ALLIED-BRUCE TERMINIX COS. v. DOBSON

Justice Breyer, delivered the opinion of the Court.

This case concerns the reach of § 2 of the Federal Arbitration Act. That section makes enforceable a written arbitration provision in "a contract evidencing a transaction involving commerce." 9 U. S. C. § 2 (emphasis added). Should we read this phrase broadly, extending the Act's reach to the limits of Congress' Commerce Clause power? Or, do the two italicized words—"involving" and "evidencing"—significantly restrict the Act's application? We conclude that the broader reading of the Act is the correct one, and we reverse a State Supreme Court judgment to the contrary.

I

In August 1987, Steven Gwin, a respondent who owned a house in Birmingham, Alabama, bought a lifetime "Termite Protection Plan" (Plan) from the local office of Allied-Bruce Terminix Companies, a franchise of Terminix International Company. In the Plan, Allied-Bruce promised "to protect" Gwin's house "against the attack of subterranean termites," to reinspect periodically, to provide any "further treatment found necessary," and to repair, up to $100,000, damage caused by new termite infestations. App. 69. Terminix International "guarantee[d] the fulfillment of the terms" of the Plan. Ibid. The Plan's contract document provided in writing that

In the spring of 1991, Mr. and Mrs. Gwin, wishing to sell their house to Mr. and Mrs. Dobson, had Allied-Bruce reinspect the house. They obtained a clean bill of health. But no sooner had they sold the house and transferred the Plan to Mr. and Mrs. Dobson than the Dobsons found the house swarming with termites. Allied-Bruce attempted to treat and repair the house, but the Dobsons found Allied-Bruce's efforts inadequate. They therefore sued the Gwins, and (along with the Gwins, who cross-claimed) also sued AlliedBruce and Terminix in Alabama state court. Allied-Bruce and Terminix, pointing to the Plan's arbitration clause and § 2 of the Federal Arbitration Act, immediately asked the court for a stay, to allow arbitration to proceed. The court denied the stay. Allied-Bruce and Terminix appealed.

The Supreme Court of Alabama upheld the denial of the stay on the basis of a state statute, Ala. Code § 8-1—41(3) (1993), making written, predispute arbitration agreements invalid and "unenforceable." 628 So.2d 354, 355 (1993). To reach this conclusion, the court had to find that the Federal Arbitration Act, which pre-empts conflicting state law, did not apply to the termite contract. It made just that finding. The court considered the federal Act inapplicable because the connection between the termite contract and interstate commerce was too slight. In the court's view, the Act applies to a contract only if "`at the time [the parties entered into the contract] and accepted the arbitration clause, they contemplated substantial interstate activity.' " Ibid. (emphasis in original) (quoting Metro Industrial Painting Corp. v. Terminal Constr. Co., 287 F.2d 382, 387 (CA2) (Lumbard, C. J., concurring), cert. denied, 368 U.S. 817 (1961)). Despite some interstate activities (e. g., Allied-Bruce, like Terminix, is a multistate firm and shipped treatment and repair material from out of state), the court found that the parties "contemplated" a transaction that was primarily local and not "substantially" interstate.

Several state courts and Federal District Courts, like the Supreme Court of Alabama, have interpreted the Act's language as requiring the parties to a contract to have "contemplated" an interstate commerce connection. See, e. g., Burke County Public Schools Bd. of Ed. v. Shaver Partnership, 303 N.C. 408, 417-420, 279 S.E.2d 816, 822-823 (1981); R. J. Palmer Constr. Co. v. Wichita Band Instrument Co., 7 Kan.App.2d 363, 367, 642 P.2d 127, 130 (1982); Lacheney v. Profitkey Int'l, Inc., 818 F.Supp. 922, 924 (ED Va. 1993). Several federal appellate courts, however, have interpreted the same language differently, as reaching to the limits of Congress' Commerce Clause power. See, e. g., Foster v. Turley, 808 F.2d 38, 40 (CA10 1986); Robert Lawrence Co. v. Devonshire Fabrics, Inc., 271 F.2d 402, 406-407 (CA2 1959), cert. dism'd, 364 U.S. 801 (1960); cf. Snyder v. Smith, 736 F.2d 409, 417-418 (CA7), cert. denied, 469 U.S. 1037 (1984). We granted certiorari to resolve this conflict, 510 U.S. 1190 (1994); and, as we said, we conclude that the broader reading of the statute is the right one.

II

Before we can reach the main issues in this case, we must set forth three items of legal background.

First, the basic purpose of the Federal Arbitration Act is to overcome courts' refusals to enforce agreements to arbitrate. See Volt Information Sciences, Inc. v. Board of Trustees of Leland Stanford Junior Univ., 489 U.S. 468, 474 (1989). The origins of those refusals apparently lie in "`ancient times,' " when the English courts fought "`for extension of jurisdiction—all of them being opposed to anything that would altogether deprive every one of them of jurisdiction.' " Bernhardt v. Polygraphic Co. of America, Inc., 350 U.S. 198, 211, n. 5 (1956) (Frankfurter, J., concurring) (quoting United States Asphalt Refining Co. v. Trinidad Lake Petroleum Co., 222 F. 1006, 1007 (SDNY 1915), in turn quoting Scott v. Avery, 5 H. L. Cas. 811 (1856) (Campbell, L. J.)). American courts initially followed English practice, perhaps just "`stand[ing] . . . upon the antiquity of the rule' " prohibiting arbitration clause enforcement, rather than "`upon its excellence or reason.' " Bernhardt v. Polygraphic Co., supra, at 211, n. 5 (quoting United States Asphalt Refining Co., supra, at 1007). Regardless, when Congress passed the Arbitration Act in 1925, it was "motivated, first and foremost, by a . . . desire" to change this antiarbitration rule. Dean Witter Reynolds Inc. v. Byrd, 470 U.S. 213, 220 (1985). It intended courts to "enforce [arbitration] agreements into which parties had entered," ibid. (footnote omitted), and to "place such agreements `upon the same footing as other contracts,' " Volt Information Sciences, Inc., supra, at 474 (quoting Scherk v. Alberto-Culver Co., 417 U.S. 506, 511 (1974)).

