INTRODUCTORY REMARKS

Peter Jaffey makes three key points in his recent contribution to this journal.¹

I engage with two of them here as a way of provoking further discussion on what I hope I show is an important subject for analytical study: the nature of legal relations involving the allocation of risk. More broadly, within the scope of this article I propose a model for dealing with the different ways in which remedial obligations and other legal consequences may arise that draws on the Hohfeldian scheme of analysis; but I also consider how adequately Hohfeld's scheme copes with these different paths to the creation of remedial obligations and other legal consequences.

Jaffey's first point is to argue that when in private law there exists an obligation on one party, D, to satisfy the remedial claim of another party, C, that obligation may arise in one of two quite distinct ways. D's obligation to provide a remedy to C (for simplicity, let us assume the remedy amounts to payment of damages by way of compensation) may arise because damage has been caused to C by D engaging in conduct that the law prohibits; D is under a prior duty not to engage in the conduct causing damage to C, which he has breached. Alternatively, D's obligation to pay damages to C may arise where damage is caused to C as a result of conduct engaged in by D but where D was under no prior duty not to engage in that conduct; the law permits D to engage in the conduct and there has been no breach of a prior duty by D.

Jaffey's second key point is to suggest that this distinction can be captured by recognizing “a primary right-duty relation” between C and D when the prior duty exists, and “a primary right-liability relation”²

between C and D when the obligation to pay damages arises without there being a breach of a prior duty.

In his conclusion, Jaffey makes the further point that the acknowledgment of a right-liability relation demonstrates that the Hohfeldian scheme of legal relations is incomplete because Hohfeld fails to recognize such a right-liability relation within his scheme.

I shall focus on the second and third of Jaffey's points. Although there may be some disagreement over the practical applications of Jaffey's first point, it cannot be denied that there do exist two quite distinct ways in which the law may determine that D is under an obligation to pay damages to C: either because D has committed a wrong against C that the law required him to avoid, or because D has been charged by the law with the responsibility of compensating C for damage caused as a result of conduct that in itself is perfectly lawful. In the latter case we may say that the law has made D assume the risk of the damage occurring³

without requiring D to avoid the conduct which carries the risk.

Perhaps the need for Jaffey to make the first point at all can be attributed to a general reluctance to admit that the law might impose an obligation to compensate upon a party who had no prior warning from the law to avoid the conduct that brought about the damage for which compensation is claimed. This might be regarded as flouting the view that the law addresses autonomous agents capable of arranging their affairs so as to avoid the law's penalties. Such a view, however, adopts an unrealistically narrow perspective on human autonomy or the law, or both. Autonomous human conduct is engaged in with a backdrop of uncertain knowledge over the exact consequences of that conduct; life is about making choices to bring about certain consequences and also about making choices to participate in events whose consequences are not altogether clear. If the law were to permit only the first type of conduct, it would necessarily outlaw a significant proportion of normal human activity. Because the law does not operate in this way, it has not only to regulate conduct but also to regulate the allocation of risk.

Alongside this general reluctance, there might be a specific reluctance to accept Jaffey's first point in the context of a particular area of law. Notably, it might be thought that in the tort of negligence with the primacy of a duty of care it would be nonsensical to suggest a remedial obligation arising without the existence of a prior duty. This sort of concern goes to the heart of understanding the area of law in question, but it cannot detract from the analytical possibility of there being two distinct ways in which a remedial obligation may arise. If, on reflection, it appears that the tort of negligence has an ambit extending over both types (i.e., sometimes a remedy is awarded in negligence based on the allocation of risk rather than the prohibition of risky conduct), then that points to the need for further reflection on our use of the concept of a duty of care. That is a matter for tort lawyers to argue about; I shall not pursue it here.

