2. Previous Explanations of the Indefinite Citation Formulae

Most ancient commentators do not consider the specifics of the introductions, focusing rather on the citation within the text. Two notable exceptions are Chrysostom and Calvin. The former states in his sermons on Hebrews that the author uses ‘someone’ in Heb 2.6 on the assumption that the readers were skilled in the Scriptures and would have found the reference familiar.Footnote ¹ For Chrysostom, therefore, the formula functions rhetorically by highlighting the familiarity of the citation to draw the author and readers together in shared assumed knowledge. Calvin admits ignorance here, saying, ‘But why he suppressed the name of David does not appear to me.’Footnote ² Yet he goes on to state, ‘Doubtless he says one, or someone, not in contempt, but for honour's sake, designating him as one of the prophets or a renowned writer.’Footnote ³ For Calvin then, the indefiniteness functions theologically, indicating the status of the speaker and therefore giving authority to the words of the citation. It is unclear, however, on what grounds he makes this claim concerning honour. Almost no ensuing commentator, moreover, has followed Calvin in arguing that the indefinite personal pronoun increases the honour of the human author.

Among older critical commentators, those who consider the indefinite formulae tend to connect the wording to the ways in which Philo introduces Scripture in his works.Footnote ⁴ They draw out the significance of this in two ways. Some take it, like Chrysostom, as a marker of common knowledge, and hence it functions rhetorically.Footnote ⁵ Others take it as a marker of common method: by using the same hermeneutical moves as Philo, the author relies on identical hermeneutical assumptions, particularly that the Scriptures (including the LXX) are the inspired words of God.Footnote ⁶ On this view then, the indefinite formulae function at least partly hermeneutically, drawing on underlying assumptions to guide understanding. Williamson, however, put a large dent in the significance of these parallels when he concluded that the similarities between Philo's writings and Hebrews are more imagined than real at multiple levels.Footnote ⁷ Yet such parallels in wording should not be overlooked, especially Philo's introduction of Gen 20.12 with the words ɛἶπɛ γὰρ πού τις (Ebr. 59–61).Footnote ⁸ Both Philo and Josephus, moreover, occasionally introduce citations with the indefinite marker of location. These authors, therefore, provide a potential counterpoint against which to check the function of the indefinite formulae in Hebrews.

Although a few recent commentators point out the connection to Philo,Footnote ⁹ most do not account for this parallel. For those who do comment on the citation formulae, they tend to make two main observations, which map to the two indefinite markers in Heb 2.6. With respect to που, some commentators point out that the indefinite location does not mean that the author did not know the location of the text within his Scriptures.Footnote ¹⁰ This is clear, especially from the second indefinite citation (Heb 4.4), which comes from Gen 2. Furthermore, the author's use of the Old Testament in the letter more generally – in both breadth and sophistication – indicates that he has spent a considerable amount of time in the study of these texts, which suggests his awareness of the location of each citation.Footnote ¹¹ Although this explanation removes a potential misunderstanding for modern readers, it does not offer a positive alternative concerning why the author would choose on two particular occasions to make the location of the citation indefinite.

With respect to τις (Heb 2.6), commentators tend to use more overtly theological logic. Like the indefiniteness of place, they often stress that the indefinite pronoun does not denote that the author claims ignorance of the human author of the text.Footnote ¹² Nor, they argue, does this imply that the author thought that the quotation from the Psalms should be heard with any less authority than other citations within the letter.Footnote ¹³ Rather, if they consider the function of the pronoun, they almost unanimously argue that the pronoun functions as a marker that draws attention away from the human speaker and towards the divine speaker of the Scriptures.Footnote ¹⁴ This, it is argued, magnifies the power and authority of the divine author by downplaying the identity of the human author.Footnote ¹⁵ The speaking of God in the Scriptures thus takes precedence over of the humans through whom God spoke, in line with other uses of the Old Testament in the letter. Yet if the author wanted to draw attention away from the human author and towards the divine one, why pronominalise the human speaker at all? Even more so, if the author wanted to focus on the divine speaking that occurs through the prophets to the forefathers (cf. Heb 1.1), why not explicitly identify the divine speaker, as he does repeatedly in the catena of quotations that proceeds Heb 2.6–8 (1.5–13), and as he regularly does throughout the epistle?

