IN THE SUPREME COURT OF TEXAS

IN THE SUPREME COURT OF TEXAS

═════════_= 52;══

No. 04-1104

═════════_= 52;══

Richard Fiess and = Stephanie=20 Fiess, Appellants,

State Farm Lloyds,=20 Appellee

═════════_= 52;══════════= ══════════ = 552;══════════= ;══════════&#= 9552;

On Certified Question from the = United=20 States

Court of Appeals for the Fifth=20 Circuit

═════════_= 52;══════════= ══════════ = 552;══════════= ;══════════&#= 9552;

Argued March 30,=20 2005

Justice Brister delivered the = opinion=20 of the Court, in which Chief = Justice=20 Jefferson, Justice Hecht, Justice=20 Wainwright, Justice = Green,=20 Justice Johnson, and = Justice Willett joined.

Justice Medina filed a = dissenting=20 opinion, in which Justice = O=92Neill=20 joined.

The = question=20 in this case is not whether insurers should provide mold coverage in = Texas, a=20 public policy question beyond our jurisdiction as a court. The question = instead=20 is whether the language in an insurance policy provides such coverage = C no more and no = less.

The = rules for=20 construing insurance policies have been around for a long time, long = before this=20 dispute arose. Those rules require us to construe a policy according to = what it=20 says, not what regulators or individual insurers thought it said. = Ambiguities in=20 the plain language must be settled in favor of consumers, but they must = appear=20 in the policy itself C=20 we cannot create ambiguities from previous policies, an agency=92s = interpretation,=20 or a =93mold crisis.=94

The = policy=20 here provides that it does not cover =93loss caused by mold.=94 While = other parts of=20 the policy sometimes make it difficult to decipher, we cannot hold that = mold=20 damage is covered when the policy expressly says it is not. Accordingly, = we=20 answer the Fifth Circuit=92s certified question =93No.=94

I. =93We do not cover loss caused by = mold=94

This case=20 comes to us on a certified question [1] from the United States Court = of=20 Appeals for the Fifth Circuit, which asks us:

&nbs= p;=20 Does the ensuing loss provision contained in Section I-Exclusions, part = 1(f) of=20 the Homeowners Form B (HO-B) insurance policy as prescribed by the Texas = Department of Insurance effective July 8, 1992 (Revised January 1, = 1996), when=20 read in conjunction with the remainder of the policy, provide coverage = for mold=20 contamination caused by water damage that is otherwise covered by the = policy?[2]

The = policy=20 provision in question provides as follows:

We do not=20 cover loss caused by:

(1) =20 wear and tear, deterioration or loss caused by any quality in property = that=20 causes it to damage or destroy itself.

(2)=20 rust, rot, mold or other fungi. =

(3)=20 dampness of atmosphere, extremes of = temperature.

(4)=20 contamination.

(5) =20 rats, mice, termites, moths or other insects.

We do cover=20 ensuing loss caused by collapse of the building or any part of the = building,=20 water damage, or breakage of glass which is part of the building if the = loss=20 would otherwise be covered under this policy.[3]

The = rules for=20 construing this provision are well settled. If a policy provision has = only one=20 reasonable interpretation, it is unambiguous and we must construe it as = a matter=20 of law.[4] If an exclusion has more = than one=20 reasonable interpretation, we must construe it in favor of the insured = as long=20 as that construction is not unreasonable.[5] A policy provision is not = ambiguous=20 merely because different parties or different courts have interpreted it = differently.[6]

As = with any=20 other contract, the parties=92 intent is governed by what they said, not = by what=20 they intended to say but did not.[7] Moreover, in cases like this = involving=20 a standard form policy mandated by a state regulatory agency, we have = held for=20 more than 100 years that the actual intent of the parties is not what = counts (as=20 they did not write it), but the ordinary, everyday meaning of the words = to the=20 general public.[8]

In = this case,=20 it is hard to find any ambiguity in the ordinary meaning of =93We do not = cover=20 loss caused by mold.=94 While the ensuing-loss clause that follows may = be=20 difficult to parse (a matter discussed below), few ordinary people would = imagine=20 that it changes the meaning of the first sentence to read =93We do = too=20 cover loss caused by mold.=94

