Documenta Praehistorica XXXVIII (2011)
O-pi e-de-i: on round building as an archetypical form of sacred space in the Aegean
Evyenia Yiannouli
Department of History, Archaeology and Cultural Resource Management, University of the Peloponnese, Kalamata, GR eyiannou@uop.gr
ABSTRACT - The archaeological record supports the general proposition that the simplest form of architecture, the round building type, variously persisted throughout the Aegean prehistory. This paper makes a case for the identification of round building as a distinct genre of sacred space in the Aegean Bronze Age, Mycenaean Crete in particular, combining evidence from architecture, iconography and the scripts. Certain archetypical features are thus delineated. It is argued that the historical understanding of the emerging LBA state requires a study of the architecture prior to the 2nd millennium BC as much as other relevant evidence issuing from the latest Aegean strata.
IZVLEČEK - Arheološki podatki podpirajo splošno predpostavko, da je najpreprostejša oblika arhitekture, stavbe okroglega tipa, vztrajala skozi vso egejsko prazgodovino. V članku predstavljam prepoznavanje okroglih stavb kot poseben žanr svetega prostora v egejski bronasti dobi, zlasti na mikenski Kreti, kjer se združujejo dokazi iz arhitekture, ikonografije in zapisov. Tako so opisane nekatere arhetipične značilnosti. V članku tudi utemeljujem, da je za zgodovinsko razumevanje nastajajoče poznobronastodobne države potrebna študija arhitekture pred 2. tisočletjem pr.n.št. kot tudi drugih pomembnih dokazov, ki izhajajo iz zadnjih egejskih plasti.
KEY WORDS - round building; Aegean Prehistory; iconography; sacred space; archetypical features
The case in point
Despite the ample, yet at times ambivalent, type of evidence regarding the identification of sanctuaries in the Aegean Bronze Age, the round building has never been discussed as a particular typological genre with concrete archaeological manifestations. The generally perishable nature of the surviving evidence, along with the conservative and unimpressive remnants of an apparently rare type, has certainly contributed to its being a rather neglected body of data. Yet, the archaeology of Aegean religion itself may conceivably be a cause of underestimation, given that it is often through types of object rather than architecture, that a space is understood as religious, cultic or ceremonial, particularly during the 3rd millennium BC. In other words, the earliest instances of the sacred are discerned within deposits that are primarily funerary (as in the Cy-cladic cemeteries or in the Messara tholoi of Crete), or less often domestic (as in EM Myrtos), or depo-
sits within a natural setting (cf. the Cretan Peak Sanctuaries), rather than by means of typical edifices or unambiguous architectural features. In the 2nd millennium BC, the relevant architectural evidence is rich, while varying greatly, both locally and chronologically, owing its standard components, hence its archaeological identification, to the formalising impetus of Creto-Mycenaean palatial cultures.
In this paper, a case will be made for the identification of round building as a distinct architectural category of sacred space in the Aegean Bronze Age, with particular reference to its cultural semantics in the Late Bronze Age.
We intend to expand on our recent findings regarding round building as an important architectural form for the prehistoric cultures of the Aegean in general. Our purpose in this paper is to establish a
framework for discussion, whereby this distinct genre of building may, on occasion, constitute a distinct realm of the sacred.
We aim specifically to show that the domestic component of round building is paramount in defining its religious character, while being inseparable from it, or closely associated with it, in the course of time; that Palatial Crete has been a catalyst in the process of defining the round building as a standardised form of religious architecture; and that Mycenaean Crete preserves the sort of evidence which supports the notion that round building may be understood as a form of sacred space with archetypical charac-teristics.1
Aegean round building and its significance: an outline of the research
Given that the round form is a simple and elementary sort of space, one needs to discern culture-specific from fortuitous attributes in the relevant architectural evidence. This is important because the evidence in question is widespread in the East and West, including the New World, sharing a good number of common properties or qualitative features, as well as significant differences: investment in perishable structures, particularly during the horizon of sedentism, when the type first occurs, are examples of the former, whereas differences in the absolute dates of foundation or the sort of stratigraphic sequences in which these buildings are embedded pertain to the latter (Yiannouli 2009. 90-91). This general character is attested not only among Mediterranean cultures, such as the Aegean and Cyprus, but also within theoretically compact spheres, as in the Aegean itself. The survey we conducted towards that end reveals that the features pertinent to the Aegean cultural sequence develop as follows.
Round building in the Aegean, but also Cyprus, appears to form part of an emerging tradition related to incipient habitation that exhibits a particular tripartite manifestation. This tripartite scheme results from combined architectural and stratigraphic correlations evidenced in the stereo of sites, which are archaeologically traceable as: (i) pits/bothroi, (ii) pit-houses with or without pits/bothroi, (iii) non-submerged round or elliptical buildings, often resting on a stone socle, with or without pits/bothroi (Yiannouli 2006a.Figs 1-6).
