POSNER, Chief Judge.

A professor of electrical engineering at the University of Wisconsin appeals from the dismissal on summary judgement of his suit against a former graduate student and against university officials for copyright infringement and employment discrimination. This curious combination of claims comes about as follows. Seshadri, the plaintiff, and Kasraian, the graduate student, had a falling out over a paper that Seshadri had submitted under both their names for publication in the Journal of Applied Physics. Seshadri accused Kasraian of academic misconduct. Cleared of the charge after a hearing, Kasraian in turn complained to the university administration, which found that Seshadri had engaged in academic misconduct. The university suspended him for one year without pay and forbade him to advise graduate students indefinitely; we assume this means he can teach graduate students but cannot supervise their doctoral dissertations.

We can dispose very quickly of the Title VII claim, which is that Seshadri was sanctioned for his adherence to a "creed [that] requires scrupulous honesty . . . in the scholarly pursuit of scientific knowledge." He claims that this is a religious creed, and he appeals to the provision of Title VII that forbids discrimination on grounds of religion. He refuses, however, to identify the religion. He claims a right not to do so, pointing out that government has no right to require a person to state his religious beliefs or affiliations. True enough; but a person who seeks to obtain a privileged legal status by virtue of his religion cannot preclude inquiry designed to determine whether he has in fact a religion. This would be obvious if Seshadri were claiming a right to a tax exemption on the ground that he is a religious institution. The Internal Revenue Service would not be required to accept his say-so. Living Faith, Inc. v. Commissioner, 950 F.2d 365, 372-74 (7th Cir. 1991); United States v. Jeffries, 854 F.2d 254, 257-58 (7th Cir. 1988); Spiritual Outreach Society v. Commissioner, 927 F.2d 335, 338-39 (8th Cir. 1991). Nor the defendant accused in a Title VII case of religious discrimination.

It is true that the EEOC, following United States v. Seeger, 380 U.S. 163, 184-85 (1965), does not think that the plaintiff in a case of religious discrimination must be a member of an authorized church or subscribe to its full menu of orthodox beliefs. 29 C.F.R. § 1605.1. We agree. Nottelson v. Smith Steel Workers D.A.L.U. 19802, 643 F.2d 445, 454 and n. 12 (7th Cir. 1981); Redmond v. GAF Corp., 574 F.2d 897, 901 n. 12 (7th Cir. 1978). For otherwise Jesus Christ, a heterodox Jew, could not be regarded as having been a victim of religious persecution. Heretics are a principal target of religious persecution. But it cannot be the law that an employee can insulate himself from discipline by his employer by claiming that his religion prevents him from meeting his employer's legitimate expectations. That would be absurd and it is blocked in some cases by two obstacles and in this case by one.

In some cases it will be obvious without entering the field of religious controversy (a field that American courts are for obvious reasons loth to enter) that the plaintiff's belief, however deep-seated, is not religious. An example is the finding of the district court in Brown v. Pena, 441 F. Supp. 1382 (S.D. Fla. 1977), that even though the ancient Egyptians worshipped cats, a belief in the deeply spiritual effects of eating Kozy Kitten People/Cat Food is not a religious belief, and the plaintiff's consumption of the cat food was therefore not a religious observance. Seshadri's creed may be as religiously spurious as Brown's, but we cannot determine that on the basis of the present record.

The blocking principle in this case, a principle familiar from disability cases (most recently Matthews v. Commonwealth Edison Co., No. 96-3665, 1997 WL 710634 at *2-3 (7th Cir. Nov. 17, 1997)), is that an employer does not violate federal antidiscrimination law if he fires (or takes other adverse personnel action against) an employee not because the employee is a member of a protected group but because the employee cannot, even with a reasonable accommodation to this condition, or will not, meet the employer's legitimate expectations; and this is true even though he could meet them if he were not disabled, old, or male, or what have you — in this case, if he were not in the grip of a creed that (improbably) required him to lodge a groundless complaint of misconduct against his student. Although Title VII imposes on the employer a duty of accommodation to an employee's religious beliefs that the religion clauses of the First Amendment ex proprio vigore do not, 42 U.S.C. § 2000e(j); Trans World Airlines, Inc. v. Hardison, 432 U.S. 63, 75 (1977); Ryan v. U.S. Dept. of Justice, 950 F.2d 458, 461 (7th Cir. 1991), we do not understand this to place a greater burden on the employer than in the cognate case of disability discrimination.

