Titulus

Speakers of Latin generally say titulus to mean the name of a book. This is not completely unfounded, but the much preferable word is inscriptio. (As far as verbs go, titulo and intitulo are late barbarisms, but it’s good Latin to say ‘liber inscribitur ___’ or ‘librum suum ___ inscripsit’.)

Titulus means, in the first place, a label written on hard material, like a tombstone or a statue. In Augustun period, it could refer to a book or chapter’s name: but even here, I think, the word almost always referred to the physical heading itself, and not the book or chapter’s name in the abstract. Thus Ovid wrote ‘nec titulus minio, nec cedro charta notetur’, and Martial:

Addita per titulos sua nomina rebus habebis;
Prætereas, si quid non facit ad stomachum.

Or else it had a slightly negative connotation, and referred to the book’s heading as an extraneous disguise of the book itself, just like we talk of judging a book by its cover. ‘Me non pænitet nullum festiviorem excogitasse titulum’, wrote Pliny, and I think there was a hint of disdain in his word titulum, by which he meant a pompous decoration which he had chosen not to put onto his book. Titulus was not used even in this sense by Cicero, who always used inscriptio for the name of a book . Estienne Dolet noticed this, and in his Commentarii linguæ latinæ (1536), under the entry Inscriptio, he wrote: ‘inscriptio est, quod vulgò titulus dicitur, & planè Barbarè marca.’ (1) ‘An inscriptio is what is popularly called a titulus, and utterly barbarously a marca.’ This is one of many places where we should obey the great master.

***

(1) Vol. I., col. 1272.

Theory of Dagesh

 There are two kinds of dagesh, lene and forte. Their identity in form is one of the defects of the Masoretic system of punctuation, because they have two completely different functions. Dagesh forte, which corresponds exactly to Arabic shaddah ( ّ ), doubles the consonant in question:

וַנֵּלֶךְ vanneleχ, אֶשָּׂא ʾessā, אֵלֶּה ʾēlle מִמֶּנּוּ mimmennū

The presence or absence of a dagesh forte is phonemic, not determined by the context, and integral to the meaning of the word. It would be like writing haṁock for hammock, or piḷow for pillow. Dagesh lene, meanwhile, indicates a stop (hard) rather than a fricative (soft) pronunciation of the letters בגדכפת. Not all of these distinctions are realized in all pronunciations of Hebrew, but here they are in their ideal form:

בָּהָר bāhār vs. וּבֵין ūβēin

גֵּרוֹ gērō vs. וְגַם veɣam

דְּרֹשׁ deroš vs. עַד ʿað

כִּי kī vs. לָכֶם lāχem

פֶּן pen vs. לִפְנֵי liɸnēi

תֶּשִׁי tešī vs. אֹתוֹ ʾoθō

Unlike dagesh forte, dagesh lene has no phonemic relevance, and its presence or absence is almost always predictable from the surrounding phonetic context. The rule is: if a בגדכפת letter is preceded by a vowel (even sometimes across a word boundary), then it does not get a dagesh lene. And if it is preceded by nothing, a consonant, or a mute schewa, then it does.

These two phenomena should really have been designated by different signs. This is because the בגדכפת letters themselves are theoretically susceptible to taking either a dagesh lene or a dagesh forte, or both. Unlike the letters אהחער, which cannot take a dagesh forte except under the rarest of circumstances, there is no reason, phonetic or otherwise, why בגדכפת cannot be doubled.

Thus there is a distinction, not marked in the Masoretic orthography, between בגדכפת letters that have a dagesh lene and the ones that have both a dagesh forte and a dagesh lene. (There is no such thing as a letter that has only a dagesh forte, for a reason that will be clear shortly.) In the six examples I listed above—בָּהָר, גֵּרו ,דְּרֹשׁ, כִּי and תֶּשִׁי—the בגדכפת letters only have a dagesh lene.

But here are some words with a double-dagesh:

כַּבֵּד
הַגּוֹי
מִדַּם
מִכֶּם
אַפּוֹ
עַתָּה

And here is my question: how are these to be pronounced? A grammar of Hebrew will tell you to pronounce a geminated version of the ‘hard’ variant of the consonant. Thus kabbeð, haggōi, middam, mikkem, ʾappō, ʿattā. And this is, as far as I know, the pronunciation of every attested Jewish tradition, both modern and medieval. 

