Upjohn Cov.New Hampshire Ins. Co.

Supreme Court of MichiganAug 26, 1991
438 Mich. 197 (Mich. 1991)
438 Mich. 197476 N.W.2d 392

Docket Nos. 86906-86908.

Argued January 10, 1991 (Calendar No. 9).

Decided August 26, 1991. Rehearing denied 439 Mich. 1202.

Denenberg, Tuffley, Bocan, Jamieson, Black, Hopkins Ewald, P.C. (by Julius Denenberg, William G. Jamieson, George F. Curran, III, and Dana L. Ramsay), for plaintiff The Upjohn Company.

Garan, Lucow, Miller, Seward, Cooper Becker, P.C. (by Thomas W. Emery); ( Philip J. McGuire, of counsel), for defendant Allstate/NESCO.

Amici Curiae:

Hooper, Hathaway, Price, Beuche Wallace (by Bruce T. Wallace); ( Wiley, Rein Fielding, by Thomas W. Brunner, Laura A. Foggan, and Robert R. Lawrence, of counsel) for Insurance Environmental Litigation Association.

John D. Noonan ( Andrew P. Buchsbaum, of counsel) for Environmental Law Society, University of Michigan, Michigan Environmental Council and Public Interest Research Group in Michigan.

Dykema, Gossett (by Roger K. Timm, Stephen D. Turner, and John A. Ferroli) for South Macomb Disposal Authority. Honigman, Miller, Schwartz Cohn (by Philip A. Grashoff); ( Covington Burling, by William H. Allen, William F. Greaney, and Dorothy C. Mitchell, of counsel) for Dow Chemical Company, Product Liability Advisory Council, American Fiber Manufacturers Association, American Petroleum Institute, Chemical Manufacturers Association, E.I. DuPont de Nemours Company, International Business Machines Corporation, and Olin Corporation.

Rosi, Olson Levine, P.C. (by James M. Olson and Barry L. Levine), for Jan and Sandra Doyen doing business as Carefree Cove Resort, and Robert L. Johnson.

Honigman, Miller, Schwartz Cohn (by Jay E. Brant, Philip A. Grashoff, Jr., Mark A. Goldsmith, and Daniel G. Helton) for Arco Industries Corporation, Bronson Plating Company, Hitachi Magnetics Corporation, Rengo Oil Company, Thomas Solvent Company, and Thermofil, Inc.

Hill, Lewis (by Richard C. Sanders); ( Anderson, Kill, Olick Oshinsky, P.C., by Eugene R. Anderson, Kevin J. O'Brien, and Bruce A. Brown, of counsel), Martha A. Churchill, General Counsel, Mid-America Legal Foundation ( Varnum, Riddering, Schmidt Howlett, by Matthew W. Zimmerman, Mark S. Allard and Rosi, Olson Levine, by Barry Levine, of counsel), for the Budd Company, the City of Clare, the City of Evart, Harrow Products, Inc., Mid-America Legal Foundation, and Traverse City Light and Power.



RILEY, J.

In this case, we are asked to interpret and apply the pollution-exclusion clause found in the comprehensive general liability policy of Allstate Insurance Company. The resolution of this issue requires us to answer two questions: (1) is the phrase "sudden and accidental" which appears in the pollution-exclusion clause and which creates an exception to the exclusion, unambiguous, and (2), if unambiguous, what is the proper meaning of the phrase "sudden and accidental."

The Allstate Insurance Company asked this Court to consider three other issues along with the pollution exclusion. These issues were as follows: (1) whether there was an "occurrence" under the policy; (2) whether recovery costs are "damages"; and (3) whether an "other insurance" clause in the policy was applicable. However, this Court limited the scope of the granted issues to the interpretation and application of the pollution-exclusion clause by an order dated December 27, 1990.

We hold that the phrase "sudden and accidental" is unambiguous. Furthermore, we find that the definition of "sudden" includes a temporal element as well as a sense of the unexpected, and that "accidental" means unexpected and unintended.

We, therefore, find that the Court of Appeals erred in holding that the pollution-exclusion clause did not apply under the facts of this case. Accordingly, we reverse the decision of the Court of Appeals and find, as a matter of law, that the pollution-exclusion clause applies and, therefore, plaintiff, The Upjohn Company, is not entitled to coverage under defendant Allstate's policy.

I. FACTS AND PROCEEDINGS

On August 13, 1982, The Upjohn Manufacturing Company (UMC), a Puerto Rico-based division of The Upjohn Company, began its annual production of clindamycin, an antibiotic. Two toxic byproducts were produced in the clindamycin campaign. These chemicals were pumped into an underground storage tank designated FA-129 which had a ten thousand gallon capacity.

Each weekday of the year, an Upjohn employee measured the level of by-product in the tank. The employee recorded the measurement on a tank farm inventory sheet and turned the sheet over to his supervisor. Each day's sheet was reviewed, and compared with previous days' sheets which were kept on file at the UMC facility. This was Upjohn's standard procedure.

Prior to August 16, 1982, the tank-level measurements remained constant at ten inches or 475 gallons. However, on August 16, 1982, the same day that Upjohn had pumped its first batch of approximately seventeen hundred gallons of the by-product into tank FA-129, the tank-level measurement read three inches or eighty gallons. Despite this discrepancy in the tank-level measurement, Upjohn continued, over the next few weeks, to pump eight more batches of by-product into tank FA-129. Approximately seventeen hundred gallons of by-product were added to tank FA-129 on each of the following days: August 18, 19, 24, 25, 26, 30 and 31, and on September 1, 1982. The daily tank measurement readings continued to show levels which did not coincide with the amount of by-product which was pumped into tank FA-129.

The data from the tank-level measurements were as follows:
FA 129 TANK LEVEL INVENTORY
DATE LEVEL GALLONS


Aug. 9 10 475 10 10 475 11 10 475 12 10 475 13 10 475 14 — — 15 — —

Aug. 16 3 80[*] 17 11.5 585 18 11.5 585 19 10.5 511 20 16 945 21 — — 22 — —

Aug. 23 8 342 24 8 342 25 10.5 511 26 8 342 27 8 342 28 — — 29 — —

Aug. 30 8.5 375 31 15.5 903

Sept. 1 25 1,794 2 18.5 1,167 3 10.5 511 4 — — 5 — —

Sept. 6 — — 7 7.5 312 8 25 1,794[**] 9 — — 10 — —

[*] First Distillate Received
[**] Partially Filled [With] Water for Washing

On September 3, 1982, Upjohn completed its monthly audit of the tank-level inventory records. Since the tank-level measurements did not coincide with what was intended to be in tank FA-129, no additional quantities of the by-product were pumped into the tank.

It was determined that tank FA-129 had three holes in it due to corrosion. Upjohn estimated that approximately fifteen thousand gallons of the toxic by-product leaked from the tank since the first batch of by-product was pumped into tank FA-129.

On January 11, 1985, Upjohn filed suit against defendant Allstate Insurance Company, claiming that Upjohn was covered under Allstate's comprehensive general liability policy for damages and expenses arising out of the leaking underground storage tank FA-129.

Upjohn commenced this action, along with subrogated property insurers who are no longer parties in this action.

Upjohn also brought suit against several other liability insurers who are no longer parties in this action.

On September 24, 1986, Upjohn moved for summary disposition. The court held that there was coverage under the policy and that coverage was not precluded by the pollution-exclusion clause contained in the policy. The Court, therefore, granted Upjohn's motion for summary disposition.

