Humans were always smart

April 11, 2026

The human nervous system, c. 1908
The human nervous system, c. 1908

Intelligence is a highly polygenic trait, meaning it's influenced by a large portion of our genetics. This suggests the growth in human cognitive potential over time has been much slower than most would assume.

When advantageous genetic mutations occur, they "sweep" through a population, because the presence of a specific sequence of DNA that makes an individual much more likely to survive and reproduce spreads quickly. Lactase persistence, the ability to digest lactose into adulthood, is the classic example of a genetic sweep. Before dairy farming first emerged around 9,000 years ago, nearly all humans were lactose intolerant. Since then, several different mutations arose independently, all affecting the same specific portion of DNA and all enabling lactase persistence. These mutations swept through populations all over the world, and today around a third of all humans are lactose tolerant. In Northwest Europe, the sweep was particularly strong, where lactase persistence first appeared only 4,500 years ago, spread slowly at first to less than 10% of the population by 3,000 years ago, and then spread very rapidly, reaching 90% today.

Lactase persistence is a monogenic trait, though — it relies on a single genetic locus. Global genetic sweeps are also uncommon. A widely-cited 2011 study found virtually no evidence of genetic sweeps as a dominant driver of human evolution in the last 250,000 years[1]. Even the widespread nature of lactase persistence is partially the result of convergent evolution, where multiple independent sweeps occurred in different populations. If a number of highly advantageous mutations affecting cognition had occurred in the last 250,000 years, they would have swept through all human populations and left a record in our genomes. Each would also be linked to a specific genetic variant with a strong influence on intelligence. This is not the case though. In reality, there are thousands of genes and a much larger number of genetic variants that are predictive of the heritability of intelligence[2]. Even the most powerfully predictive genetic loci can individually explain less than 0.02% of the variance in intelligence.

The consensus is that human intelligence evolved primarily through so-called soft sweeps, where natural selection acts on many different existing genetic variants over time, which become more frequent in human genomes as a result. But even in this case, there's no master combination of alleles that unlocked radically new levels of intelligence. If there were, this combination alone would be highly predictive of intelligence and we would expect to find a record of a highly-coordinated shift in the frequency of its alleles in historical human genomes shortly after it first appeared. That said, there are a handful of significant mutations researchers point to that influence brain development and function, but most are more than 300,000 years old and in many cases are millions of years old, predating the emergence of Homo sapiens[3].

To make this more concrete, imagine transporting a child from an uncontacted tribe into an industrialized society. Imagine also that you could transport their family and whatever other support structures this child would need to be raised in a happy, healthy, supported way. Every member of every uncontacted tribe on earth, despite lacking access to the accumulated knowledge of industrialized societies, is an anatomically modern human whose genome is at least 99.9% identical to that of every other living human. With the right teachers and resources, it's not difficult to imagine that this child could grow up to be a calculus teacher who flies airplanes in her free time.

Now imagine transporting a child from 100,000 years ago into the present day. This child would also share our anatomy and 99.9% of our genetics[4]. They, too, could likely become a calculus teacher and a pilot.

We are not fundamentally different from our very distant ancestors — or our more recent ancestors, for that matter. The word "barbarian"[5] was first used as a catch-all for foreigners, viewed as culturally and cognitively inferior due to their alien way of life and the unintelligible "bar... bar... bar..." of their speech. We make the same mistake today when we implicitly assume that very ancient humans were cognitively inferior because their way of life remains largely incomprehensible. But intelligence, just like height, resting heart rate, circadian rhythm, and a host of other human traits, is highly polygenic. It evolved slowly over a long period of time, much of which took place before Homo sapiens first appeared. Humans have been exceptionally intelligent for an exceptionally long time.


Notes:

  1. Subsequent research has found evidence for a number of regional sweeps that likely increased fitness in specific regional environments but were lost due to later mixing with other populations. See here and here.

  2. Intelligence is quantified using IQ test results, which may only be a weak proxy for intelligence. But even a perfect metric that is precisely measurable, stable throughout an individual's life, and fully isolated from culture and other environmental factors wouldn't change the underlying polygenetic nature of intelligence.

  3. There are more recent mutations that have been identified as well, notably to the ASPM and MCPH1 genes, though these mutations have been found to have no predictive power over intelligence.

  4. They may actually have a slightly bigger brain than the average human today. Humans likely had a 10% larger cranial volume, on average, 100,000 years ago. Brain size is only weakly correlated with intelligence though, and as I discussed in The archaic nature of intellect, even animals with microscopic brains, such as honeybees, are capable of impressive cognitive feats.

  5. The Greeks coined the term, originally βάρβαρος (pronounced BAR-bar-os).