dissenting.
This litigation involves the interpretation of Article 22 of the Warsaw Convention. The plain language of that Article, quoted ante, at 247, requires that the liability limits be determined by reference to the value of “gold at the standard of fineness of nine hundred thousandths” and then converted into our “national currency in round figures.”
The Court states that the Warsaw Convention’s liability limitation remains enforceable in United States courts, but that is not what the Court holds. The Court holds that the liability limitation agreed upon by the Convention is not enforceable in United States courts. Rather, a liability limitation set by Trans World Airlines, and accepted by the Civil Aeronautics Board, is held to be enforceable in United States courts, because that limitation is deemed to correspond more closely to the Convention’s “purposes” than the limitation *262actually selected by the Convention itself. Thus, instead of enforcing the Convention’s liability limitation, the Court has rewritten it.
I
A treaty is essentially a contract between or among sovereign nations. See Washington v. Washington Commercial Passenger Fishing Vessel Assn., 443 U. S. 658, 675 (1979). General rules of construction apply to international agreements. See Ware v. Hylton, 3 Dall. 199, 240-241 (1796) (opinion of Chase, J.). As with any written document, there “is a strong presumption that the literal meaning is the true one, especially as against a construction that is not interpretation but perversion . . . .” The Five Per Cent. Discount Cases, 243 U. S. 97, 106 (1917). International agreements, like “other contracts, . . . are to be read in the light of the conditions and circumstances existing at the time they were entered into, with a view to effecting the objects and purposes of the States thereby contracting,” Rocca v. Thompson, 223 U. S. 317, 331-332 (1912); see also Factor v. Laubenheimer, 290 U. S. 276, 295 (1933), and should be interpreted according to the “received acceptation of the terms in which they are expressed.” United States v. D’Auterive, 10 How. 609, 623 (1851).
The great object of an international agreement is to define the common ground between sovereign nations. Given the gulfs of language, culture, and values that separate nations, it is essential in international agreements for the parties to make explicit their common ground on the most rudimentary of matters. The frame of reference in interpreting treaties is naturally international, and not domestic. Accordingly, the language of the law of nations is always to be consulted in the interpretation of treaties. The Pizarro, 2 Wheat. 227, 246 (1817). See also Santovincenzo v. Egan, 284 U. S. 30, 40 (1931) (“As treaties are contracts between independent nations, their words are to be taken in their ordinary meaning ‘as understood in the public law of nations’ ”) (citation omit*263ted)); Society for the Propagation of the Gospel in Foreign Parts v. New-Haven, 8 Wheat. 464, 490 (1823). Constructions of treaties yielding parochial variations in their implementation are anathema to the raison d’Ure of treaties, and hence to the rules of construction applicable to them. Geofroy v. Biggs, 133 U. S. 258, 271 (1890) (“It is a general principle of construction with respect to treaties that they shall be liberally construed, so as to carry out the apparent intention of the parties to secure equality and reciprocity between them. As they are contracts between independent nations, in their construction words are to be taken in their ordinary meaning, as understood in the public law of nations, and not in any artificial or special sense impressed upon them by local law, unless such restricted sense is clearly intended”); see also Tucker v. Alexandroff, 183 U. S. 424, 437 (1902).
Finally, but most fundamentally, a treaty is positive law. Justice Story’s words concerning the judicial role in enforcing treaties are strikingly relevant to the issue facing us today:
“In the first place, this Court does not possess any treaty-making power. That power belongs by the constitution to another department of the Government; and to alter, amend, or add to any treaty, by inserting any clause, whether small or great, important or trivial, would be on our part an usurpation of power, and not an exercise of judicial functions. It would be to make, and not to construe a treaty. Neither can this Court supply a casus omissus in a treaty, any more than in a law. We are to find out the intention of the parties by just rules of interpretation applied to the subject matter; and having found that, our duty is to follow it as far as it goes, and to stop where that stops — whatever may be the imperfections or difficulties which it leaves behind.
“In the next place, this Court is bound to give effect to the stipulations of the treaty in the manner and to the *264extent which the parties have declared, and not otherwise. We are not at liberty to dispense with any of the conditions or requirements of the treaty, or take away any qualification or integral part of any stipulation, upon any notion of equity or general convenience, or substantial justice. The terms which the parties have chosen to fix, the forms which they have prescribed, and the circumstances under which they are to have operation, rest in the exclusive discretion of the contracting parties, and whether they belong to the essence or the modal parts of the treaty, equally give the rule to judicial tribunals. The same powers which have contracted, are alone competent to change or dispense with any formality. The doctrine of a performance cy pres, so just and appropriate in the civil concerns of private persons, belongs not to the solemn compacts of nations, so far as judicial tribunals are called upon to interpret or enforce them. We can as little dispense with forms as with substance.” The Amiable Isabella, 6 Wheat. 1, 71-73 (1821).
I — I í — I
Two years after Charles Lindbergh captured the world’s imagination by piloting the Spirit of St. Louis from New York to Paris, delegates from two dozen nations met in Warsaw and drafted an international agreement to encourage the establishment of a secure international civil aviation industry. The Warsaw Convention provided a uniform system of documentation for international flights, and, more significantly, a uniform limitation on international carriers’ liability to passengers and shippers. International uniformity, naturally, was the touchstone of the Convention. See Preamble of Convention, note following 49 U. S. C. § 1502.
