At the start of my undergraduate education in the fall of 2002, I took Greek with Ed Glenny. The class had a vibrant and nurturing atmosphere, owing principally to Ed’s teaching style—though the enthusiasm of my peers helped. He was particularly winsome and birthed in many of us a love for ancient languages and biblical studies. My path of study took me next to Ancient Hebrew, and soon after I naturally fell in love with the Septuagint (Old Greek or OG) as a place where Hebrew and Greek collide. During my graduate studies and now in my early career, Ed has been a significant source of encouragement and support. He has also been a wonderful interlocutor for my work on the OG, whether in print or in private conversation. I am lucky to be among Ed’s students and friends, and I am honored to contribute to this volume in his honor.

My work on the OG focuses on its value for textual criticism of the Hebrew Bible (HB);^[1] in the texts that I work with (e.g., Exodus and Psalms), my inclination is to look to the Hebrew Vorlage to explain phenomena in the OG. In contrast, Ed’s work looks to the translator and his work to explain the OG phenomena. Though the difference may stem primarily from our respective corpora (Ed works on the Minor Prophets), Ed has pushed me to give greater weight to the translator. In the following reflections on some readings in OG Ps 4, I attempt to pay significant attention to both sides of the coin.^[2] Old Greek Ps 4, like all texts in the OG, is wonderfully complex. In this piece of ancient translation, we find phenomena related to language (Hebrew and Greek), scribal copying (of the HB and OG), translation practice (from “literal” to “free”), and interpretation, all vying for our attention in a short space of text.

1. Psalm 4:7^[3]

While the Greek of Ps 4:7 follows the Hebrew closely in most respects, there is an apparent discrepancy where the OG translates נְסָה with ἐσημειώθη (“it was given as a sign”). This is amplified because the word נְסָה is itself difficult in the Hebrew. The second stich of Ps 4:7 echoes the priestly blessing of Num 6:24–26:

נְסָה־עָלֵינוּ אוֹר פָּנֶיךָ יְהוָה

Lift up the light of your face upon us, LORD.

None of Ps 4 is extant in the Dead Sea Scrolls, leaving the Masoretic Text (MT)—given above—as the only Hebrew witness. The verb נְסָה is not entirely straightforward; the exact form is found nowhere else in the HB. Given the context, and the lack of alternative explanations, this is an imperative of the verb נשׂא (“to lift up”). While the imperative, as a I-נ verb, usually drops the initial radical נ and takes the form שָׂא, the form נְשָׂא is attested in Ps 10:12, and some I-נ verbs regularly take this form (e.g., נְטֵה, “stretch”). The interchange of ס and שׂ is not infrequent, and even more so the interchange of ה and א, especially with an a-vowel at the end of a word. Though it is improbable, it seems the improbable has occurred here: the strange form נְשָׂא is used, ס is written for שׂ, and ה is written for א. Though it is possible to speculate other interpretations of the data,^[4] this theory is most compelling and reasonable.

The fact that this spelling of the verb was in widespread use is reflected in the OG’s translation, which appears to stumble over the verb:

ἐσημειώθη ἐφ᾿ ἡμᾶς τὸ φῶς τοῦ προσώπου σου, κύριε.

The light of your face was given as a sign upon us, Lord.

The words of the Greek correspond closely to the Hebrew in all respects except the initial verb. Significantly, the verb σημειόω (“to give a sign”) is only ever used here in the OG. How did the OG arrive at an aorist passive third person verb (ἐσημειώθη) for the MT’s imperative נְסָה, and why was the verb σημειόω used?

1.1 The Old Greek Ps 4:7 as Translation

In this case, it seems likely that the translator’s source text had נְסָה, as in the MT. If this were the case, one can easily imagine the translator being unsure of what to do with the unique form. Based on his translation, he evidently did not understand it to derive from the roots נוס (“to flee”) or נסה (Piel “to test”).^[5] Instead, the translator connected the verb to the noun נֵס (“banner, standard, sign”).^[6] This was probably an intuitive connection—or even possibly an intentional strategy for dealing with a difficult word.^[7] If the former, the translator may have drawn—intuitively, not conscientiously—on the relationship of some qill pattern nouns to their corresponding verbal roots: as חֵן (“grace”) is the result of the action חנן (“to be gracious”),^[8] so too the noun נֵס (as “sign”) may have been (seen as) a denominative noun from an unattested root נסס, meaning “to signal” or “to give a sign.”^[9]