But it does not follow from the rules of dagesh-placement in Masoretic Hebrew. Consider what is actually taking place when a letter is carrying a dagesh forte. It can be resolved into two like this:

ַטַּ—>טְט

The first of the geminate pair bears a mute schewa, and the second carries the vowel that the written consonant had been punctuated with. It's really no different to Italian giammai or latte; or English midday, in that it’s just a case of two identical consonants succeeding each other with no intervening vowel. Soo far so good; now let’s resolve a geminate בגדכפת letter:  

ַּדַּ—>דְד

Remember that a dagesh lene cannot exist unless it is called for by the absence of a foregoing vowel. Therefore, as the first of these two consonants is presumably preceded by a vowel, it has a soft pronunciation. But the second consonant, preceded by a mute schewa, must take a dagesh lene, just like any בגדכפת letter that appears after a mute schewa. It is to be pronounced hard.

Thus we should ideally pronounce: כַּבֵּד kaβbed, הַגּוֹי haɣgōi, מִדַּם miðdam, מִכֶּם miχkem, אַפּוֹ ʾaɸpō, עַתָּה ʿaθtā.

O Death, Where is Thy Sting?

Most of the following was well-explored in the nineteenth century, but I’ve never seen it discussed in all its aspects in one place, so here goes.

I

Hosea 14:13 goes like this:

מיד שאול אפדם ממות אגאלם אהי דבריך מות אהי קטבך שאול נחם יסתר מעיני

This is a very difficult verse, studded with several lexical difficulties. But our first guide should be its context: it is set within a prophecy of Ephraim’s ruin, so whatever interpretation we land on, it should probably be a threatening announcement of doom.

The three words that have historically caused trouble are: 

1. אהי. This is either a) an untranslatable interjection, b) the word ‘where’, akin to איה, or c) the word ‘I am’, akin to אהיה. The question is still unsettled: translations and dictionaries, even good ones and modern ones, are all at variance.

2. דבריך. ‘Your ___s.’ There is no way of knowing a priori whether דֶּבֶר, ‘plague’, or דָּבָר, ‘word/thing/affair/lawsuit’ is intended. But that דבריך comes from דֶּבֶר and not דָּבָר is made plain by the parallelism of דֶּבֶר  and קֶּטֶב at Ps 91:6. 

3. קטבך. ‘Your destruction’. This is merely a rare word, whose proper vocalization in the singular is probably קֶטֶב. 

Meanwhile there is a grammatical problem. Are the opening words מיד שאול אפדם ממות אגאלם meant to be ironic rhetorical questions (i.e. do I ransom them from the grave? or redeem them from death?)? Or are they affirmative statements that God will rescue Israel from death? And if the latter, how are they to be reconciled with the immediately following threats?

In my private view the verse is properly translated:

I ransom them from the grave, and from death I redeem them: [but] O for your plagues, O Death! O for your ruin, O grave! Comfort shall be hidden from my eyes. 

That is: ‘I save Israel from the grave and death, but only to her sorrow, for I will demand plagues from death and ruin from the grave to afflict her mercilessly withal.’ 

II

The Septuagint translates our verse like this:

ἐκ χειρὸς ᾅδου ῥύσομαι αὐτοὺς καὶ ἐκ θανάτου λυτρώσομαι αὐτούς· ποῦ ἡ δίκη σου, θάνατε; ποῦ τὸ κέντρον σου, ᾅδη; παράκλησις κέκρυπται ἀπὸ ὀφθαλμῶν μου. 

I will ransom them from the hand of Hades and redeem them from death: where is your suit, O death? Where is your sting, O Hades? Comfort is hidden from my eyes.

Δίκη, ‘(law)suit’, is a translation of דָּבָר, not דֶּבֶר. The Septuagint’s translation is thus a misapprehension of the Hebrew, especially in light of Ps 91:6 as I noted above. Moreover, it takes אהי to mean ποῦ, ‘where’, plausibly enough.

