Show someone solid evidence that a fact they hold is wrong, and plenty of people will actually budge. Show them solid evidence that a political belief is wrong, and something else happens: they’ll admit the fact underneath it looks shakier, and keep the belief anyway. That split is the finding sitting inside Neural correlates of belief change in political and non-political domains among left-wing individuals confronted with counterarguments, published Jan. 8, 2026 in Scientific Reports by Małgorzata Kossowska, Paulina Szwed, Gabriela Czarnek, Sofiia Honcharowa, Aleksandra Domagalik, and Łucja Doradzińska of Jagiellonian University, and covered Oct. 26, 2026 by PsyPost. Researchers scanned the brains of 43 left-leaning adults while they read counterarguments to political claims and ordinary factual claims. Confidence in the facts dropped hard. Confidence in the politics barely moved at all — and challenging the politics lit up a different part of the brain than challenging the facts did, a network built around thinking about yourself. William James named the mechanism underneath that exact split in 1896, in an essay about which beliefs intellect gets to settle and which ones get settled by what he called our “passional nature” instead.

The Quick Verdict

Question	What we found
What was studied?	Whether challenging political beliefs engages different brain networks than challenging non-political factual beliefs, in people with strongly held left-wing views
Who ran it?	Małgorzata Kossowska, Paulina Szwed, Gabriela Czarnek, Sofiia Honcharowa, Aleksandra Domagalik, and Łucja Doradzińska, Jagiellonian University
Where and when?	Scientific Reports, published Jan. 8, 2026; covered Oct. 26, 2026 by PsyPost
Sample	43 left-leaning Polish adults in the final behavioral sample (46 recruited), average age about 25, screened for firmly held left-wing views
Headline finding	Non-political belief ratings dropped sharply after counterarguments, from 6.13 to 5.37; political belief ratings barely moved, from 5.72 to 5.68
Mechanism	Political counterarguments produced greater activity in the Default Mode Network (posterior cingulate cortex, medial prefrontal cortex, right supramarginal gyrus) than non-political counterarguments did
What James already argued	The Will to Believe (1896): a “forced, living, and momentous” option gets decided by our “passional nature,” not pure intellect, especially once the belief is bound up with our whole way of acting

The Quick Version

A new brain-imaging study scanned 43 people with firmly held left-wing views while they read arguments against both their political beliefs and plain factual claims. They lowered their confidence in the facts. They did not lower their confidence in the politics. Resisting the political arguments lit up the brain’s self-reflection network harder than resisting the factual ones did. William James argued in 1896 that this is exactly how it should work: once a belief is wrapped up in who you are, intellect stops being the thing deciding it.

What the Study Actually Found

Kossowska’s team wasn’t trying to prove that left-leaning people are especially stubborn. They were testing something narrower and more interesting: whether the brain treats a challenge to a political belief the same way it treats a challenge to an ordinary fact, or whether it routes the two down different circuits entirely.

Forty-six strongly left-wing adults were recruited; 43 completed the full behavioral protocol, right-handed, with no neurological or psychiatric history, average age around 25. Inside an fMRI scanner, they read eight statements about politics, including “Abortion should be legal,” “Poland should agree to accept refugees,” and “Sex education should be mandatory for every student,” plus eight statements with no political charge at all, like “A college education generally improves a person’s economic prospects” and “Thomas Edison invented the light bulb.” Each statement came paired with a counterargument. Before and after, participants rated how much they believed the original claim on a 1-to-7 scale. A subset of 32 came back about a week later for a follow-up.

The behavioral gap was not subtle. Non-political belief ratings fell from 6.13 to 5.37 (a large shift, Cohen’s d of 1.13). Political belief ratings moved from 5.72 to 5.68. That’s not a smaller effect. That’s close to no effect at all, d of 0.06. People weren’t stonewalling the counterarguments outright, either. Reaction times were actually faster for political statements (5.16 seconds) than non-political ones (5.70 seconds). Whatever was happening, it wasn’t slow, effortful deliberation. If anything, it looked quicker than the honest reckoning with a fact about a light bulb.

