A method is described for the identification of possible links among dream sources and the study of their possible significance. The analysis is based on the automatic recognition of word root recurrences in text files, including dream reports and associations. Two tools are then applied: graph represen- tation and grammar analysis. Graph representation of the detected links provides a quantitative description of some of their basic features. Grammar changes for recurrent word roots can imply remarkable context changes. A plausible explanation of the identified context changes can evidence interest- ing phenomena connected to the significance of links among dream sources. Two examples of application of the method are given, 1 taken from the literature and the other from sleep lab data.
