profitable as to speak of an "intentional act" without referring to the consequences. We do not say that a man is liable for an "intentional act." We say that he is liable for those consequences which he intended. By parity of reasoning, why should we say that a man is liable for a "negligent act"? Ought we not to say that a man is liable for those consequences as to which he was negligent?

By speaking of "the legal quality of the act" we are opening new fields of liability. On p. 577 Scrutton, L.J. says:

In the present case it was negligent in discharging cargo to knock down the planks of the temporary staging, for they might easily cause some damage either to workmen, or cargo, or the ship. The fact that they did directly produce an unexpected result, a spark in an atmosphere of petrol vapour which caused a fire, does not relieve the person who was negligent from the damage which his negligent act directly caused.

If "or" is used above as a disjunctive particle, this means that a shipowner can recover for the injurious consequences of an act, which was not negligent as to the ship but which was negligent as to one of the workmen. But as Bowen, L.J. said in Thomas v. Quartermaine1:

To prove negligence it is not enough to shew that the defendant has been negligent to others, the plaintiff must shew that there has been a breach of duty towards himself. In the Polemis case the arbitrators have found that the causing of the spark could not reasonably have been anticipated from the falling of the board. To let the board fall was therefore not negligent as to that consequence. It was negligent as to other consequences which might have followed, but it clearly was not as to the actual result which occurred. The liability of the defendants is thus based not upon the actual results of their act but upon other results which might have but did not occur.

1 (1887) 18 Q.B.D. 685, 698.

A plausible criticism of the above argument must be considered here. It may be said that the consequence of injuring the structure of the ship by means other than fire and the consequence of setting the ship on fire are the same, namely, the consequence of damage to the ship. The answer is that although the ultimate consequences in both cases are identical, nevertheless the immediate consequences are different. Thus dropping the board may have two consequences: (a) Injuring the structure of the ship by means other than fire:

(b) Setting the ship afire.

As to consequence, (a), the act was negligent; as to consequence (b) the act was not negligent. The fact that consequence (c), damage to the ship, follows both from consequence (a) and from consequence (b) is immaterial. If in a series of consequences one consequence is held to be not attributable to the defendant's fault, then clearly the further consequences cannot be attributed to him.

That this division of consequences is not merely a specious distinction is shown by the arguments of counsel in the case itself.1 In the charter party there was an exception clause as to fire. The defendants claimed that they were therefore not liable for loss by fire although they would have been liable for injuring the structure of the ship by other means. The Court, however, held that the exception clause did not apply to fire negligently caused.2 The question in the Polemis case therefore was whether there was negligence in causing the fire.

It is interesting to consider what the judgment would have been if the charter party had contained an exception clause as to negligent injury to the structure of the ship by means other than fire, but

1 [1921] 3 K.B. 560, 566, 568.

2 Warrington, L.J., at p. 573: "As to the first point, the exception clause is in the following terms. [The Lord Justice read it.] There is, therefore, no express exception of loss by fire caused by negligence."

had not excepted fire caused by negligence. It is difficult to see how the charterers could have been held liable. Dropping the board would not have been legal negligence as far as injury to the structure was concerned as this would be covered by the exception clause. Nor could it be held to be negligent as to the spark under the findings of the arbitrators. The act, therefore, would not have "the legal quality of negligence." Nevertheless the act, the circumstances, the resulting consequences, and the duty of care as to those resulting consequences, would be the same as in the Polemis case. The only difference would be a difference of legal duty as to certain other consequences which might have but apparently did not result.

We have considered above the Smith case which, prior to the Polemis case, was the sole relevant authority for the "direct consequence" rule. On the other hand, there are a number of cases, for many years considered to be authoritative, which are in conflict with it. In his article, Liability for Consequences,' "Sir Frederick Pollock has shown that the Polemis case cannot be reconciled "without great difficulty" with Cox v. Burbidge, and that Sharp v. Powell "seems to be clearly overruled." In Cox v. Burbidge the defendant's horse, straying on a highway, kicked the plaintiff, a child. It was held that, although the horse was wrongfully on the highway the injury complained of was not such a consequence of the wrong as the owner was liable for. As Erle, C.J. said1:

But, even if there was any negligence on the part of the owner of the horse, I do not see how that is at all connected with the damage of which the plaintiff complains.

