There’s trouble in the Amazon Labor Union.

Yesterday, The New York Times’ Noam Scheiber reported that a rift has opened between the union’s president, Chris Smalls, and many of the union’s activists, staff, and other leaders. At issue is Smalls’s emphasis on organizing other Amazon warehouses beyond the one that successfully voted to go union—JFK8, on Staten Island—and the opposition’s emphasis on continuing to maintain workers’ support at JFK8 and building organizational infrastructure and procedures there, to strengthen the union’s hand in bargaining a contract with Amazon.

The opposition has much the better case. None of the other warehouses that Smalls has visited and pushed for a vote at have seen workers prepare the necessary groundwork for a successful unionization vote; Smalls’s cross-country trips smack more of a grand tour than an organizing strategy. Conversely, those who’ve sought to focus on building the capacity at JFK8 to successfully grapple with the company over the bargaining table brought in Jane McAlevey, one of the most accomplished union strategists and organizers, to help them build that support. But Smalls’s continued emphasis on his more far-flung and far-fetched strategy compelled McAlevey to cease her work with the union. Smalls then told staffers and other leaders that if they disapproved of his strategy, they could leave. Worse yet, he insisted that the union not hold an election for its officers among its members until a contract is reached with Amazon, thereby ensuring his continued tenure in office for—well, for the foreseeable future.

Which brings us to the real authors of the union’s troubles—the company and the weakness of labor law that permits it to delay any serious bargaining with the union. The sad fact is that employers routinely and often indefinitely delay bargaining a first contract with workers who’ve voted to go union, in the hope that the delay will cause workers to lose interest in the union and possibly vote to decertify it. A study by Cornell University’s Kate Bronfenbrenner revealed that one year after their workers have unionized, 52 percent of the companies still have not bargained a contract with them, and two years after that vote, 37 percent of those companies have yet to bargain a contract. The various efforts that Democrats have waged to restore labor law to its original potency have included provisions for mandatory mediation and contract enactment if certain time limits are exceeded, but none of those efforts have been able to clear the Senate’s 60-vote hurdle to become law.

How, then, did the massive unions that formed in the 1930s, and grew to represent over one-third of the workforce by mid-century, overcome this hurdle? The short answer is, through direct action that compelled the companies to bargain—more precisely, strikes that shut those companies down. The great strike against General Motors that autoworkers waged for recognition of their union in the first months of 1937—the breakthrough strike for American labor—took the form of occupying, fortifying, and thereby closing down two factories that produced the key components for GM’s cars, most especially Flint, Michigan’s Chevrolet Number 4, which assembled the motors for nearly all GM vehicles. Thousands of workers ringed the plant to keep company thugs from attacking it, and President Roosevelt and Michigan Gov. Frank Murphy—both of whom had received huge labor support in the November 1936 elections—declined to send in the National Guard or Army to break the strike. Given all that, GM quickly agreed on a contract with the UAW.

Today’s Amazon workers, and for that matter, the workers at the more than 200 Starbucks outlets who’ve voted to go union, face a more hazard-strewn landscape. Taking one Amazon warehouse, or even several hundred Starbucks, offline won’t shut down those companies. Alfred Sloan, GM’s president in 1937, was every bit as anti-union as Starbucks’s Howard Schultz is today, but a company with just two dozen factories, which in turn were dependent on the parts from just one factory, was much more susceptible to workers’ desire for a union than a global retailer with hundreds or thousands of facilities is today. Moreover, in the decades since the 1930s, a series of court decisions and more anti-union business practices have so undermined labor law that even a company with just a few workplaces can delay reaching a contract for months or years, as well as simply violate labor law with little or no fear of any meaningful penalty.

There’s one other significant difference between then and now that the current rift within the ALU makes clear. Unions that suddenly spring into existence often have initial leaders with the charisma to have inspired the first group of workers to go union, but have little capacity to actually run a union. That’s exactly what happened in the UAW in its first years. The small, fledgling union was initially led by Homer Martin, a former Baptist minister who could make rousing speeches. When the union suddenly ballooned to include several hundred thousand members, Martin was plainly out of his depth, and was replaced by more competent leaders. Precisely because the union did grow and establish itself—by winning contracts with the Big Three auto companies—an organization and a culture took shape that was capable of selecting appropriate leaders.

Today, Chris Smalls seems a leader in the mold of Martin. But because Amazon has successfully opposed unionization everywhere but JFK8, and has refused to bargain a contract with the union there (once the contract is in place, the union would be required to hold the election that Smalls is currently resisting), it is effectively complicit with Smalls in keeping him atop the union. Such is the state of business practices and labor law that an unlikely pairing of Jeff Bezos and Chris Smalls has now emerged. This is not good news.