R.T. Vanderbilt Co. v. Hartford Accident & Indem. Co., 156 A.3d 539, 171 Conn. App. 61 (2017)

March 7, 2017 · Connecticut Appellate Court · AC 36749; (AC 37140); (AC 37141); (AC 37142); (AC 37143); (AC 37144); (AC 37145); (AC 37146); (AC 37147); (AC 37148); (AC 37149); (AC 37150); (AC 37151)

156 A.3d 539, 171 Conn. App. 61

R.T. VANDERBILT COMPANY, INC.
v.
HARTFORD ACCIDENT AND INDEMNITY COMPANY et al.

AC 36749
(AC 37140)
(AC 37141)
(AC 37142)
(AC 37143)
(AC 37144)
(AC 37145)
(AC 37146)
(AC 37147)
(AC 37148)
(AC 37149)
(AC 37150)
(AC 37151)

Appellate Court of Connecticut.

Argued March 30, 2016
Officially released March 7, 2017

Elizabeth J. Stewart, New Haven and Jacob M. Mihm, pro hac vice, with whom were Rachel Snow Kindseth, New Haven and, on the brief, Francis J. Brady, Marilyn B. Fagelson, New Haven, Stephen Hoke, pro hac vice, and David H. Anderson, pro hac vice, for the appellant-cross appellee (substitute plaintiff).

Wayne S. Karbal, pro hac vice, with whom were Jeffrey J. Tinley, Amita Patel Rossetti, Waterbury, and, on the brief, Alan M. Posner, pro hac vice, for the appellees-cross appellants (named defendant et al.).

Michael J. Smith, pro hac vice, and Bryan W. Petrilla, pro hac vice, with whom was John F. Conway, Wallingford, for the appellees-cross appellants (defendant Mt. McKinley Insurance Company et al.).

Lorraine M. Armenti, pro hac vice, with whom were Frank H. Santoro, Hartford, Shayne W. Spencer, pro hac vice, and, on the brief, Kathleen J. Devlin, pro hac vice, and R. Cornelius Danaher, Jr., Hartford, for the appellees-cross appellants (defendant Continental Casualty Company et al.).

Michael L. Duffy, pro hac vice, with whom, on the brief, were Michael G. Albano, Hartford, and Amy R. Paulus, pro hac vice, for the appellee-cross appellant (defendant Old Republic Insurance Company).

Lawrence A. Serlin, pro hac vice, with whom was Laura Pascale Zaino, Hartford, for the appellees-cross appellants (defendant Pacific Employers Insurance Company et al.).

Robert M. Flannery, pro hac vice, with whom were William A. Meehan and, on the brief, Alexander J. Mueller, pro hac vice, and Stephen T. Roberts, for the appellees-cross appellants (defendant Certain Underwriters at Lloyd's, London, et al.).

Lawrence A. Levy, pro hac vice, with whom were Louis B. Blumenfeld, Hartford and, on the brief, Richard S. Feldman, pro hac vice, for the appellees-cross appellants (defendant Fireman's Fund Insurance Company et al.).

Kevin M. Haas, pro hac vice, with whom were Matthew G. Conway and, on the brief, MaryKate J. Geary, West Hartford and Marianne May, pro hac vice, for the appellee-cross appellant (defendant Westport Insurance Corporation).

Lawrence D. Mason, pro hac vice, with whom were John A. Lee, pro hac vice, and, on the brief, Dwight A. Kern, for the appellee-cross appellant (defendant National Casualty Company).

Kathleen D. Monnes, with whom was Erick M. Sandler, Hartford, for the appellees-cross appellants (defendant St. Paul Fire and Marine Insurance Company et al.).

Timothy G. Ronan, Stamford and Assaf Z. Ben-Atar, Bridgeport, filed a brief for the appellee-cross appellant (defendant Employers Mutual Casualty Company).

John E. Rodewald, pro hac vice, and David A. Slossberg, Milford, filed a brief for the appellee-cross appellant (defendant Munich Reinsurance America, Inc.).

Todd A. Bromberg and Laura A. Foggan filed a brief for the Complex Insurance Claims Litigation Association as amicus curiae.

Edward J. Stein, Stamford, John M. Leonard, pro hac vice, Amy Bach, pro hac vice, and Heather R. Spaide filed a brief for United Policyholders as amicus curiae.

Lavine, Beach and Bear, Js.

LAVINE, BEACH and BEAR, Js.

TABLE OF CONTENTS

Page

I. FACTS...548

A. Factual and Procedural History...548

B. Issues on Appeal...553

II. GOVERNING LEGAL PRINCIPLES...554

A. Principles of Insurance Law...554

B. Complex Asbestos Litigation...557

C. Standard of Review...557

III. ALLOCATION OF DEFENSE AND INDEMNITY COSTS...558

A. Trigger of Coverage...559

1. Whether There Is Controlling Connecticut Precedent...562

2. Whether Trigger of Coverage Is Question of Law or Fact, and Whether Expert Testimony of Dr. Kratzke Was Properly Excluded...564

3. What Trigger of Coverage Theory Governs Long-Tail Asbestos Litigation in Connecticut...571

B. Unavailability of Insurance Rule...574

1. Unavailability Rule...575

2. Equitable Exception and Testimony of Professor Priest...584

3. Whether Trial Court Properly Applied Unavailability Rule...589

4. Nonasbestos Particulates...593

C. Default Date of First Exposure and Pre-1962 Liabilities...597

1. Allocation of Past Payments under 2002 Allocation Agreement...600

2. Default Date of First Exposure for Pending and Future Actions...604

D. Mathematical Allocation Formula and Orphan Shares...606

1. Orphan Shares...608

2. Maximum Coverage Block...609

3. Additional Allocation Issues...610

E. Prospective Application of Court's Rulings...612

1. Whether Trial Court Intended to Apply its Allocation Rules Prospectively...614

2. Whether Prospective Only Application Would Be Permissible under Connecticut Law...618

3. Guidance on Remand...622

IV. SCOPE OF COVERAGE AND POLICY EXCLUSIONS...622

A. Pollution Exclusions...623

1. Standard Pollution Exclusion...623

2. Nonstandard Pollution Exclusions...642

B. Occupational Disease Exclusions...644

1. Facts...644

2. Analysis...644

C. Duty to Defend Under Continental's 1968-1977 Umbrella Policies...652

1. Facts...652

2. Analysis...654

V. OTHER CLAIMS...659

A. Admission of Posner Charts...659

B. Exhaustion of Primary Policies...664

1. Products Claims...664

2. Application of 2002 Allocation Agreement...666

C. Duty to Defend Under Old Republic's Excess Policies...667

D. Duty to Defend Under Continental's 1965-1968 and 1977-1978 Excess Policies...669 VI. CONCLUSION...672

APPENDIX A-PARTIES JOINING APPELLATE CLAIMS...673

The present action arises from thousands of underlying lawsuits alleging injuries from exposure to industrial talc mined and sold by the plaintiff, R.T. Vanderbilt Company, Inc. (Vanderbilt),¹ that purportedly contained asbestos. In this interlocutory appeal, Vanderbilt and the defendants, approximately thirty insurance companies that issued comprehensive general liability insurance policies to Vanderbilt between 1948 and 2008, are seeking, among other things, a declaratory judgment determining their respective obligations with regard to the underlying actions. Through a series of bifurcation orders, the trial court, Shaban, J. , divided the trial into four phases, and the case reaches us now, following the second phase of the trial, on the parties' appeals and cross appeals from several decisions of the court. Before the trial proceeds further, the parties ask that we address approximately twenty issues-primarily questions of law-that will significantly impact the adjudication of the remaining trial phases. These issues present a number of questions of first impression in Connecticut and, in some instances, nationally.² Although most relate to the methodology by which insurance obligations are to be allocated with respect to long latency asbestos related claims that implicate multiple policy periods, the parties also challenge the trial court's rulings with respect to the interpretation of various scope of coverage and exclusion provisions in the Vanderbilt policies, whether certain of the primary policies have been exhausted, and other evidentiary and miscellaneous issues. As detailed more fully hereinafter, we affirm in part and reverse in part the rulings of the trial court.

I

FACTS

A

Factual and Procedural History

The following facts, as found by the trial court, and procedural history are relevant to our resolution of the issues on appeal. Vanderbilt is a Connecticut corporation engaged in the mining and sale of various chemical and mineral products. In 1948, it began to produce industrial talc through its subsidiary, Gouverneur Talc Company. Vanderbilt continued to mine and sell talc until 2008, when it ceased production and sold off the last of its inventory.

Over the past several decades, thousands of underlying actions have been filed against Vanderbilt in various jurisdictions throughout the United States, many of which remain pending. Those actions alleged that talc and silica mined and sold by Vanderbilt contained asbestos or otherwise caused diseases that are correlated to asbestos exposure, such as mesothelioma, other asbestos related cancer, and asbestosis (collectively, asbestos related disease). In response, Vanderbilt has taken the position that its industrial talc does not contain asbestos. From the time that it started mining talc, Vanderbilt purchased or attempted to purchase primary and secondary comprehensive general liability insurance to cover the defense and indemnity costs of asbestos related claims.

Vanderbilt brought the present action against several insurance companies that issued it primary insurance policies between 1948 and 2008: defendant Hartford Accident & Indemnity Company (Hartford); defendant American International Specialty Lines Insurance Company (American International); and defendants Continental Casualty Company, Columbia Casualty Company, and Continental Insurance Company (collectively, Continental). Vanderbilt alleged, among other things, that Hartford and Continental (primary insurers) had breached their contractual obligations to pay their proper shares of defense and indemnity costs in the underlying actions. Vanderbilt also sought a declaratory judgment as to the parties' respective rights and responsibilities under the policies at issue.

Continental subsequently filed a third party complaint against various insurance companies that had provided secondary coverage-umbrella or excess³ -to Vanderbilt during the time that it was in the talc business. Those defendants, as well as other secondary insurers later made parties to the case, (collectively, secondary insurers) include: ACE Property & Casualty Insurance Company; American Insurance Company; Arrowood Indemnity Company; Century Indemnity Company; Employers Mutual Casualty Company; Everest Reinsurance Company (Everest); First State Insurance Company; Fireman's Fund Insurance Company (Fireman's); Government Employees Insurance Company (GEICO); Harbor Insurance Company; Mt. McKinley Insurance Company (Mt. McKinley); Munich Reinsurance America, Inc., formerly known as American Reinsurance Company; National Casualty Company (National Casualty); National Union Fire Insurance Co. of Pittsburgh, PA; Old Republic Insurance Company (Old Republic); Pacific Employers Insurance Company (Pacific); Royal Indemnity Company; St. Paul Fire and Marine Insurance Company; Travelers Casualty and Surety Company, formerly known as Aetna Casualty and Surety Company; Twin City Fire Insurance Company; Westport Insurance Corporation; Zurich International Ltd.; and Certain Underwriters at Lloyd's, London (Lloyd's), Certain London Market Insurance Companies, American International Underwriters Insurance Company, and Granite State Insurance Company (collectively, London insurers). Vanderbilt thereafter brought direct claims against these third party secondary insurers.

Vanderbilt's operative complaint contains six counts.⁴ In the first count, Vanderbilt sought a declaratory judgment regarding Continental's duty to defend and indemnify it in the underlying actions. The same declaratory relief was sought as to Hartford (count two) and the secondary insurers (count five). Vanderbilt also asserted breach of contract claims against Continental (count three), Hartford (count four), and certain secondary insurers (count six), claiming that the insurers had failed to fully defend and indemnify it against the underlying claims.

Prior to the start of trial, the trial court issued a series of scheduling orders, pursuant to which it separated the trial into four phases. In the first two phases, which were tried to the court and have been completed, the court addressed Vanderbilt's declaratory judgment claims and related counterclaims and cross claims. The primary issue before the court in those phases was how insurance obligations are to be allocated with respect to long latency⁵ asbestos related claims alleging injuries that occur over the course of years or even decades and, therefore, potentially implicate multiple insurance policy periods. Specifically, in Phase I, the court addressed the question of how defense costs for the underlying actions were to be allocated as between Vanderbilt and its insurers. That required a determination of (1) the periods during which the defendants' insurance policies were in effect and (2) whether Vanderbilt should be treated as self-insured for any period so as to create an equitable obligation to contribute to the costs of its defense. In Phase II, the court considered the same questions with respect to indemnity costs. In that phase, the court also issued rulings with respect to the meaning of various policy provisions, the exhaustion of Vanderbilt's primary policies, and related issues. In Phase III of the trial, which also will be tried to the court, the court plans to adjudicate the defendants' claims for recovery of overpayment of insurance costs. In Phase IV, Vanderbilt's breach of contract claims against its insurers are to be tried to a jury.

In addressing the allocation questions in Phases I and II, the trial court proceeded on the assumption that Connecticut follows a pro rata, time-on-the-risk approach to allocating insurance obligations in long-tail cases. See footnote 5 of this opinion. Under that allocation scheme, the court assumed that a victim of asbestos related disease suffers continuous injuries commencing at the time of initial exposure to asbestos and extending until disease manifests, and, therefore, that defense and indemnity costs must be allocated across all of the insurance policies on the risk (i.e., potentially liable) during that period (allocation block). The court further assumed that (1) the policyholder is responsible for a pro rata share of costs for any period during which it is uninsured or underinsured (proration to the insured), including so-called "orphan share" periods covered by policies that were lost, destroyed, or issued by insurers that subsequently became insolvent; but (2) Connecticut has embraced an unavailability of insurance exception pursuant to which there is no proration to the insured for periods during which insurance is not available. Applying these principles to the present case, the court held evidentiary hearings during Phases I and II to determine, among other things, whether defense and indemnity insurance coverage, respectively, was available for asbestos related claims between 1948 and 2008 and, if so, whether Vanderbilt availed itself of such coverage.

In memoranda of decision issued following the first and second phases of the trial, the court found that occurrence based⁶ comprehensive general liability insurance covering asbestos related claims was generally available for purchase from 1948 through 1955. Although Vanderbilt had purchased such policies, it was unable to locate them. The court therefore determined that Vanderbilt would be treated as self-insured and required to bear its pro rata share of defense and indemnity costs for that period. From 1956 through 1961, by contrast, Vanderbilt had purchased insurance policies that provided sufficient coverage for asbestos related claims.⁷ Accordingly, the court determined that Vanderbilt would not be treated as self-insured for those years. The court also found that, with regard to the period from 1962 through March, 1986, it was undisputed that insurance coverage for asbestos claims was generally available and that Vanderbilt had secured such coverage. Accordingly, Vanderbilt was not treated as self-insured for that period.

As to the period of 1986 through 2008, the court observed that there was considerable conflicting evidence as to whether coverage for asbestos claims was available to Vanderbilt and whether Vanderbilt in fact obtained adequate coverage. Noting that a flood of asbestos related injury claims had led the insurance industry generally to cease offering policies covering such claims after 1985, the court found that there was insufficient evidence to establish that occurrence based or claims-made policies covering asbestos related disease were available from March, 1986 to March, 1993, or after April, 2007. Accordingly, Vanderbilt was not deemed to be responsible for a pro rata share of costs during those periods, even though it did not carry any coverage for asbestos claims. By contrast, the court found that, although asbestos related coverage was not generally available to companies in the mining and chemical industries at any time after 1985, Vanderbilt was, in fact, able to purchase primary claims-made policies that provided defense-and, in one instance, indemnity-cost coverage for the period of March 3, 1993 through April 24, 2007. Because the court found that Vanderbilt was knowingly underinsured during that period of time, it concluded that Vanderbilt would be considered to be self-insured with respect to defense-but not indemnity-cost coverage for that fourteen year period.

The court also found a gap of $700,000 in indemnity coverage for the period of March 25, 1978 to April 26, 1978, for which Vanderbilt would be considered to be self-insured. It further found that some of Vanderbilt's insurers were insolvent and determined that Vanderbilt would be treated as self-insured for any liability that would have been allocated to the insolvent insurers.

With respect to the allocation of costs between Vanderbilt's insurers, the court upheld and enforced a 2002 settlement agreement between Hartford and Continental pursuant to which those primary insurers (1) allocated all obligations related to the underlying claims across their primary policies covering the 1962-1986 period and (2) applied a default date of first exposure of January 1, 1962, for underlying claims for which the actual date of initial exposure to Vanderbilt's talc could not be established. The court also appeared to adopt that default date of first exposure on a prospective basis with respect to pending and future claims against Vanderbilt.

In light of these findings and conclusions, the court determined that the allocation of defense and indemnity costs would be applied prospectively in the following manner, on the basis of a total potential exposure period of 720 months running from 1948 through 2008:⁸ (1) as to defense costs, Vanderbilt would be liable for 265 of the 720 months; (2) as to indemnity costs, Vanderbilt would be liable for ninety-six of the 720 months; and (3) Vanderbilt's responsibility as to both defense and indemnity costs would be adjusted upward for any additional periods when there was a gap in coverage or an insolvent insurer. The court applied these same findings, principles, and allocation rules to underlying actions that alleged harms arising from nonasbestos particulates such as silica. Specifically, the court credited testimony that all of the underlying actions, whether on their face or through subsequent discovery or investigation, involved claims of exposure to asbestos.

In its Phase II decision, the court also considered the applicability of two types of exclusions contained in certain of Vanderbilt's excess and umbrella policies. The court first addressed the claim by several secondary insurers that the pollution exclusion clauses contained in their policies barred coverage for the underlying actions. The court concluded that the relevant policy language was ambiguous as applied to the asbestos related claims and, therefore, that the exclusions did not preclude coverage. The court also addressed the issue of whether occupational disease exclusions contained in certain secondary policies applied only to claims brought by the policyholder's own employees. The court found that the exclusions were unambiguous and that they did, in fact, bar coverage only for claims brought by Vanderbilt's own employees.

The court also addressed the issue of whether certain primary policies issued by Hartford and Continental had been exhausted. The court determined, among other things, that Hartford's primary policies for the period of March 3, 1977 to March 3, 1986, and Continental's primary policies for the period of January 1, 1968 to March 3, 1977, had been exhausted and that the liability of any umbrella or excess carriers sitting above those policies would be determined consistently with established pro rata allocation principles.

Finally, in a separate evidentiary proceeding that was requested by the parties prior to the commencement of the Phase II trial, the court agreed to rule on the issue of whether an umbrella coverage provision contained in Continental's secondary policies covering the period from January 1, 1968 to May 17, 1977, obliged that insurer to defend the underlying claims. The court concluded that Continental had no duty to defend Vanderbilt or to pay for defense costs under those policies. At that time, the court declined to address claims relating to Continental's defense obligations under its 1965-1968 and 1977-1978 secondary policies or to make a determination as to Old Republic's defense obligations.

Following the completion of the Phase II trial, Vanderbilt and several defendants filed appeals and cross appeals, challenging approximately twenty of the court's conclusions and findings.⁹ Additional facts will be set forth as necessary.

B

Issues on Appeal

On appeal, Vanderbilt raises the following issues: 1. Did the court properly hold Vanderbilt responsible for a pro rata share of defense costs for the period of March 3, 1993 through April 24, 2007? 2. Did the court apply the correct mathematical formula in calculating Vanderbilt's pro rata share of defense and indemnity costs? 3. Did the court properly determine that Continental had no duty to defend under its umbrella policies for the period of January 1, 1968 to May 17, 1977? Vanderbilt also has set forth alternative grounds for affirmance with respect to several of the defendants' cross appeals, which are addressed to the extent necessary in this opinion.

The issues raised by Mt. McKinley¹⁰ on cross appeal are: 1. Did the court contravene Connecticut's pro rata allocation law when it (a) considered the unavailability of insurance coverage in determining the allocation of defense and indemnity costs (the unavailability rule), and (b) declined to adopt an equitable exception to the unavailability rule in light of Vanderbilt's continued sale of talc through 2008? 2. Did the court properly (a) conclude that all of the underlying claims alleged exposure only to asbestos and (b) fail to hold Vanderbilt responsible for its pro rata share of defense and indemnity costs for nonasbestos particulate claims? 3. Did the court properly determine that the pollution exclusions in the excess and umbrella insurance policies were ambiguous and inapplicable to the underlying actions? 4. Did the court properly determine that, with regard to any claims of injury where the specific dates of first exposure are unknown, the default date of first exposure would be January 1, 1962? 5. Did the court properly apply its findings prospectively from the date of its Phase II decision? 6. Did the court properly preclude the testimony of (a) a medical expert, Robert A. Kratzke, regarding when asbestos related bodily injuries take place, and (b) an insurance expert, George L. Priest, regarding the economic principles of the transfer of risk and the customs and practices in the field of insurance? 7. Did the court properly enforce the 2002 settlement agreement and determine that certain primary policies issued by Hartford and Continental had been exhausted? 8. Did the court properly admit into evidence charts and spreadsheets that were created by Vanderbilt's expert without finding that an exception to the hearsay rule supported their admission? 9. Did the court apply the correct mathematical formula in calculating Vanderbilt's pro rata share of defense and indemnity costs? The issue raised by Hartford on cross appeal is: Should the allocation rules that were set forth in the court's Phase II decision be applied retroactively to the date on which its primary coverage was exhausted to assure that Hartford's policy limits are not reopened?

The issues raised by Continental on cross appeal are: (1) Did the court properly decline to rule that Continental's obligation under its 1965-1968 umbrella-excess policy was only to reimburse Vanderbilt for the defense costs that Vanderbilt had paid for defending claims that implicated Continental's indemnification obligations? (2) Did the court properly decline to rule that Continental did not have a duty to defend Vanderbilt under its 1977-1978 excess policy? (3) Did the trial court properly create allocation blocks of 720 months for defense and indemnification costs?

The issue raised by National Casualty on cross appeal is: Did the court properly conclude that the occupational disease exclusions in certain umbrella and excess policies apply only to claims brought by Vanderbilt's own employees?

The issue raised by Old Republic on cross appeal is: Should this court direct the trial court to determine Old Republic's defense obligations prior to commencing any additional phases of the trial?

For the reasons set forth hereinafter, we reverse the rulings of the trial court (1) holding Vanderbilt responsible for defense costs for the period of March 3, 1993 through April 24, 2007, (2) applying a default date of first exposure of January 1, 1962, for pending and future claims, and (3) concluding that the occupational disease exclusions apply only to claims brought by Vanderbilt's own employees. We also modify in various respects the mathematical formula and allocation method that the court used to apportion defense and indemnity costs between Vanderbilt and its insurers, and we clarify the prospective nature of the court's determinations. Finally, we encourage the trial court, in its discretion, to determine Old Republic's defense obligations prior to commencing any additional phases of the trial. We affirm the rulings of the court in all other respects.

II

GOVERNING LEGAL PRINCIPLES

In the interests of avoiding repetition, we begin by noting certain well established principles that will be relevant to many of the issues in these appeals. In the parts of the opinion that follow, we set forth additional rules of law as relevant to particular claims.

A

Principles of Insurance Law

"An insurance policy is to be interpreted by the same general rules that govern the construction of any written contract .... In accordance with those principles, [t]he determinative question is the intent of the parties, that is, what coverage the ... [insured] expected to receive and what the [insurer] was to provide, as disclosed by the provisions of the policy.... If the terms of the policy are clear and unambiguous, then the language, from which the intention of the parties is to be deduced, must be accorded its natural and ordinary meaning.... Under those circumstances, the policy is to be given effect according to its terms." (Internal quotation marks omitted.) Lexington Ins. Co. v. Lexington Healthcare Group, Inc. , 311 Conn. 29, 37-38, 84 A.3d 1167 (2014).

"When interpreting [an insurance policy], we must look at the contract as a whole, consider all relevant portions together and, if possible, give operative effect to every provision in order to reach a reasonable overall result." (Internal quotation marks omitted.) Id., at 38, 84 A.3d 1167. "It is axiomatic that a contract of insurance must be viewed in its entirety, and the intent of the parties for entering it derived from the four corners of the policy." (Internal quotation marks omitted.) Springdale Donuts, Inc. v. Aetna Casualty & Surety Co. of Illinois , 247 Conn. 801, 805, 724 A.2d 1117 (1999).

"In determining whether the terms of an insurance policy are clear and unambiguous, [a] court will not torture words to import ambiguity where the ordinary meaning leaves no room for ambiguity .... Similarly, any ambiguity in a contract must emanate from the language used in the contract rather than from one party's subjective perception of the terms.... As with contracts generally, a provision in an insurance policy is ambiguous when it is reasonably susceptible to more than one reading." (Internal quotation marks omitted.) Lexington Ins. Co. v. Lexington Healthcare Group, Inc. , supra, 311 Conn. at 38, 84 A.3d 1167.

"[Words] do not become ambiguous simply because lawyers or laymen contend for different meanings.... Words also do not become ambiguous simply because a contract fails to define them; even when undefined, words are not ambiguous if common usage or our case law gives them a single meaning." (Citation omitted; internal quotation marks omitted.) New London County Mutual Ins. Co. v. Nantes , 303 Conn. 737, 753, 36 A.3d 224 (2012).

"[C]ontext is often central to the way in which policy language is applied; the same language may be found both ambiguous and unambiguous as applied to different facts.... Language in an insurance contract, therefore, must be construed in the circumstances of [a particular] case, and cannot be found to be ambiguous [or unambiguous] in the abstract.... In sum, the same policy provision may shift between clarity and ambiguity with changes in the event at hand ... and one court's determination that [a] term ... was unambiguous, in the specific context of the case that was before it, is not dispositive of whether the term is clear in the context of a wholly different matter." (Citations omitted; emphasis omitted; internal quotation marks omitted.) Lexington Ins. Co. v. Lexington Healthcare Group, Inc. , supra, 311 Conn. at 41-42, 84 A.3d 1167.

Our Supreme Court and this court also have recognized certain rules of construction specific to the interpretation of insurance contracts. First, "[i]t is a basic principle of insurance law that policy language will be construed as laymen would understand it and not according to the interpretation of sophisticated underwriters .... [T]he policyholder's expectations should be protected as long as they are objectively reasonable from the layman's point of view." (Internal quotation marks omitted.) R.T. Vanderbilt Co. v. Continental Casualty Co. , 273 Conn. 448, 462-63, 870 A.2d 1048 (2005) ; Nationwide Mutual Ins. Co. v. Pasiak , 161 Conn.App. 86, 96, 127 A.3d 346 (2015), cert. granted on other grounds, 320 Conn. 913, 130 A.3d 266 (2016).

Second, and relatedly, when the plain language of an insurance policy is found to be ambiguous, courts "apply the contra proferentem rule and interpret a policy against the insurer.... Indeed, our interpretation of ambiguous policy language in favor of coverage under the doctrine of contra proferentem has become near axiomatic in insurance coverage disputes." (Citation omitted; internal quotation marks omitted.) Connecticut Ins. Guaranty Assn. v. Fontaine , 278 Conn. 779, 788-89, 900 A.2d 18 (2006). "The premise behind [this] rule is simple. The party who actually does the writing of an instrument will presumably be guided by his own interests and goals in the transaction. He may choose shadings of expression, words more specific or more imprecise, according to the dictates of these interests.... A further, related rationale for the rule is that [s]ince one who speaks or writes, can by exactness of expression more easily prevent mistakes in meaning, than one with whom he is dealing, doubts arising from ambiguity are resolved in favor of the latter.... This canon ... is more rigorously applied in the context of insurance contracts than in other contracts." (Citation omitted; internal quotation marks omitted.) Israel v. State Farm Mutual Automobile Ins. Co. , 259 Conn. 503, 508-509, 789 A.2d 974 (2002).

Nevertheless, "[t]his rule of construction that favors the insured in case of ambiguity applies only when the terms are, without violence, susceptible of two [equally reasonable] interpretations ...." (Internal quotation marks omitted.) Misiti, LLC v. Travelers Property Casualty Co. of America , 308 Conn. 146, 155, 61 A.3d 485 (2013). Moreover, "the rule [of contra proferentem] should be applied as a tie breaker only when all other avenues to determining the parties' intent have been exhausted." Connecticut Ins. Guaranty Assn. v. Drown , 314 Conn. 161, 195, 101 A.3d 200 (2014) (Rogers, C.J. , concurring). Accordingly, when a policy provision is facially ambiguous, the court should first apply other tools of construction and, if relevant, consult extrinsic evidence of the parties' intentions before construing the agreement against the drafter.¹¹

Third, with respect to an insurer's duty to defend a claim brought against the insured, we note that "[u]nder the well established four corners doctrine, the duty to defend is broader than the duty to indemnify." (Internal quotation marks omitted.) Travelers Casualty & Surety Co. of America v. Netherlands Ins. Co. , 312 Conn. 714, 739, 95 A.3d 1031 (2014) (Netherlands ). "[A]n insurer's duty to defend [is to be] determined by reference to the allegations contained in the [underlying] complaint.... [T]he obligation of the insurer to defend does not depend on whether the injured party will successfully maintain a cause of action against the insured but on whether he has, in his complaint, stated facts which bring the injury within the coverage. If the latter situation prevails, the policy requires the insurer to defend, irrespective of the insured's ultimate liability.... Hence, if the complaint sets forth a cause of action within the coverage of the policy, the insurer must defend.... On the other hand, if the complaint alleges a liability which the policy does not cover, the insurer is not required to defend.... Thus, the duty to defend is triggered whenever a complaint alleges facts that potentially could fall within the scope of coverage ....

"Despite the breadth of this approach, we have recognized the necessary limits of this rule, as we will not predicate the duty to defend on a reading of the complaint that is ... conceivable but tortured and unreasonable.... Thus, although an insurer is not excused from its duty to defend merely because the underlying complaint does not specify the connection between the stated cause of action and the policy coverage ... the insurer has a duty to defend only if the underlying complaint reasonably alleges an injury that is covered by the policy." (Citations omitted; emphasis omitted; internal quotation marks omitted.) Misiti, LLC v. Travelers Property Casualty Co. of America , supra, 308 Conn. at 155-56, 61 A.3d 485.

B

Complex Asbestos Litigation

In the parts of the opinion that follow, we discuss at length the legal rules that govern insurance disputes arising from third party claims alleging long-tail asbestos related diseases. At the outset, however, we recognize that the unique logistical, jurisprudential, and financial challenges posed by asbestos litigation; see American Special Risk Ins. Co. v. A-Best Products, Inc. , 975 F.Supp. 1019, 1020-21 (N.D. Ohio 1997), aff'd, Docket Nos. 97-3919, 97-3920, 97-4004, 166 F.3d 1213, 1998 WL 833750 (6th Cir. November 19, 1998) (decision without published opinion, 166 F.3d 1213 [6th Cir. 1998] ); require that courts adopt rules that, while principled and fair, also are pragmatic and easily administered. See In re School Asbestos Litigation , 789 F.2d 996, 1011 (3d Cir.) (highly unusual nature of asbestos litigation requires novel and flexible solutions), cert. denied sub nom. National Gypsum Co. v. School District of Lancaster , 479 U.S. 915, 107 S.Ct. 318, 93 L.Ed.2d 291 (1986), and cert. denied sub nom. Celotex Corp. v. School District of Lancaster , 479 U.S. 852, 107 S.Ct. 182, 93 L.Ed.2d 117 (1986) ; Continental Casualty Co. v. Synalloy Corp. , 667 F.Supp. 1523, 1538 (S.D. Ga. 1983) (endeavoring to provide guidance "carefully, methodically and with an eye toward practical solutions"); Owens-Illinois, Inc. v. United Ins. Co. , 138 N.J. 437, 477-78, 650 A.2d 974 (1994) (asbestos related injury cases "demand special attention ... [and] specialized treatment"); Champion Dyeing & Finishing Co. v. Centennial Ins. Co. , 355 N.J.Super. 262, 271, 810 A.2d 68 (App. Div. 2002) (Champion ) (allocation mechanism must be efficient and administratively manageable). We consider the issues raised by the present appeals from that perspective.

C

Standard of Review

Unless otherwise noted, the claims at issue in these appeals concern the proper construction of insurance policies or present other pure questions of law, which we review de novo. See Misiti, LLC v. Travelers Property Casualty Co. of America , supra, 308 Conn. at 154, 61 A.3d 485. When the trial court has resolved factual disputes that underlie coverage issues, however, those findings are reviewable on appeal subject to the clearly erroneous standard. National Grange Mutual Ins. Co. v. Santaniello , 290 Conn. 81, 90, 961 A.2d 387 (2009). "Such a finding of fact will not be disturbed unless it is clearly erroneous in view of the evidence and pleadings in the whole record .... [A] finding is clearly erroneous when there is no evidence in the record to support it ... or when although there is evidence to support it, the reviewing court on the entire evidence is left with the definite and firm conviction that a mistake has been committed.... Thus ... [i]t is within the province of the trial court, when sitting as the fact finder, to weigh the evidence presented and determine the credibility and effect to be given the evidence.... Credibility must be assessed ... not by reading the cold printed record, but by observing firsthand the witness' conduct, demeanor and attitude.... An appellate court must defer to the trier of fact's assessment of credibility because [i]t is the [fact finder] ... [who has] an opportunity to observe the demeanor of the witnesses and the parties; thus [the fact finder] is best able to judge the credibility of the witnesses and to draw necessary inferences therefrom." (Citation omitted; internal quotation marks omitted.) Id., at 90-91, 961 A.2d 387.

Last, "[t]he standard under which we review evidentiary claims depends on the specific nature of the claim presented.... To the extent a trial court's admission of evidence is based on an interpretation of [law], our standard of review is plenary.... We review the trial court's decision to admit evidence, if premised on a correct view of the law, however, for an abuse of discretion." (Citations omitted; internal quotation marks omitted.) Statewide Grievance Committee v. Burton , 299 Conn. 405, 415, 10 A.3d 507 (2011). When "a trial court's admission of evidence is based on an interpretation of the [Connecticut] Code of Evidence, our standard of review is plenary." State v. Saucier , 283 Conn. 207, 218, 926 A.2d 633 (2007).

III

ALLOCATION OF DEFENSE AND INDEMNITY COSTS

We begin by addressing a series of claims raised by Vanderbilt and certain of the defendants challenging various aspects of the methodology that the trial court used to allocate defense and indemnity costs among the parties. Vanderbilt argues on appeal that the trial court improperly held it responsible for defense costs for the period March 3, 1993 through April 24, 2007, when, Vanderbilt contends, defense cost coverage was unavailable to companies with a potential exposure to third party asbestos litigation. For its part, Mt. McKinley¹² contends that the trial court improperly (1) held as a matter of law that the continuous trigger of coverage theory governs long-tail asbestos related disease claims in Connecticut; (2) applied an unavailability of insurance rule without an equitable exception for companies that continued to engage in asbestos related businesses after 1985, and also applied that rule to underlying actions alleging harms arising from exposure to nonasbestos particulates; (3) declined to consider expert testimony with respect to both the trigger of coverage question and the proposed equitable exception to the unavailability rule; (4) set an unreasonable default date of first exposure of January 1, 1962, for underlying actions that did not allege a specific date and enforced a settlement agreement between the primary insurers that applied that default date to claims arising prior to 1962; and (5) applied its findings and allocation rules on a prospective basis. In addition, both Vanderbilt and several of the defendants contend that the trial court applied an incorrect mathematical formula when calculating Vanderbilt's pro rata share of defense and indemnity costs, although Vanderbilt and the defendants challenge different aspects of the court's allocation formula.

For the reasons presented hereinafter, we agree with Vanderbilt that it should not have been allocated any post-1986 defense costs. We also agree with certain critiques of the trial court's allocation formula,¹³ and we clarify the prospective application of the court's decisions. In addition, we agree with Mt. McKinley that it is unreasonable to continue to apply a default date of first exposure of January 1, 1962, to pending and future claims. We otherwise reject the defendants' challenges to the allocation methodology applied by the trial court.

As we noted in part II B of this opinion, the adjudication of insurance coverage disputes arising from long-tail toxic torts such as asbestos related disease presents a number of complex and unique legal questions. The trial court confronted several of these questions in its Phase I and Phase II decisions, in which the court sought to determine what share of the defense and indemnity costs, respectively, should be borne by Vanderbilt as the insured and what share should be borne by its various insurers. In adopting an allocation methodology, the court addressed four general questions. First, in the case of progressive, long latency diseases such as asbestosis and asbestos related cancers that do not manifest until years or even decades after exposure to a toxic agent, when does injury occur so as to trigger insurance coverage? Second, under a pro rata allocation approach pursuant to which a policyholder is liable for periods when it self-insures, should the policyholder also be held liable for periods when coverage is unavailable? Third, what default date of first exposure should be applied when neither the complaint nor the discovery process reveals when an underlying plaintiff was first allegedly exposed to a toxic agent? Fourth, according to what mathematical formula should the respective liabilities of the insured and the various insurers be calculated?

In this part of the opinion, we address each of these substantive questions, as well as related challenges to the trial court's specific factual findings, legal conclusions, and evidentiary rulings. We also consider whether the findings and conclusions adopted by the trial court, as modified herein, apply to underlying actions alleging harms arising from exposure to nonasbestos particulates such as silica, and to what extent the court's findings and conclusions should be applied on a prospective only basis.

A

Trigger of Coverage

The first task in evaluating the parties' obligations with respect to underlying asbestos actions is to determine what event or events "trigger" an insurer's defense and indemnification obligations under the many insurance policies that Vanderbilt has purchased over the years. The trial court concluded that established Connecticut precedent compelled it to adopt a continuous trigger theory, under which every insurance policy in effect from the date a claimant is first exposed to asbestos until the date-often decades later-when the claimant manifests an asbestos related disease is on the risk for defense and liability costs. On appeal, Mt. McKinley argues that (1) the trial court was not bound by Connecticut precedent to adopt any particular trigger theory, (2) the trial court should have allowed expert medical testimony on the trigger issue, and (3) this court should adopt an injury-in-fact trigger theory with respect to asbestos related cancers,¹⁴ under which only those policies in effect when a complainant's condition becomes malignant must bear a share of liability and defense costs. We agree that our Supreme Court has not yet adopted continuous trigger as the law of Connecticut but, for the reasons discussed hereinafter, we adopt that theory of coverage as a matter of law for all asbestos related disease claims.¹⁵

Trigger of coverage is a concept used by courts to determine whether and when an event implicates a particular insurance policy. The term "trigger" does not appear in any of the policies, but is instead "a label for the event or events that under the terms of the insurance policy determines whether a policy must respond to a claim in a given set of circumstances." (Internal quotation marks omitted.) Owens-Illinois, Inc. v. United Ins. Co. , supra, 138 N.J. at 447, 650 A.2d 974 ; see also A.W. Chesterton Co. v. Massachusetts Insurers Insolvency Fund , 445 Mass. 502, 518, 838 N.E.2d 1237 (2005) ("[t]rigger of coverage is a term of art whereby the court describes what must occur during the policy period for potential coverage to commence under the specific terms of an insurance policy" [internal quotation marks omitted] ).

The comprehensive general liability policies at issue in the present case typically obligate the issuers to reimburse Vanderbilt for defense and indemnity costs for bodily injuries to third parties caused by an "occurrence."¹⁶ For example, one typical policy, a March 3, 1983 umbrella policy written by the Gibraltar Casualty Company (Gibraltar), the predecessor of Mt. McKinley, provides in relevant part: "[The insurer] will pay on behalf of the [i]nsured the [u]ltimate [n]et [l]oss, in excess of the applicable underlying or retained limit, which the [i]nsured shall become legally obligated to pay as damages because of ... [p]ersonal [i]njury ... to which this policy applies, caused by an [o]ccurrence." The policy further defines an occurrence as "an accident, a happening, an event, or a continuous or repeated exposure to conditions which results during the policy period in [p]ersonal [i]njury ...." (Emphasis added.)

According to the plain language of the policy, then, an insurer is liable only for those costs that result from injuries sustained while the policy is in effect. Indeed, it is black letter law that "[u]nder a liability policy providing coverage for each [injury] during the policy period, a risk insured against by the policy must occur during the policy period in order for coverage to be triggered." 7 S. Plitt et al., Couch on Insurance (3d Ed. Rev. 2013) § 102:23, p. 102-73.

In the case of common third party liability claims such as premises liability, dog bites, and motor vehicle accidents, there typically is little doubt whether personal injuries alleged to have resulted from an incident originated during a particular policy period. See J. Michaels et al., " The Avoidable Evils of 'All Sums' Liability for Long-Tail Insurance Coverage Claims," 64 U. Kan. L. Rev. 467, 467 (2015). With respect to progressive, long latency injuries such as asbestos related disease, by contrast, the trial court recognized that it is far more difficult to pinpoint exactly when the alleged injuries occurred and, therefore, whether they trigger coverage under any particular liability policy. See id.; see also Continental Casualty Co. v. Employers Ins. Co. of Wausau , 60 A.D.3d 128, 144-45, 871 N.Y.S.2d 48 (2008), leave to appeal denied, 13 N.Y.3d 710, 890 N.Y.S.2d 447, 918 N.E.2d 962 (2009) (decision without published opinion, 13 N.Y.3d 710, 918 N.E.2d 962, 890 N.Y.S.2d 447 [2009] ). This is true for a number of reasons.

First, asbestos exposure has been linked to various different diseases-asbestosis, mesothelioma, and lung cancer, among others-each of which has a distinct etiology, symptomology, and course of progression. See 29 C.F.R. § 1910.1001, App. G ; 29 C.F.R. § 1926.1101, App. I ; One Beacon American Ins. Co. v. Huntsman Polymers Corp. , 276 P.3d 1156, 1159 (Utah App.), cert. denied, 285 P.3d 1229 (Utah 2012) (decision without published opinion, 285 P.3d 1229 [Utah 2012] ); A. Bernstein, " Asbestos Achievements," 37 Sw. U.L. Rev. 691, 706-707 (2008). A claimant in an underlying action may allege that asbestos exposure resulted in more than one of these diseases, or may fail to specify the precise nature of the injuries alleged. See, e.g., Complaint of Catherine Mesker (alleging that decedent suffered "asbestosis, carcinoma, pneumoconiosis and emphysema"); Complaint of Antonio Ciaramitaro (alleging only that complainant "suffers from an 'asbestos-related disease' "). Accordingly, it often is difficult to assess at what point the injuries in any particular case are alleged to have commenced and it is almost impossible to generalize across thousands of underlying cases.

Second, the latency period for asbestos related diseases can be extremely long. Mesothelioma, for example, may not manifest until forty years or more after exposure. See 29 C.F.R. § 1926.1101, App. I ; Ricigliano v. Ideal Forging Corp. , 280 Conn. 723, 744, 912 A.2d 462 (2006). In addition, many individuals who develop asbestos related disease are repeatedly exposed to multiple sources of asbestos over the course of several years or more. See Continental Casualty Co. v. Employers Ins. Co. of Wausau , supra, 60 A.D.3d at 144, 871 N.Y.S.2d 48. It is not well established what threshold of exposure must be reached in order for disease to manifest. See Keene Corp. v. Ins. Co. of North America , 667 F.2d 1034, 1040 (D.C. Cir. 1981), cert. denied, 455 U.S. 1007, 102 S.Ct. 1644, 71 L.Ed.2d 875 (1982) ; A. Bernstein, supra, 37 Sw. U.L. Rev. 707. This makes it exceedingly difficult, when all that is known is the date that disease manifested or was diagnosed, to back in to a reasonable estimate of the date at which exposure crossed a threshold level of toxicity. See Continental Casualty Co. v. Employers Ins. Co. of Wausau , supra, at 148, 871 N.Y.S.2d 48 ; see also Green v. General Dynamics Corp. , 245 Conn. 66, 72-73, 712 A.2d 938 (1998) (explaining that exact time of origin of occupational diseases is necessarily obscure); 7 S. Plitt et al., supra, § 102:24, p. 102-84 ("[t]he problem of determining when the coverage-triggering event occurred is particularly acute where there is a lengthy interval between an injury from a wrongful act and the manifestation of that injury in ascertainable damage which forms the basis for a claim against the insured").

Third, and perhaps most importantly, the etiology of the primary asbestos related diseases is complex and not yet fully understood. See A. Bernstein, supra, 37 Sw. U.L. Rev. 707. As we explain more fully hereinafter, there continues to be a split of opinion, even among medical experts, as to what constitutes the initial "injury" in these types of cases.

The determination of which insurance policies are triggered in a case alleging asbestos related disease thus requires that courts answer two questions. First, at what point or points can the injury be said to occur? If, for example, a hypothetical complainant inhales asbestos dust in year one; subsequently undergoes a series of asbestos related genetic mutations; develops a latent malignancy in year fifteen; and is diagnosed with asbestos related lung cancer in year twenty when symptoms begin to appear; should any defense and indemnity costs be allocated to the insurance policy or policies in effect in (1) year one, when the complainant was first exposed to asbestos (the exposure or initial exposure theory), (2) year fifteen, when malignancy emerged (the injury-in-fact theory), (3) year twenty, when the disease manifested and was diagnosed (the manifestation theory), or (4) every year from one through twenty (the continuous or multiple trigger theory)? Second, should that question be resolved by a reviewing court as a matter of law, or by the trier of fact as a question of medical fact?

1

Whether There Is Controlling Connecticut Precedent

Before we address these questions, we briefly consider Vanderbilt's contention that they already have been resolved by our Supreme Court. Specifically, Vanderbilt asserts that, in Security Ins. Co. of Hartford v. Lumbermens Mutual Casualty Co. , 264 Conn. 688, 826 A.2d 107 (2003) (Security ), the Supreme Court adopted the continuous trigger theory as the law governing all asbestos related bodily injury claims in Connecticut. Mt. McKinley counters that the question of trigger was not before the Supreme Court in Security and remains unresolved. We agree with Mt. McKinley.

Like the present case, Security involved a dispute over the proper method of allocating defense costs for underlying asbestos actions alleging long latency bodily injury claims that potentially implicated multiple insurance policies. Id., at 690, 826 A.2d 107. The trial court in Security concluded that, under Connecticut law, insurance costs for long-tail injuries should be allocated according to a pro rata method rather than a so-called "all sums" or "joint and several" liability method. Id., at 697-98, 826 A.2d 107. In jurisdictions that follow an all sums approach, the policyholder is permitted to collect its total liability, up to the policy limit, under any policy in effect during the periods in which the progressive injuries occurred. In re Viking Pump, Inc. , 27 N.Y.3d 244, 255-56, 52 N.E.3d 1144, 33 N.Y.S.3d 118 (2016). The burden then falls to the insurer from whom the policyholder chooses to recover to seek contribution from the insurers that issued other triggered policies. Id. Under pro rata allocation, by contrast, losses are allocated equally at the outset between all of the years (or months, etc.) during which the long-tail injuries occurred. Id., at 256, 33 N.Y.S.3d 118, 52 N.E.3d 1144. Another key difference between the two allocation methods is that, under the all sums approach, a policyholder bears no liability for periods of injury during which it is uninsured; it is free to call upon one triggered policy and then the next, seriatim, until all of the policy limits are exhausted. See Security Ins. Co. of Hartford v. Lumbermens Mutual Casualty Co. , supra, 264 Conn. at 701-702, 826 A.2d 107. In most pro rata jurisdictions, by contrast, the insured is liable for costs attributable to injuries that occurred during any period when it was uninsured or underinsured. Id.

Having adopted a pro rata allocation scheme, the trial court in Security further concluded that (1) the underlying litigation "involved a continuous trigger situation such that all asbestos related injury policies issued during the extended exposure period have been triggered for coverage and all companies that issued such policies are responsible for defense costs"; (internal quotation marks omitted) id., at 696-97, 826 A.2d 107 ; and (2) defense costs should be prorated to the insured with respect to periods for which the insured had lost or destroyed its policies or otherwise assumed the obligations of an insurer. Id., at 690, 826 A.2d 107. On appeal, the policyholder did not challenge the former conclusion, but instead challenged only the trial court's adoption of a pro rata allocation scheme. Id., at 699-706, 826 A.2d 107. Moreover, our Supreme Court, which delineated the four prevailing trigger theories in a footnote, took no position as to whether the trial court had properly adopted the continuous trigger theory. See id., at 697, 826 A.2d 107 n.12. Because the question of trigger was not at issue on appeal and the Supreme Court did not opine thereon, we agree with Mt. McKinley that Security does not establish continuous trigger or any other trigger theory as the law of Connecticut. See Dept. of Public Safety v. Freedom of Information Commission , 103 Conn.App. 571, 582 n.10, 930 A.2d 739 ("[i]t is axiomatic that an appellate decision stands only for those issues presented to, and considered by, the court in that particular appeal"), cert. denied, 284 Conn. 930, 934 A.2d 245 (2007).