The Court of Appeals affirmed the trial court's grant of summary disposition. Relying on the definition of "sudden and accidental" as stated in Jonesville Products, Inc v Transamerica Ins Group, 156 Mich. App. 508, 512; 402 N.W.2d 46 (1986), the Court held that "even a continuous discharge of chemicals may be both accidental (i.e., unintended) and sudden (i.e., unexpected) and, therefore, outside the pollution exclusion." The Court held that the trial court properly concluded that there was no genuine issue of material fact that the leak was "sudden and accidental." Upjohn Co v New Hampshire Ins Co, 178 Mich. App. 706, 716; 444 N.W.2d 813 (1989).

On July 13, 1990, this Court granted Allstate's leave to appeal and ordered that it be consolidated with Polkow v Citizens Ins Co of America and Protective Nat'l Ins Co of Omaha v City of Woodhaven.

435 Mich. 862 (1990).

II. ANALYSIS

The Allstate Comprehensive General Liability Policy is subject to several exclusions which limit the broad coverage otherwise provided under the policy. The policy's pollution exclusion provides:

Justice LEVIN, in his dissent, argues that the drafting history of the pollution-exclusion clause supports the proposition that the exclusion does not limit coverage otherwise provided under the policy, but instead, merely clarifies existing coverage. Justice LEVIN states that "when a policy term is ambiguous, courts may look to extrinsic evidence of the term's meaning. In this case, the extrinsic evidence — the drafting history of the pollution-exclusion clause — supports Upjohn's claim for coverage under the CGL policy." Post, pp 224-225. Furthermore, Justice LEVIN states:


Courts that find the phrase "sudden and accidental" to be reasonably susceptible to different meanings, and therefore ambiguous, generally turn to the well-documented drafting and marketing history of the CGL'S pollution exclusion. By and large, those courts finding that the phrase is unambiguous do not address this history, even to the extent of acknowledging its existence.[27]

[27] But see Lumbermens Mutual Casualty Co v Belleville Industries, Inc, 407 Mass. 675, 682 ; 555 N.E.2d 568 (1990), where the Massachusetts Supreme Court found that the term "sudden" unambiguously carries a temporal component. Only after finding the term unambiguous did the court state that "[b]ecause the word `sudden' in the pollution exclusion clause is not ambiguous, we have no need to consider the drafting history of that clause or any statements made by insurance company representatives concerning the intention of its drafters.

[ Post, p 228.]
First, Justice LEVIN cites the rule that "when a policy term is ambiguous, courts may look to extrinsic evidence of the term's meaning," but then he condemns the majority for following the same rule. Indeed, as Justice BOYLE stated in Allstate Ins Co v Freeman, 432 Mich. 656, 712; 443 N.W.2d 734 (1989), when the policy is found to be clear and unambiguous "there is no need to resort to extrinsic evidence to ascertain the meaning of the exclusion. [See 2 Couch, Insurance, 2d (rev ed), § 15:57, pp 298-302.] (Since all prior negotiations are assumed to be merged in the written contract, the policy itself constitutes the contract between the parties, and, if the meaning is clear, it alone must be looked to in construction.)"
The majority is in accord with Justice BOYLE'S analysis and thus we do not look to the drafting history when interpreting and applying the policy terms.
Notwithstanding the above, the majority rejects Justice LEVIN'S interpretation of the drafting history. Justice LEVIN uses selected portions of the drafting history to support his conclusion that the pollution-exclusion clause was intended merely as a clarification of existing coverage. Post, pp 230-231. Furthermore, the majority refuses to treat the occurrence language and the pollution-exclusion clause as interchangeable. When reading the policy as a whole, it is clear that the two clauses have a natural and separate focus. See Polkow v Citizens Ins Co of America, 438 Mich. 174; 476 N.W.2d 382 (1991). We disagree with the assertion that the pollution-exclusion clause simply clarified the definition of an "occurrence" in the coverage section of the policy. Simply stated, it is our belief that exclusions exclude. See American Motorists Ins Co v General Host Corp, 667 F. Supp. 1423, 1429 (D Kan, 1987) ("[i]t is not a novel idea that exceptions to a broad blanket of coverage can be made"); Weedo v Stone-E-Brick, Inc, 81 N.J. 233, 237; 405 A.2d 788 (1979) (the function of an exclusion "is to restrict and shape the coverage otherwise afforded"). See also Raska v Farm Bureau Mutual Ins Co, 412 Mich. 355, 363; 314 N.W.2d 440 (1982).

This policy shall not apply: —

* * *

(f) to bodily injury or property damage arising out of the discharge, dispersal, release or escape of smoke, vapors, soot, fumes, acids, alkalies, toxic chemicals, liquids or gaseous waste materials or other irritants, contaminants or pollutants into or upon land, the atmosphere or any watercourse or body of water; but this exclusion does not apply if such discharge, dispersal, release or escape is sudden and accidental.

The pollution exclusion does not apply to releases which are "sudden and accidental." The question presented in this case is whether the release of the chemical by-product from tank FA-129 was "sudden and accidental," and, therefore, whether the pollution-exclusion clause applies, thus, precluding coverage under the policy.

"Initially, in determining whether a policy applies, we first must determine whether the policy is clear and unambiguous on its face." Metropolitan Property Liability Ins Co v DiCicco, 432 Mich. 656, 665; 443 N.W.2d 734 (1989) (opinion of RILEY, C.J.). We cannot create an ambiguity where none exists. Edgar's Warehouse, Inc v United States Fidelity Guaranty Co, 375 Mich. 598; 134 N.W.2d 746 (1965). Similarly, we reject the temptation to rewrite the plain and unambiguous meaning of the policy under the guise of interpretation. Rather, we enforce the terms of the contract as written. Eghotz v Creech, 365 Mich. 527, 530; 113 N.W.2d 815 (1962). If the language of the policy is unambiguous, it must be considered "in its plain and easily understood sense." 432 Mich. 710. See Wertman v Michigan Mutual Liability Co, 267 Mich. 508, 510; 255 N.W. 418 (1934).

We find persuasive the recent opinions of the United States Court of Appeals for the Sixth Circuit which find the terms of the pollution exclusion to be unambiguous. United States Fidelity Guaranty Co v Star Fire Coals, Inc, 856 F.2d 31 (CA 6, 1988); United States Fidelity Guaranty Co v Murray Ohio Mfg Co, 875 F.2d 868 (CA 6, 1989); FL Aerospace v Aetna Casualty Surety Co, 897 F.2d 214 (CA 6, 1990).

In his dissent, Justice LEVIN criticizes the majority for failing to further analyze the question of ambiguity. Post, p 223. However, the three cases cited in support of the majority's conclusion that the terms of the pollution exclusion are unambiguous provide the reasoning, rationale, and analysis which the majority finds persuasive. Therefore, in citing those cases, we refer the reader of the majority opinion to the analysis in those cases. "Nothing would be gained by restating here what has been persuasively stated there." New Hampshire Ins Co v H Brown Co, unpublished opinion of the Kent Circuit Court, decided September 27, 1989 (Docket No. 87-56314-CK), p 11.

We conclude that when considered in its plain and easily understood sense, "sudden" is defined with a "temporal element that joins together conceptually the immediate and the unexpected." Star Fire Coals, supra at 34. The common, everyday understanding of the term "sudden" is "`happening, coming, made or done quickly, without warning or unexpectedly; abrupt.'" FL Aerospace, supra at 219. "Accidental" means "[o]ccurring unexpectedly and unintentionally; by chance." The American Heritage Dictionary: Second College Edition, p 72. We, therefore, reject the definition of "sudden and accidental" as set forth by the Michigan Court of Appeals in Jonesville. Thus, we find that the terms "sudden" and "accidental" used in the pollution-exclusion clause are unambiguous.

