This short article is motivated by what I see as a confusion of thinking displayed in the conversation between Tom Stoppard and David Sloan Wilson in a May 22, 2015 article in The Guardian online, available here discussing issues from Stoppard's play The Hard Problem.
Dawkins' description of what we think of as the evolutionary "behaviour" of genes as "selfishness" grabbed our attention. In doing so it dragged vernacular culture some way toward thinking of ourselves and our motivations as expressions of natural processes. Unfortunately, by using the metaphor of "selfishness" he, probably without intending, invited us to import into our thinking about genes other aspects of the metaphor: a selfish entity, say a gene, is also conceived as possessing or expressing self-interested, motivated purpose.
I strongly doubt that Dawkins thinks of genes in that way at all. His metaphor should not, in my understanding, be read as including Machiavellian plotting at the genetic (or epigenetic) level.
Also potentially implied if the metaphor is taken in this way, is that the purported "selfishness" at the genetic level somehow bubbles up, through the phenotype, to the level of more-or-less conscious intentions of individuals. By this thinking, someone behaving altruistically is "really" being hypocritical because his/her "real" (i.e. genetic) intentions are towards optimising survival of (aspects of) his/her genotype, not tpwards the benefit of the object of their altruism.
I also strongly doubt that Dawkins thinks of, or intends us to think of the relationship of genes to human subjectivity in this way.
Genes, whether individually or as a whole or partial genotype, do not, indeed cannot, intend anything. It is only at the level of peri-consciousness in an individual animal or a group that "intention" can usefully be spoken of. Genes as entities are entirely passive, participating in physiological processes which are driven by physico-chemical properties of the complex molecules involved. The phenotype which they cooperate to produce is then subject to the slings, arrows and bouquets of fortune in their environment at the time and is either successful or not in raising grandchildren in that environment. If unsuccessful, the genotype's chances of propagating into the future decrease; if s/he is successful, the genotype and its component genes continues into the next generations. How far it/they survive into the future is entirely contingent on the interaction of successor phenotypes and the environment, physical, ecological and socio-cultural, in which they find themselves.
At this level of discussion, terms which imply some level of conscious subjectivity such as "intention", "altruism", "selfishness" will, if uncritically interpreted, result in more confusion than clarity.
When we look at the level of our experience and our ability to evaluate ("assess the value of") alternative choices of behaviour it becomes useful to think in terms of, for example, "intention", "altruism", "selfishness". At this level, we most definitely are able to be altruistic un-hypocritically: we are doing what we think/feel is right. The possibility that we are in some way enhancing the prospect of some aspect of our genetic heritage surviving into the future is irrelevant even if we are aware of the model in which this is thought likely. There is no direct functional link between the supposed future survival of some aspect of our genotype and our immediate experience and assignment of value.
To think, for example, that altruism is cheapened by the likelihood that it is a pro-survival characteristic for our group in evolutionary models is in my view to unnecessarily erode our feelings of self-worth, and those of our culture, on the basis of a blatant category error.
For people who don't like scare-quotes, the liberal use of them in this article is an attempt to emphasise that these quoted terms in particular (like all others) derive from models of the world which may be more or less useful to thinking about particular aspects of the world in particular contexts, but often make sense ONLY within the model from which they derive. The meta-point of this article is that many terms in conversations about genes, selfishness, altruism and other aspects of evolution, as it applies to the felt reality of being human, are commonly transplanted between models. When this is done naively, presuming that the term's meaning stays the same across models, that transplantation imports entities into their new host-model which do not make sense in the new model. "Selfishness" here being the case in point (Extrapolation based on, among others' thinking, Lakoff and Johnson, (1999) Philosophy In The Flesh, Basic Books)