CONTEXTUAL REANALYSIS OF THE ARCHITECTURAL FORM AND SOCIO-ECONOMIC STATUS OF THE PANAGIA HOUSES AT MYCENAE

Abstract

The Panagia Houses, one of the building complexes at Mycenae, located to the south-west of the Citadel, were interpreted by the excavators as a group of three independent units. They were constructed and occupied during the Late Helladic (LH) IIIB phase, with two main phases distinguishable in the archaeological record, followed by a reoccupation phase. Careful analysis of the archaeological data published by Mylonas-Shear, which focused on the arrangement of individual units, formality of layout, access and movement patterns, and visibility, suggests that the group should be interpreted rather as a single complex, developed during the period of the maximum expansion of the settlement. The building was gradually expanded to form an extended household, with primary living space surrounded by a number of additional rooms for storage, work and habitation. It was composed of a rectangular main unit, surrounded by an extension formed around a small inner courtyard. The movement between the two levels was organised through a system of connected rooftops, with trapdoors and staircases ensuring access to the various rooms of the complex. During the LH IIIB2 Early phase, Mycenae was hit by a devastating earthquake. Much of the town was left in ruin, but the Panagia Houses were rebuilt, although in a smaller form, with the main unit abandoned and the courtyard inside the extension transformed into the main room of the complex. The status of the complex probably changed, but it was still inhabited by a middle-class family, who possessed a number of valuables and took part in the palatial mobilisation system. The household probably suffered in a widespread fire which destroyed Mycenae at the end of LH IIIB. Its remains were then used as two small dwellings by the survivors of the catastrophe. The history of the complex reflects the changing vicissitudes of the Lower Town of Mycenae.

Ανάλυση του Αρχιτεκτονικού Μορφώματος και της Κοινωνικοοικονομικής Κατάστασης των Οικιών Παναγίας στις Μυκήνες

Οι Οικίες Παναγίας, ένα από τα οικοδομικά συγκροτήματα στις Μυκήνες, βρισκόμενες στα νοτιοδυτικά της Ακρόπολης, ερμηνεύθηκαν από τους ανασκαφείς ως ομάδα τριών ανεξάρτητων μονάδων. Κατασκευάστηκαν και κατοικήθηκαν κατά τη διάρκεια της Υστεροελλαδικής (ΥΕ) ΙΙΙΒ φάσης, με δύο κύριες φάσεις να διακρίνονται στα αρχαιολογικά ευρήματα, ακολουθούμενες από μια φάση επανακατοίκησης. Η προσεκτική ανάλυση των αρχαιολογικών δεδομένων που δημοσιεύτηκαν από τους Mylonas-Shear, επικεντρώθηκε στη διάταξη των ατομικών μονάδων, στη φορμαλιστική διάταξη, στα μοτίβα πρόσβασης και κίνησης και στην ορατότητα, υποδηλώνει ότι η ομάδα θα πρέπει να ερμηνευθεί μάλλον ως ένα ενιαίο συγκρότημα, που αναπτύχθηκε κατά την περίοδο της μέγιστης επέκτασης του οικισμού. Το κτίριο επεκτάθηκε σταδιακά για να σχηματίσει ένα εκτεταμένο νοικοκυριό, με τον κύριο χώρο διαβίωσης περιβαλλόμενο από μια σειρά επιπλέον δωματίων για αποθήκευση, εργασία και κατοίκηση. Αποτελούνταν από μια ορθογώνια κύρια μονάδα, περιβαλλόμενη από μια επέκταση γύρω από μια μικρή εσωτερική αυλή. Η κίνηση μεταξύ των δύο επιπέδων οργανώθηκε μέσω ενός συστήματος συνδεδεμένων οροφών, με καταπακτές και σκάλες που εξασφάλιζαν την πρόσβαση στα διάφορα δωμάτια του συγκροτήματος. Κατά τη διάρκεια της πρώιμης φάσης ΥΕ ΙΙΙΒ2, οι Μυκήνες επλήγησαν από έναν καταστροφικό σεισμό. Μεγάλο μέρος της πόλης καταστράφηκε, αλλά οι Οικίες Παναγίας ανακατασκευάστηκαν, αν και σε μικρότερη μορφή, με την κύρια μονάδα να εγκαταλείπεται και την αυλή μέσα στην επέκταση να μετατρέπεται στο κύριο δωμάτιο του συγκροτήματος. Η κατάσταση του συγκροτήματος πιθανώς άλλαξε, αλλά εξακολουθούσε να κατοικείται από μια οικογένεια μεσαίας τάξης, που κατείχε έναν αριθμό πολύτιμων αντικειμένων και συμμετείχε στο ανακτορικό σύστημα κινητοποίησης. Το νοικοκυριό πιθανώς υπέφερε σε μια εκτεταμένη πυρκαγιά που κατέστρεψε τις Μυκήνες στο τέλος της ΥΕ ΙΙΙΒ. Τα ερείπιά του χρησιμοποιήθηκαν τότε ως δύο μικρές κατοικίες των επιζώντων της καταστροφής. Η ιστορία του συγκροτήματος αντανακλά τις μεταβαλλόμενες περιπέτειες της Κάτω Πόλης των Μυκηνών.

Μετάφραση: Αναστασία Κασιμίδου

INTRODUCTION

Mycenae is one of the most important Late Bronze Age sites on the Greek mainland, and one of the most extensively excavated and thoroughly researched. Multiple archaeological campaigns at the site have uncovered around 70 Mycenaean buildings, grouped in over a dozen complexes.¹ Those are largely concentrated on the Citadel, a heavily fortified hill with the palatial complex set on top of it, which dominated the site spatially and visually. The Lower Town of Mycenae is much less known, with only a handful of locations excavated across the ridges forming the site (Fig. 1). Despite the extensive archaeological survey conducted at Mycenae in the 1990s (French et al. 2003), followed by the intensive geophysical research of the terraces located immediately below the Citadel in the early 2000s (Maggidis and Stamos 2006), our knowledge of the settlement surrounding the walls of Mycenae remains fragmentary and based mostly on the data from isolated buildings.² The present article aims to further supplement our knowledge of Late Bronze Age Mycenae through a reanalysis of the archaeological data and reinterpretation of the architectural form of the Panagia Houses as a single complex, an extended family household. This analysis leads to a discussion of the changing economic role and social position of its inhabitants in the wider context of the settlement and its community.

Fig. 1. Map of Mycenae (based on Maggidis and Stamos 2006, fig. 2, modified and updated by Katarzyna Dudlik and the author).

The Panagia Houses complex is located c. 400 m to the south-west of the Citadel, on the east slope of the Panagia Ridge, next to the Treasury of Atreus (Fig. 1:8). They were excavated under the auspices of the Greek Archaeological Service by the team supervised by George Mylonas. The works conducted in the years 1962–3 and 1965–6 uncovered the building remains, interpreted by the excavators as three separate housing units (Mylonas 1966a, 65–6; 1966b, 104–6; 1967, 94–5; 1968, 111). Additional tests were also dug in 1977, in order to obtain ceramic material which would facilitate dating of the architectural phases. The complete results of those excavations were later published by Ione Mylonas-Shear (1987, xix).

The Panagia Houses were constructed and occupied during the Late Helladic (LH) IIIB phase, which corresponds roughly to the thirteenth century BC. Two general phases of occupation can be recognised in the complex at this time, separated by a devastating earthquake that struck Mycenae during the LH IIIB2 Early phase (French 1996; French and Stockhammer 2009, table 3). Later, at the end of LH IIIB or shortly before it, the houses were destroyed in a fire and then were partially reoccupied until LH IIIC Early, that is until the first half of the twelfth century BC. Neither the architectural form nor the socio-economic status of the Panagia Houses were fixed across their entire history, reflecting the changing living conditions of the Lower Town.

