7.Parametrized Globus Model with mate selection

The selection of a partner * as an auxiliary mechanism of evolution has been a paradigm since the first developments of the theory of evolution. Darwin himself wrote The Descent of Man and Selection in Relation to Sex (1871) introducing a new factor, mate selection or sexual selection, through which females or males choose those with the most attractive qualities as their partner.

Intelligence is doubtlessly one of these desirable qualities for various reasons. However, from the viewpoint of our Globus model of intelligence evolution complexity, this does not deal with imposing such a general hypothesis given that the IQ of the mother and father is fixed.

When reflecting on the possibility of establishing another hypothesis to our model's that could improve its general adjustment while being confirmed. I have thought of trying out the idea of relevance of the difference of intelligence between the father and the mother as a selection criterion or conditioning for the initial effective acceptation of the couple's configuration, in other words, mate selection.

I sincerely think that intelligence is not a strict selection criterion when it comes to choosing a partner; but likewise, I think that a large difference in intelligence within the couple it is not very common.

This small dissertation of evolutionary psychology about sexual selection becomes more complicated when we think that, in reality, we have two intelligences: that which corresponds to each one of our progenitors and that which operates under diverse forms or conditions, as we have already established.

In short, the additional hypothesis introduced into the Globus model of mate selection is establishing as the limit of the difference in intelligence that: the most powerful chromosome of one of the couple's members has to be at least as powerful as the least powerful gene of the other and vice versa..

Additional Mate selection and intelligence
Partner compatibility

The psychological justification of the mate selection hypothesis is based on the fact that we do not demand the same intelligence from a person that we know, but to form a couple, we demand that the other person can at least acceptably follow conversation. This can be obtained with only one chromosome given that to follow an argument in a conversation, certainty is not needed; in fact, the certainty derived from genetic verification is offered by the initial argument presented, in other words, the two chromosomes of the person who presented the argument.

Although the explanation may not be very extensive, what is important here is that the Globus model's responses to selection criterion substantially improves the adjustment.

The hypothesis will affect, if introduced, only the M2 or F2 genes. These are estimated given that the measured IQs collect the power of the significant or less powerful gene, therefore, the estimations of M2 and F2 will change in light of the new information or condition introduced in the model.

The Globus model will somewhat improve with the individual variables (Otis test of fathers and mothers and Intelligence test Wechsler and Stanford-Binet scales of children), but its responses to selection criteria will be noticed much more with the centered variables. The G-MCI with the rearrangement criteria M1F1, goes from 15.61 to 17 and the maximum r² from 0.89 to 0.97 for the objective function R (see figures q063 and q077). For the objective function M & F the G-MCI is found at 17.62 while it was previously found at 17.77 and the maximum also rises from 0.89 to 0.97. The maximum values of r² almost always correspond to the variable X6 or the average of six of the children's variables.

Sensitivity analysis
without mate selection
Sensitivity analysis without mate selection

Repeating the same sensitivity analysis performed with the variable X6 in the section of internal evolution and the complexity of the Globus model, but with the additional hypothesis of partner selection, the correlation graph below of the Super Globus Model is obtained; we can highlight its following aspects.

  • Of the graph's four peaks, one is maintained and the other three are shifted upwards.

  • The correlations of the anticipated IQs for the objective function R with the function M & F are much more similar than before. The overall average of the function R slightly rises and the M & F slightly lowers.

A reliable interpretation of these results is almost impossible given the margin of sensitivity of the introduced changes and the complexity of the Globus model. We have to take into account that only ten of the seventy IQ of the function R have been affected in more than two percent of their value, but, in order to try to give a positive explanation of the two mentioned aspects we could say:

  • In respect to the first, it seems as though the model improves when its parameters are correct and that it worsens when the parameters are fictitious, which, reinforces both the model and the introduced hypothesis.

  • In regards to the second, the function R improves when additional information is incorporated in its definition, while the function M & F, in spite of its peaks, decreases because it does not incorporate in its composition the effect of the hypothesis introduced on genes M2 and F2 given that it only collects information from M1 and F1 (that are the known IQs)

On the other hand, we have to point out that the function R's results improve as an objective function but not as rearrangement criteria. This fact can be understood if we think that, because it deals with average values of different possibilities, it incorporates the differences due to the genetic combination to a greater extent than the M1F1 criteria.

The reasoning is similar to that which occurs with variable W; in many cases it has very high correlations but, as rearrangement criteria it is normally terrible because it incorporates the effects of genetic combination, of the functional limitations, and of affinity as well as the rest of the deviations due to the simulation of the errors of measurement and expression. Likewise, the children's variables tend to be very good rearrangement criterion because they do not incorporate the deviations due to genetic combination, affinity, internal evolution, or of functional limitations.

Sensitivity analysis
with mate selection
Sensitivity analysis with mate selection

It is worth pointing out that if the model's G-MCI was wrong, it could lower considerably, even for small changes. This can be confirmed in the sensitivity analysis graph in the parameters of internal evolution for the variable X6, where one percent of the increase of the potential to be transmitted by the fathers or mothers makes the cited G-MCI drop drastically.

In short, the proposed hypothesis of mate selection seems reasonably correct. The Globus model's coherence improves in general and the correlations of the variables centered with the rearrangement criterion M1F1 sensitively increase.