For several years panels of this court have used divergent standards to review decisions about sanctions under Fed.R.Civ. P. 11. Compare R.K. Harp Investment Corp. v. McQuade, 825 F.2d 1101, 1103 (7th Cir.1987), In re Central Ice Cream Co., 836 F.2d 1068, 1072 (7th Cir.1987), and Borowski v. DePuy, Inc., 850 F.2d 297, 304 (7th Cir.1988), using a deferential standard of review, with Brown v. Federation of State Medical Boards, 830 F.2d 1429, 1434 (7th Cir.1987), S.A. Auto Lube, Inc. v. Jiffy Lube International, Inc., 842 F.2d 946, 948 (7th Cir.1988), and Beeman v. Fiester, 852 F.2d 206, 209 (7th Cir.1988), using a “de novo” standard on some issues. See also Lebovitz v. Miller, 856 F.2d 902, 904 (7th Cir.1988), and FDIC v. Tefken Construction & Installation Co., 847 F.2d 440, 442-43 (7th Cir.1988), tracing developments within the circuit. We heard this ease en banc to achieve harmony. From now on, this court will use a deferential standard consistently — whether sanctions were imposed or not, whether the question be frivolousness on the objective side of Rule 11 or bad faith on the subjective side.
I
Parallel class action suits against Continental Bank N.A. (formerly Continental Illinois National Bank and Trust Co. of Chicago) led to an award of sanctions under Rule 11. William J. Tunney filed the first suit {Tunney) in state court; Mars Steel Corp. filed the second a year later in federal court. Each complaint alleged that Continental Bank had broken its contracts with, and defrauded, customers to whom it had agreed to lend money at a rate linked to the “prime rate”. The state suit alleged fraud and breach of contract, the federal suit a violation of the Racketeer Influenced and Corrupt Organizations Act (RICO), 18 U.S.C. §§ 1961-1968.
Continental’s contracts defined the prime rate as the rate it charges “for 90-day unsecured commercial loans to large corporate customers of the highest credit standing”. According to the complaints, Continental loaned money to some customers at less than the rate it had announced as its prime. Since the contractual definition of prime is the rate for customers of the “highest” credit standing, the plaintiffs contended that rates below prime are logically impossible, and that Continental had developed the concept of below-prime loans in order to avoid reducing its prime rate— for had Continental reduced that rate, it would have received less from the plaintiffs. Continental, for its part, denied that the loans at rates less than prime were “90-day unsecured commercial loans”. (Some had compensating-balance requirements, others were for terms shorter than 90 days, and so forth.)
After the state court certified Tunney as a class action, it lay dormant. Discovery proceeded in the federal case. When the evidence appeared to bear out Continental’s position, implying that the case lacked merit, Mars Steel agreed to settle the litigation on terms that were worth at most $11.5 million to the 23,000 class members. Joyce and Kubasiak, P.C., the law firm representing the Tunney class, tried to derail the agreement. Settlement negotiations between Continental and the Tunney class had collapsed over Joyce & Kubasiak’s demand for $1.25 million in legal fees — this despite having conducted no discovery, while Jerome Torshen, after aggressively *931pursuing discovery on behalf of Mars Steel, accepted $300,000 for all work in the federal suit. Although quiescent in the state case, Joyce & Kubasiak (through Edward T. Joyce, Peter B. Carey, and Steven J. Rotunno) commenced a campaign of gueril-la warfare in the federal case. It failed: the district court approved the settlement, and we affirmed. 834 F.2d 677 (7th Cir.1987). The district court “denied every motion filed by J & K except for two motions to file a brief in excess of 15 pages, a motion to conduct limited discovery before the fairness hearings and a motion to submit a corrected page.” 120 F.R.D. 53, 54 (N.D. Ill.1988). The details are unimportant, except for the two motions that led to sanctions against Tunney, Joyce, Carey, and Rotunno (the appellants).
On July 22,1986, the district court issued an order preliminarily approving the settlement and certifying the Mars Steel class for purposes of settlement only. It ordered the parties to mail a notice of settlement under Fed.R.Civ.P. 23(e) to all class members on or before August 1. On July 25 Joyce & Kubasiak filed a motion to reconsider, and on July 29 the court set a briefing schedule on this motion. Joyce & Kubasiak did not, however, seek a stay of the order directing the parties to mail notices by August 1, and the court did not issue one. The notices went out on schedule. On August 8 Joyce & Kubasiak filed a motion for a rule to show cause, arguing that counsel for Mars Steel and Continental were in contempt of court for mailing the notices, because the mailing “violated” the order setting a briefing schedule. The district court summarily denied this motion and later held that it was frivolous within the meaning of Rule 11. 120 F.R.D. at 55 n. 1.
