Some more on this never-ending debate...
In ascribing value to certain things, we acknowledge that they matter to us -- they have weight in our calculations about what we should do. But it seems to me that there are two (not mutually exclusive) ways in which we do this. The first is the way that we mean when we talk about matters of "taste." When we evaluate things in this way, the account of why we value them is autobiographical. For example, I like coffee because it tastes good to me; it makes me feel cheerful and alert; it helps me focus on tasks that I want to perform.
In order to dispute my evaluation, someone would need to similarly couch their objections in features about me. For example, I might announce that I want to eat a cheeseburger, because I like the way they taste. In order to argue with my choice, someone would need to alert me to some feature about my tastes that they believed I was failing to acknowledge. For example, they might point out that the last time I ate a cheeseburger, I felt sick, and told everyone how I never wanted to eat a cheeseburger again (this is, of course, a fictional story; cheeseburgers are delicious). The point is that with regard to discussing matters involving my own tastes, the entire focus is on me, the valuer.
The other way that people often ascribe value to things is to claim (or implicitly claim) that they are the appropriate objects of valuation, and to ascribe to them the value -- the weight in our moral calculations -- that we believe is fitting of their nature or properties. Not only are these things valued, but they are seen as valuable: to "fail" to value these things would be, in some sense, inappropriate, unbecoming, or wrong. The reason given here for an evaluation is no longer autobiographical, but cites some quality inherent in the valued object as if it were the explanation for the evaluation. For example, a parent feeds his child "because" she is hungry, and her health and wellbeing depends on being fed. The implication here is that the parent thinks that to fail to feed the child would demonstrate insufficient consideration of the fact of the child's hunger. Clearly, this sort of valuation is inherently normative: it is understood as involving an acknowledgment of how we ought to act.
When I talk about actions that are "self-interested," I refer to those actions where the account of why the action was undertaken makes reference to the first kind of valuation (the one which identifies the tastes, wants, and desires of the valuer as the explanation for the valuation). When I call an action "non-self-interested," I refer to actions taken for reasons understood in terms of the second sort of valuation (where objective features of a set of circumstances are cited to explain the valuation). As I suggested before, these sorts of valuations need not be mutually exclusive. But there is a sense in which they are taken as being separate from each other. For example, we might find someone saying "He's a really nice guy, and he's never done anything bad to me, but I just don't like hanging out with him." The implication here is that niceness and blamelessness are being seen as somehow worthy or demanding of consideration and positive valuation, but that the negative valuation conveyed by the person's "tastes" is acting as a counterweight. Now, if the person who made the above statement decided to hang out with the guy who she discussed, we might imagine her saying, "He was just so nice; when he asked, I had to say yes." Her action here would be non-self-interested. If, on the other hand, she decided not to hang out with the guy, we might imagine her saying, "As nice as he was, I just couldn't bring myself to put myself through spending another minute with him." Here, her action would be self-interested.
When I talk about something having "intrinsic value," I'm making reference to qualities or objects identified by the normative kind of valuation as providing justification for a choice. Most ethical theories (those with a "realist" component) rely on this sort of valuation to provide a basis for their claims: ethics involves the acknowledgment or recognition of the appropriate responses to certain features of things or situations. (I should mention, in passing, that one need not claim that there actually are appropriate responses or anything like that; see, for example, quasi-realism.)
It's important to distinguish here between intrinsic value and what might be called "objective value" or the "objective component" of value. When we talk about objective values, we mean that a thing can be valuable independently of anyone valuing it. This does not seem possible to me, as value seems to be, by nature, a relationship between an evaluating mind and some object (I don't mean physical object, just a "something" which is being valued). If I'm correct about this, there could be no such thing as objective value. But the concept of an objective component of value is different. This idea relies on the possibility that certain features of objects (especially where objects can be taken to also refer to concepts in our minds) cause a reaction in us which leads us to value those objects. For example, a guy might meet a beautiful, intelligent, and entertaining woman at a bar, and come to want certain things not because he chose to want them, or something more mysterious, but rather because it was natural for him to want those things upon acknowledging certain features of the woman. Her objective features caused a reaction in the man, whose value judgments would thereby be affected (though, of course, the valuation still only exists in the man's mind; it is still subjective).
It might be immediately clear that a parallel exists between the concept of an objective component of value and the concept of intrinsic value, as I alluded to in identifying the realist component of many ethical theories. But the two concepts are not necessarily identical. The man in the bar might desire a particular beer, for example, because he wants a drink and because it strikes him as satisfying the criteria of being a drink. But it would be slightly odd for him to attribute any intrinsic value to the idea of him having a drink: He would need to say, "I want that beer, and therefore it would be immoral of me to not go get it." Certainly there are people who would be comfortable saying this (anyone who would call themselves an "egoist" or a "hedonist" would immediately come to mind). But I take it that most people don't think that way. There is a difference between what we want (what is a matter of our tastes) and what we ought to want (what is a result of our attributions of intrinsic value).
There's clearly a lot more to say on this, but I figure this will do for now.