's your accommodation." If it doesn't work, the employer has a documented reason why it's unreasonable to provide that accommodation.

Reasonable accommodation is an ongoing obligation and "is not one-stop shopping," said Peter Petesch, a management attorney with Ford & Harrison.

When accommodations don't work, employers should try something else until the employer and the employee "mutually run out of options," he said. Signs that an employer really has gone far enough, according to Petesch, include:

* A real slowdown in the pipeline of ideas.

* The worker continues to not meet performance expectations.

* The employer is losing patience on performance issues that cannot be ignored.