In the night of June 20, 1947, President Harry Truman gave a nationwide radio address about the most pressing domestic issue facing the country: the deepening conflict between management and organized labor. Truman’s address publicly explained his veto earlier that day of the Taft-Hartley Act. Taft-Hartley (known formally as the Labor-Management Relations Act) was named after its two Republican co-sponsors: Congressman Fred Hartley of New Jersey, and Robert Taft, the powerful senator from Ohio and possible 1948 presidential candidate. The law was the consummation of a decade-long effort by Republican politicians, aided by Southern Democrats, business executives, and conservative writers to limit the power of the increasingly aggressive labor movement. The bill easily passed both houses of Congress, as an almost unanimous phalanx of Republicans and conservative Southern Democrats crushed the opposition of liberal Northern Democrats and a couple of maverick Republicans. Truman had made little effort to impede it.

Truman generally supported unions. But he also hated the idea that strong unions could disrupt the American economy and defy his wishes. An enraged Truman had recently threatened striking railway workers with conscription. He despised John L. Lewis, the bellicose, eloquent, and supremely powerful leader of the United Mine Workers of America. As the President wrote in a 1949 letter, he “wouldn’t appoint John L. Lewis dogcatcher.”

Yet Truman and his key advisers like Clark Clifford realized that they would need labor to win what was expected to be a tough election campaign the next year. Thus, the President gave a powerful speech defending unions and attacking the bill. Truman told the nation that Taft-Hartley was “a shocking piece of legislation” that “is unfair to the working people of this country. It clearly abuses the right, which millions of our citizens now enjoy, to join together and bargain with their employers for fair wages and fair working conditions. Under no circumstances could I have signed this bill.”

Senator Robert Wagner, the chief sponsor of the 1935 National Labor Relations Act (NLRA), was ill at the time, but he vowed that he would risk his life to cast a vote to sustain the veto if needed. It was not. The Senate easily overrode Truman’s veto, by a vote of 68-25. The next day, The New York Times front page featured a banner headline of the law’s passage, a large picture of Taft and Hartley, and no less than six articles about labor issues and strikes. Five days later, Lloyd Taft, one of Robert Taft’s four sons, got married in Cincinnati. Union pickets surrounded the church. One of the placards held aloft, addressing the groom, read: “CONGRATULATIONS TO YOU. _________________ TO YOUR OLD MAN”—the blank line conveying labor’s graphic wishes for the elder Taft.

That the Taft-Hartley Act generated such anger was not at all surprising. Unions and their allies called the act a “slave labor bill” or, in Lewis’s view, even a symptom of incipient fascism. Battles between labor and management roiled the country from coast to coast. The “labor question” generated perhaps more controversy than the debate over Obamacare has today. Republicans will have to seek to overturn Obamacare for another 20 years or so before they match labor’s determination to undo key portions of Taft-Hartley; and yet, even with its deep penetration in the workforce at the time, even at the apex of LBJ’s enormous Democratic congressional majority, labor could not repeal a comma of it.

With the long decline of the labor movement has come a parallel decline in our historical memory of its once-extraordinary influence, and of the effort to curtail that influence. Books about Truman give only passing mention to the most contentious law passed during his presidency. Taft, the son of a President and a man who might have become President himself, is barely remembered. And it is unimaginable today that a President would give a national address vociferously defending labor unions.

Ancient history? Maybe—but it’s also crucial history whose direct consequences labor and the country live with today. Taft-Hartley didn’t destroy labor. But it stopped labor dead in its tracks at a point when unions were large, growing, and confident of their economic and political power; when unions really were what The Wall Street Journal still laughably calls “big labor.” The law codified a series of legal land mines—some of which didn’t detonate for decades—that forced unions to weigh the political and economic costs of doing anything too aggressive in their efforts to grow, and, indeed, to begin fighting many rearguard actions to protect the gains they had already made. Without Taft-Hartley, it’s easy to imagine a continued increase in union density rather than a flattening followed by a gradual and then dramatic decline. Today, only 11.3 percent of American workers are unionized—and just 6.6 percent of the private sector, a level not seen since the early twentieth century.

