The presentation of Adam Brown and E. Randol Schoenberg at the International Association of Jewish Genealogical Societies’ conference in Boston in August 2013 and their articles in AVOTAYNU, Summer 2013, examine the future of genealogy research with particular attention to the needs, skills and habits of the next generation in the context of Geni.com. Schoenberg concentrates on how Geni works and on its goal of becoming an all-encompassing tree that covers everyone and merges the work of all researchers everywhere. Clearly the Geni-type tree is meant to replace other, more traditional genealogy platforms, and researchers are encouraged to record their findings directly into it. While I agree with much of what both authors say, the approach presents significant challenges, discussed below. It is more than the problem of recording erroneous findings.
Brown’s description of young people, how they work, how they think, what attracts them and what does not is excellent. As I pursue my own discussions with a 31-year-old nephew about becoming my research heir, I am aware that he would bring skills that I do not even know I need. Nevertheless, when Brown writes that the next generation is “disinclined to spend their hours conducting endless searches” and is “far more willing than past generations to share information,” I imagine errors being copied from one person to another with less than rigorous attention to source material and citations. I have no disagreement either with what Brown writes about images and metadata. The fact that connecting images to my tree is not a priority for me does not mean that I do not favor development of tools that will serve others.
Collaborative Online Genealogy
Some of my major concerns involve collaboration. I fully understand and support the idea of collaboration. When JewishGen introduced the Family Tree of the Jewish People many years ago, I asked if it did not make sense to merge these separate trees where there was obvious overlap. The organizers replied that such an approach had two problems. One was technological; the other problem, relevant here, is that the content was supplied by people on their own authority, and JewishGen did not think it was their place to meddle in that. JewishGen was correct then and that argument is still valid, even if the technology barrier is less relevant.
My Pikholz Project has been collaborative since its inception 15 years ago. The families are traced on a website I maintain that provides not just individual trees (both public and private versions), but assorted other tools and analyses. Some of the more research-minded family members use these tools to guide their own corners of research, and their input enables me to keep everything current. This is collaborative online genealogy, even if it does not allow others to make changes unilaterally.
In addressing the question of errors, Schoenberg observes, “People said Wikipedia would be full of errors,” but this is an irrelevant argument. We see ourselves as members of a proper discipline, urging academic institutions to accept genealogy as a legitimate field of study. Can legitimate academic researchers get away with citing Wikipedia? My friends in universities tell me that no one would do so, both because Wikipedia is not properly peer reviewed and also because it is an aggregation rather than actual source material. Thus, casting Geni as “Wikipedia for genealogy” does not serve his argument well. By the same logic, should a serious genealogy researcher even be citing some anonymous Geni input as a source?
Aside from the issue of accuracy, each genealogy researcher has too much of his personal self, his own personal history invested in his work to allow someone else to sweep it away. What a person submits to Wikipedia is not nearly so personal.
Many of us worry about surrendering even a tiny bit of control over our work. Yes, it is our work, not the work of some collective. Brown’s phrase “overseen by a team… deputized as ‘curators’ who roam the Geni tree like park rangers” features words that sound authoritarian. And these folks are going to “ensure the integrity and accuracy” of my work? I would rather not. In the same vein, Schoenberg writes, “Not only can individuals find the mistakes, they can also fix them—for everyone.” No, please don’t. If someone shows me an error—whether a typo or real content—tell me and I will fix it myself. I have been doing that for years. I can adjudicate my own issues, thank you. This is not just my attitude. Recently, a similar topic on one of the professional genealogists’ discussion groups showed that many others feel the same way.
Brown writes, “For genealogists concerned with having complete individual control of their data, aggregators will offer stand-alone trees with limited collaborative functionality, automatically matched to the databases described above (the present Ancestry.com and MyHeritage models).” Seasoned researchers may be forgiven if they are less than welcoming to the comprehensive system that assures, in essence, something along the line of “If you like what you have, you can keep it.”
The realm of errors includes more than just fixing them. Here are some examples. My father told me that his grandfather, whom he knew only as a young child, had an uncle named Selig Pikholz. None of the others in my father’s generation knew that bit of information, and we have found no documentation other than the similar names of their children. Knowing this fact enabled me to make significant progress in my family research. I take my work seriously and consider myself to be an accomplished genealogy researcher. By what right would any “deputized curator” tell me that this is not good enough for some institutional “integrity?” Whether or not some curator has yet done so, the very fact that they claim the authority to do so is sufficient cause for concern. Somewhere, some time, some roving park ranger may decide to mix in such things. Or some disgruntled family member will issue a challenge just because he can. Brown himself writes, “Family stories may turn out to be fabrications.” How do we know that Brown will not put Uncle Selig into that category?
Some months ago, a young man who had recently married into my family told me that he and my mother-in-law had some common lineage. When I asked for specifics he sent me to a Geni tree supervised by someone he did not know. There I saw several generations of my mother-in-law’s known ancestors, plus some additional couples whom I did not know. I went to JRI-Poland and found some supporting documents and ordered them. Some supported what the Geni site listed, but others did not. I went back to Geni and found that things had been changed—with no trace of the earlier entries and to something else that also was incorrect. The posted information was wrong, misleading and changed at will. It had nothing to do with me, except as a person who was somewhat dependent on what she had posted. Search though I did, I never found on that tree the young man who had married my cousin and who had sent me to this Geni tree in the first place.
