with whom The Chief Justice joins, dissenting.
Even if I agreed with the questionable premise that legislation can be invalidated under the Establishment Clause on the basis of its motivation alone, without regard to its effects, I would still find no justification for today’s decision. The Louisiana legislators who passed the “Balanced Treatment for Creation-Science and Evolution-Science Act” (Balanced Treatment Act), La. Rev. Stat. Ann. §§17:286.1-17:286.7 (West 1982), each of whom had sworn to support the Constitution,1 were well aware of the potential Establishment Clause problems and considered that aspect of the legislation with great care. After seven hearings and several months of study, resulting in substantial revision of the original proposal, they approved the Act overwhelmingly and specifically articulated the secular purpose they meant it to serve. Although the record contains abundant evidence of the sincerity of that purpose (the only issue pertinent to this case), the Court today holds, essentially on the basis of “its visceral knowledge regarding what must have motivated the legislators,” 778 F. 2d 225, 227 (CA5 1985) (Gee, J., dissenting) (emphasis added), that the members of the Louisiana Legislature knowingly violated their oaths and then lied about it. I dissent. Had requirements of the Balanced Treatment Act that *611are not apparent on its face been clarified by an interpretation of the Louisiana Supreme Court, or by the manner of its implementation, the Act might well be found unconstitutional; but the question of its constitutionality cannot rightly be disposed of on the gallop, by impugning the motives of its supporters.
I
This case arrives here in the following posture: The Louisiana Supreme Court has never been given an opportunity to interpret the Balanced Treatment Act, State officials have never attempted to implement it, and it has never been the subject of a full evidentiary hearing. We can only guess at its meaning. We know that it forbids instruction in either “creation-science” or “evolution-science” without instruction in the other, § 17:286.4A, but the parties are sharply divided over what creation science consists of. Appellants insist that it is a collection of educationally valuable scientific data that has been censored from classrooms by an embarrassed scientific establishment. Appellees insist it is not science at all but thinly veiled religious doctrine. Both interpretations of the intended meaning of that phrase find considerable support in the legislative history.
At least at this stage in the litigation, it is plain to me that we must accept appellants’ view of what the statute means. To begin with, the statute itself defines “creation-science” as “the scientific evidences for creation and inferences from those scientific evidences.” §17:286.3(2) (emphasis added). If, however, that definition is not thought sufficiently helpful, the means by which the Louisiana Supreme Court will give the term more precise content is quite clear — and again, at this stage in the litigation, favors the appellants’ view. “Creation science” is unquestionably a “term of art,” see Brief for 72 Nobel Laureates et al. as Amici Curiae 20, and thus, under Louisiana law, is “to be interpreted according to [its] received meaning and acceptation with the learned in the art, trade or profession to which [it] refer[s].” La. Civ. *612Code Ann., Art. 15 (West 1952).2 The only evidence in the record of the “received meaning and acceptation” of “creation science” is found in five affidavits filed by appellants. In those affidavits, two scientists, a philosopher, a theologian, and an educator, all of whom claim extensive knowledge of creation science, swear that it is essentially a collection of scientific data supporting the theory that the physical universe and life within it appeared suddenly and have not changed substantially since appearing. See App. to Juris. Statement A-19 (Kenyon); id., at A-36 (Morrow); id., at A-41 (Miethe). These experts insist that creation science is a strictly scientific concept that can be presented without religious, reference. See id., at A-19 — A-20, A-35 (Kenyon); id., at A-36 — A-38 (Morrow); id., at A-40, A-41, A-43 (Miethe); id., at A-47, A-48 (Most); id., at A-49 (Clinkert). At this point, then, we must assume that the Balanced Treatment Act does not require the presentation of religious doctrine.
Nothing in today’s opinion is plainly to the contrary, but what the statute means and what it requires are of rather little concern to the Court. Like the Court of Appeals, 765 F. 2d 1251, 1253, 1254 (CA5 1985), the Court finds it necessary to consider only the motives of the legislators who supported the Balanced Treatment Act, ante, at 586, 593-594, 596. After examining the statute, its legislative history, and its historical and social context, the Court holds that the Louisiana Legislature acted without “a secular legislative purpose” and that the Act therefore fails the “purpose” prong of the three-part test set forth in Lemon v. Kurtzman, 403 U. S. 602, 612 (1971). As I explain below, infra, at 636-640, *613I doubt whether that “purpose” requirement of Lemon is a proper interpretation of the Constitution; but even if it were, I could not agree with the Court’s assessment that the requirement was not satisfied here.
This Court has said little about the first component of the Lemon test. Almost invariably, we have effortlessly discovered a secular purpose for measures challenged under the Establishment Clause, typically devoting no more than a sentence or two to the matter. See, e. g., Witters v. Washington Dept. of Services for Blind, 474 U. S. 481, 485-486 (1986); Grand Rapids School District v. Ball, 473 U. S. 373, 383 (1985); Mueller v. Allen, 463 U. S. 388, 394-395 (1983); Larkin v. Grendel’s Den, Inc., 459 U. S. 116, 123-124 (1982); Widmar v. Vincent, 454 U. S. 263, 271 (1981); Committee for Public Education & Religious Liberty v. Regan, 444 U. S. 646, 654, 657 (1980); Wolman v. Walter, 433 U. S. 229, 236 (1977) (plurality opinion); Meek v. Pittenger, 421 U. S. 349, 363 (1975); Committee for Public Education & Religious Liberty v. Nyquist, 413 U. S. 756, 773 (1973); Levitt v. Committee for Public Education & Religious Liberty, 413 U. S. 472, 479-480, n. 7 (1973); Tilton v. Richardson, 403 U. S. 672, 678-679 (1971) (plurality opinion); Lemon v. Kurtzman, supra, at 613. In fact, only once before deciding Lemon, and twice since, have we invalidated a law for lack of a secular purpose. See Wallace v. Jaffree, 472 U. S. 38 (1985); Stone v. Graham, 449 U. S. 39 (1980) (per curiam); Epperson v. Arkansas, 393 U. S. 97 (1968).
Nevertheless, a few principles have emerged from our cases, principles which should, but to an unfortunately large extent do not, guide the Court’s application of Lemon today. It is clear, first of all, that regardless of what “legislative purpose” may mean in other contexts, for the purpose of the Lemon test it means the “actual” motives of those responsible for the challenged action. The Court recognizes this, see ante, at 585, as it has in the past, see, e. g., Witters v. Washington Dept. of Services for Blind, supra, at 486; Wallace v. *614Jaffree, supra, at 56. Thus, if those legislators who supported the Balanced Treatment Act in fact acted with a “sincere” secular purpose, ante, at 587, the Act survives the first component of the Lemon test, regardless of whether that purpose is likely to be achieved by the provisions they enacted.