Second, some initially assumed that the Federal Arbitration Act represented an exercise of Congress' Article III power to "ordain and establish" federal courts, U. S. Const., Art. III, § 1. See Southland Corp. v. Keating, 465 U.S. 1, 28, n. 16 (1984) (O'Connor, J., dissenting) (collecting cases). In 1967, however, this Court held that the Act "is based upon and confined to the incontestable federal foundations of `control over interstate commerce and over admiralty.' " Prima Paint Corp. v. Flood & Conklin Mfg. Co., 388 U.S. 395, 405 (1967) (quoting H. R. Rep. No. 96, 68th Cong., 1st Sess., 1 (1924)). The Court considered the following complicated argument: (1) The Act's provisions (about contract remedies) are important and often outcome determinative, and thus amount to "substantive," not "procedural," provisions of law; (2) Erie R. Co. v. Tompkins, 304 U.S. 64, 71-80 (1938), made clear that federal courts must apply state substantive law in diversity cases, see also Hanna v. Plumer, 380 U.S. 460, 465 (1965); therefore (3) federal courts must not apply the Federal Arbitration Act in diversity cases. This Court responded by agreeing that the Act set forth substantive law, but concluding that, nonetheless, the Act applied in diversity cases because Congress had so intended. The Court wrote: "Congress may prescribe how federal courts are to conduct themselves with respect to subject matter over which Congress plainly has power to legislate." Prima Paint, supra, at 405.

Third, the holding in Prima Paint led to a further question. Did Congress intend the Act also to apply in state courts? Did the Federal Arbitration Act pre-empt conflicting state antiarbitration law, or could state courts apply their antiarbitration rules in cases before them, thereby reaching results different from those reached in otherwise similar federal diversity cases? In Southland Corp. v. Keating, supra, this Court decided that Congress would not have wanted state and federal courts to reach different outcomes about the validity of arbitration in similar cases. The Court concluded that the Federal Arbitration Act pre-empts state law; and it held that state courts cannot apply state statutes that invalidate arbitration agreements. Id., at 15-16.

We have set forth this background because respondents, supported by 20 state attorneys general, now ask us to overrule Southland and thereby to permit Alabama to apply its antiarbitration statute in this case irrespective of the proper interpretation of § 2. The Southland Court, however, recognized that the pre-emption issue was a difficult one, and it considered the basic arguments that respondents and amici now raise (even though those issues were not thoroughly briefed at the time). Nothing significant has changed in the 10 years subsequent to Southland; no later cases have eroded Southland `s authority; and no unforeseen practical problems have arisen. Moreover, in the interim, private parties have likely written contracts relying upon Southland as authority. Further, Congress, both before and after Southland, has enacted legislation extending, not retracting, the scope of arbitration. See, e. g., 9 U. S. C. § 15 (eliminating the Act of State doctrine as a bar to arbitration); 9 U. S. C. §§ 201-208 (international arbitration). For these reasons, we find it inappropriate to reconsider what is by now well-established law.

We therefore proceed to the basic interpretive questions aware that we are interpreting an Act that seeks broadly to overcome judicial hostility to arbitration agreements and that applies in both federal and state courts. We must decide in this case whether that Act used language about interstate commerce that nonetheless limits the Act's application, thereby carving out an important statutory niche in which a State remains free to apply its antiarbitration law or policy. We conclude that it does not.

III

The initial interpretive question focuses upon the words "involving commerce." These words are broader than the often-found words of art "in commerce." They therefore cover more than "`only persons or activities within the flow of interstate commerce.' " United States v. American Building Maintenance Industries, 422 U.S. 271, 276 (1975) (quoting Gulf Oil Corp. v. Copp Paving Co., 419 U.S. 186, 195 (1974)) (defining "in commerce" as related to the "flow" and defining the "flow" to include "the generation of goods and services for interstate markets and their transport and distribution to the consumer"); see also FTC v. Bunte Brothers, Inc., 312 U.S. 349, 351 (1941). But how far beyond the flow of commerce does the word "involving" reach? Is "involving" the functional equivalent of the word "affecting"? That phrase—"affecting commerce"—normally signals Congress' intent to exercise its Commerce Clause powers to the full. See Russell v. United States, 471 U.S. 858, 859 (1985). We cannot look to other statutes for guidance for the parties tell us that this is the only federal statute that uses the word "involving" to describe an interstate commerce relation.

After examining the statute's language, background, and structure, we conclude that the word "involving" is broad and is indeed the functional equivalent of "affecting." For one thing, such an interpretation, linguistically speaking, is permissible. The dictionary finds instances in which "involve" and "affect" sometimes can mean about the same thing. V Oxford English Dictionary 466 (1st ed. 1933) (providing examples dating back to the mid-19th century, where "involve" means to "include or affect in . . . operation"). For another, the Act's legislative history, to the extent that it is informative, indicates an expansive congressional intent. See, e. g., H. R. Rep. No. 96, supra, at 1 (the Act's "control over interstate commerce reaches not only the actual physical interstate shipment of goods but also contracts relating to interstate commerce"); 65 Cong. Rec. 1931 (1924) (the Act "affects contracts relating to interstate subjects and contracts in admiralty") (remarks of Rep. Graham); Joint Hearings on S. 1005 and H. R. 646 before the Subcommittees of the Committees on the Judiciary, 68th Cong., 1st Sess., 7 (1924) (hereinafter Joint Hearings) (testimony of Charles L. Bernheimer, chairman of the Committee on Arbitration of the Chamber of Commerce of the State of New York, agreeing that the proposed bill "relates to contracts arising in interstate commerce"); id., at 16 (testimony of Julius H. Cohen, drafter for the American Bar Association of much of the proposed bill's language, that the Act reflects part of a strategy to rid the law of an "anachronism" by "get[ting] a Federal law to cover interstate and foreign commerce and admiralty"); see also 9 U. S. C. § 1 (defining the word "commerce" in the language of the Commerce Clause itself).

