Agency

by Various

 

HARVARD

LAW REVIEW

---

---

---

---

AGENCY.

---

Note.—These papers are two lectures delivered by me when I was a professor at the Law School of Harvard College, and now are printed at the request of the Editors of this Review. My other occupations have made it impossible for me to revise what I wrote more than eight years ago. I can only say that I studied the subject carefully then, and have seen no reason to change my opinions.

I PROPOSE in these lectures to study the theory of agency at common law, to the end that it may be understood upon evidence, and not merely by conjecture, and that the value of its principles may be weighed intelligently. I first shall endeavor to show why agency is a proper title in the law. I then shall give some general reasons for believing that the series of anomalies or departures from general rule which are seen wherever agency makes its appearance must be explained by some cause not manifest to common sense alone; that this cause is, in fact, the survival from ancient times of doctrines which in their earlier form embodied certain rights and liabilities of heads of families based on substantive grounds which have disappeared long since, and that in modern days these doctrines have been generalized into a fiction, which, although nothing in the world but a form of words, has reacted upon the law and has tended to carry its anomalies still farther. That fiction is, of course, that, within the scope of the agency, principal and agent are one. I next shall examine the early law of England upon every branch of the subject,—tort, contract, possession, ratification,—and show the working of survival or fiction in each. If I do not succeed in reducing the law of all these branches to a common term, I shall try to show that at least they all equally depend upon fiction for their present existence. I shall prove incidentally that agency in its narrower sense presents only a special application of the law of master and servant, and that the peculiar doctrines of both are traceable to a common source. Finally I shall give my reasons for thinking that the whole outline of the law is the resultant of a conflict at every point between logic and good sense—the one striving to work fiction out to consistent results, the other restraining and at last overcoming that effort when the results become too manifestly unjust.

A part of my task has been performed and my general view indicated in my book on the Common Law. It remains to discuss the matter systematically and in detail, giving due weight to the many difficulties or objections which are met with in the process.

My subject extends to the whole relation of master and servant—it is not confined to any one branch; so that when I choose the title "Agency," I do not use it in the strict sense just referred to, but as embracing everything of which I intend to treat.

The first question proposed is why agency is a proper title in the law. That is to say, Does agency bring into operation any new and distinct rules of law? do the facts which constitute agency have attached to them legal effects which are peculiar to it, or is the agency only a dramatic situation to which principles of larger scope are applied? And if agency has rules of its own incapable of being further generalized, what are they?

If the law went no farther than to declare a man liable for the consequences of acts specifically commanded by him with knowledge of circumstances under which those consequences were the natural results of those acts, it would need no explanation and introduce no new principle. There may have been some difficulty in arriving at this conclusion when the intervening agent was a free person and himself responsible. Speaking without special investigation, I do not remember any case in early law in which one could charge himself thus in contract or even in tort. Taking the allied case of joint trespassers, although it long has been settled that each wrong-doer is liable for the entire damages, the objection that "the battery of one cannot be the battery of the other" prevailed as late as James I.^[1] It is very possible that liability even for the commanded acts of a free person first appeared as an extension of the liability of an owner for similar acts by his slave.

But however this may be, it is plain good sense to hold people answerable for wrongs which they have intentionally brought to pass, and to recognize that it is just as possible to bring wrongs to pass through free human agents as through slaves, animals, or natural forces. This is the true scope and meaning of "Qui facit per alium facit per se," and the English law has recognized that maxim as far back as it is worth while to follow it.^[2] So it is only applying the general theory of tort to hold a man liable if he commands an act of which the natural consequence, under the circumstances known to him, is harm to his neighbor, although he has forbidden the harm. If a trespass results, it is as much the trespass of the principal as if it were the natural, though unwished-for, effect of a train of physical causes.^[3] In such cases there is nothing peculiar to master and servant; similar principles have been applied where independent contractors were employed.^[4]

No additional explanation is needed for the case of a contract specifically commanded. A difficulty has been raised concerning cases where the agent has a discretion as to the terms of the contract, and it has been called "absurd to maintain that a contract which in its exact shape emanates exclusively from a particular person is not the contract of such person [i. e., the agent], but is the contract of another."^[5] But I venture to think that the absurdity is the other way, and that there is no need of any more complex machinery in such a case than where the agent is a mere messenger to express terms settled by his principal in every detail. Suppose that the principal agrees to buy a horse at a price to be fixed by another. The principal makes the contract, not the referee who settles the price. If the agreement is communicated by messenger, it makes no difference. If the messenger is himself the referee, the case is still the same. But that is the case of an agent with discretionary powers, no matter how large they may be. So far as he expresses his principal's assent to be bound to terms to be fixed by the agent, he is a mere messenger; in fixing the terms he is a stranger to the contract, which stands on the same footing as if it had been made before his personal function began. The agent is simply a voice affording the marks provided by the principal's own expression of what he undertakes. Suppose a wager determined in amount as well as event by the spinning of a teetotum, and to be off if numbers are turned up outside certain limits; is it the contract of the teetotum?

If agency is a proper title of our corpus juris, its peculiarities must be sought in doctrines that go farther than any yet mentioned. Such doctrines are to be found in each of the great departments of the law. In tort, masters are held answerable for conduct on the part of their servants, which they not only have not authorized, but have forbidden. In contract, an undisclosed principal may bind or may be bound to another, who did not know of his very existence at the time he made the contract. By a few words of ratification a man may make a trespass or a contract his own in which he had no part in fact. The possession of a tangible object may be attributed to him although he never saw it, and may be denied to another who has it under his actual custody or control. The existence of these rules is what makes agency a proper title in the law.