Justice LEVIN, in his dissent, argues that the phrase "sudden and accidental" is susceptible to more than one reasonable interpretation or meaning. Further, he states that this being so, the phrase is ambiguous, as a matter of law, and that "[s]uch ambiguity is resolved in favor of the insured. . . ." Post, pp 224, 232.
Justice LEVIN, like those courts which find that the terms of the pollution-exclusion clause are ambiguous, relies heavily on an assertion of ambiguity buttressed by citation to the contract construction principle which construes ambiguous language in favor of the insured. The majority concludes, however, that the terms of the pollution-exclusion clause are unambiguous. See cases cited ante, p 206. Furthermore, in response to Justice LEVIN'S claim that ambiguity exists, and his reliance on rules of contract construction for such proposition, the majority cites Wozniak v John Hancock Mutual Life Ins Co, 288 Mich. 612, 615; 286 N.W. 99 (1939), which states:


"An insurance policy is a contract and should be interpreted according to its plain meaning. The court is mindful of the rule of law that where the provisions of an insurance policy are uncertain or ambiguous, or the meaning is not clear, that those terms should be given such interpretation or construction as is most favorable to the insured. This rule does not mean, however, that the plain meaning of plain words should be perverted, or that a word or phrase, the meaning of which is specific and well recognized, should be given some alien construction merely for the purpose of benefiting the insured."

See also Smith v Lumbermen's Mutual Ins Co, 101 Mich. App. 78, 83; 300 N.W.2d 457 (1980) ("[a] patently unreasonable interpretation of a contractual ambiguity will not be employed merely to allow the insured to recover his losses").
Justice LEVIN, in his dissent, attempts to support his assertion of ambiguity in two ways. First, he claims that the term "sudden" is ambiguous because it is defined in a variety of ways in the dictionary ( post, pp 227-228) and is thus susceptible to more than one reasonable interpretation. The dissent cites Just v Land Reclamation, Ltd, 155 Wis.2d 737; 456 N.W.2d 570 (1990), in support of this claim. The Just court held that "[t]he very fact that recognized dictionaries differ on the primary definition of `sudden' is evidence in and of itself that the term is ambiguous." Id. at 745.
We reject the reasoning of the Just court. Most, if not all, words are defined in a variety of ways in each particular dictionary, as well as being defined differently in different dictionaries. Similarly, different dictionaries have different ways of listing and ordering the several definitions of each particular word. If courts followed the reasoning of the Just court, it would be virtually impossible to write a contract that was unambiguous. Moreover, the majority refuses to ascribe ambiguity to words in the English language simply because dictionary publishers are obliged to define words differently to avoid possible plagiarism. Therefore, we reject the temptation to rewrite the plain and unambiguous meaning of the policy under the guise of interpretation. Rather, we enforce the terms of the contract as written. Eghotz v Creech, 365 Mich. 527, 530; 113 N.W.2d 815 (1962).
Second, Justice LEVIN claims that the fact that there has been "[s]uch profound disagreement among courts over the construction of the same phrase in standard-form insurance policies itself is evidence of ambiguity." Post, p 227. Although this serves as a convenient argument for those who support a finding of ambiguity, in our view it merely begs the question.

Having concluded that the phrase "sudden and accidental" is not ambiguous, we now turn to the proper application of the pollution-exclusion clause.

Under the facts of this case, we conclude that the release of material from tank FA-129 could not possibly be considered "sudden" because the release of by-product from tank FA-129 was not unexpected by Upjohn.

Justice LEVIN, in his dissent, criticizes as inconsistent the majority's focus on the expectancy component of the word "sudden" when applying the pollution-exclusion clause in this particular case. Justice LEVIN states:


The majority determines that the phrase "sudden and accidental" is unambiguous and properly means "happening quickly and unexpectedly." It then proceeds to review the evidence. However, rather than relying on the definition it has just adopted, the majority uses the definition that plaintiff Upjohn argues is appropriate, i.e., that "sudden and accidental" means " unexpected and unintended."

The majority holds that the release of toxic material at issue "could not possibly be considered `sudden' because the release of by-product from tank FA-129 was not unexpected by Upjohn." The majority thus resolves the issue solely on the basis of the "unexpected" component of the term "sudden," despite its determination, following [ Star Fire, supra], that "sudden" could not be defined without reference to its temporal component. [ Post, p 232. Emphasis in original.]

Justice LEVIN misreads the majority opinion and thus his argument fails. The majority does not determine that the phrase "sudden and accidental" means "happening quickly and unexpectedly." The majority determines that the word "sudden," when considered in its plain and easily understood sense, "is defined with a `temporal element that joins together conceptually the immediate and the unexpected.'" See ante, p 207. Therefore, Justice LEVIN is clearly wrong in stating that "rather than relying on the definition it has just adopted, the majority uses the definition that plaintiff Upjohn argues is appropriate, i.e., that `sudden and accidental' means ` unexpected and unintended.'" Post, p 232 (emphasis in original).
Furthermore, Justice LEVIN'S argument that the majority, by focusing on the expectancy component of the definition of the word sudden, is being inconsistent with its own determination of what sudden means is without merit. The majority states that "sudden" is defined with a temporal element that joins together conceptually the immediate and the unexpected. Under this definition, for something to be sudden, it must be both immediate and unexpected. Since, under the facts of this case, the release of by-product was not unexpected by Upjohn, it is consistent for the majority to state that the release cannot possibly be viewed as "sudden."

Each weekday of the year, tank-level measurements were taken by an Upjohn employee for all the tanks in the tank farm. The employee used a calibrated stick to measure the level of material in each tank. The employee then recorded the measurement of each tank on the tank farm inventory sheet for that day. This sheet was then turned into the employee's supervisor. Each day's sheet was reviewed and compared with previous days' sheets which were kept on file at the UMC facility. UMC has record retention schedules on all such information. These tank-level measurements were taken to determine what the level of material was in each tank and, thus, when each tank was full and needed to be emptied for disposal purposes. Such information indicated when a tank was too full to allow another batch of material to be pumped into it.

Prior to August 16, 1982, the daily tank-level measurements for tank FA-129 remained constant at ten inches or 475 gallons. This is consistent with the fact that when tank FA-129 was not being actively used, as was the case before August 16, 1982, there is an accumulation of ten inches of material or 475 gallons in the tank because that is about all that is capable of being pumped out of the tank. On August 16, 1982, the same day that the first batch of seventeen hundred gallons of by-product was received, the tank-level measurement was three inches or eighty gallons. Despite this discrepancy, Upjohn continued its production of clindamycin and added eight more batches, approximately 13,600 gallons, of the by-product to the ten thousand gallon tank. The tank level measurements taken from August 16, 1982, until early September 1982, when tank FA-129 was removed from active use, reflected that there was an unaccounted loss of the by-product.

In early September 1982, after a monthly audit of tank levels by Upjohn indicated that there was a leak in tank FA-129, no additional quantities of the by-product were pumped into the tank. It was later determined that the discrepancies in the tank-level readings were due to several holes in tank FA-129 which caused the tank to continuously leak chemicals into the ground from August 16, 1982, through September 7, 1982. Upjohn states that anywhere from twelve thousand gallons to eighteen thousand gallons of by-product leaked from tank FA-129 during the period.

Upjohn claims that although the tank level measurements for tank FA-129 were unusually low on August 16, 1982, and although a large discrepancy continued to exist in the measurements that were taken, recorded, and reviewed by Upjohn employees, Upjohn did not have enough information to expect that the chemical by-product that was in the tank and was continuously being added to the tank was escaping from a leak in the tank. Upjohn claims that they did not expect that there was a release of the by-product from tank FA-129 until a monthly audit of tank levels that was completed in early September 1982 indicated a leak. During this audit all the previously recorded and reviewed tank farm inventory sheets were reviewed by Upjohn's accounting people. It is Upjohn's contention that only after all relevant figures were compared, that UMC could even have expected that by-product was missing from Upjohn's production and storage system. Upjohn claims that the unusually low tank-level measurement for tank FA-129 taken on August 16, 1982, did not lead them to expect that chemical by-product was escaping from the tank because reduced tank levels did not themselves indicate a leak in the tank. Upjohn claims that the by-product could have been accidentally or intentionally diverted to other tanks, production could have been interrupted, or materials could have been intentionally removed from the tank for disposal.