Further, this Court has previously described the Act's reach expansively as coinciding with that of the Commerce Clause. See, e. g., Perry v. Thomas, 482 U.S. 483, 490 (1987) (the Act "embodies Congress' intent to provide for the enforcement of arbitration agreements within the full reach of the Commerce Clause"); Southland Corp. v. Keating, 465 U. S., at 14-15 (the "`involving commerce' " requirement is a constitutionally "necessary qualification" on the Act's reach, marking its permissible outer limit); see also Prima Paint Corp. v. Flood & Conklin Mfg. Co., 388 U. S., at 407 (Harlan, J., concurring) (endorsing Robert Lawrence Co. v. Devonshire Fabrics, Inc., 271 F.2d 402, 407 (CA2 1959) (Congress, in enacting the FAA, "took pains to utilize as much of its power as it could")).

Finally, a broad interpretation of this language is consistent with the Act's basic purpose, to put arbitration provisions on "`the same footing' " as a contract's other terms. Scherk v. Alberto-Culver Co., 417 U. S., at 511. Conversely, a narrower interpretation is not consistent with the Act's purpose, for (unless unreasonably narrowed to the flow of commerce) such an interpretation would create a new, unfamiliar test lying somewhere in a no man's land between "in commerce" and "affecting commerce," thereby unnecessarily complicating the law and breeding litigation from a statute that seeks to avoid it.

We recognize arguments to the contrary: The pre-New Deal Congress that passed the Act in 1925 might well have thought the Commerce Clause did not stretch as far as has turned out to be the case. But, it is not unusual for this Court in similar circumstances to ask whether the scope of a statute should expand along with the expansion of the Commerce Clause power itself, and to answer the question affirmatively—as, for the reasons set forth above, we do here. See, e. g., McLain v. Real Estate Bd. of New Orleans, Inc., 444 U.S. 232, 241 (1980); Hospital Building Co. v. Trustees of Rex Hospital, 425 U.S. 738, 743, n. 2 (1976).

Further, the Gwins and Dobsons point to two cases containing what they believe to be favorable language. In Mine Workers v. Coronado Coal Co., 259 U.S. 344 (1922), and then again in Leather Workers v. Herkert & Meisel Trunk Co., 265 U.S. 457 (1924), they say, this Court said that one might draw a distinction between, on the one hand, cases that "involve interstate commerce intrinsically," and, on the other hand, cases "affecting interstate commerce so directly as to be within the federal regulatory power." Mine Workers, supra, at 410 (emphasis added); Leather Workers, supra, at 470 (same). One could read these cases as driving a wedge between "involve" and "affecting." Yet, in these cases, the Court was not construing a statute containing the words "involving commerce." Furthermore, nothing suggests the drafters of the Act looked to these cases as a source. And, these cases themselves use the phrase "involve . . . intrinsically," not the word "involving" alone. In sum, these cases do not support respondents' position.

The Gwins and Dobsons, with far better reason, point to a different case, Bernhardt v. Polygraphic Co. of America, Inc., 350 U.S. 198 (1956). In that case, Bernhardt, a New York resident, had entered into an employment contract (containing an arbitration clause) in New York with Polygraphic, a New York corporation. But, Bernhardt "was to perform" that contract after he "later became a resident of Vermont." Id., at 199. This Court was faced with the question whether, in light of Erie, a federal court should apply the Federal Arbitration Act in a diversity case when faced with state law hostile to arbitration. 350 U. S., at 200. The Court did not reach that question, however, for it decided that the contract itself did not "involv[e]" interstate commerce and therefore fell outside the Act. Id., at 200-202. Since Congress, constitutionally speaking, could have applied the Act to Bernhardt's contract, say the parties, how then can we say that the Act's word "involving" reaches as far as the Commerce Clause itself?

The best response to this argument is to point to the way in which the Court reasoned in Bernhardt, and to what the Court said. It said that the reason the Act did not apply to Bernhardt's contract was that there was

Thus, the Court interpreted the words "involving commerce" as broadly as the words "affecting commerce"; and, as we have said, these latter words normally mean a full exercise of constitutional power. At the same time, the Court's opinion does not discuss the implications of the "interstate" facts to which the respondents now point. For these reasons, Bernhardt does not require us to narrow the scope of the word "involving." And, we conclude that the word "involving," like "affecting," signals an intent to exercise Congress' commerce power to the full.

IV

Section 2 applies where there is "a contract evidencing a transaction involving commerce." 9 U. S. C. § 2 (emphasis added). The second interpretive question focuses on the italicized words. Does "evidencing a transaction" mean only that the transaction (that the contract "evidences") must turn out, in fact, to have involved interstate commerce? Or, does it mean more?

Many years ago, Second Circuit Chief Judge Lumbard said that the phrase meant considerably more. He wrote:

The Supreme Court of Alabama and several other courts have followed this view, known as the "contemplation of the parties" test. See supra, at 269-270.

We find the interpretive choice difficult, but for several reasons we conclude that the first interpretation ("commerce in fact") is more faithful to the statute than the second ("contemplation of the parties"). First, the "contemplation of the parties" interpretation, when viewed in terms of the statute's basic purpose, seems anomalous. That interpretation invites litigation about what was, or was not, "contemplated." Why would Congress intend a test that risks the very kind of costs and delay through litigation (about the circumstances of contract formation) that Congress wrote the Act to help the parties avoid? See Moses H. Cone Memorial Hospital v. Mercury Constr. Corp., 460 U.S. 1, 29 (1983) (the Act "calls for a summary and speedy disposition of motions or petitions to enforce arbitration clauses").