We also are not persuaded by Vanderbilt's argument that our Supreme Court adopted the continuous trigger theory in Netherlands . Because Netherlands addressed a fundamentally different sort of underlying claim-the state of Connecticut sued a masonry contractor, among other defendants, for water damage resulting from negligent construction of a public building-it is not clear that any trigger of coverage theory adopted in that case would govern long-tail personal injury claims such as asbestos related disease. See John Crane, Inc. v. Admiral Ins. Co. , 372 Ill.Dec. 167, 991 N.E.2d 474 (Ill. App. 2013) (noting that different trigger theories may apply to personal injury and property damage claims), leave to appeal denied, Nos. 116207, 116209, 116211, 2013 WL 5497780, 5497781, 5497785 (Ill. September 25, 2013) (decisions without published opinions, 996 N.E.2d 11 and 996 N.E.2d 14 [Ill. 2013]); Trustees of Tufts University v. Commercial Union Ins. Co. , 415 Mass. 844, 855, 616 N.E.2d 68 (1993) (similar). In any event, we do not read Netherlands as having definitively adopted any particular trigger theory.

In Netherlands , the defendant insurance company maintained that our Supreme Court had adopted an initial exposure trigger theory in Security , whereas the plaintiff contended that Security adopted an injury-in-fact theory. Travelers Casualty & Surety Co. of America v. Netherlands Ins. Co. , supra, 312 Conn. at 752, 95 A.3d 1031. Consistent with our own analysis, however, the Supreme Court clarified that, in Security , it merely "upheld a pro rata allocation based on the trial court's unchallenged decision that the continuous trigger theory applied"; (emphasis added) id., at 754-55, 95 A.3d 1031 ; and did not endorse one exposure theory to the exclusion of others.

Id., at 754 n.33, 95 A.3d 1031. It is also noteworthy that, in resolving the dispute in favor of the plaintiffs, our Supreme Court stated that "[the plaintiffs argue] that in [Security ] we actually adopted ... an injury-in-fact trigger , under which progressive injuries that span multiple policy periods trigger all policies in effect during the progression of the injury .... We agree with [the plaintiffs], and conclude that, consistent with the continuous trigger situation addressed in Security ... the trial court properly allocated the insurers' pro rata shares ...." (Emphasis added; internal quotation marks omitted.) Id., at 752-53, 95 A.3d 1031. The court did not explain why it appeared to conflate two distinct trigger theories-injury-in-fact and continuous trigger-nor did it purport to adopt, or provide any rationale for adopting, one theory or the other as the law of Connecticut. Accordingly, we believe that the most reasonable interpretation of the Netherlands decision is that our Supreme Court merely (1) upheld the trial court's use of an injury-in-fact trigger theory as reasonable under the specific facts of that case and (2) recognized that the application of an injury-in-fact trigger by the trial court in Netherlands was not inconsistent with the application of a continuous trigger theory by the trial court in Security .¹⁷ We therefore agree with Mt. McKinley that whether continuous trigger or some other trigger theory should govern asbestos related insurance claims remains an open question in Connecticut.

2

Whether Trigger of Coverage Is Question of Law or Fact, and Whether Expert Testimony of Dr. Kratzke Was Properly Excluded

Although the parties focus their attention on the question of which trigger theory should apply to long-tail personal injury claims, we begin our analysis with the predicate question of whether that question is one of fact or law. We conclude that the efficient administration of justice requires that, like the majority of our sister courts, we resolve the question as a matter of law, at least with respect to asbestos related claims.¹⁸ For this reason, we conclude that the trial court properly excluded expert testimony on the subject.

The following additional procedural facts are relevant to our resolution of this issue. During both the Phase I and Phase II trials, Mt. McKinley offered the expert testimony of Robert Kratzke, a medical oncologist, to educate the court about recent advances in scientific knowledge regarding the etiology of cancer, including diseases such as lung cancer and mesothelioma that are strongly correlated to asbestos inhalation. According to Mt. McKinley's expert disclosure and proffers, Dr. Kratzke would have testified as follows. Since approximately the turn of the millennium, much has been learned about the genesis and progression of cancer. Cancer is largely a disease of acquired genetic mutations. Everyone experiences numerous genetic mutations as a result of random gene transcription errors during the ordinary cell division process, and people also can experience mutations as a result of exposure to carcinogens. The vast majority of mutations caused by asbestos, even mutations to cancer-relevant genes, are not harmful and do not result in injury, sickness, or death. Rather, for cancer to occur there must be multiple mutations to an individual cell line and, in Kratzke's opinion, there is neither injury to the body nor disease until the final cancer-relevant mutation takes place.

Kratzke further would have testified that it is impossible to determine when a cancerous cell line experienced its first mutation and that, if an individual suffers multiple exposures to asbestos, there is no way to determine whether any particular exposure, including the first or last exposure, caused a cancerous mutation. Finally, he would have testified that, in most cases of mesothelioma, the last necessary mutation that results in malignancy will occur no more than two to five years before the first cancer symptoms appear. For lung cancer, the corresponding period is five to ten years. Therefore, Kratzke would have opined, it is now firmly established that most types of cancer, including those correlated to asbestos exposure, do not go through a lengthy latency period after the cancer comes into existence, as was previously believed. Rather, the lengthy period between initial exposure to a carcinogen and manifestation of disease reflects the fact that initial mutations are not harmful until combined over time with subsequent mutations.

Vanderbilt filed a motion in limine to exclude Kratzke's testimony, contending that the proffered testimony was both irrelevant and prejudicial. The trial court granted Vanderbilt's motion, ruling that the testimony was either irrelevant or not materially probative of the issues before the court and, therefore, that it would not be of assistance to the court.¹⁹ See Conn. Code Evid. § 7-2 (expert testimony is admissible if it will assist trier of fact in understanding evidence or determining facts in issue); Conn. Code Evid. § 7-3 (expert testimony as to ultimate issue admissible when trier needs expert assistance). The court's ruling appears to have been predicated on its belief that, in Security , our Supreme Court adopted a continuous trigger theory as the law of Connecticut and, therefore, there was no reason for the court to entertain expert testimony that might provide factual support for alternative trigger theories.

On appeal, Mt. McKinley contends that the trial court committed reversible error by precluding Kratzke's testimony. As previously discussed, Mt. McKinley maintains that our Supreme Court has not adopted any particular trigger theory as the law of Connecticut. It further argues that, for there to be any insurance coverage for an underlying claim, Vanderbilt bears the burden of establishing, as a matter of fact, that a particular policy was triggered by bodily injury during the policy period. More generally, Mt. McKinley argues that it would be improper for us to adopt a particular trigger theory as a matter of law, divorced from contemporary scientific knowledge of the etiology and progression of asbestos related diseases. For example, if Kratzke is correct that initial exposure to asbestos fibers does no damage to the human body, and that mesothelioma does not develop until two to five years before the disease manifests, then, in Mt. McKinley's view, it would be improper for us to adopt a theory, such as initial exposure or continuous trigger, that imposes liability on insurers who provided coverage (1) at the time of initial exposure and (2) more than five years prior to the manifestation of disease. A corollary of this argument is that, in adopting a trigger of coverage theory to resolve asbestos related insurance claims, it would be improper for us to rely on decisions that other courts reached decades ago, on the basis of medical and scientific models that may since have been debunked.

Mt. McKinley's arguments are not without merit and, as we have explained; see part III A 1 of this opinion;

we agree that the trial court was mistaken in concluding that it was compelled by Security to apply a continuous trigger theory. Nevertheless, we are not persuaded that the question of which trigger theory applies to the underlying claims is a factual question about which Kratzke's testimony would have proved enlightening.²⁰ Two questions are intertwined in Mt. McKinley's argument. First, is it appropriate for an appellate court to adopt a trigger of coverage theory as a matter of law, rather than allowing the question of when injury occurs to be resolved as a matter of fact, on a case-by-case basis, in light of the best available scientific evidence? Second, if we answer the first question in the affirmative, does current scientific knowledge regarding the etiology and progression of asbestos related diseases place any restrictions on which theory or theories we may adopt? We address each question in turn.

a

Whether Trigger of Coverage Is Question of Law or Fact

Mt. McKinley contends that the process by which inhalation of asbestos fibers leads to various forms of cancer that are not diagnosable until decades later and, in particular, whether any actual injury occurs upon initial exposure, is a question of fact that is beyond the knowledge of any layperson and that can be proven only by evidence and expert testimony. For this reason, it argues, the question of what trigger theory should govern comprehensive general liability policies as applied to long-tail asbestos related claims also is one of fact that requires expert testimony. We disagree.

We begin with the standard of review. "Expert testimony should be admitted when: (1) the witness has a special skill or knowledge directly applicable to a matter in issue, (2) that skill or knowledge is not common to the average person, and (3) the testimony would be helpful to the court or jury in considering the issues." (Internal quotation marks omitted.) Weaver v. McKnight , 313 Conn. 393, 405-406, 97 A.3d 920 (2014). "We review a trial court's decision to preclude expert testimony for an abuse of discretion.... We afford our trial courts wide discretion in determining whether to admit expert testimony and, unless the trial court's decision is unreasonable, made on untenable grounds ... or involves a clear misconception of the law, we will not disturb its decision." (Citations omitted; internal quotation marks omitted.) Id., at 405, 97 A.3d 920. However, whether, as a general matter, expert testimony is required to support a particular type of claim remains a question of law that we review de novo. See Grimm v. Fox , 303 Conn. 322, 329, 33 A.3d 205 (2012).

There are three reasons why, in our view, the trial court properly concluded that it was not necessary to hear Kratzke's testimony prior to adopting a trigger of coverage rule for the underlying actions. First, a trigger of coverage rule defines what is meant by the policy terms "personal injury" or "bodily injury" in the context of a long-tail, progressive disease claim. See Keene Corp. v. Ins. Co. of North America , supra, 667 F.2d at 1043-44. It is well established that experts may not be called to instruct the court on the meaning of a nontechnical contract term. See Fuller v. Metropolitan Life Ins. Co. , 70 Conn. 647, 677, 41 A. 4 (1898) ; see also Sagamore Group, Inc. v. Commissioner of Transportation , 29 Conn.App. 292, 299, 614 A.2d 1255 (1992) (definition of term is question of law to be determined by court and about which expert witness is incompetent to testify). In the present case, one typical policy defines a covered injury simply as "bodily injury, sickness, or disease," while another insures against "bodily injury, mental injury, mental anguish, shock, sickness, occupational disease, non-occupational disease, [and] disability ...." There is no indication that the term "injury," as defined in these policies, is intended to be a medical term of art either generally or with respect to asbestos related claims in particular. Accordingly, in construing the policy language, we must look to the ordinary, dictionary definition of the term. See Heyman Associates No. 1 v. Ins. Co. of Pennsylvania , 231 Conn. 756, 771-72, 653 A.2d 122 (1995) ( Heyman ). Indeed, Mt. McKinley itself appears to recognize that Kratzke was not qualified to opine as to the meaning of "injury" in the Vanderbilt policies.

Second, to the extent that a background understanding of the nature of asbestos related diseases is helpful in selecting among trigger of coverage theories, both courts and commentators have recognized that expert testimony is not required in this context because the relevant medical facts-that (1) asbestos harms the lungs upon inhalation, (2) mesothelioma and lung cancer represent the culmination of a series of genetic mutations that accumulate over a long period of time, and (3) malignancies emerge and begin to proliferate upon the completion of the last necessary mutation-are now so widely understood and incontrovertible that many courts simply assume their truth when resolving the trigger of coverage question. See, e.g., Owens-Illinois, Inc. v. United Ins. Co. , supra, 138 N.J. at 453-54, 650 A.2d 974 ; D. Ramsey, " The Trigger of Coverage for Cancer: When Does Genetic Mutation Become 'Bodily Injury, Sickness, or Disease'?," 41 Santa Clara L. Rev. 293, 307-28 (2001). Because Kratzke's proffered testimony was fully consistent with the accepted understanding of the etiology and progression of asbestos related disease, it was reasonable for the trial court to conclude, in the exercise of its discretion, that no useful purpose would be served by its admission.

This conclusion is bolstered by the fact that, among those courts that have considered expert testimony on the question, radically different conclusions have been drawn from substantially similar medical evidence. Compare, e.g., Eagle-Picher Industries, Inc. v. Liberty Mutual Ins. Co. , 682 F.2d 12, 19 (1st Cir. 1982) (adopting manifestation trigger theory), cert. denied sub nom. Froude v. Eagle-Picher Industries, Inc. , 460 U.S. 1028, 103 S.Ct. 1279, 75 L.Ed.2d 500 (1983) ;

Ins. Co. of North America v. Forty-Eight Insulations, Inc. , 633 F.2d 1212, 1218-19 (6th Cir. 1980) (Forty-Eight Insulations ) (exposure theory), clarified on reh'g, 657 F.2d 814 (6th Cir.), cert. denied, 454 U.S. 1109, 102 S. Ct. 686, 70 L.Ed. 2d 650 (1981) ; Zurich Ins. Co. v. Raymark Industries, Inc. , 118 Ill.2d 23, 42-44, 514 N.E.2d 150, 112 Ill.Dec. 684 (1987) (unique triple trigger theory); Continental Casualty Co. v. Employers Ins. Co. of Wausau , supra, 60 A.D.3d 128, 871 N.Y.S.2d 48 (injury-in-fact theory).

Third, resolving the injury issue as a question of fact on a case-by-case basis would create uncertainty and significantly increase litigation costs. See M. Doherty, " Allocating Progressive Injury Liability Among Successive Insurance Policies," 64 U. Chi. L. Rev. 257, 258 (1997). As the United States Court of Appeals for the Sixth Circuit explained in Forty-Eight Insulations , "[i]n each case where a plaintiff sues an asbestos manufacturer, a hearing could be held to determine at what point the buildup of asbestos in the plaintiff's lungs resulted in the body's defenses being overwhelmed. At that point, asbestosis could truly be said to occur .... From then on, all companies [that] insured the manufacturer would be treated as being on the risk ....

"The only problem with this Solomonian interpretation is that no one wants it. The principal reason is cost. If medical testimony as to asbestosis' origin would have to be taken in each of the thousands of asbestosis cases, the cost of litigation would be prohibitive. This appears to be especially true since many of the asbestosis cases are settled before trial. In addition, it is almost impossible for a doctor to look back and testify with any precision as to when the development of asbestosis crossed the line and became a disease.

"The only thing on which all parties agree is that there is a need for us to arrive at an administratively manageable interpretation of the insurance policies-one that can be applied with minimal need for litigation." (Footnote omitted; internal quotation marks omitted.) Ins. Co. of North America v. Forty-Eight Insulations, Inc. , supra, 633 F.2d at 1217-18. For similar reasons, the Sixth Circuit clarified on rehearing that the same trigger theory should apply both to asbestosis and to asbestos related cancers, notwithstanding the different etiologies of those diseases. See Ins. Co. of North America v. Forty-Eight Insulations, Inc. , 657 F.2d 814, 815-16 (6th Cir. 1981), cert. denied, 454 U.S. 1109, 102 S.Ct. 686, 70 L.Ed.2d 650 (1981) ; cf. N. Andrea, " Exposure, Manifestation of Loss, Injury-in-Fact, Continuous Trigger: The Insurance Coverage Quagmire," 21 Pepp. L. Rev. 813, 850 (1994) (absence of unified approach to trigger of coverage theories adversely affects insurers and insureds). A number of other appellate courts also have resolved the trigger issue as a matter of law, without reliance on expert testimony. See, e.g., Commercial Union Ins. Co. v. Sepco Corp. , 765 F.2d 1543, 1546 (11th Cir. 1985) ; ACandS, Inc. v. Aetna Casualty & Surety Co. , 764 F.2d 968, 972 (3d Cir. 1985) ; Keene Corp. v. Ins. Co. of North America , supra, 667 F.2d at 1044 n.19, 1047 n.25 ; Owens-Illinois, Inc. v. United Ins. Co. , supra, 138 N.J. at 457, 650 A.2d 974.

For these reasons, we are persuaded that a trigger of coverage theory may be adopted as a matter of law without the assistance of expert medical testimony. We therefore turn our attention to Mt. McKinley's argument that it was nevertheless improper to exclude Kratzke's testimony because (1) whichever trigger theory we adopt must at least be compatible with current medical research and understanding of the etiology and progression of asbestos related diseases such as mesothelioma and lung cancer and (2) Kratzke's testimony, if credited, would preclude us from adopting several of the prevailing trigger theories.

b

Whether Current Medical Knowledge Contradicts Initial Exposure or Continuous Trigger Theories

We next consider whether, in adopting a trigger theory to govern long-tail asbestos related claims, we are precluded by contemporary medical knowledge from selecting any of the four prevailing theories. Specifically, Mt. McKinley argues that, if Kratzke's testimony had been admitted, it would have been clear that neither the initial exposure nor the continuous trigger theory provides a medically credible account of the injuries that actually occur as a result of asbestos inhalation.²¹ Vanderbilt responds that, even if we were to credit Kratzke's testimony, it would still be true, on the basis of the undisputed medical facts, that bodily injury occurs immediately upon inhalation of asbestos fibers and continues to occur until disease manifests. We agree with Vanderbilt and conclude that current medical knowledge does not preclude this court from adopting an initial or continuous exposure theory, as well as an injury-in-fact or manifestation theory.

Vanderbilt draws our attention to a number of key concessions that Kratzke made during his deposition. Kratzke acknowledged that, under the prevailing medical model, the lungs begin to exhibit an inflammatory response within hours or days of exposure to asbestos fibers. This inflammation occurs in individuals who later develop mesothelioma or asbestos related lung cancer and, importantly, Kratzke opined that the initial inflammation probably plays a role in causing the genetic mutations that begin to appear within a few weeks or months after exposure and ultimately lead to cancer. This process, by which asbestos related precancerous mutations are formed, takes place throughout the entire latency period of the disease. In its expert disclosure, Mt. McKinley also indicated that Kratzke would testify that this process occurs because asbestos fibers trapped in and around the lungs provoke cell division and the release of toxins from macrophages, causing nearby cells to mutate.

Kratzke also testified in his deposition that, in his view, individual cells cannot, by definition, be injured, because "injury" is a term that he reserves for major systemic diseases and traumas. Nevertheless, he acknowledged that other physicians do speak in terms of "damage" being done to cellular DNA. He also agreed that the term "injury" is too ambiguous to be medically useful when talking about changes to cellular DNA, and he acknowledged that the lung cells of a person who has been exposed to asbestos fibers are not "normal," even during the pre-malignancy period when cancer progenitor cells are accumulating mutations.

Kratzke's opinions, then, were of two types. Empirically, he conceded that asbestos fibers harm the lungs in various ways, almost from the outset. Asbestos causes tissue inflammation, cellular division, and the release of toxins, all of which contribute to and are a but-for cause of the eventual development of asbestos related malignancies in neighboring cells. Semantically, however, Kratzke indicated that oncologists such as himself would not characterize such harms as "injuries" or "diseases" until such time as the cellular damage actually manifests as a malignancy.

As we already have explained, there is no indication that the term "injury," as used in a comprehensive general liability policy, is intended to be a medical term of art, much less a term of art specific to the subspecialty of medical oncology. Moreover, as a number of our sister courts have recognized, and as Kratzke himself appeared to acknowledge, at what point cellular damage caused by a carcinogen such as asbestos begins to qualify as an injury or disease is a matter of perspective even among medical experts. From the standpoint of oncologists such as Kratzke, whose job it is to diagnose and treat active tumors, there is no injury, sickness, or disease until damage goes beyond the cellular level and begins to affect entire organ systems. Kratzke explained that he uses the term "injury" only with reference to "gross injur [ies] such as ... [when] somebody has had their finger cut off [or] somebody has been in a car accident and tissues have been grossly disrupted." By contrast, cellular biologists and pathologists, whose job it is to study cellular structures and functions, recognize that the types of damage that asbestos fibers inflict at the cellular level are also fairly characterized as injuries.²² Because there is no indication that the term "injury" as used in a standard form commercial general liability policy is intended to be limited to gross injuries of the type described by Kratzke, we conclude that his testimony as to the distinctive meaning that clinical oncologists place on that and other contract terms would not illuminate our interpretation thereof. Instead, we look to the ordinary dictionary definitions of those terms. See Heyman Associates No. 1 v. Ins. Co. of Pennsylvania , supra, 231 Conn. at 772, 653 A.2d 122.

Webster's Third New International Dictionary of the English Language Unabridged (2002) defines "injury" as "an act that damages, harms, or hurts ...." It defines "sickness" as "the condition of being ill ... a disordered, weakened, or unsound condition ... a form of disease"; and it defines "disease" as "an impairment of the normal state of the living animal ... or of any of its components that interrupts or modifies the performance of the vital functions ...." In light of Kratzke's acknowledgment that asbestos fibers cause inflammation, abnormal cellular division, and the release of toxins, and that this ongoing process increases the likelihood that malignancies will develop, we have no difficulty concluding that asbestos exposure damages, harms, hurts, weakens, and impairs the body, beginning at the time of exposure and continuing throughout the latency period until the development of malignancy and the ultimate manifestation of cancer. See Owens-Illinois, Inc. v. United Ins. Co. , supra, 138 N.J. at 454, 650 A.2d 974 (rejecting need for medical evidence because "we are satisfied, like most American jurisdictions, that medical science confirms that some injury to body tissue occurs on the inhalation of asbestos fibers, and that once lodged, the fibers pose an increased likelihood of causing or contributing to disease"). Accordingly, we conclude that current medical understanding of asbestos related cancers, to which Kratzke would have testified, is not incompatible with any of the prevailing legal theories of trigger. For that reason, it was not improper for the trial court to conclude, in its discretion, that admission of Kratzke's expert testimony would not assist the court in understanding evidence or determining facts in issue.

3

What Trigger of Coverage Theory Governs Long-Tail Asbestos Litigation in Connecticut

Having concluded that we may adopt, as a matter of law and a question of first impression, any of the four prevailing trigger of coverage theories, we now consider which of those theories should govern the pro rata allocation of insurance obligations with respect to long-tail asbestos litigation in Connecticut. Consistent with the majority of our sister states,²³ we adopt the continuous trigger theory, under which every policy in effect, beginning at the time of initial asbestos exposure and extending through the latency period and up to the manifestation of asbestos related disease, is on the risk for defense and liability costs.

Courts that apply the continuous trigger theory to extended latency, progressive disease claims typically do so for one or more of three primary reasons, each of which we find persuasive. First, courts have adopted continuous trigger as the theory most compatible with the prevailing understanding of the nature and etiology of asbestosis and asbestos related cancers. As we discussed in part III A 2 b of this opinion, it is widely accepted,²⁴ and Kratzke's expert testimony would have confirmed, that asbestos fibers become lodged in the lungs upon inhalation and, almost from the start, cause damage of various sorts within the lung tissues: inflammation, the release of toxins, and abnormal cellular division and mutation. Importantly, asbestos fibers, unlike many other carcinogens, remain permanently lodged in the lungs, where they continue to cause injury. See D. Ramsey, supra, 41 Santa Clara L. Rev. 307, 332, 337. Indeed, "asbestos-related cancers may be almost singularly unique in the degree to which they are initiated and promoted by the continuing presence of the asbestos fibers in the lung." Id., 305; see also Armstrong World Industries, Inc. v. Aetna Casualty & Surety Co. , 45 Cal.App.4th 1, 37, 52 Cal.Rptr.2d 690 (1996) ("[t]he very quality that has made asbestos useful for so long, its indestructibility, also accounts for the problems that result in asbestos-related disease" [internal quotation marks omitted] ), review denied, 1996 Cal. LEXIS 4708 (Cal. August 21, 1996). Accordingly, from a purely factual standpoint, the continuous trigger theory best reflects the fact that asbestos begins to injure the body on a cellular level within hours or days of initial exposure and contributes to the progressive worsening both of asbestosis and of precancerous conditions until the time that those diseases manifest. See Armstrong World Industries, Inc. v. Aetna Casualty & Surety Co. , supra, at 48, 52 Cal.Rptr.2d 690 ; J.H. France Refractories Co. v. Allstate Ins. Co. , 534 Pa. 29, 37, 626 A.2d 502 (1993) ; see also J. Stempel, " Assessing the Coverage Carnage: Asbestos Liability and Insurance after Three Decades of Dispute," 12 Conn. Ins. L.J. 349, 445 (2006) ("continuous trigger is the logical result of the insurance industry's own adoption of an actual injury trigger combined with the [relentless] ability of asbestos to inflict injury over many years"). Our current knowledge about asbestos related disease, then, is most consistent with the theory that the body is continuously injured by the presence of asbestos and the ongoing progression of disease, from exposure through manifestation.²⁵

The second reason that our sister courts have adopted continuous trigger is that that theory of liability also best expresses what is presently unknown about the progression of asbestos related disease. Kratzke conceded in his deposition that, in any particular case, physicians do not know with certainty: (1) how long after exposure to asbestos the lungs begin to develop inflammation; (2) how long the inflammatory response persists; (3) when precancerous mutations commence; (4) the time frame over which precancerous mutations occur; or (5) how long before diagnosis malignancy develops. As one commenter summarized, "the issues involve a series of astonishingly complex events that unfold through the interactions of vanishingly infinitesimal subcellular molecules over the course of an entire lifetime." D. Ramsey, supra, 41 Santa Clara L. Rev. 299. On a micro level, then, there is much about the progression of asbestos related disease that is not yet fully understood.

Furthermore, these epistemological gaps are magnified at the macro level, when we are confronted with a class action or action such as the present case, in which many individual lawsuits are implicated. The thousands of claimants in the underlying complaints allege various different asbestos related diseases, each of which has its own etiology and course of progression. It is not always clear from the face of the complaint which disease or diseases are alleged, and some claimants allege that they were exposed to asbestos supplied by multiple underlying defendants, at different times, in varying amounts, and through diverse mechanisms. See, e.g., Complaint of Joseph A. Campo. Underlying actions may be settled without any of these questions being squarely resolved. In most instances, then, we simply will never know exactly when a particular claimant was exposed to a particular policyholder's asbestos, how much of that policyholder's asbestos was inhaled, when that claimant contracted an asbestos related disease or diseases, and the precise relationship between these events. See One Beacon American Ins. Co. v. Huntsman Polymers Corp. , supra, 276 P.3d at 1159. The continuous trigger theory addresses this conundrum by assuming that, in each case, every exposure contributed to the ongoing worsening of the disease throughout the entire period from initial exposure to manifestation. See J. Michaels et al., supra, 64 U. Kan. L. Rev. 472 (continuous trigger relieves policyholder of burden of proving what share of damages from progressive disease occurred during each policy period); id., 487 (continuous trigger acknowledges uncertainty inherent in long-tail toxic tort claims); see also Ins. Co. of North America v. Forty-Eight Insulations, Inc. , supra, 657 F.2d at 815 ("[M]any of the underlying plaintiffs' complaints against the manufacturers allege both cancer and asbestosis. To [adopt a fact-based trigger theory that would] treat cancer and asbestosis differently would needlessly complicate settlement and defense of the individual lawsuits. We see no reason to create more difficulties for the parties than already exist in this complicated case."). Third, our sister courts have selected the continuous trigger theory as the fairest and most efficient way to distribute indemnity and defense costs among the various policies in effect over the course of a long latency disease claim. See Keene Corp. v. Ins. Co. of North America , supra, 667 F.2d at 1041 ; Montrose Chemical Corp. v. Admiral Ins. Co. , 10 Cal.4th 645, 687, 913 P.2d 878, 42 Cal.Rptr.2d 324 (1995), as modified on denial of reh'g (August 31, 1995); see also Owens-Illinois, Inc. v. United Ins. Co. , supra, 138 N.J. at 451, 650 A.2d 974 ("because it encourages all insurers to monitor risks and charge appropriate premiums, the continuous trigger rule appears to be the most efficient doctrine for toxic waste cases" [internal quotation marks omitted] ). A number of jurists and scholars have observed that repeat players in the field of asbestos litigation tend to adopt "fluctuating positions," alternatively advocating for diametrically opposed initial exposure or manifestation theories of trigger "depend[ing] upon their economic interests in a particular case ...." Keene Corp. v. Ins. Co. of North America , supra, at 1058 (Wald, J., concurring in part); accord N. Andrea, supra, 21 Pepp. L. Rev. 850. From an objective standpoint, however, we believe that the fairest and most reasonable approach is to share the burden of defense and indemnity costs among all those policies on the risk over the full course of the development and progression of a claimant's disease. Cf. J. Rawls, A Theory of Justice (1971) pp. 136-42 (just distribution of burdens is that which would be adopted by parties behind "veil of ignorance" as to their position in particular case). This approach minimizes the likelihood that any one insurer will be forced to shoulder the full expense for injuries that may predominantly have occurred either long before or long after its policy was in effect. See N. Andrea, supra, 815-16. Relatedly, a number of courts have applied a continuous trigger because that theory maximizes the total resources available for recovery. See Quincy Mutual Fire Ins. Co. v. Bellmawr , 172 N.J. 409, 434, 799 A.2d 499 (2002) ; 7 S. Plitt et al., supra, § 102:24, pp. 102-109 through 102-111; Note, "Developments in the Law-Toxic Waste Litigation," 99 Harv. L. Rev. 1458, 1578-81 (1986).

Having thoroughly reviewed the arguments for and against the alternative theories of trigger, we are persuaded to join the majority of our sister courts in adopting continuous trigger as the rule governing long-tail asbestos claims in Connecticut. We believe that continuous trigger best accounts for the progressive nature of asbestos related diseases, as both the layperson and at least some subset of medical professionals would consider the tissue damage and other harms imposed throughout the development of asbestosis and asbestos related cancers to constitute personal or bodily injuries as defined in the standard form commercial general liability policy. We also believe that continuous trigger represents the fairest and most efficient means of resolving and administering complex, multiclaimant asbestos litigation such as the present case. Accordingly, we conclude that the trial court properly applied a continuous trigger in the present case.

B

Unavailability of Insurance Rule

As we discussed in part III A of this opinion, in Security , our Supreme Court adopted a pro rata allocation method for adjudicating long latency loss claims that implicate multiple insurance policies. Security Ins. Co. of Hartford v. Lumbermens Mutual Casualty Co. , supra, 264 Conn. at 720, 826 A.2d 107. Under the pro rata approach, defense and indemnity costs are allocated among insurers on the basis of their time on the risk, but are allocated (prorated) to the insured for periods during which the insured lost or destroyed its policies or was otherwise uninsured or underinsured. See id. In this part of the opinion, we address the parties'²⁶ arguments as to whether costs should be prorated to the insured for periods when it was uninsured not by choice or negligence but because insurance coverage was unavailable. Mt. McKinley contends that there is no authority under Connecticut law for an unavailability of insurance exception to the pro rata allocation method and, therefore, that the trial court improperly exempted Vanderbilt from liability for much of the post-1985 period.²⁷ In the alternative, Mt. McKinley argues that, if we adopt an unavailability rule, we also should adopt an equitable exception to that rule pursuant to which companies that continued to engage in asbestos related business after 1985, when insurance coverage for third party asbestos claims became generally unavailable, should not be able to benefit from that rule. At the very least, Mt. McKinley contends, the trial court should have allowed expert testimony on this question. Mt. McKinley also maintains that the trial court improperly failed to hold Vanderbilt responsible for its pro rata share of costs for underlying claims alleging injuries arising from silica, talc, and other nonasbestos particulate minerals. In response, Vanderbilt submits that the trial court adopted the correct legal rules with respect to availability of insurance, but that the court applied those rules improperly in concluding that defense cost coverage was available to Vanderbilt from March 3, 1993 through April 24, 2007. We agree with Vanderbilt.

The following additional procedural history is relevant to our resolution of these claims. In its Phase I decision, the trial court addressed the question of what insurance coverage block applied to defense costs for Vanderbilt's asbestos related liabilities and for what portions of that coverage block, if any, Vanderbilt should be held responsible. The court proceeded on the assumption that, under Connecticut law, a policyholder is liable for costs attributed to long-tail losses alleged to have occurred during periods when it was self-insured, underinsured, or uninsured, but that costs should not be prorated to the insured for periods during which coverage for a particular risk cannot be acquired (the unavailability rule).

Consistent with those principles, the trial court made findings as to whether defense cost coverage for asbestos related claims was available to Vanderbilt from 1948 to 2008 and to what extent Vanderbilt availed itself of such coverage. With regard to availability, the court found that insurance companies regularly offered occurrence based defense cost and indemnity coverage for asbestos related claims until 1985 but that, by 1986, such policies had become generally unavailable to companies such as Vanderbilt that operated in the mining and chemical industries. After 1985, comprehensive general liability coverage generally was available to companies in those industries only on a claims-made basis²⁸ and only with an asbestos exclusion. Despite finding that asbestos coverage was generally unavailable to companies such as Vanderbilt after 1985, the court ultimately found in its Phase I decision that asbestos related defense coverage was available to Vanderbilt between March, 1993 and April, 2007, during which time Vanderbilt was able to obtain a limited number of primary claims-made policies but was otherwise uninsured for asbestos related claims. Accordingly, and consistent with its understanding of the unavailability of insurance rule, the court concluded that Vanderbilt would be held responsible for defense costs attributable to the March, 1993 through April, 2007, period, but not for the remainder of the post-1985 period when insurance was unavailable. In its Phase II decision, by contrast, the court found that indemnity coverage was unavailable to Vanderbilt throughout more or less the entire 1986-2008 period and, therefore, concluded that Vanderbilt was not responsible for any pro rata share of the indemnity costs for that period. The court applied these same findings with respect to underlying actions alleging harms arising from talc, silica, and other nonasbestos particulates.

1

Unavailability Rule

We begin by addressing Mt. McKinley's argument that our state's appellate courts have never adopted an unavailability of insurance rule and that we should not do so at this time. Although we agree with Mt. McKinley that there is no controlling Connecticut precedent, we conclude, as a matter of first impression, that the trial court properly determined that an unavailability rule comports with the allocation scheme adopted in Security . Accordingly, the trial court did not err in declining to prorate indemnity and defense costs to Vanderbilt for periods during which insurance was unavailable.

a

Whether There Is Controlling Connecticut Precedent

As an initial matter, the parties disagree as to whether our Supreme Court adopted an unavailability of insurance rule in Security . Vanderbilt contends that the court did adopt such a rule, whereas Mt. McKinley maintains that the issue was not before the Supreme Court in that case and that it remains an unresolved question in Connecticut.

There is no question that the trial court in Security applied an unavailability rule in the context of long-tail asbestos litigation. Relying on Stonewall Ins. Co. v. Asbestos Claims Management Corp. , 73 F.3d 1178, 1203 (2d Cir. 1995), modified on denial of reh'g, 85 F.3d 49 (2d Cir. 1996), and Keene Corp. v. Ins. Co. of North America , supra, 667 F.2d at 1058 (Wald, J., concurring in part), the court reasoned that long-tail costs are prorated to the insured only because the insured has elected to self-insure and thereby assume a portion of the risk, and that those rationales do not apply when insurance is not available. See Security Ins. Co. of Hartford v. Lumbermens Mutual Casualty Co. , Superior Court, judicial district of Hartford-New Britain at New Britain, Docket No. CV-96-0475565S, 1999 WL 545745, *8 (July 12, 1999) ("[t]he element of choice, and in turn, the conscious decision of an insured to assume the risks of being 'self-insured' is lost if the insured cannot realistically acquire the particular coverage desired"). The court also opined that "[i]t is inequitable to require an insured to assume defense costs for claims [that] it could not have insured against, when multiple insurers are already obligated to provide the insured with a defense in an action." Id.

On appeal, our Supreme Court held that the trial court had properly applied a pro rata, time-on-the-risk allocation methodology for long-tail asbestos claims that implicate multiple policy periods. Security Ins. Co. of Hartford v. Lumbermens Mutual Casualty Co. , supra, 264 Conn. at 699, 826 A.2d 107. In the course of analyzing the allocation issue, the Supreme Court recognized in a footnote that the trial court had cut off the allocation block in 1985, "at which time asbestos related injury insurance became unavailable ...." Id., at 698, 826 A.2d 107. Later in the decision, in a different footnote, the court concluded that the specific allocation methodology applied by the trial court was "reasonable ...." Id., at 720 n.17, 826 A.2d 107. The Supreme Court also suggested that the trial court had applied the correct "equitable principles" when the trial court concluded that "[i]t would be grossly inequitable to make [the insurers] bear the financial loss of [the policyholder's] actions or inactions that have rendered [the policyholder] essentially uninsured for a period when asbestos related injury insurance coverage was available ." (Emphasis added; internal quotation marks omitted.) Id., at 720, 826 A.2d 107. Vanderbilt reads these portions of Security to mean that our Supreme Court tacitly approved of the trial court's unavailability rule.

As Mt. McKinley correctly notes, however, the issue of the unavailability of insurance was not before the Supreme Court in Security and the court neither discussed the merits of nor expressly upheld an unavailability rule. Accordingly, any tacit approval of the rule in that opinion was dictum. Moreover, although certain of the rationales that the Supreme Court articulated when adopting a pro rata allocation system would seem to favor an unavailability exception, others are arguably in tension with it. Compare, e.g., id., at 719, 826 A.2d 107 (emphasizing that policyholder is party best positioned to determine whether it will be, in essence, self-insured) with id. (declining to impose on insurer "defense costs [that arose] outside of its policy period" and for which the insured never contracted). Because Security did not expressly adopt or foreclose an unavailability rule, and because language in the decision could reasonably be read to support either position, we conclude that the question remains an open one in Connecticut.

b

Whether Connecticut Applies an Unavailability of Insurance Rule

Of those jurisdictions that have adopted a pro rata, time-on-the-risk allocation scheme, a narrow majority also recognize an unavailability rule.²⁹ If one includes those jurisdictions that follow the all sums approach, the vast majority of our sister states do not hold an insured accountable for a pro rata share of long-tail losses that occur during periods when insurance is not available. For the reasons that follow, we agree that costs should not be prorated to the insured for periods during which insurance is unavailable.

The primary rationale that other courts have offered in favor of the unavailability rule is that the justifications that support prorating costs to a policyholder during periods of self-insurance or underinsurance simply do not apply when insurance is not commercially available. Proration to the insured has been justified on the grounds that (1) policyholders should be obliged to accept a share of the risk that they consciously elect to assume by self-insuring; Stonewall Ins. Co. v. Asbestos Claims Management Corp. , supra, 73 F.3d at 1204 ; (2) a policyholder would receive an undeserved windfall if it were to reap the benefits of insurance coverage that it deliberately declined to purchase; Wooddale Builders, Inc. v. Maryland Casualty Co. , 722 N.W.2d 283, 297 (Minn. 2006) ; and (3) proration creates an incentive for policyholders to maintain an unbroken chain of adequate insurance coverage, which has the socially desirable benefits of spreading risk, maximizing the resources available to respond to injuries, and ensuring that no single policy or insurer is made to shoulder a disproportionate share of the costs of a long-tail injury. See United States Fidelity & Guaranty Co. v. Treadwell Corp. , 58 F.Supp.2d 77, 105 (S.D.N.Y. 1999) ;

Owens-Illinois, Inc. v. United Ins. Co. , supra, 138 N.J. at 472-73, 650 A.2d 974 ; Crossmann Communities of North Carolina, Inc. v. Harleysville Mutual Ins. Co. , 395 S.C. 40, 63, 717 S.E.2d 589 (2011). Those rationales are largely inapplicable, however, to situations in which a policyholder desires and attempts to obtain coverage but the insurance industry declines to supply it. See Keyspan Gas East Corp. v. Munich Reinsurance America, Inc. , 143 A.D.3d 86, 93, 37 N.Y.S.3d 85 (2016).

Mt. McKinley's principal argument against the unavailability rule is that each insurer contracts to pay only for those losses that occur during its policy period and that the rule improperly allocates to insurers costs attributable to losses arising during uninsured years, for which the insurers have received no premiums. In pressing this argument, Mt. McKinley relies heavily on our Supreme Court's statement in Security that "[n]either the insurers nor the insured could reasonably have expected that the insurers would be liable for losses occurring in periods outside of their respective policy coverage periods." Security Ins. Co. of Hartford v. Lumbermens Mutual Casualty Co. , supra, 264 Conn. at 710, 826 A.2d 107.

Our Supreme Court made this statement, however, in discussing policyholders who opt not to procure insurance for extended periods when insurance is readily available; see id., at 708-10, 826 A.2d 107 ; and there is no indication that the court intended the statement to apply beyond that context. If the losses at issue in a case such as this resulted from traditional accidents such as a fire or a motor vehicle crash, then it would of course be true that a policyholder could not expect an insurer who provided coverage in, say, 1984 to defend or indemnify losses arising from a 1986 accident. The flaw in Mt. McKinley's reasoning, however, is that it fails to recognize that progressive, long latency injuries such as asbestos related disease are fundamentally different from those sorts of traditional accidents. And because the standard form comprehensive general liability policies that the insurance industry issued prior to 1986 did not anticipate or account for those differences; see Owens-Illinois, Inc. v. United Ins. Co. , supra, 138 N.J. at 468-71, 650 A.2d 974 ; courts have been forced to develop a distinct set of rules to adjudicate long-tail claims.

One important distinction is that progressive injuries caused by asbestos are indivisible and cumulative. Ins. Co. of North America v. Forty-Eight Insulations, Inc. , 451 F.Supp. 1230, 1242 (E.D. Mich. 1978), aff'd, 633 F.2d 1212 (6th Cir. 1980). This means that it is impossible to identify what portion of a claimant's bodily injury actually occurred during which policy period. Many of our sister courts, applying an all sums theory, have concluded that the indivisible nature of progressive injuries means that any insurer on the risk for any period of time can be called upon, at the discretion of the policyholder, to cover the entire claim. Connecticut has adopted a more insurer friendly pro rata allocation system, one that operates on the legal fiction that asbestos related disease occurs in equal increments commencing at the time of initial exposure and culminating with the manifestation of disease. See J. Michaels et al., supra, 64 U. Kan. L. Rev. 473 (continuous trigger theory is legal fiction). But this legal convention does not mean that the policy terms are somehow violated or coverage impermissibly broadened if the allocation rules are structured to (1) encourage policyholders to obtain the broadest possible insurance pool to respond to long-tail claims but (2) not punish those policyholders who, through no fault of their own, are unable to maintain a continuous chain of coverage.

Mt. McKinley also notes that, although most of the jurisdictions that allocate long-tail insurance costs pro rata also have adopted an unavailability rule, several of our sister courts have rejected such a rule. In particular, Mt. McKinley directs our attention to Boston Gas Co. v. Century Indemnity Co. , 454 Mass. 337, 910 N.E.2d 290 (2009) (Boston Gas ), and Sybron Transition Corp. v. Security Ins. of Hartford , 258 F.3d 595 (7th Cir. 2001) (Sybron ), two of the leading decisions to have rejected an unavailability of insurance rule. Because those cases reject the rule for very different reasons, each warrants attention.³⁰

In Boston Gas , the Supreme Judicial Court of Massachusetts explained why it disagreed with those courts that have held that long-tail losses should not be allocated to a policyholder for periods during which insurance for the risk was not commercially available. Boston Gas Co. v. Century Indemnity Co. , supra, 454 Mass. at 371, 910 N.E.2d 290. The court reasoned that the unavailability exception "effectively provides insurance where insurers made the calculated decision not to assume [a] risk and not to accept premiums. In effect, because the policyholder could not buy insurance, it is treated as though it did by passing those uninsurable losses to insured periods. This would not be equitable to insurers if the insured purchased coverage for only a few years where there was protracted damage." (Internal quotation marks omitted.) Id., at 371-72, 910 N.E.2d 290. Massachusetts' highest court thus raised both a general concern that the unavailability rule fails to honor an insurer's decision not to provide coverage as well as a specific concern that the rule may lead to inequitable results in particular circumstances. We consider each concern in turn.

The court in Boston Gas , considering the issue from the standpoint of the insurer, concluded that it would be unfair to allocate to an insurer losses that occurred after the insurer determined that it was no longer cost effective to insure against that particular risk. Other courts, by contrast, have looked at the issue from the perspective of the policyholder, concluding that it would not be fair to hold the policyholder responsible for periods during which insurance could not be purchased. See, e.g., Security Ins. Co. of Hartford v. Lumbermens Mutual Casualty Co. , supra, 1999 WL 545745, *8 ; Owens-Illinois, Inc. v. United Ins. Co. , supra, 138 N.J. at 479, 650 A.2d 974. As we have explained, however, pro rata, continuous trigger allocation is an artificial judicial construct designed to allocate costs between the various insurance policies that are on the risk during the time over which a single, indivisible injury develops. To our minds, the question of how to allocate uninsurable portions of the allocation block is not so much one of fairness but, rather, of which party should bear the risk that the insurance pool will be terminated if substantial new long-tail risks are identified after significant liabilities already have accrued.

From an equitable standpoint, either party can justifiably be assigned responsibility for ongoing asbestos related injuries after 1985.³¹ The policyholder is the one who allegedly caused the injury and, therefore, who ultimately will be financially responsible should insurance prove insufficient. At the same time, each insurer agreed to write an occurrence based policy that affords "almost unlimited prospective coverage" for future costs arising from injuries that take place during the policy period. (Internal quotation marks omitted.) Security Ins. Co. of Hartford v. Lumbermens Mutual Casualty Co. , supra, 264 Conn. at 692 n.5, 826 A.2d 107. Insurers can and do cabin that liability by including relevant exclusions in the policy language and by capping in various ways the limits that they will pay under the policy. See Champion Dyeing & Finishing Co. v. Centennial Ins. Co. , supra, 355 N.J. Super. at 277, 810 A.2d 68. Accordingly, we do not think that it would be fundamentally unfair to hold either the insurers or the policyholder responsible for portions of the allocation block during which insurance is unavailable.

Perhaps a more fruitful way of approaching the issue is to recognize that it arises only because of a unique confluence of circumstances. We assume-merely for the sake of argument-that, at the time they enter into a contract for a comprehensive general liability policy, both the policyholder and the insurer reasonably believe that the policyholder can safely engage in its chosen line of business, without undue risk of liability. If an unforeseen catastrophe such as a major oil spill occurs in a particular year, it is understood what the respective responsibilities and liabilities of each party will be for the resulting injuries. Similarly, if a well understood long-tail accident such as a gradual chemical leak comes to light, and if the policyholder has been diligent about maintaining continuous insurance coverage, then some policy will be on the risk for each portion of the allocation block. And even if a new risk associated with the business comes to light-say one of the chemical compounds utilized by the policyholder is discovered to be unstable and highly explosive-each party is free to structure its affairs going forward so as to manage that risk; insurers can opt not to renew their policies upon termination and, if the policyholder is unable to obtain other insurance, it is free either to proceed on a self-insured basis or to cease using the chemical in question. None of these conventional coverage scenarios poses an unavailability of insurance problem.

Products such as asbestos and silica, however, present a different problem. For decades, insurers provided and policyholders obtained comprehensive general liability insurance under the mutual belief that those products could be produced and sold with relative safety.³² By the time they discovered otherwise, and insurers concluded that they could no longer cost-effectively insure such risks, it was too late. Because the injuries involved were (1) long latency diseases that took decades to manifest, (2) not fully understood at the time of contracting, and (3) already so numerous and substantial by 1985 as to render them uninsurable, the parties faced a unique circumstance that was not anticipated or addressed in the policy language. See J. Stempel, supra, 12 Conn. Ins. L.J. 352-54.

The question we must resolve, then, is who should bear the risk that a new type of long-tail injury will be discovered and that the insurance market will dry up midcourse, leaving a substantial backlog of dormant liability of which the policyholder and its prior insurers were both unaware at the time of contracting and for which a continuous chain of coverage will not be available. Courts and commentators have articulated four reasons why it is more efficient and reasonable for such risks to be borne by the insurers rather than the policyholder.

First, holding the insurers on the risk collectively responsible for the full injury, up to the policy limits for which the insured has contracted, has the desirable effect of maximizing the resources available to respond to the multitude of claims facing Vanderbilt and others similarly situated. See Owens-Illinois, Inc. v. United Ins. Co. , supra, 138 N.J. at 472-73, 650 A.2d 974. Applying an unavailability rule also may encourage insurers who already are on the risk for long-tail injuries to continue to accept premiums-which presumably will go up as the risk becomes more apparent-and to make insurance available for a longer period of time, thereby spreading the risk among additional policies and generating additional resources to compensate victims of injury. See id. By contrast, a contrary rule that transfers liabilities to the policyholder as soon as insurance becomes unavailable would incentivize the insurance industry to stop offering coverage prematurely when novel risks emerge. The ensuing race to the exit would further reduce the resources available to respond to emerging risks. Cf. note, supra, 99 Harv. L. Rev. 1580.

Second, holding insurers responsible when unforeseen risks arise, and not permitting them simply to drop coverage and cut their losses, creates an incentive for insurers as well as policyholders continuously to identify and investigate previously unknown risks associated with various materials and lines of business. Proponents of the unavailability rule have pointed to the public safety benefits that would result. See, e.g., M. Doherty, supra, 64 U. Chi. L. Rev. 266 and n.52, 285; J. Stempel, supra, 12 Conn. Ins. L.J. 468.

