Josephus’s Paraphrase Style and the Testimonium Flavianum

The controversial account of Jesus in Josephus’s Jewish Antiquities 18.63–64, known as the Testimonium Flavianum , has puzzling similarities to Luke 24.18–24, a portion of the Emmaus narrative. This article proposes an explanation based on established research into Josephus’s methods of composition. Through a phrase-by-phrase study, this article finds that the Testimonium can be derived from the Emmaus narrative using transformations Josephus is demonstrated to have employed in paraphrasing known sources for the Antiquities . Precedents are identified in word adoption/ substitution and content modification. Consequently, I submit that the Testimonium is Josephus’s paraphrase of a Christian source. This result also resolves the difficulties that have raised doubts about the Testimonium’s authenticity, with implications for the understanding of the historical Jesus.

2 previous article.1 Synoptic comparison shows that these two passages share an ordered sequence of about two dozen concepts and have about a dozen word roots and some unusual vocabulary in common at parallel locations. On the other hand, even when expressing similar concepts, the passages exhibit many differences in vocabulary, syntax, and tone.
What can explain these correspondences interfused with differences? I observed that they might signify that an inventive Christian forger of the Testimonium was influenced by the Gospel of Luke; or they may simply be a series of chance resemblances arising from common traditions about Jesus. But the most provocative proposal, and the one most open to testing, is that Josephus wrote the Testimonium by paraphrasing a text very much like, if not identical to, the Emmaus excerpt. Regarding this proposal, James Carleton Paget observes it has advantages over other theories in explaining some of the Testimonium's language oddities, and the sharing of the basic structure can be the reason for the passage's puzzling brevity.2 Louis H. Feldman noted that the proposed use of a Christian text could be related to the great interest Josephus exhibits in Judaean religious and political movements.3 But these authors treat the proposal as only an interesting speculation, as does Fernando Bermejo-Rubio, who quotes John Meier's statement that all speculations about the Testimonium's source are equally unverifiable.4 And indeed, normal procedures of establishing source dependence require a greater amount of verbatim agreement than is present between the Testimonium and the Emmaus narrative. However, what has not been appreciated up to now is that there is a copious amount of research on Josephus's methods of paraphrase that can be brought to the examination of the texts. Not only the similarities but also the specific differences between the Testimonium and the Emmaus narrative could potentially be understood as products of Josephus's methods of modifying his sources. And if the Testimonium could be shown to be a Josephan paraphrase, one would have a solution to the long-standing question: Is the Testimonium 4 finds this to be the case for the majority of expressions. A comparison with other ancient accounts of Jesus demonstrates that this finding is highly unlikely to be due to chance. As a consequence of this study, I will submit that the Testimonium is indeed a paraphrase by Josephus of a text very like, or even identical to, the Emmaus narrative. This prospect, which I will call the paraphrase model, suggests there are natural and satisfying answers to the most puzzling questions about the Testimonium.
The central problem of the Testimonium is the apparent conflict between its authorship by a first-century Jewish historian and its Christian content. Beginning in the sixteenth century, commentators seeking to resolve this conflict have proposed that Christian scribes had either forged the Testimonium or had altered Josephus's original version.10 Yet the passage is attested with little variation in the Greek manuscript tradition, providing no evidence of meaningful tampering.11 While a full consideration of current theories is beyond the scope of this article, it seems fitting to briefly summarize their difficulties compared to the paraphrase model.
The proposal that the Testimonium is a complete forgery has the difficulty that the linguistic style of the passage is generally consistent with that of Josephus.12 This has been disputed, as there are also deviations from his style that have raised questions of authenticity; the paraphrase model, as will be seen, can account for these deviations in terms of influence from the Emmaus source.
The stylistic evidence has led to theories of partial editing. The most influential of these is the proposal of John Meier that the original text of Josephus can be retrieved if one deletes portions perceived to be overtly Christian in 10  content, amounting to a third of the passage (29 out of 89 words).13 A major criticism of this proposal is that the selection methodology is based not on an objective foundation, but on suppositions about Josephus's comprehension and suspicion of Christian history.14 Differences in these assumptions have thus led to many competing theories for altering the received text. In contrast, the paraphrase approach relies only on a comparison of texts. The presence of Christian elements is resolved with the solution that it was Josephus who revised the text of a Christian author, rather than the reverse. This solution would eliminate much of the need to speculate on later Christian expansions. For example, Meier's reconstruction simply produces a subset of the paraphrase model text, but Meier's largest interpolation, the resurrection sentence, is seamlessly compatible with the rest of the paraphrase (see below, sections C22-C31). It strains credulity that this can be an accident of later interpolation. The paraphrase model also has greater explanatory power, in that the proposed source provides insights into the origin of each phrase of the Testimonium-a difficult task for interpolation theories, which struggle with such matters as the brevity of the Testimonium, its overall kerygmatic structure,15 and its silence on critical matters; through the proposed paraphrase, these are readily understood as reflections of the Emmaus source. Above all, the current theories are faced with the prospect that the entire Testimonium is consistent with Josephus's paraphrase style as well as his writing style. The Alternative Explanations section of this article considers how the theories could respond to this challenge.
In the Concluding Remarks, I provide some thoughts on the implications of the paraphrase model for the study of the historical Jesus and Josephus's interaction with Christianity.

