When categorization is applied to a non-quantifiable subject such as evidence, this tendency may obscure instead of highlight valuable evidentiary conclusions. For example, if I designate a source as "primary," that designation alone may add credence to the information contained in the document wether or not the information warrants such a designation. To avoid this issue, many researchers start creating even more categories, trying to adhere to the schema of categorization despite the inability to adequately create the degree of categories necessary to describe any particular piece of evidence or any particular document.
I will start with an example of a U.S. Federal Census Record. The types of information contained, in say the 1930 U.S. Census helps to delineate the problem. Here is a list of some of the information that may be found in an enumeration page:
- Full name
- Age (can be used to calculate an approximate birth year)
- Relationship to the head of household (active military personnel in naval yards, army posts, etc. may use the term "Sailor" or list military rank rather than actual relationship to head of household)
- Birthplace of the individual and the parents (included even if the parents were not members of the household)
- Marital status (single, married, widowed, or divorced)
- Year immigrated to the United States
- Whether a naturalized citizen
- Native language if foreign-born and whether can speak English
- Whether a military veteran
- Street address and house number
Now, is a Census record a primary or secondary source? Don't we really have to look at each piece of information to make that, or any other, distinction? As a further illustration, many researchers rely on the information in the U.S. Census to add the names of family members in their ancestral lines. But how accurate is the information in the Census when it comes to names? Fairly reliable by usually incomplete and may be totally inaccurate or even misleading. It adds nothing to the accuracy of the information in the "Full Name" fields by classifying the record at all. The classification of the "record" has no bearing on the accuracy of any particular data field. So, are we looking for categories; i.e. this information is correct because it comes from a primary source document or are we looking for reliability.
My position is that categorization does not help to evaluate evidence. Many secondary sources, as designated by genealogists, are more reliable than corresponding primary sources. As a matter of fact, my conclusions as a researcher may be more valid than any of the available "primary" sources which may all be wrong (inaccurate) for some social, political or emotional reason.
So what do we do if we expunge the idea of categorization? We move to evaluation of individual pieces of evidence. We evaluate each item as it is presented in the document for consistency, reliability and accuracy. This only works if the information can be corroborated from other sources. For example, a grave marker may or may not be primary or secondary source but why is that important? Why bother classifying the sources, what is important is whether or not the information on the gravemarker is accurate. How many inaccurate gravemarkers are there out there in the world? Lots and classifying them makes little or no sense and certainly has no bearing on the reliability of any one marker.
I am all for classifying things like birds, insects, rocks and such, but I balk at classifying evidence. The evidence is what it is. Whether you believe it or not is really the question. Now, I am not talking here about proof. Proof is a highly subjective issue. If you have ever served on a jury for a murder trial you know what I mean. Whether or not any evidence is sufficient to prove a given subject is up to the researcher to decide. But whether or not a particular piece of evidence is reliable or unreliable has nothing to do with assigning it to an artificial category.