ARCO IND v. AMER MOTORISTS INS CO

BOYLE, J. (concurring).

I agree with the majority that the Court of Appeals improperly applied an objective standard of review in determining whether plaintiffs expected or intended the property damage that resulted from Arco Industries Corporation's use of volatile organic compounds (VOCS), and therefore should be reversed. I also agree that there is an inadequate basis to question the trial court's findings that American Motorists Insurance Company (AMICO) has a duty to indemnify its insureds.

I also agree with the remand order, although I would additionally direct the Court of Appeals to consider the applicability of the pollution exclusion present in three of the policies issued by AMICO. The trial court found the exclusions inapplicable, defining the phrase "sudden and accidental" as used in the pollution exclusion clauses to mean "unintended and unexpected." But see Upjohn Co v New Hampshire Ins Co, 438 Mich. 197, 207; 476 N.W.2d 392 (1991), defining "sudden" to include "a temporal element that joins together conceptually the immediate and the unexpected" (citation omitted). The Court of Appeals did not address this issue, finding that there was no coverage in the first place.

I write separately for two reasons. First, while the trial court correctly followed the direction provided by this Court in previous cases in evaluating whether the incidents at issue in this case constituted covered "occurrences" under the language of the comprehensive general liability (CGL) policy, the necessity for engaging in a two-step inquiry into whether there was an "accident" and whether damages were "intended or expected" when evaluating incidents under the standard terms of the CGL policy before us is unnecessarily duplicative. In the context of such a policy, the only analysis necessary in this regard is whether the damages were intended or expected from the standpoint of the insured. That inquiry sufficiently encompasses the threshold determination of "accident" as this Court has defined that term, and properly places the burden of proof on the question of coverage on the insured party. In addition, while intent is an element of the coverage analysis, primary emphasis in determination of coverage for environmental pollution damage is most often placed on the insured's expectation of damage. Second, I wish to make clear that my agreement with the majority that we should not disturb the trial court's finding that there was no intent to cause property damage or any expectation of such damage from 1972 to 1974, contrary to the conclusion of the dissent, is not based on a refusal to impute the collective knowledge of Arco's employees to the corporation, as I interpret the majority to do. Instead, I find the evidence on the record insufficient to question the trial court's holding on this issue.

I A

The applicable "Coverage" section of the policies issued by AMICO states that the insurer will pay damages its insureds become legally obligated to pay for covered bodily injury or property damage "caused by an occurrence." The "Definitions" sections of the policies define "occurrence" as follows:

"[O]ccurrence" means an accident, including injurious exposure to conditions, which results, during the policy period, in bodily injury or property damage neither expected nor intended from the standpoint of the insured. . . .

The majority concludes from our prior decisions that, in order to determine if there has been an occurrence, a court must evaluate whether there has been an accident, and then, if an accident has occurred, whether any resulting damage was expected or intended from the standpoint of the insured. Ante at 404-405.

I acknowledge that in other recent cases in which insurance coverage has been afforded for occurrences, we have engaged in the two-step analysis embarked upon by the majority. Auto Club Group Ins Co v Marzonie, 447 Mich. 624; 527 N.W.2d 760 (1994); Group Ins Co of Michigan v Czopek, 440 Mich. 590; 489 N.W.2d 444 (1992); Metropolitan Property Liability Ins Co v DiCicco, 432 Mich. 656; 443 N.W.2d 734 (1989). I therefore do not fault the trial court for following this direction. Federal courts applying Michigan law have similarly found the two-pronged analysis necessary in pollution coverage cases. Fireman's Fund Ins Cos v Ex-Cell-O Corp, 750 F. Supp. 1340 (ED Mich, 1990), but see Ray Industries Inc v Liberty Mutual Ins Co, 728 F. Supp. 1310 (ED Mich, 1989) (analyzing the issue of occurrence by focusing only on the insured's intent or expectation of damage), aff'd in part and rev'd in part on other grounds 974 F.2d 754 (CA 6, 1992).

In Ray Industries, the parties had stipulated that the insured did not actually intend or expect property damage. 728 F. Supp. 1315.

In all our earlier cases noted above, coverage was provided for an occurrence defined only as "an accident, including injurious exposure to conditions, which results, during the policy period, in bodily injury or property damage. . . ." It was then stated in the exclusionary section of the homeowners' policies that the insurers did not cover "bodily injury or property damage which is either expected or intended from the standpoint of the insured." DiCicco, supra at 672.

