What we said during the morning meeting will never be known completely because the tape of that conversation is the one with the 18 1/2-minute gap.

—Richard M. Nixon

On October 12, 1973, the United States Court of Appeals ordered President Richard M. Nixon to produce White House tapes and documents that had been subpoenaed in July. Ten days later, Nixon announced his compliance. District Court Judge John J. Sirica requested, among others, tapes for June 20, 1972 (three days after the burglary at the Watergate Hotel). But there was a problem. Rose Mary Woods, Nixon's personal secretary of twenty-three years, told him that she "might have caused a small gap" in the recording of a conversation that took place on June 20 with H. R. Haldeman, White House chief of staff and one of Nixon's closest advisors. Her story was unclear, but it seems that while transcribing, she reached for the telephone to place a call and inadvertently hit the delete button on the tape recorder. She did not notice the mistake until she finished her phone call of "about five minutes."

The gap actually ran for eighteen and a half minutes. Haldeman's successor, White House Chief of Staff Alexander Haig, publicly disparaged Miss Woods: "Typical of a woman," he said, "she did not know the difference between five minutes and one hour of talking." Although the charges of obstruction of justice and tampering with evidence were dropped, her reputation was damaged. A panel of scientific experts examined the tape and concluded that it was unlikely that the erasure was accidental. Did she deliberately erase the tapes? Was she willing to take the blame for someone else's actions? Or was she forced to take the blame for someone else's actions?

The covert operations of Nixon and his accessories came to be known as "Watergate" only once the burglars were apprehended. Whatever their schemes, the actions of the conspirators were brought to a halt, and all their attention turned to covering up evidence of illegal activity. Hence someone, possibly a woman, felt it necessary to erase part of the tape. The silence of eighteen and a half minutes became the focal point for debate, conjecture, and controversy.

In 1997, historian Stanley Kutler, with the help of Public Citizen, a national nonprofit public interest organization, succeeded in obtaining the release of the remaining Nixon tapes from the National Archives. Kutler edited and telescoped the conversations, eliminating what he believed to be "insignificant, trivial, or repetitious." His selection is most fascinating when it comes to his transcription of the gap. Although the tapes preserve more than an hour of conversation for the morning of June 20, Kutler chooses to transcribe only the silence:

June 20, 1972: The President and Haldeman, 11:26 a.m.-12:45 p.m., Executive Office Building

This is the highly publicized "18 1/2-minute gap." Technical and scientific investigations determined that the tape had been electronically erased by unknown persons some time after Alexander P. Butterfield revealed the existence of the Nixon Administration taping system in 1973. H. R. Haldeman's diary entry for this date talks about lengthy meetings with John Ehrlichman, John Mitchell, and John Dean, which concluded that it was necessary to keep the FBI from going "beyond what's necessary in developing evidence and that we can keep a lid on that." Haldeman said that he and the President talked about "our counterattack" and "PR offensive."

Where we hear the white noise of erasure, we read the words, "This is the highly publicized '18 1/2 minute gap.'" Notice too that Kutler's transcription is as cryptic as the silence of the erasure. The phrases that he chooses from Haldeman's diary to supplement the gap only highlight the secrecy. "Keep a lid on that" reaffirms the clandestine operations of the conspirators, while "our counterattack" suggests a counterconspiracy. Although over two hundred hours of taped conversations have been released to the public, eighteen and a half minutes of silence nevertheless conceal important information about one of the most elaborate cover-ups of the American presidency.

On November 22, 1963, in Dallas, Texas, 268 men and women witnessed the assassination of President John F. Kennedy. Twenty-two photographers, both amateur and professional, captured the event in still and moving pictures. Abraham Zapruder, with his primitive home-movie equipment, recorded the most complete film, twenty-two seconds beginning from the moment the motorcade turned from Houston onto Elm Street, until the final, fatal shot killed the president, and the limousine accelerated toward the triple underpass. The Zapruder film has become an icon of the turbulent 60s, a nationally recognized image of the violent murder emblematic of the demise of an entire generation hopeful for the social improvement advocated by Kennedy and known as the "New Frontier." For those too young to recall where they were when Kennedy was shot, the movie provides a common experience and unites the gaze of all generations of citizenry on a single focal point.

In the years following the assassination, the Zapruder film became the cornerstone of both the Warren Commission (the board of inquiry created by President Lyndon Johnson on November 29, 1963, to investigate the assassination and headed by Chief Justice Earl Warren) and its detractors. It was believed to be the most objective piece of evidence, capable of providing answers that plagued investigators. Yet, like the break of eighteen and a half minutes in the Nixon tapes, the Zapruder film too contains a significant gap. At the crucial moment of the assassination, the presidential limousine passed in front of a large street sign (now gone), reading Stemmons Freeway Keep Right, which blocked Zapruder's view. President Kennedy, Texas Governor John Connally, and their wives disappeared from sight. Each frame of the film has been scrupulously examined in conjunction with other photographs and eyewitness testimony. At frame 207, the President is seen before the street sign, waving at the crowds. At frame 224 the limousine emerges from behind the street sign, and the president's arms are raised to his neck. He is obviously hit. In the time between frames 207 and 224 (no more than four seconds), in the space behind the Stemmons Freeway sign, an event occurred that will remain unknown. Forensics, ballistics, acoustics, optics: every available scientific method has been applied and reapplied to the evidence. Nevertheless, the street sign in the middle of the screen hides crucial information.

In both examples, the information needed to complete the story and to ensure the continuity of an accurate narrative, one that represents the historical event from beginning to end, cannot be recovered. The erasure of the tape and the disappearance of the limousine behind the sign cause a gap in our knowledge of the sequence of events. In the etymological sense, the evidence, with its root in the Latin verb videre, "to see," is invisible. At these evidential blind spots, the historian, whose etymology is rooted in the Greek idein, "to see," is compelled to conjecture about what really happened. In the absence of fact, the historical accounts of Watergate and the assassination of JFK are left to the Aristotelian devices of probability and necessity. I do not claim that if the eighteen and a half minutes of tape or the four seconds of film were available, then the clouds of conspiracy would dissipate and all chains of causation would be patently clear. But it is around the gaps in the Nixon tapes and the Zapruder film that debates about the details of the conspiracies rage most fiercely. The better a historian is at negotiating these gaps, the more successful he or she is at creating a narrative that is likely to be accepted as the authoritative version, one that leaves little room for the skepticism, opinion, or imagination that can divide and thereby corrode society. A successful conspiracy narrative accounts for all the links in the chain of cause and effect and thereby contains fear and deters citizens from further unrest.