The = dissent=20 finds this policy ambiguous, primarily by construing the preceding HO-B = policy,=20 on the basis that no change was intended when that form was dropped in = 1990.=20 Evidence of prior policies is extrinsic evidence, and thus inadmissible = unless=20 this policy is ambiguous.[9] Ambiguity must be evident = from the=20 policy itself; it cannot be created by introducing parol evidence of = intent.[10] And while we have looked at = a prior=20 policy in deciding between reasonable constructions of a current one,[11] we have never done so in = lieu of=20 construing the current one at all. Given the complexities found in most=20 insurance policies, it is surely wiser to stick with our long-standing = legal=20 rule that insurance policies must be construed one policy at a time.

Nor = can we=20 agree with the dissent that this policy is ambiguous because the Texas=20 Department of Insurance advances an interpretation that, while not = convincing,=20 is a reasonable alternative to our own. It is true that courts give some = deference to an agency regulation containing a reasonable interpretation = of an=20 ambiguous statute.[12] But there are several = qualifiers in=20 that statement. First, it applies to formal opinions adopted after = formal=20 proceedings, not isolated comments during a hearing or opinions in = documents=20 like the Department=92s amicus brief here.[13] Second, the language at = issue must be=20 ambiguous; an agency=92s opinion cannot change plain language.[14] Third, the agency=92s = construction must=20 be reasonable;[15] alternative unreasonable = constructions do not make a policy ambiguous.[16] An agency=92s opinion can = help construe=20 an existing ambiguity, but it cannot create one; that the = Department=20 agrees with the Fiesses=92 construction does not make this policy = ambiguous.[17]

Moreover,=20 neglecting what this policy says in favor of what the Department says it = intended would raise a host of other problems. First, construing a = statewide=20 policy according to what a single regulator, insurer, or insured thought = about=20 it would bind many others without hearing what they might have intended. = Second,=20 even if no change was intended in 1990, that does not tell us what = anyone=20 intended before 1990, an issue we have never addressed. And = finally,=20 deriving intent from extrinsic evidence raises a fact question for = jurors, not=20 judges;[18] while ambiguous exclusions = must be=20 construed in favor of the insured as long as that construction is not=20 unreasonable,[19] ambiguous intentions are = not governed=20 by the same legal rule.

We = must of=20 course consider an insurance policy in its entirety. But in doing so, we = cannot=20 overlook the obvious =97 that the policy provision here begins by = stating=20 unambiguously, =93We do not cover loss caused by mold.=94

II. =93We do cover ensuing loss caused by water damage = . .=20 .=94

The = Fiesses=20 argue that we must disregard how this policy provision starts (=93We do = not cover=20 loss caused by mold=94) because of how it ends (=93We do cover ensuing = loss caused=20 by water damage=94). We disagree; it has again long been the rule that = we must=20 read all parts of a policy together, giving meaning to every sentence, = clause,=20 and word to avoid rendering any portion inoperative.[20]

In = Lambros=20 v. Standard Fire Insurance Co.,[21] homeowners alleged = underground water=20 cracked the slab of their home. Like the policy here, their policy = excluded=20 losses due to cracked foundations, but also stated that this exclusion = =93shall=20 not apply to ensuing loss caused by . . . water damage.=94 Justice = Cadena writing=20 for the Fourth Court of Civil Appeals found that the only reasonable=20 construction of this clause was that it applied when an excluded risk = was=20 followed by an intervening occurrence that in turn caused an ensuing = loss:

To =93ensue=94=20 means =93to follow as a consequence or in chronological succession; to = result, as=20 an ensuing conclusion or effect.=94 An =93ensuing loss,=94 then, is a = loss which=20 follows as a consequence of some preceding event or circumstance . . . . = If we=20 give to the language of the exception its ordinary meaning, we must = conclude=20 that an ensuing loss caused by water damage is a loss caused by water = damage=20 where the water damage itself is the result of a preceding cause. What = is the=20 preceding cause which gives to the exception the effect of taking the = ensuing=20 loss out of the reach of exception k [the foundation exception]? Again, = the=20 plain language of the exception compels the conclusion that the water = damage=20 must be a consequence, i.e., follow from or be the result of the types = of damage=20 enumerated in exception k. =93Ensuing loss caused by water damage=94 = refers to water=20 damage which is the result, rather than the cause, of = =93settling,=20 cracking, . . . of foundations . . .=94 [22]

