ΚΑΤΕΣΚΑΨΑΝ ΙΕΡΑΠΥΤΝΙΟΙ: THE DESTRUCTION OF POLITICAL COMMUNITIES IN THE SECOND CENTURY BC AND THE RESILIENCE OF THE CRETAN POLIS

Abstract

Recently there has been a revival of interest in both the historical and archaeological dimensions of the destruction of cities. The destruction of the principal settlement of a polis is one thing, the effective eradication of the political community quite another. What did it take to destroy a political community? The historical record is full of references to destructions of one polis by another in late Classical to Hellenistic times in Crete. And though not all of these destructions led to the end of the political community in question, some did, and between the Classical period (where we know of 49 poleis) and the Roman conquest of Metellus (where we know of only 24) the numbers of Cretan poleis were drastically reduced. The destruction of Praisos by Hierapytna between 145 and 140 BC (Strabo 10.4.12) was one such case. This seems to form part of a horizon of destructions (of Dreros, Apollonia and Phaistos) that took place between 200 and 140 BC. Florence Gaignerot-Driessen has demonstrated that there is a clear ritual dimension to this in the case of Dreros, a dimension indicated elsewhere by the use of the verb κατέσκαψαν. Excavations at Praisos have shed light on this question. This paper argues that, while there is no evidence for a widespread destruction by fire, there is clear evidence for the ending of Praisian sanctuaries and the forced abandonment of houses and (most intriguingly) the abandonment of large storage vessels. A particular archaeological ‘signature’ of these abandonments, evident at Dreros, Phaistos and Praisos, is the abandonment of the ‘household pithos’, which in many cases seem to be older than the houses in which they have been found. The paper also argues that these archaeological signatures of what had to be destroyed in order to eradicate a political community in turn shed light on what made this particular form of ‘citizen state’ so resilient.

Πρόσφατα έχει αναβιώσει το ενδιαφέρον τόσο για την ιστορική όσο και για την αρχαιολογική διάσταση της καταστροφής των πόλεων. Άλλο είναι η καταστροφή του κύριου οικισμού μιας πόλης και άλλο η αποτελεσματική εξάλειψη της πολιτικής κοινότητας. Τι χρειάστηκε για να καταστραφεί μια πολιτική κοινότητα; Οι ιστορικές πηγές είναι γεμάτες αναφορές καταστροφών μιας πόλης από μια άλλη στην Κρήτη από τους ύστερους κλασικούς έως τους ελληνιστικούς χρόνους. Ωστόσο, δεν οδήγησαν όλες αυτές οι καταστροφές στο τέλος της εν λόγω πολιτικής κοινότητας. Κάποιες όμως το πέτυχαν, και από την κλασική εποχή (όπου γνωρίζουμε για 49 πόλεις) ως τη ρωμαϊκή κατάκτηση του Μέτελλου (όπου γνωρίζουμε μόνο 24) ο αριθμός των κρητικών πόλεων μειώθηκε δραστικά. Η καταστροφή της Πραισού από την Ιεράπυτνα μεταξύ του 145 και 140 π.Χ. (Στράβων 10.4.12) ήταν μια τέτοια περίπτωση. Αυτό φαίνεται να αποτελεί μέρος ενός ορίζοντα καταστροφών (της Δρήρου, της Απελλωνίας και της Φαιστού) που σημειώθηκαν μεταξύ του 200 και 140 π.Χ. Η Gaignerot-Driessen απέδειξε ότι υπάρχει μια σαφής τελετουργική διάσταση στην περίπτωση της καταστροφής της Δρήρου, μια διάσταση που υποδεικνύεται αλλού με τη χρήση του ρήματος κατέσκαψαν. Οι ανασκαφές στην Πραισό έχουν ρίξει φως σε αυτό το ερώτημα. Αυτό το άρθρο υποστηρίζει ότι, ενώ δεν υπάρχουν στοιχεία για εκτεταμένη καταστροφή από πυρκαγιά, υπάρχουν σαφείς ενδείξεις για το τέλος των ιερών της Πραισού και την αναγκαστική εγκατάλειψη σπιτιών και (το πιο ενδιαφέρον) την εγκατάλειψη μεγάλων αποθηκευτικών αγγείων. Μια ιδιαίτερη αρχαιολογική “υπογραφή” αυτών των εγκαταλείψεων, εμφανής στη Δρήρο, τη Φαιστό και την Πραισό, είναι η εγκατάλειψη των “οικιακών πίθων”, που σε πολλές περιπτώσεις φαίνεται να είναι παλαιότεροι από τα σπίτια στα οποία έχουν βρεθεί. Το άρθρο υποστηρίζει επίσης ότι αυτές οι αρχαιολογικές υπογραφές όσον αφορά το τι έπρεπε να καταστραφεί προκειμένου να εξαλειφθεί μια πολιτική κοινότητα με τη σειρά τους ρίχνουν φως στο τι έκανε τόσο ανθεκτική αυτή τη συγκεκριμένη μορφή “κράτους πολιτών”.

Μετάφραση: Χριστίνα Χατζημιχαήλ

INTRODUCTION

Just before 140 BC the polis of Praisos in Eastern Crete was brought to an end by its neighbour, Hierapytna.¹ We can date this event with some confidence as taking place between the death of Ptolemy Philometor (145 BC; ICr 3.4.9 lines 42–4) in Egypt and the consulship of C. Laelius (140 BC; ICr 3.4.10 lines 22–23) in Rome. Before this Praisos had been a substantial polis whose territory stretched from the Aegean Sea in the north to the Libyan Sea in the south. The settlement of Praisos was not occupied thereafter, nor was its identity as a political community maintained. Praisos was not alone in this regard. The second century BC witnessed several such destructions in Crete, destructions which moreover brought an end to several political communities on the island.

Physical destructions of cities were relatively common in both Classical and Hellenistic Greece. But physical destruction of a major settlement by an enemy did not invariably bring Greek cities (that is, Greek political communities) to an end. Many cities were re-founded after apparently thorough destructions; many cities ‘destroyed’ in Classical and Hellenistic times persisted as political communities into later centuries. This is not a new observation. Edward Dodwell (1819, 243), writing in the early nineteenth century, noted:

Diodorus Siculus [B.11.c.25]² says that the Argians destroyed Mycenae, τὰς Μυκήνας κατέσκαψαν, and adds that it remained deserted to his time. Diodorus, in speaking of the destruction of sites, generally uses the word κατέσκαψαν, which supposes a complete razing; many of the cities, however, which he thus destroys, still exhibit considerable ruins, but long prior to his time.³ Livy is the destroyer of Italian cities, as Diodorus and Strabo are of those of Greece; but many of those which he [Livy] represented as ‘sine vestigiis’, still retain their walls, gates, and towers, in a state of high preservation.

Dodwell's observations are as relevant now as they were then. Several recent books have tackled the question of the destruction of cities, and whether sources like Diodorus and Strabo can be taken at face value. They have looked at how cities were destroyed and the lasting effects of these destructions (Fachard and Harris 2021b; Driessen 2013). They have, like Dodwell, noted the disparity between the literary record (where sources often suggest total eradication – κατέσκαψαν) and the physical facts, which often show that sites were re-occupied and communities re-formed or sometimes simply continuing after an apparently devastating destruction.