I do not mean to assume in advance that these rules have a common origin because they are clustered round the same subject. It would be possible to suggest separate reasons for each, and going farther still, to argue that each was no more than an application, even though a misapplication, of general principles.

Thus, in torts it is sometimes said that the liability of the master is "in effect for employing a careless servant," repeating the reason offered by the pseudo-philosophy of the Roman jurists for an exceptional rule introduced by the prætor on grounds of public policy.^[6] This reason is shown to be unsound by the single fact that no amount of care in selection will exonerate the master;^[7] but still it might be argued that, whether right or wrong, this or some other notion of policy had led to the first of the rules which I selected as peculiar, and that at most the liability of a master for his servant's torts is only a mistaken conclusion from the general theory of tort.

Then with regard to undisclosed principals in contract, it might be said that it was no hardship to hold a man bound who had commanded his servant to bind him. And as to the other and more difficult half of the doctrine, the right of an undisclosed principal to sue, it might be observed that it was first asserted in cases of debt,^[8] where the principal's goods were the consideration of the liability, and that the notion thus started was afterwards extended to other cases of simple contract. Whether the objections to the analogy and to the whole rule were duly considered or not, it might be urged, there is no connection other than a purely dramatic one between the law of agency in torts and in contracts, or between the fact of agency and the rule, and here, as there, nothing more is to be found than a possibly wrong conclusion from the general postulates of the department of law concerned.

Ratification, again, as admitted by us, the argument would continue, merely shows that the Roman maxim "ratihabitio mandato comparatur" has become imbedded in our law, as it has been from the time of Bracton.

Finally, the theory of possession through servants would be accounted for by the servant's admission of his master's present right to deal with the thing at will, and the absence of any claim or intent to assert a claim on his part, coupled with the presence of such a claim on the part of the master.

But the foregoing reasoning is wholly inadequate to justify the various doctrines mentioned, as I have shown in part and as I shall prove in detail hereafter. And assuming the inadequacy to be proved, it cannot but strike one as strange that there should run through all branches of the law a tendency to err in the same direction. If, as soon as the relation of master and servant comes in, we find the limits of liability for, or gain by, others' acts enlarged beyond the scope of the reasons offered or of any general theory, we not only have good ground for treating that relation separately, but we fairly may suspect that it is a cause as well as a concomitant of the observed effects.

Looking at the whole matter analytically it is easy to see that if the law did identify agents with principals, so far as that identification was carried the principal would have the burden and the benefit of his agent's torts, contracts, or possession. So, framing a historical hypothesis, if the starting-point of the modern law is the patria potestas, a little study will show that the fiction of identity is the natural growth from such a germ.

There is an antecedent probability that the patria potestas has exerted an influence at least upon existing rules. I have endeavored to prove elsewhere that the unlimited liability of an owner for the torts of his slave grew out of what had been merely a privilege of buying him off from a surrender to the vengeance of the offended party, in both the early Roman and the early German law. I have shown, also, how the unlimited liability thus established was extended by the prætor in certain cases to the misconduct of free servants.^[9] Of course it is unlikely that the doctrines of our two parent systems should have been without effect upon their ofispring, the common law.

The Roman law, it is true, developed no such universal doctrines of agency as have been worked out in England. Only innkeepers and shipowners (nautae, caupones, stabularii) were made answerable for the misconduct of their free servants by the prætor's edict. It was not generally possible to acquire rights or to incur obligations through the acts of free persons.^[10] But, so far as rights of property, possession,^[11] or contract^[12] could be acquired through others not slaves, the law undoubtedly started from slavery and the patria potestas.

It will be easy to see how this tended toward a fictitious identification of agent with principal, although within the limits to which it confined agency the Roman law had little need and made little use of the fiction. Ulpian says that the act of the family cannot be called the act of the paterfamilias unless it is done by his wish.^[13] But as all the family rights and obligations were simply attributes of the persona of the family head, the summary expression for the members of the family as means of loss or gain would be that they sustained that persona pro hac vice. For that purpose they were one with the paterfamilias. Justinian's Institutes tell us that the right of a slave to receive a binding promise is derived ex persona domini.^[14] And with regard to free agents, the commentators said that in such instances two persons were feigned to be one.^[15]

Such a formula, of course, is only derivative. The fiction is merely a convenient way of expressing rules which were arrived at on other grounds. The Roman prætor did not make innkeepers answerable for their servants because "the act of the servant was the act of the master," any more than because they had been negligent in choosing them. He did so on substantive grounds of policy—because of the special confidence necessarily reposed in innkeepers. So when it was held that a slave's possession was his owner's possession, the practical fact of the master's power was at the bottom of the decision.^[16]

But when such a formula is adopted, it soon acquires an independent standing of its own. Instead of remaining only a short way of saying that when from policy the law makes a master responsible for his servant, or because of his power gives him the benefit of his slave's possession or contract, it treats him to that extent as the tort-feasor, possessor, or contractee, the formula becomes a reason in itself for making the master answerable and for giving him rights. If "the act of the servant is the act of the master," or master and servant are "considered as one person," then the master must pay for the act if it is wrongful, and has the advantage of it if it is right. And the mere habit of using these phrases, where the master is bound or benefited by his servant's act, makes it likely that other cases will be brought within the penumbra of the same thought on no more substantial ground than the way of thinking which the words have brought about.

