assemblage home > issue 6 contents> Kelly-Blazeby, kapeleia
Tavernas in ancient Greece c. 475-146 BC: an archaeological perspective



The ancient Greek kapeleion, or taverna, is an institution shunned by classicists and archaeologists alike.  Preferring instead to focus on the consumption of wine within sympotic, ritual and religious contexts, this mainstay of popular life in the ancient Greek city has long been ignored.  In Pompeii, albeit at a later date, tavernas reached a density that compares to modern cities.  An assessment of their distribution in ancient Greece must, of necessity, be rather more impressionistic, but take to begin with, the laconic remark which Aristotle in the Rhetoric ascribes to Diogenes the Cynic: ta kapeleia ta Attika phiditia (‘tavernas are the canteens of Attica’). This paper discusses evidence for two possible taverna sites in Athens and Corinth.


"Just as the common messes feed and water the entire citizenry in Sparta, so the whole population of Athens can befound of an evening thronging the kapeleia".

It was this line from James Davidson's Courtesans and Fishcakes ( 1997a: 55), comparing two starkly opposed institutions; the plebeian and democratic Athenian taverna (kapeleion) [ Footnote 1and Sparta's communal dining-halls, which inspired my research into the archaeology of classical Athenian kapeleia. If, as Davidson claimed, the entire population of Athens thronged the kapeleia of an evening, which indeed the ancient sources would appear to corroborate [Footnote 2], then there must be a substantial body of archaeological evidence to betray their existence. The reality however, turned out to be completely the opposite. As I discovered from the very start, this seemingly well-attested ancient institution has been given no specific archaeological attention whatsoever. Beyond Courtesans and Fishcakes, written from the point of view of a classicist, and the reference-packed columns of Pauly and Wissowa's (Wissowa 1997) panoramic encyclopaedia of the ancient world, it is impossible to find any detailed study of the kapeleion and few scholars even bother to refer to it. To some extent, this neglect is a direct result of the prominence accorded to the symposium and its anthropological model of commensality, the symposium traditionally being treated as the classic context for the consumption of alcohol in ancient Greek society ( Davidson 1997a). The symposium carried over from the archaic period associations with the lifestyle of the wealthy aristocracy and their emulators, and however much the fifthcentury democracy might try to provide public dining rooms and civic occasions for feasting, the symposium would remain a largely private and aristocratic preserve.

Those beyond the aristocratic pale had to get their liquid refreshment somewhere other than the symposium, in the taverna or kapeleion, a far more demotic and promiscuous space than the private and selective andron or men's room of the house (Davidson 1997a).These tavernas sold wine (only barbarians drank beer) and vinegar (wine's natural by-product), and in some establishments you could have something to eat as well: tragemata (sweets)or hales (savouries) translated as "bar-snacks" by Davidson (1997:54). Wine was stored in amphorae, their shape betraying the region they came from. Once the amphora was opened, the wine would be decanted into a krater (mixing bowl) , cooled using a psykter or wine-cooler which stood inside the krater, served from oinochoai or olpai (jugs) and drunk from a variety of plain or decorated cup shapes such as the kylix , kothon and kantharos. Ancient wine was much stronger than today (around16 per cent) and was always mixed with water, one part wine to three parts water being the preferred admixture. Like today, it came in three different types, austeros (dry), glukazon (sweet) and autokratos (medium-sweet or somewhere between the two) and could be red, white or rosé. 

In order to establish a criterion against which to test the archaeological evidence, it was essential to determine a likely commonality between these establishments as a point of departure. As wine was always drunk mixed with water, an established taverna would need access to a constant supply, most likely in the form of a well or cistern (lakkos) on the premises. It was also essential to identify a likely assemblage of pottery drinking paraphernalia. All establishments needed the drinking pottery outlined above as well as, depending on the size and quality of the kapeleion, cooking wares such as escharai (braziers for grilling spitted meat), lopades (casseroles), griddles and chytrai (cooking bells) and possibly lamps, coins and items of entertainment such as flutes or knucklebones (used as gaming pieces). 

After an examination of the archaeological evidence from the ancient agoras [Footnote 3] of Athens and Corinth (two of the most comprehensively published sites in Greece), it was possible to identify only two areas which may fit into this picture. It is those areas which will be discussed in this paper. 