It is very seldom that these three form a complete sequence at the same site, as in the case of Eutresis in Boeotia. Each one, however, may appear independently as (i) or (ii) or (iii), and rarely as (i)/(iii) or
(ii)/(iii).	It is as yet difficult to judge whether this is due to the fragmentary nature of the archaeological evidence, to excavation and research biases, or to the existence of an ideal scheme whose cultural components match the peculiar archaeological picture as it materialises locally. This incipient settlement horizon appears to emerge with the onset of Neolithisation, thence re-emerging within incipient habitation phases throughout the Neolithic and Early Bronze Ages, regardless of the absolute dates of the respective sites. Any type of correlation that may be thus identified formulates a stratigraphic succession from stage (ii) to (iii) in a few sites, as well as a certain preponderance of stage (iii) during the course of the 4th and the 3rd millennia BC (Yiannouli 2009. 91-95, Fig. 1-2).
Stage (iii) or the non-submerged type of round building has so far been attested in the Middle Neolithic, continuing throughout the LN, the FN and the EB Ages in the Aegean. Features commonly shared by this body of evidence relate to the architectural structure itself (generally perishable, often resting on a stone socle), the type of deposit contained (generally domestic), inter-site characteristics (embedded in a settlement, co-occurring with mostly rectilinear houses, set individually or in a small group, occasionally in a stratigraphic succession on the same spot), and also size, varying in relation to wall width and building diameter (the Neolithic instances are generally fewer and smaller than those of the EB Age). The most recently published finds confirm and enrich the scheme outlined. Stages (i)/ (ii)/(iii) are now recorded in stratigraphic succession at an Early Neolithic site at Merenda in Attica (Kakavoyanni et al. 2009. 143-152), although the precise chronology of stage
(iii)	needs further clarification. Similarly, two huts of the transitional FN- EBI horizon, at Merenda and Houmeza (Spata), respectively (Petrou et al. 2009. 129, Fig. 3; Kakavoyanni, Douni 2009.384) add to the distribution of the relevant evidence (Fig. 1).
The most striking changes within the course of the 3rd millennium include three major transformations (Yiannouli 2009.103-5):
• An as yet ill-understood monumental variant emerges at a few EHII sites (Eutresis B, Tiryns, Voido-
1 Some of the points raised here were briefly presented in The European Association of Archaeologists, 16th Annual Meeting, 1st-5th September 2010, The Hague, Netherlands; Session 'Temples and Sanctuaries from PPN to Copper Age: Type, Content and Functionality'.
Fig. 1. Distribution of round huts built on ground level from Neolithic (underlined) and Early Bronze Age settlements. Early Bronze monumental buildings of round shape (black circles) and clay hut models from tombs (Tiryns, Lebena) are also indicated (black squares).
koilia), followed by EMIII-MMIA Crete (pre-pala-tial Phaistos, cf. also the 'bee-hive' Hypogaeum in Knossos).
•	Models of round buildings now appear - as grave good dedications in mainland Greece (Tiryns) and Crete (the EMI-IIA model from the Lebena tholos in the Messara).
•	The paucity of Neolithic and Early Bronze material from Crete, compared to sites to the north, is considerable and striking, along with the fact that the surviving EM evidence does contain both types of 3rd millennium transformations.
Palatial Crete as a cultural catalyst
The paucity of pre-palatial settlement data, although peculiar to Crete and not typical of the Aegean, seems to develop in ways that are exclusively Mi-noan during the 2nd millennium BC (Yiannouli 2009.103-105). The relevant evidence is capitulated by two major data sets, c. 18 MMIII- LMI talismanic seals (Onassoglou 1985) and c. 23 (18 prehistoric) LMIIIA2-LG/Orientalising, c. 1350-late 8th century BC, clay hut models (Hägg 1990; Mavriyannaki 1972; Mercereau 1993). Respective scholars contend that these bodies of data are homogenous, no two
examples from either set narrowly replicating each other. Instead, they share architectural analogies and structural affinities, despite the time interval between the two groups and the unstratified context of a good number of them. Their conjoined treatment is then legitimate on account of their common form (cf. the discussion of the buildings on the seals as round, Yiannouli 2006b. 41-43), of analogous structure (perishable and conservative), and aspects of temporal expanse (long time span with gaps elapsing between). That these affinities, common to both data sets, also partake in the general character of 3rd millennium concrete edifices is a perceptible as much as significant parallel, as we intend to discuss in the final section of this paper. Conversely, talismanic seal iconography and plastic representations appear to be a localised phenomenon, peculiar to Minoan customs, cogently transcribing an iconic rendering of the natural world.
The comparative analysis of the Mi-noan sets may then provide indirect, yet precise evidence for the survival of this type in Crete, albeit in iconography rather than actual architecture, which still awaits full assessment (e.g. MMII A. Photia or MMIB-MMII/IIA Myrtos-Pyrgos III).
