RootsTech 2015

Some people eat, sleep and chew gum, I do genealogy and write...

Tuesday, March 18, 2014

What are the benefits of a unified family tree?

I got the following comment about a recent post concerning user owned vs. a unified family tree. See what you think.
I believe the best result a unified tree can hope for is a collection of smaller and bigger clusters of people. Lack of evidence or proof (of parentage, for instance) will ultimately prevent all these clusters to be interconnected. I am still confused about what the real benefits would be of a unified tree, however, and most of the reasons you give in favor of this kind of collaboration are circumstantial, rather than deliberate.
My first impression is that I need to clarify the idea of a unified family tree. Every person born on the earth has a unique biological set of parents and a unique birth order. I am not going to get into the ideas of test-tube babies and all that, even in those instances there are still unique biological parents even if those parents cannot be adequately identified.



A unified tree models this reality. Each node (individual) on the unified tree only has a place for one unique individual. Now, the question of "clusters." It is possible that isolated areas of the unified tree could exist, especially between two cultures. It may also be true that the connections between these isolated cultures are so far in the past that they lack a historically verifiable connection. But that is not fatal to the concept, that is only a practical issue. You cannot abandon the concept merely because of the present difficulty of establishing connection. Let's assume a tree made up of a local ethnic group that has little contact with any other group. The fact that this sub-tree exists is not a condemnation of the concept it is in fact a confirmation of the the need for such a conceptual model. The main strength of the model is that within the "clusters" the mechanism of the tree forces the participants to agree upon the identity of the individuals in each node.

Let me illustrate a hypothetical situation. Let's suppose that I am working on identifying my ancestry for the first time and beginning with my own parents. Let's further suppose that there is a robust unified tree reflecting my particular ethnic background. What are the chances of my investigations back along my ancestry leading me to a connection with the exiting unified family tree? If the unified tree has reached a point of "critical mass" then the possibility of connecting with the tree is close to 100%. Why is that the case? Because of a basic genealogical fact called "pedigree collapse." Here is the explanation from Wikipedia:
Without pedigree collapse, a person's ancestor tree is a binary tree, formed by the person, the parents (2), the grandparents (4), great-grandparents (8), and so on. However, the number of individuals in such a tree grows exponentially and will eventually become impossibly high. For example, a single individual alive today would, over 30 generations going back to the High Middle Ages, have or roughly a billion ancestors, more than the total world population at the time. 
This apparent paradox is explained by shared ancestors. Instead of consisting of all unique individuals, a tree may have multiple places occupied by a single individual. This typically happens when the parents of an ancestor are cousins (sometimes unbeknownst to themselves). For example, the offspring of two first cousins has at most only six great-grandparents instead of the normal eight. This reduction in the number of ancestors is pedigree collapse. It collapses the ancestor tree into a directed acyclic graph.
A unified tree provides the potential for interconnections even if the practical reality cannot prove that those connections exist. Otherwise, you would have to believe that ultimately there were separate human species that had never intermingled.

What do I mean by "critical mass?" When the tree reaches a point that it has included a significant percentage of any given ethnically related community, it will contain all of the people needed to establish links with almost all families. Because of the number of people in the FamilySearch.org Family Tree, I assume that this critical mass was reached some time ago for ethnic groups coming from Western Europe. Unfortunately, there are still areas of the Family Tree program that do not function and so this fact is obscured by the program's disfunction.

OK, now to the advantages. Here is my list of advantages:

