Section 1-3 MODELS AND ENGINES



Lineage-linked model.   Figure 1 below illustrates the elements of a lineage-linked model. There are basically two ways that an individual may relate to a family genealogically. An individual may be a child in various families of orientation (FO or O-family), blood parents, adoptive parents, etc. This is the individual with siblings and parents in a nuclear family or sibship. A person may be a parent in families of procreation (FP or P-family). This is the individual with children and spouses in nuclear families. By symbolizing the relationship to the family in two different ways, and distinguishing the individuals by sex, it is possible to symbolize all the kinship terms used in a nuclear family in a simple and concise way. The family of the last paragraph (F1 of figure 1 of section 1.2) is the family of procreation for two individuals (I1 and I2 in figure 1 of section 1.2) and the family of orientation for two others (I3 and I4 in that figure).

Figure 1  Symbolizing Nuclear Family Relationships

Basic genealogical relationships..   Beyond the relationship between an individual and a sibship, i.e., family (F) there are the relationships between two individuals in the same sibship. The relationship between an individual and the parents in his family of orientation is called a child relationship. A child is either a son or a daughter of the parent. The relationship between an individual and the children in his family of procreation is called a parent relationship. A parent is either a father or a mother of the child. The relationship between an individual and other children in his family of orientation is called a sibling relationship. A sibling is either a brother or a sister of the individual. The relationship between an individual and the other parent in his family of procreation is a spouse relationship. A spouse is either a husband or a wife of the individual, depending on the individual’s sex. All these relationships may be symbolized as illustrated in the above figure 1. We have assigned distinctive symbols to each of the kinds of individuals and the kinds of relationship to a nuclear family.

Entity class transformations.   Often a particular resource is a collection of documents of a single kind. Normally a document records an event that involves individuals. Sometimes a whole collection, such as the ancestral file (AF) is lineage linked. Such a collection usually has individuals as separate records and uses the other features of events (dates, places) to identify them. If we are involved in individual linkage, it is necessary to specify whether the record that the individuals are contained in are 1) event documents or 2) family structures. These are the two common contexts of an individual — its so-called merotype. In either case there are transformations required to fill out the individual or family linkage record. One important kind of transformation relates the class of the linkage entity across records. These are the entity class transformations. They specify to the record linkage system to link the father in one family to a male child in another, for example, or the husband in a census schedule to the groom in a marriage return.

Data propagation rules.   There is another important kind of transformation. These are the so called data propagation rules. These algorithms produce virtual data — field values that were not actually in the original record. The calculation of a birth date from an age at an event is a simple example of such a rule. In some resources certain data is sometimes present, in others never. Propagation may provide data to be compared against where it was not present or only indirectly surmised. In comparing records the record linkage system does not weight agreement of virtual data as heavily as agreement of real data.

Individuals & their merotypes.   In th first paragraph of this section we discused the lineage-linked model, perhaps the most obvious of the merotypes of an individual. Other models that would contain individuals are the document and the event. This particular kind of event, say eventdoc, contain individuals in certain roles. The document contains the source citation and the eventdoc(s) that are on record. The document may contain births in a certain place. These are attributes of the document, but we want to use these events to help identfy the people in them. An individual is the merotype of such identifiers as: 1) names (including aliases and titles), and 2) vital and other events. These would be say eventind, which would not include individuals in roles, which associate the principal (or other person) in a corresponding eventdoc. Typically the principal of a birth or christening eventdoc has a birth or christening eventind of the same date and place. In the lineage-linked model individuals also have several relationships to other individuals. The eventind has its type along with its date and place. This is also the structure of marriage events, say eventfam. Hence the merotype of an event is a document, a family, or an individual. (Events may also belong to almost any other class of object that has a temporal beginning and/or end.) The date has a certain precision (on, in, about) along with the day, month, and year (depending on the calendar) and period of time. So also a place has a certain precision (at, in, near) along with the locality name and location. It is important to realize that the locality designation is appropriate or not depending on the time period of the event.