What is the relevance of all this to today? Well, Taft-Hartley isn’t going anywhere. Its land mines still detonate. And it still defines the legal and political context in which labor must operate as it tries to map out a strategy for the future. An aggressive organizing strategy, of the sort labor attempted when John Sweeney took the helm of the AFL-CIO, just doesn’t work because the smart union strategists can’t compensate for a mostly (though not entirely) uninterested working class. But labor can, without undertaking lengthy and expensive campaigns to organize new sectors, work to buttress the areas in which it is already strong, extend its alliances with other progressive groups, and even train the worker leaders of tomorrow. I call this “Fortress Unionism,” and I believe it’s labor’s best play until the day arrives, if it ever does, when the workers themselves militantly signal that they want unions.

The Power of Postwar Unions

It is difficult for a reader today to grasp how big a deal the labor movement was in postwar America—how much people, in support or opposition to unions, deeply cared about them. To understand this passion, we need to recall how rapidly unions had grown, and how often they undertook strikes. Contrasting the diminished strength of today’s labor movement against the stunning power of its postwar antecedent is thus necessary (and fascinating) so that we can also understand why, and in what precise ways, the most powerful sectors of the American polity sought to cut unions down to size.

In just the decade after the NLRA passed in 1935, union membership quadrupled from almost 3.6 million to more than 14.3 million workers. During this period, American labor dominated the daily life of much of the nation and drew the obsessive concern of politicians and the press. Even some Southern states had union membership percentages in the high teens—statewide numbers that would be among the highest in the nation today but were among the smallest then. In a six-month period in 1937 alone—the year of the great sit-down strike at General Motors (GM) in Flint, Michigan—the CIO signed up two million workers in a nation with a population of about 130 million.

After a brief lull at the start of the 1940s, union membership rose meteorically during World War II. A month after Pearl Harbor, union leaders met with President Roosevelt and agreed to a no-strike pledge; in return, the unions received compulsory government arbitration of disputes. As the war dragged on, workers (if not always with their union’s consent) violated the no-strike pledge, but the government’s all-powerful War Labor Board squeezed employers to quickly approve union organizing and contract demands. Thus union membership grew even larger.

At war’s end, nonunion and union workers feared that the end of war-driven production that had halted the Depression would result in a return to high unemployment. Workers also worried about increased inflation as Truman considered ending wartime wage and price controls. With the loss of constant overtime, take-home pay in several sectors declined by 30 percent. Economic observers—and, according to polls, the American people—believed that a postwar America might return to Depression conditions.

The American labor movement responded to these uncertainties with the greatest strike wave in the history of the United States. It started almost immediately after the war and continued right through 1946. Clerical workers walked out at the citadel of capitalism, the New York Stock Exchange. About 68,000 textile workers struck in the East, while 35,000 oil-refinery workers struck across seven states. In the Northwest, 40,000 lumber workers struck. In the Midwest, it was 70,000 Teamsters. In Oakland, a strike that started at the loading docks of two downtown department stores spread quickly to include 100,000 workers who effectively halted the city’s commerce and services for two days—the most dramatic general strike of several during the period.

At the end of 1945 and into 1946, 225,000 autoworkers struck GM—this was the great 113-day strike the brilliant leader of the United Auto Workers (UAW), Walter Reuther, used as an effort to bring European style co-determination to American labor-management relations. In spring 1946, 350,000 of Lewis’s mineworkers struck. Some 750,000 steelworkers struck for 25 days, the most workers ever engaged in a single work action. At one point, 1.6 to 1.8 million workers were on strike simultaneously.

All in all, about 10 percent of the entire American workforce withheld their labor in 1946. There were about 5,000 separate work stoppages involving about 4.6 million workers. This is over six times the number of workers involved in work stoppages over the entire eight-year period from 2005 to 2012, according to figures from the Federal Mediation and Conciliation Service, this at a time when the country had about one-third of the current nonagricultural workforce. To imagine this kind of union militancy today is to imagine 14 million workers striking in a single year.

Republicans ran hard against the strike wave during the 1946 midterm elections under the slogan “Had enough?” The GOP won substantial majorities in both houses and, with the support of Democratic conservatives in the South, moved to pass a new labor restriction bill. While Truman suggested mild changes in labor law, and proposed a labor-management commission, the leaders of America’s unions unconditionally opposed any and all limitations. This proved to be a strategic mistake, as the midterm election demonstrated that, even as the strike wave proved labor’s power (or perhaps because the strike wave proved labor’s power), it did not command sufficient support for maintaining the status quo. Leaders in the news media (including the publishers of almost every major newspaper), the business community, the Federal Reserve Board, the Republican Party, small businesses, and large corporations alike shared a determination to roll back union advances.