I shall examine successively the English authorities with regard to agency in tort, contract, ratification, and possession. But some of those authorities are of equal importance to every branch of the proposed examination, and will prove in advance that the foregoing remarks are not merely hypothetical. I therefore begin with citations sufficient to establish that family headship was recognized as a factor in regal rights and duties; that this notion of headship was extended by analogy so as to cover the relation of a master to freemen filling a servile place for the time being, and that the relations thus embraced were generalized under the misleading fiction of identity.

The familia, Bracton says, embraces "those who are regarded in the light of serfs, such as, &c. So, too, as well freemen as serfs, and those over whom one has the power of command."^[17]

In West's Symboleography,^[18] a work which was published towards the beginning of the reign of James I., and which, though mainly a form book, gives several glimpses of far-reaching insight, we read as follows:—

"The person is he which either agreeth or offendeth, and beside him none other.

"And both may be bound either mediately, or immediately.

"Immediately, if he which is bound doe agree.

"Mediately, when if he, which by nature differeth from him, but not by law, whereby as by some bond he is fained to be all one person, doth contract, or offend, of which sort in some cases be those which be in our power, as a wife, a bondman, servant, a factor, an Attourney, or Procurator, exceeding their authority."

Here we see that the patria potestas is the substantive ground, that it is extended to cover free agents, who are not even domestic servants, and that it finds its formal expression in the fiction of identity.

So, at the beginning of the next reign, it was said that an action for fire, due to the negligence of a wife, or servant, lay "vers patrem familias."^[19] The extension of the liability, as shown by West, is sometimes expressed in later books by saying that it is not confined to cases where the party stands in the relation of paterfamilias to the wrong-doer;^[20] but this only means that the rule extends to other servants besides domestic servants, and admits the analogy or starting-point.^[21]

Every one is familiar with the fiction as applied to married women. The early law dealt with married women on the footing of servants. It called both wives and servants chattels.^[22] The wife was said to be in the nature of a servant,^[23] and husband and wife were only one person in law.^[24] So far was this identification carried, so far was the persona of the wife swallowed up in and made part of her husband's, that whereas, in general, assigns could not vouch upon a warranty unless they were expressly mentioned in it,^[25] a husband could always vouch upon a warranty made to his wife before marriage. By marriage, as was said in Simon Simeon's case, "it vested in the person of the husband." That is to say, although what actually happened was that the right to enforce a contract was transferred to a stranger, in theory of law there was no transfer, because the stranger had become the same person as the contractee.^[26]

Of course the identification between husband and wife, although by no means absolute, was far more complete than that between master and menial servant, just as in the latter case it went farther than in that of an agent employed for some particlar transaction. Even in the case of villeins, while the lord might take advantage of their possession or their title, he could not take advantage of contracts or warranties made to them.^[27] But the idea and its historical starting-point were the same throughout. When considering the later cases, the reader will remember that it is incontrovertibly established that a wife was on the footing of a servant, that the consequences of the relation were familiarly expressed in terms of the fiction of identity, and, therefore, that the applicability of this fiction to the domestic relations generally must have been well known to the courts long before the date of the principal decisions, which it wall be my task to interpret.

I now take up the liability of a master for the torts of his servant at common law. This has been supposed in England to have been manufactured out of the whole cloth, and introduced by the decision in Michael v. Alestree^[28] in the reign of Charles II. In view of the historical antecedents it would be very extraordinary if such a notion were correct. I venture to think that it is mistaken, and that the principle has gradually grown to its present form from beginnings of the earliest date. I also doubt whether Michael v. Alestree is an example for the principle in question. It rather seems to me a case in which the damage complained of was the natural consequence of the very acts commanded by the master, and which, therefore, as I have said above, needs no special or peculiar doctrine to account for it. It was an action on the case against master and servant; "for that the Defendants in Lincoln's-Inn Fields, a Place where People are always going to and fro about their Business, brought a Coach with two ungovernable Horses, & eux improvide incaute & absque debita consideratione ineptitudinis loci there drove them to make them tractable and fit them for a Coach; and the Horses, because of their Ferocity, being not to be managed, ran upon the Plaintiff, and ** wounded him: The master was absent," but both defendants were found guilty. "It was moved in Arrest of Judgment, That no Sciens is here laid of the Horses being unruly, nor any Negligence alledged, but e contra. That the Horses were ungovernable: Yet judgment was given for the Plaintiff, for it is alledged that it was improvide & absque debita consideratione ineptitudinis loci; and it shall be intended the Master Sent the servant to train the Horses there."^[29] In other words, although there was no negligence averred in the mode of driving the horses at the instant of the accident, but, e contra, that the horses were ungovernable, which was the scope of the defendant's objection, there was negligence in driving ungovernable horses for the purpose of breaking them in a public place, and that was averred, and was averred to have been done negligently. Furthermore, it was averred to have been done negligently by the defendant, which was a sufficient allegation on its face, and would be supported by proof that the defendant, knowing the character of the horses, ordered his servant to break them in a public resort. Indeed, the very character of the command (to break horses) imports sufficient knowledge; and when a command is given to do the specified act complained of, it always may be laid as the act of the party giving the order.^[30]

When I come to investigate the true history of this part of the law, notwithstanding the likelihood which I have pointed out that it was a continuation and development of what I have traced in one or both of the parent systems, I must admit that I am met with a difficulty. Even in Bracton, who writes under the full influence of the Roman law, I have failed to find any passage which distinctly asserts the civil liability of masters for their servants' torts, apart from command or ratification. There is one text, to be sure, which seems corrupt as it stands and which could be amended by conjecture so as to assert it. But as the best manuscripts in Lincoln's Inn substantially confirm the printed reading, conjecture would be out of place.^[31]

On the other hand, I do find an institution which may or may not have been connected with the Anglo-Saxon laws touching the responsibility of masters, but which, at any rate, equally connects liability of a different sort with family headship.