Kapeleia in Athens

Early in the archaic period (roughly the eighth to sixth centuries BC), the central, western and southern sections of the Athenian agora were cleared of private dwellings in order to serve as the city's official civic centre. The great street of the Panathenai, named after the Panathenaic Festival in honour of Athena, cut diagonally through the agora on its way up to the Acropolis.The land east of this road remained outside the civic centre, as is evidenced by wells belonging to private structures. Only three wells in the area date to the seventh century BC, two belonging to its last quarter. In the sixth century BC this number increases significantly to eleven. Six of these wells belong to the first half of the sixth century, four to the second. One which was left unexcavated, could not be dated closely (Camp 1986). It would seem that in the sixth century BC this neighbourhood, because it now adjoined the newly formed square, had again become desirable, increasing the demand for water, which in turn led to the renewed construction of wells. 

In the late fifth century BC, when municipal buildings stood at the north-west corner and at intervals along the west and southern borders of the agora, the space to the eastof the Panathenaic Way, which cut diagonally across the square from north-west to south-east, was occupied by a patchwork of private shops and houses.It was not until the very end of the fifth century BC that public structures came to be erected on the site (Thompson 1976).Though by no means uninhabitable, this area did not offer some of the natural advantages found to the west and south. It was a little farther away from both the Acropolis and the Kerameikos, and it was more exposed, lacking the protection from wind and rain found at the foot of Kolonos Agoraios and the Areopagus (fig.1). 

fig. 1 (site in pink)

There is also evidence that well water may not have been so reliable as elsewhere (Camp 1977). Only one well dates from the fifth century, and the deliberate abandonment of the archaic wells suggest that local proprietors may not have felt encouraged to re-open their businesses in the area when they returned home after the Persian sack of Athens in 480 BC. Perhaps they found better accommodation with greater water supply, or buildings more advantageously positioned elsewhere. Some of these shopkeepers may have established themselves in a building in the south-west corner where my research is concentrated. Here, a new construction incorporating a row of five or more shops lay at a busy intersection where the Panathenaic Way joined the main road that crossed the northern edge of the agora. In later periods, this road would become the principal route to the Roman market. 

In the fifth century BC, the Panathenaic Way was the only agora thoroughfare open to wheeled traffic, as all other main roads traversing the area were broken up by flights of steps. Traffic in this locality, both wheeled and pedestrian, would therefore have been extremely busy (Young 1951). Although it was this district, lying east of the classical market square, that first came to be occupied by monumental public buildings in the Roman period, the site appears to have had a long and complex earlier history.

Wine Retailing in theSouth-east corner of the Agora

This particular region of the agora has been connected with wine-selling in earlier publications, especially by Lucy Talcott (Talcott 1935: 477-523), T. Leslie Shear (Shear 1975: 357-8) and most recently Mark Lawall (Lawall 2000) whose specific interest lies in well R13:4 and the mass of graffiti-bearing amphorae found there (fig 2.). 

fig. 2

The many amphorae, drinking-cups, mixing-bowls and cookware found within this well prompted Talcott to propose the existence of a taverna,which "flourished near the borders of the agora in the years around 440BC " (1935: 497). Apparently it met with a disaster, which caused its abandonment and the discarding of its paraphernalia somewhere around 430 BC. The contents of well R13:4 were the only significant evidence for such an establishment in the area, until excavation under the nearby Library of Pantainos in 1970-74 unearthed extensive remains belonging to classical buildings preserved beneath the floors. Altogether the plans of 14 rooms were discovered, belonging to at least three separate buildings. The rooms shared a common front wall and evidently lined the south side of an early street, which followed closely the course of the later marble-paved street of the Roman period. Underlying the eastern half of the Library are two structures which both in architecture and history were almost inextricably entwined with each other (figs.3a, 3b and 3c). 

fig. 3a

fig 3b
fig 3c

The two contiguous buildings survived in a state that has made their history fairly easy to reconstruct. The essential feature of the plan appears to have been a series of pairs of rooms and, in all its phases, a long median wall bisected the building longitudinally to create these double compartments. In one or two instances there is evidence that the front and back rooms communicated through doors in the median wall, and this is likely to have been the case generally. Most, if not all, of the front rooms had access from the street, although, due to the fact that the remains of the front wall lie immediately beneath the marble pavement of the Roman street, this could only be verified by the excavators when there were breaks in the pavement (Shear 1975). 

A peculiar aspect of the south wall was the fact that its blocks were cut away around the top of a pre-existing well U13:1 in Room 6. The existence of this well, and of another later and larger well close by in the same room, led Shear to suggest that Room 6 was in fact an "open courtyard to which the surrounding rooms had access", though he gives no further justification for this (Shear 1975 : 357-8). The fill from the larger well contained hundreds of objects which had been used and discarded by the occupants of the surrounding rooms, but it is well U13:1 that is of particular interest for the material which it contained. 

