Like I have said elsewhere..., mutually exclusive categories are okay for depicting moral ideals (or standards) but are grossly inadequate for categorizing particular actions.
This is where Merlin's argument goes off the tracks. If you categorize an action so as to align it with a moral code, then it must be done according to the characteristics of the moral code and not some different system. Actions arise out of motivations and the motivations may well arise out of the person's moral beliefs. In this case, it is the moral belief that is needed to categorize the action.
Notice that in Joe's post he said, "It [the action] is chosen based on at most one of those standards." That is the heart of the issue. People choose, and many of these choices are made on the basis of their moral beliefs.
Maybe we know the motivation, as in when we create a hypothetical that implies or explicitly states the motive. "He wanted to spend the money on his own children, but he followed the beliefs of his church and gave the money to a fund to help other children." Or, we don't know that the motivation, as in the statement, "He stopped and took a few minutes to comfort a fellow worker that was going through a hard time." Why did he do that? Was it a "Christian duty" or was it that this person was important to him?
There is nothing wrong with a category system that might be labled as, "Actions as per who benefits" along with a set of rules that define how we parse things into the categories. Are the benefits restricted to financial benefits? If not, must they be material benefits, or can positive emotions count as a benefit?
There could be a different category system that we might describe as, "Actions as per who benefits categorized by the actors primary motivation." Now we would have the same kinds of questions as before ("What do we mean by 'benefits'?") but we also need a motivation. This means we need to know the value system that was in play to generate that motivation which generated that action. And this would take us to a parsing of moral belief to their respective moral codes. Only then could we say that an action was altruistic or egoistic. And because the moral systems have to be defined ahead of time, we would have a definition of what they each are. This takes us to those moral ideals or standards (not actions) that Rand and Comte and Merlin all understand and mention as being incompatible with each other.
Because the moral ideals (moral systems) are mutually exclusive, the actions that are categorized will end up being mutually exclusive. We might end up seeing some actions that are categorized as "altruistic and both benefited", but we will never see "altruistic and egoistic and both benefited." The person has a moral belief that in a clash between choices of selfishness and selflessness will resolve in favor of one or the other. We can't get past the fact that the moral belief is where the categorization of human actions has to come from and that as moral ideals altruism and egoism are mutually exclusive.
Moral beliefs come before the motivations we are interested in, and moral beliefs generate the kinds of motivation we are looking at. Human actions of the sort we are interested in can not exist apart from a motivation. The motivations exist before the action occurs. The assignment of a given moral nature to an action can't turn around and redefine the morality according to who benefits from the action while still maintaining that there can be incompatible moral standards and ideals.
Merlin's injection of "beneficiary" as a means of attaching the tag "altruistic" or "egoistic" to an action is a fallacy. That isn't how these tags have to be understood and chosen.