Upjohn's claims cannot be reasonably supported by the record in this case. Although reduced tank levels may not themselves indicate a leak in tank FA-129, the fact that the tank level on August 16, 1982, measured three inches or eighty gallons precludes this Court from finding anything other than that Upjohn must have expected a leak in tank FA-129. Thus, we find, as a matter of law, that the Upjohn Company had sufficient information available to it on August 16, 1982, to expect that a chemical by-product was escaping from a leak in tank FA-129. Furthermore, we find that Upjohn had sufficient information available to it to expect that such release of the by-product was occurring and would continue to occur until they stopped using the tank.

The record in this case is undisputed that when the tank is not in use, there is an accumulation of ten inches or 475 gallons of by-product in the tank because that is about all that is capable of being pumped out of the tank. In essence, even when the tank is empty, there should be 475 gallons of by-product in it. In fact, the record shows that prior to August 16, 1982, the tank-level measurements for tank FA-129 were constant at ten inches or 475 gallons.

Upjohn claims that the unusually low tank-level measurement for tank FA-129 taken on August 16, 1982, did not lead it to expect that chemical by-product was escaping from the tank because reduced tank levels did not themselves indicate a leak in the tank. Upjohn claims that the by-product could have been accidentally or intentionally diverted to other tanks, production could have been interrupted, or materials could have been intentionally removed from the tank for disposal.
However, it is clear from the undisputed facts of this case that none of these claims can be reasonably upheld. A tank-level measurement of three inches or eighty gallons would not indicate an accidental or intentional diversion of by-product to another tank because if the by-product was diverted and did not reach tank FA-129, then the tank-level measurement on August 16, 1982, should still have read at least ten inches or 475 gallons. The same rationale holds true for the claim by Upjohn that production of the by-product could have been interrupted. With respect to the claim by Upjohn that the three-inch or eighty-gallon tank-level measurement could have indicated that materials could have been intentionally removed from the tank for disposal, this was not possible because the record states that even when the tank is not in use, there is an accumulation of ten inches or 475 gallons in the tank.
Therefore, the only reasonable conclusion that can be drawn from the facts presented in the record in this case is that Upjohn must have expected that chemical by-product was escaping from tank FA-129 as early as August 16, 1982.

In determining the knowledge attributable to a company, the court in United States v TIME-DC, Inc, 381 F. Supp. 730, 738 (WD Va, 1974), held:

A corporation can only act through its employees and, consequently, the acts of its employees, within the scope of their employment, constitute the acts of the corporation. Likewise, knowledge 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. Rather, the corporation is considered to have acquired the collective knowledge of its employees and is held responsible for their failure to act accordingly.

The Michigan Court of Appeals adopted this reasoning in People v American Medical Centers of Michigan, Ltd, 118 Mich. App. 135; 324 N.W.2d 782 (1982).

In Gordon Sel-Way, Inc v Spence Bros, Inc, 177 Mich. App. 116, 124; 440 N.W.2d 907 (1989), the Court recognized that "the combined knowledge of employees may be imputed to a corporation" and stated:

This case is currently on appeal before this Court on a different issue.

We agree with this rule and adopt the standard for imputed collective knowledge set forth in Copeman Laboratories Co v General Motors Corp, 36 F. Supp. 755, 762 (ED Mich, 1941):

"When a person representing a corporation is doing a thing which is in connection with and pertinent to that part of the corporation business which he is employed, or authorized or selected to do, then that which is learned or done by that person pursuant thereto is in the knowledge of the corporation. The knowledge possessed by a corporation about a particular thing is the sum total of all the knowledge which its officers and agents, who are authorized and charged with the doing of the particular thing acquire, while acting under and within the scope of their authority." [ Id. at 124-125.]

We adopt this reasoning and apply it in this case. This Court concludes that the Upjohn Company had sufficient information available to it, through its various employees and through its records kept at the UMC facility, to allow us to find, as a matter of law, that the release of chemical by-product from tank FA-129 was expected by the Upjohn Company.

Justice LEVIN, in his dissent, claims that the majority "fails to distinguish between raw data and `knowledge.'" Post, p 233. He states that only on September 3, 1982, did Upjohn determine that a leak occurred. Id., p 234. He states that "the majority errs in finding that Upjohn must have known on the first day of the clindamycin manufacturing campaign that tank FA-129 was leaking. What Upjohn possessed between August 16 and September 13, 1982, was raw, unanalyzed data, not knowledge." Id.
The majority concludes that Justice LEVIN fails to distinguish between "knowledge" and "expectancy." The majority never finds that Upjohn must have known on August 16 that the tank, in fact, was leaking. Rather, the majority finds that Upjohn had sufficient information available to it, through its various employees, to allow this Court, upon applying the imputed-collective-knowledge standard, to find, as a matter of law, that Upjohn must have expected the leak in tank FA-129. See, supra, ns 10-11 and accompanying text. The knowledge required to expect something is less than knowledge required to know the same thing. This is a distinction that Justice LEVIN fails to see. The majority need not find that Upjohn knew of the leak on August 16, 1982, to find that, as a matter of law, the release of chemical by-product was not "sudden and accidental," the majority need only find that the facts are undisputed that Upjohn expected the leak on August 16, 1982. The term "sudden" has been defined by the majority (see ante, p 207) as well as by Justice LEVIN in his dissent (see post, p 221) as having an "expectancy" component and not a "knowledge" component.

III. CONCLUSION

In summary, under the imputed-collective-knowledge standard the Upjohn Company cannot claim that it did not expect on August 16, 1982, and every day thereafter, the release of by-product from tank FA-129. 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, and that only after a monthly audit of the same information could it have expected the release. Rather, the Upjohn Company is considered to have acquired the collective knowledge of its employees, the same collective knowledge which the monthly audit revealed was available to Upjohn as early as August 16, 1982. Therefore, Upjohn must be held to have expected the release of by-product from tank FA-129 as early as August 16, 1982.

Justice LEVIN, in his dissent, would like to limit the imputed-collective-knowledge standard by holding that it is only applicable to information possessed by "employees at the appropriate level of responsibility. . . ." Post, p 234. Justice LEVIN attempts to support his limitation for so-called "appropriate" employees by citing Gordon, supra, which stated that "one of the burdens attendant upon the corporate form is that the law imputes the knowledge of individual officers and employees at a certain level of responsibility to the corporation." Id., p 233. (Emphasis added.) He adds that "[t]his principle speaks to persons representing a corporation who, acting in the scope of their employment and authority, learn or do something on behalf of the corporation." Id., pp 233-234.
After acknowledging that the employees to which the imputed-collected-knowledge standard applies are simply those employees acting in the scope of their employment and authority and who learn or do something on behalf of the corporation, Justice LEVIN attempts to limit the scope of the standard to apply only to Upjohn's auditors or officers. This is completely in conflict with the purpose behind the standard (cf. Upjohn Co v United States, 449 U.S. 383; 101 S Ct 677; 66 L Ed 2d 584 [1981]) and would allow the Upjohn Corporation to "`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'" (see ante, p 213) — something the cases cited by the majority clearly reject as the antithesis of imputed collective corporate knowledge.
The cases which endorse the imputed-collective-knowledge standard stand for the proposition that a corporation is deemed to have had knowledge of information if the means were present by which the company could have detected such information. United States v TIME-DC, Inc, 381 F. Supp. 730, 739 (WD Va, 1974). Obviously, such means were present in the case before this Court, through Upjohn's various employees and through its records kept at the UMC facility to detect information which would lead Upjohn to, at the very least, expect that tank FA-129 was leaking. "In some cases such an analysis may appear harsh, but doing business in the corporate form carries certain burdens of which this is one." Gordon, supra at 126.

