Case Study 22.2 — Han Tak Lee: The Same Bad Fire Science, and a Man Who Lived to Walk Out

Sourcing and tone. This case study draws on the public record of a Pennsylvania case (the 1989 cabin fire; the conviction for murder and arson; the release in the 2010s after the fire science was found no longer scientifically valid). It is the chapter's complementary arson case: where Willingham (Case Study 22.1) shows the discredited fire science producing an irreversible execution, Lee shows the same kind of discredited science producing a decades-long imprisonment that the courts were eventually able to undo. A man lost roughly a quarter-century of his life and his daughter died in the fire; we treat both soberly and stay within documented public facts, without inventing quotations, dates, or figures.

Background

In 1989, a fire destroyed a cabin in Pennsylvania, and Han Tak Lee's adult daughter died in the blaze. Lee was present. Investigators concluded that the fire had been deliberately set and that Lee had set it, and he was charged accordingly.

Lee was tried and convicted of murder and arson, and sentenced to life in prison. As with Willingham, the scientific spine of the case against him was the fire investigation — the era's standard origin-and-cause reasoning, built on burn-pattern interpretation and the indicator-based analysis that the field then treated as reliable knowledge. Lee maintained his innocence. He would spend roughly the next quarter-century in prison while, outside the prison walls, the science that convicted him was quietly falling apart.

The forensic evidence

The fire investigation in Lee's case relied on the same family of reasoning that this chapter has dismantled and that convicted Willingham — the debunked fire indicators (§22.3) and the interpretive habits the modern standard has repudiated:

• Burn-pattern and "pour-pattern" interpretation. Investigators read the fire's burn patterns — including patterns interpreted as showing the distribution of a liquid accelerant — as evidence that the fire had been deliberately set with poured ignitable liquid. As §22.1 and §22.4 establish, the low and irregular floor patterns the era read as "pour patterns" are routinely produced by ordinary fires that reach flashover, with no accelerant present.

• Accelerant and intensity inferences from the scene. As in the broader folklore of the period, features of the fire's behavior and the damage it left were read as indicating an accelerated, intentionally set fire — the same kind of unvalidated visual reasoning that controlled fire research has since refuted.

• The interpretive frame of the era. The investigation was conducted before the maturation of fire dynamics and before NFPA 921 had reshaped the standard of care; its conclusions reflected the confident, indicator-based practice that the field has since recognized as unreliable.

The crucial scientific fact, established over the following decades, is that fire science changed underneath the conviction. Controlled experimentation and the developing understanding of compartment fires, flashover, and ventilation established that the burn-pattern reasoning used against Lee did not reliably indicate a deliberately set fire. What had looked, in 1989, like a sound scientific basis for an arson finding was, by the 2000s and 2010s, recognized as discredited methodology.

What the evidence did — and didn't — establish

The structure of the lesson is the same as Willingham's, with one consequential difference in the ending. The fire science used against Lee did not validly establish that the fire was deliberately set. The burn patterns interpreted as "pour patterns" are produced by ordinary post-flashover fires; the era's intensity and accelerant inferences were not grounded in validated method. As fire science matured, the gap between what the original investigation claimed and what the evidence could actually support became undeniable.

In the post-conviction litigation, courts confronted that gap directly and reached a conclusion that fire-science reform had made available: the fire-science evidence underlying Lee's conviction was, in light of advances in fire science, no longer scientifically valid. This is a precise and important formulation. It does not declare, as a matter of metaphysical certainty, that the fire was accidental; it holds that the scientific basis on which the State proved arson had been discredited — that a conviction resting on that science could no longer stand. After roughly a quarter-century in prison, Han Tak Lee was released when the courts concluded the invalidated fire science could no longer sustain the conviction.

The difference from Willingham is the whole point of pairing the two cases. Lee was alive to benefit when the science changed. Because he was still imprisoned, the legal system — however slowly, however reluctantly — could reopen his case and act on the recognition that the fire science had collapsed. Willingham was executed before that recognition could save him. The same discredited methodology, the same maturing science that refuted it, two irreconcilable endings: one man walks out after decades, the other cannot be brought back. The variable that separated them was not the quality of the science used to convict — that was equally bad in both — but whether the defendant was still alive when the science was corrected.

The lesson

Han Tak Lee's case carries three lessons that complement and sharpen Willingham's:

1. The discredited fire science was not a one-jurisdiction aberration. The same indicator-based, "pour-pattern" reasoning convicted men in different states, in different decades, in different courtrooms. This was not one rogue investigator; it was a field-wide methodology that the whole discipline taught and believed — which is exactly why its correction required field-wide reform (controlled research, NFPA 921, the maturation of fire dynamics), not merely the discrediting of one expert. A junk method embedded in a discipline's standard practice produces wrongful convictions wherever that discipline operates.

2. "No longer scientifically valid" is the engine of fire-science exonerations — and it is a statement about validity, not innocence. Lee was not freed by new evidence that someone else set the fire; he was freed because the science that proved arson was withdrawn by the science itself. This is the validity spectrum (Theme 2) operating as a post-conviction remedy: when a method moves from "accepted" to "discredited," the convictions it produced become vulnerable. It is also a lesson in honest framing — the courts held that the conviction could not stand on invalid science, which is a narrower and more defensible claim than "the fire was definitely an accident."

3. The reversibility of imprisonment versus the irreversibility of execution (Theme 1, and Chapter 5, §5.1). Lee's release is, in the bleakest sense, the good outcome available when fire science corrects itself: a man wrongly imprisoned can, at least, be let out, even after a quarter-century stolen. Willingham's case is what happens when the same correction comes too late because the punishment was irreversible. Read together, the two cases are the most powerful argument in this book against allowing unvalidated forensic methods anywhere near an irreversible penalty — because science's self-correction, which is its glory, runs on a timescale the death penalty does not allow.

For the cold case, Lee and Willingham together form the standing warning that the Mill Creek fire investigation must heed: an arson finding built on "pour patterns" read by eye is the error that imprisoned Lee and executed Willingham. The cold case reaches arson the other way — on multiple origins and instrumentally confirmed ignitable liquid — precisely so that it does not repeat their fate.

Discussion questions

1. The burn-pattern reasoning used against Lee is the same kind that convicted Willingham. Using §22.1 and §22.4, explain why "pour patterns" interpreted by eye cannot reliably distinguish a deliberately set fire from a post-flashover accidental one.

2. Courts held that the fire science underlying Lee's conviction was "no longer scientifically valid." Explain why this is a more precise (and more defensible) basis for relief than "the fire was proven accidental." What is the difference between the two claims?

3. Lee was released; Willingham was executed; the fire science used against both was equally discredited. Identify the single variable that produced the different outcomes, and discuss what it implies about pairing irreversible punishments with forensic methods that may later be invalidated. (Chapter 5, §5.1.)

4. The same junk methodology convicted men in different states and decades. Why does a junk method embedded in a discipline's standard practice (rather than the misconduct of one examiner) produce wrongful convictions across many jurisdictions — and what kind of fix does that require? (Connect to Chapters 6 and 38.)

5. A skeptic says, "Lee's release just means a clever lawyer found a loophole." Refute this using the foundational-validity concept (Chapter 6, §6.5): explain why the withdrawal of the method's scientific basis is a substantive scientific development, not a procedural technicality.

6. Compare the roles fire science plays in Willingham, in Lee, and in the cold case. In which is the fire science the discredited basis of a wrongful outcome, and in which is it used validly to support an arson finding? What methodological features (affirmative evidence, confirmed ILR, NFPA 921 compliance) distinguish the valid use from the invalid ones?