This Court=20 refused the application for writ of error, thus giving the Lambros=20 opinion the same force and effect as one of our own.[23]

The = part of=20 the ensuing‑loss clause at issue in Lambros is = indistinguishable from the=20 part at issue here. The Department of Insurance asserts that the = Lambros=20 policy covered fewer water risks and the homeowners=92 claim did not = involve=20 mold. But the relevant ensuing‑loss language has changed in no = material respect;=20 that Lambros involved a different house, different homeowners, = and a=20 different insurer does not make it distinguishable. If Lambros is = still=20 the law, then this clause too applies only to losses caused by an = intervening=20 cause (like water damage) that in turn follow from an exclusion listed = in=20 paragraph 1(f).

The = Fiesses=20 and the Department make several arguments for construing = ensuing‑loss clauses=20 differently, but all would require reversing Lambros. That of = course is=20 not out of the question; our opinions are not like the law of the Medes = and the=20 Persians =97 unalterable once written.[24] But we are bound to = consider the=20 principles of stare decisis before taking such a step.

Stare decisis=20 has its greatest force in cases construing statutes, partly because our = errors=20 may be corrected by statutory amendments.[25] Although Lambros did = not=20 construe a statute, it was the next thing to it =97 a mandatory policy = form=20 promulgated by a state agency that private parties could not alter. If = our=20 policy interpretation in Lambros was wrong, it is strange that = insurance=20 regulators did nothing to change the policy for a quarter century. = Accordingly,=20 we decline the invitation to overrule it.

III. =93. . . caused by water damage . . .=94

Nor = can we=20 disregard how this policy provision starts (=93We do not cover loss = caused by=20 mold=94) because of how it ends (=93We do cover ensuing loss caused by = water=20 damage=94), as the latter is not as broad as the Fiesses suggest. By its = own=20 terms, paragraph 1(f) covers only ensuing losses from water = damage, not=20 water alone.

The = parties=20 disagree whether mold stemming from the small roof and window leaks at = issue=20 here would constitute =93water damage.=94 [26] While we have never = construed =93water=20 damage=94 in the Texas homeowners ensuing-loss clause, the legendary = Henry=20 Friendly [27] did (sitting with the Fifth = Circuit=20 by designation), and concluded that inadequate ventilation that led to=20 condensation that eventually caused a floor to rot away did not fall = within the=20 Texas ensuing-loss clause because the rot was caused by water, not = =93water=20 damage=94:

We do not=20 think that a single phenomenon that is clearly an excluded risk under = the policy=20 was meant to become compensable because in a philosophical sense it can = also be=20 classified as water damage; it would not be easy to find a case of rot = or=20 dampness of atmosphere not equally subject to that label and the = exclusions=20 would become practically meaningless. In our case the rot may have = ensued from=20 water but not from water damage, and the damage ensuing from the rot was = not the=20 damage from the direct intrusion of water conveyed by the phrase = =93water=20 damage.=94 [28]

Surely Judge=20 Friendly was correct. Mold does not grow without water; if every leak = and drip=20 is =93water damage,=94 then it is hard to imagine any mold, rust, or rot = excluded by=20 this policy, and the mold exclusion would be practically meaningless. =

The = 15 risks=20 excluded by paragraph 1(f) =97 rust, rot, mold, humidity, wear and tear, = hot and=20 cold weather, rats, termites, and so on =97 all damage a home = incrementally; when=20 they cause major damage, generally the home was not destroyed in a day. = These 15=20 risks are also very common; construing the HO-B policy to cover them all = would=20 convert it from an insurance policy into a maintenance agreement.

