In the United States Court of Appeals For the Seventh Circuit ____________________

No. 22-2556 ALUMINUM RECOVERY TECHNOLOGIES, INC., Plaintiff-Appellant, v.

ACE AMERICAN INSURANCE COMPANY, Defendant-Appellee. ____________________

Appeal from the United States District Court for the Northern District of Indiana, Fort Wayne Division. No. 1:17 CV 383 — William C. Lee, Judge. ____________________

ARGUED NOVEMBER 7, 2023 — DECIDED FEBRUARY 23, 2024 ____________________

Before EASTERBROOK, WOOD, and ST. EVE, Circuit Judges. EASTERBROOK, Circuit Judge. Aluminum Recovery Technol- ogies (ART) operates a smelter, recovering aluminum from scrap metal. It renovated and enlarged Furnace #4, which failed the very day it was put back into operation. Molten alu- minum escaped and damaged both the plant and the furnace itself. ACE American Insurance paid some of ART’s losses but not the cost of replacing the furnace’s refractory. (A refractory is a heat-resistant material that prevents molten metal from 2 No. 22-2556

damaging a furnace’s skeleton. In Furnace #4 the refractory was a mixture of ceramic and concrete.) Eventually all of the refractory from Furnace #4 was removed and replaced, at a cost approaching $400,000. In this suit under the diversity ju- risdiction the district court held that the insurer need not in- demnify ACE for that outlay. 2019 U.S. Dist. LEXIS 79873 (N.D. Ind. May 13, 2019). (Entry of final judgment was delayed while the court resolved other matters. Only the dispute about the refractory remains.) After the furnace failed, the insurer retained Engineering Systems, Inc., to determine why. Francisco Godoy, a mechan- ical engineer, conducted the investigation on its behalf. Go- doy proposed a protocol for the investigation; a team of inter- ested parties, including representatives of ART, unanimously approved that protocol. On the investigation’s first day the re- fractory nearest the site of the leak was removed. On the sec- ond day additional refractory was removed to extend the in- spection. Eventually ART had to tear out and replace all of the remaining refractory to put the furnace back in operation. When ART demanded indemnity for this expense, the in- surer invoked an exclusion in the policy. This exclusion pro- vides that: This Policy does not cover: … Any refractory lining or catalyst, except for damage or destruction directly resulting from the perils of fire, lightning, windstorm, hail, explosion … .

Engineering Systems concluded that faulty welding led the furnace’s frame to fail, allowing aluminum to escape. If this is so, then the policy’s exclusion blocks reimbursement for the cost of replacing the refractory. ART contends, by contrast, that an explosion in the furnace caused the structure’s failure. If that is so, then the exclusion does not apply. No. 22-2556 3

The district court sided with the insurer. The judge doubted that evidence would permit a reasonable jury to find that an explosion, if any, preceded the leak. One employee at the plant said that he heard a noise, which he deemed a sign of an explosion, about the time people first noticed molten metal outside Furnace #4. The insurer maintains that the sound came from the reaction of the molten aluminum with the plant’s floor, after the metal was already outside the fur- nace. Like the district court, we need not determine the se- quence, because even if we assume that an explosion of some kind came first, the policy’s exclusion applies unless that ex- plosion caused the leak. A straightforward way to show causation would be to pre- sent an expert witness who could identify the nature of the explosion and how it led the furnace’s frame to fail. Godoy, the insurer’s expert engineer, concluded that faulty welding explained the failure. William Sale of K-Industrial of Indiana, LLC, ART’s refractory contractor, agreed with this assess- ment. On the other side, ART offered … lawyers’ talk, but no evidence. The insurer maintains that all of the evidence about an ex- plosion, either before or after the aluminum’s escape, is hear- say, speculation, or both. To simplify matters we assume that an explosion occurred before the metal’s escape. The question remains: Did it cause the escape? (The contractual language is “resulting from”, a synonym for “caused by”.) ART’s princi- pal argument boils down to what in tort law would fit the Latin phrase res ipsa loquitur (“the thing speaks for itself”). Res ipsa loquitur is useful when there is only one plausible expla- nation—for example, when a piece of terra cotta falls off a building and an injured pedestrian contends that negligent 4 No. 22-2556

maintenance allowed the stonework to become loose. Yet for the failure of Furnace #4 there are at least two potential expla- nations: bad welding and an explosion powerful enough to rupture the furnace’s frame. Neither a judge nor a jury can determine which explana- tion is correct without the assistance of expert engineering ev- idence. Yet ART did not offer engineering evidence that sup- ports an inference of cause-by-explosion. Indeed, ART did not even try to estimate how powerful an explosion inside the fur- nace (likely the result of water flashing to steam) could be; Godoy thought it unlikely that much water would be inside the furnace. Sale testified in discovery that an explosion “could have” damaged the furnace but declined to venture an opinion about whether it did. The district court excluded even the “could have” part of Sale’s testimony, because ART had not disclosed Sale (or anyone else) as an expert during discov- ery. That left ART without useful evidence on the subject. At oral argument counsel for ART took a different tack, summed up by a different Latin phrase: post hoc ergo propter hoc (“after and necessarily because of”). The explosion pre- ceded the metal’s escape and therefore caused it, the argu- ment goes. This is not a good argument. Maybe an explosion occurred but was too slight to rupture the furnace. Maybe both the rupture and the explosion had a common cause, so that the explosion was a symptom of a deeper problem rather than a contributor to the failure. Post hoc ergo propter hoc is the name of a logical fallacy, not a reliable means of meeting one’s burden of proof. Whatever happens first can cause the next step in a chain of events, but more than timing is needed to support an inference of causation. We make this point fre- quently in employment-discrimination suits, holding that a No. 22-2556 5

plaintiff cannot win just by showing that a protected act pre- ceded an adverse employment action. See, e.g., Rozumalski v. W.F. Baird & Associates, Ltd., 937 F.3d 919 (7th Cir. 2019); Leon- ard v. Eastern Illinois University, 606 F.3d 428, 432–33 (7th Cir. 2010). The need for other evidence to make an inference of causation plausible is equally great elsewhere in the law. All ART offers, however, is the temporal sequence. ART has one more contention. It maintains that the insurer (or its expert, Godoy) engaged in unnecessarily destructive testing. As ART now sees things, the cost of repairing the re- fractory lining would have been under $10,000 had Godoy re- moved less of the material during his investigation. The in- surer must cover the excess expense, ART insists. This line of argument goes nowhere, however, because ART agreed to the plan of investigation, including the destructive testing. If ART’s representatives were unaware that the insurer would not cover the cost of rebuilding the refractory lining, that is ART’s own problem. Perhaps ART told its representatives to be cooperative in order to maximize the potential recovery under the policy—and the insurer did pay more than $400,000 for repair work plus extra operating expenses, balking only at the cost of replacing the refractory. Perhaps ART’s represent- atives consented because they believed (as the insurer con- tends) that recasting the whole refractory lining was inevita- ble once the furnace failed, no matter how the testing was done. We need not speculate about ART’s thinking, however. It is enough that ART agreed to Godoy’s proposed investiga- tive protocol. AFFIRMED