Madison Pierce, however, demurs from the current majority opinion on two counts. First, she argues that the Philonic parallels should be investigated, as they may provide important clues concerning the function of the indefinite citation formulae in Hebrews.Footnote ¹⁶ Second, she states that ‘the choice of the anonymous speaker is…not an attempt to distance the text from the person involved in the production’.Footnote ¹⁷ Yet to understand Pierce's own arguments regarding the function(s) of the indefinite markers, we must first examine her broader argument regarding the ancient practice of prosopological exegesis. Prosopological exegesis refers to clarification of the speakers in an ancient text in order to interpret it.Footnote ¹⁸ Although some places demonstrate a clear use of prosopological exegesis, such as Jesus’ reading of Ps 110.1 (Matt 22.41–5; Mark 12.35–7; Luke 20.41–4), Matthew Bates argues that the New Testament apostles, and especially Paul, use this kind of interpretation on a number of occasions.Footnote ¹⁹ Bates identifies three preconditions that enable the possibility of prosopological exegesis: the identity of someone in a dialogue must not be self-evident to an ancient reader;Footnote ²⁰ this ambiguity can be resolved by assigning a character within the dialogue; and the interpreter must hold that the text is inspired.Footnote ²¹ Thus, in order to identify instances of prosopological exegesis within the New Testament (or any interpretation of a text), the original text must involve speech, the person(s) assigned by the interpreter must not be trivial, and the introduction to the citation may flag for the audience that the text is being read prosopologically.Footnote ²² He also adds that if the text is read in a similar way elsewhere, this strengthens the likelihood of prosopological exegesis taking place in the citation in question.

Following on from Bates, Pierce investigates the use of prosopological exegesis in the Letter to the Hebrews. In order to identify instances of prosopological exegesis in the letter, she rearranges Bates’ two lists above into features that the ‘base text’ must present (the text must be speech, some of the participants in the speech must be ambiguous, and the text must be authoritative), and features in the interpretation that suggest prosopological exegesis (an ambiguous participant must be identified, the citation may be introduced with a formula that includes προσῶπον, and other places use of the same text in the same way).Footnote ²³ Since the προσῶπον citation marker is not found in the New Testament and parallel uses are rare, she focuses only on the identification of ambiguous figures within the citations. On this basis, Pierce shows that the author repeatedly reads his cited Old Testament texts in a prosopological way, assigning various texts to the speaking of the Father (1.5–13; 5.5–6; 7.17, 21; 8.8–12; 10.37–8; 13.5), the Son (2.12–13; 10.5–7), and the Holy Spirit (3.7–11, 15; 4.3, 5; 10.15–7; 10.30).Footnote ²⁴ For our purposes, while she argues that the author reads Gen 2.2 (Heb 4.7) as the words of the Spirit (to which I will return),Footnote ²⁵ she suggests that the indefinite pronoun in the introduction of Ps 8 leaves the speaker anonymous, which both allows the Psalm to be spoken on behalf of humanity and also differentiates the Psalm from the ‘present or timeless discourse’ of intra-trinitarian dialogue.Footnote ²⁶ She also concludes from her consideration of the function of που in the same introductory formula that it functions to distance the quotation from its original context.Footnote ²⁷

From this outline of the previous explanations of the indefinite citation criteria, three questions arise. First, with respect to Pierce, her explanation appears to leave the quotation in something of a tension: it is not part of the timeless intra-trinitarian dialogue, yet it is also not time-bound to its original context. Second, except for Pierce, commentators seldom consider whether their explanations have the same amount of explanatory power in other uses of the indefinite citation formulae, especially in Philo. Third, the common modern explanation that the indefinite personal marker glorifies God by drawing the readers’ attention away from the human author does not provide a sufficient explanation of the presence of the formulae within the context of the letter. Thus, questions remain concerning the function of these formulae within the argument of Hebrews.

