Evidence-based Genealogy - part 1: what is evidence?
15 February 2019 by Fouke Boss
15 February 2019 by Fouke Boss
Centurial claims to be true evidence-based genealogy software. In this blog we define evidence, and we demonstrate how Centurial helps you to extract evidence from sources.
Evidence-based genealogy is sometimes called source-based genealogy, and with a reason: it all starts with sources. Assuming that a genealogist does not make things up, she/he always gets her/his information from someone or something. This someone or something is called a source.
When starting with Centurial, or with genealogy for that matter, it might be a daunting prospect that every piece of information needs to come from a source. But by this definition it is perfectly ok to start off with the things you know about a person or a family: this simply makes you the source. In Centurial, this is supported by the source type Personal knowledge in the source dialog. If the information is coming from someone you discussed a person or family with, the source type is Interview.
Of course, personal knowledge or interviews might not as good a source as for example a birth certificate, but thats a completely different subject, one for a later blog post.
The next step is to identify the claims in a source. A claim is an assertion that something is true. A source might state the name of a person, a birth or a death date, of an age, or a residence or occupation, or a marriage date.
As an example, let's have a look at a page in the birth register of the town of Zundert in the Netherlands, in the year 1853:
This source contains many useful claims, including, but not limited to:
The source editor screen in Centurial allows you to enter all the information in a particular source. In this case, we would first extract the persons and relationships into the information panel like so:
Then, for each person or relationship, the claims are extracted:
However, information and claims are no evidence... yet.
Evidence could be defined as 'the available body of facts or information indicating whether a belief or proposition is true or valid'. So evidence starts by a belief or proposition; in Centurial this is called the research question. Then, information and claims only become evidence when we believe these claims to be related to a research question.
Let us clarify this with the claims we found in the birth register earlier. The register states that on 30 March 1853 a (any, non specific) child was born with the name Vincent Willem van Gogh. It does not state that this child is in fact the world-famous painter Vincent van Gogh. Only when we believe the claims to be about the painter, only when we correlate the information in the birth register to the painter, the information and claims become evidence.
This is in fact a very important part of doing genealogical research, and one that many other genealogical software applications do not support: claims become evidence only when we believe they relate to the same person. And when we later find out that they might not be about the same person, they are no longer evidence.
For example: Vincent van Gogh had one older sibling, an older brother who was born exactly one year before him (30 March 1852) who died at birth. This boy was also called Vincent van Gogh. When presented with the birth register extract of this brother, a researcher might at first believe this to be the painter himself. So the date of 30 March 1852 becomes evidence for the birth date of the painter. Only later on, when additional evidence is collected, the researcher concludes that this might not the painter after all. Centurial then allows to easily recorrelate the claims in this source to another person.
So when doing genealogy with Centurial, the big magic step in the whole process is the correlation step. After all information and claims are extracted in the source editor, the user clicks the button to start the correlation process. As every person in the source needs to be mapped to an possibly already existing person, the correlation step might become tedious and boring to the user. That's why Centurial tries to help the user by using an algorithm to map the information in the source to persons. Hence the name auto correlate.
After correlating, Centurial shows the information in the source (the smaller rectangles) on top of the persons they are correlated to (the larger rectangles):
Please note that the auto correlation algorithm is not (can not be?) perfect, it will sometimes yield less-than-perfect results. However this is not a too much of a problem, as Centurial has many facilities to improve the correlation afterwards. Auto correlation is almost always less work than doing the complete mapping manually. For more on this subject, we recommend this tutorial.
So after correlating claims to persons, they become evidence. In Centurial, evidence of a person or relationship can be viewed by selecting the person or relationship, and then selecting 'View Evidence' from the context menu under the right mouse button:
For each research question (family name, given names, gender, birth date, birth place, etc.), conclusions are drawn based on the available evidence. Clicking the analysis button opens the Analysis dialog which shows all the evidence that was gathered for that specific research question:
Note how Centurial allows the researcher to mark a piece of evidence as plausible or not plausible. This is because sources might be wrong themselves. This could be anything, from typos in original records, failing memory from people you interview (ah, the human brain!) or incorrect conclusions in authored sources. By allowing evidence to be marked as not plausible, evidence can be correlated to people even if parts of the evidence turn out to be contradictory.
Evidence can be categorized into several types, each with their own pros and cons. These types of evidence, and the way Centurial supports them, will be the subject of the second part of this blog post.