That the archaeological and literary records often do not match up is also hardly a new insight. There may be many reasons for this disparity between our two principal sources of information. One is the desire of our ancient writers to exaggerate for dramatic effect while reporting on events they had not witnessed but only heard at second hand. Another is the eagerness to infer widespread devastation of a city from limited excavation, in the hope that the literary and material records can be reconciled. This eagerness can often mislead archaeologists to make historical inferences that are, strictly speaking, unwarranted.⁴ A further difficulty is taphonomic: archaeological destruction horizons can often be elusive and are often only detectable through microstratigraphy and micromorphology (Karkanas 2021).

In this essay I want to concentrate on a fourth factor – the distinction between the physical, short-term effects of destruction and the long-term political consequences of such actions. Contributors to Fachard and Harris (2021b) note that many of the communities whose destruction they try to account for exhibit extraordinary resilience. These poleis were difficult to destroy. An examination of how political communities were brought to an end (that is, what was required to destroy not only a settlement but a polis) could therefore tell us a great deal about what originally sustained those communities – that is precisely what made them resilient. A clue to what this might have been lies in the wording that Strabo (10.4.12) uses to describe this destruction – κατέσκαψαν. Hence my title.⁵

Strabo's brief allusion to Praisos’ end might, moreover, lead us to think that a destruction is a simple event: once you have laid waste the polis’ principal settlement then the polis as a political community simply ceases to be. Here we need to make a distinction between the polis as a settlement (a town) and the polis as a political community. It was not uncommon in late Classical and Hellenistic times for the principal settlements of poleis to move (Mackil 2004). This did not necessarily entail the end of the political community – sometimes, as in the case of Myous and Miletus, the citizens of one polis chose voluntarily to merge with another; at other times they chose to move their principal settlement elsewhere. Destruction is of course a different matter, but we have to think clearly about what was being destroyed. To destroy a settlement is one thing, but to destroy a political community quite another: the κατασκαφή of a polis entails bringing the institutions that sustain it to an end. We should therefore pay particular attention to what an enemy focused on when destroying political communities such as Praisos and examine why some of these attempts were successful and others not. For the manner in which such poleis were brought to an end sheds oblique light both on the nature of those political communities and on the institutions that sustained them.

Recent scholarship has emphasised that the Cretan citizen-state⁶ differed in certain respects from other poleis in the Greek-speaking Mediterranean. These differences have been the subject of several recent books in a burgeoning field.⁷ In general, scholarship has tried to underline the differences between Cretan political communities and the ancient Greek community we know best – Athens. Athens was, of course, certainly not a typical Greek polis. It was larger than most and is much better documented. There is nonetheless a sense in which both the Athenian and Cretan political communities were citizen states.

Our understanding of what constitutes a ‘citizen state’ has changed in the past few decades. Sourvinou-Inwood (2000a; 2000b) has emphasised the role that religion played, not only in the life of the polis, but also in its constitution as a political community: a polis was a ‘community of cult’. In this respect (Morgan 2003), poleis were like ethne (ethnic confederacies such as Thessaly).⁸ Recent scholarship on Classical Athens has emphasised the centrality of religion and cult both to Athens’ sense of communal identity and their basic functioning as political communities. While Anderson (2018) sees Classical Athens as embodying an entirely different ontology (in Descola's [2013] sense) Blok (2014; 2017), more practically, sees Athens as a ‘covenant between gods and men’, a community continually sustained by communal rituals which also defined who was and was not a citizen. Cretan citizen states too had a strong communal ethos where citizenship was linked (often through forms of commensality) to the maintenance of corporate groups (Haysom 2011). In Cretan cities the commensal institution that helped to create and sustain the body of citizens was not just the sanctuary but the andreion (Seelentag 2015, 374–503; Whitley 2018a). Cretan cities too were defined by their own hiera kai hosia.⁹ In this respect scholarship on Cretan political communities in the historical period is converging with scholarship on Cretan Bronze Age polities, polities based on central courtyard complexes¹⁰ whose resilience is evidenced by their longevity. This is a point to which I shall return.

There is a broader, comparative dimension to this question. The polis was a particularly long-lived form of political community (Ober 2015). Some poleis lasted for over 1000 years. We know of 1035 such communities that date to the Archaic and Classical periods (Hansen and Nielsen 2004). Most persisted throughout Hellenistic times and retained a sense of civic identity under Roman rule well into the third century AD. Yet most poleis were not very large. If we judge the size of these poleis by ‘etic’ criteria used by anthropological archaeologists who have long worked on the comparative study of complex societies (and so states), most poleis were simply too small to be states. Not only are they small but they lack most of the qualities that anthropological archaeologists require of states.¹¹ For this reason, a parallel debate has emerged within Classics: was the polis a state?¹² Ancient historians have long simply assumed that they were, but the answer is far from clear cut. These issues are even more pressing when we turn to Crete. Cretan poleis were particularly small. We know of 49 autonomous political communities which fit Hansen's ‘emic’ criteria for a polis.¹³ This yields an average territory of 170 km² for Cretan poleis, much smaller than those to be found on other large Mediterranean islands such as Sicily, Euboea and Cyprus (Whitley 2014, table 1). If we strictly apply either Berent's (2000) or Flannery and Marcus’ (2012) criteria then no Cretan polis could count as a state.¹⁴ I believe that this conclusion is unwarranted, and that a conception of a ‘state’ based on a model derived from a comparison with early Mesopotamia and early Mesoamerica is likely to be misleading when applied more widely.¹⁵

Longevity may not be the only test of resilience. I will argue in this paper that another useful test is – how difficult was it to bring such communities to an end? Here we return to the issue of ‘destruction’. Praisos’ κατασκαφή was the last of a series of such events that took place on the island before the Roman conquest. Several Cretan cities had, over the course of the late third and second centuries BC, been razed to the ground. For some this destruction was final; for others a ‘destruction’ was simply a setback. This sequence of events has usually been seen as the province of the historian – destructions being events recorded by ancient authors whose consequences and significance are well understood. But what in the end does κατασκαφή actually entail? A ‘complete razing’ in Dodwell's terms? Does it moreover necessarily imply the intention to bring a political community to an end? Let us start with our literary evidence, before looking at the philological implications of the term κατασκαφή.

LITERARY SOURCES FOR POLEIS DESTRUCTIONS

Crete presents some rather acute source problems for the ancient historian. Very little Cretan history was written by Cretans, and no historian provides us with a continuous narrative of political events on the island. Though we are better served by our ancient sources for the Hellenistic than for earlier periods,¹⁶ the priorities of our principal sources (Polybius, Strabo, Diodorus) lie elsewhere. While Polybius’ narrative works within a solid chronological framework, he is only incidentally interested in Crete – his grander story is the rise of Rome, onto which the events of an island like Crete only occasionally intrude; Strabo (a geographer, not a historian) refers to events a century or more before his time (the reign of Augustus) without providing any kind of chronology. This problem is not, of course, peculiar to our period – Herodotus, Thucydides and Xenophon only look askance to developments in Crete. Epigraphy of course allows us to construct a general narrative account of Cretan political history in Hellenistic times (e.g. Chaniotis 1996). Relating this essentially political narrative to the archaeological record is, however, no easy task.