Our cases have also confirmed that when the Lemon Court referred to “a secular . . . purpose,” 403 U. S., at 612, it meant “a secular purpose.” The author of Lemon, writing for the Court, has said that invalidation under the purpose prong is appropriate when “there [is] no question that the statute or activity was motivated wholly by religious considerations.” Lynch v. Donnelly, 465 U. S. 668, 680 (1984) (Burger, C. J.) (emphasis added); see also id., at 681, n. 6; Wallace v. Jaffree, supra, at 56 (“[T]he First Amendment requires that a statute must be invalidated if it is entirely motivated by a purpose to advance religion”) (emphasis added; footnote omitted). In all three cases in which we struck down laws under the Establishment Clause for lack of a secular purpose, we found that the legislature’s sole motive was to promote religion. See Wallace v. Jaffree, supra, at 56, 57, 60; Stone v. Graham, supra, at 41, 43, n. 5; Epperson v. Arkansas, supra, at 103, 107-108; see also Lynch v. Donnelly, supra, at 680 (describing Stone and Epperson as cases in which we invalidated laws “motivated wholly by religious considerations”). Thus, the majority’s invalidation of the Balanced Treatment Act is defensible only if the record indicates that the Louisiana Legislature had no secular purpose.
It is important to stress that the purpose forbidden by Lemon is the purpose to “advance religion.” 403 U. S., at 613; accord, ante, at 585 (“promote” religion); Witters v. Washington Dept. of Services for Blind, supra, at 486 (“endorse religion”); Wallace v. Jaffree, 472 U. S., at 56 (“advance religion”); ibid, (“endorse . . . religion”); Committee for Public Education & Religious Liberty v. Nyquist, supra, at 788 (“ 'advancing’ . . . religion”); Levitt v. Committee for *615Public Education & Religious Liberty, supra, at 481 (“advancing religion”); Walz v. Tax Comm’n of New York City, 397 U. S. 664, 674 (1970) (“establishing, sponsoring, or supporting religion”); Board of Education v. Allen, 392 U. S. 236, 243 (1968) (“‘advancement or inhibition of religion’”) (quoting Abington School Dist. v. Schempp, 374 U. S. 203, 222 (1963)). Our cases in no way imply that the Establishment Clause forbids legislators merely to act upon their religious convictions. We surely would not strike down a law providing money to feed the hungry or shelter the homeless if it could be demonstrated that, but for the religious beliefs of the legislators, the funds would not have been approved. Also, political activism by the religiously motivated is part of our heritage. Notwithstanding the majority’s implication to the contrary, ante, at 589-591, we do not presume that the sole purpose of a law is to advance religion merely because it was supported strongly by organized religions or by adherents of particular faiths. See Walz v. Tax Comm’n of New York City, supra, at 670; cf. Harris v. McRae, 448 U. S. 297, 319-320 (1980). To do so would deprive religious men and women of their right to participate in the political process. Today’s religious activism may give us the Balanced Treatment Act, but yesterday’s resulted in the abolition of slavery, and tomorrow’s may bring relief for famine victims.
Similarly, we will not presume that a law’s purpose is to advance religion merely because it “‘happens to coincide or harmonize with the tenets of some or all religions,’” Harris v. McRae, supra, at 319 (quoting McGowan v. Maryland, 366 U. S. 420, 442 (1961)), or because it benefits religion, even substantially. We have, for example, turned back Establishment Clause challenges to restrictions on abortion funding, Harris v. McRae, supra, and to Sunday closing laws, McGowan v. Maryland, supra, despite the fact that both “agre[e] with the dictates of [some] Judaeo-Christian religions,” id., at 442. “In many instances, the Congress or state legislatures conclude that the general welfare of soci*616ety, wholly apart from any religious considerations, demands such regulation.” Ibid. On many past occasions we have had no difficulty finding a secular purpose for governmental action far more likely to advance religion than the Balanced Treatment Act. See, e. g., Mueller v. Allen, 463 U. S., at 394-395 (tax deduction for expenses of religious education); Wolman v. Walter, 433 U. S., at 236 (plurality opinion) (aid to religious schools); Meek v. Pittenger, 421 U. S., at 363 (same); Committee for Public Education & Religious Liberty v. Nyquist, 413 U. S., at 773 (same); Lemon v. Kurtzman, 403 U. S., at 613 (same); Walz v. Tax Comm’n of New York City, supra, at 672 (tax exemption for church property); Board of Education v. Allen, supra, at 243 (textbook loans to students in religious schools). Thus, the fact that creation science coincides with the beliefs of certain religions, a fact upon which the majority relies heavily, does not itself justify invalidation of the Act.
Finally, our cases indicate that even certain kinds of governmental actions undertaken with the specific intention of improving the position of religion do not “advance religion” as that term is used in Lemon. 403 U. S., at 613. Rather, we have said that in at least two circumstances government must act to advance religion, and that in a third it may do so.
First, since we have consistently described the Establishment Clause as forbidding not only state action motivated by the desire to advance religion, but also that intended to “disapprove,” “inhibit,” or evince “hostility” toward religion, see, e. g., ante, at 585 (“‘disapprove’”) (quoting Lynch v. Donnelly, supra, at 690 (O’Connor, J., concurring)); Lynch v. Donnelly, supra, at 673 (“hostility”); Committee for Public Education & Religious Liberty v. Nyquist, supra, at 788 (“‘inhibi[t]”’); and since we have said that governmental “neutrality” toward religion is the preeminent goal of the First Amendment, see, e. g., Grand Rapids School District v. Ball, 473 U. S., at 382; Roemer v. Maryland Public Works Bd., 426 U. S. 736, 747 (1976) (plurality opinion); *617Committee for Public Education & Religious Liberty v. Nyquist, supra, at 792-793; a State which discovers that its employees are inhibiting religion must take steps to prevent them from doing so, even though its purpose would clearly be to advance religion. Cf. Walz v. Tax Comm’n of New York City, supra, at 673. Thus, if the Louisiana Legislature sincerely believed that the State’s science teachers were being hostile to religion, our cases indicate that it could act to eliminate that hostility without running afoul of Lemon’s purpose test.
Second, we have held that intentional governmental advancement of religion is sometimes required by the Free Exercise Clause. For example, in Hobbie v. Unemployment Appeals Comm’n of Fla., 480 U. S. 136 (1987); Thomas v. Review Bd., Indiana Employment Security Div., 450 U. S. 707 (1981); Wisconsin v. Yoder, 406 U. S. 205 (1972); and Sherbert v. Verner, 374 U. S. 398 (1963), we held that in some circumstances States must accommodate the beliefs of religious citizens by exempting them from generally applicable regulations. We have not yet come close to reconciling Lemon and our Free Exercise cases, and typically we do not really try. See, e. g., Hobbie v. Unemployment Appeals Comm’n of Fla., supra, at 144-145; Thomas v. Review Bd., Indiana Employment Security Div., supra, at 719-720. It is clear, however, that members of the Louisiana Legislature were not impermissibly motivated for purposes of the Lemon test if they believed that approval of the Balanced Treatment Act was required by the Free Exercise Clause.
