Bread and Class in Medieval Society

Journal of Interdisciplinary History, XLVIII:3 (Winter, 2018), 335–357.

Adam Izdebski, Marcin Jaworski, Handan Üstündağ, and Arkadiusz Sołtysiak Bread and Class in Medieval Society: Foodways in Anatolia Since the time of Braudel, most historians have come to recognize the close link between social status and type of food consumed. As Braudel wrote about Mediterranean life, “The study of the grain problem takes us . . . to a greater understanding of that life in all its complexity.” The role that food and diet played in shaping the natural environment has subsequently developed into a major theme in environmental history. Further- more, economic historians have focused on human biological and

Adam Izdebski is Assistant Professor of Byzantine and Environmental History at the Institute of History, Jagiellonian University. He is the author of A Rural Economy in Transition: Asia Minor from Late Antiquity into Early Middle Ages (Warsaw, 2013); co-author, with Karen Holmgren et al., of “Realising Consilience: How Better Communication between Archae- ologists, Historians and Natural Scientists Can Transform the Study of Past Climate Change in the Mediterranean,” Quaternary Science Reviews, CXXXVI (2016), 5–22. Marcin Jaworski is an alumnus, Department of Bioarchaeology, Institute of Archaeology, University of Warsaw, and a freelance archaeologist. He is co-author, with Piotr Wroniecki, of “Magnetic Survey of the Abandoned Medieval Town of Nieszawa,” Archaeologia Polona, LIII (2015), 85–94; with Agnieszka Tomas, of “Non-Destructive Archeological Investigations in the River Sárviz Valley (Hungary), 2012,” Światowit, X (2012), 171–175. Handan Üstündağ is a faculty member, Department of Archaeology, Anadolu University. She is the author of “Human Remains from Kültepe-Kanesh: Preliminary Results of the Old Assyrian Burials from the 2005–2008 Excavations,” in Levent Atici et al. (eds.), Current Research at Kultepe-Kanesh (Atlanta, 2014), 157–167; co-author, with Marcin Jaworski and Arkadiusz Sołtysiak, of “Continuity and Change in Cereal Grinding Technology at Kültepe, Turkey,” Archaeological and Anthropological Sciences, IX (2017), 447–454. Arkadiusz Sołtysiak is Professor of Bioarchaeology, Institute of Archaeology, University of Warsaw. He is the author of “Cereal Grinding Technology in Ancient Mesopotamia: Evidence from Dental Microwear,” Journal of Archaeological Science, XXXVIII (2011), 2805–2810; “Ante- mortem Cranial Trauma in Ancient Mesopotamia,” International Journal of Osteoarchaeology, XXVII (2017), 119–128. This article presents research that Adam Izdebski undertook from 2012 to 2015 during his postdoctoral fellowship at Jagiellonian University, Krakow—funded by the National Science Centre, Poland (DEC-2012/04/S/HS3/00226)—combined with dental microwear data compiled by Marcin Jaworski. The authors thank Abdullah Deveci (Anadolu University), the director of the excavations in Akarçay Höyük; Zeynep Mercangöz (Ege University), the director of the excavations in Kadıkalesi; Akın Ersoy (Dokuz Eylül University), the director of the excavations in Smyrna Agora; and Oluş Arık (University of Ankara), the former director of excavations in Alanya Kalesi, for their support of studies examining human remains from the sites. © 2017 by the Massachusetts Institute of Technology and The Journal of Interdisciplinary History, Inc., doi:10.1162/JINH_a_01161

Downloaded from http://www.mitpressjournals.org/doi/pdf/10.1162/JINH_a_01161 by guest on 02 October 2021 336 | IZDEBSKI, JAWORSKI, ÜSTÜNDAĞ,ANDSOŁTYSIAK nutritional data from the past to reconstruct changes in general well-being and economic development. Foodways create connections between aspects of human existence that at ﬁrst glance seem completely unrelated; they also reveal the need for historians to reach beyond textual sources when biological phenomena enter their investigations. Not surprisingly, an ongoing debate among historians—featured in recent issues of this journal—concerns the consilience between historical and scientiﬁc approaches and the potential that such synergy offers for re-formulating the central questions of modern historiography. Until recently, historical studies had to rely on textual sources produced by a small, elite minority within society, thus fostering a number of methodolog- ical challenges and unanswered questions that require other types of evidence for their resolution.1 This article presents one viable approach, namely, the use of state-of-the-art bioarchaeological methods to illuminate a relatively recent historical period (the Middle Ages), combining textual sources with research about teeth found in excavations of medieval sites in Anatolia and northern Syria (modern Turkey). In a reconstruction of eating habits, we discuss the kinds of food that different social groups tended to consume, focusing particularly on access to bread, an ideal proxy for links between social class or status (understood broadly as occupation, place in the society, and wealth) and foodways. As a staple in the ancient and medieval Mediterranean, bread found its ways into textual sources as both a symbol and a topic. Medieval assumptions about bread

1 Fernand Braudel, The Mediterranean and the Mediterranean World in the Age of Philip II (New York, 1972), I, 570. For biology and economic history, see Richard H. Steckel, “Biological Measures of Economic History,” Annual Review of Economics, V (2013), 401–423; for environmental history, Neil M. Maher, “Body Counts: Tracking the Human Body Through Envi- ronmental History,” in Douglas Cazeau Sackman (ed.), A Companion to American Environmental History (Malden, Mass., 2010), 163–180; Richard C. Hoffmann, An Environmen- tal History of Medieval Europe (New York, 2014); for the consilience of science and history, Michael McCormick, “History’s Changing Climate: Climate Science, Genomics, and the Emerging Consilient Approach to Interdisciplinary History,” Journal of Interdisciplinary History, XLII (2011), 251–273; John Haldon et al., “The Climate and Environment of Byzantine Anatolia: Integrating Science, History and Archaeology,” Ibid. XLV (2014), 113–161; Izdebski et al., “Realising Consilience: How Better Communication between Archaeologists, Historians and Natural Scientists Can Transform the Study of Past Climate Change in the Mediterranean,” Quaternary Science Reviews (Special Issue: Mediterranean Holocene Climate, Environment and Hu- man Societies), CXXXVI (2016), 5–22—the subject of a review essay by Timothy Newﬁeld, and Inga Labuhn, “Realizing Consilience in Studies of Pre-Instrumental Climate and Pre-Laboratory Disease,” Journal of Interdisciplinary History, XLVII (2017), 211–240.