Instead, the=20 ensuing-loss clause provides coverage only if these relatively common = and=20 usually minor risks lead to a relatively uncommon and perhaps major = loss:=20 building collapse, glass breakage, or water damage. In construing the = last term,=20 we are governed by the traditional canon of construction noscitur a = sociis=20 =97 =93that a word is known by the company it keeps.=94[29] Accordingly, =93water = damage=94 like its=20 neighbors =93building collapse=94 and =93glass breakage=94 must refer to = something more=20 substantial than every tiny water leak or seep. Applying this = traditional rule=20 of construction, ordinary people would read paragraph 1(f) to provide = coverage=20 for the kinds of uncommon and catastrophic losses for which homeowners = obtain=20 insurance, not for the common maintenance items for which they do not. =

We = need not=20 decide today the precise scope of =93water damage=94 in the ensuing-loss = clause, an=20 issue not framed by the certified question. The issue we do decide is = that a=20 policy exclusion for =93mold=94 cannot be disregarded by simply deeming = all mold to=20 be =93water damage.=94

III. =93. . . if the loss would otherwise be covered = under this=20 policy=94

All = members=20 of the Court affirm Lambros and refuse to construe the = ensuing-loss=20 clause outside its context. But the dissent would hold that the = ensuing-loss=20 clause cancels the mold exclusion of which it is a part, because its = last phrase=20 (=93if the loss would otherwise be covered under this policy=94) = requires us to=20 disregard paragraphs 1(f), 1(g), and 1(h) of the policy. Clearly, = removing the=20 22 exclusions in those paragraphs would create mold coverage, but it = would also=20 create a different policy. To qualify as an ensuing loss, mold must = =93otherwise=20 be covered under this policy.=94

Here, the=20 first sentence of 1(f) excludes mold, and the second sentence extends = coverage=20 to ensuing losses caused by water damage. If neither sentence said = anything=20 more, the two would conflict whenever water damage (covered) caused mold = (excluded). But 1(f) resolves this potential conflict by limiting the = second=20 clause =97 the ensuing-loss clause =97 whenever it conflicts with = anything else in=20 the policy. By placing this proviso where it is, the only reasonable=20 construction is that the second sentence (covering ensuing losses) must = yield to=20 the first (excluding mold), not the other way around.

This does=20 not, as the dissent suggests, delete =93otherwise=94 from paragraph = 1(f).=20 =93Otherwise=94 when used as an adverb means =93in a different way or = manner; in=20 different circumstances=94;[30] it does not mean we should = disregard=20 the immediately preceding sentence. Assume, for example, that the flight = schedule of a certain airline stated:

We do not=20 fly from Dallas to Denver.

We do fly=20 from Dallas to all cities otherwise listed in this flight schedule.

No = reasonable=20 reader would think =93otherwise=94 means we should disregard the = preceding sentence=20 and assume the airline really does fly from Dallas to Denver.

The = dissent=20 would rewrite the ensuing-loss clause here to provide coverage =93if the = loss=20 would otherwise be covered under this policy not counting the = exclusions in=20 paragraph 1(f), 1(g), and (h).=94 But those exclusions are = part of the=20 policy; a policy without exclusions for rust, rot, mold, wear and tear, = and=20 termites is simply a different policy. This would be policy = =93construction=94 in=20 the architectural rather than the legal sense.

Moreover, the=20 upshot of the dissent=92s construction would be that the more risks = excluded=20 in a policy containing an ensuing-loss clause, the broader = coverage=20 would become. Paragraphs 1(f), 1(g), and 1(h) of the HO-B policy contain = roughly=20 22 exclusions, and each has an ensuing-loss clause listing 3 intervening = risks=20 (building collapse, water damage, and glass breakage). According to the = dissent,=20 if any one of the 22 exclusions combines with any one of the 3 = intervening risks=20 to cause any of the 22 excluded losses, the loss is no longer excluded. = This=20 would mean there are only about 1,452 possible ways to turn exclusions = into=20 coverage.[31] Thus, the more exclusions = that are=20 added, the broader coverage gets. This cannot possibly be a reasonable=20 construction.