We have also held that in some circumstances government may act to accommodate religion, even if that action is not required by the First Amendment. See Hobbie v. Unemployment Appeals Comm’n of Fla., supra, at 144-145. It is well established that “[t]he limits of permissible state accommodation to religion are by no means co-extensive with the noninterference mandated by the Free Exercise Clause.” Walz v. Tax Comm’n of New York City, supra, at 673; *618see also Gillette v. United States, 401 U. S. 437, 453 (1971). We have implied that voluntary governmental accommodation of religion is not only permissible, but desirable. See, e. g., ibid. Thus, few would contend that Title VII of the Civil Rights Act of 1964, which both forbids religious discrimination by private-sector employers, 78 Stat. 255, 42 U. S. C. § 2000e-2(a)(l), and requires them reasonably to accommodate the religious practices of their employees, § 2000e(j), violates the Establishment Clause, even though its “purpose” is, of course, to advance religion, and even though it is almost certainly not required by the Free Exercise Clause. While we have warned that at some point, accommodation may devolve into “an unlawful fostering of religion,” Hobbie v. Unemployment Appeals Comm’n of Fla., supra, at 145, we have not suggested precisely (or even roughly) where that point might be. It is possible, then, that even if the sole motive of those voting for the Balanced Treatment Act was to advance religion, and its passage was not actually required, or even believed to be required, by either the Free Exercise or Establishment Clauses, the Act would nonetheless survive scrutiny under Lemon’s purpose test.
One final observation about the application of that test: Although the Court’s opinion gives no hint of it, in the past we have repeatedly affirmed “our reluctance to attribute unconstitutional motives to the States.” Mueller v. Allen, supra, at 394; see also Lynch v. Donnelly, 465 U. S., at 699 (Brennan, J., dissenting). We “presume that legislatures act in a constitutional manner.” Illinois v. Krull, 480 U. S. 340, 351 (1987); see also Clements v. Fashing, 457 U. S. 957, 963 (1982) (plurality opinion); Rostker v. Goldberg, 453 U. S. 57, 64 (1981); McDonald v. Board of Election Comm’rs of Chicago, 394 U. S. 802, 809 (1969). Whenever we are called upon to judge the constitutionality of an act of a state legislature, “we must have ‘due regard to the fact that this Court is not exercising a primary judgment but is sitting in judgment *619upon those who also have taken the oath to observe the Constitution and who have the responsibility for carrying on government.’” Rostker v. Goldberg, supra, at 64 (quoting Joint Anti-Fascist Refugee Committee v. McGrath, 341 U. S. 123, 164 (1951) (Frankfurter, J., concurring)). This is particularly true, we have said, where the legislature has specifically considered the question of a law’s constitutionality. Ibid.
With the foregoing in mind, I now turn to the purposes underlying adoption of the Balanced Treatment Act.
r*H
A
We have relatively little information upon which to judge the motives of those who supported the Act. About the only direct evidence is the statute itself and transcripts of the seven committee hearings at which it was considered. Unfortunately, several of those hearings were sparsely attended, and the legislators who were present revealed little about their motives. We have no committee reports, no floor debates, no remarks inserted into the legislative history, no statement from the Governor, and no postenactment statements or testimony from the bill’s sponsor or any other legislators. Cf. Wallace v. Jaffree, 472 U. S., at 43, 56-57. Nevertheless, there is ample evidence that the majority is wrong in holding that the Balanced Treatment Act is without secular purpose.
At the outset, it is important to note that the Balanced Treatment Act did not fly through the Louisiana Legislature on wings of fundamentalist religious fervor — which would be unlikely, in any event, since only a small minority of the State’s citizens belong to fundamentalist religious denominations. See B. Quinn, H. Anderson, M. Bradley, P. Goetting, & P. Shriver, Churches and Church Membership in the United States 16 (1982). The Act had its genesis (so to speak) in legislation introduced by Senator Bill Keith in June *6201980. After two hearings before the Senate Committee on Education, Senator Keith asked that his bill be referred to a study commission composed of members of both Houses of the Louisiana Legislature. He expressed hope that the joint committee would give the bill careful consideration and determine whether his arguments were “legitimate.” 1 App. E-29 — E-30. The committee met twice during the interim, heard testimony (both for and against the bill) from several witnesses, and received staff reports. Senator Keith introduced his bill again when the legislature reconvened. The Senate Committee on Education held two more hearings and approved the bill after substantially amending it (in part over Senator Keith’s objection). After approval by the full Senate, the bill was referred to the House Committee on Education. That committee conducted a lengthy hearing, adopted further amendments, and sent the bill on to the full House, where it received favorable consideration. The Senate concurred in the House amendments and on July 20, 1981, the Governor signed the bill into law.
Senator Keith’s statements before the various committees that considered the bill hardly reflect the confidence of a man preaching to the converted. He asked his colleagues to “keep an open mind” and not to be “biased” by misleading characterizations of creation science. Id., at E-33. He also urged them to “look at this subject on its merits and not on some preconceived idea.” Id., at E-34; see also 2 id., at E-491. Senator Keith’s reception was not especially warm. Over his strenuous objection, the Senate Committee on Education voted 5-1 to amend his bill to deprive it of any force; as amended, the bill merely gave teachers permission to balance the teaching of creation science or evolution with the other. 1 id., at E-442 — E-461. The House Committee restored the “mandatory” language to the bill by a vote of only 6-5, 2 id., at E-626 — E-627, and both the full House (by vote of 52-35), id., at E-700 — E-706, and full Senate (23-15), id., at E-735 — E-738, had to repel further efforts to gut the bill.
*621The legislators understood that Senator Keith’s bill involved a “unique” subject, 1 id., at E-106 (Rep. M. Thompson), and they were repeatedly made aware of its potential constitutional problems, see, e. g., id., at E-26 — E-28 (McGehee); id., at E-38 — E-39 (Sen. Keith); id., at E-241— E-242 (Rossman); id., at E-257 (Probst); id., at E-261 (Beck); id., at E-282 (Sen. Keith). Although the Establishment Clause, including its secular purpose requirement, was of substantial concern to the legislators, they eventually voted overwhelmingly in favor of the Balanced Treatment Act: The House approved it 71-19 (with 15 members absent), 2 id., at E-716 — E-722; the Senate 26-12 (with all members present), id., at E-741 — E-744. The legislators specifically designated the protection of “academic freedom” as the purpose of the Act. La. Rev. Stat. Ann. § 17:286.2 (West 1982). We cannot accurately assess whether this purpose is a “sham,” ante, at 587, until we first examine the evidence presented to the legislature far more carefully than the Court has done.
Before summarizing the testimony of Senator Keith and his supporters, I wish to make clear that I by no means intend to endorse its accuracy. But my views (and the views of this Court) about creation science and evolution are (or should be) beside the point. Our task is not to judge the debate about teaching the origins of life, but to ascertain what the members of the Louisiana Legislature believed. The vast majority of them voted to approve a bill which explicitly stated a secular purpose; what is crucial is not their wisdom, in believing that purpose would be achieved by the bill, but their sincerity in believing it would be.
Most of the testimony in support of Senator Keith’s bill came from the Senator himself and from scientists and educators he presented, many of whom enjoyed academic credentials that may have been regarded as quite impressive by members of the Louisiana Legislature. To a substantial extent, their testimony was devoted to lengthy, and, to the layman, seemingly expert scientific expositions on the origin *622of life. See, e. g., 1 App. E-ll — E-18 (Sunderland); id., at E-50 — E-60 (Boudreaux); id., at E-86 — E-89 (Ward); id., at E-130 — E-153 (Boudreaux paper); id., at E-321 — E-326 (Boudreaux); id., at E-423 — E-428 (Sen. Keith). These scientific lectures touched upon, inter alia, biology, paleontology, genetics, astronomy, astrophysics, probability analysis, and biochemistry. The witnesses repeatedly assured committee members that “hundreds and hundreds” of highly respected, internationally renowned scientists believed in creation science and would support their testimony. See, e. g., id., at E-5 (Sunderland); id., at E-76 (Sen. Keith); id., at E-100 — E-101 (Reiboldt); id., at E-327 — E-328 (Boudreaux); 2 id., at E-503 — E-504 (Boudreaux).