At about the time of the Conquest, what was known as the Frithborh, or frankpledge, was either introduced or grew greatly in importance. Among other things, the master was made the pledge of his servants, to hand them over to justice or to pay the fine himself.

"Omnes qui servientes habent, eorum sint francplegii," was the requirement of William's laws. Bracton quotes the similar provisions of Edward the Confessor, and also says that in some counties a man is held to answer for the members of his family.^[32] This quasi-criminal liability of master for man is found as late as Edward II. alongside of the other rules of frankpledge, with which this discussion is not concerned. Fitzherbert's Abridgment^[33] reads as follows: "Note that if the servant (serviens) of any lord while in his service (in servicio suo existens) commits a felony and is convicted, although after the felony (the master) has not received him, he is to be amerced, and the reason is because he received him 'in bourgh.'" Bracton, in like manner, says that the master is bound "Emendare" for certain torts of his servant,^[34] meaning, as I take it, to pay a fine, not damages.

But true examples of the peculiar law of master and servant are to be found before Edward II. The maxim respondeat superior has been applied to the torts of inferior officers from the time of Edward I. to the present day. Thus that chapter of the Statute of Westminster the Second,^[35] which regulates distresses by sheriffs or bailiffs, makes the officer disregarding its provisions answerable, and then continues, "si nom habeat ballivus unde reddat reddat superior suus." So a later chapter of the same statute, after subjecting keepers of jails to an action of debt for escapes in certain cases, provides that if the keeper is not able to pay, his superior, who committed the custody of the jail to him, shall be answerable by the same writ.^[36] So, again, the eighteenth chapter of the Articuli super Chartas^[37] gives a writ of waste to wards, for waste done in their lands in the king's hands by escheators or sub-escheators, "against the escheator for his act, or the sub-escheator for his act (if he have whereof to answer), and if he have not, his master shall answer ('si respoigne son sovereign') by like pain concerning the damages, as is ordained by the statute for them that do waste in wardships." A case of the time of Edward II. interpreting the above statute concerning jailers is given in Fitzherbert's Abridgment,^[38] and later similar cases are referred to in Coke's Fourth Institute.^[39]

It may be objected that the foregoing cases are all statutory. But the same principle seems to have been applied apart from any statute except that which gave counties the power to elect coroners, to make the county of Kent answerable to the king for a coroner's default, as well as in other instances which will be mentioned later.^[40] Moreover, early statutes are as good evidence of prevailing legal conceptions as decisions are.

But again it may be objected that there were special grounds of public policy for requiring those who disposed of public offices of profit to appoint persons "for whom they will answer at their peril," in the words of another similar statute as to clerks in the King's Courts.^[41] It might be said with truth that the responsibility was greater than in the case of private servants, and it might be asked whether respondeat superior in its strict sense is not an independent principle which is rather to be deemed one of the causes of the modern law, than a branch from a common stem. It certainly has furnished us with one of the inadequate reasons which have been put forward for the law as it is,—that somebody must be held who is able to pay the damages.

The weight of the evidence seems to me to overcome these objections. I think it most probable that the liability for under-officers was a special application of conceptions drawn from the family and the power of the family head over his servants. Those conceptions were in existence, as I have shown. From a very early date, under-officers are called servants of their superior, as indeed it seems to be implied that they are, by the word "sovereign," or even "superior," in the statutes which have been cited. "Sovereign" is used as synonymous with master in Dyer.^[42] In the Y. B., 11 Edward IV. 1, pl. 1, it is said, "If I make a deputy, I am always officer, and he performs the office in my right and as my servant;" and from that day to this, not only has the same language been repeated,^[43] but, as I shall show, one of the chosen fields for the express use of the fiction of identity is the relation of superior and under-officer.

Under Edward III. it was held that if an abbot has a wardship, and a co-canon commits waste, the abbot shall be charged by it, "for that is adjudged the deed of the abbot."^[44] This expression appears to me not only to apply the rule respondeat superior beyond the case of public officers, but to adopt the fiction of identity as a mode of explaining the rule.

An earlier record of the same reign, although it turned on the laws of Oleron, shows that the King's Court would in some cases hold masters more strictly accountable for their servants' torts than is even now the case. A ship-master was held liable in trespass de bonis asportatis for goods wrongfully taken by the mariners, and it was said that he was answerable for all trespasses on board his ship.^[45]

A nearly contemporaneous statute is worth mentioning, although it perhaps is to be construed as referring to the fines which have been mentioned above, or to other forfeitures, and not to civil damages. It reads, "That no merchant nor other, of what condition that he be, shall lose or forfeit his goods nor merchandizes for the trespass and forfeiture of his servant, unless he do it by the commandment or procurement of his master, or that he hath offended in the office in which his master hath set him, or in other manner, that the master be holden to answer for the deed of his servant by the law-merchant, as elsewhere is used."^[46] The statute limits a previously existing liability, but leaves it open that the master still shall be holden to answer for the deed of his servant in certain cases, including those of the servant's offending in the office in which the master hath set him. It is dealing with merchants, to be sure, but is another evidence that the whole modern law is of ancient extraction.

