Using Semantics to Maximize Utility

I thought I'd heard all of the basic arguments against utilitarianism, but reading Tom Regan's The Case for Animal Rights, I came across the most bizarre one I've seen yet*. He claims that there's an inconsistency between the principle of utility ("maximize utility") and the utilitarian principle of equality ("count equal utilities of different individuals equally"). He writes:

For the utility of counting A's and B's interests as equal can vary from case to case, even if their interests themselves do not. Thus, if utility is our guide, we are permitted to count the same interests as equal in one case and as unequal in another.

If I'm reading Regan right, what he's proposing is -- while perhaps technically feasible -- truly bizarre. To make an analogy, it's like saying that it's inconsistent to say both "buy the biggest house you can find" and "count equal numbers of square feet in different houses equally." So sure, you could end up buying a bungalow while claiming to have bought the biggest house by declaring that ten square feet of bungalow is worth 100 square feet of mansion. But that seems like using a definitional move to evsicerate the substance of the actual goal.

There is a certain plausibility to counting equal things unequally if the measurement in question does not completely exhaust the meaning of the goal. So in the house example, if by "bigness" we mean a feeling of spaciousness, then we might be justified in counting equal square footages unequally, since square footage is not the only thing that contributes to a feeling of spaciousness -- a house with light-colored walls, for example, will feel more spacious than an identical house with dark walls. One may be forgiven for making this kind of conclusion in the case of utilitarianism as Regan defines it, since he draws a verbal distinction between the "utility" that is being maximized and the "interests" that are being counted equally. However, in utilitarianism, the thing that is being counted equally between individuals and the thing that is to be maximized (regardless of what we choose to call them) are by definition identical.

An alternate reading of Regan's passage is that his complaint is simply that if A wants one thing and B does not, we might sometimes give A what he wants but other times give B what she wants because of the preferences of additional individuals (hence the overall utility of satisfying A vs B changes between the situations even though A and B themselves remain the same). For example, say in the first case C is also affected by our action (and his preferences align with A's), while in the second case C is absent but D is affected (and her preferences align with B's). It is true that utilitarianism would advocate such an "inconsistency" in treatment between A and B. But I fail to see how that violates any moral principle or even intuition. Nobody would say it's unfair that while my dad voted for Rick Santorum in his last two elections but I voted for the Democratic challenger both times, in 2000 my dad got his way but in 2006 I got my way** simply due to the fact that there were more other Santorum-supporters in the picture in 2000. The voting example should further illustrate that even when a different side wins, it's not because any individual's interests have been counted differently the second time. It's a matter of having a different total set of people whose interests must be balanced.

* Which is perhaps not surprising, since if it was any good, lots of other people would have repeated it in the two decades since Regan's book was published.
** I actually voted in Arizona this year, but this is what would have happened if I'd stayed in Pennsylvania.