costs can be relatively high, the prospects for market sharing have seldom been realized, many valuable data sets are unavailable from public sources, and the existence of one complex database seems empirically to constitute a de facto barrier to entry that is seldom overcome. Moreover, as discussed in previous chapters, many databases in the natural sciences contain unique, nonreproducible observations that are by definition available only from a sole source. This lack of effective competition, with its inherent possibilities for discouraging add-on products and for engaging in abuses of market power, was downplayed by the European Council of Ministers in 1995, even though it had been uppermost in the minds of the European Commission's own drafters a short while earlier. Article 16 of the final Directive thus merely calls for 3-year reviews to determine whether existing antitrust laws prove inadequate to deal with the "abuse of a dominant position or other interference with free competition," in which case proposals for "non-voluntary licensing" may once again be considered.¹⁰³

The fear of market failure and of chronic underprotection that initially motivated the quest for a sui generis regime to protect electronic databases has thus given way to the creation of "mini-monopolies over information"¹⁰⁴ and to an underlying logic that is inconsistent with the public interest in the full and open flow of scientific data. The original goal of providing some incentives to augment the publishers' investment in compiling electronic databases has generated a set of norms that could render many scientific and technological undertakings prohibitively expensive. As explained below, the short-term social benefits of this so-called "extraction right" may thus conceal the long-term social costs of diminished research and development capabilities at scientific and educational institutes, including public and semipublic institutions that are already indirectly subsidizing private research and development.¹⁰⁵

When the European Commission began its deliberations concerning database protection, the climate in which intellectual property policy discussions at both the national and the international level took place differed from that prevailing today. The fate of the Uruguay Round of multilateral trade negotiations and its intellectual property component, the Agreement on Trade-Related Aspects of Intellectual Property Rights (TRIPS), remained uncertain. The U.S. intellectual property authorities had not yet begun to survey the issues posed by widespread transmission of digitized information over telecommunications networks. The Supreme Court had just denied copyright protection to telephone directories in Feist and had recently invalidated state protection of subpatentable industrial designs.¹⁰⁶ These decisions proclaimed renewed faith in a 19th-century vision of the competitive ethos without recognizing the unresolved problems of gaining returns from investments in subpatentable information goods under 21st-century