Senator Keith and his witnesses testified essentially as set forth in the following numbered paragraphs:
(1) There are two and only two scientific explanations for the beginning of life3 — evolution and creation science. 1 id., at E-6 (Sunderland); id., at E-34 (Sen. Keith); id., at E-280 (Sen. Keith); id., at E-417 — E-418 (Sen. Keith). Both are bona fide “sciences.” Id., at E-6 — E-7 (Sunderland); id., at E-12 (Sunderland); id., at E-416 (Sen. Keith); id., at E-427 (Sen. Keith); 2 id., at E-491 — E-492 (Sen. Keith); id., at E-497 — E-498 (Sen. Keith). Both posit a theory of the origin of life and subject that theory to empirical testing. Evolution posits that life arose out of inanimate chemical compounds and has gradually evolved over millions of years. Creation science posits that all life forms now on earth appeared suddenly and relatively recently and have changed little. Since there are only two possible explanations of the origin of life, any evidence that tends to disprove the theory of evolution necessarily tends to prove the theory of creation science, and vice versa. For example, the abrupt appearance in the fossil record of complex life, and the extreme rar*623ity of transitional life forms in that record, are evidence for creation science. 1 id., at E-7 (Sunderland); id., at E-12— E-18 (Sunderland); id., at E-45 — E-60 (Boudreaux); id., at E-67 (Harlow); id., at E-130 — E-153 (Boudreaux paper); id., at E-423 — E-428 (Sen. Keith).
(2) The body of scientific evidence supporting creation science is as strong as that supporting evolution. In fact, it may be stronger. Id., at E-214 (Young statement); id., at E-310 (Sen. Keith); id., at E-416 (Sen. Keith); 2 id., at E-492 (Sen. Keith). The evidence for evolution is far less compelling than we have been led to believe. Evolution is not a scientific “fact,” since it cannot actually be observed in a laboratory. Rather, evolution is merely a scientific theory or “guess.” 1 id., at E-20 — E-21 (Morris); id., at E-85 (Ward); id., at E-100 (Reiboldt); id., at E-328 — E-329 (Boudreaux); 2 id., at E-506 (Boudreaux). It is a very bad guess at that. The scientific problems with evolution are so serious that it could accurately be termed a “myth.” 1 id., at E-85 (Ward); id., at E-92 — E-93 (Kalivoda); id., at E-95 — E-97 (Sen. Keith); id., at E-154 (Boudreaux paper); id.y at E-329 (Boudreaux); id., at E-453 (Sen. Keith); 2 id., at E-505— E-506 (Boudreaux); id., at E-516 (Young).
(3) Creation science is educationally valuable. Students exposed to it better understand the current state of scientific evidence about the origin of life. 1 id., at E-19 (Sunder-land); id., at E-39 (Sen. Keith); id., at E-79 (Kalivoda); id., at E-308 (Sen. Keith); 2 id., at E-513 — E-514 (Morris). Those students even have a better understanding of evolution. 1 id., at E-19 (Sunderland). Creation science can and should be presented to children without any religious content. Id., at E-12 (Sunderland); id., at E-22 (Sanderford); id., at E-35 — E-36 (Sen. Keith); id., at E-101 (Reiboldt); id., at E-279 — E-280 (Sen. Keith); id., at E-282 (Sen. Keith).
(4) Although creation science is educationally valuable and strictly scientific, it is now being censored from or misrepresented in the public schools. Id., at E-19 (Sunderland); id., *624at E-21 (Morris); id., at E-34 (Sen. Keith); id., at E-37 (Sen. Keith); id., at E-42 (Sen. Keith); id., at E-92 (Kali-voda); id., at E-97 — E-98 (Reiboldt); id., at E-214 (Young statement); id., at E-218 (Young statement); id., at E-280 (Sen. Keith); id., at E-309 (Sen. Keith); 2 id., at E-513 (Morris). Evolution, in turn, is misrepresented as an absolute truth. 1 id., at E-63 (Harlow); id., at E-74 (Sen. Keith); id., at E-81 (Kalivoda); id., at E-214 (Young statement); 2 id., at E-507 (Harlow); id., at E-513 (Morris); id., at E-516 (Young). Teachers have been brainwashed by an entrenched scientific establishment composed almost exclusively of scientists to whom evolution is like a “religion.” These scientists discriminate against creation scientists so as to prevent evolution’s weaknesses from being exposed. 1 id., at E-61 (Boudreaux); id., at E-63 — E-64 (Harlow); id., at E-78 — E-79 (Kalivoda); id., at E-80 (Kalivoda); id., at E-95 — E-97 (Sen. Keith); id., at E-129 (Boudreaux paper); id., at E-218 (Young statement); id., at E-357 (Sen. Keith); id., at E-430 (Boudreaux).
(5) The censorship of creation science has at least two harmful effects. First, it deprives students of knowledge of one of the two scientific explanations for the origin of life and leads them to believe that evolution is proven fact; thus, their education suffers and they are wrongly taught that science has proved their religious beliefs false. Second, it violates the Establishment Clause. The United States Supreme Court has held that secular humanism is a religion. Id., at E-36 (Sen. Keith) (referring to Torcaso v. Watkins, 367 U. S. 488, 495, n. 11 (1961)); 1 App. E-418 (Sen. Keith); 2 id., at E-499 (Sen. Keith). Belief in evolution is a central tenet of that religion. 1 id., at E-282 (Sen. Keith); id., at E-312— E-313 (Sen. Keith); id., at E-317 (Sen. Keith); id., at E-418 (Sen. Keith); 2 id., at E-499 (Sen. Keith). Thus, by censoring creation science and instructing students that evolution is fact, public school teachers are now advancing religion in violation of the Establishment Clause. 1 id., at E-2 — E-4 *625(Sen. Keith); id., at E-36 — E-37, E-39 (Sen. Keith); id., at E-154 — E-155 (Boudreaux paper); id., at E-281 — E-282 (Sen. Keith); id., at E-313 (Sen. Keith); id., at E-315— E-316 (Sen. Keith); id., at E-317 (Sen. Keith); 2 id., at E-499 — E-500 (Sen. Keith).
Senator Keith repeatedly and vehemently denied that his purpose was to advance a particular religious doctrine. At the outset of the first hearing on the legislation, he testified: “We are not going to say today that you should have some kind of religious instructions in our schools. . . . We are not talking about religion today. ... I am not proposing that we take the Bible in each science class and read the first chapter of Genesis.” 1 id., at E-35. At a later hearing, Senator Keith stressed: “[T]o . . . teach religion and disguise it as creationism ... is not my intent. My intent is to see to it that our textbooks are not censored.” Id., at E-280. He made many similar statements throughout the hearings. See, e. g., id., at E-41; id., at E-282; id., at E-310; id., at E-417; see also id., at E-44 (Boudreaux); id., at E-80 (Kalivoda).