Thus the paradox that at the high-water mark of its power and size, the labor movement generated an even more powerful backlash from the nation’s power elite, which was augmented by an obsessive determination from the white South to make its region as union-free as possible. Southern elites, led by their nearly unified bloc in both houses of Congress, feared an ongoing alliance between labor and the first signs of a sustained African-American civil-rights movement, fueled by the return from the war of African-American soldiers newly emboldened to seek justice. The CIO, observing the same phenomenon, hopefully launched “Operation Dixie” in 1946, a well-funded effort to organize throughout 12 Southern states. Southern elites ruthlessly race-baited, red-baited, and intimidated poor black and white workers. Operation Dixie failed dismally, only making the Southern bloc more determined to stop unions in the region. Ira Katznelson, the great historian and political scientist, has called this implacable opposition of the South to African-American rights and the labor movement “the Southern cage.”

The Effects of Taft-Hartley

Congress held several months of hearings on the Taft-Hartley bill that were dominated by a cast of business leaders. Though the reputation of America’s executives had been battered by the Depression, they regained much of their standing by contributing to the war effort, and pleaded effectively for an end to constant commercial disruptions. Taft shrewdly played good cop to Hartley’s bad cop. The Hartley bill that came out of the House was more onerous than Taft’s Senate version. Hartley’s bill would have banned industry-wide bargaining. That would have immediately affected the national contracts that large unions like the Auto Workers and Steelworkers sought from the biggest companies in their sectors. Hartley’s bill would also have banned the union shop nationally (rather than leave it to the states, as the enacted legislation ultimately did). By modifying the bill somewhat, Taft protected it from a coalition of a small number of Republican iconoclasts joining with Democratic liberals who tried but failed to sustain Truman’s veto.

Many scholars have noted that, for the most part, Taft-Hartley did not do the frightening things that unions feared, thus disappointing its business supporters. The law merely codified case and administrative law precedents rather than introduce dramatic new transformations. Union membership dipped a tiny bit, but stayed close to its postwar high for another 20 years or so. These analyses capture part of the truth, but miss the larger point about the bill’s significance. The NLRA had been passed with the explicit goal of increasing unionization and with the understanding that unions were the democratic institutions of working-class self-organization. Taft-Hartley conceived of unions much differently, as merely a large special interest, dangerous to prosperity and stability if left unchecked. Taft-Hartley thus stands as a benchmark for labor’s inability to gain the respectful acceptance of the nation’s cultural, economic, and political elite. The bill cordoned off an aggressive, confident labor movement and enabled its subsequent decimation. To paraphrase the British historian Eric Hobsbawm’s famous 1978 line about the British labor movement, Taft-Hartley (and the almost simultaneous failure of Operation Dixie) halted the forward march of labor in America.

There are many provisions of Taft-Hartley, and the law’s scholars have argued about what the most significant ones are. I list below those that, in my view, had the most far-reaching impact:

Right to work: Probably the most well known of the bill’s clauses today, and the one that labor sought for decades to repeal. Taft resisted Hartley’s effort to eliminate union security clauses nationally. But the law permitted individual states to pass “right to work” laws. These gave workers the right not to pay union dues or an equivalent fee for union services and representation. Within five years of the law’s passage, most of the former Confederacy had passed such laws. In the states where such laws exist, they force unions to spend a good deal of administrative time pressing workers to pay dues voluntarily, even though workers know that they will receive the benefits of a union contract whether or not they pay. “Right to work” made it that much more difficult for unions to organize throughout the South, and is now legal in 24 states.

Employer’s right of free speech: Initially, the NLRA sought to make union organizing a “neutral” process, whereby workers could choose, without employer coercion, whether to have a collective bargaining agent (the union) or not. Employers kept pressing for free-speech rights, and Taft-Hartley finally codified those rights, as long they did not amount to a “threat of reprisal” or a “promise of benefit.” This undermined unionism as an independent self-executing institution of workers and made it subject to the objections of the employer. The right of employer free speech ensured that the party that signs employees’ checks—the employer—would have the most influence. From this spawned a billion-dollar a year industry of “union avoidance” consultants and law firms.

