The problem is that so much of the work is based on assumptions and faith. You have to follow the line of documents like you climb a ladder, hand-over-hand one rung at a time, trusting the steps you've already climbed to support you. But the ladder is endless. And the higher you get the more you wobble in the wind, and you keep hitting tree branches and, I don't know, power lines.
Say you have a name, the name of your great-grandfather. You know when and where he was born and when and where he died and where he and your great-grandmother lived while they were married. You find a census record for a man that fits with the information you already have, and contains the name of your grandparent during the proper time frame, so you know the man on the record is probably the man you're investigating. (Ah, probably. What a hateful word.) You use the census to find out that your great-great-grandparents were born in a different state--though you don't know their names--and you search for the censuses from the time your great-grandfather was born, looking for him as a child to establish who his parents are. You find one that looks promising--the dates are close enough, the places are right--and so you have new names, new leads.
Wash, rinse, repeat. Apply other records as necessary.
Never mind that there are about fifty bazillion "John Smith"s in that state's census for that year, or that everything is handwritten in cramped, casual cursive--"Is that a c or an a," you ask yourself, nose pressed against the computer screen. "Does that say Olive or Oline?"--and you'd be foolish to expect to find something which exactly matches the information you have. Never mind maiden names, and the ordeal of immigration records, complete with name changes and spelling discrepancies and the promise of records in foreign languages.
There is no confirmation email, no seal of approval, nothing to guarantee the accuracy of the family tree you're constructing. You just have to trust in probability, in the strength of your matches, in the fact that whether or not it's applicable the information on the documents is true. And, at least in my experience, the more documents you handle the more muddled you get. The result is that before long a sense of paranoia sets in, and I have to stop because I honestly feel like I'm going to go crazy if I continue. Wait, why am I searching naturalization records for Gladys? Oh, right, to confirm Gladys's immigration. Which I know happened because the 1890 census said Myron's parents were born in Poland. Remind me again why I care about Myron? Oh, right, Bertha's birth certificate said Myron was her dad. Remind me again why I care about Bertha? Who ARE all these people? Why do I care about any of them? Where did these names COME from? How do I know this is really who I think it is?! HOW DO I KNOW THAT'S REALLY MY MOTHER, HMMMM?! ...The house of cards will fall very, very quickly if you poke it too much.
I was at a genealogy center the other day getting help (because seriously, immigration records are a pain in the ass), and the person helping me discovered that I'd gotten someone's maiden name wrong, and so the people I had been looking for records for, Conrad and Wilhelmina, were in no way related to me. With one click of a button they were deleted from my tree, gone, and I was left with a new last name and the vast, enticing openness of possibility. I had been thinking for months that I was connected to these people. It's unsettling to think how easily familial ties, a core foundation of our society (and of most others), lauded throughout literature and media and history and common culture as A Bond Above Others, can be discredited.
And it's unsettling to realize how connected we are to people we've never even met. My grandparents on my mother's side each had seven siblings, and my great-grandma had fifteen. My aunt and I stopped putting the siblings on our tree because it was just getting ridiculous, and we weren't really interested in the siblings, we were interested in the parents. But those siblings all had families of their own, husbands and wives and children and grandchildren and sons-and-daughters-in-law who all have parents and siblings of their own, who...you get the idea. It sounds obvious, but I've never really thought about it before, about its implications, about these people as more than names. A lot of the information we've added has been based on information from other people's family trees, who had been interested in the same people we were and had already done the research on them. Because they're related to them too. My great-great-great-grandfather is also someone else's great-great-great-grandfather, some distant cousin I've never even heard of or given a thought to but who, from Simon backwards, has the exact same family tree as I do. And on their trees, they treat my great-grandma as I treat their relatives, disregarding her as an obscure aunt they don't care about, another name on the census, whereas to me she is of the Utmost Importance, because she's mine. It's even stranger when people have posted pictures of your shared relatives' graves; it really hits me that there are other people out there, strangers, who have the same eager interest in and reverence towards these predecessors, that they have the same claim to them as you do.
I recently discovered that my great-great-great-great-grandfather had the same birthday as me. He was born in 1830 in Norway. The way we attach ourselves to the dead, create links and loyalties, really is very odd, when you think about it. Yet when the genealogist helped me trace one line of my mother's family to 16th century Denmark and another line to 16th century Germany I was as proud and possessive as if I'd known these people personally. At what point do these connections become irrelevant, meaningless? Because with all of the marriages and births that would have occurred (and did occur) in five hundred years, then everyone is connected, however obscurely, to everyone else. And people certainly don't act like it sometimes.
--Flannery