Seshadri's copyright claim has a little more legal substance. Kasraian published the article in question, "Double-grating thin-film devices based on second-order Bragg interaction," 75 J. Appl. Phys. 7639 (1994), under his own name. Seshadri claims that he, Seshadri, wrote the entire article. Kasraian counters that it was a joint work and points out that the author of a joint work is a joint owner entitled to copyright it and license the copyright to a third party, subject only to a duty to account to his co-author for any profits. Erickson v. Trinity Theatre, Inc., 13 F.3d 1061, 1068 (7th Cir. 1994) (construing 17 U.S.C. § 201); United States ex rel. Berge v. Board of Trustees, 104 F.3d 1453, 1461 (4th Cir. 1997); see also Weinstein v. University of Illinois, 811 F.2d 1091, 1095 (7th Cir. 1987). Kasraian further argues that Seshadri abandoned any copyright claim he might have had.

We must decide whether, as the district judge thought, there is no genuine issue of material fact concerning Kasraian's entitlement to the status of joint author notwithstanding Seshadri's insistence in his affidavit that he indeed wrote the whole thing and never intended to abandon his copyright claim. Since the Journal of Applied Physics did not pay Kasraian a fee, and no income is anticipated from a possible reprinting of the article elsewhere, Seshadri is not seeking compensatory damages or the accounting to which he would be entitled had there been any income from the publication. The Copyright Act provides, however, for statutory damages, 17 U.S.C. § 504(c); Sony Corp. v. Universal City Studios, Inc., 464 U.S. 417, 433-34 (1984), which he is seeking along with an injunction against Kasraian's attempting to republish the article.

When a party makes damaging admissions in his deposition and then tries to retract them in an affidavit, courts give short shrift to the affidavit. Russell v. Acme-Evans Co., 51 F.3d 64, 67-68 (7th Cir. 1995); Darnell v. Target Stores, 16 F.3d 174, 176 (7th Cir. 1994); Hayes v. New York City Dept. of Corrections, 84 F.3d 614, 619 (2d Cir. 1996); Camfield Tires, Inc. v. Michelin Tire Corp., 719 F.2d 1361, 1364-65 (8th Cir. 1983). Seshadri was not deposed, but Kasraian argues that the affidavit attempts to cover a paper trail so unequivocal as to render the affidavit utterly incredible. Cf. United States v. Property Located at 15 Black Ledge Drive, 897 F.2d 97, 101-02 (2d Cir. 1990). The argument raises a general issue of some importance.

An affiant's prior statements cannot be equated to either a deposition, which is under oath, Fed.R.Civ.P. 30(c), or to a judicial admission, as in a pleading, which is binding. Soo Line R.R. v. St. Louis Southwestern Ry., 125 F.3d 481, 483 (7th Cir. 1997); Bright v. QSP, Inc., 20 F.3d 1300, 1305 (4th Cir. 1994). It is merely an evidentiary admission, or prior inconsistent statement, which does not bind the witness. Murrey v. United States, 73 F.3d 1448, 1455 (7th Cir. 1996); United States v. Blood, 806 F.2d 1218, 1221 n. 2 (4th Cir. 1986); United States v. McKeon, 738 F.2d 26, 30 (2d Cir. 1984). The opposing party can use it to try to discredit the witness's testimony, but the witness may be able to explain it away. And since a number of judicial opinions say that testimony cannot be rejected as incredible as a matter of law (that is, without need for a trial) unless it contains assertions that are contrary to the laws of nature, e.g., United States v. Zizzo, 120 F.3d 1338, 1359 (7th Cir. 1997); United States v. Dennis, 115 F.3d 524, 535 (7th Cir. 1997); United States v. Pena-Rodriguez, 110 F.3d 1120, 1123 (5th Cir. 1997), it may seem to follow that mere evidentiary admissions (as distinct from depositions) contradicted by an affidavit could never be a basis for a grant of summary judgement to the opposing party because the jury might believe the affiant when he repeated the affidavit in oral testimony.

