Section 1-6 ELEMENTS OF DATA PROPAGATION



Example of date propagation.   It may be helpful to see what may happen in sources without very complicated propagation of information. Consider this example of how a genealogist might propagate a date: Suppose the marriage of a couple is determined by the license record to have occurred subsequent to March 4, 1734, i.e., aft March 4, 1734. Suppose further that the same marriage is determined by the birth of the first (legitimate) child to be previous to June 16, 1735, i.e., bef June 16, 1735. The genealogist then says that the probable marriage date of abt March 1734. What does abt (formerly often "ca." for Lat. circa) mean and why does the genealogist prefix it to the marriage date? From the culture of this particular event the genalogist is aware that a license allows the couple to marry before the otherwise necessary publication of banns on four previous Sundays. Hence, the marriage probably occurred within the month. That the marriage actually occurred is attested by the birth of a legitimate child. We call this indication of date interpretation a date prefix.

Example of locality propagation.   Here is how a genealogist might propagate a locality for an event. Suppose the first child of a couple is determined by the church record to have occurred in the parish of St. Michaels. Suppose further that the couple is determined to have been married in St. Lukes. The genealogist then says that the probable birth place of the bride is of St. Lukes. What does of mean and why does the genealogist prefix it to the birth place? Knowing the culture of the time and place, the genealogist realizes that a couple will usually marry in the parish of the bride's residence, and further that a young single girl will hardly ever move from the parish of her birth. Hence, barring a move by her family of orientation, her birth probably occurred within the parish of her marriage. We call this indication of locality interpretation a locality prefix.

Example of personal name propagation.   Even in western cultures there are many elements comprising a person's name. In a greatly simplified form the linguist sees four obvious parts: 1) pre-positive title, such as “Mr.,” 2) given names, 3) surnames, and 4) postpositive title, such as “Esq.” Lacking a given name it is often appropriate in America to refer to a male by “Mr.,” to a married female by “Mrs.,” and to an unmarried female by “Miss” with the surname. In this country the wife usually acquires and children all inherit the surname of the husband viz. father as their family name. Particularly irksome is the novice genealogist who writes a name for the wife in her family of procreation as one acquired in full from the husband: Mrs. John Smith. We call the pre-positive title as an indication of given name interpretation a name prefix.

Elements of a semantic base for dates.   The semantic interpretation of a date is determined to be two Julian day numbers, one specifying the beginning of a time period, and the other the end. There must also be some indication of the precision of each of these real numbers. Sometimes this number is known to be accurate only to the year (365.0, i.e., ±182.5) and sometimes to the day (1.0, i.e., ±0.5). Even with precision this representation with just two day numbers should be extended. Approximations are more than the modification of the precision of the beginning and end of a time period. Normally there is a middle point between the two ends that we are fairly confident about. Usually the plus and minus tails (of a skewed probability distribution) are not symmetric. Thus we need to specify three, not two time periods. Sometimes also the two ends are a terminus post quem and a terminus ante quo, i.e., we are fairly confident only that the event took place at some time between the two dates.

Elements of a semantic base for localities.   The semantic interpretation of a place is determined to be one or more geopolitical place names as follows: one specifies the highest governmental or geographic region, and the others specify certain kinds of places so that each one would be included in the designated higher region, while all but the last would include the ones lower. The geographical location (coordinates) of the place so specified may be known to various degrees of precision. Sometimes the place is known to be accurate only to the country designation and sometimes to the town or parish. Even with precision this representation could be extended. Approximate knowledge is more than the modification of the precision of the place. Often we have a choice between two or more likely places. Usually one place is more likely than another, in other words, the probability distribution is skewed rather than symmetric. Thus we need to specify multiple, not single places. Sometimes also there are two places, one being a terminus ad quem and another the terminus de quo, i.e., we are fairly confident only that the event took place between two places — the beginning and ending of a journey, as between Liverpool and Boston, or St. Louis and Great Salt Lake.

Relationship.   The specification of a familial relationship may be important if it is known. Some other relationships (social), such as “friend” or “sponsor” may imply that P1 and P2 are possibly near the same age and/or from the same place. A rule involving these relationships would be rather imprecise. Rules for most relationships are valid in the case of birth date and place propagation, but in the case of marriage date and place propagation (in the P-family of P1) it is “self.”

Marital status.   The marital status is always conditional on the value of relationship. In the case of birth date and place propagation, where the relationship is “self,” there is a rule for each possible status: “generic,” “single,” “married,” “divorced,” “widowed,” “unknown,” etc. In the case of marriage date and place propagation (in the P-family of P1), the status must be “married.”

Sex of principal(s).   The sex of the principal may be either one (or not specified) in the case of birth date and place propagation. In the case of marriage date and place propagation (in the P-family of P1) it is always one of each.

Date precision.   There are two senses in which it is important to specify precision. One sense was referred to above. It has to do with the Gregorian calendar, which allows specification of either day, month, or year as the smallest known unit. This sense is related to the semantic interpretation of a date as a period of time. The other is in the sense of a prefix describing its general nature, such as “about,” “calculated,” “estimated,” “before,” “after,” “probably before,” etc. The rule specifies the type of event and prefix of the input date (of P1) and the prefix of the output (propagated) date belonging to the event of P2 being assumed.

Locality precision.   As with dates we may specify place precision in two senses. One sense was referred to above. It has to do with the geopolitical divisions, which allow specification of county, state, or country type localities. The smallest known unit is related to the semantic interpretation of the locality as a geographical point or polygon on the earth. The other is in the sense of a prefix describing its general nature, such as “near,” “of,” “probably near,” etc. The rule specifies the type of event and prefix of the input place (of P1) and the prefix of the output (propagated) place belonging to the event of P2 being assumed.

Conditions on age of principal(s).   There may or may not be certain conditions on the age of P1. The rule may specify whether the age may be greater than a certain number (of years, months, weeks, days) and/or less than some such number. Possible measures on such demographic constraints from which to choose are:
Adult status (gender specific)
Start of child-bearing years (for female persons only)
End of child-bearing years (for female persons only)
Average expected age at marriage (gender specific)
Average age at retirement
Average expected age at death when of adult status (gender specific)
Average expected age at death when of non-adult status
In the case of a marriage the age of the male principal may be the reference age. In certain cultures there would also be an important constraint on the female principal’s (wife’s) propagated age: total days of female cannot be more than total days of male.

Parameters needed for calculations.   The reference date is the date of the event document or vital event of P1. Any of the following parameters may be required.
Average age of a married individual
Add or subtract the average age at birth of child
Add or subtract the average age difference between spouses
Age of principal
Add or subtract the average size of a generation

Another parameter is a set of deltas — days to add and subtract from the reference date to get a time period. The positive and negative may be identical (symmetric) or different (asymmetric or skewed). There may also be constraints on the deltas. For a symmetric set of deltas you only have to specify one span for both sides and each generation of the relationship, which is then averaged.
Reference date
Child-bearing years
Same delta as the reference date
Average age of mother at the birth of a child
Average age difference between spouses
Average intergenesic interval
For an asymmetric or skewed set of deltas you only have to pick the skew side.
Adult status
End of child-bearing years
Average age at retirement
Average expected age at death
Same ending delta as the reference date
One half the child bearing years
(Absolute) age difference between spouses (2:1 OR 1:2)
In either case there may be one of the following constraints specified:
Either a maximum or a minimum extreme
Size limit to delta:
Average expected age at death
One half the intergenesic interval
Normal age for adult status (gender specific)