Despite the intent in biology to eliminate group selection, it invariably turns up as the only reasonable explanation for the cohesion of species and the behavior of large groups of animals.

Additionally, the struggle to explain altruism using kin selection and inclusive fitness is perpetually haunted by the requirement that the entire premise hinges on the existence of an actual "altruism" gene.  In other words, Hamilton's rule and inclusive fitness are meaningless if there isn't a genetic component to cooperation and altruism.

In that respect, can we be reasonably confident that genes are involved?  If we simply consider how cooperation works and examine it purely from a behavioral [not cognitive] approach, we can see cooperation/altruism (1) occurs in organisms that don't even possess a nervous system.  Therefore concepts like cooperation and altruism, are not created by the existence of cognitive processes.  However, this fails to address whether such behavior is actually genetic.

To approach this problem we need to determine whether cooperation/altruism is a trait that is actually selectable.  If so, what is the heritability of such cooperation/altruism?  

Even the cells of a multi-celled organism require a high level of cooperation and altruism to form the representative creature.  As such, one must postulate a means by which such a characteristic is passed on from generation to generation.  One can also easily imagine that as cooperation becomes a staple of the group, then those individuals that lack such a trait or characteristic will tend to either be expelled or attempt to compete independently.  Given the lower total energy costs of cooperation, it follows that the advantage cooperation provides ensures that cooperative groups tend to thrive.

This certainly wouldn't be true for all organisms, since the total energy costs of fitness would vary, so that there might be many creatures for which there is no intrinsic advantage to belonging to a group.  Similarly, other adaptations turn fitness into a pure numbers game, where by simply creating tremendously large numbers of viable offspring, sheer probability assures that some number will likely survive.

In these cases, then cooperation may actually be a liability and consequently one wouldn't expect to see it persist in a population.

However, what is more interesting is that the arguments in favor of inclusive fitness require groups.  It makes no sense to talk of altruism among different animal species, just as it makes no sense to consider cooperation as evolutionarily stable between separate species.

That this may occur isn't relevant, since it conveys neither advantage nor disadvantage on the practitioner unless such encounters produce a conflict of interest [which may often occur when altruistic or cooperative behavior changes due to one of the creatures reaching sexual maturity].

In effect, the gene-centric view in biology was a direct attempt to stop the persistence of group selection theories that suggested that evolution occurred "for the good of the group".  Specifically, no one could offer an explanation for how traits that were carried by individuals could possibly translate into group-level selection, so the idea was dismissed as mere wishful thinking.

Yet, we find that on examination, inclusive fitness is precisely the theory necessary to explain group selection.  While many may balk at such a conclusion, one must remember that the essence of the altruistic explanation is that the individual does NOT benefit.  Therefore the relevant question is to ask who does benefit?  Previously the simple answer was; the genes.  Yet, this is unsatisfactory since it is obvious that 23,000 + genes are not actively competing for propagation.  In addition, we know that survival of any species is enhanced with greater genetic diversity, not less.  Therefore the problem that evolution "solved" was how to maintain a faithful enough transfer of information to ensure sufficient uniformity so that reproduction could be maintained, but not so uniform as to risk the liabilities of amplifying negative traits.

Certainly the argument has been made that blood-related kin can benefit by passing on such genes, in kin selection theory.  Yet, this is inadequate because overwhelming almost no species is capable of recognizing kin (2), and other than the eusocial insects, acts of altruism and cooperation still occur with regularity among unrelated individuals.  In addition, many offspring have no siblings [related kin] so this fails to explain why social behavior should have developed.

Inclusive fitness extends the arguments of kin selection, by claiming that individuals within the same group will then benefit by such behaviors.  Yet, isn't this precisely what group selection claims?

Doesn't this render the altruistic act of an unrelated individual being made precisely for "the good of the group"?  Moreover, it is well known that without a group [sufficient number of breeding pairs] extinction is all but inevitable.  

Part of the problem is that Hamilton's equation is often tossed around as "evidence" of inclusive fitness explaining how an altruistic gene can persist in a population (3).

r * b > c

Where r (relatedness) times b (benefit) is greater than the c (cost).  So, the benefits must be greater than the cost where b is adjusted by the probability (r) of the relative carrying the genes [i.e. the degree of relatedness].

Yet, on examination, this begs the question, since there is NO instance of where there can ever be a benefit greater than simply not making the sacrifice.  In other words, if the altruist doesn't perform the act, then their probability of passing on their traits is 100%, compared to some lesser probability if a relative were to reproduce.