Third, an unavailability of insurance rule best comports with the reasonable expectations of the insured. Insurance involves the transfer of risk. Doucette v. Pomes , 247 Conn. 442, 459, 724 A.2d 481 (1999). Absent contractual language to the contrary, a policyholder will reasonably expect that a comprehensive general liability policy will provide full coverage, up to the policy limits, not only for known risks but also for newly recognized long latency injuries. See Champion Dyeing & Finishing Co. v. Centennial Ins. Co. , supra, 355 N.J. Super. at 277, 810 A.2d 68 ; J. Stempel, supra, 12 Conn. Ins. L.J. 471-72. If a policyholder has been diligent in its efforts to maintain a continuous stream of coverage, then it may reasonably expect that it will be able to avail itself fully of such coverage in the event that unforeseen and ongoing injuries arise. Relatedly, to the extent that the relevant standard form policies are silent on the question, this ambiguity must be construed against the insurers that drafted the policy and in favor of the policyholder. See Wentland v. American Equity Ins. Co. , 267 Conn. 592, 601, 840 A.2d 1158 (2004). This reflects the fact that insurers, who are in the business of managing risk, are better situated to anticipate such developments and, if they choose, draft policies that expressly address them. See J. Stempel, supra, 374-75, 382-83, 465.

Fourth, our sister courts have noted that insurers have a better ability to manage this sort of risk. They can continue to accept, pool, and spread the risk, pricing coverage accordingly. See Owens-Illinois, Inc. v. United Ins. Co. , supra, 138 N.J. at 472-73, 650 A.2d 974. Or they can eschew the risk at whatever point it becomes unattractive to insure, at which time the policyholder is left with no choice but to effectively self-insure going forward. See J. Stempel, supra, 12 Conn. Ins. L.J. 465-68; see also Stonewall Ins. Co. v. Asbestos Claims Management Corp. , supra, 73 F.3d at 1204 (noting that insured cannot bargain around asbestos exclusions). The unavailability rule recognizes that the policyholder, unlike the insurer, has little if any ability to manage its exposure to dormant liabilities with respect to an emerging long-tail risk. For all of these reasons, we disagree with Boston Gas that, as a general proposition, it is unreasonable for insurers to bear the risk that the insurance industry will at some point cease to offer coverage for a particular long-tail risk, leaving a coverage gap for injuries already in progress.

The court in Boston Gas also was concerned with what it identified as a more specific source of unfairness, namely, the situation in which only a few years of coverage are available to cover a substantial loss. See Boston Gas Co. v. Century Indemnity Co. , supra, 454 Mass. at 371, 910 N.E.2d 290. This might happen, for instance, if a hypothetical manufacturer entered the asbestos business in 1985 and purchased a comprehensive general liability policy from Insurer A for that year, after which time insurance became unavailable. Under those circumstances, Insurer A would have retained only one year worth of premiums, but it would be solely responsible for defending and indemnifying decades of ensuing injuries sustained by all of the claimants who were exposed to the manufacturer's asbestos during that year.

To reiterate what we already have said, the vast majority of our sister courts have concluded that, as a general matter, it is not unfair or unreasonable to allot to an insurer portions of an allocation block for which it did not write policies or receive premiums. If Insurer A insured a petroleum company for a single year during which that company was responsible for a catastrophic oil spill, Insurer A could be liable for years of ensuing cleanup costs, property damage, and other losses, despite having received only a single premium. See Champion Dyeing & Finishing Co. v. Centennial Ins. Co. , supra, 355 N.J. Super. at 276, 810 A.2d 68. It is not apparent, then, why the insurer should not also have to cover the full extent of injuries arising from asbestos exposures that occurred on its watch. Indeed, as we have discussed, many jurisdictions continue to adhere to the more policyholder-friendly all sums approach to long-tail allocation, according to which (1) the insured bears no responsibility for uninsured coverage blocks, and (2) any insurer on the risk can be called upon to bear the full costs of indemnifying and defending a long-tail tort claim, up to its policy limits. From this standpoint, it is difficult to see how the insurance industry, having received the benefits in Connecticut of a pro rata system, in which costs are uniformly allocated among policies and policyholders are responsible for uninsured periods, is treated unfairly in circumstances in which only one insurer happens to be on the risk. At worst, the liabilities of any particular insurer will be no greater than they would be in an all sums jurisdiction.

We recognize that one consequence of the unavailability rule is that, in particular cases, individual insurers may end up bearing what appears to be a disproportionate share of the financial burden. But virtually every method of allocating long-tail tort liabilities will at times result in apparent inequities of one sort or another. Jurisdictions that apply an initial exposure trigger place a greater burden on insurers who provided coverage in the early years, whereas manifestation jurisdictions call on later insurers to carry more of the load. Courts that allocate costs by policy limits as well as by time on the risk³³ disadvantage issuers of larger policies, whereas any insurer on the risk may be singled out under the all sums approach. See J. Michaels et al., supra, 64 U. Kan. L. Rev. 477 (opining that all sums approach is inequitable). Ultimately, we believe that the allocation system that we have adopted and that our Supreme Court has at least implicitly endorsed-pro rata time-on-the-risk, employing a continuous trigger and an unavailability rule-distributes the burdens equitably among all parties involved and maximizes the resources available to respond to claims while minimizing administrative hassles and transaction costs.

The second case on which Mt. McKinley relies is Sybron Transition Corp. v. Security Ins. of Hartford , supra, 258 F.3d at 595, in which the United States Court of Appeals for the Seventh Circuit articulated a very different critique of the unavailability rule. In Sybron , Judge Frank Easterbrook, writing for the court, rejected the basic premise-accepted by most other courts to have considered the question-that it is possible for insurance against a particular risk ever to be "unavailable." Id., at 599. Instead, he proceeded on the assumption that insurers will "cheerfully" underwrite any risk at the right price . Id. This is true, Judge Easterbrook maintained, even when, as was arguably the case with asbestos in the mid-1980s, the "risk [already] has come to pass and the obligation to pay is certain," so that it is no longer possible for the insurer to spread the risk among different policyholders. Id. Under those circumstances, he explained, a would-be policyholder still can participate in insurance claim administration pools, although likely for a premium in excess of the maximum outlay. Id. The fact that an insured might have to pay more than $2 million to obtain a $2 million "retroactive" asbestos policy, for example, reflects the fact that the underwriter, who actually would provide more of "a claims-administration service rather than risk-spreading ... must be reimbursed for both the expected payments to victims and the expected costs of evaluating and resolving claims ...." Id. Judge Easterbrook insisted that "[t]his kind of insurance could have been purchased for asbestos risks in the mid-1980s," and that a policyholder who opted not to join such a claims administration pool should, therefore, be treated as self-insured. Id.

There are two principal reasons-one factual and one legal-why we find the reasoning of Sybron to be unpersuasive in the context of this case. First, as a purely factual matter, there is nothing in the present record to support the assertion that, beginning in 1986, Vanderbilt could have joined the sort of insurance claim administration pool posited in Sybron . Quite the contrary; the trial court expressly found that insurance was not available to Vanderbilt or any similarly situated manufacturer. That finding was supported by substantial evidence, and Mt. McKinley does not challenge it on appeal.

Second, from a legal standpoint, even if such a pool had been available to Vanderbilt, it would not have constituted the type of insurance that is relevant for purposes of the unavailability rule. As we previously have explained; see parts III A and III A 3 of this opinion; and as the court in Sybron acknowledged; Sybron Transition Corp. v. Security Ins. of Hartford , supra, 258 F.3d at 601 ; we apply a continuous trigger, time-on-the-risk pro rata allocation system to long-tail asbestos claims in large part for epistemological reasons. Because asbestos related disease frequently follows repeated exposure to asbestos fibers, it often is impossible to know whether fibers inhaled at any particular time caused a particular disease, or what percentage of a claimant's injuries transpired in any given policy period. See id. When there is a continuous stream of coverage, then, the fairest and simplest approach is to divide the costs of defense and indemnity evenly among all of the policy periods during which injury may have occurred.

This allocation scheme may theoretically present a moral hazard problem, however, as it might have been subject to abuse by a policyholder who attempted to avoid the cost of continuous coverage. Rather than purchase an excess policy every year, for example, an asbestos producer might have opted to acquire excess coverage only once every ten years, on the assumption that any asbestos related disease will take at least ten years to manifest and, therefore, at least one excess insurance policy will be available to respond to any long-tail claim. The producer could have obtained what would amount to full coverage at one-tenth of the cost. Aside from the fact that this outcome arguably would be unfair to the once-a-decade insurer, such a strategy would defeat the benefits of a continuous trigger approach; the risk-spreading function of insurance would be diminished and the resources available to respond to potential claims would be significantly reduced. See Security Ins. Co. of Hartford v. Lumbermens Mutual Casualty Co. , supra, 264 Conn. at 711, 826 A.2d 107. For these reasons, a court might have rightly treated the policyholder as self-insured under those circumstances.

The situation is fundamentally different, however, when it becomes clear that virtually every company in a particular line of business faces a substantial backlog of dormant liability arising from its past activities, such that providing insurance to those companies will no longer perform a risk-spreading function and they are deemed to be uninsurable with respect to the risk at issue. Although Judge Easterbrook argues that they still can obtain insurance of a sort, he acknowledges; see Sybron Transition Corp. v. Security Ins. of Hartford , supra, 258 F.3d at 599 ; that the claims administration pools that he describes are a fundamentally different creature, insofar as they do not advance what our Supreme Court has identified as the fundamental purpose of insurance, namely, to protect against the risk of loss. See Rathbun v. Health Net of the Northeast , Inc. , 315 Conn. 674, 697, 110 A.3d 304 (2015) ; see also Doucette v. Pomes , supra, 247 Conn. at 459, 724 A.2d 481 ("transfer of risk ... is generally considered to be an essential element of an insurance relationship"); Brown v. State Farm Fire & Casualty Co. , 150 Conn.App. 405, 414, 90 A.3d 1054 ("[a] loss that has already occurred is not fortuitous-and is thus not insurable"), cert. denied, 315 Conn. 901, 104 A.3d 106 (2014). Even if such pools had been available to Vanderbilt after 1985, then, they do not represent the sort of insurance that is germane to the unavailability rule. See Goodyear Canada, Inc. v. American International Cos. , Docket No. CV-09-00377269, 2011 ONSC 5422, ¶¶ 57-65, 2011 WL 4348186 (Can. Ont. Sup. Ct. J. September 16, 2011).

We therefore conclude that the cases that have adopted an unavailability rule are better reasoned, represent the majority position, and more closely comport with our Supreme Court's analysis in Security . Accordingly, it was not improper for the trial court to exclude Vanderbilt from the allocation block for years in which asbestos related insurance was unavailable.

2

Equitable Exception and Testimony of Professor Priest

In the alternative, Mt. McKinley contends that, if Connecticut does apply an unavailability of insurance rule, we should recognize an "equitable exception" to that rule. Under the proposed equitable exception, costs would be prorated to the insured if the insured continues to place allegedly harmful products into the stream of commerce during a time when no coverage is available for losses attributed to those products. Under those circumstances, Mt. McKinley argues, the insured assumes any risk resulting from the ongoing production, sale, or installation of a product that is recognized to be so harmful as to be uninsurable. Although the question may be a close one, we conclude that the trial court properly declined Mt. McKinley's invitation to apply an equitable exception under the circumstances of the present case. We also conclude that the trial court did not abuse its discretion in precluding expert testimony on the issue.

The following additional procedural history is relevant to our analysis of these issues. During the Phase II trial, Mt. McKinley argued that, if the trial court applied an unavailability of insurance rule, the court also should apply an equitable exception and treat Vanderbilt as self-insured for the period from March 3, 1986, through 2008, during which Vanderbilt continued to sell talc while uninsured or underinsured. In light of the "considerable credible evidence" that Vanderbilt made a business decision to continue selling talc beyond 1986 even though it was aware of the potential legal exposure, the trial court opined that "Mt. McKinley's equitable argument possesses appeal." Nevertheless, the court declined to apply an equitable exception for three reasons. First, after reviewing the relevant case law, the court concluded that there was no authority for considering Vanderbilt's post-1985 conduct when determining whether costs should be prorated to it for that period. Second, the court reasoned that Vanderbilt's pre-1986 insurers were contractually bound to provide coverage for injuries arising from exposures prior to 1986, and that nothing in the policy language or the insurers' course of dealing with Vanderbilt indicated that that coverage would be nullified if Vanderbilt continued to sell talc after 1986. Third, the court found that, even if it were free as a matter of law to apply an equitable exception, the equities of the case did not support it. Specifically, the court found that (1) Vanderbilt consistently took the position both before and after 1985 that its talc did not contain asbestos; (2) there was no evidence in the record demonstrating that talc does contain asbestos; (3) the federal government in 1992 reinforced Vanderbilt's position in this regard by excluding talc from regulation under the Occupational Safety and Health Administration's new asbestos regulations; and (4) Vanderbilt made every effort to obtain insurance after 1986 by representing to insurers that its products did not contain asbestos. In light of these findings, the court concluded that Vanderbilt had a good faith basis on which to continue selling its talc from 1986 until 2008 and, therefore, there was no equitable justification for treating the company as self-insured and allocating it a pro rata share of the post-1985 costs arising from claims alleging a first exposure before 1986.

a

Whether Connecticut Recognizes an Equitable Exception

Whether to recognize an equitable exception to the unavailability rule is a question of first impression in Connecticut and one that, to our knowledge, only a few of our sister courts have confronted. Vanderbilt emphasizes that there is no authority, in Connecticut or elsewhere, supporting the adoption of an equitable exception. Mt. McKinley counters that this dearth of authority simply reflects the fact that other purveyors of asbestos and related carcinogens largely ceased production and distribution of those products by the mid-1980s and, therefore, our sister courts have not been confronted with the question of whether companies that intentionally continue to sell harmful products after the risks become apparent and insurance becomes unavailable should be able to benefit from the unavailability rule. Mt. McKinley further notes that, in one of the few cases in which a court has considered whether to adopt the equitable exception, Pneumo Abex Corp. v. Maryland Casualty Co. , Civil Action No. 82-2098 (JGP), 2001 WL 37111434 (D.D.C. October 9, 2001), the insured continued to sell asbestos related products for only two years after insurance became unavailable. Id., at *4. Notably, the court in that case, while declining to apply an equitable exception, expressly reserved the question of whether an equitable exception might be appropriate if a company continued to manufacture asbestos products for five or ten years after coverage ceased to be available. Id., at *10. Mt. McKinley emphasizes that Vanderbilt continued to sell its talc for more than two decades after insurance for asbestos related injuries became unavailable.

In the most recent case to address the issue, however, the Appellate Division of the New Jersey Superior Court declined to apply an equitable exception even though the policyholder had continued to manufacture materials containing asbestos for fourteen years after insurance coverage became unavailable.³⁴ See Continental Ins. Co. v. Honeywell International, Inc. , Docket Nos. A-1071-13T1, 2016 WL 3909530, *11-12 (N.J. Super. App. Div. July 20, 2016) (Honeywell ), cert. granted, Docket No. 078152, 2016 WL 7665452 (N.J. December 12, 2016). The court in Honeywell rejected as illogical the argument that the policyholder's continued sale of asbestos related products after 1987 might have increased its exposure to claims that triggered its pre-1987 policies and thereby imposed a greater financial burden on its insurers. Id., at *12. Like the policyholder in Honeywell , Vanderbilt has accepted that it bears full responsibility for all costs arising from claims alleging a date of first exposure after insurance became unavailable, but it contends that there is no reason why its post -1985 talc sales should increase its liability for claims arising from pre -1986 exposure to its products.

We agree with Vanderbilt that, in most instances, the fact that a policyholder continues to sell a potentially dangerous product after the dangers become known and after insurance ceases to be available has no bearing on the question whether that policyholder should be treated as self-insured with respect to injuries that arose from previous exposures, when insurance was in place. However, we are not as confident as the court in Honeywell that there is no possibility that continued sale of asbestos related products on an uninsured basis after 1985 had the potential to adversely impact Vanderbilt's pre-1986 insurers.³⁵ Two possible scenarios come to mind.

The first scenario would be if a new underlying complaint were filed in, say, 2017 that failed to allege a date of first exposure to Vanderbilt's talc. Given that asbestos related diseases typically manifest between twenty to forty years after first exposure, initial exposure in that scenario plausibly could have occurred either before or after March 3, 1986, thirty-one years prior to the date of this opinion. Although Vanderbilt contends that the vast majority of the underlying complaints to date have alleged a first exposure prior to 1986, it is reasonable to assume that that share will decline over time as individuals who were first exposed to asbestos related materials in the late 1980s and 1990s begin to exhibit symptoms of asbestosis and asbestos related cancers. In cases wherein the date of first exposure is not alleged and cannot readily be determined, it would be unfair always to assume that first exposure occurred before 1986; see part III C 2 of this opinion; and to hold the defendant insurers liable for all of the defense and indemnity costs, when it is quite possible that the claimants were not exposed until after 1985 and then only because Vanderbilt continued to sell asbestos related products after insurance became unavailable. Accordingly, on remand, the trial court, in devising a method for calculating a default date of first exposure for new and pending claims; see id.; is instructed to do so with a mind toward the concerns discussed herein.³⁶

The second scenario of concern would arise if a hypothetical complainant were initially exposed to asbestos related products in, say, late 1985, but continued to undergo significant and protracted asbestos exposure, solely via Vanderbilt's talc, after 1985. To the extent that the risk of developing asbestos related disease and the severity of that disease are proportional to the duration and extent of asbestos exposure, it is conceivable that, by continuing to sell talc after 1985, Vanderbilt increased the likelihood that the hypothetical complainant would develop such a disease. If so, then Mt. McKinley is correct that applying the unavailability rule without an equitable exception would afford Vanderbilt a windfall at its insurers' expense. The insurers might have to defend and indemnify claims for injuries that would not have occurred, or that would have been less severe, but for Vanderbilt's decision to continue to sell talc on an uninsured basis.

We do not foreclose the possibility that such an argument might justify the application of an equitable exception under appropriate circumstances. We agree with the conclusion of the trial court, however, that the record does not support its application in the present case. The court found that Vanderbilt had a long-standing and good faith belief that its talc did not contain asbestos and that the underlying actions were groundless. That belief appeared to be validated by federal regulators, who exempted talc from asbestos regulations; see " Occupational Exposure to Asbestos, Tremolite, Anthophyllite and Actinolite," 57 Fed. Reg. 24,310, Department of Labor, Occupational Safety and Health Administration (June 8, 1992); and by certain insurers, who agreed to insure Vanderbilt after 1993 on the basis of its representations that its talc did not contain asbestos. On the record before us, we have no basis to conclude that those representations were false or that Vanderbilt's continued sale of talc after 1985 did or will in fact increase the financial burdens on any of its pre-1986 insurers. Accordingly, we do not believe that applying the unavailability rule in the present case would afford Vanderbilt an underserved windfall. For the same reasons, one of the arguments in favor of the equitable exception-that the law should not reward policyholders for continuing to engage in risky conduct on an uninsured basis after the risks have become apparent-holds less sway in a case such as this in which the court found that Vanderbilt operated under a good faith belief that its talc was not unreasonably dangerous.³⁷ We therefore conclude that the trial court properly determined that equity does not require the carving out of an exception to the unavailability rule and the treatment of Vanderbilt as self-insured for allocation purposes from 1986 through 2008.

b

Priest Testimony

We next consider Mt. McKinley's claim that the trial court improperly excluded the expert testimony of George Priest, a professor of law and economics at Yale Law School. We conclude that the trial court did not abuse its discretion in excluding Priest's testimony.

The following additional procedural facts are relevant to our resolution of this issue. During the Phase I and Phase II trials, Mt. McKinley offered Priest's testimony to educate the court about insurance allocation and availability issues and, in particular, about the economic and insurance bases of pre-1986 comprehensive general liability policies and the methods of allocating defense and indemnity costs after 1985. According to Mt. McKinley's expert disclosure and proffers, Priest would have testified that there is no contractual basis for allocating to Vanderbilt's pre-1986 insurers any liability for post-1985 injuries. He further would have opined that applying an unavailability rule to the period after 1985 would overturn the regulatory and gatekeeping function that insurers perform with respect to risky behavior and, therefore, would encourage policyholders to engage in economically irresponsible conduct. In a nutshell, Priest's opinion would have been that, absent an equitable exception, applying the unavailability rule in the present case will send other producers of harmful products the unfortunate message that they can continue to sell such products with impunity, even after the dangers become known and even after insurers deem them too risky to insure.

Vanderbilt filed a motion in limine to preclude Priest's testimony, contending that the proffered testimony merely offered a legal analysis of the meaning of the relevant insurance policies and, therefore, was not a proper subject of expert testimony and would not be helpful to the trial court in resolving the issues before it. The trial court agreed and granted that motion.

As previously noted; see part III A 2 a of this opinion; we review for abuse of discretion the trial court's determination whether an expert's proffered testimony would impart knowledge that is not common to the average person and would be helpful to the trier of fact in considering the issues. See Weaver v. McKnight , supra, 313 Conn. at 405-406, 97 A.3d 920. To the extent that Priest intended to testify as to the meaning or proper interpretation of the standard form comprehensive general liability contract, his testimony was properly excluded. It is well established that interpretation of an insurance policy is a question of law within the exclusive province of the court. Auto Glass Express, Inc. v. Hanover Ins. Co. , 293 Conn. 218, 231, 975 A.2d 1266 (2009). Expert testimony is relevant to that endeavor only when it is necessary to shed light on contractual terms of art. See Pratt v. Dunlap , 85 Conn. 180, 186-87, 82 A. 195 (1912). There is no indication in the present case that terms of art were at issue with respect to the unavailability rule. Priest's testimony as to the meaning of Vanderbilt's insurance policies was thus inadmissible. See Lone Star Steakhouse & Saloon, Inc. v. Liberty Mutual Ins. Group , 343 F.Supp.2d 989, 1013 (D. Kan. 2004) (excluding Priest's testimony with respect to public policy goals that underlie comprehensive general liability policy because case did not present need to clarify or define any terms of art).

On the other hand, to the extent that Priest intended to testify as to the sorts of equitable and public policy considerations that a court might consider in fashioning rules for allocating long-tail toxic tort liabilities, we believe that the trial court reasonably exercised its discretion in excluding the testimony. The opinions that Priest would have offered-for example, that permitting a company to sell dangerous products with impunity will incentivize other companies to do the same-are matters of common sense that do not require the counsel of a law professor. Indeed, the fact that Mt. McKinley articulates throughout its briefs the very arguments that it indicates Priest would have offered supports the conclusion that the proffered testimony would have amounted merely to legal argument, for which expert testimony is unnecessary. See Bank of America Corp. v. SR International Business Ins. Co. , Docket No. 05 CVS 5564, 2007 WL 4480057, *11 (N.C. Super. December 19, 2007) (excluding Priest's testimony because determination of public policy is matter for courts and legislature). For these reasons, we conclude that the trial court did not abuse its discretion in excluding Priest's testimony.

3

Whether Trial Court Properly Applied Unavailability Rule

In the preceding parts of this opinion, we concluded that the trial court properly determined that (1) Connecticut applies an unavailability of insurance rule when allocating liabilities for defense and indemnity costs to the policyholder in the context of long-tail toxic tort claims that implicate multiple policy periods, and (2) it would not be appropriate to apply an equitable exception to that rule according to which Vanderbilt would be liable for a pro rata share of the costs for the period during which it continued to produce talc after coverage for asbestos related liabilities had become unavailable. We turn now to the question of whether the trial court properly applied the unavailability rule to the facts of the present case. Vanderbilt contends that the court improperly determined that (1) defense cost coverage was available to Vanderbilt from March 3, 1993 through April 24, 2007, (2) Vanderbilt failed to obtain sufficient available insurance during that time, and (3) Vanderbilt was thus responsible for a pro rata share of defense costs for that fourteen year period. We agree with Vanderbilt that it should not have been allocated any post-1985 defense costs for claims with a date of first exposure prior to March 3, 1986.

The following additional facts are relevant to our resolution of this claim. During the Phase I trial, the court held an evidentiary hearing to determine, among other things, whether defense cost coverage for asbestos related claims was available to Vanderbilt between 1948 and 2008 and, if so, whether Vanderbilt fully availed itself of such coverage. On the basis of that hearing, the court found as follows.

"Beginning in the 1970s and up through 1986, there was a dramatic increase nationwide in the number of claims for injuries related to exposure to asbestos. Up through 1985, insurance companies had regularly offered and provided occurrence based defense cost and indemnity coverage for such claims. However, by 1986, such policies became generally unavailable to companies in the mining and chemical industries, as the insurance industry had found the number of, and potential exposure for, [long-tail] loss claims to be financially unattractive to insure. The insurance industry responded to the increase in such claims by offering policies that contained exclusions for asbestos related claims. Moreover, it moved to offer to companies that carried a risk of exposure to asbestos, such as those in the mining and chemical industries, only claims-made policies rather than occurrence based policies. Occurrence based coverage remained available to other companies ... that did not actively engage in the production or distribution of asbestos-containing materials, but not to those that were directly or indirectly involved. Both [parties'] experts credibly testified that by 1986, insurance coverage was generally available to companies within those industries but only on a claims-made basis and with an asbestos exclusion....

"Since it started mining talc [in 1948, Vanderbilt] has purchased or attempted to purchase insurance to cover the defense and indemnity of asbestos related claims. [Vanderbilt] claims [to have] purchased occurrence based policies for asbestos related claims for the period 1948-1955³⁸ ... [and it obtained multiple primary] umbrella [and] excess policies from various defendants ... from 1956 through 1985....

"Beginning in 1986, because it could not find a carrier to offer it any occurrence based policies without an asbestos exclusion, [Vanderbilt] began purchasing claims-made policies (also with exclusions for asbestos related claims). This was true until 1993, when, despite the change in the industry, [Vanderbilt] was able to obtain a limited number of claims-made insurance policies in the period March, 1993 through April, 2003, that provided defense cost coverage for asbestos related claims. This coverage included policies purchased through Gerling America Insurance Co. (Gerling) covering the period March 3, 1993 to July 31, 1996, and Commerce and Industry Insurance Co. (Commerce) covering the period July 31, 1996 to April 3, 2003.... In 2003, [Vanderbilt] was able to purchase a policy through [American International] for the period April 24, 2003 to April 24, 2004, [that] provided coverage for asbestos and for which it was also able to later obtain an extended reporting endorsement effective to April 24, 2007.... However, the endorsement was offered due to the insurer electing to exclude asbestos coverage at the end of the initial term of the policy."³⁹ (Citations omitted; footnotes added.) In addition to the testimony of the parties' experts, the trial court credited the testimony of several fact witnesses as to the availability of insurance after 1985. The court credited the testimony of Mac Nadel, Vanderbilt's insurance broker, that one other insurer, ECS Reliance, offered a primary defense cost coverage policy to Vanderbilt for the year 1999-2000, the terms of which would have been consistent with the terms of the Commerce policy that Vanderbilt purchased that year. Notably, however, the court also credited testimony that Vanderbilt (1) "made strong attempts to obtain as much asbestos related coverage as possible between 1986 and 2007," and (2) never turned down any coverage that would have provided it with broader coverage than what it had been able to obtain previously. In addition, the court credited Nadel's testimony that only primary defense cost coverage for asbestos related third party liability was available to Vanderbilt after 1986 and that Vanderbilt was unable to obtain umbrella or excess policies for asbestos liability during that time.

On appeal, Vanderbilt does not challenge the trial court's factual findings. Instead, it argues that the court erred as a matter of law when it concluded, on the basis of those findings, that Vanderbilt could not take advantage of the unavailability rule and was responsible for a pro rata share of defense costs for the period from March 24, 1993 through April 24, 2007. Accordingly, we now turn our attention to the trial court's legal analysis. The court's analysis proceeded in three stages, the second and third of which Vanderbilt challenges on appeal.

In the first stage of its analysis, the trial court considered whether the unavailability rule applies whenever the relevant insurance is generally unavailable to companies in a particular industry or, rather, whether there must be evidence demonstrating that insurance was unavailable to the particular policyholder. Finding no express authority for the former standard either in Connecticut or among other jurisdictions that follow the unavailability rule, the trial court adopted the latter standard. Because Vanderbilt does not challenge this conclusion on appeal,⁴⁰ we assume without deciding that unavailability of insurance should be assessed relative to a particular policyholder, rather than as to the insurance market in general.

In the second stage of its analysis, the trial court considered whether, in light of its factual findings, defense cost coverage was available to Vanderbilt. Although it found that coverage for asbestos related defense costs was generally unavailable during the entire post-1985 period, the court nevertheless concluded that coverage was available to Vanderbilt between 1993 and 2007. The court's rationale for this conclusion was simply the fact that Vanderbilt had in fact been able to obtain limited primary claims-made coverage during that time.

In the third stage of its analysis, the trial court concluded that Vanderbilt failed to take advantage of all available defense cost coverage. The court's reasoning in this respect is not entirely clear. The court expressly found (1) that Vanderbilt made strong attempts to obtain as much asbestos related coverage as possible between 1986 and 2008, and (2) that Vanderbilt never turned down any coverage that would have provided it with broader coverage than what it obtained. Nevertheless, the court ultimately concluded that Vanderbilt "shall be considered self-insured for the period March, 1993 through April, 2007, in that it was knowingly underinsured for that period of time." Although the court did not articulate its rationale for concluding that Vanderbilt was knowingly underinsured, it may be inferred from the memorandum of decision that the court relied solely on the facts that (1) there was no evidence in the record that Vanderbilt accepted ECS Reliance's offer to furnish it with a policy for the year 1999-2000, and (2) from 1999 to 2003, Vanderbilt annually obtained approximately $100 million in claims-made coverage with asbestos exclusions but only obtained approximately 5 percent of that amount in policies that provided asbestos related defense cost coverage.

With respect to the second stage of the trial court's analysis, Vanderbilt contends that the court mixed apples and oranges in concluding that the availability of certain claims-made policies after 1993 was relevant to the allocation of liabilities arising from pre-1985 occurrence based policies. We agree. As we explained in part III A of this opinion, the challenges and complexities that surround the allocation of insurance liabilities with regard to long-tail toxic tort claims arise primarily from a distinct feature of occurrence based comprehensive general liability policies: although covered claims must result from an occurrence, such as exposure to the policyholder's asbestos related products, the event that triggers coverage under a particular policy term is not the occurrence itself but rather the resulting injury. Because progressive, long latency diseases such as asbestosis and mesothelioma continually reinjure the body for decades after initial exposure, and because those injuries are considered to be indivisible-their progression and magnitude during any particular policy period are impossible to quantify-we have adopted the continuous trigger theory and the rule that insurance liabilities are allocated pro rata on a time-on-the-risk basis. It is this judicial solution to the problems created by the intersection of long-tail disease claims with occurrence based insurance policy language that gives rise to the question of to what extent an insured should be liable for a pro rata share of the costs for periods during which no insurance is available.

By contrast, if all comprehensive general liability policies operated on a claims-made basis, there would be no need for such allocation rules. For example, if a hypothetical plaintiff brought a claim in 1997 alleging asbestos related disease, then the relevant policies in effect in 1997-and only those policies-would be on the risk, regardless of when exposure occurred and how the disease developed and manifested over time. In light of this distinction, courts and commentators have recognized that the fact that a policyholder may or may not have purchased claims-made asbestos coverage in 1997 has no bearing on whether the policyholder should be treated as self-insured with respect to long-tail illnesses that originated before 1986, when occurrence based policies were in place, so long as the underlying action is not brought in 1997. In Champion , for example, the court considered how to allocate liability for progressive injuries-in that case environmental damage-that commenced while coverage was available on an occurrence basis but continued into a period in which coverage was offered, if at all, only on a claims-made basis. Champion Dyeing & Finishing Co. v. Centennial Ins. Co. , supra, 355 N.J.Super. at 264, 810 A.2d 68. Because claims-made policies would only have been available to defend claims brought during their policy years, the court concluded that the policyholder's failure to obtain such coverage was not relevant to the unavailability determination and did not render the policyholder responsible for a pro rata share of the costs for the intervening years. See id., at 276-77, 810 A.2d 68 ; see also J. Hogg, " The Tale of a Tail," 24 Wm. Mitchell L. Rev. 515, 555-57 (1998) (explaining how applying unavailability rule across transition from occurrence to claims-made policies would be anomalous and would create " 'disappearing coverage' " problem pursuant to which diseases that take longer to manifest would be increasingly underfunded).

We agree and conclude that the trial court improperly considered the availability of claims-made coverage when allocating to Vanderbilt liability for the 1993-2007 period. Absent any evidence that occurrence based coverage was available during that period, defense costs for claims originating before 1986 should not have been prorated to the insured. But see part III D of this opinion, addressing allocation rules for claims brought while a claims-made policy is in effect.

With respect to the third stage of the court's analysis, Vanderbilt contends, and we agree, that, even if the availability of claims-made policies were relevant to the pro rata allocation methodology, the trial court's implicit findings that Vanderbilt (1) declined to purchase a claims-made policy from ECS Reliance in 1999 and (2) obtained a relatively small amount of asbestos related defense cost coverage from 1999 to 2003 do not support the court's conclusion that Vanderbilt knowingly chose to underinsure during those four years, let alone during the entire period from 1993 through 2007. Notably, the court found that excess and umbrella coverage were unavailable for asbestos related claims after 1985, and also that Vanderbilt sought to obtain as much primary asbestos related coverage as possible and never turned down any policy that would have provided it with broader coverage than what it already obtained. In light of those findings, we fail to understand the relevance of the fact that Vanderbilt was able to obtain only approximately $5 million in primary asbestos related coverage and that it did not purchase a policy in 1999 that only would have replicated its existing coverage. Absent any specific findings by the trial court that Vanderbilt could have obtained broader coverage between 1993 and 2007 but chose not to, it should not have treated Vanderbilt as underinsured and assigned it a pro rata share of defense costs for those years.

4

Nonasbestos Particulates

In this opinion, we have adopted an unavailability of insurance rule with respect to the pro rata allocation of comprehensive general liability obligations for long-tail asbestos claims. As an ancillary matter, Mt. McKinley contends that resort to that rule in the present case should be foreclosed when an underlying action includes allegations of injury from exposure to nonasbestos particulates such as talc and silica because, it argues, occurrence based coverage for such injury was available to, but not procured by, Vanderbilt from 1986 to 2004. In both its Phase I and Phase II decisions, however, the trial court found that no occurrence based coverage was available to Vanderbilt during that period. To prevail under its interpretation of the unavailability of insurance rule, Mt. McKinley must therefore establish that the court's factual finding was clearly erroneous. See part II C of this opinion.

The following undisputed facts and procedural history govern our resolution of this claim. The record before us contains a sampling of underlying actions. The court admitted into evidence 134 complaints that set forth various allegations against diverse defendants in addition to Vanderbilt. While some contain allegations of injury solely from exposure to asbestos; see, e.g., Complaint of Robert Jeter, Jr.; many allege injuries from both asbestos and nonasbestos particulates. The complaint filed in California in 2010 by Maria Olivares, Diane Cano-Casas, and Gerardo Olivares (Olivares complaint) is illustrative. Paragraph 16 of that complaint alleges in relevant part that a Vanderbilt product known as Nytal "contained [t]alc, silica, and asbestos." The Olivares complaint further alleges that those materials "entered the decedent's body through absorption and inhalation, and were a substantial factor in causing [the] decedent's lung cancer...."

At least one underlying action involves a complaint that lacks any express allegations of an asbestos injury. Nowhere in the complaint filed in Texas in 2003 by Jack W. Byrom (Byrom complaint) does the word "asbestos"

appear. Rather, that complaint alleges in relevant part that the plaintiff "while working as a kiln operator ... was exposed to harmful airborne concentrations of silica, silica-containing products, and other toxic and/or carcinogenic chemicals, which caused his injuries and disease.... [T]he products mined, manufactured or distributed by the defendants when used in the intended manner emit harmful airborne concentrations of silica dust and other toxic and/or carcinogenic chemicals into the air which were inhaled by [the] [p]laintiff causing damage and harm to his body.... [A]t all relevant times ... the defendants were aware of the fact that [the] [p]laintiff and others similarly situated were regularly being exposed to the hazards of silicosis and other lung diseases." The underlying actions thus include allegations that the complainants (1) were exposed to nonasbestos particulates as well as asbestos, and (2) developed not only asbestos related diseases such as asbestosis and mesothelioma, but also diseases such as silicosis that are unique to nonasbestos particulates.⁴¹

Mindful that an insurer's duty to defend is triggered if even one allegation of a complaint potentially falls within the policy's coverage; Capstone Building Corp. v. American Motorists Ins. Co. , 308 Conn. 760, 805, 67 A.3d 961 (2013) ; Mt. McKinley posits that the alleged availability of occurrence based coverage for nonasbestos particulate claims precludes the operation of the unavailability of insurance rule when the underlying action includes allegations of injury from exposure to nonasbestos particulates. That is to say, if a comprehensive general liability policy had been issued to Vanderbilt at some point after 1986 that provided occurrence based coverage for nonasbestos particulate injuries, the insurer would be obligated to defend with respect to injuries such as those alleged in the Olivares and Byrom complaints, irrespective of whether the complaints also alleged an injury due to asbestos exposure. See Schurgast v. Schumann , 156 Conn. 471, 490, 242 A.2d 695 (1968) ("[when] a complaint in an action against one to whom a policy of liability insurance has been issued states different causes of action or theories of recovery against the insured, and one such cause is within the coverage of the policy but others may not be within such coverage, the insurer is bound to defend with respect to those which, if proved, are within the coverage" [internal quotation marks omitted] ); 14 L. Russ & T. Segalla, Couch on Insurance (3d Ed. 2007) § 200:11, p. 200-20 ("[a]n insurer's duty to defend is expansive and arises when any part of the claim is potentially or arguably within the scope of the policy's coverage, even if the allegations of the suit are false, fraudulent, or groundless" [footnote omitted] ). Because, under the pro rata allocation system that we have adopted, a policyholder that declines to obtain available coverage is required to assume the obligations of the insurer for periods during which it opts to self-insure, Mt. McKinley contends that Vanderbilt is likewise required to shoulder its share of defense costs for underlying actions alleging post-1985 nonasbestos particulate claims.

Mt. McKinley's claim hinges principally on the testimony of its expert witness, Donald W. Bendure, an insurance underwriter. Bendure testified that occurrence based liability coverage for claims alleging bodily injury from exposure to silica, talc, and other nonasbestos particulates was "generally available" in the insurance industry "up until 2005 ...."⁴²

Although the trial court credited that testimony, that determination is not dispositive of the issue before us. In its Phase I and Phase II decisions, the court repeatedly emphasized that the critical issue for purposes of applying the unavailability of insurance rule was not whether occurrence based coverage was generally available, but rather whether such coverage was, in fact, available to Vanderbilt. Mt. McKinley has not challenged the propriety of that determination on appeal.

Fatal to the claim advanced by Mt. McKinley is the trial court's finding that Vanderbilt was unable to procure any occurrence based coverage following the sea change in insurance industry practices in 1986.⁴³ As the court noted, that change was precipitated by "the flood of asbestos related injury claims" throughout the industry. By 1986, the court continued, "the insurance industry had found the number of, and potential exposure for, long-tail latency loss claims to be financially unattractive to insure" with respect to "companies in the mining and chemical industries .... [The insurance industry] moved to offer ... only claims-made policies rather than occurrence based policies" "to companies that carried a risk of exposure to asbestos, such as those in the mining and chemical industries ...."⁴⁴ Emphasis added.) It is undisputed that Vanderbilt was such a company.⁴⁵ The court thus found that occurrence based coverage was unavailable to Vanderbilt during the post-1985 period.

That finding is substantiated by evidence in the record before us. The court heard the testimony of Joseph Denaro, a Vanderbilt employee since 1973, who served as the company's vice president, chief financial officer, and treasurer. Denaro testified that he was involved in the procurement of insurance for Vanderbilt. When asked whether, "in connection with [his] efforts to secure insurance," he was "given the opportunity after 1986 to buy occurrence based insurance," Denaro replied, "that was never offered as an opportunity.... That was unavailable." Denaro explained that, had it been available, Vanderbilt would have been "[v]ery much" interested in obtaining occurrence based coverage because it offers "greater coverage, especially for the kinds of claims we may have encountered, which are long-tail type where you don't know an injury's occurred until many years later."

The court also was presented with testimony from Nadel, a chartered property casualty underwriter at the commercial risk management insurance brokerage firm of Marsh & McLennan Companies. In 1997, Nadel began working on behalf of Vanderbilt to obtain insurance. Nadel testified that he could not secure occurrence based coverage for Vanderbilt's products because it "was not available ...." Nadel further confirmed that at no time thereafter was such coverage "available in the marketplace" for Vanderbilt. Nadel explained that "[b]ased upon the class of business that [Vanderbilt] was engaged in," only "a claims made policy ... had been [available to Vanderbilt] since 1986 ...."

Jeffrey Posner, an expert witness with more than thirty-five years of experience in the field of insurance and risk management, further corroborated that testimony. Posner testified that "occurrence based insurance was available in the marketplace after 1986 for certain clients and not available for others. Based upon the fact that Vanderbilt ... had long-tail claims, I would suspect it was probably unavailable to them, although I can't state that definitively because I wasn't there.... I think it was probably unavailable based upon my knowledge of the industry and what I've seen in this case." The trial court expressly credited the testimony of Denaro, Nadel, and Posner, as was its exclusive prerogative. See United Technologies Corp. v. East Windsor , 262 Conn. 11, 26, 807 A.2d 955 (2002) (in case tried before court, trial judge is sole arbiter of credibility of witnesses and weight to be afforded to specific testimony).

The court likewise was free to reject Bendure's opinion to the contrary. At trial, Bendure opined that occurrence based coverage "would have been generally available to [Vanderbilt] .... There is not anything remarkable about [Vanderbilt] that would have precluded them from being able to obtain occurrence based products liability" coverage after 1985. In his testimony, Bendure confirmed that an underwriter considers many factors prior to making a judgment whether to issue a policy, including the manufacturer's class of business, loss history, and financial integrity. In addition, an underwriter would consider whether asbestos related claims previously were brought against the manufacturer. On cross-examination, Bendure acknowledged that thousands of asbestos related claims had been filed against Vanderbilt over the years. Although he steadfastly maintained that whether a policy ultimately issues "would depend on the underwriter and the type of coverage provided," he acknowledged that an "average underwriter" would not provide coverage to a manufacturer in Vanderbilt's predicament. In addition, despite his years of experience as an underwriter, Bendure conceded that he had no experience dealing with insurance policies that involved talc. Moreover, he admitted that he had no information as to whether Vanderbilt in fact had been offered any occurrence based coverage after 1985. In light of the foregoing, the trial court reasonably could discount Bendure's testimony that such coverage was available to Vanderbilt. See Kervick v. Silver Hill Hospital , 309 Conn. 688, 717-18, 72 A.3d 1044 (2013) (trial court free to accept or reject testimony in whole or in part).

Last, we note that no party offered any testimonial or documentary evidence from an insurer indicating that it would have offered occurrence based coverage to Vanderbilt during the years in question. Although dozens of insurance companies from around the world are involved in this litigation, none submitted such evidence at trial.

The evidence credited by the trial court supports its finding that occurrence based coverage was unavailable to Vanderbilt from 1986 until it ceased production of industrial talc in 2008. That finding, therefore, is not clearly erroneous. Mt. McKinley thus cannot demonstrate that a departure from the unavailability of insurance rule is warranted in this case.⁴⁶

C

Default Date of First Exposure and Pre-1962 Liabilities

In part III A of this opinion, we determined that third party liability claims alleging a long-tail toxic tort such as asbestos related disease trigger defense and liability coverage running continuously from the date of first exposure to the alleged toxin until the manifestation of disease. In this part, we address two related questions raised by Mt. McKinley. First, did the trial court properly enforce a 2002 settlement agreement between Vanderbilt's primary insurers pursuant to which pre-1962 liabilities were allocated to the post-1962 period? Second, what date of first exposure applies when an underlying complaint does not allege a date and when it would not be possible or practical to establish the date through discovery? We answer the former question in the affirmative. With respect to the latter question, we affirm in part and reverse in part the ruling of the trial court that a default date of first exposure of January 1, 1962, shall apply to all asbestos related claims against Vanderbilt. Although we agree that it was reasonable to adopt that default date with regard to underlying claims that already have been paid by Vanderbilt's primary insurers, we conclude that it is not reasonable to do so indefinitely for future claims. Accordingly, we remand the case to the trial court with direction to adopt a more reasonable default rule for pending and future claims.

The following additional facts-as found by the trial court unless otherwise noted-and procedural history are relevant to these issues. It is undisputed that Continental and Hartford, respectively, provided occurrence based primary coverage to Vanderbilt from January 1, 1962 until March 3, 1977, and from March 3, 1977 until March 3, 1986. Although Vanderbilt consistently has contended that Continental began providing primary coverage on January 1, 1956, Vanderbilt has been unable to locate copies of any Continental policies covering the 1956 through 1961 period and, prior to the commencement of the present action, the existence and terms of those policies were unresolved and disputed factual issues. During the first two phases of the trial, however, Continental and Vanderbilt entered into two stipulations pursuant to which Continental acknowledged that it had issued primary policies to Vanderbilt covering the period from January 1, 1956 to January 1, 1962.

For many decades, Hartford and Continental independently defended and indemnified Vanderbilt in various underlying actions alleging bodily injury resulting from exposure to Vanderbilt's talc. In 1995, Continental filed an action against Hartford seeking contribution for costs that it had incurred defending and indemnifying Vanderbilt for asbestos related injury claims. Hartford and Continental settled their dispute in 2002, executing an agreement in which they reallocated between themselves various defense and indemnity payments that they had made to Vanderbilt as of that date. For underlying complaints that had alleged a date of first exposure to Vanderbilt's talc between January 1, 1962 and March 3, 1986, the allocation agreement provided that Hartford and Continental would allocate defense and indemnity obligations on a pro rata, time-on-the-risk basis running from the alleged date of first exposure until March 3, 1986. The primary insurers also agreed that, for underlying claims wherein the alleged date of first exposure to Vanderbilt's product was before January 1, 1962, or was unknown, the pro rata allocation coverage block for their purposes would run from January 1, 1962 to March 3, 1986. In other words, under the agreement, Continental did not assume responsibility for defense or indemnity costs arising between 1948 and 1961 (because coverage could not be established), nor did the primary insurers assign any liability to Vanderbilt for the pre-1962 period (because they were unsure whether proration to the insured was permitted under Connecticut law).

In addition to reallocating payments that Hartford and Continental had made on behalf of Vanderbilt, the 2002 allocation agreement also provided that the primary insurers would negotiate in good faith to establish an allocation arrangement for future defense and indemnity costs. Although the trial court made no express findings with respect to this allocation arrangement, there was undisputed testimony that, pursuant to that provision, the primary insurers did in fact reach an informal understanding according to which they continued to allocate payments on pending and future claims according to the 2002 methodology, that is, on the basis of a 1962-1986 allocation block.

During the Phase II trial, the question of whether the primary insurers had properly allocated responsibility for Vanderbilt's pre-1962 asbestos liabilities arose with respect to (1) underlying actions for which the primary insurers already have provided coverage, and (2) pending and future underlying actions for which insurance funds are yet to be expended. As to the former, Continental and Hartford contended that their respective primary insurance policies running from 1968 to 1977 and from 1977 to 1986 had been exhausted as a result of past payments, and they asked the trial court to find accordingly. Vanderbilt's excess and umbrella insurers, whose policies are not triggered until the primary policies are exhausted, objected. They argued that, under the continuous trigger, pro rata allocation system, the primary insurers should have applied a default date of first exposure of 1948, when Vanderbilt began producing talc. They further argued that Hartford and Continental wrongfully failed to allocate a pro rata share of indemnity costs to Vanderbilt for the 1948 through 1956 period and to Continental for the 1956 through 1961 period. By artificially compressing the potential allocation block into a twenty-four year period running from 1962 to 1986, the secondary insurers argued, the primary insurers prematurely exhausted certain policies in effect during that time and thus prematurely brought the secondary insurers on the risk.⁴⁷

The trial court rejected the argument of the secondary insurers. The court concluded that the primary insurers had acted reasonably in adopting a default date of first exposure of January 1, 1962, and it declined to reallocate any payments already allocated under that agreement. On the basis of that decision, the court also found that all of Vanderbilt's primary policies in effect between January 1, 1968 and March 3, 1986, had been exhausted.