The question whether damage occurred during the policy period was not at issue in these cases, nor has it been argued to this Court in the case at bar.

I do not mean to suggest that the separate occurrence definition and intended or expected injury exclusion will always be present in homeowner's policies, only that this division was present in the homeowner's policies analyzed in prior cases.

While neither Auto-Owners Ins Co v Churchman, 440 Mich. 560; 489 N.W.2d 431 (1991), nor Allstate Ins v Freeman, the companion case to DiCicco, provided coverage for occurrences, instead simply stating that coverage was provided for bodily injury, personal injury, or property damage, both policies provided within their exclusion sections that expected or intended property damage or personal injury was not covered, under either an objective or subjective standard. See also Buczkowski v Allstate Ins, 447 Mich. 669; 526 N.W.2d 589 (1994) (interpreting an objective standard exclusion).

By articulating a definition for occurrence — which provides coverage — separate from a determination of expected or intended injury for exclusion of coverage, the policies in the cases noted dictated a two-pronged inquiry. We first had to determine whether coverage existed, applying the occurrence definition, before considering whether coverage was excluded because damage was intended or expected. "[T]he proper construction of a contract requires that we first determine whether coverage exists, and then whether an exclusion precludes coverage." DiCicco, supra at 668 (RILEY, J.).

In the present case, the definition of occurrence has been expanded to include what was exclusionary in the homeowner's policies. The question under the contract agreed to by the parties therefore is whether the damage for which Arco has become liable is covered by the policy, because it was caused by an "occurrence," i.e., an accident, including injurious exposure to conditions, that resulted, during the policy period, in property damage neither expected nor intended from Arco's standpoint.

This single inquiry into coverage under a CGL policy like the one before us allows a court to collapse the two-part investigation into the existence of an accident and whether damage was expected or intended, into a single question: whether the insured subjectively expected or intended damage. The two-pronged accident and intentional or expected damage test, under the current definition of accident employed by this Court, requires analysis of (1) whether there were unintended acts (wherein the results would clearly be unintended) or intentional acts with unintended or unexpected results, i.e., an accident, then (2) whether the damage resulting from the accident was subjectively intended or expected. The same question is therefore being asked twice. Where the sole issue is one of coverage, I do not find this circular analysis necessary. Where the standard definition of an occurrence in a CGL policy is used, as in the present case, I would only require the insured to demonstrate that damage transpired during the policy period and that it was neither expected nor intended by the insured in order to prove that coverage exists. "The newer definition of occurrence and accident [articulated in the post-1966 CGL policies] eliminates the need for an exact finding of the cause of damages so long as they are neither expected nor intended from the standpoint of the insured." 7A Appleman, Insurance Law Practice (rev ed), § 4493, p 49.

In Frankenmuth Mutual Ins Co v Piccard, 440 Mich. 539; 489 N.W.2d 422 (1992), the two-pronged analysis used by the majority here was applied to the same CGL occurrence definition as in the present case; however, no majority of this Court agreed with the substantive analysis in Piccard. It therefore is not binding precedent. "Plurality decisions in which no majority of the justices participating agree as to the reasoning are not an authoritative interpretation binding on this Court under the doctrine of stare decisis." Negri v Slotkin, 397 Mich. 105, 109; 244 N.W.2d 98 (1976).

As the majority correctly notes, an accident has been defined by this Court to mean "`"an undesigned contingency, a casualty, a happening by chance, something out of the usual course of things, unusual, fortuitous, not anticipated, and not naturally to be expected."'" Ante at 404-405. This definition incorporates both unintended and unexpected results. See 10 Couch, Insurance, 2d (rev vol), § 41:7, p 9.

The question when the damage occurred is one of whether coverage was "triggered." This normally contentious matter in pollution cases has not yet been addressed by this Court, and is not before us in the present case. See, e.g., Ray Industries Inc v Liberty Mutual Ins Co, 974 F.2d 754 (CA 6, 1992) (describing the four theories of trigger, and concluding that this Court would adopt a continuous trigger theory on the facts of that case, which concerned insurance coverage for environmental pollution); Ex-Cell-O Corp, supra (adopting the exposure theory in an environmental pollution case). Application of the "injurious exposure to conditions" clause in the occurrence definition has also not been argued in this case.