Air transportation was then viewed as dangerous. The liability limitation was deemed necessary in order to enable air carriers “to attract capital that might otherwise be *265scared away by the fear of a single catastrophic accident.” Lowenfeld & Mendelsohn, The United States and the Warsaw Convention, 80 Harv. L. Rev. 497, 499 (1967).1
In settling on a particular figure for a liability limitation, the delegates at Warsaw first had to agree upon a common standard of value. They did so in unambiguous terms: the standard of value was gold of a stated fineness. Art. 22 (4). The liability limitation for cargo was set at 250 units of fine gold weighing 6514 milligrams per unit. Arts. 22(2), 22(4).2 The delegates chose an international standard of value — a standard which would be the same in Paris as it would be in New York.
The Convention, while using gold as a standard of value, recognized that all nations did not use it as a medium of ex*266change. Therefore, payment in specie was not required or anticipated by the Convention. Rather, the sum of gold set forth in the Convention could be converted into national currencies in round figures. Art. 22(4). All that making the conversion would entail is knowledge of a fact: the amount of that national currency that would exchange for the specified sum of gold.
The unsuitability of relying upon a national currency as a standard of value was demonstrated as a point of actual fact at the outset of the Convention’s deliberations on the liability limitation. The draft proposal prepared prior to the Warsaw proceedings by the Comité International Technique d’Experts Juridique Aériens — the functional equivalent of a bill reported to the floor of a legislative assembly by a committee — suggested a limitation of 100 gold francs. Prior to the Convention, however, the franc had been fluctuating in value. France had suspended the convertibility of its currency into gold since the Great War, but returned to a convertible, devalued franc in 1928. The first order of business at the Warsaw proceedings concerning the liability limitation, therefore, was to convert the liability limits contained in the draft proposal, expressed in gold francs defined by a law of August 5, 1914, to the new stabilized French franc defined by the law of June 25,1928, to be 65.5 milligrams of fine gold. This was a simple matter, for the standard of value of the two statutes was the same — gold—and there was a l-to-5 ratio between the former gold franc and the new French franc. Since the French currency had stabilized, and since the law of June 25, 1928, defined the French franc in terms of gold, the French proposed that the Convention could eliminate the phrase “the values hereabove are gold values” from the draft proposal. App. 159.
The Swiss delegate submitted a proposal corresponding to the International Railroad Convention for the calculation of values. He stated:
*267“Naturally, when we prepared our text, the French franc was variable, it has been stabilized since. But the fact that a currency has been stabilized does not imply that it is a final thing; a law can always modify another law. For this reason, in Switzerland, we have preferred to stick to the gold standard, which is the same in all countries, since there is but one quality of gold.
“We would not be opposed to refer to the French franc, but to the gold French franc, that is to say, based on a weight of gold at such and such one thousandth.
“Naturally, one can say ‘French franc’ but the French franc, it’s your national law which determines it, and one need have only a modification of the national law to overturn the essence of this provision. We must base ourselves on an international value, and we have taken the [gold] dollar. Let one take the gold French franc, it’s all the same to me, but let’s take a gold value ... as a basis of calculation, be it American or French.” Id., at 161-162 (emphasis supplied).
The French delegate resisted the Swiss proposal, pointing out that the particular formula proposed by the Swiss delegate was based on the definition of the 1914 gold franc, and then stating:
“If this Convention of air law is to be applied during one or two centuries, I would perhaps share the fears of [the Swiss delegate], but it’s a question of stabilization which was done in practically every country, for a Convention which is drawn for a few years, and I believe that when you will have fixed the present French franc, you will add nothing in saying ‘gold franc.’ What fear can you have? It is evident that the definition will correspond to the present franc.” Id., at 162.
The delegate from Great Britain inquired what would happen if the French franc were to be revalued again. The *268French delegate responded that the Convention would apply to the French franc as defined in the 1928 statute. The Swiss delegate then stated no objection to using the French franc, but objected to referring to the French statute, and proposed: “insert in our international convention the same formula as that which you have in France and we accept it.” Id., at 163. The French delegate relented, and the Convention agreed to the proposal.
Once an international standard of value had been agreed to, the remaining order of business was to select the quantitative limitation on liability in units of that standard. A liability limitation of 500 new French francs for cargo was contained in the draft proposal. The French, however, proposed a limit of 100 new French francs per kilogram. The French delegate argued that the draft proposal’s limit was too high, based upon air carriers’ representations that the 100 franc figure represented the actual value of cargo, on the average. That is, the air carriers divided the declared value of cargo by the number of kilograms of cargo for the year 1928, which yielded a figure of approximately 130 francs per kilogram. Id., at 160. The German delegate argued that the French figure was far too low, and proposed a figure of 250 francs, which was ultimately agreed upon.