We have no way of knowing, of course, how many legislators believed the testimony of Senator Keith and his witnesses. But in the absence of evidence to the contrary,4 we *626have to assume that many of them did. Given that assumption, the Court today plainly errs in holding that the Louisiana Legislature passed the Balanced Treatment Act for exclusively religious purposes.
B
Even with nothing more than this legislative history to go on, I think it would be extraordinary to invalidate the Balanced Treatment Act for lack of a valid secular purpose. Striking down a law approved by the democratically elected representatives of the people is no minor matter. “The cardinal principle of statutory construction is to save and not to destroy. We have repeatedly held that as between two possible interpretations of a statute, by one of which it would be unconstitutional and by the other valid, our plain duty is to adopt that which will save the act.” NLRB v. Jones & Laughlin Steel Corp., 301 U. S. 1, 30 (1937). So, too, it seems to me, with discerning statutory purpose. Even if the legislative history were silent or ambiguous about the existence of a secular purpose — and here it is not — the statute should survive Lemon’s purpose test. But even more validation than mere legislative history is present here. The Louisiana Legislature explicitly set forth its secular purpose *627(“protecting academic freedom”) in the very text of the Act. La. Rev. Stat. § 17:286.2 (West 1982). We have in the past repeatedly relied upon or deferred to such expressions, see, e. g., Committee for Public Education & Religious Liberty v. Regan, 444 U. S., at 654; Meek v. Pittenger, 421 U. S., at 363, 367-368; Committee for Public Education & Religious Liberty v. Nyquist, 413 U. S., at 773; Levitt v. Committee for Public Education & Religious Liberty, 413 U. S., at 479-480, n. 7; Tilton v. Richardson, 403 U. S., at 678-679 (plurality opinion); Lemon v. Kurtzman, 403 U. S., at 613; Board of Education v. Allen, 392 U. S., at 243.
The Court seeks to evade the force of this expression of purpose by stubbornly misinterpreting it, and then finding that the provisions of the Act do not advance that misinterpreted purpose, thereby showing it to be a sham. The Court first surmises that “academic freedom” means “enhancing the freedom of teachers to teach what they will,” ante, at 586— even though “academic freedom” in that sense has little scope in the structured elementary and secondary curriculums with which the Act is concerned. Alternatively, the Court suggests that it might mean “maximizing] the comprehensiveness and effectiveness of science instruction,” ante, at 588— though that is an exceedingly strange interpretation of the words, and one that is refuted on the very face of the statute. See § 17:286.5. Had the Court devoted to this central question of the meaning of the legislatively expressed purpose a small fraction of the research into legislative history that produced its quotations of religiously motivated statements by individual legislators, it would have discerned quite readily what “academic freedom” meant: students’ freedom from indoctrination. The legislature wanted to ensure that students would be free to decide for themselves how life began, based upon a fair and balanced presentation of the scientific evidence — that is, to protect “the right of each [student] voluntarily to determine what to believe (and what not to believe) free of any coercive pressures from the State.” Grand *628Rapids School District v. Ball, 473 U. S., at 385. The legislature did not care whether the topic of origins was taught; it simply wished to ensure that when the topic was taught, students would receive “‘all of the evidence.”’ Ante, at 586 (quoting Tr. of Oral Arg. 60).
As originally introduced, the “purpose” section of the Balanced Treatment Act read: “This Chapter is enacted for the purposes of protecting academic freedom . . . of students . . . and assisting students in their search for truth.” 1 App. E-292 (emphasis added). Among the proposed findings of fact contained in the original version of the bill was the following: “Public school instruction in only evolution-science . . . violates the principle of academic freedom because it denies students a choice between scientific models and instead indoctrinates them in evolution science alone.” Id., at E-295 (emphasis added).5 Senator Keith unquestionably understood “academic freedom” to mean “freedom from indoctrination.” See id., at E-36 (purpose of bill is “to protect academic freedom by providing student choice”); id., at E-283 (purpose of bill is to protect “academic freedom” by giving students a “choice” rather than subjecting them to “indoctrination on origins”).
If one adopts the obviously intended meaning of the statutory term “academic freedom,” there is no basis whatever for concluding that the purpose they express is a “sham.” Ante, *629at 587. To the contrary, the Act pursues that purpose plainly and consistently. It requires that, whenever the subject of origins is covered, evolution be “taught as a theory, rather than as proven scientific fact” and that scientific evidence inconsistent with the theory of evolution (viz., “creation science”) be taught as well. La. Rev. Stat. Ann. § 17:286.4A (West 1982). Living up to its title of “Balanced Treatment for Creation-Science and Evolution-Science Act,” §17.286.1, it treats the teaching of creation the same way. It does not mandate instruction in creation science, § 17:286.5; forbids teachers to present creation science “as proven scientific fact,” § 17:286.4A; and bans the teaching of creation science unless the theory is (to use the Court’s terminology) “discredited] ‘. . . at every turn’” with the teaching of evolution. Ante, at 589 (quoting 765 F. 2d, at 1257). It surpasses understanding how the Court can see in this a purpose “to restructure the science curriculum to conform with a particular religious viewpoint,” ante, at 593, “to provide a persuasive advantage to a particular religious doctrine,” ante, at 592, “to promote the theory of creation science which embodies a particular religious tenet,” ante, at 593, and “to endorse a particular religious doctrine,” ante, at 594.
The Act’s reference to “creation” is not convincing evidence of religious purpose. The Act defines creation science as “scientific evidenc[ej,” § 17:286.3(2) (emphasis added), and Senator Keith and his witnesses repeatedly stressed that the subject can and should be presented without religious content. See supra, at 623. We have no basis on the record to conclude that creation science need be anything other than a collection of scientific data supporting the theory that life abruptly appeared on earth. See n. 4, swpra. Creation science, its proponents insist, no more must explain whence life came than evolution must explain whence came the inanimate materials from which it says life evolved. But even if that were not so, to posit a past creator is not to posit the eternal and personal God who is the object of religious veneration. *630Indeed, it is not even to posit the “unmoved mover” hypothesized by Aristotle and other notably nonfundamentalist philosophers. Senator Keith suggested this when he referred to “a creator however you define a creator.” 1 App. E-280 (emphasis added).