It must be remembered, however, that the cases in which the modern doctrines could have been applied in the time of the Year Books were exceedingly few. The torts dealt with by the early law were almost invariably wilful. They were either prompted by actual malevolence, or at least were committed with full foresight of the ensuing damage.^[47] And as the judges from an early day were familiar with the distinction between acts done by a man on his own behalf and those done in the capacity of servant,^[48] it is obvious that they could not have held masters generally answerable for such torts unless they were prepared to go much beyond the point at which their successors have stopped.^[49] Apart from frauds^[50] and intentional trespasses against the master's will^[51] I only know of one other case in the Year Books which is important to this part of my subject. That, however, is very important. It is the case concerning fire,^[52] which was the precedent relied on by Lord Holt in deciding Turberville v. Stampe,^[53] which in its turn has been the starting-point of the later decisions on master and servant.^[54] I therefore shall state it at length.

Beaulieu sued Finglam, alleging that the defendant so negligently guarded his fire that for want of due guard of the same the plaintiff's houses and goods were burned. Markham [J.], A man is held to answer for the act of his servant or of his guest (hosteller) in such case; for if my servant or my guest puts a candle on a beam, (en un pariet,) and the candle falls in the straw, and burns all my house, and the house of my neighbor also, in this case I shall answer to my neighbor for the damage which he has, quod concedebatur per curiam. Horneby [of counsel], Then he should have had a writ, Quare domum suam ardebat vel exarsit. Hull [of counsel], That will be against all reason to put blame or default in a man where there is (il ad) none in him; for negligence of his servants cannot be called his feasance. Thirning [C. J.], If a man kills (tue ou occist) a man by misfortune he will forfeit his goods, and he must have his charter of pardon de grace. Ad quod Curia concordat. Markham, I shall answer to my neighbor for him who enters my house by my leave or my knowledge, or is entertained (hoste) by me or by my servant, if he does, or any one of them does such a thing, as with a candle (come de chandel), or other thing, by which feasance the house of my neighbor is burned; but if a man from outside my house, against my will, puts the fire in the straw of my house, or elsewhere, by which my house is burned and also the houses of my neighbor are burned, for that I shall not be held to answer to them, etc., for this cannot be said to be through ill-doing (male) on my part, but against my will. Horneby then said that the defendant would be ruined if this action were maintained against him. "Thirning [C. J.], What is that to us? It is better that he should be undone wholly, than that the law should be changed for him.^[55] Then they were at issue that the plaintiff's house was not burned by the defendant's fire."

The foregoing case affords some ground for the argument which was vainly pressed in Turberville v. Stampe, that the liability was confined to the house.^[56] Such a limit is not unsupported by analogy. By the old law a servant's custody of his master's things was said to be the master's possession within his house, but the servant's on a journey outside of it.^[57] So an innkeeper was liable for all goods within the inn, whether he had the custody of them or not.^[58] So in the case which has been mentioned above, a master was said to be responsible for the acts of his servants on board ship. It will be noticed also that the responsibility of a house-holder seems to be extended to his guests. From that day to this there have been occasional glimpses of a tendency to regard guests as part of the familia for the purposes of the law.^[59] And in view of the fact that by earlier law if a guest was allowed to stop in the house three days, he was called hoghenehine or agenhine, that is, own hine or servant of the host, it may be thought that we have here an echo of the frithborh.^[60] But with whatever limits and for whatever occult causes, the responsibility of the head of the house for his servants was clearly recognized, and, it would seem, the identification of the two, notwithstanding a statement by counsel, as clear as ever has been made since, of the objections to the doctrine.

The later cases in the Year Books are of wilful wrongs, as I have said, and I now pass to the subsequent reports. Under Elizabeth a defendant justified taking sheep for toll under a usage to have toll of strangers' sheep driven through the vill by strangers, and if he were denied by such stranger driving them, to distrain them. The defendant alleged that the plaintiff, the owner of the sheep, was a stranger, but did not allege that the driver was. But the court sustained the plea, saying, "The driving of the servant is the driving of the master; and if he be a foreigner, that sufficeth."^[61]

I leave on one side certain cases which often have been cited for the proposition that a master is chargeable for his servant's torts, because they may be explained otherwise and make no mention of it.^[62]

The next evidence of the law to which I refer is the passage from West's Symboleography which was given in full at the outset, and which gives the modern doctrine of agency as well as the fiction of identity in their full development. There are two nearly contemporaneous cases in which unsuccessful attempts were made to hold masters liable for wilful wrongs of their servants, in one for a piracy,^[63] in the other for a fraud.^[64] They are interesting chiefly as showing that the doctrine under discussion was in the air, but that its limits were not definitely fixed. The former sought to carry the rule respondeat superior to the full extent of the early statutes and cases which have been referred to, and cited the Roman law for its application to public affairs. The latter cites Doctor and Student. West also, it will have been noticed, indicates Roman influence.

Omitting one or two cases on the liability of the servant, which will be mentioned shortly, I come once more to a line of authorities touching public officers. I have said that although there was a difference in the degree of responsibility, under-officers always have been said to be servants.

Under Charles II. this difference was recognized, but it was laid down that "the high sheriff and under-sheriff is one officer," and on that ground the sheriff was held chargeable.^[65] Lord Holt expressed the same thought: "What is done by the deputy is done by the principal, and it is the act of the principal," or, as it is put in the margin of the report, "Act of deputy may forfeit office of principal, because it is quasi his act."^[66] Later still, Blackstone repeats from the bench the language of Charles's day. "There is a difference between master and servant, but a sheriff and all his officers are considered in cases like this as one person." So his associate judge, Gould, "I consider [the under-sheriff's clerk] as standing in the place of, and representing the very persons of ... the sheriffs themselves."^[67] Again, the same idea is stated by Lord Mansfield: "For all civil purposes the act of the sheriffs bailiff is the act of the sheriff."^[68] The distinction taken above by Blackstone did not prevent his saying in his Commentaries that under-officers are servants of the sheriff;^[69] and in Woodgate v. Knatchbull,^[70] Ashurst, J., after citing the words of Lord Mansfield, adds, "This holds, indeed, in most instances with regard to servants in general;" and Blackstone says the same thing in a passage to be quoted hereafter.