We have already examined those architectural features that recur systematically within and between sets, reaching the conclusion that these representations convey either the reality of an existing building type or the form of a particular building (Yiannouli 2006b.38-43; 2009.99-103). The inferred relation between architectural representation and concrete reality accords with previous research on Neolithic and Early Bronze Age building models (Ma-rangou 1992) as well as between models and architecture, or between iconography and architecture in the case of Minoan Crete (discussion in Yiannouli 2009.103). For the sake of argument, we shall here summarise the main features of the resulting typology, first as a distinct genre of building, then as a particular realm of the sacred.
In all cases, a demarcation of floors in the form of euthenteria is depicted. At times, this is possibly wooden, given the analogous treatment between
floor demarcation and the grooving or juncture between walls and roof, so that wood might have been easier or imperative to depict. On some seals in particular, floors may exhibit vertical strokes in the manner typical of a hut on piles.
Doors are employed in both sets, by way of a handle signal on seals, but concrete, detachable referents in models. It is then conceivable that it is imperative to depict a wall boundary, implying the existence of an opening on the seals, contrary to the models, where it is imperative to stress building permeability instead. If the wall boundary and permeable opening are the two basic constituents of all building, then the two sets stand in inverse relation to each other. This difference may be understood as two different ways of considering the same type of structure, whereby the former stands closer to the experience of the viewer, whereas the latter stands closer to the experiential reality of built space. The importance of the door as an elemental unit per se is highlighted by the fact that it constitutes a distinct architectural motif in other seals of the talisma-nic group (Schiering 1984). However, detaching a part from its generic context, as in abstracting a door from its building, seems reasonable with the inferred convention within the seals, contrary to that of models. In other words, the treatment of doors in abstractio, be they isolated or visually signaled, stands in inverse relation to the act of permeating itself, which is instead conveyed in concretio by the door being modeled in clay.
Walls are rendered as vertical or convex or concave in both sets, plus flaring outwards in certain models only. On seals they consist of uprights, either in groups of two or consecutively in a row, or with an interwoven lattice in between. On models, a tripartite colour code, consisting of red, black and white, is transmitted through the combined use of clays, slips or pigments; and coarse gravel inclusions.
Roofs are invariably pitched with straight or slightly convex or concave sides that at times terminate in eaves on seals, but not models (contrary to the employment of eaves in both 3rd millennium models from Tiryns and Lebena), and bear evidence of thatching (cross-hatching, concentric incisions or even wheel-marks on models). The sprawling straw end of the pitch on a few seals and the flattened disc arrangement on the pitch of some models are reminiscent of an analogous arrangement of the Linear B 123-4 hut ideograms, as well as the much earlier Le-bena model, emphasising thus the conservative ar-
chitecture of this type of building. However, the obtuse roof carination on seals, usually 50° -70°, and the comparatively oblique on models, usually 25°-30°, indicates a live tradition, whereby formal conservatism materialises in culture-specific ways of conveyance.
The conjoined treatment of the evidence in question reveals a distinct architectural typology of huts (for the aforementioned documentation and analysis of material structure, cf. Yiannouli 2006b. 38-46):
•	Partly submerged huts, the latter with either vertical or outward slanting basement walls.
•	Buildings at ground level.
•	Pile-huts (also evidenced in Hieroglyphic, Linear A and Linear B scripts, Evans 1909.198, no. 43; Evans 1921.Fig. 477.S52, and p. 637, n. 2).
According to the structural material represented, these types may be subdivided in terms of superstructures made of:
•	Wood and branches (in seals).
•	Clay / pise or rubble combining into a tripartite, i.e. red-black-white, colour code (in clay models). For the sake of argument, one should further add a particular type of vessel to the aforementioned building types, i.e. a vessel type that retains the form of a hut, preserving an extravagantly sized sort of handle positioned somewhere other than that expected in conventional door representation.
Note that these patterned similarities give cohesion to the data, irrespective of the wooden or clay structure of the individual examples in each set. Attributes such as the presence or absence of eaves, or roof carination and wall inclination types are cases in point. On the other hand, the very medium and corresponding techniques of manufacture, such as incision on the stone seals and the plastic rendering of clay in models, differentially convey certain intended properties, such as the abundance of construction details in the superstructure of the former and the tripartite colour code on the surfaces of the latter.
That this type of building was also employed as a particular realm for the sacred is, however, denoted not merely by way of resisting radical change in general form, shape, and appearance, but through its overt association with motifs, themes or arrangements that are typical of Minoan religion.
On a few seals, these buildings constitute the focal centre in a tripartite arrangement, laterally flanked with wooden platforms serving as stands for plants
and horns of consecration (Onassoglou 1985.Tab. X. 9, 11; Yiannouli 2006b.45). The association of hut, plants, and bucrania are, however, already associated in the Hieroglyphic script and Linear A (the hut ideogram, ox-head and spray, Evans 1909.198), thus transferring a concrete reality from a scriptural to an iconic medium of representation.
