The Navy That Denies

In a September 4 essay in Foreign Affairs, Holmes Liao and Dennis Blair argue that Taiwan should build its surface fleet around that distinction. Rather than replace aging destroyers and frigates with another generation of large combatants, they propose more stealthy, missile- and drone-armed corvettes, operating alongside shore-based weapons, aircraft and submarines. The objective is not to possess a miniature version of China's navy. It is to make an amphibious crossing prohibitively difficult.

This is a policy proposal, not a Taiwanese government decision or proof of how a war would end. Liao is a senior adviser at the Taiwan Space Agency; Blair is a former commander of U.S. Pacific Command and former director of national intelligence. Their experience gives the argument weight. It does not turn its predictions into observations.

The strongest part of their case is the mismatch it identifies. An invasion force must transport people and equipment across water. A defender does not have to transport an equivalent invading army in the opposite direction. The two sides therefore need not purchase matching answers.

The difficult part begins after that insight. A navy built to prevent a landing is not automatically a navy that can keep commerce moving, resist a quarantine or preserve civilian life through a prolonged crisis. Taiwan needs to ask what each investment makes possible—and what it leaves unanswered.

Denial Is Not Dominion

Sea control and sea denial describe different ambitions. Control means being able to use an area for one's own purposes. Denial means preventing an opponent from using it effectively. They can overlap, but they are not interchangeable.

Liao and Blair's argument concerns the second. They propose corvettes in the 700–1,000-ton range, dispersed across a defense network rather than concentrated into a few large hulls. Their premise is that smaller, harder-to-target combatants would be more likely to survive an opening attack and remain able to threaten ships carrying an invasion force.

That is a proposed advantage, not a law of displacement. Small ships can be detected and destroyed. Their usefulness depends on weapons, crews, surveillance, communications and support, not simply on being small. A collection of surviving hulls that cannot locate targets or coordinate action is not the same thing as a surviving combat capability.

The authors recognize this: their proposal includes improvements to sensors, command networks and uncrewed systems. The corvette is the visible component of a larger arrangement. Replacing a large ship with several smaller ones while neglecting the connections among them would reproduce the original error in a different size.

The strategic attraction is nevertheless clear. A defender may be able to place uncertainty across an attacker's route without matching its fleet ship for ship. Deterrence would rest on an adversary's doubt that it could complete the crossing, not on the defender winning every category in a naval inventory.

That is alignment with a particular task, rather than imitation of an opponent's scale. But the word “particular” matters. It limits the claim as much as it explains the opportunity.

The Ship and the Institution

Liao and Blair identify institutional resistance as a central obstacle. Large ships, they argue, offer politicians visible evidence of national achievement and give naval leaders familiar platforms. Smaller, dispersed systems are harder to display as a single accomplishment.

There is a genuine accountability problem here, although the essay does not establish that prestige explains any particular Taiwanese procurement decision. A large vessel is not useless because it photographs well. Nor is an unfamiliar system effective because its advocates call it transformative. Those are rival sales pitches, not evaluations.

The more useful question is whether the public demonstration and the military function have become substitutes for each other. A commissioning ceremony proves that a ship has entered service. It does not prove the ship would remain useful under the attack conditions for which it was purchased. The ribbon is not a survivability test.

The same scrutiny must apply to the reformers. A dispersed force can acquire its own prestige vocabulary: agile, networked, autonomous. None of those adjectives shows that communications will work when disrupted, that crews can sustain operations or that replacement equipment will arrive when needed. Modernization is not exempt from the institutional habits it criticizes.

The relevant public argument should therefore concern alternatives under common assumptions. What mission is being prioritized? What threats does each design assume? What support does it require? What happens when part of that support disappears? Which other missions receive less funding as a result?

Not every operational detail can be public. But the trade being made should be. Asking citizens to accept a different force structure requires more than telling them that admirals are conservative or that smaller ships are cheaper. It requires explaining what protection their money is intended to buy, including the protection it cannot buy.

A Network Has Dependencies, Too

Dispersal promises that an attacker cannot disable a whole force by destroying a few concentrated assets. Whether it delivers that promise depends on where the dependencies remain concentrated.

Several ships may share the same vulnerable source of information. Mobile weapons may still require maintenance and resupply. Independent units must have a way to distinguish useful orders from confusion. These are questions any proposal for distributed defense must answer, not findings that Taiwan's present arrangements necessarily fail.