The imaging explains the gap better than the behavior alone does. Challenges to political beliefs produced relatively greater activity in regions of the Default Mode Network (the posterior cingulate cortex, the medial prefrontal cortex, and the right supramarginal gyrus) than challenges to non-political beliefs. Greater resistance to changing a belief also tracked with more activity in the dorsomedial prefrontal cortex and less in the orbitofrontal cortex, a region more associated with updating judgments in light of new information. Kossowska told PsyPost the team expected emotion-processing regions to be doing the heavy lifting here, the usual story about political beliefs running hot. “We did not replicate these associations,” she said. “Instead, our results pointed more strongly toward brain systems involved in self-referential processing.”

What Did James Mean By This?

In The Will to Believe (1896), James writes: “Our passional nature not only lawfully may, but must, decide an option between propositions, whenever it is a genuine option that cannot by its nature be decided on intellectual grounds.”

James wasn’t handing out a license to believe whatever feels good. His argument was narrower and did real work: he defined a “genuine option” as one that’s forced (you can’t dodge the choice by waiting), living (it actually grips you, not a hypothetical you’d shrug off), and momentous (something real is riding on it, and the chance to decide won’t come around again in the same way). Most everyday factual questions don’t clear that bar. Whether Edison invented the light bulb doesn’t touch who you are or how you live. But a belief about abortion, or refugees, or what children should be taught: that kind of belief is load-bearing. It’s wired into identity, group loyalty, a whole way of acting in the world. James’s claim was that once a question crosses into that territory, intellect alone was never going to be the thing deciding it. Something else always was. He just didn’t have a scanner to point at what.

Read the Kossowska numbers against that distinction and the fit is closer than a slogan. “Thomas Edison invented the light bulb” is James’s non-genuine option: revisable, low-stakes, not bound to anything about who you are. “Abortion should be legal” is a textbook genuine option in his sense: forced in practice (you live inside a society that resolves it one way or another), gripping, and tied to identity in a way a trivia fact never will be. The brain networks tracked the same line James drew by argument alone, a century and change before anyone could watch it happen on a screen.

Why Don’t Facts Change Political Beliefs?

1. Facts get processed as information. Political beliefs get processed as self. The Default Mode Network (the posterior cingulate cortex, medial prefrontal cortex, and right supramarginal gyrus) is associated with self-referential thought generally, not political thought specifically. A challenge to a political belief seems to recruit the same machinery you’d use to think about who you are.
2. Resistance isn’t slow or effortful here: it’s fast. Participants answered political items faster than factual ones, not slower. Whatever is protecting the belief isn’t a deliberate, grinding defense. It looks closer to a belief that doesn’t need to be defended because it was never purely intellectual in the first place.
3. More dorsomedial prefrontal activity, less orbitofrontal activity, tracked more resistance. The orbitofrontal cortex is tied to updating judgments against new evidence. Less activity there, specifically on political items, lines up with less actual updating.
4. The pattern showed up without comparing left and right. Kossowska was explicit that this study doesn’t test whether left-wing views are more or less flexible than right-wing ones. It tests only that political beliefs, in this sample, behaved differently from non-political ones. The asymmetry worth noticing is belief-type, not party.

Why the Study and the Philosophy Rhyme

The honest version of this comparison isn’t “a 19th-century pragmatist predicted a 2026 fMRI study.” It’s narrower, and it holds up better for being narrower: James’s specific claim was that intellect only governs the beliefs that don’t touch identity, and that once a belief does touch identity, something else takes over the deciding — not as a flaw in reasoning, but as how genuine options always work. Kossowska’s team wasn’t testing James. They were testing whether political and non-political belief challenges recruit different brain systems in the same people. But the shape of what they found is James’s shape: one category of belief yielding to evidence, another holding firm while routing through the brain’s sense of self.

