ARTICLE ORGANISATIONAL CULTURE AND THE LIMITS OF TRANSFORMATION: The Implementation of an Adult Education Programme on South Africa's Mines Penny Vinjevold and Brahm Fleisch The past two years have seen a significant shift in debates about the transformation of state and society in South Africa The focus has moved from concerns about the nature of the class struggle and strategies of opposition to questions about policy options. While this shift is clearly a necessary development, the debate needs to go one step further. We need to begin to theorise and research the problems of implementing transformation. This paper attempts to contribute to this discussion by offering an implementation case-study. Given recent interest in private sector provision of adult education, we set out to examine the implementation of a mining house literacy programme. Ten mines were visited over a six month period in 1991.l Given the limits of our research method, our preliminary findings suggest that the programme was unevenly implemented in the ten sites. While some mines had attempted to follow the spirit of the head office guideline, most had implemented the programme only in a pro forma way. Our findings confirm Timar's hypothesis (1989) that when policy initiatives are imple- mented in decentralised institutional structures, significant local adaptation general- ly occurs. In addition, he asserts that the character of such local adaptations was determined in large part by the organisational culture of the local institutions. Local mine culture and interests, our research suggests, shapes the nature of a particular mine programme. The resulting unevenness in implementation can be linked to current economic and political transition in South Africa From a methodological perspective, the case-study illustrates the usefulness of implementation analysis in the developing field of policy analysis in South Africa. Although beyond the scope of this paper, the research points to the need for third generation policy analysis which links macro and micro level analysis (Lerner, 1986; and McLaughlin, 1987). Context Prior to the late-1970s, mining houses in South Africa did not themselves provide adult literacy programmes. However, outside literacy organisations were en- couraged to provide basic education for mine workers. In particular, the Bureau of Literature and Literacy (BLL) operated in mine hostels. Mine management per- ceived these kinds of organisations as providing a 'welfare' service (Brown, 1988). 36 TRANSFORMATION 20 (1992) VINJEVOLD & FLEISCH ARTICLE By the early 1980s, with the scrapping of job reservation laws, mining houses began actively to promote and finance literacy and numeracy programmes with the specific aim of developing a leadership corp in the migrant labour force. Interest in literacy promotion faded in the mid-1980s only to be revived after 1989. In March 1991, at the biannual manpower conference of one of South Africa's leading mining houses, the chairman announced a corporate objective: by the turn of the century, eight out of every ten employees should be literate. In rather vague terms, the mining house's in-house magazine connected this objective with 'a changing South Africa'. Objectives of Adult Literacy Programmes The publicly stated aims of the Mining House's literacy guideline stressed a number of interconnected goals or objectives: the reconstruction of some form of consensus or culture on the mines; meeting technical requirements associated with current shifts in mining processes; and the development of a system of job promotion and occupational mobility. The objective of literacy training, in the words of the in-house mining magazine, 'is to improve communication at the workplace, improve the quality of life and enable them to further their studies'. In an interview, the Chief Executive of Management Resources stated that the literacy programme was part of the process of creating 'an identity of interest with our employees and building healthy relationships with all the communities with whom we come into contact'. Literacy in his view would 'step up communication in order that employees may improve their understanding and perspective of the progress, goals and problems experienced by the company and get more involved in finding a solution to problems'. Literate employees, proficient in English lan- guage, would enable the company to realise its strategic goals of enhancing employee participation and involvement in the activities of the company. The objective of the literacy programme is also expressed in terms of upgrading technical skills. A mine magazine points out that the existing lingua franca of the mines, Fanakalo, does not have the necessary vocabulary to disseminate technical information. It is not literacy per se that is stressed, rather literacy in English. This commitment to higher levels of English proficiency, a prerequisite for the upgrading of technical skills, is reflected in the policy of newly opened mines. They are beginning to demand a minimum of Standard Six and even a Standard Eight for all employees as a prerequisite for employment The Mining House also connects literacy