Jerome translated the verse as follows for the Vulgate: 

De manu mortis liberabo eos de morte redimam eos ero mors tua o mors ero morsus tuus inferne consolatio abscondita est ab oculis meis.

This differs in several points from the Septuagint. Jerome explained himself like this:

...In eo loco, in quo LXX transtulerunt, ubi est causa tu? et nos diximus, ero mors tua: Symmachus interpretatus est, ero plaga tua: quinta editio et Aquila: Ubi sunt sermones tui? quod hebraice scribitur DABARACH: legentes DABAR, hoc est, verbum pro DEBER, quod interpretatur mors...Pro aculeo quoque, quem nos morsum transtulimus, Symmachus ἀπαντημα, id est occursum, Theodotion et quinta editio, plagam, et conclusionem interpretati sunt.

For the word דְבָרֶיךָ, which the Septuagint translated as where is your suit?, we as ‘I will be your death’, Symmachus as: ‘I will be your plague’; and the Quinta and Aquila as: ‘where are your words?’ —the Septuagint read דָּבָר; that is, ‘word’, rather than ‘ דֶּבֶר’, which is translated ‘mors’. .. And for aculeo [קטבך], which we translated as ‘morsus’ [bite], Symmchus had ἀπαντημα, i.e. ‘meeting’, Theodotion ‘plague’, and the Quinta ‘shutting-up’.

Thus Jerome correctly noticed the Septuagint’s error at דְבָרֶיךָ; probably not independently, but following on Symmachus’ unique translation of the verse. Apparently only Symmachus had rendered דְבָרֶיךָ as ‘plagues’ rather than as one of the variants of דָּבָר. Jerome happened to prefer his translation to that of the Septuagint and all of the other hexaplar translators—which was fortunate, as it happens to have been the right one. Meanwhile, he had the good sense to reject Symmachus’ strange translation of קטבך.

[By the way, here are Frederick Field’s back-translated reconstructions of the hexaplar variants at this verse:

Aquila. ἔσομαι ῥήματά σου, θάνατε, ἔσομαι δηγμοί σου, ᾅδε.

Symmachus. ἔσομαι πλήγή σου ἐν θανάτῳ, ἔσομαι ἀκηδία σου ἐν ᾅδῃ.

Theodotion. καὶ ἔσται ἡ δίκη σου ἐν θανάτῳ, καὶ πληγή σου ἐν ᾅδῃ.

Quinta. ποῦ οἱ λόγοι σου...]

III

That would seem to be all, but I have not yet mentioned what is by far the most famous ancient commentary on this verse: Paul’s quotation of it at I Corinthians 15:54–6. In the Nestlé-Aland edition we read:

ὅταν δὲ τὸ φθαρτὸν τοῦτο ἐνδύσηται ἀφθαρσίαν καὶ τὸ θνητὸν τοῦτο ἐνδύσηται ἀθανασίαν, τότε γενήσεται ὁ λόγος ὁ γεγραμμένος· κατεπόθη ὁ θάνατος εἰς νῖκος. ποῦ σου, θάνατε, τὸ νῖκος; ποῦ σου, θάνατε, τὸ κέντρον; τὸ δὲ κέντρον τοῦ θανάτου ἡ ἁμαρτία, ἡ δὲ δύναμις τῆς ἁμαρτίας ὁ νόμος.

But when this corrupt thing puts on incorruption, and this mortal thing puts on immortality, then will come about the saying that is written: Death is swallowed up into victory [Isaiah 25:8]. 

O death where is thy victory? O death, where is thy sting? The sting of death is sin, and the power of sin is the Law.