Because Tunney, a member of the Mars Steel class, objected to the settlement in response to the notice, the district court held a hearing on the fairness of the disposition. Mars Steel and Continental offered seven affidavits. Joyce & Kubasiak objected to these on the ground that affidavits denied them the opportunity to cross-examine the affiants. Concluding that hearsay is admissible during a fairness hearing, the district court overruled the objection and admitted the affidavits. After the hearing Joyce & Kubasiak moved to strike the affidavits on a different ground: because some of the affiants had been available to testify, the affidavits were inadmissible. The new ground fared no better than the old, and the district court held that this motion too violated Rule 11. 120 F.R.D. 55-56 n. 2. The court pointed out that if it may admit hearsay in a fairness hearing the presence of the affiant changes nothing.
These two transgressions led to an award of $1,648 to cover the parties’ reasonable costs of responding. Although the district judge was tempted to hold that Joyce & Kubasiak’s entire course of conduct was objectively frivolous and conducted in bad faith to boot, he held back because he believed that this court, by discussing Joyce & Kubasiak’s arguments in affirming his approval of the settlement, had implicitly determined that the arguments were neither frivolous nor pressed in bad faith. 120 F.R.D. at 55-57. Mars Steel and Continental have not appealed from the order to the extent it denied their request for additional sanctions, so we shall not have to determine whether the judge’s inference from our opinion was appropriate.
II
Rule 11 provides that a lawyer’s or party’s signature on any paper filed in district court
constitutes a certificate by the signer that the signer has read the pleading, motion, or other paper; that to the best of the signer’s knowledge, information, and belief formed after reasonable inquiry it is well grounded in fact and is warranted by existing law or a good faith argument for the extension, modification, or reversal of existing law, and that it is not interposed for any improper purpose, such as to harass or to cause unnecessary delay or needless increase in the cost of litigation.
This has both a subjective and an objective component. A paper “interposed for any *932improper purpose” is sanctionable whether or not it is supported by the facts and the law, and no matter how careful the pre-fil-ing investigation. The objective component is that a paper filed in the best of faith, by a lawyer convinced of the justice of his client’s cause, is sanctionable if counsel neglected to make “reasonable inquiry” beforehand. This objective component is the big change in the 1983 amendment to Rule 11. “An empty head but a pure heart is no defense. The Rule requires counsel to read and consider before litigating.” Thornton v. Wahl, 787 F.2d 1151, 1154 (7th Cir.1986). Counsel may not drop papers into the hopper and insist that the court or opposing counsel undertake bothersome factual and legal investigation.
Rule 11 is not a fee-shifting statute in the sense that the loser pays. It is a law imposing sanctions if counsel files with improper motives or inadequate investigation. Under the American Rule on the allocation of attorneys’ fees, each side bears its own expenses. Rule 11 ensures that each side really does bear the expenses of its own case — that the proponent of a position incurs the costs of investigating the facts and the law. In re TCI Ltd., 769 F.2d 441, 446 (7th Cir.1985). Its focus is ex ante (what should have been done before filing) rather than ex post (how things turned out). How much investigation is justified (i.e., “reasonable”) in light of the costs depends on the circumstances of the case, and Rule 11 does not allow an award of sanctions just because things went poorly after an investigation that was adequate in light of what was known (and how much time was available) before the paper was filed. FDIC v. Elefant, 790 F.2d 661, 668 (7th Cir.1986); District No. 8 v. Clearing, 807 F.2d 618, 622 (7th Cir.1986); Frantz v. United States Powerlifting Federation, 836 F.2d 1063, 1067-68 (7th Cir.1987). Sanctuary as a result of reasonable investigation ensures that counsel may take novel, innovative positions— that Rule 11 does not jeopardize aggressive advocacy or legal evolution.
Unlike the traditional fee-shifting statute, Rule 11 focuses on inputs rather than outputs, conduct rather than result. Stephen B. Burbank (reporter), Rule 11 in Transition: The Report of the Third Circuit Task Force on Federal Rule of Civil Procedure 11 10-25 (American Judicature Society 1989); Thomas E. Willging, The Rule 11 Sanctioning Process 35-65 (Federal Judicial Center 1988). It establishes a new form of negligence (legal malpractice). E.g., Central Ice Cream, 836 F.2d at 1072 (“the Rule speaks of ‘reasonable’ pre-filing inquiry, the language of tort law”); Classic Components Supply, Inc. v. Mitsubishi Electronics America, Inc., 841 F.2d 163, 165 (7th Cir.1988); Hays v. Sony Corp., 847 F.2d 412, 418 (7th Cir.1988) (“Rule 11 defines a new form of legal malpractice”). Rule 11 creates duties to one’s adversary and to the legal system, just as tort law creates duties to one’s client. The duty to one’s adversary is to avoid needless legal costs and delay. The duty to the legal system (that is, to litigants in other cases) is to avoid clogging the courts with paper that wastes judicial time and thus defers the disposition of other cases or, by leaving judges less time to resolve each case, increases the rate of error. Rule 11 allows judges to husband their scarce energy for the claims of litigants with serious disputes needing resolution.