Once coverage is shown, the determination whether coverage is precluded by an exclusion in the policy would be required.

See Marzonie, supra at 652-662 (LEVIN, J., dissenting).

Justice LEVIN noted the circularity of the two-pronged accident and expected or intended injury analysis earlier this term in Marzonie, supra at 652-662 (LEVIN, J., dissenting). I will not repeat Justice LEVIN's analysis of this point, but note that I disagree with his conclusion that the presence of the expected or intended language in the exclusionary section of homeowner's policies like that in Marzonie, as opposed to its presence in the definition of a covered occurrence in the CGL policy in the present case, is of no significance. Id. at 659, n 24. It is the general rule of insurance contract interpretation, applied by this Court, that an insured bears the burden of proving coverage, while the insurer must prove that an exclusion to coverage is applicable. Williams v Detroit Fire Marine Ins Co, 280 Mich. 215, 218; 273 N.W. 452 (1937); Roddis Lumber Veneer Co v American Alliance Ins Co, 330 Mich. 81, 88; 47 N.W.2d 23 (1951); Ex-Cell-O, supra at 1348, 1350. Under the CGL policy language before us, the burden is on the insured to prove that damage was neither expected nor intended. This issue is all encompassed within the issue of coverage, with any successive questions of exclusion being the burden of the insured to prove. Under a homeowner's policy requiring the two-pronged analysis, it would be the insured's burden to prove that an accident occurred, and the insurer's burden to prove that damage was expected or intended. While the question may be the same to both parties in analyzing application of the homeowner's policy to a particular set of facts, the insured and insurer would have successive burdens, dictated by the policy, to prove the intent or expectation of damage.

A commentator analyzing the definition of "occurrence" soon after it was added to CGL policies made the following observation, applying an objective standard of expectation:

Let us assume for the purpose of argument that the expression "which results, during the policy period, in bodily injury or property damage neither expected nor intended from the standpoint of the insured" modifies the word "accident." We could then make an interesting observation. An "accident" is a negligent act which is not intentional and which causes unprecedented or unforeseeable injury or damage. It is obvious that "accident" is not something wilfully or deliberately caused. If the words "neither expected nor intended" refers to some conditions or damage accidentally caused but not caused by accident, then the words are nonsense. [Even, The corporate insurance administrator — Problems with the 1966 revised liability policy, 3 Forum 95, 104-105 (1967).]

But see Shell Oil Co v Winterthur Swiss Ins, 12 Cal.App.4th 715, 748-752; 15 Cal.Rptr.2d 815 (1993) (holding, inter alia, that "accident" in a CGL policy is the direct and immediate causal event of resulting property damage, and constitutes a separate inquiry from whether damages were intended or expected). I do not address the merits of the California court's construction of the accident definition, because this Court has clearly held that an "accident," analyzed in terms of whether there has been an occurrence, includes intentional acts resulting in unintended harm. DiCicco, supra at 670. "If the resulting damages can be viewed as unintended by a fact finder the result constitutes an `accident' for purposes of the liability insurance policy; it is the quality of result rather than the quality of the cause that is controlling." 7A Appleman, supra, § 4492.02, p 31 (citations omitted).

In the present case, the issue of the standpoint from which the happening of an occurrence should be viewed has been clearly stated in the policy as being that of the insured. Czopek, supra at 608-609 (BOYLE, J., concurring).
I agree with the majority in the present case that a majority of this Court has recently found that, where an insurance policy is silent with regard to the standpoint from which an accident should be evaluated, such as in the homeowner's policies in DiCicco, Czopek, and Marzonie, the standpoint of the insured should be applied. Ante at 405. If an accident in such instances was viewed from the standpoint of the injured party, however, see Czopek, supra at 608-612 (BOYLE, J., concurring), the two-pronged inquiry dictated by the homeowner's policies would examine two distinct questions: 1) was there an accident from the injured party's perspective, and 2) was damage intended or expected from the standpoint of the insured.