At the time when the Panagia Houses were constructed, the Lower Town of Mycenae developed intensively, and expanded to the north, west and south. The entire settlement reached its maximum size of c. 25 ha (Zeman and Dudlik 2023, 141–2). Apart from the Panagia Houses, a number of other structures from this period were excavated across the site, outside the Citadel walls (Fig. 1). These included a building complex constructed around the Cyclopean Terrace Building on the western slope of Pezoulia (French 1961; Fig. 1:3), the House of the Tripods Tomb and the ‘Workshop’, both located to the north of the Citadel (Onasoglou 1995; Danielidou 2008; Fig. 1:4–5), the Plakes House built further north behind the Kokoretsa ravine (Iakovidis 2013; Fig. 1:6), the West House Group which gradually expanded on the east slope of Panagia, in front of the Citadel (Tournavitou 1995; Fig. 1:7), and the House of Lead located further south, on the tip of the Panagia ridge (Wace 1956, 119–21; Fig. 1:9).

The following analysis is part of a much larger study of Mycenaean palatial towns, which seeks to comprehensively analyse LH palatial sites as urban settlements. Within this study I use a comparative urbanism approach (Smith 2007) to define settlement zones and groups of coordinated buildings forming complexes, such as the Panagia Houses. Although the structure has been traditionally recognised as three separate units, following the initial assessment by the excavators (Mylonas-Shear 1968, 3; Hiesel 1990, 77–9, 121, 149; Darcque 2005, plans 22–5), a change of interpretation is enabled by abandoning the typological approach to Mycenaean architecture that has previously dominated the study of LH urbanism and town planning.³

Post-processual urban archaeology, despite the changes caused by the abandonment of many of the classificatory culture historical paradigms, was long impeded by the continued search for categories that would describe ancient cities in terms of the presence or absence of certain traits, even if those were no longer necessary to define a settlement as urban. This was probably most apparent in an approach to urban planning, which has for a long time centred around a false dichotomy of defining planned versus unplanned cities. This approach was largely derived from Classical archaeology and its study of Greek and Roman towns that were often built with orthogonal layouts, but such a dichotomy was also sought in other parts of the world, including Mesoamerica and Asia (see for example Owens 1992, 1–30; Morris 1994, 8–10; van de Mieroop 1999, 93–4; Crouch and Johnson 2001, 210–13). A response to this approach was formulated by Michael Smith (2007). He defined the idea of comparative urbanism, and this constitutes the foundation of contemporary urban archaeology. This approach, which uses the functional definition of urbanism,⁴ rejects the dichotomy of planned versus unplanned cities, and replaces it with a much more complex consideration of spatial principles that formulates a series of ordinal scales and defines various degrees of urban planning, instead of acknowledging only the presence or absence of a certain characteristic. It is based on the two main components that are to be considered: coordination among buildings and spaces in a city and standardisation among cities. The standardisation among cities is discussed by Smith (2007, 25–9) in relation to architectural inventories (groups of similar buildings among cities of the same archaeological culture, civilisation or historical political entity), spatial patterns, orientation and metrology. However, it is the other component of the comparative urbanism approach that is crucial for the present study, namely the coordination among buildings and spaces, so the arrangement of the architectural features with reference to one another. The key aspects to analyse alongside coordination in regard of a single complex are formality and monumentality of layout (Smith 2007, 8–12). Formality of layout refers to a plan with clearly observable organisational principles (Taylor 1981, 65–8), while monumentality is a characteristic of buildings that are larger than they need to be for utilitarian purposes (Trigger 1990). Other factors which should be taken into account are access and movement patterns in and between the buildings, and visibility. The latter refers to two aspects of visual perception: outward viewshed, defining the area that can be seen from a given point, and inward viewshed, defining the areas from which a given point can be seen (Smith 2007, 23–4). Smith has also tied his approach closely with Rapoport's idea of levels of meaning in the built environment (Rapoport 1982), using analysis of the coordination and standardisation among buildings and cities, to recognise differing levels of meaning in settlement plans (Smith 2007, 30).

The comparative approach to urbanism has rarely been used in Aegean archaeology, especially in regard to mainland Greece. Quentin Letesson and Carl Knappett published an important volume on Minoan architecture and urbanism, with many of the papers drawing from comparative urbanism, focusing on patterns of spatial arrangements and planned developments (Letesson and Knappett 2017). In this volume they also published, together with Smith, a summary of the comparative perspective on Minoan urbanism, focusing on the organisation of residential spaces, town planning and regional settlement patterns and hierarchies (Letesson, Knappett and Smith 2017). The comparative approach to urbanism was also used by Kalliopi Efkleidou (2021) for the study of Mycenae during the palatial period. She recognised various elements of urban planning at the site, analysing mainly location, visibility and coordination of structures, and tried to recognise middle-level meanings behind various building activities, to explain changes occurring throughout LH IIIA and IIIB (Efkleidou 2021, 33–5).

THE HOUSES AND THEIR CONTENTS

The summary of the archaeological data published by Mylonas-Shear (1987) is presented below. Her division of the complex into three units – House I (Fig. 2:1–7), House II (Fig. 2:8–21) and House III (Fig. 2:27–38) – is followed for consistency with earlier literature and because it reflects the topographical distribution of the architecture. It should be emphasised that despite the criticism and an attempt to reinterpret certain elements discussed by Mylonas-Shear, this study was possible only thanks to the exceptional quality of her publication.

Fig. 2. Plan of the Panagia Houses with walls of LH IIIA date shaded in grey (after Mylonas-Shear 1987, modified by the author; courtesy of the Penn Museum).

House I

House I was constructed at the beginning of LH IIIB1, as suggested by the construction deposit with multiple sherds with vertical whorl shells (Mylonas-Shear 1987, 68). The building covered some earlier LH IIIA2 architecture with clay floors, of unknown form and function (Mylonas-Shear 1987, 26). There were also two substantial terraces built to the west and south of this house, most probably to support other houses, which were not preserved. No pottery was found which could date the terraces, but they must have been built either before or together with House I, because the later constructions in the area were adjusted to them (Mylonas-Shear 1987, 2).

The building consisted of an external courtyard (Fig. 2:1) with two rows of rooms around a long corridor (Fig. 2:6), although only the part west of the corridor was fully excavated. On this side of the corridor, a megaron-type unit was located, with a vestibule (Fig. 2:3), serving as the main entrance to the house, a main room (Fig. 2:5), and a rear basement room (Fig. 2:7) behind the latter (Mylonas-Shear 1987, 16–24). The staircase in the corridor led to an upper storey, that must have existed above room 7, which had no ground floor doorway (Mylonas-Shear 1987, 20–1).

The house had a modest form, with walls of mudbrick on rubble masonry set directly on bedrock. Pieces of mud-brick were found across the building. The walls were plastered and at least partially painted. Some rooms could have been decorated with patterns, although the distribution of the plaster fragments does not allow the reconstruction of any of their details (Mylonas-Shear 1987, 136–41). All the floors were of clay, with no doors, stone or wooden thresholds, or any other complex features.

Room 5 was equipped with a hearth, around which a terracotta female figurine, a chimney pot, a large pithos, and three cooking tripods (Furumark Shape [FS] 320) were found (Mylonas-Shear 1987, 17–19), along with a group of other pottery vessels which included a mug (FS 226) decorated with running spirals (Furumark Motif [FM] 46), a Group A deep bowl (FS 284), a red monochrome krater (FS 281), a small jug (FS 120) with concentric half circles (FM 43) on the shoulder, a miniature ladle (FS 236), three small decorated stirrup jars (FS 171), and one coarse undecorated jug (FS 74), all of which can be dated to LH IIIB (Mylonas-Shear 1987, 68).

The eastern part of the building was much worse preserved than the western part, although there was a drain found in room 2. A single female terracotta figurine was also found in this room (Mylonas-Shear 1987, 22). No finds were associated with room 7, while the arrangement of the space above it remains uncertain (see below).

The house was probably destroyed in an earthquake, as no traces of fire were found in the remains, while a skeleton covered by stones was found in the doorway of the main room.⁵ It probably belonged to a woman killed by the collapsing structure. The building was never rebuilt, and thus this part of the complex was abandoned. However, the shallow stratigraphy does not allow a reconstruction of the appearance of the collapsed ruins during the later history of the complex (Mylonas-Shear 1987, 19).