As in tort law, the event sometimes speaks for itself. That is, Rule 11 no less than common law recognizes the doctrine of res ipsa loquitur. A lawyer who founds his suit on Plessy v. Ferguson, 163 U.S. 537, 16 S.Ct. 1138, 41 L.Ed. 256 (1896), has revealed all we need to know about the reasonableness of the pre-filing inquiry. Szabo Food Service, Inc. v. Canteen Corp., 823 F.2d 1073, 1080-82 (7th Cir.1987). If the legal point is obscure, though, even an absurd argument may not be sanctionable, because a “reasonable” inquiry does not turn up every dusty statute and precedent. Elefant, 790 F.2d at 667. An objectively frivolous legal position supports an inference that the signer did not do a reasonable amount of research, but an inference, no matter how impressive, is still no more than an inference. In most cases the amount of research into legal questions that is “reasonable” depends on whether *933the issue is central, the stakes of the case, and related matters that influence whether further investigation is worth the costs.
The focus of Rule 11 on conduct rather than result, its close affiliation to tort law, and the fact that objectively frivolous filings support but do not compel an inference of unreasonable investigation, mean that each Rule 11 case in the district court is unique, just as every tort suit is unique. How much investigation should have been done in a given case becomes a question of line-drawing, as much a matter of “fact” as is the purpose behind the paper. Decisions in tort cases are reviewed on appeal under a deferential standard: findings of fact may be disturbed only if “clearly erroneous”, and resolutions requiring the application of judgment may be undone only if the district court has abused its discretion. The panel assembled under the auspices of the Third Circuit Judicial Council believes that this uniformly deferential standard is appropriate under Rule 11, see Rule 11 in Transition at 45-49, and we agree. We therefore follow the six other courts of appeals that employ deferential review across the board. Kale v. Combined Insurance Co., 861 F.2d 746, 756-58 (1st Cir.1988); Teamsters v. Cement Express, Inc., 841 F.2d 66, 68 (3d Cir.1988); Stevens v. Lawyers Mutual Liability Insurance Co., 789 F.2d 1056, 1060 (4th Cir.1986); Thomas v. Capital Security Services, Inc., 836 F.2d 866, 871-73 (5th Cir.1988) (en banc); Century Products, Inc. v. Sutter, 837 F.2d 247, 250 (6th Cir.1988); Adamson v. Bowen, 855 F.2d 668, 673 (10th Cir.1988). We respectfully disagree with the Ninth Circuit, which has adopted a multi-tier approach to review, e.g., Zaldivar v. City of Los Angeles, 780 F.2d 823, 828 (9th Cir.1986) (findings of historical fact reviewed on clearly erroneous standard, decision whether counsel violated Rule 11 reviewed de novo, and choice of sanction reviewed for abuse of discretion), and the three circuits that apply the de novo standard to the objective components of Rule 11, Donaldson v. Clark, 819 F.2d 1551, 1556 (11th Cir.1987); Westmoreland v. CBS, Inc., 770 F.2d 1168, 1174-75 (D.C.Cir.1985); Eastway Construction Corp. v. City of New York, 762 F.2d 243, 254 n. 7 (2d Cir.1985).
De novo review would be contrary to the spirit of many recent cases, such as Pullman-Standard v. Swint, 456 U.S. 273, 102 S.Ct. 1781, 72 L.Ed.2d 66 (1982); Anderson v. Bessemer City, 470 U.S. 564, 105 S.Ct. 1504, 84 L.Ed.2d 518 (1985); and Icicle Seafoods, Inc. v. Worthington, 475 U.S. 709, 106 S.Ct. 1527, 89 L.Ed.2d 739 (1986), reiterating the message that appellate courts should use deferential review for fact-intensive disputes, including the application of legal rules to facts. See also, e.g., Scandia Down Corp. v. Euroquilt, Inc., 772 F.2d 1423, 1427-29 (7th Cir.1985); Mucha v. King, 792 F.2d 602, 605-06 (7th Cir.1986); Tonya K. v. Board of Education, 847 F.2d 1243, 1248 (7th Cir.1988); Foy v. First National Bank of Elkhart, 868 F.2d 251, 254 (7th Cir.1989). Whether counsel did an appropriate amount of pre-filing investigation, and whether a legal position is far enough off the mark to be “frivolous”, are fact-bound. Whether the lawyer “went too far” (or “didn’t do enough”) is inevitably a judgment call.