Moreover, that interpretation too often would turn the validity of an arbitration clause on what, from the perspective of the statute's basic purpose, seems happenstance, namely, whether the parties happened to think to insert a reference to interstate commerce in the document or happened to mention it in an initial conversation. After all, parties to a sales contract with an arbitration clause might naturally think about the goods sold, or about arbitration, but why should they naturally think about an interstate commerce connection?

Further, that interpretation fits awkwardly with the rest of § 2. That section, for example, permits parties to agree to submit to arbitration "an existing controversy arising out of" a contract made earlier. Why would Congress want to risk nonenforceability of this later arbitration agreement (even if fully connected with interstate commerce) simply because the parties did not properly "contemplate" (or write about) the interstate aspects of the earlier contract? The first interpretation, requiring only that the "transaction" in fact involve interstate commerce, avoids this anomaly, as it avoids the other anomalous effects growing out of the "contemplation of the parties" test.

Second, the statute's language permits the "commerce in fact" interpretation. That interpretation, we concede, leaves little work for the word "evidencing" (in the phrase "a contract evidencing a transaction") to perform, for every contract evidences some transaction. But, perhaps Congress did not want that word to perform much work. The Act's history, to the extent informative, indicates that the Act's supporters saw the Act as part of an effort to make arbitration agreements universally enforceable. They wanted to "get a Federal law" that would "cover" areas where the Constitution authorized Congress to legislate, namely, "interstate and foreign commerce and admiralty." Joint Hearings 16 (testimony of Julius H. Cohen). They urged Congress to model the Act after a New York statute that made enforceable a written arbitration provision "in a written contract," Act of Apr. 19, 1920, ch. 275, § 2, 1920 N. Y. Laws 803, 804. Hearing on S. 4213 and S. 4214 before the Subcommittee of the Senate Committee on the Judiciary, 67th Cong., 4th Sess., 2 (1923) (testimony of Charles L. Bernheimer). Early drafts made enforceable a written arbitration provision "in any contract or maritime transaction or transaction involving commerce." S. 4214, 67th Cong., 4th Sess., § 2 (1922) (emphasis added); S. 1005, 68th Cong., 1st Sess. (1923); H. R. 646, 68th Cong., 1st Sess. (1924). Members of Congress, looking at that phrase, might have thought the words "any contract" standing alone went beyond Congress' constitutional authority. And, if so, they might have simply connected those words with the later words "transaction involving commerce," thereby creating the phrase that became law. Nothing in the Act's history suggests any other, more limiting, task for the language.

Third, the basic practical argument underlying the "contemplation of the parties" test was, in Chief Judge Lumbard's words, the need to "be cautious in construing the act lest we excessively encroach on the powers which Congressional policy, if not the Constitution, would reserve to the states." Metro Industrial Painting Corp., supra, at 386 (concurring opinion). The practical force of this argument has diminished in light of this Court's later holdings that the Act does displace state law to the contrary. See Southland Corp. v. Keating, 465 U. S., at 10-16; Perry v. Thomas, 482 U. S., at 489-492.

Finally, we note that an amicus curiae argues for an "objective" ("reasonable person" oriented) version of the "contemplation of the parties" test on the ground that such an interpretation would better protect consumers asked to sign form contracts by businesses. We agree that Congress, when enacting this law, had the needs of consumers, as well as others, in mind. See S. Rep. No. 536, 68th Cong., 1st Sess., 3 (1924) (the Act, by avoiding "the delay and expense of litigation," will appeal "to big business and little business alike, . . . corporate interests [and] . . . individuals"). Indeed, arbitration's advantages often would seem helpful to individuals, say, complaining about a product, who need a less expensive alternative to litigation. See, e. g., H. R. Rep. No. 97-542, p. 13 (1982) ("The advantages of arbitration are many: it is usually cheaper and faster than litigation; it can have simpler procedural and evidentiary rules; it normally minimizes hostility and is less disruptive of ongoing and future business dealings among the parties; it is often more flexible in regard to scheduling of times and places of hearings and discovery devices . . ."). And, according to the American Arbitration Association (also an amicus here), more than one-third of its claims involve amounts below $10,000, while another third involve claims of $10,000 to $50,000 (with an average processing time of less than six months). App. to Brief for American Arbitration Association as Amicus Curiae 26-27.

We are uncertain, however, just how the "objective" version ofthe "contemplation" test would help consumers. Sometimes, of course, it would permit, say, a consumer with potentially large damages claims to disavow a contract's arbitration provision and proceed in court. But, if so, it would equally permit, say, local business entities to disavow a contract's arbitration provisions, thereby leaving the typical consumer who has only a small damages claim (who seeks, say, the value of only a defective refrigerator or television set) without any remedy but a court remedy, the costs and delays of which could eat up the value of an eventual small recovery.

In any event, § 2 gives States a method for protecting consumers against unfair pressure to agree to a contract with an unwanted arbitration provision. States may regulate contracts, including arbitration clauses, under general contract law principles and they may invalidate an arbitration clause "upon such grounds as exist at law or in equity for the revocation of any contract." 9 U. S. C. § 2 (emphasis added). What States may not do is decide that a contract is fair enough to enforce all its basic terms (price, service, credit), but not fair enough to enforce its arbitration clause. The Act makes any such state policy unlawful, for that kind of policy would place arbitration clauses on an unequal "footing," directly contrary to the Act's language and Congress' intent. See Volt Information Sciences, Inc., 489 U. S., at 474.