Since the release of chemical by-product was not unexpected, as a matter of law it cannot be "sudden and accidental." Therefore, the pollution-exclusion clause does apply, and the Upjohn Company is not entitled to coverage under the Allstate Insurance Company policy. Accordingly, we reverse the decision of the Court of Appeals.

Chief Justice CAVANAGH, dissenting, argues that "having defined `sudden' to have a temporal element, it would seem prudent to determine the quickness with which this leak occurred." Post, p 218. He states that if the leak developed quickly it could be declared "sudden." Id. However, a leak, even if it developed quickly, could not be declared "sudden" if it were not also unexpected. Under the majority's definition of sudden, with which Chief Justice CAVANAGH agrees, a leak must be both quick and unexpected. See n 9. Therefore, it is not necessary, as Chief Justice CAVANAGH suggests, to determine the quickness with which a leak occurred when it has been determined, as in the majority, that the leak was not unexpected.
Chief Justice CAVANAGH, however, believes that there is a genuine issue of material fact regarding whether the leak was expected by Upjohn. Chief Justice CAVANAGH argues that the majority relies on comparisons of daily tank farm inventory sheets to show that Upjohn had the information sufficient to expect a leak. Id. Such an argument misreads the majority's application of the facts in this case. Indeed, the majority insists that the tank level reading on August 16 alone was enough to provide the necessary information to the Upjohn Company, to allow this Court to find that Upjohn expected the leak. See ante, p 212. Therefore, any dispute regarding daily comparisons is not a genuine issue of material fact.

BOYLE, GRIFFIN, and MALLETT, JJ., concurred with RILEY, J.


CAVANAGH, C.J. (concurring in part and dissenting in part).

I agree with the majority that the phrase "sudden and accidental" contained in the exception to the pollution exclusion clause is unambiguous. I also agree that the word "sudden" includes a temporal element. I write separately because the grant of summary disposition in this case was inappropriate and because I am not convinced that this leak could not possibly, as a matter of law, have been sudden. Whether the underground leak occurred gradually or "suddenly" upon the first day's influx of by-product is a question of fact, and the case should be remanded to allow the lower court to apply the definition developed by this Court.

The lead opinion declares that the release "could not possibly be considered `sudden' because the release of by-product from tank FA-129 was not unexpected by Upjohn." Ante, p 209 (emphasis added). This conclusion is based on Upjohn's process of monitoring tank levels. The lead opinion describes the monitoring process as follows:

Each weekday of the year, an Upjohn employee measured the level of by-product in the tank. The employee recorded the measurement on a tank farm inventory sheet and turned the sheet over to his supervisor. Each day's sheet was reviewed, and compared with previous days' sheets which were kept on file at the UMC facility. [ Id., p 202.]

According to defendant Upjohn, however, the recordings were kept on separate sheets and were not "compared with previous days' sheets," until the monthly audit was performed. Since Upjohn disputes the daily comparison, there is a genuine issue of material fact, and it is inappropriate for this Court to decide which version of the facts it prefers.

Furthermore, having defined "sudden" to have a temporal element, it would seem prudent to determine the quickness with which this leak occurred. If the holes in the tank developed simultaneously upon the first day's influx of by-product, it could be declared a "sudden" occurrence. In fact, the majority opinion confirms my belief that the leak could have been sudden:

Prior to August 16, 1982, the tank-level measurements remained constant at ten inches or 475 gallons. However, on August 16, 1982, the same day that Upjohn had pumped its first batch of approximately seventeen hundred gallons of the by-product into tank FA-129, the tank-level measurement read three inches or eighty gallons. [ Id.]

This lends credence to the theory that the leak occurred "suddenly." The by-product apparently ran out of the tank as rapidly as it was put in. Even if it were proper for this Court to make this factual determination, since the tank had maintained its structural integrity over the year since its last use, I cannot at this stage accept the factual conclusion that Upjohn expected this leak.

While Upjohn may have been negligent in its monitoring, the possibility of negligence is the very reason for purchasing comprehensive, general liability insurance. The lapse of time between the beginning of the leak and the discovery of the leak is not determinative of the question whether the leak was sudden or accidental. In Wagner v Milwaukee Mutual Ins Co, 145 Wis.2d 609; 427 N.W.2d 854 (1988), rev'd on other grounds 155 Wis.2d 737; 456 N.W.2d 570 (1990), a leak in a gas pipe which continued over a three-year period was held to have been sudden. The pipe had been damaged in 1981 when cement footings were poured around it, but the leak was not discovered until 1984. The court reasoned that "reliance on the period of time that elapsed between the event that damaged the pipe and the discovery of the leak is misplaced. . . . The length of time that elapsed before the leak was discovered is irrelevant as to the suddenness of the discharge." 145 Wis.2d 616. This analysis applies here; there should be a determination whether the holes in the tank all gave way at once, producing a "sudden" discharge, rather than focusing on the length of time which elapsed before Upjohn discovered the leak.

The Wisconsin Supreme Court modified the appellate court's definition of "sudden and accidental" to mean "unexpected and unintended." This definition permits recovery for unexpected damage even if it occurs over time and removes any need for a determination how "quickly" the discharge occurred. 155 Wis.2d 746.

I would remand for a factual determination regarding the "sudden and accidental" nature of this leak.

BRICKLEY, J., concurred with CAVANAGH, C.J.


LEVIN, J. (dissenting).

The principal question presented concerns the construction and application of the pollution-exclusion clause of the comprehensive general liability (CGL) policy, and in particular the phrase "sudden and accidental."

The majority holds that "sudden and accidental" is unambiguous, that "when considered in its plain and easily understood sense, `sudden' is defined with a `temporal element that joins together conceptually the immediate and the unexpected,'" and that "`[a]ccidental' means `[o]ccurring unexpectedly and unintentionally; by chance.'"

RILEY, J., ante, p 207.

Id., pp 207-208.

Then, focusing solely on the definition of "sudden," the majority finds that the leak from Upjohn's underground storage tank "could not possibly be considered `sudden' because the release of by-product from tank FA-129 was not unexpected by Upjohn."

Id., p 209 (emphasis added).

Upjohn's daily inventory records indicated that tank FA-129 contained less fluid after the first batch of by-product was pumped into it than it had before. On the basis of this record, the majority finds that "the fact that the tank level on August 16, 1982, measured three inches or eighty gallons precludes this Court from finding anything other than that Upjohn must have expected a leak in tank FA-129."

Id., p 212 (emphasis added).

The majority finds "as a matter of law" that Upjohn, under the theory of the "imputed-collective-knowledge standard," possessed sufficient information on the first day of the campaign to "expect" that tank FA-129 was leaking and would continue to leak until Upjohn stopped using it. The majority concludes that, because the release of the by-product was not "sudden," in that it was not unexpected, it does not fall within the "sudden and accidental" exception, with the result that coverage under the policy is not triggered.

Id., p 213.

Two of my colleagues concur in part and dissent in part. They concur in the holdings that the phrase "sudden and accidental" is unambiguous and that the definition of the word "sudden" includes a temporal element. They dissent because "the grant of summary disposition in this case was inappropriate and because [they are] not convinced that this leak could not possibly, as a matter of law, have been sudden." They would remand for a factual determination of whether the leak was "sudden and accidental."

CAVANAGH, C.J., ante, p 217.

I would hold that the phrase "sudden and accidental" is ambiguous, and that it means "unexpected and unintended." I would, with my dissenting colleagues, remand for trial. I join in their expression of disagreement with the extent to which the majority has acted as finder of fact to determine "as a matter of law" that the leak could not have been "sudden" or "unexpected."