Fact-intensive disputes, those whose resolution is unlikely to establish rules of future conduct, are reviewed under a deferential standard because the role of appellate courts in establishing and articulating rules of law is not at stake. District judges have the best information about the patterns of their cases, information appellate judges could duplicate only at great cost in time. As our chief judge wrote not so long ago, “because the trial court alone has an intimate familiarity with the relevant proceedings, its decision whether counsel has conducted the kind of inquiry required by Rule 11 and taken a position reasonable in light of the facts and governing law is reviewable only where there has been an abuse of discretion.” R.K Harp Investment Corp., 825 F.2d at 1103. Even when the dispute may be resolved by examining documents, the claim that the court of appeals is in as good a position as the district judge to review a written record disregards the division of labor (why should three judges redo the work of one?), *934the sense of the situation that is valuable on all occasions, and the fact that having redone all of the work the court of appeals would not produce a valuable precedent. Anderson, 470 U.S. at 573-76, 105 S.Ct. at 1511-13.
Whether a particular position lacked reasonable support in the facts and law is one of these non-repeating matters. Positions rarely are frivolous in the abstract. There is first a question of degree — how far from the correct legal position does one have to travel to commit this sin? — and next a question of application: a statement might be “frivolous” as applied to factual pattern X but not as applied to pattern Y. Appellate judges could come to grips with these, but the effort to do so would be wasteful because every case would still be a little different from the last, and the process would have to be repeated. There is little to gain, and much to lose, in duplicating the efforts of district judges. Anderson, 470 U.S. at 574-75, 105 S.Ct. at 1511-12. One standard for frivolousness is risibility — if you start laughing when repeating the argument, then it’s frivolous. Cf. Sanford Levinson, Frivolous Cases: Do Lawyers Really Know Anything at All?, 24 Osgoode Hall L.J. 353 (1987). Laughter often depends on context (“You had to be there”), and no less so under Rule 11. Because identifying an unsupported factual or legal position simply starts a chain of inference — does the feeble position demonstrate that the pre-filing investigation was unreasonable? — the district court’s role is even greater, and appellate judges farther removed from the scene.
Today’s case illustrates the point. Whether the filing of a motion to strike affidavits was supported by reasonable investigation is a factual matter (how much investigation, if any, did counsel do?) before it is a legal one. Whether counsel did “enough” investigation before filing depends on the context, and our resolution of it is unlikely to be useful to any other litigants. The district judge has a comprehensive vision that we lack. The motion was only one document among a blizzard of papers filed by Joyce & Kubasiak, and its reasonableness (as well as the motive behind it) cannot be understood without regard to the rest of the filings and the firm’s litigating strategy. Each of the three judges on a panel of the court of appeals could sink many hours into the case without duplicating the district court’s feeling for the situation, and without generating an opinion having the slightest significance for tomorrow’s disputes.
Although we think that deferential review is appropriate for reasons we have given, Pierce v. Underwood, — U.S. -, 108 S.Ct. 2541, 2546-49, 101 L.Ed.2d 490 (1988), would provide a powerful impetus toward that conclusion even if we were otherwise disposed. Underwood holds that courts of appeals should use uniformly deferential review when assessing a district court’s conclusion under the Equal Access to Justice Act, 28 U.S.C. § 2412(d), that the government’s litigating position is (or isn’t) “substantially justified”. Underwood concluded that “the sound administration of justice”, 108 S.Ct. at 2547, quoting from Miller v. Fenton, 474 U.S. 104, 114, 106 S.Ct. 445, 451, 88 L.Ed.2d 405 (1985), calls for committing case-specific decisions concerning the degree of factual or legal support for a litigating position to district courts, with review for clear error. That deferential review is the right standard for factual conclusions is something every member of this court accepts. Underwood said this about legal questions:
In some cases ... the attorney’s fee determination will involve a judgment ultimately based upon evaluation of the purely legal issue governing the litigation. It cannot be assumed, however, that de novo review of this will not require the appellate court to invest substantial additional time, since it will in any case have to grapple with the same legal issue on the merits. To the contrary, one would expect that where the Government’s case is so feeble as to provide grounds for an EAJA award, there will often be (as there was here) a settlement below, or a failure to appeal from the adverse judgment. Moreover, even if there is a merits appeal, and even if it occurs simultaneously with (or goes to *935the same panel that entertains) the appeal from the attorney’s fee award, the latter legal question will not be precisely the same as the merits: not what the law now is, but what the Government was substantially justified in believing it to have been. In all the separate-from-the-merits EAJA appeals, the investment of appellate energy will either fail to produce the normal law-clarifying benefits that come from an appellate decision on a question of law, or else will strangely distort the appellate process....
