Art can do polemics, but polemics are not what art does best. What art does best is ambivalence, the intersection of opposing, even hostile concepts: between knowledge and creativity, feeling and fact, conscience and impulse, commerce and gift--and, most visibly in the last two or so years, between art and science.

It's rich territory that hints at some divisions between things that may be more perceived than real: for example, objectivity and subjectivity, and truth and anything else. This is something to bear in mind as you walk through Gene(sis): Contemporary Art Explores Human Genomics. The exhibition is more than simply a meet-and-greet between the two fields; it's a kind of extensive and ongoing syncretism that may not be strictly contemporary. You would do best to start at a small exhibition of photographs mounted as part of the Short Stories curatorial series. They date from the 1820s, around the time that art and science were seen to diverge, and the concept of subjectivity emerged. These images--taxonomic, precise, serious--are also highly aesthetic, as if the cleaving of art and science had not been the neat slice philosophers imagined.

Similarly, Gene(sis) asks questions that are hard, complicated, nuanced, and, at times, absurd. And while these questions are concerned with morality and ethics, there is no particular slant. This, curator Robin Held told me, was deliberate, and what emerges is a set of visual theories played out across all media.

One question is that of authorship and responsibility--which I imagine is constantly fought on the frontiers of research--as if artists haven't always been the primary hand behind their own works. But when an artist creates genetic material out of arbitrarily translated code, as Eduardo Kac has done, a Pygmalion-like border is crossed. Catherine Chalmers' large-scale portraits of mice bred for specific diseases give a heightened sense of obligation to the control of actual lives, without resorting to easy emotions. These mice have a larger-than-life dignity; they very nearly dare you to feel sorry for them.

The most rigorous work in Gene(sis) is Chimera Obscura, a game by Shawn Brixey and Richard Rinehart that almost beggars description ("game" hardly scratches the surface). It's a maze, based on a thumbprint, that you access and navigate on the Internet, and as you move through it, you deposit homemade bits of media (such as video, text, audio, called "memes") throughout not only in the latitudes and longitudes of the maze, but up and down a z-axis to an infinite depth. The complication, as if things weren't complicated enough, is a mutating "Minotaur" that wanders through the maze, changing the memes at will. According to Brixey, this work won't get truly interesting for months or even years, but when it does, it will be like one of those Internet fantasy games that takes on subcultures, languages, and inside jokes of its own. In the gallery, you can watch a robotic navigator move about in response to a player somewhere out there in the world--an emphatic division of "hands-on" and "remote," of human agency, with all its frustrating blindness, and technological precision.

Not all of the work rises to Chimera Obscura's level of involvement and inquiry, and some of it is attractive without adding depth, such as Joan Fontcuberta's Hemograms--enlarged photographic images of his family's blood samples. But the best of Gene(sis) recalls and refines Emily Dickinson's dictum that "Nature is a haunted house; art is a house that tries to be haunted," with work that looks at both the essential and the abstract. What could be more so than DNA mapping (as seen by both Iñigo Manglano-Ovalle and Jaq Chartier)? Or traces of biological evidence (photographed as elegant minimalism by Orit Raff, and made into cloning kits by Larry Miller)? Or the rich biological worlds built by Susan Robb out of spit and food and Play-Doh?

The Henry has wisely and thoroughly planned a great deal of programming (lectures, films, performance) around Gene(sis), which means that the usual chatter about "starting a dialogue" isn't just empty gesturing. These are hard questions, without answers. Which doesn't make it any less compelling to try. Which is what art does best.