Strabo's description of Praisos’ destruction moreover raises some tricky philological questions. We have first to explain Strabo's choice of language, κατέσκαψαν. Liddell and Scott (LSJ⁸ 911) define κατασκάπτω as (in its primary sense) ‘digging down’, and in its wider usage as ‘destroy utterly, raze to the ground’, and κατασκαφή primarily as a ‘razing to the ground, destruction’.¹⁷ There are reasons to believe, however, that the verb κατασκάπτω and noun κατασκαφή imply more than a simple act of destruction. Conor argues that this verb is first generally applied specifically to the destruction of houses. Such destruction is a form of punishment with a distinct ritual dimension – not only the person but the household is deliberately and ritually defaced (Conor 1985). An example is a law from Locris οἰκία κατασπαπτέθο.¹⁸ Κατασκαφή then had both a ritual and a punitive dimension to it – symbolic punishment for a crime committed by a member of a household. What happens when this action is applied not simply to a household but to a political community as a whole?

Conor discusses two cases where this verb is applied by ancient chroniclers to the destruction of a city.¹⁹ The first does not necessarily imply a total destruction of either the material or the social fabric of a political community: Xenophon uses the verb in his account of the consequence of Athens’ defeat at the end of the Peloponnesian War – καὶ τα τείχη κατέσκαπτον ὑπ’αὐλητρίδων πολλῇ προθυμίᾳ (‘and they tore down the walls with great enthusiasm to the sound of flutes’; Xenophon, Hellenica 2.2.23). Here it is the Athenian exiles who joyfully take part in this destruction, which they understand to be the liberation of Greece rather than the eradication of ‘the Athenians’. That Athens is being punished, and that this has a ritual dimension, is however undeniable.

The second however implies something more drastic. In Arrian's account of Alexander's sack of Thebes in 335 BC he uses the phrase τὴν πόλιν δὲ κατασκάψαι ἐς ἔδαφος – ‘razed the city to the ground’ (Arrian, Anabasis 1.9.9). Here the sense of κατασκαφή as ‘destruction as ritual punishment’ is clear cut (since the Thebans had broken their word). Arrian goes on to describe how Alexander enslaved the women, children and men (presumably male citizens) and distributed Thebes’ land amongst its neighbours. Other sources for this event use similar language. Plutarch (Life of Alexander 12.5) says ἡ δε πόλις ἥλω καὶ διαρπασθεῖσα κατεσκάφη (‘the city was taken and once seized razed’); Diodorus Siculus (16.14.4) says the king (Alexander) τὴν μὲν πόλιν κατασκάψας (having razed the city) went on to Athens. These descriptions imply a total eradication of the political community (the citizen body and its capacity to renew itself), as well as the physical destruction of the city – clearly this was both an act of punishment as well as an act of terror.²⁰

Need some form of ritual punishment (the destruction of walls and houses) necessarily entail an intention on Alexander's part to eradicate Thebes as a political community? Our sources imply as much. His father Philip II had destroyed both Methone (in 354 BC) and Olynthus (in 348), and these settlements were not occupied thereafter.²¹ And yet in 316–15 BC Thebes was re-founded and its polis reborn. Most of our sources (e.g. Pausanias 9.7.1–2; Diodorus 19.53.1–2) attribute this to the agency of one man – Cassander, then ruler of Macedon– a man moreover with a particular personal grudge against Alexander. But attributing the re-foundation to one powerful individual does not quite explain how, within one generation, ‘the Thebans’ had managed to stage a remarkable comeback. Many other cities – including Athens – contributed to this re-foundation and were not simply motivated by a desire to curry favour with Cassander (Kalliontzis and Papazarkadas 2019). Alexander himself may have allowed this to happen since in sparing some Thebans (and in refraining from destroying its temples or committing other acts of impiety) he was at least allowing for the possibility that ‘the Thebans’ could reconstitute themselves at some future date.²² If there are still some citizens of Thebes – that is, some Thebans – then the Theban citizen-state can be reborn.

Κατασκαφή then need not, necessarily, imply the final end of a political community. Or rather it may imply the intention to do so on the part of an Alexander (or anyone else) – while also indicating that this intention was very rarely fulfilled. Political communities – at least poleis – were very difficult to destroy. This was in part because the collective identity of political communities did not invariably depend on there being a principal settlement which provided the focus of the polis, and which (physically) could be destroyed. Political communities of the ‘citizen state’ type could be astonishingly resilient. To stick just with Boeotia, the settlement of Plataea (Plataiai) was destroyed more than once during the Classical period, but the political community as such was never brought to an end; Thespiai endured a similar pattern of attempted destruction and re-foundation.²³ But the most extreme case, in the Classical world, is that of the Messenians (Luraghi 2002; 2008). Though the territory of Messenia had been incorporated into the Spartan state at some time in the seventh century BC, ‘the Messenians’ managed to reconstitute themselves in the fifth, basing their revolt on an older settlement which had existed on Mt Ithome (the so-called Third Messenian War).²⁴ Though there was some kind of settlement here, Mt Ithome cannot be considered a ‘principal settlement’ of Messenia in the same way that Athens, Thebes or Corinth were the principal settlements of Attica, Boeotia or the Corinthia. Habitation in or around Mt Ithome was intermittent, and its occupation during the ‘Third Messenian War’ temporary. The end of the war and the reassertion of Spartan control did not, however, bring this political community to an end. Messenian exiles continued to exist and think of themselves as Messenians. Conflict continued throughout the fifth century and beyond, conflict which resulted in several Messenian victories, one of which is commemorated in the most illustrious of victory dedications we know of from the Classical world.²⁵ In brief, Messenians continued to constitute themselves as some kind of ‘citizen state’ even when they had neither city (polis/asty) nor territory (chora). When Messenia was re-founded with the help of Thebes after the Spartan defeat at Leuktra there were Messenians to populate the new city.

To return to our case: if we follow Conor (1985), the verb κατέσκαψαν seems to imply that Hierapytna was punishing Praisos for some crime – though Strabo's bare account gives no hint of what this was punishment for (impiety?). Punishment of this kind need not result in the total eradication of the political community – but in this case that appears to have been the result. Were there peculiar circumstances in Hellenistic Crete that might have led to this?

CRETE FROM THE ARCHAIC TO THE HELLENISTIC: THE RATE OF DESTRUCTIONS

Crete was an island that, in earlier Archaic times, seems to have supported many autonomous political communities (see Fig. 1). Perlman (2004a) counted 49 for the Archaic and Classical periods, but we can be sure that before 500 BC there were more. In Homer, Crete has a ‘hundred cities’ (Homer Iliad 2.649: Κρήτην ἑκατόμπολιν), and while no-one has yet to find 100 names of Cretan cities, the names of at least 74 potential political communities have been identified (Faure 1960). Some of these names might be associated with the large number of nucleated settlements (all 10 ha or more in size) established before 900 BC which have been identified by Wallace (2010a; 2010b). One such settlement was certainly a polis. This is Prinias (almost certainly not Rhitten), which had been abandoned in the late sixth century. Prinias’ status as a major polis is not in doubt, as its numerous legal inscriptions confirm.²⁶ The case of Azoria, destroyed (or abandoned) around 475 BC, is more dubious. But even if we lack evidence to specify Azoria's political status, the community established on the hill of Azoria in Archaic times seems to have had some form of civic identity.²⁷ During the fifth and fourth centuries the number of political communities remained stable (49) – there was warfare to be sure but not final destructions. This relative stability seems to have persisted throughout most of the third century BC. Our principal evidence here is twofold: a) the treaty that Eumenes of Pergamon made with various Cretan cities, datable to around 183 BC; and (b) the cities known to be issuing coins around this time.²⁸ This inscription records at least 30 political communities who were both able and willing to make a treaty with Pergamon, with two more names that we cannot decipher. It necessarily excludes Itanos (an ally, willing or not, of the Ptolemies; Spyridakis 1970, 68–103), Kydonia (which stood aloof) and the various Πόλεις Υπήκοοι (subordinate or dependent poleis).²⁹ Other poleis which did not make a treaty with Eumenes were still minting coins in the first part of the second century BC (Sanders 1982, 9–13 and fig. 2). Given all these factors, Sanders’ estimate of 43 poleis still in existence by the end of the third century BC seems reasonable (see Table 1).