The connection to נֵס (“sign”) seems likely. But why did the translator render the consonants נסה as a passive verb? The reasons are not entirely clear. If the verb were active in Hebrew and the noun phrase יְהוָה (“LORD”) were the subject, it would be difficult to view אוֹר פָּנֶיךָ (“light of your face”) as the object of a verb meaning “to signal.”^[10] Yet, the OG’s understanding of אוֹר פָּנֶיךָ as the subject of a passive verb implies that with an active verb אוֹר פָּנֶיךָ could have been the object—the subject of a passive verb is the object of the corresponding passive.^[11] The OG, then, must have understood our hypothetical verb נסס to mean “to make something into a signal.” Perhaps the OG began by understanding יְהוָה to be a vocative (κύριε) and needed אוֹר פָּנֶיךָ to be the subject (τὸ φῶς); or perhaps the OG translated word-by-word, without an eye to the end of the verse, and did not consider the possibility that יְהוָה could be the subject. Either way, with אוֹר פָּנֶיךָ as the subject, the verb was consequently understood as passive. The interpretation of אוֹר פָּנֶיךָ (“light of your face”) as the subject of the verb rules out the possibility of taking the Hebrew verb as an imperative; instead, it is understood as qatal and past tense (thus the Greek aorist is used).

All these contextual considerations beg the question: how could the OG take the consonants נסה to be a passive qatal verb from the root נסס? The initial נ of a Nifal geminate verb would not assimilate in the qatal form;^[12] such a form would be written ננס. Perhaps the translator did not know this, or perhaps he “fudged” because the text was difficult. Then again, it is possible that the translator, again via intuition, viewed the root as נסה, given that geminate verbs sometimes have III-ה by-forms.^[13] The Nifal qatal form of this root would be נִסָּה (“it was made a signal”), consistent with the consonants of the MT’s text. The intuitive analysis of the unknown form נסה, then, would have followed several concurrent steps—an association with the noun נֵס (“banner, signal, sign”), semantics following the relationship of some qill nouns to their corresponding roots, and finally the interchange of a geminate root (נסס) with a III-ה by-form (נסה).

All things considered, it is more likely that the OG did not make all of these grammatical associations—even intuitively—in trying to deal with what was a difficult text. Rather, not knowing what to make of the spelling in context, he did his best by choosing something that both works contextually but also shows some etymological fidelity to the source text. This required him to skirt around some aspects of the grammar, but he had little other choice.

One interesting implication of this example is that the translator here was reading the Hebrew source himself. Theo van der Louw has argued that the translation of some books of the OG came about by one person reading the Hebrew text aloud and another person translating based on what they hear.^[14] While the theory has merit in other cases, in this example the person who translated into Greek must have been the same person who read the Hebrew. Otherwise, the translator would have heard nissā (if Nifal qatal) or nəsā (if Qal imperative) and would have assumed the root נשׂא—in this context, the root נסה would be indistinguishable.

1.2 The Old Greek Ps 4:7 as Witness to the Hebrew

If one wanted to venture beyond the MT’s text in search of a Vorlage for OG’s σημειόω (“to give a sign”), one would need to look to related nouns for translation evidence—because σημειόω occurs only here in the OG. Though the evidence is somewhat sparse, the noun σημεῖον (“sign”) is rigidly paired with אוֹת (“sign”) in the Psalms—אוֹת is translated only by σημεῖον, and σημεῖον is never used to translate any other word.^[15] One might, then, consider reconstructing a verb of the root אוה (“to sign, mark”). However, another related noun, σημείωσις (“indication, notice”), also provides important evidence. It occurs just once in Ps 60:6, where it translates the only occurrence of נֵס (“banner, standard, sign”) in Psalms. Furthermore, the translation of נֵס outside of Psalms is important, since OG Psalms sometimes looked to books in the OG Pentateuch for lexical equivalences. In OG Numbers, the three occurrences of נֵס are all translated by σημεῖον.^[16] If one were to speculate about a retroversion for the verb σημειόω, then, the (scanty) evidence points to a verb related to נֵס. If the MT’s text had something other than נְסָה, I would decline to retrovert a Vorlage in this instance, because there is insufficient data. But as it is, our extant Hebrew evidence suggests that נסה was in the OG’s Vorlage.