As to pending and future actions for which insurance funds have yet to be expended, the parties asked that the trial court provide guidance as to the appropriate allocation methodology. The court did not expressly address this issue in its Phase II decision. However, the court's ultimate conclusion as to the default date of first exposure is worded broadly enough that it appears to encompass pending and future underlying actions as well as actions already resolved by Vanderbilt's primary insurers: "As to any claims [that] have an unknown date of first exposure, the default date of first exposure shall be January 1, 1962, consistent with the allocation agreement entered into by and between Hartford and [Continental], which agreement the court has found to be reasonable."⁴⁸ The secondary insurers challenge both determinations on appeal. We consider each issue in turn. 1

Allocation of Past Payments under 2002 Allocation Agreement

We first consider whether the trial court properly enforced the 2002 allocation agreement pursuant to which Hartford and Continental allocated all of their payments on Vanderbilt's asbestos related liabilities over the 1962 to 1986 period. Mt. McKinley contends that, with respect to underlying actions that alleged a date of first exposure prior to 1962, it was unreasonable and unfair for the primary insurers to allocate pre-1962 liabilities to the 1962 to 1986 coverage block, resulting in the premature exhaustion of primary policies in effect during that period. We are not persuaded.

The following additional facts and procedural history are relevant to our resolution of this issue. At trial, representatives of the primary insurers testified regarding the genesis of and rationales for the 2002 allocation agreement. They testified that, at the time that Hartford and Continental entered into the settlement agreement, they did not have complete copies of any Continental policies issued to Vanderbilt between 1956 and 1961, and they were unable to confirm the existence or terms of any such policies. In addition, as of 2002, our Supreme Court had not yet published its Security decision, and the primary insurers were then of the opinion that it would be improper for them to allocate to Vanderbilt a pro-rata share of indemnity costs for the pre-1962 period. For those reasons, they proceeded on the assumption that Vanderbilt did not have any comprehensive general liability insurance prior to 1962 and that the coverage block for long-tail claims against Vanderbilt could not begin until that time.

The trial court also heard testimony from several expert witnesses as to the reasonableness of this agreement. The experts opined that, although the allocation agreement did not strictly adhere to the pro rata allocation rules that our Supreme Court later established in Security , the agreement was fair and reasonable under the circumstances, given what Hartford and Continental knew and believed to be true at the time. The experts reached that conclusion notwithstanding their awareness that Continental had written several letters during the 1980s and 1990s in which it acknowledged that it had insured Vanderbilt from 1956 through 1977. This expert testimony was unchallenged.

Consistent with those expert opinions, the trial court concluded that the allocation scheme adopted under the 2002 settlement agreement was objectively reasonable at the time of adoption. Accordingly, the court enforced the agreement and the consequent allocations, despite the fact that the allocation scheme did not fully comport with the pro-rata allocation methodology that the court established.

The court offered several justifications for its decision to uphold and enforce the settlement agreement. First, the court credited the testimony of Continental's witnesses that, as of 2002, they were unable to verify the terms of any pre-1962 policies issued to Vanderbilt.

The court expressly found that there was no evidence that Continental took any action to conceal the existence of those policies, and that both parties entered into the agreement in good faith. The court also found that, as of the time of trial, the parties still had been unable to locate copies of any pre-1962 policies, and that there continued to be a dispute over the terms of those policies. Moreover, the court credited the expert testimony of two insurance risk management consultants, Donald Huffer and Jeffrey Posner, both of whom opined that the allocation method that the primary insurers adopted in 2002 was reasonable and consistent with industry standards.

Second, the trial court recognized that "it is the sound public policy of Connecticut to encourage parties to settle their disputes and to avoid protracted litigation." Monti v. Wenkert , 287 Conn. 101, 125, 947 A.2d 261 (2008). Relying on Metropolitan Life Ins. Co. v. Aetna Casualty & Surety Co. , 249 Conn. 36, 55 n.21, 730 A.2d 51 (1999), the court assumed that the reasonableness of an insurance settlement is to be assessed on the basis of the objective facts known to the parties at the time. The court therefore concluded that the 2002 allocation agreement could not have been rendered unreasonable by subsequent developments, including the release of the Security decision in 2003 and the parties' stipulations at trial regarding the Continental policies issued between 1956 and 1961.

Third, the court concluded that it would be unreasonable and impracticable to compel Hartford and Continental to retroactively reallocate the millions of dollars that the primary insurers already had paid to resolve thousands of underlying actions over the course of several decades. The court explained that "[t]o effectuate a reallocation to include, for example, the period 1956-1962 would require the court and the parties to engage in mathematical calculations that would not only be extremely arduous, time-consuming, and to some degree subjective, they would also be of marginal utility and ultimately undermine the prior significant efforts of the parties to compromise their differences over the allocation of the payments between themselves and which were made for the benefit of their insured." "Such an undertaking," the court concluded, "would lie between arduous and Sisyphean."

On appeal, Mt. McKinley focuses on several purported defects in the trial court's reasoning. Mt. McKinley argues that the record does not support the conclusion of the trial court that it was reasonable for the primary insurers to agree in 2002 to allocate all of their indemnity payments to the post-1961 period. Specifically, Mt. McKinley points to several documents in the record that indicate that, during the 1980s and 1990s, Continental repeatedly confirmed the existence and limits of its 1956-to-1961 policies. Relatedly, Mt. McKinley argues that neither the trial court nor any of the experts whose testimony the court credited were able to explain how it could be reasonable for the primary insurers to enter into an agreement by which they effectively transferred their own obligations to the secondary insurers. Because those expert opinions were wholly conclusory, Mt. McKinley contends, it was improper for the trial court to rely on them. Finally, Mt. McKinley emphasizes that Hartford and Continental agreed only among themselves to allocate liabilities exclusively to the post-1962 period, and that none of the secondary insurers were party to that agreement.

a

Standard of Review

Whether the trial court properly determined that the scheme established by the 2002 allocation agreement was reasonable and enforceable presents a mixed question of fact and law. To the extent that the court's determination hinged on its finding credible the expert testimony of Huffer and Posner regarding insurance industry standards, that determination will not be disturbed unless clearly erroneous. See Wyszomierski v. Siracusa , 290 Conn. 225, 243-44, 963 A.2d 943 (2009). Whether the court applied the correct legal standard with regard to the enforceability of settlement agreements, however, is a question of law subject to plenary review. Costantino v. Skolnick , 294 Conn. 719, 730, 988 A.2d 257 (2010). b

Whether Court's Reliance on Expert Testimony Was Clearly Erroneous

With respect to the expert testimony, Mt. McKinley relies on the principle that "[n]o weight may be accorded to an expert opinion [that] is totally conclusory in nature and [that] is unsupported by any discernible, factually based chain of underlying reasoning ...." (Internal quotation marks omitted.) Commissioner of Transportation v. Larobina , 92 Conn.App. 15, 26, 882 A.2d 1265, cert. denied, 276 Conn. 931, 889 A.2d 816 (2005) ; see also State v. Asherman , 193 Conn. 695, 716, 478 A.2d 227 (1984) ("[i]n order to render an expert opinion ... there must be a factual basis for the opinion"), cert. denied, 470 U.S. 1050, 105 S.Ct. 1749, 84 L.Ed.2d 814 (1985). Mt. McKinley contends that neither Huffer nor Posner offered any support for their opinions that the allocation agreement was reasonable. Mt. McKinley further argues that the experts failed to explain how it could have been reasonable for Continental and Hartford to proceed on the assumption that there was no coverage before 1962 when Continental had acknowledged in several letters, written prior to the allocation agreement, that it provided coverage to Vanderbilt beginning in 1956.

We disagree that the expert testimony on this subject was so conclusory as to render the trial court's reliance on it clearly erroneous. At least one of the experts, Donald Huffer, was aware of the various letters written during the 1980s and 1990s in which representatives of Continental had acknowledged that Continental provided liability insurance to Vanderbilt beginning in 1956. Nevertheless, it was Huffer's opinion that, according to the customs, practices, and standards of the insurance industry, it was not unreasonable for the primary insurers to agree in 2002 to use an allocation block commencing in 1962. Huffer based his opinion on the principle that, according to industry standards, Hartford and Continental were obligated to allocate payments over confirmed coverage periods only and that, without having in hand an actual policy by which the precise terms of coverage can be ascertained, a mere acknowledgement letter stating in general terms that an insurer provided coverage during a particular year does not suffice to confirm coverage for the types of claims at issue. Although Huffer's testimony in this respect was somewhat tentative and not as specific as it might have been, he did proffer a factually based chain of reasoning for his opinion. Accordingly, we cannot say that Huffer's testimony was so conclusory that the court's reliance thereon was clearly erroneous.⁴⁹

c

Whether Allocation Agreement Was Enforceable Against Secondary Insurers

We next consider whether, as a matter of law, the trial court properly concluded that it was not improper for the primary insurers to allocate payments in a manner that may have prematurely triggered Vanderbilt's excess and umbrella policies, notwithstanding that the secondary insurers were not party to the allocation agreement. The following principles govern our resolution of this claim.

As a general rule, a secondary insurer is not bound by the coverage decisions of a primary insurer; it remains free to contest coverage under its own policy. See, e.g., Allmerica Financial Corp. v. Certain Underwriters at Lloyd's, London , 449 Mass. 621, 633, 871 N.E.2d 418 (2007) (explaining that "[a]n excess carrier's intent to incorporate the same words used in a separate agreement between the primary insurer and the insured does not imply an intent by the excess carrier to accept decisions made by the primary carrier about the extent of its obligations under its own agreement"). At the same time, however, "an excess liability insurer cannot avoid or reduce liability under its own policy by challenging a separate insurer's decision to settle or pay out claims at a prior layer of insurance. In the absence of a specific contractual arrangement, the primary insurer owes no duty to the excess insurer with respect to claims within the parameters of [the] excess carrier's coverage.... An excess liability insurer, particularly a follow-form insurer, assumes the risk that the primary insurer will adjust claims in a certain manner and in such a way that triggers the potential for liability ... under the excess policy." (Citation omitted; internal quotation marks omitted.) Edward E. Gillen Co. v. Ins. Co. of Pennsylvania , Docket No. 10-C-564, 2011 WL 1694431, *4 (E.D. Wis. May 3, 2011) ; see also Allmerica Financial Corp. v. Certain Underwriters at Lloyd's, London , supra, at 634, 871 N.E.2d 418 (recognizing right of insurer to make coverage and settlement decisions independent of third parties, including other insurers). Nevertheless, when a primary insurer negotiates settlement of a claim in an amount that invades excess coverage, it "owes a duty of good faith to the excess carrier .... It must consider in good faith the interests of the excess insurer ... and must conduct the defense of the litigation, including settlement negotiations, so as not to expose the excess insurer to unwarranted liabilities." (Internal quotation marks omitted.) RLI Ins. Co. v. Superior Court of Ventura County , Docket No. B173316, 2004 WL 1171649, *5 (Cal. App. May 26, 2004) ; cf. Infinity Ins. Co. v. Worcester Ins. Co. , Superior Court, judicial district of Hartford, Docket No. CV-00-0597436 (December 4, 2000) (28 Conn. L. Rptr. 478, 2000 WL 1890126) (suggesting that excess insurer may maintain action for equitable subrogation against primary insurer stemming from primary insurer's settlement with insured).

We find the reasoning of the District Court in E.R. Squibb & Sons, Inc. v. Accident & Casualty Ins. Co. , 860 F.Supp. 124 (S.D.N.Y. 1994), to be persuasive on this point. The plaintiff in that case brought an action against its excess carriers for coverage in connection with product liability claims arising out of injuries incurred by users of the product diethylstilbestrol. The plaintiff and its primary carriers entered into a number of settlement agreements, which resulted in the exhaustion of the primary policies. The excess insurers challenged these agreements, but the court enforced them, holding that "exhaustion [of primary insurance] may occur by payment or settlement, provided the settlement is noncollusive and at arm's length." Id., at 126. The court offered the following illustration to clarify its holding: "[A]ssume that insurer A agrees with [a plaintiff] in a settlement agreement that in return for various immediate payments or other consideration, only manifestation or diagnosis will trigger that insurer's coverage. No other insurer, excess or otherwise, can avoid coverage because of this restriction on A's liability in the settlement between [the plaintiff ] and insurer A ." (Emphasis added.) Id.

Consistent with these principles, our sister courts have refused to enforce agreements among primary insurers or between primary insurers and the insured when it is apparent that the intent of the agreement is to transfer to excess insurers costs that clearly should have been borne by the primary coverage. For example, in Rees v. Viking Ins. Co. , 77 Wash.App. 716, 718-19, 892 P.2d 1128 (1995), the plaintiffs and the primary insurer agreed to a settlement according to which liability was stipulated to be $600,000 but the primary insurer was required to pay only $421,250 of a $500,000 policy limit, leaving the excess insurer on the hook for the balance. Concluding that the agreement was "little more than an artifice designed to support a claim against [the] excess policy," the Court of Appeals of the state of Washington affirmed the judgment of the trial court dismissing the plaintiffs' claim against the excess insurer. Id., at 719-20, 892 P.2d 1128 ; see also New York Dock Co. v. Ernest W. Brown, Inc. , 272 N.Y. 176, 180, 5 N.E.2d 190 (1936). Noncollusive settlements, by contrast, have been enforced as promoting the public policy in favor of settlement. See, e.g., United States Fidelity & Guaranty Co. v. Treadwell Corp. , supra, 58 F.Supp.2d at 107 ("treating a primary insurer's settlement with an insured as binding for allocation purposes, at least in the absence of evidence of collusion to defraud an excess insurer, furthers the strong public interest in promoting settlement").

In the present case, although there was some evidence that Continental previously had acknowledged providing coverage to Vanderbilt prior to 1962, the trial court found that neither the coverage nor the parameters thereof could be confirmed as of 2002.⁵⁰ The court also found that Continental and Hartford entered into the allocation agreement in good faith. Absent any evidence that the primary insurers intentionally allocated payments so as to shift their obligations onto the secondary insurers, we cannot conclude that it was improper for the trial court to uphold and enforce the allocation agreement.

2

Default Date of First Exposure for Pending and Future Actions

We next consider what default date of first exposure should apply to pending and future claims against Vanderbilt. Mt. McKinley contends that, even if it was reasonable for the primary insurers to use a 1962 default date for purposes of their 2002 settlement agreement, it is not reasonable to continue to use that date going forward in light of (1) Continental's recent stipulations that it provided coverage from 1956 through 1961, and (2) our Supreme Court's determination in Security that, under Connecticut law, defense and indemnity costs should be prorated to the insured for any periods during which the policyholder loses its policy or opts to self-insure. Mt. McKinley emphasizes that the principal rationales on which the trial court relied in upholding the 2002 agreement -the public policy favoring settlement and the practical difficulties that would attend reallocating past insurance payments using a different default date-do not apply with respect to pending or future claims for which payments have yet to be made. For purposes of those claims, Mt. McKinley contends, there simply is no reason to continue to bind all of the parties to the terms of a settlement agreement that was predicated on an erroneous understanding of Connecticut law and to which only the primary insurers were party. Rather, Mt. McKinley submits that obligations to defend and indemnify pending and future claims should be allocated on the basis of a 1948 default date of first exposure because that is the earliest date that a claimant could have been injured by Vanderbilt's talc. We agree with Mt. McKinley's premises but not with its conclusion.

The trial court's determination that the 2002 settlement agreement reasonably adopted a 1962 default date of first exposure was predicated on the fact that, at the time of the agreement, Vanderbilt could not prove that Continental had issued any policies prior to 1962 and it was not yet established that Vanderbilt, as the policyholder, was responsible for a pro rata share of the costs for years in which it was uninsured. However, if it had been understood at that time that Vanderbilt could be held accountable for the 1948 to 1955 period and Continental for the 1956 to 1961 period, then it would not have been reasonable for the primary insurers to enter into an agreement that would absolve Vanderbilt and Continental of their responsibility for fourteen years worth of defense and indemnity costs and, by transferring those costs to the 1962-1986 primary policies, prematurely exhaust those policies and shift the burden to the secondary insurers. See, e.g., New York Dock Co. v. Ernest W. Brown, Inc. , supra, 272 N.Y. at 180, 5 N.E.2d at 191 ; Rees v.

Viking Ins. Co. , supra, 77 Wash.App. at 719, 892 P.2d 1128. By the same logic, now that the parties' respective responsibilities for the pre-1962 period have been established, we see no reason why they should continue to rely for pending and future actions on a 1962 default date that (1) was chosen only because that was when the coverage block was believed to commence, and (2) would continue to benefit Vanderbilt and the parties to the 2002 settlement agreement at the expense of secondary insurers who did not consent to be bound thereby.

On the other hand, we perceive no rationale for moving the default date of first exposure for future claims fourteen years further into the past , to 1948, as Mt. McKinley proposes. If an underlying plaintiff is diagnosed in 2017, a default exposure date of 1948 would imply a latency period of nearly seventy years, which far exceeds the typical latency period for any asbestos related disease. Moreover, if that hypothetical plaintiff was first exposed to Vanderbilt's talc at the age of twenty-five and his disease in fact took forty years to manifest, then Mt. McKinley's proposed rule would require that we apply a presumptive date of first exposure before he was even born. The absurd consequences of such a rule would be amplified over time, as it will become increasingly unlikely that any new claim originated with an exposure during the Truman administration.

Mt. McKinley's principal argument in favor of a default date of first exposure of 1948-the year that Vanderbilt entered the talc business-is that in Security our Supreme Court chose as a default date 1951, which was the year that the defendant in that case began applying fireproofing materials containing asbestos. See Security Ins. Co. of Hartford v. Lumbermens Mutual Casualty Co. , supra, 264 Conn. at 691, 826 A.2d 107. We read Security differently. The question of the default date of first exposure was not at issue in that appeal. Rather, our Supreme Court merely noted that the parties had agreed that the defendant's potential liability commenced on the date that it entered the asbestos business. The court neither analyzed the reasonableness of that agreement nor established that, as a general rule, the default date of first exposure in a long-tail asbestos related disease claim should be the date that the defendant entered the business, even when that date is decades earlier than a claimant reasonably could have been exposed. We therefore reject Mt. McKinley's argument that the trial court should have adopted a default date of 1948.

The question, then, is what constitutes a reasonable default date of first exposure for pending and future claims against Vanderbilt that do not allege a date of first exposure and for which a date cannot be readily established through discovery. Continental proposes that we adopt an approach whereby the default date for underlying actions would simply be the average of the dates that are alleged in all of the underlying actions brought during the preceding five years. Under that approach, the default date of first exposure would continue to refresh for newly filed claims and could be expected to remain reasonably accurate.

Although Continental's proposal appears sensible, neither the other parties nor the trial court have addressed it in any depth,⁵¹ and the record before us is inadequate to assess its merits vis-à -vis other possible approaches. It would be helpful, for example, to know the following: (1) whether a statistically meaningful comparator could be obtained by consulting only those peer cases brought during the year or two preceding the claim at issue, rather than going back five years; (2) whether it would be preferable simply to consult and average the exposure dates alleged in some statistically meaningful number of immediately preceding actions, rather than averaging all of the actions filed during some arbitrary time period; (3) whether only actions alleging the same disease as that alleged in the claim at issue should be included; and (4) whether it would be preferable to back into a default date of first exposure on the basis of expert testimony as to the average latency period for the disease alleged. We believe that these and other questions; see, e.g., part III B 2 a of this opinion; can best be answered in the first instance by the trial court. Accordingly, we remand the matter to the trial court to take additional testimony if warranted, make any necessary findings, and articulate what it considers to be the most reasonable method of establishing a default date of first exposure for pending and future underlying actions in which a date is not alleged and cannot practically be established through discovery.

D

Mathematical Allocation Formula and Orphan Shares

The next set of issues concerns the mathematical formula that the trial court used during the Phase II trial to calculate Vanderbilt's pro rata share of defense and indemnity costs. Almost all of the parties share the belief that the trial court miscalculated Vanderbilt's share, but they offer differing proposals as to the proper formula. We take this opportunity to clarify the proper approach.⁵²

The following additional procedural history is relevant to our analysis. In its decisions following the Phase I and Phase II trials, the trial court determined during which years between 1948 and 2008 Vanderbilt should be treated as self-insured for purposes of allocating asbestos related defense and indemnity obligations, respectively. Having made those determinations, the court then proceeded to outline the general formula by which Vanderbilt's total liability was to be calculated. The court's starting point was footnotes 13 and 17 in Security , in which our Supreme Court concluded that the following allocation method was reasonable: "[P]ro-ration-to-the-insured [is determined] by obliging [the policyholder] to pay a share of each claim represented by a fraction that has as its denominator the number of years of the injury up to 1985, [the time at which insurance companies ceased providing asbestos related comprehensive general liability coverage], and as its numerator the number of those years in which [the policyholder] was uninsured ...." (Internal quotation marks omitted.) Security Ins. Co. of Hartford v. Lumbermens Mutual Casualty Co. , supra, 264 Conn. at 698 n.13, 826 A.2d 107 ; see id., at 720, 826 A.2d 107. To determine a policyholder's pro rata share of defense and indemnity obligations, then, Security instructs the trier of fact to determine two numbers: (1) the number of months in the coverage block and (2) the number of those months that the policyholder is effectively self-insured.⁵³ The policyholder's obligations are then calculated by dividing the latter (the numerator) into the former (the denominator), and multiplying the result by the judgment or settlement amount in a particular case. The same approach applies to calculating the obligations of the policyholder's various insurers; each is responsible, up to the policy limits, for a percentage share of the total allocation block corresponding to its time on the risk.

Applying these rules to the present case, the trial court determined that the coverage block for asbestos related claims against Vanderbilt runs from 1948 through 2008-a total of 720 months.⁵⁴ This represents the denominator in the allocation formula. With respect to defense costs, the court determined that Vanderbilt is responsible for at least 265 of these 720 months, or 36.8 percent. With respect to indemnity costs, the court determined that Vanderbilt is responsible for at least ninety-six of these 720 months, or 13.3 percent. The difference reflects the court's finding that Vanderbilt was responsible for defense but not indemnity costs between 1993 and 2007, a period during which asbestos related coverage was generally unavailable but when Vanderbilt was able to obtain some limited claims-made coverage. The court also concluded that, in addition to those self-insured months, Vanderbilt would be held responsible for a gap period between March 25, 1978 and April 26, 1978, when the company was underinsured, as well as for any period during which its insurers had become insolvent.⁵⁵

On appeal, we review the trial court's legal conclusions de novo and its factual determinations for clear error. We note at the outset that, with respect to the general allocation formula for long-tail toxic tort claims, we agree with the trial court that Security provides the proper framework for allocating costs between insurers and policyholders. Defense and indemnity costs are to be allocated evenly over the period of injury pursuant to a continuous trigger theory, and each insurer is then responsible, up to its policy limits, for the periods during which it was on the risk. For its part, the policyholder is liable for those periods during which it was effectively self-insured, as well as for any residual amounts after the primary and secondary policies in any given year have been exhausted. See Champion Dyeing & Finishing Co. v. Centennial Ins. Co. , supra, 355 N.J.Super. at 277, 810 A.2d 68 ; M. Doherty, supra, 64 U. Chi. L. Rev. 281.

With respect to the numerator in the allocation formula, and specifically as to Vanderbilt's share of the costs, Vanderbilt does not challenge the court's finding that it is responsible for both defense and indemnity costs for the ninety-six month lost policy period from 1948 through 1955, as well as for the 1978 gap period and any periods of insurer insolvency. As we explained in part III C of this opinion, however, Vanderbilt contends, and we agree, that the trial court went astray in concluding that Vanderbilt was liable for a pro rata share of defense costs for the years 1993 through 2007. The court's conclusion in Phase II that Vanderbilt was liable for 265 months of defense costs, which includes the 1993-2007 period, was thus incorrect.

Turning next to the denominator, the parties generally agree that the trial court, having adopted a continuous trigger theory of injury and an unavailability of insurance rule, should not have used an allocation block running from 1948 through 2008. They contend that the allocation block applied by the trial court (1) would result in the creation of "orphan shares" for which no party is responsible and (2) represents the maximum possible range of injury but will exceed the actual allocation block to be applied in most cases. We agree with both contentions.

1

Orphan Shares

We first address the issue of orphan shares. With respect to the period after March 3, 1986, when Vanderbilt's comprehensive general liability policies expired and asbestos coverage was generally unavailable, the trial court concluded that Vanderbilt was not self-insured as to indemnity insurance at any time and, with respect to defense cost coverage, that Vanderbilt was to be deemed self-insured only from March 3, 1993 through April 24, 2007. In part III C of this opinion, we held that the trial court incorrectly determined that Vanderbilt was self-insured for defense costs from March 3, 1993 through April 24, 2007. Therefore, with respect to underlying actions alleging a date of first exposure prior to March, 1986, Vanderbilt should not be assigned any share of the post-1985 allocation block.

It is also the case, however, that, with a few limited exceptions, none of Vanderbilt's insurers issued policies for the post-1985 period that covered asbestos related claims. Accordingly, if the court were to employ a pro rata allocation scheme in which the allocation block runs from 1948 through 2008, then there would be a period of more than twenty-two years, from 1986 to 2008, in which costs were allocated to months for which neither Vanderbilt nor any of its insurers bore responsibility. In other words, more than 37 percent of the total costs would be orphan shares.

For this reason, we agree with the parties that periods during which no insurer issued coverage and the policyholder is not treated as self-insured should not be included in the total allocation block. This approach comports with the methodology that our Supreme Court adopted in Security ; see Security Ins. Co. of Hartford v. Lumbermens Mutual Casualty Co. , supra, 264 Conn. at 698 n.13, 826 A.2d 107 ; and we believe that it is a necessary corollary of the unavailability of insurance rule. Accordingly, with certain exceptions to be discussed hereinafter, no indemnity or defense costs arising from pre-1986 exposures should be allocated to the post-1985 period. The total allocation block should run from January 1, 1948, when Vanderbilt began producing talc, until its occurrence based policies expired on March 3, 1986, or just over 458 months. Vanderbilt is responsible for at least ninety-six months-roughly 21 percent-of that period, representing its period of self-insurance from 1948 through 1955. Vanderbilt also will be responsible for any additional periods of self-insurance or underinsurance, including periods of insurer insolvency, as determined by the trial court.

2

Maximum Coverage Block

Turning to the parties' second contention, we also agree with Vanderbilt and many of the defendants that this allocation represents the maximum period of injury for allocation purposes, and Vanderbilt's share of that maximum,⁵⁶ but that the actual allocation block and the parties' shares thereof will vary on a case-by-case basis. That is to say, in the event that an underlying plaintiff was first exposed to Vanderbilt talc on January 1, 1948, and did not manifest asbestos related disease until after March 3, 1986, then that plaintiff's injuries as defined by the continuous trigger theory would span the entire allocation block, and each party's respective share of liability for that particular plaintiff would be the same as its share of the total allocation block. In most instances, a particular plaintiff's injuries will span only a portion of the total allocation block, either because they will have been exposed after January 1, 1948, or because they will have manifested disease before March 3, 1986. The shares of defense and liability costs attributable to Vanderbilt and to its various insurers in such instances are not fixed; they will vary in each case. For example, Vanderbilt, which is treated as self-insured for the 1948 through 1955 period, would be responsible for the lion's share of costs arising from an underlying action alleging a date of first exposure in 1949 and manifestation of disease in 1959, whereas Hartford would shoulder more of the responsibility for a claim alleging first exposure in 1980 and manifestation sometime after 1986. 3

Additional Allocation Issues

Finally, we address three additional issues that the parties have raised with respect to the allocation formula and that are likely to confront the trial court on remand or during the next phase of the trial. First, the parties dispute whether and how the allocation block should be expanded to account for the fact that Vanderbilt was able to obtain claims-made coverage between 1993 and 2007. As we explained in part III B 3 of this opinion, the fundamental differences between occurrence based and claims-made policies imply that it would not be appropriate to include in the total allocation block all of the years in which Vanderbilt obtained claims-made policies after occurrence based coverage for asbestos related disease had become unavailable. A question arises, however, as to how costs should be allocated when an underlying plaintiff is first exposed to a potential toxin prior to 1986, while occurrence based policies were in effect, but is diagnosed with asbestos related disease and files suit after 1985, in a year in which a claims-made policy was in effect. Under those circumstances, there is no doubt that the claims-made policy is also on the risk. As our sister courts have recognized, "the issue [of pro rata allocation of long-tail tort liabilities] becomes infinitely more complicated when a risk is covered by a succession of policies written first on an occurrence and, in later years, on a claims-made basis ...." Champion Dyeing & Finishing Co. v. Centennial Ins. Co. , supra, 355 N.J.Super. at 268, 810 A.2d 68.

In such a scenario, Vanderbilt contends, and we agree, that the most sensible approach is simply to expand the allocation block to include the claims-made policy period during which the claim is brought. For example, if an underlying plaintiff brought an action in January, 1995, alleging a date of first exposure to Vanderbilt's talc in January, 1965, then the allocation block would comprise approximately 266 months-the period from January, 1965 through March, 1986, plus the twelve months of 1994-1995 during which the Gerling policy was in effect. This approach is fair to all parties, comports with their reasonable expectations, and maximizes the resources available to respond to claims.⁵⁷

Second, and relatedly, the parties disagree as to how the allocation formula is to be applied to years such as 2003, when Vanderbilt was able to obtain a primary insurance policy but when excess and umbrella insurance covering asbestos related claims was not available. Vanderbilt submits that no costs should be allocated to that year because the only coverage it obtained-the 2003 American International primary policy-has been exhausted. In any event, Vanderbilt argues, no costs should be allocated to it for 2003 because it purchased all available insurance for that year. Mt. McKinley, by contrast, contends that 2003 should be treated the same as any other year in which insurance was available, and that whether the policy has exhausted is irrelevant to the allocation formula.

As a general rule, Mt. McKinley is correct that long-tail liabilities are allocated evenly across the relevant allocation block without regard to whether particular policies in particular years have been exhausted.

Once a sum has been allocated to a particular year or policy period, then any policies on the risk for that period must absorb the costs, pursuant to the policy terms, until exhausted, and once all policies are exhausted the insured is liable for any remainder.

The present case presents a problem, however, insofar as Vanderbilt was able to obtain limited primary coverage but no secondary coverage in 2003. We already have concluded that it would be improper to treat Vanderbilt as self-insured for that year because it obtained all available insurance. See part III B 3 of this opinion. But if we were to allocate to 2003 an equal share of the liabilities for claims brought in that year, the effect would be the same. Once the relatively low primary policy limit exhausts, Vanderbilt would be liable for a potentially sizable residual. This would run counter to the rationales that underlie the unavailability rule, and also would create a disincentive for policyholders to make every effort to obtain any insurance coverage during periods when adequate coverage is not readily available. Cf. Olin Corp. v. Ins. Co. of North America , 221 F.3d 307, 327 (2d Cir. 2000).

In general, then, we believe that the fairest and most reasonable approach, ab initio, would have been to include 2003 in the allocation block up to the policy limits for that year, but then to divide the remainder among the years for which insurance was fully available.⁵⁸ In the present case, however, because the 2003 American International policy is already exhausted, the outcome effectively will be that sought by Vanderbilt: 2003 should not be included in the allocation block with respect to pending and future liabilities.

Third, we consider Vanderbilt's argument, raised in its reply brief, that, if the trial court determines that it was uninsured or underinsured during certain months or years prior to 1986, it should be held responsible for a pro rata share of the costs for those periods only if the defendants can prove that insurance was generally available. Vanderbilt contends that the defendants bear the burden of establishing the availability of insurance and that they have failed to satisfy that burden. Therefore, because Vanderbilt's failure to obtain insurance during certain pre-1986 periods might simply reflect the fact that insurance was unavailable at the time, there should be no proration to the insured for those periods. We are not persuaded.

In its Phase I and Phase II decisions, the trial court found that it was undisputed that insurance for the defense and indemnification of asbestos related injury claims was available through 1985, and noted that the parties had stipulated accordingly. The court specifically noted that it was in 1985 that asbestos exclusion clauses were first incorporated into the standard form comprehensive general liability policy. Other courts and commentators likewise have observed that there was no issue with respect to the availability of coverage for asbestos related claims prior to the mid-1980s. See, e.g., Stonewall Ins. Co. v. Asbestos Claims Management Corp. , supra, 73 F.3d at 1204 n.18 ; J. Stempel, supra, 12 Conn. Ins. L.J. 464-65. We therefore disagree with Vanderbilt that its insurers bore the burden of establishing the availability of insurance coverage prior to 1986.⁵⁹

Moreover, prior to filing its reply brief in this court, Vanderbilt repeatedly conceded the accuracy of the court's findings. Both at trial and in its initial appellate brief, Vanderbilt acknowledged that the trial court was correct in finding that insurance for the defense of claims of asbestos related injury was available from 1948 until March, 1986. Having permitted the trial court to rely on that representation, Vanderbilt will not now be heard to contradict it. See Dougan v. Dougan , 301 Conn. 361, 372-73, 21 A.3d 791 (2011). Accordingly, we will not disturb the finding of the trial court that coverage for asbestos related injuries was available prior to 1986, nor its conclusion that Vanderbilt must be treated as self-insured for any portion of that time for which it was uninsured or underinsured.

E

Prospective Application of Court's Rulings

We next consider claims by certain insurers that the trial court improperly determined that its findings and allocation rules would be applied on a prospective basis from the date of the Phase II memorandum of decision. Such an approach would result, among other things, in freeing Vanderbilt from any responsibility for a pro rata share of its past defense and indemnity costs. Although we agree that a prospective only application of the court's rulings in the present case would be presumptively improper, we do not understand that to have been the court's intention.

The following additional procedural history is relevant to this issue. At the end of its Phase II decision, the court ruled that its "findings in Phases I and II lead the court to the conclusion that the allocation of defense and indemnity costs shall be applied prospectively, and consistent with the principles set forth in Security ...." The court did not elaborate on its meaning or intent in stating that its findings would be applied in a prospective manner.

In a counterclaim it filed, Mt. McKinley had requested that the court reallocate prior defense and indemnity expenditures on a pro rata basis and order reimbursement by Vanderbilt in accordance therewith.⁶⁰ Following the release of the Phase II decision, Mt. McKinley filed a motion to "reargue, reconsider and/or clarify" that decision, in which it asked the court to clarify its prospectivity ruling in light of "the fact that all payments for defense and indemnity Mt. McKinley has made over the past several years [have] been subject to a full and express reservation of rights ...." Mt. McKinley averred that "nowhere does the Phase II decision identify or define what is meant by 'prospectively.' Nor does the Phase II decision identify any legal basis for 'prospective' application." Mt. McKinley further argued that "there certainly is no basis in law or fact to apply pro rata allocation" from the date of the Phase II decision.

In denying the motion, the court reiterated that "[t]he issues as to ... any claim of damages for recovery from the plaintiff for the overpayment of defense and indemnity costs by [Continental] and other carriers, or any claims of reimbursement between carriers, were reserved to later phases of the trial.... Any claims and defenses relative to the obligation of specific policies that have not already been addressed by the court in ... Phases I and II ... remain under consideration pending the completion of the later phases of the trial." With particular respect to its prospectivity ruling, the court stated only that it contained no ambiguity "that would necessitate clarification." That ruling did not address the issue of Mt. McKinley's reservation of rights.

On appeal, Vanderbilt contends-and many of the defendants agree-that the use of the term "prospectively" indicates the court's intention that its pro rata, time-on-the-risk allocation methodology will apply only to future defense and indemnity expenditures. The parties disagree, however, as to whether such a prospective only application of the court's findings and allocation methods would be appropriate. Vanderbilt contends that to go back and reallocate past payments on thousands of underlying claims would be impracticable. Mt. McKinley, by contrast, contends that application of the correct allocation methodology solely from the date of the court's Phase II decision would eviscerate its reservation of rights and confound a central purpose of the pending Phase III proceeding. Mt. McKinley thus argues that the allocation methodology should apply retroactively, rather than from the March 28, 2014 date of the Phase II decision. For its part, Hartford argues that, regardless of what the trial court intended, the court's allocation rules should be applied retroactively to December 2, 2011-the date when Hartford's primary policies exhausted-to assure that Hartford's policy limits are not reopened.⁶¹ The proper application of the allocation methodology adopted by the trial court, as refined by this court on appeal, presents an issue of law over which we exercise de novo review. Despite the breadth of that review, we are mindful of the trial court's observation that the present case is decided in "the unique context of mass long latency losses, a situation which ... requires a specialized response insofar as allocation is concerned."

1

Whether Trial Court Intended to Apply its Allocation Rules Prospectively

We first consider whether the parties have correctly construed the court's Phase II decision and, specifically, whether the court intended that all of its findings and rulings would apply on a prospective only basis. Viewed in light of the entirety of the Phase II decision and related orders, we conclude that the court intended otherwise.

Relying primarily on the common usage of the term, Vanderbilt argues that "prospectively," as that term is used in the court's decision, "means from the date of the Phase II decision ...." The common understanding of the term "prospective" is not in dispute, which is defined as "relating to or effective in the future." Merriam-Webster's Collegiate Dictionary (11th Ed. 2003) p. 998; see also Black's Law Dictionary (9th Ed. 2009) p. 1342 (defining "prospective" as "[e]ffective or operative in the future"); Ballentine's Law Dictionary (3d Ed. 1969) p. 1014 (defining "prospective" as "[l]ooking to the future"). In Vanderbilt's view, the court's use of that term in the concluding sentence of its seventy-seven page memorandum of decision indicates that all of the court's allocation rulings are to be applied on a prospective only basis, thereby foreclosing reimbursement claims for defense and indemnity expenditures incurred prior to March 28, 2014.

It is well established, however, that, "an opinion must be read as a whole, without particular portions read in isolation, to discern the parameters of its holding." Fisher v. Big Y Foods, Inc. , 298 Conn. 414, 424-25, 3 A.3d 919 (2010) ; see also Matza v. Matza , 226 Conn. 166, 187, 627 A.2d 414 (1993) ("[t]he two sentences upon which the defendant hangs her hat must be read in the context of the entire memorandum of decision"); McGaffin v. Roberts , 193 Conn. 393, 408, 479 A.2d 176 (1984) ("[t]he memorandum of decision must be read as a whole to put these statements in fair context"), cert. denied, 470 U.S. 1050, 105 S.Ct. 1747, 84 L.Ed.2d 813 (1985). Vanderbilt's reading of the sentence at issue is in tension with multiple other portions of the Phase II decision, as well as with the court's prior and subsequent procedural rulings.

The principal infirmity in Vanderbilt's interpretation is that it largely ignores the unique procedural posture of this multiphase litigation. From the outset, various defendants have asserted claims for reimbursement for past payments stemming from their participation in the defense and indemnity of Vanderbilt with respect to the underlying actions. Some, such as Mt. McKinley; see footnote 60 of this opinion; specifically requested that the court reallocate the respective obligations of Vanderbilt and the insurers on a pro rata basis and order reimbursement accordingly. In response, the court repeatedly apprised the parties that it would adjudicate claims for reimbursement of past defense and indemnity payments in a future phase of this litigation, and entered orders to that effect.

On June 4, 2012, for example, the court entered the first of many bifurcation orders. In its order, the court noted in relevant part that "all claims relative to the issues of allocation ... reimbursement and/or overpayment of defense costs ... shall ... be tried to the court." Continental thereafter filed a motion "to further bifurcate from the court trial any evidence regarding claims for damages on the issue of recovering from [Vanderbilt] any reimbursement and/or the overpayment of defense and indemnity costs by all parties." The court granted that motion and ordered that "[t]hose issues shall be tried at a later dated to be scheduled by the court." In its July 18, 2012 order, the court confirmed that "[u]pon resolution of the coverage block issue through the written opinion of the court, the parties shall then ... try the remaining issues left for trial to the court," including "the claims of damages for the recovery of overpayments by [insurers] of [Vanderbilt's] defense and indemnity costs ...." (Citations omitted.) Approximately two months after issuing its Phase I decision, the court again noted that such claims would remain the focus of later proceedings. In its May 3, 2013 order, the court stated: "[I]t is ordered that the issue of damages, including reimbursement and/or overpayment of indemnity and defense costs, shall be tried (in a third phase) ... following the court's ruling on all other nonjury issues to be heard in the second phase of the trial ...."

That the court intends in Phase III to address reimbursement claims for defense and indemnity expenditures incurred prior to March 28, 2014, is further evidenced throughout the Phase II decision itself. At the outset of the decision, the court emphasized that "[t]he issues as to ... any claim of damages for recovery from [Vanderbilt] for the overpayment of defense and indemnity costs by [Continental] and other carriers, or any claims of reimbursement between carriers, have been reserved to later phases of the trial." Later in the opinion, the court noted that "any specific ruling as to the exact financial obligation for defense or indemnity costs under the individual policies issued by each carrier is left to Phase III of this trial as part of the calculation of damages and reimbursement of such costs.... [S]uch issues [will] be fully addressed in Phase III. That latter phase shall necessarily determine the scope of any defense or indemnity obligations in order to determine the amounts due, if any, under any policy not already addressed by the court in this phase." (Emphasis added.) Even further into the decision, the court clarified that "the claims set forth in [Continental's] cross claim and [American International's] counterclaim are matters intended to be left for a later phase of the trial dealing with claims of reimbursement for overpayment of indemnity/defense costs/expenses ...."⁶² The court likewise stated in the concluding sentence to that decision that "[c]onsistent with the court's prior order ... Phase III of this trial on the issue of damages, including the liability for and amount of reimbursement and/or overpayment of indemnity and defense costs, shall commence no later than forty-five days from the date hereof." Thus, the court's decision undoubtedly contemplates consideration of claims of overpayment and reimbursement for past defense and indemnity payments during the Phase III proceeding.

Most significant is the court's August 21, 2014 order-issued almost five months after its Phase II decision. In responding to a motion for reargument and reconsideration filed by Continental, the court indicated that the pending claims regarding overpayment and reimbursement for past defense and indemnity expenditures would be addressed at the upcoming Phase III trial. As the court's order states: "The issues as to ... any claim of damages for recovery from [Vanderbilt] for the overpayment of defense and indemnity costs by [Continental] and other carriers, or any claims of reimbursement between carriers, were reserved to later phases of the trial. Upon motion of the parties, the court agreed to bifurcate Phase II of the trial so as to remove the issues of damages and reimbursement of any overpayment of defense and indemnity costs which the court had directed be addressed. Those issues are to be considered following the completion of Phase II." (Emphasis added.)

Vanderbilt's interpretation of the court's prospectivity ruling is difficult to reconcile with the court's repeated orders expressly deferring consideration of the issue of overpayment and reimbursement for past payments until the Phase III proceeding. If Vanderbilt's reading is correct, and the allocation methodology adopted by the court has no application to defense and indemnity costs incurred prior to the March 28, 2014 Phase II decision, then little remains of the reimbursement and overpayment claims reserved for the Phase III trial. The court's August 21, 2014 order demonstrates in convincing fashion that the court intended otherwise.

Vanderbilt's interpretation of the court's prospectivity ruling also is in tension with the legal authorities on which the court relied in its Phase II decision. In the ruling at issue, the court stated that its "findings in Phases I and II lead the court to the conclusion that the allocation of defense and indemnity costs shall be applied prospectively, and consistent with the principles set forth in Security ...." (Emphasis added.) Certain principles enunciated in Security , therefore, merit additional attention.

In Security , our Supreme Court adopted a pro rata allocation method for adjudicating long latency loss claims that implicate multiple insurance policies. Security Ins. Co. of Hartford v. Lumbermens Mutual Casualty Co. , supra, 264 Conn. at 720, 826 A.2d 107. Significantly, the Supreme Court held, among other things, that the trial court in that case "properly recognized a cause of action for equitable contribution and reimbursement by an insurer against its insured." Id., at 699, 826 A.2d 107. As the court explained, "[a] cause of action for reimbursement is cognizable to the extent required to ensure that the insured not reap a benefit for which it has not paid and thus be unjustly enriched. Where the insurer defends the insured against an action that includes claims not even potentially covered by the insurance policy, a court will order reimbursement for the cost of defending the uncovered claims in order to prevent the insured from receiving a windfall. Consistent with the pro rata method of allocation, we have concluded that time on the risk is a reasonable means of prorating defense costs for periods of self-insurance. Those costs allocable to periods of self-insurance are not even potentially covered by the insurer's policies.... Thus, the insured would be unjustly enriched were we to conclude that there is no claim for reimbursement for the cost expended by the insurers in defending periods of self-insurance. Accordingly, we conclude that, where the pro rata method of apportionment applies, there is a cause of action for reimbursement by an insurer against its insured." Id., at 717-18, 826 A.2d 107. It is precisely such claims that the trial court reserved for the Phase III proceeding in the present case.

Moreover, as the Supreme Court noted, the trial court in Security "allocated [past] defense costs to [the manufacturer] on a pro rata basis and ordered contribution and reimbursement" to the plaintiff insurer. Id., at 699, 826 A.2d 107. Our Supreme Court upheld the propriety of that retroactive application of the pro rata allocation methodology, stating in relevant part that "the trial court properly ordered [the manufacturer] to reimburse [the plaintiff insurer] for its pro rata share of the cost of defending the ... asbestos litigation" for certain time periods. Id., at 718, 826 A.2d 107 ; accord Continental Casualty Co. v. Indian Head Industries, Inc. , 666 Fed.Appx. 456, 466-67, 2016 WL 7321362, *8 (6th Cir. 2016) (pro rata allocation appropriate for previously incurred defense costs). Security thus instructs that a trial court employing pro rata allocation among multiple insurance policies properly may apply that methodology in retroactive fashion by entertaining claims for reimbursement by insurers. The trial court's express reliance on Security throughout its Phase II decision provides further evidence that the court did not intend all of its allocation rulings to apply on a prospective only basis.

What, then, was the court's intention with respect to prospectivity? We perceive at least two possibilities. First, the court may have used the term "prospectively" merely in a procedural sense, to indicate that it planned to address those matters that remained pending and apply the new allocation rules in future phases of this multiphase litigation. On this reading, the court's use of the term "prospectively" is synonymous with "effective or operative in future phases of these proceedings." That construction comports with the express reliance in the court's prospectivity ruling on "the principles set forth in Security ," which recognized an insurer's right to reimbursement for past defense and indemnification payments. It further acknowledges the numerous orders and rulings of the court dedicating that issue to the future Phase III proceeding. Finally, that construction explains the court's silence on the substance of Mt. McKinley's reservation of rights, as that issue involves a question of overpayment and Mt. McKinley's corresponding right to reimbursement.

Alternatively, the court's reference to prospective application may have been limited to payments that were made pursuant to the primary insurers' 2002 settlement agreement, which the court previously had determined would not be retroactively reallocated. As we explained in part III C 1 of this opinion, in 1995, more than one decade prior to the commencement of the present litigation, Continental filed an action against Hartford seeking contribution for defense and indemnity payments that it had made on underlying claims involving Vanderbilt. Those parties settled that dispute by executing the allocation agreement in 2002, an agreement that the trial court concluded was reasonable and enforceable. The court's only analysis of retroactive versus prospective application in the Phase II decision was in that context.⁶³ Accordingly, it is reasonable to assume that the court's concluding, unelaborated reference to prospective application also was intended to apply in that specific context.⁶⁴ Tellingly, the court concluded its discussion of the exhaustion issue by noting that "the liability of any umbrella or excess carriers above those [primary] policies are to be determined consistently with the pro rata allocation principles set forth in Security ." In so doing, the court implicitly distinguished the allocation of obligations among the primary insurers pursuant to the allocation agreement-which the court refused to "compel Hartford and [Continental] to retroactively reallocate"-from the allocation of obligations of the secondary insurers, to which, the court determined, it would apply the pro rata allocation methodology.

We reiterate that this is an interlocutory appeal. On remand the court will have an opportunity to clarify its intentions in this regard, consistent with the legal principles discussed in part III E 2 of this opinion.

2

Whether Prospective Only Application Would Be Permissible under Connecticut Law

In the event that we have misperceived the court's intentions, and it intended to apply its allocation findings and rules solely to future defense and indemnity payments, we now turn our attention to the question whether Connecticut law would permit the court to apply its findings and conclusions on a prospective only basis. We conclude that such an approach would run afoul not only of the general presumption that judicial decisions apply retroactively but also of our law's specific sanctioning of reservations of rights in insurance coverage disputes.

a

Judicial Decisions Are Presumptively Retroactive

In Connecticut, "[t]he general rule is that judicial decisions apply retroactively to pending cases ...." (Citations omitted.) Campos v. Coleman , 319 Conn. 36, 62, 123 A.3d 854 (2015) ; accord Harper v. Virginia Dept. of Taxation , 509 U.S. 86, 94, 113 S.Ct. 2510, 125 L.Ed.2d 74 (1993) (discussing "fundamental rule" of retrospective operation of rulings in civil cases); Kuhn v. Fairmont Coal Co. , 215 U.S. 349, 372, 30 S.Ct. 140, 54 L.Ed. 228 (1910) (Holmes, J., dissenting) ("[j]udicial decisions have had retrospective operation for near a thousand years"). In light of that general rule of retroactivity, our Supreme Court in the civil context has rejected a litigant's contention that the court "should not make [its ruling] available to the parties in the present case but should apply it prospectively only." Campos v. Coleman , supra, at 61, 123 A.3d 854 ; see also George v. Ericson , 250 Conn. 312, 326, 736 A.2d 889 (1999) ("[w]hen this court renders a decision, the general rule is that the decision will apply to the parties involved in the case in which the decision was reached"); accord Harper v. Virginia Dept. of Taxation , supra, at 97-98, 113 S.Ct. 2510 (judicial opinion "is properly understood to have followed the normal rule of retroactive application and must be read to hold ... that its rule should apply retroactively to the litigants then before the Court" [internal quotation marks omitted] ); Bradley v. Richmond School Board , 416 U.S. 696, 711, 94 S.Ct. 2006, 40 L.Ed.2d 476 (1974) ("a court is to apply the law in effect at the time it renders its decision").