House II

The area to the north and north-west of House I was overbuilt probably shortly after the latter was constructed, with the same form of construction, namely low rubble stone foundations topped by a mud-brick superstructure. House II covered some earlier remains of LH IIIA date, with only a few walls identified, mainly the thick corner underneath the floors of room 8 (Mylonas-Shear 1987, 66; Fig. 2). There are two construction deposits for the building: one below the earlier floor in room 15 with LH IIIB material, and the other one below the floor of room 18 composed of a mix of LH IIIA and IIIB pottery (Mylonas-Shear 1987, 36, 46). Although their dating is broad, the architectural sequence is clear even without the pottery, as the walls of House II are built in relation to the north-west corner of House I, and thus the construction of the former building must have followed that of the latter (Mylonas-Shear 1987, 68).

House II was composed of a series of simple, poorly arranged rooms, with scarce evidence for plastering of the walls (Mylonas-Shear 1987, 136).⁶ Originally the south part of the building probably consisted of a series of five rooms (Fig. 2:8,10–13) built around space which Mylonas-Shear interpreted as room 9, but which is, during this phase, more likely to have formed an internal open courtyard (Fig. 2:9), with a large hearth and access to the drain. There might have been a stone pavement in the south-east corner of the room already during this phase. Room 10, located behind the west wall of the courtyard, did not have any doorway leading to the latter, and had communicated through a doorway with a rough stone threshold with the rooms to the south, assigned by the excavators to House III (Mylonas-Shear 1987, 31). To the north of the courtyard there was a narrow corridor (Fig. 2:14), which turned east and then again north. The northernmost part of this corridor in my opinion most probably formed a long ramp, which served as the main entrance to this part of the complex, leading to a doorway in the north wall of room 8 and to the corridor joining with the courtyard.⁷ To the west of the ramp, five additional rooms were arranged (Fig. 2:15–17 as one room, 18,19,21,23 as individual rooms), which must have been accessed through rooftops.

The finds associated with the first phase were not abundant. The only object which might be used during this earlier phase from the southern part of the building was the lower half of the ivory figurine found embedded into the floor of the second phase in the doorway between rooms 8 and 9 (Mylonas-Shear 1987, 31). Some objects from the fill located below the second phase floors of rooms 15 and 16 might have belonged to the inventory of room 15–17 of the first phase. Those included a terracotta basket, a faience bead, two fragments of bronze and a bronze arrowhead, a miniature terracotta chair, two stone buttons, and two bronze tools (Mylonas-Shear 1987, 40). Room 21 had a modest child burial set underneath the bench built of stones and soil, probably during this phase. The finds from below the second phase floor of this room might have been used during the first phase. They included a small bronze tool, a bone needle, two stone buttons, a decorated jug (FS 74) and a small bowl with pictorial animal decoration (FM 5–6) (Mylonas-Shear 1987, 43–4). Rooms 10, 11 and 12 had no deposits which could belong to them and were covered by a fill washed down from the hill and the terrace above (Mylonas-Shear 1987, 31–5).

The first phase ends with a destruction which must have been the same earthquake that was behind the collapse of House I. It is evidenced by the moved foundations of the east wall of room 11 and lack of fire traces on any of the earlier floors. Moreover, the following rebuilding and alterations of the plan suggest that a significant part of the building collapsed (Mylonas-Shear 1987, 31, 44–5).

The building was soon rebuilt and its plan changed. In my opinion only at this point was the courtyard (Fig. 2:9) roofed and transformed into the main room of the house. This is suggested by the fact that only one floor level was found in the room, despite testing. The poor quality floor was made of clay, and thus was rather unlikely to remain in use after the house was damaged. It is far more probable that it was added only during the second phase of occupation. This is also suggested by the pottery fragments found in the fill below this floor, some of which should be dated ‘well into the LH IIIB period’ (Mylonas-Shear 1987, 31). The earlier hearth was covered with a new one of the same type. The two columns located next to the hearth were also added to the arrangement only during this phase. This is shown by the fact that their stone bases were embedded in the clay floor of the room. Their addition is best explained by the need to support a roof over a large area that was previously open. The floor of the second phase was not preserved around the pavement of flat stones set in the south-east corner of the room, which probably continued to function (Mylonas-Shear 1987, 29). The floor may have also simply never been laid if the pavement occupied the area since the first phase. I would suggest that the short east wall of room 9, going over the drain which originated in House III, was also added only at this point, to close the gap and form the actual room. This is indicated by the fact that this wall does not bond with other walls around it (Mylonas-Shear 1987, pl. 8A).

The floor deposit of room 9, found within the layer of burned debris, was composed of a small bronze tool, pieces of melted lead, fragments of glass paste inlay, carbonised olive seeds, and a cooking tripod (FS 320), with a collection of other ceramic vessels (Mylonas-Shear 1987, 30). Those included a krater (FS 281) decorated with stemmed spirals (FM 49) set within a tricurved arch (FM 62), a Group A deep bowl (FS 284), an undecorated carinated kylix (FS 267), and a small decorated stirrup jar (FS 171), all of which should be dated to LH IIIB (Mylonas-Shear 1987, 68–9). Fragments of a pithos and a terracotta stand were also found in the room. An ivory ring was found in the fill above the floor, and it did not necessarily originate from House II. This is also true of the broken seal-stone which was found in room 8. However, the two almost complete vases from this room – a decorated stirrup jar (FS 171) and a decorated jug (FS 120) – probably belonged to the floor deposit (Mylonas-Shear 1987, 28–31).

In rooms 10, 11, 12, 18 and 23, there were no new floors laid, and no traces of fire which ended this phase. They were thus probably abandoned after the first destruction. The north part of the building was rebuilt in a different arrangement, consisting of a series of smaller storerooms (Fig. 2:15–17), and an additional room built on the terrace above (Fig. 2:20), while a new floor was laid in room 21. Of those rooms only one, 17, had a doorway leading into the corridor, thus the rest must have still been accessed through the roofs (Mylonas-Shear 1987, 31–43). There was probably no actual full-scale second floor above those rooms, as the walls are rather thin, and no staircase was identified in the area. Stratigraphy also does not support the existence of a second floor, because the clay fragments of roof were found on top of the destruction deposit, which was also not very deep (Mylonas-Shear 1987, 38–41).

All of the rooms to the north were filled with a layer of burned building material and other debris. The richest deposits were found in rooms 15, 16 and 17. They included a glass ornament, a piece of a bent gold wire, two terracotta figurines, a spindle whorl, fragments of melted lead, an obsidian blade, carbonised lentils, three clay sealings located on the floor of room 15 and made from two different seal stones, an unbaked clay pithos and a collection of ceramic vessels. The latter formed a standard Mycenaean domestic group of LH IIIB. It included two small decorated stirrup jars (FS 171), one medium-sized decorated stirrup jar (FS 170), three amphorae (FS 67, 68 and 69), two unpainted hydrias (FS 128), a few other jugs and jars, two unpainted conical kylikes (FS 265), one unpainted carinated kylix (FS 267), two unpainted shallow angular bowls (FS 295), and two Group A deep bowls (FS 284) (Mylonas-Shear 1987, 37–41). Many of the vases found within the destruction deposit were damaged or located in the upper parts of the fill, while fragments of the same vases were found between the pottery from various rooms. Therefore, at least some of the vessels found in the rooms most likely did not belong to their inventory, but fell inside from the roofs during the collapse (Mylonas-Shear 1987, 40–2).

Another group of LH IIIB pottery was covered by burned debris in room 20 (Mylonas-Shear 1987, 44). A floor deposit from that room included a glass paste bead, an unpainted lekane (FS 290), an unpainted conical kylix (FS 267), a Group A deep bowl (FS 284), a medium-sized decorated stirrup jar (FS 170), a coarse jug (FS 74), and a small jug (FS 105) with linear decoration. Sherds of a few other open vessels were found in the burned fill of the room, and thus might have fallen inside from the collapsed roof. Those included a small cup (FS 220) decorated with a wavy line (FM 53), an unpainted angular bowl (FS 222), an unpainted carinated kylix (FS 267), and two more deep bowls (FS 248) – one of Group A and another with pictorial animal decoration (FM 2).

The second phase ended with a fire which left the burned destruction deposits mentioned in association with many of the rooms (Mylonas-Shear 1987, 46–7). Most of the building was subsequently abandoned.

