The Ancient Greek Pharmakos Rituals

This article examines the role of mistrust – especially in constructions of purity, impurity, and purification  – in ancient Greek religion. It begins by examining so-called scapegoat or pharmakos rituals, in which an individual was expelled from the city, apparently as a purificatory offering to the gods. Recent analyses have argued that these rituals were outlets for community aggression, and/or were resonant with myths of self-sacrifice. This article will suggest a different analysis of the evidence. I offer an alternative way of interpreting these rituals that sets them in a wider context of Greek ritual and belief: it suggests that the ritual of the pharmakos arose in a context of social and spiritual insecurity. This created, I argue, a prevailing dynamic of social and spiritual mistrust, within which the pharmakos ritual emerged – and which it exacerbated.

1 One notable period of scholarly focus, at least in the Anglophone world was the early 2000s, as noted by Hosking, who begins his Introduction "In the western world there is a crisis of trust " (2006: 95). As he points out, citing Onora O'Neill's Reith lectures (2002: 19): "Perhaps claims about a crisis of trust are really evidence of an unrealistic hankering after a world in which safely and compliance are total, and breaches of trust are totally eliminated." Thus, there are four key parts to this article. The first section explores the evidence for pharmakos rituals, evaluating scholarly approaches that assimilate these rituals to either killing and (self-)sacrifice or the expulsion of demons. In the following section, examination of evidence for the pharmakos ritual at Athens leads to the suggestion that pharmakos rituals were intended to cleanse the community of sources of impurity that were hidden within it. Underpinning that sense of hidden threat, I argue, was a crucial lack of secure knowledge, of two key types. The first type of insecure knowledge is addressed in the third section of this article, which builds on the perspectives of two key theorists -Julia Kristeva (1982) and Michel Foucault (2013) -to explore the role of different forms of knowledge in the development of concepts of purity and impurity, arguing that it was the frailty of mortal knowledge, and the consequent social dynamics that it generated, that allowed a threat of impurity to develop -and that this reached beyond a sense of external disorder. This then leads to the final section, which examines the second type, which arises from the implications of the frailty of mortal knowledge in a larger context of ancient Greek religion. This focuses on the idea of the unknowability of the gods: it suggests that "spiritual insecurity" (e.g., Ashforth 1998) gave rise to a crucial dynamic of "mistrust" in relations between mortals and gods, which ritual activity was intended to, but did not, ameliorate. In this context of social and spiritual mistrust, it concludes, the ritual of the pharmakos emerged, both a product and a component of this prevailing dynamic.

"Scapegoats"
There is evidence for ancient Greek rituals that involved the expulsion of a pharmakos being held regularly at Abdera, Athens (at the Thargelia festival), and Leukas; but the sources also note ceremonies held in response to specific crises.2 In general, in scholarship, these rituals are often called "scapegoat rituals" (e.g., Parker 1985: 257-270), but, as Mary Douglas (2004: 40-41) among others has shown, the original scapegoat of Jewish ritual had little in common with the "scapegoat" of Greek scholarship insofar as the latter has come to develop a very specific set of associations with persecutory behaviors of shame, blame, self-sacrifice, and death.3 This has not prevented the term from being widely used, along with a persistent strand of scholarship that has argued that the scapegoat of ancient Greek ritual was, on the one hand, the "scum" of the city, in a ritual that evoked a human sacrifice that, on the other hand, was made explicit in myths that described kings or virgins who were (often voluntarily) sacrificed for the city.4 These arguments draw on theories that emphasize the fundamental violence of humankind and concomitant role of sacrifice as a form of release for that violence.5 Réné Girard (1977) argued that the scapegoat ritual was an outlet for the cycle of interhuman aggression that, unless released by a death, led to an endless cycle of violence. In his work on ancient Greek religion, Burkert (1979Burkert ( : 64-66, 1996) also built on the notion of innate human violence, emphasizing the killing of animals in the ritual of sacrifice.
While highlighting the pharmakos ritual as one of purification, he discusses the scapegoat as a sacrificial victim, and the ritual as a result of "aggression excited by fear" (1985: 83). Examination of the ancient evidence, however, raises some questions about the basis of these theories and the ways in which they have been developed. The only classical ancient Greek source that refers explicitly to the status of the pharmakos is a comment by the Chorus in Aristophanes's Frogs: Just so with our citizens: the ones we acknowledge to be well-born, wellbehaved, just, fine, and outstanding men, men brought up in wrestling schools, choruses, and the arts, we treat them shabbily, while for all purposes we choose the coppers, the aliens, the redheads, bad people with bad ancestors, the latest arrivals, whom formerly the city would not readily have used even as pharmakoi. 727-733, trans. Henderson 1998 3 The scapegoat of Jewish ritual "is not attacked, shamed or harmed" as Douglas argues (2004: 41). She blames the elision on Réné Girard's interpretation of the pharmakos, and his argument that "slaughter, blood, and killing have always been the central elements of religion." But as this article argues below, the evidence indicates that this was not the focus of the pharmakos ritual either. 4 Bremmer 1983: 307 = Bremmer 2008: "Summing up, we conclude that in historical reality the community sacrificed the least valuable members of the polis, who were represented, however, as very valuable persons. In the mythical tales one could pass this stage and in the myths we always find beautiful or important sons, although even then these scapegoats remain marginal figures: young men and women, and a king." 5 Graf 2012 offers a useful overview. This is clearly Aristophanes commenting on contemporary political events. For the purposes of this article, we can note that it provides a reflection on the nature of those chosen to be a pharmakos: the Chorus describes the physical attributes, specifically of noncitizens, as well as referring to those with bad moral character. It is, in contrast, the scholia to these lines that provide a far more explicit and detailed description of displeasing physical characteristics of the victims, describing them as "plotted against by nature," which presumably means ugly or somehow disabled.6 Similar attention to this kind of detail is found in two other dramatic scholia, appended to lines of verse that do not refer to pharmakoi. For example, Aeschylus's Septem l. 680 is about the expiation of bloodshed in battle as opposed to the pollution arising from family members killing one another. The scholion's language recalls the scholia to Aristophanes's Frogs in its addition of details about the pharmakos as "distasteful, maimed and lame."7 In Aristophanes's