The Trail of Two Georges

When it comes to errors in genealogical documentation, I guess the main question is: how badly does that bother you?

You know that squishy feeling. You look at a record or two concerning your target ancestor, then take in someone else's work, and think, "This doesn't look right." Perhaps you can't articulate why, but you don't get a good feeling about the splat of it all.

It is so easy to grab the first record you find in the family's county of residence—say, the 1880 census—and decide, "Yep. This is it. I've found my family."

That may not be the case, at all. But you didn't look far enough to determine that, yet.

When it comes to a surname like Gordon, it pays to look twice. While not as common a surname as Smith or Jones, the surname Gordon occurs frequently enough to make me want to proceed with caution on this research trail.

Couple that with ethnic tendencies of past centuries to employ traditions such as naming patterns, and the likelihood that extended families may have emigrated en masse to their new frontier home—each married brother naming his firstborn son exactly the same—and you have just doubly doomed yourself to the possibility of same name relatives living within close proximity of each other.

A case like that prompts me to come up with alternate hypotheses when looking for family records. And those hypotheses are formulated to attempt to prove myself wrong.

So, when I came upon my ancestor in question today—George Gordon and his wife Sarah, living in Perry County, Ohio—I didn't want to prove that he is my man when I found him in the census. I want to prove that he isn't my guy. In other words, I want to search for someone else who can prove Find Number One to be in error. If I don't succeed in that forced false response, only then can I consider my discovery safe.

Let's take a look at George Gordon. As far as I know, I'm looking for the youngest son of William Gordon and his second wife, Mary Cain. Since George's father William died before the 1850 census, it's no surprise to see youngest child George in his widowed mother's household at that point. Since there is no reference to relationships in that census, and since Mary is admittedly also a common name, such a record may also be prone to duplication. So we can't rest our laurels on this one discovery. There are too many missing details we still need to fill in.

Still, we can glean from that record that our George (if this was the right George) was unmarried, eighteen at the time of the 1850 census, and was born in Ohio. That's a start.

Let's assume, for the moment, that his age was reported correctly—yes, a huge assumption, as census enumerations go—which makes his year of birth approximately 1832. And let's take that hypothesis out for a spin with the brand new census model in 1860.

It seems like smooth sailing on this road out to Reading Township in Perry County, location of the old William Gordon farm. There, right next to his older brother Mark, is the entry for George Gordon. Just like clockwork, the man has aged precisely ten years, showing his current age now to be twenty eight. As can be expected from a marriage record giving the year of their wedding as 1858, he and the former Sarah Jane Dittoe are now proud parents of a one year old son. In honor of George's father, and with a secondary possible nod to Sarah's father's name (which I've yet to uncover), they named their child William Jacob Gordon.

All was not perfect with this record, though. True, George's widowed mother Mary was still with them—although her place of birth had somehow morphed from Maryland to Pennsylvania. But that was an understandable minor detail. More puzzling was the presence of another child—a four year old named Harvey Taylor. Admittedly, two households away from the Gordon farm lived a family by the name Taylor, but what was their (supposedly) four year old son doing here? The most likely scenario in that time period was that there would be some sort of family relationship—but none that I've discovered, so far.

Setting aside those small discrepancies, let's move onward to 1870. There, once again, we find George and Sarah in Reading township. Gone was his mother, Mary Cain Gordon, who had died three years earlier. Gone also was any mention of that mystery Taylor boy—or the neighbors by the same name to which he might have been related. Nor was George's brother Mark in the vicinity, either—although in his place was the young family of Adam Gordon, George's orphaned nephew who had been in Mary's household along with George in the 1850 census.

Thankfully, positioning near these other Gordon relatives had helped persuade me that I found the right George Gordon, as one other missing detail from this 1870 census entry was the only descendant whose record I had gleaned from the earlier census. Where William Jacob was gone, he had been replaced with stair-step siblings: Mary, Rebecca, Martha, Charles and Thomas.

All would seem convincingly complete, if we concentrated solely on the assumption that, having found one entry, it was the only entry possible. However, nine pages further into the enumeration for the very same township—Reading—what do we find but another entry for a George Gordon. And yes, this one was also complete with a wife named Sarah. This couple also had a son they named Jacob, followed by two other young children, Israel and Mary. Had I found this George and Sarah first—and not felt prompted to continue looking, after this discovery—I might have made the wrong assumption. After all, how many times do we see census discrepancies in locations or dates of birth?

To follow through and insure we haven't stumbled upon a fluke—you know those inscrutable census enumerators—let's check out the 1880 census, just to make sure.

And there they both were, residing in Reading township again—the two Georges with their respective wives Sarah. One—which used to have a son Jacob—now without a Jacob, while the other household has one. Both households with a son Thomas. And a whole lot of other children.

This is one of those instances where the absence of the 1890 census is keenly felt. Whatever. Let's see what the 1900 census brings, after that twenty year record absence. 

This is where we run into road hazards in Perry County. More than that, we have an entire fork in the road. One of our George Gordon households is entirely gone—now living in Douglas County, Illinois, it appears. Because of the twenty year gap, it is hard to compare the names of the children to the previous census record. The only possible overlap occurs if the hard-to-read 1900 entry for "Chris V." Gordon is actually the Charles Gordon of the 1880 census.

The George Gordon family remaining in Perry County includes two children too young to have been included in the 1880 census—one tantalizingly named William H. Gordon, same as our George's oldest full brother, father of the Adam Gordon George grew up with.

What to do? The two Georges reported birth years varying, at the most, within a five year range of each other. They each had wives named Sarah. Each had a son named Jacob, and later another son named Thomas.

Of course, I can always reconstruct the history of each of those children, delineating each child by the maiden name of his or her mother. But that does little to confirm which Sarah belongs to the George who was son of William and Mary. And that is the key. After all, I'm tracing the line of George, son of William, son of John, the immigrant.

Of course, it may turn out that both Georges are descended from John. It is very possible, considering the path for both of them led to Ohio through Pennsylvania. But you know I can't rest until I have each George in his proper place in the Gordon family tree. Any less would never do.