The first constitution of the American Federation of Labor, adopted at its founding in 1886, declared the new organization open to the membership of any “seven wage workers of good character, and favorable to Trade Unions, and not members of any body affiliated with this Federation.” Tens of thousands of such groups applied for and received direct affiliation with the national federation–afterward, though sometimes long afterward, typically migrating to one or another international union.

The tactic was particularly prevalent during peak periods of union organization, such as the turn of the twentieth century and again in the 1930s, when workers who did not fit well into their established forms sought to join unions. During these periods another union formation was also widespread: “minority” or “members only” unions, which offered representation to workers without a demonstrated pro-union majority at their worksite. Such nonmajority unions were critical to organizing new sectors of American industry, providing a union presence in the workplace well before an employer recognized a collective-bargaining unit. Most of the early organizing of the industrial trades, for example, and of early industrial unions like the mineworkers and steelworkers, was achieved through such minority unions.

After World War II, however, unions effectively abandoned both “direct affiliation” and “minority unionism” as common practices. Over the past half-century, union membership has come to mean membership in an organization that has demonstrated majority support among workers at a particular worksite, recognized by an employer as the exclusive representative of workers for purposes of collective bargaining. Labor is not as open in its membership, in admitting different configurations of workers, as it was in the past.

We believe this self-imposed limit on the meaning of membership today poses an unnecessary barrier to union influence and growth, and it should be reconsidered. There are tens of millions of nonunion workers–many times the size of the existing labor movement–who want better representation at work or better representation of workers’ interests politically, but who remain cut off from the benefits of union membership. Unions can and should seek to change this by reforming labor law or by increasing their organizing efforts. In addition, however, organized labor should open itself to a wider range of members.

Pro-union workers who do not make up a majority at their workplace are not irrelevant to building a labor movement. They have simply not yet achieved one particular measure of union strength–not even necessarily the most important one. These workers have much to offer labor and much to gain from labor. Today as in the past, nontraditional members in nonmajority settings can give labor an immense boost in its reach, leverage and access to strategic information on employer behavior. Adding nonmajority or otherwise nontraditional workers to union membership need not, moreover, conflict with the goal of traditional majorities-only organizing. To the contrary, such new members would provide natural ballast for the legal and policy reforms and organizing committees that unions need to succeed in such organizing.

Opening up to these new members would entail some administrative challenges. Many unionists will worry about the cost of servicing workers outside union security clauses and regular dues collection by employers. But the economics of the Internet have changed this cost equation in fundamental ways. At essentially zero marginal cost, unions can communicate with an ever-expanding number of new members, and they can deliver all manner of services to them through the Internet.

A labor movement that embraced this vision–taking its own historical lessons with diversified membership seriously and relying more heavily on the Internet in membership communication and servicing–would be practicing what we call “open-source unionism” (OSU).

The case for OSU begins by recognizing that traditional unionism and strategies for advancing it are not succeeding. Seven years after John Sweeney’s “new voices” team took over at the AFL-CIO, only 9 percent of private-sector workers belong to unions–a lower proportion than when he took over, indeed lower than a century ago. Unions look healthier in the public sector, but public-sector unionization has natural boundaries on its importance. Public employment is only 15 percent of total employment, and public-sector wage and work norms cannot be maintained indefinitely at sharp odds with the private economy. To give workers greater say in the American economy, unions must increase their power vis-à-vis private employers. This they have failed to do.

The failure is by no means because workers reject unionism. American unions operate under a labor law that is the least favorable to collective worker action in the developed world. They are pummeled daily by a powerful business community uniquely hostile to unions. And for all their PAC giving and get-out-the-votedrives on behalf of Democrats, labor suffers from a party that gets more excited about fighting for free-trade agreements and the interests of high-tech companies than fighting for worker rights.

Admitting all this, however, tells us little about what labor should do. Should it lobby once again for a labor law reform that Congress failed to deliver when unions had a larger share of the work force? Persuade business that labor can be its friend? Reinvent the Democratic Party? Not likely, at least not anytime soon, and almost certainly not without first growing the membership base that could create movement on these fronts. A declining union movement falls into a vicious downward spiral, as lower density reduces resources and ability to reverse the fall. To break the spiral, unions need more bodies and more broad public support.

It seems very unlikely that unions can achieve the necessary scale and recognition through traditional majorities-based organizing alone. Because of work-force growth and steady churning in the job base, unions must organize hundreds of thousands of workers annually merely to maintain their present private-sector density–far more than they currently do. To increase density a percentage point, they need to organize about 1 million per year. To get back to the position they were in when Ronald Reagan took office, they would need to do that for about twelve years running.