Finally, a=20 =93yes=94 answer to the certified question today would give ensuing-loss = clauses in=20 Texas a different meaning from what they have in most other American=20 jurisdictions. These clauses are common in all-risk policies, and while = rarely=20 identical they share more similarities than differences.[32] Accordingly, we should = strive for=20 uniformity in construing them.[33] But the Fiesses=92 argument = that an=20 ensuing-loss clause can make an excluded loss reappear as a covered loss = has=20 been rejected by courts in Alabama,[34] Arizona,[35] California,[36] Florida,[37] Illinois,[38] Massachusetts,[39] Minnesota,[40] New York,[41] North Carolina,[42] New Hampshire,[43] Ohio,[44] Pennsylvania,[45] Vermont,[46] Washington,[47] and Wisconsin.[48] There would have to be = something very=20 peculiar about the Texas ensuing-loss clause for its results to be so = very=20 different from similar clauses used everywhere else.

V. Conclusion

Courts adhere=20 to prior precedents for reasons of efficiency, fairness, and = legitimacy.[49] For more than a century = this Court=20 has held that in construing insurance policies =93where the language is = plain and=20 unambiguous, courts must enforce the contract as made by the parties, = and cannot=20 make a new contract for them, nor change that which they have made under = the=20 guise of construction.=94 [50] If the political branches = of Texas=20 government decide that mold should be covered in Texas insurance = policies, they=20 have tools at their disposal to do so; Texas courts must stick to what = those=20 policies say, and cannot adopt a different rule when a =93crisis=94 = arises.[51]

Accordingly,=20 for the reasons and to the extent stated in this opinion, we answer the=20 certified question =93No.=94

________________________________

Scott=20 Brister

Justice

OPINION = DELIVERED: August=20 31, 2006

^[1] See Tex.=20 Const. art. V, =A7 3‑c; Tex. R.=20 App. P. 58.1.

^[2] 392 F.3d 802, 811‑812 (5th Cir. = 2004).

^[3] We do not address personal property coverage = under=20 paragraph 9 (accidental discharge, leakage or overflow of water) of the = HO-B=20 policy because the Fifth Circuit Court of Appeals concluded that the = Fiesses=20 failed to appeal that issue. Id. at 805 n.5.

^[4] Texas Farm Bureau Mut. Ins. Co. v.=20 Sturrock, 146 S.W.3d 123, 126 (Tex. 2004); E. Texas Fire Ins. Co. = v.=20 Kempner, 27 S.W. 122, 122 (Tex. 1894).

^[5] Nat=92l Union Fire Ins. Co. v. Hudson = Energy Co.,=20 Inc., 811 S.W.2d 552, 555 (Tex. 1991); Glover v. Nat=92l Ins.=20 Underwriters, 545 S.W.2d 762, 763 (Tex. 1977); Continental Cas. = Co. v.=20 Warren, 254 S.W.2d 755, 761 (Tex. 1953).

^[6] Kelley‑Coppedge, Inc. v. Highlands = Ins. Co., 980=20 S.W.2d 462, 465 (Tex. 1998); Grain Dealers Mut. Ins. Co. v. = McKee, 943=20 S.W.2d 455, 459 (Tex. 1997).

^[7] Balandran v. Safeco Ins. Co. of Am., = 972 S.W.2d=20 738, 741 (Tex. 1998) (=93Our primary goal, therefore, is to give effect = to the=20 written expression of the parties=92 intent.=94); Nat=92l Union Fire = Ins. Co. v.=20 CBI Indus., 907 S.W.2d 517, 520 (Tex. 1995); Univ. C.I.T. Credit = Corp. v.=20 Daniel, 243 S.W.2d 154, 157 (Tex. 1951).

^[8] See Progressive County Mut. Ins. Co. v. = Sink,=20 107 S.W.3d 547, 551 (Tex. 2003) (=93[W]here the policy forms are = mandated=20 by a state regulatory agency . . . we look to determine the ordinary, = everyday=20 meaning of the words to the general public.=94); U.S. Ins. Co. of = Waco v.=20 Boyer, 269 S.W.2d 340, 341 (Tex. 1954) (=93While undoubtedly in the = early days=20 of the insurance business the actual intent of the immediate parties to = the=20 contract was material, now with the insurance business regulated and the = policy=20 forms prescribed by a State Insurance Commission, the court in = construing a=20 policy determines the everyday meaning of the words to the general = public=97the=20 meaning of the words =91in common parlance=92=97 =91the usual and = popular understanding=20 of the term.=92=93); Mutual Life Ins. Co. v. Simpson, 31 S.W. = 501, 502 (Tex.=20 1895) (=93[C]ontracts of insurance, like other contracts, are to be = construed=20 according to the sense and meaning of the terms which the parties have = used, and=20 if they are clear and unambiguous, their terms are to be taken and = understood in=20 their plain, ordinary and popular sense.=94).