Knights the Chorus describes the treatment of Athenian politicians: "Your ministers, then, are your victims (demosioi), whom you nourish and feed up expressly in the Pnyx, so that, the day your dinner is ready, you may immolate the fattest and eat him" (1135( -1140( , trans. E. O'Neill 1938. It is not necessary to accept that this refers to pharmakoi, and, indeed, the specific term demosioi is understood by two of the scholia as referring to animals sacrificed by the city: one adds the alternative of the pharmakos, one refers only to the pharmakos (see Hughes 1991: 149-150). Again, these scholia add specific information about the low social status of the pharmakos, in this case describing him as "exceedingly low-born, penniless, and useless."8 Other sources may provide earlier or more specific evidence that could explain the scholiasts' focus on status. The source describing the ritual at Abdera notes that "a bought person" (ὠνητὸς ἄνθρωπος) was used for the ritual. It is not clear whether that phrase refers to an enslaved person or could indicate a person who has offered to play this role for some reward, as appears to be the case in the ritual at Massilia, where the individual is specifically aliquis de egentissimis -"someone from among the poorest" -who volunteered to be fed for a year before being expelled from the city.9 At Leukas, also, the status of the person chosen is unclear: they are described as τινα τῶν ἐν αἰτίαις ὄντων, which is translated by H. L. Jones as "some criminal." But the Greek is 6 Schol. ad Ar. Frogs 733a: φαύλους καὶ παρὰ τῆς φύσεως ἐπιβεβουλευμένους. 7 Schol. ad Aesch. Septem 680 l (there are ten other scholia listed for this phrase, none of which refer to pharmakoi). 8 Schol. ad Ar. Knights 1136a and c. 9 Petronius fr. 1; Lactant. on Stat. Theb. 15.793;Schol. Luc. 10.334. more puzzling than that: it seems rather to identify "one of those who are held responsible" or one who is "being blamed."10 The multiple descriptions of the Thargelia festival in Athens do not mention the status of those chosen for the ritual, although one source does emphasize ugliness as a criterion for their selection.11 Thus, the arguments for the status and nature of the pharmakos rests largely on the evidence of scholia that were composed far later than the presumed dates of the rituals themselves. The statements tend to be attached to unconnected phrases, to be very general with little indication of what ritual is being described, and to use very similar phrasing: none of this inspires confidence.
They also describe extreme events: the scholia to Frogs, Knights, and Seven Against Thebes go on to mention the sacrifice of the pharmakos and how this cleanses the city. Most scholars have rejected the possibility that humans were sacrificed in ancient Greece, and I will not rehearse their arguments here (see, for example, Hughes 1991;Bonnechere 1994). Scholars have explored the kinds of associations that may have given rise, over time, to these extreme versions of the ritual. For example, as Dennis Hughes (1991: 142-143) has illustrated, Tzetzes's assertion (Chil. 5.735-6) that the pharmakos was burned on a pyre and his ashes scattered in the sea is likely to have developed in response to other rituals for destroying potentially polluting objects.12 In a couple of examples, it is possible to see how a later commentator has "improved on" an original version of the ritual. For example, until 1934, based on the evidence of some verses of Ovid's Ibis and the detailed scholia, it was thought that the scapegoat ritual in Abdera involved the victim being stoned to death.13 But a papyrus with a fragment of Callimachus was found that stated unequivocally that the scapegoat was chased over the border with stones rather than being killed.
In Abdera a bought person, who is used as a means of purification, stands on a piece of grey brick and enjoys an abundant meal, and when he is completely full he is brought forward to the so-called Prurian Gates. After that he goes round in a circle outside the wall thereby purifying the town 10 Jones 1928. Radt (2004  Cf. Bonnechere 1994. 13 Ov. Ibis 467-8 and schol. ad loc. = Call. fr. 90b (Harder). on all sides and then he is stoned by the king and the others, until he is driven from their territory.
Call. fr. 90a (Harder) We can compare the sources for the scapegoat ritual in Massilia: during a plague, a man offered himself to play this role; he was feasted for a year then cast out of the city.14 It is only according to later scholia, on Statius' Thebaid, that he was stoned to death.15 The idea that the pharmakos must die, although it appears to be a later development of Byzantine scholia, has nevertheless been read back into concomitant evidence, supported, as noted above, by theories of human behavior. Walter Burkert briefly mentions those myths in which kings (Kodros, Oedipus) save a city that is suffering misfortune, summarizing: "The pharmakos ritual turns out to be equivalent to a king's tragedy" (1979: 65).16 He then identifies myths that concern the sacrifice of virgins as "tale types" based on the same "scapegoat pattern."17 As already noted above, other scholars, perhaps drawing on the scholia, have built on this approach to argue for a more specific relationship, in which the myths of savior kings and/or their daughters sacrificing themselves in return for the salvation of a city were the underlying, extreme model, which the pharmakos ritual (employing poor, ugly or, maimed individuals) evoked.18 Yet the rituals themselves -according to our sources -were not explicitly or even implicitly related to stories of the deaths of kings or sacrificed virgins. Instead, tracing the association of ideas, a process which has helped to explain the scholia's interpretations, may also be useful in the case of these more recent explanations. It is possible to identify an association of story types that resonate with modern cultural models of human behavior. Indeed, alongside his observations about primordial human behavior, Burkert explicitly discusses how apparently related story patterns were associated, although he is not always clear about how those associations can be made.19 Similarly, Robert Parker analyses "the ideas that emerged in connection with the ritual and mythical scapegoats," observing for example that there is no secure evidence that "the sacrifice of the innocent was anything more than a traditional legendary motif " (1985: 264). In contrast, he sees plenty of evidence for the story type of the sacrificed leader, noting that, in every period of Greek history, the figure of the leader was to blame (1985: 265). This is, of course, a phenomenon that is not limited to ancient Greek culture: the human brain seeks out patterns in the world.20 These patterns can, in turn, entangle with and reinforce each other: for example, the sacrifice of the innocent and the sacrifice of the leader may be seen to come together in the self-sacrificing martyr, which can be understood as "a post-Biblical interpretation of sacrifice" that resonates with Christian understanding of self-sacrifice (Janowitz 2011).21 A similar process of entangled patterns may help to explain another approach to the scapegoat, put forward by, among others, Christopher Faraone (2004). Drawing on Burkert's work, Faraone argued that "scapegoat ceremonies are, in fact, merely a subset of a larger group of Greek rituals designed to expel or drive out evil demons or pollution" (2004: 216).22 Faraone's argument for associating a demon expulsion ritual and a pharmakos ritual leans heavily on a passage that Burkert also mentions in this context: according to Plutarch (Quaest. conv. 6.8.1), a ritual took place at Chaeronea in which the demon Boulimos ("Famine"), represented by a slave, was chased away with rods of the 19 Burkert 1979: 75: "Somehow related are all those tales in which a girl is handed over as a bride to some god or demon … with complex rituals of girl sacrifice in the background." 