The Court cites three provisions of the Act which, it argues, demonstrate a “discriminatory preference for the teaching of creation science” and no interest in “academic freedom.” Ante, at 588. First, the Act prohibits discrimination only against creation scientists and those who teach creation science. §17:286.40. Second, the Act requires local school boards to develop and provide to science teachers “a curriculum guide on presentation of creation-science.” § 17:286.7A. Finally, the Act requires the Governor to designate seven creation scientists who shall, upon request, assist local school boards in developing the curriculum guides. § 17:286.7B. But none of these provisions casts doubt upon the sincerity of the legislators’ articulated purpose of “academic freedom” — unless, of course, one gives that term the obviously erroneous meanings preferred by the Court. The Louisiana legislators had been told repeatedly that creation scientists were scorned by most educators and scientists, who themselves had an almost religious faith in evolution. It is hardly surprising, then, that in seeking to achieve a balanced, “non-indoctrinating” curriculum, the legislators protected from discrimination only those teachers whom they thought were suffering from discrimination. (Also, the legislators were undoubtedly aware of Epperson v. Arkansas, 393 U. S. 97 (1968), and thus could quite reasonably have concluded that discrimination against evolutionists was already prohibited.) The two provisions respecting the development of curriculum guides are also consistent with “academic freedom” as the Louisiana Legislature understood the term. Witnesses had informed the legislators that, because of the hostility of most scientists and educators to creation science, the topic had been censored from or badly misrepresented in elementary *631and secondary school texts. In light of the unavailability of works on creation science suitable for classroom use (a fact appellees concede, see Brief for Appellees 27, 40) and the existence of ample materials on evolution, it was entirely reasonable for the legislature to conclude that science teachers attempting to implement the Act would need a curriculum guide on creation science, but not on evolution, and that those charged with developing the guide would need an easily accessible group of creation scientists. Thus, the provisions of the Act of so much concern to the Court support the conclusion that the legislature acted to advance “academic freedom.”
The legislative history gives ample evidence of the sincerity of the Balanced Treatment Act’s articulated purpose. Witness after witness urged the legislators to support the Act so that students would not be “indoctrinated” but would instead be free to decide for themselves, based upon a fair presentation of the scientific evidence, about the origin of life. See, e. g., 1 App. E-18 (Sunderland) (“all that we are advocating” is presenting “scientific data” to students and “letting [them] make up their own mind[s]”); id., at E-19— E-20 (Sunderland) (Students are now being “indoctrinated” in evolution through the use of “censored school books. . . . All that we are asking for is [the] open unbiased education in the classroom . . . your students deserve”); id., at E-21 (Morris) (“A. student cannot [make an intelligent decision about the origin of life] unless he is well informed about both [evolution and creation science]”); id., at E-22 (Sanderford) (“We are asking very simply [that] . . . creationism [be presented] alongside . . . evolution and let people make their own mind[s] up”); id., at E-23 (Young) (the bill would require teachers to live up to their “obligation to present all theories” and thereby enable “students to make judgments themselves”); id., at E-44 (Boudreaux) (“Our intention is truth and as a scientist, I am interested in truth”); id., at E-60— E-61 (Boudreaux) (“[W]e [teachers] are guilty of a lot of *632brainwashing. ... We have a duty to . . . [present the] truth” to students “at all levels from gradeschool on through the college level”); id., at E-79 (Kalivoda) (“This [hearing] is being held I think to determine whether children will benefit from freedom of information or if they will be handicapped educationally by having little or no information about creation”); id., at E-80 (Kalivoda) (“I am not interested in teaching religion in schools. ... I am interested in the truth and [students] having the opportunity to hear more than one side”); id., at E-98 (Reiboldt) (“The students have a right to know there is an alternate creationist point of view. They have a right to know the scientific evidences which suppor[t] that alternative”); id., at E-218 (Young statement) (passage of the bill will ensure that “communication of scientific ideas and discoveries may be unhindered”); 2 id., at E-514 (Morris) (“[A]re we going to allow [students] to look at evolution, to look at creationism, and to let one or the other stand or fall on its own merits, or will we by failing to pass this bill. . . deny students an opportunity to hear another viewpoint?”); id., at E-516 — E-517 (Young) (“We want to give the children here in this state an equal opportunity to see both sides of the theories”). Senator Keith expressed similar views. See, e. g., 1 id., at E-36; id., at E-41; id., at E-280; id., at E-283.
Legislators other than Senator Keith made only a few statements providing insight into their motives, but those statements cast no doubt upon the sincerity of the Act’s articulated purpose. The legislators were concerned primarily about the manner in which the subject of origins was presented in Louisiana schools — specifically, about whether scientifically valuable information was being censored and students misled about evolution. Representatives Cain, Jenkins, and F. Thompson seemed impressed by the scientific evidence presented in support of creation science. See 2 id., at E-530 (Rep. F. Thompson); id., at E-533 (Rep. Cain); id., at E-613 (Rep. Jenkins). At the first study commission hearing, Senator Picard and Representative M. Thompson ques*633tioned Senator Keith about Louisiana teachers’ treatment of evolution and creation science. See 1 id., at E-71 — E-74. At the close of the hearing, Representative M. Thompson told the audience:
“We as members of the committee will also receive from the staff information of what is currently being taught in the Louisiana public schools. We really want to see [it]. I. . . have no idea in what manner [biology] is presented and in what manner the creationist theories [are] excluded in the public school[s]. We want to look at what the status of the situation is.” Id., at E-104.
Legislators made other comments suggesting a concern about censorship and misrepresentation of scientific information. See, e. g., id., at E-386 (Sen. McLeod); 2 id., at E-527 (Rep. Jenkins); id., at E-528 (Rep. M. Thompson); id., at E-534 (Rep. Fair).
It is undoubtedly true that what prompted the legislature to direct its attention to the misrepresentation of evolution in the schools (rather than the inaccurate presentation of other topics) was its awareness of the tension between evolution and the religious beliefs of many children. But even appel-lees concede that a valid secular purpose is not rendered impermissible simply because its pursuit is prompted by concern for religious sensitivities. Tr. of Oral Arg. 43, 56. If a history teacher falsely told her students that the bones of Jesus Christ had been discovered, or a physics teacher that the Shroud of Turin had been conclusively established to be inexplicable on the basis of natural causes, I cannot believe (despite the majority’s implication to the contrary, see ante, at 592-593) that legislators or school board members would be constitutionally prohibited from taking corrective action, simply because that action was prompted by concern for the religious beliefs of the misinstructed students.
In sum, even if one concedes, for the sake of argument, that a majority of the Louisiana Legislature voted for the Balanced Treatment Act partly in order to foster (rather *634than merely eliminate discrimination against) Christian fundamentalist beliefs, our cases establish that that alone would not suffice to invalidate the Act, so long as there was a genuine secular purpose as well. We have, moreover, no adequate basis for disbelieving the secular purpose set forth in the Act itself, or for concluding that it is a sham enacted to conceal the legislators’ violation of their oaths of office. I am astonished by the Court’s unprecedented readiness to reach such a conclusion, which I can only attribute to an intellectual predisposition created by the facts and the legend of Scopes v. State, 154 Tenn. 105, 289 S. W. 363 (1927)-an instinctive reaction that any governmentally imposed requirements bearing upon the teaching of evolution must be a manifestation of Christian fundamentalist repression. In this case, however, it seems to me the Court’s position is the repressive one. The people of Louisiana, including those who are Christian fundamentalists, are quite entitled, as a secular matter, to have whatever scientific evidence there may be against evolution presented in their schools, just as Mr. Scopes was entitled to present whatever scientific evidence there was for it. Perhaps what the Louisiana Legislature has done is unconstitutional because there is no such evidence, and the scheme they have established will amount to no more than a presentation of the Book of Genesis. But we cannot say that on the evidence before us in this summary judgment context, which includes ample uncontradicted testimony that “creation science” is a body of scientific knowledge rather than revealed belief. Infinitely less can we say (or should we say) that the scientific evidence for evolution is so conclusive that no one could be gullible enough to believe that there is any real scientific evidence to the contrary, so that the legislation’s stated purpose must be a lie. Yet that illiberal judgment, that Scopes-in-reverse, is ultimately the basis on which the Court’s facile rejection of the Louisiana Legislature’s purpose must rest.