Fig. 1. Map of Crete showing major cities and sanctuaries. Prepared by Kirsty Harding.

Table 1. Number of know poleis on Crete the late Archaic, Classical, Hellenistic and Roman periods (information from sources given in text).

It is only in the second century BC that the process of elimination of political communities seems to accelerate: Apollonia/Apellonia, Rhaukos, Phaistos, Dreros and Praisos (all major players in Cretan politics in earlier times) were eradicated (Chaniotis 1996, 44–60; see Fig. 1). Such a rate of destruction appears unprecedented. By Roman times the number of autonomous communities was much reduced. Sanders counts only 24.³⁰

Historians here might point to external factors – of which the rise of Rome and the series of wars that established Rome's dominance of the Aegean by 146 BC with the destruction of Corinth (Williams et al. 2021) were the principal causes. I will not attempt here to evaluate this point. My focus is different. I am interested in what the destruction of Praisos and similar destructions entailed. Or, to put it another way, what did it take to destroy a political community? A common-sense answer to this question is that you destroy that community's principal settlement, after which it ceases to exist. But as we saw in the cases of ‘the Thebans’ and ‘the Messenians’, this answer is unsatisfactory.

In Crete resilience is a feature of many of the 49 political communities we know of from textual and epigraphic sources. The apparent archaeological disappearance of a settlement moreover must be distinguished from the ending of a political community. So, for example, at least one major polis seems to disappear from the archaeological record for almost a century (Knossos in the sixth century).³¹ This gap in the record, however, does not seem to have entailed the disappearance of the ‘Knossians’ as this polis survived as an independent state until the Roman conquest (Whitley 2023, 148–62).

Table 2 shows the number of known destructions in Crete between the middle of the third and the latter part of the second century BC (see Fig. 1). Not all of these destructions entailed the end of a political community. Let us start with Lyktos/Lyttos, one of the oldest, largest and most important of Cretan poleis. Two attempted destructions of Lyttos are known from the historical record. Diodorus notes (16.62.3) that at some point in the mid-fourth century BC the Knossians πόλιν κατελάβοντο τὴν καλουμένην Λύκτον (took the city named Lyktos) – but Lyttos was not thereby brought to an end, nor was it conquered by Knossos. Later, in the third century (probably around 220 BC), Polybius (4.54.1; see Chaniotis 1996, 36–8) records that the Knossians managed to take the settlement itself (the Lyttian army being elsewhere):

συνοήσταντες οἱ Κνώσιοι τὸ γεγονὸς καταλάμβανονται τὴν Λύττον, ἔρημον οὖσαν τῶν Βοηθησόντων καὶ τὰ μὲν τέκνα καὶ τὰς γυναῖκας εἰς Κνωσὸν ἀπέπεμψαν τὴν δε πόλιν ἐμπρήσαντες καὶ κατασκάψαντες καὶ λωβήσαμενοι κατὰ πάντα τρόπον ἐπανῆλθον.

Table 2. Sequence of actual and attempted destructions of settlements and poleis from the late 4th to 2nd century BC in Crete.

In Waterfield's (2010, 266) idiomatic translation:

The Cnossians response was to seize the now defenceless Lyctus and remove the children and womenfolk to Cnossus. Then they set fire to the town, razed it to the ground and did everything they could to turn the place into a ruin.

Polybius’ language here suggests that it was the Knossians’ intention to eliminate Lyktos as a polis. The Knossians certainly managed to cause extensive damage to the city itself. A clear destruction horizon has been picked up in several separate locations in rescue excavations conducted by the Greek archaeological service.³² Yet the political community – the Lyktians – persisted. There were several reasons for this. The principal settlement was both large (between 60 ha and 100 ha) and loose, a fact which made it too difficult a task for the Knossians to raze the city completely.³³ The abduction of women and children, and the extensive burning of the town, evidently damaged but did not thereby bring an end to the polis of the Lyktians – the male citizens managed to rebuild their community thereafter. The recently discovered treaty of alliance between Lyttos and Olous mentions a Lyttian festival that commemorates the re-foundation of the city sometime after 220 BC (SEG LXI 722, lines 4–8). The independent polis of Lyttos survived until the conquest of Metellus, existed in some form in the time of Strabo and persisted in some form until Late Antiquity.³⁴

Smaller sites were not so lucky. Between Knossos and Lyktos we know from archaeological excavation that the site of Prophitis Elias above Archalochori was destroyed in the mid to late third century BC – though we know nothing of the circumstances, nor even the ancient name of the site.³⁵ Around 185 BC Knossos appears to have destroyed the settlement of Lykastos, a settlement not regarded as a polis by Perlman.³⁶

In the second century BC the pace of the destruction of major political communities picked up. At some point probably in the early second century BC, Dreros was destroyed by the newly re-founded Lyktos. There is no direct literary evidence for this destruction whose date and nature we infer from a mix of archaeological and epigraphic data.³⁷ But Florence Gaignerot-Driessen is surely right to draw attention to its ritual dimension – a true example of κατασκαφή.³⁸ Though there was no ‘fire’ destruction, the Archaic inscriptions for which Dreros is so famous³⁹ were removed from their position in the walls of the temple of Apollo (Perlman 2004b, 191–5) and placed in a large pit/cistern in the Agora area (Demargne and Van Effenterre 1937a, 27–32; 1937b). As is the case in other Cretan cities, the destruction horizon is marked by the abandonment of large storage vessels of apparently Archaic date within houses (Zographaki et al. 2020, 174–6, figs 11 and 12). Gaignerot-Driessen sees both the destruction of the city's laws and the destruction of the household unit (as represented by the pithoi) as a necessary ritual element in the destruction of Dreros as a political community. She argues that this ritual dimension to ‘community destruction’ is connoted by the Greek noun κατασκαφή (and so is indicated in any passage where κατέσκαψαν or κατασκάψαντες is used, as in the cases of Praisos and Lyktos respectively).