20 This is part of predictive processing: patterns are used not only to process incoming information but also to predict what will happen next. See Lynn et al. 2020, which draws attention to the way in which, during this process, the brain balances complexity and accuracy. 21 Bonnechere 1994 suggests that later sources were influenced not only by the assimilation of initiatory and salvation myths, but also by the similarities of these stories to Christian devotion. 22 Burkert draws attention to this ritual (1979: 65) and notes: "Myth may introduce demons who must be pacified" (72). The parallel of the pharmakos ritual to the expulsion of a demon is also explored in Versnel 1977: 41-42. agnus castus. But while there is no doubt that there were civic rituals both for the expulsion of demons and for the removal of pharmakoi, examination of the details of each ritual type suggests that there were significant differences between them. As Faraone has himself acknowledged, in Plutarch's account the expelled person is described as "the one who is pursued," not as a pharmakos, and the ceremony is referred to as a "traditional sacrifice."23 This may seem irrelevantafter all, the same can be said of the pharmakos ritual at Leukas (which is described as "the annual sacrifice in honor of Apollo");24 but there are further inconsistencies in the case of the ritual at Chaeronea. The ritual is clearly meant to drive the slave out of a building and the slave is identified with a specific abstract quality. Neither of these occur in evidence for other pharmakos rituals, where the individuals are driven out of the city and are significantly mortal (as well as being described as taking on some kind of responsibility). The other parallels suggested are also not convincing. For example, Faraone suggests that the flee formulas on gemstones, which were intended to repel demons, were similar to pharmakos rituals. But the rings' formulae make no explicit mention of purity and pollution, which are the concern of pharmakos rituals; while the pharmakoi rituals do not use the terminology of fleeing that he highlights on those gems (2004: 216-218). He argues that we find politicians abused as both pharmakoi and demons, suggesting this indicates an equivalence (ibid.: 239). But while these may both be relevant ways of abusing politicians and highlighting their bad qualities, it does not mean that demons are the same as pharmakoi in a ritual context. As Parker notes, when a politician attacks his rival as a pharmakos, "he is primarily denouncing his opponent's policies and their practical consequences, but also seeking to suggest that with such an impure rogue in charge, afflictions of every kind are likely to follow" (1985: 268).
Faraone's argument certainly highlights a basic paradigm of removal -a high-level story pattern of "removing something harmful from somewhere."25 And while it does not indicate that the pharmakos was a demon, it does draw attention to another aspect, perhaps more useful for understanding the 23 Faraone argues that the best known of such expulsion scenes is in Euripides' Alcestis, in which Herakles confronts Thanatos ("Death") when he comes to take the queen away, but to describe this as the expulsion of a demon is surely to overlook the intended magnitude of Heracles' achievement; there is no evidence that it should be viewed as a pharmakos ritual. 24 Leukas: Strabo 10.2.9, Ampelius 8. 25 And this makes it difficult to see how the two ideas -that demons and pharmakoi were related, and that pharmakoi rituals have a mythic counterpart in stories of self-sacrificial kings and high-born women -align. pharmakos ritual: that is, that the pharmakos was somehow representative of harm. In what follows, I will explore this idea in the evidence for the pharmakos ritual in Athens at the Thargelia festival.26

Reading the Evidence
It is unassailable that the evidence for pharmakos rituals indicates that they represented a cleansing of mortal wrongdoing that had posed or might pose a problem for the community in its relationship with the gods. This association is made in the earliest references to the pharmakos, in surviving fragments of the sixth-century poet Hipponax of Kolophon, a blame poet who, like many of his genre, used the imagery of the pharmakos to describe his own enemies, and those of his city.27 What kind of wrongdoing this might be, and what kind of person was responsible for it, can, I think, be elicited from the aition of the pharmakos ritual at the Thargelia at Athens as recorded by Harpocration, drawing on Istros, which apparently commemorated the story of the original Pharmakos.28 Pharmakos: Lysias in his speech Against Andokides on the charge of impiety, if authentic. Two men at Athens were driven out of the city as katharsia at the Thargelia festival, one of them representing the males and one the females. The fact that the proper name is the Scapegoat, who, after stealing the sacrificial bowls of Apollo, was caught and stoned by the followers of Achilles, and that the expulsion of those (sc. the men) at the 26 I have selected this ritual because it provides specific evidence for analysis. That does not mean that I am assuming that all pharmakos rituals in all communities were the same or were conducted for the same reasons. Nevertheless, as I will go on to argue, it seems to me that a reinterpretation of the Athenian ritual raises questions in turn about the traditional interpretations of the other rituals. 27 Hipponax fr. 5-11 (West  Bremmer (1983Bremmer ( : 303n26 = 2008 gives Neanthes FGrH 84 F 16 and Diog. Laert. 1.110 for other aitia of the Thargelia, which both tell the story of the purification of Athens by the Cretan seer Epimenides by means of the deaths of young men. Neanthes reports that one man willingly sacrificed himself followed by the death of his lover; Diogenes Laertius describes two men being put to death. Neither account mentions the Thargelia or pharmakoi, or identifies these events as the origin of the festival or ritual.
Thargelia festival are a remembrance of this, Istros has said in the first book of his Apollo's Epiphanies.29 The account describes how Pharmakos was caught and stoned to death by the followers of Achilles after he stole the sacrificial bowls of Apollo. This raises the question of whether we should regard this Pharmakos as the original pharmakos. Some have assumed so (e.g., Hughes 1991: 152;Burkert 1979: 72;Bremmer 1983: 308), but the details of the ritual do not seem to map neatly onto this explanation.30 On the one hand, it is because of the killing of Pharmakos that the city is endangered. On the other hand, at the festival itself, two men, not one, were expelled. The obvious question this leads to is whether the ritual pharmakoi represent Pharmakos, or whether, in fact, they represent those who committed the crime, the followers of Achilles.