*635Since the existence of secular purpose is so entirely clear, and thus dispositive, I will not go on to discuss the fact that, even if the Louisiana Legislature’s purpose were exclusively to advance religion, some of the well-established exceptions to the impermissibility of that purpose might be applicable— the validating intent to eliminate a perceived discrimination against a particular religion, to facilitate its free exercise, or to accommodate it. See supra, at 617-618. I am not in any case enamored of those amorphous exceptions, since I think them no more than unpredictable correctives to what is (as the next Part of this opinion will discuss) a fundamentally unsound rule. It is surprising, however, that the Court does not address these exceptions, since the context of the legislature’s action gives some reason to believe they may be applicable.6
*636Because I believe that the Balanced Treatment Act had a secular purpose, which is all the first component of the Lemon test requires, I would reverse the judgment of the Court of Appeals and remand for further consideration.
I — < J — I I — i
I have to this point assumed the validity of the Lemon “purpose” test. In fact, however, I think the pessimistic evaluation that The Chief Justice made of the totality of Lemon is particularly applicable to the “purpose” prong: it is “a constitutional theory [that] has no basis in the history of the amendment it seeks to interpret, is difficult to apply and yields unprincipled results . . . Wallace v. Jaffree, 472 U. S., at 112 (Rehnquist, J., dissenting).
Our cases interpreting and applying the purpose test have made such a maze of the Establishment Clause that even the most conscientious governmental officials can only guess what motives will be held unconstitutional. We have said essentially the following: Government may not act with the purpose of advancing religion, except when forced to do so by the Free Exercise Clause (which is now and then); or when eliminating existing governmental hostility to religion (which exists sometimes); or even when merely accommodating gov-ernmentally uninhibited religious practices, except that at some point (it is unclear where) intentional accommodation results in the fostering of religion, which is of course unconstitutional. See supra, at 614-618.
But the difficulty of knowing what vitiating purpose one is looking for is as nothing compared with the difficulty of knowing how or where to find it. For while it is possible to discern the objective “purpose” of a statute (i. e., the public good at which its provisions appear to be directed), or even the formal motivation for a statute where that is explicitly set forth (as it was, to no avail, here), discerning the subjective motivation of those enacting the statute is, to be honest, almost always an impossible task. The number of possible *637motivations, to begin with, is not binary, or indeed even finite. In the present case, for example, a particular legislator need not have voted for the Act either because he wanted to foster religion or because he wanted to improve education. He may have thought the bill would provide jobs for his district, or may have wanted to make amends with a faction of his party he had alienated on another vote, or he may have been a close friend of the bill’s sponsor, or he may have been repaying a favor he owed the majority leader, or he may have hoped the Governor would appreciate his vote and make a fundraising appearance for him, or he may have been pressured to vote for a bill he disliked by a wealthy contributor or by a flood of constituent mail, or he may have been seeking favorable publicity, or he may have been reluctant to hurt the feelings of a loyal staff member who worked on the bill, or he may have been settling an old score with a legislator who opposed the bill, or he may have been mad at his wife who opposed the bill, or he may have been intoxicated and utterly ^motivated when the vote was called, or he may have accidentally voted “yes” instead of “no,” or, of course, he may have had (and very likely did have) a combination of some of the above and many other motivations. To look for the sole purpose of even a single legislator is probably to look for something that does not exist.
Putting that problem aside, however, where ought we to look for the individual legislator’s purpose? We cannot of course assume that every member present (if, as is unlikely, we know who or even how many they were) agreed with the motivation expressed in a particular legislator’s pre-enactment floor or committee statement. Quite obviously, “[w]hat motivates one legislator to make a speech about a statute is not necessarily what motivates scores of others to enact it.” United States v. O’Brien, 391 U. S. 367, 384 (1968). Can we assume, then, that they all agree with the motivation expressed in the staff-prepared committee reports they might have read — even though we are unwilling to *638assume that they agreed with the motivation expressed in the very statute that they voted for? Should we consider postenactment floor statements? Or postenactment testimony from legislators, obtained expressly for the lawsuit? Should we consider media reports on the realities of the legislative bargaining? All of these sources, of course, are eminently manipulate. Legislative histories can be contrived and sanitized, favorable media coverage orchestrated, and postenactment recollections conveniently distorted. Perhaps most valuable of all would be more objective indications —for example, evidence regarding the individual legislators’ religious affiliations. And if that, why not evidence regarding the fervor or tepidity of their beliefs?
Having achieved, through these simple means, an assessment of what individual legislators intended, we must still confront the question (yet to be addressed in any of our cases) how many of them must have the invalidating intent. If a state senate approves a bill by vote of 26 to 25, and only one of the 26 intended solely to advance religion, is the law unconstitutional? What if 13 of the 26 had that intent? What if 3 of the 26 had the impermissible intent, but 3 of the 25 voting against the bill were motivated by religious hostility or were simply attempting to “balance” the votes of their impermissi-bly motivated colleagues? Or is it possible that the intent of the bill’s sponsor is alone enough to invalidate it — on a theory, perhaps, that even though everyone else’s intent was pure, what they produced was the fruit of a forbidden tree?
Because there are no good answers to these questions, this Court has recognized from Chief Justice Marshall, see Fletcher v. Peck, 6 Cranch 87, 130 (1810), to Chief Justice Warren, United States v. O’Brien, supra, at 383-384, that determining the subjective intent of legislators is a perilous enterprise. See also Palmer v. Thompson, 403 U. S. 217, 224-225 (1971); Epperson v. Arkansas, 393 U. S., at 113 (Black, J., concurring). It is perilous, I might note, not just for the judges who will very likely reach the wrong result, *639but also for the legislators who find that they must assess the validity of proposed legislation — and risk the condemnation of having voted for an unconstitutional measure — not on the basis of what the legislation contains, nor even on the basis of what they themselves intend, but on the basis of what others have in mind.
Given the many hazards involved in assessing the subjective intent of governmental decisionmakers, the first prong of Lemon is defensible, I think, only if the text of the Establishment Clause demands it. That is surely not the case. The Clause states that “Congress shall make no law respecting an establishment of religion.” One could argue, I suppose, that any time Congress acts with the intent of advancing religion, it has enacted a “law respecting an establishment of religion”; but far from being an unavoidable reading, it is quite an unnatural one. I doubt, for example, that the Clayton Act, 38 Stat. 730, as amended, 15 U. S. C. §12 et seq., could reasonably be described as a “law respecting an establishment of religion” if bizarre new historical evidence revealed that it lacked a secular purpose, even though it has no discernible nonsecular effect. It is, in short, far from an inevitable reading of the Establishment Clause that it forbids all governmental action intended to advance religion; and if not inevitable, any reading with such untoward consequences must be wrong.