The Paraphrase Study
The following study asks two questions for each phrase in the Emmaus excerpt: 'Is this concept represented in a Testimonium expression at a similar position? 6 If so, is there evidence in Josephus's sources and works that he would revise the Emmaus vocabulary into that seen in the Testimonium? ' The study seeks source-to-Antiquities transformations in Josephus's paraphrase that exemplify these revisions. It is to be understood that the source-to-Antiquities transformations thus found are not unique, as Josephus has a number of different ways of dealing with particular source words, including omission. The study can only establish that Josephus has at times made a particular transformation, not that he must.
An important limiting factor is that some Emmaus expressions have little or no representation in Josephus's other sources, so there can be no paraphrase parallels. In these cases, the study will examine if he associates similar expressions within his own works with the vocabulary used in the Testimonium. These associations are necessarily weaker evidence than direct paraphrase, but address the same question as to whether Josephus would rewrite a phrase a certain way.
For brevity, the Testimonium will be referred to as the 'tf' , and the Emmaus excerpt as simply 'Emmaus. ' Out of necessity, the study uses the text from the Gospel of Luke, without demanding that the Gospel must have been the actual source. I will remark on manuscript variations of either passage when they may bear on the discussion.16

Preservation of Ordered Content
In the study that follows, I have aligned the texts to display phrases where the tf could arguably be based, at least in part, on the same underlying concept, or logical template, as Emmaus. In the discussion, these correspondences are numbered C1, C2, and so on.
As will be seen, the study identifies thirty-one shared concepts at analogous positions in the Emmaus narrative and the tf. These concepts comprise almost the entirety of the tf. 16 The Greek text of Luke used in this article is that of the 28th edition of Nestle-Aland. The Greek text of the As noted in the introduction, this structure-vocabulary changes within a general preservation of ordered content-is the primary characteristic of Josephus's paraphrases when a reference text meets his needs.

Content Modifications
Josephus's paraphrases can exhibit changes in content to achieve certain characteristic goals. These changes can be grouped, following Christopher Begg, into three categories: rearrangements, omissions/condensations, and additions/expansions.17 Josephus's motives in making alterations are collected in the two volumes by Louis Feldman on Josephus's rewriting of the Bible.18 The major content differences are discussed in the following study, where it is seen they can all be understood in terms of these known motives.

Linguistic Modifications
To study linguistic modifications, it is often necessary to employ the portions of the Antiquities that are known to be based on Greek texts, rather than on the Hebrew or Aramaic Bibles. The most unambiguous Greek source is the Letter of Aristeas, which is paraphrased in Ant. 12.7-12.118; André Pelletier has performed a careful comparison of these texts.19 Karl-Friedrich Pohlmann has shown Josephus's close dependence on 1 Esdras (in Ant. 11.1-158),20 and numerous writers have identified a similar reliance on 1 Maccabees (Ant. 12.240-13.214).21 In addition, for some biblical books Josephus is believed to have relied primarily on the Septuagint, e.g., Kings and Chronicles; many of these revisions have been analyzed by Christopher Begg.22 Also useful are Josephus's revisions of his own earlier writing in the Judean War. The modifications fall into three categories. First are the shared word roots: Josephus's revisions typically retain a few words from their sources, often with a change in inflection or conversion into a cognate.23 Out of the thirty-one correspondences, the following discussion reveals twelve shared word roots in the proper context; each adaptation has paraphrase support. Next are substitutions with direct paraphrase support: I find there are eight expression substitutions in Emmaus-tf that have identifiable precedents in source-to-Antiquities paraphrases. Then there are substitutions with indirect support: there are at least eight correspondences that do not have clear paraphrase parallels, but which are seen in Josephus's writings as expression pairs that he treats as near synonyms. Last are replacements with little or no support: the remaining three pairs have less explicit evidence of association, but do employ Josephan language. The expression pairs in these four categories are listed in the Summary.