A further aetiological story about the Athenian pharmakoi by Helladius may offer further insights.31 Because it was the custom at Athens to lead two pharmakoi … This cleansing served to ward off plagues of disease, and it took its beginning from Androgeus the Cretan, because the Athenians were afflicted with a plague of disease when he died unjustly in Athens, and this custom began to be in force, to always cleanse the city with pharmakoi.32 This story has also been interpreted as meaning that Androgeus, through being murdered, became the "scapegoat hero," leading to the instantiation of the pharmakos ritual in Athens to purify the city after his death.33 But the 29 Φαρμακός· Λυσίας ἐν τῶι Κατ᾽ ᾽Ανδοκίδου ἀσεβείας, εἰ γνήσιος. δύο ἄνδρας ᾽Αθήνησιν ἐξῆγον καθάρσια ἐσομένους τῆς πόλεως ἐν τοῖς Θαργηλίοις, ἕνα μὲν ὑπὲρ τῶν ἀνδρῶν, ἕνα δὲ ὑπὲρ τῶν γυναικῶν. ὅτι δὲ ὄνομα κύριόν ἐστιν ὁ Φαρμακός, ἱερὰς δὲ φιάλας τοῦ ᾽Απόλλωνος κλέψας καὶ ἁλοὺς ὑπὸ τῶν περὶ τὸν ᾽Αχιλλέα κατελεύσθη, καὶ τὰ τοῖς Θαργηλίοις ἀγόμενα τούτων ἀπομιμήματά ἐστιν, ῎Ιστρος ἐν ᾱ τῶν ᾽Απόλλωνος ᾽Επιφανειῶν εἴρηκεν. 30 Parker (1985Parker ( , 1996Parker ( , 2005 does not discuss the origins of the Athenian pharmakos, but he notes that Jacoby (on Istros FGrH 334 F 50) questioned whether Istros described the Attic rite, because of a number of inconsistencies, including "the singular number φαρμακός" (2005: 482). For this reason, he suggests that this is an aition for an unknown Ionian city combined with an account of the Attic rite; cf. also Bremmer (1983Bremmer ( : 308 = 2008 text describes the pharmakoi of the ritual in the plural. In turn, this raises the question of whom they represent, and again, it seems possible that the answer may be that rather than representing Androgeus, who is the innocent murdered party, the pharmakoi represent the two impure parties: the killers of Androgeus. Together, these etiological stories of plural killers, coupled with the duality of the pharmakoi in the ritual, cast the aims of the ritual in a new light: rather than evoking the sacrifice of a single individual on behalf of the community, the ritual plays out the cleansing of the city of the impurity of those individuals who, through their actions, were likely to have endangered it and its inhabitants. It is understandable that this ritual attracts parallels with deaths of kings and expulsions of demons. The pharmakos possesses a dual role that can be seen in the language used to describe it. On the one hand, the removal of the pharmakoi explains why we find the pharmakoi referred to as "offscourings" or katharmata (or, at Leukas, "be thou our peripsema!").34 When the fifthcentury bce comedy writer Eupolis sets it in the mouth of an unknown character in a fragment that refers to the degeneration of political figures ("we elect as our generals the katharmata") he is using it to refer, literally, to what is left once something has been cleansed.35 The same is the case when the fourthcentury bce orator and politician Demosthenes uses it of his political rivals Meidias or Aeschines, or when Aeschines uses it of Demosthenes (Dem. 21.185,198,and 18.128. Aesch. 3.211). But, on the other hand, as the Suda notes of the related term peripsema, this means both what is thrown away, and, in the ritual context, it also represents "salvation and redemption."36 This dual role, this ambiguity, is employed for comic effect in Aristophanes's Knights (1402-1408), when Demos uses the term to describe the seat to which he invites the Sausage Seller. The seat is an honor, a special dining privilege accorded both to those who brought great honor to the state and also to the pharmakoi driven out of Athens during the Thargelia. While the double-edged nature of the offer is funny, it underlines the ambiguous nature of the pharmakos.
This quality extends beyond the figure of the pharmakos to the ritual itself, which exposed the harm hidden within the community. This can be seen in the aitia of the Athenian Thargelia: while the murder was the crime that was committed in each case, for the purposes of the ritual, the significant aspect was the nature of these offences, which remained concealed by an apparently good deed. In the story of the original Pharmakos, his murderers seemed at first to be justified in their actions; what followed revealed that they were wrong. Similarly, the death of Androgeus seemed at first to be righteous, but turned out to be so wrong that it endangered the city of Athens.37 That this idea of a harm hidden in a good deed was widely understood as part of the meaning of the pharmakos can be seen in the use of this term in the lawcourts. When Lysias (6.53) describes Andokides as a pharmakos who must be expelled, he is describing how Andokides's actions, which might otherwise seem the height of piety, are in fact impure and therefore endanger the city. Similarly, it is this aspect of being harmful, of threatening potential damage that Demosthenes invokes when he describes the brother of Aristogeiton with this noun (Dem. 25.80. cf. also Ar. Fr. 655 in Henderson 2008); he has paid some of Aristogeiton's debts, and his brother holds him responsible for the rest (for which Eunomos is now prosecuting Aristogeiton), but he is himself a very shady character.38 While we only have sufficient evidence to examine the Athenian Thargelia in detail, I would like to raise the possibility that this reading may also deepen our understanding of the rituals held in other cities. We have seen above that, while the accounts of the pharmakos rituals indicate a process of purification, they seldom explain the specific offence that may have caused an impurity. It is possible that some pharmakos rituals may have involved the assimilation of some kind of crime to the pharmakos, justifying their expulsion. This might explain the phrase used of the pharmakos at Leukas (Strabo 10.2.9), which seems to allude to an undefined responsibility. And, indeed, the references to the individual being ugly and low-born, which are found in later sources, may be interpreted as post-hoc details that were added in order to fill this apparent gap.39 But I want to suggest that instead of being concerned with the impurity 37 May have been envious competitors: Apollodoros 3.15.7; killed in battle: Hyginus Fables 41; killed by Aegeus, king of Athens, who sends Androgeus out to face the bull of Marathon, but the hero is killed by the monster: Apollodoros 3.15.7. 38 The idea that one could -indeed, in certain circumstances, one must -be mentally pure can be argued to be a refinement of a long-held idea that a person's "impurity" could be concealed and might well not be known about until it was too late. 39 This notion of assumed impurity may lead to another explanation for the necklaces of figs (of white and black), in the Athenian Thargelia festival; and the figs are also noted as part of the pharmakos ritual described by Hipponax (Hipponax 5-11 West). Scholars have argued that these may have indicated the final aim of the ritual of ensuring future fertility for the city: Lin Foxhall (1995: 109n32) has observed that the modern custom of throwing necklaces of "wild" male figs on to fig trees to fertilize domestic figs is matched by similar practices from antiquity (Theophr. Causes of Plants 2.9.5-15 and History of Plants 2.8.1-3). But if we accept the evidence that there was a law against the smuggling of figs outside created by a specific offence or type of crime, the rituals of the pharmakos were deliberately ambiguous. They represented the salvation of the city from a far more frightening phenomenon: the potential threat to the community that was hidden within it, the apparent purity/lack of a crime that was in fact impurity/a concealed offence against the gods, which must be removed to restore the safety of the city. Underpinning that sense of hidden danger, I argue, was a crucial lack of secure knowledge of two key types. The first type, addressed in the next section, concerned the potential dangers that could threaten oneself and one's city; these included, importantly, knowledge of what or who was impure. The second, which will be addressed in the final section, started from the unknowability of the gods.