In her argument Gaignerot-Driessen (2013, 288) invokes Conor (1985) where he says that κατασκαφή (and corresponding verbs) imply an intention to punish which also has a ritual dimension. Evidence for this ritual dimension can be found in another much-discussed inscription, the ‘Oath of the Drerian Ephebes’ (ICr 1.9.1, side A), datable to the late third century BC. In this oath, the agelai (young men who were candidates for citizenship) of Dreros swear, not by one god but by many gods, eternal enmity with Lyktos and corresponding eternal friendship with Knossos. The extreme language of this oath is something commentators have struggled to explain (Van Effenterre 1937). The date of this inscription, just after Knossos’ unsuccessful attempt to seize and destroy Lyktos, suggests that this is not just an oath but a provocation, one which the Lyktians might well have seen as an act of impiety.⁴⁰ Further support for this is provided by the inscription (SEG LXI 722) which mentions the seizure of Dreros by Lyttos (lines 9–10), an event which Kritzas (2011) argues must have taken place between 217/6 and 183 BC.⁴¹ This is a treaty between the Olountians and Lyttians and mentions joint festivals which commemorate both the seizure of Dreros by Lyktos and the Lyttian success in re-founding their city after its partial destruction by the Knossians. Such festivals imply that there must have been a ritual dimension to the destruction of Dreros, a feeling that the gods were mocked by the oath of the Drerian youths and that the Lyktians had justice on their side in bringing this political community to an end.

No single source records the Dreros destruction. We cannot then quite assess how it was destroyed – how, that is, the Lyktians succeeded in bringing this community to an end where the Knossians had singularly failed in the Lyktian case. To assess what was entailed in a successful destruction we have to turn to that of Apollonia/Apellonia for which we have both literary and archaeological evidence. The degree of violence involved in the elimination of this city seems to be of a different order. Polybius (28.14) seems to have been genuinely shocked by what happened:

παρασπονδήσαντες τοὺς Ἀπολλωνιάτας κατελάβοντο τὴν πόλιν καὶ τοὺς μεν ἄνδρας κατέσφηξαν, τὰ δ’ ὑπαρχοντα διήρπασαν, τὰς (δὲ) γυναῖκας καὶ τὰ τέκνα καὶ τὴν πόλιν καὶ τὴν χώραν διανειμάμενοι κατεῖχον.

A translation might read:

Breaking their treaty with the Apollonians they [the Kydonians] took their city and slaughtered the men, plundered their goods, and, having divided the women, children, city and territory amongst themselves, held onto [all these things].

The wholesale destruction of Apollonia/Apellonia by Kydonia in 171/0 BC then required the unilateral breaking of treaties on the part of Kydonia (with whom they shared συμπολιτεία), the slaughter of all male citizens, the seizure of women and children, the occupation of the city and the division of the land amongst Kydonian citizens. Apollonia ceased to function as a political community after this. As with Lyktos, there is some archaeological confirmation of the date of destruction, which in this case involved the complete razing of at least one major public building (interpreted by the excavators as an andreion).⁴²

Apollonia, then, provides one template for how you eradicate a long-standing political community. You must be both ruthless and treacherous. You must kill the male citizen-soldiers and deport the women and children. You must permanently occupy both the city and its land and divide it amongst your own citizens. Soon after the Kydonians eliminated Apollonia, the Knossians and Gortynians joined forces to bring to an end Apollonia's near neighbour, Rhaukos. The victors divided the territory between them.⁴³

This brings us to the two mid-century destructions alluded to by Strabo, where he uses the verb κατέσκαψαν. Gortyn destroyed Phaistos – κατέσκαψαν Γορτύνιοι (Strabo 10.4.14) – an event we can only date in very general terms to between 170 (the Apollonia destruction) and 150 BC. Excavations of Phaistos in several locations (to the west of the Palace itself and in the Chalara area of the site) by the Italian School have again provided archaeological confirmation of this destruction.⁴⁴ As with the site of Apollonia, much of this evidence takes the form of Hellenistic fine wares whose closest parallels are those from the Little Palace well in Knossos. But along with the fine wares are some quite spectacular antiques. These take the form of large storage pithoi, one manufactured in the seventh century BC, and one inscribed Ἐρπετιδαμο Παιδοπιλάς οδε, which seems to date to around 700 BC.⁴⁵ These pithoi must then have been over 400 years old at the time of the site's destruction.

Such discoveries are not confined to Phaistos. Both Rethemiotakis and Lebessi had discovered at least nine seventh/sixth-century pithoi in the Lyktos destruction horizon of 220 BC.⁴⁶ Recent excavations at the site of Prophitis Elias above Archalochori in Central Crete (between Knossos and Lyktos) have revealed several sixth-century Archaic pithoi in association with a destruction horizon of the mid-third century BC (Galanaki, Papadaki and Christakis 2015); apparently Archaic pithoi have recently been found in the destruction/abandonment horizon of Dreros itself (Zographaki et al. 2020, 174–6, figs 11 and 12). Much the same is true of sites in Eastern Crete. The early second-century BC abandonment/destruction horizon of the site of Trypetos close to Sitia is marked by the abandonment in situ of various pithoi which can be assigned to Brisart's Afrati group (and thus date to a little after 600 BC).⁴⁷ These pithoi then are considerably older than the Hellenistic (third-century BC) houses in which they were found.

Or so I would argue. Many have, however, doubted that most, or indeed any, such pithoi found in Hellenistic destruction horizons can have been quite so old. Such critics suggest implicitly that most ‘Archaic’ pithoi are in fact ‘archaising’. I disagree, but this point requires discussion.

THE ANTIQUITY AND LONGEVITY OF CRETAN PITHOI

It is not quite true to say that we do not know what Hellenistic Cretan pithoi looked like. There are many examples of pithoi with minimal decoration (except for raised bands) that have been found in Hellenistic levels in Crete. There is one example from a third-century BC floor level from the Stratigraphical Museum Excavation at Knossos (Warren 1985, 128, fig. 17); a plain example inscribed ΠΑΝΣΩΝΟΣ in Ionic script (ICr 3.6.29; see Whitley 2011, 28–9 n. 21) comes from Bosanquet's (1901–2, 269, pl. XII) excavation within the Almond Tree House at Praisos; another (uninscribed) example was found in the 2007 excavations just below this structure (Whitley 2011, 33 and 38, fig. 40); and plain pithoi associated with the destruction/abandonment horizon have been found alongside ‘Archaic’ ones from the recent excavations at Dreros (Zographaki et al. 2020, 174–6, figs 11 and 12). This does not, of course, preclude the possibility that such plain pithoi might have been contemporary with ‘archaising’ examples of Hellenistic date, since we know so little about how this vessel form evolved from Archaic to Hellenistic times.⁴⁸ This is part of a more general archaeological problem where the study of plainer vessels is neglected in favour of more highly decorated ones, and it is certainly the case that highly decorated ‘daedalic’ pithoi loom large both in the literature and in the surface finds of major Cretan cities (Savignoni 1901).

Could any of these ‘Archaic’ pithoi found in Hellenistic destruction horizons in fact be ‘archaising’? In some cases this would be unlikely if not impossible. Pithoi inscribed with Archaic letter forms, such as examples from Lyktos (Lebessi 1974, 493–9, fig. 512ab; Rethemiotakis 1984, 58), Phaistos (Guarducci 1952–4; La Rosa and Chiara Portale 1996–7, 256–7), Archalochori (Kotsonas 2022, 145, fig. 6:8) and Azoria (Haggis et al. 2011, 58, fig. 42; see discussion in West III 2007; 2015), must be Archaic – no-one would argue that the Geometric example from Phaistos inscribed with the name ‘Erpetidamos’ (Levi 1969), reckoned to be the earliest alphabetic inscription from Crete (Jeffery 1990, 468, no. 8a), is Hellenistic. There are moreover some technical considerations which would have made it very difficult for Hellenistic pithos makers to produce convincingly Archaic-seeming pithoi in an ‘archaising’ style. Cretan pithoi are decorated using moulds of a similar type to those used to produce the distinctive range of Archaic Cretan terracottas (as in Pilz 2011). They often share the same iconography of griffins and sphinxes. While new moulds sharing the same iconography (and style) can be made from old terracottas by a process long recognised in coroplasty (Nicholls 1952), it is much more difficult to make a new mould (for pithos decoration) from an old pithos to use in the production of a newer one than it is to fashion a new mould from an old terracotta plaque. If such a process were being used, we could detect a ‘series’ in the same way we can detect a ‘series’ in coroplasty. No-one has yet detected such a series in the case of pithoi.