After the building of the second phase was burned, habitation moved further up the slope to the west, but in my opinion there was another, third, phase of House II.⁸ The previously abandoned room 23 was re-established with new walls, together with a new corridor (Fig. 2:22), and probably a small staircase which provided access to a new room (Fig. 2:25) built higher up on the slope. Area 22–3 was rebuilt with the use of the repaired north wall of rooms 11 and 12, and the west wall of room 21 (Mylonas-Shear 1987, 64–5). This reoccupation occurred shortly after the fire which ended the previous phase. There was a small floor deposit in room 23, which contained two LH IIIB–C semiglobular cups (FS 215) with linear decoration and monochrome interiors.⁹ Other finds from that deposit included a bronze fibula, a bronze pin or needle, three disc-shaped stone buttons, two pieces of glass paste inlay, two beads and a whetstone (Mylonas-Shear 1987, 66). However, in general the north-west area of excavations was extremely poorly preserved, with no actual floor or destruction deposits, and only partially uncovered, and thus the interpretation of the architectural sequence is not certain. No reason or date for abandonment of this last phase of the building is known, but it probably shared the occupational history of House III.¹⁰

House III

This building was in the south-west part of the complex.¹¹ It comprises a series of small rooms with clay floors and simple arrangement without any decoration. The mud-brick superstructure in this part of the complex left no traces, but it must have existed.¹² The shallow, mostly eroded fill and poor preservation of the walls, especially in the east part of the building, makes all interpretations problematic. This part of the complex was probably originally developed in LH IIIB1, together with House II or shortly after it, not long after the original construction of House I. The sherds found under the earliest floors in rooms 29 and 31 were of general LH IIIB date, but the coordinated construction is indicated by the presence of multiple walls abutting the east wall of House I, the south wall of House II, and the terrace to the west (Mylonas-Shear 1987, 52, 56). The building covered some earlier remains, probably of LH IIIA date, preserved fragmentarily in the form of the west drain in room 30 and walls below room 35 (Mylonas-Shear 1987, 55, pl. 15A).

The first phase of this part of the complex consisted of a series of rooms arranged along the terrace to the west (Fig. 2:28–31), which communicated with each other and had a direct access to room 10 of House II, and another set of basement rooms (Fig. 2:35–8), covering the large central drain originating in room 31, and providing access to the space above room 7 through the roof. In my opinion there was most probably a small staircase (Fig. 2:34), or maybe even just a trapdoor with a ladder, providing access from the roof to the floor level of room 33.¹³ The latter led to room 29 and to the open courtyard formed between the storerooms to the west and House I to the east (Fig. 2:27,32).

Later, probably only after the earthquake which destroyed House I and parts of House II, House III went through significant alterations. New clay floors were laid in rooms 29–31 and 33, which was previously equipped only with a floor of packed earth. Sherds found underneath the new floor in room 31 date to LH IIIB (Mylonas-Shear 1987, pl. 51B). An additional west wall in room 29 and a buttress in room 28 were built probably to support the construction (Mylonas-Shear 1987, 53–6). The west part of the building was altered, with courtyard space limited to area 27, and a new room (Fig. 2:32) built to the south of it, while the doorway in the east wall of room 29 was blocked (Mylonas-Shear 1987, 54). Room 32 was equipped with a stone bench, built with the use of clay mortar, a technique used also to construct the west wall of room 29. This new room provided then the only access to room 33, which remained in use (Mylonas-Shear 1987, 56–8). I would suggest that the small staircase 34 also remained in use, and still provided access to the rooftop over rooms 35–8, which was now most probably connected with the newly built roof of room 9. The old west wall of room 36 has slipped eastwards off its foundations, probably due to the earthquake. However, it was rebuilt, which suggests that despite its modest appearance, it was of key structural importance, probably for the support of the roof. It might have also been used to hold the collapsed rubble which filled much of room 36. The rubble buried the central drain, which was not found in this area, and probably was already not used during this phase (Mylonas-Shear 1987, 58–9). This is suggested also by the fact that the stone slabs which were originally covering the drain in rooms 30 and 31 were partially removed (Mylonas-Shear 1987, 55–6).

There were no floor deposits preserved in the building. Finds from the fill located below the third floor in room 30, a stone tool and an unpainted carinated kylix (FS 267) may have belonged to the second phase deposit of the room (Mylonas-Shear 1987, 55).

According to the excavator, the fire which burned House II did not reach House III, as there were no traces of it in the latter building. It might have been abandoned for a short period of time, but then clearly continued to be occupied. It is possible that some of the alterations to the building described above took place then, but this is unclear, and it seems more probable that all of the east part of the building was abandoned, and occupation of the third phase concentrated to the west, in rooms 28–31, where new floors of packed earth were laid (Mylonas-Shear 1987, 55–6). The painted sherds found below the latest floors of rooms 29 and 31 were of general LH IIIB date, but the group from the former room included a fragment of an open shape with panelled decoration enclosing spirals and monochrome interior, which can date to LH IIIB2 (Mylonas-Shear 1987, 54–6, 89; Iakovidis 1986, 250–1). This last phase of occupation might have lasted until the end of LH IIIC Early. There are no floor deposits to support this, but the lack of Close Style pottery fragments of LH IIIC Middle in the fill suggests that the area was abandoned by that period (Mylonas-Shear 1987, 3).

The central drain

The central drain, mentioned above several times, must have been built before the construction of Houses II and III, as their walls were clearly coordinated with the course of the canal. At least part of the drain might have originally belonged to an earlier arrangement dating to LH IIIA, as suggested by the fact that it runs parallel to the walls of this phase located below room 35 (Mylonas-Shear 1987, 49; Fig. 2).¹⁴ Alternatively, it could have been constructed in LH IIIB, immediately before Houses II and III, as the first step in a pre-planned and coordinated development.

The drain was mostly carved in the bedrock, but stones were occasionally packed at the bottom or used to strengthen the sides. The beginning of the canal was split into two branches, originating in rooms 30 and 31. Construction of those branches must have been coordinated with the raising of the rooms. They were overlaid by stone slabs, installed most probably to cover the recesses and fill the floors. The branches then joined into one open central drain, running east under rooms 36–8, into an opening in the south wall of courtyard/room 9 and then further into the space between rooms 7 and 8. An almost complete terracotta animal figurine was found in the part of the drain which ran through room 38 (Mylonas-Shear 1987, 55, 59–60).

Yet another part of the central drain was built to the east of the south-east corner of room 8.¹⁵ This part did not join directly with the main part of the drain, but was clearly positioned to collect not only the rainwater from the roof of room 8, but also the waste coming into the space between rooms 7 and 8 (Mylonas-Shear 1987, 48).

It seems that after the collapse of some of the walls of House III, probably due to an earthquake which brought an end to its first phase, at least the western part of the central drain was abandoned. The part of it running through room 36 was completely buried by the rubble, and could not be found during excavations. Many of the slabs covering the branches originating in rooms 30 and 31 were pushed aside. This regards especially those in room 31, where the drain was left entirely open (Mylonas-Shear 1987, 55–6, 59). However, the eastern part of the central drain might have still been in use. Probably only at this point did the short east wall of room 9 fill the space between rooms 7 and 8, but the opening in that wall was constructed to facilitate use of the drain.

At a point c. 2.35 m from the east wall of room 8, beyond the end of the drain marked with a stone, there was a hollow formed by the bedrock. A large group of whole or almost whole vases was excavated within the fill inside of the hollow. This group included 15 painted kylikes (FS 258) decorated mostly with vertical whorl shells (FM 23), three unpainted conical kylikes (FS 265), six unpainted carinated kylikes (FS 267), two ladles (FS 311) with burning marks identified by the excavator as lamps, a stemmed bowl (FS 305) with monochrome interior, a krater (FS 7–9) decorated with vertical whorl shells (FM 23), and three shallow cups (FS 220) – one unpainted and two others decorated, one with stemmed spiral (FM 46) and the other with the N pattern (FM 60). The entire group, despite the presence of a few earlier vases, should be dated to LH IIIB1, and must be associated with the first phase of the complex, preceding the earthquake which destroyed House I and damaged the other two buildings (Mountjoy 1986, 93–120; Mylonas-Shear 1987, 49–51). Such a marked concentration of complete vessels strongly suggests that the deposit was formed during a single event, and did not accumulate inside the hollow over a longer period of time.