[An additional reason for deferential review is] that the question whether the Government’s litigating position has been “substantially justified” is precisely such a multifarious and novel question, little susceptible, for the time being at least, of useful generalization, and likely to profit from the experience that an abuse-of-discretion rule will permit to develop.
108 S.Ct. at 2547-48. These considerations are equally applicable to Rule 11. Deciding whether a party’s legal proposition is “frivolous” does not clarify the law; it merely marks the boundaries of departures from the law. “Degrees of silliness” do not establish legal rules — though an invitation to appeal in order to get a de novo call on every close question contains great potential for litigation.
Undoubtedly there are differences between the EAJA and Rule 11. The district judge awards fees under the EAJA “unless the court finds” substantial justification, while under Rule 11 the court “shall” award sanctions if it finds a violation. Brown, 830 F.2d at 1434 n. 3; Central Ice Cream, 836 F.2d at 1072. Differences of this kind are immaterial for our purpose, however. The EAJA is stacked against the government, but the deferential standard applies even if the government wins in the district court; the bias in the rule does not govern the standards on appeal. Rule 11 is biased against an award of fees, for unless the judge finds unreasonable conduct he can’t impose fees. This is the same sense in which tort law is biased against a finding of liability; would it follow that appellate review should be more searching in tort eases than in EAJA cases? In civil rights cases? The law is otherwise. Pullman-Standard, v. Swint.
Our question is not whether a court “shall” impose a sanction when it finds a violation; it must; the question is instead how to determine whether there was a violation. Some of the early cases under Rule 11, e.g., Westmoreland, inferred de novo appellate review from the “shall impose” language, but more recent cases, e.g., Kale, 861 F.2d at 757-58, have recognized that the obligatory sanction does not prescribe the relation between trial and appellate judges. It is a rule of law that if the defendant is negligent, the court shall award a remedy. Does it follow that issues concerning the reasonableness of the defendant’s conduct are reviewed de novo in tort cases? The requirement: “if violation then remedy” doesn’t help a judge find violations or prescribe which judges (trial or appellate) are principally responsible for the decision.
Those seeking to distinguish the EAJA from Rule 11 also observe that the average fee under the EAJA is about $3,000, while the stakes in a Rule 11 case include not only money but also reputation. We do not know the average award under Rule 11 (most awards in this circuit have been small; the $1,648 in our case is typical). As for reputation: not “substantially justified”, the way Underwood construed that term (to mean “unreasonable”), is scarcely different from not preceded by “reasonable investigation” into the law under Rule 11. An award of EAJA fees imputes professional delicts to the lawyer for the government. More to the point, appellate review does not become more searching as the stakes of the case rise. The stakes may influence the standard of proof (although courts cheerfully award $50 million in antitrust cases using the preponderance-of-evidence standard) but do not speak to the role of appellate judges. Underwood itself involved $300,000, and the precedential effect of the case resolved many similar disputes. Yet the Court applied the same deferential review in Underwood that it called for in $2,000 cases.
*936Many judges and lawyers are concerned about the lack of uniformity in the application of Rule 11. See William W. Schwarzer, Rule 11 Revisited, 101 Harv. L. Rev. 1013 (1988); Georgene M. Vairo, Rule 11: A Critical Analysis, 118 F.R.D. 189 (1988). Legitimate though this concern is, it is in the end no more serious than the same concern among physicians about uniformity in medical malpractice cases, or among manufacturers of drugs about uniformity in failure-to-warn cases. Fact-bound resolutions cannot be made uniform through appellate review, de novo or otherwise. Thomas, 836 F.2d at 873. Justice Holmes believed that courts would (at least, should) slowly reduce all of tort law to objective, readily applied rules. Law in Science and Science in Law, 12 Harv. L. Rev. 443 (1899), reprinted in Collected Legal Papers 234-39 (1920); cf. The Common Law 112-15 (1881). This is not viewed today as one of his more astute predictions. To the extent there are enduring differences among judges about how to evaluate the propriety of lawyers’ conduct, shifting decisions to appellate courts will just change the locus of confusion. Panels of this court are capable of discord to the same extent as different district judges, and because of the fact-specific nature of the inquiries no “rule of law” will ensure that the outlook of the first three judges of this court will be shared by the rest. We can multiply the occasions for disagreement much more easily than we can bring harmony. Appellate confusion is worse: counsel and district judges must spend time trying to harmonize and apply our opinions. This is one of the reasons why ease-specific decisions should be made in the trial courts and reviewed on a deferential standard. The judgment of three appellate judges is not necessarily better than the judgment of one district judge, but it is assuredly more costly to obtain and interpret.