There are two possible exceptions to this. One is the plainer pithoi apparently manufactured in the Afrati workshop (found at Trypetos and Praisos; Brisart 2007). These must be later than the seventh century, but not so much later that they can be considered ‘Hellenistic’. The other are the pithoi decorated with raised bands and stamped rosettes, which have been found in Praisos (Savignoni 1901; Whitley, Prent and Thorne 1999, 248, fig. 14, no. 591.4.10) and at the recent excavations at Dreros (Zographaki et al. 2020, 174–6, figs 11 and 12). The continued use of stamped rosettes is not subject to the same technical constraints as more elaborate depictions of sphinxes or griffins, and it is possible that some of these might conceivably be ‘archaising’ Hellenistic rather than Archaic proper. Even if some of these examples are ‘archaising’, their decoration is deliberately conservative; they were made to look old. For all these reasons the majority of the Archaic-seeming pithoi found in these Hellenistic destruction/abandonment horizons must indeed have been Archaic in date.

Almost all of them are also to be found in houses. Pithoi are particularly connected to the household in Greece in general and Crete in particular (Ebbinghaus 2005; Whitley 2018b, 60–63). Pithoi were not abandoned simply because they were difficult to move – though they were that – since the pithoi at Trypetos (Vogeikoff-Brogan 2011a; 2011b) must have been moved into these houses from elsewhere. People may move houses, but they bring their pithoi with them. Pithoi were heirlooms in a strong sense of that term, heirlooms with strong symbolic links to the household.

In listing what was destroyed and what was left behind after a destruction we are beginning to look at what needed to be destroyed to bring a political community to an end. This brings us back to the destruction of Praisos itself.

THE DESTRUCTION OF PRAISOS

Praisos was the principal city of the far East of Crete (the Siteia peninsula; Perlman 2004a, 1183–4, no. 984). Bounded by Hierapytna in the west and Itanos to the north-east, its territory (Fig. 1) was large by Cretan standards and comprised at least two subordinate communities – Stalai (near modern Makriyialos) and Dragmos (whose location we do not know).⁴⁹ Though the territory was large, the urban core at around 28 ha was not – Strabo (10.4.6) calls it a πολίχνιον (‘little city’).⁵⁰ There is, moreover, an additional ethnic dimension to Praisos and its destruction. The ancient sources begin with a passage in the Odyssey, where Odysseus (pretending falsely to be a Cretan) names the five peoples of ancient Crete – which ends with the great-hearted ‘Eteocretans’ or ‘true Cretans’ (Hom. Odyssey 19.175–7). This is followed by a passage in Herodotus, which recounts a story recounted to him (it is implied) by the people of Praisos, of how King Minos led an expedition to Sicily (Herodotus 7.170–1). This expedition turned out disastrously. The whole centre of the island was emptied of King Minos’ subjects (the true Cretans?) apart from the Polichnitai (in the West) and the Praisioi (in the East). Strabo, basing his statements on Staphylos of Naukratis (writing in the fourth century BC), links the Praisioi with Homer's Eteocretans (10.4.6). And it is Strabo who tells us of Praisos’ destruction, κατέσκαψαν δ’ Ἱεραπύτνιοι (Strabo 10.4.12).

The implication here is that, in destroying this political community the Hierapytnioi were also bringing to an end a distinctive ethnic group. The discovery of the ‘Eteocretan’ inscriptions by Halbherr and Bosanquet (inscriptions written in Greek letters but not in the Greek language) confirmed, in these scholars’ eyes, that the people of Praisos formed an aboriginal survival from the time before Minos.⁵¹ The retention of what might well have been simply a ‘ritual’ language (i.e. Eteocretan) down into the fourth century BC, and the ‘myth of descent’ that Herodotus’ tale represents, is sufficient in some scholars’ eyes to confirm that Praisos was, at the time of its destruction, a distinct ethnic as well as political community.⁵² As both an ethnic as well as a political community it would then surely be more, not less, difficult to bring to an end.

Praisos has been investigated by many archaeologists of many different nationalities (Whitley 2015). Excavations have been extensive, but fitful (that is, major excavations have not been sustained in the same way as they have been for major Bronze Age Cretan sites). Four areas investigated by Halbherr, Bosanquet and others are particularly relevant to understanding what κατασκαφή might have entailed in both material and social terms: the sanctuary deposit near the ‘spring at Vavelloi’; the principal sanctuary on the Third Acropolis (or Altar Hill); recent excavations of a store-room associated with a possible sanctuary near the summit of the First Acropolis; and excavations both of and close to what Bosanquet called ‘an andreion’ or ‘Almond Tree House’ on the north-west slopes of the First Acropolis (Fig. 2).

Fig. 2. Map of Praisos showing areas affected by the Hierapytnian destruction of 145–140 BC. Drawn by Kirsty Harding, after an original by Howard Mason.

The ‘spring at Vavelloi’ is a site that lies below modern ‘Nea Praisos’ (still referred to locally as Vavelloi) and about 0.5 km south of the three hills of ancient Praisos. Though both Halbherr and Bosanquet attempted to investigate the rich votive deposit here, it had already been thoroughly looted before either of them arrived. The deposit comprises many terracotta plaques dating from the Geometric period through to the Early Hellenistic (Halbherr 1901, 384–92; Forster 1901–2, 280–1; 1904–5; Prent 2005, 306–8, no. B.46). Though these plaques are now dispersed through many of the world's great museums, recent re-evaluations have been able to identify which plaques originate from this votive deposit. These reappraisals underscore the close similarity in the types and iconography of the plaques found here and those found at another ‘spring shrine’, that at Anoixe near Roussa Ekklesia.⁵³ The majority of plaques from both these spring shrines comprise Forster's (1904–5) type 8, man with staff – a masculine image with a long history (represented by three stages in a mould series).⁵⁴ Both sanctuaries have examples of terracotta plaques with a distinctly masculine iconography, that of a ‘warrior abducting a youth’.⁵⁵ The masculine terracotta plaques from Vavelloi itself show the greatest degree of iconographic continuity, which can be traced over four centuries in two distinct images. The first is that of a (nude) male with hand on hip,⁵⁶ the second of ‘warriors’ with plumed helmets and shields facing left (Fig. 3).⁵⁷ It is not too much, I think, to suggest that all these plaques may have been related to a kind of initiation ceremony for young men – a process by which they became citizens and warriors through a ritual not dissimilar from the famous account of Ephorus/Strabo (Strabo 10.4.20-1; Ephorus, FGrHist 70 F 149).⁵⁸

Fig. 3. Sequence of terracotta ‘warrior’ plaques from the sanctuary of Vavelloi and from trenches below the Almond Tree House (both at Praisos). After Whitley 2016, 259, fig. 2. Prepared by Kirsty Harding.

