SUPERNATURAL IN HONG KONG 147 practices do not work. We will never understand the possible interrelations between belief, practice, and efficacy if we do not ask. A stance of rigid rejection of the natural sciences will fail to discover the fascinating possible links between culture and biology that medical scientists are beginning to explore In the case of The Chinese University ghosts, even though students disagree as to whether the ghosts are supernatural, I can analytically say they are supernatural because they involve spirits, an order of existence beyond the visible observable universe, and appear to transcend the laws of nature. Using the concept of supernatural in cases of spiritual agency allows a cultural analysis to become the primary, not secondary, issue. It is precisely because the ghost stories analysed above are'supernatural'(by my definition, and regardless of how they are defined by students who tell them) that it is appropriate to study them as cultural products and to interpret them, rather than to study them as actual physical manifestations(see, on the other hand, Emmons 1982, who uses Hong Kong ghost stories to identify real aspects of their manifestations aside from their cultural differences). I do not accept the occultist view that the stories are literally true(see Daoyijushi 200la, b), but one cannot prove ghosts do not exist. In a sense, because they are believed, they are true and become real, but in a different sense from that imagined by the believers. As Lohmann(this issue puts it, they are empirical but imaginary. Not everything people believe is true is physically true; in some cases, by 'true, people do not necessarily mean physically true. I have analysed the stories as natural phenomena, the natural product of social and psychological tensions. If I believed there were anything natural to the ghosts themselves, I would be obliged to go beyond just a symbolic terpretation. Ghosts may be physically real or natural; I cannot prove they are not In saying they are supernatural, however, I am taking a stand of scepticism and saying I will only investigate the social and cultural levels The concept of supernatural is, as Klass notes, a boundary that has been drawn in the West. It is not arbitrary, however; it is based on current scientific understandings. The boundary between natural and supernatural may shift, but it is still an important and useful boundary Conclusion All concepts are culturally defined. Though the category of supernatural is not always a native concept, and though the boundary between what we understand as natural and supernatural has shifted, and is still shifting with research on the placebo effect and on traditional healing practices, the concept is a useful research tool for certain questions. It marks a type of phenomenon that involves spirit or magical agency. Without such a concept, we are forced to accept th existence of spirits and phenomena transcending the laws of nature. Yet, we must not make the boundary too sharp, as many supernatural phenomena haveSUPERNATURAL IN HONG KONG 147 practices do not work. We will never understand the possible interrelations between belief, practice, and efficacy if we do not ask. A stance of rigid rejection of the natural sciences will fail to discover the fascinating possible links between culture and biology that medical scientists are beginning to explore. In the case of The Chinese University ghosts, even though students disagree as to whether the ghosts are supernatural, I can analytically say they are supernatural because they involve spirits, ‘an order of existence beyond the visible observable universe’, and appear to transcend the laws of nature. Using the concept of supernatural in cases of spiritual agency allows a cultural analysis to become the primary, not secondary, issue. It is precisely because the ghost stories analysed above are ‘supernatural’ (by my definition, and regardless of how they are defined by students who tell them) that it is appropriate to study them as cultural products and to interpret them, rather than to study them as actual physical manifestations (see, on the other hand, Emmons 1982, who uses Hong Kong ghost stories to identify real aspects of their manifestations aside from their cultural differences). I do not accept the occultist view that the stories are literally true (see Daoyijushi 2001a, b), but one cannot prove ghosts do not exist. In a sense, because they are believed, they are true and become real, but in a different sense from that imagined by the believers. As Lohmann (this issue) puts it, they are empirical but imaginary. Not everything people believe is true is physically true; in some cases, by ‘true’, people do not necessarily mean physically true. I have analysed the stories as natural phenomena, the natural product of social and psychological tensions. If I believed there were anything natural to the ghosts themselves, I would be obliged to go beyond just a symbolic interpretation. Ghosts may be physically real or natural; I cannot prove they are not. In saying they are supernatural, however, I am taking a stand of scepticism and saying I will only investigate the social and cultural levels. The concept of supernatural is, as Klass notes, a boundary that has been drawn in the West. It is not arbitrary, however; it is based on current scientific understandings. The boundary between natural and supernatural may shift, but it is still an important and useful boundary. Conclusion All concepts are culturally defined. Though the category of supernatural is not always a native concept, and though the boundary between what we understand as natural and supernatural has shifted, and is still shifting with research on the placebo effect and on traditional healing practices, the concept is a useful research tool for certain questions. It marks a type of phenomenon that involves spirit or magical agency. Without such a concept, we are forced to accept the existence of spirits and phenomena transcending the laws of nature. Yet, we must not make the boundary too sharp, as many supernatural phenomena have