A school of red herring in a sea of bias

The vitriol spewing forth from the mouths of this year’s edition of authoritarian U.S. presidential candidates is a veritable school of red herring swimming in a vast sea of bias.

We have actual problems to solve that this gang is deliberately avoiding. There is a climate crisis, full stop. Our income inequality and level of social unrest rivals that of the Ancien Régime in 1788. Don’t even get me started on the problem of gun violence: I’m from Philadelphia. And the number of hate groups that have proliferated? Don’t get me going on that one either: I’m from Pennsylvania.

Hate groups feel empowered when politicians feed them rhetorically blood-red meat. And, we must be mindful this is not just a right-wing phenomenon. — in Philly, a city that bleeds blue, we have a Democratic mayoral candidate advocating stop-and-frisk. Instead of dealing with actual problems, hyper-cynical politicians indulge in culture-war, caring not that this feeds into the cycle of hate. If you want people to vote against their own self-interest, pander to their biases. Feed them plate after plate after plate of delicious red herring about immigrating Mexican rapists, drag-queens reading books to children in libraries, or a transgender girl playing sports for the sole purpose of achieving a hormonal advantage.

So, OK: what does this have to do with GEDCOM?

Bias

Social psychologists have a term for the pervasive hidden prejudices that every person carries with them: implicit bias. It is one of the less endearing aspects of the human mind, but we all carry these unseen biases around with us.

Implicit bias is manipulated by politicians using the aforementioned tropes. But it doesn’t stop there. It intrudes on almost every aspect of social decision-making. That includes design. And in the case of software design, GEDCOM is a particularly egregious example of bias by design.

What is GEDCOM?

For those who don’t know what GEDCOM is, it’s a database standard for genealogical data. This allows data to be imported to and exported from any genealogy software that supports the standard, which is almost all of them.

How does GECOM function as a standard? Let’s have a look under the hood:

This is the entry for Theodore James Cromar in the GEDCOM file I created, first in Ancestry, then in GRAMPS. This file goes on for line after line after dizzying line: 87,286 lines of data for 4967 individuals listed in Peter Cromar’s descendancy, an average of over 17 lines per person.

It doesn’t look very biased, does it?

Aside from sheer dullness, it’s not a difficult standard to grasp. You can quickly figure out that NAME is a name, GIVN is the “given name” portion of that name, SURN is the surname portion, SEX is gender, BIRT is birth, DATE is a date, PLAC is a place, DEAT is death, and so on. A three-to-four letter designation for every kind of datapoint one would care to add for an individual in the database. Some of it gets a bit obscure: the FAMS @F1234@ designation assigns a person to a family group, for example, and this is auto-generated. But lurking in those all-caps three-to-four letter data designations is bias.

Sex versus gender

Let’s start with sex.

That comes off a bit forward doesn’t it? Let me rephrase: let’s discuss the difference between sex and gender. I and others have written before about general gender bias in genealogy, but GEDCOM hard-wires the phenomenon by equating sex with gender. If you are attempting to record a gay person who has a same-sex partner, or a transgender person doing just about anything, you’re out of luck. This conflation of sex with gender creates a high-contrast, low-resolution binary in GEDCOM: a male husband may only have a female wife.

Now, in my direct family, we have gay partnerships, we have transgender individuals (including my son), and some persons with a bisexual orientation. I’m willing to bet your family does too. In the most recent Gallup poll, LGBT identification in the US has plateaued for the past 2 years at just over 7%. The chance that you only find cisgender individuals in your family is vanishingly small: one out of every 14 persons on average will not conform to the binary standard imposed by GEDCOM. It’s a standard that cannot correctly identify nearly 360 persons in Peter Cromar’s descendancy! Given the simplicity of the GEDCOM standard, how difficult would it be to introduce a four letter designation like SPOU for the gender-neutral “spouse” as just one example?

Cultural bias

As you dig deeper into GEDCOM, note the proliferation of Judeo-Christian rituals such as christenings, baptisms, bar (but not bat, with gender bias rearing its head again) mitzvahs, circumcisions, confirmations, first communions, excommunications (wowzers!), ordinations, marriage banns… the list goes on.

I’m not Muslim, but I’m willing to bet a once-in-a-lifetime hajj is a ritual one would want to indicate. If I’m Buddhist, would I want a way to note dharma transmission? In our descendancy chart, I was surprised to see a Cromar branch in India, emerging out of Ron Cromar’s line. I’ve wondered what kind of traditions they might follow that GEDCOM does not recognize.

But it’s even more insidious. Language, we know, is one of the strongest manifestations of culture. And when one examines the languages supported in a variety of genealogy softwares, it’s surprising how limited options beyond English can be. According to Tom MacWright, developer of the gedcom parser that allows one to create an interesting force-directed interactive out of a GEDCOM file, the language bias in GEDCOM is structural:

Though it is switching to Unicode, it previously adopted ANSEL, a character set that had no mechanism for representing names in non-Western scripts.

Tom MacWright

Roots of GEDCOM

Tom gets even more impatient with GEDCOM when he writes:

GEDCOM continues to be developed by the Church of Jesus Christ of Latter-day Saints, which has bigoted views of homosexuality, so it’s unlikely that these things will change.

Tom MacWright

He’s not wrong when he points to the bigoted perspective of the LDS movement when it comes to gender non-conformity. However, in a back-handed defense of LDS ideology, I’ll hasten to point out this is not the only Western religious tradition that is obsessed with this. And it seems to be a shibboleth that infects most major world religions. Having said that, it should come as no surprise that GEDCOM’s religious roots should encode its hostility toward diversity.

Is there any way out?

I’ve explored with great interest the alternative created by Gramps, known as GrampsXML (extended markup language, kind of akin to HTML). Tom mentions it in his critique of GEDCOM, and in the Gramps project wiki, there is a discussion comparing the two standards. But, like opportunistic politicians fanning the flames of historical biases to gather votes, it’s hard to get rid of a standard once it becomes a standard. GEDCOM rules the waves.

The Ron vs. Don show may the most obvious and blatant manifestation of bias our culture has to offer at the moment, but it’s not the only one, nor the most dangerous. Designed objects like GEDCOM are more subtle and therefore more insidious. Such a small, hidden thing can keep the flame of hatred of the Other alive. Genealogy really needs to pay attention!