The major sanctuary of Praisos was, however, the Altar Hill, or Third Acropolis. This is the only sanctuary in the vicinity with clear evidence of animal sacrifice which took place around an open-air altar.⁵⁹ Finds include examples of bronze armour, both full-scale and in miniature (Bosanquet 1901–2, 258, pl. X; Hutchinson, Eccles and Benton 1939–40, 57). The altar was marked out by a balustrade decorated with several large terracotta figures. It was on this balustrade that the laws of this political community – those written in both Greek and Eteocretan – were originally displayed. Halbherr and Bosanquet found three inscriptions damaged on the top of the Altar Hill itself, one to the north-west and seven to the south-west – that is, outside the boundaries of the city.⁶⁰ Further investigation by Davaras (1982) revealed more possible debris from this destruction on the lower northern slopes of the Altar Hill. That the majority of inscriptions were found not on the hill itself but just below it was, for Bosanquet (1901–2, 258, pl. X), sufficient evidence that this sanctuary had been deliberately targeted and ransacked.⁶¹ Care was then taken by the Hierapytnians ostentatiously to destroy the city's laws. This is destruction as performance.

The third area is the summit of the First Acropolis, where investigation first by Nikos Papadakis (2001) and then by Chryssa Sophianou (2013; 2014) has revealed a large store-room (5.15 x 2.15 m, to a depth of 1.92 m) full of ‘Archaic’ pithoi on stone bases associated with Hellenistic pottery, amphoras with pointed feet, lamps and loomweights.⁶² Sophianou (2010) argues that this store-room is associated with a small sanctuary of Kybele, a terracotta representation of whom was among the finds.⁶³ This room too must then be associated with the Hierapytnian destruction of the city, and the pithoi (which I have seen) do seem ‘Archaic’ in that they are decorated with rosette bands – though in this case they are not associated with the household.

The fourth area is that in and around a large structure on the north-west flanks of the First Acropolis, first investigated by Bosanquet (1901–2, 259–70). This Classical structure, with an imposing façade of ashlar limestone blocks, is much larger and more complex than most Cretan houses, a fact which suggests it played some civic role (Westgate 2007, 440–1). Bosanquet called it by two names – the ‘Almond Tree House’ and ‘the Andreion’.⁶⁴ In 2007, further excavations took place on the terrace immediately below this structure, with the intention of reaching some Classical/Hellenistic houses (Whitley 2011, 8–37). This entailed excavating through Bosanquet's dump.

Only in one area did excavation manage to reach a clear floor level. This was context 216, which contained a hearth, a warming stone and pottery left in situ. There was no destruction horizon as such – no signs of wholesale destruction by fire. There were also very few fine wares, such as those comparable to finds from Phaistos. The only firm dating evidence was provided by a coin of Praisos containing a ‘winged thunderbolt’, of a type which ought to date to the late third or early second century BC.⁶⁵ It is therefore compatible with our known historical destruction date. There were, however, two pithoi, almost certainly Archaic, one of which is from Brisart's (2007) Afrati group. It is very similar to examples from Trypetos, studied by Natalia Vogeikoff-Brogan (2011a; 2011b; see also Whitley 2011, 27–32).

This evidence was interpreted as indicating not so much a wholesale ‘destruction’ as a forced abandonment of the central part of the settlement (at 28 ha, much smaller than Lyktos). Finds here do help answer part of our question – how do you destroy a political community? Well, in order to accomplish this, you have to destroy the household. This can be achieved by abduction of the women and children – as was the case in Apollonia and Lyktos. To do this the artefact type through whose agency the household is maintained over several generations (the household pithos), must be left behind, as the women and children will have to be incorporated into new households in Hierapytna.

But what of the men – the warrior citizens of Praisos? As we saw in the case of Lyktos, if the men are not killed or enslaved, then the political community can reconstitute itself. At Praisos we have some idea about how the male citizen body reproduced itself from generation to generation. The evidence for this takes the form of terracotta plaques, two of which were found in the dump fill (the debris from Bosanquet's excavations) in trench 200 above our abandonment horizon. The masculine iconography of these plaques deserves some scrutiny (Whitley 2011, 16–19). One (trench A 202.6 object 5) is of a young man with his hand on his hip (Forster 1904–5, type 25, 26 or 27); the other (Fig. 3, far left: 205.7 object 6) shows a warrior with a plumed helmet and a shield in the shape of a ram's head. This too can be related to earlier types (10 and 11), two in fact, which date first to Late Geometric and then Late Archaic times (Fig. 3), some of which have been found in the ‘Fountain of Vavelloi’. These plaques, though not in anything like primary contexts, must be associated with the institution of (one of) the Praisos’ andreia, where participation in communal messes seems to have been part of both the privileges and the duties of citizenship.

Though no examples of our warrior series can be found demonstrably later than the fourth century, the youth with hand on hip is a series which only ends in Hellenistic times. No such plaques can be dated later than the early second century BC – indeed no plaques which belong iconographically to the highly distinctive series of Praisos plaques can be dated to after the time of the city's destruction. Eastern Crete, which had been the most prolific source for terracottas during Archaic, Classical and Early Hellenistic times, abruptly ceases to produce anything that can be dated to late Hellenistic times. The majority of these plaques seem to have been deposited at two spring shrines – Vavelloi (near to the city) and at Anoixe/Roussa Ekklesia. The published finds from Roussa Ekklesia (Erickson 2009) are mainly Archaic, but deposition of male terracottas of distinctive Praisos type (Forster 1904–5, types 25, 26 and 27), which form a series, seems to have continued after the Classical period and into middle Hellenistic times; there are no late Hellenistic examples from here.

None of the known finds from these sanctuaries (Altar Hill, Summit of First Acropolis, Vavelloi and Roussa Ekklesia) can then be dated to after the middle of the second century. There is no evidence of late Hellenistic or Roman reuse of any of these cults.

This brings us to the other part of the answer to the question ‘how do you destroy a political community?’ It is not enough just to enslave the women and kill the men. You have to destroy its institutions: you start with the household and family and go on to the public and ritual sphere – that is, the city's cults. With the important exception of the sanctuary of Dictaean Zeus at Palaikastro, none of the urban or extra-urban cults in the territory of Praisos has finds which post-date 140 BC. The series of terracotta plaques, with an iconography clearly linked to the initiation of male citizens, stops abruptly around this time. In destroying institutions – specifically households, andreia, the city's laws and the city's cults – you also have to bring to an end the ‘symbolic means of social and political reproduction’ represented (in part) by the pithos (for the household), the plaques (for the andreia), the cults themselves (through animal sacrifice and votive deposition) and finally the city's laws. These finds (or their absence) then provide a kind of archaeological signature for the end of a political community.

There remains the problem of exactly how the city itself was sacked. Here we suffer from a poor understanding of Cretan warfare in Hellenistic (and earlier) times.⁶⁶ Unlike mainland cities, few if any Cretan cities of Hellenistic date were completely surrounded by walls (Coutsinas 2013) – with the important exception of Itanos with its Ptolemaic garrison (Coutsinas 2010; 2013, 401–4). Warfare seems to have relied more on archers and slingers – evidence for which comes (Kelly 2012) in the form of stamped lead slingshots. The destruction of Praisos then (when it came) would have involved neither a siege of a walled city nor any hoplite battle on the plain. Instead skirmishes by a mixture of slingers and archers (perhaps supported by some heavy infantry) must be envisaged. Quite how in military terms a city was first taken and then destroyed remains unclear.