^[9] Sharp v. State Farm Fire and Cas. Ins. = Co., 115=20 F.3d 1258, 1262 (5th Cir. 1997); Balandran, 972 S.W.2d at 741;=20 CBI, 907 S.W.2d at 520-21.

^[10] See CBI, 907 S.W.2d at = 521.

^[11] See Balandran, 972 S.W.2d at=20 741‑42.

^[12] See Chevron U.S.A. Inc. v. Natural = Resources Defense=20 Council, Inc., 467 U.S. 837, 842-44 (1984); City of Corpus = Christi v.=20 Public Utility Com'n of Texas, 51 S.W.3d 231, 261 (Tex. 2001).=20

^[13] See Christensen v. Harris County, 529 = U.S. 576,=20 587 (2000) (refusing to defer to agency policy statements, agency = manuals, and=20 enforcement guidelines lacking the force of law); see also Reno v. = Koray,=20 515 U.S. 50, 61 (1995); EEOC v. Arabian American Oil Co., 499 = U.S. 244,=20 256‑258 (1991); Martin v. Occupational Safety and Health Review = Comm'n,=20 499 U.S. 144, 157 (1991).

^[14] See Pretzer v. Motor Vehicle Bd., 138 = S.W.3d=20 908, 915 (Tex. 2004); Continental Cas. Co. v. Downs, 81 S.W.3d = 803, 807=20 (Tex. 2002); City of Corpus Christi, 51 S.W.3d at 261; = Stanford v.=20 Butler, 181 S.W.2d 269, 273 (1944).

^[15] Downs, 81 S.W.3d at 807 (Tex. 2002); = Tarrant=20 Appraisal Dist. v. Moore, 845 S.W.2d 820, 823 (Tex.1993); = Stanford,=20 181 S.W.2d at 273; cf. Christensen, 529 U.S. at 587 (stating that = agency=20 opinions are entitled to respect to the extent of their power to=20 persuade).

^[16] See Balandran v. Safeco Ins. Co. of = Am., 972=20 S.W.2d 738, 741 (Tex. 1998). As we have noted before, reasonable people=20 sometimes disagree about what is reasonable. See City of = Keller v.=20 Wilson, 168 S.W.3d 802, 828 (Tex. 2005).

^[17] See Downs, 81 S.W.3d at 807 (=93[T]hat = the=20 Commission agrees with Continental=92s construction of the statute does = not make=20 that construction any more persuasive.=94).

^[18] See Couch on Insurance 3d =A7 = 21.13 (=93If,=20 however, ambiguous words are to be construed in light of the extrinsic = evidence,=20 or of the surrounding circumstances, the meaning of such words becomes a = question of fact for the jury.=94); see also Foreca, S.A. v. GRD Dev. = Co.,=20 758 S.W.2d 744, 746 (Tex. 1988).

^[19] See Hudson Energy Co., 811 S.W.2d at=20 555.

^[20] Balandran, 972 S.W.2d at 741; = Liberty=20 Mut. Ins. Co. v. Am. Emp. Ins. Co., 556 S.W.2d 242, 245 (Tex. 1977); = Pan=20 Am. Life Ins. Co. v. Andrews, 340 S.W.2d 787, 790 (Tex.=20 1960).

^[21] 530 S.W.2d 138 (Tex. Civ. App.=97San Antonio = 1975, writ=20 ref=92d).

^[22] Id. at 141 (citations omitted) = (emphasis=20 added).