The banning of secondary boycotts: Secondary boycotts were an especially effective tactic by which workers at a company not directly involved in a labor dispute could refuse to handle goods from another firm embroiled in a union fight. The Teamsters had used this strategy to great effect in organizing much of the trucking industry. There are some loopholes in the restriction that allow unions to work around it, but Taft-Hartley caused a vast decline in its use.

The exclusion of foremen and supervisors from labor law: This is one of the least remembered provisions today, but one of the most consequential. At the time of the bill’s passage, the Foreman’s Association of America was growing rapidly and had aligned itself with the CIO. Theoretically, front-line supervisors who did not set company policy could join forces with the unionized workers they ostensibly bossed. Employers viewed this practice, as historian Nelson Lichtenstein notes, as nothing short of “industrial anarchy.” The banishment of the supervisors’ organization flipped these midlevel employees back to the side of the employer and ensured that, in subsequent decades, millions of employees in emerging white-collar segments of the economy would be statutorily denied the right to unionize.

The non-Communist affidavit: Today, we would call this provision the creation of a wedge issue designed to provoke intra-union conflict. Taft-Hartley required officers of unions to sign an affidavit asserting that they were not Communists, either by affiliation or belief. If the officers did not comply, their unions would lose the right of standing before the National Labor Relations Board (NLRB). At the time of the bill’s passage, perhaps one million workers were affiliated with unions controlled by or sympathetic to the Communist Party. Communist trade unionists were, indeed, followers of the Moscow line, but were also cutting-edge supporters of union militancy and civil rights. Many union leaders, including non-Communists, saw the affidavits as odious and opposed them (in 1965, the Supreme Court held them to be unconstitutional). But others, like Reuther, who was fighting a strong Communist faction within the UAW, avidly pressed for compliance. The more conservative unions of the AFL supported the affidavit clause, as it pressured and embarrassed its rival, the more radical CIO, with which almost all the Communist unions were affiliated. Thus the affidavit, as Lichtenstein put it, “depoliticized” labor, “curbed inter-union solidarity,” and was a factor that led to the expulsion of 11 Communist-influenced unions from the CIO in 1950.

The passage of Taft-Hartley encapsulated the geographical, numerical, and ideological boundaries of American unionism. Labor protested the bill and not only at the wedding of Lloyd Taft. More than 100,000 marched in protest in New York City. The UAW and other unions engineered a massive work stoppage and rally in Detroit, each drawing hundreds of thousands. Still, union leaders such as Phillip Murray, the president of the CIO, chose not to attempt more militant and long-running protests; there would be no repeat of the 1945-46 strike wave in opposition to Taft-Hartley. Union workers were strong, and could strike by the millions. But they were not strong enough to repeal a federal law supported by most of the country’s most powerful people.

Business leaders and their political allies obviously feared the nationally famous union “bosses” like Lewis and Reuther who, with a single command, might order hundreds of thousands of workers to paralyze a major industry. Yet they perhaps feared one possibility even more: that those same workers might be unrestrained even by their leadership and might create civil unrest themselves. The obstacles that Taft-Hartley put before bosses and rank and file alike were simple in the way that pouring honey over a butterfly is simple. Taft-Hartley bureaucratized labor unions. Unions required more and more lawyers—and more and more union stewards adept at labor law—to untangle the welter of laws, board decisions, judicial decisions, and contractual obligations that now ensnared the modern labor organization. This pervasive legalistic framework made the labor titans increasingly cautious, and it drained the energy and creativity out of the members and their rank-and-file leadership—the idea was to wait for the lawyers to tell them what would fly before the NLRB or the courts. A Lewis who could close the coal mines and defy the President and the courts was a great danger. Thousands of Oakland bus drivers, dockworkers, and retail clerks taking over the city were a less likely but even greater danger. Taft-Hartley unmistakably signaled that anti-union political, economic, and cultural elites could contain both the leaders and their ranks. The bill shaved a risk-taking edge off labor that, perhaps, it didn’t realize it needed until subsequent moments of institutional crisis, like Ronald Reagan’s firing of striking air-traffic controllers in 1981.

