I'd just like to jump in here and try to explain the Lakoff portion of these comments a little. First, let me say that Lakoff stresses the imaginative nature of human thought. Rather than being strictly logical and literal, concepts tend to be metaphorical. This is very important because it means that oftentimes meaning is determined not by a dictionary definition, but by how you frame a given situation and what metaphors you choose. With that in mind let's turn to morality.
Lakoff believes that there are basically two competing moral systems at work in society today (with many variations). "Strict Father" Morality and "Nurturant Parent" Morality. The significance of these systems is that they rank the various moral metaphors as being more or less important.
Strict Father Morality values very highly the metaphor of Moral Strength. The mapping is as follows:
Physical Strength -> Moral Stength
Physical Weakness -> Moral Weakness
There are also entailments of this metaphor. Note that physical strength must be built up. So, according to the metaphor, humans must "build up" moral strength lest they be morally weak. Because of this entailment, children (and adults) are viewed as inherently "weak" and must be strictly disciplined from an early age if they're ever to become "strong". There's a a folk version of Behaviorism that goes along with this that says that people are purely motivated my gain and deterred by loss. Combining these two concepts presents a view of humanitiy that is innately selfish. People seek to satisfy their desires and unless they have built up moral strength in the form of self-discipline, they will step on others to do so.
Nurturant Parent Morality values different metaphors, and presents an alternate view of humanity. The primary metaphor of Nurturant Parent Morality is Moral Nurturance:
Physical Nurturance -> Moral Nurturance
Physical Well-Being -> Moral Well-Being
In this metaphor, humans become moral beings by receiving love and nurturance and learning to love and nurture others (rather than by acquiring self-discipline). People are not seen as being inherently self-interested in this model but rather as beings that are capable of providing moral nurturance and receiving it. Indeed, following your own "self-interest" in this model would entail that you automatically seek to maximize the well-being of those around you -- for you cannot maximize your well-being unless those who nurture you have maximized theirs as well. As you can see, this conceptualization presents a radically different view of what it means to be selfish.
I'm not sure how much all this relates directly to what was being discussed above but I can say that any abstract concept (moral or otherwise) such as selfishness is not so cut and dry. The meanings of the words involved deeply depend on the underlying conceptual structure and framing methodology that you use.
As a side note, perhaps the most damning evidence against Behaviorism (and the notion that people are motivated primarily by reward and punishment) is the work of Kahneman and Tversky. The above study, as well as many others they've done, confirms that humans simply do not consider gains and losses logically. How can one act based on self-interest when one is often not aware what one's self-interest is?