Taiwan Isn’t Giving Up, It’s Pulling Apart

In a recent article for The Diplomat, Mor Sobol offered an even-handed assessment of the debate set off by Simon Shuster’s piece in The Atlantic, which asked whether Taiwan might simply “give up” under growing pressure from China. Sobol argued that both Shuster’s pessimistic answer and his critics’ confident rebuttals rest on thinner evidence than either side recognizes. 

Sobol pointed out that telling a pollster “I would defend Taiwan” is very different from making that decision when missiles are falling, and that Taiwan is not a unitary actor. The question worth asking is how the trends in public opinion, military recruitment, defense spending and other metrics move over time, not what any single snapshot shows.

I agree. As a scholar who studies Taiwanese defense attitudes with survey data, I want to add evidence showing why the much-discussed “willingness to fight” has, to some degree, misled outside observers. The problem with polling is not that attitudes can’t be measured. It’s that pollsters keep asking the most superficial question. “Would you fight?” captures a sentiment. What decides whether that sentiment translates into action is whether citizens believe the conditions for sacrifice are in place. Those conditions can be measured, tracked over time, and broken down by group. 

When we do that, the picture is more troubling than either the optimists or the pessimists suggest, though for different reasons.

Citizens never accept the risks of war unconditionally. They take part in what I call a Defense Social Contract, which rests on three pillars, all of them vitally important.

The first pillar is trust in the state. People will sacrifice for a government they believe governs competently and fairly, and for a country they believe has a future worth defending. When citizens feel the state has failed them, its claim to extraordinary sacrifice loses legitimacy.

The second pillar is trust in one another. Defense is a public good, and public goods invite free-riding. I will hold the line only if I believe my neighbors will too. That means the burden must be shared, rather than some Taiwanese being expected to serve while others are exempt. Resistance must also be the common cause of all Taiwanese, not the property of one political camp.

The final pillar is trust that resistance can work. Facing a far more powerful adversary, citizens of a small democracy must believe their own military is capable and that allies will arrive in time. Sacrifice that looks futile turns into fatalism.

If any one pillar gives way, professed willingness to fight stays a sentiment and never becomes action.

What the Surveys Show

All three pillars can be measured. To do so, I used two nationally representative waves of the Taiwan National Security Survey (TNSS), fielded in June 2024 and September 2025, with 2,745 respondents in total. This makes it possible to do what Sobol recommends: track change over time and look inside the aggregate. The timing deserves emphasis. The 2024 survey was fielded about a month after President Lai Ching-te took office. The 2025 survey was fielded about a month after the final round of Taiwan’s Great Recall votes, a national effort to recall all legislators from the Kuomintang (KMT) that ultimately failed.

The TNSS directly asks respondents what they would actually do if China invaded. One option is to “let nature take its course” (shun qi zi ran). That answer is not surrender, and it is not a rejection of Taiwan. It is withdrawal: wait and see. I call it fatalistic passivity. Between the two waves, the share of Taiwanese choosing it rose from about 26 percent to nearly 37 percent.

From 2024 to 2025, the TNSS results showed that all three pillars weakened or came under strain.

On trust in the state, the share of respondents who said the national economy had gotten worse rose from 31 to 48 percent. Negative economic evaluations are strongly associated with passivity, although the association weakens once other factors are taken into account.

The share who doubt Taiwan’s military could defend the island on its own rose from 69 to 74 percent. Confidence in the military and the expectation of U.S. intervention are among the strongest and most consistent predictors of lower passivity across every model. Passivity is highest among those pessimistic on both counts.

Finally, there’s trust in fellow Taiwanese. Respondents who believe “most Taiwanese would resist” are much less likely to opt to “let nature take its course.” Predicted passivity falls from about 36 percent among those who think their compatriots definitely would not resist to about 25 percent among those who think they definitely would.

The most important finding, however, is not the numbers themselves. It’s the breakdown of the survey results along political lines. Among respondents who identified with the KMT, expressions of passivity rose from 22 percent in 2024 to 40 percent in 2025, an 18-point jump. That difference holds after controlling for national identity, unification-independence preferences, confidence in the military, expectations of U.S. assistance, economic evaluations, and demographics. Among Democratic Progressive Party (DPP) identifiers, it rose from 26 to 30.5 percent, a change statistically indistinguishable from zero. The 13.6-point gap between the two trajectories is statistically significant.