Like its neighbours, well U13:1 was full of amphora sherds datable to the early years of the fourth century BC. Around 390BC, a bedrock collapse of the walls seems to have ended the well's use as a source of water and started its life as a rubbish dump. A massive deposit from the upper part of the shaft produced no fewer than 716 tins of broken pottery and 455 miscellaneous catalogued objects. That this enormous quantity of material was a homogeneous dump is evident, as fragments of the same pots were found dispersed from top to bottom in the deposit. This also makes a strong case for the accumulation of the dump happening over a short space of time. Shear in his report onthe excavations believed that a large proportion of the pottery was clearly refuse cast out from a nearby kitchen (Shear 1975), and it is true that many of the familiar, so called "domestic" shapes are represented: lopades, escharai, mortars, lekanai (mixing bowls), chytrai (cooking bells) and an assortment of jugs. However, the quantity in which they were found must surely exclude the possibility that they came from the kitchen of a private house. In addition to the four restored lopades and eleven lids, there were fragments of at least 100 more casseroles of this type, eleven complete mortars and parts of sixteen others. Fragments of no less than 400 lekanai were counted among the deposit and 27 of these were nearly complete, while pieces of at least 76 different escharai were also recognised. This all suggests cooking on a massive scale, and it would not be unreasonable to suggest that a large part of the building functioned as a prosperous kapeleion, perhaps even the one which Lucy Talcott believed was flourishing in the vicinity. Other types of table-ware, likewise found in abundance, were plates, small bowls and salt-cellars, which presumably would have come from the same source. As many as 79 wine jars were mended and catalogued and fragments of at least 280 others were counted. Judging by the types of amphora, this kapelos kept a good cellar which specialised in imported wines such as Mendean, Chian, Corinthian, Samian and Lesbian, in addition to the local Attic vintage. 

A taverna would also be the most obvious source of the large quantity of drinking- cups of various types which predominated among the fine ware from the well. The serving vessels, the kraters and oinochoai of the finer wares unfortunately only survived in many hundreds of small fragments.

Also included among the deposit was a large amount of animal bone and shell comprising cattle, sheep, goat, pig, equid, dog, turtle, tortoise, red deer, murex, arca oyster and mussel, suggesting that the menu was of the same quality and diversity as the wine cellar. Under examination, large numbers of the cattle, pig, sheep and goat bones showed signs of butchering, either in connection with the preparation of food for the taverna or, as Shear believed, as refuse from a neighbouring butcher's shop (Shear 1975: 357-8). There was however, a distinct lack of vertebrae and ribs and most of the bones show signs of the meat being stripped off, an act which I prefer to see not as evidence for a butcher, but as proof that the kapelos was buying in cuts (perhaps the cheaper ones) suitable for the little kebabs which would have grilled on the eschara. If indeed there was a butcher's shop in the area, surely a greater variety of bone types would have been discarded. Allied to this is the fact that there were several skulls in which holes had been cut to extract the brain. This remains a favourite meal in Greece today and individuals, especially in the countryside, buy the whole head. This argues strongly for brain being on the taverna menu, bought-in whole and cracked open on the premises. A smaller amount of bones of sheep and goat survived; Shear claimed that this was because they were sold whole. If that was the case, then it may also point to consumption on the premises as opposed to sale from a butcher's shop. The small amount of sheep and goat bones found their way onto the menu and subsequently into the deposit. If they were sold whole, as Shear believed, their inclusion in the deposit at all would have to be explained. 

Likewise, large numbers of fish bones were recovered, the majority of which were long thin bones associated with the inedible fins of large fish. It is questionable whether or not fish would have been sold in a butcher's shop, so they might also be considered as debris from the taverna menu. A bone flute and "several lamps" from layer 1 are recorded in the excavator's notebook, though they are not mentioned in the final Hesperia publication. 