If we accept Jaffey's first point to the extent that we acknowledge the analytical possibility of remedial obligations arising in these two distinct ways, there remains work to be done in considering whether Jaffey has clarified these positions effectively by concentrating on liability as an explanatory device and in considering whether the implications he draws for the shortcomings of Hohfeld's scheme are sound. These two tasks are interrelated. As the Hohfeldian scheme purports to assist in the elucidation of practical legal relations,⁴

one cannot really avoid referring to Hohfeld in testing whether Jaffey's proposal is convincing as well as referring to any insight Jaffey might provide in testing whether Hohfeld's scheme is acceptable. Having said that, the criticism I make below of Jaffey's position is not aimed merely as a vindication of Hohfeld and is not dependent on a general consensus of how Hohfeld's scheme should be understood or how successful it should be regarded.⁵ I suggest below that his scheme requires a certain amount of extension and revision. Nevertheless, on any understanding of Hohfeld, I seek to show that Jaffey's criticism of his scheme is misdirected.

PROBLEMS WITH LIABILITY

The intuitive rejection of liability as a primary concept stems from regarding it as a second-best situation to be in: liability is something to be avoided. The best situation to be in is one where all the right steps have been taken to avoid liability. This viewpoint ties in with the prior-duty path to establishing D's obligation to provide a remedy to C. Comply with the duty, and no obligation, no liability, will occur. The liability is secondary to the prior duty. It is, indeed, contingent upon the nonperformance of that duty.

Does that mean that if we recognize the alternative path to establishing an obligation to provide a remedy (as we have) then, in the absence of a prior duty, the liability becomes primary, as Jaffey suggests? In order to explore this issue, it is necessary to clarify the terms and ideas involved. An immediate problem lies in the use of liability as a term to convey quite different ideas.

The intuitive use of liability raised at the beginning of this section depends on taking liability to carry with it a negative connotation in conveying a position of enforceable indebtedness to another party. Another use of liability is to convey a potential state of a party: a person may be under a liability to experience something or suffer something, given that events work out in a particular way. This use of liability is also usually negative in connotation: “You are liable to catch a cold if you go out in this weather without a coat.”

In the legal context, both uses are recognizable in closely related situations. So we may ask the question, “Which party is liable to bear the loss in the case of the goods being destroyed before delivery?”—conveying a potential state. We may also answer that question with the statement, “In these circumstances the seller is liable to the purchaser to compensate him for the loss of the goods”—conveying a position of enforceable indebtedness. The fact that the potential state in the first case manifests itself in a state of indebtedness once the potential materializes in the second case should not prevent us from recognizing these two quite distinct uses of liability.

The ambiguity produced by these two uses is particularly compressed in the statement, “D is liable for negligence.” This may be understood to mean that if damage is negligently caused, then D's position will change to place him under an obligation to C to provide a remedy. It may also be understood to mean that D is under an obligation to C to provide a remedy because damage has been negligently caused by him.

The opportunity for confusion is increased further when we bring Hohfeld's technical use of liability into the picture. Hohfeld uses power to capture a potential in one party to change the legal relations of another party, and in treating liability as capturing the correlative position of that other party, Hohfeld is clearly using liability to express a potential position rather than an existing obligation to provide a remedy. Moreover, for Hohfeld it is clear that liability could refer to being subject to change in one's legal relations irrespective of whether that change was for the better or the worse. So, for Hohfeld, the owner of property has the power to change the legal relations of another party because he has the legal capacity to bestow ownership of that property on the other by way of gift. The position of the other party is described as being under a liability to be made owner by way of gift.⁶

So liability conveying a potential state does not, for Hohfeld, necessarily bear a negative connotation.

In the technical sense in which Hohfeld uses liability—as indicating a party's potential to have his legal relations changed by another party whether that will produce a disadvantage or a benefit to him—two things are clear. First, the intuitive objection noted above to regarding liability as a primary concept does not bite, because a liability here is not a second-best position to be in that the party could have avoided. For one thing, it might be an advantageous position to occupy. For another thing, it might be a position that a party occupies without having been able to do anything to bring it about or to avoid it. Both points hold in the case of being under a liability to be made owner by way of gift.