Downloaded from http://www.mitpressjournals.org/doi/pdf/10.1162/JINH_a_01161 by guest on 02 October 2021 BREAD AND CLASS | 337 shed light on general stereotypes about food and the social hierarchy, whether in Christian or Islamic society. As it turns out, the stereotypes favored in the written sources are only partially true: The conclusions that emerge from the research herein show that social-biological reality was much more complex than the written sources suggest.2 Research on enamel microwear patterns in human teeth has been used successfully to reconstruct the diet of extinct hominids and to trace the transition from foraging to farming. Past research also applied this method to more fine-grained problems, such as the development of sophisticated grinding tools in late antiquity, as well as the timing of the weaning process. This approach derives from the observation that particular facets on molar crowns have specific functions during the mastication process. Some of them, mainly external ones, are in contact with their counterparts during the first (vertical) phase of mastication, while others (mainly facets close to the grooves) are affected during the second (horizontal) phase, the grinding of food particles between upper and lower teeth. Therefore, hard particles—mineral grit and at least some plant phytoliths—can leave traces on these facets; both the size and pattern of these traces is determined by the size and number of these particles. Although initially the analysis of enamel microwear patterns met with enthusiasm, the limitations of the method soon became evident. Many kinds of food (meat or dairy products) do not have much of an effect on enamel; external mineral grit is primarily responsible for observed damage on the enamel surface. None- theless, in specific cases, the damage is enough to permit insight into food-preparation practices in past human populations, especially those that relied on ground grains—flour in particular—for the vast portion of their calorie intake. Enamel microwear is particularly helpful in the study of Mediterranean Graeco-Roman and medieval foodways. Since all our tooth samples come from detailed excavations, we are able to determine the social status of individuals involved and thus to make connections between their

2 For the role of bread in medieval societies in general, see Hoffmann, Environmental History, 115–116; Remi Esclassan et al., “A Panorama of Tooth Wear During the Medieval Period,” Anthropologischer Anzeiger, LXXII (2015), 185–199; Clark Spencer Larsen, Bioarchaeology: Inter- preting Behavior from the Human Skeleton (New York, 2015).

Downloaded from http://www.mitpressjournals.org/doi/pdf/10.1162/JINH_a_01161 by guest on 02 October 2021 338 | IZDEBSKI, JAWORSKI, ÜSTÜNDAĞ,ANDSOŁTYSIAK socio-economic status and the kinds of bread that they consumed. In more technical terms, we assumed that the higher was the frequency of small nonlinear features in individuals’ tooth enamel, the better was the quality of their food during the two to three weeks before death. In this context, quality of food is defined by the amount of large-grit particles (the fewer, the better) in flour.3 Medieval Anatolia, together with other parts of the Middle East, offers an ideal case study for historical-scientificresearch. The wealth of textual sources enables reconstruction of the stereotypes about food and social class that the region’s medieval (literate) inhabitants shared. Moreover, Turkey boasts one of the longest histories of modern excavation in the world, making it possible to select sites that aptly represent diverse cultures and types of settlements in almost any period. Likewise, medieval Anatolia allows for a comparison between Christian and Islamic attitudes and cultural dispositions about food and social status; the predom- inantly Greek Orthodox region experienced a major invasion by, and migration of, Muslim Seljuk Turks during the eleventh century. Prior to that period, Anatolia was the heartland of the Eastern Roman (Byzantine) Empire, though its eastern parts still felt the effects of the Arab conquest. In the tenth century, the Byzantines re-conquered the whole of Anatolia as well as the northern parts of Syria. Their dominance in the region, however, did not last long, collapsing in the wake of the major Seljuk victory in the battle of Manzikert (1071). The Seljuk state, with its capital located

3 Peter S. Ungar, Jessica R. Scott, and Christine M. Steininger, “Dental Microwear Differ- ences between Eastern and Southern African Fossil Bovids and Hominins,” South African Journal of Science, CXII (2016), 134–138; Laura Mónica Martínez et al., “Testing Dietary Hypotheses of East African Hominines Using Buccal Dental Microwear Data,” PLoS ONE, 11(11) (2016), e0165447; Patrick Mahoney, “Dental Microwear from Natuﬁan Hunter-Gatherers and Early Neolithic Farmers: Comparisons within and between Samples,” American Journal of Physical Anthropology, CXXX (2006), 308–319; Sołtysiak, “Cereal Grinding Technology in Ancient Mesopotamia: Evidence from Dental Microwear,” Journal of Archaeological Science, XXXVIII (2011), 2805–2810; Rachel M. Scott and Siân E. Halcrow, “Investigating Weaning Using Dental Microwear Analysis: A Review,” ibid.: Reports, XI (2017), 1–11; Jing Xia et al., “New Model to Explain Tooth Wear with Implications for Microwear Formation and Diet Reconstruction,” Proceedings of the National Academy of Sciences, CXII (2015), 10669–10672; Kristin L. Krueger et al., “Dental Microwear Textures of ‘Phase I’ and ‘Phase II’ Facets,” American Journal of Physical Anthropology, CXXXVII (2008), 485–490; Jonathan M. Hoffman, Danielle Fraser, and Mark T. Clementz, “Controlled Feeding Trials with Ungulates: A New Application of in Vivo Dental Molding to Assess the Abrasive Factors of Microwear,” Journal of Experimental Biology,CCXVIII (2015), 1538–1547.

Downloaded from http://www.mitpressjournals.org/doi/pdf/10.1162/JINH_a_01161 by guest on 02 October 2021 BREAD AND CLASS | 339 in the inland city of Konya, introduced a new Islamic culture to the region. Despite the conquest of Constantinople by Western Europeans in 1204, the ultimate disintegration of the Byzantine Empire, and the major defeat of the Seljuk sultanate by the Mongols in 1243, the cultural divisions of twelfth-century Anatolia survived until the early fourteenth century.4 Our investigation of medieval Anatolian and Middle Eastern foodways unfolds in three stages; (1) a discussion of the stereo- typical Byzantine and Islamic links between food and social hierarchy in textual sources; (2) an enamel-microwear analysis of four populations from different parts of modern Turkey, including materials, methods, and results; (3) an explanation of the differing “views” of the two types of evidence about the correlation between social status and bread quality.

BYZANTINE AND ISLAMIC SOURCES FOR BREAD AND SOCIAL STATUS Most environmental historians today share a strong conviction that bread quality was closely connected to personal wealth. As Braudel observed, “There was a bread for the rich and a bread for the poor; only the former was made from wheat. In Lisbon, grain from the north, when intended for rich men’stables,was ﬁrst carefully sieved to remove stones and other impurities. The women of Lisbon could be seen on their doorsteps engaged upon this task.” Nor have historians’ understanding of this phenomenon changed since the publication, in the late 1940s, of Braudel’s mag- isterial study of the premodern Mediterranean. In a recent survey of medieval environmental history, Hoffmann wrote, “[F]rom the early to the central Middle Ages, the choice of bread and meat was more a matter of wealth and status than simply where a person lived, what they could afford, and what was needed for nutritional purposes. Elites wanted meat and elites wanted white bread.”5 The testimonies of medieval sources, whether Byzantine or Islamic, agree on this matter. Several Byzantine texts suggest that

4 Jean Claude Cheynet, “Byzance entre le Turcs et les Croisés,” in idem (ed.), Le monde byzantin (Paris, 2004), II, 43–65; Jonathan Shepard (ed.), The Cambridge History of the Byzantine Empire c. 500–1492 (New York, 2008). For the signiﬁcance of Anatolia in the context of medieval environmental history, see Haldon et al., “Climate and Environment.” 5Braudel,Mediterranean and the Mediterranean World,I,571;Hoffmann,Environmental History, 115.