^[23] See Tex.=20 R. App. P. 56.1(c); Hubenak v. San Jacinto Gas Transmission = Co.,=20 141 S.W.3d 172, 181 (Tex. 2004); Texas Utils. Elec. Co. v.=20 Timmons, 947 S.W.2d 191, 199 (Tex. 1997).

^[24] See Daniel 6:12; 4 Matthew Henry=92s Commentary = 1067 (=93The=20 Persians magnified the wisdom of their king, by supposing that whatever = law he=20 solemnly ratified it was so well made that there could be no occasion to = alter=20 it, or dispense with it, as if any human foresight could, in framing a = law,=20 guard against all inconveniences.=94).

^[25] See Grapevine Excavation, Inc. v. Md. = Lloyds, 35=20 S.W.3d 1, 6 (Tex. 2000) (Gonzalez, J., concurring) (=93[E]nacting = statutes is=20 within the unique province of the Legislature, and as to statutes, the = ultimate=20 interpretation is within their hands=94); accord, Quill Corp. v. = North=20 Dakota, 504 U.S. 298, 320 (1992) (Scalia, J., concurring) (citing=20 Patterson v. McLean Credit Union, 491 U.S. 164, 172‑173 = (1989)=20 (=93Considerations of stare decisis have special force in the area of = statutory=20 interpretation, for here, unlike in the context of constitutional=20 interpretation, the legislative power is implicated, and Congress = remains free=20 to alter what we have done.=94)).

^[26] The Fiesses recovered under a separate policy = for mold=20 caused by Tropical Storm Allison, and failed to preserve a claim for = mold caused=20 by air‑conditioning and plumbing leaks. See Fiess, 392 F.3d = at 804, 807.=20

^[27] See Finley v. U.S., 490 U.S. 545, 565 = n.18=20 (1989) (noting Friendly =93is universally recognized...as one of our = wisest=20 judges=94); Paul Freund, In Memoriam: Henry J. Friendly, 99 Harv. L.R. 1709, 1715 (noting = Friendly=20 =93was not merely a legend in his own time...=94); Richard A. Posner, = id. at=20 1724 (=93He was the greatest federal appellate judge of his time ‑ = in analytic=20 power, memory, and application perhaps of any time. His opinions have = exhibited=20 greater staying power than that of any of his contemporaries on the = federal=20 courts of appeals.=94).

^[28] See Aetna Cas. & Sur. Co. v. = Yates, 344 F.2d=20 939, 941 (5th Cir. 1965) (construing clause stated in part stating that=20 exclusions =93shall not apply to ensuing loss caused by . . . water = damage=94).=20

^[29] Gustafson v. Alloyd Co., 513 U.S. 561, = 575=20 (1995) (noting that the purpose of this rule is =93to avoid ascribing to = one word=20 a meaning so broad that it is inconsistent with its accompanying = words.=94);=20 see also Black=92s Law=20 Distionary 1087 (8th ed. 2004) (=93[Latin for =91it is known by = its=20 associates=92] A canon of construction holding that the meaning of an = unclear word=20 or phrase should be determined by the words immediately surrounding=20 it.=94).

^[30] Webster=92s New=20 Collegiate Dictionary 835 (9th ed. 1985).

^[31] That is, 22 x 3 x 22 =91 1,452. See = Sheldon Ross, A First Course in Probability = 2‑3 (7th=20 ed. 2006) (noting that multiplying possibilities of each of a sequence = of events=20 produces the total possible outcomes). It is true that some combinations = are=20 unlikely, such as wear-and-tear followed by glass breakage that causes = mice. But=20 with 1,452 to choose from, no doubt plenty of options = remain.