An analogous case can be made with reference to the extravagant S-shaped handles attached to certain huts on the seals, namely that they depict not real buildings, but building models instead. The association of hut and vessel is far from foreign to the Minoan idiom, as seen in the 'ritual vessel' motif, often topped with a lid in the form of a thatched roof (Kenna 1969.Pls. 3-4) or even in the repertory of hut vessels measuring palatial produce, as in the Linear B 123-4 hut ideograms in the G-Series (Chad-wick 1973.Fig. 10).
The formalising impetus of Palatial Crete seems to survive the end of that era. For one may recall the basic structural affinity between two chronologically (but not geographically) distant cases, namely the hut and cups association of the model from a sanctuary context in Gortyn (Orientalising Period), having as precedence the circular basin filled with cups from Malia, House E (Yiannouli 2006b.43-45, 48).
But this is far from an exceptional parallelism between prehistoric and early historical Crete. The Mi-noan horns of consecration, one of the most typical symbols of palatial religion and cult, would be an example par excellence. This symbol sustains its basic formal attributes and ritual associations during the flourishing palatial era on the island of Crete, but also during its long aftermath from LMIII (Karphi, Kavousi, Vrokastro) to PG-G times (Vrokastro, Patsos, Syme Viannou, Dreros Aphrati?), as evidenced in the respective cult deposits (Yiannouli 2006c. 305). In fact, a most eloquent analogy for our purposes derives from the hut models themselves. The LMIIIC-Subminoan clay hut model from Knossos and the Protogeometric B clay hut model from Ar-chanes (see further below) both focus on the symbolism of a goddess with upraised arms at the very centre of their interior.
The significance of the Creto-Mycenaean component in the LM III relevant evidence
The Cylindrical model from Knossos (LMIIIC-Sub-minoan) comes from the Spring Chamber Shrine of the Palace (Evans 1921.128-133), retaining overt
affinities with the PGB Archanes model, allegedly from a tholos at Phythies (Hägg, Marinatos 1991; Sakellarakis 1987). The Knossos and the Archanes models, however, share what we believe to be a significant feature with some other examples, two of which come from the town of Chania, one possibly LMIIIB? (Hägg 1990.96; Mercereau 1993. Cat. Nos. 3-4) and a third from a house in the town of Kastri-Palaikastro of LMIIIB-C (Mercereau 1993. Cat. No 6; also, Cat. Nos. 18-19, for the Knossos and the Archanes models, respectively).
The horizon of LMIIIB emerges as a likely post-quem for all five in this particular group, which, despite some undated finds (Chania) or the stylistic dating of others (Archanes), seems to range broadly within LMIIIB?-PGB or roughly 1300?- 850/800 BC. However, certain examples of the evidence in question do date to LMIIIB-C and LMIIIC-Subminoan (Kastri-Palaikastro, Knossos), so that one may reasonably infer the continuous employment of the features under discussion. These do not so much relate to the famous goddess, nor its typical stance, nor even the combination of clay hut and the goddess with upraised arms, although the latter is evidenced in LMIII (discussion in Yiannouli 2009.106-107). Rather, it is the amorphous lump protruding at the centre of these models, as well as the way that the goddess herself is incorporated at the analogous spot that compare cogently with two Linear B strands of evidence from the palatial centres of Mycenaean Greece.
On the one hand, the Linear B 123-4 hut ideograms of the G-series, attested in the major palaces of Knossos, Pylos, and Mycenae, are employed in order to measure industrial produce for condiments, as in cyperus and coriander. At Mycenae, the coriander sign forms a ligature with the hut sign, whereas in Pylos it is part of the unguent production process. Consequently, the Linear B 123-4 hut ideograms of the G-series are associated with earth produce, typifying widespread ways of measuring industrial production of particular flora that falls within the interests of palatial administration.
On the other hand, the tablet An 1281 form Pylos, relates to personnel or craftsmen responsible for workshop operations for offerings to Potnia, as in po-ti-ni-ja i-qe-ja do-so]-mo o-pi e-de-i the significant phrase being o-pi e-de-i and translated as 'at her shrine', but literally meaning at her seat (Chadwick 1973.483-4, 565). However, in our previous discussion of the word e-de-i < edos <
eSoq (seat), we traced the evidence according to which £5oq constitutes the basic semantic stratum of eSaqioq (earth/ground), for eSa^oq = £5oq + a^rj (ground + touch), eSa^oq being also neutral after £5oq. Etymologically, the fundamental semantic stratum of the word £5oq 'seat' becomes referent, reason and impetus for formulating subsequent semantics, in the manner that 'seat' is a sort of prerequisite for 'abode or residence', i.e. the place of a shrine, later extending to include the foundation or base for statues in the historical period (Yiannouli 2006b. 47-48).