Downloaded from http://www.mitpressjournals.org/doi/pdf/10.1162/JINH_a_01161 by guest on 02 October 2021 340 | IZDEBSKI, JAWORSKI, ÜSTÜNDAĞ,ANDSOŁTYSIAK people from different social backgrounds consumed distinct types of bread. Elite authors show a strong dislike for the “dirty bread,” often made of barley rather than wheat, that the poor tended to eat.Thesamedelightin“pure bread” is conspicuous in every genre of Byzantine text—be it a medical treatise, a piece of poetry, a saint’s life, or a work of history. In fact, two late Byzantine historians, George Akropolites (thirteenth century) and Nicephorus Gregoras (fourteenth century), used bread as a topos that allowed them to draw a clear distinction between the upper and the lower classes of Byzantine society. Both wrote complex, highbrow his- toriographies in the classical style that the educated elite of byzantine society always admired.6 Akropolites employed bread and food as a shorthand for the divide between the cultivated elite and the uneducated poor. For example, he ridiculed Constantine Margarites, a military commander, as “a peasant born of peasants, reared on barley and bran and know- ing only how to grunt.” His use of the Greek word pituron (“husks” or “bran”) implied that Margarites’ bread was composed of grain and ﬂour that no aristocrat would ever consume. The author’s intended elite audience would have immediately understood this metaphor and its offensive intention.7 Gregoras’ example is not personally derogatory, but it uses the same stereotype. In a long narrative describing war heroes looking for food and shelter in a village, he wrote, “And if they say that for a hungry person every (kind of ) bread is pleasant, so we too found it pleasant and humane to wallow in ashes.” The ashes in this passage are the impurities contained in peasant bread, commonly baked in homes and often containing ashes added un- intentionally during the baking process. Thus, this excerpt, again referring to a stereotype, intimates that a member of the upper class would eat only pure, unpolluted bread. Such an allusion to commonplaces about bread and class in the context of military adventures is surprising, however, given that Byzantine soldiers,

6 Phaidōn Koukoules, “Onomata kai eide arton kata tous Buzantinous chronous,” Epeteris Hetaireias Byzantinon Spoudon, V (1928), 36–52; John L. Teall, “The Grain Supply of the Byzantine Empire, 330–1025,” Dumbarton Oaks Papers, XIII (1959), 87–139; Ruth Macrides (ed.), George Akropolites: The History: Introduction, Translation and Commentary (Oxford, 2007); Gregoras (trans. Jan Louis van Dieten and F. H. Tinnefeld), Rhomäische Geschichte (Stuttgart, 1973), V. 7 Acropolita (ed. P. Wirth and A. Heisenberg), Opera (Leipzig, 1978), I, 123, 7–9. For the English translation, see Macrides, George Akropolites, 297.

Downloaded from http://www.mitpressjournals.org/doi/pdf/10.1162/JINH_a_01161 by guest on 02 October 2021 BREAD AND CLASS | 341 including officers on campaigns, ground their own grain and baked their own bread, which typically suffered from the same defects as peasant bread. Gregoras’ disregard for the actual conditions of soldiers’ life shows the strength of the connection between bread and social status in the literature composed by and for the Byzantine elite.8 The Christian and Islamic worlds in medieval Anatolia shared the same truisms about bread and social status. Several Seljuk and later Islamic Anatolian sources emphasize that pure bread was a valued commodity, serving as a metaphor for products of fine quality, much as it had in the Byzantine literature. Nor were these the only Islamic societies of the medieval Middle East to hold such notions. For instance, manuals intended for Cairo’smarketin- spectors advised officials to be on the alert for ground peas, broad beans, and chickpeas that added weight to bread, and they identified the impurities—insects, straw, hair, or bodily fluids—that often tainted bread during the dough-kneading process. Even though Egypt, unlike almost any other region of the Mediterranean, had easy access to wheat flour, the kind and caliber of the bread consumed there depended on the presence or absence of non- wheat or even non-cereal admixtures and impurities. In this respect, Mamluk Egypt resembled Abbasid Iraq. The most comprehensive Abbasid cooking manual, written by Ibn Sayyāral-Warrāq, shows a strong concern for the quality of bread and flour. Interestingly, although al-Warrāq recommended bread baked at home (rather than in the market), he preferred flour purchased from millers to flour that was ground in rotary querns at home, which could add impurities to the final product.9 Medieval textual sources from the Middle East—Byzantine, Abbasid, Seljuk, and Mamluk—leave no doubt that the best bread (made of finely ground wheat flour with no impurities)

8 Gregoras (ed. Ludwig Schopen), Nicephori Gregorae Byzantina historia (Bonn, 1829), I, 379, 6–8 (VIII 14.5); Haldon, Warfare, State and Society in the Byzantine World, 565–1204 (London, 1999), 168–169. “Pure bread” was already a topos in the works of Greek and Roman writers. See Koukoules, “Onomata kai eide arton kata tous Buzantinous chronous.” 9 Nicolas Trépanier, Foodways and Daily Life in Medieval Anatolia: A New Social History (Austin, 2014), 86; Amalia Levanoni, “Food and Cooking during the Mamluk Era: Social and Political Implications,” Mamluk Studies Review, IX (2005), 201–222; Paulina B. Lewicka, Food and Foodways of Medieval Cairenes: Aspects of Life in an Islamic Metropolis of the Eastern Mediterranean (Leiden, 2011), 155–158; David Waines, “Cereals, Bread and Society: An Essay on the Staff of Life in Medieval Iraq,” Journal of the Economic and Social History of the Orient, XXX (1987), 255–285.

Downloaded from http://www.mitpressjournals.org/doi/pdf/10.1162/JINH_a_01161 by guest on 02 October 2021 342 | IZDEBSKI, JAWORSKI, ÜSTÜNDAĞ,ANDSOŁTYSIAK was reserved for the elite. The hierarchy of foods corresponded to the hierarchy of social groups.

ENAMEL-MICROWEAR ANALYSIS People and Places How does the vision of food’srolein medieval society in written sources produced by literate elites compare with the results of a bioarchaeological study of enamel microwearontheteethofindividuals from four medieval sites? Three of these sites were Byzantine towns. The ﬁrst, Kadıkalesi, was a fortress located on the Aegean coast that protected the har- bor of Anaia, and the other two—Kalon Oros (Alanya Kalesi, a tenth-century cemetery) and Smyrna (Agora, twelfth through thirteenth centuries)—were relatively important regional centers. The fourth site, Akarçay Höyük, was a Muslim rural community at the northernmost outskirts of Syria, between the Seljuk Sultanate and the Levantine polities, now belonging to Anatolia (see Fig- ure 1). Üstündağ analyzed the remains of the individuals from

Fig.1 Map of Byzantine-Seljuk Anatolia with the Study Sites

NOTE The dotted frontier line represents the maximum territorial extent of Byzantine power in twelfth-century Anatolia (under Manuel I) and provides an approximation of cultural divisions in Anatolia during most of our study period (twelfth through thirteenth centuries).