To say that appellate review cannot bring about a set of rules defining allowable conduct is not to say that courts of appeals will look at each case as unconnected from the past and irrelevant to the future. District courts must make findings and exercise discretion against a background of legal rules. If they err in appreciating or applying legal rules, their mistakes will be undone. All of the “pure” questions of law implicit or explicit in the district judge’s findings are reviewable for error, and over time these decisions should build up a reasonably stable and useful body of rules, to the extent sanctions decisions are amenable to governance by rule. Moreover, even when the case does not contain a controlling legal issue,
Review under the abuse of discretion standard does not mean no appellate review. Rule 11 sanctions have significant impact beyond the merits of the individual case. Concerns for the effect on both an attorney’s reputation and for the vig- or and creativity of advocacy by other members of the bar necessarily require that we exercise less than total deference to the district court in its decision to impose Rule 11 sanctions. As we have stated: “Despite the increased license to impose sanctions, judges should always reflect seriously upon the nuances of the particular case, and the implications the case has on the nature of the legal representation, before imposing sanctions.”
In re Ronco, Inc., 838 F.2d 212, 217-18 (7th Cir.1988), quoting from Brown v. National Board of Medical Examiners, 800 F.2d 168, 173 (7th Cir.1986). Deferential review will not prevent this court from ensuring that district judges reflect seriously, and consider fully, before imposing (or denying) sanctions. Until discretion has been taken seriously in the district court and exercised conscientiously, the basis for deference is missing. Frantz, 836 F.2d at 1067-68; Szabo, 823 F.2d at 1083-84; United States v. Beasley, 809 F.2d 1273, 1279 (7th Cir.1987).
We end up where the Supreme Court did in Underwood: “although ... our resolution of this issue is not rigorously scientific, we are satisfied that the text ... permits, and sound judicial administration counsels, deferential review of a district court’s decision regarding attorneys fees”, 108 S.Ct. at 2549. Our exploration of the subject has been assisted by the briefs of six bar associations as amici curiae, filed *937at the court’s invitation when the case was set for argument en banc. We thank the associations for giving of their time and experience, and we look forward to their continued assistance when in the future we seek their counsel on issues of mutual interest to bench and bar.
Ill
Judge Aspen’s conclusion that two motions filed by Joyce & Kubasiak violated Rule 11 is not based on clear errors of fact and does not abuse his discretion.
The Motion for a Rule to Show Cause. By order of July 22, 1986, the court directed the parties to mail the Rule 23(e) notices to the class by August 1,1986. Joyce & Kubasiak, on behalf of Tunney and another of the named plaintiffs in the state suit, asked the court to rescind this order. The judge filed a minute order stating in full:
Enter Tunney and Weisbrook’s motion to reconsider this Court’s July 22, 1986 order. Any party wishing to respond, may do so, on or before August 5,1986. Tunney and Weisbrook to reply by August 8, 1986. Tunney and Weisbrook’s motion for leave to conduct discovery before fairness hearing is denied.
Nothing in this order rescinds the deadline for sending the notices. Mars Steel and Continental Bank accordingly were not only entitled but also required to mail them by August 1, as they did.
Joyce & Kubasiak do not contend that they did even a smidgeon of research before asking the district court to hold counsel for Mars Steel and Continental Bank in contempt of court for violating a nonexistent order putting off the deadline. Their motion cites nary a case. It argued only that the motion to reconsider automatically stayed the order. Observing that a motion to reconsider a final decision under Fed.R.Civ.P. 59 tolls the time to appeal under Fed.R.App.P. 4(a)(4), Joyce & Kubasiak insisted that the motion to reconsider the order stayed it. Rule 4 deals with the time to appeal, not with the effectiveness of the order under attack. Unless the party obtains a stay from the district court under Fed.R.Civ.P. 62(b) or from the court of appeals, an outstanding order must be obeyed, whether or not a Rule 59 motion or a notice of appeal has been filed. We imposed sanctions in Thornton, 787 F.2d at 1153, on an almost identical claim made in a brief filed in our court. It was not an abuse of discretion for Judge Aspen to impose sanctions when the same assertion was made in his.
Joyce & Kubasiak asked us to review Rule 11 decisions de novo in large measure to protect attorneys from undeserved calumny. Counsel for Continental Bank cogently observed at oral argument that this represents something of a change of heart. Joyce & Kubasiak showed no similar concern for the reputations of members of the bar when it asked Judge Aspen to hold its adversaries’ attorneys in contempt of court for violating a nonexistent order.