The Failure of Comprehensive Campaigns

So, what is to be done? What can be done? Recently, the distinguished historian of labor Melvyn Dubofsky wrote a disquieting essay, “Does Organized Labor Have a Future?” for the journal Logos. Dubofsky—who thinks that the baneful influence of Taft-Hartley on labor has been overplayed—walks the reader through all the reasons, both external and self-inflicted, that American unions have declined: the containment, which Taft-Hartley abetted, of the movement to the Northeast, Midwest, and West Coast; globalization; automation; the failure to embrace women and workers of color more quickly; labor racketeering, and, just as important, government and corporate accusations of labor racketeering; the vehement opposition of employers and the implicit and explicit support for that opposition from the government—for example, the firing of the air-traffic controllers.

In answer to the question posed by his title, Dubofsky concludes: probably not. He writes, “Given the current alignment of forces domestically and globally, I find it hard to conceive of any tactics or broader strategy through which the labor movement might re-establish its former size, place, and power.”

I think Dubofsky is right about labor’s crisis today (if, in other work, too dismissive of the impact of Taft-Hartley), but, lest his remark sound fatalistic, it’s important to understand exactly what he means. Dubofsky is not saying the labor movement is doomed. Rather, he is saying that its current institutional expression—today’s existing unions—cannot, via a creative conceptual breakthrough (“tactics or broader strategy”), engender a vast growth in union strength comparable to its former peak. In short, “organized labor” can no longer create a space for workers to join their organizations by the millions—but the workers themselves might.

Before getting to what could someday succeed, we have to understand what hasn’t succeeded. And we have to define what success would mean. I use the word in the sense that, I think, most of the smartest union members, union staff, and union officers use it. Success would be to organize enough workers so that union density increases significantly, or at least keeps pace with workforce growth.

The decades-long decline in density, despite gargantuan efforts by many in labor, is the surest sign that Dubofsky is right. For the past 30 years, and with even greater intensity since 1995, when SEIU President John Sweeney took over the AFL-CIO pledging internal reform that would spur aggressive growth, unions have created ingenious “comprehensive campaigns” to organize new members and defend existing units. These campaigns sidestepped the legal limitations Taft-Hartley had put on unions by engaging corporate issues not directly pertaining to worker’s wages, benefits, and working conditions. They analyzed every facet of a company’s business strategy—not merely its labor-relations policies, but also its relationships to key institutional shareholders, the litigation and regulatory proceedings the firm faced, its plans to expand or dispose of key subsidiaries—in an effort to focus public and shareholder attention on the entirety of its business model and record of corporate citizenship.

Some of these campaigns led to significant victories, resulting in the organizing of thousands of workers. The SEIU’s brilliantly conceived “Justice for Janitors” (J4J) campaign, which launched in Pittsburgh and Denver 27 years ago and spread to a couple dozen cities around the country, used “street” tactics, such as civil disobedience and sophisticated challenges to zoning ordinances, to make commercial real-estate owners accountable for their janitorial staffs. As a result, SEIU’s urban janitorial workers have increased their standard of living and bargaining power. A particularly noteworthy victory in 2006 was the organizing of 5,000 janitors in Houston, one of the most important union victories in the South in decades.

Yet even the J4J model has proven unable to stop labor’s slide. Despite Sweeney’s 1995 pledge, private-sector unionism has shrunk 40 percent since then. It has continued to decline under his dynamic successor, Richard Trumka. In 2005, several large, aggressive organizing unions, including the SEIU and the Teamsters, grew impatient with labor’s continued deterioration under the Sweeney administration and left the AFL-CIO to form an alternative federation, Change to Win. The breakaway federation promised to put together organizing campaigns targeting 500,000 workers in its first year. Despite the hard work of its staff (which included me from 2005 to 2011), it has never come remotely close to achieving that goal, and overall union membership continues its downward trend.

The Difficulty of Organizing Today

The sheer logistics of organizing workers in major economic sectors today is vastly different than it was 75 years ago. The GM plants at Flint, Michigan, site of the famous sit-down strike that riveted the nation, employed 47,000 workers in 1937—about as many workers as the number of janitors that J4J has organized in its 23 years. That was about 10 percent of all of the autoworkers in the nation then. Ford’s River Rouge plant in Dearborn was even larger: It employed 60,000 workers in peacetime, 80,000 during the war. Today? Just 6,000.

