The defendant appeals from a judgment awarding damages to the plaintiff — a waiter upon one of its dining cars — for injuries suffered while in its service on February 16, 1943. The action was brought under the Federal Employers’ Liability Act, §§ 51-60, Title 45 U.S.C.A.; but the only question raised upon this appeal concerns the validity of two releases, dated August 23, 1943, by which the plaintiff released all claims against the defendant upon the payment of $600. The plaintiff had already executed a release for $150 in the same words on March 19, 1943, and one of the two releases of August 23, recited a payment of $750, the sum of the two payments; but, as that release plays no part in the result we shall ignore it and speak as though only the second release had been given in August. The plaintiff testified that he executed the first release after a talk with one, Brown, the defendant’s claim agent, who told him that the payment of $150 was only for the tips and wages which he had lost; and that, relying upon this representation, he did not read the release, but signed it as Brown told him to do. Between that time and July he made some efforts to work but still felt incapacitated; and towards the end of that month, or early in August, he went again to Brown asking for more money. They could not agree, and he left, saying to Brown: “Well, I will have to get somebody to get all my tips and everything, my salary, because that is what I am getting.” He then went to an attorney named Reich, who, after talking to Brown on the telephone, later brought to the plaintiff the release drawn on the defendant’s form and told him: “I was just to sign a receipt for the amount of money that I got for the time I was off, my tips also included, to go back to work and they won’t have anything against you * * * he had the word from the Pennsylvania that I will be taken care of.” On his redirect he somewhat amplified this. Reich had told him: “the $600 is just for my earnings and my tips, because that would be better such and such. He *759said he did not want to sue. He said that would be better, and the company wants you to sign. That is big money. If you do, they won’t have anything against you. Since you stay in the company, you have eligibility to be retired. You will have full retirement pay.” Relying upon this, and again not reading the release, he signed it and the defendant paid him $600. The plaintiff’s wife also testified that Reich had said that “the money was for his back pay and tips.”

This version of the transaction the defendant denied. It called Reich, who said that Brown, when Reich consulted him, agreed to pay $600 for a complete release; and that all this Reich explained to the plaintiff when the release was executed. The judge charged the jury that, if the plaintiff executed the release “without fraud or misrepresentation, and understanding what he was doing,” it bound him, but that if he “signed these papers as receipts for wages, if it was as his understanding that that was all he was signing, that he did not sign any general release, then, of course you take up the question of damages.” Again: “Was it represented to the plaintiff by his lawyer that the papers he signed on August 23, 1943, were releases of all claims or only for lost wages?” The defendant made several requests to charge, but in none of them did it suggest that the jury should find whether the plaintiff retained Reich to settle all claims he had against the railroad, or whether — as the plaintiff testified — Reich’s retainer was limited to collecting wages and tips.

The right of action here in suit was created by act of Congress, and it is abundantly settled that its interpretation is a matter of federal law and not governed by state decisions, even when it speaks in the words of the common-law. Chesapeake & Ohio R. Co. v. Kuhn, 284 U.S. 44, 52 S.Ct. 45, 76 L.Ed. 157. It would not inevitably follow that, after such a right had come into existence, the legal effect upon it of a transaction within a state — as here, of a release — was also to be treated as matter of federal law; conceivably, its fate might be left to the law of the state. However, as we read Garrett v. Moore-McCormack Co., 317 U.S. 239, 63 S.Ct. 246, 87 L.Ed. 239, this is not so. The right of action was there under the Jones Act, but the action had been brought in a state court, which had held that the burden of proof of establishing a release was governed by the law of Pennsylvania. This the Supreme Court denied, holding that the admiralty rule controlled; and it would seem to follow that the validity of the release at bar is to be decided by the common-law, to be gathered from the same sources which, before Erie Railroad Co. v. Tompkins, 304 U.S. 64, 58 S.Ct. 817, 82 L.Ed. 1188, 114 A.L.R. 1487, we used to employ in cases depending on diversity of citizenship.