The distinction matters because “networked” is often offered as the solution to the limits of smaller platforms. A network can distribute information. It can also distribute dependence on a connection that an opponent will try to interrupt.

Liao and Blair's emphasis on improving Taiwan's surveillance and command systems is consequently not a secondary recommendation. It is part of the burden of proof. Their proposed fleet must be assessed as an operating system of people, equipment and decisions, not as a shopping list of corvettes.

Their historical examples serve a similar purpose but cannot settle that assessment. They invoke Ukraine's Black Sea campaign as evidence that a weaker maritime force can constrain a stronger one without building a comparable fleet. That is their strategic interpretation, not an independent demonstration here that Taiwan could reproduce the result. Geography, outside assistance and the adversary's objectives have to be examined rather than carried across with the analogy.

The transferable question is whether an attacker must expose something essential that the defender can still threaten. The transferable answer is not that small platforms always win. History is useful when it identifies a condition to test. It becomes decoration when it supplies victory in advance.

The Opponent Can Choose Another Question

The most important counterargument to a fleet organized around invasion denial is not that invasion is harmless. It is that invasion is not the only form of coercion.

A June 2024 CSIS brief by Bonny Lin and colleagues explores hypothetical Chinese maritime quarantines. It distinguishes a law-enforcement-led quarantine from a military-led blockade. In its scenarios, the aim is not necessarily to seal Taiwan off completely. It is to impose controls on commerce and obtain compliance that demonstrates Beijing's claimed authority over the island.

These are scenarios, not descriptions of an operation now underway. The authors also acknowledge that such an undertaking would be complex, risky and dependent on the responses of Taiwan and other countries. Their analysis nevertheless exposes a different kind of vulnerability from an amphibious crossing.

If shipping companies comply with announced inspections because they fear disruption or penalties, coercive leverage could develop without an invasion fleet sailing. The dispute would involve commercial decisions, law-enforcement encounters and uncertainty about outside intervention, not just the ability to destroy military transports.

An anti-invasion weapon does not by itself answer what Taiwan should do when a purported inspection regime changes the behavior of shipping companies. Nor does the existence of that scenario establish that large frigates are the answer. It establishes that procurement has to be judged against more than one contingency.

This is the limit of denial as a complete political theory. Preventing an opponent's use of the sea does not automatically secure one's own access to it. Keeping an army from landing and keeping a society supplied are related tasks, but they are not identical.

A serious assessment of Liao and Blair's proposal would therefore ask what complementary capabilities are needed for coercion below open war, and what would be lost by reducing the larger surface fleet. Calling these questions distractions would make the argument tidier by making Taiwan's actual problem smaller.

Winning Is an Inadequate Measure

There is another reason to resist the confident language of naval solutions: even a successful defense could leave a society devastated.

In its January 2023 report The First Battle of the Next War, CSIS describes running a Chinese amphibious-invasion wargame 24 times. In most scenarios, the United States, Taiwan and Japan defeated the invasion and maintained an autonomous Taiwan. The same summary reports severe losses of ships, aircraft and service members, alongside devastation of Taiwan's economy.

Those are modeled outcomes under the project's assumptions, not a forecast, a probability estimate or a validation of the fleet Liao and Blair now propose. Their relevance is more basic. An outcome labeled “successful defense” can contain extraordinary destruction.

The civilian population is not an inconvenient qualification attached to the military result. Its ability to continue living and governing itself is the reason for the defense. A strategy that counts surviving launchers while treating civilian survival as somebody else's spreadsheet has mistaken its instruments for its purpose.

That does not make defensive preparation pointless. It makes the prevention of war, preservation of political choice and maintenance of essential civilian functions part of the evaluation. Deterrence should be assessed by whether it makes coercion less attractive, not by how satisfyingly a hypothetical engagement ends.

Nor should outsiders treat Taiwan's population as a demonstration project for their preferred naval doctrine. The people who bear the risks must remain participants in choosing them. Strategic coherence achieved by excluding their questions would be merely a well-organized imposition.

Liao and Blair offer a serious challenge to the assumption that a threatened navy must resemble the navy threatening it. Their proposal deserves testing against survivability, sustained operations, alternative forms of coercion and civilian consequences. It does not deserve either automatic adoption or dismissal because it disrupts a familiar fleet.

Taiwan need not win the fleet review to make an invasion fail. The harder requirement is to ensure that what survives the military comparison is the society the navy exists to protect.