A gold clause such as that contained in the Convention was common in treaties, see, e. g., Article 262 of The Treaty of Versailles; Article 214 of The Treaty of St. Germain, and Article 197 of the Treaty of Trianon, and other international agreements. The very year the Convention was drafted, a major controversy on the world financial scene was resolved respecting gold clauses which were not drawn as artfully as that contained in the Convention. The Permanent Court of International Justice, in Payment of Various Serbian Loans Issued in France, 2 Hudson W. C. 340 (1929), and Payment in Gold of Brazilian Federal Loans Issued in France, 2 Hudson W. C. 402 (1929) (Serbian and Brazilian Bond Cases) *269(July 12, 1929), held that simple references to the gold franc in certain international obligations were intended to represent a gold standard of value, 2 Hudson W. C., at 365. The court continued:
“As this standard of value was adopted by the Parties, it is not admissible to assert that the standard should not govern the payments because the depreciation in French currency was not foreseen, or, as it is insisted, could not be foreseen at the time the contracts were made. The question is not what the Parties actually foresaw, or could foresee, but what means they selected for their protection. To safeguard the repayment of the loans, they provided for payment in gold value having reference to a recognized standard [of weight and fineness].
“The conclusion at which the Court has thus arrived is not affected by the fact that, for more or less extended periods gold specie in francs or a franc at gold parity was not quoted on the money market, as was the case at the time when the loans were issued; for the value can always be fixed either by comparison with the exchange rates of currency of a country in which gold coin is actually in circulation, or, should this not be possible, by comparison with the price of gold bullion. Once the gold value is fixed, it is its equivalent in money in circulation which constitutes the amount which is payable . . . .” Id., at 366-367 (emphasis supplied).
“[To construe the bonds otherwise], in substance, would eliminate the word ‘gold’ from the bonds. The contract of the Parties cannot be treated in such a manner. When the Brazilian Government promised to pay ‘gold francs’, the reference to a well-known standard of value cannot be considered as inserted merely for literary effect, or as a routine expression without significance. *270The Court is called upon to construe the promise, not to ignore it.
. It was depreciation in value that was the object of the safeguard, not in this or that particular currency, and it was evidently for this reason that the reference was made to the well-known stability of gold value.” Id., at 422-423 (emphasis supplied).
This language is not only significant in terms of the light that it casts on the law of nations and the delegates’ contemporaneous intention, but it is also, quite simply, good law. The dissent argued that the intention of the parties was simply not clear enough from the use of the term gold franc, observing it “was possible, for instance, to stipulate, and this is frequently done, that the payments must be effected in gold francs of the same weight and standard as that of the franc in circulation at the time in France . . . .” Id., at 431 (Bustamante, J., dissenting). This, of course, is what the Convention did. Indeed, at the insistence of the Swiss delegate, the Convention went even further, and eliminated the reference to French law altogether. Moreover, the gold clause did not depend on signatory nations adhering to a gold standard. Indeed, such clauses were used in contemplation of a nation going off the gold standard as a method of describing and measuring payment with the intention of guarding against fluctuations in the value of a domestic currency. Feist v. Société Intercommunale Beige D’Électricité, [1934] A. C. 161, 171-173.
The intention of the Convention simply could not be more manifest. The plain language of the Convention, the deliberations of the delegates, and the contemporary law of nations leave no question as to the the intent of the gold clause: “Here what was intended was to assure the payment of a money debt in dollars of a value as constant as that of gold. Norman v. Baltimore & Ohio R. Co., [294 U. S. 240,] *271302 [1935]; cf. Feist v. Société Intercommunale Beige D’Électricité, L. R. [1934] A. C. 161, 172, 173. . . . Weasel words will not avail to defeat the triumph of intention when once the words are read in the setting of the whole transaction.” Holyoke Water Power Co. v. American Writing Paper Co., 300 U. S. 324, 336 (1937). The Convention was of the view that its standard of value must have intrinsic value, and that gold, valued throughout the world, was a most suitable measure of value. See 1 A. Smith, The Wealth of Nations 23-28 (1911); J. Mill, Principles of Political Economy 484-485 (1936); see also The Legal Tender Cases, 12 Wall. 457 (1871) (overruling Hepburn v. Griswold, 8 Wall. 603 (1870)); 12 Wall., at 624 (Clifford, J., dissenting); id., at 647, 650 (Field, J., dissenting); Bronson v. Rodes, 7 Wall. 229, 246, 249 (1869); see generally Hepburn v. Griswold, 8 Wall., at 607-608; Ogden v. Saunders, 12 Wheat. 213, 265 (1827). Gold was money: “an universal medium or common standard, by a comparison with which the value of all merchandise may be ascertained, or it is a sign which represents the respective values of all commodities.” 1 W. Blackstone, Commentaries *276. The fluctuating and uncertain value of paper money, a fact long taken as gospel, e. g., J. Mill, Principles of Political Economy 542-563 (1936); The Legal Tender Cases, 12 Wall., at 679 (Field, J., dissenting); Bronson v. Rodes, 7 Wall., at 246; Craig v. Missouri, 4 Pet. 410, 432 (1830), and a point hit home by the rampant inflation which had been decimating the paper currencies of Europe, including France, during the post World War I era, led the Convention expressly to reject such a standard of value.