3. Hebrews 4.4

The less complex use of an indefinite citation formula comes in the introduction to Gen 2.2: ɛἴρηκɛν γάρ που πɛρὶ τῆς ἑβδόμης οὕτως (Heb 4.4). Here, the author adduces the statement about God's rest within his extended reflection on Ps 94.8–11 LXX (95.7b–11 MT). The author begins his explanation of the Psalm with the connective διό (3.7),Footnote ²⁸ which signals that he is going to draw out the implications of the audience as the household of God.Footnote ²⁹ In order to accomplish this, the author introduces Ps 94 LXX with the words καθὼς λέγɛι τὸ πνɛῦμα τὸ ἄγιον. This introduction denotes not only that the Holy Spirit speaks generally throughout the Scriptures, but also that the author to the Hebrews reads these words as the Spirit speaking directly to the new covenant community. Although Pierce considers in part whether such a prosopological reading arises from the work of the Spirit recounted in the Psalm,Footnote ³⁰ it seems probable that the author turns to Ps 94 LXX at least in part because of the citation's opening word, σήμɛρον. The only two places in the LXX of the Psalms where this word occurs are Ps 2.7, which leads the opening catena (Heb 1.5), and here in Ps 94.5 LXX. Thus, the author may view Ps 94.5 LXX as instructions on living in the ‘today’: not the eternal today that exists between past and future, but the salvation-historical today of the Son's enthronement.Footnote ³¹ This leads the author to consider the negative example of unbelief and disobedience that the wilderness generation provides for the readers (3.12–9).

Yet the author does not end with this negative example. Because the vast majority of the wilderness generation missed out on the rest, the author asserts that God's promise of his people entering his rest still stands (4.1). This leads the author to draw a parallel between the reception of the Gospel by the wilderness generation and his audience, with the difference between the two coming only in the faith of the latter in the promises of God. Griffiths here argues that the rest remains open for the people on the basis of God's promise to give a land to the descendants of the patriarchs (Num 14.23), which fits with the allusions to Num 14 in the preceding context.Footnote ³² Yet the author's reading of the Psalm itself provides a simpler explanation. With the requoting of Ps 94.11 LXX (Heb 4.3), the author implicitly demonstrates that the Holy Spirit would not speak to those within the salvation-historical ‘today’ and warn them to avoid the fate of the wilderness generation unless the possibility remained open that those in the ‘today’ could succeed where the wilderness generation did not.

The parallel message to the two groups, moreover, implies that the goal of the promise must be the same in each case. The author, thus, considers the nature of ‘my rest’ with the help of Gen 2.2. Here, he employs the Jewish exegetical technique of gezerah shawah in connecting the noun κατάπαυσις (Ps 94.11 LXX) and the cognate verb καταπαύω (Gen 2.2) – a lexical connection present in the Septuagint but absent from the Masoretic Text.Footnote ³³ While this may suggest why the author introduces the Genesis text, this exegetical technique cannot explain its function within the argument. Rather, Gen 2.2 demonstrates that God did not make a rest as a part of his works; if this were the case, then the rest would exist within the created realm. Rather his rest comes into existence through his own resting at the end of creation, and therefore exists in the heavenly realm. The connection of Ps 94 LXX and Gen 2.2, therefore, demonstrates that the rest of which the Holy Spirit speaks in the Psalm is something that God himself enjoys in the heavenly realm and something into which he invites others.Footnote ³⁴

For the author, therefore, the wilderness generation and the readers represent two instances of the same reality, albeit salvation-historically displaced: people summoned by the Gospel into God's own presence to participate with him in the rest that he has had since the end of the first week of creation. The author goes on to underline this point in the verses that follow. Even the next generation, which did enter the promised land under Joshua, did not enter God's rest. For if they had, this would negate the need for the Holy Spirit through David, centuries later, to promise a ‘today’ when God's rest would still be available (4.7–8). The rest in Ps 94.11 LXX does not equate to the physical land of Israel but rather to dwelling with God; this is why God calls it ‘my rest’.Footnote ³⁵ It is this rest, the author to the Hebrews argues, from which God excluded the wilderness generation.Footnote ³⁶

