If there's some perceived benefit in legal recognition of a romantic union, then there's no reason to exclude same-sex unions.
Basically, there are a few consistent positions:
1. Children are important. Outcome: either incentivize children, or incentivize fertile couples.
2. Romantic unions are important. Outcome: legally recognize all of them, including same-sex couples.
3. Freedom and small government is important. Outcome: stop recognizing marriage and let people handle it privately.
4. Religion is important. Outcome: legally recognize marriage under something like the rules of your religion.
We can argue for and against the merits of these individual possibilities, but what I can't understand is when someone uses an argument for #1 (as the other commenter here is) or #2 (as the site being discussed is) but somehow ends up with the outcome from #4.
Well, that is what the other commenter is arguing that we already do, at least in theory. The laws are from a different time, so there are many modern-day loopholes for sure, but when you look at the conditions at the time when the laws were enacted marriage would be essentially limited to those who could (to the best of their knowledge) and were willing to have children, intended to create an environment to make having children amiable.
Even divorce law is predicated on the idea that the spouse will stay home to take care of the children. Since that leaves one member of the household without an income for many years, in the event of a dissolution of the partnership one member will be left out in the cold, so to speak, so we created rules to share the proceeds gained in the process to ensure that each party is compensated for their efforts.
I think we can find all kinds of arguments as to why that is flawed in 2017, but if we just stick to the assumption that the subsidy provides exactly the expected outcome: Why does it need to change at all?
But, since nobody seems to be incorporating that data in the discussion, I'm not sure we can really fault the other commenter for believing that we are getting the desired results out of the existing subsidies. The results haven't shown up as a clear negative with respect to the intended outcome, so it is natural to assume, without better information, that it is working just fine. And, as you said, if it is working just fine there is no need for change.
Absolutely, more data is needed in this discussion to evaluate the effectiveness of what we have and to compare against alternatives. But to outright dismiss the unmeasured effectiveness of a subsidy because a handful of people have proven the willingness to do the same without that subsidy seems to be the only silly part of this discussion.
To be clear, I'm not sure it is an effective subsidy to begin with, but that's a moot point as far as this discussion goes.