For these reasons, we accept the "commerce in fact" interpretation, reading the Act's language as insisting that the "transaction" in fact "involv[e]" interstate commerce, even if the parties did not contemplate an interstate commerce connection.

V

The parties do not contest that the transaction in this case, in fact, involved interstate commerce. In addition to the multistate nature of Terminix and Allied-Bruce, the termite-treating and house-repairing material used by Allied-Bruce in its (allegedly inadequate) efforts to carry out the terms of the Plan, came from outside Alabama.

Consequently, the judgment of the Supreme Court of Alabama is reversed, and the case is remanded for further proceedings not inconsistent with this opinion.

It is so ordered.

Justice O'Connor, concurring.

I agree with the Court's construction of § 2 of the Federal Arbitration Act. As applied in federal courts, the Court's interpretation comports fully with my understanding of congressional intent. A more restrictive definition of "evidencing" and "involving" would doubtless foster prearbitration litigation that would frustrate the very purpose of the statute. As applied in state courts, however, the effect of a broad formulation of § 2 is more troublesome. The reading of § 2 adopted today will displace many state statutes carefully calibrated to protect consumers, see, e. g., Mont. Code Ann. § 27-5—114(2)(b) (1993) (refusing to enforce arbitration clauses in consumer contracts where the consideration is $5,000 or less), and state procedural requirements aimed at ensuring knowing and voluntary consent, see, e. g., S. C. Code Ann. § 15-48-10(a) (Supp. 1993) (requiring that notice of arbitration provision be prominently placed on first page of contract). I have long adhered to the view, discussed below, that Congress designed the Federal Arbitration Act to apply only in federal courts. But if we are to apply the Act in state courts, it makes little sense to read § 2 differently in that context. In the end, my agreement with the Court's construction of § 2 rests largely on the wisdom of maintaining a uniform standard.

I continue to believe that Congress never intended the Federal Arbitration Act to apply in state courts, and that this Court has strayed far afield in giving the Act so broad a compass. See Southland Corp. v. Keating, 465 U.S. 1, 21-36 (1984) (O'Connor, J., dissenting); see also Perry v. Thomas, 482 U.S. 483, 494-495 (1987) (O'Connor, J., dissenting); York International v. Alabama Oxygen Co., 465 U.S. 1016 (1984) (O'Connor, J., dissenting from remand). We have often said that the pre-emptive effect of a federal statute is fundamentally a question of congressional intent. See, e. g., Cipollone v. Liggett Group, Inc., 505 U.S. 504, 516 (1992); English v. General Elec. Co., 496 U.S. 72, 78-79 (1990); Schneidewind v. ANR Pipeline Co., 485 U.S. 293, 299 (1988); Rice v. Santa Fe Elevator Corp., 331 U.S. 218, 230 (1947). Indeed, we have held that "`[w]here . . . the field which Congress is said to have pre-empted' includes areas that have `been traditionally occupied by the States,' congressional intent to supersede state laws must be `clear and manifest.' " English, supra, at 79, quoting Jones v. Rath Packing Co., 430 U.S. 519, 525 (1977). Yet, over the past decade, the Court has abandoned all pretense of ascertaining congressional intent with respect to the Federal Arbitration Act, building instead, case by case, an edifice of its own creation. See Perry v. Thomas, supra, at 493 (Stevens, J., dissenting) ("It is only in the last few years that the Court has effectively rewritten the statute to give it a pre-emptive scope that Congress certainly did not intend"). I have no doubt that Congress could enact, in the first instance, a federal arbitration statute that displaces most state arbitration laws. But I also have no doubt that, in 1925, Congress enacted no such statute.

Were we writing on a clean slate, I would adhere to that view and affirm the Alabama court's decision. But, as the Court points out, more than 10 years have passed since Southland, several subsequent cases have built upon its reasoning, and parties have undoubtedly made contracts in reliance on the Court's interpretation of the Act in the interim. After reflection, I am persuaded by considerations of stare decisis, which we have said "have special force in the area of statutory interpretation," Patterson v. McLean Credit Union, 491 U.S. 164, 172-173 (1989), to acquiesce in today's judgment. Though wrong, Southland has not proved unworkable, and, as always, "Congress remains free to alter what we have done." Ibid.

Today's decision caps this Court's effort to expand the Federal Arbitration Act. Although each decision has built logically upon the decisions preceding it, the initial building block in Southland laid a faulty foundation. I acquiesce in today's judgment because there is no "special justification" to overrule Southland. Arizona v. Rumsey, 467 U.S. 203, 212 (1984). It remains now for Congress to correct this interpretation ifit wishes to preserve state autonomy in state courts.

Justice Scalia, dissenting.

I have previously joined two judgments of this Court that rested upon the holding of Southland Corp. v. Keating, 465 U.S. 1 (1984). See Volt Information Sciences, Inc. v. Board of Trustees of Leland Stanford Junior Univ., 489 U.S. 468 (1989); Perry v. Thomas, 482 U.S. 483 (1987). In neither of those cases, however, did any party ask that Southland be overruled, and it was therefore not necessary to consider the question. In the present case, by contrast, one of respondents' central arguments is that Southland was wrongly decided, and their request for its overruling has been supported by an amicus brief signed by the attorneys general of 20 States. For the reasons set forth in Justice Thomas' opinion, which I join, I agree with the respondents (and belatedly with Justice O'Connor) that Southland clearly misconstrued the Federal Arbitration Act.

I do not believe that proper application of stare decisis prevents correction of the mistake. Adhering to Southland entails a permanent, unauthorized eviction of state-court power to adjudicate a potentially large class of disputes. Abandoning it does not impair reliance interests to a degree that justifies this evil. Primary behavior is not affected: No rule of conduct is retroactively changed, but only (perhaps) the forum in which violation is to be determined and remedied. I doubt that many contracts with arbitration clauses would have been forgone, or entered into only for significantly higher remuneration, absent the Southland guarantee. Where, moreover, reliance on Southland did make a significant difference, rescission of the contract for mistake of law would often be available. See 3 A. Corbin, Corbin on Contracts § 616 (1960 ed. and Supp. 1992); Restatement (Second) of Contracts § 152 (1979).