It is no doubt daring—and intentionally startling—to begin a book on conspiracy narratives in Roman history with a discussion of Watergate and the assassination of JFK, but these irresistible modern American events clearly illustrate the problem that compels this study. On the one hand, history is a forum in which to exhibit the deeds of men and women, so that they not fade into oblivion. But because conspiracy is a hidden, secret event, it resists—defies—exposition. In recording any conspiracy, important facts always remain in the shadows; to tell the tale of a conspiracy is to guess at a very great deal. So how does one reveal something that is deliberately kept secret? How does one speak with any authority on matters about which one knows little or nothing for certain? Of course, all historians face uncertainty and ignorance about their subject matter at some point. For all these reasons, I maintain that a conspiracy is an ideal circumstance in which to observe how a historian confronts the limits of knowledge.

***

Conspiracy plagued Rome from its beginning. According to Livy, the mysterious death of Rome's founder Romulus gave rise to rumors that the senators had plotted against him (1.16.4). The disaffected sons of Rome's fourth king, Ancus Marcius, conspired to assassinate Tarquinius Priscus because he was grooming his adopted son Servius Tullius for the throne. Servius was, they argued, nothing but a foreigner and the son of a slave woman (1.40). Immediately upon the assassination of Tarquinius Priscus, his wife Tanaquil concealed the death until the position of her stepson Servius was secure (1.41). The change from kingdom to republic also occasioned a conspiracy, for members of the family of the expelled Tarquinius Superbus plotted to restore him as king. They sent envoys supposedly to recover Superbus' property, but they in fact enlisted conspirators in secret. A slave, overhearing conversations and intercepting letters meant for Superbus, betrayed the conspirators to the consuls (2.3-5).

At the height of Roman expansion, the senate enacted a senatus consultum that called for the dissolution of the large-scale worship of Bacchus. Livy reports that more than seven thousand men and women were participants in the so-called Bacchanalian conspiracy (39.17.6). The restive last days of the republic were fertile ground for malcontents like Catiline who recruited the aid of foreigners. Calpurnia tried in vain to warn Caesar of the Ides of March; his assassination precipitated a course of events culminating in the reign of Augustus, who was himself vulnerable to repeated threats of conspiracy.

Of course, the successive Julio-Claudians fell prey to murderous plots. The suspicious circumstances surrounding the death of Augustus, Livia's concealment of it, and her schemes to secure the position of her stepson Tiberius echo the death of Tarquinius Priscus and the way Tanaquil jockeyed for her stepson's position. Josephus' lengthy account of the assassination of Caligula attributes the incipient conspiracy in part to the maltreatment of a woman. Agrippina the Younger plotted to kill her husband Claudius; she was, in turn, murdered by her own son Nero. Domitian was assassinated in a plot that involved his wife. According to the late antique collection of biographies known as the Historia Augusta, the tyrannical actions of the emperor Commodus drove his sister to conspire to assassinate him; according to Dio, her motives were suspect. The plan backfired, resulting in her exile and execution.

Clearly conspiracies and assassinations in Roman history are not hard to find, and many more examples could be added to this abbreviated catalogue. Roman politics appear to be synonymous with intrigue, and no conspiracy is complete, it seems, without the involvement of a woman, a slave, or a sometimes even a foreigner. Conspiracy is a particularly dangerous crisis of legitimacy, because the conspirators' clandestine actions run counter to the most fundamental principle of res publica: that all actions concerning Rome be conducted in public. The secrecy of conspiracy completely undermines the general operation of Roman politics and society. Therefore, conspiracy carries especially heavy emotional and moral burdens. This book seeks to understand how the Roman historians talk about conspiracy; how they articulate, in the open and public forum of history writing, the closed and secret event of conspiracy. As we shall see, conspiracy is betrayed—and thereby revealed to the public—by people trusted with access to the private chambers of the conspirators: bedfellows and slaves.

This principle is demonstrated in three of the most famous conspiracies of ancient Rome: the betrayed Catilinarian, Bacchanalian, and Pisonian conspiracies. Sallust's Bellum Catilinae, Livy's Bacchanalian affair (39.8-19), and Tacitus' Pisonian conspiracy (Annales 15.48-74) are strikingly similar. In each account, the historian struggles to construct a continuous chain of causality of an event that is shrouded in secrecy and silence. Moreover, women play an important role in each conspiracy narrative. By comparing the depiction of women, their various actions and motives, we can see how conspiracy was a corruption of all that Roman aristocratic life stood for.

But these conspiracies were betrayed. What happens when a conspiracy is successful in achieving its aim? The assassinations of Julius Caesar and the emperor Caligula provide instructive contrasts. While Sallust, Livy, and Tacitus tried to construct a coherent narrative about a secret event, Josephus (Jewish Antiquities 19.1-273) was at liberty to describe the assassination of Caligula in its entirety, precisely because the conspiracy was neither detected nor suppressed; its secrets were exposed only once the deed was successfully completed. Our study concludes with an analysis of Appian's narrative of the murder of Julius Caesar (Civil Wars 2.111-117), perhaps the most famous event in Roman history, which haunted the Romans and structured their perception of conspiracy, tyranny, and freedom.

In these five conspiracy narratives, a rhetoric of conspiracy contributes to what I call a strategy of containment. Conspiracy is dangerous and threatening, morally, politically, economically, and socially. It arouses fear, both in those in power, who risk being overthrown, and in those who conspire, who risk being discovered and punished. A historical account of a conspiracy controls these fears by solidifying them in written word and disengaging other possible narrative outcomes. Conspiracy narratives operate as palliatives; well-constructed accounts of the worrisome events assure the reader that the conspiracy was a rare exception and will not happen again, if everyone remains vigilant. In this sense, conspiracy narratives both contain fear and deter future attempts at revolution. Other types of events in Roman historiography are celebratory, lauding identity and power. The foundation stories of Aeneas, Romulus, and Lucius Junius Brutus, for example, extol the establishment of Roman rule. Great battles, sacks of cities, kings routed and captured, internal discord between consuls and tribunes, agrarian legislation, the struggle of the orders—these were the topics that made for good old-fashioned history, according to Tacitus (Ann. 4.32.1). There is no room for internal threats of conspiracy in this list. Closest to civil war, conspiracy reveals Roman society at its worst and Roman politics at its weakest. But unlike civil war, when difference explodes into fraternal bloodshed like a violent volcanic eruption, conspiracy is more like an earthquake, whose unseen forces suddenly and unexpectedly shift the ground beneath one's feet. Hovering between stability and revolution, conspiracy is most important for understanding how the Romans maintained continuous authority in the face of internal threats of violence and disruption.