Using Haldane's famous quote about saving two drowning brothers, the statement is problematic.  In the first place, both brothers would need to be drowning to balance the equation, which is already somewhat more unlikely.  However, there is no argument that can be made for saving the brothers that doesn't produce a greater benefit to the initial actor than simply standing on the shore.  In other words, the altruistic act invariably falls short of propagating the desired genes regardless of the degree of relatedness.

More importantly, one doesn't observe altruistic behavior that fits into Haldane's quote.  An altruist will likely risk sacrificing themselves for considerably less lofty results, so the postulate granting such a gene becoming pervasive based only on individual genetic success is unsatisfactory.

The only way to make sense of this situation is if the benefit [fitness benefit] is assigned to the group rather than the individual.  In that situation, if the group survives, then it will provide benefit for considerably more members than the singular act addresses, and therefore any altruist's cost will always make sense within that context. 

On the other hand, if the cost is zero, then ALL actions are justified since any action, regardless of how small, is beneficial.  This is also an argument that makes sense if one views the cost as merely a risk-taking venture, where the fitness cost is not assured, but merely a probability.  In those cases, then any success will automatically produce a benefit with the cost being zero.

What makes this important is that inclusive fitness, as it is currently defined, does not adequately account for those situations in which it is not beneficial to incur a cost.  Despite popular belief, it is often better for a parent to sacrifice their children than to risk themselves.  Since the parent can produce more children, but the children may not survive without the parent, to behave altruistically places both at risk, and has a slim probability of success.  Therefore, Hamilton's Rule is necessarily restricted to those instances of where there is no dependency between the individuals if one is not capable of reproducing.

As a result, this also turns the altruism attributed to eusocial insects on its head, since there is no benefit in an insect sacrificing itself for another sterile insect.  Therefore the only context that makes sense is group selection.  The sacrifice is beneficial because it conveys benefit to the group.

In fact, even if we were to argue from the gene-centric position that somehow the objective of the gene is to be propagated into future generations, there is little benefit in a gene supporting an individual versus supporting the group.  There is far greater benefit (to the gene) to be conserved within the group, rather than struggling to be passed on to a single individual. 

If anything, the gene-centric view, might offer some insight into how novel genes are introduced into a population, but it contributes nothing to the notion of conserved genes already present within a population.  After all, if they are already pervasive throughout the population, then what's the point in evolving altruism, since every member of the group already carries the gene?

Instead, we have to consider that this explanation is insufficient for anything except genes that haven't gone to fixation within a population. 

In short, the most likely explanation is a combination of all these ideas, to varying degrees, so that at individual levels we find cooperation producing economies of scale that render it beneficial, coupled with altruistic behaviors that promote the survival of the group.  Once group survival is assured, then competition between individuals within groups becomes possible.   In many cases, the development of cooperate strategies can "bypass" phenotypic developments by providing a group level benefit that is beyond the individuals themselves.   It is clear that the gene is insufficient to account for all the traits and criteria for successful survival of individuals, therefore one must consider that the solution exists within the myriad number of groups that constitute life.

The greater the benefit of cooperation, the greater the dependency of individuals to group survival.  The individual cells of a multi-celled organism have completely become committed to "group" survival, while other organisms may experience varying degrees of such a dependency.  So, while the gene is certainly a crucial element for conveying information into the next generation at the individual level, it is not the only or necessarily the most significant factor shaping the next generation in an evolving population.

Additional reading:

(1) Admittedly one of the difficulties occurs in defining "altruism" itself.  The conventional definition is "any act that increases the fitness of an organism while decreasing the fitness of the actor".  A portion of this post is to refute whether such a definition is actually accurate.

(2) Interestingly enough, there is some suggestion that while organisms may not recognize kin, their microbiota may be capable of such recognition, or at least recognition of similar microbes.

(3) In a previous article I noted that altruism didn't appear likely to be related specifically to fitness in humans anyway. 

However, there is a more fundamental problem with the idea of an altruistic gene.  How does an altruist become the beneficiary of an altruist?  The concept is contradictory, since the nature of the general view of altruism is that all members are self-sacrificing.  Therefore, the logic conclusion of a population filled with altruists, is a suicidal group.

It is for this reason that I argue that altruism is a risk-taking behavior with the beneficiary being the individual, but primarily the group at large.  This ensures that the "altruist" is not merely a self-sacrificing individual [since such individuals are destructive to the group], but rather someone that can take a risk versus a huge benefit.