The second thing that is clear about Hohfeld's liability is that it is not seeking to capture a remedial position that the party is obliged to discharge. It may so turn out that the potential state captured by Hohfeld's concept is linked to a remedial obligation, but that is just one of the many different kinds of legal positions that may be expressed in a potential form by the liability and is not intrinsic to the Hohfeldian liability itself. To put it simply: one may be under an Hohfeldian liability (H-liability) to be under a remedial liability (R-liability) to pay damages, but if that is the case, then the H-liability expresses the potential of being under an R-liability contingent upon the exercise of an H-power by another party.⁷

A CASE STUDY: VINCENT V. LAKE ERIE

It should already be apparent that there is a danger that Jaffey's proposed analysis may have confused or even merged an H-liability with an R-liability or, less technically, one colloquial use of liability in conveying a potential state of a party with the other colloquial use in conveying a position of enforceable indebtedness. Unlike Hohfeld, for whom a liability in the first sense is not necessarily linked to a remedial obligation, Jaffey persistently skews his analysis to make the link.⁸

And it is certainly possible to find Jaffey slipping from one use of liability to the other.⁹ In order to flush out the ambiguity of liability and test Jaffey's analysis more fully, it will be useful to consider a case study, which Jaffey himself helpfully provides. The case of Vincent v. Lake Erie¹⁰ seemingly involves a situation where D is under an obligation to pay damages to C without being in breach of a prior duty. The situation arises when due to the necessity of avoiding the peril of a violent storm, D keeps his boat at C's dock without permission and, due to the boat crashing against the dock through the force of the storm, causes damage to the dock. It is held that D acted lawfully in mooring his boat without permission but is nevertheless under an obligation to pay damages to C.

Jaffey argues that because the obligation of D to pay compensation to C arises without D having been under a prior duty to avoid mooring at the dock, then D should be recognized as being under a liability that may itself be regarded as primary. Moreover, he links this liability of D with a correlative right in C to be compensated. I suggest below that a rigorous analysis of the Vincent v. Lake Erie scenario detracts from both aspects of Jaffey's analysis.

In order to undertake a rigorous and clear analysis, it is necessary to clarify our terms. I adopt the convention already introduced of distinguishing an H-liability from an R-liability and I also use the H prefix to indicate further Hohfeldian concepts. It does not matter for present purposes whether the understanding I employ of Hohfeld's fundamental conceptions is generally accepted as authentic, as long as it can be seen that the H-terms signify distinct legal positions. I take up the particular challenge Jaffey makes against the Hohfeldian scheme in the following section. We can then analyze the various legal positions within Vincent v. Lake Erie as follows:

• (A) Prior to storm:

• (1) D is under an H-duty to avoid mooring his boat at C's dock without permission;
• (2) C enjoys a correlative H-right that D avoid mooring his boat.
• (B) On event of peril caused by violent storm, the legal relations of D and C change to:

• (1) D is under an H-liberty to moor his boat at C's dock without permission;
• (2) C has a correlative H-(no-right) that D avoid mooring his boat;
• (3) D has an H-power to change [his and] C's legal relations by mooring at the dock such that in the event of damage to the dock [he will be under an H-duty to compensate C and] C will enjoy an H-right to receive compensation from D;
• (4) C is under a correlative H-liability to have his legal relations changed so as to enjoy an H-right that D pay him damages in such circumstances.
• (C) On D mooring his boat at C's dock, D exercises his liberty (B)(1) but also exercises his power (B)(3) so that:
• (D) On the event of D's boat crashing against C's dock and causing damage:

• (1) D is under an H-duty to pay damages to C;
• (2) C enjoys a correlative H-right that D pay him damages.

In the light of this analysis, a number of observations can be made: (i) There is indeed at (D)(1) an obligation of D to pay compensation to C arising without D having been under a prior duty to avoid mooring at the dock. D's H-liberty at (B)(1) is the negation of such a prior H-duty. (ii) This obligation could in the terminology introduced above be described as an R-liability; but in Hohfeldian terminology it is properly described as an H-duty. (iii) To the H-duty of D at (D)(1) there is a correlative H-right of C at (D)(2). Hence (iv) Jaffey's use of liability to describe the position of D at (D)(1) is both unnecessary and misleading—particularly where he seeks in general to contrast liability with duty. Moreover, (v) the correlativity of D's and C's positions at (D) can be expressed as involving a right-duty relationship, rendering Jaffey's insistence on recognizing a right-liability relationship redundant.