This passage displays strong textual variation in the manuscripts, including an inconsistent ordering of the twin ποῦ σου θάνατε clauses. To keep to one of them, I think that we should probably not read (θάνατε, τὸ) νῖκος, victory, but rather …τὸ νεῖκος, suit; controversy. In order to clarify this error, look at the Hebrew original of Isaiah 25:8, which Paul quotes immediately before our verse:

בִּלַּע המות לנצח

Κατεπόθη ὁ θάνατος εἰς νῖκος is indeed a correct translation of this. (Κατέπιεν ὁ θάνατος ἰσχύσας, in the Septuagint, is an acceptable alternative.) Εἰς νεῖκος, meanwhile, cannot be the translation of לנצח by any stretch.  But as for Paul’s following citation, אהי דבריך מות cannot easily be translated as ποῦ σου, θάνατε, τὸ νῖκος; and a much better translation is νεῖκος, which is in line with the Septuagint’s understanding of דבריך as a suffixed form of דָּבָר. Paul was probably relying on a manuscript of the Septuagint that had νεῖκος instead of its synonym δίκη. (Where he found the variant translation of Isaiah 25:8 is something that I’d be very happy to learn.) Anyway, I suspect that the two words νῖκος and νεῖκος in I Corinthians have been confused in most manuscripts as a result of cross-contamination between proximate verses. This is extremely understandable on palaeographic grounds; so much so, in fact, that it would be surprising if the manuscript tradition had managed to get the two words straight.

Pier Vettori, writing in the sixteenth century, observed that Jerome’s translations implied that he had read  νεῖκος in both places in his manuscripts of I Corinthians. Vettori also cited some other authors who had been under the same impression, and a manuscript that had the same reading. Most of our received texts, in contrast, tend to have νῖκος in both places. It might be tempting to conclude immediately that Paul originally wrote both νῖκος and νεῖκος, which is the correct reading as far as conformity to (the Septuagint’s reading of) the Hebrew is concerned. But this is only one of three possibilities:

1. Paul wrote νῖκος / νεῖκος, but the earliest copies of I Corinthians merged them both into either νῖκος or νεῖκος.

2. The manuscripts used by Paul had νεῖκος / νεῖκος or νῖκος / νῖκος, and he copied what he saw.

3. The manuscripts used by Paul had νῖκος / νεῖκος, and he made a mistake.

All the same, given the existence of both variants in the manuscripts of Corinthians, I think that possibility 1 is the most likely after all. Paul probably wrote ‘κατεπόθη ὁ θάνατος εἰς νεῖκος. ποῦ σου, θάνατε, τὸ νῖκος;’, and variations of this represent scribal corruptions.

Apart from this, there are some other variants. For instance, the first θάνατε is replaced in some manuscripts by ᾅδη. But I am inclined to think that this and several other variants result from scribal attempts to bring Paul’s phrasing into line with the Septuagint’s. The double θάνατε, both for its awkwardness and for its marked deviation from the Septuagint, is the harder reading. And as we have already seen in the case of the Isaiah 25:8 citation, there is no reason to assume that Paul was copying faithfully from any single manuscript of the Septuagint or any other Greek translation of Hosea or Isaiah.

Conjecture at Horace, C. I.xxxii.15

Horace, C.I.xxxii.13–16, according to the mss.:

O decus Phœbi et dapibus supremi
Grata testudo Iovis o laborum

Dulce lenimen mihi cumque salve
Rite vocanti.

This is gibberish. Lachmann made things better, writing obiter in his comment on Lucr. V.311: 

…præterea cumque nisi cum relativis coniunctum lingua Latina non agnoscit: neque Horatius in carminum 1,32 potuit dicere mihi cumque salve (volunt enim hoc esse ‘salve quotiens te advoco’: at cur lyra alias ei non salveat?), sed scripsit o laborum Dulce lenimen medicumque salve Rite vocanti. 

Now I propose:

O decus Phoebi et dapibus supremi
Grata testudo Iovis, o laborum
Dulce lenimen melicumque salve
Rite vocanti.

O grace of Apollo, O lyre beloved at the feasts of exalted Jupiter, O sweet and songful lightener of suffering, rejoice in him who duly calls thee.

Update: A loyal friend of mine has pointed out that T. J. B. Brady made this very emendation in the late nineteenth century. My rabbi once told me that you shouldn’t curse your luck in such a situation for being preempted, but thank God that you were allowed to attain the same insight as one of your elders and betters.  

New word submitted to the OED: Gargalism

Gargalism, n.

Something gargled; a mouthwash. 

 It is attested in the funeral sermon for Lancelot Andrewes, 1626:

And true Religion is no way a gargalisme onely, to wash the tongue and mouth, to speake good words: it must root in the heart, and then fructifie in the hand; else it will not cleanse the whole man.

