'Darwinian Gastronomy: why we use spices' (1999, PDF-link) by Sherman and Billing is a nice and early example of an attempt to use statistical analysis of cookbooks to reveal deeper patterns about what we eat and why. The paper theorizes that there is an evolutionary benefit to eating spices: "by cleansing food of pathogens before consumption, spice users contribute to the health, longevity and fitness of themselves, their families and their guests." There is more disease in the tropics and this is also where most spices are added to food, or so the paper seems to argues. Personally I think the argument runs the risk of putting the horse behind the carriage. Spices predominately grow in tropical areas and it makes sense to expect that this is where they eat them most.
There are some very nice graphs showing spice-use in 36 countries and I can easily appreciate how much data (93 cookbooks) went into making them. But we are never told what those books were. How can someone who knows about these things ever make a judgement on how soundly food traditions/cuisines are represented? A cookbook is not a neutral source, but a vehicle of someone's dietary ideology. Something advertised as traditional may be less than a week old, something advertised as national may be produced by a radical fringe. What and where is the baseline? The paper, in its conclusion, recommends cookbooks "from different eras" as "a written record of our coevolutionary races against foodborne diseases." This I seriously doubt: the history of cookbooks is not old enough to pick up deep evolutionary changes. If we do eat spice as an antidote to unclean food we, in the West, could do without them, our ever growing love for spicy food shows the reverse.
A related paper explains why "vegetable recipes are less spicy" (PDF-link) from which we learn a debatable assumption made in the other paper: it treats fish as a meat.