I know a person who says they are one-fifth dutch. Is it possible for a person to be one-fifth anything? Can’t you only be 1/2, 1/4, 1/8, 1/16 etc.??
Perhaps he finds it easier to say one fifth (as an approximate) rather than the more exact three sixteenths.
…or 13/64.
It’s quite possible to be any percentage you like in terms of ancestry. Imagine for example that my parents are double 2nd cousins, not an uncommon thing especially in the past. That means that I effectively only have 5 great grandparents. Provided that one of those great grandparents is Dutch I am 1/5 Dutch.
Of course by going even further back we can produce any other ration you like. 3rd cousins can produce 7ths and 9th, 4th cousins 11ths and so forth via various combination with a half Dutch ancestors and so forth.
Remember that most people have family trees with intertwined branches within about 5 generations.
Um, no, Blake, that’s not quite right. Suppose that person A’s parents are B and C. B’s parents are D and E. C’s parents are F and G. D’s parents are H and I. E’s parents are J and K. F’s parents are also H and I. G’s parents are L and K. (This is as close as I can come to your example.) It’s true that A has only five distinct great-grandparents, but you would expect that on average that A got one-eighth of his genetic material from each of J and L and one-quarter of his genetic material from each of H, I, and K. I would say that if one of J or L is Dutch then A is one-eighth Dutch and if one of H, I, or K is Dutch then A is one-quarter Dutch. No matter how far back you go then, your ancestry could only be expressed as a sum of fractions with denominator a power of 2, and 1/5 cannot be expressed this way.
You get those sorts of convergences when you start tallying up fractional ancestries. For example, I show up as a mixture of English (with Welch mixed in), French, and Dutch. But when you start tracking it back, they come out in about equal proportions – my French-descent great-grandfather married an English-descent woman, my English-descent great-grandfather married a Dutch-descent woman, and so on. So while there’s no way I can be 33.3% each, the numbers converge towards equal thirds.
I’ll bet that helps when you get yourself into a jam…
I’d just like to add that I’m 100% Dutch.
I’m also 1/8 French, and 1/4 Surinam. The Surinam part is made up out of 50 % Polish Jew, and 25 % Ethiopian.
Does that clear things up? 
Berry funny!
The kind of jam I had in mind is when you accidentally type a C for an S.
Maybe I’ll start using Cymric, or perhaps Cambrian… :dubious:
What I always wanted to understand is how Gilgamesh’s ancestry could have been two-thirds god, and one-third human. The same arithmetic problem: how do come up with an odd-numbered denominator? Perhaps his granny and grampa were brother and sister? Like in the novel Middlesex by Jeffrey Eugenides?
Suppose you have A, B, C, D, E who are pure-blood of five different ancestries. A, C, and E are male, B and D are female.
A and B have a male child F. B and C have a female child G. C and D have a male child H. D and E have a female child I. (Clearly a rather non-traditional intermixing of parents, but certainly possible).
F and G have a male child J. H and I have a female child K.
J and K have a child L.
Clearly L has exactly one fifth of each of the ancestries of A, B, C, D, and E.
Now I know that this exact scenario is extremely unlikely, but the fact that it is theoretically possible demonstrates that the proportions don’t have to be 1/(power of 2), and more complex combinations involving more distant ancestors could also cause someone to have one fifth of some ancestry.
No. B, C and D all are doubled ancestors 3 generation back, whereas A and E only occupy one slot - 1/8 A, 1/8 E, 1/4 B, 1/4 C, 1/4 D. However you draw a tree of ancestors, your inheritance from any one ancestor will have a power of 2 in the denominator.
Not so clear, dmartin. A and E have one child each, while the others each have two children. So A’s and E’s genes will be less represented in subsequent generations. F and I will be likewise penalized, as will the end parents in each generation. If you carry it all the way through, then L will be one sixteenth A, four sixteenths B, six sixteenths C, four sixteenths D, and one sixteenth E. Uncoincidentally, 1 4 6 4 1 is the fifth row of Pascal’s triangle: If you tried the same thing with six in the starting generation, you’d have 1 5 10 10 5 1, and if you started with seven, you’d get 1 6 15 20 15 6 1.
I think I’m right on this one - 1/8, 1/8, 1/4, 1/4, 1/4. Draw it out:
A B B C C D D E
F G H I
J K
L
C doesn’t contribute more than B or D. In general, the rule is to follow a path back to an ancestor, dividing by two each step of the way. If an ancestor appears multiple times, add the paths together.
The indentation got screwed up - put F, G, H and I under the proper pairs in the top row, J under the F,G pair, K under the H,I pair.
No no no. You misunderstood, I love me. This is what he meant by one fifth Dutch.
One Dutch fifth, you mean.
Ha! Ketel One is better! Plus, it comes fom Schiedam (near Rotterdam) where I used to live.
Oops. I was adding another layer, with G also mating with H, then the three grandkids producing two great-grands, and finally a great-great-grand. Which would, admittedly, be rather incestuous. I guess I got sort of lost in all the letter people.
Can we summarize this by saying that Mr. One-Fifth Dutch is either bad at math or has a family tree so questionable that makes one wonder why he’d talk about ancestry at all?