We disagree. The "laws of nature" formulation is an exaggeration designed to underscore the limited scope of appellate review of determinations of credibility, rather than a precise and exhaustive statement of the test for determining that scope. The example we gave earlier of an affidavit that contradicts the witness's deposition shows that the test of physical impossibility is not exhaustive. The Supreme Court has noted that "factors other than demeanor and inflection go into the decision whether or not to believe a witness. Documents or objective evidence may contradict the witness' story; or the story itself may be so internally inconsistent or implausible on its face that a reasonable factfinder would not credit it. Where such factors are present, the court of appeals may well find clear error even in a finding purportedly based on a credibility determination." Anderson v. City of Bessemer City, 470 U.S. 564, 575 (1985). This was said in the context of appellate review, but the principle is equally applicable to summary judgment, and we may state it thus: testimony can and should be rejected without a trial if, in the circumstances, no reasonable person would believe it. See Winchester Packaging, Inc. v. Mobil Chemical Co., 14 F.3d 316, 319 (7th Cir. 1994); Bullard v. Sercon Corp., 846 F.2d 463, 466 (7th Cir. 1988); United States v. Property Located at 15 Black Ledge Drive, supra, 897 F.2d at 101-02. So let us look at the circumstances.

Kasraian was a Ph.D. student in Professor Seshadri's department. Seshadri was his supervisor. The two had collaborated on four previous papers published in leading journals, and in all four instances Kasraian's name had been listed first. In November of 1992 Seshadri submitted the double-grating article (a highly mathematical description of a process for strengthening laser beams) to the Journal of Applied Physics, listing as the authors "Masoud Kasraian and S.R. Seshadri." Included with the submission was an agreement signed by Kasraian and Seshadri assigning the copyright to the publisher of the journal upon the acceptance of the article for publication in it. The assignment became ineffective, however, when (as we are about to see) before acceptance Seshadri withdrew the manuscript from the journal. When Kasraian resubmitted the article under his own name he signed a new copyright assignment. Seshadri's contention is that in doing this Kasraian infringed Seshadri's copyright; that Kasraian, not being a joint author, had no copyright in the article that he could license to the journal.

Here is the background of the falling out between professor and student. Shortly after the initial submission of the article to the Journal of Applied Physics, Kasraian had decided, against Seshadri's advice, not to take a course offered by him. Seshadri became angry, either because he thought the course essential to Kasraian's development or because the defection caused the enrollment in the course to drop below the minimum number of students that would entitle Seshadri to teaching credit. At all events, in March of 1993 Seshadri wrote Kasraian that "not only did I write the major part of the manuscript, I revised it completely when you were vacationing in France." The next month Seshadri wrote the Journal of Applied Physics, stating, "Masoud Kasraian was my graduate student until recently and he is no longer my graduate student. We carried out the theoretical analysis for the above-mentioned manuscript and Masoud Kasraian carried out the numerical work, prepared the figures and typed the manuscript. The preparation of the manuscript was based on certain erroneous information given by Masoud Kasraian." The letter goes on to apologize for Seshadri's "not having been very careful to check all aspects of the manuscript before submission." He asks the journal to delete his name "from the authorship of the manuscript" and to permit him to withdraw the manuscript from consideration for publication by the journal but adds that if the journal agrees to the second request (withdrawal) it need not return the copies to him. In response, the journal returned the copies to Seshadri. Explaining to Kasraian what he had done, Seshadri said that Kasraian had "informed me of having obtained certain results while in fact you did not obtain those results nor, I suspect, did you actually compute them at the time the manuscript was written."

Kasraian resubmitted the paper to the journal under his own name, and the journal agreed to publish it. In a subsequent letter to Kasraian, Seshadri contended that all but two sections of the paper had been written entirely by himself, and that those two sections "were written by me from the drafts that you supplied. With my guidance, you carried out the entire numerical computations and obtained all the figures in the manuscript. . . . Since the collaborative work began only with the applications reported in [those two sections], and since all the numerical results and the figures were obtained by you, although with my active participation, because of the possible unethical practice involved, I do not want to associate my name with those parts of the work."