and English language proficiency to job promotion. The possibility of further technical training for higher paid jobs would make literacy training more attractive for the existing work force. Decentralised Structure of the Mining House The Mining House is part of a larger entity. This larger financial structure has autonomous arms including an insurance company, a forestry company and a large metal processing and manufacturing arm. Under the Chief Executive Officer, the Mining House evolved a highly decentralised organisational system with each TRANSFORMATION 20(1992) 37 ARTICLE VINJEVOLD&FLEISCH business unit administering its own affairs. However, head office retained control over specialised functions, in what they called 'centres of excellence'. Individual mines enter into contracts with head office for services like engineering or human resources. A fee is payable to head office based on labour-hours used by the local mine or division. The top executive believed that these head office centres should be 'competitive suppliers to their clients' - the individual mines within the group. According to the in-house magazine, 'the objective... is to create a competitive market for centrally provided services, which otherwise run the risk of becoming arrogant, complacent bureaucracies'. More generally then, the function of head office is to lay down guidelines which the individual mines must follow. Implementation of these guidelines is left to the discretion of the individual mines. Local mine management jealously guards this autonomy. Within the context of the adult literacy programme, head office provides the guideline, in this case 80% of employees to be literate by the turn of the century. The head office Adult Education Unit is responsible for technical assistance to local mines. The decision to make use of this technical service is ultimately that of the local management teams and the mine manager. The head office Adult Education Unit, a subdivision of Training and Development, which itself forms only a very small subsection of the Division of Management Resources, developed a modular literacy programme which consists of four modules: Breakthrough to Literacy, Survival English, Conversational English, and Operational English. Breakthrough to Literacy teaches reading and writing in vernacular languages, the other three modules are graded English courses. In the initial stage of the implementation of the guideline, personnel from the Adult Education Unit are responsible for disseminating information about the head office programme to the managements at the various mines. Although the local mines are under no obligation to purchase or use the head office's literacy programme, none of the mines in the company had, at the time the research was conducted, adopted an alternative literacy programme. Not only did the head office Adult Education Unit develop the programme, it provides assistance with the implementation of the programme if requested to do so by individual mines. Implementation of the Adult Literacy Programmes Once a local mine management has undertaken to implement the literacy programme, the Manpower Department selects a supervisor whose responsibility it is to coordinate and monitor the programme. The supervisor hires tutors, who are then sent on a week long training programme. The tutors return and together with the supervisor, select students for their classes and begin teaching the Adult Educa- tion Unit package. Reasons for Uneven Implementation There are two broad areas in which the literacy programmes were particularly uneven - supervision and tuition. 38 TRANSFORMATION 20 (1992) VINJEVOLD & FLEISCH ARTICLE • Supervision At all ten mines visited, the Labour/Manpower Department was assigned respon- sibility for the literacy programme. At seven of the ten mines, full-time adult education coordinators were appointed to supervise the programme. At the other three mines, existing Manpower/Labour personnel were assigned the responsibility. All literacy supervisors who were selected were white and six of the seven full-time coordinators were ex-teachers. Different factors seemed to determine the appointment of supervisors. First, decisions concerning the appointment of supervisory personnel for the literacy programme appeared to have been dictated in no small way by the financial position of the individual mine. The platinum mines and two of the gold mines in the sample were profitable, while the coal and silicon mines were under considerable financial stress. The gold and platinum mines appointed full-time adult education coor- dinators; the coal mine that appointed a full-time coordinator appears to have done so in name only, as the appointee was often called away from her work, including her teaching, to perform other duties. At the financially-troubled coal and silicon mines, the Manpower and Labour personnel had been made responsible for im- plementing the literacy programme. In interviews, the supervisors from these mines claimed that their other pressing responsibilities prevented them from becoming meaningfully involved in the literacy programme. Second, on a number of mines, personal