The purpose and effect of gold clauses such as that contained in the Convention were not only well known in this country at the time the United States adhered to the Convention, such gold clauses were actually made unlawful in 1933. Congress declared that “every provision contained in or made with respect to any obligation which purports to give the obli-*272gee a right to require payment in gold or a particular kind of coin or currency, or in an amount in money of the United States measured thereby, ... to be against public policy .. . .” Joint Resolution of June 5, 1933, 48 Stat. 113, 31 U. S. C. § 463(a) (1976 ed.).3 The important point is that there simply was no question as to the purpose and effect of such a clause, absent some valid exercise of governmental power over the currency to subvert it. It “calls for the payment of value in money, measured by a stated number of gold dollars of the standard defined in the clause. Feist v. Société Intercommunale Beige d’Électricité, [1934] A. C. 161, 170-173; Serbian and Brazilian Bond Cases, P. C. I. J., series A., Nos. 20-21, pp. 32-34, 109-119. In the absence of any further exertion of governmental power, that obligation plainly could not be satisfied by payment of the same number of dollars, either specie or paper, measured by a gold dollar of lesser weight, regardless of their purchasing power or the state of our internal economy at the due date.” Perry v. United States, 294 U. S. 330, 358-359 (1935) (Stone, J., concurring); see also Norman v. Baltimore & Ohio R. Co., 294 U. S. 240 (1935); Nortz v. United States, 294 U. S. 317 (1935); id., at 364 (McReynolds, J., dissenting) (gold clauses intended to afford definite standard or measure of value and thus guard against fluctuations in currency); see generally Guaranty Trust Co. v. Henwood, 307 U. S. 247 (1939).
I — I HH h-l
The United States did not participate in the Warsaw proceedings, though it did send an observer. The Aeronautics *273Branch of the Commerce Department studied the Convention and communicated with United States air carriers, who universally gave it their strong support. Lowenfeld & Mendel-sohn, The United States and the Warsaw Convention, 80 Harv. L. Rev. 497, 502 (1967). The President submitted the Convention to the Senate in 1934, which gave its advice and consent by voice vote without committee hearings, committee reports, or floor debate. 78 Cong. Rec. 11582 (1934).
The Warsaw Convention’s limitation on liability for damage to cargo has not been altered in the past half-century so far as the United States is concerned: the United States continues to adhere to the Convention as drafted in 1929.
Conditions in the world of aviation are, of course, radically different than they were 50 years ago. The Spirit of St. Louis, and the age of aviation it represents, are relics of the past. What was then a startling and daring feat for “Lucky Lindy” is now a humdrum occurrence for millions of travelers. Air travel is among the safest forms of transportation, and the fledging venture of a half century ago is a major, established international industry today. Nevertheless, though application of the Warsaw Convention’s liability limitation is anachronistic in today’s world of aviation, we are obliged to enforce it so long as the political branches of the Government adhere to the Convention. The maxim that cessante ratione legis, cessat et ipsa lex, applicable to the common law, does not govern the judiciary in cases involving application of positive law.
Conditions in the world of finance are, of course, also radically different than they were 50 years ago, when the nations of the world were attempting to reinstitute the international gold exchange standard, abandoned at the outbreak of World War I. The gold standard as a stable medium of international exchange, and the use of gold as an official standard of value for domestic purposes, are relics of the past. What has not changed, however, is the concept of a standard of value, and the effect of adopting one item as the standard as *274opposed to another. A relic though a gold standard may be, it was the standard adopted by the Convention. The delegates to that Convention were schooled not in the theories of John Maynard Keynes, but rather, in the accepted learning of John Stuart Mill. We are as obliged to apply the standard of value agreed upon by the Convention as we are obliged to apply the liability limitation.4 Indeed, of course, it is mean*275ingless to attempt to speak of one without the other: a liability limitation has no meaning without reference to a standard of value.
“The value of a thing is what it will exchange for: the value of money is what money will exchange for; the purchasing power of money.” J. Mill, Principles of Political Economy 489 (1936). The United States, of course, clung to a gold standard until recently. That is, gold was at least in theory an official standard of value for the currency and hence the number of dollars which would exchange for a given amount of gold was set by law. When the “price” of gold was fixed by law, the conversion of French francs, merely a sum of gold, into United States dollars, also merely a sum of gold, was determined by law. See generally The Collector v. Richards, 23 Wall. 246 (1875). Gold has now been demonetized. But the Convention’s standard of value remains and the concept of value has not changed. The price of gold is simply no longer fixed by law. Gold, however, still will exchange for dollars. The rate at which a domestic currency exchanges for gold was and is the only “conversion” permitted or anticipated by the Convention. That figure is the liability limitation of the Warsaw Convention.