Conspiracy in the Roman Literary Imagination

These five conspiracies may at first seem a rather disparate collection, ranging from 186 B.C.E. to 65 C.E., and concerning economic distress, religious worship, party politics, and imperial disaffection; indeed, each conspiracy narrative is its own constellation of substantive and stylistic factors. I have selected these five based on three interconnected criteria: their historical importance, their narrative extent, and the variety of sources that document them. So the texts merit comparison for several significant reasons. First of all, Sallust, Livy, Tacitus, Josephus, and Appian represent the temporal range of Roman historiography from the end of the republic through the age of the Antonines. Considered diachronically, the conspiracy narratives exhibit development and change over time. These five historians also represent a range of socioeconomic backgrounds, from the senatorial to the provincial. In spite of differences in native language and place of origin, these historians nevertheless choose to construct narratives in the face of the limits of knowledge that conspiracy imposes. The narratives themselves manifest similar elements, including a list of conspirators, reasons for discontent, secret meetings and intrigue, revelation of the plot or perpetration of the murder, torture or punishment, and reward of the informants.

Furthermore, these conspiracies are attested in more than one source, allowing us to compare the historians' narratives with other accounts. Sallust's full-length monograph, the Bellum Catilinae, is supplemented by Cicero's four speeches In Catilinam, the speeches Pro Murena and Pro Sulla, together with the commentary of Asconius on Cicero's speeches In Toga Candida, In Pisonem, and Pro Cornelio. In Latin, the summaries of Velleius Paterculus (2.34-35) and Annius Florus (2.12) supplement the biographical material in Suetonius' Divus Iulius (14, 17); in Greek, Diodorus Siculus (40.5), Appian (Civil Wars 2.1), Dio (37.29-42), and Plutarch's Life of Cicero (14-22) also recount the events. Yet, while the Catilinarian conspiracy may be one of the most fully documented events in republican history, the sources remain irreconcilable on several counts. This is precisely because the secretive nature of the event precludes full disclosure.

The famed Bacchanalian affair of 186 B.C.E. is presented by Livy as a conspiracy to upset the stability of Rome by means of frenzied Bacchic worship introduced by foreigners from the East. The senatorial decree banning the cult survives in an inscription (CIL I2 581 = ILS 18) and provides a primary source against which to measure Livy's lurid account. Tacitus' account of the aborted attempt on Nero's life is supplemented by the brief mentions in Suetonius' Nero, the epitome of Dio (62.24-28), and Plutarch's De Garrulitate (Moralia 505c-d).

A generation before Tacitus, Josephus wrote his account of the death of Caligula, an event that surely occupied Tacitus in the lost portion of the Annales. Our understanding of the assassination of Caligula is aided by Suetonius' Caligula and Dio (59.29). The assassination of Julius Caesar is documented in the works of three historians: Velleius Paterculus (2.56.3-59.1), Appian (Civil Wars 2.111-117), and Dio (44.12-22). These are supplemented by Plutarch's and Suetonius' biographies of Julius Caesar. In addition, Nicolaus of Damascus' biography of Augustus (the oldest surviving source) was written during the 20s B.C.E. and presumably reflects the version of the assassination given by Augustus in his Memoirs. The interconnections of the ancient sources demonstrate the historical and literary importance of these five conspiracies. Throughout the book, I draw on these other sources to clarify what is distinctive about the narratives of Sallust, Livy, Tacitus, Josephus, and Appian.

Most importantly, however, these conspiracies fascinated the Romans themselves. Catiline became emblematic of evil. Vergil commemorates him on the shield of Aeneas as one of Rome's most famous criminals. Civil war, like conspiracy, pits fellow citizens against one another; Horace, Lucan, and Martial compare the destruction of civil war with the wrath of Catiline. In his account of the conspiracy of patrician women who concocted poison in secret, Livy names Sergia and Cornelia. The allusion to the evil conspirators Lucius Sergius Catilina and the members of the house of the Cornelii—Publius Cornelius Cethegus, Publius Cornelius Lentulus Sura, Publius and Servius Cornelius Sulla—would not have been lost on Livy's audience. The moralizing satirist Juvenal includes the conspirators in an abbreviated, but pointed, catalogue of infamous historical figures. Avidius Cassius, who conspired against Marcus Aurelius, is called a "second Catiline," while Clodius Albinus, who plotted against Septimius Severus, is called the Catiline of his times.

The achievements of Cicero, and especially his suppression of the conspiracy, were eulogized in an epic poem by Cornelius Severus, and Lucan invokes the eloquence of Cicero, "beneath whose authority fierce Catiline trembled" (2.541). Seneca, Tacitus, and Martial attest the canonical status of Cicero's speeches In Catilinam. Along with the speeches, Sallust's monograph was also widely read in antiquity. Valerius Maximus twice draws upon Sallust's account for anecdotes. His example of A. Fulvius as a severe parent who visits harsh punishment upon his son derives no doubt from Sallust, as does his example of Catiline as a man of luxury and lust. Quintilian notes that Lentulus was the third Cornelius to rule Rome, a detail found in both the Bellum Catilinae and Cicero's third Catilinarian; in Ampelius, the conspirators are proverbial. Orosius, contemporary of Augustine of Hippo, declares in his history that the Catilinarian conspiracy was well known to all because of Sallust's monograph. Eventually the year 63 B.C.E. became synonymous with the conspiracy. Suetonius erroneously reckons that Augustus was born "on the day the conspiracy of Catiline was before the senate." The error belies the degree to which the Catilinarian conspiracy dominated the year 63 and cast a shadow across the imagination of Roman writers for centuries.

The worship of Bacchus also holds a prominent place in Latin literature, where it is depicted as unbecoming of a proper Roman. The frenzied Bacchanalia, characterized by madness, intoxication, and extreme physicality, appears often in the slapstick comedies of Plautus. Vergil twice compares the love struck Dido to a Bacchant, a convention taken up by later epic poets. The unbridled lasciviousness of the Bacchic revelers enrages the satirist Juvenal, who contrasts proper moral behavior with the wanton revelry of the Bacchanalia. Tacitus describes the Dionysiac garden party at which Messalina, wife of Claudius, behaved like a maenad (Ann. 11.31). Pliny's letter to Trajan on the Christians recasts the republican problems of religious worship under the empire; his description of worship recalls the dangers of such secretive societies. On the other hand, later Christians also co-opted the Bacchanalian affair for their own moralizing discourse. Clearly, the severity of the senate's response to Bacchic worship had lasting repercussions on the Roman mindset.