I

Upjohn annually manufactures the antibiotic clindamycin in two-month-long "campaigns." Toxic by-products produced during the manufacturing process are pumped into a 10,000-gallon storage tank. During the first twenty-two days of the 1982 campaign, virtually all the 15,000 gallons of toxic by-product that were pumped into the tank leaked out. The leak was discovered during a monthly audit. According to Upjohn officials, daily tank level readings were taken by employees, recorded on separate sheets of paper, and turned in at the end of each day. These readings were audited monthly.

This tank was specially manufactured of materials that would not be corroded by waste products.

Immediately upon discovering the leak, Upjohn began a cleanup that included extracting the contaminant from the ground around the tanks and from groundwater. Upjohn provided drinking water for surrounding communities where wells had been contaminated. In 1987, the EPA entered a consent order requiring Upjohn to continue monitoring, and in 1989 a cleanup order was entered. Upjohn expended $6.7 million on the cleanup.

Allstate is one of Upjohn's excess liability insurers. While the policy's indemnification provision reimburses Upjohn for both damages and expenses, and thus affords broader coverage than the standard CGL policy, the pollution exclusion is standard. Upjohn commenced an action against Allstate seeking to obtain coverage under the excess liability policy. The circuit court granted Upjohn's motion for summary judgment, holding that coverage was not excluded by the pollution exclusion because the leak was "sudden and accidental." The Court of Appeals affirmed. This Court granted leave to appeal and ordered this case consolidated with Polkow v Citizens Ins Co of America and Protective Nat'l Ins Co of Omaha v City of Woodhaven.

Other coverage included property insurance, which reimbursed $3 million for cleanup costs, and primary liability insurance, which reimbursed almost $1 million.

The circuit court also held that, while Upjohn was not entitled to reimbursement for damage to its own property it could recover for expenditures, including cleanup costs, to repair damage to the property of third parties, including groundwater, which is governmental property.

435 Mich. 862 (1990).

II

The majority construes and applies the pollution-exclusion clause in the standard CGL policy issued by Allstate. In so doing, the majority asks and answers two questions: First, is "sudden and accidental" ambiguous? Second, if the phrase is unambiguous, what is the meaning?

Addressing the first question, the majority looks to three decisions of the United States Court of Appeals for the Sixth Circuit "which find the terms of the pollution exclusion to be unambiguous." Each of these decisions holds that there is a singular, "plain everyday" or commonsense meaning of the word "sudden." This everyday meaning is found to incorporate a temporal element that joins the concepts of immediacy and the unexpected. "Accidental" is defined as "unexpected or unintended." The majority declares that the three Sixth Circuit decisions are persuasive, adopts their definition of "sudden and accidental," and holds that the phrase is unambiguous.

RILEY, J., ante, p 207.

United States Fidelity Guaranty Co v Star Fire Coals, Inc, 856 F.2d 31 (CA 6, 1988); United States Fidelity Guaranty Co v Murray Ohio Mfg Co, 875 F.2d 868 (CA 6, 1989); FL Aerospace v Aetna Casualty Surety Co, 897 F.2d 214 (CA 6, 1990).

According to Star Fire, supra, p 34, " it is [ not] possible to define `sudden' without reference to a temporal element that joins together conceptually the immediate and the unexpected." (Emphasis added.)

III

I would hold that the pollution exclusion, in particular the phrase "sudden and accidental," may be construed in more than one way by reasonable laypersons. It is well established in insurance law that a policy term is ambiguous when it is susceptible to more than one reasonable definition. The reasonableness of a definition is to be assessed from the layperson's point of view. In the presence of two differing interpretations, only one of which results in liability to the insurer, this Court has consistently declared that the policy should be read to provide rather than to exclude coverage. Further, as discussed in part IV, when a policy term is ambiguous, courts may look to extrinsic evidence of the term's meaning. In this case, the extrinsic evidence — the drafting history of the pollution-exclusion clause — supports Upjohn's claim for coverage under the CGL policy.

The pollution exclusion states as follows:
This policy shall not apply: —
* * *


(f) to bodily injury or property damage arising out of the discharge, dispersal, release or escape of smoke, vapors, soot, fumes, acids, alkalies, toxic chemicals, liquids or gaseous waste materials or other irritants, contaminants or pollutants into or upon land, the atmosphere or any watercourse or body of water; but this exclusion does not apply if such discharge, dispersal, release or escape is sudden and accidental.

See, e.g., Hecla Mining Co v New Hampshire Ins Co, 811 P.2d 1083, 1091-1092 (Colo, 1991); Morgan v Prudential Ins Co, 86 Wn.2d 432; 545 P.2d 1193 (1976); Prete v Merchants Property Ins Co, 159 W. Va. 508; 223 S.E.2d 441 (1976); Shadbolt v Farmers Ins Exchange, 275 Or. 407; 551 P.2d 478 (1976); Garriguenc v Love, 67 Wis.2d 130; 226 N.W.2d 414 (1975); Alvarez v Southwestern Life Ins Co, 86 N.M. 300; 523 P.2d 544 (1974); Clark v Prudential Ins Co, 204 Kan. 487; 464 P.2d 253 (1970); Wachovia Bank Trust Co v Westchester Fire Ins Co, 276 N.C. 348; 172 S.E.2d 518 (1970); English v Old American Ins Co, 426 S.W.2d 33 (Mo, 1968).

See, e.g., Marston v American Employers Ins Co, 439 F.2d 1035 (CA 1, 1971) (Puerto Rican law); Morgan v Prudential Ins Co, n 16 supra; C H Plumbing Heating, Inc v Employers Mutual Casualty Co, 264 Md. App. 510; 287 A.2d 238 (1972); Reserve Ins Co v Staats, 9 Ariz. App. 410; 453 P.2d 239 (1969); Logan v Victory Life Ins Co, 175 Kan. 88; 259 P.2d 165 (1953).

Powers v DAIIE, 427 Mich. 602; 398 N.W.2d 411 (1986); Ebert v Prudential Ins Co, 338 Mich. 320; 61 N.W.2d 164 (1953); VanZanten v Nat'l Casualty Co, 333 Mich. 28; 52 N.W.2d 581 (1952); Francis v Scheper, 326 Mich. 441; 40 N.W.2d 214 (1949); Hooper v State Mutual Life Assurance Co, 318 Mich. 384; 28 N.W.2d 331 (1947); Pietrantonio v Travelers Ins Co, 282 Mich. 111; 275 N.W. 786 (1937); Boesky Bros Corp v USFG Co, 267 Mich. 628; 255 N.W. 307 (1934).

Chicago Bd Options Exchange v Connecticut Gen'l Life Ins Co, 713 F.2d 254 (CA 7, 1983); Aetna Ins Co v Getchell Steel Treating Co, 395 F.2d 12 (CA 8, 1968); Great West Casualty Co v Truck Ins Exchange, 358 F.2d 883 (CA 10, 1966); Fidelity Casualty Co of New York v Seven Provinces Ins Co, 345 F.2d 227 (CA 6, 1965); Union Ins Society v William Gluckin Co, 353 F.2d 946 (CA 2, 1965); Gribaldo, Jacobs, Jones Associates v Agrippina Versicherunges AG, 3 Cal.3d 406; 91 Cal.Rptr. 6; 476 P.2d 406 (1970); Transport Indemnity Co v Dahlen Transport, 281 Minn. 253; 161 N.W.2d 546 (1968); Prince v Universal Underwriters Ins Co, 143 N.W.2d 708 (ND, 1966). See also Vigil v Badger Mutual Ins Co, 363 Mich. 380; 109 N.W.2d 793 (1961).