Downloaded from http://www.mitpressjournals.org/doi/pdf/10.1162/JINH_a_01161 by guest on 02 October 2021 BREAD AND CLASS | 343 all four sites using standard protocols for sex and age-at-death assessments.10 The Hellenistic-Byzantine citadel of Kalon Oros, today’s Alanya, provides the oldest tooth samples, dating back to the times of the Byzantine re-conquest of Anatolia. During recent archaeological excavations, a tenth-century cemetery with eight burials containing twenty-seven individuals was discovered in the northern nave of a collapsed late antique basilica. The deceased were buried on an west–east axis, facing east (as was the custom for Christians), with their arms on their abdomens or with one hand on their chest and the other on their abdomen. The pit graves were covered by bricks, and one of the graves had a pearl cross on it. It is impossible, however, to determine the social status of these individuals solely from their graves or the goods found in them. But judging from the relatively high prevalence of caries on their teeth, which suggests the carbohydrate diet typical of an urban population relying largely on plant-related food, they seem to represent the middle range of Alanya’s tenth-century urban population. During this period, Kalon Oros was an important supply base for the Byzantine maritime theme (military and administration unit) of Kibyrraiotes. The Book of Ceremonies,an important Byzantine text attributed to Emperor Constantine Porphyrogenites, attests to Kalon Oros’ role in the preparation of naval expeditions during this period. Hence, the cemetery appears to have been a burial site for an important military and urban base.11 The town of Smyrna now falls within the modern metropolis of İzmir. Archaeological work was possible only in the ancient agora, which after the Seljuk invasion served as a cemetery for Byzantine city dwellers (as it has continuously into Ottoman times).

10 Jane A. Buikstra and Douglas H. Ubelaker (eds.), Standards for Data Collection from Human Skeletal Remains (Fayetteville, 1994). 11 M. Oluş-Arık, “Alanya Kalesi 2004 Yılı Çalişmalari,” 27. Kazı Sonuçları Toplantısi, 2. Cilt (2006), 213–228; Haldon, “Theory and Practice in Tenth-Century Military Administration: Chapters II, 44 and 45 of the Book of Ceremonies, Text and Translation,” Travaux et Mémoires, XIII (2000), 201–352; Hansgerd Hellenkemper and Friedrich Hild, Lykien und Pamphylien (Wien, 2004), 50; Üstündağ and F. Arzu-Demirel, “Alanya Kalesi Kazılarında Bulunan İnsan İskelet Kalıntılarının Osteolojik Analizi,” Türk Arkeoloji ve Etnoğrafya Dergisi,VIII(2008),79–90; idem, “Alanya Kalesi İskelet Topluluğunda AğızveDiş Sağlığı,” Hacettepe Üniversitesi Edebiyat Fakültesi Dergisi, XXVI (2009), 219–234. For the correlation between low protein diet and dental caries, see Larsen, Bioarchaeology.

Downloaded from http://www.mitpressjournals.org/doi/pdf/10.1162/JINH_a_01161 by guest on 02 October 2021 344 | IZDEBSKI, JAWORSKI, ÜSTÜNDAĞ,ANDSOŁTYSIAK Remains of thirty-three Christians (heads facing east), dated from the twelfth to thirteenth century, were discovered during excavations in 2007. Again, burials alone offered no clues to their social status, but the percentage of individuals in this group who show pathological signs of physical stress is greater than that of any other sample group in this study. The high frequency (74 percent) of periosteal reactions on long bones may have resulted from infection. The prevalent porotic hyperostosis (73 percent), due to iron or B12 deficiency, suggests malnutrition. As in Alanya/Kalon Oros, dental caries were relatively common (15.9 percent). Altogether, this profile suggests that Byzantine Smyrna had an average urban population on socio- economic terms.12 According to ongoing excavations, Kadıkalesi, the last of our Byzantine sites—a Byzantine stronghold on the southern Aegean coast of Anatolia with good access to Mediterranean trade networks—was relatively rich and well-stocked with luxury items. Fifty-eight burials were unearthed there in two locations, close to the church and along the stronghold walls. Because the skeletons were dated to the twelfth or thirteenth century, they most likely belonged to the Byzantine town dwellers (buried in pit graves covered by bricks and heads oriented eastward). One burial from the church contained a silver garment clasp, indicating an aged female, clearly of high social status. Pathological indicators of physical stress were rare compared to those in other sample groups—porotic hyperostosis observed in only 17 percent and periosteal reactions in only 26 percent of individuals. Only dental caries were as common as elsewhere (15 percent), possibly reflect- ing the higher percentage of females in this group. In short, the

12 Akın Ersoy, “2007 Yılı Smyrna Antik Kenti Kazısı,” 30. Kazi Sonuçlari Toplantisi, 2. Cilt (2009), 33–46; Markus Kohl and idem, “Agora d’Izmir, travaux 2007,” Anatolia Antiqua, XVI (2008), 345–353; Ersoy, Gülten Çelik, and Seçil Yılmaz, “2010 Yılı Smyrna Antik Kenti Kazısı Raporu,” 33.KaziSonuçlariToplantisi,2.Cılt (2012), 179–204. For the history of Smyrna during this period, see Hélène Ahrweiler, “L’Histoire et la Géographie de la région de Smyrne entre les deux occupations turques (1081–1317),” Travaux et Mémoires, 1 (1965), 1–204; for the link between malnutrition and porotic hyperostosis, Philip L. Walker et al., “The Causes of Porotic Hyperostosis and Cribra Orbitalia: A Reappraisal of the Iron- Deﬁciency-Anemia Hypothesis,” American Journal of Physical Anthropology, CXXXIX (2009), 109–125; for the indicators of stress on bones, Alan H. Goodman et al., “Biocultural Perspectives on Stress in Prehistoric, Historical, and Contemporary Population Research,” ibid., XXI (1988), 169–202.

Downloaded from http://www.mitpressjournals.org/doi/pdf/10.1162/JINH_a_01161 by guest on 02 October 2021 BREAD AND CLASS | 345 living conditions of those buried in Kadıkalesi were better than those in the two other Byzantine sites.13 Akarçay Höyük, our only Islamic site, is located on the eastern bank of the contemporary Carchemish Dam Reservoir. People lived there from the Bronze Age to the Hellenistic-Roman period; it became a cemetery from the thirteenth to the fourteenth century after several centuries of abandonment. During this period, it came under the political control of European Crusaders, Seljuks, Mongols, and, most importantly, Egyptian Ayyubids and Mamluks. Age-at- death and sex distribution suggests that Akarçay Höyük was a regular attritional cemetery for the local Muslim rural population (either agriculturalists or Turkic herders). The burials follow the common Islamic custom: simple pit inhumations marked by oval rings consisting of a single row of stones; graves in an east–west orientation, with heads toward the east and faces/bodies turned southward; and both hands placed on the abdomen. According to the pathological indicators, this rural Muslim population was significantly healthier than the Byzantine urban groups, with low percentages of dental caries (8 percent) and porotic hyperostosis (6 percent).14 Forty individuals with lower second molars were selected as suitable for the present analysis—ten from Kadıkalesi, ten from Smyrna Agora, eleven from Akarçay Höyük, and nine from Alanya Kalesi (see Table 1 for details). As shown in Figure 2, Akarçay and Kadıkalesi had a similar health profile, different from those identified at both Alanya and Smyrna. The low frequency of porotic hyperostosis and periosteal reactions in Akarçay and Kadıkalesi suggests that these communities enjoyed better living conditions than the other two locales. The low dental-caries rate in Akarçay is probably