Emmaus-tf Parallels in the Antiquities
In the following, the passages are compared in table format, sentence by sentence. On the left is Emmaus and on the right is the tf, with somewhat literal translations beneath. Expressions are compared in order except for a few displacements, which are indicated by an asterisk in the discussion heading.  The things happening in it in these days-And he said to them, What things?-And they said to him, the things about Jesus the Nazarene, There happens about this time Jesus,

(Omission) Dialog Framework
The tf does not use the framework of a conversation on the road. Yet it conveys the core of what was said. There is a parallel in Josephus's treatment of the Letter of Aristeas. As Pelletier points out, Josephus extracted the substance of Aristeas's letter to a companion and erased all the elements that would recall a written conversation with a friend.24
The exact tf opening, Γίνεται δὲ, is found twenty-nine times at the beginnings of sentences in his works, but only three fall in Antiquities passages where sources are available. Two of these are derived from sources that, in the Septuagint version, do have 'happens/becomes' (although with the connotation of birth).25 The most similar revision is Gen. 21.2-7, 'the son that had come (lxx γενομένου) to him' , versus its condensation in Ant. 1.213, 'Now there comes (Γίνεται δὲ) to Abraham a son' . The inflection change-aorist participle to present indicative-and shift in position to the head of the sentence is formally nearly identical to the Emmaus-tf difference. In the second example, the revision of Gen. 36.4-5 to Ant. 2.4, ἐγένοντο is given a similar treatment.

(C5) prophet/wise (προφήτης/σοφὸς)
In Emmaus, Jesus is a 'prophet man' (ἀνὴρ προφήτης), and in the tf, a 'wise man' (σοφὸς ἀνήρ). Josephus typically eliminates noun appositions where they appear in his biblical sources. For example, 'warrior man' (lxx ἀνὴρ πολεμιστὴς, mt ish milhamah) in 1 Samuel 16.18 becomes 'warrior' (πολεμιστ ής) in Ant. 6.167; 'prophets men' (lxx ἄνδρας προφήτας) in 1 Kgs 18.4 becomes 'prophets' in Ant. 8.330 (although this may directly reflect the mt nevi'im). This treatment applied to the Emmaus expression would have produced 'prophet' . But it is well established that Josephus avoids using 'prophet' for persons who lived after the biblical era.29 There are no biblical verses where Josephus alters 'prophet' to 'wise' , but biblical prophets do not present the problem of the Emmaus text; what are relevant are post-biblical Antiquities sources, and none of these use 'prophet' . As described earlier (under 'The Paraphrase Study'), in cases where source representation is lacking, I instead ask if he elsewhere associates the words 'prophet' and 'wise' . One does see this association, for example, in reference Daniel, whom he calls both a 'wise man' (Ant. 10.237, drawn from Dan. 5.11) and a 'prophet' (Ant. 10.268, without biblical precedent).30 Josephus both associates and distinguishes these terms. He does tell us that 'wisdom' among his people is only credited to scholars capable of interpreting the Holy Scriptures (Ant. 20.264). What distinguishes a 'prophet' , on the other hand, as seen throughout his writings, is a power that comes directly from the divine spirit.31 If he did alter the Emmaus 'prophet' to 'wise' , then, he did not transmit the source with complete faithfulness; the implication would be that Josephus had a separate understanding of Jesus as a learned interpreter of Scripture. Perhaps that is why the next phrase could be interpreted as Josephus apologizing for inaccurately rendering his source.

(Addition) if indeed one should call him a man (εἴγε ἄνδρα αὐτὸν λέγειν χρή)
The remark 'if indeed one should call him a man' has no explicit parallel in Emmaus. Although often considered a forgery, similar vocabulary is used by 29  before God and all the people, of people such as received the truth with pleasure, and he won over many Jews as well as many of the Greeks; (C6) powerful in deed/doer (δυνατὸς ἐν ἔργῳ /ποιητής) Emmaus elucidates 'prophet man' with 'powerful in deed and word' . The tf similarly expands wise man with an analogous pair of concepts: a 'doer' (ποιητής) of 'deeds' and a 'teacher' . Geza Vermes has called attention to this similarity between Lk. 24.19 and the tf; he has proposed that the shared depiction of Jesus as a wonder-worker and teacher recalls 'the earliest, pre-Pauline Judaeo-Christian portrait' of Jesus as a prophet. 34 When the Bible uses the Emmaus expression 'powerful in…' (lxx δυνατὸς ἐν…), Josephus endeavors to capture its underlying connotations (e.g., 'hoplites' , 2 Chr. 25.6/Ant. 9.188); but there are no cases where he revises 'powerful' (δυνατὸς) to 'doer' (ποιητής). There are only three sourced instances of ποιητής in the Antiquities, and these are all copied from Aristeas and signify 'poet' .35 The verb 'to do/make' (ποιέω), however, occurs over a thousand times in the Antiquities and is used to connote any ability, as well as in conjunction with 'deed' (e.g., Ant. 6.243, 1.306). In an example that shows Josephus can link it to δυνατὸς, Josephus revises Aristeas 39, 'able to translate [δυνατοὺς ἑρμηνεῦσ αι]' , to Ant.12.49, 'be able to make a translation [δυνήσονται τὴν ἑρμηνείαν…πο ιήσασθαι]' .