I want to stress what this does not mean, because the finding is easy to misread. On its face, it might seem to show that KMT supporters love Taiwan less or are less willing to defend it. That’s not accurate. In 2024, KMT identifiers were actually slightly less fatalistic than DPP identifiers. In other words, at the start of Lai’s presidency, KMT supporters showed no sign of giving up on Taiwan’s defense simply because the DPP had won another term. 

In neither wave is the partisan gap on its own statistically significant. Both surveys took place under a DPP government, with KMT supporters in opposition throughout. The divergence between June 2024 and September 2025 therefore suggests that something changed over those 15 months in how the opposition camp relates to the DPP government. 

Between the two surveys, Taiwan went through more than a year of fierce executive-legislative conflict, ending in the 2025 Great Recall. Advocates framed the campaign to remove 31 KMT legislators as a defense of national security against Chinese infiltration. Opponents saw it as an attempt to overturn an election result. All 31 recall votes failed. The September 2025 survey was fielded weeks later. 

The partisan gap was already there. In both waves of the TNSS, KMT identifiers were about 20 points less likely than DPP identifiers to believe most Taiwanese would resist (roughly 61-63 percent versus 80-82 percent). In both the DPP and the KMT, the share who believe “most Taiwanese would resist” barely moved from 2024 to 2025. A substantial share of the electorate had already lost confidence that the burden would be shared.

But people’s own willingness to act did change, and it fell most among supporters of the KMT, whose political opponents had been presenting national security as a partisan cause.

Survey data cannot prove that any single event caused the shift. But when the language of national security becomes a weapon in domestic political combat, this pattern should come as no surprise. 

What We Should Be Tracking

My reading is that the problem for Taiwan’s defense is not a lack of will but a lack of unity. It is very hard to ask citizens to risk their lives alongside people they already dislike and distrust in everyday politics, or to sacrifice for a government they see as a political adversary. The TNSS does not directly measure cross-partisan trust, so this interpretation needs further testing. It fits the evidence, though: a society that still sees itself as one nation but is increasingly unsure it can fight as one.

To gauge Taiwan’s willingness to fight, Sobol argued for watching behavior: recruitment, retention, reserve training, and civil defense participation. I strongly agree. I would add three sets of attitudinal indicators that come before this behavior and help explain it.

Do supporters of every party believe that the government asking for sacrifice governs fairly and competently, and that the country has a future?

Do citizens believe the burden is shared? Conscription reform, reserve obligations, and civil defense must be seen as universal, not selective. National security must stop being used against political opponents. Leaders of all parties should stop branding one another as Beijing’s proxies and should negotiate defense budgets and reforms as cross-party agreements.

Are allied signals consistent? Taiwanese must believe they can hold out until help arrives. Mixed messages from Washington or Tokyo damage the very pillar my data identify as most critical. Allies should also engage Taiwan’s major parties evenly, so that outside support is not seen as backing one camp.

These indicators should be broken down by party, age, gender, and region, and tracked across survey waves. If behavior and attitudes move together, across various segments of society, we can trust them more. If the aggregate holds steady while part of society quietly pulls back, we are looking at what I call “fragile resilience.”

So is Taiwan giving up? The evidence says no. Taiwanese still identify with their country, and most still expect their compatriots to resist. What the data show is a society pulling apart. Fatalism rose most sharply among those who identify with the KMT, whose legislators had just been targeted in a campaign that cast national security as a partisan cause.

This polarization is harder to see; the headline number on willingness to fight won’t capture it. It’s also dangerous. Beijing doesn’t need Taiwanese to embrace unification. It only needs enough of them to decide that the war belongs to someone else.

But pulling apart is also easier to fix than giving up. It is a political problem, and political problems have political remedies: a government that earns trust across party lines, a defense burden that everyone shares, and allies whose signals don’t waver. The fate of 23 million people depends on whether Taiwan’s leaders can rebuild a defense contract that citizens across the political spectrum believe in.

No one can save a small democracy that cannot stand together, however many polls say it is willing to fight.