So, the analogy between 'seat ^ abode/residence' of a god in direct relation to the semantics of 'ground' or 'earth' stresses the fundamental role of £5oq, which also seems to apply to the clay hut models in question. The amorphous lump at the centre (Cha-nia, Karphi-Palaikastro) or the very goddess seated on it (Archanes) or emanating from the ground-seat of her place in a manner affixed and inseparable from it (Knossos) succinctly shape the imagery of the fundamental semantic strata briefly exposed.
This chthonic frame of reference is basic, constant and consistent with the multitude of available evidence: hut architecture (perishable structure of wood or earth), its association with particular kinds of flora or earth produce in Linear B, its overt and sustained associations with cemeteries (the Lebena and Tiryns models, the seals from Sphoungara and Pla-tanos, the Archanes model), no less than associations with cult (Hägg 1990), particularly the clay hut models from LMIIIA-B Amnisos being next to an inverted cup covering a triton shell, the LMIIIC-Sub-minoan Spring Chamber Shrine from Knossos, and perhaps the find from contemporary Karphi (Mer-cereau 1993.22, Tab. 1).
Archetypical features in LMIII hut model architecture and the open-ended question of origins
An archetype is a mental construct that pertains to the typical character of an original formation. The term is used here quite literally, rather than in its classic psychological sense, namely as a means to bridge the span between original formations and subsequent manifestations, while assessing the degree, frequency, kind, and modes of such a typical recurrence in time.
Discerning archetypical features in architecture is thus difficult, because these are often confounded with typical formal characteristics. Whereas arche-
typical features need to be tangible and constant, generally on a par with the formation of typical characteristics, the latter do not invariably lead to the identification of the former. Conversely, archetypical features are impossible to discern irrespective of the recurrent emergence of specific attributes that materialise the fundamental semantics of tangible form. Throughout the Aegean Bronze Age, the common elements of the round building type, i.e. a rare or atypical settlement type, its generally perishable structure and its conservative form, are also typical of concrete edifices as much as of round building representations. What is commonly shared (Yiannouli 2006a; 2006b; 2009) is then an important prerequisite for assessing what different cultures pinpoint as archetypical in this common repertoire of characteristics.
Turning the question on its head, one may just as well propose that it is precisely the very constancy of the particularities of architectural structure, namely its being conservative, generally perishable and rare, while common to both building and building representation, that capitulate and materialise the elementary constituents of emerging archetypical attributes.
This notion is hard to contest, given that these common attributes can be traced in the relevant evidence over millennia. Indeed, a round hut with a perishable structure of wood or clay, built on the ground or as partly submerged edifice, rarely encountered and, so atypical, compared to contemporary evidence of often greater or even exceptional sophistication, capitulate the standardised character of the material in either real or representational forms. In fact, this archaeologically tenable state of things is tantamount to a corollary of our discussion so far: namely, that the general typology of round building representation after the 2nd millennium evidence echoes the general typology of actual huts in the settlements of the preceding millennia, irrespective of the means and material used in their construction, or conceivable contacts among people from the sites in question. It is, indeed, futile to suppose that people from, say, 3rd millennium settlements dictated in any way the requirements of round building manifestation in the cultures of the 2nd millennium. We have already proposed that the persistent adoption of a particular architectural form is archaeologically evidenced through the local adaptation of certain intrinsic properties: "...that patterned elements may be diversified locally [indicating] that [an] inferred tradition, or its patterned
imprint is not predictable or homogenous, but alive and so assimilated according to a local idiom, hence the regional differences encountered even within culturally compact spheres, such as the Aegean" (Yiannouli 2009.107). The inferred parallel state, whereby 2 nd millennium round building representation conforms to the architectural typology of earlier settlement huts, does adhere to the basic requirements of archetypical features, while retaining the individual character of each specific case. These features, however, being common to all the evidence in question, do not sufficiently differentiate the respective cultural horizons in that sense. One should then construe what may pertain to LMIII, as currently known, in addition to the preceding exposition.
© There is an emphasis on the vertical axis of architecture in round hut representation.
We have indicated that the emphasis on the vertical axis of architectural structure, i.e. above and below ground, becomes as fundamental to incipient settlement habitation as the horizontal axis of internal and external space common to all building (Yiannouli 2006a.30, 35-36). This tendency is portrayed in our stages (i) pits/bothroi and (ii) pit-houses, which construe two of the three ways that an incipient habitation horizon is archaeologically manifest. In the Aegean, the relevant Neolithic and Early Bronze Age sites further record a few cases with the succeeding stage (iii) non-submerged round houses, while many more proceed to mode (iii) directly, as if by-passing (i) and (ii). According to our current knowledge, the vertical axis of building appears to be fundamental in primary architecture, but eventually dispensable with regard to the horizontal axis, which became common in built space. Consequently, the inferred representations of semi-submerged huts with vertical or sloping basement walls (models from Gortyn and Phaistos, respectively, cf. discussion in Yiannouli 2006b.43) capitulate a basic quality in the act of building in its primary sense. Similarly, the amorphous lump at the centre of models or the placement of the Minoan Goddess at this very spot, even rising from it in the case of Knossos, cogently signal the importance of the vertical axis in incipient habitation settlement.