Not all courts have followed this rule [that an accident should be evaluated from the standpoint of the insured]. It has been traditional to deny insurance protection for deliberate, outrageous or irresponsible behavior as a form of punishment to the wrongdoer. One judge suggested that such a theory follows "the sporting theory of justice." It ignores the fact that most of the actors involved in irresponsible behavior are also financially irresponsible and that it is the injured victim that is really being punished. If punishment is the objective of these rulings surely some other way of doing so can be found. To look at only the insured's behavior ignores the fact that insurance also is for the benefit of injured victims of accidents. They are the third party beneficiaries of the insurance contracts and their situation should not be ignored. The interpretation of an insured's conduct is not only a dispute between the insured and his insurance company. Some courts have recognized this and consider the resulting injury from the point of view of the victim. If it was accidental from that point of view, the loss will be covered by the liability insurance issued to the actor. From the standpoint of the aggressor an injury willfully inflicted upon another is not an accident but from that of the victim of an uninvited and unprovoked aggression, the injury is accidental. [7A Appleman, supra, § 4492.02, pp. 27-28.]

This perspective should not require the insurer to indemnify an insured for liability it did not contract to cover. For example, where damage is the result of inherently dangerous behavior engaged in with criminal intent, exclusion of coverage due to expectation, if not intentional infliction of injury may be apparent, even if the policy does not include an express criminal acts exclusion. See, e.g., Continental Western Ins Co v Toal, 244 N.W.2d 121 (Minn, 1976). In Toal, the Minnesota Supreme Court found error in a jury instruction that an occurrence definition identical to that in the present case barred recovery for injuries that were the natural and probable consequences of the insured's intentional act, or that merely foreseeable injuries were excluded. Nonetheless, the court held that intent to inflict injury, precluding insurance coverage, could be inferred as a matter of law, where the insureds participated in the armed robbery of a bowling alley that resulted in the death of an employee. Id. at 125-126. "[T]he insured need not intend the actual bodily injury inflicted in order to fall within the instant exclusionary clause. It is sufficient that the factfinder conclude the insured subjectively expected some type of harm reasonably foreseeable from the insured's standpoint." DiCicco at 731, n 11 (ARCHER, J., concurring in part and dissenting in part). Assertions by an insured in such a case that damage was not intended or expected appears to be a denial of expected harm that "flies in the face of all reason, common sense and experience. . . ." Id. at 720 (BOYLE, J., concurring in part and dissenting in part) (citation omitted). But see Marzonie, supra at 666, n 36 (LEVIN, J., dissenting) (citing cases in which insurance coverage for injuries inflicted by insureds committing criminal acts was a question of fact).

B

As noted above, I do not fault the trial court in the present case for engaging in the two-pronged analysis of accident and intended or expected damage that this Court has previously stated is necessary. We provided that direction, however, without making a distinction between policies that separate the accident question and the intended or expected damage question between issues of coverage and exclusion, and policies with a standard occurrence definition encompassing both issues. The trial court's analysis of the case under this two-pronged test resulted in findings that satisfy the single intended or expected damage inquiry. While the trial judge did not expressly reference the burden of proof, he made an affirmative holding that property damage was not expected or intended by the insureds, rather than finding that there was insufficient proof of intent or expectation. I interpret the trial court's statement that it found no intent to contaminate the environment to mean that it found that Arco did not know that the leaching process it employed would contribute VOCS to the ground water. I thus agree with the majority that the Court of Appeals incorrectly reversed the trial court's decision.

If the trial court had found only that Arco did not intend or expect the precise contamination that actually resulted from its process, but did expect some property damage, then the policies would not have provided coverage for Arco's liability. A literal "intent to contaminate the environment" was not necessary to find a subjective intent or expectation of property damage under the contract.

II

In determining whether damage was caused by an occurrence where the insured is a commercial entity, the focus is primarily on whether the insured expected damage. Rynearson, Exclusion of expected or intended personal injury or property damage under the occurrence definition of the standard comprehensive general liability policy, 19 Forum 513, 514 (1984). In the context of environmental pollution, it will be the rare case in which the proofs will include testimony that an insured actually intended to cause property damage.

A determination whether an insured subjectively expected damage to occur from its actions clearly takes the inquiry out of the realm of mere reasonable foreseeability. In determining subjective expectation, we need not rely solely on the representations of the insured with regard to his subjective state of mind. Expectation can be inferred from the nature of the act in question and the high degree of probability, i.e., a substantial probability, of resultant damage. It is in such cases that an insured's claim that no injury was expected or intended "flies in the face of all reason, common sense and experience. . . ." DiCicco, supra at 720.