(C8) before…all the people/surprising (ἐναντίον…παντὸς τοῦ λαοῦ/παραδόξων)
Feldman has drawn attention to Josephus's strong 'psychologizing tendency' , the inclination to add thoughts and emotions that are only implied in the source.36 Thus, where the Bible has an expression only denoting observation, such as Emmaus's 'before all the people' , Josephus will often substitute the people's reaction, including the tf's 'surprising' (παράδοξος In the tf, Jesus's words are not claimed to be true, only that he had followers who enjoyed truth. For this detail, as with 'teacher' , Josephus would have probably needed another source.
When 'all the people' is clearly unrealistic-as it is in Emmaus-Josephus uses instead 'most' (οἱ πλείους) as in 1 Sam. 13.22/Ant. 6.105), or 'the multitude' (τὸ πλῆθος), as in Jer. 26.8 (mt)/Ant.10.90) and 1 Sam. 14.39-40/Ant. 6.124-5. But there is no example where Josephus divides 'all' into two ethnic groups, as in the tf. But for an implausibly large number other than 'all' , there is a close parallel found in Josephus's revision of his own prior work: the perhaps unbelievable 'six thousand colonists' in War 1.403 are divided, in Ant. 15.296, into 'many [πολλοὺς μὲν] of his allies in the war as well as many [πολλοὺς δὲ] of the neighboring populations' .

(C12) people/people & Judaeans & Greeks (λαοῦ/ἀνθρώπων & Ἰουδαίους & Ἑλληνικοῦ)
Where Emmaus states Jesus' audience is all the 'people' (λαοῦ), the tf characterizes, first, the 'people' (ἀνθρώπων) who enjoyed hearing truth, and then the diversity of his followers, 'Judaeans and 'Greeks' . As the latter has little gospel support, it has been proposed that Josephus has drawn on his own knowledge of the mixed composition of the Jewish-Christian community in Rome.41 (The asterisk indicates displacement.) The opinion of Jesus' Emmaus audience that 'we were hoping he would liberate Israel' establishes he had firm followers, which is also implied by his powerful speaking ability. Josephus appears to capture this concept with 'won over' or 'led' (ἐπηγάγετο). The word appears over one hundred times in his works, but there are no known instances where it replaces the Emmaus ἠλπίζομεν; the latter occurs only nine times in his lxx sources (see C20). The displacement is discussed in the next section.