The $9.07-per-pound limit set in TWA’s tariff is void under Article 23 of the Convention, which nullifies “[a]ny provision tending to relieve [a] carrier of liability or to fix a lower limit than that which is laid down in this convention.” That tariff is no less void because it was accepted by the CAB. E. g., 49 U. S. C. § 1502 (CAB must exercise its authority over tar*276iffs “consistently with any obligation assumed by the United States in any treaty, convention, or agreement that may be in force between the United States and any foreign country or foreign countries . . .”).5
The drafters of the Convention would surely have agreed that the “least covert of all. modes of knavery . . . consists in calling a shilling a pound, that a debt of one hundred pounds may be cancelled by the payment of one hundred shillings.” J. Mill, Principles of Political Economy 486 (1936). Basically, TWA invites us to call a dime a dollar, in order to can*277cel a debt of 80,000 dollars by the payment of 80,000 dimes. We should not accept that invitation.
<
The approach of the Court to this litigation is quite different from mine. Rather than attempting to ascertain the intent of the Convention and then applying the liability limitation thought appropriate by the Convention, the Court considers its function to make one up, with the aid of the Civil Aeronautics Board, so long as its “choice” is “sufficiently consistent” with the broad “purposes” of the Convention. Ante, at 256.6
*278The purported textual basis in the Convention for the Court’s freewheeling approach to this case is Article 22(4), which permits the sums of gold specified therein to be con*279verted into an equivalent amount of any national currency. Ante, at 260. The Court, of course, does not convert those sums into dollars; the Court converts the standard of value selected by the Convention into a standard of value expressly rejected by the Convention. In this way, it substitutes a fixed $9.07-per-pound liability limit for a liability limit of a sum of gold per pound. The only relationship between the two figures is a historical one: that is to say, one which no longer exists.7
The Court tells us that the limit agreed to by the Convention is “fluid, uncertain, and altogether inconvenient” in *280today’s world. Ante, at 260.8 If the Convention as drafted is unworkable in today’s world, that should not be surprising. To paraphrase the majority, it was written for a few years, not for a half century of the most rapid and fundamental changes in the history of the planet. The majority takes the Convention written for a few years in the era of the Spirit of St. Louis, and rewrites it in the hope, I presume, that it will last a few more years into the age of the Space Shuttle. Just why it does so escapes me. The question whether that needs to be done and the question whether that should be done are simply not decisions for this Court to make.
The Court reaches its singularly peculiar result by concluding that the method of limiting liability chosen by the Convention would fail to effect any purpose of the Convention’s framers in light of the contemporary domestic and international monetary structure. Ante, at 255.9 Ironically, in *281essence, the Court agrees with the rationale of the Court of Appeals: that the limit is unenforceable on grounds of frustration of purpose. The Court differs with the Court of *282Appeals only in terms of the remedy. Whereas the Court of Appeals thought the appropriate remedy to be rescission of the agreement, the Court thinks the appropriate remedy is reformation of the agreement. Of course, if the premise of the Court is correct, and the liability limitation is unworkable in today’s world, there is but one remedy: amendment of the Convention by the parties.10
V
Some students of the Court take for granted that our decisions represent the will of the judges rather than the will of the law.11 This dogma may be the current fashion, but I remain convinced that such remarks reflect a profound misunderstanding of the nature of our work. Unfortunately, however, cynics — parading under the banner of legal realism — are given a measure of credibility whenever the Court bases a decision on its own notions of sound policy, rather *283than on what the law commands.12 It does so today.13 The task of revising an international treaty is not one that this Court has any authority to perform. I respectfully dissent.
These fears were epitomized by the crash of the Hindenberg in 1937, though the Warsaw Convention’s liability limitation could not save the dirigible — then a significant mode of international air transportation — from rapid extinction. See generally L. Ege, Balloons and Airships 176-221 (1973); D. Robinson, Giants in the Sky 250-315 (1973).
The liability limitation, it may be noted, was not limited to damages arising from special risks of an aircraft crash, nor did it fully protect carriers from those special risks. With respect to parties contracting with air carriers to ship goods, the Convention wrote with a broad brush: limiting liability as well for the kind of damage to goods that would be as likely to occur in any mode of transportation — such as the loss of a single item of cargo in connection with a safe flight. On the other hand, the Convention provided air carriers with no special protection whatsoever vis-á-vis third parties who might be injured by air crashes.
These units, it so happened, corresponded to the French franc as defined by a 1928 French statute. It was thus convenient to call them francs, and the Convention did so. That French statute, however, could have been repealed the day after the delegates adjourned their meeting, and it would not have affected the liability limitation. For the delegates had selected as the standard of value a commodity with a value independent of any one nation’s control; indeed, a commodity perceived to have “intrinsic” value — a commodity individuals had valued before there were nations, and would value whether or not national governments made it an official medium of exchange.
Since the Convention was adhered to by the United States subsequent to the passage of this statute, presumably the Convention was an exception to this prohibition, unless we are to indulge the inference that Congress simultaneously abrogated and ratified the Convention. In any event, this statute has been repealed with respect to obligations incurred after October 27, 1977. 31 U. S. C. § 5118(d)(2). See 123 Cong. Rec. 33219-33220 (1977).