Having the function of Gen 2.2 within the author's argument now in place, two questions regarding the introductory formula may now be considered. First, since the author does not provide a nominative in the formula, who does the author think spoke these words? In the preceding context, the most immediate speaker is God the Father in the oath requoted in Heb 2.3. Most English translations imply this understanding, by translating it generically as ‘he has spoken’. Yet compared to Gen 2.2 LXX, the author inserts ὁ θɛός into the citation, and this suggests that he sees a distance between the speaker of Gen 2.2 and the one being spoken about in the passage.Footnote ³⁷ If this is the case, then the author likely sees the Holy Spirit as the speaker of Gen 2, just as he sees the Spirit as the speaker of Ps 94 LXX. The same Spirit, therefore, testifies both from the beginning that God's rest exists in the heavenly realm and also exhorts those in the ‘today’ to endure in their pilgrimage towards that rest.

The second question is central to this investigation: why does the author include the indefinite spatial marker in his formula at all? The first possibility is that its inclusion has no particular significance. This would imply that the quote from Genesis and its introduction would mean exactly the same thing had the indefinite που been omitted. Yet the author does not customarily introduce Scripture in this way (from the limited sample size of the letter), and this indicates that the author chose to include it. This choice carries meaning with it.Footnote ³⁸ The second possibility is that που functions more like the English ‘somewhere else’. In this case, the author uses the indefinite marker to indicate to his readers that he is going to draw from elsewhere in his Scriptures before returning to the main text under consideration. While this does introduce a level of redundancy into the text, this does not rule out such a possibility, especially if the text of Hebrews started out as an oral communication.Footnote ³⁹ The third possibility is more explicitly hermeneutical: the author uses the indefinite location marker because, unlike the words of Ps 94 that speak to the ‘today’, the words of Gen 2.2 relate to matters that have been true since the beginning of creation and will continue from the ‘yesterday’ through the ‘today’ and into the ‘forever’. In this sense, the time of which Gen 2.2 speaks transcends the divisions of salvation history that play a part in the interpretation of Ps 94 LXX. Either of these second or third options gives a plausible explanation to the function of the word, and further consideration must wait until its use in Heb 2.6 is examined below.

Overall, therefore, Gen 2.2 does not reshape the understanding of Ps 94 LXX.Footnote ⁴⁰ Rather, the author argues that the Psalm was always spoken by the Holy Spirit to the community of God's people who live in the ‘today’. He shows that implicitly within the warning concerning the wilderness generation lies the promise that the same rest remains open for another generation. This rest, as Gen 2.2 shows, was not found in the land of Canaan but rather in the heavenly presence of God himself as he rests from his works of creation. The inclusion of που in the citation formula that introduces Gen 2.2 fits within this argument, either as a marker of bringing a secondary text in to explain the primary text or as a marker that Gen 2.2 does not speak about a particular time in salvation history as Ps 94 LXX does.

4. Hebrews 2.6

Similar to the instance considered above, the author introduces his citation of Ps 8 with the words διɛμαρτύρατο δέ πού τις λέγων (Heb 2.6). Yet unlike his use of Gen 2.2, the author gives an extended reflection on Ps 8, which appears to begin a new stage in the argument of the epistle rather than coming within an argument about another passage.Footnote ⁴¹ In the preceding context, the author assembles a catena of seven quotations (Heb 1.5–13). Together these citations function to differentiate the Son from angels, particularly in the entrance of the former into the coming world to sit down at the Father's right hand until the final consummation.Footnote ⁴² Importantly for the purposes of the current investigation, the author achieves this aim by consistently using prosopological exegesis – he takes all seven citations as words spoken by the Father, with one citation addressed to the angels (1.7) and the other six addressed to the Son (1.5–6, 8–13). The seventh and final citation brings the catena to a climax, showing that the Son now sits at the right of the Father while all his enemies are brought under his feet.Footnote ⁴³ According to the author, this is the present state and location of the Son.