So organizing GM in Flint alone meant that the UAW gained a substantial foothold in the most important industry of mid-twentieth-century America. No city’s cohort of janitors or any other category of workers can have an impact remotely like that. Consider Wal-Mart, today’s socioeconomic parallel to the great car and steel companies of that time, the quintessential corporation of our historical moment. It is said that we have a “Wal-Mart economy” today in the same way we had a “Fordist economy” at midcentury. Wal-Mart employs about 1.4 million in the United States, about 1 percent of the entire workforce, a bit lower than the ratio that the auto workers represented in the U.S. workforce of 1940. The difference is that Wal-Mart has more than 4,200 stores in America today, and GM and Ford together had perhaps 160 auto plants in 1940. The auto plants of that day averaged perhaps 2,500 workers each. Steel plants were similarly large; Wal-Mart stores average about 300 workers. Given that Wal-Mart’s anti-union animus is as fierce as that of the great carmakers during the Depression, it would be as difficult today to organize a single store of 300 as it was then to organize a giant auto plant. The recent courageous activism of several hundred workers at Wal-Mart stores around the country only underscores the overwhelming challenge of organizing the entire company.

The even larger problem, however, is the theory implicit behind the comprehensive campaign model. Even though unions naturally want to win these campaigns as quickly as possible, they undertake these fights assuming that any one of them could go on for years, or, like J4J, remain “on line” for decades as an incremental expediter of union growth.

The trouble with this kind of thinking is not only that the opportunity costs increase greatly the longer the campaign lasts (Andy Stern, the former president of SEIU, has observed that it can cost up to $3,000 to organize just one worker in the private sector). More important is this: No successful labor movement anywhere in the Western world has ever significantly expanded its membership in this kind of slow but steady fashion. Harvard labor economist Richard Freeman demonstrated this in his important 1998 essay, “Spurts in Union Growth: Defining Moments and Social Processes.” Since the late nineteenth century, private-sector union growth has occurred only in large “spurts” in almost every advanced Western country during roughly the same historical periods (he identifies five, the most recent ending in 1945). During all other periods, union membership erodes. And Freeman further demonstrates that the process happens from the “bottom up,” not the “top down.” The National Labor Relations Act didn’t trigger worker militancy. Worker militancy triggered the passage of the act.

So to carefully plot a campaign in a union office in Washington whose goal is to organize 10,000 to 20,000 workers—even 50,000 workers—in two to four years doesn’t result in meaningful union growth. That’s too small a number of workers within a workforce of more than 140 million people to generate such an increase, unless the number is replicated dozens of times over in several different industries and sectors. But this is economically and logistically impossible for unions to do; organizing the workers (if the union ever does) takes too much time, and it costs too much in money and staff resources to try to do so over that long period of time.

I’m not saying that the era of the comprehensive campaign was wasted. Unions are, and were, desperate to stem decline, and the campaigning strategy responded creatively to the problem, with some substantial victories to its credit. Comprehensive campaigns were worth a try. But we’re 30 years into them now, and they haven’t worked on a scale sufficient to reverse the trend. Unions have undertaken a natural experiment in whether large multiyear, comprehensive campaigns can significantly increase union density. They can’t. Labor should pocket whatever successes the campaign era brought and move on.

Waiting for the Next Mass Movement

So what is to be done? I propose what I call “Fortress Unionism.” (I am speaking here only about private-sector unions, which face a deep crisis. Public-sector unions have their own well-known dilemmas, but require a completely different discussion.) Fortress Unionism would buttress the remaining strengths of labor. The fortress would remain open; labor’s effort to build coalitions with other progressive forces should continue. Unions, however, should not undertake long, expensive comprehensive campaigns outside their core areas of strength. Today, less would be more. In sum:

Defend the remaining high-density regions, sectors, and companies. They include, respectively: Las Vegas, Los Angeles, the Bay Area, Seattle, and New York City; the auto industry, large supermarket chains, several hospital chains, building services in major cities, and convention-sized hotels in major cities; UPS (Teamsters), and the telecom companies (Communications Workers). Strong labor movements in metropolitan areas are especially important to sustain, as they are labor-liberal bulwarks of economic and political strength. The labor movement has been particularly effective in jointly mobilizing with Latinos in Las Vegas and Los Angeles. There is no contradiction between organizing around class issues and so-called “identity politics.” It was called something else then, but identity politics as part of union organizing has been around since the first German-American and Irish-American workers unionized in antebellum New York and Philadelphia.