Our Supreme Court nonetheless has, on multiple occasions, applied a three part test for determining whether prospective application is appropriate. "A common-law decision will be applied nonretroactively only if: (1) it establishes a new principle of law, either by overruling past precedent on which litigants have relied ... or by deciding an issue of first impression whose resolution was not clearly foreshadowed ... (2) given its prior history, purpose and effect, retrospective application of the rule would retard its operation; and (3) retroactive application would produce substantial inequitable results, injustice or hardship." (Citation omitted; internal quotation marks omitted.) Ostrowski v. Avery , 243 Conn. 355, 378 n.18, 703 A.2d 117 (1997) ; see also Campos v. Coleman , supra, 319 Conn. at 61-64, 123 A.3d 854 ; Neyland v. Board of Education , 195 Conn. 174, 179-83, 487 A.2d 181 (1985) ; cf. In re Daniel N. , 323 Conn. 640, 653, 150 A.3d 657 (2016) (describing three part test as "the usual test for retroactivity"). Notably, our Supreme Court has described this test as "an exception to [the] general rule" that judicial decisions "apply retroactively to pending cases ...." (Citations omitted; emphasis in original.) Campos v. Coleman , supra, at 62, 123 A.3d 854 ; accord Garcia v. Pasquarell , 117 Fed.Appx. 337, 339 (5th Cir. 2004) ("retroactive application of judicial decisions is overwhelmingly the norm, and any exceptions to this rule that might still exist are limited to the extremely unusual and unforeseeable case" [internal quotation marks omitted] ); Stonehill College v. Massachusetts Commission Against Discrimination , 441 Mass. 549, 569-70, 808 N.E.2d 205 (applying similar three part test for exception to general rule of retroactivity and specifically considering "[p]rinciples of equity and fairness"), cert. denied sub nom. Wilfert Bros. Realty Co. v. Massachusetts Commission Against Discrimination , 543 U.S. 979, 125 S.Ct. 481, 160 L.Ed.2d 356 (2004). In the present case, the Phase I and II proceedings involved several issues of first impression whose resolution was not clearly foreshadowed, including the adoption of the unavailability of insurance rule and the continuous trigger theory of coverage, a point underscored by their extensive treatment in this opinion. Thus, we may assume that the first prong of the retroactivity exception test is satisfied.

Nevertheless, the trial court has not articulated any basis on which we can conclude that the second and third prongs of that test are satisfied. The court's lengthy and thorough Phase II decision contains no legal analysis or explication of the court's prospectivity ruling. Furthermore, when asked to clarify the basis of that ruling, the court declined to do so. In ruling on Mt. McKinley's motion to "reargue, reconsider and/or clarify", the court stated simply that, in its view, no ambiguity was present that necessitated clarification of its decision.

Nor do we perceive any rationale for concluding that retroactive application of the court's rulings would retard their operation or produce substantial inequitable results, injustice or hardship.⁶⁵ Given the multifaceted nature of this litigation, and the court's consistent reservation of "any claim of damages for recovery from [Vanderbilt] for the overpayment of defense and indemnity costs by [Continental] and other carriers, or any claims of reimbursement between carriers" to the Phase III proceeding, it is difficult to fathom how application of the allocation methodology to such claims would impair its operation or produce inequitable results. This is particularly true in light of the fact that parties such as Mt. McKinley furnished a defense and indemnification of Vanderbilt pursuant to a reservation of rights. Thus, an exception to the general rule of retroactivity does not appear to be warranted with respect to the obligations of the secondary insurers and Vanderbilt.

b

Prospective Only Application Would Vitiate Mt. McKinley's Reservation of Rights

Mt. McKinley also contends, and we agree, that applying the trial court's findings and rulings on a solely prospective basis would eviscerate the reservation of rights that Mt. McKinley issued when it came on the risk and began defending and indemnifying Vanderbilt. The following additional undisputed facts and procedural history are relevant to this issue.

At trial, Mt. McKinley offered documentary and testimonial evidence indicating that it undertook the defense and indemnification of Vanderbilt subject to a reservation of rights. Lillian Philburn, a claims manager who handled Vanderbilt's account on behalf of Mt. McKinley, testified at the Phase II trial. Philburn confirmed that Mt. McKinley began "defending and indemnifying [Vanderbilt] with regard to underlying talc claims" in late 2011, though always subject to "a complete and full reservation of rights ...." As she explained, "[w]e advised [Vanderbilt] that we will participate in their defense and indemnity under a reservation subject to certain terms and conditions, such as we reserve the right to reallocate any payments that we have made, to recover any payments that we have made. We reserve the right to confirm that there has been bodily injury resulting in occurrence during our policy period. We also reserved on proper allocation of any payments that have been made in the past and going forward."⁶⁶ Philburn's testimony was corroborated by Thomas Radcliffe, an attorney who served as Vanderbilt's national coordinating counsel for underlying actions. Radcliffe confirmed in his testimony that Mt. McKinley had agreed to defend Vanderbilt pursuant to a reservation of rights. Consistent with this undisputed testimony, the trial court found that Mt. McKinley had, in fact, proceeded subject to a reservation of rights.

"A reservation of rights agreement serves to furnish temporary protection to an insured, even though ... it may turn out that the insured was not entitled to such protection." (Internal quotation marks omitted.) Signature Development Cos., Inc. v. Royal Ins. Co. of America , 230 F.3d 1215, 1220 (10th Cir. 2000) ; see also Globecon Group, LLC v. Hartford Fire Ins. Co. , 434 F.3d 165, 176 (2d Cir. 2006) (unasserted defenses not waived where insurer "repeatedly and expressly reserved its rights in its communications" with insured). Reservations of rights exist "to protect both the insurer and the insured by allowing an insurer who is uncertain of its obligations under the policy to undertake a defense while reserving its rights to ultimately deny coverage following its investigation or to file a declaratory judgment action to determine its obligations." 14 L. Russ & T. Segalla, Couch on Insurance (Cum. Supp. 2016) § 202:38, p. 350 n.71. We share Mt. McKinley's concern that to apply the trial court's findings and rulings on a solely prospective basis would essentially eviscerate its reservation of rights and, as a result, discourage other insurers from defending their insureds under reservation.⁶⁷ At the very least, the fact that Mt. McKinley defended Vanderbilt under an express reservation of rights suggests that it would not be inequitable to apply the court's pro rata allocation rules to Vanderbilt on a retroactive basis. 3

Guidance on Remand

In the preceding parts of this opinion, we concluded on the basis of the record before us that prospective only application of the trial court's findings and allocation rulings would offend both the general presumption in favor of retroactivity and the law's specific preference for reservations of rights in insurance coverage disputes. Nevertheless, in light of the interlocutory nature of these appeals and unique character of long-tail toxic tort litigation, prudence dictates that we not foreclose further consideration of these issues by the trial court. We are particularly mindful of the trial court's superior vantage point as fact finder and gatekeeper in this quintessential example of complex litigation. In this case, the trial court has made it abundantly clear that "any specific ruling as to the exact financial obligation for defense or indemnity costs under the individual policies issued by each carrier is left to Phase III of this trial as part of the calculation of damages and reimbursement of such costs." The precise nature of those obligations, and the attendant claims of overpayment and entitlement to reimbursement, all entail factual determinations. Given those critical unresolved issues of fact, we believe it would be imprudent at this stage to foreclose the possibility that the retroactivity exception might be implicated at a later date. Should the trial court, therefore, become convinced that a departure from the rule of retroactivity is warranted with respect to the parties' overpayment and reimbursement claims, then further analysis of the latter two prongs of that test will be appropriate. Because these appeals are interlocutory in nature, an avenue of further appellate review of any such determination remains following the completion of the proceedings before the trial court.

IV

SCOPE OF COVERAGE AND POLICY EXCLUSIONS

We next consider claims relating to three types of clauses contained in certain of Vanderbilt's primary and secondary insurance policies. First, Mt. McKinley⁶⁸ contends that the trial court improperly determined that pollution exclusions in its policies do not apply so as to bar coverage for most asbestos related disease claims. Second, National Casualty challenges the court's determination that occupational disease exclusions in its policies preclude coverage only for claims brought by Vanderbilt's own employees. Third, Vanderbilt contends that the trial court improperly determined that certain of Continental's secondary insurance policies do not provide for a duty to defend when the underlying insurance has exhausted. For the reasons that follow, we affirm the rulings of the trial court with respect to the pollution exclusion and duty to defend clauses but reverse with respect to the occupational disease exclusions.

In addition to the legal principles discussed in part II of this opinion, the following principles govern our resolution of these claims. The "rule of construction favorable to the insured extends to exclusion clauses." (Internal quotation marks omitted.) Travelers Ins. Co. v. Namerow , 257 Conn. 812, 827, 778 A.2d 168 (2001), superseded in part on other grounds, 261 Conn. 784, 807 A.2d 467 (2002). The insurer bears the burden of proving that an exclusion clause applies. Capstone Building Corp. v. American Motorists Ins. Co. , supra, 308 Conn. at 788 n.24, 67 A.3d 961. The insured, however, has the burden of proving that an exception to an exclusion reinstates coverage. Id. ;

Buell Industries, Inc. v. Greater New York Mutual Ins. Co. , 259 Conn. 527, 551, 791 A.2d 489 (2002).

A

Pollution Exclusions

We first consider the applicability of the so-called pollution exclusion clause to claims arising from alleged asbestos exposure. Most of Vanderbilt's policies written after 1970 contain a clause that excludes coverage for bodily injury or property damage resulting from the release of pollutants. Mt. McKinley contends that the plain language of these exclusions unambiguously applies to and bars coverage for the underlying claims. Vanderbilt responds that the pollution exclusions, on their face, apply only to "traditional" environmental pollution and do not bar coverage for asbestos related claims, most of which allege harms arising from exposure to asbestos dust released in indoor environments in the course of routine manufacturing or construction activities. In the alternative, Vanderbilt contends that the trial court properly concluded that the contractual language is ambiguous with respect to alleged asbestos related disease and that the policies should, therefore, be construed in favor of coverage. The question whether a pollution exclusion clause in a comprehensive general liability policy bars coverage for claims arising from exposure to toxic substances such as asbestos in indoor environments and/or in the course of their intended use is one of first impression for Connecticut's appellate courts. Recognizing that the question is a close one, over which our sister courts are sharply divided, we conclude that the pollution exclusions bar coverage only when the exposure arises from traditional environmental pollution, such as when the dumping of waste materials containing asbestos causes asbestos fibers to migrate onto neighboring properties or into the natural environment.

The following additional procedural history is relevant to this claim. In 2010, a number of the secondary insurers moved for summary judgment, contending that pollution exclusion clauses contained in their Vanderbilt policies exclude coverage for asbestos related claims and that all of the underlying actions fall within the scope of the exclusions. In a preliminary decision issued in July, 2011, the trial court denied the defendants' motions, concluding that, at least with respect to one representative underlying action, the policy language did not unambiguously apply to claims of asbestos related disease. The court affirmed this result in its Phase II decision, concluding that the various exclusions at issue were ambiguous as applied to the underlying actions and, therefore, must be construed in favor of coverage.

1

Standard Pollution Exclusion

Most of the policies at issue in this case contain what has come to be known as the standard pollution exclusion. The language of one typical policy⁶⁹ reads: "Exclusion (Contamination or Pollution)

"It is agreed that the insurance does not apply to personal injury or property damage arising out of the discharge, dispersal, release or escape of smoke, vapors, soot, fumes, acids, alkalis, toxic chemicals, liquids or gases, 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."

We first consider whether the trial court properly concluded that this standard pollution exclusion does not apply so as to bar coverage for the types of claims typically raised in the underlying complaints. We then consider whether a different result obtains with respect to the nonstandard pollution exclusion clauses contained in several of the defendants' policies, which contain language that differs materially from the standard exclusion.

a

Connecticut Precedent

As an initial matter, we address the dispute between the parties as to whether the resolution of this question is dictated by the decision of our Supreme Court in Heyman Associates No. 1 v. Ins. Co. of Pennsylvania , supra, 231 Conn. 756, 653 A.2d 122. In Heyman , the defendant insurers denied coverage to the plaintiff, on the basis of a pollution exclusion clause, for damages caused when a large quantity of fuel oil leaked from the plaintiff's property into Stamford Harbor, potentially impacting the water, shoreline, and wildlife, and creating liability under the Federal Water Pollution Control Act, 33 U.S.C. § 1251 et seq.Heyman Associates No. 1 v. Ins. Co.

of Pennsylvania, supra, at 759-60, 653 A.2d 122. The precise question before our Supreme Court was "whether the language of each policy clause clearly and unambiguously defines 'pollutant' to include fuel oil released into a waterway such as the Stamford Harbor."⁷⁰ Id., at 771, 653 A.2d 122. To answer this question, the court began by reviewing the plain language of each policy provision and looked to three sources to construe its meaning: (1) dictionary definitions of the ordinary meaning of "pollutant" and related contract terms; id., at 772, 653 A.2d 122 ; (2) statutory definitions of these terms; id., at 773, 653 A.2d 122 ; and (3) decisions from courts in other jurisdictions that addressed the question of whether a fuel oil spill constitutes pollution for purposes of a pollution exclusion clause. Id., at 774, 653 A.2d 122. Because each of these sources indicated that a fuel oil spill into a public waterway constitutes pollution, the Supreme Court concluded that the policy language clearly and unambiguously applied to bar coverage for the claims at issue. Id., at 774-76, 653 A.2d 122. Accordingly, the court declined to consider extrinsic evidence, such as the drafting history of the pollution exclusion clauses, that might vary or distort the plain meaning of the policy language. Id. On three separate occasions, however, the court in Heyman emphasized that the language of the pollution exclusion was clear and unambiguous "as applied to the facts of [that ] case ." (Emphasis added.) Id., at 779, 781, 788, 653 A.2d 122 ; see also Schilberg Integrated Metals Corp. v. Continental Casualty Co. , 263 Conn. 245, 276-78, 819 A.2d 773 (2003) ( Schilberg ) (declining to consider parol evidence because language of pollution exclusion was "clear and unambiguous as applied to the present facts").

Vanderbilt contends, and the trial court agreed, that the present case is factually distinguishable from Heyman . Whereas a large oil spill into a public waterway in violation of federal environmental legislation represents a classic example of environmental pollution, Vanderbilt argues, it is at best ambiguous whether the types of harms alleged in the underlying complaints, such as that of a woman who inhales or otherwise ingests asbestos while washing her husband's work clothing, can be said to arise from the release of a pollutant.⁷¹ See, e.g., Complaint of Alan C. Thorp, Sr. The defendants counter that in Yale University v. Cigna Ins. Co. , 224 F.Supp.2d 402 (D. Conn. 2002) (Yale ), the federal District Court interpreted Heyman more broadly to mean that a pollution exclusion clause bars coverage for any harm arising from the inhalation of any substance that satisfies the general dictionary definition of a "pollutant," "irritant," or "contaminant." (Internal quotation marks omitted.) Id., at 421-22. In Yale , as in the present case, the question was whether a pollution exclusion barred coverage for asbestos related damage. Reasoning that "there can be little doubt that" asbestos poses a health threat and that the release of asbestos dust renders a building " 'unfit for use by the introduction of unwholesome or undesirable elements' and/or 'physically impure or unclean,' " the District Court concluded that "[a] straightforward application of Heyman ... to the facts of this case militates against coverage for Yale's asbestos abatement costs." Id., at 422. The federal court further interpreted Heyman as having rejected the conclusion, reached by courts in other jurisdictions, that a pollution exclusion clause in a comprehensive general liability policy bars coverage only for traditional environmental pollution such as oil spills or the dumping of hazardous waste into the natural environment. Id., at 422-23.

Yale , however, was decided prior to our Supreme Court's decision in Allstate Ins. Co. v. Barron , 269 Conn. 394, 848 A.2d 1165 (2004) ( Barron ). In Barron , the Supreme Court concluded; id., at 420-24, 848 A.2d 1165 ; that poisoning resulting from the inhalation of carbon monoxide gas released during a residential fire did not fall under the auspices of a pollution exclusion clause that expressly barred coverage for bodily injury arising from the "discharge, dispersal, release or escape of vapors, fumes ... toxic gasses ... or other irritants, contaminants or pollutants." (Internal quotation marks omitted.) Id., at 420 n.17, 848 A.2d 1165. That conclusion arguably was dicta in Barron ; see id., at 420 n.18, 848 A.2d 1165 ; and, in any event, the court reached its conclusion on the basis of other policy provisions that are not contained in any of the policies at issue in the present case. See id., at 422-23, 848 A.2d 1165 (explaining that policy expressly covered smoke damage and that reasonable policyholder would not expect policy to cover injuries caused by smoke but not by toxic fumes and chemicals that can be components of smoke). Importantly, however, the court also addressed in a footnote the general question of the scope of the pollution exclusion with respect to toxic chemicals. See id., at 421-22 n.19, 848 A.2d 1165. In that footnote, the Supreme Court acknowledged that courts in other jurisdictions have concluded that toxic substances such as carbon monoxide, formaldehyde, and lead paint are not pollutants within the meaning of a pollution exclusion clause because (1) such clauses are intended to exclude coverage only for environmental damage caused by active or intentional industrial polluters, and (2) dispersal of those substances within the confines of a building does not constitute release into the " 'atmosphere,' "; id., at 422 n.19, 848 A.2d 1165 ; as required by the standard pollution exclusion. Id., at 421-22 n.19, 848 A.2d 1165. Having set forth this more limited interpretation of the pollution exclusion, our Supreme Court concluded the discussion by expressly leaving open the question whether that interpretation is the proper one: "[b]ecause we conclude that, under the specific terms of the policy at issue in the present case, a reasonable policyholder would believe that 'smoke' was expressly excepted from the pollution exclusion clause, we need not consider whether a reasonable policyholder would believe that smoke from a house fire or carbon monoxide contained in that smoke would be excluded pollutants in the absence of such an exception." Id., at 422, 848 A.2d 1165 n.19.

In light of this discussion in Barron , we agree with the trial court that Heyman stands only for the limited and uncontroversial proposition that a pollution exclusion clause bars coverage for an oil spill in a public waterway. See Island Associates, Inc. v. Eric Group, Inc. , 894 F.Supp. 200, 203 (W.D. Pa. 1995) (cases applying pollution exclusion to prototypical environmental pollution do not resolve whether exclusion also applies to chemical exposure within confined work-site). We further conclude that our Supreme Court has left open the question whether such clauses apply only to traditional environmental pollution, such as the dumping of hazardous waste, or whether they apply more broadly to circumstances such as the release of asbestos dust and similar toxic industrial products within a building when used as intended. Accordingly, it falls now to us to answer this question.

We begin by looking at the plain language of the standard pollution exclusion clause through the three lenses used by the court in Heyman : the ordinary meaning of the contract language, any technical meaning as expressed in relevant statutes and regulations, and the legal conclusions drawn by our sister courts upon review of similar policy language. Only if we conclude from this tripartite analysis that the language of the standard pollution exclusion is ambiguous as applied to asbestos exposure may we proceed to consider parol evidence, such as the history and purpose of the pollution exclusion.

b

Ordinary Meaning

We first examine the ordinary meaning of the contract terms. See Heyman Associates No. 1 v. Ins. Co. of Pennsylvania , supra, 231 Conn. at 771, 653 A.2d 122. As previously noted, the defendants, in arguing for a broad interpretation of the standard pollution exclusion, rely on our Supreme Court's analysis of a similar pollution exclusion clause in Heyman . The court's analysis of the contract language in that case was succinct, relying on the primary definitions of two key terms in a 1986 dictionary: "The plain language of the absolute pollution exclusion clause ... makes clear that a liquid is an excluded 'pollutant' if it may be characterized as an 'irritant' or 'contaminant.' The dictionary defines 'contaminant' as 'something that contaminates,' and it defines 'contaminate' as 'to soil, stain, corrupt, or infect by contact or association' or 'to render unfit for use by the introduction of unwholesome or undesirable elements.' Webster's Third New International Dictionary (1986). Similarly, the dictionary defines 'pollutant' as 'something that pollutes ... a polluting substance, medium, or agent,' and it defines 'pollute' as 'to ... impair the purity of ... to make physically impure or unclean.' Id.

"There is no question that the introduction of fuel oil into a waterway such as Stamford Harbor 'soils,' 'corrupts,' 'infects,' and/or 'renders unfit for use' the affected water.... Moreover, it cannot seriously be disputed that the introduction of fuel oil served to 'impair the purity of' the water in the harbor.... Thus, an ordinary, lay definition of 'pollutant' includes fuel oil spilled in a waterway such as Stamford Harbor." Heyman Associates No. 1 v. Ins. Co. of Pennsylvania , supra, 231 Conn. at 772-73, 653 A.2d 122. The defendants in the present case contend that friable asbestos dust, which is undisputably a toxic substance, likewise soils and renders impure any air into which it may be released, and thus constitutes a pollutant as defined in Heyman .

The defendants' argument, although facially attractive, falters upon closer scrutiny. As we have explained, the basis for the underlying claim in Heyman -a large oil spill into a public waterway-was a classic case of environmental pollution that fell squarely within any reasonable definition of the policy terms "pollution" and "contamination." Accordingly, there simply was no need for the Supreme Court in that case to conduct a probing analysis of the policy language. By contrast, cases such as the present appeals, in which reasonable minds may differ as to the applicability of the pollution exclusion; see part IV A 1 d of this opinion; require that we conduct a more comprehensive analysis and consider factors and issues that fell beyond the ambit of the Heyman decision. To understand whether the language of the pollution exclusion unambiguously applies to the underlying complaints, we must, for example: (1) look to dictionary definitions from the period when the standard pollution exclusion was initially drafted; (2) evaluate the ordinary meaning of terms such as "pollution" not only in isolation but also within the context of the full pollution exclusion clause and related policy provisions; (3) consider not only whether asbestos dust constitutes a pollutant, contaminant, or irritant, but also whether the circumstances alleged in the underlying complaints constitute the "discharge, dispersal, release or escape [of asbestos] into or upon land, the atmosphere or any watercourse or body of water"; and (4) consider whether adopting the defendants' proposed interpretation of the policy language would yield unreasonable or absurd results. Accordingly, we now turn our attention to the key policy terms.

i

"Pollutants"

Following the lead of our Supreme Court in Heyman , we begin by defining the key term "pollutant," as well as cognate terms such as "pollute" and "pollution." It is well established that "[t]o ascertain the commonly approved usage of a word, it is appropriate to look to the dictionary definition of the term." (Internal quotation marks omitted.) Buell Industries, Inc. v. Greater New York Mutual Ins. Co. , supra, 259 Conn. at 539, 791 A.2d 489. Furthermore, because we seek to discern the intent of the parties at the time the policies were drafted, we refer to dictionaries in print at that time. Id. ; see also State v. Menditto , 315 Conn. 861, 866, 110 A.3d 410 (2015). Although it was not unreasonable for the court in Heyman , in construing insurance policies issued in 1989; see Heyman Associates No. 1 v. Ins. Co. of Pennsylvania , supra, 231 Conn. at 760 n.3, 653 A.2d 122 ; to rely on dictionary definitions from 1986, in the present case we must look to dictionaries published fifteen years prior. It was, after all, in 1970 that the standard pollution exclusion was initially drafted,⁷² and policies issued in the present case began to incorporate that language as early as 1971.

Although it is true that the most generic definition of "pollute" is simply "to make impure or unclean," dictionaries published during the late 1960s and early 1970s-the period during which Congress enacted several of the landmark federal environmental statutes-invariably offered an example, parenthetical, or secondary definition that called to mind what has come to be thought of as traditional environmental pollution, such as the knowing release of exhaust or hazardous waste into public air, water, or soil resources. The following represents a sampling of such definitions:⁷³

• pollutant: "something that pollutes; a polluting substance, medium, or agent < domestic wastes ... are another chief source of pollutants> < the great pollutants are industrial plants and oil burners ...>" Webster's Third New International Dictionary of the English Language Unabridged (1971).

• pollute: "1: to render ceremonially or morally impure ... 2: to make physically impure or unclean: befoul, dirty, taint < pollute a water supply by the introduction of sewage>" Id.

• polluted: "1: made unclean or impure: morally corrupt or defiled: physically tainted < change ... a pure stream into polluted and poisoned ditch>" Id.

• pollution: "2: the action of polluting or the state of being polluted: defilement, desecration, impurity, uncleanness < streams subject to pollution by ... mill wastes> < the dilution of atmospheric pollution ...>" Id.

• pollutant: "that which pollutes: Rivers are full ofpollutants from the factories and cities along their banks ." Random House Dictionary of the English Language (1966).

• pollute: "1. to make foul or unclean; dirty: to pollute the air with smoke ." Id.

• pollutant: "Anything that pollutes; especially, any gaseous, chemical, or organic waste that contaminates air, soil, or water." American Heritage Dictionary of the English Language (1969).

• pollution: "1. The act or process of polluting or the state of being polluted. 2. The contamination of soil, water, or the atmosphere by the discharge of noxious substances." Id.

• "pollute: Destroy the purity or sanctity of; make foul or filthy; contaminate or defile (man's environment) ...." Concise Oxford Dictionary (6th Ed. 1976).

In addition, the Oxford English Dictionary provides the following examples of typical usages from that time period:

• "1970 ... Mercury is now the most dangerous environmental pollutant." 12 Oxford English Dictionary (2d Ed. 1991).

• "1966 ... The absence of sulphur ensures that the products of combustion are non-corrosive [and] do not pollute the atmosphere." Id.

• "1969 ... The danger of 'thermal pollution' is greatest where electric and other power plants return to rivers and streams water that has been heated by between six and 16 degrees Centigrade. This often proves deadly to fish." Id.

• "1970 ... At American universities, pollution has been a student rallying cry for some months now." Id.

It is clear, then, that at that time in our nation's history the principal connotation of the terms "pollutant" and "pollution" was with reference to contamination of the natural environment by industrial and other hazardous wastes. Accordingly, we conclude, consistent with the decisions of a number of our sister courts,⁷⁴ that it is at best ambiguous whether the use of these terms in the standard pollution exclusion clause was intended and would reasonably have been understood to extend beyond those contexts, to scenarios such as the inhalation or ingestion of asbestos dust released in small quantities in an indoor environment during everyday activities such as manufacturing, laundering, or remodeling.

ii

Other "Irritants" and "Contaminants"

Nor are we persuaded that asbestos unambiguously qualifies as "smoke, vapors, soot, fumes, acids, alkalis, toxic chemicals, liquids or gases, waste materials or other irritants [or] contaminants" in the context of the exclusion language to be applied to the allegations of the underlying complaints. Once again, on the most general level there is little doubt that asbestos is a toxic substance that may irritate the lungs and the ingestion of which can be said to contaminate the human body.⁷⁵ To that extent, it plainly falls under the policy language. On the other hand, a strong argument can be made that this list is merely intended to provide examples of the sorts of substances that constitute "pollutants" and, therefore, that our previous determination that asbestos is not unambiguously a pollutant when applied to the allegations of the underlying complaints is dispositive.

We begin by recognizing that, in construing the language of a contract, "[m]eaning is inevitably dependent on context. A word changes meaning when it becomes part of a sentence, the sentence when it becomes part of a paragraph." 2 Restatement (Second), Contracts § 202, comment (d), p. 88 (1981); see also MacKinnon v. Truck Ins. Exchange , 31 Cal.4th 635, 649, 73 P.3d 1205, 3 Cal.Rptr.3d 228 (2003) (courts that construe pollution exclusion as unambiguous commit fallacy of examining key policy terms in isolation, divorced from context); 17A Am. Jur. 2d 362-63, Contracts § 375 (2004) ("A contract must be construed as a whole, and the intention of the parties is to be ascertained from the entire instrument. The contract's meaning must be gathered from the entire context and not from particular words, phrases, or clauses, or from detached or isolated portions of the contract."). In particular we recognize that, "[when] a provision contains two or more words grouped together, we often examine a particular word's relationship to the associated words and phrases to determine its meaning pursuant to the canon of construction noscitur a sociis [a word is known by the company it keeps]." Cantonbury Heights Condominium Assn., Inc. v. Local Land Development, LLC , 273 Conn. 724, 740, 873 A.2d 898 (2005) ; see also Smedley Co. v. Employers Mutual Liability Ins. Co. of Wisconsin , 143 Conn. 510, 514-15, 123 A.2d 755 (1956) (finding ambiguity upon applying noscitur a sociis canon to insurance policy exclusion).

In the present case, when we consider the language of the standard pollution exclusion as a whole, and in the context of neighboring policy provisions, there are at least five reasons to believe that the parties did not intend to exclude coverage for harms inflicted by all toxic chemicals, irritants, and contaminants in the most broadly literal sense of those words. First, as several of our sister courts have recognized, the various substances enumerated in the exclusion, "smoke, vapor, soot, fumes, acids, alkalis, chemicals, and waste-are either products used to operate equipment or machinery or byproducts of the operation of equipment or machinery." Lefrak Organization, Inc. v. Chubb Custom Ins. Co. , 942 F.Supp. 949, 956 (S.D.N.Y. 1996). One plausible interpretation of the policy language, then, is that it was intended to exclude only those harms and injuries resulting from the dross of industrial production. See, e.g., Westchester Fire Ins. Co. v. Pittsburg , 768 F.Supp. 1463, 1470 (D. Kan. 1991) ("[t]he terms 'irritant' and 'contaminant,' however, cannot be read in isolation, but must be construed as substances generally recognized as polluting the environment").

Second, we note that most of the standard pollution exclusion clauses at issue in the present case have titles such as "seepage & pollution endorsement clause" or "pollution and contamination exclusion." The fact that the term "pollution" appears in the title of each exclusion clause and that the parties-consistent with industry practice-refer to such clauses as "pollution exclusions" suggests that pollution represents the primary concern of these provisions.⁷⁶ See Connecticut Ins. Guaranty Assn. v. Drown , supra, 314 Conn. at 204, 101 A.3d 200 (McDonald, J ., dissenting) (title of insurance provision illuminates meaning).⁷⁷

Third, in a number of the relevant policies, the standard pollution exclusion language previously cited is merely the first paragraph of a two paragraph "contamination or pollution" or "seepage & pollution" exclusion. The second paragraph of one of those exclusions reads: "It is further agreed that, if with respect to operations described in this endorsement there is a discharge, dispersal, release or escape of oil or other petroleum substance or derivative (including any oil refuse or oil mixed with wastes) into or upon any watercourse or body of water, the insurance does not apply to bodily injury or property damage arising out of such discharge, dispersal, release or escape whether or not sudden and accidental." In other words, whereas the first paragraph of the standard exclusion provides that sudden and accidental releases of pollutants into the air, water, or land are covered under the policy, this second paragraph carves out an additional exclusion for waterborne oil spills, even accidental incidents of which are exempted from coverage. Pursuant to the noscitur a sociis canon of construction, the fact that the second paragraph of the exclusion is addressed specifically to oil spills is further indication that the standard exclusion is directed toward traditional types of environmental contamination, rather than to routine exposure to potentially harmful building materials such as asbestos. Fourth, in many of the policies at issue in these appeals, the pollution exclusion immediately precedes or follows other exclusions directed toward other types of environmental harms. For example, the pollution exclusion clause in the Fireman's 1976 policy directly follows a "Nuclear Energy Liability Exclusion Endorsement" that bars coverage for personal injuries and destruction caused by the hazardous properties of nuclear materials contained in spent fuel or waste. Similarly, the pollution exclusion clause in a 1978 policy issued by the Puritan Insurance Company, the predecessor of Westport, immediately precedes a clause excluding coverage for property damage arising from the subsidence of land, while the Lloyd's policies from the 1970s place the pollution exclusion on the same policy page as a "Radioactive Contamination Exclusion Clause." Once again, the noscitur a sociis canon of construction suggests that the scope of the pollution exclusion may be limited to these sorts of federally regulated environmental dangers.

Fifth, we agree with those courts that have concluded that a literal interpretation of the list of substances in the standard pollution exclusion language would render the clause so broad as to be meaningless, and would lead to irrational and absurd consequences. Many courts reaching this conclusion cite to the following analysis from the United States Court of Appeals for the Seventh Circuit: "The terms irritant and contaminant, when viewed in isolation, are virtually boundless, for there is virtually no substance or chemical in existence that would not irritate or damage some person or property.... Without some limiting principle, the pollution exclusion clause would extend far beyond its intended scope, and lead to some absurd results. To take but two simple examples, reading the clause broadly would bar coverage for bodily injuries suffered by one who slips and falls on the spilled contents of a bottle of Drano, and for bodily injury caused by an allergic reaction to chlorine [splashed from] a public pool. Although Drano and chlorine are both irritants or contaminants that cause, under certain conditions, bodily injury or property damage, one would not ordinarily characterize these events as pollution." (Citation omitted; internal quotation marks omitted.) Pipefitters Welfare Educational Fund v. Westchester Fire Ins. Co. , 976 F.2d 1037, 1043 (7th Cir. 1992).

We need not resort to hypotheticals, however, to recognize the absurd consequences that would result from a broadly literal reading of the pollution exclusion. In fact, a number of our sister courts have confronted exactly the sort of scenario envisioned by the Seventh Circuit. In Regent Ins. Co. v. Holmes , 835 F.Supp. 579 (D. Kan. 1993), for example, the insurer sought to deny coverage pursuant to a pollution exclusion when a three year old girl accidentally spilled a bottle of the policyholder's carpet-testing acid, resulting in burns to her leg. Id., at 580. Notwithstanding that the provision in question barred coverage for bodily injuries arising out of the release of any pollutant, defined to include irritants and acids, the court held the exclusion inapplicable, reasoning that the policy language in question reasonably could be construed to apply only to releases of toxic chemicals into the natural environment. Id., at 581-82. Similarly, in Mellin v. Northern Security Ins. Co. , 167 N.H. 544, 115 A.3d 799 (2015), the Supreme Court of New Hampshire declined to apply a pollution exclusion to damage resulting from the migration of cat urine odor in the air between two condominiums, despite the fact that such damage was encompassed under a literal reading of the policy language. "Applying these definitions in a purely literal interpretation ... surely stretch[es] the intended meaning of the policy exclusion," the court reasoned, "and could lead to absurd results contrary to any reasonable policyholder's expectations ...." (Citation omitted; internal quotation marks omitted.) Id., at 552, 115 A.3d 799 ; see also MacKinnon v. Truck Ins. Exchange , supra, 31 Cal.4th at 648, 650-56, 3 Cal.Rptr.3d 228, 73 P.3d 1205 (exclusion did not bar coverage for death of tenant exposed to pesticides sprayed to eradicate yellow jackets at her apartment building because few individuals would consider such injuries to arise from "pollution," and literal interpretation of policy language would lead to absurd and overbroad results).

The fact that the standard pollution exclusion bars coverage for damage and injuries resulting from the release or dispersal of both "acids" and "alkalis" is especially noteworthy in this respect, and strongly counsels against a strictly literal interpretation of the enumerated terms. We may take judicial notice of the scientific fact that most liquid solutions are either acidic or alkaline (base) to some extent. Accordingly, if read literally, the pollution exclusion would bar coverage for harms resulting from virtually any liquid spill. Moreover, the fact that pure water is pH neutral, and hence neither acidic nor basic, points to additional absurd results that would flow from a literal reading of the exclusion. For instance, the policy would provide coverage for burns caused by the release of aqueous steam from a pipe, but not burns caused if the same pipe contained chlorine gas (an acid) or ammonia gas (an alkali). Although the parties were certainly free to draft such a policy, one is hard-pressed to envision a plausible rationale for drawing such a distinction. See Suffield Development Associates Ltd. Partnership v. National Loan Investors, L.P. , 97 Conn.App. 541, 560, 905 A.2d 1214 (court should avoid reading contract so as to render it illogical), cert. denied, 280 Conn. 942, 943, 912 A.2d 479 (2006) ; South End Plaza Assn., Inc. v. Cote , 52 Conn.App. 374, 378, 727 A.2d 231 (1999) ("[i]n giving meaning to the language of a contract, we presume that the parties did not intend to create an absurd result" [internal quotation marks omitted] ). For all of these reasons, we conclude that it is at best ambiguous whether asbestos dust is an irritant, contaminant, or other pollutant as defined in the standard pollution exclusion.

iii

"Discharge ... [I]nto or [U]pon [L]and, the Atmosphere or [A]ny [W]atercourse"

As the Supreme Court of Illinois has recognized, the standard pollution exclusion applies only when three elements are satisfied. "To exclude coverage pursuant to the pollution exclusion, the alleged property damage must arise out of (1) some form of discharge or release (2) of a contaminant or pollutant and (3) into or upon land, the atmosphere or any water course or body of water." United States Fidelity & Guaranty Co. v. Wilkin Insulation Co. , 144 Ill.2d 64, 79, 578 N.E.2d 926, 161 Ill.Dec. 280 (1991). Even if we were to assume that friable asbestos dust unambiguously qualifies as an irritant, contaminant, or pollutant, as defined in the standard pollution exclusion, the question would remain whether the inadvertent dissemination of small quantities of the substance inside a building in the course of routine activities such as manufacturing, remodeling, or laundering constitutes "the discharge, dispersal, release or escape of ... [asbestos] into or upon land, the atmosphere or any watercourse or body of water ...."⁷⁸ Several of our sister courts, while concluding (or assuming for the sake of argument) that toxic chemicals such as asbestos do qualify as pollutants or irritants for purposes of the exclusion, nevertheless have found the exclusion language to be ambiguous or not to be implicated with respect to the first and third elements of the exclusion. See, e.g., Island Associates, Inc. v. Eric Group, Inc. , supra, 894 F.Supp. at 203-204 ; Porterfield v. Audubon Indemnity Co. , 856 So.2d 789, 800 (Ala. 2002) ; Essex Ins. Co. v. Avondale Mills, Inc. , 639 So.2d 1339, 1341 (Ala. 1994) ; United States Fidelity & Guaranty Co. v. Wilkin Insulation Co. , supra, at 79, 161 Ill.Dec. 280, 578 N.E.2d 926 ; Board of Regents v. Royal Ins. Co. of America , 517 N.W.2d 888, 892-94 (Minn. 1994) ; Belt Painting Corp. v. TIG Ins. Co. , 100 N.Y.2d 377, 384, 795 N.E.2d 15, 763 N.Y.S.2d 790 (2003) ; Continental Casualty Co. v. Rapid-American Corp. , 80 N.Y.2d 640, 653-54, 609 N.E.2d 506, 593 N.Y.S.2d 966 (1993).⁷⁹

Although a few published opinions have questioned whether the migration of a toxic substance over the distance of a few feet prior to inhalation constitutes "discharge, dispersal, release or escape" as those terms are ordinarily used,⁸⁰ the primary focus of attention has been on whether the air inside a confined space such as a building-the setting for much alleged asbestos exposure-qualifies as "the atmosphere." Once again, we consider the policy terms both individually and when read as a whole.

We understand the underlying actions primarily to allege that personal injuries occurred when asbestos dust became airborne, creating a risk that it would be inhaled or otherwise ingested by individuals working with or otherwise exposed to the substance. Accordingly, the parties appear to agree that the issue is whether such exposure arises from a release into "the atmosphere ...." Looking again to dictionaries in print at about the time the standard pollution exclusion was drafted, two meanings of the term "atmosphere" predominate. The following definition is typical: "atmosphere: 1. The gaseous mass or envelope surrounding a celestial body, especially that surrounding the earth, and retained by the body's gravitational field. 2. The atmosphere or climate in a specific place." American Heritage Dictionary of the English Language, supra. Because the term "atmosphere" reasonably can refer either to the interior air at a particular location or to the earth's natural atmosphere, we conclude that the third element of the pollution exclusion also is ambiguous as applied to the underlying claims. See Continental Casualty Co. v. Rapid-American Corp. , supra, 80 N.Y.2d at 653-54, 593 N.Y.S.2d 966, 609 N.E.2d 506.

We further note that this ambiguity is largely resolved, in favor of Vanderbilt, when we consider the term "atmosphere" in the context of the phrase in which it appears: "into or upon land, the atmosphere or any watercourse or body of water." United States Fidelity & Guaranty Co. v. Wilkin Insulation Co. , supra, 144 Ill.2d at 79, 161 Ill.Dec. 280, 578 N.E.2d 926 ; see Essex Ins. Co. v. Avondale Mills, Inc. , supra, 639 So.2d at 1341-42. There are several reasons to think that the term atmosphere, as used in that phrase, is not intended to be synonymous with the air in a particular place. On the one hand, if the terms "land," "atmosphere," and "water" are construed broadly, then the third element of the pollution exclusion becomes superfluous; any dispersal of a toxic chemical will, broadly speaking, be into the air, land, or water. It is well established, however, that "the law of contract interpretation ... militates against interpreting a contract in a way that renders a provision superfluous." (Internal quotation marks omitted.) Ramirez v. Health Net of the Northeast, Inc. , 285 Conn. 1, 14, 938 A.2d 576 (2008).

On the other hand, if those terms are construed more narrowly, then we are confronted with a conundrum. Why would the contracting parties have chosen to exclude coverage for harms arising from the release of asbestos into all manner of outdoor environments-lawns, fields, rivers, lakes, and the air-and also into the air within a building, but not for claims arising, say, from the direct release of that same asbestos onto a table or floor, swimming pool, article of clothing, or other item of realty or personal property? As we previously have discussed, although the parties are free to adopt such a policy, we should avoid construing the contract language in a manner that has no rational explanation, especially when a more reasonable interpretation is readily available.

Here, as one court has explained, the interpretation that makes the most sense is that "the exclusion is worded broadly to encompass the natural resources of this planet in their natural setting, namely, land, the atmosphere, and bodies of water.... Significantly, the pollution exclusion does not use the generic term 'water' but rather the phrase 'any watercourse or body of water,' a description indicative of water in streams, ponds or lakes. The use of the term 'land,' instead of 'property,' whether real or personal, likewise appears directed at land as a natural resource. And, within this context, the term 'atmosphere,' we think, refers to the ambient air. We are not saying here that air inside a building differs from the air outside, or that the inside and outside air do not intermingle. Rather, within the context of the pollution exclusion, the distinction is not in the air itself but where the air happens to be. When the air supply within a building becomes contaminated, it is harmful to the controlled environment of that building; but the contamination of the air in a building is not harmful to the surrounding natural environment, at least not until it escapes into that environment so as to cause personal injury or property damage ...." Board of Regents v. Royal Ins. Co. of America , supra, 517 N.W.2d at 892-93 ; see also Essex Ins. Co. v. Avondale Mills, Inc. , supra, 639 So.2d at 1342 (phrase "suggest[s] contamination of a broad natural environment rather than the environs of a building").

We agree with those decisions and conclude that a reasonable insured would not expect such claims to be barred by the standard pollution exclusion. At best it is ambiguous whether the underlying allegations of asbestos exposure can reasonably be said to arise from release of asbestos dust "into or upon land, the atmosphere or any watercourse or body of water ...."

c

Environmental Terms of Art

Having reviewed the plain language of the pollution exclusion, we agree with Vanderbilt that (1) the most reasonable interpretation of the contract language, read in context and taken as a whole, is that the plain meaning of the exclusion does not bar coverage for the underlying claims, and (2) at the very least, it is ambiguous whether asbestos dust constitutes an irritant, contaminant, or pollutant as defined in the policies at issue, and also whether its release as alleged in the underlying complaints is into the "atmosphere." In this part of the opinion, we consider an additional source of support for these conclusions, namely, the determination by a number of our sister courts that the language of the pollution exclusion is intended to be understood not according to its ordinary, lay meaning but, instead, as technical language comprised of environmental terms of art.

In Heyman , our Supreme Court indicated that, in determining whether the language of an insurance exclusion is facially ambiguous, it is appropriate to consider not only the use of key terms in everyday parlance, but also whether those terms are components of and defined by a statutory or regulatory regime. See Heyman Associates No. 1 v. Ins. Co. of Pennsylvania , supra, 231 Conn. at 773, 653 A.2d 122. In considering the question presented by the present appeals, many of our sister courts have concluded that the distinct phrasing of the standard pollution exclusion mirrors the language used in federal and state environmental legislation and regulation, and, therefore, that the exclusion should be construed as applicable only in that context. See, e.g., Porterfield v. Audubon Indemnity Co. , supra, 856 So.2d at 795-804 (discussing authorities holding that "terms discharge, dispersal, release, and escape [are] terms of art in environmental law which generally are used with reference to damage or injury caused by improper disposal or containment of hazardous waste" [internal quotation marks omitted] ); MacKinnon v. Truck Ins. Exchange , supra, 31 Cal.4th at 646, 651-53, 3 Cal.Rptr.3d 228, 73 P.3d 1205 (collecting authorities and concluding that language of dispersal, when used in conjunction with term "pollutant," typically references environmental pollution); Richardson v. Nationwide Mutual Ins. Co. , 826 A.2d 310, 324-29 (D.C. 2003) (noting that pollution exclusion is "replete with language used in environmental statutes and regulations" and finding "considerable support in reason and authority" for view that exclusion contains terms of art in environmental law), vacated pursuant to settlement, 844 A.2d 344 (D.C. 2004) ;

Motorists Mutual Ins. Co. v. RSJ, Inc. , 926 S.W.2d 679, 681 (Ky. App. 1996) ("drafters' utilization of environmental law terms of art ... reflects the exclusion's historical objective-avoidance of liability for environmental catastrophes related to intentional industrial pollution"); Belt Painting Corp. v. TIG Ins. Co. , supra, 100 N.Y.2d at 387, 763 N.Y.S.2d 790, 795 N.E.2d 15 ("terms used in the exclusion to describe the method of pollution ... are terms of art in environmental law" [internal quotation marks omitted] ); West American Ins. Co. v. Tufco Flooring East, Inc. , 104 N.C.App. 312, 325-26, 409 S.E.2d 692 (1991) (key terms of exclusion are defined by federal environmental statutes), overruled in part on other grounds by Gaston County Dyeing Machine Co. v. Northfield Ins. Co. , 351 N.C. 293, 303, 524 S.E.2d 558 (2000).

The following represents a brief sampling of the myriad instances in which key language in the standard pollution exclusion is defined or used in federal and state environmental statutes and regulations, as well as in technical dictionaries:⁸¹

• "The term 'pollutant or contaminant' shall include, but not be limited to, any element, substance, compound, or mixture, including disease-causing agents, which after release into the environment and upon exposure, ingestion, inhalation, or assimilation into any organism, either directly from the environment or indirectly by ingestion through food chains, will or may reasonably be anticipated to cause death, disease, behavioral abnormalities, cancer, genetic mutation, physiological malfunctions (including malfunctions in reproduction) or physical deformations, in such organisms or their offspring; except that the term 'pollutant or contaminant' shall not include petroleum ...." 42 U.S.C. § 9601 (33).

• "The term 'release' means any spilling, leaking, pumping, pouring, emitting, emptying, discharging, injecting, escaping, leaching, dumping, or disposing into the environment (including the abandonment or discarding of barrels, containers, and other closed receptacles containing any hazardous substance or pollutant or contaminant) ...." 42 U.S.C. § 9601 (22).

• "Discharge when used without qualification means the 'discharge of a pollutant.' Discharge of a pollutant means: (a) Any addition of any 'pollutant' or combination of pollutants to 'waters of the United States' from any 'point source,' or (b) Any addition of any pollutant or combination of pollutants to the waters of the 'contiguous zone' or the ocean ...." (Emphasis omitted.) 40 C.F.R. § 122.2.

• " 'Air pollution' means the presence in the ambient air of one or more air pollutants or any combination thereof in such quantities and of such characteristics and duration as to be, or likely to be, injurious to public welfare or the environment, to the health of human, plant or animal life, or to property, or as unreasonably to interfere with the enjoyment of life and property." Regs., Conn. State Agencies § 22a-174-1 (6). " 'Ambient air' means that portion of the atmosphere, external to buildings, to which the general public has access." Regs., Conn. State Agencies § 22a-174-1 (9).

• " 'Emission' means the release or discharge of an air pollutant into the ambient air from any source." Regs., Conn. State Agencies § 22a-174-1 (38).

• " 'Release' means any discharge, as defined in 40 [C.F.R. §] 260.10, or any migration of substances from a waste or combination of wastes into the environment." Regs., Conn. State Agencies § 22a-449 (c)-100 (c) (25).