  1. The unified tree forces collaboration. If you and I share a common ancestor and we both are trying to fill that node on the tree, we must come to a common consent as to the information about that ancestor. However, most users of a unified tree see this as detrimental rather than beneficial. They are so wrapped up in the supposed accuracy of their own research that they cannot countenance the existence of contrary opinions. They also do not want to take the time to educate their supposedly "unreasonable" relatives. 
  2. The unified tree ultimately avoids duplication. There is an assumption made that somehow the same sloppy, undocumented garbage that exists in user-owned family trees will spread to a unified tree. This is the case initially, but as the tree matures, there will be less and less space for this type of problem to occur. If you judge the utility of a unified tree by the accuracy of the initial data set, you are lacking an understanding of the mechanism of how the tree evolves. The initial data set is arbitrary. It could be gibberish. What matters is the involvement of those who have the "correct" information. As time passes and more and more changes are made to the tree, the information becomes more and more accurate. When I am talking about duplication, I do not mean merely duplication of individuals, which is not possible, but duplication of research and effort. Once my grandfather is in the tree and documentation, photos, stories etc. are included, there is no need for each generation to repeat that process as is happening today. 
  3. The unified tree provides a reference point for serious research. Anyone who is interested in any given family line merely has to go to the tree to see what has already been done and what needs to be done. Presently, in the FamiySearch.org Family Tree there relatively very few documented entries. As that changes over time, it will be possible to determine the state of research of almost any family line included in the tree.
If you do not see these points as benefits then you still do not understand how a unified tree will work. The major issue is frustration with people adding or changing information with no accountability. The weakest link in the present FamilySearch.org Family Tree is the ability of users to make changes without having current contact information in the program. The second weakest link is the ability to make changes without providing a source. If these two issues were resolved, a measurable amount of the frustration with the tree would disappear. 

6 comments:

  1. "When the tree reaches a point that it has included a significant percentage of any given ethnically related community, it will contain all of the people needed to establish links with almost all families. Because of the number of people in the FamilySearch.org Family Tree, I assume that this critical mass was reached some time ago for ethnic groups coming from Western Europe"

    Hmm. I am very dubious about that. Thinking about the people in a photo of my GG GF's brother, taken in 1916 in San Francisco, I'm finding it difficult to find them in FS FT. The brother is indeed in FS FT in his birth family back in Dundee, Scotland. Now that's good because I didn't put them in, someone else must have. ("FamilySearch" did, of course! Gee thanks, please see all previous conversations!) But there's no sign of his wife, who is Baltimore born, in FS FT. And, of course, none of their 2 daughters. No way either of linking the California Robert to the Dundee Robert.

    No sign of son-in-law number 1, nor his daughters. Son-in-law 2 *is* in FS FT - slightly shambolic in the data, but he's in with his birth family, unlike son-in-law 1's birth family which is missing.

    So, if I take you literally and just look at FS FT, we are rather short of critical mass - though obviously this is hardly a statistically significant test. But what *is* true is that the relevant data is there in the US censuses, which I presume are all in FS Records.

    So, I think the point is that descendants of Robert from California cannot stick bits together inside FS FT. They have to do old-fashioned genealogy with records, outside FS FT. Not that you or I have a problem with that!

    I don't have a problem with the idea of a critical mass - it's just that I think that (outside the Mormon community maybe???) we're light years short of it in FS FT. But not necessarily if you look at US and UK censuses in FS Records.

    Adrian

    ReplyDelete
    Replies
    1. I guess I wasn't very clear. What I meant to say was if you keep following your ancestry back a few generations, you will almost always connect with the tree if your family comes from Western Europe. Not every line but there will be a connection.

      Delete
    2. OK - I can happily accept that there will almost certainly be *a* connection somewhere. As I said, Robert is there back home in Dundee (you just have to do a bit of work to get there). I also have a line linking into what I might call the heartlands of FSFT as one of my Bate relatives from Cheshire ended up in Utah. Ironically, that line doesn't show FSFT at its best as it's affected by the issues with the link to the LDS membership software, so every few months 4G GF becomes his own father in FSFT. Or is it his own son?

      Delete
  2. Your "arguments" for a unified family tree are mighty. I have shared them with my class and have remarked that I hope someone at FamilySearch will read the last paragraph over a few times. Thank you for your input, always!

    ReplyDelete
  3. Some edits in FS FT would also be nice. I had one where the birth year was entered incorrectly so she died before she was born. That said, both Puzzilla and CreateFan have tools to assist with the data cleanup. I think FS FT is a good place to share research, but I plan to keep my own copy of my tree.

    ReplyDelete