Although the law was at one time otherwise, at least in this country (Friedlander v. Texas & Pacific R. Co., 130 U.S. 416, 9 S.Ct. 570, 32 L.Ed. 991), it is now settled both in the federal system (Gleason v. Seaboard Airline R. Co., 278 U.S. 349, 49 S.Ct. 161, 73 L.Ed. 415), and in England (Lloyd v. Grace, Smith & Co., [1912] A.C. 716), that an agent does not cease to be acting within the scope of his authority when he is engaged in a fraud upon a third person. That has probably always been the more generally accepted doctrine. Fifth Avenue Bank v. Forty-Second Street & G. St. Ferry R. Co., 137 N.Y. 231, 33 N. E. 378, 19 L.R.A. 331, 33 Am.St.Rep. 712; Ripon Knitting Works v. Railway Express Agency, 207 Wis. 452, 240 N.W. 840; Tollett v. Montgomery Real Estate & Insurance Co., 238 Ala. 617, 622, 193 So. 127; Restatement of Agency, § 262. We can see no distinction in principle between that situation and one in which the agent deceives, not the third person, but his principal. The reason in each case for holding the principal is that the third person has no means of knowing that the agent is acting beyond his authority, and it is a matter of entire indifference whether the agent adds deception of his principal to deception of the third person; for it is obviously true that for the agent consciously to exceed his powers is to deceive the third person. Hence we may assume in the case at bar that, if the plaintiff had retained Reich to settle any claim he might have against the defendant, the plaintiff would have been bound, if Reich had procured the execution of the release by deceiving him as to its contents.

On the other hand, if the plaintiff retained Reich merely to collect his wages and tips, as to which he had failed to come to a satisfactory agreement with Brown in July or early August, the release was invalid; for, whatever may be the law in England, it is well settled in this *760country that an attorney has no implied authority to compromise a claim. Holker v. Parker, 7 Cranch 436, 452, 453, 3 L.Ed. 396; United States v. Beebe, 180 U.S. 343, 351, 352, 21 S.Ct. 371, 45 L.Ed. 563; Glover v. Bradley, 4 Cir., 233 F. 721, Ann.Cas.1917A, 921; McFarland v. Curtin, 4 Cir., 233 F. 728; Barber-Colman Co. v. Magnano Corp., 1 Cir, 299 F. 401; Jacob v. City of New York, 2 Cir, 119 F.2d 800 (reversed as to a different cause of action in Jacob v. City of New York, 315 U.S. 752, 62 S.Ct. 854, 86 L.Ed. 1166); Countryman v. Breen, 241 App.Div. 392, 394, 271 N.Y.S. 744, affirmed 268 N.Y. 643, 198 N.E. 536. The release would still be invalid even though the plaintiff signed it without reading it, for he would have been justified in relying upon what his lawyer told him of the contents. The theory upon which a document binds one who signs it, but who does not read it, is that either he accepts it whatever may be its contents, or that he has been careless in choosing his informant. Foster v. Mackinnon, L.R. 4 C.P. 704; Pimpinello v. Swift & Co., 253 N.Y. 159, 170 N.E. 530; Chapman v. Rose, 56 N.Y. 137, 15 Am.Rep. 401; Williston, § 95A. In the case at bar, whatever we may think about the correctness of the verdict on the facts, the jury could have interpreted the plaintiff’s testimony to mean that he only told Reich to collect his wages and tips; which incidentally was consistent with his parting words to Brown. Indeed, it was not really possible to believe the words which he put in Reich’s mouth, unless he had retained him for that limited purpose. At any rate, if the defendant had meant to raise the question whether Reich had been broadly commissioned, it should have done so by a proper request. This it did not do; and the result is that upon this, the turning point in the case — the extent of Reich’s authority — the only question that remains upon this appeal is of the sufficiency of the evidence to support a verdict. It is true that the plaintiff did not offer to return the money received in March and in August, 1943; since however the defendant did not raise this point either in its answer, at the trial, or in this court, we have not considered it.

Judgment affirmed.

(concurring).

1. .Plaintiff, as a result of a railroad accident which occurred while he was working as an employee of the Pennsylvania Railroad Company, suffered personal injuries which turned out to be so serious that the jury returned a verdict in his favor for $7,500, which the Railroad Company does not contest — except on one ground, i. e, that his claim was barred by a release he gave the company on payment to him of one-tenth that sum, or $750. That smaller sum represents merely the approximate amount of his lost earnings up to the date of the release. When he signed the release, he could not read it because of the effects of the accident, and without negligence on his part — since he relied on his own lawyer who mis-informed him — he understood that it purported to be only a receipt for payment of those lost earnings.

