For nearly thirty years geographers have been producing models of Latin American urban land use. Beginning with Marvin Baker's insightful four-stage model of Mexican cities (1970), which formed part of his dissertation but which, unfortunately, was never published, and continuing with various efforts by German geographers in the 1970s and early 1980s, most especially Jurgen Bahr and Gunter Mertins, the number of models available grew quickly (Bahr 1976; Borsdoff 1976, 1982; Mertins 1980; Bahr and Mertins 1981; Gormsen 1981). New offerings and modifications of older models have continued to appear through the 1990s. Some designers refer to their creations as models of "urban structure," an ill-defined term the essence of which remains foggy. Whatever the claims, generally models look almost exclusively at land use, and they focus overwhelmingly on residential land use. The standard work among these productions remains that by Ernst Griffin and Larry Ford (1980), which can be found in innumerable introductory and upper-division textbooks in human, urban, and Latin American geography. Strangely, none of the other models, to my knowledge, had been published in any of these English-language texts until that of Daniel Arreola and James Curtis (1993, reprinted in Blouet and Blouet 1997, 178).
That so many divergent models exist means that those who have looked at the scene have viewed it differently. Because models are abstractions of reality, investigators have varied in their choices of what land-use elements should be included. Although every model builder claims the best conception of the Latin American city, dearly no prescription exists for building a land-use model. A model builder must decide on the number of categories of land use to include, how to represent them cartographically, and how to explain their spatial relationships. Visually a model can be either very simple or very complex. The simpler the model the more easily a viewer grasps its thrust, but simple models may also be at a further remove from reality than more complicated ones. The anticipated audience for a model - the classroom versus other academics - also should play a role in design. Finally, cities change, and so too should models of them. The Latin American city of 1998 boasts elements distinct from the Latin American city of 1968 or 1978. We should not remain time locked in model construction (as we have been with North American urban models, or at least as we had been until Michael Dear and Steven Flusty's 1998 breakthrough). If a model offered in 1996 differs from one in 198o it may mean that urban metamorphosis required a new formulation of land-use arrangement. It may also mean that the model's author missed something in 1980.
Griffin and Ford's model, though generally praised, has been criticized on two general grounds: It ignored the work of others published earlier, and it omitted too much. Certainly, part of its strength lay in its authors' familiarity, at the time, with a number of Latin American cities. It was also the first Latin American urban land-use model published in English. It showed how Latin American cities differed from the so-called classical models that had dominated Enghsh-language textbooks for decades but that were based only on limited experiences in the United States.
Those Germans who preceded, and others who followed, Griffin and Ford's publication offered more complex models, which included additional categories and were at times dynamic. Works by Peter Hoffman (1983), David Howell (1989), Peter Ward (1990), Arreola and Curtis (1993), and William Crowley (1995) built on Griffin and Ford's work, as well as other offerings, and acknowledged their contributions while suggesting points of difference. Some models were meant to be fairly specific - Ward spoke only to Mexico City, Arreola and Curtis only to U.S.-Mexico border cities. Others, as in my essay (Crowley 1995), reached for the same level of generality as did Griffin and Ford. …