The author then uses these two facts as the dual basis for the urgent need to pay more careful attention to the words of the Son. After this first warning (2.1–4), the author supports his warning with a return to the nature of the Son and his relationship with those whom he saves.Footnote ⁴⁴ Here he adds a third basis for the need for obedience to the Son's speaking: just as the Father has not spoken to the angels to enthrone them (1.5, 13), neither has he subjected the coming world to these same angels (2.5). This raises the question that drives the ensuing argument: to whom did God subject the coming world?Footnote ⁴⁵ The answer to this question is not immediately obvious, however, since the context is finally balanced at this point. On the one hand, the climax of the catena shows the Son as the one who reigns over subjected things (1.13), and both 1.5–14 and 2.5–9 portray an antithetical relationship between the Son and angels.Footnote ⁴⁶ Yet on the other hand, the more immediate context, and the one to which this statement is syntactically linked (γάρ), concerns the salvation of the people of God. God, moreover, has confirmed this salvation with works of the Spirit, including those that demonstrate authority over the created order.Footnote ⁴⁷ This creates uncertainty in the expectation of the audience, and into this ambiguity, the author cites Ps 8.5–7 LXX using the exegetical technique of gezerah shawah.Footnote ⁴⁸

Yet Ps 8 does not immediately appear to resolve this ambiguity. This can be seen in the most discussed aspect of the author's use of the Psalm – the debate concerning whether the author interprets the Psalm anthropologicallyFootnote ⁴⁹ or Christologically.Footnote ⁵⁰ Much of this debate has centred around the meaning of υἱὸς ἀνθρώπου, and whether translating the phrase as ‘mortal’ loses something essential in the argument of the author. Commentators debate whether the author and the audience were likely to have known of Jesus’ sustained self-referential use of the title.Footnote ⁵¹ Yet the connection of the Psalm to the argument does not concern υἱὸς ἀνθρώπου even though the language of Son and angels provides a degree of lexical cohesion with both the preceding and ensuing contexts.Footnote ⁵² Instead, the connection comes from the statement that God subjects enemies to a third party. This link can be seen both in the use of the same idea already quoted in Ps 109.1 LXX (1.13) and the focus on subjugation on both sides of the citation (2.5, 8).

Yet the author does not resolve this ambiguity even in the immediately following exposition since the referent of αὐτῷ (2.8) is not clear: the present lack of subjugation of all things could apply equally to humanity in general or to Jesus in particular. It is only in Heb 2.9 that the author names Jesus – for the first time in the letter – and shows that he in his incarnation is the fulfilment of Ps 8. It is here that those who argue for a Christological reading gain their strongest support, especially in the arc that Jesus lives out, which mirrors the narratival shape of the citation as it stands in the LXX and as edited in the epistle.Footnote ⁵³ Yet as the author continues his argument, Jesus fulfils this Psalm precisely because he became a human, becoming like his brothers and sisters in every way (2.17). The author chooses, moreover, to begin the citation with a line that explicitly places the focus on humanity. While this neither prohibits Jesus from fulfilling the Psalm nor entirely resolves the ambiguity in the context, it does draw the focus towards the nature of humanity. It is this nature that Jesus lives out, and therefore, the Son who sits at God's right hand in Ps 109 LXX can only do so because he is at the same time the one who suffers as the exemplary human in Ps 8.Footnote ⁵⁴ For this reason, although many scholars argue for a ‘both-and’ interpretation of the Psalm (rather than either anthropological or Christological),Footnote ⁵⁵ it is likely better to understand that the author presents the Psalm anthropologically and then takes that anthropological thought up within a Christological framework in Jesus’ fulfilment of the Psalm.Footnote ⁵⁶

