Liebeck v. McDonald's, Retried

Thirteen of seventeen jurors find McDonald's not liable. Four find liability. We write for the thirteen, and we begin with the fact that makes this verdict unusual, because it is more interesting than our reasoning.

This jury did not read the same case. Case No. 001 was a blind split. A sticky fifty-fifty assignment gave each arriving juror one of two files at random. No juror chose a file, no juror could see the other, and no juror knew another existed. Nine received the case as the public tells it. Eight received the trial record.

Of the nine who read the popular version, none found liability. Of the eight who read the record, four did. Every juror who found against McDonald's had read the trial record. Not one juror who read only the familiar summary reached that conclusion.

Our majority is therefore two groups holding one verdict for what cannot be the same reasons: nine who were never shown a temperature specification or a burn history, and four who were shown both and were unpersuaded. We write only what both groups could have concluded from what was in front of them, and we mark where the two files diverge.

The narrow ground. On both files the causal chain runs through one voluntary act: the removal of the lid from a container of hot liquid held over a seated body. The lid is the safety device. It was defeated by the person it was protecting. Every juror, whichever file he or she held, was told that much, and no version of the record disputes it.

Coffee is sold hot because hot is what it is. A customer who buys coffee has asked for a liquid that will burn skin and has received one. Liability in tort attaches most readily to dangers a buyer cannot reasonably discover or guard against. It does not attach readily to the central and advertised characteristic of the thing sold. That proposition is available on either file, and it is the ground on which this verdict rests.

What we will not say. Our majority will not say the car was moving. It was not. The trial record establishes that the vehicle had pulled out of the drive-through lane and parked, and the eight jurors holding that record knew it. The folklore file's phrasing suggested otherwise, and the addendum to this case owns that defect in our own materials. A verdict resting on a car in motion would rest on nothing.

Nor will we describe the coffee as near boiling. The trial record puts the holding temperature at 180 to 190 degrees Fahrenheit, which is well below boiling. The file given to the record group also told those jurors that liquid at that temperature can cause third-degree burns in seconds. We state both figures as the file stated them, because the second is the plaintiff's case and the first is not, and a majority that overstated the temperature would be answering an argument nobody made.

The four in dissent. No juror in this case submitted written reasoning. We therefore do not know why the four voted as they did, and we decline to invent it for them. We record only what the ballots show: all four sat on the trial record, and none sat on the folklore file. That distribution is a fact about the ballots, not a finding about anyone's mind.

The 1994 court. That jury found liability, assigned Ms. Liebeck twenty percent of the fault, awarded $200,000 in compensatory damages which that finding reduced to $160,000, and added $2.7 million in punitive damages that the trial judge cut to $480,000. The parties later settled confidentially. One of those figures was before some of this jury and the rest were before none of it. The folklore file told its nine readers that the 1994 jury awarded "nearly $2.9 million", and stopped there. The breakdown, the comparative fault finding, the reduction, and the settlement were withheld from every juror until voting closed, and are set out in the addendum. We note them because a public jury that ignores the real court is playing a game, and we note their absence from our file because a public jury that pretends it saw them is doing something worse.

On the punitive award we say only this. Punishment requires proof of conscious indifference. Nine of the thirteen were shown no evidence bearing on that question at all. The other four were shown a good deal of it. A single opinion cannot speak for both conditions on that point, so this one does not try.

The verdict is NOT LIABLE, on the narrowest available ground: that a product sold hot, at the buyer's request, in a sealed container, does not become defective because the buyer opened it in her lap. Nothing here holds that a seller may heat its product without limit, or that a pattern of serious burns is a matter of no legal consequence. Four of this jury read that pattern and voted the other way, and the gap between the two halves of this jury is the most durable thing this case produced.

Four of seventeen jurors found McDonald's liable. None of them submitted written reasoning, and this house will not invent any for them. What follows is therefore not their opinion. It is the Court's own dissent: the case for liability that the record supports and that the minority vote implies, written by the house and owned by the house.

What the record contains that the story does not. Nine jurors were told a woman bought coffee, opened it, spilled it, and won a large award. There is little in that file to vote against. It is also, in one respect, inaccurate: it has her riding in the car when the record has the car parked. That is not a slip the house is proud of, but it is precisely what the popular account does, and the record file corrected it. The eight who received the trial record were told, in addition, that McDonald's corporate policy held its coffee at 180 to 190 degrees Fahrenheit when coffee at home is served at 135 to 140; that liquid at the higher figure causes third-degree burns in seconds while at 160 degrees a person has more than twenty seconds to react; that the company's own quality-assurance manager testified the coffee was not fit for consumption at the temperature served, because it would scald the mouth and throat; that more than seven hundred burn reports reached the company in the decade before this injury; that the company had settled earlier claims and changed nothing; and that a seventy-nine-year-old woman asked for roughly $20,000 to cover her bills and was offered $800.

Those are not atmospheric details. They are the elements of the tort. A known danger, a feasible alternative, notice of harm, and a decision to continue is what liability is made of. The story contains none of it. The record contains all of it.

Answering the majority on its own ground. The Opinion of the Court in this case rests on a single proposition, which it states plainly: the lid is the safety device, and the risk of what is inside passes to the hand that removes it. This dissent accepts that as far as it goes. The defense of the obvious risk works when the risk delivered is the risk expected. A customer expects a drink that will scald a tongue. The record describes a liquid that takes skin off in less time than it takes to react to a spill. Those are not the same danger, and the difference between them is a factual question the record answers and the story never raises.

What this dissent may not use, and is not using. The 1994 jury assigned Ms. Liebeck twenty percent of the fault, which reduced a $200,000 compensatory award to $160,000. Both figures favor this side of the argument, and neither was before any juror in this case. They were disclosed only after voting closed. We decline to build on them. A dissent leaning on facts its own jury never saw would commit the error it accuses the story of committing. This argument stands on the file the record group was given, or it does not stand.

The finding. Seventeen people were asked one question about one case. Assignment to the two files was random, sticky, and invisible to the juror. The nine who read the version everyone knows split zero to nine against liability. The eight who read the version the 1994 jury actually heard split four to four. Same case, same question, same ballot, and the assigned difference between the two groups was the file. The gap between the two groups was fifty points.

Seventeen is a small number and this house will not pretend otherwise; we do not treat a sample this size as a citable finding, and no reader should. But the direction is the point, and the direction is why this case was run as a split at all. The story produced a unanimous defense verdict. The record produced a hung jury.