It is then imperative for some LMIII clay hut models to resort to features (the vertical axis) that pertain to the earliest stages of a construed tripartite scheme regarding incipient habitation, even if the vertical axis is eventually rendered obsolete by the movement along the horizontal.
The vertical axis of architecture and the position of the goddess are then homologous in place, hence the archetypical adaptation of a typically Minoan symbol (the Minoan goddess) to a spatial relation that is fundamental in incipient settlement (vertical axis), as reaffirmed in a combination specific to LMIII (clay hut model and Minoan goddess).
© An emphasis on the original type of context, i.e. settlement context, is associated with LMIII clay hut models.
Our survey of the Aegean evidence leaves no doubt about the earliest associations of round huts, basement or not, with the earliest habitation sites and with generally domestic deposits. A major transformation occurs in the 3rd millennium, whereby the aforementioned state of things is modified by the erection of monumental round buildings in a few settlements and by the first appearance of clay models in graves as offerings on both the mainland and Crete. Obviously, the association of round building architecture with a funerary context might have had far greater consequences for Crete than the mainland (Yiannouli 2009.95-97, 103-105). However, it is from the Palatial Eras onwards that a multitude of contexts preserve the surviving evidence, as in funerary and sacral, including certain palatial sites themselves (Yiannouli 2006b.46-47). Palatial interest in specific types of produce appears to employ the type in vessel form and/or actual huts, further illuminating industrial aspects of the Linear B administration. It is then significant that the majority of the LMIII evidence comes from settlements, whereas the PGB models are attributed with a tomb or sanctuary provenance (Mercereau 1993.Tab. 1).
It may not then be inconceivable that a settlement context, being original to this type of architecture, recurs persistently, thus reaffirming an original type of relation, in the manner that the vertical axis is fundamental to the incipience of building.
© The semantics of £5oq, associating the meanings of 'earth/ground' - 'abode' - 'female divinity', is significantly portrayed in the Knossos and Archanes models and, elliptically, in the group of five models discussed earlier.
The direct relation between č5oq and eSaqioQ, e5a-90Q being formed after £5oq, transmits that the semantics of £5oq 'seat' is a linguistic precedent for the semantics of eSaqiOQ 'ground/earth', literally resulting after the association between 'seat' and
'touch' (= £5oq + a^rj). In terms of the most basic semantic stratum then, 'ground/earth' is to be found where 'seat + touch' are narrowly confounded.
The basic semantics in an etymon are culturally significant, in that the etymological meaning is parallel to a sort of 'stereo' in semantic stratigraphy, after a culture-specific, hence non-arbitrary, iconic relation between language and referent (discussion in Yiannouli 1992.CH. 2). It is then of consequence that material culture preserves an analogous configuration, whereby 'seat' is narrowly replicated locally as 'centre' and also as the seat of a female divinity in the closest possible connection, so that 'seat' and 'earth/ground' are inseparable. The amorphous appearance of the lump itself at the centre of floors retorts the imagery of inseparability or low level distinction between place ('earth/ground') and configuration ('seat'), shaping it after the unmediated image of common natural formation. So the coarse and technically poor rendering of the basic referents, 'seat' and 'earth/ ground', are conveyed so that they reaffirm in material culture the elementary semantics configured in language.
Associating basic semantic elements (seat-earth/ ground) with architectural principles typical of incipient settlement (round hut, vertical axis) may then materialise the archetypical features in these LMIII representations.
It is important to repeat at this point that it is the conjoined appearance of all the above features that define the LMIII intent, rather than each feature individually. For seated figures or an unmediated relation to the earth are also basic constituents of Mi-noan religion in general. The barefoot females in the plethora of processional scenes in Minoan frescoes and signet rings reproduce in tangible form the unmistakable, corporeal quality of such an unmedia-ted relation. The barefoot and bare-breasted female seated on a rocky outcrop on the LMI lentoid from A. Triada (Younger 1995.173, Cat. No. 129, Fig. LXIIIa) or the ivory plaque from Mycenae (Younger 1995.176, Cat. No. 141, Fig. LXIVb) succinctly convey such precedents, along with the purport of these ideas in the repertoire of the formal arts. Conversely, the impoverished, coarse and technically basic means of conveyance, as well as a possible rustic air in imagery, are better explicable as cues towards the semantic stereo in a cultural palimpsest of custom. So that the coarseness of a representation, technically, materially or in imagery, is not, in our view, a reflection of an artistic regression, let alone license,
but the intent to configure and reaffirm in form and/or execution aspects of elementary semantics. This would explain the 'natural' representation of 'seat' in Neopalatial Crete (A. Triada) or during LMIII (Mycenae plaque, the group of five hut models), despite the ability to produce and reproduce magnificent stools or thrones amidst the great variety of such items, including the scripts (Younger 1995. 168-178, 188-194).