Schoenberg reports that Geni now includes Neil Rosenstein’s The Unbroken Chain, but at the IAJGS conference in 2007, Rosenstein spoke about significant errors in his work because of incorrect interpretations of certain punctuation and antecedents. Is Geni, six years later, working with corrected data? If Geni is sweeping up information from published works, can users be certain that everything is the latest version? In my summer 2013 article in AVOTAYNU, I mentioned a biographical sketch that has an incorrect birth year, and I repeat the comment I made then. “It’s written in a book, so who checks it?” Recently I had to make a change in my database; a Page of Testimony had been submitted to Yad Vashem in 1955 for a woman who we later learned died in Warsaw in 1977. I have no document to prove her date of death, but two people who knew her, one of them her grandson, said so. When Geni chooses to nourish its “super-tree” from unfiltered and self-authorizing sources such as Yad Vashem’s Pages of Testimony, it easily may insert errors into other peoples’ well-researched and well-documented work as a result of copying faulty information.
The problem does not always involve major institutions. More commonly, the disagreements are near to home. For example, two of my close family members have placed wildly incorrect information on publicly available trees. For years I have been trying and failing to get a cousin to fix his errors on the Family Tree of the Jewish People, with no success. Convincing submitters to make corrections is not always easy. I prefer not to involve Geni’s curators in these family squabbles.
Schoenberg describes one of the differences between Separate and Mergeable collaborative trees. In the first case, algorithms “detect matches and suggest them to the customers,” while in the latter case Geni has “the ability to merge separate trees.” That “ability” sounds very much like more of the authoritarianism discussed above. I do not believe that serious genealogy researchers want that kind of authoritarianism. Some casual genealogists may enjoy the convenience, but undoubtedly the results will suffer.
In my summer 2013 AVOTAYNU article, I wrote of seven children born to Isak and Feige Pikholz and strongly suggested that there may be two such couples, not one. If I am not blocking the door, what prevents the merging algorithm from deciding on its own that these seven are siblings? Even if the merging algorithm is right 95 percent of the time, within the other five percent, it can be 100 percent wrong. Names are tricky things. They repeat in families in a single location, but they also have variants that make them look different when they are not. How can we trust that Geni’s merging algorithm is sufficiently sensitive to the nuances of Hebrew and Yiddish names and nicknames to make correct decisions? As a serious genealogist, I do not want to transfer authority to make interpretation to anyone else. Suggestions are welcome, but the final decisions must rest with the researcher.
Public Trees, Privacy and Security
Schoenberg twice writes that one of the advantages of Geni over traditional platforms is that public trees are searchable on the Internet. He says that others can find a genealogist’s work on Geni because it offers searchability. In fact, my own website, which I maintain with no special keyword tricks, is searchable too, and every few months I receive a note from someone saying, for example, he found his grandmother’s birth on my website and wants me to know that she grew to adulthood and produced three generations of descendants. These messages relate not only to the more complex parts of my website, but also to the simplest family charts. Invariably they have found me via a simple Google search.
When Schoenberg writes, “There always will be smaller trees that people have not shared or have not been migrated to Geni,” he is implying that this is not the way he wants it to be. Brown’s statement, discussed above, the essence of which is something like, “If you like what you have you can keep it,” is part of the same phenomenon.
Schoenberg addresses the issue of turning everything over to Geni to a collaborative project with the not entirely reassuring “Thus far, we do not see much risk of any of the major collaborative tree projects disappearing.” The words “Thus far,” “much” and “major” are not reassuring. And the word “disappearing” could just as well be replaced by “deciding to charge $100 a year.” He concludes that paragraph with the following, in which several words raise (or should raise) red flags: “…it seems unlikely that the data ever will disappear.…most of the platforms allow at least a limited ability to download a GEDCOM file of some size to protect your work.” The equivocation and limitations expressed here are not reassuring.
My database and my website are not the same. I use the term database here to refer to the program that holds my basic family data. It exists solely on my computer and its backups. In my case, it runs on Brother’s Keeper. Other such programs include Family Tree Maker by Ancestry.com, Dorotree, Reunion and Legacy Family Tree. The term website, of course, refers to what I put online for others to see. The database is definitive. What is there is fully vetted and should contain proper citations. I have an entry for the Mordecai Pikholz who married Taube and had a series of documented children, and I have a separate entry for the birth and death of what must be that very same man based on a death record. Since I cannot prove that it is the same man, a note in my database says “almost certainly the same as….”
I have mentioned Uncle Selig whom my father said was his grandfather’s uncle. I have no doubt this is correct and my website reflects that—but my database does not. His entry in Brother’s Keeper has a note that says “almost certainly the brother of…” and cites my father as my source. My website is not definitive. It is meant to offer a snapshot of the family structure, sometimes more general than specific, depending on the context. There, Mordecai is listed simply as “~1805–64” and I can put Uncle Selig where I believe he belongs. My database has no place for dotted, broken or colored lines denoting various levels of possibility, probability or uncertainty. My website—when properly updated—does have these features. For example, I have drawn some conclusions based on DNA that will appear in my website but not in my database. I think I know what the DNA shows me, but there is some uncertainty, so it cannot be recorded as definitive.
If a genealogist chooses to put all his or her work on Geni or some other similar site, and only has a website that is not even his own, how can he maintain those two different levels? Can a collaborative online tree even begin to deal with this? Do curators even want to think about mediating among researchers who have differing levels of certainty? Such is what comes of the Collaborative Online Tree where people are encouraged to find information or documents and to enter them straight into the collaborative site. When a genealogist’s database and website are one and the same, one of the jobs is not being done properly. The definitive database and the illustrative website are different. They serve different purposes. Relying on something like Geni pretty much devastates that concept.
Geni and sites similar to it are useful, but we must not allow them to ride roughshod over those of us who prefer not to use them as a repository for our work. That kind of self-control likely will become more difficult over time because of the implied pressure from Geni that collaborative sites are the way of the future. The general genealogy community myst not accept collaborative online trees as the default standard assumed for everyone. Let it remain one tool among many, appropriate for some but not for all.
Also published at http://www.pikholz.org/Articles/Geni.html.