By Daniel Hubbard | March 17, 2013
I remember hearing a story years ago of a man who sat in an archive, in Spain or perhaps Portugal, sniffing documents. He was trying to detect the subtle, lingering sent of vinegar. Why? Long ago people attempted to prevent the spread of disease by sprinkling mail with vinegar. A letter might claim “we are all well here” but if its smell had that hint of sharpness, someone knew, or at least suspected, otherwise.
Records can supply genealogical information in many ways. Usually we don’t need to smell them to extract that information. What we normally do is read the contents. The contents of a document generally supply most of the information, but that document also has a context. Sometimes we forget to consider the context. Other times we might use both contents and context without thinking about it. What information comes from the contents of the document and what information comes from the context? Thinking about that question can solve problems.
A Move that Wasn’t
I’ve gotten to the point where I know a family that I am tracing fairly well. I have all sorts of documents that pinpoint their location at various times, often only months apart. Early on, I had found them in the Kansas state census in 1915. It was an interesting hint about the instability of this family’s living situation. I found them fairly far from where I expected and they could not have remained there for very long. It might have been the kind of move made by someone struggling to find work.
As I worked on the family, gradually, while collecting other documents, that 1915 location was becoming stranger and stranger. When I stopped to look at the big picture, that 1915 location started to seem impossible. Even if they hadn’t been there on the official census day and had only been caught by the enumerator later, it just didn’t seem to fit. I decided to look at that state census again. There they were. Exactly were they were supposed to be and yet where they really couldn’t have been.
The answer to the problem is probably contained in contents versus context. I looked at the contents of the document. There were the names, the ages, the birthplaces—everything that one needs to declare a perfect match. There was something missing though—the place. This census page did not itself have the place written on it. Neither did the previous page. In fact none of the pages I examined had the place where the enumeration occurred written on them. The place could only come from the context not the contents. Context can change. I looked for evidence that the context in which I found this document might be wrong.
I found that evidence. First when I checked how many pages supposedly filled with name after name by one tiny little place, it was suspiciously large, as if many pages were swallowed by a context that was really much too small to hold them. Then I started to check names that seemed easy to read and unusual enough that they should be easy to find without any ambiguity. Where were those people in 1910 and 1920? They were all somewhere else in 1910 and in the very same somewhere-else in 1920. That somewhere-else happened to be where I would expect to have found my family on March 1, 1915 and were I am now reasonably sure they were all along.
They didn’t move. Their census pages did the moving. The contents were found residing within the wrong context.
When information comes from context not contents, beware of altered context.Twitter It!