Consequently, it is not the individual occurrence, but the convergence of the selected elements, namely, principles of incipient architecture (round building/vertical axis), the iconic rendering of the relation between £5oq and eSaqioQ (seat/earth protrusion), and the goddess herself seated at, or emanating from the 'seat'- centre of floors - that conjoin to adapt the chthonic aspect of Minoan religion (the inseparability or the unmediated corporeality between Goddess and earth/ground) to the architectural and linguistic strands in question (Fig. 2).
The round building as an archetypical form of sacred space in the Aegean may then be definitely traced in the particular LMIII palimpsest, while Palatial Crete itself may currently set a conventional post quem for the inception of round building as a form of sacred space. For it is an analogous hut of perishable wood that repeatedly combines with motifs typical of Minoan cult, as in sprays and horns or bucrania, that we see in both talismanic iconography and in the Hieroglyphic script.
However, the question of original associations between a particular building form and a sacred space may not be answered, in our view, solely on account of standardised, hence recognisable manifestations, as those of Palatial Crete. On the one hand, standardisation in religious terms presumes a long formative period. On the other, such an important convergence may long remain latent in the perishable and conservative character of the structures in question. One then needs to delve closely and systematically into the evidence prior to that of the 2nd millennium, partly because the 3rd millennium is an (as yet) insufficiently understood turning point in the course of the round building tradition on both the mainland and Crete, and partly because the surviving or bulk of the known architectural evidence comes from pre-2nd millennium sites.
However, as yet we have studied these cases in terms of common chronological, architectural, structural, typological, stratigraphic and general contextual fea-
Goddess with upraised arm;
EÖCKpOw Eöo: + a (pi]
Fig. 2. Schematic association of basic semantic elements (%8oq > äSoupoq) with principles of incipient architecture (round hut - vertical axis), as typified in LMIII (clay hut model, the Goddess with upraised arms).
tures on a regional or area level, rather than in terms of individual characteristics, let alone of differential attributes within the respective settlements. Consequently, examining original associations between a particular building form and a sacred space within specific sites defines a distinct research topic resulting from the foregoing analysis and one of the basic objectives of our ongoing study of this genre of building in the Aegean.
Thus, in addition to the primary data, one need not simply search for a round building model in a sanctuary complex or even for a round building used as sanctuary, but for those cases whereby the 'domestic' and the 'sacred' are confounded notions in an archaeological context. For the overall character of this type, be it in concrete architecture or architectural representations, is that of an atypical and infrequent building structure that originates in settlements, per-
sisting over a good number of centuries and resisting drastic change in general form, but not in size or specific formal characteristics. No particular example is replicated in any other in a given group, concrete or iconographic, modest or monumental, yet a limited set of regularities, along which each case seems to have been constructed, has already been set in evidence. This paper in particular concentrated on such a limited set of regularities, singled out as archetypical among the typical properties of LMIII clay hut models. Yet, even a cursory glance at the surviving pre-2nd millennium cases conveys the ample scope for such research. For instance, why do atypical buildings contain typically domestic deposits, as these assemblages are generically identified? Are all 'domestic deposits' of the same order on account of being identified as such, regardless of their being recovered within typical or atypical buildings in a settlement? Is it significant that some round buildings are found in stratigraphic succession at some sites?
Consequently, the singularity of each case and the regularities that provide the cohesion within and between different groups are complementary prerequisites for future discussion with regard to the pre-2nd millennium settlement evidence.
This dual level of reasoning, however, applies to round buildings and their position within sites as much as to architecture in general, particularly as a primary act of settlement. For an incipient habitation horizon does not set a mere marker of seden-tism, but the onset of an irreversible cultural shift with a distinctive architectural component whose bearings are locally peculiar, yet potentially prevalent in culture specific ways well after this incipient change of course.
REFERENCES
CHADWICK J. 1973. Documents in Mycenaean Greek. Cambridge University Press. Cambridge.
EVANS A. 1909. Scripta Minoa I. The Clarendon Press. Oxford.
1921. The palace of Minos at Knossos, vol. II. Mac-millan. London.
HÄGG R. 1990. The Cretan hut-models. Opuscula Athe-niensia XVIII6:95-107.
HÄGG R., MARINATOS N. 1991. The Giamalakis model from Archanes. Between the Minoan and the Greek worlds. In D. Musti et al. (eds), La transitione dal Mice-neo all'alto arcaismo. Atti del Concegno Internazionale, Roma 14-19 Marzo 1988. Consiglio Nazionale delle Ri-cerche, Roma: 301-308.