2. Psalm 4:3

Psalm 4:3 is significantly different between the Hebrew and the Greek. The Hebrew text is extant only in the MT:

בְּנֵי אִישׁ עַד־מֶה כְבוֹדִי לִכְלִמָּה תֶּאֱהָבוּן רִיק תְּבַקְשׁוּ כָזָב

Sons of man, how long [will] my glory [be] for reproach? You love vanity, you seek a lie.

The first stich lacks a verb, though it is possible to read a covert copula (“to be”). The lamed preposition in לִכְלִמָּה seems best read as communicating transformation, whether “for reproach” or “as reproach;” the point is that the “sons of man” turn the psalmist’s glory into reproach (perhaps only in their estimation, or they denigrate the psalmist’s reputation within the community broadly). Alternatively, lamed could be understood to communicate direction or goal here: the “sons of man” cause the psalmist’s glory to lead to or result in reproach. The second stich has overt verbs, but may potentially be gapping עַד־מֶה from the first stich: “[How long] will you love vanity and [how long] will you seek a lie?”

The OG clearly derives from some version of the Hebrew as found in the MT.^[17] However, whereas OG Psalms usually corresponds to the MT quite closely, there is significant difference in the middle of the verse:

υἱοὶ ἀνθρώπων, ἕως πότε βαρυκάρδιοι; ἵνα τί ἀγαπᾶτε ματαιότητα καὶ ζητεῖτε ψεῦδος;

Sons of man, how long [will you be] heavy-hearted? Why do you love vanity and seek falsehood?

Instead of the psalmist’s glory and reproach, the OG refers to the psalmist’s opponents as heavy-hearted, and asks why (ἵνα τί) they behave as they do. Placing the OG alongside the MT, most aspects of the texts pair in a manner typical in the Psalms. The words that do not align are כְבוֹדִי לִכְלִמָּה (“my glory for reproach”) and βαρυκάρδιοι ἵνα τί (“heavy-hearted, why”).

2.1 The Old Greek Ps 4:3 as Translation

Septuagintalists and text critics, when confronted by the data in this verse, conclude that the OG read a Hebrew text at variance with the MT. Before exploring that possibility, however, how might the OG be explained as a translation of the MT? If כְבוֹדִי לִכְלִמָּה were in the OG’s Vorlage, it would not have been difficult for the OG to translate it isomorphically: in the OG, כָּבוֹד (“glory”) is nearly always δόξα (“glory”), simple prepositions like lamed are represented by a preposition, and the noun כְלִמָּה (“reproach”) and related verb כלם (Niph “to be humiliated,” Hiph “to humiliate”) are regularly translated by ἐντροπή (“humiliation”), ἐντρέπω (“to put to shame”), and καταισχύνω (“to dishonor”). Something like δόξα μου εἰς ἐντροπήν (“my glory for humiliation”) is what would be expected for כְבוֹדִי לִכְלִמָּה.

Based on translation evidence elsewhere, it is clear that the OG translator understood the individual Hebrew words in the phrase.^[18] Did they make sense, however, together in their context in Ps 4? The lack of a verb in the first stich could have made the text difficult to the translator. The initial noun phrase and interrogative are rendered in Greek with a noun phrase and interrogative; following these two phrases, the translator faced yet another noun phrase, without a verb to glue the elements together into a meaningful utterance. If this were the case, the translator could draw on at least two different strategies. First, he may have ignored the difficult Hebrew and filled in the blanks from context within the Greek target text.^[19] In other words, the translator may have taken the Hebrew he understood—“sons of man, how long … you love vanity, you seek a lie”—and filled in the ellipses in Greek. In the preceding verse, the psalmist alludes to his “distress” and asks God to heed his prayer. Given the near context, then, the translator would naturally read the “sons of men” as adversaries, filling in the blank before their “loving vanity” with appropriately negative content. There is a strong tradition of negative figures in the HB having stubborn hearts (e.g., Pharaoh), and the theme appears also in the Psalms (95:8; 119:70). Calling the “sons of man” hard of heart, then, would be one way to smooth over the difficult section of this verse. Notably, the word βαρυκάρδιος (“heavy-hearted”) occurs nowhere else in the OG,^[20] perhaps suggesting that the translation is “free” or target-oriented at this point.