Given that the author shapes his discourse in this way, his use of an indefinite citational formula can now be examined. Although it is seldom done, each of the indefinite markers may be considered separately. As already demonstrated, που could denote that the author will introduce a secondary text to aid with his ongoing explanation of a primary text. Such a denotation is possible here with respect to Ps 109 LXX, although the extended consideration of Ps 94 LXX creates a significant textual distance between the uses of Ps 109 LXX (Heb 1.13; 5.6). More promising is the other explanation considered with respect to Gen 2.2: που signifies that the text does not speak of or into one particular era of salvation history. Whereas the author has just noted that he is speaking about the τὴν οἰκουμέμην τὴν μέλλουσαν (2.5), he now guides the reader to interpret the words of Ps 8 as things true from the creation of the world, and therefore, equally applicable in both the current world and the coming one. In this case, the author guides his audience to understand the quote as an accurate assessment of the nature of humanity at all times in creation, not just from the time of the incarnation onwards. The pronominalisation of an indefinite speaker in this introduction, moreover, also supports this explanation, since this ‘someone’ could have spoken these words at any point in time, and they would still be true for every point in the salvation history of both the present and the new worlds.

This much Pierce also argues in her work on prosopological exegesis.Footnote ⁵⁷ Yet the indefinite τις likely plays another, more important role. The author uses prosopological exegesis in his reading of each of the seven citations of the catena in the preceding context (1.5–13), as well as the three quotes (2.12–3) that follow. The base text of Ps 8 LXX, moreover, contains all of the elements that both Bates and Pierce identify as precursors for a prosopological reading: the Psalm as a prayer necessarily contains a dialogical shape; both the author and the readers hold the text to be authoritative since they view it as Scripture; and most importantly, the self-reference of the speaker as the υἱὸς ἀνθρώπου presents a particularly inviting opportunity for the Christian community to read the words of Ps 8 prosopologically as the words of the incarnate Christ. Yet the use of the Psalm in the argument shows that the author does not interpret the Psalm in this way. The use of the indefinite pronoun of speaker then guards against seeing these words as the words of the Son: it is not the Son who speaks these words to the Father, but ‘someone’.Footnote ⁵⁸

The indefiniteness of τις in Heb 2.6 is, therefore, not intended to magnify the glory of the divine speaker by obscuring the identity of the human agent. Rather, the author draws attention to the human speaker, albeit indefinitely, in order to distinguish his interpretation of the text from the texts surrounding it that convey the intra-trinitarian conversation.Footnote ⁵⁹ This is not to argue that the author would reject the understanding of Ps 8 as divine discourse any more than the current prevailing explanation would force the author to deny that the Psalm is also human discourse.Footnote ⁶⁰ The point instead is hermeneutical: as a skilled communicator, the author guides his readers to the particular interpretation of the text. Hence, with the introductory formula διɛμαρτύρατο δέ πού τις λέγων, the author signals two facts that the audience must keep in mind in their interpretation of Ps 8: it is a testimony that applies to every era of salvation history, and these are not the words of the Son in response to the Father. The author thus uses the citation formula to guide his audience to hear Ps 8, a solemn and authoritative testimony from Scripture on the place of humanity within God's created and recreated order. For the author, Jesus uniquely fulfils this place, along with his brothers and sisters for whom he acts as their pioneer and priest (2.10–1).

5. Indefinite Citation Formulae in Other Second Temple Works

As already noted, some contemporaneous authors also use indefinite citation formulae in their works, and the purpose of this section is to compare their uses to those of the author to the Hebrews discussed above. The closest parallel that has been identified comes when Philo introduces a citation from Gen 20.12 with the words ɛἴπɛ γάρ πού τις (Ebr. 61). In the context, Philo discusses the nature of folly, and as part of his argument, he allegorises women as a symbol of those who determine their lives solely by the needs of the body and the customs of the world rather than by reason (Ebr. 46–56). Yet he also finds promise in the example of Sarah, ‘for she is represented by the oracles as one who has ceased all feminine things’ (αὕτη γὰρ ɛἰσάγɛται διὰ τῶν χρησμῶν τὰ γυναικɛῖα πάντ' ἐκλιποῦσα Ebr. 60, citing Gen 18.11).Footnote ⁶¹ As support for why Sarah could have left behind the things of women, Philo quotes from Gen 20.12, where Abraham testifies καὶ γὰρ ἀληθῶς ἀδɛλφή μού ἐστιν ἐκ πατρός, ἀλλ’ οὐκ ἐκ μητρός (Ebr. 61).