A staff memorandum addressed to the Civil Aeronautics Board on March 18, 1980, by the Chief, Policy Development Division, contains this telling comment on the background of the Warsaw liability limits and their contemporary relevance:
“The Warsaw Convention was negotiated during the late 1920’s when the aviation industry was in its infancy. The minutes of the negotiations show that the primary concerns of the drafters are no longer of great importance to the industry. In addition, their assumptions about how the liability limitation mechanism would work were erroneous.
“In 1929, air travel was perceived by the public and, more importantly, by insurance companies to be an extremely risky mode of transportation. A major justification for limiting liability was that, unless carriers could present potential insurers with some degree of predictability in estimating damages from aircraft accidents, they would have great difficulty in obtaining coverage. Furthermore, the delegates had little sympathy for anyone foolish enough to board an airplane without enough personal insurance to provide for his widow (or her widower) and children should the plane crash. Over the years, air travel has become one of the safest modes of transportation, and airlines, even those operating under circumstances where they cannot limit their liability for death or personal injury, have no special difficulties finding insurers.
“The minutes also reflect the delegates’ rationale for using gold as the unit of reference in determining carrier liability limits, instead of pegging the limits to some particular currency like the dollar or the franc, with no reference to the metal. . . .
“Warsaw has become an anachronism, yet various attempts to amend it have become stalled by the ratification process. While the Guatemala and Montreal Protocols would have revised the limits upward, the proposed limits were still relatively low, and none of the proposals contained a mechanism that would provide for periodical adjustments to compensate for inflation. Article 22 tied to gold, however, overcompensates for inflation to the point that the industry may view it as a sort of passengers’ affirmative *275action plan that is supposed to make up for years of unreasonably low limits. Although gold-based limits may not be the most rational approach to allocating risks between carriers and consumers of international air transportation, the framers of Warsaw deliberately adopted this approach— expressly rejecting a dollar or some other currency-based system — and probably left us no flexibility as long as Warsaw remains in effect.” App. 49-51 (footnote omitted).
The majority asks why the CAB was “disqualified from setting a similar conversion rate one year after Congress stopped doing so.” Ante, at 255, n. 26 (emphasis supplied). So framed, the majority’s question seems to answer itself. What the majority ignores is that the powers delegated to the CAB, 49 U. S. C. § 1502, previously quoted supra, must be exercised consistently with any convention in force, including the Warsaw Convention.
Although the Court makes a ritual of referring to the “CAB’s choice” in its opinion (implying that it might not have made the same choice independently), ante, at 255, 256, 257, 258, 259, this is a suit between two private parties, not a declaratory judgment action challenging the CAB’s exercise of its authority over tariffs. If it were, one would frame the question as whether the CAB has exercised its authority consistently with the Convention, though that is not how the majority frames the question presented at the outset of its opinion. Ante, at 245. In any event, the answer to that question, of course, turns on what the Convention means. On this matter, the CAB’s views are not entitled to any special deference. While the Convention is a limitation on the CAB’s powers, the CAB is not the governmental organ charged with enforcing the liability limitation — that responsibility rests with the courts of the United States. The Solicitor General correctly observes: “The Warsaw Convention is a self-executing treaty that provides a source of rules of decision applicable in United States courts without requiring enactment of any supplementary implementing legislation by Congress.” Brief for United States as Amicus Curiae 16. Courts, not the CAB, render money judgments, and in rendering those money judgments, courts must apply the rules of decision provided by the Convention. Furthermore, even if some deference were to be accorded to the CAB’s views, the CAB’s position would have to be rejected since it conflicts with the plain meaning of the Convention.
The Court does speak of intent at the close of its opinion. Ante, at 260. Its approach to ascertaining intent is novel, to say the least. It postulates the general “purposes” of the liability limitation. It then divines the effects of a gold-based liability in today’s world. It finds the general purposes inconsistent with a gold-based limitation, and concludes that the Convention did not intend to adopt a gold-based limitation. The rather telling deficiency with this conclusion is that the Convention did adopt a gold-based limitation. The Court itself must recognize this defect, for it tells us that to achieve their purposes, “the Convention’s framers chose an international, not a parochial, standard, free from the control of any one country.” Ante, at 256 (footnote omitted). Inexplicably, however, the Court then proceeds to ignore the standard the Convention selected. It does rely upon the practice of the signatories between 1934 and 1978. Ante, at 259. Our practice was consistent with the Convention so long as the price of gold was set by law. It is no longer. One commentator argued prior to the total demonetization of gold that the market price of gold should have been used in light of the economic reality of the early 1970’s, see Heller, The Warsaw Convention and the “Two-Tier” Gold Market, 7 J. World Trade L. 126 (1973); Heller, The Value of the Gold Franc — A Different Point of View, 6 J. Mar. L. & Com. 73 (1974). The response to that argument rested on the fact that the price of gold was set by law— a fact that no longer obtains. See Boehringer Mannheim Diagnostics, Inc. v. Pan American World Airways, Inc., 531 F. Supp. 344, 353, n. 46 (SD Tex. 1981), appeal docketed, No. 81-2519 (CA5, Dec. 30, 1981).