interest in the promotion of literacy influenced the appointment of supervisory personnel. At a platinum mine, the manpower manager, a self-styled liberal, took great interest in the literacy programme and ensured that the largest portion of the human resources budget was allocated to this programme. This meant that the mine was able to provide for an additional literacy administrative assistant to the coordinator. In contrast to this situation, another platinum mine's manpower manager was a sports enthusiast He allocated significant amounts of the discretionary funds to mine-sponsored sports activities to the detriment of the literacy programme. At this particular mine, the adult education coordinator was also the sports coordinator. This individual appeared to have spent much of his working hours on sports administration. Third, the choice of personnel for supervision of the literacy programme seemed to have been influenced by mine managements' tendency to equate literacy classes with schools. Six of the seven full-time coordinators are ex-school teachers, despite the fact that the role of the coordinator was to administer rather than to teach on the literacy programme. A fourth factor that influences the choice of supervisory personnel was the accepted, but unwritten practice of the mines of not appointing black employees to managerial positions. All ten supervisors on the mines visited were white. The six coordinators who were ex-teachers supervised black tutors who often had far greater teaching experience than their superiors. The manpower managers who were inter- viewed about this practice claimed that this was not discriminatory but that blacks were not yet ready to be managers. Literacy supervisors have the following tasks: budgeting, planning, tutor selection, TRANSFORMATION 20(1992) 39 ARTICLE VINJEVOLD & FLEISCH monitoring, evaluating, and reporting. For Weatherly and Lipsky (1977; 1980) lower level organisational managers, who they refer to as 'street level bureaucrats' have considerable discretionary powers, and find ways of accommodating policy direc- tives within existing institutional cultures. These bureaucrats are forced to 'remake' policy directives to meet broader organisational priorities. The notion of a street level bureaucrat helps us to explain the organisational context which underlies the unevenness with which the literacy programme was implemented. The supervisors' expertise, terms of employment, authority, budgetary constraints, their ideology or value system as well as that of the particular mine's culture all influence this process of 'remaking'. Many of the supervisors had very little expertise with regard to the literacy programme. In interviews with the supervisors they generally displayed an incom- plete understanding of the principles, content, and methodology of the literacy programme. The only avenue to acquire any level of expertise in the use of this particular programme was through attendance of the tutors' training course. Five of the ten literacy supervisors had not attended either of the tutors' training courses, and none of the supervisors had attended a Breakthrough to Literacy (BTL) course. Four of the five supervisors claimed that they could not afford to be away from work for a week because of their other responsibilities. It is, however, possible that the supervisors' lack of interest or lack of regard for the literacy programme was the reason they had not attended these courses. In addition to having a limited understanding of the literacy programme, the supervisors generally did not speak the languages used in the programme. None of the ten literacy supervisors spoke the vernacular languages used in the BTL courses and only one was a mother tongue English speaker. This not only militated against the supervisors teaching on the literacy programme, but hampered their ability to monitor and evaluate the programme. The appointment of literacy supervisors who did not speak the languages of the literacy programme suggests, at best, a lack of understanding of the programme, and at worst, a lack of regard for the programme. The supervisors' terms of employment discouraged them from directly supervising the programme. All the supervisors of the literacy programme were appointed as management personnel. In terms of their employment, management personnel were required to work regular office hours. The vast majority of the literacy programmes were conducted after office hours. Although there were no obstacles to prevent supervisors from attending the afternoon and evening classes this would have constituted unpaid overtime work. At only two sites was there any after hours direct supervision. On one site, the coordinator regularly visited the afternoon classes and another visited the after hours classes once a week. The first mentioned coordinator had negotiated a two hour midday break to compensate for his late-afternoon visits. The second coordinator had been given the use of a company vehicle on the night she visited the literacy classes. Lack of direct monitoring did not affect the implementation of the literacy programmes in any uniform way. At the mine where