(C14) the Christ (τὸν Χριστὸν/ὁ χριστὸς)
When his biblical source contains 'the Lord's anointed' , Josephus never adopts the Septuagint's noun χριστὸς ('one anointed'), but replaces it instead with a more common expression, such as 'appointed by God' .42 He does use the verb 'anoint' (χρίω) ten times in kingship ceremonies where he makes clear its function, and at least once he avoids χριστὸς by using the verb instead. Because he had a standard way of rendering it that elucidates its sense, not doing so for the tf would indicate Christos was a term with its own significance to his readers and cannot be avoided. The meaning of Christos to the average Roman reader of Josephus in the 90s ce might be gleaned from Tacitus, writing some two decades later (Ann. 15.44), which indicates then Josephus's readers would have been aware that the Christians had been named after a real person in Judea, Christos. It isn't unreasonable to suppose Josephus composed the tf with the majority of his readers in mind.
Remarks of the same composition as ὁ χριστὸς οὗτος ἦν are found in Josephus's works. The presence of the definite article coupled with a familiarity with Christian history has led most translators to render the tf's ὁ χριστὸς as a title, 'the Christ' , or even 'the Messiah,' even though it can also be taken as a name, in accordance with definite article usage in Greek. The proposal that Christos in the tf is simply intended to be understood as a personal name goes back at least to William Whiston.44 The question is whether the article originated with Emmaus, where it also appears. Pelletier has studied the question of whether Josephus follows his source in using the definite article. Where the Letter of Aristeas includes the definite article with the name of a person, Pelletier has found that, ninety percent of the time, Josephus also does so in his corresponding paraphrase. 'The Demetrios' (ὁ Δημήτριος), for example, appears twice in Aristeas (301 and 308), and is copied both times by Josephus (Ant. 12.103 and 12.107). Pelletier gives 10 such examples, including Aristeas 304/12.105 ('the' Dorotheus) and 41/12.51 ('the' Eleazar), while only once is Aristeas's article omitted. Conversely, in sixteen out of the seventeen instances where Aristeas does not use the definite article, Josephus again follows suit.45 This practice shows Josephus would have no problem with copying the article from Emmaus while allowing the tf phrase to be understood as 'This was Christos' , in which he calls on the reader's recognition of the name.
I will comment here on just two of the questions that have been raised concerning this phrase. Many commentators have noted it seems out of context, which some have taken as evidence of interpolation, others of genuineness.46 I note that Emmaus has two descriptions of the followers' opinion of Jesus, that they 'hoped' he would 'redeem Israel' and that they are 'to believe' (πιστεύειν) the prophecies that he was 'the Christ' . It would have been characteristic for Josephus to condense these and group them thematically with his remark about Jesus's teaching a receptive audience.47 This process can have led to C12, C13 and the Christos phrase in their current locations. I discuss further the 'redeem Israel' phrase below (after C20). This leads to the next question. Jerome's Latin version of the Testimonium (De viris illustribus 13) translates as, 'He had many Judeans and gentiles as his followers and was believed to be [credebatur esse] Christus.'48 Many commentators have proposed that this may reflect an original form of the tf and that a later Christian editor removed 'believed to be' .49 As just noted, Emmaus also has 'believe' in association with Christos, suggesting Josephus can have adopted it and arrived at the construction referenced by Jerome. If so, however, there are reasons to conjecture that Josephus himself was the one to later alter the text. First, the received text conforms to his style. More significantly, Jerome's version does not make much sense outside of Jewish and Christian circles. For if the majority of Josephus's readers did indeed consider Christos to be the name of a historical person, then they would be perplexed by the assertion that one real person (Jesus) was 'believed to be' another real person (Christos). A puzzled audience response in public readings, then, can have led Josephus to make the most expedient change, the removal of 'believed to be' before official publication. While a better solution may have been 'he was called ho Christos,' similar to Ant. 20.200,50 this has neither textural witnesses nor support as an Emmaus paraphrase; and it still requires the activity of an unknown editor to obtain the current text. If Jerome's does represent an earlier version, the proposal of a final edit by Josephus is the parsimonious solution.

(C15) delivered/indictment (παρέδωκαν/ἐνδείξει) There are no Antiquities examples where a source's 'delivered' (Emmaus παρέδωκαν) is replaced by 'indicted' (tf ἐνδείξει).
There is a passage that treats the verbs as near-synonyms: at War 7.47, a magistrate's son in Antioch has 'denounced' (ἐνεδείκνυτο) a number of his fellow Jews 'and also delivered up' (παρεδίδου) some foreign Jews in a plot to burn the city. Josephus elsewhere uses ἐνδείξει for an accusation in fifteen instances.

(C16) chief priests and leaders/principal men (οἱ ἀρχιερεῖς καὶ οἱ ἄρχοντες/τῶν πρώτων ἀνδρῶν)
The accusers of Jesus are, in Emmaus, 'the chief priests and leaders' (οἱ ἀρχιερεῖς καὶ οἱ ἄρχοντες) and in the tf 'the principal men' (τῶν πρώτων ἀνδρῶν); the latter is one of Josephus's stock expressions for members of the ruling class.51 There are 14 examples of 'principal men' in the Antiquities, but only two are in passages whose sources are known; in these, the Septuagint has 'chiefs' (ἀρχηγοὶ; Num. 16

(C17) our/among us (ἡμῶν/παρ᾽ ἡμῖν)
The Emmaus narratives refer to 'our [ἡμῶν] leaders'; in the tf equivalent, 'the principal men among us' (παρ᾽ ἡμῖν), the first person is an aberration, for, in historical narrative, Josephus takes care to write in the third person. For example, compare the more typical Ant. 14.165, where the accusers of Herod are 'the principal ones of the Judaeans [οἱ πρῶτοι τῶν Ἰουδαίων]' . Louis Feldman termed the tf's uncharacteristic use of the first person a 'stylistic peculiarity' that raises suspicions of forgery.52 Josephus, as a rule, only employs the first person when he digresses from historical narrative to comment directly to the reader on current Judaean cultural realities-'our laws' , 'our festivals' , etc.53 The tf is not of this type.
On the other hand, there are odd points where Josephus relates reported speech in the third person, and suddenly shifts to the first person; the lack of quotation marks makes it uncertain if these are intentional or mistakes.54 The most apt example is Ant. 11.143, where the third person is used consistently except at one point: 'he was ashamed because of the people, who had put out of their minds all the things that had befallen our fathers [πατράσιν ἡμῶν] because of their impiety.' One would have expected 'their fathers' , as, for example, Ant. 8.127. The explanation is evidently in the source, 1 Esdr. 8. 70-8.74. In this speech, Ezra confesses that 'ever since the time of our fathers [πατέρων ημών] we have been and are in great sin.' It appears evident that with 'our fathers' Josephus was careless while converting direct to indirect speech. But it was a thoughtful error-the inflection was changed to match the context.
In the same way, the first person in the Emmaus speech can have been transmitted into the tf. This also required a thoughtful adaptation, from 'our' (ἡμῶν) to 'among us' (παρ᾽ ἡμῖν). Possibly the expression was so natural to him that he did not notice that in this instance it was out of context.