Strengthen existing union locals. Many local unions have atrophied. Staff and a cohort of committed members often run local unions on behalf of a large silent majority of members, who view union membership as something like an insurance policy, paid for by their union dues, rather than a rank-and-file driven activist organization. Train more workers and hire more staff to enforce contracts and teach workers their rights. Invest heavily in worker education programs, everything from knowledge about occupational safety and health to labor history courses. Workers who feel connected and engaged with their local union will someday help organize new members.

Ask one key question about organizing drives: Will they increase the density or power of existing strongholds? Try, for example, to organize remaining nonunion casinos in the labor powerhouse of Las Vegas. (The Culinary Workers Union in Vegas is, arguably, the strongest, most militant local union in the country.) Continue “bargain to organize” efforts, in which unions gain new organizing rights as a condition of collective bargaining agreements for current members. But for the time being, do not try to organize, via multiyear campaigns, currently nonunion or de minimis union sectors.

Sustain coalition work with other progressive organizations. Post-New Left egalitarians fill top leadership positions across the labor movement and are pushing the movement beyond the white-male iconography of the Taft-Hartley era. They have urged unions, with increasing success, to reach out to environmentalists, community organizations, immigration reformers, racial justice advocates, feminists, gay rights activists, and political reformers to pursue policy changes like limiting the filibuster and protecting voting rights. Unions should make the most of these alliances. They expose unions to creative thinking from outside of organized labor and put union money and staff to use behind important projects. Unions, whenever appropriate, should yield control to other organizations and advocates, and play a supporting and facilitative role. Labor’s time in the spotlight is during those great upsurges of high growth. This is not such a time.

Invest heavily in alt-labor organizations, especially Working America. Alt-labor is the name given to efforts to organize disparate workers outside the conventional one-union to one-workplace structure. The AFL-CIO’s 3.2 million-member Working America, led by legendary “9 to 5” organizer Karen Nussbaum, is the largest and best funded of these efforts. The logic of alt-labor is to find the potential leaders of tomorrow’s mass union organizing and organize them today around discrete, achievable demands. It’s exactly the right idea. As AFL-CIO president Trumka said in The Nation recently, “We hope that we will have the seed planted for people to understand the importance of collective action.” Seed away.

And then…wait. Wait for the workers to say they’ve had enough. When they demand in vast numbers collective solutions to their problems, seize upon that energy and institutionalize it.

That is how massive union growth occurs—workers take matters into their own hands and then unions capture that energy like lightning in a bottle. The workers risk their jobs, and sometimes even their lives, to form a union. It has happened this way all over the world. The workers will signal—loudly—when they want to organize.

In short, union growth occurs when working-class activism overwhelms the quotidian strictures of civil society, forcing political and economic elites to accept unionization as the price of civil peace. During episodes of massive union growth, the workers don’t confine themselves to the careful strategies of union staff—they disregard them, and force the union to play catch up. Conflict spreads quickly from worksite to worksite. If the Wal-Mart demonstrations in November 2012 had followed the pattern of the great railroad strike of 1877 or the 1894 Pullman strike or the 1934 Toledo Auto-Lite strike or any other signature struggle of American labor history, it would have sparked unrest in stores all over the country. Win or lose, workers would have risked their jobs, maybe their physical well-being, and challenged private and public authority. In the summer of 1937, 18 workers died and dozens more were wounded in union-organizing-related violence, shot and beaten by company-hired private security and local police forces.

As the San Francisco Chronicle editorialized in opposition to the 1946 Oakland general strike, sustained worker activism disrupts “the orderly process of daily life.” This may sound melodramatic, but there is no substitute for it. And when the workers do signal, the existing unions and their memberships should stand ready to help. Unions were invented at the same time as modern capitalism. The system generates problems for employees that only collective representation (or the threat of it) can mitigate. An ostensibly democratic capitalism without unions is barely more thinkable than it would be without capitalists. The workers are willful when they want unions. Keep your eye on them. The unions will follow.

Correction
This article originally stated the Justice for Janitors campaign started in Los Angeles 23 years ago. It started 27 years ago in Pittsburgh and Denver. We regret the error.