In Sharp v. Powell, the facts of which are too

1 L.Q.R., Vol. xxxviii., p. 165.

2 (1863) 13 C.B.N.S. 430, 32 L.J.C.P. 89.

3 (1872) L.R. 7 C.P. 253.

4 At p. 436.

I

well known to need repetition, Keating, J. said1:

It may be said, and indeed it has been said, that, if the defendant had not washed his cart where he did, and allowed the water to run down to the drain, it would not have accumulated and frozen there, and so the accident would not have happened. If that is to render the defendant liable, where is the liability to stop? The damage in question, not being one which the defendant could fairly be expected to anticipate as likely to ensue from his act, is in my judgment too remote.

In this case there was a wrongful act, and damage as a direct consequence of that act. A unanimous court, however, held that the defendant was not liable because he could not "fairly be expected to anticipate" the result.

Perhaps we might add to these cases the judgment of Lord Esher, M.R. in Cobb v. G.W.R.Co.

He said:

With regard to the second head of complaint, there was according to the statement of claim, a breach of duty. It was the duty of the defendants not to allow their carriage to be overcrowded. But then it is necessary to show that the alleged damage was such as would naturally and ordinarily result from such breach of duty. It cannot be considered as the probable and ordinary result of allowing a compartment of a railway carriage to be overcrowded that a passenger should be robbed by his fellow-passengers. This case can be distinguished from Metropolitan Ry. Co. v. Jackson, the "thumb" case, in which the House of Lords held that the accident might no less have happened if the carriage had not been overcrowded at all. In the Cobb case the overcrowding had a causal but not a "probable and ordinary" connection with the damage.

It is therefore submitted that the "direct consequence" rule in torts of negligence is in conflict with former precedents of great weight. From the logical point of view, too, it is open to the grave objection that it is based on the fallacy that consequences "flow from negligence." Consequences are the result of an act. That act may be negligent 1 At p. 261. 2 [1893] 1 Q.B. 459, 463. 3 3 App. Ca. 193.

3

as to certain consequences and not as to others; it is, therefore, unscientific to talk of an act as having "the legal quality" of negligence. Whether the "direct consequence' rule is equally open criticism in torts other than negligence it is not within the scope of this paper to consider.

open to

What rule can be suggested in its place? In torts of negligence there is an obvious one. A tortfeasor should only be held liable for those consequences as to which he was actually negligent, i.e. as to those consequences which a reasonable man placed in his position would have foreseen as possible and would have avoided by due care. As to other consequences he was not negligent and therefore not a tortfeasor.1 This suggested rule would also make unnecessary the much criticised "natural and probable" consequence rule. Foresight is a conception which differs totally from probability. I may have foresight that a thing may happen even though it is exceedingly improbable. Thus a man who throws a book out of a window has foresight that the book may strike someone, although in the ordinary connotation of the word there is no probability of it, as the chances may be a thousand to one against. Whether a man has used reasonable care to avoid a foreseeable consequence will depend upon the facts in each particular case.

By introducing the idea of foreseeability rather than probability, a point raised in Dulieu v. White and Sons can be explained. Kennedy, J. said2:

It may be admitted that the plaintiff in this American case would not have suffered exactly as she did . . . if she had not been pregnant at the time; and, no doubt, the driver of the defendants' horses could not anticipate that she was in this condition. But what does that fact matter? If a man is negligently run over, . . . it is no answer to the sufferer's claim for damages that he would have suffered less injury, or no injury at all, if he had not had an unusually thin skull or an unusually weak heart.

1 See Pollock, Liability for Consequences, L.Q.R., Vol. xxxviii., pp. 165, 167.

2 [1901] 2 K.B., at p. 679.