Having thus followed down the fiction of identity with regard to one class of servants, I must now return once more to Lord Holt's time. In Boson v. Sandford,^[71] Eyres, J., says that the master of a ship is no more than a servant, "the power which he hath is by the civil law, Hob. 111, and it is plain the act or default of the servant shall charge the owner." Again, in Turberville v. Stampe,^[72] Lord Holt, after beginning according to the Roman law that "if my servant throws dirt into the highway I am indictable," continues, "So in this case, if the defendant's servant kindled the fire in the way of husbandry and proper for his employment, though he had no express command of his master, yet the master shall be liable to an action for damages done to another by the fire; for it shall be intended, that the servant had authority from his master, it being for his master's benefit." This is the first of a series of cases decided by Lord Holt^[73] which are the usual starting-point of modern decisions, and it will be found to be the chief authority relied on by cases which have become leading in their turn.^[74] It therefore is interesting to note that it only applied the principles of Beaulieu v. Finglam, in the Year Book 2 Henry IV., to a fire outside the house, that the illustration taken from the Roman law shows that Lord Holt was thinking of the responsibility of a paterfamilias, and that in another case within three years^[75] he made use of the fiction of identity.

I may add, by way of confirmation, that Blackstone, in his Commentaries, after comparing the liability of the master who "hath the superintendence and charge of all his household" if any of his family cast anything out of his house into the street, with that of the Roman paterfamilias,^[76] further observes that the "master may frequently be answerable for his servant's misbehavior, but never can shelter himself from punishment by laying the blame on his agent. The reason of this is still uniform and the same; that the wrong done by the servant is looked upon in law as the wrong of the master himself."^[77]

There is another line of cases which affords striking and independent evidence that the law of master and servant is a survival from slavery or other institution of like effect for the present purpose, and that the identification of the two parties was carried out in some cases to its logical result. If a servant, although a freeman, was treated for the purposes of the relation as if he were a slave who only sustained the persona of his master, it followed that when the master was liable, the servant was not. There seems to have been a willingness at one time to accept the conclusion. It was said under James and Charles I. that the sheriff only was liable if an under-sheriff made a false return, "for the law doth not take notice of him."^[78] So it was held in the latter reign that case does not lie against husband and wife for negligently keeping their fire in their house, "because this action lies on the ... custom ... against patrem familias and not against a servant or a feme covert who is in the nature of a servant.^[79] So Rolle says that "if the servant of an innkeeper sells wine which is corrupt, knowing this, action of deceit lies not against the servant, for he did this only as servant."^[80] So as to an attorney maliciously acting in a case where he knew there was no cause of action. "For that what he does is only as servant to another, and in the way of his calling and profession."^[81]

Later this was cut down by Lord Holt to this rule that a servant is not liable for a neglect (i.e., a nonfeasance), "for they must consider him only as a servant;" "but for a misfeasance an action will lie against a servant or deputy, but not quatenus a deputy or servant, but as a wrong-doer."^[82] That is to say, although it is contrary to theory to allow a servant to be sued for conduct in his capacity as such, he cannot rid himself of his responsibility as a freeman, and may be sued as a free wrong doer. This, of course, is the law to-day.^[83] Yet as late as Blackstone's Commentaries it was said that "if a smith's servant lames a horse while he is shoeing him, an action hes against the master, and not against the servant.^[84]

I think I now have traced sufficiently the history of agency in torts. The evidence satisfies me that the common law has started from the patria potestas and the frithborh,—whether following or simply helped by the Roman law, it does not matter,—and that it has worked itself out to its limits through the formula of identity. It is true that liability for another as master or principal is not confined to family relations; but I have shown partly, and shall complete the proof later, that the whole doctrine has been worked out in terms of master and servant and on the analogies which those terms suggested.

O. W. Holmes, Jr.

[To be continued.]