Even those courts that advocate an objective inquiry into intent and expectation reject the suggestion that a mere reasonable foreseeability of damage is sufficient to bar coverage. City of Carter Lake v Aetna Casualty Surety Co, 604 F.2d 1052, 1058-1059 (CA 8, 1979). In City of Carter Lake, the court held that coverage should be excluded because of the objective expectation of damage only where "the actor knew or should have known that there was a substantial probability" of damage, defining substantial probability as indications of damage strong enough to warn the reasonably prudent actor that damage was "highly likely" to occur. The amici curiae in the present case who advocate that this Court adopt the objective standard of inquiry acknowledge that the policy dictates that the reasonable person be one who stands in the shoes of the insured, i.e., views the likelihood of damage "from the standpoint of the insured." Such observations lead me often to conclude that the objective versus subjective expectation debate, where subjective expectation can be proven by circumstantial evidence, generates more heat than light. Neither standard relies solely on the admissions of the insured in order to find an expectation of damage. In City of Carter Lake, the court held that when sewage backed up once and no action was taken by the insured to head off future incidents, there was a substantial likelihood of the occurrence of the second through sixth backup so as to preclude coverage for the subsequent incidents, because of an objective expectation of damage. It seems to me the court could have just as easily found that the evidence was sufficient (one backup, and nothing done to correct the problem) to find a subjective expectation of damage, despite representations by the insured to the contrary. See also Bituminous Casualty Corp v Tonka Corp, 9 F.3d 51 (CA 8, 1993) (the insured was found to have objectively expected property damage where the insured routinely dumped contaminant containing sludge on site, the insured admitted knowledge of contaminants in the sludge, and the insured and its employees had knowledge of the harmful character of the contaminants).

The question of "intended or expected" damage from the standpoint of the insured thus essentially mirrors the definition of "intent" in tort law. 1 Restatement Torts, 2d, § 8A, p 15 (defining "intent" to mean "that the actor desires to cause consequences of his act, or that he believes that the consequences are substantially certain to result from it").

In the present case, the trial court found that there was no subjective expectation of damage for, two reasons. First, the court believed the insureds' testimony that they had no such expectation, and, second, the court found that there was insufficient knowledge possessed during 1968 through 1974 by either the insureds or the general community of experts in environmental pollution of the likelihood of property damage caused by release of VOCS to be able to infer that such expectation was subjectively possessed by the insureds. Therefore, even taking into account the collective knowledge of all of Arco's employees, Upjohn Co v New Hampshire Ins Co, 438 Mich. 197, 213; 476 N.W.2d 392 (1991), there was insufficient knowledge of the dangers of VOC contamination to provide sufficient evidence of expectation of property damage. I do not find the admission by Arco's chemist that he knew in 1972 that release of VOCS into the company's seepage lagoon could result in property damage sufficient evidence of an expectation that property damage would occur to question the conclusion of the trial court.

III

In summary, I agree that the judgment of the Court of Appeals should be reversed, the findings of the trial court reinstated, and the case remanded to the Court of Appeals for further consideration. Where a CGL policy, including within its definition of a covered occurrence the requirement that there be an accident and that damage be neither intended nor expected from the standpoint of the insured, the only issue in determining if coverage exists, other than whether damage occurred during the policy period, is whether the insured has fulfilled the burden of proving that it neither subjectively expected nor intended the damage. When the insureds are commercial parties, this analysis will normally focus on the insureds' expectation of damage. In the present case, I find insufficient evidence to question the trial court's finding that the insureds neither expected nor intended property damage during the entire period covered by the policies issued by AMICO.

RILEY, J. (dissenting).

The dispositive issue in this case is whether Arco intended or expected to contaminate the environment through its actions during 1968 through 1974, the period in which defendant, American Motorists Insurance Co (AMICO), insured it. The majority found that Arco did not intend or expect that its actions within this time frame would contaminate the environment. As a result, the majority concluded that an occurrence within the definition of Arco's insurance policy occurred and that AMICO should be liable for coverage.