The core of the observations so far amounts to a rejection of Jaffey's insistence that a new concept is needed to express D's position at (D)(1). It transpires that Jaffey has simply introduced one of the colloquial uses of liability at this point, where technically it is unnecessary. Far from ushering in a new concept, the terminology of “D being under a liability to pay damages to C” is synonymous with “D being under a remedial obligation to pay damages to C,” which is synonymous with “D being under a duty to pay damages to C.” All that Jaffey is doing here is using an R-liability as a synonym for an H-duty in this context.

But does this analytical commentary on Vincent v. Lake Erie adequately deal with Jaffey's insistence on recognizing a distinct legal position where the remedial obligation to pay damages has arisen as it were de novo, without there being a prior duty that D has breached? The brief retort to that is to make the following further observations: (vi) At (D)(1) D's H-duty to pay damages to C does not arise de novo but is the consequence of events at (C) and (D). (vii) Although these prior events do not comprise the breach by D of a prior duty, they do constitute the basis of the remedial obligation, the H-duty of D at (D)(1). (viii) There is accordingly no need to recognize a distinct legal position, which Jaffey seeks to label a “primary liability,” on the ground that the remedial obligation is now the principal basis for the remedial claim: in both cases it is secondary in the sense that it is the consequence of previous legally significant events; the only difference lies in whether those previous events involve the breach of a duty by D or not.

As Jaffey concentrates his discussion on the change in the parties' positions and emphasizes more the other colloquial use of liability in conveying a potential state of a party, the confusion he falls into can be regarded as a failure to respect the proper sequence of events that these further observations indicate. Notably, in his depiction of C's H-right to recover damages as being connected with D's H-liability,¹¹

Jaffey confuses the stage at (B) where an H-liability can be found with the stage at (D) where the subsequent H-duty materializes. For a fuller analysis at (B) would not only recognize the H-liability of C mentioned at (B)(4) but also acknowledge an additional H-liability of D himself which is implied by the square brackets of D's H-power at (B)(3). That is to say that an H-power in D may not only correlate with an H-liability in C but also at the same time involve an H-liability in D himself, since the exercise of the power will bring about a change in D's legal relations as much as C's.¹² We could add the latent line of the analysis at (B) as:

• (5) D is under a reflexive-correlative H-liability to have his legal relations changed so as to be under an H-duty to pay damages to C.

But this H-liability of D in (B)(5) does not correlate with the H-right of C in (D)(2). The former coheres with the H-power of D in (B)(3), which correlates with the H-liability of C in (B)(4); while the latter clearly correlates with the H-duty of D in (D)(1). That is to say that both parties enjoy potential legal positions at (B) and both parties enjoy materialized legal positions at (D), the potential positions both being H-liabilities, and the materialized positions being an H-right and an H-duty. Nowhere do we get a right-liability correlation.

This confusion is evident when Jaffey seeks to introduce his right-liability relation. He states, “‘Liability’ is used in the sense that someone subject to a liability is susceptible to a change in his legal position, here the change being the accrual of a claim against him.”¹³

But the accrual of the claim, as illustrated at (D)(2) in the above case study, occurs when the change materializes; the liability arises beforehand, at (B)(5), when the claim is potential and contingent, when someone is only “susceptible” to the change. It is not possible to correlate the claim, that is, remedial H-right, with the H-liability into a right-liability relation. At the stage of the liability, the claim has not materialized and may in fact never do so.

However, Jaffey also suggests in this part of his discussion that the correlative right-liability relationship may attach the liability not to the specific remedial right (or claim) but to a general property right: “C's primary right, the right by virtue of which his claim [i.e., remedial right] arises … his right of ownership or an aspect of it, is correlated with … a liability.”¹⁴

From this perspective the correlativity of C's remedial right at (D)(2) with D's duty to pay damages at (D)(1) could be admitted by Jaffey.¹⁵ The place for the novel correlation of right with liability now appears to occur at an earlier stage when we are to consider C's “primary right of ownership.” It is this right of C that Jaffey focuses upon in order to obtain a correlation with D's liability.