A useful rule of thumb puts the cost of acquiring a new union member at $1,000; some estimates are as high as $2,000-$3,000 per new member. So a million new members would cost at least $1 billion, or about 20 percent of unions’ annual income. It was on this reckoning that Sweeney, upon taking office, challenged AFL-CIO affiliates to dedicate at least that share of their budgets to organizing. But nobody has seen organizing on the scale of millions of workers since the 1930s, and only a handful of unions have come even close to meeting Sweeney’s benchmark. If current trends hold, then, density will continue to decline.

What is needed is a larger transformation in strategy that would change the broader balance of forces in the organizing equation by getting a lot more workers into the labor movement, and spreading labor’s influence more widely in society. Labor needs to open itself up. OSU would accomplish that, while complementing the traditional powers that labor still retains.

To clarify the direction we believe labor should go, let’s contrast the proposed open-source union model more explicitly with the existing one. Under the current model, workers typically become union members only when unions gain majority support at a particular workplace. This makes the union the exclusive representative of those workers for purposes of collective bargaining. Getting to majority status–in the trade, “50 percent + 1”–is a struggle. The law barely punishes employers who violate it, and the success of the union drive is typically determined by the level of employer resistance. Unions usually abandon workers who are unsuccessful in their fight to achieve majority status, and they are uninterested in workers who have no plausible near-term chance of such success.

Under open-source unionism, by contrast, unions would welcome members even before they achieved majority status, and stick with them as they fought for it–maybe for a very long time. These “pre-majority” workers would presumably pay reduced dues in the absence of the benefits of collective bargaining, but would otherwise be normal union members. They would gain some of the bread-and-butter benefits of traditional unionism–advice and support on their legal rights, bargaining over wages and working conditions if feasible, protection of pension holdings, political representation, career guidance, access to training and so on. And even in minority positions, they might gain a collective contract for union members, or grow to the point of being able to force a wall-to-wall agreement for all workers in the unit. But under OSU, such an agreement, which is traditionally the singular goal of organizing, would not be the defining criterion for achieving or losing membership. Joining the labor movement would be something you did for a long time, not just an organizational relationship you entered into with a third party upon taking some particular job, to expire when that job expired or changed.

OSU would engage a range of workers in different states of organization rather than discrete majorities of workers in collective-bargaining agreements. There would be traditional employer-specific unions, but there would likely be more cross-employer professional sorts of union formations and more geographically defined ones. Within any of these boundaries, the goal of OSU would not be collective bargaining per se but broader worker influence over the terms and conditions of work and working life. Because OSU unions would typically have less clout inside firms or with particular employers, they would probably be more concerned than traditional unionism with the political and policy environment surrounding their employers and employment settings. They would be more open to alliance with nonlabor forces–community forces of various kinds, constituencies organized around interests not best expressed through work or even class (here think environmental, feminist, diversity or work/family concerns)–that might support them in this work. As a result, labor as a whole would likely have a more pronounced “social” face with OSU than it has today.

How realistic is this vision? Nobody knows for sure. But there is evidence to suggest that it is feasible–evidence of unmet demand for unionism among workers, evidence of legal support for minority unionism and evidence that the Internet can be a vehicle for low-cost provision of information, communication and work-related services.

The Market

Approximately 100 million private-sector American workers–including 91 percent of the total–have no collective representation at work. Our mid-1990s survey of worker attitudes found that most workers want some organization–ranging from unions to workplace committees of various forms–speaking to their everyday concerns at work: wages and benefits, statutory rights, technology and training, safety, work/family scheduling conflicts, etc. Applying our results to today’s work force, about 42 million workers want an organization with elected representatives and arbitration of disputes with management. Another 42 million or so want an organization more focused on information, career assistance or consultation with management, but still operating independent of management. Together these roughly 85 million workers–a group twelve times the size of present private union membership–are the market for open-source unionism. Capturing even a small share of this market could massively expand the American labor movement and vastly extend its reach.

The Law

Many union and business leaders believe that pro-union workers without a workplace majority have no collective rights–that they exist in a sort of legal black hole devoid of the protections our national labor laws afford concerted activity. That is not the case. In fact, all the basic rights and protections of that law apply to workers acting together in nonmajority situations: protection from discrimination against union activity, the right to strike without being discharged, the right to present demands and request negotiations with management, the right to designate union stewards and the right to bargain and make a collective agreement for members, among others. Not only does minority unionism have historical antecedents in the private sector, but it has strong roots in the public sector, which accounts for an ever larger share of the union movement. Most public-sector unions in the 1960s and ’70s first developed from minority representation. Teachers’ unions, for example, emerged through agreements negotiated only with members of non-majority associations.