INTERPRETATION OF THE ARCHAEOLOGICAL DATA

The interpretation of this complex by Mylonas-Shear as a group of three separate, independent houses of similar plans was heavily influenced by her typological approach to Mycenaean architecture (Mylonas-Shear 1968; 1987, 144–56). Problems in confirming this view in the archaeological record have been previously noticed in the literature (see for example Darcque 2005, 317; Burns 2007, 113–14). Although Mylonas-Shear (1987, 156) clearly imagined the three houses as belonging to a unified social group, a ‘clan’ or even a ‘single family group’, and thus recognised the coordination between the three buildings, she had still tried to recognise a distinct megaron unit in each of the houses (Fig. 3).¹⁶ This has not only affected her reconstructions of Houses II and III, but also her interpretations of the architectural development, internal organisation, and planning level of the entire complex.

Fig. 3. The restored plans of the Panagia Houses according to Mylonas-Shear (1987, figs 2–5; courtesy of the Penn Museum).

In general Mylonas-Shear did not consider the fact that the supposedly separate houses shared multiple walls and a large central drain. Moreover, she assumed the presence of multiple additional elements of construction, which would be necessary to form three separate megaron-type units, but they otherwise did not appear in the archaeological record. In order to interpret rooms 27–38 as a separate house, a doorway between room 31 and room 10 has to be filled, and the entire upper floor of a megaron-type plan has to be reconstructed, together with an external staircase in area 27 (Mylonas-Shear 1987, 61). For rooms 8–18 to be interpreted as a separate house of a standardised plan, an external staircase has to be added to the south of room 8, while another internal staircase has to be put between rooms 9 and 10. Mylonas-Shear also put a doorway in the east wall of room 8, and filled the missing part of its north wall, in order to further reinforce the restoration of the central axis of the building (Mylonas-Shear 1987, 28–9, 32). She also excluded rooms 22–5 from her interpretation of House II.

These shortcomings provide good reasons to propose a complete reinterpretation of the complex, focused on a careful study of the formality and functionality of the layout, as well as possible access, movement and visibility patterns. Abandonment of the search for specific building types and megaron-type arrangements allows us to limit the number of assumptions necessary to reconstruct the appearance of the complex and frames the discussion around the actual archaeological remains and the biography of the building they represent.

The Panagia Houses are adjusted to the space formed by the two terraces to the west and north. House I, with a simple and clear plan of two wings built around the courtyard and a long corridor, was clearly set on the same general north–south orientation as those two terraces, and thus it was probably planned and built together with them (Mylonas-Shear 1987, 15). Although the rooms of Houses II and III retained this general north–south orientation, and were built around another courtyard, their development was clearly planned according to the constraints imposed by the presence of older structures (Iakovidis 1986, 249–50). In the case of those rooms the more formal plan with a clear central axis and identical orientation of the rooms, which characterised House I, was abandoned in favour of maximum use of space. Despite these differences in spatial planning, the construction of the new rooms was clearly coordinated, with many of their walls abutting the terraces or walls of House I. Understanding the functionality of this coordination requires the analysis to focus on access and movement patterns.

The key to understanding all the three houses together and their identification as parts of a larger complex is the architectural sequence of House III described above. This is due to the crucial location of House III, which linked the two other houses together, facilitated the movement between them, and served as a key access point. This sequence is evidenced by a number of detailed architectural and stratigraphic observations, which are incompatible with excavators’ interpretation of this building that depended entirely on a supposed existence of an upper floor, with a megaron-type Mycenaean house (Mylonas-Shear 1987, 61–3). Instead, this building was rather a series of small rooms forming the south part of the extension of the complex, with a basement area supporting the roof which was linked to the roof of room 7, thus connecting the main unit of the complex to the extension.

Most importantly, House III was structurally linked to rooms 10–12, interpreted by the excavators as parts of House II, together with them forming the upper part of the extension of the complex, located on the terrace built above the main unit – House I, and above the lower part of the extension formed to the north (House II without rooms 10–12 on Fig. 2). Thus, the complex could have been divided into three general parts – the main unit (Fig. 2:1–7), the lower extension (Fig. 2:8–9,13–21) and the upper extension (Fig. 2:10–12,27–38), albeit defined differently than on the original plans by Mylonas-Shear (1987, figs 2–6). This is indicated by the differences in levels of the rooms, visible on the sections from the excavations, with rooms 10–12 and 27–37 on a similar level, and significantly above the neighbouring areas (Figs 4, 5). Section A-A of the complex (Fig. 4) suggests that a roof over room 7 would be on a level that could provide access to the space to the west, located higher on the slope. This is especially probable if the area of rooms 35–8 was built over only with low, basement-type rooms, as suggested by their size and appearance. None of them had doors on the ground level, and thus they must have been accessed through trapdoors, if they were used at all and did not serve simply as a structural support for the roof. In such a case they could have even been filled in to function as terraces (Wright 1980, 62–4). Those rooms constituted the only part of the building which could have supported an upper floor, the existence of which is, however, not evidenced by any corresponding destruction debris in the fill. Moreover, area 34 is too small to form a regular room, while it is perfectly sized and located to function as a small staircase. Finally, House III was directly connected to House I by the east wall of room 33, which bonds with the buttress on the south wall of 35. The former wall directly joined with the west wall of room 5, probably to further merge the roofs of two parts of the complex, as the space to the east of room 35, formed by this wall, was too small and isolated to serve any specific function.

Fig. 4. Section A-A through the south part of the Panagia Houses, looking north (Mylonas-Shear 1987; courtesy of the Penn Museum).

Fig. 5. Section D-D through the north part of the Panagia Houses, looking north (Mylonas-Shear 1987; courtesy of the Penn Museum).

There are two arguments that this arrangement was formed already during the first phase of the complex. There was an earlier wall below the south wall of room 36, which was abutted by the central drain built definitely already during the first phase (Mylonas-Shear 1987, 59–60). There was also an earlier wall below the east wall of space 34, which bonded with the south wall of room 35, proving that the entire arrangement of area 34 and room 35 also belongs to phase 1.¹⁷ At some point, most probably after the earthquake in LH IIIB2 Early, this area was reconstructed and reinforced, similarly to rooms 28–9 (see above). The widening of the north wall of 35 and the construction of walls dividing rooms 36 from 37 and 37 from 38 suggest that this part of the building needed additional structural support (Mylonas-Shear 1987, 60). At this time, after the abandonment of House I, the roof above rooms 35–8 must have been connected to the roof over room 9 and the rest of House II.

The difference in the number of the floors strongly indicates a difference in the occupational history of the rooms of House III. Only in rooms 29–31 were there three floor levels, indicating a longer use of the area in comparison to other rooms of the building. Moreover, despite the fact that most of the pottery from the Panagia Houses can be given only a general LH IIIB date, at least one sherd found below the latest floors of those rooms can be dated to LH IIIB2, thus indicating that the floors were laid long after the original construction of the building (see above; Iakovidis 1986, 250–1; Mylonas-Shear 1987, 69). No preserved floor was reported from room 28, but the threshold of the east doorway is covered in the same type of earth-packed layer that forms the latest floor of 29, which suggests that the two had a shared occupational history. Room 32 had only one floor level, and the bench in this room resembled the west wall of room 29 (see above), which indicates that the room was occupied only during the second phase. Finally, room 33 apparently had two floor levels. The first one was made of hard-packed earth, while the second one was laid in clay. This might suggest that the room was originally located close to the outside, and only later became a more internal part of the building when room 32 was built (Mylonas-Shear 1987, 56–8). This is also indicated by the fact that it was cut off from room 29 during the second phase. Thus room 33 was probably part of the original construction, which remained in use throughout the first two phases, but it was probably not reoccupied, as it missed the third floor level found in rooms 29–31.

Phase 1

Following the above interpretation of House III, the reinterpreted history of the Panagia Houses can be discussed. The complex started to develop at the beginning of LH IIIB1, when a new building – House I – was constructed in the area (Fig. 6). The date is indicated by the construction deposit of scattered sherds, many of which were decorated with vertical whorl shells (Mylonas-Shear 1987, pl. 45B–D). The building was a simple single-storey house and the main residential unit of the first phase. It is unclear if some of the LH IIIA remains in the area (one of the drains in room 30, walls below rooms 35, 8, 15 and 16 [see Fig. 2]) belong to structures which functioned together with this unit, and were simply dismantled to make room for the new establishment, or if they form an entirely separate earlier phase of the local occupational sequence, which would then last probably until the end of LH IIIA. The former interpretation is supported by the fact that the distribution of those remains resembles the plan of the later extension of the complex. It is thus possible that there were additional storage and working areas organised around the main unit from the very beginning of the complex's existence.