Kapeleia in Corinth

Despite extensive excavations of the town of Olynthus in northern Greece and of settlement sites elsewhere, the only other possible taverna so far recognised lies in the South Stoa at Corinth, a large commercial building which lined the southern edge of the agora (fig. 4). 

fig. 4

During the excavation season 1946-47 (Broneer 1947:239-41), several shop wells dating to the third century BC were cleared, exposing an elaborate series of thirty-one connected wells (with the exception of two, one was situated in each of the front shops). The channel which supplied the wells with water was linked with the Peirene spring system and lay at a depth of nearly 12 metres beneath the shops. None of the excavated well curbs display any tell-tale rope marks made during the act of drawing up water, leading the excavators to believe that these wells had been used predominantly for cooling foodstuffs. They state that they "furnish the best example of ancient refrigeration preserved in Greece" ( Broneer 1947: 239). The objects from the lowest levels, most of which had found their way into the water while the wells were still in use, help to determine the kind of business carried on in the shops prior to their destruction. What is clear from the material recovered from the fill is that many of these shops were tavernas and that the wells were probably used not only for their water, but also for chilling wine. Pottery recovered came, in the most part, from Knidian and Rhodian amphorae and from amphorae "from other parts of Greece and even from more distant centres", though exactly where is not made explicit (Broneer 1947: 240).  Drinking-cups were numerous and belong to the full range of shapes, the most typical being the kantharos. Many lamps were recovered, most intact and showing signs of use. This prompted the excavators to suggest that although these tavernas would have been lit chiefly through the open doors of the stoa, the lamps that found their way into the wells unbroken would have "fallen into the water by accident, having been placed on the well curb or on the wooden contrivance for lowering the wine and victuals into the wells" ( Broneer 1947: 241).

Table tops of red and white marble were claimed from several of the wells (though from which we are not told) along with large numbers of knucklebones, all smooth and blackened from frequent use (one single deposit contained an impressive fifty-eight pieces). Further evidence for entertainment comes from four fragments of bone or ivory flutes. The four remnants came from different wells, and the excavators believed that they belonged to more than one instrument.


In this article, I hope to have demonstrated convincingly that although no hard archaeological evidence has been recognised as yet for classical kapeleia, there is definite scope for further research on the subject. In the past, when a substantial body of "drinking" pottery was encountered, it has almost always been misinterpreted as debris from a public dining-place, domestic kitchen or wineshop, with only LucyTalcott (1935) and Mark Lawall (2000) recognising the taverna as a viable archaeological entity. 

Excavators need to recognise that there is a further option when faced with a large and anomalous "drinking" assemblage. When I met with the directors of the Athenian Agora and Corinth excavations, my query regarding possible tavernas was met with a degree of wonderment, yet both of these excavations have turned up the most likely kapeleia sites. Wine-shops were mentioned, as were public dining-rooms and ritual drinking practices, but nobody was prepared to consider the existence of a taverna.

The basis for this misinterpretation may be partly due, as I mentioned in the introduction, to the prominence accorded to the symposium in traditional accounts of classical Greek drinking practices. Unsurprisingly, tavernas do not sit easily with the old-school classicists' perception that Greek "culture" was everything civilised and noble, a view which is slowly beginning to change. 

Thanks are due to Mark Lawall of Manitoba University for discussing his work on well R13:4 with me and
allowing me to use his plan of the area, and to Graham Shipley of Leicester University for his keen proof-reading skills. Also a big thank you to John Camp and all the staff of the Athenian Agora Excavations for giving me such generous access to their library and time (and cream-cheese and strawberry sandwiches...!).


[1] From the Greek taberna, a commercial drinking and eating place. 

[2] In the comedies of Aristophanes tavernas appear as an already well-worn feature of the urban environment; their staff in particular are the frequent target of jokes. See also Plato (Gorgias), Hyperides (Against Patrocles), and Gager (1992) for evidence from curse tablets (katadesmoi).

[3]The agora (literally 'gathering place') was the civic and commercial heart of the ancient city where market-place and law court would rub shoulders.


Austin,M. M. and Vidal-Naquet, P. 1977. Economic and Social History of AncientGreece. Berkeley: University of California Press.

Broneer, O. 1947. Investigations at Corinth, 1946-1947. In Hesperia 16. pp. 239-41, American School of Classical Studies.

Camp, J.M. 1977. The Water Supply of Ancient Athens from 3000-86 BC.  Dissertation.  Princeton University.

Camp, J.M. 1986. The Athenian Agora.  London: Thames & Hudson.

Casson, L. 1984. Ancient Trade and Society. Detroit: Wayne State University Press.

Davidson, J. 1997a. Courtesans and Fishcakes; The Consuming Passions of Classical Athens. London: Fontana Press.

Davidson, J. 1997b. A Ban on Public Bars in Thasos? In Classical Quarterly 47(ii). pp. 392-5. 

DeCoulanges, F. 1991. The Ancient City. Baltimore: The John Hopkins University Press.

Douglas, M. 1987. Constructive Drinking: Perspectives on Drink from Anthropology .Cambridge: Cambridge University Press.