13 Zeynep Mercangöz, “Ostentatious Life in a Byzantine Province: Some Selected Pieces from the Finds of the Excavation in Kuşadası,Kadıkalesi/Anaia (Prov. Aydin, TR),” in Falco Daim and Jörg Drauschke (eds.), Byzanz - das Römerreich im Mittelalter, 2 (Mainz, 2010), 181– 198; IşılTalu,“Classiﬁcation and Visual Analysis of Weathering Forms of Stone in Kadıkalesi, Kuşadası,” unpub. M.A. thesis (İzmir Institute of Technology, 2005); Üstündağ, “Kuşadası Kadıkalesi/Anaia kazısında bulunan insan iskelet kalıntılar,” 24. Arkeometri Sonuçları Toplantısı (2009), 209–228. 14 Abdullah Deveci, “2001 Excavations at Akaraçay Höyük,” in Numan Tuna, Jean Greenhalgh, and Jale Velibeyoğlu (eds.), Salvage Project of the Archaeological Heritage of the Ilısu and Carchemish Dam Reservoirs, Activities in 2001 (Ankara, 2004), 279–291; idem and H. Kübra Ensert, “The 2002 Excavations at Akarçay Höyük,” in Tuna and Owen Doonan (eds.), Salvage Project of the Archaeological Heritage of the Ilısu and Carchemish Dam Reservoirs Activities in 2002 (Ankara, 2011), 204–214.

Downloaded from http://www.mitpressjournals.org/doi/pdf/10.1162/JINH_a_01161 by guest on 02 October 2021 346 | IZDEBSKI, JAWORSKI, ÜSTÜNDAĞ,ANDSOŁTYSIAK Table 1 General Description of the Analyzed Sample

ID SITE TAG SEX AGE-AT-DEATH A1 Alanya Kalesi 2004 M1 Ind2 M 50+ A2 Alanya Kalesi 2004 M4 Ind1 M 40–45 A3 Alanya Kalesi 2006 M8 Ind3 M 30–35 A4 Alanya Kalesi Alanya No ID. M 20–25 A5 Alanya Kalesi 2005 M8 Ind1 F 20 A6 Alanya Kalesi 2006 M8 Ind5 M 30–35 A7 Alanya Kalesi 2005 M6 Ind2 M 25 K1 Kadıkalesi 2005 D18 M3 F 18–19 K2 Kadıkalesi 2005 D18 M8 F 20 K3 Kadıkalesi 2006 U32 M10 F 25–30 K4 Kadıkalesi 2007 S23 M21 F 35–40 K5 Kadıkalesi 2006 U32 M11 F 55 K6 Kadıkalesi 2005 D18 M6 M 35–40 K7 Kadıkalesi 2007 P23 M8 F 20 S1 Smyrna Agora 2007 M5 Ind1 M 30–35 S2 Smyrna Agora 2007 Av1.14 Ind1 F 20 S3 Smyrna Agora 2007 M6 Ind2 M 20 S4 Smyrna Agora 2007 M6 Ind4 F 25–30 S5 Smyrna Agora 2007 M6 Ind7 M 35–40 S6 Smyrna Agora 2007 Av1.08 F 20–25 S7 Smyrna Agora 2007 M6 Ind8 M 25–30 S8 Smyrna Agora 2007 Av1.14 Ind4 F 35–40 S9 Smyrna Agora 2007 M6 Ind10 M 40–50 H1 Akarçay Höyük 2005 F 25y F 25 H2 Akarçay Höyük 2001 I-XIVb KT28 M1 F 20–25 H3 Akarçay Höyük 2006 M1 F 25–35 H4 Akarçay Höyük 2001 JXIV A1 M4 M 20–25 H5 Akarçay Höyük KT 2002 KT3 F 25–30 H6 Akarçay Höyük 2001 KT11 M3 Ind3 F 35–40 H7 Akarçay Höyük 2006 M2 M 50+

evidence of a superior access to animal protein, and the graves at Kadıkalesi clearly contain individuals who were wealthier than the norm. Methods Since the average wear of ﬁrst molars is usually higher than that of second molars, we selected lower second molars with a wear degree of the protoconid cusp between 2 and 5 in Scott’s scale to maximize the sample size. If an individual’s right and left molars did not differ signiﬁcantly in degree of dental wear, we selected the right one. We cleaned the teeth with water and soft brushes and after drying them, made polyurethane resin

Downloaded from http://www.mitpressjournals.org/doi/pdf/10.1162/JINH_a_01161 by guest on 02 October 2021 BREAD AND CLASS | 347 Fig.2 Distribution of Paleopathological Indicators of Diet and Physiological Stress (in Percentage)

SOURCES Üstündağ andF.A.Demirel,“Alanya Kalesi Kazılarında Bulunan İnsan İskelet Kalıntılarının Osteolojik Analizi,” Türk Arkeoloji ve Etnoğrafya Dergisi, VIII (2008), 79–90; idem, “Alanya Kalesi İskelet Topluluğunda AğızveDiş Sağlığı,” Hacettepe Üniversitesi Edebiyat Fakültesi Dergisi, XXVI (2009), 219–234; Üstündağ, “Kuşadası Kadıkalesi/Anaia kazısında bulunan insan iskelet kalıntılar,” 24. Arkeometri Sonuçları Toplantısı (2009), 209–228.

(RenCast FC52) casts using silicone negatives (Gumosil B poli- condensing rubber).15 SEM pictures were taken with an LEO 1430VP microscope at the Biology Dept., University of Warsaw. We selected as our surface the protoconid facet x, a phase II facet that reliably shows microwear patterns differentiated by diet and entails a relatively low morphological variation of the protoconid, compared to hypoconid and hypoconulid. Depending on the observed size of facet x, usually one or two, but occasionally three or four, micrographs were taken under magniﬁcation of ×300 (see Figure 3).16 At the ﬁrst stage, all micrographs were reviewed to select those suitable for an analysis of microwear patterns. If three or four pictures were available, we selected the two with the lowest evidence of remaining external dirt. The second picture was used for estimation of measurement error. The many micrographs in which postmortem erosion was evident were rejected.

15 Eugenie C. Scott, “Dental Wear Scoring Technique,” American Journal of Physical Anthro- pology, LI (1979), 213–218. 16 See Krueger et al., “Dental Microwear Textures of ‘Phase I’ and ‘Phase II’ Facets.”