I shall not in the future dissent from judgments that rest on Southland. I will, however, stand ready to join four other Justices in overruling it, since Southland will not become more correct over time, the course of future lawmaking seems unlikely to be affected by its existence, cf. Pennsylvania v. Union Gas Co., 491 U.S. 1, 34-35 (1989) (Scalia, J., concurring in part and dissenting in part), and the accumulated private reliance will not likely increase beyond the level it has already achieved (few contracts not terminable at will have more than a 5-year term).

For these reasons, I respectfully dissent from the judgment of the Court.

Justice Thomas, with whom Justice Scalia joins, dissenting.

I disagree with the majority at the threshold of this case, and so I do not reach the question that it decides. In my view, the Federal Arbitration Act (FAA) does not apply in state courts. I respectfully dissent.

I

In Southland Corp. v. Keating, 465 U.S. 1 (1984), this Court concluded that § 2 of the FAA "appl[ies] in state as well as federal courts," id., at 12, and "withdr[aws] the power of the states to require a judicial forum for the resolution of claims which the contracting parties agreed to resolve by arbitration," id., at 10. In my view, both aspects of Southland are wrong.

A

Section 2 of the FAA declares that an arbitration clause contained in "a contract evidencing a transaction involving commerce" shall be "valid, irrevocable, and enforceable, save upon such grounds as exist at law or in equity for the revocation of any contract." 9 U. S. C. § 2; see also § 1 (defining "commerce," as relevant here, to mean "commerce among the several States or with foreign nations"). On its face, and considered out of context, § 2 draws no apparent distinction between federal courts and state courts. But not until 1959—nearly 35 years after Congress enacted the FAA—did any court suggest that § 2 applied in state courts. See Robert Lawrence Co. v. Devonshire Fabrics, Inc., 271 F.2d 402, 407 (CA2 1959), cert. dism'd, 364 U.S. 801 (1960). No state court agreed until the 1960's. See, e. g., REA Express v. Missouri Pacific R. Co., 447 S.W.2d 721, 726 (Tex. Civ. App. 1969) (stating that the FAA applies but noting that it had been waived in the case at hand); cf. Rubewa Products Co. v. Watson's Quality Turkey Products, Inc., 242 A.2d 609, 613 (D. C. 1968) (same). This Court waited until 1984 to conclude, over a strong dissent by Justice O'Connor, that § 2 extends to the States. See Southland, supra, at 10-16.

The explanation for this delay is simple: The statute that Congress enacted actually applies only in federal courts. At the time of the FAA's passage in 1925, laws governing the enforceability of arbitration agreements were generally thought to deal purely with matters of procedure rather than substance, because they were directed solely to the mechanisms for resolving the underlying disputes. As then-Judge Cardozo explained: "Arbitration is a form of procedure whereby differences may be settled. It is not a definition of the rights and wrongs out of which differences grow." Berkovitz v. Arbib & Houlberg, Inc., 230 N.Y. 261, 270, 130 N. E. 288, 290 (1921) (holding the New York arbitration statute of 1920, from which the FAA was copied, to be purely procedural).1 It would have been extraordinary for Congress to attempt to prescribe procedural rules for state courts. See, e. g., Ex parte Gounis, 304 Mo. 428, 437, 263 S. W. 988, 990 (1924) (describing the rule that Congress cannot "regulate or control [state courts'] modes of procedure" as one of the "general principles which have come to be accepted as settled constitutional law"). And because the FAA was enacted against this general background, no one read it as such an attempt. See, e. g., Baum & Pressman, The Enforcement of Commercial Arbitration Agreements in the Federal Courts, 8 N. Y. U. L. Q. Rev. 428, 459 (1931) (noting that the FAA "does not purport to extend its teeth to state proceedings," though arguing that it constitutionally could have done so); 6 S. Williston & G. Thompson, Law of Contracts 5368 (rev. ed. 1938) ("Inasmuch as arbitration acts are deemed procedural, the [FAA] applies only to the federal courts . . ." (footnote omitted)); cf. Southland, 465 U. S., at 25-29 (O'Connor, J., dissenting) (describing "unambiguous" legislative history to this effect).

Indeed, to judge from the reported cases, it appears that no state court was even asked to enforce the statute for many years after the passage of the FAA. Federal courts, for their part, refused to apply state arbitration statutes in cases to which the FAA was inapplicable. See, e. g., California Prune & Apricot Growers' Assn. v. Catz American Co., 60 F.2d 788 (CA9 1932). Their refusal was not the outgrowth of this Court's decision in Swift v. Tyson, 16 Pet. 1 (1842), which held that certain categories of state judicial decisions were not "laws" for purposes of the Rules of Decisions Act and hence were not binding in federal courts; even under Swift, state statutes unambiguously constituted "laws." Rather, federal courts did not apply the state arbitration statutes because the statutes were not considered substantive laws. See California Prune, supra, at 790 ("It is undoubtedly true that a federal court in proper cases may enforce state laws; but this principle is applicable only when the state legislation invoke[d] creates or establishes a substantive or general right"). In short, state arbitration statutes prescribed rules for the state courts, and the FAA prescribed rules for the federal courts.

It is easy to understand why lawyers in 1925 classified arbitration statutes as procedural. An arbitration agreement is a species of forum-selection clause: Without laying down any rules of decision, it identifies the adjudicator of disputes. A strong argument can be made that such forumselection clauses concern procedure rather than substance. Cf. Fed. Rules Civ. Proc. 73 (district court, with consent of the parties, may refer case to magistrate for resolution), 53 (district court may refer issues to special master). And if a contractual provision deals purely with matters of judicial procedure, one might well conclude that questions about whether and how it will be enforced also relate to procedure.

