Last week I wrote a post in which I suggested I was going to explore further the nature of the obligation that Catholics have to "vote pro-life." Assuming that we do have such an obligation, what does that mean? Let's review:
Some think the obligation is only that we cannot prefer the pro-choice position. We may (reluctantly) vote for Pro-Choice Charlie over Pro-Life Louie if Charlie has a better stance on other issues and if we think those issues are more important. This is the answer which, if correct, restricts our voting behavior the least.
Others say, no, we are obligated to consider life issues as most important, and we may not vote for any pro-choice candidates over any viable pro-life candidates: if Pro-Choice Charlie runs against Pro-Life Louie, we may not vote for Charlie. However, if there are no pro-life candidates who have a chance of winning -- if we have to choose between Pro-Choice Charlie and Pro-Choice Chester -- we may vote for Charlie if his stance on other issues is better. This form of the negative obligation would be a bit more restrictive than the first.
A third interpretation would be that we commit actual sin if we cast a vote for any supporter of legal abortion at all. If there are no viable pro-life candidates, the only moral choice is to withhold our vote or "throw it away" on a non-viable pro-life candidate. If Charlie is on the ballot running against Chester, then we are stuck with writing in Louie or else not voting. This negative obligation is still more restrictive, as it strikes some candidates permanently off our potential list of choices.
And a fourth interpretation is that we must orchestrate our support so as to maximize the possibility that a pro-life candidate will be elected in the general election. This is a different, positive obligation, and also quite restrictive. (In the case that no pro-life candidate can possibly be elected, the positive obligation disappears and you're left with one of the two latter negative obligations.)
I think it is instructive to consider exactly how our voting behavior is constrained under each of the four interpretations of the obligations of a Catholic voter. So I want to consider a stripped-down model of a race for one office: two election days, two parties, two issues.
Here are the rules:
There are two political parties. Call them the Canine Party (C) and the Feline Party (F). Each party holds a primary election to decide which candidate will win the nomination for the general election. A voter may vote in either the Feline primary or the Canine primary but not both, and then of course can go on to vote in the general election.
There are two political issues. In principle these could be almost any two features that influence voters to pick one candidate over the other -- you might think of them as two quite specific policy questions ("Shall we promote single-payer health care?" "Shall we build a wall across the Mexican border?"), or you might think of them as aggregations of related policy questions ("The Economy" and "Social Issues"), or you might think of them as features of the candidates themselves ("Policies" and "Character"). For the purposes of setting up the model, though, they are simply two generic political issues: Issue 1 and Issue 2. Each issue has two possible policy positions, which we will call Yes or No.
In the model, party affiliation and policy positions are the only meaningful distinctions among the candidates. It just so happens that all possible unique candidates happen to be running in this race. So four candidates will compete in the Canine party, representing the four policy positions:
Yes on 1, Yes on 2 (candidate yyC)
Yes on 1, No on 2 (candidate ynC)
No on 1, Yes on 2 (candidate nyC)
No on 1, No on 2 (candidate nnC)
And, of course, there are four candidates competing in the Feline party, which four are yyF, ynF, nyF, nnF. Eight candidates. The winner of each primary competes in the general election. That's sixteen possible head-to-head pairings in the general election.
(My more smartypants readers have probably already noticed that this model can be generalized to the case of p political parties and n political issues, each with m possible policy positions. You can even generalize it to a multi-level model election that will have little model people setting up "November Madness" brackets and placing bets around little model water coolers. Go ahead and set up the equations if you like. I'm not going to.)
Now, let's imagine a rational voter: Alice. Alice doesn't need anyone to tell her how to vote in the general election, because she's so very rational that she could tell you not only who are her favorite candidates, but she can rank all eight in order of preference, and so given any of the possible sixteen head-to-head combinations, Alice knows who will get her vote. This is because Alice follows a very rational algorithm to choose between two candidates:
(1) Always choose "Yes on 1" over "No on 1."
(2) If the candidates have the same position on Issue 1, choose "Yes on 2" over "No on 2."
(3) If the candidates have the same position on Issues 1 and 2, choose Felines over Canines.
Clearly Alice's favorite candidate is yyF and her least favorite is nnC, but Alice is not completely specified until the three features (issue 1, issue 2, and party) are ranked in order of importance. Her friend Bob also loves candidate yyF and hates candidate nnC, but his voting behavior is different because he always chooses a Feline over a Canine first, then moves on to consider Issue 1, and finally Issue 2. Mathematically attentive but non-smartypants readers may have noticed that there are 2 x 2 x 2 x 6 = 48 different kinds of rational voters in this election. (My smartypants readers, working out the general case, are deep in matrix algebra by now.)
This model offers us a lot of ways to play. We could, for example, populate the electorate with so many rational voters like Alice, so many like Bob, so many like Carol, and so on. We could sprinkle in a bunch of irrational voters, too, who don't follow any discernible rules in choosing one candidate or another. We could pretend that we had poll results that allow us to predict the outcomes of the sixteen hypothetical general-election face-offs. And then -- we can ask ourselves interesting questions, like: "Given her preferences and this set of predicted general-election results, whom should Alice support in the primary election?"
I haven't brought any Catholicism into the model yet -- I've just tried to set up a simplified structure for thinking about it, one that perhaps can be enlarged upon later. More on that in a subsequent post, in which we consider the constraints introduced by the various interpretations of the "Catholic voter's obligations."