A second strategy, which could be used in tandem with the first strategy of contextual smoothing, is etymologizing.^[21] If the translator did not know how to plug the second noun phrase, כְבוֹדִי (“my glory”), into the clause, he may have looked for other senses connected to the root כבד. The verb is used to describe the dulling of the senses—the eyes or ears, for example—and, by extension, the dulling of the conscience—whether using language of ears or heart (Gen 48:10; Isa 6:10; 59:1; Zech 7:11). This is the verb that is used to describe the hardening of Pharaoh’s heart in Exodus. The translator may have looked to these connections to come up with a Greek translation of the difficult text.

If the translator were familiar with biblical poetry, particularly in the Psalms, it may have been natural for him to adjust the feel of the second stich from a declarative statement (“you love vanity”) to a rhetorical question that implies the same sentiment (“why do you love vanity?”). Similar rhetorical questions are used, for example, in Pss 2:1; 10:1; 43:2; 44:24–25; etc.

2.2 The Old Greek Ps 4:3 as Witness to the Hebrew

A more straightforward explanation of the OG is that it read the Hebrew כִּבְדֵי לֵב לָמָּה (“hard of heart”)—at the end of the first stich, and “why” at the beginning of the second stich—instead of MT’s כְבוֹדִי לִכְלִמָּה (“my glory for shame”).^[22] The two readings are extremely close from a text-critical perspective. Though spaces between words were used in manuscripts during the Second Temple period, these spaces were often inconsistent—thus the different divisions in these two readings. The use of matres lectionis, like ו in כְבוֹדִי, were not consistent during this period—thus כְבוֹדִי could be written כבדי. Finally, the letters ב and כ were often mistaken for one another, thus the ב in OG’s לֵב versus the כ in MT’s לִכְלִמָּה.

This reconstruction of the OG’s Hebrew Vorlage is not just a nifty solution to the discrepancy between the OG and the MT. Even if there were no Hebrew evidence extant for this verse, the translation data from the rest of OG Psalms would suggest this reconstruction. “Heart” (לֵב) is nearly always translated by καρδία, and where the root כבד does not have the sense of “glorify/ied, honor/ed,” it is translated with the root βαρυ* (Pss 32:4; 38:5). So while βαρυκάρδιος is a hapax legomenon, the Hebrew it most likely represents—following these translation patterns—is כִּבְדֵי לֵב, itself a hapax legomenon. For the last word of the phrase, לָמָּה is nearly always translated by Greek ἵνα τί or διὰ τί. Based on translation evidence alone, then, כִּבְדֵי לֵב לָמָּה is the most likely Hebrew behind the Greek.

Did the translator—or the person reading the text aloud to the translator—misread the consonantal Hebrew text, committing scribal errors of graphic confusion and word mis-division?^[23] Or, was he simply reading the Hebrew text as it was written in his source? This is often a difficult question to answer.^[24] Pietersma points to the fact that “[OG] clearly knows the meaning of כלמה, which he regularly (six times) renders by ἐντροπή (‘embarrassment’/’disgrace’),” to argue that the OG’s Vorlage must not have had the MT’s כלמה.^[25] However, even though the OG would not have struggled with the word כלמה, this does not rule out that he misread the כ as a ב: when reading a text, the mind processes whole words, not individual letters.^[26] The larger interpretation of the phrase and division of words, therefore, would have been a much greater factor than the identification of a single letter as כ or ב, with contextual pressures pushing him to read לב למה without scrutinizing the second letter.

Regardless, Pietersma is correct that the variant probably stems from a Vorlage that differs from the MT, and that the translator did not misread the Hebrew. This is because the Hebrew reading behind the OG here is likely an earlier reading than the MT.^[27] The MT is a little strange, while the OG’s reading fits contextually and poetically. MT’s clause בְּנֵי אִישׁ עַד־מֶה כְבוֹדִי לִכְלִמָּה (“sons of man, how long my glory into shame”) lacks a contextually appropriate verb—whether explicit or elided. There is no verb in the preceding line that is gapped (as is common in Hebrew poetry), and context does not indicate to the reader what the verb ought to be (including a copula “is”).^[28] The following clause, תֶּאֱהָבוּן רִיק (“you love vanity”), also seems to be missing an element. It is possible that עַד־מֶה (“how long”) is gapped from the earlier clause; however, together with the missing verb in the preceding clause it raises red flags.

