In a nineteenth century novel, a character says "For your sake I have endured the Nibelungen Cycle from soup to nuts—or I should say, after Horace, 'from the egg to the apple.'" Horace's Latin phrase, ab ovo usque ad mala, was shorthand for the standard Roman table service, as the American phrase "soup to nuts" sums up an older American idea of the standard European service, which for a very long time has begun with soup. If it did not always end with dessert, it was more likely to include (after the dessert) smaller treats such as the medieval French candies known as draguées and the American substitution (for the kind of very large meal rarely served today) might well have been nuts.
If the two phrases are symmetrical, it should also be clear that they refer to different orders - which is to say that somewhere between the Romans and a later time, the European idea of order in table service changed. Which of course raises the question: how?
The Roman courses began, not only with eggs, but lettuce and other light foods, meant to stimulate appetite; this course was called the gustatio or promulsis. The main course was called the coena, which was also the name of the meal (typically in the evening) which was most likely to be divided into courses. This however could be divided into a number of fercula, a term which referred both to the service and the platter on which it was brought in (ferculum). Typically there were between three and six of these; Juvenal speaks pointedly of a man who had seven when dining privately
Or which of our Forefathers far'd so well,
As on seven Dishes, at a private Meal?
The last principal course was the mensae secundae ("second tables"). which was essentially dessert, including cakes, sweets and fruit.
In Gaul, even after the Franks took over, there are hints that some version of the Roman order survived; not in the eggs, but in the idea of having something green - except that the lettuce became vegetables. The poet-bishop Fortunatus describes one meal which began with "vegetables, drizzled in honey". As a member of the high clergy, Fortunatus lived in a Gallo-Roman, rather than a Germanic, milieu. He also lived luxuriously. But Gregory de Tours describes a Gallo-Roman couple, humble enough for the wife to be cooking, serving a meal in several courses. Again, the first dish is of vegetables.
This is a simple enough detail. But even this distinguishes the order from those found later, when vegetables were rarely served as a first course. (Eggs and lettuce seem never to have reappeared.)
Later in the early medieval period, it is virtually impossible to know what, if any order, was used. Charlemagne's first biographer says that he ate four courses (or possibly dishes) and a roast, but provides no further details. This at least tells us that meals continued to be served in separate courses, at least among people of rank. Einhard also makes it plain that this was the Emperor's daily fare and intends this to show Charlemagne's personal simplicity; for greater occasions, he knew how to impress.
During Charlemagne's own lifetime, an astounding figure, Abu l-Hasan ‘Ali Ibn Nafi‘ (789-857), came from Iraq and north Africa to Spain (then al-Andalus). Known to history as "Ziryab", he is credited with a long list of cultural innovations, including laying the basis for flamenco and establishing seasons in fashion. By some accounts, he also set the order for courses in a meal and this innovation ultimately made its way to Europe, though only much later and probably by the roundabout route of post-Crusade influence from the Mideast. However, as will be seen, this claim does not square with the available evidence.
Meanwhile, for several centuries after Charlemagne, no specific account of a European meal survives. The first may come from the twelfth century, when a bishop outlined a special feast for monks. But this order (which was probably not typical) bears no hint of those that would become standard later. It begins with hams, feet and a head of pork, followed by sausages and other parts made from pork offal, then smoked beef on a bed of a cabbage; two types of bacon, peppered, roasted and grilled pork, a young pig garnished with game meats, fat bacon with strong mustard, then a plate of millet with eggs, milk and pig blood, and ends with a roasted and larded shoulder of pork.
This looks very much like a "baconic", a meal later said to have been made up entirely of pork (though the word does not appear in period accounts). But even if that was the intent, there would certainly have been ways to adapt it to the later standard courses, such as beginning with a pork soup, especially since this version makes room for millet and smoked beef. It seems, rather, that that order was not yet standard at this point.
Otherwise, until the fourteenth century, the sequence of courses can only be deduced from the fact that in 1294 Philip the Fair found it necessary to limit people to having a dish and an entremets (itself a new concept) for an ordinary meal and two with a bacon soup for a major meal. If this implies that people were already having far more courses, it also shows that he thought it reasonable to restrain diners to the point that a three or four course meal would have been impossible. Nor did he seem to think soup an essential course in an ordinary meal.