However, Jaffey's vocabulary does not make a clean break between the remedial right or claim on the one hand and the primary right of ownership on the other. The property right or “right to a thing” is treated by Jaffey as being in part identified with the owner's ability to “claim payment for its unauthorized use.”¹⁶

From a Hohfeldian analysis, “the right of ownership” is anathema. The legal recognition of ownership does not take the form of a unitary right but is seen as a bundle of rights (claims, liberties, powers, and immunities) which together constitute all the different legal advantages which ownership provides. Jaffey appears to show some deference to this view by talking of “his right of ownership or an aspect of it.”¹⁷ Some of these advantages of ownership or aspects of it are ordinarily contingent. For example, as owner of Whiteacre I enjoy a liberty to use reasonable force in order to eject you if you trespass on my land. That liberty is contingent because it will not arise unless you do trespass, but it is still an important aspect of my ownership of Whiteacre. The point to stress is that the contingent liberty exists as an aspect of my ownership and should not be regarded as in some way secondary to it, so that we would have to talk first of the “primary right” of ownership and only after recognizing that of some sort of secondary contingent liberty to eject a trespasser.

So if Jaffey is making a liability of D correlative with an aspect of C's ownership, then it is that aspect of ownership that needs to be examined. And since the liability of D is now taken to convey the potential state of his legal position, the correlative aspect of C's ownership will similarly be found in a potential form: the contingent H-liberty of the landowner to use reasonable force to eject a trespasser, in the example we have just been discussing; or the contingent H-right of the dock owner to receive compensation for the damage to his dock, in our case study.

In Hohfeldian terminology, the potential state of a party's legal position finds expression as an H-liability. In our case study, it is C's H-liability at (B)(4) that correlates with D's H-power at (B)(3) combined with D's H-liability at (B)(5), which coheres with that H-power. We do not need to posit a separate right of ownership in C in order to be able to capture the legal relationship between C and D in its contingent state. In the case-study analysis, the contingency has been analyzed by focusing on D's H-power and the mutually correlative H-liabilities of C and D. Even if we do not deal with such technically refined Hohfeldian power-liability relationships in the analysis and find that we are speaking merely of the contingent positions of C and D, as we did in the discussion of the landowner's liberty to eject a trespasser, we get the (contingent H-liberty) of L to use reasonable force to eject T, correlating with the (contingent H-(no-right)) of T that L not use reasonable force to eject him, again with no need to have recourse to some prior general right of ownership in L.

If a general “right of ownership” comes into the picture anywhere, it comes in at a much earlier stage altogether: as a broad social (or political or moral) conception of ownership that is advanced prior to the determination of precise legal relations, not as part of a prior or “primary” legal relation. From a particular view of ownership in society, it may be argued that such and such legal relations are required, but once the legal relations are established, then we have no need in our legal analysis to visit that prelegal conception. It may be true that in the incomplete and dynamic state of the law, in practice lawyers have to move adroitly between analyzing established legal relations and arguing for the legal recognition of favored social conceptions.¹⁸

Even so, that does not mean that we should confuse the level of arguing in favor of a particular conception of ownership with the level of analyzing an established legal relation.

These rather abstract remarks can be made more readily intelligible by further illustrations taken from the case study we have been examining, which, as has been noted and is recognized by Jaffey, involves the allocation of risk. The “legal” argument about the allocation of risk in Vincent v. Lake Erie in fact involves two competing social conceptions of ownership as applied specifically to the ownership of a dock, upon which determinate legal judgment had yet to be made. Ought the ownership of the dock to be viewed as sufficiently strong that it should be protected against the risk of damage incurred when a boat lawfully seeking shelter out of necessity from a perilous storm is dashed against the dock by the force of the storm? Or should a weaker conception of ownership prevail so that the interests of the owner are subordinated to the interests of the owner of the boat in peril, and the risk is borne by the owner of the dock as “accidental” damage?¹⁹

Once the majority of the court decided in favor of the former conception of ownership for the dock owner, then the legal relations as analyzed above in (A) to (D) come to represent fully the state of the law. It follows that the H-liability of D at (B)(5) (or of C at (B)(4)) expresses the law that has been established, because an allocation of risk has been made to favor one social conception of ownership or general “right of ownership” as opposed to another. It also follows that the law so established expresses the social conception of ownership or general “right of ownership” that has been favored by the majority of the court. It does not follow that one part of the legal relationship established at (B) does in a peculiar way relate to that social conception of ownership or general “right of ownership” to the exclusion of the other part of the legal relationship; it does not follow that the H-liability of D is the correlative of the general right of ownership of C.