To be sure, getting robust minority union presence in the private sector will not be easy, as employers may react with the same opposition to minority unions as to traditional ones. But there is nothing in the law that prevents or even discourages exploring what minority unionism might look like today.

The Technology

A longstanding objection to more open-ended and diverse union membership is that with relatively low density in any given place, the members would be too costly to service: The economics of servicing require a collective-bargaining agreement and the accompanying dues and union security. But here we think the Internet is changing the economics of membership servicing in fundamental ways [see sidebar opposite].

The Internet reduces to near-zero the marginal cost of providing information, advice and some direct services to members. And Net usage in America is approaching 80 percent of households or workplaces. What this means is that unions can be continuously communicating with even a vast membership, at a cost that is basically independent of the number of members. Servicing and coordination of a mass labor movement, drawing on membership more varied and dispersed than present membership, is economically feasible today in a way it was not just a few years ago.

Of course, most workers will want human contact and direct exchange in addition to advice and guidance through the web. These relationships require some shared physical space, which is one reason open-source unionism would have a strong geographic component. But it does not gainsay the degree to which the Net can support alternative organizing, especially from a minority position of strength. The best evidence of this is what workers are already doing along these lines. As the examples in the sidebar indicate, whether job-based, occupation-based, geographically based or international/local-union-based, workers can be mobilized and organized through the Net, which can also connect labor with broader communities at a speed and cost unimaginable even a few years ago.

The Opportunity

If unions were to combine open membership, minority representation and low-cost, Net-based servicing and coordination–perhaps including more “direct affiliation” of new worker organizations to the national AFL-CIO, or regional bodies, or existing internationals–we believe that over the long run they would expand membership substantially. They would also enjoy immediate gains in labor’s public image and political effectiveness.

The AFL-CIO takes great pride in its recent political program, claiming that it has dramatically increased the union household share of the active electorate even as its share of the working population has declined. Upon closer inspection, this claim proves exaggerated, an artifact of exit-poll procedures and inconsistency in question wording. But what is clear from the polls is that the number of nonmembers now in the electorate who express great support for unions is vastly greater than the number of union members who express such support–three to four times greater. A political program centered on labor’s interests, with manifest general benefit, would find an audience among these voters. Especially when coupled with human contact and presence locally–provided, for example, by a well-organized central labor council or state federation–this sort of diffuse political support could greatly affect state and local as well as national elections.

Of course, admitting new sorts of members to its ranks–or better coordinating with outsiders on politics–would disrupt established labor routines. New unions would form, jurisdictional boundaries would be crossed and union alliances with nonunion community and advocacy groups would give rise to a different labor politics disturbing to the status quo. For some within labor, that may be enough reason not to try it.

But the open-source idea is eminently scalable. It can start small. And it can start in part of the movement. Labor, like other progressive organizations, sometimes acts as if it cannot coordinate on anything until it agrees on everything. That is not necessary here. A single state federation, or central labor council, or international could initiate it–anywhere there is a consensus to allow for experimentation.

Some traditionalists in labor may argue that the new workers brought in through OSU will not look like or have the same concerns or organize themselves the same way as “traditional” union members. And they would be right. How could new members from throughout the American economy and society, drawn together largely by different means, be replicas of current members? All great surges in organizing have been preceded by fears that the new members will be different from the old, and confusion about the right form of union–craft versus industrial, general versus narrow jurisdictions, public-sector associations versus “real unions.” What we know from this history is that forms must adjust to workers and the broader economy, and nobody knows in advance which new forms will turn out to be enduring.

Labor currently has more support for its values in American society than it is harnessing and mobilizing, either through its political program or organizing. Workers want a connection to unions far greater than they have now. Present organizing is not keeping pace with economic changes and a nearly lethal employer and policy environment. Turning labor around will require more than simply doing more of what unions have been doing over the past decade. It will require a broader–if also, at least in part, shallower–membership base and stronger alliance between labor and those outside itself. That will not be achieved through rhetoric. It necessitates changes in membership, and the routines for servicing and mobilizing those members. What we need in America today is a labor movement that workers can join easily, without going to war with their employers; a labor movement that welcomes support anywhere it finds it, and is able to crystallize what is now diffuse support into real membership and shared action; and a movement that will offer support anywhere workers are struggling to build power. “Open-source unionism” describes the structure and ambitions of a labor movement that seeks to do these things–“The new union movement, we come from everywhere.” It has a nice ring, doesn’t it?