Lets consider a current cultural institution
and examine some alternatives. The institution is our naming system.
The traditional American method stinks. I propose some alternative
solutions. First, we should note that some other countries do things more fairly. For example, Pablo Alliende writes that in spanish countries, a married woman keeps her two surnames, and children receive one surname from each parent. The only unfairness is that paternal surnames went first, so a woman's grandchildren would no longer have any part of her name, however a law was recently passed in Spain allowing people to reverse this if so desired.
Proceeding, we assume that the point of last or "family" names
is to indicate something about the lineage of the individual,
so that is what we will try to do. Note that there are two issues
- how to modify names upon marriage and what to name the kids.
Also remember that these algorithms (which take the parent's names
and tell you the kid's name) must be "lossy", that is
they must throw away some information. Since we are trying to
encode clan information in strings of characters, and the number
of ancestors is very large, we can only represent a small portion
of that data. Every generation needs to take a double set of this
data and reduce it to a single one (its like a cryptographic hash
function). By definition, if names don't grow, we get to keep
only half the data from the previous generation.
Solutions:
The conventional American method is that one spouse takes the
family name of the other. This seems to imply one-sidedness and
domination, a bad thing for a relationship (although if one spouse
really detests their name, that seems a reasonable exception).
Also this makes female clan and line of descent unnoticed and
irrelevant, when it is contributes just as much genetically as
the male line. The standard alternative for those who object is
for both spouses (if the plural of mouse is mice, what is the
plural of spouse?) to retain their names. While this removes the
inequity, it means that the couples names have not changed to
reflect their marriage. Since much of the purpose of marriage
is to produce children, thus uniting two genetic lines ("clans"),
this seems unsatisfactory as well.
Selective Combination is one simple method. When they marry, a
couple comes up with some unique combination of their family names.
For example, when my friends Rob Prestegard and Anna Harzog married,
many of us encouraged them to change their last name to Prestezog,
which for some odd reason they resisted. This does not discriminate
against either spouse, and it lets them to pick something they
like, but it doesn't allow much clan continuity.
My favorite solution is for each person to have two last names.
One comes from the mothers side, the other from the fathers. You
get your mothers mothername (matrilineal clan), and your fathers
fathername (patrilineal clan). To make order fair, we can have
daughters go patri matri and sons go matri patri (or the reverse,
depending on whether first or second denotes more important).
So a woman's daughters passes on her family name, and a man's
sons his. A man's mothername will not be passed on. Both names
are used in formal settings, but perhaps only one is used the
rest of the time. An example.
Mother: Anna Kiraly Edes [Anna Edes for short, Kiraly
is father-clan, Edes is mother-clan]
Father: Jason Smith Jones [Smith is mother-clan,
Jones is father-clan]
Daughter: Krista Jones Edes
Son: Marcus Edes Jones
So brother and sister just have reversed last names, so you can
still tell their relation. Unfortunately, while this allows sibling
identification, it does not put information about husband/wife
in their names. There are some possible patches for this. For
example, when marriage occurs, the man replaces his matronymic
(which he isn't going to pass on anyway) with that of his wife,
and she does the same with her patronymic. This has a good balance
to it. Men always have the patronymic of their male clan, which
is perpetuated eternally down the generations. When unmarried,
they have their mother's matronymic, and when married they have
their wife's. Perhaps a modification could be made to indicate
which is currently true. For example, the prefixes ma =
Mother's, pa = Father's, sa- = Spouses. Now we have:
Anna pa-Kiraly Edes (?ma-Edes?) marries
Jason ma-Smith Jones
They become:
Anna sa-Jones Edes
Jason sa-Edes Jones
Their children are:
Krista pa-Jones Edes
Marcus ma-Edes Jones
Notice that we can tell by the prefixes whether two people are
husband & wife or brother & sister. This is necessary
since the well-known phenomenon of "attribute drift"
in couples, where the personality, speech patterns, clothing style,
and even physical appearance of the two become homogenous, renders
observable similarities insufficient to distinguish the relationship.
I think this is a good system for retaining a reasonable amount
of clan information. For very formal occasions, we could reach
a generation higher to get 4 (8, 16...) last names, with special
prefixes to show which is which.
As an interesting aside, what happens to
this system if we throw out assumptions about marriages being
between a man and a woman? Consider two men:
Bruce ma-Tlaski Hedding marries
Lance ma-Frommel Muntz.
They can just take each others patronymics to replace their matronymics,
becoming:
Bruce sa-Muntz Hedding
Lance sa-Hedding Muntz
We can tell they are married, but there
is no matrilineal information present, since there are no females
in the marriage. There are quite a number of possibilities for
naming their children. One can picture extending this to group
marriages, with perhaps one surname denoting the group and two
others indicating the biological parents (if that is considered
important).
Anyway, this is a great example of an element of culture where
everyone just goes along doing it in the traditional way and most
people never think about it, but it can be interesting to dissect
and ponder improvements.
Last Modified: March, 2000
Patri
Friedman / patri@izzy.com