In the aftermath of the Pisonian conspiracy, Nero's brutal retaliation against some of Rome's most talented writers of the day came to symbolize the harsh circumstances of life under tyranny. Three of the greatest names of Neronian literature, Seneca, Lucan, and Petronius, were forced to commit suicide. Their works are a testimony to their literary genius and their precarious position in the Neronian circle. The late antique biography of Clodius Albinus reproduces a letter of the emperor Septimius Severus to the senate in which he chastises them: "You have not returned to me the gratitude which your forefathers showed in the face of the Pisonian conspiracy against Nero, which they likewise showed Trajan, and which they showed lately in opposing Avidius Cassius." The letter is no doubt the product of the biographer's imagination. Nevertheless, the statement is good evidence that the aborted Pisonian conspiracy had become legendary enough to be an effective point of comparison.

Like the foundation myths and legends that occupied Vergil in the Aeneid and Livy in Book 1, the prominence of the Catilinarian, Bacchanalian, and Pisonian conspiracies in Roman literature is testimony to their significance in the Roman literary imagination. As the ancient myths of Aeneas and Romulus allowed Romans to celebrate their origins, so the stories of the conspiracies allowed Romans to commemorate their preservation from the most pernicious danger of all: one's own fellow citizens.

It is by the hands of those closest to him that Caligula fell. Yet his assassination is not commemorated in the literature to the same extent that the Catilinarian, Bacchanalian, and Pisonian conspiracies were, no doubt because immediately upon his death, the senate desired to disgrace Caligula and refuse him honors. Although Claudius formally rejected a motion declaring Caligula a hostis (a public enemy), nevertheless he caused statues of the former emperor to be removed. Ample evidence of the defacement of Caligula in statues, inscriptions, and coins suggests a popular renunciation of him. As a Julio-Claudian, Caligula was assured his place in history, but he never seems to have been a favorite of the poets.

The assassination of Julius Caesar, on the other hand, fascinated the ages and provided ample material for literary productions. In the Georgics, Vergil gives one of the most extensive lists of portents in ancient literature, the omens attending the assassination of Caesar. Indeed, the foreboding indications of his death captivated the ancients. They also seem to have had a morbid fascination with the actual number of wounds. Appian, Suetonius, Livy, Florus, Zonaras, Eutropius, Valerius Maximus, and Plutarch all record that Caesar was stabbed twenty-three times. Twice Ovid speaks of the conspirators; in the Metamorphoses, Venus spies the incipient conspiracy (Met. 15.763), while in the description of the temple of Mars Ultor in the Fasti, Augustus vows to avenge Caesar's death at the hands of the conspirators (5.518). In the aftermath, Brutus and Cassius were canonized as the Liberators, as evidenced in the Tacitean speech of the Tiberian historian Cremutius Cordus: "There will be plenty of people who will remember not only Cassius and Brutus but even me, should destruction fall upon me." We shall see that the Flavian historian Josephus alludes to the murder of Caesar in his account of the assassination of Caligula. Indeed, the memory of the Ides of March runs deep in the Roman collective memory.

The Vocabulary of Conspiracy

Conspiracy is a paradoxical phenomenon, for it can be identified and named as such only once it has been brought to light. Disclosing a secret plot and calling it a "conspiracy" gives substance to the covert event, enabling it to be narrated. In this sense, conspiracy resides in the space between concealment and revelation, between silence and speech. This interstitial nature is manifested fundamentally in the vocabulary of conspiracy.

The Latin conspiracy narratives featured in this study denote the event with the noun coniuratio. Etymologically and in its earliest attested meaning, the verb coniurare refers to the act of taking an oath or joining in league with others for a common purpose. Poets use the term to designate the oath sworn by the Greeks against the Trojans at Aulis at the outset of the Trojan war. Coniurare is also the verb used by Statius for the oath that the women of Lemnos swore against their husbands. In a speech in favor of the Oppian law, Livy's Cato alludes to the Lemnian conspiracy.

In military contexts, coniuratio denotes the oath that soldiers took before a campaign. In the republic, there were two distinct military oaths. The consul administered to all soldiers upon enlisting the compulsory sacramentum, which had religious implications; violation incurred the wrath of the gods. This primary oath bound the soldier to the republic, his commander, and his comrades. A soldier took a second oath before a tribune when he actually joined his unit; in the ius iurandum the soldier swore to uphold discipline and to obey his immediate commanders. According to Livy, before the battle of Cannae in 216 B.C.E., the two oaths were combined for the first time and administered by the tribune. In his commentary on the Aeneid, Servius preserves a vivid description of pageantry of this bygone ceremony:

Surely if there is some uprising, namely, the Italian or Gallic war, causing great fear because of the proximity of danger, since there is no time to ask each [soldier] individually, the one who would have led the army used to proceed to the Capitol, holding up two standards; one red, calling forth the infantry, the other blue, summoning the cavalry (for blue is the color of the sea, from whose god it is commonly held the horse originates). Then he used to say, "Who so wishes the republic to be safe, let him follow me." And those who were present used to swear allegiance all together [at the same time]. So this type of campaign is called a coniuratio.

Coniuratio was thus initially a legitimate measure taken by law-abiding citizens dedicated to defending the republic. It was a course of action followed in extreme circumstances, when sudden danger made it impossible to enlist men individually. It was a public act by which members of the community affirmed their commitment to the preservation of the Roman state and to each other. The soldiers who assembled under a coniuratio were legitimized to fight the war at hand and would automatically disband at its conclusion.

In contrast to the military coniuratio, Caesar uses the term in the Bellum Gallicum to designate the treacherous uprisings of Gauls against him. The Remi (the only Belgian tribe to submit to Caesar at once and remain loyal to the end) deny participating in a conspiracy against Caesar (2.1.1, 3.3). As Caesar prepares to cross the channel to Britain, the coastal tribe of the Veneti swears an oath to take no separate action but to endure whatever happens among themselves (3.8.3). In response, Caesar takes harsh measures, justifying his action by equating the coniuratio with iniuriae, rebellio, and defectio (3.10.2). The Britains (4.30.3), the Gauls (5.27.4), and the Senones and Carnutes (6.44.1) are all said to have conspired against Caesar. Even the continuator of the Bellum Gallicum, Aulus Hirtius, emphasizes the ever-present threat of foreign conspiracy in the first sentence of Book 8: "Several tribes at that time were reported to have renewed their war councils and to have joined in conspiracies" (cf. 8.2.2, 23.3).

When Caesar established the colony of Urso in the months just before his assassination, a provision was made that "no colonist of the colonia Genetiva, established by the order of G. Caesar the dictator, shall get together any assemblage or meeting or conspiracy. . . ." Caesar had good reason to guard against conspiracy from this particular quarter; the people of Urso had been sympathetic to the Pompeian cause. Although the word coniuratio never appears in the Bellum Civile, the inscription of the charter of Urso testifies to the growing possibility of citizens engaging in conspiracy. Thus, in the course of Caesar's career, the use of coniuratio shifts from foreign to domestic contexts.

