Two hypotheses about natural intelligence

We assume that the natural intelligence (human, particularly) is equivalent to a large inferring structure, which took shape in the last 400/500 million years. Then two hypotheses, about this structure and its development, are put forward for consideration. The first one concerns the transmission, from one generation to another, of the structure: we propose that this passage is done by direct transfer, mother to children, during pregnancy (“maternal download”). The second hypothesis regards the structure evolution: now the acquired improvements can be transferred to the descendants, so it is possible to envisage a governed evolutionary process (“evolution by improvements”). Introduction Following [1] and, to some extent [3], we assume that the natural intelligence is equivalent to an inferring structure (an articulate and organized object able to infer and compute), presently very large and complex, at least for the mammals. Because all such structures are in a sense equivalent [1, 2], this assumption is not really objectionable; the only uncertain point is the structure size. This size will be evaluated employing, as unit of measure, the usual programming instructions, but every other method could be considered appropriate. A size, for today’s humans, of 10 instructions seems reasonable and, in any case, compatible with the physical characteristics of the brain. Moreover, the whole structure (or great part of it) must have taken shape, in some manner, by some evolutionary process, in the last 400/500 million years. Now, we are asking ourselves: (A) how the inferring structure passes from one generation to another; (B) in which manner it has reached the current size and complexity, at least for some species. In the first and second section we propose an answer, based upon two, strictly connected, hypotheses: (A) the generation-to-generation transmission is done by direct 10 9 10 ÷