The policy does not define "sudden and accidental." When determining the meaning of words, definitions in recognized dictionaries may be considered. Authoritative dictionaries differ on the meaning of "sudden." Webster's Third New International Dictionary, p 2285, defines "sudden" in a number of ways: first, as "happening without previous notice . . . occurring unexpectedly . . . not foreseen." Then it lists synonyms for "sudden" that include "prompt" and "immediate." The Random House Dictionary of the English Language (2d ed), p 1900, defines "sudden" in a temporal sense as "happening, coming, made, or done quickly." Black's Law Dictionary (5th ed), p 1284, defines "sudden" as "[h]appening without previous notice or with very brief notice; coming or occurring unexpectedly; unforeseen; unprepared for."

Some courts have found that the failure of an insurance policy to define crucial terms may in itself render those terms ambiguous. See, e.g., Buckeye Union Ins Co v Liberty Solvents Chemicals, 17 Ohio App.3d 127, 132-133; 477 N.E.2d 1227 (1984); State Farm Mutual Automobile Ins Co v Shelton, 368 S.W.2d 734 (Ky, 1963).

As urged by Allstate, the sequence in which definitions of a word appear in different dictionaries may correspond to the date of a particular definition's first use rather than its currency or predominance. See Just v Land Reclamation, Ltd, 155 Wis.2d 737, 760; 456 N.W.2d 570 (1990) (Steinmetz, J., dissenting). Nonetheless, such differences in meaning are indicative of ambiguity.

Judicial efforts to define "sudden and accidental" reflect this diversity of definition. In New Castle Co v Hartford Accident Indemnity Co, 933 F.2d 1162 (CA 3, 1991), the United States Court of Appeals for the Third Circuit said that the numerous decisions construing the pollution exclusion appear to be evenly divided between the competing constructions of the phrase, with half the cases holding, as does the majority here, that the clause bars coverage, and half holding that it does not.

The cases are categorized and cited in ns 60 and 61, p 1195. See also Hecla Mining Co, n 16 supra, p 1091.

In Just v Land Reclamation, Ltd, 155 Wis.2d 737; 456 N.W.2d 570 (1990), the Wisconsin Supreme Court found that "sudden and accidental" was ambiguous because the term "sudden" has different meanings. The court noted that Webster's gives the primary meaning of "sudden" as unexpected, and a secondary meaning as "prompt," while the Random House Dictionary gives the primary meaning as "quickly" and the secondary meaning as "unexpected." This disparity, the court observed, is evidence of ambiguity. The court then quoted the Georgia Supreme Court:

"Perhaps, the secondary meaning is so common in the vernacular that it is, indeed, difficult to think of `sudden' without a temporal connotation: a sudden flash, a sudden burst of speed, a sudden bang. But, on reflection one realizes that, even in its popular usage, `sudden' does not usually describe the duration of an event, but rather its unexpectedness: a sudden storm, a sudden turn in the road, sudden death. Even when used to describe the onset of an event, the word has an elastic temporal connotation that varies with expectations: Suddenly, it's spring. See also Oxford English Dictionary, at 96 (1933) (giving usage examples dating back to 1340, e.g., `She heard a sudden step behind her'; and, `A sudden little river crossed my path As unexpected as a serpent comes.'" [Emphasis added.]

155 Wis.2d 737, 745-746; 456 N.W.2d 570, quoting Claussen v Aetna Casualty Surety Co, 259 Ga. 333, 335; 380 S.E.2d 686 (1989).

The Wisconsin Supreme Court concluded that because "sudden and accidental" is reasonably susceptible to more than one meaning, including abrupt and immediate as well as unexpected and unintended, it is ambiguous as used in the pollution-exclusion clause.

The Supreme Courts of Georgia and Colorado agree. Claussen, n 23 supra, p 335; Hecla Mining Co, n 16 supra.

In short, the scope of the pollution-exclusion clause, particularly the phrase "sudden and accidental," has been the subject of intense and frequent litigation since adoption by the insurance industry in the early 1970s. Such profound disagreement among courts over the construction of the same phrase in standard-form insurance policies itself is evidence of ambiguity. See cases cited in Just and New Castle Co, and see 13 Appleman, Insurance Law Practice, § 7404, p 337, and cases cited in n 96. The extensive debate concerning the meaning of the phrase indicates that determining the correct construction is not "simple, direct [and] straightforward" as Allstate contends.

Several courts have concluded that the substantial conflict in authority regarding the meaning, even if not dispositive, suggests that there is indeed ambiguity. These include United States Fidelity Guaranty Co v Thomas Solvent Co, 683 F. Supp. 1139, 1155-1156 (WD Mich, 1988) (Michigan law), Just, n 21 supra, 759-760, Grinnell Mutual Reinsurance Co v Wasmuth, 432 N.W.2d 495, 499 (Minn App, 1988), New Castle Co, supra, pp 1193-1194, and Hecla Mining Co, n 16 supra, p 21. The same conclusion has been reached in cases construing other insurance policy terms: Michigan Mutual Ins Co v Combs, 446 N.E.2d 1001, 1007 (Ind App, 1983) ("respectable lines of authority producing conflicting interpretations of the word `upon'" contribute to a finding that the term is ambiguous as used in automobile policies); General Credit Corp v Imperial Casualty Indemnity Co, 167 Neb. 833, 843-844; 95 N.W.2d 145 (1959) (a clause requiring a mortgagee to pay premiums on a real-property policy could be read as either a condition or a covenant; conflicting authority indicates ambiguity); Olmstead Co v Metropolitan Life Ins Co, 118 Ohio St. 421, 426-427; 161 N.E. 276 (1928) (same); Gould Morris Electric Co v Atlantic Fire Ins Co, 229 N.C. 518, 520; 50 S.E.2d 295 (1948) (the phrase "collision of the conveyance on which the goods are carried" was ambiguous because it was reasonably susceptible of two meanings, as reflected in conflicting authority).

IV

Courts that find the phrase "sudden and accidental" to be reasonably susceptible to different meanings, and therefore ambiguous, generally turn to the well-documented drafting and marketing history of the CGL'S pollution exclusion. By and large, those courts finding that the phrase is unambiguous do not address this history, even to the extent of acknowledging its existence.

I acknowledge that, as Allstate contended at oral argument, the drafting history of the pollution-exclusion clause was not made part of the record. I note, however, that industry representations were in some instances a condition precedent to approval of the revised phrasing by state insurance commissioners and other regulatory bodies, and that the proceedings leading to approval are matters of public record of which judicial notice may be appropriate. Courts that have considered the drafting history as extrinsic evidence of the meaning of "sudden and accidental" have done so without addressing the question whether the drafting history should have been made part of the record. If Allstate were to offer evidence that the version of the drafting history set forth here is inaccurate or incomplete, Allstate would be entitled to an opportunity to present such evidence.

But see Lumbermens Mutual Casualty Co v Belleville Industries, Inc, 407 Mass. 675, 682 ; 555 N.E.2d 568 (1990), where the Massachusetts Supreme Court found that the term "sudden" unambiguously carries a temporal component. Only after finding the term unambiguous did the court state that "[b]ecause the word `sudden' in the pollution exclusion clause is not ambiguous, we have no need to consider the drafting history of that clause or any statements made by insurance company representatives concerning the intention of its drafters."

The CGL policy's drafting history has been extensively documented in many law review articles, and summarized in many cases. See Just, supra, p 747; Claussen v Aetna Casualty Surety Co, 259 Ga. 333; 380 S.E.2d 686 (1989); New Castle Co, supra. In brief, industrywide revisions of the CGL standard-form policy occurred in 1966 and in 1973. Before 1966, the standard CGL policy covered bodily injury and property damage "caused by accident." The term "accident" was not defined in the policy, although insurers argued that the term encompassed only "brief catastrophic events" and not "gradual damage." Courts generally rejected the insurers' arguments, holding that "accident" encompassed "unintended injury or damage resulting from, among other things, extended exposure to pollutants."