The same pattern of usage, from foreign to domestic, is evident in the speeches of Cicero. In only two of his pre-Catilinarian speeches does Cicero use the word coniuratio. The voluminous speeches against Verres, governor of Sicily tried for res repetundae (extortion) in 70 B.C.E., were not delivered in their entirety before a jury, but their subsequent publication provided a powerful venue for Cicero to voice his views on the corruption within both the provincial and judicial system, while at the same time promoting his own self-image among the Roman élite.

In the fifth book of the second speech, he refutes the argument that Verres ought to be acquitted because he was a good general, on the grounds that he had in fact suborned slave conspiracies while in Sicily. Verres had learned of a slave revolt in Triocala; the participants were convicted and condemned to death by crucifixion. They were bound to the stake, when Verres ordered their release and return to their master—for a handsome price, Cicero alleges. Just one year later, Cicero defended a former provincial governor from the charge of extortion. In the speech delivered on behalf of Fonteius, former governor of Transalpine Gaul, Cicero demonstrates his rhetorical agility. He argues that Fonteius is not guilty of res repetundae because he did not extort money from other Romans living in the province but only from Gauls. Furthermore, in a xenophobic tirade that forms the basis for his entire defense, Cicero discredits the witnesses, on the grounds that Gauls are, in general, untrustworthy. He insinuates that they are greedy, irrational, sacrilegious, and capable of plotting conspiracies.

In both of these speeches, Cicero depends on the prevailing negative attitudes towards slaves and foreigners. By casting opprobrium on the band of fugitivi (slaves who have revolted) and by casting the suspicion of conspiracy on foreigners, he reinforces the stereotype that was eventually to form the foundation of his attack against Catiline. Cicero vilified his fellow citizen by comparing him to a foreign enemy. Of course, Catiline was a greater danger precisely because he dwelled among the Romans. Thus, in the course of Cicero's speeches, the direction of the meaning of conspiracy changed, from an external attack by foreigners to an internal attack by citizens.

Habinek argues that in the speeches against Catiline, Cicero mobilizes the polyvalence of the word coniuratio and its ability to call to mind both the positive, salutary acts of law-abiding citizens and the negative, destructive acts of bandits and criminals. Relying on the remark of Servius that "coniuratio can be used of good things, for the word has a middling sense," Habinek maintains that the word is "ethically neutral," referring to runaway slaves and bandits as readily as to legitimate Roman citizens. The dual nature of coniuratio is captured on republican coinage dating to the first year of the Hannibalic war (218/217 B.C.E.), depicting a ceremonial oath-taking on one side and the head of Janus on the other. Inherent in the portraits of Janus on the coins are his associations with roads and journeys, with beginnings in general, and most significantly for understanding the iconography of these coins, with "the sacrifice preliminary to any important ceremony." Thus, his presence on the coins officiates the military coniuratio, bestowing his benediction on the oath-taking. The tumultus that necessitates the coniuratio puts the community in a state of uncertainty; the sudden onset of danger meets its swift oppression, in a moment that looks both forward and back. The duality of the portrait of Janus also suggests the possibility of a counterconspiracy; for every obverse there is a reverse, for every oath sworn to protect the state there is the possibility of an oath sworn to destroy it.

The summer of 32 B.C.E. saw a radical redefinition of communal oaths:

The whole of Italy voluntarily took an oath of allegiance to me and demanded me as its leader in the war which I afterwards won at Actium. The same oath was taken by the provinces of Gaul, Spain, Africa, Sicily, and Sardinia.

The speaker is, of course, Caesar Augustus, retrospectively recording his career in the monumental inscription known as the Res Gestae. In a characteristic ploy, Augustus used an already existing military practice, the act of swearing an oath of allegiance, in a novel way. Not only soldiers but citizens as well felt compelled to pledge their faith to Augustus. But this was not the coniuratio of Hannibalic times, taken by soldiers in the face of sudden tumult. Sworn by citizens, this oath was extra-constitutional and did not have the force of law. Surely the notion of tota Italia was suspect; according to Suetonius, Augustus excused the people of Bononia from swearing the oath because they had for so long been dependents of Antony (Aug. 17.2). The exemption of some suggests the compulsion of others. The phrase tota Italia does indicate the globalization of an otherwise localized event. A conspiracy is to be feared when a few swear oaths in secret or when a few swear allegiance to a particular leader, thereby undermining the constitution. The oath to Augustus, on the other hand, assuaged fear because everyone (tota Italia) swore the oath openly, indeed, voluntarily (sua sponte). Augustus manipulated the concept of "swearing together" so completely and effectively that thereafter a similar kind of oath was taken on the accession of new emperors.55 But in the turbulent years before Actium, at the moment of his break with Antony in that summer of 32, Augustus claimed that stability was achieved by the very act—an oath of allegiance—that had, within his own lifetime, produced instability and had in fact been deployed simultaneously by his enemy: "Such was the zeal on both sides alike that they made alliances with each leader by oaths of allegiance" (Dio 50.6.6). The oath recorded in the Res Gestae (together with its legacy) demonstrates quite forcefully the binary capacity of coniuratio, a state of affairs between danger and safety, between conflict and resolution, between old ways and new, between ancestral custom (mos maiorum) and revolution (res novae).

Approaching Conspiracy

A historical study of conspiracy would consider, among other aspects, the causes of the conspiracies, the social implications of rewards and punishments, and the political and economic trends surrounding the phenomenon of conspiracy. While this book makes efforts toward understanding these broader historical issues and may inspire such further work, it is not, in its conception, a historical contribution. Rather, as a literary study, this book is concerned with the way conspiracies are portrayed. Therefore, I approach each conspiracy individually for several reasons. Since we shall be dealing with five different events ranging from 186 B.C.E. to 65 C.E. and involving a wide cast of characters in very different situations, it will simply be easier to engage in the specifics of one conspiracy at a time. Furthermore, each author should be taken on his own terms; the circumstances of production for each were unique, and should remain so.

Individual chapters on Josephus and Appian underscore their importance in the development of Roman historiography; they are treated as essential, not tangential, contributors to the fabric of Roman imperial literature. Josephus and Appian remind us that the Rome of the emperors was a polyglot society that had successfully incorporated the diverse peoples of the Mediterranean basin within its political as well as cultural realm. Josephus and Appian compose their histories in the Greek language but clearly follow lost Latin sources. Thus, they offer a unique opportunity to observe the way Latin literature influences the composition of Greek works. Finally, it is important to take each conspiracy as a whole, for only then can we observe how narrative continuity is achieved by each author.