Impurity and Uncertainty
Recent work on purity and impurity is moving its focus from earlier formulations that turned on strict categorizations -Mary Douglas's "matter out of place" or "disorder" (1966: 84) -to more fluid notions of purity and impurity as "flexible metaphors" (Carbon 2018: 20), with particular focus on the individual, and the perceived relationship of body and mind (Carbon and Peels-Mathey 2018; Petrovic and Petrovic 2016). Such approaches have explored the development of the notion of the impurity of the mind in addition to that of the body. Angelos Chaniotis (2018: 41), for example, argues that this developed in parallel with and was prompted by changes in secular law, specifically laws of homicide (an idea that also occurs in the theories of Michel Foucault, as discussed below). Building on these formulations, I want to suggest a slightly different perspective, which approaches these concepts not only in terms of the development of ideas of purity, but also in terms of their underlying social and supernatural relationality, focusing particularly on the role, and nature, of knowledge. To guide this approach, I will draw on the studies made by Julia Kristeva and Michel Foucault into the origins of the idea of the sacred, and the development of notions of purity and impurity. Foucault's examination is historical and considers sociopolitical dynamics. Kristeva's explorations take a psychoanalytical approach, and focus on the subjective origins of these categories.
Athens, it may be that the necklace of figs that adorned the pharmakoi as they were expelled from the city ensured that that individual was clearly responsible for an offence in that moment of expulsion.
Numen (2022)  In Foucault's lectures on "The Will to Know," written in 1970/71, he performed an analysis of discursive events to chart how "the truth founds its site of emergence, function, distribution, and necessary form in Greek society" (2013: 56).40 He argued that changes in political structures led to a juridical definition of the individual: this was achieved via control of legislation concerned primarily with the event, and consequences, of death, accompanied by changes in forms of worship, which became dominated by the State. Through these processes, the State gained control over its citizens, and also claimed control over the concepts of purity and impurity (ibid.: 167-175). The idea that a crime (such as homicide) could cause an impurity in an individual emerged, and alongside it the importance of identifying and excluding that individual.41 But as Foucault emphasizes, there is no clear statement in a Greek text that indicates how impurity is transmitted or what effects it has (ibid.: 180). The State instead placed emphasis on the idea of truth and the establishment of fact: truth allowed the creation of boundaries between what was pure and impure, and the exclusion of the impure from the space and nomos (laws and customs) of the city. It was, thus, not the fact of impurity that created the practice of exclusion, rather, the practice of exclusion was constitutive of the pure -impure division in Greek practice (ibid.: 180). In this way, the relation between impurity and the nomos of the city was established: the individual who committed an impurity must have done so because they were ignorant of or blind to the nomos of the city -and to be so excluded, they must already have been impure (188). We can see something of this dynamic in the preceding case study of the pharmakos, where no specifics about the crime -no definition of impurity -seems to have been provided as part of the ritual. Rather, the idea of impurity itself appears to have been generated through the very process of the ritual.
To illustrate his approach, Foucault turns to the story of Oedipus, which he sees not as an exploration of desire, as Freud does (1997: 155-158), but rather as indicating "the historical constraints of our system of truth" (Foucault 2013: 192). The fable of Oedipus reveals how the city, through its juridical and social processes, achieves purification by establishing truth, "posited as visible, ascertainable, measurable, compliant with laws similar to those governing the order 40 I am guided by the work of Adey and Duschinsky 2014, in particular their reflections on Foucault's "The Will to Know" lectures. 41 He summarizes: "The juridical-religious superimposition of crime and purity entails a new relationship to the truth. In fact: α -impurity is now a quality of the individual constituted by the crime; β -this impurity is the source of dangerous contacts which spread throughout the space of the city; γ -it is therefore important [sic] know if the crime has been committed and by whom" (Foucault 2013: 198). of the world" (ibid.: 213). But it also -and importantly for this article -shows the multiplicity of forms of knowledge involved, and the different procedures for acquiring them. Oedipus is aware of divine knowledge, the knowledge held by his equals, and the knowledge of common people and slaves. As Foucault emphasizes, the problem for Oedipus is that he does not trust any of these; he only trusts himself. And it is in insisting on his own inquiry that he discovers the truth: "Oedipus is forced to hear what he did not want to listen to, to see what he did not want to see" (ibid.: 253). Oedipus conforms to Foucault's image of the impure individual, that is, the individual who was impure because he was ignorant of the city's nomos, and was ignorant of the nomos because he was impure.
Foucault's analysis concentrates on how discourses of purity and impurity become institutionally embedded as discourses of power, but he does also briefly allude to the experience of the individual. This aspect is the focus of Julia Kristeva's approach, which examines the individual psychology underlying the concepts of purity and impurity. Whereas for Foucault, impurity is a barely identified notion against which purity can be categorized, for Kristeva, purity and impurity discourses are rooted in a subjective experience, which she calls "abjection," which is part of an individual's psychosexual development. As Kristeva puts it, abjection "preserves what existed in the archaism of pre-objectal relationship, in the immemorial violence with which a body becomes separated from another body in order to be" (1982: 10). The experience of abjection thus evokes the breakdown of meaning that threatens when there is any loss of the distinction between subject and object.
