Wednesday, November 6, 2019

Genealogy and Bayes

Let's do a recap to set the basis for the post. We are addressing a case that concerns a descendant of John Alden and Priscilla Mullins; he is buried in a small cemetery in Nebraska. Lyman Porter is his name. He is #37963 in the database offered at He married Caroline Hopwood. His parents were Noah Porter #20029 and Nabby Cumins. Lyman has a brother, Reuben, with descendants in the Mayflower organization (MF). Lyman has a daughter, Anna, who has descendants in the MF.

As an aside, we are building a reference database for this work on WikiTree using the great-granddaughter of Lyman and Caroline (Gladys Helen (Gill Defibaugh) Long Murray). Gladys gave MF a little problem, too, as did her mother. Chloe and her daughter, Myra, both died young. Chloe's granddaughter (Gladys) was an infant when Myra died; she was raised by her father's (Arthur's) mother, Phylura, until Phylura died; then, Gladys was adopted by friends (Defibaughs - also neighbors) of Phylura; Gladys married twice. But, all of this is documented. It's Chloe generation where care needs to be taken (more below).

In the fall of 2015, we saw a line from Alden to Porter being mentioned in a letter to the editor of a heritage group publication. Someone had noted that they had joined the MF organization via Anna, daughter of Lyman and Caroline. Okay, we matched up Anna's sister, Chloe, and asked the applicant to put in the original query. You know, fill in the form, pay, and then wait. So, we were contacted and assigned a researcher. The interaction and work was all done via email and DropBox.2 There was a flurry of activity in early 2016 as we gathered documents. By summer of 2016, we were essentially done and had a slew (lots and lots) of stuff.

Well, things bogged down with respect to the MF interfacing. On the suggestion of the MF person, we did further work by doing a probate search. We found out some additional information, but that did not get us off the bottleneck.3 For us, though, it was a learning experience. Too, the Court records actually confirmed up some of the information. So, that is a positive (see below discussion). Some material shows that Chloe's family and that of her husband, Thomas, were quite prominent in the count and state of their residence when she married, had a baby, and died. How did Chloe get forgotten?

We are at 2019 which is three years later after the last bit of work. Things are just sitting in limbo. In 2016, we wrote the first article as this seemed to be a case of problems related to western expansion not being as appreciated as they might be. Some romanticized the whole notion. Others were such as to leave lasting lessons in endurance. In fact, at the time (2016), there were some talking as if history, and especially that of 400 years, was of no consequence. Too, though, my thought was that the MF ought to be looking for descendants of their ancestors rather than playing membership games (do you see me now? more below). The first article was published; a second one provided another look and was published; finally, we will get to the details with another article.4 The articles, to date, are: Flyover Country, Pseudo Wall (not brick at all), and, at least, one more that is in the works.

So, we have our recap with this brief review.

Now we can get to the gist of the issue. What is the hold up? Well, in retrospect, we probably ought to have worked through the Alden group (more below) since they do consider the whole family. But, let's look at some reasoning about this.

Chloe is the daughter in question. She is not documented with respect to birth. Neither was her younger brother, George. But, George died young and is recorded, at death, as the son of Lyman. Chloe died young, too, but as the mother of a young child. Chloe is not even recorded for marriage or death. We have a photo of Chloe's grave stone which names Thomas. We have Chloe and Thomas with the baby, Myra, in the 1870 Census (US). BTW, Chloe (name misspelled) is in the 1850 and 1860 Census (US) with her folks and siblings.

One question is: can we take those people in the Census (US) as related? May we look at the milieu? If we look at the Census (US) of this family, they were consistent in their reporting. Really, as we would expect a citizen to do to help the government. George was in one Census; he died; in the next one, he's not there. Chloe's younger sister, Emmira, was not in the first; she is born and documented as daughter of Lyman and Caroline; she's in the next Census (US) as we expect, birth order.

BTW, the State of Chloe's birth did not record these events until after she was born. Be that as it may, in the 1850 Census (US), when Chloe is an infant, there are some other people in the household. Some we know as her siblings by later census records. But, there was an older woman, Permelia, and five younger people. Who were these folks? Given what we know, they were Chloe's grandmother and Chloe's cousins. The husband of this family (Chloe's grandfather) is recorded as having died. Guess what? We are talking Caroline's siblings (and her mother) in her household with Lyman. One of the sibs was four years old. Permelia was 44; Caroline, herself, was 23.

20th and 21st century genealogists, at their little tubes, have no clue about the reality of these situations. However, one thing to note. The birth order was there. As in, filling in these things, there were ground rules, which the Census (US) taker would have known. That's a strength as it persists across several Census (US) occurrences.