Since women participate in each conspiracy, it is necessary to identify some basic assumptions that underlie and shape our understanding of their role in the narratives. First, the representations of women are not expected to be wholly accurate portrayals of lived experience but rather fictional constructions designed along certain organizing principles. Women are assigned actions in the sources to serve the interests and the rhetorical strategies of male authors. For instance, in most texts, "woman" is a definitional tool that operates as the opposite of "man." Woman's actions and behavior are always contrasted to man's and are usually found wanting. When women usurp power beyond their means, they constitute a threat that must be controlled.

At the same time, it is important to remember that the ancient sources are notoriously unreliable about women's sexual misconduct. Political invective regularly deploys adultery as a weapon against women; the tropes of satire depend on the stereotypical sexual deviance of women. Before this premise was exposed, the study of ancient literary sources merely reproduced the male-authored point of view. But recognition of this premise allows us to observe the political, sexual, and moral ways in which women serve male discourse. For example, it is not coincidental that the women of the Julio-Claudian dynasty are consistently painted unfavorably by authors who tend to criticize the regime. Thus, attention to women in conspiracy narratives tells us more about male discourse on conspiracies. Any assertion otherwise would simply reinscribe male perceptions rather than expose them.

Second, there exists a persistent tension between the general representation of women and the details of their daily existence. Any reading of women in Roman conspiracies must acknowledge where the text follows certain generic rules for representing women and where the text deviates from the rules to represent a specific woman. These deviations are also constructed along certain organizing principles, for instance, the need to accommodate different social strata, political structures, temporal and spatial settings, and perhaps most obviously, differences in immediate circumstances. In each conspiracy narrative, the participation of specific women is a historical fact; yet the portrayal of these specific women is, to a certain degree, commensurate with the portrayal of women in general.

In this sense, the content of the five conspiracy narratives is not arbitrary or trivial; the presence of women in these conspiracies is not an accident or a rhetorical flourish but an essential fact of the event. Thus, while male authors control the way women are portrayed, the actual presence of women in the conspiracies appears to flex the rules of generally accepted gendered discourse. No doubt women have prominent roles in the outcome of Roman conspiracies. They are portrayed as having made a difference because they actually did influence the course of events.

To this extent, women in Roman conspiracy narratives demonstrate the tendency of Roman discourse on sex to be engrossed with departures from the established norms. Conspiratorial women are anomalies both sexually and morally, and these irregularities serve as symbolic frameworks for identifying and denigrating the fundamental nature of conspiracy. The singularity of women in conspiracy narratives serves to underscore the singularity of the phenomenon of conspiracy. By focusing on the exceptionality of conspiracy, an author can mitigate its effects, tame, and contain it.

Limitations are obvious: I say little of other women in Sallust, Livy, Tacitus, Josephus, and Appian. I do not assume that the other unstudied women are variations of the conspiratorial woman, nor do I want to make conspiratorial women emblematic of women in Roman historiography. My point is to suggest that the strategy of containment evident in Roman conspiracy narratives is strengthened by the portrayal of conspiratorial women.

In an article on Sallust's famous portrait of Sempronia, G. M. Paul remarks in passing that it is customary to include women, especially those of the lower class, on the list of degenerates ready to foment violence against the state. He goes on to claim that "the ancient reader . . . was fully susceptible to the idea of finding a woman, especially one with sexual charms, at the center of a conspiracy." Rather than take for granted the participation of women in conspiracies, I ask why they were drawn into conspiracies, what purpose their presence served, how they contributed to the overall continuity of the narrative, and most significantly, how they concealed or revealed secrets. While certain common features are readily perceived, individual characteristics emerge. I explore the dialectic between the general notion of a "whore with a heart of gold" and the particular woman's situation in each conspiracy. Some of the women taken up in this study are certainly not whores; others have hearts of lesser metal. It behooves us to pay attention to these differences, for the commonplace is not devoid of meaning. Although such critical practice can be condemned as merely making mountains out of molehills, nevertheless our understanding of Roman women stands to gain from attention to the smallest scraps preserved in an already fragmentary record.

As an intermediate state of affairs, conspiracy confounded categories not only of gender but also of status. As women were privy to the secrets of conspirators by virtue of their intimate contact, so too slaves had access to the secrets of conspiracy. Slave labor provided the basis for the ancient economy; yet slaves' "separate subjectivity," to use McCarthy's terminology, was the very thing that made them both valuable and dangerous. Although slaves could be subdued by the promise of manumission, the threat of slave resistance, both active and passive, was never far from the surface of society. For instance, certain intimate tasks afforded slaves dangerous avenues of contact with the lives of their masters. The barber or doctor had access to the body (indeed, the throat) of the master. A messenger could distort or destroy the entrusted communication. A slave's purposeful idleness or laziness, although difficult to pinpoint, detracted from his value to the master.

Although the threat of punishment deterred such behavior, nevertheless the possibility of revolt remained, giving rise to a persistent undertone of fear and mistrust between slave and master, as a letter of Pliny the Younger attests: "You see with how many dangers, insults, and ridicules we are surrounded; no one can be safe if he is lax and lenient." The constant presence of slaves in the daily lives of their masters gave slaves access to knowledge and so to power that threatened the equilibrium of the slave-master relationship.

Slaves were privy to everything the master said and did. Only their silence ensured the master's privacy, but this silence depended upon either the slave's loyalty or the master's threat of punishment. Between the extremes of loyalty and punishment, the ancient system of slavery interposed judicial torture, the legal, physical, destruction of the slave's silence at the expense of his ever-present body. As noncitizen chattel, slaves were subject to corporal punishment and were permitted to give testimony only under judicial torture. In general, slaves could not be compelled to testify against their owners. Slave owners in ancient Rome, however, had the right to call for the torture of their slaves, both male and female, for evidence on criminal matters; Bradley demonstrates the continuity of the practice from early republic to late empire. Although the slave owner risked damaging his own property, presumably the ends justified the loss.

A slave's access to spaces both public and private corresponds to the conception of his body as comprised of a deceptive exterior and a truth-bearing interior. So judicial torture was necessary to excavate the truth hidden within the interior of the slave body. Such state-sanctioned, institutionalized violence against slaves afforded the Romans one more way to gloss the unease with the power inherent in the dual nature of the slave's subjectivity and to deny the power caused by this duality.