Finley, M. I. 1999. The Ancient Economy. Berkeley: University of California Press.

Fisher,N. R. E. 1992. Hybris: A Study in the Values of Honour and Shame inAncient Greece. England: Warminster.

Gager, J. G. 1992. Curse Tablets and Binding Spells from the Ancient World. England: Oxford University Press.

Garnsey, P. 1999. Food and Society in Classical Antiquity. Cambridge: Cambridge University Press.

Glotz, G. 1926. Ancient Greece at Work. London.

Grace, V. 1979. Amphoras and the Ancient Wine Trade. Athenian Agora Picture Books. New Jersey: American School of Classical Studies.

Hunter, V. J. 1994. Policing Athens. Princeton: Princeton University Press.

Immerwahr, H. R. 1948. An Athenian Wine Shop. In Transactions of the American Philological Association, vol. 79, pp. 184-190

Lawall, M. 2000. Graffiti, Wine Selling, and the Reuse of Amphoras in the Athenian Agora, ca. 430 to 400 BC. Hesperia 69, number 1, pp. 3-90, American School of Classical Studies.

Meijer, F. and von Nijf, O. 1992. Trade, Transport & Society in the Ancient World. London: Routledge.

Murray, O and Tecu?an, M (eds.). 1995. In Vino Veritas. Oxford: Alden Press.

Parkins, H and Smith, C (eds.). 1998. Trade, Traders & the Ancient City. London: Routledge.

Rotroff, S. I. and Oakley, J. H. 1992. Debris from a Public Dining Place in the Athenian Agora. In Hesperia: Supplement XXV. New Jersey: American School of Classical Studies.

Seltman, C. 1957. Wine in the Ancient World. London: Routledge.

Shear, T. L. (Jr). 1975. The Athenian Agora: Excavations of 1973-4. In Hesperia 44, 346-361. American School of Classical Studies.

Slater, W. J. (ed.). 1991. Dining in a Classical Context. Michigan: The University of Michigan Press.

Sparkes, B. A. 1962. The Greek Kitchen. In Journal of Hellenic Studies 82, pp. 121-37.

Sparkes, B. A. and Talcott, L. 1951. Pots and Pans of Classical Athens . Athenian Agora Picture Books. New Jersey: American School of Classical Studies.

Talcott, L. 1935. Attic Black-Glazed Stamped Ware and other Pottery from a Fifth Century Well. In Hesperia 4, pp 477-523. American School of Classical Studies.

Talcott, L. 1936. Vases & Kalos Names from an Agora Well. In Hesperia 5, pp 333-354. American School of Classical Studies.

Thompson, H. A. 1976. The Athenian Agora: A Guide to the Excavation and Museum. Greece: Ekdotike Hellados.

Tredennick, H. (trans.). 1970. Xenophon: Memoirs of Socrates & the Symposium. England: Penguin Books Ltd

Wissowa, G. 1997. Paulys Realencyclopaedie der Classichen Altertumwisserschaft Gesamtregister. Stuttgart: Metzler

Young, R. 1951. An Industrial District of Ancient Athens. In: Hesperia 20, pp. 135-288.  American School of Classical Studies.

Ancient Sources

1973 The Wasps/The Poet and the Women/TheFrogs. Penguin Classics, A. Sommerstein trans.

1978 The Knights/Peace/The Birds/The Assembly Women/Wealth. Penguin Classics, D. Barrett & A. Sommerstein trans.

1973 Lysistrata/The Acharnians/The Clouds.Penguin Classics, A Sommerstein trans.

1871 Gorgias. G. Long ed., 518b, p. 156. London: Whittaker & Co

1970 Memoirs of Socrates and the Symposium.Penguin Classics, H. Tredennick trans. 

Clare Kelly-Blazeby
Clare Kelly-Blazeby is currently a PhD student at the University of Leicester where she is expanding on research undertaken for her MPhil (distinction) into the archaeology of Classical Greek tavernas (kapeleia). Clare is very interested in all feedback/comments and thoughts on the subject of commercial wine consumption in Ancient Greece. She can be contacted at:

School of Archaeology & Ancient History
University of Leicester
University Road


© C. Kelly-Blazeby 2001
© assemblage 2001
Issue 6 contents

assemblage - the University of Sheffieldgraduate journal of archaeology
Research School of Archaeology andArchaeological Sciences 
University of Sheffield 
2 Mappin Street, 
Sheffield S1 4DT 
Tel: (0114) 222 5102 Fax: (0114) 272 7347
page updated: 14.8.2001