Judge HAND says (and I entirely agree) that the evidence sufficiently shows, that the lawyer acted beyond his authority-Accordingly, it is as if a non-lawyer, carefully selected by plaintiff,. had erroneously interpreted the release to him. The case thus comes within the category, described by Williston,1 of non-negligent unilateral mistakes preventing the formation, of valid contracts. Accepting Williston’s analysis,. Judge HAND’S rationale seems to me to be impregnable.

But I am not content to rest the decision on that analysis, because I think that that analysis leads to needless complexities which will confront us in future cases. The Supreme Court recently, in a case (cited by Judge HAND) relating to a release by a seaman, Garrett v. Moore-Mc-Cormack Co., 317 U.S. 239, at page 248, 63 S.Ct. 246, 87 L.Ed. 239, note 17, has. broadly hinted that the courts should treat non-maritime employees, with respect to. releases of personal injury claims, just as they treat seamen. I think we should take that hint, and, in doing so, should reject many of the finespun distinctions made by Williston and expressed in the Restatement of Contracts. Since I believe that not only is an important social policy involved but also that a good opportunity offers itself to uncomplicate an excessively complicated! set of legal rules, I shall state, in some detail, my reasons for this conclusion.

2. In the early days of this century a struggle went on between the respective proponents of two theories of contracts, (a) the “actual intent” theory- — or “meeting of the minds” or “will” theory — and *761(t>) the so-called “objective” theory.2 Without doubt, the first theory had been carried too far: Once a contract has been validly made, the courts attach legal consequences to the relation created by the contract, consequences of which the parties usually never dreamed — as, for instance, where situations arise which the parties had not contemplated.3 As to such matters, the “actual intent” theory induced much fictional discourse which imputed to the parties intentions they plainly did not have.

But the objectivists also went too far. They tried (1) to treat virtually all the varieties of contractual arrangements in the same way, and (2), as to all contracts in all their phases, to exclude, as legally irrelevant, consideration of the actual intention of the parties or either of them, as distinguished from the outward manifestation of that intention. The objectivists transferred from the field of torts that stubborn anti-subjectivist, the “reasonable man”;4 so that, in part at least, advocacy of the “objective” standard in contracts appears to have represented a desire for legal symmetry, legal uniformity, a desire seemingly prompted by aesthetic impulses.5 Whether (thanks to the “subjectivity” of the jurymen’s reactions and other factors) the objectivists’ formula, in its practical workings, could yield much actual objectivity, certainty, and uniformity may well be doubted.6 At any rate, the sponsors of *762complete “objectivity” in contracts 7 largely won out in the wider generalizations of the Restatement of Contracts 8 and in some judicial pronouncements.9

Influenced by their passion for excessive simplicity and uniformity, many objectiv-ists have failed to give adequate special consideration to releases of claims for personal injuries, and especially to such releases by employees to their employers. Williston, the leader of the objectivists, insists that, as to all contracts, without differentiation, the objective theory is essential because “founded upon the fundamental principle of the security of business transactions”.10

He goes to great lengths to maintain this theory, using a! variety of' rather desperate verbal distinctions to that end, Thus he distinguishes between (1) a unilateral non-negligent mistake in executing an instrument (i. e., a mistake of that character in signing an instrument of one kind believing it to be of another kind) and (2) a similar sort of mistake as to the meaning of a contract which one intended to make.11 The former, he says, renders the contract “void”;12 the latter does not prevent the formation of a valid contract. Yet in both instances “the fundamental principle of the security of business transactions” is equally at stake, for there has been the same “disappointment of well-founded expectations.”13 More than that, Williston concedes that a mistaken idea of one party as to the meaning of a valid contract (Willis-ton’s second category) “may, under certain circumstances," be ground for relief from enforcement of the contract.” But he asserts that (a) such a contract is not “void” but “voidable,” and (b) that the granting of such relief is no exception to the objective theory, because this relief “is in its origin equitable,” and “equity” does not deny the formation of a valid contract but merely acts “by subsequently * * * rescinding” it.14 His differentiation, moreover, of “void” and “voidable” has little if any practical significance: He says that a “voidable” contract will be binding unless the mistaken party sets up the mistake as a defense;15 but the same is obvi*763•ously true of agreements which (because of unilateral mistakes affecting their “validity”) he describes as “wholly void.” 16