Torture did not necessarily succeed in elucidating the truth; indeed, some slaves would say anything, even if it were not true, to stop the torture. Others would maintain obdurate silence in the face of cruelty, even to the point of death. According to Tacitus, the slaves of Octavia (wife of Nero) demonstrated the range of responses: examined under torture, some were forced by the intensity of the agony to admit falsehoods, but most persisted to uphold their mistress. When one of the faithful did speak, she chastised her tormentor with an insult, ironically all the more true precisely because delivered under torture: "Octavia's genitals are more pure than your mouth." Thus, evidentiary torture, a formal legal mechanism intended to safeguard the rights and property of the ruling élite, elicits not a dependable outcome of truth but a range of responses that bespeak the duality inherent in the institution of slavery itself.

While the line of demarcation between the slave, subject to evidentiary torture, and the citizen, immune from torture, was clearly discernible in the Republic, the torture of free witnesses is attested under Tiberius, Caligula, Claudius, and Nero. Indeed, the victims, both men and women, were nearly always incriminated in acts of treason, and conspiracy in particular. Garnsey summarizes:

When treason was the charge, no man was safe from torture, whether as a punishment or as a means of securing the names of possible confederates. The theory was that whoever threatened the life of the Emperor had forfeited his rights and privileges: he could be treated as a slave.

The torture of free persons by emperors was clearly an abuse of power for which they were not held accountable. Thus, the tyrannical response to the suspicion of conspiracy blurred the distinction between free and slave. As conspirators sought to change the political order, so the emperors effaced the social order.

No doubt, élite male conspirators had unlimited access to operate in all spheres of Roman society. But to preserve the integrity of the conspiracy, and indeed to preserve the appearance of their own moral integrity, they could not conduct the plot in open, public spaces. Upstanding Roman citizens led their lives in full view of the public eye. But conspirators had to keep out of sight. Hidden away in private quarters, the conspirators operated in the places women and slaves tended to occupy. Thus, women and slaves were privy to information otherwise inaccessible to the public. As a result, they were both valuable and dangerous to conspirators, for they could either keep or betray secrets. This Januslike quality made women and slaves the ideal participants in the duplicitous act of conspiracy.

The final part of this introduction sets forth my approach to conspiracy in the Roman historians. Although some scholars have paid attention to the rhetorical aspects of the narratives, there has been no unified exploration of the typology of conspiracy narrative. At work in these texts (as I hope to show) is more than literary allusion; resonances among the texts demonstrate more than erudite engagement, competitive emulation, or even self-deprecating admission of secondariness to one's predecessors. Rather, the conspiracy narratives reveal several different tactics for conquering fear and distrust and several different rhetorical strategies for depicting the struggle. I compare and contrast these five conspiracies to understand how the Roman historians talked about the perilous phenomenon of conspiracy and how they managed to construct coherent, continuous narratives of secret events so as to contain fear and deter future revolution. To do this, I employ a narratological method that seeks to discover, describe, and explain the mechanics of conspiracy narratives, the elements responsible for their form and functioning. Narratology takes into consideration three components of the reading experience: author, text, and reader. Attention to all three components yields a composite understanding of the narrative.

By author, I mean the five Roman historians (Sallust, Livy, Tacitus, Josephus, and Appian) who produced historical texts that contain descriptions of conspiracies. In each conspiracy narrative, the presence of the historian is evident in the text. For example, the use of the first-person pronoun, deictic markers of space or time (e.g., "in my day") and even directional signals (e.g., "as I stated above") signal the act of writing and remind the reader of the status of the narrative as written document. The intrusiveness of a given historian helps gage his degree of self-consciousness and his reliability, and this profoundly affects interpretation and the reader's perception of the historian. Does he appear to be objective or subjective? Is he to be regarded with suspicion or is he generally trustworthy?

A definition of text is somewhat more complicated. For my purposes, it refers to the written word that the author produced. The text is further subcategorized. "Story" refers to the events that are to be depicted, while "plot" refers to the chain of causation that dictates how the events are to be linked. But beyond the story and the plot, "narrative" is the revelation of events in a particular mode for the purpose of storing knowledge used to create and maintain a collective identity. In classical Greek and Latin texts, the revelation of events occurs either by showing or by telling. This is best exemplified by the presentation of the Iliad: the epic advances either by description or by speech. Furthermore, the epic is a storehouse of information that informs a particular cultural identity. Likewise, the historians constructed their storehouses of cultural-historical information on this twofold principle of revelation, namely, speech often in the form of dramatic mimesis, and description in the form of the author's own voice. This duality could also be thought of in terms of scene and summary: scenes in which characters speak for themselves, and summary in which the author speaks.

Despite this simplicity, such representation admits of complex literary productions because of the mode of representation, for it is in the choice of a particular mode that the author exercises his control over the text. There are several ways the author can manipulate the presentation of a narrative.

First, the author can present material either explicitly or implicitly. Either information is imparted in such a way as to be verifiable, or information is communicated by suggestion, innuendo, insinuation, or implication. In conspiracy narratives, implicit information is often more abundant than explicit information.

Second, the author makes presuppositions about what his audience knows. Presuppositions allow the author to shape the reader's world. The author emphasizes certain characters and incidents and not others. Some information is thereby considered shared, common, and therefore not new or revelatory, while other information is perceived as new and even startling.

Third, the author chooses his mode of discourse. In direct discourse, the author indicates that a character speaks. In Latin and Greek, postpositive verbs of speaking serve to mark oratio recta (direct discourse), given in an independent clause. In indirect discourse, the character does not speak. Instead, the author reports what the character said. In English, quotation marks are suppressed, and the conjunction that is inserted before the reported speech. Personal and possessive pronouns are shifted from the first or second to the third person. The tenses of verbs shift back, converting present tense to past. In Latin and Greek, the introductory verb is followed by a dependent clause; this is traditionally referred to as oratio oblique (indirect discourse). A third category, free indirect discourse, suppresses or delays the reporting verb of saying or thinking. McHale provides a clear example in English: "Oh, no, she was fine, she was just going to stay in bed all day, Mary answered in a dead voice."

Different modes of discourse allow for information to be processed in different ways. Speech that is highlighted by direct discourse or decentralized in free indirect discourse can emphasize certain themes or characterizations, lend a degree of suspense, contribute to the overall sense of aesthetic pleasure, and appeal to the reader's emotions more or less directly. Sometimes the mode of discourse complicates the reader's ability to discern easily who exactly is speaking and whose ideas are being represented. The degree of directness and indirectness of discourse provides a vantage point from which to observe the degree of the author's involvement in the statements being made.