In the past we have attempted to justify our embarrassing Establishment Clause jurisprudence7 on the ground that it *640“sacrifices clarity and predictability for flexibility.” Committee for Public Education & Religious Liberty v. Regan, 444 U. S., at 662. One commentator has aptly characterized this as “a euphemism . . . for . . . the absence of any principled rationale.” Choper, supra n. 7, at 681. I think it time that we sacrifice some “flexibility” for “clarity and predictability.” Abandoning Lemon’s purpose test — a test which exacerbates the tension between the Free Exercise and Establishment Clauses, has no basis in the language or history of the Amendment, and, as today’s decision shows, has wonderfully flexible consequences —would be a good place to start.
Article VI, cl. 3, of the Constitution provides that “the Members of the several State Legislatures . . . shall be bound by Oath or Affirmation, to support this Constitution.”
Thus the popular dictionary definitions cited by Justice Powell, ante, at 598-599 (concurring opinion), and appellees, see Brief for Appel-lees 25, 26; Tr. of Oral Arg. 32, 34, are utterly irrelevant, as are the views of the school superintendents cited by the majority, ante, at 595, n. 18. Three-quarters of those surveyed had “[n]o” or “[llimited” knowledge of “creation-science theory,” and not a single superintendent claimed “[e]x-tensive” knowledge of the subject. 2 App. E-798.
Although creation scientists and evolutionists also disagree about the origin of the physical universe, both proponents and opponents of Senator Keith’s bill focused on the question of the beginning of life.
Although appellees and amici dismiss the testimony of Senator Keith and his witnesses as pure fantasy, they did not bother to submit evidence of that to the District Court, making it difficult for us to agree with them. The State, by contrast, submitted the affidavits of two scientists, a philosopher, a theologian, and an educator, whose academic credentials are rather impressive. See App. to Juris. Statement A-17 — A-18 (Kenyon); id., at A-36 (Morrow); id., at A-39 — A-40 (Miethe); id., at A-46 — A-47 (Most); id., at A-49 (Clinkert). Like Senator Keith and his witnesses, the affiants swear that evolution and creation science are the only two scientific explanations for the origin of life, see id., at A-19 — A-20 (Kenyon); id., at A-38 (Morrow); id., at A-41 (Miethe); that creation science is strictly scientific, see id., at A-18 (Kenyon); id., at A-36 (Morrow); id., at A-40 — A-41 (Miethe); id., at A-49 (Clinkert); that creation science is simply a collection of scientific data that supports the hypothesis that life appeared on earth suddenly and has changed little, see id., at A-19 (Kenyon); id., at A-36 *626(Morrow); id., at A-41 (Miethe); that hundreds of respected scientists believe in creation science, see id., at A-20 (Kenyon); that evidence for creation science is as strong as evidence for evolution, see id., at A-21 (Kenyon); id., at A-34 — A-35 (Kenyon); id., at A-37 — A-38 (Morrow); that creation science is educationally valuable, see id., at A-19 (Kenyon); id., at A-36 (Morrow); id., at A-38 — A-39 (Morrow); id., at A-49 (Clinkert); that creation science can be presented without religious content, see id., at A-19 (Kenyon); id., at A-35 (Kenyon); id., at A-36 (Morrow); id., at A-40 (Miethe); id., at A-43 — A-44 (Miethe); id., at A-47 (Most); id., at A-49 (Clinkert); and that creation science is now censored from classrooms while evolution is misrepresented as proven fact, see id., at A-20 (Kenyon); id., at A-35 (Kenyon); id., at A-39 (Morrow); id., at A-50 (Clinkert). It is difficult to conclude on the basis of these affidavits — the only substantive evidence in the record — that the laymen serving in the Louisiana Legislature must have disbelieved Senator Keith or his witnesses.
The majority finds it “astonishing” that I would cite a portion of Senator Keith’s original bill that was later deleted as evidence of the legislature’s understanding of the phrase “academic freedom.” Ante, at 589, n. 8. What is astonishing is the majority’s implication that the deletion of that section deprives it of value as a clear indication of what the phrase meant — there and in the other, retained, sections of the bill. The Senate Committee on Education deleted most of the lengthy “purpose” section of the bill (with Senator Keith’s consent) because it resembled legislative “findings of fact,” which, committee members felt, should generally not be incorporated in legislation. The deletion had absolutely nothing to do with the manner in which the section described “academic freedom.” See 1 App. E-314 — E-320; id., at E-440-E-442.
As the majority recognizes, ante, at 592, Senator Keith sincerely believed that “secular humanism is a bona fide religion,” 1 App. E-36; see also id., at E-418; 2 id., at E-499, and that “evolution is the cornerstone of that religion,” 1 id., at E-418; see also id., at E-282; id., at E-312 — E-313; id., at E-317; 2 id., at E-499. The Senator even told his colleagues that this Court had “held” that secular humanism was a religion. See 1 id., at E-36, id., at E-418; 2 id., at E-499. (In Torcaso v. Watkins, 367 U. S. 488, 495, n. 11 (1961), we did indeed refer to “Secular Humanism” as a “religio[n].”) Senator Keith and his supporters raised the “religion” of secular humanism not, as the majority suggests, to explain the source of their “disdain for the theory of evolution,” ante, at 592, but to convince the legislature that the State of Louisiana was violating the Establishment Clause because its teachers were misrepresenting evolution as fact and depriving students of the information necessary to question that theory. 1 App. E-2 — E-4 (Sen. Keith); id., at E-36 — E-37, E-39 (Sen. Keith); id., at E-154 — E-155 (Boudreaux paper); id., at E-281 — E-282 (Sen. Keith); id., at E-317 (Sen. Keith); 2 id., at E-499 — E-500 (Sen. Keith). The Senator repeatedly urged his colleagues to pass his bill to remedy this Establishment Clause violation by ensuring state neutrality in religious matters, see, e. g., 1 id., at E-36; id., at E-39; id., at E-313, surely a permissible purpose under Lemon. Senator Keith’s argument may be questionable, but nothing in the statute or its legislative history gives us reason to doubt his sincerity or that of his supporters.
Professor Choper summarized our school aid cases thusly:
“[A] provision for therapeutic and diagnostic health services to parochial school pupils by public employees is invalid if provided in the parochial school, but not if offered at a neutral site, even if in a mobile unit adjacent to the parochial school. Reimbursement to parochial schools for the expense of administering teacher-prepared tests required by state law is invalid, but the state may reimburse parochial schools for the expense of administering state-prepared tests. The state may lend school textbooks to parochial school pupils because, the Court has explained, the books can be checked in advance for religious content and are ‘self-policing’; but the *640state may not lend other seemingly self-policing instructional items such as tape recorders and maps. The state may pay the cost of bus transportation to parochial schools, which the Court has ruled are ‘permeated’ with religion; but the state is forbidden to pay for field trip transportation visits ‘to governmental, industrial, cultural, and scientific centers designed to enrich the secular studies of students.’ ” Choper, The Religion Clauses of the First Amendment: Reconciling the Conflict, 41 U. Pitt. L. Rev. 673, 680-681 (1980) (footnotes omitted).
Since that was written, more decisions on the subject have been rendered, but they leave the theme of chaos securely unimpaired. See, e. g., Aguilar v. Felton, 473 U. S. 402 (1986); Grand Rapids School District v. Ball, 473 U. S. 373 (1985).