H/T: Grammar requires less brainpower than falsificative sapienthood
This H/T is based on the notion that brains work by being connective in ways that allows bigger, more interconnected brains to make more precise distinctions, as outlined in can a diagnosis be a falsifiable H/T? and reductio ad absurdum. This is similar to the effects of interferometry, in which multiple telescopes connected to a central image generating computer gives a sharper picture than any of the telescopes could do on its own. The background is supported by Pavlov's research on discrimination in conditioned learning, in which large and connective brains show their more precise distinctions even when the sensory organs are more sensitive in the smaller-brained organism.
An example is that human test subjects only salivated when the metronome ticked at the same rate as when they were fed while the dogs salivated regardless of how fast the metronome ticked. This was the case even though the greater number of receptor cells and different, more sound amplifying structure of the small bones in the dog's inner ear give dogs an ability to hear sounds that are too low volume for humans to hear. This, in turn, shows that the sharpness of the senses cannot be placed on a linear scale that is separate for each sense. Instead, it establishes a "discrimination quotient" or more appropriately a discrimination coefficient that make experiences sharper in all senses and is determined by the overall connectivity of the brain, and that it is distinct from the detection thresholds that are determined by the sensory organs separately for each sense.
The fact that humans can distinguish billions of odors despite their small "olfactory lobes" contradict the claim that olfaction is an "island" in the brain. An encapsulated "domain specific" brain area the small size of a human olfactory lobe cannot possibly distinguish billions of odors, such fine distinctions require a larger neural network which implies that human olfaction relies on brain-wide connectivity and is not biologically isolated from other brain processes. This agrees with the fact that odors are expressible in some languages, https://www.ncbi.nlm.nih.gov/pubmed/24355816 . As it is possible to experience and think things without expressing them in language, the notion that language affects what can be expressed in language has nothing to do with the claim that language determines what it is possible to think. To confuse the concept that different things are possible to express in different languages on one hand with the Sapir-Whorf hypothesis on the other is a silly straw man. The discrimination coefficient (DC) or "discrimination quotient" ("DQ") is evolutionarily possible to generalize in a way that the human-centered concept of IQ is not.
In addition to that, this H/T also states that distinguishing a reductio ad absurdum from a straw man fallacy or a false belief, proper sapientness, requires more precise distinctions than those required to build grammatically correct sentences. This H/T, being based on the concept of precise distinction being the limiting factor of intelligence, does not predict that limits to formal structure as such restricts grammar. While the hypothesis do predict that it takes some discrimination to classify real words into word classes which can then be placed in grammatical sentences, the hypothesis also predict that the discrimination required to do that is of a lower degree than that required to use proper reductio ad absurdums. A difference in biological system requirements of the brain, implying that not all beings that are capable of grammatical language are even biologically capable of acquiring the scientific method.
See also H/T: Toxin resistant hybridization to human intelligence for a possible explanation of evolution across the gap between grammar and critical thought despite the evolutionarily disadvantageous increase in nutrient cost with bigger brains.
Testable on Earth
This H/T predicts that syntactically correct grammar is not a sufficient condition for the ability to distinguish reductio ad absurdums from straw men or false beliefs. It predicts a minimum discrimination coefficient required for independent identification of reductio ad absurdum that is higher than that required for syntactic language. It therefore predicts that some (possibly exclusively future) stages of Artificial Intelligence (AI) research and design should produce AIs that can generate formally correct grammar, but are still incapable of reliably distinguishing reductio ad absurdums as a category. If this prediction is correct, AI that is capable of correct grammar will be invented before AI that is capable of the scientific method.
The H/T also predicts the current existence of biological brains that can generate formally correct grammar but lack the true intelligentness to distinguish reductio ad absurdums. These may exist in the form of children, mild retardation, brain connectivity defects that bear some resemblance to retards but is faster and pass as average or above average on IQ tests, and early stages of dementia that conceal the more obvious early symptoms by memory training.
The H/T also predicts the existence of formal structures in behavior mathematically equivalent of recursive grammar in brains incapable of placing real words in grammatical sentences. That is, apes and maybe even pigeons should be able to place colored objects with colors that represent different classes of words (nouns, verbs and so on) in series that formally correspond to grammatical sentences.