I agree that Arco did not expect or intend to injure the environment during 1968 through 1972, but I have come to a different conclusion regarding the years spanning 1972 through 1974. In 1972, Arco hired Ed Koperdak, a chemist, who was aware that the volatile organic compounds (VOCS) used at the plant posed a serious danger to the environment. This knowledge was then imputed to the corporation under Upjohn Co v New Hampshire Ins Co, 438 Mich. 197, 213; 476 N.W.2d 392 (1991). Although Koperdak may not have been aware that these VOCS were being dumped into the drains that led to the seepage lagoon, other employees were aware, and their knowledge likewise was imputed to the corporation. Consequently, from 1972 through 1974, Arco was aware that it was dumping harmful VOCS into the lagoon. As a result, an occurrence within the definition of Arco's insurance policy did not occur, and AMICO should not be held liable for the damage caused to the environment during those years.

I

The insurance policy that this case revolves around was issued by AMICO to Arco in 1968. The policy was a comprehensive general liability policy which provided:

[AMICO] will pay on behalf of the insured all sums which the insured shall become legally obligated to pay as damages . . . caused by an occurrence, and the company shall have the right and duty to defend any suit against the insured seeking damages on account of such bodily injury or property damage, even if any of the allegations of the suit are groundless, false or fraudulent, and may make such investigation and settlement of any claim or suit as it deems expedient, but the company shall not be obligated to pay any claim or judgment or to defend any suit after the applicable limit of the company's liability has been exhausted by payment of judgments or settlements. [Emphasis added.]

Under the language of this policy, AMICO's liability is triggered whenever an "occurrence" transpires. Fortunately, "occurrence" is defined within the policy as:

[An] accident, including injurious exposure to conditions, which results, during the policy period, in bodily injury or property damage neither expected nor intended from the standpoint of the insured. [Emphasis added.]

Unfortunately, the pivotal term "accident" is not defined within the policy. This is not an uncommon problem because in many insurance policies "accident" is not defined. When this happens, generally the "commonly used meaning" simply controls. Group Ins Co of Michigan v Czopek, 440 Mich. 590, 596; 489 N.W.2d 444 (1992). In fact, this Court in Auto Club Group Ins Co v Marzonie, 447 Mich. 624, 631-632; 527 N.W.2d 760 (1994), recently analyzed the common meaning of "accident" and found: "`an accident is an undesigned contingency, a casualty, a happening by chance, something out of the usual course of things, unusual, fortuitous, not anticipated, and not naturally to be expected.'"

Hence, the first question becomes whether or not the VOCS found their way to the seepage lagoon by means of an accident. The answer to this question is no. Ruth Kelly, an employee at Arco, testified that from about 1971 until 1973 she worked in the tool crib area cleaning glue guns. She would clean these guns by disassembling them and placing them into buckets of MEK, a powerful solvent. When these buckets became too dirty, she would simply dump their contents into the drains as her foreman had instructed her to do.

A. We used MEK for cleaning the glue guns.

Q. For cleaning the glue guns?

A. Yes.

Q. Was there a machine that cleaned the glue guns?

A. No. I did it myself. You tear 'em down and remove the o-rings and stuff that can't go in the MEK. And then you put the metal parts into the gallon pails of MEK to soak the glue off of 'em.

* * *

Q. And you — would put the glue gun into the pail. Is that correct?

A. Right.

* * *

Q. Would you have to replace the MEK in the pails, periodically?

A. Yes.

Q. And why was that?

A. Because it got dirty.

Q. With . . .

A. With the glue in it.

* * *

Q. How would you dispose of that dirty MEK?

A. The foreman in there told me to dump it in the drain.

Q. Which drain?

A. This one right here.

* * *

Q. Do you recall how often these gallon pails would get dirty and have to be replaced? The solvent, I mean.

A. I would say maybe every other week. It would depend on how many times you had to clean the gun.

The fact that Kelly was intentionally dumping these solvents every other week into the drains that led to the lagoon is by no means an "undesigned contingency" or "happening by chance." It was a conscious, intentional choice on her part and on her foreman's part because he told her to dump the buckets in this manner.

However, this was not the only time that Kelly intentionally brought about the flowing of VOCS to the lagoon. In 1973, Kelly began work on the epoxy machine where she had more experiences with VOCS. While working on the epoxy machine, one of her jobs was to use perchlor to clean up the epoxy that dripped from the machine every day. She deliberately poured this solvent onto the floor in order to loosen the epoxy from the floor. She would then mop up this epoxy-perchlor mixture right over the drains.

