TWO CONTEMPORARY ACCOUNTS OF BOUNDED DIRECT HORIZONTALITY

A. Van der Walt

In The Horizontal Effect Revolution and the Question of Sovereignty, Johan van der Walt develops an account of horizontality that is founded on a root-and- branch critique of the assumptions underlying the default vertical approach.^[295] To begin with, recall the third assumption underlying default verticality, according to which individuals in the ‘private sphere’ face each other on terms of abstract equality (ie within the same ‘horizontal plane’).

Default verticality’s assumption that private relationships are horizontal is accomplished, Van der Walt points out, through the conscious ‘erasure of baseline questions’^[299] (that is, questions relating to the distribution of economic and social power) from the legal enquiry.

A combination of these two assumptions - abstract freedom and individ­ual responsibility - yields a situation where actually existing inequalities and hierarchies within the ‘private' domain are deemed normatively irrelevant for the application of the constitutional rights framework. Thus, for Van der Walt, what horizontality is really about is ‘the restriction and even the undoing of private hierarchies. It concerns the constitutionally required levelling and hori- zontalisation of these hierarchies.'^[300]

This framing is important, because it suggests that the key challenge for horizontality is not about how best to ‘transplant' constitutional rights from the public to the private sphere (the problem that, as we saw, indirect horizontality struggles with^[301] and something that Van der Walt expressly criticises it for), but rather, how to determine which forms of power relations or inequalities within the ‘private' sphere are normatively significant with respect to the application of the rights framework (‘re-horizontalisation'). Because, as Van der Walt himself concedes, some form of principled limitation is required, so that the ‘the practice of constitutional review [does not turn into] the higher review of all private law disputes'.^[302]

What principle does Van der Walt adopt? He argues that there are certain private relationships and transactions where ‘disputes between individuals... have social-political significance that extends beyond the relations of those directly involved’ (emphasis added).^[303] Drawing upon the work of Gunther Teubner and other systems theorists, he defines the purpose of horizontality as to ‘thematise conflicts in which human rights of individuals are asserted against structural violence',^[304] because in certain ‘individual forms of action'^[305] what is actually implicit is ‘institutional conflict'.^[306]

As an example, Van der Walt takes the famous instance of Shelley v Kraemer^[307] and the constitutional validity of racially restrictive covenants in a society that is itself scarred by racial domination and subjugation.

Two Contemporary Accounts of Bounded Direct Horizontality 75 to be private’ (emphasis added).^[308] Thus, the collective refusal of white property owners to sell to Black homebuyers is no longer simply a private decision about how an individual intends to dispose of their property, but, rather, should be understood as a set of ‘private endeavours to promote or sustain public inequality’ (emphasis added).^[309] Van der Walt does not explain what he means here by ‘public inequality’, but presumably it is inequality that goes beyond the specific private relationship at issue. This is evident from the fact that Van der Walt goes on to refer to disputes between individuals that have a ‘social-political significance that extends beyond the relations of those directly involved’ (empha­sis added).^[310] Van der Walt draws upon Laurence Tribe’s insights in his analysis of Shelley which Tribe describes as the subjugation thesis: private discrimina­tion is permitted except where it has an ‘overall effect of confining minorities to their current disadvantaged positions in society’.^[311]

While Shelley and racially restrictive covenants present straightforward examples, Van der Walt still needs a final argument to complete his account: how do we identify in general those private relationships that ‘thematise... an institutional conflict’? Van der Walt answers this question in two parts. He begins by identifying the first assumption underlying default verticality: the idea of undivided sovereignty. As long as there continues to exist ‘an opposi­tional monopoly as far as threats to liberty are concerned’,^[312] departures from default verticality will continue to have to go through the state (whether through state action or indirect horizontality) in order to get to the private relationship at issue.

As we can therefore see, Van der Walt’s model of horizontality turns upon three axes: (i) a rejection of the three assumptions underlying default verticality; (ii) locating horizontality in those private transactions or disputes that reflect not simply the two (or more) private parties involved, but embody ‘institutional conflict... within individual forms of action’; and lastly (iii) specifying that this ‘institutional conflict’ is that which exists between ‘social majorities and minorities’ at any given time.

B. Thomas

Thomas’s starting point is similar to Van der Walt’s.^[313] She begins by noting that default verticality is a historically contingent phenomenon, whose origins lie in the unique form of power that was once vested in the state.^[314] However, the existence of other power structures in the modern world - many of which are themselves the products of private law - that are capable of violating rights to an extent and a degree that is not all that dissimilar to what the state can do, requires a clear departure from default verticality.^[315]

Thomas’s priority, therefore, is the rights bearer. Thus, the first question relevant for an application of the rights framework is to ask which entities are capable of violating rights, and how. Like Van der Walt, however, Thomas also concedes that this may be a necessary, albeit insufficient, condition for horizon­tality, as it would mean that ‘the potential applicative scope of those [horizontal] rights expands exponentially’.^[316] In such a situation, ‘each right generates a number of duties, such as non-interference, protection against violation by third parties, and possibly also positive measures to diminish the likelihood that rights will be violated’.^[317]

Thomas is also concerned with the ‘transplant problem’, noting, as I argued in chapter three, ‘the legal presumption that non-state actors are all the same, so far as rights are concerned’.^[318] Thus, much like Van der Walt, Thomas clari­fies that the purpose of the horizontal rights framework is ‘not to apply public law rights to private relations directly’,^[319] but to articulate a principled basis for determining which private entities are subject to horizontal obligations and which private relationships ought to be brought within the framework.