Nor can the liability of D be regarded as in any sense “primary.” From a sociopolitical or moral perspective, it merely expresses at the level of legal relations the outworking of the primary social conception that has been favored. From a legal perspective, it is true to say that the liability is prior at (B) to the remedial obligation at (D), but that does not make it necessarily primary. If anything it is the substantial remedial obligation at (D) that is primary; the contingent state of the liability at (B) makes it as yet unsure of its ever contributing to a material legal position.

These further reflections on the case study accordingly raise doubts over Jaffey's consistency in his search for what he labels a “primary” legal relation. He states that “the whole point of designating a right or relation as primary” is to identify it as the position out of which the remedial claim arises.²⁰

This is quite a different matter if we are seeking the primary social conception that informs this area of law or if we are seeking the primary legal relation that provides a rationale for connected legal relations, and even in the latter case there may be different ways of approaching the issue.

The charge of inconsistency is reinforced if we turn to the alternative path to establishing a remedial obligation: where there has been breach of a prior duty. In this case, Jaffey wishes to mark the prior right-duty relation as primary, which covers the behavior that the law requires D to avoid. Yet in this scenario we can also find a general “right of ownership” and also an H-liability in D, which in Jaffey's misjoined alliance of a right-liability relation have both been accorded “primary” status. As illustration, consider another feature of the example discussed above of T's duty not to trespass on Whiteacre, owned by L. T can be analyzed prior to the act of trespass as having both an H-duty not to trespass, correlative to L's H-right, and also an H-power to change [his and] L's legal relations by entering Whiteacre such that in that event [he will be under an H-duty to compensate L and] L will enjoy an H-right to receive compensation from T—with correlative H-liabilities in L and (reflexively) in T.²¹

Moreover, the precise legal relations established for L and T here can be taken to express the prevailing social conception of ownership of land, or general “right of ownership,” which will determine just where the conflict of interests between landowners and others wishing to pass over their land will be resolved.

But if those features regarded by Jaffey as “primary” in the no-prior-duty path to remedial obligation and as being distinctive to that path are also to be found in the prior-duty path to remedial obligation, then his efforts to shed illumination on the two distinctive ways in which a remedial obligation may arise by employing the device of a “primary liability” must be regarded as a failure. From our consideration of a variety of ways of understanding Jaffey's position, we may also conclude that there remain no grounds for expressing a correlative relationship between C and D in terms of right-liability.

THE ADEQUACY OF HOHFELD'S SCHEME

It would follow from the conclusion that Jaffey's attempt to demonstrate a “primary right-liability” relationship is unfounded, that his criticism of Hohfeld for failing to recognize a right-liability relationship must be misplaced. That is not to say that Hohfeld's scheme is adequate in every respect. One inadequacy has already been noted in Hohfeld's simple bilateral power-liability correlation, which has been expanded to allow for the expression of a further reflexive liability in the power holder—introduced as (B)(5) in the case study above.²²

It is worth considering whether the different ways in which an obligation to provide a remedy may arise have any other repercussions for Hohfeld's scheme.

One positive point to make is that a Hohfeldian analysis expanded to recognize reflexive liabilities may help us to see clearly just where the difference lies between the two paths of establishing a remedial obligation. In fact there are two key differences that emerge. If we take the analysis from the case study of a no-prior-duty (NPD) path in the previous section, we note that at (B) we have coexisting legal relations between D and C comprising: (i) an H-liberty in D at (B)(1) permitting him to engage in the relevant activity together with correlative H-(no-right) in C at (B)(2); and (ii) an H-power in D at (B)(3) cohering with a reflexive H-liability in D at (B)(5) together with the correlative H-liability in C at (B)(4), which covers the contingent state of the potential remedial relationship. The first difference, then, is that corresponding to (i) in a prior-duty (PD) path case we have an H-duty in B with a correlative H-right in C forbidding D to engage in the relevant activity.