Q. In this process would some of the epoxy or vinyl drip or spill onto the floor?

A. Yes.

Q. How often?

A. Every day.

Q. And how close was the epoxy machine to the drain?

A. It sat right behind the drain. . . .

Q. And as a [sic] epoxy machine operator was it your responsibility to clean the floors around the machine?

A. Yes. It was.

Q. How often would you have to clean those floors?

A. Sometimes two or three times a day.

Q. And do you recall whether you used a solvent to do that?

A. Yes. At the time I think it was Perchlor we were using?

Q. Perchlor?

A. Yes. Q. Where would you get that Perchlor?

A. From Compounding.

Q. When you would mop the floors with the Perchlor would that Perchlor go down the drains?

A. Oh, yeah. Because it sat right there where you were mopping over the drain again.

Kelly's act of mopping up the floors over the drains with perchlor was not uncommon — many of the other employees also cleaned the floors in this manner. In fact, the Court of Appeals in its decision wrote:

Former Arco employees also testified that VOCS were used to mop the plant floors from at least 1964 to 1979. The mopping was often performed on and around the drains that led directly to the unlined lagoons. Given the frequency and volume of mopping required to clean the plant floors, some of the VOCS would invariably go into the drains and be washed into the lagoons. [ 198 Mich. App. 347, 352; 497 N.W.2d 190 (1993).]

Koperdak also testified about the use of volatile organic compounds to mop the floors of the plant.

Q. They would come into compounding and sometime use a requisition to indicate we want the mop water; is that correct?

A. Yes.

Q. And they would be given a volatile organic compound as that mop water, correct?

A. Yes.

Q. And this is from when you started at Arco in 1972 at least up until 1979; is that correct?

A. Yes.

These employees, through their actions, were intentionally bringing about the placement of the VOCS into the seepage lagoon. Their acts were not simply "happenings by chance" or "undesigned contingencies." As a result, their acts were not accidents.

II

The next question becomes, Could Arco have expected the harm that occurred? The answer to this question lies in whether or not Arco expected that the VOCS were harmful and whether they expected that they were going into the seepage lagoon.

A

Arco, without question, knew that these VOCS were harmful because in 1972 it hired Ed Koperdak as a chemist. Koperdak admitted at trial that when he was hired he knew that the VOCS being used at the plant posed a distinct danger to the environment.

Q. You understood, did you not, when you came to Arco that it was important that volatile organic compounds not be discharged in the seepage lagoon, correct?

A. Yes.

Q. And you knew that . . . if those volatile organic compounds were discharged into the seepage lagoon they could degrade the environment; isn't that correct?

A. Yes.

The fact that Koperdak possessed this knowledge is critical to the outcome of the case because under Upjohn this knowledge can be imputed to the corporation.

"[K]nowledge acquired by employees within the scope of their employment is imputed to the corporation. In consequence, a corporation cannot plead innocence by asserting that the information obtained by several employees was not acquired by any one individual employee who then would have comprehended its full import." [ Upjohn, supra at 213, quoting United States v TIME-DC, Inc, 381 F. Supp. 730, 738 (WD Va, 1974).]

Since Koperdak was employed as a chemist, any knowledge he acquired in regard to these VOCS would be within the scope of his employment and imputed to Arco. Hence, Arco knew that the VOCS were harmful to the environment.

B

Arco also knew that the VOCS were going into the seepage lagoon. Koperdak prepared a memorandum on July 27, 1977, indicating that such VOCS should not be dumped into the pond.

Q. To you, the purpose of this memorandum was to insure, was it not, that particular compounds didn't make their way into the Arco pond; is that correct?

A. Yes.

Q. It wasn't because you were worried about — this wasn't primarily designed out of a concern for accidents within the plant that might harm employees; is that correct?

A. Yes.

Q. So your concern about ammonia was also — or your primary, key concern with ammonia was it getting into the drains and getting into the pond?

A. One of the concerns, yes.

Q. And your concern with solvents was solvents getting into the drains and getting to the pond?

A. Yes.

However, Koperdak claims that he prepared this as just a general warning and not in response to any problem he was aware of.

Q. In other words, when you came out with this memorandum you had never heard from anyone or had any knowledge or employees dumping any of these compounds in the drains?

A. No.

* * *

Q. Did you also send around memos telling employees not to break windows or tip over machines, even though that had never occurred in the past?

