Abstract
This monograph is an attempt to describe basic quantitative regularities of context. The main hypothesis is that all relevant features of a unit are reflected by its textual context.
For the purpose of quantitative description this approach adopts three main measures: variability of context (ie. number of different words on a single context position), proximity of two units (ie. their mean distance) and obligatoriness (ie. percentage of the instances of a unit reserved for its existence in combination with other unit). Properties of these three measures are discussed as well as their mutual relationship.