The OG’s Vorlage, on the other hand, reads well. The first stich is verbless but the predication is clear: בְּנֵי אִישׁ עַד־מֶה כִּבְדֵי לֵב (“sons of man, how long [will you be] heavy-hearted?”).^[29] The following stich asks a parallel question whose semantics mirror the first (allowing the reader to dwell on and explore the thought^[30]): לָמָּה תֶּאֱהָבוּן רִיק (“why do you love vanity?”).^[31] Some argue that the Hebrew behind the OG is not good biblical poetry, but these concerns are unfounded.^[32] Barthélemy, for example, objects to the parallelism of OG’s Vorlage, seemingly on the grounds that parallel stichs should communicate synonymous ideas.^[33] Granted, the MT’s text presents a line-to-line semantic parallel between the psalmist’s adversaries tarnishing his reputation in the first stich and telling lies about him in the second. However, it has long been recognized that such line-to-line semantic parallelism is not, as was once thought, at the core of poetic parallelism. Correspondence at all levels of language—morphology, consonance, prosody, syntax, vocabulary, etc.—can be involved in parallelism between stichs and also at closer and further distances.^[34] Moreover, the narrow tripartite typology of “synonymous,” “antithetic,” and “synthetic” line-to-line parallelism is neither adequate descriptively, nor accurate linguistically.^[35] The parallelism of the OG’s Vorlage is, in fact, not at all out of keeping with what is found elsewhere in the Psalms. Barthélemy also argues that the pejorative designation “heavy-hearted” is not appropriate for those who are told to trust God in v. 6 (בִטְחוּ אֶל־יְהוָה, “put your trust in the LORD”). This, however, assumes that the addressees of v. 6 are the same as those of v. 3, which is neither necessary nor probable—even in the MT, v. 3 is negative toward its addressees.

To be sure, sometimes in textual criticism the more difficult reading is earlier—but in this case, the difficult reading is uncharacteristic of Hebrew poetry. Moreover, the less difficult reading cannot be explained simply as an improvement of the difficulty—it is tightly woven into the fabric of the psalm. The most plausible explanation, then, is that the MT’s reading results from scribal error. If the OG’s Hebrew is earlier, it could not have arisen from the translator misreading; it must have been the OG’s Hebrew Vorlage.

3. Psalm 4:8

The OG understands the relationship between the two stichs of Ps 4:8 differently than modern interpreters. In Hebrew, the verse is best understood as employing enjambment—the clause beginning in the first stich is finished in the second:

נָתַתָּה שִׂמְחָה בְלִבִּי מֵעֵת דְּגָנָם וְתִירוֹשָׁם רָבּוּ

You have put joy in my heart, more than the season during which their grain and wine abound.

The phrase מֵעֵת (“from the time”) and its relationship to the following words are crucial here. The noun עֵת (“time”) takes the same form whether it is bound or free. It could be taken as bound to the noun דְּגָנָם (“their grain”), together meaning “time of their grain;” however, it makes better sense contextually to read עֵת as modified by an unmarked relative clause. Unmarked relatives are not uncommon, particularly in the Psalms; the noun עֵת could be bound to the entire relative clause—its boundness signaling a restrictive relative clause^[36]—or it could be free with the relative signaled purely by context. Either reading works well. In the prepositional phrase מֵעֵת, a null element is understood from context (“more than [the joy of] the season”). This, too, is not uncommon, and the reader understands that the comparison (signaled by מִן) is not between the psalmist’s joy and a time, but between his joy and the joy of “them” when their grain and wine abound.^[37]

Although the translator of OG Psalms usually spots unmarked relatives, especially when the head-noun is bound, here the OG reads the second stich as an independent clause without a relative:

ἔδωκας εὐφροσύνην εἰς τὴν καρδίαν μου· ἀπὸ καιροῦ σίτου καὶ οἴνου καὶ ἐλαίου αὐτῶν ἐπληθύνθησαν