The corresponding situation at (ii) in a PD path case is similar in that it, too, covers the contingent state of the potential remedial relationship by means of a similar H-power and H-liabilities arrangement.²³

Yet a second difference emerges if we take note of the fact that in our NPD-path case study, there is an intervening stage at (D) after the exercise of the power at (C) that amounts to the activity of an inanimate agent (the storm) and comprises the realization of the risk²⁴ necessary for the materialization of the remedial relationship. By contrast, in a PD-path case there is no intervening stage: once the power is exercised by breaching the prior duty (entering Whiteacre), the remedial obligation automatically follows.

We may summarize these differences that a Hohfeldian analysis reveals by concentrating on the positions of D, and for the moment assuming the H-prefix, in the following manner:

The italicized sections clearly bring out the two key differences mentioned. As to whether it is justifiable to employ these different paths to remedial obligation in different circumstances, that is a debate to be exercised over which conception of ownership should prevail in a society. Certainly the fact that the NPD path is still subject to a realization of risk might be regarded as making it fairer, in some circumstances, to impose a remedial obligation on an activity that the law permits.

Having said that, we should also recognize that not every NPD path to a remedial obligation, or at least some other form of financial indebtedness, involves the allocation of risk. Jaffey provides a good example in the case of a license fee payable for use of somebody else's intellectual property, where that use was permitted on payment of a fee. Here there is no risk of the obligation to pay materializing. It follows as a certainty, as in the PD path, on engaging in the activity—but there remains the commonly recognized distinction that we are not dealing with a consequence of what the law prohibits (by a prior duty) but a consequence of an activity that the law permits. This means that a complete analysis would have to differentiate between an NPD₁ path, as above, involving the allocation of risk, and an NPD₂ path involving the determination of a fee:

Such a path might also be used to analyze the payment of income tax on the permitted activity of engaging in gainful employment (tax being a certainty rather than a risk) but, as useful as this further path might be, it takes us beyond the conventional instances of remedial obligations.²⁵

So much for the positive contribution that a Hohfeldian analysis can make. On the negative side, once we have made proper provision for the expression of reflexive liabilities, there is another aspect of the analysis we have developed above where the Hohfeldian scheme may be regarded as inadequate. This occurs at the intervening stage (D), which in the case study we have seen amounts to the activity of an inanimate agent (the storm) and is necessary to bring about the realization of the risk before the remedial relationship can materialize. Although this is a crucial stage in the development of the legal relations we have recognized, it is not itself expressible in terms of legal relations and does not fit neatly into the Hohfeldian scheme.²⁶

The problem here is more convoluted than the simple need to expand Hohfeld's analysis in order to recognize nonhuman factors, or inanimate forces, as having legal consequences. In our discussion of contingent legal positions, we have noted that these can be dealt with either by locating them as the subject of H-powers and H-liabilities within Hohfeld's scheme or by attaching a contingent status to standard conceptions found within Hohfeld's scheme. However, it is important to recognize that where the contingency is dependent on the activity of an inanimate force, then the choice between these two representations of contingency on grounds of style or analytical detail is no longer available. H-powers and the related H-liabilities are restricted to cases where the potential change in legal positions is dependent on the human conduct of the power holder.

Where the change occurs as a result of the operation of an inanimate force, as it does, for example, at (B) in the case study in the previous section (the legal relations of D and C being changed by the operation of the violent storm), we cannot employ an analysis revolving around the exercise of an H-power. We need to acknowledge here that the law recognizes the operation of (inanimate) conditions as well as the exercise of (human) powers as being capable of bringing about a change in legal relations. As for describing the potential state of the positions of the parties prior to the operation of such a condition, it is possible to deploy the alternative device we have noted of attaching a contingent status to the Hohfeldian conceptions that will, after the event, capture the parties' positions.²⁷

We could, for example, speak of a (contingent H-liberty) of D at (A)(3) to moor his boat at C's dock without permission in the event of peril caused by a violent storm, and a correlative (contingent H-(no-right)) of C at (A)(4). These are properly analyzable as part of the broader legal positions (bundle of legal relations) held by D and C at (A), although we tend to overlook most of the vast number of conditions or exceptions that apply to our legal positions as being redundant to our immediate concerns.

