It may be wondered why I am so confident that genius is group-selected (apart from the fact that Michael A Woodley persuaded me!) considering that the status of group selection remains controversial within evolutionary biology.
That groups selection remains controversial is largely, perhaps, the result of the influence of Richard Dawkins's 1976 book The Selfish Gene; which hugely popularised the kin selection/ inclusive fitness ideas of WD Hamilton - which were presented as if they refuted the case for group selection; while omitting that WD Hamilton was himself a lifelong believer in the reality of group selection (as may be seen from his three volume collection Narrow Roads of Gene Land).
Note added: What is the essence of group selection of genius?
It is simply that genius is essentially for the benefit of the group and not for the benefit of the genius himself.
In terms of natural selection, this means the work of genius is 'for' the genetic benefit of the group; in broader terms this means the works of genius are 'for' general social benefit (therefore, not for the specific private benefit of the genius himself).
The basic reason for entertaining the hypothesis of group selection for genius, is that on average genius has poor reproductive success - with an average of less than two offspring per genius from most estimates; indeed many geniuses seem to have had no children at all (Newton, Beethoven) - and especially among the relatively few women geniuses (e.g. Jane Austen, George Eliot, Emily Dickenson).
Certainly, there is near-zero evidence of enhanced reproductive success for geniuses - indeed this is extremely implausible given that geniuses invest far less effort and resources into social and sexual goals than average people do - due to their investing so much more effort and resources into the subject matter of their genius.
Therefore, the 'standard' mechanism of natural selection is apparently not the cause of geniuses, and we must look at indirect mechanisms.
The theoretical problem with group selection is that if an allele occurs by chance which reduces differential reproductive success (RS) in direct descendants while enhancing the success of the group overall, then this mutation will tend to go extinct over not many generations.
This is true enough, so far as it goes; but does not eliminate the possibility that a population may produce an enhanced number of geniuses by group selection for at least several generations - if there is some mechanism to over-compensate for the reduced differential RS of genius - i.e. the decline in the proportion of the 'genius alleles' in the population may be overcompensated by a mechanism leading to an increase in the absolute number (not proportion) of genius alleles.
The answer is that a reduced proportion of 'genius alleles' in the group can be over-compensated by the massive effect of even one world-historical genius on the group as a whole - and more so by the presence of several major geniuses.
So if the presence of a genius results in the expansion of the population of his group (or arrests the decline of a sufficiently large group), then genius alleles might survive in increased numbers for several generations.
What seems to have happened in England, for example, is that during the Medieval period, natural selection acted to increase the average intelligence of the society - which had the effect of increasing the capability of the geniuses - so that their work became of world-historic (rather than merely local) importance.
The initial effect, from about the seventeenth century, seems to have been an increase in the proportion and number of geniuses, leading to a strengthening and expansion of the kind of society which had led (by various means) to that initial high proportion of geniuses.
English geniuses began to accumulate so quickly, and make so many innovations in agriculture and industry, that the English population expanded very rapidly; and the English 'group' began to spread around the world to India, North America, Australasia, South Africa, and so on.
In other words, the genius alleles in England were probably a relatively-slowly dwindling proportion of a very-rapidly expanding population - thus genius alleles were (for several generations) increasing in number, even as they reduced in proportion. Genius preserved itself, for a while; even though the genius alleles led to on-average reduced reproductive success.
The reason why genius may confidently be assumed to be potentially group selected is therefore that the cultural impact of a genius upon the group is so vast that it dwarfs any potential genetic influence.
So one single genius who provides a breakthrough that massively enhances something like food production, military technology, building, the cohesive benefits of religion, effectiveness of societal organization, price or swiftness or trade or transportation, or economic efficiency... can enormously expand the size and domination of this genius's group - such that the rising tide (of culture) floats all (genetic) boats (at least for a few generations).
In other words, the genius's large 'group' of cultural partners will experience a tremendous boost to their reproductive success - thereby expanding the same kind of society which (through whatever combination of genes and culture and whatever else) produced, sustained and recognized the genius in the first place.
In sum a genius will tend to enhance the genius-friendly society which produced him.
On top of this, it may be that the conditions which favoured genius in the first place, do so because (in that particular, contingent, time and place) groups can survive and grow only by creative innovation. Clearly this has not been the usual or typical case in the history of most places; but it certainly looks as if it may have been the case in and around Europe throughout the Middle Ages.
Since the personality trait of Psychoticism seems to provide the creativity aspect of genius; then all that is required was for natural selection to lead to a positively-skewed and sexually differentiated distribution of Psychoticism trait - so that the population would contain a small (but vital) proportion of High Psychoticism Males - as the specialist creative innovators.
This would explain the distribution of High Psychoticism - in a minority because the majority are needed for reproduction; and in males, because there is a relative surplus of males, and through most of history the majority of reproduction has been done by a small minority of optimal fitness males selected by a strong selective sieve of high mortality rates, and validated by intensive male-versus-male competition for mates and matings.
On the other hand, what might be termed a bad, or destructive genius - will also have a massively damaging effect to the culture which produced him. The work of a negative genius will tends to damage the size, effectiveness, efficiency, reproduction, cohesion or some other useful attribute of the group which produced him.
Therefore, when a genius is bad, the amplification effect works in the opposite direction to wreak wholesale destruction.
Indeed, since it is easier to destroy than to create; it might be expected that bad geniuses will have a larger effect than good ones. It is plausible that Napoleon (building on the French Revolution) single-handedly inflicted permanent damage on the French nation by the mass maiming and slaughter of the most able, and able-bodied, men. Hitler perhaps did much the same to Germany; Lenin and Stalin to Russia. From England, the genius of Karl Marx and other 'reformers', socialists, feminists, sexual revolutionaries, atheists and radicals - unleashed truly massive group damage; initially locally, but affecting by now the whole world.
In summary, genius is an unstable product of nations.
A nation with conditions that produce many geniuses may find itself undermined both by the success of its good geniuses (because such is the amplification effect on genius, that useful innovations are difficult to keep secret, and tend to spread from the originating group to that group's enemies); and by its bad geniuses.
Genius may, for a while, lead to more genius - by group selection. But sooner or later, the massive amplification effect of genius is likely to destabilize and destroy the system that engendered it; or else the fitness-reducing effect of genius alleles will tend towards genetic extinction.