It is little wonder that a considerable number of competent legal scholars have criticized the extent to which the objective theory, under Williston’s influence, was carried in the Restatement of Contracts.17 One of them, Whittier, says that the theory, in its application to the formation of contracts, is a generalization from the exceptional cases; he points out that the theory of “actual mutual assent” explains the great majority of the decisions, so that it would be better, he believes, to adhere to it, creating an exception for the relatively few instances where one party has reasonably relied on negligent use of words by the other. “Why not,” asks Whittier, say that actual assent communicated is the basis of ‘mutual assent’ except where there is careless misleading which induces a reasonable belief in assent ?”18 There may be much in that notion: Willis-ton admits that “the law generally is expressed in terms of subjective assent, rather than of objective expressions * * *”;18a and that “a doctrine which permits the rescission of a contract on account of a unilateral mistake approaches nearly to a contradiction of the objective theory * * * ” 19 As able a judge as Cuthbert Pound said, not long ago, “The meeting of minds which establishes contractual relations must be shown.” 20

Another critic21 suggests that, in general, Williston, because he did not searchingly inquire into the practical results of *764many of his formulations, assumed, unwarrantably, without proof, that those results must invariably have a general social value, although (as Williston admits as to the objective theory) they are “frequently harsh.”21a

In other realms of thought, attempted over-simplification has yielded complexities in practice.22 So here, as appears from the following.

Fortunately, most judges are too common-sensible to allow, for long, a passion for aesthetic elegance, or for the appearance of an abstract consistency, to bring about obviously unjust results.23 Accordingly, courts not infrequently have depart*765ed from the objective theory when necessary to avoid what they have considered an unfair decision against a person who, for a small sum, signed a release without understanding either the seriousness of his injury or the import of the words of the release, provided (1) he was not “negligent” and (2) the other party (the releasee) had not, in reliance on the release, importantly “changed his position.”24 Some courts, in some of the mistake cases, frankly abandon the “objective” test, saying boldly that a non-negligent unilateral mistake justifies cancellation or rescission of a contract.25 As New York, a lively center of commerce, at least to some extent allows relief for such unilateral mistakes,26 it should be obvious that, contrary to Williston & Co., any deviation from the objective theory is not fatal to the functioning of business.

Some courts, however, escape marring the verbal symmetry of the objective theory, while actually abandoning it, thus: They say that a mistake by one party about a striking fact (sometimes called “palpable”) 27 must be deemed to have been known to the other party, that, even if the evidence fails to show that he knew it, yet he had “reason to know it” and is therefore to be treated as if he did; so that there results, by this device, which comes close to a fiction, a “mutual mistake of fact.”28 This court, in pre-Erie-Tomp-*766kins days, in effect .adopted that rule in a case where the plaintiff, before executing the release, had consulted her own physician.28a That thesis has been utilized especially when an employee has given a release to his employer29 of all claims for an injury in consideration of a sum which approximates his lost wages (or his lost wages and medical expenses) and no more. Many courts have said that, on such a state of facts, it is impossible to believe that the releasor and releasee had in mind serious consequences of the injury which became apparent after the release was given, and therefore there was a “mutual mistake of fact.” 30

Williston, realizing the desire of the courts to escape the objective theory in such cases, describes them as follows: “Thus, where a release is given by one injured in an accident and'more serious injuries develop than were supposed to exist at the time of settlement, it is a question of fact whether the parties assumed as a basis of the release the known injuries, or whether the intent was to make a compromise for whatever injuries from the accident might exist whether known or not. On a fair interpretation not only of the

language of the instrument, but of the intention of the parties, the latter position is more likely, but presumably out of tenderness for injured plaintiffs some courts have gone very far in accordance with the former possibility.”31 (In the instant case, plaintiff’s lawyer testified that he had received the report of a physician to the effect that plaintiff’s injury was not serious. The evidence was, then, probably sufficient to bring this case within the mutual mistake rulé; but I think the trial judge’s instructions were such as not to leave that issue to the jury.)