In his interpretation of Abraham's words, the absence of the possessive pronoun μου from the two final prepositional phrases allows Philo to connect Sarah to his broader discourse in an absolute sense: ‘[she is] from a father but not from a mother’.Footnote ⁶² Pierce calls this a ‘legitimate, albeit creative, reading’ of the LXX tradition, and she argues that the indefinite citation formula facilitates this reading by abstracting the text from its context.Footnote ⁶³ Unlike Heb 2.6, Philo does not use this formula to signal a shift in his exegetical technique since he immediately allegorises Gen 20.12 as a statement that Sarah is born only of the unchanging Cause of all things (Ebr. 61). Yet like Hebrews, Philo uses the formula to remove the citation both from the speech of either Abraham or the oracles (τις) and from the historical situation (που).

Beyond this formula, Philo also occasionally uses the indefinite adverb of place που within various instances of his citation formulae. In some examples, he also nominalises a speaker, such as καθάπɛρ καὶ ὁ ὑμνωδὸς ɛἶπέ που (‘just as also the psalmist says somewhere’ Deus 74; see also Plant. 108) and even names οἱ ποιηταί (the poets) in some instances (Agr. 24; Somn. 2.52; Prov 2.13). On other occasions that parallel the formula in Heb 4.4, he leaves the speaker implicit, often with the formula ɛἶπɛ γάρ που (Plant. 90; Congr. 176; Fug. 197; see also Agr. 51; Conf. 52; Migr. 182; Praem. 111). On most of these occasions, Philo uses the cited text as support (marked by the use of γάρ) for a broader point that he is making from another passage. Yet given that he uses the prepositional phrase ἐν ἑτέροις to introduce a citation on several occasions (Leg. 1.51; 2.27, 35, 48; 3.4, 42, 65, 142, 186; Sacr. 67; Det. 103; Post. 26, 89, 142; Gig. 49; Deus 77; Migr. 131; Her. 117; Fug. 58; Somn. 1.74; 2:222; Decal. 38; Spec. 1.25, 104; 4:123; Praem. 113; QG 2.59; QE 2.10), it is less likely that που carries the meaning of ‘somewhere else’ in these formulae. Rather, as with the author to the Hebrews, the indefinite adverb of location appears to function as a device that deanchors a statement from its context and allows it to be read in an absolute way, although unlike Hebrews the broader scriptural point may lie several lines, or pages, prior. In all these instances, therefore, Philo uses indefinite citation formulae for hermeneutical purposes.

Beyond these instances, most of the Second Temple uses of indefinite citation makers tend to be more idiomatic. Philo, for instance, occasionally speaks about the supposed derivation of a word using some combination of δηλόω and που, for example, ὡς καὶ αὐτό που δηλοῖ τοὔνομα (‘as also the same name declares somewhere’ Deus 103; cf. Deus 42; Her. 161; Abr. 230; Spec. 4.83; Legat. 112–3). In these instances, Philo most likely would not know the circumstances surrounding the development of the particular word in question. Yet for the purposes of his argument, these circumstances do not matter, and thus using που once again, allows him to remove the derivation from its context and use it in a more absolute sense. Perhaps the most stylised uses, however, come from Josephus, who references earlier material in his works with variations of πρότɛρόν που δɛδηλώκαμɛν (‘we have formerly made clear somewhere’ J.W. 7.7.4 §244; cf. Ant.11.341; 12.390; 20.239). On two of these occasions, Josephus has not mentioned his topic at an earlier point, and thus the Loeb editors assume that Josephus has erred, mistaking a source as his own work or not remembering his earlier discourse.Footnote ⁶⁴ Yet from what I have argued above, it is more likely that the indefinite marker does not signify a lack of precision of memory, but rather that the location of the writing – possibly in another non-extant work – does not matter for the point being made. Thus, even in these idiomatic uses of the indefinite adverb of place with verbs of speaking, the adverb functions in the same way as in the scriptural citational formulae of both the author to the Hebrews and Philo. The author to the Hebrews differs only in the context removed by που: not only the immediate literary context but also the salvation-historical context that he establishes at the beginning of his work (Heb 1.1–2).