1. Sampson v. Cranfield, 1 Bulstr. 157 (T. 9 Jac.).
2. In Tort: Y. B. 32 Ed. I. 318, 320 (Harwood); 22 Ass. pl. 43, fol. 94; 11 H. IV. 90, pl. 47; 9 H. VI. 53, pl. 37; 21 H. VI. 39; 4 Ed. IV. 36; Dr. & Stud., II. c. 42; Seaman & Browning's Case, 4 Leon. 123, pl. 249 (M. 31 Eliz.). Conveyance: Fitz. Abr. Annuitie, pl. 51 (H. 33 Ed. I.), where the maxim is quoted. Account: 4 Inst. 109.
3. Gregory v. Piper, 9 B. & C. 591. Cf. The Common Law, 53, 54, and Lect. 3 and 4.
4. Bower v. Peate, 1 Q. B. D. 321.
5. Thöl, Handelsrecht, sect. 70, cited in Wharton, Agency, sect. 6.
6. Parke, B., in Sharrod v. London & N. W. Ry. Co., 4 Exch. 580, 585 (1849); 1 Austin, Jurisprudence, Lect. 26, 3d ed., p. 513. Cf. The Common Law, 15, 16.
7. Dansey v. Richardson, 3 El. & Bl. 144, 161.
8. Scrimshire v. Alderton, 2 Strange, 1182 (H. 16 G. II.). Cf. Gurratt v. Cullum (T. 9 Anne, B. R.), stated in Scott v. Surman, Willes, 400, at p. 405 (H. 16 G. II.), and in Buller, N. P. 42.
9. The Common Law, 9, 15-20.
10. Inst. 2, 9, § 5; D. 44, 7, 11; D. 45, 1, 126, § 2.
11. Inst. 2, 9, esp. §§ 4, 5. Cf. D. 41, 1, 53.
12. Inst. 3, 17; D. 41, 1, 53; D. 45, 1, 38, § 17.
13. D. 43, 16, 1, §§ 11-13.
14. Inst. 3, 17, pr. 18, in the older editions.
15. D. 45, 1, 38, § 17, Elzevir ed. Gothofred, note 74. Cf. D. 44, 2, 4, note 17.
16. The Common Law, 228.
17. "Et etiam familiae appellatio eos complectitur qui loco servorum habentur, sicut sunt mercenarii et conductitii. Item tarn liberi quam servi, et quibus poterit imperari." Bract., fol. 171 b.
18. Lib I., Sect. 3, ad fin. "Of the Fact of Man."
19. Shelley & Barr's Case, 1 Roll. Abr. 2, pl. 7 (M. 1 Car. I.).
20. Bac. Abr., Master and Servant, K; Smith's Master & Servant, 3d ed., 260.
21. Laugher v. Pointer, 5 B. & C. 547, 554 (1826). Cf. Bush v. Steinman, 1 Bos. & P. 404 (1799).
22. Y. B. 19 H. VI. 31, pl. 59; 2 Roll. Abr. 546 (D).
23. 1 Roll. Abr. 2, pl. 7.
24. Dial, de Scaccario II., c. 18; Bract., fol. 429 b; Y. B. 22 H. VI. 38, pl. 6; Litt. §§ 168, 191; 3 Salk. 46; Com. Dig. Baron & Feme (D); 1 Bl. Comm. 442.
25. The Common Law, 375, n. 2, 401, n. 1.
26. Simon Simeon's Case, Y. B. 30 Ed. III. 14; s. c. ib. 6; 29 Ed. III. 48. I have seen no reason to change the views expressed in The Common Law, Lecture XI., to meet the suggestions of Prof. Ames in 3 Harv. Law Rev. 388, n. 6. Undoubtedly the letter of credit was known in the reign of Henry III. Royal letter, Hen. III. 315. But the modern theory of contract applied to letters of credit, in my opinion, was not the theory on which assigns got the benefit of a warranty. Norcross v. Ames, 140 Mass. 188.
27. Y. B. 22, Ass. pl. 27, fol. 93; Co. Lit. 117 a.
28. 2 Levinz, 172; s. c. 3 Keble, 650, 1 Ventris, 295 (T. 28 Car. II.).
29. 2 Lev. 172.
30. Sup. p. 346, n.
31. Bract., fol. 115 a.
32. "Tenebitur ille, in quibusdam partibus, de cujus fuerint familia et manupastu." Bract., fol. 124 b; i.e., for the persons under his patria potestas. LL. Gul. I. c. 52; LL Edw. Conf. c. 21 (al. 20).
33. Corone, pl. 428 (8 Ed. II. It. canc.).
34. Bract., fol. 158 b, 171 a, b, 172 b. Cf. Ducange, "Emenda."
35. St. 13 Ed. I., St. I, c. 2, § 3.
36. c 11, ad finem. "Et si custos gaole non habeat per quod justicietur vel unde solvat respondeat superior suus qui custodiam hujusmodi gaole sibi commisit per idem breve."
37. St. 28 Ed. I., c. 18.
38. Dette, pl. 172 (M. 11 Ed. II.).
39. 4 Inst. 114; "45 E. 3, 9, 10. Prior datife et removeable suffer eschape, respondeat superior. 14 E. 4. Pur insufficiency del bailie dun libertie respondeat dominus libertatis. Vid. 44 E. 3, 13; 50 E. 3, 5; 14 H. 4, 22; 11 H. 6, 52; 30 H. 6, 32."
40. See the writ of H. 14 Ed. III. ex parte Remem. Regis, rot. 9, in Scacc. in 4 Inst. 114, and less fully in 2 Inst. 175. "Et quia ipse coronator electus erat per comitatum juxta formam statuti, etc. ita quod in defectu ejusdem coronatoris totus comitatus ut elector et superior, etc. (tenetur), habeant regi respondere, praecip (praeceptum fuit) nunc vic' quod de terris et tenementis (hominum) hujusmodi totius comitatus in balliva sua fieri fac." etc. See the other references in 4 Inst. 114, and further Y. B. 49 Ed. III. 25, 26, pl. 3.
41. St. 2 H. VI., c. 10.
