When the Federal Trade Commission proposed its near-total ban on non-compete agreements in 2023, the business press responded as though the agency were dismantling something modern capitalism had carefully constructed. Executives warned of innovation theft. Lawyers cited precedent. Economists argued about labor market efficiency. Almost no one pointed out that the argument being rehearsed — that a trained worker's knowledge belongs, in some meaningful sense, to the person who paid for the training — is approximately as old as writing itself.
The non-compete clause is not a corporate invention. It is a permanent feature of how skilled labor markets behave when left to their own psychology.
The Babylonian Precedent
The earliest surviving labor contracts from ancient Mesopotamia — clay tablets dating to roughly 2000 BCE — already contain provisions that would be recognizable to any employment attorney today. Masters who trained apprentices in weaving, metalworking, or scribal arts frequently inscribed obligations requiring the apprentice to remain in service for a fixed term, often five to seven years, and in some cases prohibiting the newly trained worker from practicing the craft independently within the same city district for a period after completion.
The legal rationale encoded in these tablets is strikingly familiar: the master had invested resources in the apprentice's development and was entitled to recoup that investment before the apprentice was free to compete. The word "invest" is anachronistic; the concept is not. Babylonian law understood that skill transfer was an economic transaction, and it structured the aftermath of that transaction accordingly.
What the tablets also reveal, however, is that apprentices violated these agreements with remarkable consistency. Complaint records from the same period document masters pursuing former apprentices who had set up competing workshops across town. The enforcement problem, it turns out, is as old as the contract itself.
The Medieval Workshop as Total Institution
By the height of the European guild system — roughly the twelfth through sixteenth centuries — the legal architecture of labor restriction had become considerably more elaborate. A master craftsman in Florence or Bruges or Cologne did not merely bind an apprentice's labor. He bound the apprentice's entire social existence.
Guild contracts of this era commonly prohibited apprentices from marrying without the master's consent, from leaving the city without permission, from working for any other craftsman in the same trade, and from discussing the methods of the workshop with outsiders. The latter provision is particularly striking: it is, in functional terms, a non-disclosure agreement, and it predates the modern corporate version by several centuries.
The psychological logic was coherent, if brutal. A master glassblower in Venice, for instance, operated in a market where his competitive advantage was almost entirely contained in techniques that could not be patented, only guarded. The Venetian Republic was so determined to protect the island glassblowers of Murano that it extended their families noble status — and simultaneously threatened death to any glassblower who shared trade secrets with a foreign power. The state had simply nationalized the non-compete.
Why Workers Kept Signing Anyway
The historical record of resistance to these agreements is long and well-documented. Apprentices fled. Workers formed coalitions. English common law courts were, by the seventeenth century, already developing the doctrine that overly broad restraints of trade were unenforceable — a legal position that would eventually evolve into modern American case law on the subject.
And yet the contracts kept appearing. Every generation of workers that successfully pushed back against the most extreme restrictions found that the next generation faced a somewhat modified version of the same structure.
The reason, visible across five thousand years of data, is straightforward: the power asymmetry that produces these agreements does not disappear when the agreements are struck down. A newly trained worker almost always needs access to clients, equipment, supply chains, or reputational networks that the master or employer controls. That dependency creates leverage, and leverage produces contracts. The specific legal form of those contracts is a detail. The underlying relationship generates them reliably.
This is what the FTC's proposed ban — subsequently blocked in federal court — failed to reckon with fully. Non-compete clauses are a symptom. The condition is the structural imbalance between those who hold access to markets and those who hold skill but lack that access.
The Modern Iteration
The twentieth-century American version of this story has its own distinctive features. California's long-standing refusal to enforce non-competes is frequently cited as a driver of Silicon Valley's unusual talent mobility and innovation density — the argument being that when engineers can freely move between companies, knowledge circulates faster and new ventures form more readily. The empirical evidence supports this view to a meaningful degree.
But California is the exception that proves the rule. In most American states, non-competes have remained enforceable against workers at nearly every income level — a fact that attracted significant public attention when journalists began documenting their use against sandwich shop employees and hair salon workers, people with no trade secrets to protect and no meaningful leverage to resist signing.
That extension of the mechanism to workers with no specialized knowledge is itself historically familiar. Medieval guilds applied similar restrictions to journeymen whose skills were, by definition, already standardized. The contract was never purely about protecting genuine competitive advantage. It was also about maintaining the structural position of those who controlled access to work.
What Five Millennia Actually Reveal
The persistence of labor restriction contracts across radically different legal systems, economic structures, and cultural contexts suggests something important about the psychology driving them. They are not primarily the product of corporate greed, though greed participates. They are the product of a genuine human anxiety about investment and return — the feeling, experienced by masters and employers across recorded history, that training someone and then watching them walk away is a fundamental injustice.
That feeling is not irrational. It is, however, consistently weaponized beyond any proportionate relationship to the actual investment at stake. The Venetian glassblower's master had a real interest in protecting genuinely rare technique. The fast-food franchise operator binding a teenager to a non-compete does not. The emotional logic is identical; the application has sprawled far beyond its original justification.
Workers have always known this. The tablets, the guild records, the court cases, and the modern litigation all show the same pattern: people sign these agreements under conditions of unequal power, resist them when they can, and fight them in court when the courts are available. The agreements return in modified form. The resistance continues.
Five thousand years of this cycle have not produced a stable equilibrium. They have produced a permanent negotiation — one that will continue regardless of what any particular regulatory body decides, because the psychology that generates it does not require a specific legal framework to operate. It only requires that some people know things that other people paid to teach them.
