Every genealogist knows, or should know, that keeping track of sources is a must. Each fact should have a source.
There is something hidden in that last statement. It assumes we have a clear idea of what a “fact” is. Do we?
Slicing and Dicing
How small can a fact be? Is someone’s name a fact or is it many facts? I would say that it is many. A typical American for the last few centuries has three parts to their name. We don’t always have a single source that clearly gives all three. If a person is a bit of a mystery and three sources list him as John Public, another lists J. Q. Public, and a newly discovered source has J. Quail Public, what is the man’s full name?
If we assume that all the sources refer to the same person, we might say that his name is John Quail Public but is that really one single fact? Are there five sources for that name? Five sources support the first initial “J.” Three of those support the name “John.” The middle initial “Q” is supported by two sources but only one gives the full middle name of “Quail.” Different sources support different parts of the name. That name isn’t so much one fact as several facts.
Can you trust those sources to the same level? Probably not. Different parts of the name are probably not just supported by different amounts of evidence but by different levels of quality. If you look at the name “John Quail Public” as one fact and think that it is supported by five sources, it may slip your mind, and go undocumented in your records, that the middle name that we might use to tell him apart from all the other John Q. Publics in the area was only supported by one source. If that source is weak as well, we might be in trouble. If a grandson’s memory that his grandfather’s middle name was “Quail” was wrong, so are we. What if we felt confident in our “well supported” name and used it to filter away a few things turned up about John Quince Public? What if Quince was the middle name we should have been targeting?
Ambiguity
How well defined is any one fact? Is it a fact that a child was born on August 6, 1781 if there is some evidence that suggests that he was born on the 16th or in 1782? What if there is some weak evidence that implies that the family had a child of that name that died and you just might have the birth of the first but you need the birth information for the second?
What sources do you attach to your “fact”? Perhaps you work only with those that cluster together to support August 6, 1781. You list those sources as all supporting that date. Perhaps, as above, different sources support different parts of the date. Beyond that, what do you do with the sources that imply otherwise? A source that says August 6, 1782 could be said to support August 6 so you could use it but it also casts doubt on 1781. Do you simply refer to it, or do you keep separate track of the support for the month and the day and the ambiguity for the year? Do you keep track of those sources that give a different date, the ones that might correspond to a second child of the same name? Should they be buried in a file or should they be referenced directly from the birth you are trying to understand?
The question becomes, do we think of a fact as an actual value or is it a placeholder for something that we hope to know? Everyone was born, we just don’t know when or where without doing some work. We start out with an empty box in a form or in a database interface. Does there need to be a well supported value there and then only the sources that support that value or do the sources instead belong to the space where we would like to record a value? If the sources belong with the space, then we can attach to that space all the ambiguous references that may or may not be relevant but that should not be forgotten.
That Empty Box
I look at that empty box as two things. It is a place to put the best distillation of the information presented by the sources. The sources that go along with that value are naturally the ones that support it.
The box is also a thing in itself. Many sources can have something to say about what might be entered in that box. All those sources should be connected to that blank whether they agree with the current thinking or not.
Sources tend not to present an amalgam. They say one thing or they say another. Each will present its version of the facts. Each supports something, though perhaps only part of what we could call a “large-scale” fact. That large-scale fact might be a birthdate but a source might give a month of birth and nothing more. In order to really understand the status of a large-scale fact, it is wise to keep track of what sources support what parts of that fact and what sources give alternate possibilities. Some sources will match the large-scale fact in all its glory, others only one or more of its parts and finally, some will contradict the whole large-scale fact or some of its parts. Only a full view of all that complexity will help when another source comes along or a conclusion based on our large-scale fact fails to be true.
Very true. For years I’ve been hoping for the ability to source the day, month, and year of a date separately.
What I do these days is to include a short note with every source that gives the information that actually comes from the source because it often is not exactly what the combined information from all the sources implies.
I struggle with this issue. My grandfather was John Sheppard Groce but he always went by J.S. This is a fact because I’ve seen both names in original documents from my dad’s files and baby book ca 1901. Grandfather was mentioned in newspaper articles, county histories, advertisements etc. online.
For ancestors farther back, it would make my teeth itch to put five sources, with explanations, for a name and an aka in my database.
I’m thinking of sprucing up my research logs (or creating from notes) and sourcing there. In the database, a “see” footnote. For 15 years of research an awful task but needs to be done.