KAKAVOYANNI O., DOUNI K. 2009. MiKpeq epeuveq Kai avaoKaq>e<; oe npoioxopiKei; Beoen; ot a Me-oo^eia, 1994-2004. (Short investigations and excavations in prehistoric sites in Mesogeia, 1994-2004). In V.
B. Vassilopoulou and S. Katsarou-Tzeveleki (eds.), From Mesogeia to Argosaronikos. B' Ephorate of Prehistoric and Classical Antiquities, The Works 1994-2003, Proceedings of a Conference, Athens 18-20 December 2003. The Municipality of Markopoulo of Mesogaia, Athens: 383-398.
KAKAVOYANNI O., TSELEPI E. and KATSAVOU C. 2009. OiKio|oq xnq Ap%aiöxepnq Neo^iBiKTf q Kai oiKia xnq Neoxepnq NeoA.i9iKTfq Eno%rfq oxn Mepevxa. (An Ancient Neolithic settlement and a Late Neolithic house at Merenda). In V. B. Vassilopoulou and S. Katsa-rou-Tzeveleki (eds.), From Mesogeia to Argosaronikos, B' Ephorate of Prehistoric and Classical Antiquities, The Works 1994-2003, Proceedings of a Conference, Athens 18-20 December 2003. The Municipality of Markopoulo of Mesogaia, Athens: 143-157.
KENNA V. 1969. The Cretan talismanic stone in the Late Minoan Age. Studies in Mediterranean archaeology XXIV. P. Äström, Lund.
MARANGOU C. 1992. EIAOMA, Figurines et miniatures du Neolithique Recent et du Bronze Ancien en Grece. BAR International Series 576. BAR Oxford.
MAVRIYANNAKI C. 1972. Modellini fittili di costruzioni circolari della Creta minoica. Studi Micenei ed Egeo-Ana-tolici 15 (Incunabula Graeca 55): 151-170.
MERCEREAU R. 1993. Cretan cylindrical models. American Journal of Archaeology 97:1-47.
ONASSOGLOU A. 1985. Die "talismanischen" Siegel. CMS Beiheft 2. Gebr. Mann Verlag. Berlin.
PETROU A., SALAVOURA E. PITSIKOULIS A., SKLAVOS M. and ARGYROPOULOS B. 2009. MuKnvalKof xcf poi Kai KoivoxiKcf veKpoxapefa YeraiexpiKrav, ap%atKcfv Kai K^aoiKcfv xpovrav oxn Mepevxa. (Mycenaean graves and Geometric, Archaic and Classical cemeteries). In V. B. Vassilopoulou and S. Katsarou-Tzeveleki (eds.), From Mesogeia to Argosaronikos, B' Ephorate of Prehistoric and Classical Antiquities, The Works 1994-2003, Proceedings of a Conference, Athens 18-20 December 2003.
The Municipality of Markopoulo of Mesogaia, Athens: 127-142.
SAKELLARAKIS Y. 1987. Ap%ai0^0YiKTf fpeuva Yia |fa apxaioKann^fa to 1949 oxnv Kprfxn. (An archaeological investigation of plundering in Crete, 1949). In Words of Amity, Hommage to G. Mylonas. The Archaeological Society at Athens, Athens: 37-70.
SCHIERING W. 1984. Bemerkungen zu den sogenannten architektonischen Motiven der minischen Siegelkunst. In Aux Origines de l'hellenisme, la Crete et la Grece, Hommage a Henri van Effenterre. Sorbonne, Paris: 4-71.
YIANNOULI E. 1992. Reason in Architecture: The Component of Space. A Study of Domestic and Palatial Buildings in Bronze Age Greece. Unpublished doctoral dissertation. Cambridge University. Cambridge.
2006a. XKeyeiq Yia xiq oiKieq-opuY|j.axa uno xo praq xrav 5e5o|j£va>v xnq ©eooa^fa. (On pit houses in the light of Thessalian remains). 1st Archaeological Works of Thessaly and Sterea Hellas, Volos 27.223.3.2003. Ministry of Culture, Athens: 29-43.
2006b. Perishable, yet perennial: formal and semantic properties of round building architecture in prehistoric Aegean. Mediterranean Archaeology and Archaeometry 6(2): 37-51.
2006c. Kfpaxa KaBooicf oeraq. EK5o%eq evoq Mi-vraiKou ou|ßö^ou oxiq MuKTfveq. (Horns of consecration: Aspects of a Minoan symbol in Mycenae). Proceedings of the 9th Cretological Conference, Elounda 1-7 October 2001. Society of Cretan Historical Studies, Herakleion: 297-311.
2009. The emergence and the development of a round building tradition in the Aegean and Crete. Mediterranean Archaeology and Archaeometry 9(1): 89-113.
YOUNGER J. 1995. The iconography of rulership: A conspectus. In P. Rehak (ed.), The role of the ruler in Prehistoric Aegean. Aegeum 11. Universite de Liege, Liege: 151-211.