We could itemize other strengths. That is things that confirm, albeit, not overwhelmingly. But, these can be thought of as additive. For instance, Lyman notes that Myra is his granddaughter. In fact, on both her marriage and her death records, Myra is referenced as the daughter of Chloe and Thomas. Court records indicate that Myra is the heir of Thomas. And, Court records indicate that Myra, after her father's death, was living with Lyman and Caroline in Nebraska. Then, there is a Census that names Myra in the household of Lyman and Caroline.

There are many little tidbits, of various types, that add to the notions about Chloe's family. So, let's transition, a little, before looking at those.

Rev. Bayes (his bio at St. Andrews - see footnote for a look at the use of Bayes and its importance to the modern world's complexity)5 has made a great impression on the modern age. Computing? The past decade plus, we have seen systems relying on his approach. In fact, lots of the problem now can be attributed to this being taken too far, however, Bayes will continue to be of importance. What the Rev. argued is that we can improve our view of a probability by considering related information. Take death.  An older person can be said to have a higher probability of dying than a younger one. That is, given a number of other conditions being left out (accident, disease, ...). However, at some point, some older persons actually start to have their probability of 'not dying' (in a particular year) go up. Hence, we have the very old.

Bayes' ideas were around for a long time. Why their attraction now versus before? Well, I was using his ideas five decades ago. But, it was hard to do this stuff by hand. Typical situation. The evolution of the computer got to where there were easier methods (algorithms improved). Too, the need came up; Bayes fit the need; the modern minds over fit the situation. Another story altogether.

In genealogy, this case is an example. We might say that we're looking at the probability of Chloe's membership in her family. Let's look at the positive side; then, we'll go to the other.

You know, certain minds would take an official document as 100%. Ah, we know of problems there. Transcription errors (even wrong recordings). Illegible mess that is not readable. There is a whole lot more. Essentially, we go from 100% to 99% or less. The thing is that the 'lead feet' syndrome of expecting complete certitude is not realistic (some relaxation might actually be productive).

Given that we do not know when Chloe was born via official views, we can use the accumulation of material to build up some 'belief' in her status. That is why I used positive influences, above. In the last go-around, when dealing with details, this will be more thorough. Actually, having Chloe's grave stone says a lot. it gives her name and middle initial. Also, it names her husband. And, it gives her age. She was 24 years old when she died and left her daughter.

Flyover country
Why might she have been forgotten? A step-mother came into the picture. Then, Thomas died, too. So, that is the story of Guardianship for Myra and the Probate process. This can all be told and taken from the view of Chloe being the daughter of Lyman and Caroline.

As fortune may have it, we were able to get the Probate records for this case. They paint a fairly good picture of the situation. However, we know that the brother of Thomas was named guardian of Chloe's daughter. We know that another brother of Thomas handled (was executor for) the estate in the Probate dealings. As I said, these guys were well known. The step-mother? She fled but got a settlement. Myra came out okay in the deal. She graduated from Monmouth College. She taught music. Ah, but, Myra married a railroad man. Know where this is going? Luckily, this Arthur's mother took Chloe's daughter in. But, Phylura was elderly.

Now, are there other ways that one could think of this case other than that old thing of checking boxes? You bet. Remember, we ranted at the MF people for not respecting the wishes of their elders who would have wanted to know their offspring. Too, I called it dissing. Three generations of women, scoffed at. However, we now know that the family organizations connected with MF have extended the work. Such as, the Alden group has data out to eight generations. And, we can feed them more information.

But, let's just pause a moment. We will put four alternate views here. And, look at them. What will we see, then?
  • Chloe was found in a cabbage patch? Oops. Did we write that? Or, she dropped from the sky? Why this? We're talking a pioneering family. They were not sitting on their fat behinds back on the east coast of Plymouth and surrounds. Chloe was her mother's oldest daughter. Caroline lived long enough to die in LA with her daughter at her side. That is, Chloe's sister. Lyman had died (his lonely grave in the countryside - broken stone). So, Caroline saw early Hollywood. We can tell stories about that. You know, Chloe named her daughter after her sister who is documented, by the way. That is, the sister that Caroline was living with in LA. 
  • Chloe was dropped off by another family? We know of these things. We can talk a case of a woman not knowing her ancestry as her father had dropped her off at a western household (she was an infant; her mother died) saying that he would return. Never did. We can dig deeper into the local history. Oh yes. The MF person says to me: "you're interested in the history (fuller picture); we want so show pedigree" (sure, would be my response, with antiquated thinking - we can write that since the NEHGS has weighed in to help MF. Recent paper where NEHGS did the heavy work of handling the circumstantial information. Oh yes. MF took it. Well, MF, there are other researchers than NEHGS around. Pro-bono, too.). 
  • Chloe was Caroline's sister? You see, Permelia was of child-bearing age. Her husband died in 1849. What was going on? Lots to look at there. But, it does not pertain to the MF stalemate. There were other siblings of Caroline. Perhaps, they were too much for her mother. 
  • Chloe had a father other than Lyman? Bite your tongue. Caroline was the backbone of this family way into the 20th century.   
You see how ridiculous all of this gets? They would have varying probabilities but would weigh on the negative side. Actually, they are too close to zero to consider. The accumulation of the positive material (there is a slew, as said) would push out a fair description of this family unit, of whom one is wrapped already in the MF arms - is that hug worth anything?

