If you browse the geography student textbooks of the 1960s and 1970s, you may be surprised now to see how many were focused on the broad subject of ‘location analysis’. The classic works of Haggett (1965), Harvey (1969), Haggett et al. (1977) and Wilson (1970; 1974) are perhaps the most cited of this group and were the core readings in many geography departments at this time. They built on classic foundations that now seem to be rarely taught at undergraduate level— Christaller, Lösch, von Thunen, Weber, Palander, Reilly, etc. Although often perceived as theoretically and mathematically complex (Peter Gould (1972) termed Wilson’s 1970 modelling text as ‘the most difficult I have ever read’), they were arguing for an essentially applied geography—the use of these models to help industrialists, farmers, land-use planners, retailers, etc. to find optimal sites to locate and reach their markets given spatial variations in the cost of various factors of production (remember von Thunen himself was a farm estate manager interested in improving the efficiency of production and distribution). Unfortunately, this type of economic geography rapidly fell from grace. As Pacione notes in the introduction to this volume, in relation to geography as a whole, the moral high ground was captured by ‘new’ geographies that emphasised the wider theoretical frameworks concerned with underlying capitalist social relations and shifts in national and international political economies at the expense of location models per se (see Massey and Meegan 1985 for a good illustration). All the sub-fields of economic geography were subsequently influenced by this theoretical shift. Retail geography, for example, provides a good illustration of this transformation. Despite notable exception (see Fenwick 1978, Davies and Rogers 1984, and Ghosh and MacLafferty 1987 on site location methods), the number of new texts on quantitative methods for site location rapidly diminished, while newer retail geographies based on culture, consumption and capital came to the fore (Wrigley and Lowe (1996) provide the best illustration). While these new retail geographies are in themselves important additions to the literature, they deliberately omit location models, GIS, etc. as if they are no longer important. Smith (1989) summarises this ‘progress’: ‘instead of trying to identify the optimum site for a new supermarket many geographers turned to identify the broader processes in which whole landscapes were made and remade’ (p. 142). Research on site location was now seen as politically incorrect. Clarke (1996) provides some explanation for this. He suggests that GIS and models produce knowledge that proves most useful to those already powerful groups in society within a descriptive neo-classical economic geography still largely underpinned by central place theory (CPT).
It could be argued that much of this damning critique is simply wrong. First, models can be