Downloaded from http://www.mitpressjournals.org/doi/pdf/10.1162/JINH_a_01161 by guest on 02 October 2021 348 | IZDEBSKI, JAWORSKI, ÜSTÜNDAĞ,ANDSOŁTYSIAK Fig.3 Examples of Enamel Microwear in (a) Alanya Kalesi, (b) Kadıkalesi, (c) Smyrna Agora, and (d) Akarçay Höyük.

NOTE Scale bars 100 μm.

The features were counted and measured using Microwear 4.02 software and then classiﬁed manually into four groups, depending on their shape and size—small linear features (LS,upto 5μm in breadth), large linear features (LL,morethan5μmin breadth), small nonlinear features (NS,upto20μm in diameter), and large nonlinear features (NL, more than 20μmindiameter). The standard deviation for average relative orientation and average length of all linear features was calculated. Pictures were processed in random order, without prior reference to their origin and tags to avoid systematic error. Differences in frequencies between two micrographs of the same facet were tested using the χ2 test to estimate the measurement error. For testing differences in distribution of speciﬁed variables between three subsets, we used Kruskal- Wallis ANOVA and chose Correpondence Analysis (CA) to see the overall pattern of feature frequencies. All statistics were calculated using Statistica 10 software.17

17 Peter S. Ungar, Microware Software, Version 4.02: A Semi-Automated Image Analysis System for the Quantiﬁcation of Dental Microwear (Fayetteville, 2002).

Downloaded from http://www.mitpressjournals.org/doi/pdf/10.1162/JINH_a_01161 by guest on 02 October 2021 BREAD AND CLASS | 349 Results From the initial sample of forty teeth, ten were not suitable for further analysis due to postmortem erosion. For thirteen of the teeth, we could take only one micrograph since the available area of the facet x was too small. The data are presented in Table 2. In the case of teeth with two micrographs, both microwear patterns were usually similar; only once was the difference significant at the 0.05 level, and twice 0.1>p>0.05. For further statistical analysis, micrographs of all teeth were divided into two subsets (A and B). If two micrographs of one facet were available, one was assigned to the subset A and the other to the subset B. If only one micrograph was taken, it was assigned to both subsets. The analyses were then performed for both subsets, thereby providing insight into repeatability of the results. Among four defined features, the differences in LS were significant (Kruskall-Wallis test, p<0.05) in both subsets: Teeth from Kadıkalesi and Akarçay Höyük clearly exhibited a higher number of small linear features than teeth from two other samples (see Table 3). Additionally, large nonlinear features were more common in teeth from Alanya Kalesi, and this difference is significant in the subset B. On the other hand, there are no clear differences in the frequency of large linear and small non-linear features between teeth from four sites. Orientation and average length of linear features exhibit considerable differences between subsets. The overall pattern of differences was checked using CA for all micrographs (total χ2=502.9, p<0.001). The results are presented as two bi-plots, separately for the subsets A and B (Figure 4a and b), after exclusion of two outliers with an abnormally high frequency of nonlinear features (A7 and S6). The first dimension (c. 45 percent of inertia) distinguishes between LS and all other categories of features; the second dimension (c. 31 percent of inertia) discrimi- nates between all categories, especially between NL and LL. The relatively varied mineral content that individuals from these four populations consumed allows for a few interesting obser- vations. First, people in the Byzantine regional centers of Alanya/ Kalon Oros and Smyrna generally ate bread with coarser grit, as indicated by their greater frequency of significant microwear features. As explained above, both samples represent typical urban populations, who were generally neither high on the social scale nor possessed of particularly good health. It is hardly surprising that their bread (which must have formed a crucial element of their diet,

Downloaded from http://www.mitpressjournals.org/doi/pdf/10.1162/JINH_a_01161 by guest on 02 October 2021 Downloaded from http://www.mitpressjournals.org/doi/pdf/10.1162/JINH_a_01161 by guest on02 October 2021

Table 2 Frequency of Enamel Microwear Features in Studied Sites

2 ID LS LS% LL LL% NS NS% NL NL% ALL ORIENT. LENGTH χ P A1a 116 53.2 50 22.9 38 17.4 14 6.4 218 20.13 130.21 0.553 0.907 A1b 112 52.1 48 22.3 43 20.0 12 5.6 215 25.55 150.41 A2a 57 39.0 51 34.9 28 19.2 10 6.8 146 55.58 117.95 7.845 0.097 A2b 72 51.1 44 31.2 13 9.2 12 8.5 141 50.36 104.67 A3a 43 47.8 18 20.0 22 24.4 7 7.8 90 22.01 106.25 1.165 0.761 A3b 38 41.3 23 25.0 25 27.2 6 6.5 92 33.69 118.65 A4 27 32.1 37 44.0 10 11.9 10 11.9 84 30.24 120.18 A5 18 33.3 19 35.2 8 14.8 9 16.7 54 15.65 133.93 A6 21 35.0 34 56.7 5 8.3 0 0.0 60 34.38 128.19 A7 22 40.0 2 3.7 22 40.0 9 16.4 55 31.35 84.93 K1a 63 59.4 27 25.5 7 6.6 9 8.5 106 36.63 101.62 7.614 0.055 K1b 45 59.2 26 34.2 5 6.6 0 0.0 76 23.22 111.99 K2a 114 63.0 45 24.9 20 11.0 2 1.1 181 27.48 92.09 5.286 0.152 K2b 95 70.9 32 23.9 7 5.2 0 0.0 134 31.13 120.06 K3a 76 58.5 33 25.4 16 12.3 5 3.8 130 22.05 135.1 3.851 0.278 K3b 59 51.8 25 21.9 22 19.3 8 7.0 114 21.53 136.25 K4a 45 40.9 33 30.0 30 27.3 2 1.8 110 27.97 176.16 3.400 0.334 K4b 35 42.7 30 36.6 14 17.1 3 3.7 82 19.37 169.98 K5 61 71.8 18 21.2 3 3.5 3 3.5 85 32.08 128.32 K6 88 47.1 67 35.8 19 10.2 4 2.1 187 17.96 126.40 K7 64 39.5 53 32.7 32 19.8 13 8.0 162 37.23 97.63 S1a 25 48.1 17 32.7 5 9.6 5 9.6 52 22.15 166.69 3.636 0.304 S1b 35 44.9 25 32.1 15 19.2 3 3.8 78 48.86 149.87 Downloaded from http://www.mitpressjournals.org/doi/pdf/10.1162/JINH_a_01161 by guest on02 October 2021