It connects to the piece on Spinoza and partisan wishful thinking, which found belief about who’ll win an election tracking desire rather than evidence, a different kind of political belief, the same refusal to let intellect run the show alone. It rhymes with the other study covered here on James, where a bodily reaction nobody consciously noticed still shaped political identity; here, a belief gets consciously defended, but by a network built for self-concept rather than argument. And it sits in real tension with the piece on James and self-boundary collapse, where a thinner sense of self tracked with more peace (the Default Mode Network, in that study, going quiet rather than lighting up). Here, a thicker, more defended sense of self is exactly what’s standing between a person and a belief they’d otherwise have to revise. It’s also worth reading next to the piece on Sextus Empiricus and the limits of one study: the honest lesson from a single Polish sample of left-leaning twenty-somethings is narrower than the headline, and James’s own caution about overclaiming applies here as much as it does to the neuroscience.

The Gap James Didn’t See Coming

James built his argument from introspection, philosophy of religion, and the stakes of belief in God. He was defending the right to a religious commitment that couldn’t be settled by evidence either way. He had no reason to expect the mechanism to show up as a specific, locatable pattern of blood flow, and no reason to expect it to generalize from personal metaphysical commitments to something as mundane as an abortion-policy statement on a laptop screen in a scanner.

He also had no way to anticipate the asymmetry this specific study couldn’t test: whether the passional nature he described operates the same way, with the same intensity, across the political spectrum. Kossowska was careful on this point: her sample was left-wing only, by design, and she said directly that the findings “should not be interpreted as suggesting that left-wing individuals are particularly resistant to changing their political beliefs.” James’s passional nature was a claim about minds in general. Whether it runs hotter, cooler, or identically on the other side of the aisle is a question this study was never built to answer, and one James, writing about belief in the abstract, never had to confront at all.

What This Study Doesn’t Show

Be precise about the size of this. The framing invites more confidence than the data earns in either direction.

One Political Side, No Comparison

Every participant was recruited for strongly left-wing views. There is no right-wing comparison group in this study. Nothing here tells you whether the same DMN pattern would show up, show up more, or show up less, in people with opposite politics.

Not Permanent, Not Predetermined

Kossowska was explicit about this too: “Our findings do not mean that political beliefs are impossible to change or that people’s responses are neurologically predetermined.” A brain pattern that shows up under one experimental challenge in one sitting isn’t a life sentence on anyone’s capacity to ever update a political view.

Eight Statements, Chosen for Poland in 2026

The political items (abortion, refugees, mandatory sex education) reflect live debates in Poland at the time of the study. A different set of political statements, in a different country, might not produce an identical split between “political” and “non-political” processing.

No Real-World Behavior Measured

Nobody’s voting, canvassing, or arguing with family was tracked. This is belief-rating inside a scanner, immediately after a counterargument. Whether the same resistance holds up a week, a month, or an election cycle later is a separate question the design doesn’t answer.

What Actually Follows From This

1. Don’t expect a fact-check to do a political belief’s job. If the study and James are both pointing at the same mechanism, a correction aimed at the factual layer underneath a political belief can succeed completely (people did lower their confidence in facts here) and still leave the political conclusion standing untouched.
2. Notice when a disagreement has stopped being about evidence. Faster reaction times on the political items, not slower, suggests the defense isn’t effortful reasoning. If a conversation about politics feels suspiciously quick and settled on your end, that speed might be the tell, not a sign you’ve thought it through.
3. Ask James’s actual question before you argue the facts again: is this a genuine option for the other person? Forced, living, momentous, bound up with identity: if so, better evidence alone was never going to be the lever. Something closer to trust, relationship, or shared stakes might move more than one more statistic will.

James never ran a regression and Kossowska’s team wasn’t out to vindicate a 130-year-old essay about the ethics of belief. But a 2026 fMRI study built to test whether political and factual beliefs get processed the same way landed on the exact split James argued for in 1896: facts yielding to evidence, and beliefs bound up with who we are yielding to something else entirely — a self, not a syllogism, doing the deciding.

---

This post discusses a study published Jan. 8, 2026 in Scientific Reports (Kossowska, Szwed, Czarnek, Honcharowa, Domagalik, and Doradzińska, Jagiellonian University), covered Oct. 26, 2026 by PsyPost, and William James’s The Will to Believe (1896). The study’s sample was limited to 43 left-leaning Polish adults with no comparison group from other political orientations, and the researchers themselves caution against reading the findings as evidence that political beliefs are immune to change.