The majority is correct in stating that I have no problem with the conclusion that our practice was consistent with the Convention when the price of *278gold was set by law. Ante, at 264-255, n. 26. We need not speculate on that score, for at the outset of the Warsaw proceedings the drafters performed the very kind of conversion they anticipated would occur when a national currency was based on the gold standard — they converted the sums expressed in gold francs in the draft proposal into equivalent sums of new French francs.
The Convention did not, of course, deprive a signatory sovereign nation of its control over its own currency. Indeed, it was a recognition of that authority which led the Swiss delegate to insist that a sum of gold — rather than a reference to the 1928 French statute — be specified in the Convention. Congress has plenary power over our currency. See, e. g., The Legal Tender Cases, 12 Wall. 457 (1871). In exercising that power, it may define the dollar in terms of gold. In doing so, and in periodically adjusting that relationship, it necessarily affects legal interests in many contractual areas. See, e. g., Norman v. Baltimore & Ohio R. Co., 294 U. S. 240 (1935). For example, when Congress passed the Par Value Modification Act in 1972, setting a new price of gold, Congress “did not suggest that the CAB should thereafter use a different conversion factor” for Warsaw Convention purposes, but “the CAB did, of course, use that price to translate the Convention’s gold-based liability limit into dollars.” Ante, at 255. This was the only “flexible implementation” of the treaty which occurred from 1934 to 1978. Ante, at 259. The CAB, of course, had no flexibility— air carriers were bound by the Convention, and bound by Congress’ decisions on monetary policy, see, e. g., Norman v. Baltimore & Ohio R. Co., supra. Congress has most recently exercised its authority over the currency by demonetizing gold completely. That has legal consequences as well. One of those consequences is that the rate of exchange between dollars and gold is no longer determined by law, and it is irrelevant that when Congress repealed the Par Value Modification Act it “did not suggest that the CAB should thereafter use a different conversion factor” for Warsaw Convention purposes. Congress has not delegated its authority over the currency to the CAB. And Congress clearly has not delegated authority to the CAB to violate Article 23 of the Convention — it has expressly stated that the CAB’s authority must be exercised consistently with our treaty obligations. Congress itself, of course, does not have the authority to violate Article 23 short of repudiating the Convention. Congress naturally could repudiate the Convention and set its own liability limitation through domestic legislation, but has not done so.
The Court might as well have selected a figure corresponding to the dollar value of 12 grams of gold in 1929, or 1934, as that existing in 1978. Indeed, if the Court fancies itself competent to decide the proper liability limit necessary to effect the “purposes” of the Convention, one wonders why it selects a figure which happens to correspond to a value of a given amount of gold at any point in time. Since it adopts such a freewheeling approach to the subject, the Court could, for example, compute the liability limit in the same basic manner as did the Convention — dividing the declared value of all air cargo by the tonnage, and then doubling that figure to arrive at a per-pound limitation.
Any fixed dollar figure lower than the actual damage figure, of course, would meet all of the purposes postulated by the Court as well as $9.07 per pound. But it appears that the Court will be willing to modify the limit it has set from time to time, for it states that its limit “might require periodic adjustment... to account both for the dollar’s changing value relative to other Western currencies and, if necessary, for changes in the conversion rates adopted by other Convention signatories.” Ante, at 256. The only reason the Court apparently does not make such an adjustment is that since 1978, “no substantial changes of either type have occurred.” Ibid. Of course, if the standard adopted by the Convention were enforced, rather than ignored, no such adjustments would need to be made by courts — for as the Court candidly admits, “[sjince gold is freely traded on an international market its price always provides a unique and internationally uniform conversion rate.” Ante, at 258. The fact that gold has always had a uniform value — since, to use the words of the Swiss delegate, there is but one quality of gold — was, of course, a major reason why it was chosen as a standard of value in an international agreement setting an internationally enforceable limitation on liability.
Compare ante, at 253 (“[W]hen the parties to a treaty continue to assert its vitality a private person who finds the continued existence of the treaty inconvenient may not invoke the doctrine [of rebus sic stantibus] on their behalf”).
The response of the framers to this assertion by the majority, I am quite sure, would be “au contraire.” Indeed, on the face of the majority opinion, only one of the purposes the Court identifies is not served by the gold standard — that of a stable and predictable limit. The other purposes, setting some limit, setting an internationally uniform limit, and linking the limit to a real value, are all achieved by the limit adopted by the Convention. In fact, the Court concedes the first purpose is served by any standard, ante, at 256, and seems to agree that the latter two purposes are better achieved by the limit adopted by the Convention, ante, at 258-259. Hence, even if I were to adopt the freewheeling approach of the Court to the question before us — that is, an independent judicial determination of which of several “choices” would “best effect” the “different and to some extent contradictory” general “purposes” of the Convention — I would find the Court’s “choice” unpersuasive.