S2a 45 43.3 26 25.0 31 29.8 2 1.9 104 21.12 112.44 2.412 0.491 S2b 40 41.2 33 34.0 22 22.7 2 2.1 97 23.39 111.42 S3a 32 35.2 34 37.4 24 26.4 1 1.1 91 22.92 208.86 3.323 0.344 S3b 36 32.7 52 47.3 22 20.0 0 0.0 110 41.27 146.91 S4a 47 43.5 36 33.3 20 18.5 5 4.6 108 32.59 193.99 4.632 0.201 S4b 54 55.7 26 26.8 16 16.5 1 1.0 97 44.23 144.76 S5a 57 46.3 47 38.2 18 14.6 1 0.8 123 35.23 136.02 10.638 0.031 S5b 42 29.4 67 46.9 27 18.9 7 4.9 143 38.92 122.84 S6a 41 34.7 20 16.9 50 42.4 7 5.9 118 37.38 110.43 2.212 0.697 S6b 55 39.6 15 10.8 61 43.9 8 5.7 139 28.98 109.16 S7 21 24.7 47 55.3 17 20.0 0 0.0 85 40.86 176.13 S8 46 34.6 55 41.3 25 18.8 7 5.3 133 53.71 128.04 S9 56 43.1 65 50.0 9 6.9 0 0.0 130 34.52 159.60 H1a 75 55.6 30 22.2 22 16.3 8 5.9 135 28.49 120.60 2.469 0.48 H1b 48 48.5 24 24.2 23 23.2 4 4.0 99 42.06 118.64 H2a 63 65.6 18 18.8 15 15.6 0 0.0 96 43.12 88.88 2.227 0.328 H2b 47 57.3 15 18.3 20 24.4 0 0.0 82 38.33 104.98 H3a 56 64.4 21 24.1 10 11.5 0 0.0 87 34.64 85.16 2.657 0.265 H3b 54 58.1 20 21.5 19 20.4 0 0.0 93 32.53 116.87 H4a 44 55.7 23 29.1 10 12.7 2 2.5 79 28.08 71.10 2.252 0.522 H4b 38 53.5 17 23.9 15 21.1 1 1.4 71 39.21 141.13 H5 58 55.2 31 29.5 10 9.5 6 5.7 105 27.26 113.31 H6 36 47.4 18 23.7 18 23.7 4 5.3 76 40.87 90.65 H7 31 63.3 16 32.7 2 4.1 0 0.0 49 35.91 131.79

NOTES LS=small linear features; LL=large linear features (>5 μm in breadth); NS=small nonlinear features; NL=large nonlinear features (>20 μm in diameter); orient=standard deviation of linear-feature orientation; length=mean length of linear features. 352 | IZDEBSKI, JAWORSKI, ÜSTÜNDAĞ,ANDSOŁTYSIAK Table 3a Kruskal-Wallis Test Results for Observed Microwear Features (Subset A, N=30)

ALANYA KALESI KADIKALESI SMYRNA AKARÇAY (N=7) (N=7) AGORA (N=9) HÖYÜK (N=7)

FEATURE MEAN S.D. MEAN S.D. MEAN S.D. MEAN S.D. HP

LS% 40.1 7.8 54.3 12.1 39.3 7.5 58.2 6.6 14.90 0.002 LL% 31.1 17.3 27.9 5.2 36.7 11.7 25.7 4.9 5.39 0.145 NS% 19.4 10.4 13.0 8.1 20.8 10.9 13.3 6.1 3.93 0.269 NL% 9.4 6.0 4.1 3.0 3.3 3.3 2.8 2.8 7.26 0.064 All 101.0 60.7 137.3 39.8 104.9 25.8 89.6 26.7 5.81 0.121 Orientation 29.9 13.2 28.8 7.2 33.4 10.5 34.1 6.4 2.76 0.430 Length 117.4 17.0 122.5 29.1 154.7 35.2 100.2 21.9 9.44 0.024

Table 3b Kruskal-Wallis Test Results for Observed Microwear Features (Subset B, N=30)

ALANYA KALESI KADIKALESI SMYRNA AKARÇAY (N=7) (N=7) AGORA (N=9) HÖYÜK (N=7)

FEATURE MEAN S.D. MEAN S.D. MEAN S.D. MEAN S.D. HP

LS% 40.7 8.1 54.7 13.0 38.4 9.3 54.8 5.6 12.83 0.005 LL% 31.2 16.9 29.5 6.8 38.3 13.9 24.8 4.8 5.95 0.114 NS% 18.8 11.4 11.7 6.9 20.8 9.7 18.1 8.0 5.00 0.172 NL% 9.4 6.0 3.5 3.1 2.5 2.4 2.3 2.6 7.98 0.046 All 100.1 59.1 120.0 43.0 112.4 24.5 82.1 19.1 6.00 0.165 Orientation 31.6 10.5 26.1 7.4 39.4 9.4 36.6 5.2 9.64 0.022 Length 120.1 21.0 127.2 22.6 138.7 22.5 116.8 16.6 3.87 0.276

as suggested by the high frequency of dental caries) was not of high quality. Second, most of the people from the Byzantine town of Kadıkalesi and all the Muslim villagers buried at Akarçay Höyük had access to superior bread. Accordingly, the Kadıkalesi sample consists of numerous healthy, possibly even elite, individuals, and the Akarçay sample consists of a relatively healthy rural population with better access to animal protein. Third, the distinction between people who ate the purest bread and those who did not in Kadıkalesi during this period

Downloaded from http://www.mitpressjournals.org/doi/pdf/10.1162/JINH_a_01161 by guest on 02 October 2021 BREAD AND CLASS | 353 corresponds with their burial location either near the church (higher status) or outside the citadel walls (lower status). Observed contrasts in the dental-microwear patterns between the four compared populations are most likely the consequence of differences in their grinding tools, since the other parameters— availability of various cereal species and the method of baking ﬂat bread in ovens—were roughly the same throughout Anatolia and northern Syria during that period.18

MEDIEVAL TEXTS AND FOODWAY REALITIES Reliance on the written sources alone would lead to an obvious conclusion: The rich and powerful ate better than the poor and less privileged. Bioarchaeo- logical research, however, reveals that medieval food realities were much more complex. On one hand, the Byzantine town of Kadıkalesi seems to confirm some aspects of the stereotype: The richer and more influential individuals, buried close to the church, indeed ate higher-quality bread than the poorer individuals buried outside the town walls. On the other hand, the main difference between our study populations follows an urban–rural divide that runs contrary to the notion that peasants ate the kind of “dirty” bread that left strong damage on teeth enamel. The group that, on average, ate the best bread was the villagers from northern Syria. In fact, the peasants buried at Akarçay Höyük ate bread similar to that of the urban elite from Kadıkalesi. The lower strata of the urban communities in Smyrna, Kalon Oros, and Kadıkalesi emerge as the most underprivileged in terms of diet. The stereotype may have reflected certain social-biological conditions— those of the two groups that were in immediate contact, namely, city-based elites and average city dwellers—but not those of the society in general.

18 Waines, “Cereals, Bread and Society”; Naomi F. Miller, “The Crusader Period Fortress: Some Archaeobotanical Samples from Medieval Gritille,” Anatolica, XVIII (1992), 87–99; John Moore, Tille Höyük 1: The Medieval Period (London, 1993), 19–54; Miller, “Patterns of Agriculture and Land Use at Medieval Gritille,” in Scott Redford (ed.), Archaeology of the Frontier in the Medieval Near East: Excavations at Gritille, Turkey (Philadelphia, 1998), 211–252; Noor Mulder-Heymans, “Archaeology, Experimental Archaeology and Ethnoarchaeology on Bread Ovens in Syria,” Civilisations: Revue internationale d’anthropologie et de sciences humaines, XLIX (2002), 197–221; John M. Marston, “Agricultural Strategies and Political Economy in Ancient Anatolia,” American Journal of Archaeology, CXVI (2012), 377–403.