Two approaches have been suggested which diverge from that of Williston and the Restatement but which perhaps come closer to the realities of business experience. (1) The first utilizes the concept of an “assumption of risk”: The parties to a contract, it is said, are presumed to undertake the risk that the facts upon the basis of which they entered into the contract might, within a certain margin, prove to be non-existent; accordingly, one who is mistaken about any such fact should not, absent a deliberate assumption by him of that risk, be held for more than the actual expenses caused by his conduct. Other*767wise, the other party will receive a windfall to which he is not entitled.32 (2) The second suggestion runs thus: Business is conducted on the assumption that men who bargain are fully informed as to all vital facts about the transactions in which they engage; a contract based upon a mistake as to any such fact as would have deterred either of the parties from making it, had he known that fact, should therefore be set aside in order to prevent unjust enrichment to him who made the mistake; the other party, on this suggestion also, is entitled to no more than his actual expenses.33 Each of those suggestions may result in unfairness, if the other party reasonably believing that he has made a binding contract, has lost the benefit of other specific bargains available at that time but no longer open to him.34 But any such possibility of unfairness will seldom, if ever, exist in the case of a release of liability for personal injury whatever the nature of the mistake (i. e., whether it fits into one or the other of Williston’s categories).

In short, the “security of business transactions” does not require a uniform answer to the question when and to what extent the non-ncgligcnt use of words should give rise to rights in one who has reasonably relied on them. That the answer should be favorable to the reliar when the words are used in certain kinds of contracts, does not mean that it should also be when they are used in a release of a claim for personal injury; and there may be still further reasons for an unfavorable answer when the claim is by an employee against his employer.

In all likelihood, it is because the courts have sensed the differentiated character of releases of personal injury claims that the “modern trend,” as Wigmore describes it, “is to * * * develop a special doctrine * * * for that class of cases, liberally relieving the party who signed the release.” 35 Surely much is to be said for that liberality, especially in a case where an employee has given a release of personal injury liability, without the fullest comprehension of what he was about, for a relatively small sum which turns out to he wholly inadequate.

In the admiralty cases, such relief has long been accorded seamen; the courts, calling them “wards of admiralty,” have regarded them, in many of their dealings with their employers, as necessitous persons, under strong economic pressures, who, because of their helplessness, are to be protected from hard bargains,36 just as “equity,” for similar reasons, protects mortgagors and beneficiaries of spendthrift trusts. The usual non-maritime employees, because they .are under similar economic pressures, are no less helpless in their trafficking with their employers. It can truthfully be said of them what the admiralty decisions say of seamen: “They are,” remarked Mr. Justice Story, “considered as tplaced under the dominion and influence of men, who have naturally acquired a mastery over them; and as they have little of the foresight and caution belonging to persons trained in other pursuits of life,37 the most rigid scrutiny is instituted into the terms of every contract, in which they engage. If there is any undue inequality in the terms, any disproportion in the bargain, any sacrifice of rights of one side which are not compensated by extraordinary benefits on the other, the judicial interpretation of the transaction is that the bargain is unjust and unreasonable, that advantage has been taken of the situation of the weaker party, and that pro tanto the bargain ought to be set aside as inequitable.” 38 To men like plaintiff here, the following comment about seamen fully applies : “They are * * * placed too much in the power of the owners [i. e., employ*768ers] to be able to negotiate with them on equal terms.” 39

It is not surprising, then, that many courts — although without such direct expressions as those which adorn the seaman cases — have in fact in the release cases-manifested, although obliquely, a not dissimilar guardianship of employees of large corporations. As. already noted, the Supreme Court recently gave a broad hint that th.e admiralty doctrine is not as exceptional as is sometimes supposed; see Garrett v. Moore-McCormack Co., 317 U.S. 239, 248, 63 S.Ct. 246, 87 L.Ed. 239; note 17.40 For reasons previously indicated, I think we should take that hint. It seems to me that the timq has come to give up the elaboration of distinctions found in the judicial opinions relieving non-admiralty employees of their releases.41 I believe that the courts should now say forthrightly that the judiciary regards the ordinary employee as one'who needs and will receive the special protection of the courts when, for a small consideration, he has given a release after an injury. As Mr. Justice Holmes often urged, when an important issue of social policy arises, it should be candidly, not evasively, articulated.42 In other contexts, the courts have openly acknowledged that the economic inequality between the ordinary employer and the ordinary individual employee usually means the absence of “free bargaining.” 43 I think the courts should do so in these employee release cases.44 And the federal *769courts, I think, should so hold in respect to liability pursuant to the Federal Employers’ Liability Act.45 I think, therefore, that we should treat the plaintiff here as we would if he were a seaman.46

