AMCU & Others v Chamber of Mines of SA & Others (CCT87/16)  ZACC 3 (21 February 2017)
- The 'workplace' is not the place where any single employee works - it is where employees collectively work. In determining that, 'location' is not the primary factor, 'functional organisation' is. A 'workplace' may be a single location or a number of locations, based on the independence of those operations in the light of their size, function and organisation.
- Majoritarianism is a recurrent theme throughout the LRA. Whilst its application may limit the right to strike, this is justified through benefitting orderly collective bargaining.
The Chamber of Mines, acting on behalf of its members in the gold mining sector (including Harmony Gold, AngloGold Ashanti and Sibanye Gold), negotiated wages and working conditions with unions representing the majority of workers in the sector, namely NUM, Solidarity and UASA. The resulting collective agreement expressly made it applicable to all the companies' employees in terms of s23(1)(d) - ie even those not members of those unions.
AMCU did not accept the employers' offer and was not a party to the agreement. In January 2014, it notified the three companies mentioned above that its members were going on strike from 23 January 2014 at 5 specific mines at which it had majority membership. AMCU was not however the overall majority union at any of the mining companies who own those mines. In response, the Chamber obtained a Labour Court interdict in terms of s.65(1) and (3), which prohibit striking by anyone who is bound by a collective agreement that either prohibits a strike or even regulates the issue in dispute. The LC accepted that AMCU's members at those mines were bound by the collective agreement concluded with the majority unions.
AMCU appealed the LC decision to the LAC, but failed. This was then referred to the Constitutional Court. The crisp issue facing the CC was whether the collective agreement negotiated with the unions having an overall majority in the sector, applied at the 5 mines at which AMCU had a majority. This required the CC to consider and apply the definition of a "workplace" in s213, which provides as follows:
"If an employer carries on or conducts two or more operations that are independent of one another by reason of their size, function or organisation, the place or places where employees work in connection with each independent operation, constitutes the workplace for that operation".
The CC accordingly had to decide whether each mine where AMCU had a majority was an "independent operation" by reason of its "size, function or organisation"? In summary, AMCU argued that the collective agreement should not have been extended in terms of s23(1)(d) to apply at those 5 mines at which it had majority membership, as the other unions did not have a majority at those workplaces - those mines should have been regarded as "the workplace" for the purposes of the LRA. Further, the application of s23(1)(d) in this instance was unconstitutional, as it interfered with the constitutional right of AMCU's members to strike.
The CC did not agree. The CC confirmed that for the purposes of the LRA, a 'workplace' is not the place where any single employee works - like that individual's workshop or assembly line or desk: it is where employees collectively work. And in determining that, 'location' is not the primary factor, 'functional organisation' is. This then means that a 'workplace' may be a single location or a number of locations, based on the independence of those operations in the light of their size, function and organisation.
On the facts of this case, the CC agreed with both the LC and the LAC that the 5 mines at which AMCU had majority membership, were not independent operations. The CC found that each mining company constituted a single industry-wide workplace, despite the fact that at some of the individual mines, the companies had concluded separate recognition agreements with AMCU.
The CC noted that majoritarianism is a recurrent theme throughout the LRA. The CC recognised that its application in this instance limited the right to strike, but that this was justified in that majoritarianism, in this context, benefitted orderly collective bargaining.
The key then to determining the workplace is whether the operation is 'functionally independent' - not where it is located. This will be determined on the facts of each case. Parties intending on persuading a court to accept their interpretation of the 'workplace' in their particular circumstances, will need to lead convincing evidence relating to the independence of that workplace on the basis of its size, function and organisation.
Extract from the judgment:
 Two things are immediately notable about the way the statute defines "workplace". The first is its focus on employees as a collectivity. The second is the relative immateriality of location. Both signal that "workplace" has a special statutory meaning.
 First, "workplace" is not the place where any single employee works - like that individual's workshop or assembly line or field or desk or office. It is where "the employees of an employer", collectively, work. The statute approaches the concept from the point of view of those employees as a collectivity. This accords with the role the term "workplace" plays in the LRA. This sees workers as a collectivity, rather than as isolated individuals. And that in turn squares with the statute's objects. The promotion of orderly bargaining by workers, collectively, is one of the statute's express primary objects. That the focus of the definition of "workplace" is on workers as a collectivity rather than as separate individuals fits.