To sum up then, the κατασκαφή of a Cretan citizen state involves the actions given in Table 3, which have clear material (archaeological) correlates. These then define the institutions that sustained Cretan citizen states. How much had such states changed since Archaic times? Though Perlman (Gagarin and Perlman 2016) and Seelentag (2015) have sharply different approaches to understanding these citizen states, and so to understanding their evolution (Seelentag preferring to work back from largely Hellenistic evidence to the Archaic period, Perlman working forward), both agree that they were more democratic than Aristotle implies and that they maintained a strong corporate ethos, based on initiation into citizenship. The distinctly conservative, political, iconography of the Praisos plaques (Fig. 3) suggests that the fundamental institutions of the polis of Praisos did not change that much between 700 and 250 BC (Whitley 2016; 2018a).

Table 3. Archaeological signatures for the destruction of a political community relating to the household, cults and the city's laws.

The destruction of Praisos was the last such event in Hellenistic times. There is little in the epigraphic record to indicate that the Praisians were aware of the threat posed by their neighbour. Praisos had established friendly relations with her other neighbour to the north-east, Itanos, around 164/3 BC (ICr 3.4.9, lines 60–5; Chaniotis 1996, 303–6); border issues between Hierapytna and Praisos seem to have been settled in the decades before the Hierapytnian takeover (ICr 3.4.9, lines 65–70; Chaniotis 1996, 306–7). The final phase of the Hellenistic period in Crete (circa 140–67 BC) is one where political communities had achieved a measure of stability, and where territorial disputes (such as that between Hierapytna and Itanos over control of the sanctuary of Dictaean Zeus of Palaikastro; ICr 3.4.9) were settled by mediation and arbitration. Increasingly Romans were involved in resolving these disputes (e.g. ICr 3.4.10). The Roman presence in Crete had been felt as early as the late third century. Around 217–209, Ptolemy IV Philopator had put a Roman, one Lucius Gaius, in charge of his garrison at Itanos (ICr 3.4.18). Apparently benign Roman interest in the border issues between Cretan poleis increased markedly after 113 BC (Chaniotis 1996, 329–30). Cretans may have had intimations that this power was of a different order than the Hellenistic kingdoms of Egypt, Syria, Macedon and Pergamon, powers which Cretan cities had hitherto been able to ignore.

CRETAN POLITICAL COMMUNITIES COMPARED: STATES IN THE SECOND AND FIRST MILLENNIA BC

When did Iron Age political communities emerge on Crete? Saro Wallace (2010a; 2010b) has suggested that they emerged soon after the major phase of settlement nucleation on the island in the middle of the tenth century BC, and Anna Lucia D'Agata (2012) has argued (on different grounds) that ‘warrior citizens’ in ‘proto-states’ emerged at least around Sybrita at this time. The emergence of strong corporate groups can be detected in the burial record of Knossos around 900 to 850 BC – these groups persisted until the ‘Archaic gap’ of the sixth century (Wallace 2010a; Whitley 2023, 139–45). All in all, there is a strong case for political communities with a corporate ethos to have been in existence well before the eighth century. That so many of them managed to persist not only into Hellenistic times but beyond is an indication of their resilience.

Years ago, Runciman (1990) argued that ‘the polis’ was an evolutionary dead end. Runciman defined polis as I am defining it, as a citizen-state, and it was an ‘evolutionary dead end’ in the sense that it failed to adapt (and so persist) in the changed conditions of the third and second centuries BC (by which Runciman means the rise of Rome). In the sense that, eventually, there were no more poleis by (at the very latest) the third century AD, Runciman's statement is certainly true. I am, however, here talking not so much about evolutionary adaptation as resilience. Hellenistic kingdoms (which lasted barely 300 years) were certainly no more successful in ‘evolutionary’ terms than the citizen-states of the Greek-speaking Mediterranean world; citizen-states did not simply disappear with the advent of these kingdoms (Ma 2009; Ober 2015). Judged simply by their longevity the poleis were both successful and resilient. Longevity is moreover not the only measure of their resilience – poleis remained very difficult to eradicate as political communities. They had a tendency to reappear, if given half the chance.

This reappraisal of Iron Age and later Cretan citizen states should be seen in the broader context of Cretan history, that is, in relation to a thoroughgoing reappraisal of that earlier form of Bronze Age political community, the so-called Cretan palaces (Hatzimichael and Whitley 2012; Whitley 2023, 79–127). Cretan (unlike mainland Mycenaean) palaces lasted for a very long time, taking shape around 1900 BC – or perhaps even earlier. If Peter Tomkins is right, the central courtyard in Knossos takes shape around 2700 BC.⁶⁷ These structures – which in their earlier phases are rightly referred to as courtyard complexes, following Jan Driessen (2002) – go through major phases of rebuilding. At Knossos (following Tomkins) the first of these is in 2700 BC, the next around 1900 (Middle Minoan [MM] IB), and the next in the middle of MMIII, a rebuilding of a structure conceived as a whole and centred on processions and cult. In some of these earlier phases, especially MMII, these central courtyards and west courts seem to have been the focus of major feasting events undertaken by constituent corporate groups (Macdonald and Knappett 2007, 57–68, 161–65; Whitley 2023, 81–8). Similar corporate groups making their presence felt through feasting have been detected by Donald Haggis (2007) around Petras in Eastern Crete. It is only from MMIIIB that the palace of Knossos begins to look like a palace. What happens in Late Minoan (LM) IB remains controversial (Driessen and Macdonald 1997, 105–15, esp. 108–9). It is only in its final phase (LMII–LMIIIA2) that we know that Knossos was the residence of a ruler and the centre of administration of a unitary state (Bennet 1990).

This essentially Mycenaean monarchical state did not last long (little more than 100 years). The destruction of (in my view) the only true palace in the sequence of rebuildings of the central courtyard structures at Knossos must have been quite a spectacle – it has left traces of extensive burning that were obvious to the early twentieth-century excavators of the site (Evans 1900). These traces are the very opposite of the elusive destruction horizons that Karkanas (2021) has documented for many mainland cities that were ‘razed’ (κατέσκαψαν) in historical times. Once destroyed, the final palace at Knossos was not rebuilt. Our understanding of the relationship between Bronze Age and Iron Age states in the Aegean has been bedevilled by the terms ‘palace’ and ‘polis’. But the palaces were not real palaces: Cretan protopalatial and neopalatial ‘palaces’ such as those to be found at Phaistos, Mallia and Zakro were never the residences of rulers; and Mycenaean palaces such as Pylos lacked the staying power of their Near Eastern counterparts. Instead they were, in Susan Sherratt's (2001) terms, Potemkin palaces, and (on Crete) lasted for about as long as the original Potemkin villages. Both before these palaces and after them, there were on Crete states based on communal religion and composed of corporate groups defined by seasonal gatherings (if not actual feasting). It is to the similarities between these two forms of Bronze Age and Iron Age political community (rather than the misleading contrast between ‘palace’ and ‘polis’) that research should now turn.