^[32] For example, decisions in a number of = jurisdictions=20 address ensuing-loss clauses containing the same three intervening = causes=20 (building collapse, water damage, and glass breakage) as the HO-B = policy.=20 See, e.g., Souza v. Corvick, 441 F.2d 1013, 1016 = (D.C. Cir.=20 1970); N.Z. Ins. v. Lenoff, 315 F.2d 95, 95 n.1 (9th Cir. 1963); = Beach=20 v. Middlesex Mut. Ass. Co., 532 A.2d 1297, 1298 n.1 (Conn. 1987); = Phoenix=20 Ins. Co. v. Branch, 234 So.2d 396, 398 (Fla. Dist. Ct. App. 1970);=20 Nationwide Ins. Co. v. Warren, 675 S.W.2d 402, 403 (Ky. Ct. App. = 1984);=20 Shields v. Pa. Gen. Ins. Co., 488 So.2d 1252, 1253 (La. Ct. App. = 1986);=20 Cantrell v. Farm Bureau Town & Country Ins. Co. of Mo., 876 = S.W.2d=20 660, 662 (Mo. Ct. App. 1994); Umanoff v. Nationwide Mut. Fire = Ins., 442=20 N.Y.S.2d 892, 893 (N.Y. Civ. Ct. 1981).

^[33] See Nat=92l Union Fire Ins. Co. v. CBI = Indus., 907=20 S.W.2d 517, 522 (Tex. 1995).

^[34] See Schloss v. Cincinnati Ins. Co., 54 = F. Supp.=20 2d 1090, 1098 (M.D. Ala. 1999), aff=92d, 211 F.3d 131 (11th Cir.=20 2000).

^[35] See Cooper v. Am. Family Mut. Ins. = Co., 184 F.=20 Supp. 2d 960, 964 (D. Ariz. 2002).

^[36] See Murray v. State Farm Fire & Cas. = Co.,=20 219 Cal. App. 3d 58, 65 (Cal. Ct. App. 1990).

^[37] See Church of the = Palms‑Presbyterian=20 (U.S.A.), Inc. v. Cincinnati Ins. Co., 404 F. Supp. 2d 1339, 1342 = (M.D. Fla.=20 2005).

^[38] See Bd. of Educ. of Maine Township v. = Int=92l Ins.=20 Co., 684 N.E.2d 978, 984 (Ill. App. Ct. 1997).

^[39] See Ames Privilege Assocs. Ltd. P=92ship = v. Utica Mut.=20 Ins. Co., 742 F.Supp. 704, 708 (D. Mass. 1990).

^[40] See Myers v. State Farm Fire & Cas. = Co.,=20 2002 WL 1547673 *6 (Minn. Ct. App. July 16, 2002).

^[41] See Narob Dev. Corp. v. Ins. Co. of N. = Am., 631=20 N.Y.S. 155 (N.Y. App. Div. 1995).

^[42] See Alwart v. State Farm Fire & Cas. = Co.,=20 508 S.E.2d 531, 533-34 (N.C. Ct. App. 1998).

^[43] See Weeks v. Coop. Ins. Cos., 817 A.2d = 292, 296=20 (N.H. 2003).

^[44] See Boughan v. Nationwide Prop. & Cas. = Co.,=20 2005 WL 126781 *3 (Ohio Ct. App. Jan. 24, 2005).

^[45] See Banks v. Allstate Ins. Co., 1993 = WL 40113 *5=20 (E.D. Pa. Feb. 12, 1993).

^[46] See Vt. Elec. Power Co., v. Hartford Steam = Boiler=20 Inspection and Ins. Co., 72 F. Supp. 2d 441, 445 (D.Vt.=20 1999).

^[47] See McDonald v. State Farm Fire & Cas. = Co.,=20 837 P.2d 1000, 1005-06 (Wash. 1992).

^[48] See Richland Valley Prod. = Inc. v. St.=20 Paul Fire & Cas. Co., 548 N.W.2d 127, 133 (Wis. Ct. App. 1996). = But=20 see Phillips v. United Services Auto. Ass'n, 146 S.W.3d 629, 635-36 = (Tenn.=20 Ct. App. 2004) (holding rot, though excluded clause, was covered by = ensuing-loss=20 provision).

^[49] See Weiner v. Wasson, 900 = S.W.2d 316, 320=20 (Tex. 1995).

^[50] E. Texas Fire Ins. Co. v. Kempner, 27 = S.W. 122,=20 122 (Tex. 1894).

^[51] Weiner, 900 S.W.2d at 320 (=93[T]he = legitimacy of=20 the judiciary rests in large part upon a stable and predictable decision = making=20 process that differs dramatically from that properly employed by the = political=20 branches of government.=94).