John and Priscilla are watching, we might say. How long before they can see the offspring from Chloe added to the 'official' MF roles as if that really meant anything?


1. The group has presented its Caveats about use. Our own caveats: we are not genealogists, but John knows modern research methods including how to program and to solve mathematical/logical problems in a slew of technical domains. We started this work in 2009. This blog is meant as a record of work and findings. It is partially indexed at our portal to truth via images:
2. This summer, we helped a DAR applicant with her eApp. It was all done at the DAR website and the phone. The approach was fantastic. We finished the application in September and submitted it. By the end of October, it had been verified (there was a wait in the queue). A week later, the person was a member. That was the latest example. We have filled in, successfully, all sorts of applications for many friends for many of the organizations under the Hereditary Society Community. Frankly, we started doing applications to confirm our work (all pro-bono). Some genealogists who reviewed the applications noted how thorough these were.
3. We just had all of the material of this case reviewed by another person who saw no hole in the presentation. In fact, the term 'preponderance of evidence' necessary was used. Our next step is to document this with all of the material shown. The two articles mentioned in the text are overviews. The next will be highly detailed. Why? Gladys and her two prior forebears (Chloe and Myra).
4. These articles were published in The Gardner Annals and are in print form. See "Publications" for how to order.
5. There has been a lot written about Rev. Bayes and the use of his ideas. For a simple introduction, we can look at this example (Remarks 11/08/2019 - common sense and Bayes - flirting) or a better one from D. Joyce (of genealogical fame - his was a site that we visited often as we got into the work in 2009). For a more advanced look, Griffiths of Berkeley and Princeton offers a nice view chosen since it predates the latest trends: Bayesian models in Cognition. We'll get to the more recent (at some point, for now: Manifesto from 2014).

Remarks: Modified: 08/07/2020

11/07/2019 -- Gladys (see above) will be the focus of a major study. She has Alden through Porter. But, she has Brewster through Gill on her father's side. So, that's a merge under the western sky. ... Say "Genealogy and Bayes" quickly (or with the proper tongue) and hear genealogy and bias. This study can be used to weigh in on truth engineering as it pertains to the mischief out of 'silicone' valley.

11/07/2019 -- Added links to material, both introductory and more advanced, on Bayesian ways and means. Leaving aside, for now, the question of how all of this applies: Cultural heritage.

11/08/2019 -- Went looking for a good example and found several. This one stood out: Flirting -  An exercise in Bayesian statistics - at Medium. If you don't want to read the intro to Bayes' thoughts, go down about 1/2 of the page to the "flirting" part. It starts after this: "Every time you change your mind because of new evidence, you’re using Bayes’ Rule." I pulled out this piece (please read the whole post):
    So what does this have to do with flirting? You’ve probably guessed it by now. Flirting is a high-stakes, emotionally-charged manifestation of Bayes’ Rule. There are a couple of ways to think about this.

    First: you want to figure out if your love interest likes you back. You start out with a relatively uninformed prior, just a random guess as to whether or not they like you. Maybe you had a bad first impression, and you set your prior at P(they like me) = 25%. That’s way to uncertain for your taste to go declaring your love. You want more information first.

    So you hang out more, you go on dates, you see each other more often, and in each of those instances you’re collecting data. Maybe they hugged you tightly — the probability of that happening if there was a 25% chance they liked you is pretty low — maybe you underestimated yourself in the beginning and now you update your prior of 25% to a posterior of 27%. Maybe you touched them and they instinctively recoiled in disgust → update that prior and reduce it to maybe 5%. And on and on we go.
What we have is the need to update our priors so that we can have a better posterior. Tricky. Now, in reference to the theme of this post, how many have been dissed?

11/11/2019 -- Changed to using names to reduce the reference stacking.

08/07/2020 -- Cumberland Pass in the west has the same name as that in the east known as the Gap. However, they represent the movement across the country which established the country. So many stories. Too, they created instances like poor Chloe being dissed by Plymouth people squatted in eastern Massachusetts. It's a large country out here, folks. So, I have now seen oodles of families with the same problem from the tip (Canadian border) all the way to the lower part of Texas. Has anyone really awakened to this? Not that I can see, otherwise I would not have the need for this type of post (Genealogy and Bayes).

No comments:

Post a Comment