The context of § 2 confirms this understanding of the FAA's original meaning. Most sections of the statute plainly have no application in state courts, but rather prescribe rules either for federal courts or for arbitration proceedings themselves. Thus, § 3 provides:

Section 4 addresses the converse situation, in which a party breaches an arbitration agreement not by filing a lawsuit but rather by refusing to submit to arbitration:

The Act then turns its attention to the covered arbitration proceedings themselves, treating the arbitration forum as an extension of the federal courts. Section 7, for instance, provides that the fees for witnesses "shall be the same as the fees of witnesses before masters of the United States courts"; it adds that if a witness neglects a summons to appear at an arbitration hearing,

Likewise, when the arbitrator eventually issues an award, either party (absent contrary directions in the agreement) may apply to "the United States court in and for the district within which such award was made" for an order confirming the award. § 9. The district court may also vacate or modify the award in a few specified circumstances, §§ 10-11, but generally it will simply enter a confirmatory judgment, § 9, which is then docketed and given the same effect as a judgment in an ordinary civil case, § 13.

Despite the FAA's general focus on the federal courts, of course, § 2 itself contains no such explicit limitation. But the text of the statute nonetheless makes clear that § 2 was not meant as a statement of substantive law binding on the States. After all, if § 2 really was understood to "creat[e] federal substantive law requiring the parties to honor arbitration agreements," Southland, supra, at 15, n. 9, then the breach of an arbitration agreement covered by § 2 would give rise to a federal question within the subject-matter jurisdiction of the federal district courts. See 28 U. S. C. § 1331. Yet the ensuing provisions of the Act, without expressly taking away this jurisdiction, clearly rest on the assumption that federal courts have jurisdiction to enforce arbitration agreements only when they would have had jurisdiction over the underlying dispute. See 9 U. S. C. §§ 3, 4, 8. In other 4, one words, the FAA treats arbitration simply as means of resolving disputes that lie within the jurisdiction of the federal courts; it makes clear that the breach of a covered arbitration agreement does not itself provide any independent basis for such jurisdiction. Even the Southland majority was forced to acknowledge this point, conceding that § 2 "does not create any independent federal-question jurisdiction under 28 U. S. C. § 1331 or otherwise." 465 U. S., at 15, n. 9. But the reason that § 2 does not give rise to federalquestion jurisdiction is that it was enacted as a purely procedural provision. For the same reason, it applies only in the federal courts.

The distinction between "substance" and "procedure" acquired new meaning after Erie R. Co. v. Tompkins, 304 U.S. 64 (1938). Thus, in 1956 we held that for Erie purposes, the question whether a court should stay litigation brought in breach of an arbitration agreement is one of "substantive" law. Bernhardt v. Polygraphic Co. of America, Inc., 350 U.S. 198, 203-204. But this later development could not change the original meaning of the statute that Congress enacted in 1925. Although Bernhardt classified portions of the FAA as "substantive" rather than "procedural," it does not mean that they were so understood in 1925 or that Congress extended the FAA's reach beyond the federal courts.

When Justice O'Connor pointed out the FAA's original meaning in her Southland dissent, see 465 U. S., at 25-30, the majority offered only one real response. If § 2 had been considered a purely procedural provision, the majority reasoned, Congress would have extended it to all contracts rather than simply to maritime transactions and "contract[s] evidencing a transaction involving [interstate or foreign] commerce." See id. , at 14. Yet Congress might well have thought that even if it could have called upon federal courts to enforce arbitration agreements in every single case that came before them, there was no federal interest in doing so unless interstate commerce or maritime transactions were involved. This conclusion is far more plausible than Southland `s idea that Congress both viewed § 2 as a statement of substantive law and believed that it created no federalquestion jurisdiction.

Even if the interstate commerce requirement raises uncertainty about the original meaning of the statute, we should resolve the uncertainty in light of core principles of federalism. While "Congress may legislate in areas traditionally regulated by the States" as long as it "is acting within the powers granted it under the Constitution," we assume that "Congress does not exercise [this power] lightly." Gregory v. Ashcroft, 501 U.S. 452, 460 (1991). To the extent that federal statutes are ambiguous, we do not read them to displace state law. Rather, we must be "absolutely certain" that Congress intended such displacement before we give pre-emptive effect to a federal statute. Id., at 464. In 1925, the enactment of a "substantive" arbitration statute along the lines envisioned by Southland would have displaced an enormous body of state law: Outside of a few States, predispute arbitration agreements either were wholly unenforceable or at least were not subject to specific performance. See generally Note to Williams v. Branning Mfg. Co., 47 L. R. A. (n.s.) 337 (1914) (detailed listing of state cases). Far from being "absolutely certain" that Congress swept aside these state rules, I am quite sure that it did not.

B

Suppose, however, that the first aspect of Southland was correct: § 2 requires States to enforce the covered arbitration agreements and pre-empts all contrary state law. There still would be no textual basis for Southland `s suggestion that § 2 requires the States to enforce those agreements through the remedy of specific performance—that is, by forcing the parties to submit to arbitration. A contract surely can be "valid, irrevocable, and enforceable" even though it can be enforced only through actions for damages. Thus, on the eve of the FAA's enactment, this Court described executory arbitration agreements as being "valid" and as creating "a perfect obligation" under federal law even though federal courts refused to order their specific performance. See Red Cross Line v. Atlantic Fruit Co., 264 U.S. 109, 120-123 (1924).2

To be sure, §§ 3 and 4 of the FAA require that federal the plaintiffs to their promise not to take their claims straight to court, while § 4 holds the defendants to their promise to submit to arbitration rather than making the other party sue them. Had this case arisen in one of the "courts of the United States," it is § 3 that would have been relevant. Upon proper motion, the court would have been obliged to grant a stay pending arbitration, unless the contract between the parties did not "evidenc[e] a transaction involving [interstate] commerce." See Bernhardt, supra, at 202 (holding that § 3 is limited to the arbitration agreements that § 2 declares valid). Because this case arose in the courts of Alabama, however, petitioners are forced to contend that § 2 imposes precisely the same obligation on all courts (both federal and state) that § 3 imposes solely on federal courts. Though Southland supports this argument, it simply cannot be correct, or § 3 would be superfluous.

