There's a secret to anti-trust law, but learning it isn't likely to reassure a high tech investor pondering the implications of the Microsoft verdict. The secret is that the most basic terms used in the field, concepts crucial to great issues of industrial structure, billions of dollars in treble damages, and years of jail time, have no certain meanings. As used by the enforcers, analysts, and judges, their definitions are varying and often contradictory. Take the most important term of them all, "competition." Since the raison d’etre of anti-trust is protection of "competition" and "the competitive system," reasonable people might expect the concept to be well-defined. They would be wrong. Competition can refer to the structure of an industry or market. It can mean something like the economists’ model of perfect competition, a system in which all economic entities are so atomized that no one has power over anything. Or, since this world cannot exist outside of a textbook, competition can mean nothing more than non-monopoly—that there is more than one firm, or perhaps that there are several firms. Or perhaps, somewhat more artfully, that there are "enough" firms to make it difficult for them to act in parallel. Or perhaps only that there are other firms that might enter, or that the firms already in an industry only think that others might enter. On the other hand, "competition" is also used to describe conduct, not structure. It may refer to a system in which rivals try to dish each other. In this context, "anti-competitive" might mean any action that reduces rivalry between any two firms. On the other hand, sometimes it might not mean this at all. Perhaps reducing rivalry between the two firms would help the combination dish other firms, in which case an action that seems anti-competitive is actually pro-competitive. Unless, of course, the combination might overwhelm some third firm, which could reconvert the anti-competitive action that was actually pro-competitive back into anti-competitive. Or it might not, depending on the possible impacts on firms four and five. The companies will find out whether they were pro- or anti- only through a fact-intensive inquiry conducted according to no clear decision rules by battalions of $300 per hour lawyers and $500 per hour economists. The decision will be made by five political appointees called the FTC, or by a federal judge whose training is totally in the abstractions of the legal system, not in business or economics. And do not forget the state attorneys general, whose whims are increasingly accorded solemn deference. If a monopoly is involved, rules change. Promoting competition means not dishing rivals. In fact, the anti-trust laws can be converted into an Americans with Disabilities Act for small business, with large firms forced to provide reasonable accommodation. But the exact nature of the monopoly’s duties to its rivals is never clear. Monopoly exists when a firm has too much market power. And what is "market power"? According to DOJ/FTC guidelines, it is "the ability profitably to maintain prices above, or output below, competitive levels for a significant period of time." And what is a "competitive level" or a "significant period of time"? And under what assumptions about the strategic reactions of other firms, or about the definition of "profitability"? Would you be surprised to learn that these questions have no clear answers? New uncertainties are being added to the old ones. The latest all-purpose justification used by antitrust enforcers eager to expand into high tech is "network effects," the idea that in some cases, such as for a telephone system, the whole is more than the sum of the parts. A useful line of inquiry, but it is already morphing from analytic concept into a buzzword. As Professor Lawrence White recently noted, "[T]he term network industry has become an expansive, all inclusive term that appears to embrace almost any composite good or service embodying complementary components." Reading the Microsoft opinion with these word games in mind is illuminating, and sobering. One realizes that the judge has no solid grasp of the "competition" he is supposed to protect. Sometimes "competition" means structure (of unspecified sort), sometimes it means conduct. Sometimes his objection is to excessive rivalry, sometimes to insufficient rivalry. Monopoly is certainly found, but the explanation of why Microsoft had market power against the massed phalanxes of IBM, Oracle, Sun, a host of other companies and a huge pool of floating and avaricious venture capital is never clear. The most disturbing dimension is the possible implications for other companies. The high tech world is full of firms that cooperate in some ways and compete in others, that act as rivals, customers and suppliers all at once. The term "partial integration" is sometimes used, or the jazzier "co-opetition." Competition is often characterized by successive waves of technology, with the winner at each stage blessed with huge market share. So here is another game. Given the uncertain meanings of the crucial concepts, try to think of successful companies that could consider themselves safe from anti-trust action, should the government choose to pursue them. This game is not an easy one.