In May 2001, the Swartz Foundation sponsored a workshop titled "Can a Machine Be Conscious", the fifth in a series of meetings we've sponored at the Banbury Center of Cold Spring Harbor Laboratory. Prior to the last couple of years, studies of consciousness were the sole domain of philosophers. Consciousness was definitely not considered, by most, as a ripe target for serious neurobiological investigation. Even among the initiated, difficulties of all sorts arise when people convene to talk about consciousness. Conundrums are posed and semantic Gordian knots appear. Metalanguage is invented to aid discussion, and distinctions are made between the "hard question" and the apparently not-so-tough questions. So little consensus exists at any level of the problem that some ask "what's the big deal", and others question whether consciousness exists at all. Given these considerations, any meeting on consciousness co us.nsciousness threatens to be a parlous, if not actually perilous, exercise.

Past Swartz Foundation workshops at Banbury have addressed mainline, data-driven issues in neurobiology. They were designed to be small meetings, rooted in data, for intensive exchange on important current topics. In 1998, our first Banbury conference, "Neural Computation", was reportedly of scientific value to the participants and served as an opportunity for the newly-formed Foundation to participate in exploring its chosen field: where physical, computer and brain science converge. The second Banbury meeting, 1999's "Cortex and Thalamus", dealt largely with the organization and function of reciprocal connections that exist in that system. Discussions of how activity in these connections might be involved in supporting consciousness came up as a particularly interesting sidebar.

The Swartz Foundation's charter points to "cutting edge" issues in neurobiology that we want to explore. Indeed, consciousness is specifically mentioned. The "Cortex and Thalamus" meeting, suggested that a workshop on consciousness could be workable if it were organized around some (putative) component feature already being studied by neurobiologists. We thought that this operational approach could be an anodyne to the "hard questions," while still allowing a productive workshop that included consciousness as a topic.

With this in mind, our Banbury meeting in 2000 was called "Experimental Animal Models of Attention and Consciousness". The presentations included data-driven ones, cleaving firmly to the "attention" theme, and other, more speculative, but still neurobiologically-responsible, talks on frameworks and measures for studying consciousness. In this forum, there was serious consideration of how issues of consciousness were starting to be addressed neurobiologically, and what approaches might be successfully used in future.

Moving to the May 2001 Banbury, Christof Koch and I had previously discussed the idea of organizing a workshop along the general theme of Man and Machine. It was now clearer that consciousness could be a manageable subject in the context of a Banbury meeting. What was needed was a "lensing" question to focus the discussion: "Can a Machine Be Conscious?" The organizers' (Christof, Ron Goodman, David Chalmers, Owen Holland and I) invited a mixed group of neurobiologists, computer scientists, engineers and philosophers. The invitation proposed an enfilade of thoughtful questions and ended with the promise that it would be "an enjoyable workshop."

In fact, "Can a Machine Be Conscious", was one of the first interdisciplinary symposia to bring together theoreticians and practitioners; biologists with researchers from "hard" science; engineers and philosophers; to mutually inform and to discuss practical approaches to this difficult and provocative subject. The Swartz Foundation thanks the organizers and the participants. It was, as promised, "an enjoyable workshop."