Alabama law brings these issues into sharp focus. Citing "public policy" grounds that reach back to Bozeman v. Gilbert, 1 Ala. 90 (1840), Alabama courts have declared that predispute arbitration agreements are "void." See, e. g., Wells v. Mobile County Bd. of Realtors, 387 So.2d 140, 144 (Ala. 1980). But a separate state statute also includes "[a]n agreement to submit a controversy to arbitration" among the obligations that "cannot be specifically enforced" in Alabama. Ala. Code § 8-1—41 (1975). Especially in light of the Gregory v. Ashcroft presumption, § 2—even if applicable to the States—is most naturally read to pre-empt only Alabama's common-law rule and not the state statute; the statute does not itself make executory arbitration agreements invalid, revocable, or unenforceable, any more than the inclusion of "[a]n obligation to render personal service" in the same statutory provision means that employment contracts are invalid in Alabama. In the case at hand, the specific-enforcement statute appears to provide an adequate ground for the denial of petitioners' motion for a stay.

II

Rather than attempting to defend Southland on its merits, petitioners rely chiefly on the doctrine of stare decisis in urging us to adhere to our mistaken interpretation of the FAA. See Reply Brief for Petitioners 3-6. In my view, that doctrine is insufficient to save Southland.

The majority (ante, at 272-273) and Justice O'Connor (ante, at 283-284) properly focus on whether overruling Southland would frustrate the legitimate expectations of people who have drafted and executed contracts in the belief that even state courts will strictly enforce arbitration clauses. I do not doubt that innumerable contracts containing arbitration clauses have been written since 1984, or that arbitrable disputes might yet arise out of a large proportion of these contracts. Some of these contracts might well have been written differently in the absence of Southland. Still, I see no reason to think that the costs of overruling Southland are unacceptably high. Certainly no reliance interests are involved in cases like the present one, where the applicability of the FAA was not within the contemplation of the parties at the time of contracting. In many other cases, moreover, the parties will simply comply with their arbitration agreement, either on the theory that they should live up to their promises or on the theory that arbitration is the cheapest and best way of resolving their dispute. In a fair number of the remaining cases, the party seeking to enforce an arbitration agreement will be able to get into federal court, where the FAA will apply. And even if access to federal court is impossible (because § 2 creates no independent basis for federal-question jurisdiction), many cases will arise in States whose own law largely parallels the FAA. Only Alabama, Mississippi, and Nebraska still hold all executory arbitration agreements to be unenforceable, though some other States refuse to enforce particular classes of such agreements. See Strickland, The Federal Arbitration Act's Interstate Commerce Requirement: What's Left for State Arbitration Law?, 21 Hofstra L. Rev. 385, 401-403, and n. 93 (1992).

Quoting Arizona v. Rumsey, 467 U.S. 203, 212 (1984), Justice O'Connor nonetheless acquiesces in the majority's judgment "because there is no `special justification' to overrule Southland. " Ante, at 284. Even under this approach, the necessity of "preserv[ing] state autonomy in state courts," ibid., seems sufficient to me.

But suppose that stare decisis really did require us to abide by Southland `s holding that § 2applies to the States. The doctrine still would not require us to follow Southland `s suggestion that § 2 requires the specific enforcement of the arbitration agreements that it covers. We accord no precedential weight to mere dicta, and this latter suggestion was wholly unnecessary to the decision in Southland. The arbitration agreement at issue there, if valid at all with respect to the particular claims in dispute, clearly was subject to specific performance under state law; indeed, the state trial court had already compelled arbitration for all the other claims raised in the complaint. See Southland, 465 U. S., at 4; Cal. Civ. Proc. Code Ann. §§ 1281.2, 1281.4 (West 1982). Accordingly, the only question properly before the Southland Court was whether § 2 pre-empted a separate state law declaring the arbitration agreement "void" as applied to the remaining claims. See 465 U. S., at 10 (discussing Cal. Corp. Code Ann. § 31512 (West 1977)). The same can be said for Perry v. Thomas, 482 U.S. 483 (1987), in which we again held that § 2 pre-empted a California statute that (as we had observed in a prior case, see Merrill Lynch, Pierce, Fenner & Smith, Inc. v. Ware, 414 U.S. 117, 133 (1973)) made certain arbitration clauses "unenforceable." We have subsequently reserved judgment about the extent to which state courts must enforce arbitration agreements through the mechanisms that §§ 3 and 4 of the FAA prescribe for the federal courts. See Volt Information Sciences, Inc. v. Board of Trustees of Leland Stanford Junior Univ., 489 U.S. 468, 477 (1989). Cf. McDermott Int'l, Inc. v. Lloyds Underwriters of London, 944 F.2d 1199, 1210 (CA5 1991) ("We conclude from the Supreme Court's opinions that state courts do not necessarily have to grant stays of conflicting litigation or compel arbitration in compliance with the FAA's sections 3 and 4"). In short, we have never actually held, as opposed to stating or implying in dicta, that the FAA requires a state court to stay lawsuits brought in violation of an arbitration agreement covered by § 2.

Because I believe that the FAA imposes no such obligation on state courts, and indeed that the statute is wholly inapplicable in those courts, I would affirm the Alabama Supreme Court's judgment.