Since bibliometrics is an inherently retrospective method, it remains an open question whether it is possible to detect trends for the future by eavesdropping on today's scientific communication. This contribution tries to clarify this question by examining historic scientific trends and looking for specific patterns, in order to distinguish between different types of emerging topics. It will be demonstrated that it is indeed possible to distinguish between different kinds of emerging topics by means of bibliometric quantities. Furthermore, it will be discussed whether, and if so how it is possible to establish a scheme for the classification of emerging topics using these quantities. The proposed approach is named "Trend Archaeology".
