This paper describes some of the theory behind the technology of record linkage. Record linkage is a fancy term for the determination that two or more records relate to the same entity, such as individuals, events, places, etc. In genealogy it is natural to have records for people and for sibships, i.e., nuclear families — for groups of closely related individuals. Different records of birth and death may belong to the same person. Different records that may be used as evidence for a family, e.g., marriages, divorces, and births of children may belong in the same family. The genealogist's goal is to reconstruct certain families of certain ancestors whose only traces are in records. Through experience with records and a knowledge of languages and cultures she may make judgements about which records are of interest to her particular goal.

After introducing the objects of linkage, we will discuss how the studies of probability and statistics may be brought to bear on modeling the way a genealogist makes record linkage judgements. This constitutes the study of probabilistic record linkage. In four additional chapters we discuss some of the concomitant challenges in some practical applications of probabilistic record linkage.