While it relates in the psychoanalytical context to separation from the mother, and while recognizing the entanglement of this experience with the feminine, Kristeva stresses how abjection arises from "a coming face to face with an unnamable otherness" (1982: 59), which we both fear and are drawn towards. She gives, as one example, the traumatic experience of seeing a human corpse. It confronts a person with their own materiality, their own mortality. They must accept it as the body they possess, but also thrust it aside, repress knowledge of it, in order to live: "The corpse, seen without God and outside of science, is the utmost of abjection. It is death infecting life. Abject" (ibid.: 4). As this suggests, for Kristeva, the abject is "what disturbs identity, system, order. What does not respect borders, positions, rules" (ibid.: 4), but it is also "something rejected from which one does not part, from which one does not protect oneself as from an object … it beckons to us and ends up engulfing us" (ibid.: 4-5).
Kristeva explores a number of sources (the Hebrew Bible and New Testament) as well as different cultures (India, New Guinea) to chart the ways in which communities have identified and dealt with the abject, establishing reactions of disgust and rites of defilement and pollution that purify through procedures of exclusion or taboo (1982: 17). As Kristeva herself emphasizes, while her approach to purity and impurity shares with the theories of Mary Douglas and others a concern with "matter out of place" and "by-products of an ideal of order,"42 her insight into abjection deepens our understanding of the subjective experience that gives rise to these concepts (ibid.: 65-67). It draws attention to the threat not only from outside oneself -lack of cleanliness or health -but also from "what disturbs identity, system, order" (ibid.: 4). The danger is not (only) to categories, but more disturbingly, to one's very sense of self.
This aspect of subjective experience adds a further dimension to this article's analysis of purity and impurity. Where Foucault evokes the story of Oedipus to explain how different forms of knowledge shape the power discourses about purity and impurity, Kristeva explores Oedipus as a figure of abjection (1982: 84): his ejection from the city allows the city to survive. But knowledge -in its different forms -plays a significant role here, too. Oedipus's abjection "is due to the permanent ambiguity of the parts he plays without his knowledge, even when he believes he knows." We have seen a similar ambiguity in the preceding case study of the pharmakos, who plays the role of both savior and offscouring. What Kristeva says next about Oedipus resonates further with this reading of the pharmakos ritual: "The mainspring of the tragedy lies in that ambiguity; prohibition and ideal are joined in a single character in order to signify that the speaking being has no space of his own but stands on a fragile threshold as if stranded on account of an impossible demarcation" (ibid.: 84-85). Similarly, knowledge of the pharmakos can never be secure: by its very nature it retains this ambiguity, this impossibility of definition.
This emphasis on the frailty of human knowledge, and its implications, which emerges from both Foucault's and Kristeva's theories, has implications for the context of human relationality in which the pharmakos ritual took place. In both cases, we see how different forms of knowledge -a lack of secure knowledge -both about others and about oneself, maintains a constant risk that impurity will emerge. In Foucault's account, the historical dynamics he describes depend on a tacit assumption by the State that its citizens cannot 42 Douglas's argument that the idea that pollution was "matter out of place", implied two important conditions: "a set of ordered relations and a contravention of that order" (1966: 44). Parker, although not specifically citing Douglas at this point, apparently builds on this argument when he argues that "a culture's beliefs about pollution … are rather byproducts of an ideal of order" (1985: 326).
know each other; indeed, in the case of Oedipus, that assumption extends even to him. And, in this respect, Foucault's approach resonates with the more subjective focus of Julia Kristeva. In her approach, I believe we find a similar dynamic: a lack of knowledge not only of others, but also, importantly, of oneself. As Kristeva portrays it, Oedipus is a figure of ambiguity: he cannot know others, he cannot be known, and he also cannot know himself. In such a situation, it might be argued, rites of purification could provide some sense of security. But here I part company with Foucault and Kristeva to argue that this was not the case. The structure of ancient Greek relations with the gods meant that such security remained always somehow just out of reach. In the next section I turn to the nature of relations between mortals and gods in the context of ancient Greek religion, and to the key dynamic of mistrust.

Spiritual Insecurity and Mistrust
This final section examines the larger context in which the concept of impurity -and the ritual of the pharmakos -emerged. It suggests that mistrust was pervasive in ancient Greek society, as a product of the context of mortal and metaphysical uncertainty and danger that was generated by the fundamental unknowability not only of one's fellow citizens, but also of the gods. This lack of secure knowledge made impossible the "acceptance of vulnerability" that is a key element of trust (as described in the initial quotation of this article). It meant that there was no secure knowledge of one's own relationships with the divine, or with other mortals, and no securely positive expectations of the intentions or behavior of the divine towards mortals. This created, I argue, a prevailing dynamic of mistrust, within which the pharmakos ritual emerged, and which it exacerbated. In making this argument, this article will work with the concept of mistrust used by the anthropologist Matthew Carey (2017), who introduces the dynamic of mistrust in his discussion of communities in the High Atlas mountains, exploring how it both responds to and shapes a social environment that emphasizes the opacity of others, pervasive uncertainty, and a sense of contingency.43 As he argues, trust is generally taken to be a social technology 43 Modern examinations of trust are revealing a more persistent focus on what we might think of as the other side of trust. Whereas it used to garner little attention and be treated as "an absence" or "deficiency" (scholarly analysis has tended to regard mistrust as "un simple défaut ou manque de confiance, comme une absence qui demanderait à être comblée pour maintenir ou restaurer un équilibre -social, mental, politique ou économique" [Allard, Carey, and Renault 2016: 1]). Nowadays, scholars are beginning to explore its that depends on others and creates a redistribution of control (2017: 7). In contrast, mistrust arises from a strong sense of a lack of control. Carey regards mistrust as different from distrust because, he argues, distrust is more likely to be based on a specific past experience, whereas mistrust is "a general sense of the unreliability of a person or a thing" (ibid.: 7-8). Mistrust is "not simply a strategy or a matter of choice, but also a disposition," that is, "a general attitude of mistrust" (ibid.: 7). Like trust, mistrust allows us to think about the future.44 It allows for a certain level of anticipation, and thus of action, and this canlike trust -serve as "a basis for practical conduct" (Simmel 1950: 318-319). But Carey argues that whereas trust assumes (in the words of Simmel 1990: 179) "that there exists between our idea of a being and the being itself a definite connection and unity," mistrust assumes the opposite, creating a fundamentally different set of core social assumptions, practices, and social forms (Carey 2017: 3).