To arrive at this basis, Thomas begins by asking herself what it is that char­acterises the rights framework in the vertical context (ie between the individual and the state).

This is evident from how Thomas develops her argument. First, she exam­ines what it is that is being protected (a set of core enumerated interests). Next, she argues that what characterises rights is their ‘relational context’^[320] - ie they depend upon the specific relationship between the two parties in question. In other words, they ‘cannot be understood exclusively in reference to the inter­ests of individual subjects’.^[321] This relationship, Thomas claims, is that of ‘interpersonal dependency’.^[322] ‘Dependency’, she argues, ‘involves an increased vulnerability of one party to another... it is a relational structure that links power and need, or power and vulnerability, so that the vulnerability of one person is specific to the influence of another.’^[323] We may notice here that the condition of dependency addresses the ‘transplant problem’ that we identified with indirect horizontality in the previous chapter. Instead of simply ‘trans­planting’ the rights framework to the private context and then engaging in an intuitive ‘balancing’ of rights and duties between two formally equal parties, the condition of dependency links the application of the rights framework to the specific nature of the relationship between the private parties.

Thomas refines this idea further by noting that dependency must be specific, both in terms of who one is dependent on and for what. In particular, a relation­ship of interpersonal dependency exists where ‘one party substantially controls the goods of another’ (emphasis added).^[324] Drawing once again from the exam­ple of the vertical rights framework, evidence of such control can be found, for example, in the existence of a monopoly (akin to the state’s monopoly over legitimate coercion), the state’s ability to affect the ‘dignitary status’ of everyone in its territory because (in part) of the absence of feasible exit options, and so on. This, according to Thomas, is

the moral context of the vertical relationship between the individual and state... [which] means that the special protection associated with that relationship can be replicated on a principled basis within the private sphere so that not all relational contexts within that sphere are treated as though they were the same.^[325]

The last bit, of course, speaks to the principled limitations of the model: only those relationships within the ‘private sphere’ that replicate the ‘moral context’ of the individual-state relationship will be subject to the rights framework.

Finally, while the existence of a dependency relationship is necessary for constitutional horizontal rights to apply, it is not a sufficient condition. Thomas still needs a reason for why obligations ought specifically to attach to the private duty bearer; the mere existence of dependency appears not to be enough. Consequently, she argues that this dependency must, in addition, arise in the course of an ‘undertaking’ - ie a course of action that the more powerful party must have voluntarily undertaken, which provides the background context to the existence of dependency (eg the act of hiring an employee). This provides the last link in the chain, justifying not only the application of the rights framework, but also ensuring that it is that specific private party who is bound to it. Thus,

the more powerful person’s obligations will be limited in two main ways. First, they will arise only in respect of those vulnerable persons whose vulnerability arises in respect of the undertaking itself. This means that we cannot extend the obligations to say that the more powerful supervisor [ie a supervisor on the factory floor] will owe all the obligations in question (protection, aid, etc.) to everyone he passes on the street at the end of the workday. Second, the more powerful party’s obligations will arise only in virtue of those interests in respect of which the other person’s vulnerabil­ity is caused or significantly exacerbated by the undertaking itself. The dependency analysis, therefore, operates within a sphere of the undertaking that is the nexus or fulcrum of the relational structure.^[326]

In sum, therefore, Thomas’s horizontality model depends upon (i) identifying that an individual’s access to a core good is at stake; (ii) identifying the existence of a relationship of ‘dependency’ between the two parties, where the dominant party exercises substantial control over the weaker party’s access to the good; and (iii) demonstrating that that relationship has arisen within the context of an ‘undertaking’ involving both the parties.

What unites Van der Walt’s and Thomas’s bounded horizontality accounts is a prima facie rejection of the three assumptions underlying default verticality. Like that of Van der Walt, the basis of Thomas’s model lies in the understanding that default verticality’s ‘flattening’ of the ‘private sphere’ - so that all ‘private relationships’ are deemed to take place within the same ‘horizontal plane’ between equally free individuals - is deeply flawed. And this visualisation of the private sphere and the horizontal plane is linked to the unitary conception of sovereignty, which forms the basis of limiting rights claims to the individual­state relationship. Thomas addresses this constraint by arguing that it is not sovereignty that is the normatively significant factor in the rights framework, but a specific kind of power difference (‘dependency’) - a difference that is also found in relationships that do not involve the state and its sovereign power. Thus, while Van der Walt propounds a new idea of sovereignty that exists in the ‘private’ domain, Thomas displaces the concept of sovereignty from the rights framework itself.

There is also, however, a more significant difference between the two. Van der Walt and Thomas differ on the question of how to identify the private

relationships that will be subjected to the rights framework. For Van der Walt, the key lies in locating an institutional background that frames an individual dispute. This, in turn, he defines as conflicts involving social majorities and minorities. For Thomas, on the other hand, what matters most is the relative position of the two parties vis-a-vis each other: what is crucial, apart from an ‘undertaking’, is the existence of a specific form of power and control (over basic goods) by the stronger party, and the existence of vulnerability or dependency on the part of the rights bearer, in a manner that approximates the ‘moral context’ characterising the relationship between the individual and the state.

III.