Gold was selected because it would continue to be an internationally recognized standard of value, irrespective of what national governments did with respect to their currencies. It was also selected to guard against the fluctuating values of currencies. Gold continues to be an internationally *281recognized standard of value. The legacy of eons is not readily discarded. See generally 1 Report to Congress of the Commission on the Role of Gold in the Domestic and International Monetary Systems 98 (Mar. 1982); M. Friedman & A. Schwartz, A Monetary History of the United States 473, 684 (1971). Its nominal value is still overwhelmingly its value in exchange, not its value in use — it is not simply another commodity. Fort Knox does not house any pork bellies. Currencies are worth far less in relation to gold than they were several years ago, that is, in perhaps bygone jargon, the currencies have been substantially devalued. The variation in the relative value of gold to currencies is not the kind of variation in value that the framers of the Convention wanted to avoid — it is one that they desired. Certainly, however, the extent of the current variation is more than they could have anticipated, and currencies have not devalued in overall purchasing power to the extent that they have in relation to gold.
The Court, in support of the standard it has selected — the dollar — thinks that the relative value of various currencies has remained relatively stable in recent years (a conclusion one cannot reach based on the data relied upon by the majority). Even if that were true, all that would mean is that the currencies have not depreciated in relation to one another — it does not mean that the currencies have not devalued. The real value of those currencies has not been stable, it has been declining. The real value of gold has not been stable either, but it has not declined since 1929. The same manifestly cannot be said of the dollar.
The interests in stability and predictability, of course, do favor the Court’s choice. But the fact that the market price of gold fluctuates on a daily basis does not seem to me to present particularly serious problems. Indeed, the existence of a well-recognized daily price provides a simple point of reference for computing the exact amount of the limit on the date of shipment. The calculation of insurance rates would have to take into account the variable character of the gold market, but that is hardly á matter that underwriters are incapable of evaluating realistically. And, of course, the problem of computing their contingent tort liability for cargo would remain less difficult than the problem they confront in computing their contingent tort liability to third parties.
The idea that air transport users and carriers are forced to become unwilling speculators in the gold market is mere rhetoric. It stands history on its head even to suggest that the Warsaw Convention’s liability limit *282has been foisted upon air carriers against their will. If, because of the demonetization of gold, the gold standard makes air carriers “speculators” in the gold market, and if the air carriers find this situation unacceptable, they can urge modification of the Convention. But, of course, they have already failed to secure Senate approval of the Montreal Protocols. Today, their attorneys win the battle that their lobbyists lost.
Incredibly, the majority apparently derives some comfort from the fact that the liability limit it adopts is approximately equivalent in purchasing power to that contained in the Montreal Protocols, doing so in spite of the fact that the Montreal Protocols were rejected by the Senate. See ante, at 256-257. Indeed, the CAB accepts tariffs based on the Montreal Protocols, which passes by the majority without criticism. Ante, at 257, n. 29.
For example, one law professor recently wrote in a doctrinaire fashion: “Modem jurisprudence regards the distinction Marshall Court justices sought to make between the ‘will of the judge’ and the ‘will of the law’ as a distinction without a difference. The ‘legal’ decisions of judges are, in the modern consciousness, necessarily personal and creative. . . .
“. . . All of the modern canons for judicial behavior, and all of the modem theories of judicial performance, presuppose that judges ‘make’ law and that judicial will and legal will are thus inseparable.” White, The Working Life of The Marshall Court, 1815-1835, 70 Va. L. Rev. 1, 49-50 (1984).
Judges, of course, must perform a lawmaking function, even in cases involving statutory construction. See, e. g., R. Pound, The Spirit of the Common Law 170-175 (1921). But the limits of our authority and our ability to develop the law should always be respected. As Justice Cardozo explained:
“No doubt there is a field within which judicial judgment moves untrammeled by fixed principles. Obscurity of statute or of precedent or of customs or of morals, or collision between some or all of them, may leave the law unsettled, and cast a duty upon the courts to declare it retrospectively in the exercise of a power frankly legislative in function. . . . We must not let these occasional and relatively rare instances blind our eyes to the innumerable instances where there is neither obscurity nor collision nor opportunity for diverse judgment. ... In countless litigations, the law is so clear that judges have no discretion. They have the right to legislate within gaps, but often there are no gaps. We shall have a false view of the landscape if we look at the waste spaces only, and refuse to see the acres already sown and fruitful. I think the difficulty has its origin in the failure to distinguish between right and power, between the command embodied in a judgment and the jural principle to which the obedience of the judge is due. Judges have, of course, .the power, though not the right, to ignore the mandate of a statute, and render judgment in despite of it. They have the power, though not the right, to travel beyond the walls of the interstices, the bounds set to judicial innovation by precedent and custom. None the less, by that abuse of power, they violate the law.” B. Cardozo, The Nature of the Judicial Process 128-129 (1921).
Perhaps the majority would insist that it is merely deferring to the CAB’s notions of wise policy. If so, this would mean that the CAB could have made another choice to which the majority would have deferred. But surely the majority would not countenance selection of the choice made by the Convention itself, for the majority believes that choice would fail to effect any of the Convention’s purposes. Moreover, I cannot see how the majority could defer to a choice of the limit set forth in the Montreal Protocols, given the fact that the Senate has recently rejected the Montreal Protocols. But see n. 10, supra. Nor can one imagine the Court “deferring” to the CAB’s judgment if it selected the value of the current French franc. In short, it seems rather clear that the only potential choice of the CAB to which the majority would “defer” is the one it selected.