Turning to the ancient Greek context, we can analyze certain contexts to see these dynamics -trust and mistrust -at work among mortals. Steven Johnstone has noted how, for example, the Athenian lawcourts offered a ritual process of decision-making, which allowed for a reinforcement of citizen relationships, overcoming the strong sense of powerlessness that clever arguments provided. Instead, he argues, rhetoric in the Athenian lawcourts affirmed "a generalized trust in the system" because it "constituted relationships among citizens -not just between a speaker and a listener, but among the members of the judging audience -that were abstract impersonal, and powerful" (Johnstone 2011: 169). But perhaps all was not as straightforward as this argument suggests. For example, binding spells from a judicial context suggest a very different affect: not so much trust as profound mistrust. They indicate that litigants recognized the potential for their colleagues to request the intervention of more-thanhuman forces, which could distort these ritual processes of decision-making. They imply that citizen relationships were not so tightly aligned, and that individuals did feel mistrustful, both of each other and of their relationships with the gods (Eidinow 2013: 160-194). Such a sense of mistrust, I would argue, was not limited to this context, but was more generally a feature of mortal relations with the gods, both generated and produced by a prevailing spiritual insecurity.45 In what follows, I will explore some of the ways in which the social role. There is still some uncertainty about how to describe it (mistrust/distrust?), and there has even been the coining of new terms in the identification by some scholars as "untrust" (Marsh and Dibben 2005: 20). 44 For mistrust as the "functional equivalent" of trust, see Luhmann 2006: 83. 45 Drawing on Ashforth 1998Ashforth , 2005Ashforth , 2011; see also Eidinow 2019. nature and level of this spiritual insecurity may have varied, focusing here on cultural explanations of misfortune. That variation is, of course, not limited to what I describe in the outline below, and the particular valence of spiritual insecurity in any situation would have been influenced by many different factors, but I want to suggest that spiritual insecurity was a prevalent feature of ancient Greek religion. In ancient Greek culture there were a number of ways to explain experiences of misfortune, many of which turned on the possible activities of a multitude of autonomous "invisible beings," acting from a variety of motivations: you might think you were under supernatural attack by someone in the community, or you might consider that the gods had targeted you for some specific reason. Different types of epistemological processes would be used in these different contexts. Thus, if you thought you were under magical attack by your fellow mortals, you might identify the role of human agency using an inductive process that involved assembling social knowledge into a plausible pattern.46 Making it more difficult was the shifting nature of this social knowledge, which was likely to comprise not only one's own ideas about interpersonal relations and local politics, but also those of others (Goldman and Blanchard 2016); gossip would play a key role in sifting and making sense of this information. In contrast, in trying to identify a divine intervention unmediated by mortals, you could use a deductive approach. The evidence demonstrates that there existed a fundamental understanding of what behavior was required in relating to the gods. Ancient Greek men and women established clear parameters for the conduct of relations between communities and/or individuals and particular divinities. These systems of implicit knowledge underpinned rituals, informed individuals and communities about how to behave, and indicated how particular behaviors would lead to certain outcomes.
But this is not to argue that everything was straightforward. Gods could also be imagined as behaving willfully (throughout Homeric epic), or cruelly, as in tragedy, or simply inexplicably, as, for example, in Herodotus's (1.31) account of the rewarding of the good and pious service of Cleobis and Biton with an early death.47 The deductive approach to divine intervention did not mean certainty -there were too many interpretative possibilities. And I want to suggest that the lack of a single, reliable epistemological framework of interpretation to facilitate the understanding of more-than-human forces in creating 46 This aligns with the arguments of Eidinow 2013 concerning the role of the social construction of risk in the writing of binding spells, in which it is social knowledge about what is dangerous that becomes the basis for the choice to use certain processes. 47 Parker 1997 observes how different views of the gods are found in different genres.
experiences of misfortune would have resulted in some epistemic anxiety, and a resulting "spiritual insecurity." For this concept, I am drawing on Adam Ashforth's observations, arising from his work in Soweto, South Africa, where, he argues, people must try to make sense of the conflicting schemes of interpretation propounded by medical authorities, healers, and religious leaders. Ashforth describes it as "an existential condition marked by epistemic anxiety produced by ignorance of, uncertainty about, and/or disagreement among relevant authorities over the proper and effective modes of managing relations with agencies deemed capable of causing harm as well as those deemed responsible for the subject's safety and well-being" (2011: 136). I have argued elsewhere that this is a helpful paradigm for considering mortal -divine relations in ancient Greece, where there was a plethora of different ritual authorities, and a variety of explanations of and potential solutions for misfortune (Eidinow 2019).48 This is not an argument that the ancient Greeks experienced epistemological chaos. Rather, it is an attempt to explore in more detail the implications of what Christiane Sourvinou-Inwood identified as "the darkness and unknowability of the cosmos and the divine" (2003: 408).49 Certainly, ritual activities instantiated a set of activities and behaviors that, as it were, scaffolded that darkness and unknowability with a sense of reassurance; they were attempts to communicate with the divine. But they were not, I would suggest, considered to be foolproof. The evidence for this is manifold. Stories from across ancient Greek culture, told across its different genres, would have constantly reminded ancient Greek men and women that there was always the possibility that the gods would not respond, or would respond in ways that were unexpected, and even perhaps, unwelcome.
The fundamental relationship of charis, literally "grace" or "favor," provides a useful example: the term is most frequently used to describe the feelings of individuals in a relationship of reciprocity. Thus, the giver of a gift is demonstrating kindness or goodwill, while the recipient of the gift is feeling a sense of gratitude. Between gods and mortals, charis was elicited from the gods towards mortals by the provision of regular gifts, of the right kind, at the right time, to 48 This is, obviously, not to argue that the two contexts map directly on to each other. My intention with this comparison is to explore the potential impact on an individual of the availability of multiple possible explanations of the causes of misfortune and myriad approaches to resolving the problems. 49 This is an apposite study for this article, insofar as it argues that religious problematization was an increasing focus of tragedy from 430 B.C.E. While it suggests that tragedy may have set those issues "in a wider, and ultimately reassuring, perspective," it nevertheless indicates that the unknowability of the gods was an issue that required reassurance.
the right divinity. But as modern scholars trying to understand the mindset of an ancient individual or community, it is important not to look for a model that resembles the Christian relationship with the divine, or overestimate the extent to which the notion of charis provided a sense of reliable security and certainty.