 The second point follows. It is that location is not primary: functional organisation is. The definition encompasses one or more "place or places where employees of an employer work". This means that "the place or places" where workers work may constitute a single workplace. That entails the intrinsic possibility of locational multiplicity for a single "workplace". Right at the outset this eliminates any notion, which the ordinary meaning of "workplace" might encourage, that each single place where a worker works is a separate "workplace".
 The first part of the definition creates a default rule that, regardless of the places, one or more, where employees of an employer work, they are all part of the same workplace. The second part superimposes a proviso in the form of an exception - regardless of how many places where employees work, different "operations" may be different workplaces only if they meet the criteria the definition specifies. The key is whether an operation is independent - not where it is located. Yet again, no significance is attached to the "places" where employees work, since the term features in both parts of the definition. Each independent operation, which constitutes a separate "workplace", may itself be at one or more separate locations.
 Hence the proviso determines not so much whether separate physical places of work are separate workplaces, but rather whether independent "operations", however geographically dispersed, are separate workplaces. The pivotal concept is independence. If there are two or more operations and they are "independent of one another by reason of their size, function or organisation" then "the place or places where employees work in connection with each independent operation, constitutes the workplace for that operation". This is a test of functional organisation, and not geography or location.
 Both features of the definition - its approach to workers as a collectivity, and its de-emphasis of geography - have a practical bite. They signal that for purposes of the LRA "workplace" doesn't have its ordinary meaning: the legislature has assigned a special meaning to the term. It follows that AMCU's contention that the ordinary meaning of "workplace" applies, namely the geographical places of work of its members, at their individual mines, faces into a conceptual windstorm. It must battle against not only the specified statutory wording, but the entire statutory context that supports that meaning and in which it is embedded.
 It is this statutory definition the Labour Court and the Labour Appeal Court applied. Was each AMCU-majority mine a separate "workplace"? That depends not on the mines' geographic location or where the individual workers worked, but on the functional signifiers of independence the definition lists. It requires one to determine whether the employer companies conduct two or more operations "that are independent of one another by reason of their size, function or organisation".
 On this question, the facts before the Labour Court and the Labour Appeal Court were not in dispute. They related to the organisational methodology and practicalities of each mining company. The Labour Court and the Labour Appeal Court both found, in conclusory terms, that the individual AMCU-majority mines did not constitute independent operations. They were not swayed by the fact that, at some of the individual mines, the companies had concluded separate recognition agreements with AMCU. Each mining company constituted a single industry?wide workplace.
 The question is not whether a single mine can constitute a "workplace". It obviously can. The definition expressly provides for that. Instead, the critical issue is whether any of the five AMCU-majority mines was an independent operation by reason of size, function or organisation. Both the Labour Court and the Labour Appeal Court determined that each mining house operated integrally as a single workplace, and that each AMCU-majority mine was not an independent operation. Even upholding AMCU's argument that the application of the statutory definition is not a purely factual enquiry does not lead to a different finding. No reason in constitutional principle, legal analysis or factual assessment provides a reason for this Court to overturn those findings. To this one should add that the findings of the Labour Court and the Labour Appeal Court are owed special consideration since they operate as specialist tribunals.
 Once majoritarianism is recognised as a founding principle of the LRA, the statute must unavoidably determine some practical way in which the principle operates. Without a constituency that defines it, there cannot be a collectivity. AMCU's complaint is not that the majority counts, but how to define the constituency within which the majority counts. And thence flows its constitutional grievance.
 AMCU is right that the codification of majoritarianism in section 23(1)(d) limits the right to strike. The key question is whether the principle provides sufficient justification for that limitation. Both the Labour Court and the Labour Appeal Court gave detailed and extensive consideration to this. I do not seek to improve their reasoning. In short, the best justification for the limitation the principle imposes is that majoritarianism, in this context, benefits orderly collective bargaining.
 Perhaps a different definition of "workplace" might have worked equally well, or maybe even better, or been fairer to smaller or emergent unions. AMCU makes a plangent case for saying so. But that is not the question before us. Our task as judges is not to pick and choose between the rights and wrongs, advantages and disadvantages, of different constituency models. Our responsibility is much narrower. It is to determine whether the model Parliament has in fact chosen passes scrutiny under the Bill of Rights.