The Motion to Strike. After the fairness hearing, Joyce & Kubasiak renewed a motion to strike that had been presented orally and overruled on the spot. The district judge concluded that this renewal lacked factual and legal support. As before, appellants do not contend that they did any significant research before renewing the motion or changing the grounds on which it was based. We conclude that the district court did not abuse its discretion in finding the motion sanctionable.
Some of the colloquy at the time of the oral motion might have been understood to leave open the possibility of further deliberations on the question. That a question has not been forever decided does not imply, however, that without further inquiry counsel may take any old position that comes to mind. Reserved questions are to be reopened after further reflection and illumination. Judge Aspen was entitled to believe that Joyce & Kubasiak offered neither. The five-page motion to strike cited only two cases, Galella v. Onassis, 487 F.2d 986 (2d Cir.1973), and Hooker v. New York Life Insurance Co., 66 F.Supp. 313 (N.D.Ill.1946). Both predate the Federal Rules of Evidence; neither deals with fairness hearings under Rule 23(e). The mo*938tion paid no attention to cases saying that when deciding whether a settlement of a class action should be approved, the court can use “whatever is necessary to aid it in reaching an informed, just and reasoned decision.” Glicken v. Bradford, 35 F.R.D. 144, 148 (S.D.N.Y.1964). See also Flinn v. FMC Corp., 528 F.2d 1169, 1173 (4th Cir.1975); Cotton v. Hinton, 559 F.2d 1326, 1331 (5th Cir.1977).
To say that Joyce & Kubasiak did not construct a plausible legal argument is not to say that the firm couldn’t have. The Federal Rules of Evidence “govern proceedings in the courts of the United States ... to the extent and with the exceptions stated in rule 1101.” Rule 101. The exceptions, listed in Rule 1101(d), include proceedings such as extradition and sentencing in criminal cases. Fairness hearings conducted under Fed.R.Civ.P. 23(e) are not among the proceedings excepted from the Rules of Evidence. No case of which we are aware holds that Rule 1101(d) suspends the usual rules of evidence for fairness hearings; no case expressly holds that affidavits are admissible in such hearings, although several cases mention their use without deciding the propriety of that use. E.g., Handschu v. Special Services Division, 787 F.2d 828, 834 (2d Cir.1986).
Because Rule 11 is addressed to conduct (the adequacy of the pre-filing investigation) rather than to results, a motion may be sanctionable even though something could have been said in its behalf. Ronco, 838 F.2d at 218-19 & n. 10; Frazier v. Cast, 771 F.2d 259, 265 (7th Cir.1985). Litigants “may not pretend that the law favors their view and impose on the court or their adversaries the burden of research to uncover the basic rule.” Central Ice Cream, 836 F.2d at 1073. That is all Joyce & Kubasiak did. Counsel demanded that Judge Aspen reopen the subject and induced attorneys for the other side to dig into the books, without troubling to do so themselves. Judge Aspen did not abuse his discretion in concluding that Joyce & Kubasiak had violated Rule 11.
IV
Mars Steel and Continental Bank ask us to award attorneys’ fees and double costs as a sanction under Fed.R. App.P. 38 for pursuing a frivolous appeal. Rule 38 provides that if a court of appeals “shall determine that an appeal is frivolous, it may award just damages and single or double costs to the appellee.” Attorneys’ fees are a form of “damages” under Rule 38. An appeal is “frivolous” when the result is foreordained by the lack of substance to the appellant’s arguments. See Williams v. Postal Service, 873 F.2d 1069, 1075 (7th Cir.1989); Ross-Berger Cos. v. Equitable Life Assurance Society, 872 F.2d 1331, 1340-41 (7th Cir.1989); Mays v. Chicago Sun-Times, 865 F.2d 134, 138 (7th Cir.1989); Clearing, 807 F.2d at 623; Coleman v. CIR, 791 F.2d 68 (7th Cir.1986); Spiegel v. Continental Illinois National Bank, 790 F.2d 638, 650 (7th Cir.1986); Indianapolis Colts v. Mayor and City Council of Baltimore, 775 F.2d 177, 184 (7th Cir.1985); Reid v. United States, 715 F.2d 1148, 1154-55 (7th Cir.1983). The standard depends on the work product: neither the lawyer’s state of mind nor the preparation behind the appeal matter. Rule 38 also allows the court of ap peals to decline to impose sanctions even if the appeal is frivolous; it says “may award” while Rule 11 says “shall”. Brooks v. Allison Division of General Motors Corp., 874 F.2d 489 (7th Cir.1989). Still, “discretionary choices are not left to a court's ‘inclination, but to its judgment; and its judgment is to be guided by sound legal principles.’ ” Albemarle Paper Co. v. Moody, 422 U.S. 405, 416, 95 S.Ct. 2362, 2371, 45 L.Ed.2d 280 (1975), quoting from United States v. Burr, 25 F.Cas. 30, 35 (No. 14,682d) (CC Va.1807) (Marshall, C.J.).
