Tsai Shih-Chieh's Perspective: 'The Dragon Slayer Becomes the Dragon' Part 1 – Poisoned Oil on the Table, Chiang Wan-an's Outrage on Ketagalan Boulevard

"He who fights with monsters should be careful lest he thereby become a monster. And if you gaze long into an abyss, the abyss also gazes into you." – Nietzsche, Beyond Good and Evil, Section 146

The so-called 'dragon slayer' refers to those who once demanded accountability from power; the 'dragon' is the one who, after gaining power, refuses to be measured by the same standard. This food safety crisis has revealed not just a single isolated scale from the dragon’s birth, but the layered armor of power—formed through shifting standards, delayed explanations, and evaded responsibilities. This first chapter peels back one scale from eight batch numbers: once critics enter the center of power, do they still accept the very yardstick they used to measure others?

Why should the central government be held responsible for the Chung Lien toxic oil incident? Because it controls national risk assessment, cross-county data integration, and product listing rules, it holds the 'final decision-making authority.' Whoever holds the final decision-making power must personally explain their decisions and bear the consequences. The central government cannot position itself at the top when setting rules, yet retreat behind local administrative divisions when accountability is demanded. Simply put, the flip side of final decision-making authority is the inability to transfer explanation duties and political responsibility to local authorities when problems arise.

Chapter One: Eight Batch Numbers, Four Turning Points – How a Single Drop of Oil Became a National Crisis

On June 30, 2026, Chung Lien Oils triggered a benzopyrene (BaP)超标 incident. The real question is not merely that one batch of oil tested positive for carcinogens, but why it triggered responses from three major oil manufacturers, hundreds of downstream businesses, school meal programs, local and central governments, judicial investigations, and street mobilizations. If a drop of oil were just oil, it wouldn’t shake society. A drop of oil becomes a storm because, as it flows through the supply chain, it gradually exposes systemic cracks layer by layer.

The first danger of this incident lies not only in BaP exceeding the legal limit for edible oils. It is even more dangerous because once the oil leaves the factory, it is processed into different products, incorporated into various recipes, and dispersed across brands, distribution channels, restaurants, schools, and households. Food safety cannot rely solely on numbers in oil tanks; batch numbers must be linked to products, and products traced to their distribution paths. The system must maintain an unbroken data chain; once data breaks, responsibility finds an escape route.

Therefore, the eight key batch numbers must not be compressed into a flat list: 315-1150404, 313-1150512, 315-1150510, 318-1150406, 314-1150410, 314-1150401, 313-1150408, and 315-1150413. Each of these numbers reveals different aspects: initiating the incident, overturning the single-batch explanation, exposing delayed reporting or missing entries in lists, and outlining the contours of seven batches of substandard oil. 314-1150401 was added to the preventive recall list due to underreporting and false registration, and must not be mistakenly written as the eighth confirmed超标 batch. Product超标, delayed reporting, list omissions, and criminal liability must be layered separately for responsibility to be properly assigned.

Batch numbers are not mere lists—they are nodes of responsibility.

Local authorities must answer for oversight within their jurisdictions; once the incident crosses county lines, distribution channels, and school systems, the central government must answer for national standards, information integration, and political accountability. The more concentrated the power, the less it can transfer responsibility downward.

The fury on Ketagalan Boulevard cannot be explained solely through political rhetoric. It must return to the institutional cross-section left by the eight batch numbers: how upstream crude oil, downstream products, end-user flows, local reporting corrections, and central decision-making connect—and where they break. This article uses only one yardstick: not allowing science to become a veil for power, not letting anger replace evidence, and not letting evidence lose its questioning of power.

To interpret the eight batch numbers, the first cut must be made on the supply chain: the first layer is the upstream raw material batches manufactured by Chung Lien; the second layer is the products processed, blended, and bottled by downstream businesses; the third layer is the end-use settings such as distribution, retail, restaurants, and schools. These three layers must be precisely matched; upstream batch numbers must not be used as substitutes for downstream product numbers, nor should flow volume be directly equated with consumption volume.

Once the supply chain layers are separated, the roles of the eight batch numbers become clear:

First Turning Point: The incident began with one batch of oil, but did not end with one batch

315-1150404 was the first batch publicly identified with超标 BaP (benzopyrene) levels, marking the beginning of the entire incident. This batch of Chung Lien soybean salad oil was manufactured on April 4, 2026, with approximately 1,300 metric tons distributed: about 588 tons to Fumao, 421 tons to Fushou, and 291 tons to Taishan. What turned this batch from a shipment record into a food safety alert was the subsequent test result: a BaP level of 8.1 μg/kg, exceeding the legal limit of 2.0 μg/kg for edible oils. Chung Lien Oils reported this to the Food and Drug Administration (FDA) on the evening of June 30, 2026, and provided additional clarification the next day (July 1). Since then, the figure 8.1 μg/kg has become the most prominent warning in the case; however, it is also easily misused to claim 'all Chung Lien oil is不合格.' A more rigorous interpretation of the evidence should be: the first batch opened the door to the incident, but each subsequent batch must be reconnected to its own test report, manufacturing date, and downstream flow.

