concurring in the judgment.
I concur only in the judgment of the Court because I believe that the text of Articles XIX and XXI of the Convention between the United States and Canada Respecting Double Taxation, Mar. 4, 1942, 56 Stat. 1405-1406, T. S. No. 983, is completely dispositive of respondents’ claim under the agreement. The Court apparently agrees. See ante, at 365-366. Given that the Treaty’s language resolves the issue presented, there is no necessity of looking further to discover “the intent of the Treaty parties,” ante, at 366, and special reason to avoid the particular materials that the Court unnecessarily consults.
I
Of course, no one can be opposed to giving effect to “the intent of the Treaty parties.” The critical question, however, is whether that is more reliably and predictably achieved by a rule of construction which credits, when it is clear, the contracting sovereigns’ carefully framed and solemnly ratified expression of those intentions and expectations, or rather one which sets judges in various jurisdictions at large to ignore that clear expression and discern a “genuine” contrary intent elsewhere. To ask that question is to answer it.
One can readily understand the appeal of making the additional argument that the plain language of a treaty (which is conclusive) does indeed effectuate the genuine intent as shown elsewhere — just as one can understand the appeal, in statutory cases, of pointing out that what the statute provides (which is conclusive) happens to be sound social policy. But using every string to one’s bow in this fashion has unfortunate implications. (“It would be wrong; and besides, it wouldn’t work.”) Here the implication is that, had the extrinsic evidence contradicted the plain language of the Treaty it would govern. That is indeed what we mistakenly said in the earlier case that the Court cites as authority for its approach. In Sumitomo Shoji America, Inc. v. Avagliano, *372457 U. S. 176, 180 (1982), we stated that “ ‘[t]he clear import of treaty language controls unless ‘application of the words of the treaty according to their obvious meaning effects a result inconsistent with the intent or expectations of its signatories.’ . . . Maximov v. United States, 873 U. S. 49, 54 (1963).” The authority quoted for that proposition in fact does not support it. In Maximov, confronted with an argument appealing to the “intent or expectations” of the signatories, we responded that “[t]he immediate and compelling answer to this contention is that . . . the language of the Convention itself not only fails to support the petitioner’s view, but is contrary to it.” Maximov v. United States, 373 U. S. 49, 54 (1963). We then continued: “Moreover, it is particularly inappropriate for a court to sanction a deviation from the clear import of a solemn treaty . . . when, as here, there is no indication that application of the words of the treaty according to their obvious meaning effects a result inconsistent with the intent or expectations of its signatories.” Ibid, (emphasis added). The import of the highlighted adverb is, of course, that it would be inappropriate to sanction a deviation from clear text even if there were indications of contrary intent. Our Sumitomo dictum separated the last clause of this quotation from its context to support precisely the opposite of what it said. Regrettably, that passage from Sumitomo is already being quoted by lower courts as “[t]he general rule in interpreting treaties.” Rainbow Navigation, Inc. v. Department of Navy, 686 F. Supp. 354, 359, n. 25 (DC 1988).
Notwithstanding the Sumitomo dictum to which the Court alludes, our traditional rule of treaty construction is that an agreement’s language is the best evidence of its purpose and its parties’ intent. In Rocca v. Thompson, 223 U. S. 317 (1912), it was urged upon us that a Treaty granting consuls the right “to intervene in the possession, administration, and judicial liquidation of the estate of the deceased” also granted them the right to administer the property of the deceased, *373since that would effectuate the Treaty’s “objects and purposes.” We conducted no separate inquiry into the intent or expectations of the signatories beyond those expressed in the text, but said simply:
“[Tjreaties are the subject of careful consideration before they are entered into, and are drawn by persons competent to express their meaning and to choose apt words in which to embody the purposes of the high contracting parties. Had it been the intention to commit the administration of estates of citizens of one country, dying in another, exclusively to the consul of the foreign nation, it would have been very easy to have declared that purpose in unmistakable terms.” Id., at 332.
That is the governing principle of interpretation. Only when a treaty provision is ambiguous have we found it appropriate to give authoritative effect to extratextual materials. See, e. g., Air France v. Saks, 470 U. S. 392, 400 (1985); Nielsen v. Johnson, 279 U. S. 47, 52 (1929).
t — I 1 — (
Even, however, if one generally regards the use of pre-ratification extrinsic materials to confirm an unambiguous text as an innocuous practice, there is special reason to object to that superfluous reference in the present case. What is distinctive here is the nature of the extratextual materials to which the Court unnecessarily refers. To discover Canada’s and the United States’ “intent and expectations,” the Court looks solely to the United States Senate floor debates that preceded the President’s ratification of the treaty. Ante, at 366-368, and nn. 7-8. The use of such materials is unprecedented. Even where the terms of the treaty are ambiguous, and resort to preratification materials is therefore appropriate, I have been unable to discover a single case in which this Court has consulted the Senate debate, committee hearings, or committee reports. It would be no more appropriate for me than it is for the Court to use the present case as the occa*374sion for pronouncing upon the legitimacy of using such materials, but it is permissible to suggest some of the arguments against it. Using preratification Senate materials, it may be said, is rather like determining the meaning of a bilateral contract between two corporations on the basis of what the board of directors of one of them thought it meant when authorizing the chief executive officer to conclude it. The question before us in a treaty case is what the two or more sovereigns agreed to, rather than what a single one of them, or the legislature of a single one of them, thought it agreed to. And to answer that question accurately, it can reasonably be said, whatever extratextual materials are consulted must be materials that reflect the mutual agreement (for example, the negotiating history) rather than a unilateral understanding. Thus, we have declined to give effect, not merely to Senate debates and committee reports, but even to an explicit condition of ratification adopted by the full Senate, when the President failed to include that in his ratification. We said:
“The power to make treaties is vested by the Constitution in the President and Senate, and, while this proviso was adopted by the Senate, there is no evidence that it ever received the sanction or approval of the President. It cannot be considered as a legislative act, since the power to legislate is vested in the President, Senate and House of Representatives. There is something, too, which shocks the conscience in the idea that a treaty can be put forth as embodying the terms of an arrangement with a foreign power or an Indian tribe, a material provision of which is unknown to one of the contracting parties, and is kept in the background to be used by the other only when the exigencies of a particular case may demand it.” New York Indians v. United States, 170 U. S. 1, 23 (1898).