The first thorough look at medieval courses comes in the fourteenth century Menagier de Paris, which includes a list of twenty four menus of meals "for great lords and others". Right off, then, it is clear that these were not for standard meals. The most striking aspect of the menus is that not only do they follow nothing like the later order in European dining, they differ greatly between them. Most begin with a heterogeneous assortment of dishes, often including one or more pies and only sometimes including soup. Here (in Janet Hinson's translation) is the first course for the second dinner (today, lunch): "Pies of veal chopped small in grease and marrow of beef, pompano pies, black-puddings, sausages, forcemeat, and rich pies de quibus". And for the second supper: "Capons with herbs, a cominy, daguenet [?Danish (JH)], peas, loach in yellow sauce, venison in soup."
The food historian Jean Louis Flandrin examines these menus closely in his L'ordre des mets (in English, "Arranging the Meal: A History of Table Service in France"). He tries to show that there is a certain logic to these seemingly random sequences; notably that the roast always plays a central role and that the courses themselves tend to be symmetrical in number of dishes. It is a valiant, but unconvincing, effort. The fact remains that a modern cook would be hard put to organize a meal using the structure of these as a guide.
The work offers two far grander menus after the numbered ones, and these at least name the first course "soup", though the dishes listed do not obviously match that description; for instance, "capons in fricassee, pomegranate and red sugared almonds on top." But this at least suggests that the later European order was taking shape in this period, even if it was far from established.
Another view can be deduced from the early manuscripts of the Viandier of Taillevent, the first really famous cookbook. These start with a chapter on pottages. Today potage has has the same meaning as soupe in French. But in this period it was a much thicker, more complex production. Was beginning a meal with pottage the same as beginning it, later, with soup? One can at least see a linear evolution, especially since the next chapter is on roasts, followed by one on entremets (which Prescott translates, reasonably enough, as "subtleties"). Since Taillevent's work was intended for the highest ranks of society, it is possible that the grandest meals were (as in the Menagier) beginning to follow the sequence which would later become standard.
For another look at courses in the fourteenth century, we can also turn to a curious document from Humbert, the Dauphin of Vienne. In a moment of financial difficulty, Humbert carefully laid out what would be served at each of the meals in his household. He probably ate more ornately when better off, but a restricted meal for a Dauphin would still have been a good one for the time. For Mondays and Wednesdays, then, he wanted "a purée of peas or of broad beans, with two pounds of salted pork; then good tripes, cooked in water. For the second service, two portions (rotulos) of beef and mutton, boiled and served with a warm pepper sauce; and, as a roast, six capons, or six fat hens." Dessert was of cheese and fruit (which would largely remain the case for centuries and was not, in fact, too far from the Roman model - except that pastries would take some time to return.)
For Saturday, he did in fact begin with two soups with (again) a purée of broad beans and almonds, seasoned with onion juice and olive oil; fish, "if there is any"; twelve poached eggs, with a good sauce; tarts of greens, and eight hard-boiled eggs.(Note that here the eggs have moved beyond the initial place assigned to them by the Romans.) But for his supper (every day?), he wanted only "a half-portion of roast beef; beef feet, prepared in vinegar with parsley; and grilled beef tongues, with cameline sauce." It seems then that soup at the start of a meal still remained only one of several possible options.
The overall impression left by all this data is that as the Roman sequence of service declined in France, at least, it was not replaced for a long time by anything particularly organized. The one clear tendency was to eat far more meat and fowl - even the heartiest Roman meals would have been far more varied than the later meals seen here. Finally, in the fourteenth century, the idea of starting with soup (or pottage?) began to take hold among the grand and, as often happens with fashion, later made its way to the less exalted ranks of society. (To complicate matters still further, even in later centuries, the sequence sometimes shifted. But that is another story,)
Can these first glimmerings in fourteenth century France be traced back to Ziryab? A thirteenth century work defines eight courses for meals in Arab north Africa and Spain, and these do not specifically begin with soup. In the East itself, apparently courses were all served at once in this period. So it seems unlikely that the Western sequence had its roots in a model from either side of the Arab world. But the question offers fertile ground for future exploration.
FOR FURTHER READING
Gregory (st, bp. of Tours.), Les livres des miracles et autres opuscules de Georges-Florent Grégoire
Charles Gérard, L'ancienne Alsace à table: étude historique et archéologique sur l'alimentation, les moeurs et les usages épulaires de l'ancienne province d'Alsace 1877
Historische und Antiquarische Gesellschaft zu Basel, Basel im vierzehnten Jahrhundert:Geschichte Darstellungen zur fünften saculargeier des erdbebens am S. Lucastage 1356 herausgegeben von der Basler Historischen Gesellschaft
Charles Perry, "The Medieval Arab Meal, East and West", The Meal: Proceedings of the Oxford Symposium on Food and Cookery, 2001 edited by Harlan Walker