Are Differences in Religious Belief Rooted in Differences in Unconscious Information Processing?

We found that individual differences in implicit learning of visuospatial patterns - an evolved mechanism for learning about our environment - predicted religious belief in two culturally and religiously distinct samples.

Like Comment
Read the paper

How can religious belief be so universal (it’s an element of nearly all human cultures) and at the same time so varied (many do not believe, and levels of belief vary widely)? According to several accounts in cognitive anthropology, beliefs about God emerged from universal, bottom-up neurocognitive processes. If belief is indeed rooted in fundamental elements of human information processing, then it would be likely to emerge wherever humans are found. Likewise, individual differences in information processing might account for the variability in belief found in every culture. 

Our research explored these questions, focusing especially on the bottom-up process of implicit pattern learning, in the United States and in Afghanistan – countries that differ dramatically along several religious and cultural dimensions. Unsurprisingly, collecting data on belief in God in Afghanistan presented a number of methodological and ethical challenges.

For instance, how does one ask a participant if s/he believes in God, when it is illegal to say “no”? To denounce God in Afghanistan is considered a crime in Sharia law, punishable with a maximum penalty of death. The solution –we realized after some time – was to phrase the measures such that participants would report their kind of belief without any implication of doubt about God’s existence.

We also had to give careful consideration to the appropriateness of all our experimental measures. For example, we considered using a task that asked participants to imagine the “hand of God” in a scenario where a man is described as having been narrowly missed by a speeding vehicle. But such a measure couldn’t work in Afghanistan, where it is forbidden to depict god in any human form, even as a hand. We considered alternatives, such as depicting god as a cloud, or as wind, but ultimately, meaningful comparisons to Western measures proved elusive.

There were additional cultural challenges related to the optics of conducting psychological research on religion in Kabul. In one instance, our lead Afghan researcher was approached by a local mullah who insisted that we must have ulterior motives. Surely, he insisted, we were seeking to convert Muslims or challenge Islam. Providing an explanation of our research questions and hypotheses did not put his mind at ease. This kind of reaction was not uncommon and it should be considered that the scientific method is not widely taught in Afghan schools. Indeed, we were doing something quite strange by asking questions about religious belief in a scientific context.

Even after all the measures were adequately translated, data collection introduced its own set of challenges. Participants and researchers were gender-matched to accommodate local norms. Substantial efforts had to be taken to mitigate concerns around security and the privacy of our participants, including mobile testing that enabled flexibility in data collection sites.

These challenges afforded valuable learning, even beyond the data, especially for those on the U.S. side of our team who were largely naïve to Afghan culture. And the end result was well worth the fretful moments of uncertainty. We found that individual differences in bottom-up, implicit pattern learning predicted strength of belief in an intervening God as well as increased strength of belief from childhood to adulthood. These findings replicated across the U.S. and Afghanistan, indicating that implicit pattern learning may be a shared neurocognitive basis of religious belief formation, and of variability in belief.


Weinberger, A.B., Gallagher, N.M., Warren, Z.J. et al. Implicit pattern learning predicts individual differences in belief in God in the United States and Afghanistan. Nat Commun 11, 4503 (2020).

Adam Weinberger

Postdoctoral Scientist, Georgetown University

No comments yet.