• " 'Discharge' means the emission of any water, substance or material into waters of the state whether or not such substance causes pollution ...." General Statutes § 22a-38 (10).

• " 'Watercourses' means rivers, streams, brooks, waterways, lakes, ponds, marshes, swamps, bogs and all other bodies of water, natural or artificial, vernal or intermittent, public or private, which are contained within, flow through or border upon this state or any portion thereof ...." General Statutes § 22a-38 (16).

• "Contaminant: In the natural environmental context, a substance introduced into a natural ecosystem by human agency. It may change the system in some way but, unlike a pollutant, a contaminant does not necessarily impair or harm organisms." Encyclopaedic Dictionary of Environmental Change (2003) p. 119.

• "Pollutant: A substance introduced into a natural system by human agency and which impairs the system or harms organisms. A contaminant becomes a pollutant when there is damage or adverse effects." (Emphasis in original.) Id., p. 496.

• "Dispersal: The breaking up, spreading out, or distribution of some material released from a concentrated source to a more diffuse distribution within the environment." Facts on File Dictionary of Environmental Science (3d Ed. 2007) p. 125.

• "Release: A spill, leak, escape, or loss of a regulated chemical agent into the environment, including air, water, or land." Id., p. 355.

Particularly noteworthy is the fact that a New York statute enacted in 1971, the same year that the exclusion began to appear in the policies at issue in these appeals, contains language almost identical to that of the exclusion. Chapter 765 of the 1971 edition of McKinney's Session Laws of New York, was entitled, "An Act to amend the insurance law, in relation to prohibiting coverage against environmental pollution." (Emphasis added.) It provided in relevant part: "Policies issued to commercial or industrial enterprises providing insurance against the legal liabilities specified in this subdivision shall expressly exclude therefrom liability arising out of pollution or contamination caused by the discharge, dispersal, release or escape of any pollutants, irritants or contaminants into or upon land, the atmosphere or any water course or body of water unless such discharge, dispersal, release or escape is sudden and accidental." (Emphasis added.) 1971 McKinney's Session Laws of New York, c. 765, § 1 (13). The accompanying memorandum of the state executive department explained that the purpose of the act was to "prohibit commercial or industrial enterprises from buying insurance to protect themselves against liabilities arising out of their pollution of the environment." Memorandum of State Executive Department, 1971 McKinney's Session Laws of New York, c. 765, p. 2485. In his signing statement, Governor Nelson A. Rockefeller further explained that the act was part of his environmental program and was intended to supplement the state's "stringent standards to prohibit despoiling our environment through the discharge of noxious substances into the water and air." Signing Statement of Governor Nelson A. Rockefeller, June 25, 1971, 1971 McKinney's Session Laws of New York, c. 765, p. 2633. He elaborated that "[m]any insurance companies have voluntarily initiated action to protect the environment by refusing to insure against liability arising out of environmental pollution," and that the new law would further that purpose by ensuring that the availability of pollution insurance did not undermine this public policy. Id.

Similarly, 49 C.F.R. § 387, a 1981 federal regulation mandating that transporters of "hazardous materials, hazardous substances, or hazardous wastes"; 49 C.F.R. § 387.3 (b) ; be able to demonstrate sufficient financial responsibility for the environmental consequences of any accidents resulting from their negligent operation, provides that "[e]nvironmental restoration means restitution for the loss, damage, or destruction of natural resources arising out of the accidental discharge, dispersal, release or escape into or upon the land, atmosphere, watercourse, or body of water of any commodity transported by a motor carrier...." (Emphasis added.) 49 C.F.R. § 387.5.

These authorities lend strong support to Vanderbilt's argument that the policy language, when read as a whole, is intended to exclude coverage only for traditional environmental pollution, such as the intentional disposal or negligent release of industrial and other hazardous waste into the public air, land, or water resources. As the District of Columbia Court of Appeals concluded after reviewing the relevant authorities, "the similarity between the language of the pollution exclusion and the terminology of environmental statutes, regulations, and judicial decisions is sufficiently striking to render a coincidence improbable." Richardson v. Nationwide Mutual Ins. Co. , supra, 826 A.2d at 328.

d

Decisions in Other Jurisdictions

Following the guidance of our Supreme Court in Heyman , we next consider how courts in other jurisdictions have interpreted the pollution exclusion. Although the question is one of first impression for Connecticut's appellate courts, dozens of other jurisdictions have confronted the issue of whether a pollution exclusion clause-in most instances the standard one⁸² -bars coverage for claims arising from exposure to toxic substances that are (1) released indoors or in a confined space, (2) encountered when used as intended in the ordinary course of business, and/or (3) otherwise alleged to have caused harms not associated with traditional environmental pollution. Generalizing about the decisions of those courts is complicated by the fact that policy language varies from case to case and the underlying complaints are highly fact specific. Nevertheless, it safely can be said that our sister courts are sharply divided on this question and that no clear consensus has emerged.⁸³ As one court has explained, "[o]ur review and analysis of the entire body of existing precedent reveals that there exists not just a split of authority, but an absolute fragmentation of authority. Cases may be found for and against every issue any litigant has ever raised, and often the cases reaching the same conclusion as to a particular issue do so on the basis of differing, and sometimes inconsistent, rationales." Porterfield v. Audubon Indemnity Co. , supra, 856 So.2d at 800 ; see also Nationwide Mutual Ins. Co. v. Richardson , 270 F.3d 948, 954 (D.C. Cir. 2001) ("[c]ourts across the nation are hopelessly divided"); Bituminous Casualty Corp. v. Sand Livestock Systems, Inc., 728 N.W.2d 216, 220 (Iowa 2007) (recognizing " 'a dizzying array of results' ").

Our own research bears out these conclusions. The relatively small number of state appellate courts and federal district courts to have considered the pollution exclusion with reference to asbestos contamination in particular are more or less evenly divided. Compare Great Northern Ins. Co. v. Benjamin Franklin Federal Savings & Loan Assn. , 793 F.Supp. 259 (D. Or. 1990), aff'd, Docket No. 90-35654, 953 F.2d 1387, 1992 WL 16749 (9th Cir. January 31, 1992) (decision without published opinion, 953 F.2d 1387 [9th Cir. 1992] ), American States Ins. Co. v. Zippro Construction Co. , 216 Ga.App. 499, 455 S.E.2d 133 (1995), petition for cert. dismissed sub nom. Conley v. American States Ins. Co. , 1995 Ga. LEXIS 598 (Ga. May 15, 1995), Cincinnati Ins. Co. v. German St. Vincent Orphan Assn., Inc. , 54 S.W.3d 661 (Mo. App. 2001), and Selm v. American States Ins. Co. , Docket No. C-010057, 2001 WL 1103509 (Ohio App. September 21, 2001) (exclusion applies to release of asbestos containing material), with In re Asbestos Products Liability Litigation (VI) , Docket No. Civ. A. 96-968, 1997 WL 539916 (E.D. La. September 2, 1997), Essex Ins. Co. v. Avondale Mills, Inc. , supra, 639 So.2d 1339, United States Fidelity & Guaranty Co. v. Wilkin Insulation Co. , supra, 144 Ill.2d at 80, 161 Ill.Dec. 280, 578 N.E.2d 926, and Continental Casualty Co. v. Rapid-American Corp. , supra, 80 N.Y.2d at 654-55, 593 N.Y.S.2d 966, 609 N.E.2d 506 (exclusion does not apply); see also Board of Regents v. Royal Ins. Co. of America , supra, 517 N.W.2d at 888 (outcome depends on exact wording of exclusion). Looking to the broader universe of cases addressing chemicals and chemical products such as lead paint, carbon monoxide, insecticides, industrial chemicals, and other toxic substances that are either dispersed indoors or present health risks when used as intended, a narrow majority of our sister states appear to have concluded that the pollution exclusion applies so as to bar coverage only in the context of traditional environmental pollution. See MacKinnon v. Truck Ins. Exchange , supra, 31 Cal.4th at 642 n.2, 3 Cal.Rptr.3d 228, 73 P.3d 1205.

Mt. McKinley contends, and the court in Yale concluded, that decisions from other jurisdictions that limit the applicability of the pollution exclusion to traditional environmental contamination do not constitute persuasive authority in Connecticut because those decisions rely on extrinsic evidence, such as the drafting history and purpose of the pollution exclusion, without first determining that the plain language of the exclusion is ambiguous. See Yale University v. Cigna Ins. Co. , supra, 224 F.Supp.2d at 423 (noting that, "[u]nder Connecticut law ... any ambiguity in a contract must emanate from the language used in the contract rather than from one party's subjective perception of the terms" [internal quotation marks omitted] ). We disagree.

Although it is true that a number of the cases from other jurisdictions that interpret the pollution exclusion narrowly consider the history of and intent behind the exclusion, most do so only after first reviewing the policy language and finding it to be independently ambiguous. See, e.g., Continental Casualty Co. v. Rapid-American Corp. , supra, 80 N.Y.2d at 652, 593 N.Y.S.2d 966, 609 N.E.2d 506 ; West American Ins. Co. v. Tufco Flooring East, Inc. , supra, 104 N.C.App. at 323, 409 S.E.2d 692. In addition, many of the cases that conclude that the exclusion does not bar coverage do so solely on the basis of a plain language analysis, without ever considering extrinsic evidence of drafting history or intent. See, e.g., In re Asbestos Products Liability Litigation (VI) , supra, 1997 WL 539916 ; Lefrak Organization, Inc. v. Chubb Custom Ins. Co. , supra, 942 F.Supp. at 949 ; Regent Ins. Co. v. Holmes , supra, 835 F.Supp. at 579 ; Essex Ins. Co. v. Avondale Mills, Inc. , supra, 639 So.2d 1339 ; Mellin v. Northern Security Ins. Co. , supra, 167 N.H. at 544, 115 A.3d 799. Moreover, we agree with the trial court that the cases that adopt a narrow reading of the standard pollution exclusion tend to be more convincing than those that do not, insofar as the former interpret key terms such as "contaminant" and "irritant" in context and with reference to whether the sorts of claims raised in the underlying complaints constitute "release ... into or upon land, the atmosphere or any watercourse or body of water ...." (Emphasis omitted.) Accordingly, we find the cases supporting Vanderbilt's interpretation of the pollution exclusion to be more persuasive.

e

Drafting History and Purpose

To summarize, our review of the policy language tends to support Vanderbilt's position that a reasonable insured would not expect the standard pollution exclusion to apply outside the context of traditional environmental pollution. To the extent that the language is ambiguous, as the trial court concluded, our review of extrinsic evidence of the drafting history and purpose of the standard pollution exclusion also favors Vanderbilt's interpretation of the policy language.

The drafting history and original purpose of the standard pollution exclusion have been extensively reviewed by other courts and commentators, virtually all of whom have concluded that its initial intent was to preclude coverage only for cleanup costs and other liabilities associated with intentional environmental pollution.⁸⁴ As the New Jersey Supreme Court explained, "[comprehensive general liability] policies prior to 1966 afforded liability coverage for bodily injury and property damage caused by accident, the term accident being undefined in the standard policy. Courts generally construed the term accident to encompass ongoing events that inflicted injury over an extended period provided that the injury was unexpected and unintended from the insured's standpoint.

"In 1966 the insurance industry revised its standard-form [comprehensive general liability] policy to afford coverage based on an occurrence, which the policy defined as an accident, including injurious exposure to conditions, which results, during the policy period, in bodily injury or property damage that was neither expected nor intended from the standpoint of the insured.... The 1966 revision of the [comprehensive general liability] policy was generally understood to cover pollution liability that arose from gradual losses ....

"Foreseeing an impending increase in claims for environmentally-related losses, and cognizant of the broadened coverage for pollution damage provided by the occurrence-based, [comprehensive general liability] policy, the insurance industry drafting organizations began in 1970 the process of drafting and securing regulatory approval for the standard pollution-exclusion clause. The insurer[s'] primary concern was that the occurrence-based policies, drafted before large scale industrial pollution attracted wide public attention, seemed tailor-made to extend coverage to most pollution situations.... Commentators attribute the insurance industry's increased concern about pollution claims to environmental catastrophes that occurred during the 1960s. Pollution claims burst on the insurance scene following the Torrey Canyon disaster and the Santa Barbara off-shore drilling oil spills in 1969.... Other commentators observe that the insurance industry, concerned about public reaction to environmental pollution, desired to clarify and publicize its position that [comprehensive general liability] policies did not indemnify knowing polluters.... Consistent with that objective, the [p]resident of [one insurance company] announced [that] ...

"We will no longer insure the company which knowingly dumps its wastes. In our opinion, such repeated actions-especially in violation of specific laws-are not insurable exposures. Moreover, we are inclined to think that any attempt to provide such insurance might well be contrary to public policy....

"The end-product of the [Insurance Rating Board's] drafting effort was the standard pollution-exclusion clause .... According to one member of the drafting committee, the pollution-exclusion clause allowed the underwriters to perform their traditional function as insurers of the unexpected event or happening and yet ... [did] not allow an insured to seek protection from his liability insurers if he knowingly pollute[d].... The New York State legislature apparently shared that view of the pollution-exclusion clause's purpose, enacting in 1971 a statute requiring policies issued to commercial or industrial enterprises to include the standard form pollution-exclusion clause ... and offering this explanation for its adoption:

"For example, a polluting corporation might continue to pollute the environment if it could buy protection from potential liability for only the small cost of an annual insurance premium, whereas, it might stop polluting, if it had to risk bearing itself the full penalty for violating the law....

"After industry approval, the [Insurance Rating Board] and the Mutual Insurance Rating Bureau ... sought state regulatory approval to add the pollution-exclusion clause as an endorsement to standard [comprehensive general liability] policies, apparently submitting to most if not all states in which approval was sought a standard explanatory memorandum that read in part as follows:

"Coverage for pollution or contamination is not provided in most cases under present policies because the damages can be said to be expected or intended and thus are excluded by the definition of occurrence. The above exclusion clarifies this situation so as to avoid any question of intent." (Citations omitted; internal quotation marks omitted.) Morton International, Inc. v. General Accident Ins. Co. of America , 134 N.J. 1, 31-36, 629 A.2d 831 (1993), cert. denied sub nom. Ins. Co. of North America v. Morton International, Inc. , 512 U.S. 1245, 114 S.Ct. 2764, 129 L.Ed.2d 878 (1994).

We agree with our sister courts that this drafting history makes abundantly clear that the insurance industry drafted the pollution exclusion in 1970⁸⁵ to address new liabilities that had arisen in conjunction with the advent of the modern environmental regulatory system in the 1960s, and that the exclusion was intended to bar coverage only for liabilities arising out of traditional environmental pollution such as the intentional dumping of hazardous waste and other toxic materials into the natural environment. The clause was never intended to apply to situations in which a commercial or industrial product is discovered to pose health threats to individuals who manufacture, apply, or are otherwise exposed to it in the ordinary course of business. Accordingly, having reviewed the relevant drafting history, we remain persuaded that a reasonable insured would not expect the standard pollution exclusion to bar coverage for claims of asbestos related disease that do not qualify as traditional environmental pollution.

f

Conclusion

In conclusion, we agree with the trial court that the standard pollution exclusions do not, as a general matter, bar coverage for the underlying claims. To the extent that certain underlying actions may allege traditional environmental contamination, for example, that the outdoor dumping of silica waste permitted asbestos fibers to become airborne and disperse onto neighboring properties or into the natural environment, we understand that the trial court intends to make such factual determinations during a subsequent stage of the proceedings.

2

Nonstandard Pollution Exclusions

Having concluded that the standard pollution exclusion does not, as a general matter, bar coverage for the underlying claims, we now turn our attention to the nonstandard exclusions contained in some of the defendants' policies. These are of three general types.

First, in keeping with industry practices, certain of the defendants beginning in 1985 issued what have come to be known as absolute pollution exclusions. Absolute exclusions omit from the standard exclusion the exception that "this exclusion does not apply if such discharge, dispersal, release or escape is sudden and accidental."⁸⁶

None of the defendants contend that this omission is relevant to the question of whether the pollution exclusion applies outside of the traditional environmental context.

Second, certain policies make minor changes to the list of substances to which the pollution exclusion applies. After 1979, for example, rather than including a separate exclusion provision governing oil and other petroleum substances, Gibraltar added those substances to the list of pollutants contained in the standard pollution exclusion. Its policies provide: "This policy shall not apply to ... smoke, vapors, soot, fumes, acids, alkalis, oil or other petroleum substance [s ], toxic chemicals, liquids or gases, waste materials or other irritants, contaminants or pollutants ...." (Emphasis added.) Again, none of the defendants contend that this omission is relevant to the question before us.

Third, the 1985 policies issued by National Casualty and Lloyd's contain exclusions that, while incorporating some language from the standard exclusion, make substantial material additions to and deletions from the standard exclusion language. National Casualty identifies its pollution clause as a "Total Pollution Exclusion"⁸⁷ whereas Lloyd's is dubbed an "Industries, Seepage, Pollution and Contamination Exclusion."⁸⁸ The London insurers, alone among the defendants, maintain that their policy language differs materially from the standard exclusion. Specifically, they contend that the trial court improperly concluded that the term "contamination" in the Lloyd's 1985 policy is ambiguous with respect to alleged asbestos exposure. We disagree.

For the most part, our analysis in the preceding parts of this opinion applies with equal force to these nonstandard exclusions.

With respect to the Lloyd's 1985 policy in particular, we reiterate that "pollution" frequently is defined as more or less synonymous with "contamination," both in dictionaries; see, e.g., American Heritage Dictionary of the English Language, supra; and in environmental regulations. See, e.g., 42 U.S.C. § 9601 (33) ( "[t]he term 'pollutant or contaminant' shall include ..."). Moreover, technical dictionaries indicate that "contamination," like "pollution," can be an environmental term of art. See, e.g., Encyclopaedic Dictionary of Environmental Change, supra, p. 119 ("Contaminant:

In the natural environmental context, a substance introduced into a natural ecosystem by human agency. It may change the system in some way but, unlike a pollutant, a contaminant does not necessarily impair or harm organisms."). We note as well that "seepage," the other term used in the Lloyd's exclusion, also can be a term of art used with reference to the release and remediation of environmental contaminants. See Regs., Conn. State Agencies § 22a-133k-1 (53) (" '[r]elease' means any discharge, spillage, uncontrolled loss, seepage , filtration, leakage, injection, escape, dumping, pumping, pouring, emitting, emptying, or disposal of a substance" [emphasis added] ). Accordingly, we have no difficulty concluding that it is at least ambiguous whether the term "contamination," both taken alone and, especially, in the context of the phrase "seepage, pollution or contamination," applies so as to bar coverage for the underlying claims. Although we have no specific information as to the drafting history of the Lloyd's provision, we agree with the trial court's conclusion that, in the absence of extrinsic evidence, we must construe this ambiguous provision in favor of the insured. We therefore conclude that the trial court properly determined that none of the pollution exclusion clauses at issue in this case apply so as to bar coverage outside the context of traditional environmental pollution.

B

Occupational Disease Exclusions

We next turn our attention to clauses in certain of Vanderbilt's secondary insurance policies that exclude coverage for occupational disease. Addressing a question of first impression not only in Connecticut but also nationally,⁸⁹ the trial court concluded that those clauses bar coverage only for occupational disease claims brought by a policyholder's own employees and that the exclusions do not apply to complainants who developed occupational disease while using the policyholder's products in the course of working for another employer. On appeal, National Casualty contends, and we agree, that the trial court construed the occupational disease exclusions too narrowly, and that they unambiguously apply so as to bar coverage for any underlying actions whose allegations meet the standard definition of occupational disease. Although the definition of "occupational disease" may be derived from workers' compensation law, it does not follow that the term applies only to workers' compensation claims brought against one's own employer.

1

Facts

The following additional facts as found by the trial court, undisputed insurance policy language gleaned from the record, and procedural history are relevant to our disposition of this claim. At trial, several of Vanderbilt's secondary insurers either sought declaratory judgments determining or raised special defenses or claims alleging that occupational disease exclusions in their policies precluded coverage for some of the underlying actions. Two versions of the occupational disease exclusion are at issue.⁹⁰

The first policy at issue, Lloyd's policy number 77/18503/1/PNB21250D, was in effect from May 17, 1977 through March 3, 1979. The policy contains an endorsement clause stating in relevant part that "this policy shall not apply ... to personal injury (fatal or nonfatal) by occupational disease." Several other defendants issued secondary policies following form to the Lloyd's policy.⁹¹

The second policy at issue, Pacific policy number XMO017535 (NCA15), was in effect from March 3, 1985 through March 3, 1986. It contains the following endorsement clause: "This policy does not apply to any liability arising out of: Occupational Disease." National Casualty, which has taken the lead in challenging the trial court's rulings regarding the occupational disease exclusions, issued an excess policy, number XU000233, which follows form to the Pacific policy. Lloyd's also issued an excess policy that follows form to the Pacific policy. None of the relevant policies defines the term "occupational disease."

In addition to these occupational disease exclusions, the Lloyd's and Pacific policies contain employers' liability exclusions. The Lloyd's policy provides that "this policy shall not apply ... to the liability of employees." The Pacific policy provides that "[t]his policy does not apply to personal injury to any employee of the insured arising out of and in the course of his employment by the insured or to any obligation of the insured to indemnify another because of damages arising out of such injury." In addition, National Casualty's excess policy, while following form to the Pacific policy, also includes its own "employers liability exclusion," which is somewhat broader than the one in the Pacific policy. It provides in relevant part: "[T]his policy shall not apply to any liability for bodily injury, sickness, disease, disability or shock, including death at any time resulting therefrom ... sustained by any employee of the insured and arising out of and in the course of his employment by the insured." Last, both the Lloyd's and Pacific policies contain exclusions for obligations for which the insured may be held liable under workers' compensation, unemployment compensation, or disability benefits laws.

To facilitate the trial court's resolution of the issue, the parties stipulated during the second phase of the trial that none of the claimants in the underlying actions are or ever were Vanderbilt employees. The parties further stipulated that the underlying complaints fall into three categories: those that allege (1) exposure to Vanderbilt products solely through the workplace of another employer, (2) exposure both in and outside the workplace, and (3) exposure solely outside the workplace. Accordingly, if the occupational disease exclusions do apply to nonemployees of Vanderbilt, they likely will bar coverage for some but not all of the underlying complaints during the relevant policy years.⁹²

In its Phase II decision, the trial court concluded that the occupational disease exclusions apply only to claims brought by Vanderbilt's own employees.

Because the policies themselves do not define the term "occupational disease," the court looked to the Workers' Compensation Act (act), General Statutes § 31-275 et seq., for a definition of the term. Section 31-275 (15) provides that " '[o]ccupational disease' includes any disease peculiar to the occupation in which the employee was engaged and due to causes in excess of the ordinary hazards of employment as such, and includes any disease due to or attributable to exposure to or contact with any radioactive material by an employee in the course of his employment." The trial court concluded that the term, as defined in the statute, was unambiguous, and that it applied solely to employees of the insured. The court rejected the defendants' argument that such a construction would render the occupational disease exclusion superfluous, insofar as the employers' liability exclusions in the policies already preclude coverage for any claims of workplace injury or disease by employees of the policyholder. The court reasoned that the act draws a distinction between occupational diseases; General Statutes § 31-275 (15) ; and " '[p]ersonal injur[ies]' "; General Statutes § 31-275 (16) ; and that the policies at issue incorporate that distinction-whereas the occupational disease exclusion applies to employees of an insured who allege occupational diseases, the employers' liability exclusion applies to employees who allege that they have suffered sudden personal injuries while on the job.

Because the court agreed with Vanderbilt that the occupational disease exclusions do not apply to any of the underlying claims, the court did not address Vanderbilt's alternative arguments that (1) in the event that the policy language is determined to be ambiguous, the exclusions should be construed in favor of the insured pursuant to the doctrine of contra proferentem, and (2) certain of the defendants have waived their right to invoke the exclusions.

2

Analysis

Although the trial court's reasoning is not entirely clear, the court appears to have assumed that the term "occupational disease," which is not defined in the policies, is a legal term of art that derives its meaning from Connecticut's workers' compensation laws. The court further assumed that, if the phrase is a term of art peculiar to workers' compensation law, then, because workers' compensation law governs only workers' claims against their own employers, it necessarily follows that the policy exclusions also apply solely to such claims. We disagree with both steps of the court's analysis.

The following principles guide our resolution of this claim. "In construing the terms of a [contract], the court is required to give the document's language its common and generally accepted, or plain and ordinary meaning, unless a technical or special meaning is clearly intended." (Internal quotation marks omitted.) Keeper's, Inc. v. ATGCKG Realestate, LLC , 146 Conn.App. 789, 798, 80 A.3d 88 (2013), cert. denied, 311 Conn. 913, 84 A.3d 881 (2014). "A phrase is a legal term of art if it has acquired a peculiar and appropriate meaning in the law requiring it to be construed and understood accordingly." (Internal quotation marks omitted.) Id., at 799, 80 A.3d 88. "If the language used in a [contract] is technical or constitutes terms of art, the general rule is that such language is to be given its common technical meaning .... Words with a fixed legal or judicially settled meaning must be presumed to have been used in that sense." (Citation omitted; internal quotation marks omitted.) Id., at 798-99, 80 A.3d 88 ; see also 2 Restatement (Second), supra, § 202 (3) (b) (terms of art are to be given their technical meaning when used in contractual transaction within their technical field).

Although it is true that a term that is not defined in a contract may be a term of art that is definable by reference to a governing statute or regulation; see Schmidt v. O.K. Baking Co. , 90 Conn. 217, 220, 96 A. 963 (1916) ; there is no reason to assume that the parties to a contract intended to incorporate a definition provided in an unrelated statute. Rather, because "[i]t is a fundamental principle of contract law that the ... terms of a contract are to be determined from the intent of the parties"; (internal quotation marks omitted) Auto Glass Express, Inc. v. Hanover Ins. Co. , supra, 293 Conn. at 225, 975 A.2d 1266 ; "the mutual understanding of the parties [ordinarily] prevails even [when] the contractual term has been defined differently by statute or administrative regulation." 2 Restatement (Second), supra, § 201, comment (c), p. 84; see also id., illustration (3), pp. 84-85 (when contractual term is undefined, prevailing trade use trumps statutory definition); id., § 202, comment (f), p. 90 (technical phrases may be used in contract in different, nontechnical senses).

Applying these principles to the present case, we begin by observing that the plain language of the occupational disease exclusions is stated in broad, general terms, and nowhere indicates that coverage is barred only for claims brought by a policyholder's own employees. The Pacific policy, for example, provides simply that "[t]his policy does not apply to any liability arising out of: Occupational Disease." (Emphasis added.) The onus is thus on Vanderbilt to establish that, notwithstanding the plain language of the policy, the occupational disease exclusions contain some latent ambiguity or implicitly apply only to employee lawsuits.

Vanderbilt attempts to meet this burden by establishing that "occupational disease" is a term of art peculiar to Connecticut workers' compensation law. That theory hits an immediate snag, however, insofar as there is nothing in the record to suggest that the exclusions at issue were included only in insurance policies sold in Connecticut. In fact, when the National Casualty policy was issued, it listed an address for Vanderbilt in Lawrence, New York. Notably, that policy contains a "New York Amendatory Endorsement," and notice stamps on each page of the policy indicate that the policy qualified for "New York State Free Trade Zone" exemptions from New York's insurance regulations,⁹³ suggesting that the parties may have understood the agreement to be governed by New York law.⁹⁴ Accordingly, it is not apparent to us that the parties intended or expected that undefined terms would be defined with reference to Connecticut statutes.

There also is no indication on the face of the policies that the parties intended that the phrase "occupational disease" would be construed as a term of art of workers' compensation law. Although the policies do contain brief references to "obligation[s] ... under any workmen's compensation, unemployment compensation or disability benefits law," those references are located in the main, boilerplate portion of the policy, among other standard form exclusions. The occupational disease exclusion, by contrast, is contained in a separate endorsement that is printed in a distinct typeface and indicates that it was prepared as an addendum, "subsequent to the preparation of the policy." Accordingly, we see no reason to conclude that the parties, who entered into a comprehensive general liability insurance agreement, intended that the policy terms would be construed according to workers' compensation law, rather than according to ordinary usage.

Vanderbilt's primary argument in response is that "the term 'occupational disease' is so interwoven with the concept of workers' compensation and other claims by an employee against his employer as to be meaningless outside of that particular context." In other words, Vanderbilt's position is that "occupational disease" is a term of art that is a creature of workers' compensation law and that has no ordinary meaning outside of that arena. Vanderbilt claims to find support for this theory in sources such as Black's Law Dictionary, supra, and H. Rubin, Dictionary of Insurance Terms (5th Ed. 2008), both of which define occupational disease with reference to workers' compensation law.

We agree that the term "occupational disease" is frequently used and has obtained a peculiar meaning in the context of workers' compensation law. In Ricigliano v. Ideal Forging Corp. , supra, 280 Conn. at 723, 912 A.2d 462, for instance, our Supreme Court traced the origin of the phrase, " 'first manifestation of a symptom of the occupational disease,' " to a 1927 amendment to our state's workers' compensation laws. Id., at 732, 912 A.2d 462. We also have no cause to question the conclusion of the trial court that § 31-275 (15) provides a reasonable definition of the term "occupational disease." What is at issue in the present dispute, however, is not the meaning of that phrase but, rather, its application . Vanderbilt contends that coverage is barred only when a complainant sues his or her employer for " 'disease[s] peculiar to the occupation in which the employee was engaged,' " whereas National Casualty contends that the preclusion applies when a complainant sues any policyholder for such a disease. Nothing in § 31-275 (15) itself purports to address that question.

Distilled to its essence, Vanderbilt's argument is that workers' compensation programs represented the sole or primary use of the term "occupational disease" at the time the relevant policies were drafted. We disagree. Rather, our research reveals that, between the late 1970s and mid-1980s, "occupational disease" had a common and ordinary meaning within the legal and insurance fields. In 1979, for example, a note was published in the Harvard Law Review entitled " Compensating Victims of Occupational Disease," 93 Harv. L. Rev. 916 (1979). The note explained that, at that time, "relief for the victims of occupational diseases [could] come from workers' compensation or the tort system ." (Emphasis added; footnote omitted.) Id., 916. Observing that "[p]roducts liability actions filed by those contracting occupational diseases have proliferated in recent years," the note explained that individuals who were barred by workers' compensation laws from suing their employers were instead "su[ing] the manufacturer or seller of a product used in the workplace if that product caused the illness." Id., 926. Notably, the primary example that the note gave of this proliferation of private occupational disease litigation as an alternative to workers' compensation remedies was that "a worker exposed to asbestos while employed by an installer of insulation can sue the company that supplied the asbestos to his employer." Id.

Other legal sources published at the time of drafting likewise evidence a concern or recognition that the growing prevalence of asbestos related occupational diseases was giving rise to an increase in private lawsuits against manufacturers, outside of the workers' compensation arena.⁹⁵ Moreover, although Vanderbilt is correct that certain contemporary reference sources define "occupational disease" with respect to workers' compensation; see, e.g., Black's Law Dictionary, supra, p. 1184 (stating that "[e]mployees who suffer from occupational diseases are eligible for workers' compensation"); H. Rubin, supra, p. 346 (stating that "[c]overage for [occupational disease] is found under workers compensation insurance"); other sources define the phrase generally, without reference to workers' compensation. See, e.g., 10 Oxford English Dictionary (2d Ed. 1991) p. 682 (defining "occupational disease" as "[a] disease to which a particular occupation renders a person especially liable"); see also C. Crobaugh, Handbook of Insurance (1931) pp. 929-30 (discussing occupational disease problem and noting that, at time of writing, such diseases were not covered by workers' compensation programs in most states). Still other sources define the phrase with specificity, yet contain no reference to workers' compensation. See, e.g., Ballentine's Law Dictionary, supra, p. 879 (defining "occupational disease" as "[a] disease which develops gradually and imperceptibly as a result of engaging in a particular employment and is generally known and understood to be a usual and natural incident or hazard of such employment.... A disease caused by or especially incident to a particular employment.... Something other than an accidental injury. But none the less a personal injury, the injury being regarded as sustained when the employee becomes unable to work." [Citations omitted.] ).

Accordingly, our review of the use of the term "occupational disease" at the time the relevant policies were drafted does not persuade us that the parties intended to use the term in a limited, technical manner. Rather, the most reasonable reading of the policy language, particularly in light of the cited legal scholarship from that era, is that the insurance industry was concerned over the emerging proliferation of private litigation by workers who, having developed long latency diseases after exposure to asbestos and other alleged industrial toxins, sought to circumvent the workers' compensation system and sue manufacturers of those products.

We also agree with National Casualty that reading an implied restriction into the occupational disease exclusions would violate several recognized canons of contract construction. It is well established, for example, that "in construing contracts, we [must] give effect to all the language included therein, as the law of contract interpretation ... militates against interpreting a contract in a way that renders a provision superfluous." (Internal quotation marks omitted.) Ramirez v. Health Net of the Northeast, Inc. , supra, 285 Conn. at 14, 938 A.2d 576. The contracts at issue contain other exclusions that (1) bar coverage for workers' compensation claims and (2) bar coverage for any liabilities arising out of injuries to the policyholder's own employees. If we were to construe the occupational disease exclusions to apply only to claims against one's employer, as Vanderbilt proposes, then the exclusions would be rendered redundant and superfluous because these other provisions already bar coverage for such claims.

The trial court, in rejecting this argument, relied on what it understood to be a distinction in the Workers' Compensation Act between two types of employment related injuries: "occupational diseases," which are defined in § 31-275 (15), and " '[p]ersonal injur[ies],' " which are defined in § 31-275 (16) (A). Specifically, the court understood personal injuries to be sudden injuries such as industrial accidents, which occur on a particular date, whereas occupational diseases are disabilities that emerge gradually over time. The court reasoned that the occupational disease exclusions and the employer liability exclusions in the Vanderbilt policies were distinct exclusions that tracked this distinction between "occupational disease" and "personal injury" in the act. In other words, the employer liability clauses bar employee suits arising from sudden, accidental injuries, whereas the occupational disease exclusions bar employee suits arising from diseases gradually developed.

We perceive several flaws in this analysis. First, it is not clear to us that the act does in fact draw a sharp distinction between personal injuries and occupational diseases. Section 31-275 (16) (A) provides: " 'Personal injury' or 'injury' includes, in addition to accidental injury that may be definitely located as to the time when and the place where the accident occurred, an injury to an employee that is causally connected with the employee's employment and is the direct result of repetitive trauma or repetitive acts incident to such employment, and occupational disease ." (Emphasis added.) The trial court, focusing on the first section of the definition, appears to have overlooked the latter portion, which suggests that personal injury is a blanket concept that encompasses not only accidental injuries but also repetitive stress injuries and occupational diseases. That occupational diseases represent a subset of personal injuries for purposes of the act finds further support in § 31-275 (16) (B), which provides in relevant part: " 'Personal injury' or 'injury' shall not be construed to include ... (ii) [a] mental or emotional impairment, unless such impairment ... arises from a physical injury or occupational disease ...." (Emphasis added.) The plain language of the act thus does not support the court's effort to distinguish occupational diseases from personal injuries, and neither the court nor Vanderbilt has cited any authority to support a contrary interpretation of Connecticut law.

The second defect in the court's analysis is that it assumed that the parties intended to import the act's definition of the term "personal injury." See Commercial Union Ins. Co. v. Porter Hayden Co. , 116 Md.App. 605, 697-701, 698 A.2d 1167 (courts should not assume that statutory distinction between personal injury and occupational disease applies outside of workers' compensation context), cert. denied, 348 Md. 205, 703 A.2d 147 (1997). The employer's liability exclusion in the Pacific policy, for example, provides that "[t]his policy does not apply to personal injury to any employee of the insured ...." But the term "personal injury" is defined at some length in the "Definitions" section of the policy, which provides in relevant part that " 'personal injury' means, (a) bodily injury ...." "Bodily injury," in turn, is defined to mean "bodily injury, sickness or disease sustained by any person which occurs during the policy period ...." (Emphasis added.) By the policy's express terms, then, personal injury includes not only sudden, accidental injury, but also sickness and disease. This forecloses the trial court's interpretation of the employer's liability exclusion as speaking only to nondisease injuries.

The court's analysis is even less compelling when applied to the National Casualty policy. The employer liability exclusion in that policy provides in relevant part that "this policy shall not apply to any liability for bodily injury, sickness, disease , disability or shock, including death at any time resulting therefrom ...." (Emphasis added.) In this case, the exclusion clause itself refers to illness and disease as well as injury, precluding the possibility that the drafters intended to import a distinction between accidental injury and occupational disease.⁹⁶

The third shortcoming in the court's analysis is that it is inconsistent with another well established canon of construction. "It is a general rule of contract interpretation that if a contract includes a level of specificity in one context and then omits that specificity in a similar context, such an omission is purposeful and should be given meaning." O'Connell v. Liberty Mutual Fire Ins. Co. , 43 F.Supp.3d 1093, 1097 n.3 (D. Mont. 2014) ; see also State v. Heredia , 310 Conn. 742, 761, 81 A.3d 1163 (2013). In the present case, the employer liability exclusion clauses in the Pacific and National Casualty policies are labeled as such, and each expressly states that it excludes coverage only for injuries sustained by "any employee of the insured ...." The fact that the occupational disease exclusions are framed broadly and do not contain any similar language of limitation strongly suggests that the parties did not intend that they would be qualified in that manner. In other words, if the parties had intended that the occupational disease clause would bar coverage for occupational disease by employees and the employer liability clause would bar coverage for accidental personal injuries by employees , then it is reasonable to assume that the drafters would have expressly referenced employees of the insured in both instances or (perhaps) neither, but not in the latter instance but not the former.

For all of these reasons, we conclude that the trial court construed the occupational disease clauses too narrowly, and that those exclusions unambiguously⁹⁷ bar coverage for occupational disease claims brought not only by employees of Vanderbilt but also by individuals who contracted an occupational disease in the course of their work for other employers. On remand, the court is instructed to consider Vanderbilt's alternative argument that certain defendants are precluded from invoking the exclusions because they failed to timely plead the exclusions as a special defense.

C

Duty to Defend Under Continental's 1968-1977 Umbrella Policies

We next consider Vanderbilt's claim that the secondary insurance policies that it obtained from Continental between 1968 and 1977 unambiguously obligate Continental to defend Vanderbilt in the underlying actions for years in which Vanderbilt's primary policies are exhausted, and that the trial court erred in holding to the contrary. In the alternative, Vanderbilt contends that the relevant policy language is ambiguous and should be construed in favor of coverage. We disagree and conclude that the trial court properly construed the policy provisions at issue, which unambiguously do not afford excess defense cost coverage.

1

Facts

The following additional undisputed facts, procedural history, and policy language as gleaned from the record are relevant to Vanderbilt's claim. At trial, Vanderbilt and Continental agreed that Continental had issued primary comprehensive general liability policies to Vanderbilt from January 1, 1968 through March 3, 1977, and that Continental already had paid out the full coverage limits on those policies. The parties disagreed, however, as to whether, upon the exhaustion of those primary policies, Continental was obliged to contribute to defense costs in the underlying actions pursuant to certain secondary policies it had issued to Vanderbilt.

At issue were four secondary insurance policies: RDU 9004526, effective from January 1, 1968 to January 1, 1971; RDU 8046898, effective from January 1, 1971 to January 1, 1974; RDU 1251453, effective from January 1, 1974 to January 1, 1977; and RDU 1863573, effective from January 1, 1977 to January 1, 1978.⁹⁸ Each of these Continental policies is identified on its face as an "Umbrella Excess Third Party Liability Policy."

Each policy contains dual coverage grants. The first grant, entitled "COVERAGE A-EXCESS LIABILITY INDEMNITY," states in relevant part: "The company will indemnify the insured for loss in excess of the total applicable limits of liability stated in the schedule of underlying insurance. The provisions of the immediate underlying policy are, with respect to Coverage A, incorporated as a part of this policy except for any obligation to investigate and defend and pay for costs and expenses incident to any of the same , the amounts of the limits of liability, an 'other insurance' provision and any other provisions therein which are inconsistent with this policy."⁹⁹ (Emphasis added.) In light of the highlighted language, Vanderbilt has conceded that it cannot obtain a defense under Coverage A.

The second coverage grant contained in Continental's secondary policies, entitled "COVERAGE B-EXCESS LIABILITY INDEMNITY OVER RETAINED LIMIT," provides in relevant part: "The company will indemnify the insured, with respect to any occurrence not covered by underlying insurance, or with respect to damages not covered by underlying insurance but which results from an occurrence covered by underlying insurance, for ultimate net loss in excess of the insured's retained limit which the insured shall become obligated to pay as damages by reason of liability imposed upon the insured by law or assumed by the insured under contract because of personal injury, property damage, or advertising injury to which this policy applies, caused by an occurrence.

"The company, with respect to an occurrence not covered in whole or in part by underlying insurance or to which there is no other insurance in any way applicable , shall have the right and duty to defend any suit against the insured seeking damages on account of such personal injury, property damage or advertising injury, 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." (Emphasis added.)

Prior to the second phase of the trial, the parties requested that the trial court resolve a dispute as to whether the Coverage B provisions provide for a duty to defend under the circumstances of the present action. Specifically, Vanderbilt argued that if the underlying primary policies were exhausted, those policies no longer " 'covered' " and were no longer " 'applicable' " to the underlying actions. Therefore, the injuries alleged in the underlying actions now qualify as "occurrence[s] not covered in whole or in part by underlying insurance or to which there is no other insurance in any way applicable ...."

Continental countered that the question of whether underlying insurance covers an occurrence for purposes of Coverage B focuses on the nature of the occurrence, not the amount of the primary policy limit. In other words, Continental's view was that, because the underlying actions indisputably involve the types of claims " 'covered in whole or in part' " by Vanderbilt's primary policies, Coverage B is not implicated regardless of whether those policies have been exhausted. Rather, Continental took the position that Vanderbilt can obtain only excess indemnity coverage for the underlying actions, and only under Coverage A.

In a supplemental memorandum of decision, the trial court agreed with Continental. The court concluded that the plain and unambiguous language of the policies does not create a duty to defend under Coverage B when the underlying primary policies have been exhausted. Finding no relevant Connecticut authority on point, the court reviewed the handful of sister state decisions to have considered similar policy language and found those favoring Continental's position to be more persuasive. The court also concluded that interpreting Coverage B to provide a duty to defend would render Coverage A inconsistent and superfluous because Coverage A expressly provides that the insurer does not have a duty to defend with respect to losses in excess of the underlying primary policy limits. Vanderbilt challenges these conclusions on appeal.

2

Analysis

In debating the meaning and scope of the Coverage B defense coverage provisions, Vanderbilt and Continental largely talk past each other. Vanderbilt makes much of the specific policy language providing that Continental shall have a duty to defend with respect to occurrences " 'not covered in whole or in part by underlying insurance' " (Coverage Clause) or " 'to which there is no other insurance in any way applicable' " (Applicable Clause), arguing that the use of the disjunctive "or" means that each clause must be given independent meaning. Continental, by contrast, focuses on the overall structure of the policy and the purposes of the different coverage grants. It distinguishes between Coverage A, which Continental understands to provide excess coverage over the primary policy limits, and Coverage B, which it understands to provide drop-down or primary umbrella coverage for types of claims that fall completely outside the ambit of the primary insurance. For its part, the trial court primarily considered how sister courts in other jurisdictions have construed policy language that is similar but, in most instances, not identical to that at issue in this case.

We begin by examining in greater detail the arguments of the parties. On appeal, Vanderbilt now appears to concede that the Coverage Clause contained in Coverage B does not apply to circumstances such as those in the present case, wherein an underlying primary policy has responded to certain claims up to the policy limits, so that the primary policy is now exhausted and no longer available to respond to other such claims. There is no dispute, then, that new underlying claims can be said to be " 'covered in whole or in part' " by the primary policies and, therefore, they are not entitled to a defense under the Coverage B Coverage Clause.

Vanderbilt maintains, however, that we must assume that the Coverage Clause and the Applicable Clause have independent meaning and that, because the clauses are presented in the disjunctive, only one need be satisfied in order to implicate the duty to defend. A necessary corollary, Vanderbilt contends, is that the policy language envisions scenarios in which a claim is covered in whole or part by the underlying insurance and yet there is no insurance in any way applicable. Although this interpretation of the policy language appears to be paradoxical, Vanderbilt posits that the exhaustion of the primary policy limits would be an example of such a scenario: once Vanderbilt's 1971 primary policy pays out its limit on covered asbestos claims, for instance, then, Vanderbilt contends, there is no longer any insurance in any way applicable to new claims alleging injuries during that year and, therefore, Continental must provide a defense under the Applicable Clause of Coverage B. Vanderbilt's theory thus relies on an implicit temporal distinction between the Coverage and Applicable Clauses. On its reading, the Coverage Clause refers to policies that provide coverage for a particular occurrence on their face, at any time, whereas the Applicable Clause refers only to those policies that are inapplicable at the time a claim is made and a coverage decision must be reached.

Vanderbilt's creative argument that Coverage B distinguishes between primary policies that never "covered" a claim and those that provided coverage but are no longer "applicable" to the claim sidesteps in large measure the case law on which the trial court relied. The court cited to a number of cases for the proposition that a drop-down umbrella insurance policy does not provide a duty to defend when the underlying policy is exhausted. See, e.g., Mission National Ins. Co. v. Duke Transportation Co. , 792 F.2d 550 (5th Cir. 1986) ; American Special Risk Ins. Co. v. A-Best Products, Inc. , supra, 975 F.Supp. at 1019 ; Continental Casualty Co. v. Roper Corp. , 173 Ill.App.3d 760, 123 Ill.Dec. 360, 527 N.E.2d 998 (1988). Vanderbilt contends that those decisions are distinguishable from the present case because, in each case, either the policies at issue did not contain the "in any way applicable" language or, if they did, the deciding court did not discuss that language but instead focused its analysis exclusively on whether the Coverage Clause afforded a defense. More directly on point, Vanderbilt contends, is Continental Casualty Co. v. Synalloy Corp. , supra, 667 F.Supp. at 1523, in which the District Court suggested-albeit in dicta and without any discussion or analysis-that the Applicable Clause does provide a duty to defend when underlying insurance has exhausted.¹⁰⁰ Id., at 1541.

Continental, by contrast, takes a broader perspective, focusing its argument on the different functions allegedly served by the Coverage A and Coverage B provisions in its umbrella policies. As the Supreme Court of Ohio recently explained, "[u]mbrella policies are different from standard excess insurance policies, since they provide both excess coverage (vertical coverage) and primary coverage (horizontal coverage).... The vertical coverage provides additional coverage above the limits of the insured's underlying primary insurance, whereas the horizontal coverage is said to drop down to provide primary coverage for situations where the underlying insurance provides no coverage at all." (Citation omitted; internal quotation marks omitted.) Granger v. Auto-Owners Ins. , 144 Ohio St.3d 57, 62, 40 N.E.3d 1110 (2015) ;¹⁰¹ accord M. Taylor et al., Connecticut Insurance Law (2d Ed. 2013) pp. 226-27. Continental contends that the Coverage A provision in its policies serves the excess, or vertical, insurance function, following form to an underlying primary policy and affording additional indemnity-but not defense-coverage once the limits of the primary policy have been reached. Coverage B, by contrast, serves a horizontal, gap-filling function, acting in place of primary indemnity and defense coverage with respect to actions or damages that fall completely outside the scope of the underlying primary policy and any other applicable policies. Continental relies on cases such as Roper Corp . and A-Best Products, Inc ., which have concluded that these types of gap-filling provisions are primarily intended to respond to claims seeking punitive damages or alleging causes of action such as malpractice or worldwide operations liability, which typically are not covered by comprehensive general liability policies. See American Special Risk Ins. Co. v. A-Best Products, Inc. , supra, 975 F.Supp. at 1022 ; Continental Casualty Co. v. Roper Corp. , supra, 173 Ill.App.3d at 764, 123 Ill.Dec. 360, 527 N.E.2d 998. According to this view, Coverages A and B represent different, mutually exclusive types of insurance, and providing additional coverage once an underlying policy has exhausted clearly falls under the purview of the former and not the latter. See American Special Risk Ins. Co. v. A-Best Products, Inc. , supra, at 1025 (distinguishing between coverage, which comprehends set of risks, and collectibility, which comprehends whether damages for covered risk are recoverable from insurer in light of its policy limits).

We agree with Continental that, in construing insurance policies, we are instructed to "look at the contract as a whole, consider all relevant portions together and, if possible, give operative effect to every provision in order to reach a reasonable overall result." (Internal quotation marks omitted.) Lexington Ins. Co. v. Lexington Healthcare Group, Inc. , supra, 311 Conn. at 38, 84 A.3d 1167. In the present case, several features of the policy support Continental's interpretation. First, we note that the excess policies incorporate various provisions of the underlying primary insurance policies "with respect to Coverage A," whereas Coverage B expressly applies to occurrences or damages not covered by underlying insurance. The policies also indicate that Coverage A follows form to the terms of the underlying insurance, whereas Coverage B does not.¹⁰² This supports Continental's theory that only Coverage A sits atop the underlying insurance.