A. No.

In my opinion, Koperdak was aware that these VOCS were leaking into the pond and it was this knowledge that inspired him to prepare the memorandum. In any event, even if he had no knowledge of these spills the result should remain the same since other employees were aware that these VOCS were being placed into the seepage lagoon.

As stated before, Arco's employees were placing VOCS into the drains that led out to the lagoon by either mopping over the drains with perchlor or dumping MEK directly into the drains. This knowledge was then imputed to Arco under Upjohn.

The majority, however, seems to argue that a specific employee has to know both that VOCS are harmful and that they are being discharged into the lagoon in order for Arco to be able to expect that such harm will occur.

Although it is tempting to infer intention or an awareness of the dangers of VOCS, there simply was no testimony that any of Arco's employees released VOCS into the environment with the intention to harm the environment. Furthermore, there was no testimony that anyone should have expected that harm would result from their actions. [ Ante, p 415.]

This belief by the majority that an employee has to know of the harm and of the discharges in order for Arco to expect the damage, however, is simply not true. Arco is a corporation, and under Upjohn, a corporation, in essence, is an entity that possesses the combined knowledge of all its employees.

Furthermore, we reject Upjohn's assertion that the information obtained by several of its employees was not acquired by any individual employee who then would have comprehended its full import. . . . Rather, the Upjohn Company is considered to have acquired the same collective knowledge of its employees. . . . [ Upjohn, supra at 215-216.]

Since Arco knew, through Koperdak, that these VOCS were harmful to the environment and since Arco knew, through Ruth Kelly and the other employees, that the VOCS were going to the seepage lagoon, it knew and expected that such harm would occur.

This expectation of harm is very similar to the expectation possessed by the defendant in Marzonie. In that case, Oaks fired a shotgun at an automobile. Unfortunately, the driver of the vehicle was struck by the blast and severely injured. Oaks indicated that he did not intend to hurt anyone and as a result should be covered under his homeowner's insurance policy because the event occurred on his premises. This Court, however, found otherwise concluding that Oaks could have expected that his act of firing a shotgun at an occupied vehicle would injure someone. "[W]e find that Oaks should have expected bodily injury to result from firing this shotgun at Marzonie's occupied vehicle." Marzonie, supra at 644. As a result, an occurrence within the definition of Oaks' insurance policy did not happen. Much like Oaks, Arco intentionally set loose something very dangerous — VOCS. These VOCS, like the pellets from Marzonie's shotgun, were headed toward a place where they would cause great harm, the environment.

The majority argues, however, that the facts here are similar to Frankenmuth Mutual Ins Co v Piccard, 440 Mich. 539; 489 N.W.2d 422 (1992). In that case, Piccard intentionally set fire to his own store. While firemen were fighting the blaze, a fire fighter fell and was severely injured. This fireman then brought suit against Piccard, and Piccard's insurance company brought the action to determine its liability to defend Piccard, its insured. The majority writes that this Court held in Piccard:

that the intentional act of setting the fire could still be defined as an "occurrence" as long as the act was not committed with the intention or expectation that the injury to the fireman would occur. [ Ante, p 416.]

The majority then attempts to draw a parallel from Piccard to the present case.

Similarly, in the instant case, even though the act of introducing VOCS into the environment may have been intentional, there was no evidentiary support establishing that the acts were intended or expected to harm the environment. [ Id.]

The majority, however, is incorrect. Arco knew that these VOCS were harmful and were headed toward the unlined seepage lagoon. If Piccard knew that by setting the fire a fireman would fall from a roof and be injured, then surely he would have expected the injury and no accident would have occurred. However, that was not the case and that is why Piccard is distinguishable.

III

CONCLUSION

It is my opinion that Arco did not intend or expect to harm the environment from 1968 until 1972, so AMICO should be held liable for the damages that occurred during those years. However, from 1972 through 1974, Arco knew that VOCS were being deliberately placed into the drainage ditches that led out to the lagoon. It also knew that these VOCS posed a danger to the environment. Hence, it knew that its actions would cause serious environmental harm. As a result, the damage that happened to the environment from 1972 until 1974 was expected by Arco and was not an accident. Thus, an occurrence within the definition of Arco's insurance policy never happened and AMICO's liability was never triggered during those years.

WEAVER, J., took no part in the decision of this case.