the supervisor regularly visited the literacy classes the number of BTL classes dropped from 16 to four in the six 40 TRANSFORMATION 20(1992) VINJEVOLD & FLEISCH ARTICLE weeks that he was away from work. The number of BTL classes returned to 16 within two months of his return. However, on two other mines, where there was no supervision and little management interest in the literacy programme, the programme ran relatively regularly and efficiently. Although the effects of the lack of adequate monitoring were uneven, programmes were generally better when there was regular and systematic supervision. The lack of direct supervision meant that supervisors relied almost entirely on the integrity of tutor reports for planning, administering and reporting on the literacy programme. Although it was beyond the scope of this study to estimate the prevalence of inaccurate reporting on the part of tutors, the researcher found that reports submitted by tutors on after hours classes rarely corresponded to actual practice. The areas of discrepancy included the attendance numbers and the rate of progress made through the literacy modules. The fact that many of the literacy supervisors colluded with the tutors by submitting these reports to mine management can partly be explained by the fact that tutors and supervisors were both judged by their superiors on the number of students who pass through the literacy programme. The following vignette illustrates collusion between supervisors and tutors. At the mine, the Human Resources manager had been assigned responsibility for the literacy programme. In an interview in the initial stages of the research, he indicated that he had little time to devote to this aspect of his job. During a site visit to the mine, the Human Resources manager informed the researcher that there were three literacy classes at the mine hostel - one Survival English class, one Conversational English class, and one Operational English class; and that there was a total of 39 students in these classes. He expressed concern about the relatively low number of students in the literacy programme. The researcher subsequently conducted inter- views with the tutors at the mine hostel. The researcher was informed by both the literacy tutors and a black labour officer that many workers were keen to attend literacy classes but there were not sufficient tutors or classrooms for extra classes at the time. Both the tutors and the labour officer claimed that there were in fact 54 students who regularly attended literacy classes but that the tutors' attendance was irregular partly because tutors lived in the township some distance from the hostel and transport was scarce when classes finished. At the end of the interview the researcher requested permission to observe all the classes. When the researcher arrived at the hostel for a 16h00 to 17hOO class, she found only seven students present. Each student had a Operation English Book open to Lesson One. There was then some dispute between the three tutors present as to who would teach the class. One said that he had taught that morning and so it was not his turn. One said that he had not taught that year and the third explained that it was not his turn that day. One tutor then volunteered to teach, and taught the difference between the present simple, and the present perfect tenses. After half an hour another of the tutors took over and explained an exercise on page eight of the Operation English course book. It soon became clear from the students' responses that they were not being taught the appropriate literacy module. The course had, according to the tutors' and the management records, started on 15 October 1990, and so should have been com- TRANSFORMATION 20(1992) 41 ARTICLE VINJEVOLD & FLEISCH pleted in March 1991. Instead, lesson one was being taught in April 1991. In a meeting with the tutors after the lesson, the researcher enquired about the Survival English course and the Conversational English course and was told that there was no Survival course and the Conversational course had been suspended for a month while the fourth tutor was away on leave. When the researcher enquired why one of the other tutors had not taken over the class, the tutors admitted that the class had never begun. The researcher met the Human Resources manager the following day; he indicated that he had suspected that very little was happening in the literacy classes. Because of his suspicion he arbitarily reduced the student attendance figures. This 'appearance management' emerged as a response to the enormous pressure experienced by the supervisors. A simple calculation tells them how many workers must receive literacy training per year over the next eight years to achieve this goal. At all ten mines visited, the number of workers trained annually was well below the number required to reach the corporate objective. Supervisors had to make difficult budgetary decisions. One key set of allocation choices related to the number and type of tutors employed. In part, the financial position of the mine seemed to have had some influence on the options. Supervisors at the four profitable platinum