The appeal from the award for filing the motion for a rule to show cause had no prospect of success no matter what standard of review this court used. Joyce & Kubasiak could not seriously have thought that this court would hold that they were justified in contending that a motion for rehearing, or a briefing schedule, automatically stays a prior order. *939Thornton had ridiculed a functionally identical claim. Appellants’ brief in this court ignored Thornton. When appellees’ briefs dwelled on that case and demanded sanctions, appellants filed a reply brief that did not mention Thornton. We have told lawyers that “[t]he ostrich-like tactic of pretending that potentially dispositive authority against a litigant’s position does not exist is as unprofessional as it is pointless.” Hill v. Norfolk & Western Ry., 814 F.2d 1192, 1198 (7th Cir.1987). Hill added that the tactic is sanctionable under Rule 38, and this case is no different. Appellants devoted four pages of their reply brief to the request for sanctions under Rule 38 and did not ask for a hearing on that question. None is necessary. Hill, 814 F.2d at 1201-03.
Objections to the award on account of the motion to strike the affidavits stand on a different footing. Resolution of this contention at least potentially turned on our standard of review, an unsettled issue. That aspect of the appeal was not frivolous and is not sanctionable under Rule 38.
Rule 38 is not the only ground for awarding attorneys’ fees and costs to the prevailing party. Sanctions under Rule 11 frequently are designed to make the adverse party whole — to put the party opposing the motion in as good a position as it would have occupied had the motion never been made. If the wronged party must bear its own legal costs on appeal in order to defend its award, it may end up substantially in the hole, may end up, indeed, worse off because of Rule 11. For the costs of defending on appeal an award of $1,648 will substantially exceed that sum. Pyrrhic victories are the stuff of history but hardly balm for legal wounds. It would be unfortunate if aggrieved parties abandoned defense of their awards in order to hold down their own legal costs; as we have emphasized, enforcement of Rule 11 is important to the functioning of the legal system and the protection of the rights of parties in other cases. Those to whom sanctions have been awarded should be induced to defend their awards, without additional costs to themselves. We accordingly held in Hays, 847 F.2d at 419, that a party prevailing in the defense of an award of fees under Rule 11 is presumptively entitled to a further award from the district court, so that the appeal makes him no worse off. See also Hamer v. County of Lake, 871 F.2d 58, 60 (7th Cir.1989); Gorenstein Enterprises v. Quality Care-USA, Inc., 874 F.2d 431 (7th Cir.1989); Borowski v. DePuy, Inc., 876 F.2d 1339 (7th Cir.1989), all reaffirming and following Hays. See also, e.g., Danik, Inc. v. Hartmarx Corp., 875 F.2d 890, 897 (D.C.Cir.1989); Muthig v. Brant Point Nantucket, Inc., 838 F.2d 600, 607 (1st Cir.1988) (following this approach).
Hays, Hamer, Gorenstein, and Borowski would support an award under Rule 11 without regard to our conclusion under Rule 38 that at least one aspect of the appeal is not frivolous. Today’s case illustrates, however, the need for discretionary judgments under Rule 11 concerning the amount of sanctions. Tefken, 847 F.2d at 443 n. 3; Flip Side Productions, Inc. v. Jam Productions, Ltd., 843 F.2d 1024, 1036 (7th Cir.1988); Cheek v. Doe, 828 F.2d 395, 397 (7th Cir.1987). Mars Steel and Continental Bank are the beneficiaries of what may have been a legal error in admitting the hearsay. Joyce & Kubasiak’s shallow motion ran up the costs, but the fact remains that the proponents of the settlement have dodged a bullet. Moreover, the costs of defending this appeal have been augmented through no fault of the appellants. This court set the case for reargument and called for further briefs. Costs of the reargument probably exceed the costs of the original briefing and argument — part of which appellees will receive under Rule 38 anyway. Counsel for the appellees should write off the remaining costs to the pro bono publico aspects of their work as members of our bar, just as counsel for the six amici curiae bar associations will have done. Hays sets the norm; today we see the exception.
The judgment is affirmed, with sanctions under Rule 38. Counsel for Mars Steel and Continental Bank have 15 days to file statements with the court showing the legal fees reasonably incurred in preparing that *940portion of the first round of briefs addressed to the motion for a rule to show cause.