Downloaded from http://www.mitpressjournals.org/doi/pdf/10.1162/JINH_a_01161 by guest on 02 October 2021 354 | IZDEBSKI, JAWORSKI, ÜSTÜNDAĞ,ANDSOŁTYSIAK Though reminiscent of the modern understanding of medieval cities as ecological black holes—particularly regarding the less privileged—the patterns displayed in this article derive from surprising differences between social groups and their access to technology. As far as the enamel microwear pattern is con- cerned, the quality of bread depends on the available ﬂour, the contingencies of baking (labor costs, ovens sealed from ashes, etc.), and the milling process. Watermills were better than hand mills at producing ﬂour—due to the weight of their stones, which affected the size of the microwear content—and their lower costs (minimal human labor). Watermills could grind

Fig.4a Correspondence Analysis Bi-plot for Data from Table 2 (subset A)

NOTES Dimension 1 distinguishes between Small Linear (LS) features in enamel-microwear pattern and all other features. LS features are associated with bread (ﬂour) of higher quality. Dimension 2, less important, distinguishes between linear and nonlinear features—nonlinear being the most frequent in some teeth from Alanya Kalesi. Note bi-modal distribution at Kadıkalesi, in which some teeth are dominated by small linear features, but others have many more large lines and nonlinear features.

Downloaded from http://www.mitpressjournals.org/doi/pdf/10.1162/JINH_a_01161 by guest on 02 October 2021 BREAD AND CLASS | 355 Fig.4b Correspondence Analysis Bi-plot for Data from Table 2 (subset B)

NOTE With some minor differences, this bi-plot shows the same pattern as the bi-plot for subset A, indicating that observed differences between sites are not accidental. Despite the differences, both bi-plots show roughly the same pattern. Teeth from Kadıkalesi and Akarçay Höyük cluster around LS, though the distribution for Akarçay Höyük is clearly narrower. The lesser homogeneity of the teeth microwear pattern in Kadıkalesi is due to the burial sites that provided the samples: The two teeth with enamel-microwear patterns most visibly displaying a diet with bread of poorer quality belonged to people of lower social rank buried outside the stronghold walls (2007 S23 M21 and 2007 P23 M8 in Table 1); those who ranked higher in the town community’s hierarchy were buried close to the church. In contrast, teeth from Alanya Kalesi and Smyrna Agora cluster around LL AND NS, suggesting a diet based on low quality bread. Higher frequency of large nonlinear features (NL) at Alanya Kalesi indicates an even poorer quality of bread. However, only the difference between the two pairs of sites (Kadıkalesi and Akarçay Höyük vs. Alanya Kalesi and Smyrna) in the ﬁrst CA dimension is clear, based on the signiﬁcant difference in LS as tested using Kruskall-Wallis ANOVA. This pattern holds for both subsets.

the same grain several times, thereby ensuring the quality of the ﬁnal product.19 Hand mills (more speciﬁcally, rotary hand querns) were the dominant milling technology in Byzantine cities, especially for

19 Hoffmann, Environmental History, 227–237; Britta Padberg, DieOaseausStein:humanökologische Aspekte des Lebens im mittelalterlichen Städten (Berlin, 1996), 59–80.

Downloaded from http://www.mitpressjournals.org/doi/pdf/10.1162/JINH_a_01161 by guest on 02 October 2021 356 | IZDEBSKI, JAWORSKI, ÜSTÜNDAĞ,ANDSOŁTYSIAK flour prepared at home (poorer households tended to prepare their dough at home and then bring it to a bakery to be finished). Yet bakers in Constantinople probably used animal-powered mills capable of producing excellent flour. Even in other cities and towns, the servants of the wealthier inhabitants could usually obtain good flour for their masters’ tables. Furthermore, the ubiquity of watermills across the Byzantine countryside is evident in the Nomos Georgikos, a Byzantine legal collection of regulations dealing with rural life, as well as in the reports of a number of archaeological projects.20 Watermills also continued to dominate the Anatolian and northern Syrian countryside after the Byzantine state lost its grip on Anatolia. Medieval Muslim villages usually had one in close proximity if not within their own borders. Recent archaeological surveys leave no doubt that watermills remained an element of the Turish landscape throughout Ottoman times and even under the early Republic. Early Ottoman documents show that watermills were present in the vicinity of Akarçay Höyük during the six- teenth century. Thus, an important biological and dietary difference between elite, urban, and rural populations in medieval Anatolia, also perceived as a crucial marker of social status, is trace- able not just to the availability of high-quality raw products, which were typical for medieval cities, but also to inequalities in access to key food-processing technology.21 It is not difficult to explain the failure of the stereotypes in the textual sources to reflect social-biological circumstances prop- erly. The written sources, which we owe to the literacy of elites, inescapably reflect narrow perceptions of food and class, show- ing what elites considered to be an important marker of their social status. They use bread as a symbol of their superiority over

20 Michael Decker, “Agriculture and Agricultural Technology,” in Elizabeth Jeffreys, Haldon, and Robin Cormack (eds.), The Oxford Handbook of Byzantine Studies (New York, 2008), 397–406; Igor P. Medvedev (ed.), Nomos georgikos: Vizantiiskii zemledel’cheskii zakon (Leningrad, 1984), 126–128. 21 Trépanier, Foodways and Daily Life,51–54; K. Donners, M. Waelkens, and J. Deckers, “Water Mills in the Area of Sagalassos: A Disappearing Ancient Technology,” Anatolian Studies, LII (2002), 1–17. For Akarçay Höyük, see Mehmet E. Üner, “Osmanlı Döneminde Urfa’da Hanımlar Tarafından Kurulmuş Vakıﬂar,” Şanlıurfa Kültür Sanat Tarih ve Turizm Dergisi (2014), 8–18; Abdurrahman Acar and M. Mesut Ergin, “Halep’te Memlüklü Dönemine ait Medrese Vakıﬂarı,” Dicle Üniversitesi İlahiyat Fakültesi Dergisi, IV (2012), 1–36.

Downloaded from http://www.mitpressjournals.org/doi/pdf/10.1162/JINH_a_01161 by guest on 02 October 2021 BREAD AND CLASS | 357 uneducated, primitive town and village dwellers. This elite self- understanding is most easily articulated through a comparison with the lifestyle that contrasted drastically with their own— that of peasants. Nonetheless, the details of these stereotypes had more to do with the everyday realities of urban life than with divisions concerning foodways (and technology) in a wider society. This study demonstrates that environmental historians should always be aware that description in medieval sources often sacriﬁces accu- racy to other goals. The application of scientiﬁc archaeological methods is one of the solutions to this problem; combined with textual sources, it has the potential to reveal the complexities be- neath the simplistic world of medieval stereotypes.