mines and one of the gold mines appointed only full-time tutors. Budgetary considerations also seemed to influence the level of academic qualification of those who were employed to teach. In the mid- to late-1980s mines employed ex-teachers as full-time tutors. At that time teachers' salaries were well below what the mines could offer. In the last four years, teachers' salaries have risen considerably. By 1990, teachers' salaries were above what any of the mines was prepared to pay tutors. In the last two years only mine employees have been recruited to teach on the literacy programme. The supervisor's value system was an important influence on the selection of tutors. The two qualities that supervisors identified as important in their selection of tutors were proficiency in spoken English and trustworthiness. In practice the latter was of much greater importance. Despite the fact that the Adult Education Unit recom- mended that only those candidates who had passed a specific language proficiency test be sent for the tutors' training course, interviews with the course presenters revealed that supervisors often sent candidates who did not have the language skills to teach these courses. There was also evidence that the mines did not select the best candidates available. On at least two mines visited candidates who received high scores on the language proficiency test were rejected because they were trade unionists or 'troublemakers'. The selection of supervisors and system of supervision resulted in uneven im- plementation of the literacy programme on the ten mines. First, the choice of supervisor was not determined by educational or administrative expertise, but rather by budgetary, racial, and parochial factors. Second, supervisors did not have, nor demanded, adequate authority, time, resources and access to expert knowledge. The choice of supervisors and the support given to the literacy programme reflects the low status of literacy programme on the mines. 42 TRANSFORMATION 20 (1992) VINJ EVOLD & FLEISCH ARTICLE • Tuition This section examines the selection of tutors, the training of tutors, classroom practices and the tutors' remuneration and working conditions. These factors directly affect the implementation of the literacy programme. The majority of tutors were drawn from the pool of mine employees; only a small minority were working as educators before assuming positions as tutors on the literacy programme. Interviews with the tutors revealed that they were 'approached' to teach on the programme. Although some were genuinely interested in doing this type of work, most felt obliged to participate because of the status of the person who approached them, and the way in which they were approached. This is illustrated by the procedure followed on one mine. On this mine, all 59 employees who had completed Standard Nine or Ten were, without warning or explanation give a language proficiency test. The labour manager then interviewed the 20 employees with the highest scores, and chose ten to be sent on a tutors' training course. At the beginning of the course all ten indicated that they did not know what the course was about or why they had been sent on it. Once a tutor had been engaged he was sent on the tutors' training course. Tutors attended a week-long course for Break Through to Literacy (BTL) programme and another week-long course for the English modules. At the end of the course the presenters sent reports to the relevant mines on each of the participants. Ap- proximately 20 percent of the tutors who were teaching on the literacy programme were not considered competent to teach by the course presenters. Nevertheless the supervisors engaged them as teachers presumably because of the high cost attached to tutor training (approximately Rl 000 per tutor.) The company which trained tutors did not provide follow-up services but the Adult Education Unit provided this service. The cost of this service is not borne by the individual mines. The lack of interest in this service may be explained by the fact that tutors often did not know this service was available and supervisors did not know that in-service support was necessary. In addition supervisors may not have wanted to expose their programmes to Head Office scrutiny. The actual teaching practice and methodology of the full-time tutors was of an uneven quality. Some of the tutors observed during the course of the site visits had adopted the communicative, participative approach to language teaching advocated by the training courses. But many reverted to teaching methods they previously practised or experienced as learners. In many instances this involved long explana- tions at the chalkboard or endless drills. Other tutors adopted some of the advocated teaching procedures but did not carry through the entire process. Although it is beyond the scope of this study to provide an analysis of the efficacy of the methodology, research (Sabatier and Mazmanian, 1980; Johnson and O'Connor 1979) cautions the policy analyst about assuming that teachers, in this case the tutors, were 'resistant to change' or were just lazy, ignorant or subversive. Researchers have