Three responsibility lines must be distinguished: first, reporting—whether production was immediately halted, recalls initiated, and legal reporting conducted upon detecting anomalies; second, the product—whether the specific batch and sample超标; third, criminal liability—whether actions meet the elements of a crime. Administrative penalties can connect with criminal investigations but must not replace each other.

313-1150512 is the second batch. On July 7, the Changhua County Health Bureau received self-testing results from Taishan: 'Taishan Haoli Blended Oil 0.6L' tested positive for BaP at 2.5 μg/kg. Tracing back through procurement and formula records, the issue pointed to the same batch of raw material produced by Chung Lien on May 12. This batch was supplied to Fumao (369.49 tons), Fushou (529.19 tons), and Taishan (410.84 tons), totaling 1,309.52 tons. The alert thus expanded from upstream raw materials to downstream bottled products already on the market, and from a single oil tank to the products, inventory, and market flows of three oil manufacturers.

This traceability line also revealed Taishan's reporting responsibility. The Changhua County Health Bureau determined that Taishan knew of the anomaly as early as June 15 but failed to report it immediately, and did not fully disclose information during an inspection on July 1. The Changhua County Government therefore imposed a fine of NT$3 million for delayed reporting and another NT$3 million for inadequate cooperation with the investigation, totaling NT$6 million. This penalty targets the failure to report and cooperate after becoming aware of the issue, not a duplicate penalty for the 2.5 μg/kg test result.

315-1150510 is the third batch, emerging just two days after the second batch was revealed. On the evening of July 9, 2026, the FDA received a notification from Nan Chiao: their Chung Lien soybean salad oil, manufactured on May 10, tested positive for benzopyrene at 2.9 μg/kg, exceeding the legal limit. The Taichung City Government later announced that this batch was supplied to Fumao (353.8 tons), Fushou (706.25 tons), and Taishan (245.8 tons), totaling 1,305.85 tons, and led to 11 derivative products and 26 downstream product batch numbers. This is not just another batch number—it is the critical moment when the 'single batch' explanation officially collapses.

Here arises a notable temporal inversion: the third batch, 315-1150510, manufactured on May 10, predates the second batch, 313-1150512, manufactured on May 12. The terms 'first,' 'second,' and 'third batch' refer to the order of discovery and reporting, not manufacturing dates. As batches from April 4, May 10, and May 12陆续 emerge, raw material acceptance, process management, inspection frequency, and supply chain reporting must all be re-examined.

After the third batch emerged, the Ministry of Health and Welfare and the FDA judged that the incident might no longer be limited to a single batch, and thus expanded preventive recalls and sampling inspections. When the full scope of the incident is still unclear, the conservative principle of public health fears not acting too quickly, but allowing potentially contaminated oil to continue circulating. Subsequent comprehensive testing confirmed additional超标 batches, proving the concerns were not unfounded. When risks have already spread downstream, halting potentially problematic products should be the central authority’s first response.

Second Turning Point: The oil flowed out first, but test results reached health authorities late

The issue now shifts from 'which batch超标' to another timeline: who obtained abnormal information, when, and how long did it take before health authorities intervened? If test results remain internal to a company, they are merely quality alerts; only when they enter the legal reporting system do they become public actions to protect markets and consumers.

Nan Chiao’s self-testing is the first alarm on this timeline that cannot be overlooked. On April 14, Nan Chiao initiated raw material self-testing; on May 13, they obtained a BaP result of 4.6 μg/kg; on June 4, an outsourced test detected 8.0 μg/kg. Both results exceeded the legal limit of 2.0 μg/kg for edible oils. By May 13 at the latest, a clear warning had emerged within the supply chain; the second report on June 4 did not rule out the anomaly.

The subsequent notification sequence reveals the gap between corporate communication and public reporting. Nan Chiao notified Fushou on June 10; Fushou notified Chung Lien the next day. Three nodes must be distinguished: on June 30, Chung Lien reported 315-1150404 to the FDA; on July 2, Nan Chiao reported the same batch anomaly to the Taipei City Health Bureau, which promptly conducted an inspection and launched an 'Oil Product Testing Special Project'; on July 9, Nan Chiao reported another batch, 315-1150510, to the FDA, with BaP detected at 2.9 μg/kg. The first two nodes involve the initial batch and delayed