The event occurring at (D) in the case study, although it, too, involves the operation of the storm in bringing about a change to the legal relations of D and C, cannot be treated in quite the same way as the event at (B). The event at (B) was in itself a discrete condition for bringing about the change in the legal relations of D and C: before the storm there is no liberty to moor at the dock without permission; due to the storm the liberty comes into existence. The event at (D) does not work in exactly the same way because this event completes the potential change initiated at (C), when D keeps his boat at the dock. As is clear from the NPD₁-path analysis above, the operation of the storm at (D) realizes the risk that D has taken upon himself by his activity at (C) in keeping his boat at the dock. Upon a more careful analysis, we do not have here legal relations that change either side of the condition (as did the liberty above) from a state of being denied to a state of being recognized purely because of the satisfaction of the condition. What we have either side of the condition being satisfied at (D) is action by one of the parties to bring about the potential for a remedial duty and then fulfillment of the inanimate condition so as to realize that potential in bringing about the duty. We then have to conclude also that the H-power exercised by D at (C) is not fully an H-power because, unlike an H-power proper, it is not sufficient upon exercise by D to bring about a change in legal relations (as did the H-powers in the PD and NPD₂ analyses above); even after the apparent H-power is exercised at (C), it remains ineffectual until the realization of risk at (D).

The point at which the Hohfeldian scheme does require major revision then turns out to be in dealing with the incurring of risk: where one party by his conduct takes on the risk of damage to another party's interests. That situation is neither analyzable as a contingency captured in power-liability relations nor analyzable as a contingency simply captured by contingent relations subject to the fulfillment of a condition. It involves the compound {(exercise of power)(satisfaction of condition)} to bring about the change in legal relations. We have here a complex double contingency.

Yet as the allocation of risk as well as the prescription of conduct forms the work of a modern legal system, the Hohfeldian scheme may be regarded as less than comprehensive in failing to give sufficient recognition to the particular legal relations involved in the incurring and realization of risk. The revision required is significant in that it undoes the simple assumptions of Hohfeld's scheme that all legal relations involve either what is required or not required of one party's conduct by another (in the right-duty and liberty-(no-right) relations) or what one party can or cannot do to change what is required (in the power-liability and immunity-disability relations).²⁸

We now have to accept additionally that the business of the law is concerned with the risk of what might be required and that one party can change not just what is required but who incurs the risk of what might be required.

In proposing a model based on differentiating the PD path, the NPD₁ path, and the NPD₂ path to deal with the different ways in which remedial obligations and other legal consequences may arise, I have drawn on the benefits in terms of analytical clarity that Hohfeld's scheme provides. By arguing that this produces a more intelligible approach than Jaffey's attempt to offer an explanation for the different routes to remedial obligations, in a manner that poses a fundamental threat to Hohfeld's scheme of analysis by seeking to establish the need for a right-liability relation, I have accordingly been to a significant extent engaged in a vindication of Hohfeld. Nevertheless, accepting Jaffey's challenge to confront the particular characteristics of legal relations involving remedial duties that are not associated with the breach of a prior duty (or that can be understood as involving the allocation of risk) has revealed the need to extend Hohfeld's scheme and to revise it.

The extensions have been made in order to accommodate the expression of reflexive liability (in the power holder) and with that to recognize mutually correlative liabilities (in the power holder and the other party); and also to acknowledge the significance of (nonhuman) conditions for contingent Hohfeldian relations. The major revision to Hohfeld's scheme technically forces us to recognize what we might refer to as a weaker form of H-power alongside the established conception, where the exercise of power is ineffectual without the satisfaction of a further condition (realization of risk) to bring about a change in legal relations. The recognition of this particular role of legal conditions in the exercise of legal powers might provoke yet further reflection on the significance of legal conditions and their relationship to legal powers.²⁹