recognized that educators' uneven responses to new methodologies may be inter- preted as tutors' best efforts to do their job and to provide the best they can for their students. Failure to follow a new methodology may signal their assessment that new TRANSFORMATION 20(1992) 43 ARTICLE VINJEVOLD & FLEISCH practices are not as good as the existing ones given the many limitations imposed on them in the particular learning situation. The researcher found that student absenteeism and dropout rates were not related to tutor practice in any direct or simple way. Many factors contributed to these rates but the role of the tutor was not negligible either. Tutors who consistently had high attendance rates appeared to have taught regularly and moved at a steady pace through the literacy programme. The particular type of teaching practice observed may partially be attributable to tutor's low commitment to the literacy programmes. Interviews with tutors revealed that they were unmotivated and disgruntled. Two factors contributed to tutors' attitudes, their working conditions and remuneration. Working conditions appeared to have affected teaching practices on the literacy programme. With the exception of two mines which employed full-time tutors, the mines expected their tutors to clock in for the mine's office hours. This meant that on three mines the full-time tutors were 'on-duty' for two to four hours longer than other mine employees, including the literacy supervisors. On one of the mines there was no teaching during office hours while on four mines the full-time tutors taught one class during office hours. It is difficult to ascertain what the tutors did during all the hours that they were at work but not teaching. Most tutors indicated that they prepared lessons, marked students' work and completed administrative work. The adult education coordinators claimed that there was not much more than half an hour's administrative work per day. The researcher's observation of the tutors' lessons suggests that little time was spent on the preparation of lessons and marking students' work. By the time after hours literacy classes started the full-time tutors had been at work for eight hours with little to do. Part-time tutors had a different set of problems. They worked in their regular mine jobs from 07h30 to 15h30 and are generally exhausted when they began teaching at 16h00 or 16h30. Remuneration was another source of tutor grievance. Full-time tutors on the mines earned between R1700 and R2700 a month. They seldom received merit increases and did not receive bonuses. In many cases they were expected to be at work for more than ten hours a day but were informed that they were not eligible for increases because they only worked, that is taught, for a few hours a day. There was also no opportunity for promotion for full-time tutors. Theremunerationof part-time tutors was also a source of considerable controversy. At a silicon mine, the mine secretary and labour manager believed that their tutors were providing a community service and so they should not be paid. The Adult Education Unit has persuaded them that the tutors should be rewarded in some way and the two tutors were given a R100 bonus every month. At one of the gold mines all the tutors were paid a standard overtime rate of R7 per hour. At other mines tutors were paid according to the overtime rate of their job category. In other words, a Category 1 employee who tutored on the literacy programme earned as little as R2 an hour while a Category 8 employee who taught in the same programme could have earned in excess of RIO per hour. In sharp contrast, the one mine which employed 44 TRANSFORMATION 20 (1992) VINJEVOLD & FLEISCH ARTICLE white part-time tutors paid all its tutors substantially more than the other mines, that is, R30 per hour. Part-time tutors' long working hours and the low remuneration clearly affected the quality of teaching on the literacy programmes. A small minority of tutors had left after a month or two of teaching. Many others showed their dissatisfaction through frequent absenteeism or by simply not teaching when they were present Tutor selection, inadequate tutor training, the poor working conditions and remuneration of tutors all influenced the implementation of the literacy programme. All these factors indicate that local mine managements have not made literacy a priority. Despite the public relations hype that some mines have given their literacy programme, they have not paid it the attention they would technical training. Why? One possible explanation is that a comprehensive and successful literacy programme would fundamentally alter mine culture. Successful implementation would require a significant reallocation of resources, changes in the shift system, and even the structure of the managerial hierarchy. While we did not encounter any overt opposition to literacy training, it is clear that local management is not ready or willing to undertake the structural changes that the successful implementation of a literacy programme requires. In fact, such structural changes may run counter to their long- term interests. Conclusion This study