Kasraian wrote Seshadri that he had inserted in the article the following acknowledgment: "The author is grateful to Professor S.R. Seshadri for his help in preparation of the manuscript and for many fruitful discussions on second-order Bragg interaction." The letter asks whether Seshadri would prefer to have the acknowledgment deleted. Seshadri replied, "You are not permitted to publish any of my ideas and words as your own. The research leading to the manuscript was suggested, supervised, and guided by me. The research was carried out jointly with me. . . . Since that manuscript is tainted by unethical practice and contains errors, you were not allowed to use my name." Kasraian deleted the acknowledgment. The paper was published with only minor changes from the version submitted back in November of 1992. There is no evidence that the paper contains errors; and the citation history of the article does not reveal any criticisms of it.

Anyone reading the correspondence between Seshadri and Kasraian would conclude that the article was indeed a joint work. Seshadri not only submitted it under both names but listed Kasraian's first, and while K does precede S in the alphabet, it would be odd for a senior professor to list a graduate student's name before his own if the student had contributed nothing more to the article than the usual assistance that a research assistant provides. This is not conclusive, of course; Seshadri may just be unusually generous toward his students, though his subsequent conduct rather belies this suggestion. The assistance that a research assistant or secretary or draftsman or helpfully commenting colleague provides in the preparation of a scholarly paper does not entitle the helper to claim the status of a joint author. Childress v. Taylor, 945 F.2d 500, 507 (2d Cir. 1991). To be a joint author, an assistant or collaborator must contribute significant copyrightable material. Id. at 506-07; Erickson v. Trinity Theatre, Inc., supra, 13 F.3d at 1070-71; M.G.B. Homes, Inc. v. Ameron Homes, Inc., 903 F.2d 1486, 1493 (11th Cir. 1990).

This is merely a default rule; by contract the sole creator of a work can transfer "authorship" in whole or part to another. Childress v. Taylor, supra, 945 F.2d at 507; 17 U.S.C. § 201(b). But it is a sensible rule, because ordinarily the person who does not contribute significant copyrightable work does not expect to have a copyright interest, and one who does does. As the acknowledged author of the first draft of two sections that occupy almost 5 pages of a 13-page article, as well as of the technical apparatus of the article, which includes a number of complicated graphs, Kasraian certainly qualifies as a joint author. His contribution was both copyrightable and significant.

If a joint work is marred by errors reflecting unfavorably on his coauthor, with quantifiable adverse effects on the coauthor's career, the coauthor might conceivably have some legal remedy, but it wouldn't be under the Copyright Act. We don't know what it would be under: possibly the law of contracts; in Europe it might be a violation of the author's "moral right" (droit moral), the right to the integrity of his work; and there are glimmers of the moral-rights doctrine in contemporary American copyright law. See Lee v. A.R.T. Co., 125 F.3d 580, 582-83 (7th Cir. 1997); Weinstein v. University of Illinois, supra, 811 F.2d at 1095 n. 3; WGN Continental Broadcasting Co. v. United Video, Inc., 693 F.2d 622, 625 (7th Cir. 1982); Carter v. Helmsley-Spear, Inc., 71 F.3d 77, 81-82 (2d Cir. 1995); Gilliam v. American Broadcasting Cos., 538 F.2d 14, 24 (2d Cir. 1976). But all that matters here is that a joint author does not lose his copyright by being a lousy scholar; were that the rule, rights of joint authorship would be in legal limbo.

We do not understand Seshadri to disagree with anything we have said so far about his copyright claim. His argument rather is that we must for purposes of summary judgement accept the assertions in his affidavit (actually affidavits, but we can ignore that detail). There he claims to have written every word in the article, as well as in the previous four articles "jointly authored" by the two of them. As for portions of the article that are based on manuscript in Kasraian's handwriting, Seshadri explains that Kasraian's handwritten portions "are words that Seshadri spoke and wrote down on paper" and that Kasraian simply copied those words. But if every word was Seshadri's, where did the errors that so exercise him come in? The answer Seshadri gives is that he used numbers that Kasraian had given him, that the numbers were wrong, and that he missed the mistakes the first time around because he failed to check the accuracy of Kasraian's work — classic research assistance, as Seshadri describes it — carefully.