42. Alford v. Eglisfield, Dyer, 230 b, pl. 56. The passage will be cited later in dealing with factors. See also Y. B. 27 H. VIII. 24, pl. 3.
43. Parkes v. Mosse, Cro. Eliz. 181 (E. 32 Eliz.); Wheteley v. Stone, 2 Roll. Abr. 556, pl. 14; s. c. Hobart, 180; 1 Bl. Comm. 345, 346.
44. Y. B. 49 Ed. III. 25, 26, pl. 3.
45. Brevia Regis in Turr. London, T. 24 Ed. III., No. 45, Bristol, printed in Molloy, Book 2, ch. 3, § 16.
46. St. 27 Ed. III. St. 2, cap. 19.
47. The Common Law, 3, 4, 101-103. I do not mean as a matter of articulate theory, but as a natural result of the condition of things. As to very early principles of liability see now Dr. Brunner's most learned and able discussion in Sitzungsberichte der kön. Preuss. Akademie der Wissensch. xxxv. July 10, 1890. Über absichtlose Missethat im Altdeutschen Strafrechte. Spme of the cases mentioned by him, such as Beowulf, 2435, had come to my notice.
48. See, e.g., Gascoigne in Y. B. 7 H. IV. 34, 35, pl. l.
49. Cf. Dr. & Stud. Dial. 2, c. 42 (A.D. 1530).
50. Y. B. 9 H. VI. 53, pl. 37.
51. Y. B. 13 H. VII. 15, pl. 10. Cf. Keilway, 3 b, pl. 7 (M. 12 H. VII.).
52. Y. B. 2 H. IV. 18, pl. 6.
53. Carthew, 425, shows that the Year Book was cited. And the language of Lord Holt, reported in 1 Ld. Raym. 264, shows that he had it before his mind.
54. Brucker v. Fromont, 6 T. R. 659; M'Manus v. Crickett, 1 East, 106; Patten v. Rea, 2 C. B. N. S. 606.
55. Y. B. 2 H. IV. 18. pl. 6.
56. See also 1 Bl, Comm. 431; Noy's Maxims, c. 44.
57. Y. B. 21 H. VII. 14, pl. 21; The Common Law, 226.
58. Y. B. 42 Ass., pl. 17, fol. 260; 42 Ed. III, 11, pl. 13.
59. Y. B. 13 Ed. IV. 10, pl 5; Southcote v. Stanley, 1 H. & N. 247, 250.
60. Bract., fol. 124 b; LL. Gul. I., c. 48; LL. Edw. Conf., c. 23.
61. Smith v. Shepherd, Cro. Eliz. 710; M. 41 & 42 Eliz. B. R.
62. The most important is Lord North's case, Dyer, 161 a (T. 4 & 5 Phil. & M.); but there the master was a bailee bound to return at his peril (cf. The Common Law, 175-179). In Dyer, 238 b, pl. 38 (E. 7 Eliz.), a customer of a port was said to be liable to the penalties for a false return, although he made it through the concealment of his deputy. One or both of these cases are cited in Waltham v. Mulgar, Moore, 776; Southern v. How, Popham, 143; Boson v. Sandford, 1 Shower, 101; Lane v. Cotton, 12 Mod. 472, 489, etc.
63. Waltham v. Mulgar, Moore, 776 (P. 3 Jac. I.).
64. Southern v. How, Cro. Jac. 468; s. c. Popham, 143; 2 Roll. Rep, 5, 26; Bridgman 125, where the special verdict is set forth.
65. Cremer v. Humberston, 2 Keble, 352 (H. 19 & 20 Car. II.).
66. Lane v. Cotton, 1 Salk. 17, 18; s. c. 1 Ld. Raym. 646, Com. 100 (P. 12 W. III.).
67. Saunderson v. Baker, 3 Wilson, 309 s. c. 2 Wm. Bl. 832; (T. 12 G. III. 1772).
68. Ackworth v. Kempe, Douglas, 40, 42 (M. 19 G. III. 1778).
69. 1 Bl. Comm. 345, 346.
70. 2 T. R. 148, 154 (1787).
71. 1 Shower, 101, 107 (M. 2 W. III.).
72. 1 Ld. Raym. 264 (M. 9 W. III.); s. c. 3 id. 250, Carthew, 425, Com. 32, 1 Salk. 13, Skinner, 681, 12 Mod. 151, Comb. 459, Holt, 9.
73. Jones v. Hart, 2 Salk. 441; s. c. 1 Ld. Raym. 738, 739 (M. 10 W. III.); Middleton v. Fowler, 1 Salk. 282 (M. 10 W. III.); Hern v. Nichols, 1 Salk. 289.
74. Brucker v. Fromont, 6 T. R. 659; M'Manus v. Crickett, 1 East, 106; Patten v. Rea, 2 C. B. N. S. 606 (1857).
75. Lane v. Cotton, 1 Salk. 17, 18.
76. See also Noy's Maxims, c. 44.
77. Bl. Comm. 431, 432.
78. Cremer & Tookley's Case, Godbolt, 385, 389 (Jac. I.); Laicock's Case, Latch, 187 (H. 2 Car. I.).
79. Shelley & Burr, 1 Roll. Abr. 2, pl. 7 (M. 1 Car. I). Cf. 1 Bl. Comm. 431; Com. Dig., Action on the case for negligence, A. C.
80. Roll. Abr. 95 (T.), citing no authority, and adding, "Contra, 9 Hen. VI. 53 b." The contradiction is doubtful.
81. Anon., 1 Mod. 209, 210 (H. 27 & 28 Car. II.). Cf. Barker v. Braham, 2 W. Bl. 866, 869.
82. Lane v. Cotton, 12 Mod. 472, 488, T. 13 W. III. Cf. Mors. v. Slew, 3 Keble, 135 (23 & 24 Car. II. 1671, 1672); also Mires v. Solebay, 2 Mod. 242, 244 (T. 29 Car. II.), for an exception by Scroggs, C. J.
83. Sands v. Childs, 3 Lev. 351, 352; Perkins v. Smith, 3 Wilson, 328 (1752).
84. 1 Bl. Comm. 431; Bac. Abr., Master & Servant, K. It is enough simply to refer to the law as to the liability of married women.