(C18) judgment of death/sentenced (κρίμα θανάτου/ἐπιτετιμηκότος)
The Emmaus 'judgment of death' (κρίμα θανάτου) is paralleled by the tf's condemned (ἐπιτετιμηκότος). The latter word with regard to punishment is found only twice where sources are known (but fourteen times in all, six in Book 18) and neither of these revise a source 'judgment of death' .55 Oddly, the combination 'condemn death [for oneself]' (θάνατον ἐπιτιμᾶν) occurs in the next story after the tf (Ant. 18.68), the only Antiquities instance of this association.
Unlike Emmaus, the tf names Pilate and correctly distinguishes his role from that of the principal men.

(C19) crucified/cross (ἐσταύρωσαν/σταυρῷ)
Emmaus uses the verb 'crucified' , the tf the cognate noun 'cross' . Josephus employs 'cross' and 'crucified' over forty times, sixteen where sources are known, almost always as a rendering of 'tree' and 'hanged' , and retaining the part of speech as noun or verb. There is only one instance of 'crucify' in the Septuagint, Esther 7.10, 'Let him be crucified thereon' (lxx σταυρωθήτω, mt t'luhu); this is revised at Ant. 11.267, 'ordered him at once to be hanged on that same cross [τοῦ σταυροῦ κρεμασθέντα].' But this may not be a transition from verb to noun, as it might reflect the prior sentence's 'tree' (lxx ξύλου, mt ets).

(C20) we were hoping that he/those first loving him (ἡμεῖς δὲ ἠλπίζομεν ὅτι αὐτός /τὸ πρῶτον ἀγαπήσαντες)
A notable content correspondence is that both texts, after relating the crucifixion, reference the previous high regard his followers had developed for Jesus. But in terms of vocabulary substitution, forms of the Emmaus verb 'hoping' (ἠλπίζομεν) occur only nine times in Josephus's sources, none of them in a comparable context, so there is no direct evidence he would make this rephrasing. In some indirect evidence he would associate these words, the exact tf form of 'loving' ἀγαπήσαντες appears only twice elsewhere in Josephus, and in one of these, Life 24, the people have the attribute of hope: 'those loving war' (οἱ τὸν πόλεμον ἀγαπήσαντες) 'hoped' (ἤλπισαν) to vanquish the Romans. Curiously, the 'hope' here is the same as in Emmaus, national liberation. Elsewhere, because John of Gischala would 'hope for' (ἐλπίσαι) great things he 'loved' (ἀγαπᾶν) war (War 4.85).

(Omission) the one about to redeem Israel
The tf does not have the Emmaus narrators' hope 'that he is the one about to redeem Israel' (ὅτι αὐτός ἐστιν ὁ μέλλων λυτροῦσθαι τὸν Ἰσραήλ). In Josephus's
josephus's paraphrase style Journal for the Study of the Historical Jesus 20 (2022) 1-32 sources, the verse that is the most similar to that of Emmaus is 1 Macc. 4.11; in fact, Luke seems to be referencing it here. In that verse, Judah Maccabee speaks to his troops at Emmaus-in Church tradition, the same site as Jesus's appearance-and prays that heaven will grant them victory against Syria, 'and all the nations will know that there is one who redeems and saves Israel' (ὅτι ἔστιν ὁ λυτρούμενος καὶ σῴζων τὸν Ισραηλ). In Josephus's paraphrase (Ant. 12.307) he omits this line of Judah's speech. As plain meaning is a desire for national liberation, the omission is in keeping with his tendency to avoid prophecies against the ruling empire.56 By analogy, he would also omit the very similar Emmaus remark. Moreover, if Josephus had included the hope for liberation in the tf, he would have been unjustly depicting contemporary Christians as adherents to an anti-Roman ideology.