Fig. 6. Plan of the Panagia Houses, with walls of the first phase indicated (based on Mylonas-Shear 1987, modified by the author; courtesy of the Penn Museum).

House I must have been a residential building. The main room was room 5, where a collection of domestic pottery was found. The presence of the hearth and the cooking pottery indicates it was used for food preparation. The function of the east part of the building remains unknown, as not much of it was preserved, but in my opinion some production might have taken place in room 2, where a drain was located. I would suggest that room 7, a basement cell located behind room 5, was a storeroom used for the storage of perishable goods, probably in non-ceramic containers, although it is possible that it was used for other goods and was simply emptied before the destruction.¹⁸ The arrangement of the space above room 7 remains uncertain. It must have been accessed through the stairway at the north end of corridor 6. The location at the back of the house supports interpreting this as a private space, possibly a sleeping chamber, as suggested by the excavator.¹⁹ However, it is also possible that the function of the space above room 7 was different, or that it was only a roof, serving as a passage to other roofs of the complex.

Sometime later, but still during LH IIIB1, the building was developed into a larger complex (Fig. 6). The south part of the extension comprised a set of five rooms built against a terrace wall to the west (Fig. 6:28–31,10), connected to the original unit by a small staircase (Fig. 6:34) and low basement rooms (Fig. 6:28–35), which covered the central drain added to the complex and linked the roofs of the extension to the original house. The staircase led to the vestibule (Fig. 6:33), and to another small external courtyard (Fig. 6:27,32). To the north, there was another set of rooms constructed around an open internal courtyard (Fig. 6:9) with a hearth. The access to the drain and the hearth, and the stone pavement probably already present in the south-east corner, all suggest that this courtyard might have served as a communal working area. Around it, rooms for storage (Fig. 6:11,12) and habitation (Fig. 6:13) were constructed. Another large vestibule room (Fig. 6:8), with doorways to the west and north, limited the courtyard from the east. A narrow corridor went around the north-west part of this room, and behind its corner it met with the entrance ramp, leading to the main doorway in the north wall of room 8. Along this ramp another part of the complex was arranged. A large room (Fig. 6:15–17) was set against another terrace wall to the north-west and probably served for habitation or work, while smaller rooms set around it were designed for storage (Fig. 6:18,19,21,23). All of these rooms must have communicated through the roofs and trapdoors (Mylonas-Shear 1987, 146), with the main communication route going along the terrace wall to the west (Fig. 6:28–31,10), through the roof link over the drain (Fig. 6:35–38), on to the roof of the main house (Fig. 6:7).

During phase 1 the basic residential unit of a single family – House I – was expanded with additional storage, working and probably also habitation areas. The reason behind this expansion might be an extension of the family, elevation of its status, or a need to expand economic activities. Nevertheless, the status of the inhabitants during this phase remains uncertain. The finds associated with phase 1 were not numerous, and they consisted mostly of the rather modest domestic deposit of House I, although there were also a few metal and stone objects found below the floors of rooms 15, 16 and 21 which probably belonged to phase 1. The complex was built using simple architectural means. The floors were made of clay, and only in House I were the walls decorated with painted plaster. However, the amount of storage space available to the inhabitants is significant, and the entire complex stands out in size in comparison to other similar buildings of that period known from the Lower Town of Mycenae, occupying 477 m².²⁰ Moreover, despite the fact that the rooms of the expansions were squeezed between older buildings and their orientation did not follow perfectly that of House I, the formality of the layout was preserved to some degree by the appearance of the main east façade of the complex, the only one which would be visible from the road leading to the Citadel. This façade, composed of the east wall of House I, the east wall of room 8, and the east wall of the ramp leading to room 8 from the north, formed a clear architectural boundary of the complex, with the movement directed around it and not through it. The main access points of the complex were thus positioned at the two ends of the east façade, namely on courtyard 1 in the south, and on the ramp leading to room 8.

Phase 1 ends with an earthquake, which completely destroyed the main house, leaving a female victim crushed in a doorway of the main room (Fig. 6:5) by the collapsing structure. Some of the walls in the rest of the complex also collapsed (Fig. 6:36,11–12), or even moved out of alignment with their foundations (Fig. 6:11). The only certain destruction deposit is located in the main room of the original house (Fig. 6:5). It is of general LH IIIB date. However, there is another group of vessels which should be associated with phase 1, namely the rich deposit of full vessels excavated at the west end of the central drain. Although the exact origins of the group remain unknown, it might have been discarded shortly after the earthquake, during the cleaning of the rubble. If 15 decorated kylikes (FS 258) from this deposit are added to the pottery associated with the end of phase 1, it should not be dated later than LH IIIB2 Early (Vitale 2006, 197). This further strengthens the synchronisation of the end of the first phase of the Panagia Houses with the LH IIIB2 Early earthquake, which affected multiple other buildings across the site (French and Stockhammer 2009, table 3).

Phase 2

House I was left in ruins after the catastrophe. The fact that one of its victims was left under the debris is of interest, but the interpretation of this fact is uncertain. Although it could suggest that the new inhabitants of the complex were somehow disassociated from the family who lived in House I, there are also other possible explanations of the situation, of which the most obvious is the inability to retrieve the body from underneath the collapsed building.

The second phase of the complex began when it was rebuilt on a new plan (Fig. 7), after the original housing unit (Fig. 7:3–7) was abandoned, together with some of the other rooms (Fig. 7:10–12,18,19,23). The south part of the complex was re-established, with new floors laid in the rooms along the west terrace (Fig. 7:28–31), and an entire new room built (Fig. 7:32), which limited the space of the courtyard there (Fig. 7:27). The central drain was partially buried under the collapsed wall and the use of its west part was abandoned. There are no finds to suggest any specific function for those rooms, but they formed an easily accessible row of ground floor cells which might have been used for processing some materials. Working areas in this part of the complex, previously probably located in the rooms with access to the drain (Fig. 7:30,31), could have been moved to the southernmost rooms, equipped with benches (Fig. 7:28,32). A buttress in room 28 and an additional thick wall in room 29 were constructed to strengthen the south-west corner of the building, possibly damaged by the earthquake (Mylonas-Shear 1987, 53–4). A new wall was also built in the central area (Fig. 7:34,36), above the abandoned drain, probably to remake the staircase and form a roof support. The latter was important for a communication with the north part of the complex. It remains unclear if the basement rooms (Fig. 7:35–8) in this area were actually used for storage at the time, due to their small size and irregular form. It seems the most probable that they were used only as structural support for the roof (see above).

Fig. 7. Plan of the Panagia Houses, with walls of the second phase indicated (after Mylonas-Shear 1987, modified by the author; courtesy of the Penn Museum).

The new communication route led to the roof of the new main unit, transformed from an open courtyard into a proper room (Fig. 7:9). It was still equipped with a hearth, and a stone pavement, and had an access to the mouth of the central drain, which was probably its only part that remained in use. Room 9 was the largest room and the only one where the hearth probably served as a central communal space, associated also with food preparation. In this regard it probably took over the function previously served by room 5. This is indicated not only by the central location of the room and presence of the hearth, but also by the similarities between the floor deposits of the rooms, which both included sets composed of a deep bowl (FS 284), a krater (FS 281), a small decorated stirrup jar (FS 171), and tripod cooking pots (FS 320) (Mylonas-Shear 1987, 68).

In the north, the space was divided with a set of new, small storage rooms (Fig. 7:15–17,20,21), only one of which had doorways (Fig. 7:17), while others had to be accessed through the roofs. Many of the pottery vessels found inside the storerooms must have been originally kept on those roofs, which were probably used not only for communication but also for various daily activities.