Charis was certainly part of the deductive scaffolding that I have described. There is ample evidence that it offered individuals and communities a formulation for constructing a relationship with an unknowable god -but it was far from a guarantee. This is clear from, for example, the questions posed at the oracle of Dodona, asking for instruction about which god or hero to address, which illustrate the concern that one might get any aspect of this ritual wrong in the future (or have got this wrong in the past).50 Success was not only contingent on getting one's own gift right, but, as Herodotus's account of the story of Croesus illustrates (Hdt 1.91), it could depend on the gods' judgment of your distant ancestor's behaviors. Or perhaps the gods simply had different plans in mind; a programmatic example occurs in the Iliad (6.297-311), when Athena does not grant the desperate prayers of Theano and the women of Troy. Thus, there were no guarantees, and the potential presence in ancient Greek culture of the resulting "spiritual insecurity" provides a more nuanced understanding of the myriad ritualized activities of individuals and communities in this ancient culture. In particular, it raises questions about the roles played by rituals in both relieving, and exacerbating, this context.
The performance of rituals was intended to maintain relations with the gods. Research on ritual indicates that ritual performance likely provided a sense of reassurance and comfort (Hobson et al. 2018). In a world directed by more-than-human beings -whose responses were opaque and uncertainritual activity provided a way of reaching out to those beings and attempting to construct a relationship. But while this may well be the case, the performance of those rituals could also have been a reminder of the potential unreliability of those relationships.51 This is not to argue that the performance of ritual, in general, does not produce, at least briefly, less anxiety and a greater sense of group connection from its performance. But rather it is a suggestion, based on recent research (see further below), that there is more to say about ritual, the 50 Eidinow 2013: chapter 5, also discussed in terms of Xenophon's approach to the gods in the Anabasis ( context in which it takes place, and its effects, which may provide a greater understanding of ancient Greek ritual activity. Research on cognitive responses to ritual indicates that rituals tend to be conducted in contexts in which there is a sense of a lack of control, increased anxiety, and affiliative deficit (e.g., Hobson et al. 2018); these are symptoms that are associated, in modern psychological research, with high Intolerance of Uncertainty (IU) (Jacoby 2020). For individuals with a high IU, while rituals may soothe and comfort these feelings, lowering the perceived threat level, that relief is likely to be only temporary, since safety cannot be guaranteed (Jacoby 2020). My suggestion here is that the social and spiritual mistrust that an ancient Greek man or woman experienced could have resulted in a similar kind of context and response.52 The requirement for a continuous course of rituals running throughout the year is itself evidence that relationships with the gods and other supernatural powers were viewed as needing continual mortal maintenance.
Thus, I want to suggest here that ritual activities, such as the ritual of the pharmakos, which were intended to provide some sense of reassurance about the attitude of the gods to mortals, may well, paradoxically perhaps, have implicitly underlined the opposite: the fundamental unknowability of the gods. Indeed, as the preceding analysis of the ritual of the pharmakos suggests, it was perhaps particularly susceptible to this characterization. In a context of social and "spiritual insecurity," the ritual identification and expulsion of impurity appeared to offer some sense of security, but, simultaneously, the very ritual itself, as I have argued, underlined the impossibility of attaining that security. Like the figure of the pharmakos himself, the ritual of the pharmakos was not what it seemed.

Conclusion
Like most Greek rituals, the ritual of the pharmakos was a response to a context of social and "spiritual insecurity,"53 which gave rise to a crucial emotion, overlooked in scholarship on ancient Greek religion: a sense of mistrust. This context emerged from a combination of dynamics, between mortal and mortal, and between mortals and gods. In the case of the pharmakos ritual, as I have argued, lack of secure knowledge about one's fellow citizens laid open the 52 As above, this is not to say that modern psychological research maps neatly on to ancient minds -it is an attempt to attain insight based on a comparison. 53 Drawing on Ashforth 1998Ashforth , 2005Ashforth , 2011; see also Eidinow 2019. possibility that some hidden impurity might emerge and endanger the community. We can, perhaps, glimpse in the evidence for the ancient pharmakos rituals, the dynamic that Foucault describes, in which the specific impurity of an individual emerged from the State ritual of exclusion that punished it. In turn, Kristeva's examination may offer some insight into the psychological underpinning of these rituals, and the ways in which the curious ambiguity of the historical pharmakos ritual illuminates the crucial role it played for the community. Both Foucault's and Kristeva's studies suggest that rituals of defilement were intended to restabilize the community. But I have suggested that in the case of the ritual of the pharmakos -and more generally across ancient Greek ritual activity -this may not have been the case. Instead, the profound unknowability of the gods meant that ancient Greek men and women may have experienced a sense of spiritual insecurity, which gave rise to a crucial dynamic of "mistrust" in the relations between mortal and mortals, and between mortals and gods. While the performance of rituals was intended to maintain relations with the gods and reduce anxiety, research into cognitive responses to rituals indicates that they may have provided only temporary relief.
This then was the context of social and spiritual mistrust in which the ritual of the pharmakos took place. The pharmakos was not a demon to be expelled, nor did it represent a high-status individual willing to be sacrificed to ensure salvation. The pharmakos was rather, and importantly, an ambiguous figure: a representation of the hidden dangers that, at any time, threatened the community, and might engulf each individual.54 The ritual of the pharmakos was similarly ambiguous. It may have asserted some temporary sense of safety by creating clear categories: pure/impure, responsible/innocent, self/other. By driving this individual outside the city, both literally and figuratively, the community drove to the margins its haunting insecurity. But even as it did so, the ritual provided a constant reminder of the mistrust and ambiguity on which those fragile borders depended. 54 Cf. Kavaja 2017: 8, who sees the Thargelia as a routine expulsion of potential pollution. Versnel (1977: 38) argues that the pharmakos is 'identical with ἄγος ἐλαυνειν' (the driving out of pollution), citing Thuc. 1.126. But that phrase is used at 1.127.1 to refer to a curse; and at 1.126.12 a different formulation (ἤλασαν…τοὺς ἐναγεῖς) refers to accursed people. In each case, there is a clearly identifiable source of concern; in contrast, my argument turns on the uncertainty that the pharmakos represents.