Basic principles of classification

Abstract

Classification is a systematic ordering of the object of research, in this case, ecosystems at the earth’s surface or, in other words: landscape units as ‘holons’. As for general principles of classification, we can learn a lot from the best-known classification, the taxonomical classification of species. This has functioned as an example for similar classifications of land attributes, such as soil or vegetation.

For a classification of ecosystems, we must select diagnostic characteristics from the large number of ecosystem properties, for which selection guiding principles are an aid. Examples are given of the most important guiding principles and possible diagnostic characteristics are discussed. Also, it is argued that typifications of land attributes are the best diagnostic characteristics for ecosystem classification.

Two different approaches to classification are compared, viz. by agglomeration, which leads to the most pure typification, and by sub-division, which is always connected with mapping. These two approaches are related to two different hierachies. A hierarchy of agglomeration is related to classification by agglomeration with abstract boundaries in a typification, whereas a hierarchy in space is related to classification by sub-division with concrete boundaries in the field. A map’s legend, for example, is essentially a classification by subdivision. However, units can be described by means of the units of a typification.

Finally, it is questioned whether a world embracing typification is worthwhile to strive for. It is concluded that for most applied surveys an ad-hoc classification by sub-division is the most appropriate, but using land attribute typifications for defining the legend units.