This sketch of the drafting history is taken from the courts' summaries in Claussen, Just, and New Castle Co.

In 1966 the insurance industry acknowledged case law and revised the standard language to provide "occurrence based" coverage. The revised standard-form policy defined "occurrence" as

an accident, including continuous or repeated 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 1966 standard occurrence-based policy thus explicitly covered property damage resulting from gradual pollution. Courts generally extended coverage to all pollution-related damage, even if it arose from the intentional discharge of pollutants, unless the ultimate loss was either expected or intended. At the time this policy change was implemented, representatives of the insurance industry stated that it was to be viewed as "a broadening of coverage," and that under the new policy, an insured would be covered "`until such time as [he] becomes aware that the damage was being done.'"

Lyman J. Baldwin, Jr., Secretary of Underwriting, Insurance Company of North America, quoted by Just, supra, p 749.

In the early 1970s, the standard-form policy was revised to add the pollution exclusion at issue in this case. Under this provision, only pollution-related losses that arose from occurrences both "sudden" and "accidental" would be covered. One writer explains that the exclusion was designed to decrease claims for losses caused by expected or intended pollution by providing an incentive to improve manufacturing and disposal processes. Unintentional or unexpected damages would still be covered as an "occurrence" under the policy.

Greenlaw, The CGL policy and the pollution exclusion clause: Using the drafting history to raise the interpretation out of the quagmire, 23 Colum J of L Soc Prob 233 (1990).

The insurance industry submitted this revision to state regulatory authorities for approval. In West Virginia and Georgia, the dialogue between the insurance industry and the regulatory authority is a matter of public record.

The West Virginia Insurance Commissioner approved the terms of the pollution exclusion only on the basis of representations, made orally and in writing by the Mutual Insurance Rating Bureau, one of the two major insurance trade associations, that they were "`merely clarifications of existing coverages as defined and limited to the definitions of the term "occurrence," contained in the respective policies to which said exclusions would be attached.'"


The Mutual Insurance Rating Bureau, which assisted in the draft of the exclusion endorsement, in a submission to the West Virginia Commissioner of Insurance, explained that the intent of the clause was to clarify "`that the definition of occurrence excludes damages that can be said to be expected or intended.'" [ Just, supra, pp 750-751.]

In Georgia, the Insurance Rating Board, the other major insurance trade association, represented to the Insurance Commissioner that the clause was intended to shut out only intentional polluters, that the change would have no effect on the vast majority of risks, and that it was intended only as a clarification of existing coverage.

Claussen, supra, pp 335-336.

The insurance industry described the pollution exclusion to its agents as follows:

"In one important respect, the exclusion simply reinforces the definition of occurrence. [The policy] will not cover claims where the `damage was expected or intended' by the insured and the exclusion states, in effect, that the policy will cover incidents which are sudden and accidental — unexpected and not intended."

Excerpted from The Fire, Casualty Surety Bulletin, as quoted in Just, supra, p 752 (emphasis added).

It is against this backdrop that at least half the cases construing the pollution-exclusion clause have found it to be ambiguous. Of these, some have read the clause to provide coverage for injury or damage caused by an unintentional and unexpected event. See New Castle Co, supra, and cases there cited. Other cases, which find the clause ambiguous, hold that "sudden and accidental" is simply a restatement of the definition of "occurrence" and that policies incorporating the pollution-exclusion clause cover claims where the alleged injury or harm was "neither expected nor intended." See Just, supra, and cases there cited.

V

Once the phrase "sudden and accidental" is found to be reasonably susceptible of more than one meaning, it is ambiguous as a matter of law. Such ambiguity is resolved in favor of the insured: the exception to the pollution clause, intended by its drafters as a restatement of "occurrence," covers "unexpected or unintended damage or results." Taken as a whole, the pollution-exclusion clause and exception thus preclude coverage only when the insured intentionally or recklessly causes injury or damage.

VI

The majority determines that the phrase "sudden and accidental" is unambiguous and properly means "happening quickly and unexpectedly." It then proceeds to review the evidence. However, rather than relying on the definition it has just adopted, the majority uses the definition that plaintiff Upjohn argues is appropriate, i.e., that "sudden and accidental" means "unexpected and unintended."

The majority holds that the release of toxic material at issue "could not possibly be considered `sudden' because the release of by-product from tank FA-129 was not unexpected by Upjohn." The majority thus resolves the issue solely on the basis of the "unexpected" component of the term "sudden," despite its determination, following United States Fidelity Guaranty Co v Star Fire Coals, Inc, 856 F.2d 31 (CA 6, 1988), that "sudden" could not be defined without reference to its temporal component.

RILEY, J., ante, p 209.

See n 1.

The majority assumes the role of factfinder in applying "sudden and accidental" to the "facts." The majority focuses on whether the leak was "unexpected," and concludes, as a matter of law, that it was not. That conclusion is flawed, because the majority first fails to distinguish between raw data and knowledge, second, misapplies agency principles, and, third, turns a question of fact into a question of law.

The majority states, " as a matter of law, that the Upjohn Company had sufficient information available to it on August 16, 1982, to expect that a chemical by-product was escaping from a leak in tank FA-129." Upjohn may be charged with this expectation, the majority says, because, under the "imputed-collective-knowledge standard," information collected by the employees who recorded the level of fluid in the tank was immediately imputed to the corporation. The leak was expected because the company knew, on the first day the tank was used on the clindamycin manufacturing run, that fluid levels in the tank went down instead of up after the chemical was pumped into the tank.

Ante, p 212 (emphasis added).

Id., p 215.

The "imputed-collective-knowledge" standard employed by the majority stems from agency principles applicable to knowledge attributable to a corporation. In Gordon Sel-Way, Inc v Spence Bros, Inc, 177 Mich. App. 116, 124; 440 N.W.2d 907 (1989), cited by the majority for this "imputed-collective-knowledge" standard, the Court of Appeals considered what a corporation is deemed to know and when it is deemed to know it. The Court said that one of the burdens attendant upon the corporate form is that the law imputes the knowledge of individual officers and employees at a certain level of responsibility to the corporation. This principle speaks to persons representing a corporation who, acting in the scope of their employment and authority, learn or do something on behalf of the corporation.

Currently pending before this Court on other issues.

According to Upjohn, the tank level inventory record, consisting of the separate slips of paper on which the daily readings were recorded, was audited on a monthly basis. The records covering the 1982 clindamycin campaign were audited, pursuant to standard Upjohn procedures, between September 1 and September 3, 1982. Only on September 3, 1982, did Upjohn employees at the appropriate level of responsibility determine that a leak had occurred. Further, Upjohn executives testified on deposition that tank FA-129 was specially manufactured of materials not subject to corrosion by clindamycin by-products.

Upjohn executives also testified on deposition that because tank FA-129 was part of a "tank farm," a reduced tank level did not necessarily indicate a leak because materials could have been accidentally or unintentionally diverted to other tanks, production could have been interrupted, or materials may have been intentionally removed from the tank for disposal.

In conclusion, the majority errs in finding that Upjohn must have known on the first day of the clindamycin manufacturing campaign that tank FA-129 was leaking. What Upjohn possessed between August 16 and September 3, 1982, was raw, unanalyzed data, not knowledge. Upjohn could indeed have audited and analyzed the data and drawn conclusions on a daily basis. Those conclusions, if known to a person "at a certain level of responsibility," might be imputed to the corporation. The adequacy of Upjohn's procedures for auditing and analyzing the fluid level of the tank is a question of reasonableness under the circumstances, and therefore is not a question of law.

I would reverse and remand for trial.