(C21) and with all these things & some women among us*/did not cease (ἀλλά γε καὶ σὺν πᾶσιν τούτοις… /οὐκ ἐπαύσαντο)
At the analogous location to the tf's 'did not cease' (οὐκ ἐπαύσαντο), Emmaus depicts the followers' despair and then their encouragement by the women's story. There is an interesting parallel for this way of abbreviating continuing activity: a list of heretical actions by the northern kingdom (2 Kgs 7.7-23) is condensed by Josephus with the remark, 'they did not stop [οὐ…ἐπαύσαντο] their lawlessness' (Ant. 9.266).57 The avoidance of the women's story is consistent with Josephus's tendency to omit certain material he finds problematic, such as angels.58 καὶ ἀπῆλθόν τινες τῶν σὺν ἡμῖν ἐπὶ τὸ μνημεῖον, καὶ εὗρον οὕτως καθὼς καὶ αἱ γυναῖκες εἶπον, αὐτὸν δὲ οὐκ εἶδον. καὶ αὐτὸς εἶπεν πρὸς αὐτούς Ὦ ἀνόητοι καὶ βραδεῖς τῇ καρδίᾳ τοῦ πιστεύειν ἐπὶ πάλιν ζῶν today since these things occurred; and some women astounded us who had been early at the tomb, and not finding the body came saying they had seen a vision of angels who said he lived.
And some of those went with us to the tomb and found as the women said; but saw him not. And he said to them, "Oh, senseless ones and slow of heart to believe in living again

(C22) said to them*/appeared to them (αὐτὸς εἶπεν πρὸς αὐτούς/ἐφάνη γὰρ αὐτοῖς)
The dramatic turn in Emmaus is, 'And he said to them'; in the tf, 'he appeared [ἐφάνη] to them' . The difference reflects the tf's absence of the dialog framework. Surprisingly, one finds this association of terms in other paraphrases: there are a number of biblical verses where God or an angel speaks and Josephus, in his paraphrase, inserts a word for an appearance (with the tf's 24 root φαίνω).59 An example is 2 Sam. 7.4, 'That night the word of the Lord came to Nathan saying'; paraphrased at Ant. 7.92 as, 'But that night God appeared (φανέντος) to Nathan and told him' . In 1 Kgs 17.1, Elijah states that rain will come only 'at my word' (lxx λόγου μου, mt d'bari); in Ant. 8.319, there will be no rain until 'he appeared' (φανέντος).
In contrast to Emmaus's vivid encounter, the tf gives a tepid report. Josephus does have a tendency to lessen miracles.60 And the account is consistent with Greek historiographical practice; as Lucian would later advise, 'an extraordinary tale should be narrated simply for the audience to make of it what they will, so that the author does not take a risk by inclining to either side.'61

(C23) third day [accusative case] (τρίτην…ἡμέραν/τρίτην…ἡμέραν)
In Emmaus, 'there brings this third day' (τρίτην…ἡμέραν ἄγει); in the Testimonium, Jesus appears 'having a third day' (τρίτην ἔχων ἡμέραν). These verb phrases with the accusative 'third day' are distinct from the traditional dative, 'on the third day' (τῇ τρίτῃ ἡμέρᾳ).62 In fact, Emmaus and the tf are the only texts with the resurrection 'third day' as the object of a verb in all of ancient literature. Support for an Emmaus influence is that Josephus does have a strong tendency to duplicate the grammatical case of his source in the matter of day periods. He does this in all eight instances where he takes 'day(s)' from 1 Esdras; for example, at Ant. 11.148, the dative 'in two or three days' is an exact copy of 1 Esdr. 9.4.63 In the paraphrase of Aristeas, there are three instances where he duplicates the inflection versus two where he changes it.64 In revising 1 Maccabees, there are nine duplications versus one change.65 Thus, for these known Greek sources, 87% of the time Josephus does copy the day period inflection.

(C24) brings/having (ἄγει/ἔχων)
There are no Antiquities paraphrases replacing 'brings' (ἄγει) with 'having' (ἔχων). But Josephus is known to use either verb with time. There are five cases of 'have' (ἔχω) with 'day(s)' as its object; in each the source has a different verb. The most similar to the tf is Ant. 7.1, 'David passing two days in Sikella' (δύο ἡμέρας ἔχοντος), based on lxx 2 Sam. 1.1, 'stayed two days' (ἐκάθισε…ἡμέρας δύο and countless other wonders about him having foretold.

(C26) prophets (προφῆται & προφητῶν/προφητῶν)
The word 'prophets' occurs twice in the Emmaus narrative and once in the Testimonium. In both texts, what the prophets foretold encourages the followers after the crucifixion. Unsurprisingly, Josephus frequently adopts 'prophets' where it occurs in his biblical source; its adoption here, in a scriptural context, would not present the problem seen with 'prophet man' (C5). As a point of style, tf 'divine prophets' (τῶν θείων προφητῶν) recalls the characterization in Ant. 10.35 of Isaiah, and twelve others, as a 'divine prophet' (προφήτης θεῖος).
The content modifications have been seen to be consistent with Josephus's characteristic alterations. The main rearrangement is (C14) ho Christos. The significant omissions/condensations are the dialog framework, the women at the tomb, and 'the one to redeem Israel' . The major additions/enhancements are Pilate (C18), Greeks (C12), received truth with pleasure (C10), if one should call him a man, and the closing statement.

