In this contribution the main features of the concept of the 'double bind' are described. Although the theoretical status of this concept is highly abstract, the theory from which it originates has important and immediate implications for the practice of family- and other forms of therapy. However, the concept belongs to the so-called 'high inference' categories; therefore it is difficult to operationalize. Some other variables of family interaction are more easy to investigate. Recent research suggests that it can be fruitful to investigate family interaction in connection with the problem of schizofrenia.