A large vestibule room of this part of the complex (Fig. 7:8) was probably still in use in this phase, despite the fact that there is no evidence for a new floor. However, the good quality clay floor of the first phase could have simply remained in use. Also, from an architectural perspective, its wall must have still been an important part of the structure. The same regards the smaller room to the north of the main room (Fig. 7:13), the walls of which must have been in use to support the roof. The corridor (Fig. 7:14) and an entrance ramp east of rooms 15–17 remained in use, also to give access to the doorway of the northernmost storeroom (Fig. 7:17).

The complex was probably still a dwelling for a single family, albeit not necessarily the same one which occupied it during phase 1. The finds associated with phase 2 are relatively numerous, and apart from pottery, terracotta figurines and spindle whorls they included bronze tools, pieces of lead, an obsidian blade, fragments of glass jewellery, and even some gold, as testified by the piece of gold wire found in room 15 (Mylonas-Shear 1987, 41). Thus, it is clear that the inhabitants had access to some valuable resources, as well as the ability and opportunity to store a larger amount of goods in multiple storage rooms. The fact that the inhabitants possessed metalwork is particularly suggestive, and testifies that they certainly did not belong to the lowest strata of Mycenaean society (Aulsebrook 2020, 256). The inhabitants also participated in the palatial system of distribution of goods, or some other system of organised exchange, as indicated by three sealings found in room 15 and possibly also a broken seal-stone from room 8. However, it remains uncertain whether the storage or production capabilities of the complex exceeded the needs of a single household (contra Mylonas-Shear 1987, 62–3; Efkleidou 2021, 29). It is possible, given the amount and arrangement of space, but the number of finds which would support such a hypothesis is small. There were few storage vessels or tools found in the buildings, and no unfinished objects or significant amounts of raw materials. Although it is certain that the inhabitants processed various materials, this was probably only small-scale household production and repair.²¹

It seems that the status of the inhabitants might have changed between the first and second phases of occupation, which might have been connected with a possible change of the family occupying the complex. The small number of finds associated with the first phase, which were located almost exclusively in House I, prevents any analysis of changes of the content of the rooms. However, the change of the architectural form may suggest some difference of status. After the rebuilding the complex was much smaller than in the original arrangement (270 vs 477 m²), which might be associated with a lower number of inhabitants. However, the formality of the layout of the complex had also clearly decreased, with the ruins of House I standing in front of the east façade, which was visible from the main road leading to the Citadel. Thus, the main architectural boundary of the complex was not as clear as during phase 1, although the general north–south orientation of the structure was preserved. The ramp leading to room 8 from the north must have still served as the main access point, but movement around the south part of the complex is difficult to reconstruct. It is not excluded that access from the east was also possible, going through or around the ruins of House I, into courtyard 27. Abandonment of the inner courtyard 9, which was turned into the main room, also must have affected the living comfort of the residents, cutting off the light and fresh air from the northern part of the complex. Moreover, after the collapse of House I, no part of the complex had even simply decorated walls. In conclusion, the relatively high status of the inhabitants suggested by the finds from the second phase described above may have been even higher during the first phase, when the complex was much larger and had a more formal layout.

The Panagia Houses did not stand in isolation. A few walls of LH IIIB date were located to the south and north of the complex, and together with the existence of a solid terrace wall to the west, and another one to the north, they suggest that it was a part of a larger settlement area, probably extending along the entire east slope of the Panagia Ridge (Mylonas-Shear 1987, 66). A relation of the houses to the nearby Treasury of Atreus is intriguing, but difficult to interpret. The construction of the tomb is conventionally dated to LH IIIA2 (French 1964), but it must have been used at least once during LH IIIB, as indicated by fragments of a stemmed bowl (FS 305) found underneath the threshold (Wace et al. 1923, 348, fig. 76a). Elizabeth French (2002, 69) proposed that this material was deposited during remodelling of the tholos, and that it dates to the second period of its use.²² In such a case, it is possible that the reopening of the tholos affected the spatial planning of this part of the settlement, and it might have prevented the full rebuilding of the surrounding houses after the earthquake. This situation could have also affected the potential loss of status by the Panagia Houses discussed above. However, the fact that the complex was still partially rebuilt and reoccupied remains puzzling, and indicates that many factors must have influenced the redevelopment of the Lower Town during this period (Dudlik and Zeman 2020, 27–9; Zeman and Dudlik 2023, 143–4).

The second phase ended with a violent fire, which left a burned destruction deposit across the north part of the complex (Fig. 7:9,15–17,20). The date of the event is uncertain, as the pottery from the destruction deposit is not particularly informative and its chronological assessment is not obvious.²³ The problem boils down to the question of whether the Panagia Houses complex was burned down at the very end of LH IIIB, simultaneously with the widespread destruction horizon which at the time affected Mycenae (French and Stockhammer 2009, table 3), or slightly earlier. Although the complete disappearance of decorated kylikes (FS 258) confirms that phase 2 should be dated to an advanced stage of LH IIIB2, the scarcity of deep bowls (FS 284) prevents any more refined dating (Vitale 2006, 178, 197–8). Only seven such vessels were found inside the complex, and not a single one of those was a Group B deep bowl that is normally a characteristic feature of LH IIIB2 (Mylonas-Shear 1987, 84–7). Unfortunately, the pottery associated with phase 3 also does not provide an indisputable answer to the problem (see below).

Phase 3

The third occupational phase manifests itself with new floors set in the south-west area, along the terrace wall (Fig. 8:29–31, probably 28 too). The four rooms there are probably the only ones of the previous phase still in use, while the rest of the complex was abandoned. However, new rooms were constructed to the north-west (Fig. 8:22–5), with the use of repaired walls destroyed in the earthquake (north wall of rooms 11–12) and later in the fire (west wall of rooms 20–1). It is uncertain if the two areas still in use during this phase continued to function together as one complex. Communication between them could be then established through the doorway in room 31.

Fig. 8. Plan of the Panagia Houses, with walls of the third phase indicated (after Mylonas-Shear 1987, modified by the author; courtesy of the Penn Museum).

The two semiglobular cups (FS 215) with linear decorations and monochrome interiors which were part of the floor deposit in room 23 are common during LH IIIC Early, but they can sometimes appear already in LH IIIB2 Late (Mylonas-Shear 1987, 77; Vitale 2006, 197–8). Therefore, in the absence of a certain dating of the end of phase 2 (see above), phase 3 could have started in the final years of the palace period, or just after its end. The latter seems more probable, but there is no definite answer. No reason for the final abandonment of the reoccupied rooms is known, but it must have occurred before the beginning of LH IIIC Middle, which is indicated by the lack of any Close Style pottery fragments in the fill which covered the remains of the complex.²⁴

CONCLUSIONS

Following his early work on the site, Christos Tsountas suggested that Mycenae was composed of isolated hamlets, spread on the hills around the site and inhabited by family groups or clans, which built their chamber tombs around their houses (Tsountas and Manatt 1897, 32–3). This theory was still very much valid to Mylonas-Shear when she published Panagia Houses (Mylonas-Shear 1987, 4), and it affected her image of Mycenae and its spatial and social organisation. However, Tsountas’ theory and the search for groupings of houses and their tombs have long been abandoned, mostly due to the results of the intensive archaeological survey conducted at the site in 1990s (French et al. 2003, 22). An abundance of Mycenaean remains was identified during the survey, which testified to the existence of a single extensive settlement surrounding the Citadel. Nevertheless, this settlement remains only fragmentarily explored, and analysis of its organisation is usually limited to discussions of its few known individual buildings. A comparative approach to urbanism and focus on the coordination of structures and movement patterns enables not only the proper identification of multi-building complexes such as the Panagia Houses, but also the identification of mutually dependent settlement zones that make up the urbanised palatial town (Dudlik and Zeman 2020; Zeman and Dudlik 2023). One of such zones was the Lower Town, which spread over the Pezoulia slope and Panagia ridge to the west, and along the Kokoretsa ravine to the north, and developed mostly organically along the main ridges of the site and roads leading to the Citadel (Zeman and Dudlik 2023, 141; Fig. 1).

