The present contribution deals with the grammatical category of number in Czech nouns. On the basis of empirical investigation, we propose the introduction of a new semantic distinction within the forms of nouns, namely the distinction of a simple quantitative meaning versus a pair/group meaning. There are nouns in Czech that typically refer to a pair or to a (commonly encountered) group of entities, not merely to a large amount of these entities (e.g. ruce ‘arms’, vlasy ‘hair’, or sirky ‘matches’). These nouns are combined with set numerals rather than with the basic ones. In the paper, we argue that the above mentioned subcategorization is grammaticalized. A brief outline of the expressions of the pair/group meaning in Czech as compared to some other languages (German, English, Slavonic) is also provided.