In the case of possible extraterrestrials
In the case of any future discoveries of extraterrestrial complex life on multiple worlds by means of interstellar space travel (not to be confused with radio SETI, or optical SETI that may or may not have discovered extraterrestrial technological civilizations as per H/T: Periodic light pulsations from 234 stars are created by technological civilizations which could only detect high technology), this H/T predicts that in addition to alien sapients capable of creating technological civilizations, there should also be alien pre-sapients capable of grammatical language that cannot do science and therefore cannot create technological civilizations. The H/T predicts that in some cases, such a limitation in extraterrestrial presapient thought should be due to brainpower restrictions and not simply culture, and that it should show up in an inability to conceptually distinguish reductio ad absurdum from straw men. The H/T also predicts that all technological civilizations should be created by entities capable of that distinction.
The hypothesis also predict that those capable of reductio ad absurdum should have higher discrimination coefficients than those capable of grammar but not of reductio ad absurdums. These differences should be detectable by discrimination tests of conditioned learning similar to those used by Pavlov. Aliens capable of grammar but not reductio ad absurdum should have higher discrimination coefficient than Pavlov's dogs but lower than adult humans, possibly similar to those of the 8-9 year olds Pavlov examined in his studies.
Distribution gaps and evolutionary fluxes
A challenged prediction
Since more precisely discriminating brains cost more nutrients, high cost brains require more advantages to select for even more precise brains in evolution, until critical thresholds become necessary for any further increase by means of natural selection. That implies that it will take something other than a gradual increase in an isolated population to cross the gap between grammar and reductio ad absurdum. This means that populations with averages between the grammar threshold and the reductio ad absurdum threshold (except those slightly above the grammar threshold that are explainable by redundancy in the case of genetic variation and wear and tear of the brain) should only exist for short periods changing either upward or downward and, in the case of upward, only if some individuals are already at the critical threshold for reductio ad absurdums. Therefore, the hypothesis predicts that there will be a range of discrimination coefficients that end slightly higher than the threshold for grammatical language, and a gap between that range and a separate range that begins at or slightly above the threshold of reductio ad absurdum. The hypothesis also predict that the very few populations in the mostly empty gap will all be in rapid evolutionary flux (rapid natural selection on individual hereditary variation that already exists within a population that may possibly but not necessarily have formed by hybridization, as opposed to the long waiting for new mutations that is the limiting factor of long term evolutionary speed) and that none of them will display fossil or genetic evidence of being a part of slow evolutionary change.
The hypothesis do allow for some differences in the distribution of how and when the brain consume the most nutrients. It does permit the existence of some beings in whom the brain consume the most nutrients at the pre-sapient stage by forming more well-insulated nerve connections in the brain that make neuron firing cheaper in adulthood, and other beings whose brains consume less nutrients before sapience and/or mature faster but at the price of less insulated nerves that make neuron firing in adulthood more expensive in nutrients. It does not, however, change the fact that more precise brains increase nutrient cost in one form or another, and so the predictions about a gap in distribution and the flux remain.
The prediction challenge note
Brain damage can reduce brainpower and make redundant capacity an advantage. It may be possible for populations with high risk of brain damage to evolve towards higher discrimination levels even in the gap between grammar and the ability to distinguish actual outcomes from predicted outcomes. This challenges the prediction of a gap between populations.
Test data on Earth, fossil humans
There is archaeological evidence from hominins combining modern sapiens features with more archaic characteristics that a modern brain case did not go hand in hand with modern vocal tracts. While the 300000 year old human fossils in Jebel Irhoud combine modern human faces and mouths with an archaic brain casing as is the empirical content of this article, there are other fossils in Africa more than a hundred thousand years old that combine modern human brain cases with an archaic human mouth. This shows that a modern human vocal tract was selected for in at least one population that still had lower brain connectivity than modern humans.
While both populations lived in Africa, the fact that they lived on the same continent does not prove that they were the same gene pool with the same characteristics. After all, chimpanzees and bonobos both live in Africa, as different populations that differ in their characteristics. There are also many plant and animal species that have population ranges distributed on both sides of the line on the map between Africa and the Middle East. Some of them are primates (baboons).