Second, we note that the fourth section of each umbrella policy, entitled "Limits of Liability," lays out the effect that the exhaustion of the underlying primary insurance will have on the insurer's obligations under Coverage A . This section provides in relevant part: "With respect to Coverage A, if the applicable limit of liability of the underlying insurance is less than as stated in the schedule of underlying insurance because the aggregate limit of liability of the underlying insurance has been ... exhausted, this policy ... replaces the limit of liability exhausted ...." By contrast, nothing in the policy expressly instructs how Coverage B is to be applied in the event that underlying primary insurance is exhausted. This notable lacuna bolsters Continental's argument that Coverage B is solely a drop-down umbrella provision that is not implicated by the exhaustion of underlying insurance.

Third, it is difficult to understand how the parties to an agreement could reasonably expect that one provision, Coverage A, would disclaim any obligation by the insurer to provide a defense once the underlying policy is exhausted while the very next provision, Coverage B, imposes a duty to defend under the very same circumstances.¹⁰³ Because Vanderbilt's interpretation of the policy language results in a conflict between the Coverage A and B provisions, any equally plausible interpretation that gives harmonious effect to all of the contractual covenants is to be preferred. See Colorado Milling & Elevator Co. v. Chicago, Rock Island & Pacific Railroad Co. , 382 F.2d 834, 836 (10th Cir. 1967) ; 2 Restatement (Second), supra, § 202, comment (d), p. 88; 3 S. Williston, Contracts (Rev. Ed. 1936) § 618, p. 1779.

Fourth, we note that there is nothing in the grammar or structure of the relevant Coverage B language to support Vanderbilt's view that the Applicable Clause is to be construed at the time that coverage is sought, whereas the Coverage Clause is to be construed in a temporal vacuum. Why, in other words, should the relevant language-"with respect to an occurrence not covered in whole or in part by underlying insurance or to which there is no other insurance in any way applicable"-be understood to mean that an exhausted policy continues to cover a claim but is no longer applicable to the claim? Particularly in light of the fact that coverage under Coverage B is limited to situations in which there is no other insurance in any way applicable, it seems odd to interpret the Applicable Clause as narrowly as Vanderbilt proposes. Surely, a primary insurance policy that on its face provides coverage for a particular claim is in some way applicable to that claim, even if the policy limits have been reached.¹⁰⁴

Finally, we recognize that other courts, in parsing similar contractual provisions, have adopted Continental's view of the policy structure, referring to Coverage A as providing "excess" coverage and Coverage B as providing "umbrella" coverage. See, e.g., Tri-State Construction, Inc. v. Columbia Casualty Co./CNA , 39 Wash.App. 309, 313, 692 P.2d 899 (1984). Although Vanderbilt correctly notes that the word "excess" appears in the title of both coverage provisions, the full phrase "excess ... over retained limit" used in the Coverage B title is understood to be a term of art by which the insurance industry describes drop-down umbrella coverage. See T. Novak, " The Defense Obligation of Excess and Umbrella Liability Insurance Policies," 36 Brief 12, 14 (2006) ; see also United States v. Security Management Co. , 96 F.3d 260, 267-68 (7th Cir. 1996) (drop-down umbrella provision provided "[c]overage [o]ver [r]etained [l]imit"); Fireman's Fund Ins. Co. v. National Bank for Cooperatives , 849 F.Supp. 1347, 1366 (N.D. Cal. 1994) (same). We also have cautioned that courts should not interpret contracts on the basis of section headings when doing so would conflict with the plain meaning of the text. See FirstLight Hydro Generating Co. v. First Black Ink, LLC , 143 Conn.App. 635, 642, 70 A.3d 174, cert. denied, 310 Conn. 913, 76 A.3d 629 (2013) ; but see footnote 77 of this opinion and accompanying text.

If Continental is correct that (1) Coverage A provides excess indemnity coverage over and above the primary policy limits and (2) Coverage B serves a fundamentally different purpose, dropping down to serve as primary indemnity and defense coverage for claims that fall-and always would have fallen-completely outside the ambit of the primary policy, then the question becomes whether the Coverage B language at issue can be construed in a manner that is at least as plausible as Vanderbilt's reading but that better reconciles all of the various policy provisions. Specifically, why are the two coverage clauses of Coverage B written in the disjunctive, so as to suggest that there are a class of claims that are covered in whole or in part by the underlying insurance and yet to which there is no insurance (other than the umbrella policy) in any way applicable?

Although the policy literally states that defense coverage is available when the Coverage Clause or the Applicable Clause is implicated, both this court and our Supreme Court have recognized that the word "or" may on occasion be understood in the conjunctive rather than disjunctive sense. See, e.g., State v. Angell , 237 Conn. 321, 329, 677 A.2d 912 (1996) ; D'Occhio v. Connecticut Real Estate Commission , 189 Conn. 162, 169-70, 455 A.2d 833 (1983) ; Bordonaro v. Senk , 109 Conn. 428, 430, 147 A. 136 (1929) ; Gelinas v. West Hartford , 65 Conn.App. 265, 281, 782 A.2d 679, cert. denied, 258 Conn. 926, 783 A.2d 1028 (2001). Here, given that Coverage B is undoubtedly an umbrella provision that provides coverage for claims and damages not covered by other insurance, the most reasonable reading of the policy language is that defense coverage is available when a claim is not covered by underlying primary insurance and when no other coverage is applicable.¹⁰⁵ This reading is supported by the language stating that defense coverage is available only when no other insurance is applicable "in any way." Notably, at least two other courts that have construed substantially identical policy language have either explicitly or implicitly construed the "or" in Coverage B in this manner. See Continental Casualty Co. v. Synalloy Corp. , supra, 667 F.Supp. at 1540 (insurer's denial of coverage under primary policy triggers Coverage B duty to defend if there also is no other applicable insurance); Caldwell Freight Lines, Inc. v. Lumbermens Mutual Casualty Co. , 947 So.2d 948, 957 (Miss. 2007) (concluding that policy language unambiguously "provides that where neither the terms and conditions of the [underlying policy nor] any other policy apply, then Coverage B ... may apply"). By contrast, we are not aware of, and Vanderbilt has not identified, any court that has construed the policy language in the manner that Vanderbilt proposes.

Indeed, even if we were to understand the word "or" in its traditional disjunctive sense, the Applicable Clause need not be read as expanding the class of occurrences for which defense coverage is available. Instead, the language at issue simply may have been an inartful way of expressing the idea that defense coverage is available under the policy only for occurrences that are not covered (1) in whole by underlying insurance, or (2) in part by underlying insurance, or (3) in any way by other applicable insurance.

In this sense, the disputed language in the second paragraph of Coverage B may be understood as paralleling the indemnity provision in the first paragraph of Coverage B. The first paragraph states that the insurer will provide indemnity coverage when (1) occurrences are not covered by underlying insurance or (2) occurrences are covered by underlying insurance but the damages claimed are not. The relevant language in the second paragraph tracks that dichotomy and may be understood to provide defense coverage under the same two circumstances: (1) when an occurrence is not covered by underlying primary insurance; or (2) when the occurrence is covered but the damages claimed are not covered and no other relevant policies are available to respond. See footnote 104 of this opinion. The fact that the second paragraph of Coverage B affords the insurer the right as well as the duty to defend under those two circumstances is consistent with the idea that the language at issue in the second paragraph simply tracks the two circumstances under which the first paragraph affords indemnity coverage.

Although this interpretation requires that we read something into the policy language, it is no less fettered to the plain language of the contract than is the interpretation proposed by Vanderbilt. Because Continental's reading of the policy language is equally plausible on its face and best comports with other specific provisions of the policy, the overall policy structure, and the conclusions reached by our sister courts, we conclude that the trial court properly determined that Coverage B unambiguously does not require the insurer to defend claims that would have been covered by underlying insurance but for its exhaustion.

V

OTHER CLAIMS

Finally, we consider: (1) Mt. McKinley's claim that the trial court improperly admitted certain charts into evidence; (2) two additional claims raised by Mt. McKinley regarding exhaustion of the primary policies issued by Hartford and Continental; (3) Old Republic's claim that the trial court should have determined the extent of Old Republic's defense obligations pursuant to its excess insurance policies during the Phase II trial; and (4) Continental's claim that the trial court should have concluded that Continental has no duty to defend Vanderbilt under certain other excess insurance policies. We find no error.

A

Admission of Posner Charts

Mt. McKinley challenges the admission into evidence of certain charts prepared by Posner, who testified as an expert witness on behalf of Vanderbilt. Mt. McKinley claims that the court improperly admitted those materials as substantive evidence for the truth of the information contained therein, in contravention of § 7-4 (b) of the Connecticut Code of Evidence. By contrast, Vanderbilt submits that the court properly admitted the charts for the limited purpose of explaining the methodology and information relied on by Posner in formulating his expert opinion. On our review of the record, we agree with Vanderbilt.

The following additional facts are relevant to this claim. At the time of trial, Posner possessed more than thirty-five years of experience in the field of insurance and risk management. Posner testified that he had performed allocations among multiple insurance carriers for "long-tail claims," such as asbestos and environmental claims, "hundreds of times." During the Phase II proceeding, Posner provided expert testimony on the allocation of indemnity payments by Continental and Hartford.¹⁰⁶ His expert opinion was predicated in part on information documented in four charts that Posner prepared prior to trial.¹⁰⁷ The first chart, known as plaintiff's exhibit V-781-285, was captioned "Preliminary Data for Allocation." Referred to at trial as the "master document," Posner described that chart as "an Excel spreadsheet, which I put together in order to capture important information from which I [then] was going to perform an allocation."¹⁰⁸ That chart detailed, among other things, the names of numerous claimants, their "exposure begin dates and exposure end dates," and payment amounts by certain entities. The second and third charts at issue depict "allocation model[s]" that Posner created utilizing the information contained in the master document, albeit with different default dates of first exposure.¹⁰⁹ Posner described the fourth chart as a "reconciliation chart" that "summarizes the amounts that I allocated [and] that [Continental] and Hartford allocated, and it explains ... why there are differences between what I allocated and what they allocated."

During Posner's testimony at trial, Vanderbilt's counsel moved to admit the master document known as plaintiff's exhibit V-781-285 into evidence. At that time, counsel for Mt. McKinley objected on the ground that "there's no basis for its admissibility, it's irrelevant; information relied upon by an expert doesn't come in unless there's some kind of independent basis. And this contains hearsay." In response, Vanderbilt's counsel noted that the chart documented the methodology by which Posner reached his expert opinion, stating that it demonstrated "the steps he took to analyze and form that opinion ...." The court overruled the objection and admitted that chart into evidence. In so doing, the court expressly noted that the defendants reserved the right to challenge the substance of that underlying information, stating: "Defendants certainly [are] entitled to, on cross-examination, challenge any accuracy issues, the information used in formulating an opinion, any other documents that were relied upon in formulating that opinion."

When Vanderbilt subsequently sought to introduce Posner's "reconciliation chart" into evidence, counsel for Mt. McKinley again objected, stating in relevant part: "Hearsay, again ... this is a document an expert relied upon, it's not admissible unless there's an independent hearsay exception." In response, the court inquired, "Mr. Posner, let me ask you one more question. This material that you reviewed, did you rely upon it in formulating your opinions?" Posner answered in the affirmative. The court then overruled the objection and permitted the reconciliation chart to be admitted into evidence. Vanderbilt later moved to introduce the two charts containing Posner's allocation models with differing default dates of first exposure, and Mt. McKinley once again objected on hearsay grounds. The court overruled that objection, noting that "these [two charts] were born of the master document that has been admitted into evidence already." Mt. McKinley now challenges the propriety of those evidentiary determinations.

As a general matter, we note that our trial courts are afforded "wide discretion in determining whether to admit expert testimony and, unless the trial court's decision is unreasonable, made on untenable grounds ... or involves a clear misconception of the law, we will not disturb its decision." (Internal quotation marks omitted.)

Fleming v. Dionisio , 317 Conn. 498, 505, 119 A.3d 531 (2015). Nevertheless, a crucial predicate to the admissibility of such testimony is the showing of sufficient foundational facts for the expert's opinion. Conn. Code Evid. § 7-4 (a) ; see also Viera v. Cohen , 283 Conn. 412, 444-49, 927 A.2d 843 (2007).

Titled, "Bases of opinion testimony by experts," § 7-4 (b) of the Connecticut Code of Evidence provides: "The facts in the particular case upon which an expert bases an opinion may be those perceived by or made known to the expert at or before the proceeding. The facts need not be admissible in evidence if of a type customarily relied on by experts in the particular field in forming opinions on the subject. The facts relied on pursuant to this subsection are not substantive evidence, unless otherwise admissible as such evidence." The commentary to that rule of evidence explains that § 7-4 (b)"expressly forbids the facts upon which the expert based his or her opinion to be admitted for their truth unless otherwise substantively admissible under other provisions of the Code. Thus, subsection (b) does not constitute an exception to the hearsay rule or any other exclusionary provision of the Code. However, because subsection (a) requires disclosure of a sufficient factual basis for the expert's opinion, and because the cross-examiner often will want to explore the expert's factual basis further, subsection (b) does not preclude the trial court, in its discretion, from admitting the underlying facts relied on by the expert for the limited purpose of explaining the factual basis for the expert's opinion." Conn. Code Evid. § 7-4, commentary; accord 2 C. McCormick, Evidence (6th Ed. 2006) § 324.3, pp. 417-18 ("[a]n expert often should be allowed to disclose to the [trier of fact] the basis for an opinion because otherwise the opinion is left unsupported with little way for evaluation of its correctness").

Consistent with that authority, our appellate courts have construed that rule of evidence to permit the admission of otherwise inadmissible hearsay evidence "for the limited purpose of explaining the factual basis for the expert's opinion." (Emphasis in original.) Milford v. Maykut , 117 Conn.App. 237, 241 n.7, 978 A.2d 570, cert. denied, 294 Conn. 906, 982 A.2d 1080 (2009) ; see also Milliun v. New Milford Hospital , 310 Conn. 711, 728, 80 A.3d 887 (2013) ("[t]he fact that [an expert's] report includes hearsay statements ... would not bar the report's admission on that basis unless those statements were being offered for substantive purposes, i.e., the truth of the matter asserted"); C. Tait & E. Prescott, Connecticut Evidence (5th Ed. 2014) § 7.9.4 (b), p. 475 ("such hearsay information is used solely to show the basis for the expert opinion and is not independently admissible unless within a recognized hearsay exception").

Accordingly, the critical inquiry in the present case entails consideration of Vanderbilt's purpose in offering the charts into evidence. As this court has noted, "[i]nformation on which an expert relied that is not offered for its truth but is offered to show that the expert relied on it is not hearsay and may be the subject of proper cross-examination to test the basis of that expert's opinion." State v. Henry , 27 Conn.App. 520, 529, 608 A.2d 696 (1992). When a party introduces evidence "to show the basis of [the] expert opinion," rather than for the truth of the matter asserted, the evidence properly may be admitted, as such evidence does not constitute hearsay. Id., at 530, 608 A.2d 696 ; compare State v. O'Brien-Veader , 318 Conn. 514, 536, 122 A.3d 555 (2015) (trial court declined to admit medical records "because they were not facts relied upon by [the expert] in forming his opinion"), with Milliun v. New Milford Hospital , supra, 310 Conn. at 728-29, 80 A.3d 887 (medical reports admissible because offered as evidence of information on which experts relied in formulating opinion). When Mt. McKinley first objected to the introduction of the "master document" chart, its counsel began by acknowledging that "this is data the expert relied upon"-precisely the point later articulated by opposing counsel. Mt. McKinley nonetheless argued that the chart contained hearsay and that "no basis for its admission" existed. In response, Vanderbilt did not proffer a hearsay exception or request that the chart and its contents be admitted for the truth of the matter asserted. Instead, its counsel informed the court that the chart documented the methodology and information underlying Posner's expert opinion, stating that "the expert has testified that he had an opinion and [this chart details] the steps he took to analyze and form that opinion ...." In so doing, Vanderbilt plainly indicated to the court that it was offering the master document as evidence of the basis of Posner's expert opinion. Such evidence is admissible in the discretion of the trial court. Pickel v. Automated Waste Disposal, Inc. , 65 Conn.App. 176, 192-93, 782 A.2d 231 (2001) ; see also Conn. Code Evid. § 7-4, commentary (§ 7-4 [b] "does not preclude the trial court, in its discretion, from admitting the underlying facts relied on by the expert for the limited purpose of explaining the factual basis for the expert's opinion"); C. Tait & E. Prescott, supra, § 7.9.3, p. 474 (discretion afforded to court pursuant to § 7-4 [b] may be "exercised to admit such facts on direct examination").

The question, then, is whether the court abused its discretion in admitting the master document and other charts into evidence. We conclude that it did not. Because those materials were introduced as evidence of the information on which Posner relied in formulating his expert opinion, they were not hearsay. State v. Henry , supra, 27 Conn.App. at 529-30, 608 A.2d 696. Moreover, the court, in exercising its discretion to admit that evidence, informed the parties, consistent with appellate precedent, that such underlying information was subject to cross-examination to test the basis of the expert's opinion. See id., at 529, 608 A.2d 696.

Although it is true that the court's admission of the master document and other charts was not accompanied by reference to the limited nature of such admissions, the court earlier had indicated to the parties that it appreciated that distinction. Prior to the introduction of any of the charts at issue, the court apprised the parties as follows: "[T]his is a court trial, and to that extent ... I'll reiterate that the parties are not precluded from making any sort of objection, but certainly there will be times that there may be either opinions offered from an expert or from a fact witness that may either bleed over into an area that might be deemed [inadmissible] or something that's not properly an area of inquiry on, but sitting as the trier of fact, I think I can say with some measure of confidence that the court pretty much is able to discern what is proper evidence before it and what is not.... [A]s a general matter, I think the court usually is able to sift through what should be properly before it and what isn't.... I'm not saying that there shouldn't have been any objection, please don't misunderstand me at all; this is sort of just guidance for the future from here on out." The court also informed the parties that "as I sit here and listen to testimony and evidence ... I'm also sifting through it and [considering it] from a standpoint of, is this properly admissible to the court or not." As a result, the potential "danger," as one commentator termed it; 2 C. McCormick, supra, § 324.3, p. 419; of jurors improperly utilizing otherwise inadmissible information for the truth of the matter asserted was absent in this court trial. Furthermore, at no point in the proceedings before the trial court did any party bring § 7-4 (b) of the Connecticut Code of Evidence or the issue of limited admissibility of the charts to the court's attention. "It is well established that where the record is silent, a reviewing court will not presume error." State v. Brown , 145 Conn.App. 174, 187 n.13, 75 A.3d 713, cert. denied, 310 Conn. 936, 79 A.3d 890 (2013). In the face of such silence, we presume, consistent with its earlier advisory to the parties, that the court, in accordance with § 7-4 (b), properly admitted the charts as evidence of the basis for Posner's expert opinion, and not for the truth of the matter contained therein. See State v. Carter , 122 Conn.App. 527, 533, 998 A.2d 1217 (2010) ("[a]bsent evidence to the contrary, we presume that the court properly applied [the] law"), cert. denied, 300 Conn. 915, 13 A.3d 1104 (2011). That presumption is buttressed by the fact that, when Mt. McKinley later objected to the admission of the "reconciliation chart," the court immediately asked Posner whether it constituted material that he relied upon in "formulating" his expert opinion. Only after Posner answered in the affirmative did the court rule that the chart was admissible.

In determining whether the trial court has abused its discretion, "we make every reasonable presumption in favor of upholding the trial court's ruling, and only upset it for a manifest abuse of discretion." (Internal quotation marks omitted.) Chief Information Officer v. Computers Plus Center, Inc. , 310 Conn. 60, 98, 74 A.3d 1242 (2013). No manifest abuse of discretion emanates from the record before us. Accordingly, we conclude that the court properly admitted the charts into evidence for the limited purpose of explaining the factual basis of the expert's opinion, in adherence to § 7-4 of our Code of Evidence.

Even if we were to conclude that the admission of the charts was improper, Mt. McKinley still has not met its burden of establishing reversible error. See Brookfield v. Candlewood Shores Estates, Inc. , 201 Conn. 1, 7, 513 A.2d 1218 (1986) ("[t]he burden is on the appellant to prove harmful error"). "When a court commits an evidentiary impropriety, we will reverse the trial court's judgment only if we conclude that the trial court's improper ruling harmed the plaintiffs.... In a civil case, a party proves harm by showing that the improper evidentiary ruling likely affected the outcome of the proceeding." (Internal quotation marks omitted.) Doe v. Hartford Roman Catholic Diocesan Corp. , 317 Conn. 357, 396-97, 119 A.3d 462 (2015).

No such showing has been made in the present case. First, we note that the trial court assured the parties prior to the introduction of the charts in question that, although there might arise evidentiary matters "that might be deemed [inadmissible] or something that's not properly an area of inquiry," the court "sitting as the trier of fact ... can say with some measure of confidence that [it] is able to discern what is proper evidence before it and what is not." That advisory suggests that the court both appreciated the delicacy of the evidentiary issues before it and remained vigilant in its consideration of that evidence as trial proceeded. Second, to paraphrase this court's decision in DeNunzio v. DeNunzio , 151 Conn.App. 403, 414, 95 A.3d 557 (2014), aff'd, 320 Conn. 178, 128 A.3d 901 (2016), our review of the record reveals no evidence that the court improperly considered inadmissible hearsay with respect to the charts, but rather considered their contents as information that Posner considered in formulating his opinion. In particular, our review of the court's thorough memorandum of decision in the Phase II proceeding discloses no reliance by the court on those charts. We thus presume that the court did not improperly consider any such inadmissible hearsay as substantive evidence. Third, much of the information contained in the charts prepared by Posner is merely cumulative of other documentary and testimonial evidence in the record. Last, to the extent that Mt. McKinley and other defendants have articulated a concern that Vanderbilt might seek to utilize the charts as substantive evidence for the truth of the matter asserted in future phases of this litigation, our decision today should dispel that notion. We have concluded that the trial court properly admitted those charts into evidence for the limited purpose of explaining the basis of Posner's expert opinion, and not for the truth of the information contained therein. At present, those charts cannot be utilized in any other manner without running afoul of § 7-4 (b) of the Connecticut Code of Evidence. We therefore conclude that Mt. McKinley has not established that the admission of the charts likely affected the outcome of the proceeding, and thus cannot demonstrate harmful error.

B

Exhaustion of Primary Policies

In part III of this opinion, we considered certain claims that pertained to the exhaustion of Vanderbilt's primary insurance policies. We now address two additional claims raised by Mt. McKinley regarding the trial court's finding that Continental's 1968-1977 primary policies and Hartford's 1977-1986 primary policies have been exhausted. Mt. McKinley contends that the court failed to determine whether all of the underlying claims constituted "products" claims, which erode the aggregate policy limits, or whether some might have constituted premises/operations claims, which do not erode the limits in the Continental policies and which are not covered by the Hartford policies. Mt. McKinley also challenges the court's conclusion that Hartford's allocation to indemnity of a portion of the contribution payment that it made to Continental pursuant to the 2002 allocation agreement was reasonable.¹¹⁰ We do not agree.

1

Products Claims

Mt. McKinley maintains that the court failed to determine whether all of the underlying claims constituted "products" claims under the primary policies. The record before us includes copies of the policies in question. Those policies indicate that the Hartford policies at issue provide coverage only for claims falling within the "completed operations" and "products" hazards specified therein. Similarly, only payments on claims covered by the "completed operations" and "products" hazards count against the aggregate limit of the Continental policies. In its principal appellate brief, Mt. McKinley distinguishes such "products" claims from those that "fall outside of the products-completed operations hazards," which it refers to as "nonproducts" claims.¹¹¹ It contends that either the policyholder or the primary insurers bear the burden of establishing that the primary policies have been exhausted before the secondary policies can be triggered, and that it is impossible to know whether the primary policies were exhausted until the court determines whether all of the payments allocated to those policies were in fact for covered products claims.

At the Phase II trial, the primary insurers proffered evidence indicating that the underlying claims for which they provided defense and indemnity coverage involved products claims. The court heard testimony from Peter A. Pogue, a claims consultant with Resolute Management, Inc., which administered asbestos claims on behalf of Continental. Pogue handled the Vanderbilt account and testified that he had examined the primary insurance policies issued by Continental. In reviewing documentation of the payments made by Continental for underlying claims, Pogue detailed the payments made "for asbestos related and ... bodily injuries" pursuant to the "completed operations" and "products" hazards. Pogue explained that Continental predicated its determination that the aggregate limits of its policies were reached on those payments.

The court also heard testimony from Lawrence Farber, an assistant vice president in Hartford's complex claims group who oversees "various types of complex claims, including asbestos claims, pollution claims ... typically claims that may span multiple policy periods." Questioned about Hartford's duties to defend and indemnify, Farber explained that "[e]very claim is a case-by-case basis .... [I]f we looked at our policy and the facts presented to us, if we determine we had a duty, we would uphold that duty." Farber testified that Vanderbilt had tendered bodily injury claims to Hartford under the primary insurance policies, which he confirmed were "products claims." Farber explained that his knowledge of Vanderbilt's account "comes from my review of all the claim files. It comes from talking to the [claims] handlers that were involved with the handling of the claim. It comes from my review of the policies, review of the loss runs."¹¹² Farber confirmed that he also conducted a review of "the underlying claims file ... to determine whether Hartford understood that they were paying products claims versus other types of claims with respect to the Vanderbilt account." On that review, Farber concluded "that they were products claims."

In its Phase II decision, the court found that "Farber was credible and persuasive in demonstrating ... that the total amount of indemnity payments made by Hartford were sufficient to exhaust its policies ...." The court emphasized that "[n]o evidence contradicted either Farber's testimony or Hartford and [Continental's] loss runs.... There is no suggestion anywhere on the record that Hartford and [Continental] did anything other than act reasonably and in good faith throughout the duration of their defense and indemnity of the plaintiff." The court also credited Pogue's testimony "that [Continental] has paid indemnity well in excess of" the aggregate limits of its primary policies and Farber's testimony that Hartford "had paid indemnity in excess of [the limits of its policies] throughout the years." Accordingly, the court found that "the Hartford and [Continental] primary policies for the periods March 3, 1977 to March 3, 1986, and January 1, 1968 to March 3, 1977, respectively, are exhausted ...." Although the court did not explicitly address the distinction between products and nonproducts claims, implicit in its decision is the determination that those policies were exhausted through payments for covered products claims. The testimony of Farber and Pogue, which the court expressly credited in finding the policies to be exhausted, substantiates that determination.

"In Connecticut, our appellate courts do not presume error on the part of the trial court.... Rather, the burden rests with the appellant to demonstrate reversible error." (Citation omitted; internal quotation marks omitted.) Jalbert v. Mulligan , 153 Conn.App. 124, 145, 101 A.3d 279, cert. denied, 315 Conn. 901, 104 A.3d 107 (2014). Mt. McKinley has not met that burden. It furnished no evidence at trial, and has identified no evidence on appeal, indicating that any of the underlying claims fell outside the completed operations and products hazards of the primary policies.¹¹³ Lillian Philburn, a claims manager who handled Vanderbilt's account on behalf of Mt. McKinley, confirmed at trial that Mt. McKinley received "all of the claim[s] files" from Hartford in 2009-years prior to the Phase II trial. She further testified that Mt. McKinley had a third party review those files. Yet Mt. McKinley provided no evidence at trial on which the court could find that any of those claims did not qualify for coverage under the policies at issue. On the evidence before it, therefore, the court reasonably could determine that the underlying claims constituted products claims pursuant to the primary policies.

2

Application of 2002 Allocation Agreement

Mt. McKinley also contests the court's determination that the primary insurers reasonably allocated Hartford's settlement payment to Continental pursuant to their 2002 allocation agreement. Under that agreement, a portion¹¹⁴ of Hartford's payment was allocated to indemnity, and that share then was allocated evenly across all of the Hartford policies at issue. Mt. McKinley contends that this allocation scheme was unreasonable and, specifically, that the indemnity payments should have been allocated on a claim by claim basis in light of the allocation block and date of first exposure unique to each underlying action.

This claim largely is subsumed by our discussion in part III C 1 of this opinion, in which we affirm the trial court's decision to uphold the allocation agreement, which agreement encompasses the contribution payment made by Hartford. In so doing, we concluded that the trial court properly determined that the allocation agreement was objectively reasonable at the time of its adoption, in accordance with industry standards, and was undertaken in good faith.

In rejecting Mt. McKinley's claim, the trial court indicated that it was "unwilling to second-guess Hartford's allocation to indemnity of a portion of the settlement" with Continental. The court credited Farber's testimony that such allocation was done reasonably and in good faith. Asked how that allocation was reached, Farber explained that "[i]nformation was exchanged between [Hartford and Continental], and it was reviewed and we determined that the [specific sum] would be our appropriate share" of the indemnity expenses incurred by Continental. The court also credited Huffer's testimony that the allocation of the indemnity portion of the contribution payment "across the coverage block" was "fair and reasonable." In addition, the court relied on the loss runs of Continental and Hartford, which it found "provided evidence that both insurers determined the moneys from the settlement allocated to indemnity represented a good faith estimation of the indemnity payments Hartford owed to [Continental]." Last, the court emphasized that "there was insufficient credible evidence presented by the excess and/or umbrella insurers to demonstrate that Hartford's allocation of the settlement amount was improper ...." We concur with that assessment. Our review of the record reveals no evidence that the primary insurers intentionally allocated payments so as to prematurely exhaust their primary policies and shift their obligations onto the secondary insurers. We therefore refuse to disturb the court's determinations in this regard.

C

Duty to Defend Under Old Republic's Excess Policies

Apart from joining several issues raised by Mt. McKinley and Continental in their cross appeals, Old Republic presents one distinct matter for our consideration in its capacity as a cross appellant. Consistent with a stipulation it has entered into with Vanderbilt, Old Republic asks this court to direct the trial court on remand to consider the issue of its defense obligations prior to commencing the Phase III trial.

The following additional procedural facts are relevant to this request. Old Republic issued two high level excess insurance policies to Vanderbilt for the periods of March 3, 1981 through March 3, 1982, and March 3, 1982 through March 3, 1983, respectively. When this litigation ensued, Continental named Old Republic as a defendant in its third party complaint.

Prior to the commencement of the Phase I trial, Old Republic joined a motion for summary judgment filed by defendant Certain London Market Insurance Companies. That motion alleged in relevant part that the policies in question do not "provide for a defense obligation" and only require those excess insurers "to reimburse [Vanderbilt] for defense costs within limits for covered claims." By order dated July 13, 2012, the court denied that motion, stating simply that on "a review of [the] evidence properly presented, the court finds that there remains a genuine issue relative to one or more material facts."

Trial in Phases I and II of this proceeding followed. As noted in its memorandum of decision, the court, upon motion of the parties, "agreed to bifurcate Phase II of the trial so as to remove the issues of damages and reimbursement of any overpayment of defense and indemnity costs [which] will be considered following the completion of Phase II."

The Phase II trial commenced on May 15, 2013, and continued over the course of fourteen days. Old Republic thereafter filed a posttrial brief "concerning the issue of how [its] policies pay defense costs within limits for covered claims." In that brief, Old Republic maintained that its "policies are excess indemnity policies that contain no duty to defend, and only provide for indemnification or reimbursement of certain loss and expense payments associated only with covered claims.... [T]he Old Republic policies are stand-alone policies that are governed by their own terms, conditions and exclusions, and are distinct from the underlying primary and umbrella policies. Thus, while the Old Republic policies 'follow form' to the [Mt. McKinley] umbrella policies to a limited extent, they do not have the same obligation to pay concerning the defense of underlying claims, including the payment of defense costs as incurred for uncovered claims." Relying on the distinction between liability and indemnity policies drawn by our Supreme Court in Cohn v. Pacific Employers Ins. Co. , 213 Conn. 540, 546-47, 569 A.2d 544 (1990),¹¹⁵ Old Republic claimed that the plain language of the excess policies it issued to Vanderbilt indicate that they imposed "an 'indemnity only' obligation." Old Republic thus asked the court to find, among other things, that its policies "contain no affirmative duty to [defend] or to pay defense costs as incurred" by Vanderbilt, and that its policies "only indemnify for defense expenses incurred and paid for covered claims ...." The court declined to address those claims in its Phase II memorandum of decision.

In its preliminary statement of issues in AC 37145, Old Republic averred that the court improperly (1) "denied Old Republic's motion for summary judgment declaring defense obligations and payment of defense costs within limits for covered claims," and (2) "failed to address the issue of defense obligations and payment of defense costs in the Old Republic policies during Phase II of trial, but decided that it would do so during a later phase of trial ...." Vanderbilt subsequently moved to strike those issues from Old Republic's interlocutory appeal, claiming they were beyond the scope of what the trial court permitted the parties to raise. This court disagreed, and thus denied that motion.

Prior to oral argument before this court, Old Republic and Vanderbilt entered into a stipulation, whereby Old Republic agreed to withdraw those two issues. Vanderbilt, in turn, agreed not to object to Old Republic's request that the trial court on remand consider the substance of those two issues "prior to any additional [p]hase of [t]rial for issues not yet addressed during the first two [p]hases of [t]rial." The stipulation further provides that "[b]oth parties reserve all substantive rights as to issues of law and facts." That stipulation and its corresponding withdrawal were filed with this court on August 6, 2015.

Old Republic now asks this court to direct the trial court accordingly. As it states in its principal appellate brief, "it is not seeking any substantive ruling from this court regarding any defense obligations that [its excess policies] may or may not have .... Old Republic respectfully is seeking an order from this court to require resolution of this issue prior to the commencement of any further phases of trial.... Consistent with the parties' and the trial court's prior representations, briefing and orders ... the issue of any defense obligations of Old Republic should be decided on relevant briefs, and prior to any further phase of trial in this matter.... Moreover, there are no claims pending against Old Republic that involve the reimbursement and/or overpayment issues that have been specified for Phase III of trial in this matter, and thus such a ruling likely would facilitate the streamlining of this matter, as Old Republic would not have to be present ... at that phase of trial." (Citations omitted.)

We appreciate the rationale for Old Republic's request. Should the trial court conclude that the excess insurance policies Old Republic issued to Vanderbilt are indemnity only policies that impose no duty to defend, that conclusion certainly would alter the court's analysis of Old Republic's obligations, and hence role, in this litigation. See Cohn v. Pacific Employers Ins.

Co. , supra, 213 Conn. at 546-47, 569 A.2d 544. Moreover, we note that these interlocutory appeals involve dozens of parties. Since the stipulation was filed almost eighteen months ago, no party has raised any issue or objection thereto. In light of the foregoing, the trial court on remand is strongly encouraged to consider, in its discretion, the extent of Old Republic's defense obligations pursuant to its excess policies prior to the commencement of the Phase III trial.

D

Duty to Defend Under Continental's 1965-1968 and 1977-1978 Excess Policies

Continental next asks this court to conclude that it has no duty to defend Vanderbilt under the excess insurance policies known as RDU 9433526 and RDX 3652404, which it issued to Vanderbilt for the periods of January 1, 1965 through January 1, 1968, and May 1, 1977 through March 3, 1978, respectively. For two reasons, we decline that request.

First, the record amply demonstrates that the trial court, in exercising its discretion to bifurcate the issues at trial, expressly deferred consideration of the claim now advanced by Continental in this interlocutory appeal. The following procedural history is relevant to our analysis. Prior to the commencement of the Phase I trial, Continental filed a motion for summary judgment, in which it sought, among other things, a declaration that the aforementioned excess policies "do not require [Continental] to pay Vanderbilt's defense costs relative to the underlying silica, talc, and/or asbestos related bodily injury suits ...." Vanderbilt opposed that motion with respect to policy RDU 9433526, raising multiple objections thereto. At the same time, Vanderbilt voiced no objection with respect to policy RDX

3652404. As it stated in its memorandum of law in opposition to Continental's motion for summary judgment, "Vanderbilt acknowledges that the [Continental] excess policy from May 1, 1977 to March 3, 1978 (RDX 3652404), contains language that clearly provides that [Continental] has no defense obligation to Vanderbilt." In these appeals, Vanderbilt has not articulated any opposition to Continental's claim with respect to policy RDX 3652404.

The trial court did not rule on Continental's motion for summary judgment. Rather, the court entered a series of orders that divided the proceeding into distinct phases. As the court recounted in its Phase I memorandum of decision, it "bifurcated the action so as to remove from the court trial any claim of damages for recovery from [Vanderbilt] for the overpayment of defense and indemnity costs." The court further stated: "To be clear, the issues of which parties were specifically obligated to provide for the defense ... of underlying claims in any particular time period is left to later phases of the trial."

The court also did not consider what it termed "issues of defense obligations" during the Phase II trial. As the court expressly indicated in its Phase II memorandum of decision, "such issues are beyond the limited scope of issues considered in Phase II and are better suited to be fully addressed in Phase III. That latter phase shall necessarily determine the scope of any defense ... obligations ...." Continental thereafter filed a motion to reargue the issue of its defense obligations under the RDU 9433526 excess policy. The court denied that motion, stating in relevant part that Continental's "motion asks the court to issue a ruling which the court has expressly deferred, for the purpose of judicial economy, until the completion of a later phase of the trial. The relief sought relative to the court's findings on the limited issues designated for consideration in Phase II seeks to elicit a substantive expansion of the scope of the court's findings."

It is well established that the trial court is vested with ample discretion to bifurcate a civil trial. "Pursuant to General Statutes § 52-205¹¹⁶ and Practice Book § 15-1,¹¹⁷ the trial court may order that one or more issues that are joined be tried before the others. The interests served by bifurcated trials are convenience, negation of prejudice and judicial efficiency.... Bifurcation may be appropriate in cases in which litigation of one issue may obviate the need to litigate another issue.... The bifurcation of trial proceedings lies solely within the discretion of the trial court." (Citation omitted; footnotes in original; internal quotation marks omitted.) Barry v. Quality Steel Products, Inc. , 263 Conn. 424, 448-49, 820 A.2d 258 (2003) ; see also Kervick v. Silver Hill Hospital , supra, 309 Conn. at 716, 72 A.3d 1044 (review of trial court decision on motion to bifurcate governed by abuse of discretion standard). "Accordingly, appellate review is limited to a determination of whether this discretion has been abused.... In reviewing claims that the trial court abused its discretion [in bifurcating certain issues at trial] the unquestioned rule is that great weight is due to the action of the trial court and every reasonable presumption should be given in favor of its correctness; the ultimate issue is whether the court could reasonably conclude as it did ...." (Citation omitted; internal quotation marks omitted.) Saczynski v. Saczynski , 109 Conn.App. 426, 428, 951 A.2d 670 (2008).

The record before us indicates that the trial court, in dividing this civil litigation into multiple phases, deliberately excised the "issues of defense obligations" from the Phase I and Phase II trials. The court further explained that those issues were "better suited to be fully addressed in Phase III" of the proceedings. At the time that it entered the bifurcation orders, the court was confronted by a veritable mountain of pleadings, motions, and requests. Brimming with a multitude of parties and intricate issues, this matter is the quintessence of complex litigation. In such instances, our rules of practice permit the trial court to "enter any appropriate order which facilitates the management of the complex litigation cases." (Emphasis added.) Practice Book § 23-14.

The court, in fashioning its bifurcation orders, emphasized that there was "a need to provide additional procedures ... to resolve [the parties'] claims and to promote convenience, negation of prejudice and judicial efficiency ...." The court also observed that the division of the matter into multiple phases "may obviate the need to litigate" certain remaining issues. Given the magnitude-both in terms of scope and substance-of the matters before it, we cannot conclude that the court abused its discretion in dividing this civil litigation into multiple phases.¹¹⁸ We likewise perceive no abuse of that discretion in the court's decision to defer consideration of the "issues of defense obligations" until the Phase III proceeding, and thus we refuse to disturb that determination on appeal.¹¹⁹

Continental's invitation for this court to decide the issue of its duty to defend under the excess policies in question is problematic for another reason, as it is procedurally improper. These interlocutory appeals were commenced pursuant to the strictures of Practice Book § 61-4. That rule of practice permits such appeals "only if the trial court makes a written determination that the issues resolved by the judgment are of such significance to the determination of the outcome of the case that the delay incident to the appeal would be justified, and the chief justice or chief judge of the court having appellate jurisdiction con curs...."¹²⁰ (Emphasis altered.) Practice Book § 61-4 (a). As one judge noted, "the purpose of § 61-4 is to create a narrow exception to our final judgment rule for those rare and special cases where interlocutory review of a trial court's pretrial ruling will resolve or greatly streamline the resolution of the entire case. In those limited circumstances, the purpose of the final judgment rule-to promote efficiency in the handling of cases by avoiding the added cost, delay and administrative burden of piecemeal litigation-is better served by granting the right to an immediate appeal than, as usual, postponing any appeal until the rights of all parties have been fully adjudicated in the trial court. Only if the trial judge, who knows the case personally and understands the interplay among its several claims, and the chief judge of the appellate court having jurisdiction, who knows the current status of his or her appellate docket, are mutually satisfied that the possible benefits of early appellate review exceed the likely costs and burdens of such review should the motion be granted." Share-America, Inc. v. Ernst & Young, LLP , Superior Court, judicial district of Waterbury, Docket No. CV-93-0150132-S (July 23, 1999) (Sheldon, J .) (25 Conn.L.Rptr. 160, 162 ).

In the present case, the issue of Continental's duty to defend under the excess insurance policies in question was not resolved by the trial court. Rather, the court deliberately and expressly deferred consideration of that issue, which we previously have concluded did not constitute an abuse of its discretion. The court likewise did not furnish a written determination that the issue was of such significance to merit the extraordinary review provided by Practice Book § 61-4. The issue of Continental's duty to defend under those excess policies, therefore, is not properly part of these interlocutory appeals. We therefore decline its request to decide that issue. Like Old Republic, with its somewhat related claim regarding its duty to defend, Continental will be able to pursue this issue on remand in the trial court.¹²¹ See, e.g., 36 DeForest Avenue, LLC v. Creadore , 99 Conn.App. 690, 705, 915 A.2d 916 ("[i]nsofar as [an] appeal is interlocutory, if the plaintiff wants to pursue this claim, there remains the opportunity to do so"), cert. denied, 282 Conn. 905, 920 A.2d 311 (2007).

VI

CONCLUSION

The rulings of the trial court are reversed only with respect to the determinations that (1) Vanderbilt is responsible for defense costs for the period of March 3, 1993 through April 24, 2007, (2) a default date of first exposure of January 1, 1962, applies to pending and future claims, and (3) the occupational disease exclusions in certain secondary policies apply only to claims brought by Vanderbilt's own employees; the proper allocation methodology and the prospective application of that methodology are clarified as set forth herein; and the case is remanded for further proceedings consistent with this opinion. The rulings are affirmed in all other respects. APPENDIX A

1 The action was filed by R.T. Vanderbilt Company, Inc. During the trial court proceedings, the court granted that company's motion to substitute its successor, Vanderbilt Minerals, LLC, as the party plaintiff. For convenience, we refer to both entities as "Vanderbilt" throughout this opinion.

2 Moreover, unlike most appeals, this case does not come to us with a fully developed factual record and a final judgment. This presents special challenges in crafting an opinion and can be expected to present special challenges to the trial court, which will be tasked with applying the principles set forth herein on remand and during the next phase of the proceedings.

3 See footnote 101 of this opinion and accompanying text. Many of the secondary policies follow form to the Hartford and Continental primary policies. See footnote 91 of this opinion.

4 The operative complaint is the modified seventh amended complaint.

5 Throughout this opinion, we use the terms "long latency," "long-tail," and "progressive injury" interchangeably. Those terms refer to the fact that toxic tort claims typically allege that exposure to toxins such as asbestos causes a series of continuing, indivisible injuries that develop gradually over time but may not manifest for many years. See State v. Continental Ins. Co., 55 Cal.4th 186, 195-96, 281 P.3d 1000, 145 Cal.Rptr.3d 1, as modified, 2012 Cal. LEXIS 8788 (September 19, 2012); Y. Colon, "Pay It Forward: Allocating Defense and Indemnity Costs in Environmental Liability Cases in California," 14 Cal. Ins. L. & Reg. Rep., no. 3, 2002, n.24 and accompanying text.

6 See footnote 28 of this opinion.

7 Although Vanderbilt lost those policies as well, Continental stipulated at trial to their existence.

8 See footnote 54 of this opinion.

9 Everest filed an immediate appeal from the trial court's Phase I and Phase II rulings on the ground that the rulings constituted a final judgment as to it. Vanderbilt and other defendants were subsequently granted permission to file interlocutory appeals pursuant to Practice Book § 61-4 (a), which provides in relevant part that an interlocutory ruling is considered to be an appealable final judgment when "the trial court makes a written determination that the issues resolved by the judgment are of such significance to the determination of the outcome of the case that the delay incident to the appeal would be justified, and the chief justice or chief judge of the court having appellate jurisdiction concurs...."

For purposes of briefing, Vanderbilt has been designated as the appellant, and defendants Mt. McKinley and Everest have been designated as the lead cross appellants-appellees. The other defendants have been designated as cross appellants-appellees.

We also note that approximately two years after the commencement of these appeals, Clearwater Insurance Company was substituted for Mt. McKinley Insurance Company. TIG Insurance Company thereafter was substituted for Clearwater Insurance Company. For clarity, we refer to that party as Mt. McKinley throughout this opinion.

10 Many of the claims raised on cross appeal have been joined by multiple defendants. For clarity and brevity, we refer in this opinion only to the primary party raising each issue, except in those instances when other parties have independently addressed an issue in such a manner as to merit discrete treatment, or when otherwise necessary for purposes of this opinion. For a complete list of which parties have joined in which appellate claims, see Appendix A of this opinion.

11 See Travelers Casualty & Surety Co. of America v. Netherlands Ins. Co., 312 Conn. 714, 740, 95 A.3d 1031 (2014) ("If the policy is ambiguous, extrinsic evidence may be introduced to support a particular interpretation.... [I]f an ambiguity arises that cannot be resolved by examining the parties' intentions ... the ambiguous language should be construed in accordance with the reasonable expectations of the insured when he entered into the contract." [Internal quotation marks omitted.] ); Israel v. State Farm Mutual Automobile Ins. Co., supra, 259 Conn. at 510-11, 789 A.2d 974 (applying other canons of construction prior to construing policy in favor of drafter); Fiallo v. Allstate Ins. Co., 138 Conn.App. 325, 340, 51 A.3d 1193 (2012) (collecting sources); 2 S. Plitt et al., Couch on Insurance (3d Ed. Rev. 2010) § 22:16, pp. 22-93 through 22-94 ("since the rule of strict construction of an ambiguous policy against insurer is a rule of last resort, and not to be permitted to frustrate parties' expressed intention if such intention could be otherwise ascertained, where there is extrinsic evidence of parties' intention, which is proffered and admissible, and which resolved ambiguity, albeit in favor of noncoverage, the rule of strict construction need not be applied"); M. Taylor et al., Connecticut Insurance Law (2011) § 2-5:1, p. 35 ("[o]nce a determination is made that the policy is ambiguous, then the court may consider any relevant evidence which demonstrates the intent of the parties at the time that they entered into the policy"); but see Lexington Ins. Co. v. Lexington Healthcare Group, Inc., supra, 311 Conn. at 77, 84 A.3d 1167 (Eveleigh, J., concurring and dissenting) (arguing that, at least in some instances, rule should be applied in lieu of consulting extrinsic evidence of parties' intentions); 1 New Appleman on Insurance Law, Library Edition (J. Thomas & F. Mootz III eds., 2011) § 5.02, p. 5-7 (similar).

12 See footnote 10 of this opinion.

13 See footnote 52 of this opinion.

14 Mt. McKinley does not take any position as to which trigger of coverage theory should apply to claims alleging asbestosis and other noncancer diseases.

15 We note that none of the parties to the present dispute expressly identified the trial court's adoption of the continuous trigger theory as a distinct appellate issue in their briefs. See Practice Book § 67-4. Nevertheless, we conclude that the question of which trigger theory governs the present appeals is properly before this court because (1) it has been briefed and argued by the parties, and (2) its resolution is a necessary prerequisite to our resolution of other issues that the parties have expressly raised, including the admissibility of Dr. Kratzke's testimony and whether the trial court applied the correct availability rules, allocation formula, and default date of first exposure. See Stein v. Tong, 117 Conn.App. 19, 21 n.1, 979 A.2d 494 (2009) ; United Technologies Corp. v. American Home Assurance Co., 989 F.Supp. 128, 152 (D. Conn. 1997).