Such a ruling will not produce legal uncertainty but will promote certainty- — as anyone can see who reads the large number of cases in this field, with their numerous intricate methods of getting around the objective theory.47 Such a ruling would simply do directly what many courts have been doing indirectly. It is fairly clear that they have felt, although they have not said, that employers should not, by such releases, rid themselves of obligations to injured employees, obligations which society at large will bear — either, by taxes, through the government or, by donations, through private charitable organizations.

3. The Pennsylvania Railroad Company warns us that, if a release given by an employee, advised by his own lawyer, is disregarded in a case like this, in the absence of fraud on the part of the employer, then employers will never hereafter settle with their employees who, to their grave disadvantage, will always in the future be forced to sue even for minor personal injury claims. That is a glib prediction based upon no evidence and intended to frighten the court. Sometimes judges have been persuaded by such prophecies which later events have shown to have been unfounded. So Choate, in Pollock v. Farmers’ Loan & Trust Co., 1895, 157 U.S. 429, 532, 15 S.Ct. 673, 39 L.Ed. 759, seemingly alarmed the majority of the Court by his forecast that a federal income tax would usher in a communist regime in this country. And it is well to recall Lord Abinger’s dire prediction when in 1837 he enunciated the fellow-servant rule which the Employers Liability Act has wiped out: “If the master be liable to the servant in this action, the principle of liability will be found to carry us to an alarming extent * * * The inconvenience, not to say the absurdity of these consequences, afford, a sufficient *770argument against the application of this principle to the present case * * * In fact, to allow this action to prevail would be an encouragement to the “servant to omit that diligence and caution which he is in duty bound to exercise on the behalf of his masters, to protect him against the misconduct of others who serve him, and which diligence and caution, while they protect the master, are a much better security against any injury the servant may sustain by the negligence of others engaged under the same master, than any recourse against his master for damages could possibly afford.” 48 Certainly, that prophecy went astray.

In New York, the rule as to releases is precisely that to which the Pennsylvania Railroad here objects; yet I venture to guess that thousands of settlements yearly are made in New York by employers who take the risk that, on a proper showing, the releases will be judicially disregarded49 Where the amount paid in settlement is relatively small, very likely most employers are willing to take such a risk.

One need not be highly imaginative or unduly cynical 50 to surmise that the Pennsylvania Railroad, in seeking here to be rid of a liability of $7,50051 to this employee for a payment of $750, is not too strongly motivated by a philanthropic regard for its other employees.

(dissenting).

As I understand Judge HAND’S opinion he differentiates between the case where a lawyer is engaged by an injured plaintiff to settle any and all claims arising out of the accident and the case where the lawyer is engaged to settle only a claim for loss of wages due to the accident; in the former case a release signed by the plaintiff will bind him even though the lawyer represented to him that it was only a receipt for wages, in the latter it will not. So far I agree with my brother. But I cannot agree to affirmance of the judgment on the ground that the jury found that the plaintiff engaged Mr. Reich under a limited retainer. I think that affirmance on this ground rests upon an issue not argued by the parties, not submitted to the jury and not supported by the evidence. It appears from the plaintiff’s own testimony, as well as from Mr. Reich’s, that after the latter was retained he caused the plaintiff to be examined by a doctor to ascertain the extent of his injuries. This seems wholly inconsistent with the present theory that Mr. Reich was to present to the defendant only a claim for wages and tips. When an injured employee engages his own attorney, and the latter has a doctor investigate the extent of the injuries, and then makes a complete settlement and gets his client to sign a release in full, that ought to end the matter, unless the employer practiced fraud on the attorney. In my opinion the judgment should be reversed and the cause remanded for a new trial in which the issue of what claims the attorney was authorized to settle shall be clearly submitted to the jury.