You gave^[38] gladness in my heart; they^[39] multiplied from the season of their grain and wine and oil.^[40]

Perhaps the OG translator overlooked the unmarked relative because it is the first to occur in the Psalms. By Ps 7, the OG recognizes these relative constructions and translates them as such.^[41] Instead, עֵת is understood as bound to דְּגָנָם (“their grain”), the preposition מִן is taken as denoting source instead of comparison, and the second stich is taken as a standalone clause with its main verb at the end.^[42]

Curiously, not a single OG manuscript in Greek has the word καιροῦ (“time”) as in the text of Rahlfs’s edition, given above.^[43] Instead, all the Greek manuscripts have καρποῦ (“fruit”). Many early Old Latin witnesses and church fathers writing in Latin, however, have the word tempore here, reflecting καιροῦ, rather than fructu, reflecting καρποῦ. Because the Old Latin tradition was translated from the Greek and is not corrected to a Hebrew tradition—at least not early on—the best explanation for the Latin’s agreement with the MT is that it reflects the original OG translation better than the Greek manuscripts themselves.^[44] This is why Rahlfs has corrected to καιροῦ (“time”). Moreover, there is a straightforward explanation for the diversity of the evidence: the words καιροῦ and καρποῦ are very similar, most of the letters being identical—especially in uncial script, where the two differing letters are mistakable, ΙΡ and ΡΠ.^[45] The contextually motivated scribal error must have occurred very early in the transmission of the Greek, with the result that the original reading is preserved only in Latin.

Despite this excellent explanation of the OG and Latin evidence, there is no harm in exploring the possibility that the variation stems from variant Hebrew. Indeed, one possible explanation, however compelling, does not rule out the possibility of other explanations, and as a point of methodology a scholar cannot stop when they discover one plausible explanation. What if the Old Latin was corrected to a Hebrew manuscript (or, more likely, a recension of the OG like Aquila), resulting in the reading tempore, as an alternative to postulating inner-Greek scribal error? If the original OG were indeed καρποῦ (“fruit”) not καιροῦ (“time”), it would reflect Hebrew פְּרִי (“fruit”) or יְבוּל (“produce”). It is not obvious how the variant Hebrew readings עֵת and פְּרִי or יְבוּל would arise; but if taw (ת) were written poorly, without much of a serif on its left stroke, it could be mistaken for resh plus yod (רי), and if ayin (ע) were written poorly, with the left vertical stroke too far to the right, it could be mistaken for peh (פ). Although these graphic interchanges are not common, neither are they impossible, especially considering that graphic error is rarely merely about mistaking letters, but also the impact of context on the word or words that the reader’s brain expects to see.^[46] The following words “grain,” “wine,” and “oil,” are foodstuffs, which could have induced the translator or a scribe to view this word as another foodstuff. A shift from מעת (to מערו) to מפרי is thus not completely unreasonable. This variant Vorlage might also explain why the OG, which notices unmarked relatives elsewhere, does not read one in Ps 4:8—with the noun פְּרִי instead of עֵת, an unmarked relative is improbable if not impossible^[47]—and takes מִן as indicating source rather than comparison.

The point of the preceding paragraph is to explore what is possible. To be clear, I do not think this is a plausible explanation. For one thing, the phrase פְּרִי דְּגָנָם, “fruit of their grain,” makes little sense: the “grain” is already the fruit/produce, and as such a phrase like this is never found in the HB.^[48] More importantly, the alternative explanation is much more convincing: the OG missed the unmarked relative but translated עֵת, with the text changing subsequently in the transmission of the Greek. From there, the straightforward mechanism of change (ΙΡ to ΡΠ in the context of reading) and external evidence (the earliest Old Latin testifies to tempore) are very compelling.

4. Conclusion

The OG is truly a fascinating collection of texts. At the crossroads of Hebrew and Greek, it is complex linguistically. And its complexity extends to the possible explanations for its phenomena. In the OG we encounter a variety of translation styles and techniques, scribal errors and scribal revisions, text-critical complexities in Hebrew and Greek manuscripts, and even the occasional exegetical move. And though generations of scholars have explored many of these possibilities, the OG’s depths have not been plumbed; it holds many further treasures yet to be discovered.