16 We note that scores of different insurance policies are at issue in this case and, although the policies are generally similar in most respects material to the present appeals, there are minor variations in the policy language and terms. We address those variations herein only to the extent that the parties have identified them as potentially relevant or we otherwise deem them to be so.

17 Implicit in the latter conclusion is the fact that there may be no practical difference between these two trigger theories with respect to the property damage at issue in Netherlands. Unlike the situation with asbestos related disease, where there is a lengthy premalignancy latency period and an individual arguably can be exposed to asbestos fibers without suffering any injury; but see part III A 3 of this opinion; the Supreme Court in Netherlands may have been operating under the assumption that building damage begins to occur from the moment that water ingress begins, so that exposure and injury are contemporaneous. See GenCorp, Inc. v. AIU Ins. Co., 104 F.Supp.2d 740, 742, 746-48 (N.D. Ohio 2000).

18 We do not foreclose the possibility that, with respect to other types of long-tail losses, it might be both possible and practical to determine, as a factual matter, what portion of the injury occurs during each policy period. See Owens-Illinois, Inc. v. United Ins. Co., supra, 138 N.J. at 458-59, 650 A.2d 974 ; M. Doherty, "Allocating Progressive Injury Liability Among Successive Insurance Policies," 64 U. Chi. L. Rev. 257, 260 n.14 (1997).

19 During Phase II of the trial, the court denied a motion by Mt. McKinley renewing its proffer.

20 To the extent that the trial court excluded Kraztke's testimony solely on the basis of the court's belief that Security rendered any expert testimony on the trigger issue irrelevant, we affirm the ruling of the court on the alternative grounds presented by Vanderbilt. See Pelletier Mechanical Services, LLC v. G & W Management, Inc., 162 Conn.App. 294, 301, 131 A.3d 1189, cert. denied, 320 Conn. 932, 134 A.3d 622 (2016).

21 We assume without deciding that Mt. McKinley is correct that it would be improper for this court to adopt, as a legal fiction, a trigger of coverage theory that was incompatible with current scientific knowledge regarding the nature of the injuries at issue. We also note that Mt. McKinley's arguments with respect to the Kratzke testimony are limited to asbestos related cancers ; Mt. McKinley does not contend that it would be improper to apply a continuous trigger theory to claims alleging asbestosis.

22 See Eagle-Picher Industries, Inc. v. Liberty Mutual Ins. Co., supra, 682 F.2d at 18 (noting different perspectives of research scientists and treating physicians); Ins. Co. of North America v. Forty-Eight Insulations, Inc., supra, 633 F.2d at 1218 (similar); Zurich Ins. Co. v. Raymark Industries, Inc., 145 Ill.App.3d 175, 183, 494 N.E.2d 634, 98 Ill.Dec. 512 (1986) (explaining that, for cellular pathologists, "injury" includes physical or chemical damage that may be detectable only on microscopic or subclinical level, including the alteration of cellular structure or function, and disease is simply ongoing process of repairing or reacting to such injuries, whereas clinicians define "injury" in terms of some kind of noticeable harm), aff'd, 118 Ill.2d 23, 112 Ill.Dec. 684, 514 N.E.2d 150 (1987).

23 See Stonewall Ins. Co. v. Asbestos Claims Management Corp., 73 F.3d 1178, 1196 (2d Cir. 1995) ("most courts that have analyzed the issue have found the continuous injury trigger of coverage applicable to the standard [occurrence based comprehensive general liability] policy" [internal quotation marks omitted] ), modified on denial of reh'g, 85 F.3d 49 (2d Cir. 1996) ; Montrose Chemical Corp. v. Admiral Ins. Co., 10 Cal.4th 645, 677, 913 P.2d 878, 42 Cal.Rptr.2d 324 (1995), modified on denial of reh'g (August 31, 1995) (similar); J. Michaels et al., supra, 64 U. Kan. L. Rev. 471-72 (similar).

24 See Ins. Co. of North America v. Forty-Eight Insulations, Inc., supra, 633 F.2d at 1218 ; Owens-Illinois, Inc. v. United Ins. Co., supra, 138 N.J. at 453-54, 650 A.2d 974.

25 By contrast, the injury-in-fact theory championed by Mt. McKinley, under which no coverage obligations attach until malignancy emerges, ignores the overwhelming weight of authority acknowledging that bodily injury occurs throughout the period that asbestos is lodged in the lung tissue. See Lloyd E. Mitchell, Inc. v. Maryland Casualty Co., 324 Md. 44, 62, 595 A.2d 469 (1991) ; Owens-Illinois, Inc. v. United Ins. Co., supra, 138 N.J. at 454, 650 A.2d 974. Other trigger theories such as initial exposure and manifestation likewise turn a blind eye to the realities of asbestos related disease: during all stages of the disease process, a person's body is progressively compromised by the toxin to which he or she has been exposed.

26 Both of the amici-the United Policyholders and the Complex Insurance Claims Litigation Association-have submitted briefs addressed to this issue. Although we have considered the arguments of the amici, we do not address them herein except to the extent that they overlap or bolster the arguments presented by the parties. See Dow & Condon, Inc. v. Brookfield Development Corp., 266 Conn. 572, 595, 833 A.2d 908 (2003).

27 Courts and commentators generally have recognized that comprehensive general liability insurance was unavailable for companies engaged in asbestos related businesses after 1985. Because Vanderbilt's relevant policies did not expire until March 3, 1986, however, the issue in the present case is most accurately stated as whether such insurance was available to Vanderbilt after March 3, 1986. Nevertheless, for purposes of brevity we will refer simply to the availability of insurance during the "post-1985" and "pre-1986" periods.

28 A claims-made policy provides coverage only for those claims made during the policy period. By contrast, an occurrence based policy provides coverage for any injuries that take place during the policy period, even if the claim arising from those injuries is not made until long after the policy period has ended. A claims-made policy thus allows an insurer both to cabin its potential liability and to better predict its long-term exposure relative to occurrence based policies, which afford "almost unlimited prospective coverage" up to the policy limits. (Internal quotation marks omitted.) Security Ins. Co. of Hartford v. Lumbermens Mutual Casualty Co., supra, 264 Conn. at 692 n.5, 826 A.2d 107.

29 Compare Chemical Leaman Tank Lines, Inc. v. Aetna Casualty & Surety Co., 177 F.3d 210, 231 (3d Cir. 1999), Stonewall Ins. Co. v. Asbestos Claims Management Corp., supra, 73 F.3d at 1204, Keene Corp. v. Ins. Co. of North America, supra, 667 F.2d at 1058 (Wald, J., concurring in part), Pneumo Abex Corp. v. Maryland Casualty Co., Civil Action No. 82-2098 (JGP), 2001 WL 37111434, *10 (D.D.C. October 9, 2001), Hercules, Inc. v. AIU Ins. Co., 784 A.2d 481, 493 (Del. 2001), Baltimore v. Utica Mutual Ins. Co., 145 Md.App. 256, 313-14 n.54, 802 A.2d 1070 (2002), cert. dismissed by petitioner before oral argument, 2003 Md. LEXIS 176 (Md. March 6, 2003) (decision without published opinion, 374 Md. 81, 821 A.2d 369 [2003] ), Wooddale Builders, Inc. v. Maryland Casualty Co., 722 N.W.2d 283, 297 (Minn. 2006), and Owens-Illinois, Inc. v. United Ins. Co., supra, 138 N.J. at 479, 650 A.2d 974 (adopting unavailability rule), with Sybron Transition Corp. v. Security Ins. of Hartford, 258 F.3d 595, 600 (7th Cir. 2001), AAA Disposal Systems, Inc. v. Aetna Casualty & Surety Co., 355 Ill.App.3d 275, 287-88, 821 N.E.2d 1278, 290 Ill.Dec. 704 (2005), leave to appeal denied, 2005 Ill. LEXIS 296 (Ill. January 26, 2005) (decision without published opinion, 213 Ill.2d 553, 829 N.E.2d 786, 293 Ill.Dec. 861 [2005] ), Boston Gas Co. v. Century Indemnity Co., 454 Mass. 337, 371, 910 N.E.2d 290 (2009), Crossmann Communities of North Carolina, Inc. v. Harleysville Mutual Ins. Co., 395 S.C. 40, 66 and n.16, 717 S.E.2d 589 (2011), and Bradford Oil Co. v. Stonington Ins. Co., 190 Vt. 330, 342, 54 A.3d 983 (2011) (rejecting rule); see also Keyspan Gas East Corp. v. Munich Reinsurance America, Inc., 143 A.D.3d 86, 93-95, 37 N.Y.S.3d 85 (2016) (outcome depends on specific policy language).

30 The other cases to which Mt. McKinley directs our attention reject the unavailability rule for substantially the same reasons as did Boston Gas and Sybron. See footnote 29 of this opinion.

31 Throughout this part of the opinion, we refer only to those injuries alleged to have resulted from a first exposure to Vanderbilt's talc prior to 1986 (or, more generally, during an insured period) but to have manifested after 1985 (or, more generally, to have spanned a period during which insurance was unavailable). The parties agree that Vanderbilt is solely responsible for injuries allegedly arising from an initial exposure to its talc after 1985.

32 We reiterate that we make this assumption merely for purposes of analysis. We take no position on the factual question of the extent to which the risks allegedly associated with asbestos or talc were known or foreseeable prior to the mid-1980s, when insurance coverage for those products ceased to be generally available.

33 See, e.g., Owens-Illinois, Inc. v. United Ins. Co., supra, 138 N.J. at 475-76, 650 A.2d 974 (prorating costs by time-on-the-risk and degree of risk assumed).

34 The other cases cited by Vanderbilt as evidence that our sister courts have not applied an equitable exception are not instructive, as they involve policyholders who discovered, years after comprehensive general liability policies began incorporating a pollution exclusion; see part IV A of this opinion; that their facilities had been inadvertently polluting the natural environment. See, e.g., Champion Dyeing & Finishing Co. v. Centennial Ins. Co., supra, 355 N.J.Super. 262, 810 A.2d 68.

35 For purposes of this discussion, we assume, solely for the sake of argument, that Vanderbilt's talc may have contained asbestos or had the potential to cause asbestos related injuries. We recognize that Vanderbilt fervently denies this, but also that juries in certain of the underlying actions have found against Vanderbilt. Because this factual question was not before the trial court, we must consider the proposed equitable exception in light of both possibilities.

36 For example, the court might conclude that, for any underlying action for which the court's chosen method would result in a default date of first exposure less than five years prior to March 3, 1986, Vanderbilt would bear the burden of proving that initial exposure did in fact occur during the time when the company was insured.

37 Vanderbilt also argues that it would be improper to apply an equitable exception in the present case because the relevant policies did not contain exclusions precluding coverage in the event that Vanderbilt continued to sell its talc. This argument misses the mark. None of the elements of the pro rata allocation approach that this court and our Supreme Court have adopted-continuous trigger, time-on-the-risk allocation, proration to the insured, or the unavailability rule-appear as provisions in the standard form comprehensive general liability policies at issue in this case. Rather, it is precisely because those policies failed to anticipate and provide an allocation methodology for long-tail toxic tort claims potentially implicating multiple policies that courts have been compelled to develop rules for distributing defense and indemnity obligations. The proposed equitable exception is no different.

38 During the Phase I trial, Vanderbilt acknowledged that it had lost the relevant policies covering these years and, on appeal, it does not challenge the trial court's conclusion that it must be treated as self-insured and allocated a pro rata share of defense costs for the period from January 1, 1948 through December 31, 1955.

39 Vanderbilt contends that, even if we uphold the trial court's determination that it was liable for some defense costs after 1993, the trial court improperly determined that its liability terminated on April 24, 2007 (the end of the extended reporting period), rather than on April 24, 2004 (the end of the initial policy term). Because we conclude that Vanderbilt was not liable for any post-1985 defense costs, we need not resolve this claim. For the same reason, we need not consider Vanderbilt's argument that insurance should not have been deemed available between 1993 and 2007 for purposes of the underlying actions because its ability to obtain limited coverage during that period was predicated on its representation to its insurers that its products did not contain asbestos.

40 Vanderbilt does contend, however, that it would be perverse to "punish" an insured who, through extraordinary efforts, is able to obtain limited coverage during a time when such coverage is generally unavailable. To construe that limited success as evidence that insurance was available to the particular policyholder, and then allocate a pro rata share of costs to the policyholder because it failed to obtain even more coverage, would, Vanderbilt argues, discourage such efforts and reduce the total resources available to respond to long-tail injuries.

41 Silicosis is "[a] form of pneumoconiosis resulting from occupational exposure to and inhalation of silica dust over a period of years; characterized by a slowly progressive fibrosis of the lungs, which may result in impairment of lung function ...." Stedman's Medical Dictionary (28th Ed. 2006) p. 1773; see also McClain v. Metabolife International, Inc., 401 F.3d 1233, 1239 (11th Cir. 2005) (discussing both "toxins like asbestos, which causes asbestosis and mesothelioma," and "silica, which causes silicosis"). Furthermore, the insurance industry has crafted distinct exclusions for silica and asbestos related injuries. See, e.g., Liberty Mutual Ins. Co. v. Lone Star Industries, Inc., 290 Conn. 767, 811, 967 A.2d 1 (2009) (noting that policy in question contained both asbestos exclusion and silicosis exclusion).

42 Bendure did not elaborate on precisely what "generally available" constituted, apart from confirming that, in his opinion, the issuance of even one such policy in a given year could meet that metric.

43 The trial court found that the parties had furnished no evidence that occurrence based coverage was available to Vanderbilt after 1986 that "it elected not to purchase." The court also found that Vanderbilt "neither elected to decline to purchase available [occurrence based] indemnity insurance, nor did it purchase an insufficient amount of (available) insurance."

44 The trial court found, and no party to these appeals disputes, that "[f]rom 1999-2003, [Vanderbilt] annually obtained approximately $100 million in claims-made coverage with asbestos exclusions."

45 As Bendure acknowledged in his testimony, more than 2000 asbestos related claims had been filed against Vanderbilt.

46 Because a necessary factual predicate to Mt. McKinley's claim is lacking-namely, a finding by the court that occurrence based coverage was available to Vanderbilt during the years in question-we do not further consider the substantive merits of Mt. McKinley's claim regarding the contours of the unavailability of insurance rule. Specifically, we express no opinion as to Mt. McKinley's assertion that the general rules regarding the expansive obligations of insurers to defend actions in which any of the allegations of the complaint potentially fall under the scope of the policy also apply to a policyholder deemed to be self-insured for purposes of pro rata allocation.

47 Consider, for example, a hypothetical claimant who alleges that she developed mesothelioma in 1962 but for whom a date of first exposure cannot be established. Assume that her claim settles for $1.5 million. Under the methodology advocated by the secondary insurers, the court would apply a default date of first exposure of 1948 and would allocate the indemnity payment over fifteen years, from 1948 through 1962. Vanderbilt would be responsible for the 1948 through 1955 period, during which it was effectively self-insured, and Continental for the 1956 through 1961 period, for which it now has stipulated that it provided coverage, as well as for 1962. Under that approach, only $100,000 (one-fifteenth of $1.5 million) of the settlement would be allocated to Continental's $300,000 1962 policy. By contrast, under the primary insurers' allocation agreement, all $1.5 million would have been allocated to the 1962 policy, resulting in its exhaustion and leaving a substantial remainder for the 1962 excess carriers.

48 After the court issued its Phase II decision, Mt. McKinley moved for reargument, reconsideration, or clarification of that decision, seeking, among other things, an explanation as to why the court had elected to apply the default date of January 1, 1962, on a prospective basis. The court denied this request.

49 Apart from Huffer's testimony, the court also was presented with Posner's expert testimony on the issue. Posner also indicated that, in his opinion, "the allocation performed by Hartford and Continental was reasonable under the customs and practices in the industry" at that time.

50 Mt. McKinley argues that the court ignored Continental's prior acknowledgments that it had issued coverage to Vanderbilt from 1956 to 1962. That claim is untenable. The court specifically found that "at the time that the allocation agreement was executed, the existence of those [pre-2002] policies and any obligations thereunder was an unresolved and disputed factual issue, which was not determined until addressed by this court" in the Phase I and II decisions.

That finding is substantiated by evidence in the record before us. Peter A. Pogue, a claims consultant who administered asbestos claims on behalf of Continental, testified that 1962 was "the date [of] the very first policy that [Continental] had a complete copy of" and reflected the "first confirmed coverage." Lawrence Farber, an assistant vice president in Hartford's complex claims group, similarly testified that Continental's "first confirmed coverage" was under the 1962 policy. On that testimony, the court found that the 1962 policy "was the first policy that [Continental] had a complete copy [of] to confirm and verify coverage." The record contains no evidence to the contrary.

51 Continental first suggested such an approach in a footnote to its Phase II posttrial brief, but the trial court did not address the proposal in its decision.

52 For certain of the errors alleged in this part, we agree with the parties as to the proper approach to calculating Vanderbilt's share of the costs, but it is unclear to us whether the trial court departed or intended to depart from that approach. In light of the unique procedural posture of this case, we think the most prudent path is simply to explain what approach the trial court is to employ on remand and during the next phases of the trial without first attempting to determine whether the court actually strayed from that approach.

53 Although throughout this section we refer to the allocation block in terms of months or years for purposes of brevity and clarity, the precise coverage block likely will have to be calculated in terms of days because the full allocation block-at least with respect to Vanderbilt's occurrence based policies-is presumed to begin on January 1, 1948, and to end on March 3, 1986.

54 To the extent that the court intended to calculate the number of months between January 1, 1948, and December 31, 2008, which it presumed to be the respective dates on which Vanderbilt first and last sold talc products, the length of the sixty-one year allocation block should have been 732 months rather than 720. Any such miscalculation is immaterial, however, in light of our determination that the maximum allocation block does not extend to 2008.

55 In a posttrial motion, Mt. McKinley asked the trial court also to consider the effect on Vanderbilt's pro rata liabilities of the court's determination that the umbrella policies Continental issued between 1968 and 1977 did not cover defense costs. The court denied this request.

56 In referring to Vanderbilt's share of the maximum allocation block, we do not preclude the possibility that the company's share of the total allocation might rise in the future if, for example, an additional insurer were to become insolvent.

57 We note that Vanderbilt was able to obtain limited claims-made coverage between 1993 and 2007 conditioned on its representation that its talc products did not contain asbestos. The question of whether those policies nevertheless are available to respond to the underlying actions and should, therefore, be included in the allocation block is not presently before us.

58 Assume for example that an underlying action alleging a date of first exposure in March, 1977, and filed in 2003, resulted in a settlement of $10.1 million, and that the policyholder purchased a $100,000 primary claims-made policy for 2003 but that additional insurance was unavailable. Under those circumstances, $100,000 would be allocated to 2003 (assuming the policy limits were otherwise untapped) and $1 million would be allocated to each of the policy years from March, 1977 through March, 1986. Although the policyholder would not be responsible for any residual for 2003, it would be responsible for any residual in the remaining years after all primary and secondary policies were exhausted.

59 Our sister courts are divided as to the general question of which party bears the burden of establishing the availability or unavailability of insurance for purposes of pro rata allocation of long-tail costs. Compare Decker Mfg. Corp. v. Travelers Indemnity Co., 106 F.Supp.3d 892, 898 (W.D. Mich. 2015) ("[t]he insured bears the burden of proving that insurance was not reasonably available to it"), with Chemical Leaman Tank Lines, Inc. v. Aetna Casualty & Surety Co., 177 F.3d 210, 231 (3d Cir. 1999) ("the insurers should bear the burden of proving that insurance coverage was available"). Recognizing that the burden of proof in an insurance coverage dispute does not always follow ordinary contract law; M. Taylor et al., Connecticut Insurance Law (2d Ed. 2013) p. 71; and that the doctrine at issue is an equitable one, we believe that the most reasonable approach is to require the policyholder to prove that it was unable to obtain asbestos coverage prior to 1986, when such insurance was generally available, but to require the insurer to prove that insurance was available to the policyholder after 1985, when exclusion clauses were incorporated into the standard form comprehensive general liability policy. In other words, the widely accepted fact that occurrence based asbestos coverage was generally available through 1985 but not thereafter creates a rebuttable presumption that the situation was no different with respect to any individual policyholder. This rule does not require either party to prove a negative and, in both instances, places the burden on the party that is arguably best positioned to make the proof. See Buell Industries, Inc. v. Greater New York Mutual Ins. Co., 259 Conn. 527, 551, 791 A.2d 489 (2002) ; Fiallo v. Allstate Ins. Co., 138 Conn.App. 325, 349, 51 A.3d 1193 (2012) (Borden, J., concurring).

60 Mt. McKinley's counterclaim states in relevant part that "[t]o the extent that Mt. McKinley [is] held to have any coverage obligations to [Vanderbilt] ... with respect to any underlying claims ... then [Mt. McKinley seeks] a declaration that [Vanderbilt] should be allocated a pro rata share of the defense and indemnity costs for each underlying claim at issue ... because [Vanderbilt] failed to obtain insurance ... and/or knowingly continued to mine and distribute talc ... and/or because [Vanderbilt] has lost or is missing its insurance policies ... and/or because [Vanderbilt] was uninsured, underinsured or self-insured, has insolvent insurer policy periods and/or periods where the insurer(s) has no defense obligation.... To the extent that any of the defense or indemnity costs for any of [Vanderbilt's] previously resolved underlying claims at issue were allocable to [Vanderbilt] and [Vanderbilt] failed to pay such defense or indemnity costs, then [Mt. McKinley seeks] a declaration that those amounts are reallocated to [Vanderbilt], requiring [Vanderbilt] to reimburse the insurer(s) that previously paid such costs ...."

61 Continental also submits, in a footnote to its principal appellate brief that is devoid of any analysis or citation to legal authority, that, despite the court's intent to apply its findings prospectively, Continental "still has a claim for reimbursement" because it "has been paying more than its allocable share of Vanderbilt's defense and indemnity within the 1962-1986 allocation block ...." We decline to consider that bald assertion. See Knapp v. Knapp, 270 Conn. 815, 823 n.8, 856 A.2d 358 (2004) ("We consistently have held that [a]nalysis, rather than mere abstract assertion, is required in order to avoid abandoning an issue by failure to brief the issue properly.... [A]ssignments of error which are merely mentioned but not briefed beyond a statement of the claim will be deemed abandoned and will not be reviewed by this court.... Where the parties cite no law and provide no analysis of their claims, we do not review such claims." [Internal quotation marks omitted.] ).

62 In its counterclaim, Continental claimed that "Vanderbilt is ... liable to contribute to and/or indemnify and reimburse Continental for the disproportionate share of liability Continental has paid on Vanderbilt's behalf." American International similarly sought in its counterclaim reimbursement "for the full amount of the per occurrence deductible which [it] has paid on Vanderbilt's behalf."

63 As the trial court found in part III C 2 of its Phase II decision, "[t]o effectuate a reallocation to include, for example, the period 1956-1962 would require the court and the parties to engage in mathematical calculations that not only would be extremely arduous, time-consuming and to some degree subjective, they would also be of marginal utility and ultimately undermine the prior significant efforts of the parties to compromise their differences over the allocation of the payments between themselves and which were made for the benefit of their insured. Beyond the issue of such recalculations, to reopen the limits of the allocation agreement would work contrary to this state's public policy in favor of the settlement of civil litigation."

Finding that the allocation agreement between Continental and Hartford "was reasonable at the time it was entered into ... and was taken in good faith," the court concluded that it "will not force those parties to reallocate" the millions of dollars that those primary insurers already had paid to resolve thousands of underlying actions over the course of several decades. We have affirmed the propriety of that determination in part III C 1 of this opinion.

64 For this reason, we are not persuaded by Vanderbilt's reliance on an isolated remark of the court regarding the challenging nature of allocating past defense and indemnity payments. As support for its interpretation of the court's prospectivity ruling, Vanderbilt points to the court's observation, in the exhaustion section of its Phase II decision, that requiring "courts and litigants to go back and recalculate precisely what amounts were paid on which claims, during which periods of time, whether allocation to that time frame was appropriate, and whether the payments were made reasonably and in good faith [would be] an undertaking [that] would lie between arduous and Sisyphean." That commentary, however, was made not with respect to the general retroactive application of allocation rules but, rather, in the specific context of the court's determination that the primary insurance policies issued by Continental and Hartford had been exhausted and, in particular, that the 2002 allocation agreement was enforceable.

65 Vanderbilt argues that "reimbursement of past defense and indemnity costs by Vanderbilt, when Vanderbilt did not control the decision-making, would be entirely unjustified." Vanderbilt has provided neither legal authority nor analysis to substantiate that bald assertion. "We repeatedly have stated that [w]e are not required to review issues that have been improperly presented to this court through an inadequate brief.... Analysis, rather than mere abstract assertion, is required in order to avoid abandoning an issue by failure to brief the issue properly." (Internal quotation marks omitted.) Taylor v. Mucci, 288 Conn. 379, 383 n.4, 952 A.2d 776 (2008) ; see also Northeast Ct. Economic Alliance, Inc. v. ATC Partnership, 272 Conn. 14, 51 n.23, 861 A.2d 473 (2004) ("[i]nasmuch as the plaintiffs' briefing of the ... issue constitutes an abstract assertion completely devoid of citation to legal authority or the appropriate standard of review, we exercise our discretion to decline to review this claim as inadequately briefed"); Russell v. Russell, 91 Conn.App. 619, 635, 882 A.2d 98 (parties must analyze relationship between facts of case and applicable law), cert. denied, 276 Conn. 924, 925, 888 A.2d 92 (2005). We therefore do not further consider Vanderbilt's claim in this regard.

66 Philburn testified that Mt. McKinley so advised Vanderbilt through the issuance of "reservation of rights letters." At trial, multiple such letters were introduced into evidence. Philburn confirmed that the issuance of reservation of rights letters was standard practice in responding to claims tendered to Mt. McKinley by Vanderbilt.

67 It cannot be forgotten that these are interlocutory appeals, in which the trier of fact has not yet made any determinations regarding the applicability or reach of Mt. McKinley's reservation of rights. In both its counterclaim and a cross claim, Mt. McKinley sought "a proper allocation of the loss or damages among [Vanderbilt], any insured, [Mt. McKinley] and all other insurers," and asserted "a claim for contribution, reimbursement, indemnification, set off, subrogation, equitable relief and/or any other appropriate relief ... to the fullest extent permitted by law." Mt. McKinley's reservation of rights bears directly on its claims for reimbursement. The court's dedication of the Phase III proceeding to that issue suggests that Mt. McKinley's reservation of rights properly will be a subject thereof.

68 See footnote 10 of this opinion.

69 See footnote 16 of this opinion.

70 Although the insurance policies in Heyman contained nonstandard pollution exclusion clauses, the present defendants contend that the Supreme Court's analysis in that case applies to and governs all of the pollution exclusion clauses at issue in this case.

71 Buell Industries, Inc. v. Greater New York Mutual Ins. Co., supra, 259 Conn. at 530-32 and nn.1-2, 791 A.2d 489, and Schilberg Integrated Metals Corp. v. Continental Casualty Co., supra, 263 Conn. at 247-48, 248 n.1, 819 A.2d 773, may be distinguished on similar grounds. In Buell Industries, Inc., in which a pollution exclusion was held to bar coverage, an environmental hazardous waste audit revealed that waste oil and degreasing chemicals had leaked from a company's wastewater lagoon into neighboring groundwater. In Schilberg, the Pennsylvania Department of Environmental Resources (now Department of Environmental Protection) brought an administrative action when soil and groundwater contamination were discovered near an unauthorized hazardous waste disposal facility. Both cases, like Heyman, represent paradigmatic examples of environmental pollution.

72 See American States Ins. Co. v. Koloms, 177 Ill.2d 473, 490-91, 687 N.E.2d 72, 227 Ill.Dec. 149 (1997).

73 In each instance, emphasis is as in the original and information regarding pronunciations, word origins, and parts of speech has been omitted.

74 See, e.g., Maryland Casualty Co. v. W.R. Grace & Co., 794 F.Supp. 1206, 1229 (S.D.N.Y. 1991), rev'd in part on other grounds, 23 F.3d 617 (2d Cir. 1993), cert. denied, 513 U.S. 1052, 115 S.Ct. 655, 130 L.Ed.2d 559 (1994) ; West American Ins. Co. v. Tufco Flooring East, Inc., 104 N.C.App. 312, 322, 409 S.E.2d 692 (1991), overruled in part on other grounds by Gaston County Dyeing Machine Co. v. Northfield Ins. Co., 351 N.C. 293, 303, 524 S.E.2d 558 (2000).

75 Even if we were to conclude that asbestos is an irritant within the scope of the policy exclusion, we still hold that the pollution exclusion does not bar coverage pursuant to the dispersal clause. See part IV A 1 b iii of this opinion.

76 Also, we observe that later iterations of the pollution exclusion, like the one at issue in Heyman, make explicit the fact that the various enumerated irritants and contaminants listed in the clause are all merely intended to be examples or categories of pollutants. National Casualty's 1985 policy, for example, provides in relevant part: "[This policy shall not apply] ... to 'bodily injury' or 'property damage' arising out of the actual, alleged or threatened discharge, dispersal, release or escape of pollutants .... Pollutants means any solid, liquid, gaseous or thermal irritant or contaminant, including smoke, vapor, soot, fumes, acids, alkalis, chemicals and waste material."

Although the terms of one defendant's 1985 policy certainly cannot dictate the meaning of policies issued a decade earlier by different defendants, this revision to or clarification of the standard form policy language certainly is consistent with the conclusion that, from the outset, the pollution exclusion clause was addressed principally to pollution.

77 We recognize that, in Drown, a majority of our Supreme Court cautioned that "[t]he title of an insurance policy cannot ... be used to create ambiguity within the plain and unambiguous terms of the contract." Connecticut Ins. Guaranty Assn. v. Drown, supra, 314 Conn. at 191 n.14, 101 A.3d 200. In the present case, however, in which we already have determined that at least one key policy term is ambiguous as applied, the title of the pollution exclusion is merely one of many factors we consider in assessing the scope and extent of that ambiguity.

78 We note that not all of the underlying complaints expressly allege aerial dispersal or inhalation of asbestos. Because we answer this question in the negative, we need not determine whether the pollution exclusion even potentially applies to claims alleging that asbestos dust was ingested or absorbed through means other than inhalation. See Continental Casualty Co. v. Rapid-American Corp., supra, 80 N.Y.2d at 654, 593 N.Y.S.2d 966, 609 N.E.2d 506 (noting that asbestos fibers can be transmitted by direct contact with clothing or skin, rather than through inhalation).

79 In Yale University v. Cigna Ins. Co., supra, 224 F.Supp.2d at 402, on which the defendants rely, the District Court failed to consider whether the third element of the exclusion was satisfied, focusing solely on the question of whether asbestos dust constitutes a pollutant, irritant, or contaminant. It is not even clear whether the policy at issue in that case included the "atmosphere" language. See id., at 421.

80 See, e.g., Lumbermens Mutual Casualty Co. v. S-W Industries, Inc., 39 F.3d 1324, 1336 (6th Cir. 1994) ; MacKinnon v. Truck Ins. Exchange, supra, 31 Cal.4th at 650-51, 3 Cal.Rptr.3d 228, 73 P.3d 1205 ; Belt Painting Corp. v. TIG Ins. Co., supra, 100 N.Y.2d at 387-88, 763 N.Y.S.2d 790, 795 N.E.2d 15 ; see also In re Asbestos Products Liability Litigation (VI), Docket No. Civ. A. 96-968, 1997 WL 539916, *7 (E.D. La. September 2, 1997) ("release" and "disperse" may not apply to predictable exposure, such as inhalation during asbestos remediation).

81 We recognize that asbestos has been identified as an environmental pollutant. See Selm v. American States Ins. Co., Docket No. C-010057, 2001 WL 1103509, *3 (Ohio App. September 21, 2001) ; see also Regs., Conn. State Agencies §§ 22a-209-1, 22a-209-8 (i), 22a-430-4, App. D. The question that we address here, however, is whether, consistent with the governing regulations, the pollution exclusion is intended to deny coverage only for environmental contamination resulting from its release. It is noteworthy in this regard that a number of the environmental regulations governing asbestos use expressly bar the "discharge" of asbestos only "to the outside air" and, therefore, presumably would not apply to the indoor exposure scenarios alleged in many of the underlying complaints. See, e.g., 40 C.F.R. § 61.142 (a) ; 40 C.F.R. § 61.149 (b) and (c).

82 In light of the large number of relevant sister state cases and the fact-specific nature of those cases, we do not distinguish in this part of the opinion between cases construing standard versus nonstandard pollution exclusion clauses.

83 Some courts have reasoned that this diversity of judicial opinion itself is evidence that the policy language is ambiguous. E.g., Motorists Mutual Ins. Co. v. RSJ, Inc., supra, 926 S.W.2d at 681 ; see also C & H Electric, Inc. v. Bethel, 312 Conn. 843, 855, 96 A.3d 477 (2014) (appearing to accept similar argument).

84 See, e.g., Nationwide Mutual Ins. Co. v. Richardson, supra, 270 F.3d at 952 ; MacKinnon v. Truck Ins. Exchange, supra, 31 Cal.4th at 643-45, 3 Cal.Rptr.3d 228, 73 P.3d 1205 ; American States Ins. Co. v. Koloms, 177 Ill.2d 473, 489-93, 227 Ill.Dec. 149, 687 N.E.2d 72 (1997) ; Doerr v. Mobil Oil Corp., 774 So.2d 119, 126-28 (La. 2000) ; Morton International, Inc. v. General Accident Ins. Co. of America, 134 N.J. 1, 31-42, 629 A.2d 831 (1993), cert. denied sub nom. Ins. Co. of North America v. Morton International, Inc., 512 U.S. 1245, 114 S.Ct. 2764, 129 L.Ed.2d 878 (1994) ; Belt Painting Corp. v. TIG Ins. Co., supra, 100 N.Y.2d at 384-86, 763 N.Y.S.2d 790, 795 N.E.2d 15 ; R. Burke, "Pollution Exclusion Clauses: The Agony, the Ecstasy, and the Irony for Insurance Companies," 17 N. Ky. L. Rev. 443, 446-51 (1990) ; 3 R. Long, The Law of Liability Insurance (1976) App. 30, App. 58, App. 68C; C. Salisbury, "Pollution Liability Insurance Coverage, the Standard-Form Pollution Exclusion, and the Insurance Industry: A Case Study in Collective Amnesia," 21 Envtl. L. 357, 381-86 (1991).

85 See Maska U.S.A., Inc. v. Kansa General Ins. Co., 198 F.3d 74, 80 (2d Cir. 1999) ; Buell Industries, Inc. v. Greater New York Mutual Ins. Co., supra, 259 Conn. at 539, 791 A.2d 489 ; S. Hurwitz & D. Kohane, "The Love Canal-Insurance Coverage for Environmental Accidents," 50 Ins. Couns. J. 378 (1983).

86 The question of whether the sudden and accidental exception applies to the underlying claims is not at issue in the present appeals.

87 The National Casualty exclusion provides: "[It is agreed that the insurance does not apply]:

"1. To 'bodily injury' or 'property damage' arising out of the actual, alleged or threatened discharge, dispersal, release or escape of pollutants:

"a. at or from premises you own, rent or occupy;

"b. at or from any site or location used by or for you or others for the handling, storage, disposal, processing or treatment of waste material.

"c. which are at any time transported, handled, stored, treated, disposed of, or processed as waste by or for you or any person or organization for whom you may be legally responsible; or

"d. at or from any site or location on which you or any contractors or subcontractors working directly or indirectly on your behalf are performing operations:

"(i) to test for, monitor, clean up, [remove, contain, treat, detoxify] or neutralize the pollutants, or

"(ii) if the pollutants are brought on or to the site or location by or for you.

"2. [To] any loss, cost or expense arising out of any governmental direction or request that you test for, monitor, clean up, remove, contain, treat, detoxify or neutralize pollutants.

"Pollutants means any solid, liquid, gaseous or thermal irritant or contaminant, including, smoke, vapor, soot, fumes, acids, alkalis, chemicals and waste material. Waste material includes materials which are intended to be or have been recycled, reconditioned or reclaimed.

"Provided however, that this exclusion does not apply to bodily injury, or property damage which is within the products hazard as defined in this policy nor to such discharge, dispersal, release or escape directly caused by fire, explosion, vandalism and malicious mischief, lightning, windstorm or upset or collision of a motor vehicle."

88 The Lloyd's exclusion provides: "This Insurance does not cover any liability for:

"(1) Personal Injury or Bodily Injury or loss of, damage to or loss of use of property directly or indirectly caused by seepage, pollution or contamination.

"(2) The cost of removing, nullifying or cleaning-up seeping, polluting or contaminating substances.

"(3) Fines, penalties, punitive or exemplary damages."

89 For this reason, Vanderbilt's argument that the defendants are unable to cite to any cases in support of their position is unavailing. Vanderbilt is in the same boat.

90 The trial court found that the minor variations in policy language between the two versions are not relevant to the question of whether the occupational disease exclusions apply to nonemployees of the policyholder. On appeal, the parties do not challenge this finding or argue that the two provisions are materially different.

91 "The phrase 'follow form' refers to the practice, common in excess policies, of having the second-layer coverage follow substantively the primary layer provided by the main insurer ...." (Citation omitted.) Insituform Technologies, Inc. v. American Home Assurance Co., 566 F.3d 274, 278 (1st Cir. 2009) ; see also SFA Group, LLC v. Certain Underwriters at Lloyd's, London, Docket No. CV 16-04202-GHK, 2016 WL 5842180, *4 (C.D. Cal. September 29, 2016) (following form "means that the excess policies are subject to the same terms, exclusions, conditions and definitions as the primary policy" [emphasis omitted; internal quotation marks omitted] ), appeal docketed, No. 16-56467 (9th Cir. October 6, 2016).

92 For this reason, we reject Vanderbilt's argument that the defendants' interpretation of the occupational disease exclusions would render much of the coverage afforded by the policies "illusory." At the very least, the exclusions would not bar coverage for claims brought by complainants in category 3.

We note in this respect that the parties have neither briefed nor asked us to resolve the question of whether, if the occupational disease exclusions do apply to nonemployees, they bar coverage for underlying actions in category 2, which allege both workplace and nonworkplace exposure. That question will fall to the trial court on remand to address in the first instance.

93 See N.Y. Ins. Law § 6301 (McKinney 2016) ; N.Y. Comp. Codes R. & Regs. tit. 11, § 16.0.

94 We note that, for purposes of the present appeals, none of the parties has argued that any of the contracts at issue should be construed under the law of any state other than Connecticut.

95 See, e.g., W. Viscusi, "Structuring an Effective Occupational Disease Policy: Victim Compensation and Risk Regulation," 2 Yale J. on Reg. 53, 65 (1984) (noting that, "[o]ver the past decade, [occupational] disease victims have attempted to circumvent the restrictions of workers' compensation programs by bringing products liability claims against the manufacturers of hazardous materials used in the workplace"); American Bar Association, ABA Blueprint for Improving the Civil Justice System: Report of the ABA Working Group on Civil Justice System Proposals (1992) p. 53 (recounting recommendations of 1983 committee "with respect to claims arising out of occupational diseases ... such as asbestosis").

96 The Lloyd's policy does not use the term "personal injury" in its employee liability exclusion clause. It does, however, contain a definition of "personal injuries" that also includes sickness and disease.

97 We are not persuaded by Vanderbilt's argument that a Maryland case, Commercial Union Ins. Co. v. Porter Hayden Co., supra, 116 Md.App. 605, 698 A.2d 1167, stands for the proposition that any use of the term "occupational disease" outside of the workers' compensation context is inherently ambiguous. Rather, we read Porter Hayden Co. as cautioning against the very misstep that the trial court committed here, namely, assuming that any special role that the term "occupational disease" plays in the context of workers' compensation law necessarily applies in the distinct context of tort law and insurance litigation. See id., at 699-701, 698 A.2d 1167.

98 This last policy was superseded by policy RDX 3652404, which was effective from May 17, 1977 to March 25, 1978.

99 The text of the Coverage A and Coverage B provisions contained in several of the policies at issue was amended by a New York State endorsement clause in ways not material to the current dispute.

100 Curiously, none of the parties refers us to Continental Marble & Granite Co. v. Canal Ins. Co., 785 F.2d 1258, 1259 (5th Cir. 1986), or to Caldwell Freight Lines, Inc. v. Lumbermens Mutual Casualty Co., 947 So.2d 948, 957 (Miss. 2007), both of which reach the opposite conclusion with respect to materially similar provisions.

101 It should be noted that the term "umbrella coverage" is often used not only with reference to policies that offer both excess coverage and primary drop-down insurance, but also specifically to the drop-down portion of such policies.

102 For example, § 5, entitled "Policy Period," provides: "Coverage A-This coverage applies to injury or destruction which occurs during this policy period in the places stated in the immediate underlying policy....

"Coverage B-This coverage applies to personal injury, property damage or advertising injury taking place during this policy period." (Emphasis added.)

103 To the extent that Vanderbilt argues that facts such as the size of the policy premiums support its position that it reasonably expected Continental to provide defense coverage under Coverage B, the trial court made no findings that would bear out those arguments.

104 We note in this respect that, if the parties had intended that the policy would provide defense coverage when the underlying primary insurance is exhausted, they easily could have said so expressly. See, e.g., Cambridge Mutual Fire Ins. Co. v. Ketchum, Docket No. 3:11-cv-00743 (VLB), 2012 WL 3544885, *1 (D. Conn. August 16, 2012) (umbrella policy expressly provided that insurer would provide coverage if occurrence was "either: 1. not covered by any primary insurance ... or any other insurance which applies; or 2. [c]overed by a primary policy ... which ... has been exhausted" [internal quotation marks omitted] ). It is true that, under either party's interpretation, the policy language is inartful and could have been more carefully drafted. We are hard-pressed to understand, however, why the parties would have used the language they did to express the meaning that Vanderbilt ascribes to it.

105 At oral argument before this court, Continental explained by way of example that an insured might purchase a primary policy containing a pollution exclusion and then purchase a separate pollution policy or obtain pollution coverage under a separate subcontractor policy. In that instance, Coverage B would not respond to a pollution claim because, although the underlying insurance does not provide defense coverage, there is other applicable insurance. Continental maintains that the Applicable Clause in its policies was drafted with that sort of scenario in mind.

106 In his testimony, Posner opined that the manner in which Continental and Hartford allocated their indemnity payments was reasonable.

107 The charts contain information subject to a confidentiality agreement between the parties. We therefore describe those materials and their contents in general fashion.

108 Earlier in his testimony, Posner was asked about his process in performing an allocation. Posner explained that "the first thing ... is, you need to gather ... the underlying data, which is the amounts that have been paid on the cases that you're allocating. You need to obtain the exposure dates. A lot of the times, these dates are obtained through counsel, even if you have a client that has literally hundreds of thousands of asbestos claims. A process is set up sometimes where the information will come into a central place, will be put into a computer database, and ultimately you'll be left with a database of information that will give you the claimant's name, information from which to determine what the exposure dates are and the amounts that need to be allocated.... [The information would be entered] into a database, and ultimately ... we would use that information to perform the allocation."

109 Whereas the allocations in plaintiff's exhibit V-781-282 utilized a default date of first exposure of January 1, 1948, the allocations in plaintiff's exhibit V-781-283 utilized a default date of first exposure of January 1, 1956.

110 Mt. McKinley also argues that the 2002 settlement agreement improperly failed to allocate any payments to Continental's 1956-1961 policies and that the primary insurers' continued use of the allocation methodology adopted pursuant to that agreement after 2002 was unfair to the secondary insurers. We have addressed these claims in part III C of this opinion.

111 Although the parties did not further detail the nature of such " 'nonproducts' " claims at trial, the policies themselves contain numerous exclusions. For example, Continental policy numbers CCP9024038R, in effect from January 1, 1974 through January 1, 1977, and CCP3000112, in effect from January 1, 1977 through March 1, 1977, exclude from coverage property damage to "property owned or occupied by or rented to the insured ...." Hartford policy number 10JPRB46801E, in effect from March 3, 1977 through March 3, 1978, excludes from coverage bodily injury or property damage due to "any act of the [insured's] Vendor which changes the condition of the products" and "any failure [on the part of the insured's vendor] to maintain the product in merchantable condition ...."

112 A loss run is a report provided by an insurance company that documents claim activity on an insured's policy. See, e.g., North American Capacity Ins. Co. v. Brister's Thunder Karts, Inc., Docket. No. CIV. A. 97-0330, 1998 WL 259966, *1 n.3 (E.D. La. May 20, 1998) (loss runs "are generated by an insurer relative to a particular customer and summarize the loss/claim data for the period or periods of coverage"); Viking Pump, Inc. v. Century Indemnity Co., Docket No. 10C-06-141 FSS CCLD, 2012 WL 5383100, *1 (Del. Super. September 18, 2012) (describing loss run as "a payment report" of processed claims).

113 Although Mt. McKinley, in its April 16, 2014 motion to "reargue, reconsider and/or clarify," raised multiple issues regarding the court's Phase II decision, that motion is silent as to the present issue.

114 Because the payment in question was made pursuant to a "confidential settlement" that resolved the contribution action between Continental and Hartford, the court sealed certain specifics thereof in the proceeding at trial. Consistent with that order, we describe the details of that payment in general terms.

115 In Cohn, the Supreme Court explained: "Where the terms of a policy are clear and unambiguous, they will be given their plain and ordinary meaning.... Whether an insurance contract is a liability policy or an indemnity policy depends upon the intention of the parties, as evidenced by the phraseology of their agreement .... The chief difference between a liability policy and an indemnity policy is that under the former a cause of action accrues when the liability attaches, while under the latter there is no cause of action until the liability has been discharged, as by payment of the judgment by the insured." (Citations omitted; internal quotation marks omitted.) Cohn v. Pacific Employers Ins. Co., supra, 213 Conn. at 546-47, 569 A.2d 544.

116 General Statutes § 52-205 provides: "In all cases, whether entered upon the docket as jury cases or court cases, the court may order that one or more of the issues joined be tried before the others."

117 Practice Book § 15-1 provides: "In all cases, whether entered upon the docket as jury cases or court cases, the judicial authority may order that one or more of the issues joined be tried before the others. Where the pleadings in an action present issues both of law and of fact, the issues of law must be tried first, unless the judicial authority otherwise directs. If some, but not all, of the issues in a cause are put to the jury, the remaining issue or issues shall be tried first, unless the judicial authority otherwise directs."

118 Indeed, at least one of the bifurcation orders was entered at the behest of Continental, which filed a motion requesting such action pursuant to § 52-205 and Practice Book § 15-1.

119 Continental also suggests that, absent a determination by this court as to its duty to defend under the excess policy known as RDU 9433526, an orphan share will be created for the January 1, 1965 through January 1, 1968 time period. That contention is fraught with assumption, as the trial court has not yet decided whether the primary insurance policy in place for that time period has been exhausted, nor has it determined whether Vanderbilt failed to procure available excess insurance for defense costs during that period. Without those determinations, the existence of an orphan share is little more than sheer speculation, which has no place in appellate review. See New Hartford v. Connecticut Resources Recovery Authority, 291 Conn. 502, 510, 970 A.2d 578 (2009).

On a more basic level, Continental's concern that resolution of its duty to defend issue by this court is necessary to properly allocate defense costs is unfounded. Our decision today articulates the appropriate allocation methodology to be employed by the trial court on remand. See part III of this opinion. On remand, the trial court is tasked with applying that methodology to the issues before it, including the issue of Vanderbilt's insurance coverage from 1965 through 1968. At that time, Continental remains free to pursue its claims regarding any defense obligations under the RDU 9433526 and RDX 3652404 excess policies.

120 The requirement that the issue must have been resolved by the trial court comports with a fundamental tenet of appellate review in this state. Our appellate courts generally "will not address issues not decided by the trial court." Willow Springs Condominium Assn., Inc. v. Seventh BRT Development Corp., 245 Conn. 1, 52, 717 A.2d 77 (1998) ; see also Crest Pontiac Cadillac, Inc. v. Hadley, 239 Conn. 437, 444 n.10, 685 A.2d 670 (1996) (claims "neither addressed nor decided" by trial court are not properly before appellate tribunal); Atwood v. Jarrett, 81 Conn. 532, 533, 71 A. 569 (1909) ("[t]he record nowhere ... discloses that the claim embodied in the remaining assignment of error was ... passed upon by the court below [and] therefore [it is] not properly before us for consideration"); McGuire v. McGuire, 102 Conn.App. 79, 87, 924 A.2d 886 (2007) ("[w]e have repeatedly held that this court will not consider claimed errors on the part of the trial court unless it appears on the record that the question ... was ruled upon and decided by the court adversely to the appellant's claim" [internal quotation marks omitted] ).

121 As with Old Republic's claim regarding its defense obligations, there is nothing that precludes the trial court on remand from considering Continental's claim regarding its duty to defend under the RDU 9433526 and RDX 3652404 excess policies prior to the formal commencement of the Phase III trial.