Although professors and heads of laboratories are sometimes accused of taking credit for work done by their junior colleagues and graduate students, we have never heard of a case in which a professor listed a graduate student as the lead author of an article (in fact five articles) every word of which had been written by the professor, the student's contribution being limited to doing some computations. (In the words of the affidavit, "All of the ideas, concepts, words, expressions, figures and tables are the work and expressions of Seshadri.") But in any event, we do not think that an affidavit should be allowed without explanation to controvert the affiant's written admissions, albeit made in documents rather than in a deposition — indeed documents that having been composed before the litigation would carry more conviction than a deposition, even though a deposition is given under oath. The documents acknowledge Kasraian as joint author on the basis of his having drafted a substantial portion of the manuscript. Even if, as we doubt, Seshadri rewrote those portions completely, he was working from Kasraian's draft, and Kasraian is acknowledged to have performed all the computations and supplied all the tables in an article that contains 92 numbered equations and numerous tables, figures, and other computations.

We said earlier that a party might be able to explain away his prior inconsistent statements. But here we note the corollary: if he makes no attempt at explanation, he is stuck with those statements. The statements are numerous and unequivocal, and Seshadri's only explanation — that when he wrote the Journal of Applied Physics that "We carried out the theoretical analysis for the above-mentioned manuscript" he was using the royal "we" is both incredible and, in light of all the other statements showing joint authorship, incomplete.

It is not our competence to find facts. But a motion for summary judgement should be granted when the record is such that if it were the record of a trial no reasonable jury could find for the party opposing the motion. Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 249 (1986); Bombard v. Fort Wayne Newspapers, Inc., 92 F.3d 560, 562 (7th Cir. 1996); Haynes v. Alfred A. Knopf, Inc., 8 F.3d 1222, 1233-34 (7th Cir. 1993); Lawrence v. National Westminster Bank, 98 F.3d 61, 65 (3d Cir. 1996). Evidence that is not "significantly probative" will not justify a trial. Anderson v. Liberty Lobby, Inc., supra, 477 U.S. at 249-50. That is why the "law of nature" formula for determining when testimony may be disbelieved on summary judgement is an exaggeration, and why summary judgement was rightly granted here. The plaintiff's admissions show that Kasraian was the joint author of the article, and he could not by filing a blind affidavit — one that failed to explain barefaced inconsistencies with his prior statements — retract those admissions. The only novelty in our ruling on summary judgement on the copyright count, therefore, is in refusing to allow what we have called a blind affidavit to retract admissions even if not embodied in a deposition; and it is a novelty thoroughly in accord with the statement in the Bessemer City case that we quoted earlier.

We need not consider whether Seshadri abandoned his copyright claim when he refused to allow his name to be used in conjunction with the article. But we add for completeness that this issue could not be resolved on summary judgement on the basis of the record compiled to date. Had Seshadri authorized Kasraian or the Journal of Applied Physics to publish the article under Kasraian's sole name, that would be abandonment — a statement or other act that demonstrates an intention of relinquishing any copyright interest in a work. Fantastic Fakes, Inc. v. Pickwick Int'l, Inc., 661 F.2d 479, 486 n. 6 (5th Cir. 1981); National Comics Publications, Inc. v. Fawcett Publications, 191 F.2d 594, 598 (2d Cir. 1951) (L. Hand, J.). Authorizing another to publish under his sole name would amount to a public disclaimer of authorship, and while authors are not the only people who can hold copyrights, there is no suggestion of any other basis for Seshadri's claim.

But that is not the inevitable reading of the correspondence, although it is the most plausible one. An alternative reading is that Seshadri simply did not want his name associated with a work poisoned by his collaborator's errors. It would be an assertion rather than an abandonment of copyright to tell a journal not to publish one's article because it contained errors introduced by someone else. Implicit in the copyright holder's exclusive right to distribute copies of his work to the public, 17 U.S.C. § 106(3), is the right not to publish the work, Harper Row, Publishers, Inc. v. Nation Enterprises, 471 U.S. 539, 551 (1985); Forward v. Thorogood, 985 F.2d 604 (1st Cir. 1993); and that as we say is a possible interpretation of what Seshadri told the Journal of Applied Physics when he withdrew the article. As for his refusal to authorize the acknowledgment, that too need not be read as a license to publish the article without it. So there was no abandonment, at least as a matter of law. But as there was joint authorship, Seshadri's only right under the copyright law was to the accounting that he has not sought.

Affirmed.