Closely related to the mode of discourse is what has been traditionally referred to as point of view and more recently as focalization. There are two types that the author can adopt in his presentation. From the unrestricted point of view, the author-narrator tells more than any and all the characters could know or tell at the time of the situation described. Such a stance is characteristic of the so-called omniscient narrator. From the internal point of view, however, information is presented strictly in terms of the knowledge perception of one or several characters. The author tells the story through the eyes of one or more characters who know and can reveal only so much at any given time. Our Roman historians present much of the conspiracy narratives from the unrestricted point of view; however, they do shift to the internal point of view. Such shifts to internal focalization alert the reader to the limits of knowledge. Who speaks and who perceives the information that is presented shape the reader's understanding of the progress of the narrative.

Finally, the author chooses his mode of description. For instance, description can take the form of ekphrasis or allusion. The author perspicuously chooses the sequence, the temporal ordering, in which he presents events. Indeed, historians are by no means bound to strict chronological order; elsewhere I have demonstrated the effect, for example, of Tacitus' temporal ordering of events in the first book of the Annales. Ekphrasis, allusion, indeed all the tropes of description contribute to the overall aesthetic pleasure of the reading experience. No doubt pleasure is a fundamental aspect of literature. Yet the amusement of the reader is seldom innocuous; Horace links delight to admonition. Thus, in a conspiracy narrative, a pleasurable mode of presentation is an effective way to discourage thoughts of further uprising. By anesthetizing the reader to the realities of the violence of conspiracy, the aesthetics of a conspiracy narrative can also defuse the fear that a conspiracy engenders. Once reading a conspiracy narrative becomes fun, the conspiracy is no longer dangerous, fearful, or insidious.

Above all, the author chooses the degree to which the progress of the events is impeded. Delay, detour, digression, diversion—these techniques add an element of pleasure by entertaining the reader, and an element of suspense, or even frustration, by withholding information from the reader. For Barthes, the suspension of information is one of the key features of the hermeneutic principle, and it is largely on this principle that my investigation of conspiracies rests. A hermeneutic reading of a text interrogates the suggestions or assertions that there is a mystery to be solved. To what extent does the author formulate that mystery? Does the author propose a possible solution? To what extent does information presented contribute or hinder the solution of the mystery? As we examine the modes in which the Roman historians choose to reveal or suspend information, we will be forced to consider a more pressing question: What happens when the events themselves are withheld from the author who is trying to create a narrative?

At last we come to the third and final component of the act of reading, the reader. All the techniques of representation at the disposal of the author have profound affects on the reader's interpretation. Interpretation depends to a large extent on the mode of representation; at a fundamental level, the reader's understanding is constrained not only by his or her ability to decode the text but also by the content of the text. A text that presents information implicitly is more opaque than a text that presents information explicitly. Presuppositions, indirect discourse, internal focalization, temporal displacement, and extensive use of delay complicate the task of interpretation by making it difficult for the reader to answer certain questions and so to gain knowledge from the narrative.

Most readers assume that the author strives to be understood, that he chooses modes of presentation that limit interpretive possibilities. Indeed, I argue that Sallust, Livy, Tacitus, Josephus, and Appian all strive to limit interpretive possibilities and thereby contain the phenomenon of conspiracy within a strict code, so as to deter readers from entertaining the notion of conspiring against Rome. Yet the historian's ability to construct such an airtight narrative depends on his own knowledge of events and his ability to present the events—secret, hidden events at that—in a fully plausible way. In other words, the strategy of containment rests upon the historian's ability to construct a continuous narrative.

Studies of beginnings and endings in literature seem to take for granted the basic question of how an author gets from the opening to the last words. The inaugural moment of a text breaks a silence; closure is a retreat back into silence. Generic considerations often determine the ways an author treats these moments. An epic poet traditionally declares his subject up front; sometimes lyric poets employ the priamel (or preamble, a series of statements that gradually ascend to the goal of the poem) as an introductory tool. Historians often begin by justifying their endeavor in superlatives—"the greatest war ever waged"—and by stating claims of impartiality—"with neither animus nor eagerness." Often endings are signaled with such closural language as references to night, mourning, or death. But once the author has solved his problem of how to begin and end his text, the more abiding problem of how to maintain narrative continuity presses at every turn.

Continuity is the quality of a text that creates the sense of advancement from beginning, to middle, to end, following a chain of causality that gives the impression that the story "hangs together." Continuity is achieved in several ways. Repetition of information keeps the subject in the forefront of the reader's mind and assures the reader that the subject of the narrative is united. Yet variations in the repetitions advance the narrative; the same material is re-presented but with new qualifications that add to the reader's knowledge. Anaphora is a particularly useful rhetorical device for achieving continuity; the repetition of a word or phrase at the beginning of successive paragraphs helps keep the reader's place in the narrative sequence. Transitional words and phrases help ease the reader from one subject to the next. Structural and logical phrase and clause markers lend cohesiveness to a text by accounting for the succession of propositions that lead to a conclusion. The historian determines the connections between discrete events that bind them together in a continuous narrative; in short, he asserts causality.

In sum, three principles govern my approach to conspiracy narratives in the Roman historiography. First, the author controls the mode of representation of the narrative. The author chooses to give information explicitly or implicitly. He chooses his presuppositions. He decides whether information is transmitted by description or by dialogue, and if by dialogue, whether it is reported in direct or indirect discourse. He selects the point of view and can shift it within the course of the narrative. He decides the order for presenting information. Second, the hermeneutic principle accounts for the constraints placed on the reader's knowledge; it also explains the creation of suspense in the narrative. The author suppresses or delays information so as to guide the reader toward a particular conclusion. Until that conclusion is reached, the reader remains in a state of uncertainty. Finally, according to the principle of continuity, the narrative unfolds from beginning to end, logically. Making sense of conspiracies in narrative is a key step in the process of averting fear and keeping conspiracies from surfacing in real life.

After comparing how Sallust, Livy, and Tacitus construct their accounts of the betrayed Catilinarian, Bacchanalian, and Pisonian conspiracies, the three chapters of Part I culminate in a typology of conspiracy narrative. The continuous narrative of a secret event depends upon the transmittal of information from a private setting to the public sphere, and women are the ideal conduits for the transfer of sensitive information. Because the conspiracies to assassinate Caligula and Julius Caesar were not betrayed, the focus of Part II shifts to elucidate the hermeneutics of assassination, that is, the ways Josephus and Appian maintain suspense throughout their accounts, in spite of the reader's prior knowledge of the outcome of the events. Taken together, the narratives of the Catilinarian, Bacchanalian, Pisonian conspiracies and the assassinations of Caligula and Julius Caesar demonstrate that the capacity to expose, and so to narrate a secret conspiracy depends on either the betrayal that halts it or the success that brings it to fruition.