It seems that throughout the palatial period, the main type of buildings constructed in the Lower Town of Mycenae was single-family houses. They had a variety of forms, from large monumental establishments to rather small simple units, but were all characterised by plans with a central room or courtyard, and multiple smaller rooms, which always included an abundance of storage space. The architectural forms were adjusted to the terrain. The Panagia Houses are a perfect example of such an establishment. During the first phase it was a relatively large complex, centred around the inner courtyard which was surrounded with multiple storerooms (Fig. 6). The household was adjusted to the slope, being constructed on three levels (Figs 4 and 5), which were connected through rooftops and stairways. Moreover, the form of the expanded complex was constrained by the surrounding terraces, which forced the lack of symmetry of the plan. Nevertheless, the Panagia Houses were designed as a functional and well-integrated household. It was one of a variety of buildings which must have served as living quarters for a single family, belonging to the Mycenaean ‘middle-class’, which inhabited at least parts of the Lower Town of Mycenae. Those were the buildings of a solid, but not monumental or elaborately decorated form, the inhabitants of which possessed at least some valuables, including metals, and stored relatively many supplies. They were the members of the community who were not the elite, but also not the low-status workers, servants and slaves, who occupied the lowest step of the social pyramid (Jung 2016, 564). This must have been a diverse group of craftsmen, administrators, and palatial labourers, to various extents involved in the palatial economy of selective mobilisation of resources and labour (Nakassis, Galaty and Parkinson 2010, 244). The activities of some of the individuals of this class, including receiving raw materials, transferring finished goods or obtaining land allocation, were recorded in the extensive palatial Linear B archives at Pylos and Knossos (Nakassis 2013; 2015). Finds of Linear B tablets at Mycenae are scarce, and they are mostly concentrated in the West House Group (Fig. 1:7), but nevertheless their presence confirms the activity of palatial bureaucracy at the site (French 2002, 123–8). However, individuals involved in the Mycenaean palatial system still belonged to various economic and social networks and did not necessarily share a strong common identity or a sense of common interest. The level of involvement of the general population in the Mycenaean collective identity is unknown, as the composition of the population of the palatial polities must have been heterogeneous. Nevertheless, a variety of social activities organised by the ruling elites, including religious processions and palace sponsored feasts, served not only to meet the needs of Mycenaean religious ideology but also reinforced a shared sense of identity (Wright 2004; Maran 2006; Whittaker 2015; Maran and Wright 2020, 115–16).

Despite the abandonment of Tsountas’ theory of clustering individual residential areas with their chamber tombs, it seems that the latter were the dominant form of burial among the inhabitants of Mycenae during the palace period, used probably also by the inhabitants of the Panagia Houses. The town was surrounded by more than 200 chamber tombs, grouped into 27 cemeteries spread over the area of 300 ha around the settlement (French et al. 2003, 35–9). Their setting depended mostly on geological conditions, with mild slopes of soft rock being necessary for their construction. However, the distribution of cemeteries might also relate to other factors, such as distance from the roads and pathways which provided an optimal route for funerary procession, religious beliefs, social and political dependencies, and possible individual landholdings (Cavanagh and Mee 1990; Mee and Cavanagh 1990; French and Shelton 2005, 181; Efkleidou 2019). Chamber tombs typically served as multi-generational family tombs, and their widespread use demonstrates the promotion of the nuclear family as the basic political unit of Mycenaean society during the palace period (Wright 2008). Moreover, the tombs were becoming increasingly simple and modest throughout the palatial period, thus testifying to the spread of this burial form across various social groups of the local community (French et al. 2003, 38).

The residences on the Panagia ridge were associated by Tsountas and later by Mylonas-Shear with the so-called Third Kilometre Cemetery, located down the hill, c. 100–200 m south from the Treasury of Atreus (Mylonas-Shear 1987, 4). However, this association is doubtful because of the exceptional proximity of the tombs to this particular settlement area, and the fact that the cemetery was established already in LH II, while the earliest occupation of the Panagia ridge dates to LH IIIA, when the structure excavated below the Panagia Houses was constructed.

The Panagia Houses were often contrasted against the nearby West House Group (see for example Mylonas-Shear 1987, 150–4; Tournavitou 1995, 292–6; French 2002, 68; Burns 2007), a large (c. 1600 m²), multi-functional elite complex, with an extraordinary collection of pottery, tools and raw materials, including an abundance of ivory and stone (Tournavitou 1995). Although the two complexes certainly differed significantly in status and function, they should also not be discussed as two extremes.²⁵ They were both following the same general idea of development, with additional working, storage and habitation areas gradually expanded around the original main residential unit, but on two different scales. However, the architectural concepts behind the expansions of two complexes implemented this idea in two opposite ways, which must have been a result of marked differences in both the status and the function of both complexes. The buildings expanding the West House Group were constructed against the constraints of the terrain, with a monumental and highly formal layout of massive terraces of similar plan and size, set on exactly the same orientation and with a strong central axis implemented. They were built on a sloping ground, transforming it into a flat plateau, then overbuilt by superstructures raised over the terraces (Burns 2007, 115–18). The Panagia Houses, on the other hand, were built according to the constraints of the terrain and the nearby older structures. The terracing of the area was thus adjusted to the sloping ground, which resulted in the irregular three-levelled layout of the complex. The formality of the plan was not a priority, and thus the orientation and spatial planning of the rooms was adjusted to the surroundings. However, a noteworthy similarity between the plans of both complexes is the presence of a formal east façade, which in both instances was the main architectural boundary of the complex directed towards the road leading to the Citadel.

It is doubtful that any storage or production which occurred at the Panagia Houses exceeded the needs of the household, contrary to the West House Group, which formed the most important known economic establishment of the Lower Town, and one of the most significant complexes of the entire palatial town, smaller only than the palace. Nevertheless, the inhabitants of the Panagia Houses definitely had some economic capabilities, and at Mycenae there is a good example of a simple household which was inhabited by people of probably noticeably lower status – the House of the Tripods Tomb (Onasoglou 1995). It was composed of a series of small ground-level rooms built around a corridor (c. 230 m²), with only a modest domestic deposit and no metal finds associated with the occupation of the building. The fact that buildings of both higher and lower status are known from Mycenae further testifies to the social position of the inhabitants of the Panagia Houses.

Most of the Lower Town of Mycenae must have been left in ruins after the earthquake, as there are very few examples of occupation in the later part of LH IIIB2. Apart from the Panagia Houses only the ‘Workshop’ (Danielidou 2008, 345–6) and the House of the Tripods Tomb (Onasoglou 1995, 149) were definitely inhabited during this period, while all the other known structures were abandoned. The Lower Town clearly lost its economic importance and political status, and at the same time the Citadel was significantly expanded thanks to the ambitious architectural programme that characterised the last part of the palace period (Maran 2006, 80–4; 2009; 2022, 238–9; Wardle 2015, 591–4; Maran and Papadimitriou 2020, 698; Zeman and Dudlik 2023, 143–4). The Panagia Houses were rebuilt on a simpler and smaller plan, with remains of House I left in ruins in front of the new complex (Fig. 7). It must have been still inhabited by a single middle-class family, although not necessarily the same one, and possibly of a status slightly lower than during the first phase. The living, working, and storage spaces decreased, but finds of the second phase testify to the status of the inhabitants. They lived in a changed landscape, surrounded by ruined walls, and under the shadow of a massive tholos tomb – the Treasury of Atreus – reopened and rebuilt nearby.

This final phase of the history of the palatial town ends with a severe conflagration, when the entire palatial complex was burned together with most of the Citadel (Mylonas 1966c, 224–9). It is uncertain whether signs of this destruction horizon can also be found in the Lower Town, although it may include the end of the phase 2 destruction deposit at the Panagia Houses. Although those destructions resulted in the fall of the palatial administration at the site, the Lower Town was not initially abandoned, with the Panagia Houses, the ‘Workshop’, and the House of the Tripods Tomb all inhabited until the end of LH IIIC Early (Mylonas-Shear 1987, 156–7; Onasoglou 1995, 149; Danielidou 2008, 303–4). During this period the Panagia Houses complex served as a modest dwelling of the survivors of the fall of the palace (Fig. 8). Its final abandonment marks an important change in the occupational pattern of the Lower Town, which occurred in LH IIIC Middle, when the East House was constructed (Tournavitou 2015; Fig. 1:12). The ridges, which were the focal areas of the Mycenaean Lower Town, were ultimately abandoned, and the new settlement started to grow on the terraces below the Citadel, on the banks of Chavos. This change turned out to be long-standing, as evidenced by the later remains of the historical period (French 2002, 141–50).