Twenty years ago, when I first came to South Africa, I worked at an advice center called the Industrial Aid Society (IAS) on Jeppe Street in downtown Johannesburg. This was around the time of the first democratic elections, but before post-apartheid labor law reform began in 1995. Every early morning, long queues and quiet clusters of workers waited outside the door. We would serve some 50 workers a day. If a worker was a member of a union, we would send her back to her union to resolve her issues. The paralegals accompanied workers to what were then known as industrial councils to sit framed by polished wood and to argue for their protections. If the worker did not fall within an industrial council, the advice office would phone the employer to negotiate the basic conditions laid out in a wage determination, or fill out the paperwork for the Unemployment Insurance Fund, or tick the boxes detailing the body parts lost or injured for occupational health benefits. It was through working at this advice center for three-and-a-half years that I realized how mutually constitutive precarious labor and regulation were.

In 2012, I returned to another worker advice center, the Casual Worker Advice Office (CWAO) in Germiston. It was set up specifically to assist casual, contingent, and contract workers. In the period between the IAS and the CWAO, South Africa has undergone two decades of democracy with labor law reform a primary lineament of expanded protection and participation (Adler and Webster 2000; Webster and Von Holdt 2005). Indeed, employment has been definitive of the ANC's (African National Congress) imagination of social citizenship in post-apartheid South Africa (Barchiesi 2011).

Yet while labor reforms extended the definition of employee to many workers previously not covered, it also structured labor protections around a set of normative assumptions about who that worker was: ‘a male, permanent, full-time, long-term employee with employment benefits (medical aid and pension fund) working at one fixed workplace where formal representation by a nationally organized union is available’ (Page-Shipp 2003: 11; see also Kenny 2009; Theron 2005). Many precarious workers did not fit into this model. Moreover, basic protections that did apply were simply de facto subverted by employers or not pursued by unions.