Of course the Senate has unquestioned power to enforce its own understanding of treaties. It may, in the form of a reso*375lution, give its consent on the basis of conditions. If these are agreed to by the President and accepted by the other contracting parties, they become part of the treaty and of the law of the United States, see Northwestern Bands of Shoshone Indians v. United States, 324 U. S. 335, 351-352 (1945); see also Restatement (Third) of Foreign Relations Law of the United States §314 (1987). If they are not agreed to by the President, his only constitutionally permissible course is to decline to ratify the treaty, and his ratification without the conditions would presumably provide the basis for impeachment. Moreover, if Congress does not like the interpretation that a treaty has been given by the courts or by the President, it may abrogate or amend it as a matter of internal law by simply enacting inconsistent legislation. La Abra Silver Mining Co. v. United States, 175 U. S. 423, 460 (1899); Head Money Cases, 112 U. S. 580, 599 (1884). But it is a far cry from all of this to say that the meaning of a treaty can be determined, not by a reservation attached to the President’s ratification at the instance of the Senate, nor even by formal resolution of the Senate unmentioned in the President’s ratification, but by legislative history of the sort that we have become accustomed to using for purpose of determining the meaning of domestic legislation.
The American Law Institute’s Restatement of the Foreign Relations Law of the United States would permit the courts to refer to materials of the sort at issue here. See Restatement (Third) of Foreign Relations Law of the United States §314, Comment d (1987); id., §325, Reporter’s Note 5. But despite the title of the work, this must be regarded as a proposal for change rather than a restatement of existing doctrine, since the commentary refers to not a single case, of this or any other United States court, that has employed the practice. The current version of the Restatement provides no explanation for (or even acknowledgment of) this curiosity. An explanation was provided in the Proposed Official Draft of the Second Restatement, which is of some interest:
*376“There is virtually no precise decisional authority on this matter, probably because of the domestic interpretative rule, stated in § 155, that executive interpretations of international agreements are given great weight by courts in the United States or because, as explained in Comment a to this Section, the courts wish to avoid if possible creating disharmony between the international and the domestic meanings of international agreements.” Restatement (Second) of Foreign Relations Law of the United States §154, Comment b(ii) (Prop. Off. Draft 1962).
This is not the case in which to commit ourselves to an approach that significantly reduces what has hitherto been the President’s role in the interpretation of treaties, and commits the United States to a form of interpretation plainly out of step with international practice.
It can hardly have escaped the Court’s attention that the role of Senate understanding in the treaty ratification process has recently been the subject of some considerable dispute between the Senate and the Executive. See Washington Post, Mar. 19, 1988, p. All, col. 1 (discussing disagreement on the importance to be accorded to Senate understanding of the Anti-Ballistic Missile Treaty at the time of advice and consent to the President’s ratification); Washington Post, Feb. 17, 1988, p. A17, col. 1 (same); Washington Post, Feb. 6, 1988, p. Al, col. 6 (same). The first (and, as far as I am aware, the only) federal decisions relying upon preratification Senate materials for the interpretation of a treaty were issued by the District Court for the District of Columbia, in successive phases of the same controversy, last May, see Rainbow Navigation, Inc. v. Department of Navy, 686 F. Supp. 354 (1988), and last November, see Rainbow Navigation, Inc. v. Department of Navy, 699 F. Supp. 339 (1988). In the first of those cases, the court rejected the Government’s contention that its representations to the Senate regarding the meaning of a treaty are not binding as to *377the treaty’s interpretation. See 686 F. Supp., at 357-358, n. 17.* In the second of them, the Government conceded that “authoritative Executive branch representations concerning the meaning of a Treaty which form part of the basis upon which the Senate gives advice and consent are entitled to be accorded binding weight as a matter of domestic constitutional law, and the Executive branch fully accepts that it is bound by such statements.” 699 F. Supp., at 343 (quoting Defendants’ Reply Brief and Opposition to Plaintiff’s Cross-Motion for Summary Judgment 2, n. 2.). It is not clear that this latest position taken by the Government in District Court is correct, or would even be the position taken before us by the Solicitor General. It is even less clear, however, assuming that position to be correct, that Senate understandings which are not the product of Executive representations in the advice-and-consent hearings should have any relevance. It is odd, to say the least, that in the present case, where the language of the Treaty is clear, where the role of Senate reports and debates has not even been argued, and where the Solicitor General has not been requested to give us the benefit of his views on that subject, we should reach out to use such materials for the first time in two centuries of treaty construction.
The court relied in part upon testimony — reproduced in The ABM Treaty Interpretation Resolution, Report of the Committee on Foreign Relations of the United States Senate, S. Rep. No. 100-164, p. 49 (1987)— by none other than the reporter for the Restatement of Foreign Relations Law, Professor Louis Henkin. Thus, by self-exertion, so to speak, there is now at least one case that the Restatement almost restates. The qualifier is needed because, as I discuss later in text, even that case does not go as far as the Restatement (and the Court’s opinion today) would do.