attempts to document and analyse the uneven implementation of a private sector literacy programme. Although the evidence presented clearly points to the influence of local organisational culture in implementation, the mine cultures themselves need to be situated in a broader context. The unevenness with which the literacy programme was implemented may reflect an underlying contradiction in South Africa's mining industry in this period of transition. There is some indication that corporate leaders are attempting to shift the organisa- tional culture of South African mining from 'low skill, low equilibrium' (LSE) to 'high skill, high quality equilibrium' (HSE). In South African terms this means a shift from baasskap: unskilled work, race-based classification of jobs, and a migrant labour system, to a semi-skilled/skilled work force that is resident with families in privately-owned homes. The mining houses seem to be speaking about a new set of social relations on the mines. In the new organisational culture there will be greater cooperation between the representatives of labour and management One can only speculate that this shift in social relations is a response to increasingly effective trade unions, falling mineral prices, and international management trends. In the new political and economic climate, mining would be skilled work, mine workers would receive on-going technical training, and job promotion. A new atmosphere of cooperation would emerge between the lean and sophisticated work force and mine management An important component of the new mine management discourse and style is an emphasis on decentralisation. The trend towards greater local autonomy is aresponse TRANSFORMATION 20(1992) 45 ARTICLE VINJEVOLD & FLEISCH to the criticism that mine management is rigid and autocratic rather than individualis- tic, innovative and imaginative. But in the South African context, decentralisation of mine organisations translates into the perpetuation of existing local mine cultures. In this situation, the failure of nearly all of the ten mines' literacy programmes to realise the corporate goal of educating the work force may be interpreted as local resistance to fundamental organisational change. On individual mines, lower level mine management and white working class' interests run counter to the new corporate objectives. An effective literacy programme in conjunction with technical training and ability-based promotions - key elements in the new social relations of production - would inevitably undermine entrenched white privilege. The weak- nesses of the literacy programme are thus attributable not necessarily to self-con- scious white opposition, but rather to the recalcitrant nature of organisational cultures on the mines which have historically maintained white working class privilege. We acknowledge that the above argument remains at the level of conjecture: future research is needed to link local-level implementation research to broader structural changes. The problem South African policy analysts face, in this period of transfor- mation, is the need to link the lived world of the supervisor/bureaucrat and local organisational cultures to systemic processes and macro-economic shifts. NOTES 1. To achieve a cross-section of mines within the mining house, the following factors were used as selection criteria: the relative profitability of the mine, the type of mineral mined, the level of union activism, and the perceived success of the programme. Although the mines were not chosen at random, we have no way to determine possible biases in the sample. Another limitation of the study is that much of the evidence from the site visits is anecdotal and based on personal observation. Many of the interviewees' perceived the researcher to be from head office. 2. To protect the identities of individuals we have chosen not to reveal the name of the mining house. REFERENCES Brown, D (1988) - 'The basements of babylon: language and literacy on South African gold mines', in Social Dynamics, 14(1). lipslcy, M (1980) - Street-level bureaucrats (New York: Russel Sage Press). Johnson, RW and RE O'Connor (1979) - 'Infra-agency limitations on policy implementation: You can't always get what you want, but sometimes you get what you need', in Administration and Society. Kraak, A (1991) - Beyond the Market: comprehensive institutional restructuring of South African labour markets, education, training and work (NEPI Human Resources Group). Marshall, C (1988) - 'Bridging the chasm between policymakers and educators', in Theory into Practice, 27. McLaughlin, MW (1987) - 'Learning from experience: lessons from policy implementation', in Educational Evaluation and Policy Analysis, 9. Sabatier, P and D Mazmanian (1980) - "The implementation of public policy: a framework of analysis' y in Policy Studies Journal, 8. Timar, T (1989) - 'A theoretical framework for local responses to state policy: implementing Utah's career ladder program' in Educational Evaluation and Policy Analysis, 11. Weatherly, R and M Lipsky (1977) - 'Street-level bureaucrats and educational innovation', in Harvard Educational Review, 47. 46 TRANSFORMATION 20(1992)