Alternative Explanations
The most salient alternative to a paraphrase explanation is that the correspondences are a series of coincidences. This raises the question of probability. Are thirty-one ordered content parallels a significant number to find in texts of this size? How many Antiquities paraphrase parallels could one expect to find? The most straightforward way to answer these questions is to compare other descriptions of Jesus of a similar length to Emmaus, in order to determine if they, too, can be judged consistent with Josephus's paraphrase style. The full results are too extensive for this article; what follows is a very brief summary. First, compare the Nicene Creed of 381 ce (174 words). For the simplest evidence, the word roots shared with the tf, there are 6: Jesus, Christ (but without the article), crucified, Pilate, on the third day (but dative), and prophets. There are perhaps a total of 9 content correspondences, not necessarily in order, far less than Emmaus, and many content differences.
The Acts speeches have respectively, 4, 4, 1, 7 and 4 word root matches with the Testimonium, an average of 4. In content correspondences, but disregarding position, I count 13, 13, 4, 17, and 14 potential matches with the tf (granting that others might count differently), for an average of 12.2. In both of these measures, Emmaus (12 roots, 31 content) far exceeds their average and is about twice their maximum (Acts 10 in both cases, which has many similarities to Emmaus).
In the category of substitutions with paraphrase support, the speeches average 2.2, versus 7 for Emmaus. Of these, the only substitution that does not also occur in Emmaus is hanged from a tree (Acts 5 and 10), a Septuagintism that Josephus typically converts to cross/crucify. Expressions without paraphrase support are rejected, proclaimed, and raised, among others. These examples show that the Antiquities is not so large that one can find evidence that any given expression can be paraphrased into a tf term. Most of all, synoptic comparisons of the speeches with the tf are poor compared to the foregoing Emmaus-tf exhibits.
Justin Martyr and Ignatius of Antioch also have several short descriptions of Jesus.73 Suffice it to say that the results of these comparisons fall between those of the Acts speeches and the Nicene Creed.
One can now attempt to answer the probability question. For the given measures, the Emmaus text is at least 3.7 standard deviations above the mean of the other texts. By the simplest estimate (a normal distribution), the probability that the Emmaus-tf correspondences are due to chance is about one in ten thousand. I do not want to put too much weight on statistical arguments with this type of sample set. But this estimate does give some quantitative perspective to the synoptic evaluations.
If coincidence is then excluded, an alternative explanation is that a sophisticated forger deliberately rewrote Luke 24.18-27 using Josephus's vocabulary. But it is counterproductive for a forger to closely follow the template of one writer in order to create an imitation of another-it would debase the feeling of authenticity. And there are many changes to Emmaus that are natural for Josephus to make for the tf that are difficult to understand coming from a Christian writer: the treatment of Septuagintisms, the intermittent adoption of words such as the non-creedal form of the 'third day' , the lowered tone, the mild expansion of 'mighty in deed and word' without added specifics such as healing, the addition of a large Greek following in contradiction to the gospel, the omission of the empty tomb, and so on. It seems one gains only problems and no advantages with this alternative.

Concluding Remarks
This study has shown that there is a path Josephus can have taken to revise the Emmaus narrative into the Testimonium by employing his known methods of revision. How confident can one be in this model? It must be acknowledged that the evidence is not of uniform strength. Where Emmaus is vague, e.g. 'mighty in word and deed' , when the tf is more specific, the argument of parallel content is necessarily weaker; one must postulate Josephus's use of a second source. Other additions and omissions marked out notable differences between the texts, and these were argued to be reasonable based on parallels in Josephus's writing. If the plausibility of these modifications is granted, then one can say that the sequence of the Emmaus template is followed quite closely.
The majority of the vocabulary modifications have good parallels among Josephus's known paraphrases, but others rely for support on additional examples of Josephus's usage, as shown in the Summary given above. The evidence is not of equal quality, as the number of parallels that can be found is subject to the vagaries of the resemblances between Emmaus and Josephus's other sources. The same variations of evidence could be expected in examining any of Josephus's paraphrases; an analysis of this point would be helpful. And some substitutions are more distinctive of Josephus than others. Even given occasional weaknesses, the cumulative evidence of substitutions strikes one as surprising. As observed above, comparison with other ancient descriptions of Jesus confirms the significance of this amount of evidence.
In the Introduction I outlined the principal advantages of the paraphrase model over current theories of the tf: that it is simple, powerful, and free of josephus's paraphrase style Journal for the Study of the Historical Jesus 20 (2022) 1-32