John Buckeridge, A Sermon Preached at the Funeral of the Right Honorable and Reverend Father in God Lancelot, Late Lord Bishop of Winchester, in the Parish Church of St. Savior in Southwarke, On Saturday being the XI. of November, A. D. MDCXXVI (London: Richard Badger, 1629), p. 7.

This word would seem to have the same meaning of gargarism, n., and could perhaps be listed with it as a variant form. Its figurative connotation, however, is not present in any of the examples listed under sense 1 of gargarism.

Letter to the OED on Torpedo

I’ve written this letter to the Oxford English Dictionary about part of their entry on torpedo:

 b. figurative. One who or that which has a benumbing influence.

a1593   C. Marlowe Edward II (1594) sig. C2v   Faire Queene forbeare to angle for the fish..I meane that vile Torpedo, Gaueston.

1762   O. Goldsmith Life R. Nash 34   He used to call a pen his torpedo, whenever he grasped it, it numbed all his faculties.

c1855   B. S. Hollis Hymn-bk. C'tess Huntingdon's Connecticut Pref.   The torpedo of formality had benumbed the churches.

Sense 1b of the existing entry ‘torpedo’ does not exist as a coherent category, and the quotations listed under it need to be resolved into at least two clearly distinguishable senses. The first of these is merely ‘numbness’, and is derived directly from the primitive sense of Latin torpēdō, defined by the Oxford Latin Dictionary as ‘A state of inertness, sluggishness, lethargy’. The citation from Hollis belongs plainly to this sense, as ‘torpedo’ does not refer to any agent, abstract or concrete, but is merely in (genitive) apposition to ‘formality’. The citation from Goldsmith about Nash is on the borderline, I think, and might belong to the distinct sense ‘that which has a benumbing influence’; but should probably just be folded into the general sense of ‘numbness’. Torpedo was already conceived of in Latin as having an active benumbing force on a person; cf. Sallust’s ‘si tanta torpedo animos obrepsit’ or Tacitus’ ‘tanta torpedo invaserat animum’. Incidentally, the Goldsmith quotation seems to be related somehow to the following passage in Boswell’s Life of Johnson, which should be added to the list of attestations:

‘It has been circulated, I know not with what authenticity, that Johnson considered Dr. Birch as a dull writer, and said of him, “Tom Birch is as brisk as a bee in conversation; but no sooner does he take a pen in his hand, than it becomes a torpedo to him, and benumbs all his faculties.”’ (ed. 1791, pp. 85–6.)

I think it is highly unlikely that Johnson or Nash (if either of them indeed said this) had the electric ray in mind; far likelier they were just using a Latin word for oppressive numbness. Therefore it is questionable whether the sense of the Nash and Hollis attestations should be described as ‘figurative’, as it has nothing to do with fish and does not rely for its meaning on the abstract application of any concrete image. 

As for the citation from Marlowe, it does not belong to the other two at all, because it is unmistakably a reference to the electric ray. However, that this reference is mischaracterized by the definition ‘one who or that which has a benumbing influence’ is made clear by the full context:

Faire Queene forbeare to angle for the fish,
Which being caught, strikes him that takes it dead,
I mean that vile Torpedo, Gaueston,
That now I hope flotes on the Irish seas.
(ed. 1594, fol. C2 v.)

Marlowe’s torpedo doesn’t benumb you—it strikes you dead! In any case, I am not convinced that this attestation deserves even to be called figurative, as it rather seems to be a literal reference to the fish, to which Gaveston is likened in a simile. Thus it can arguably be listed under sense 1a.

Parsing Problems

It was a frigid January day in 2015, and I was sitting in my Herodotus seminar in college. It was my turn to translate. I was very proud of how well I’d prepared, and began confidently reading the Greek aloud. But after three words of my English translation, I was stopped short by the august professor. ‘Mr Nathan’, he said, ‘you have a parsing problem’.

I have hardly passed a day of my life without a parsing problem. It started when I realized that I had no way of knowing whether ‘the skies are not cloudy all day’ meant that the skies stayed cloudless for the whole day, or whether they were only cloudy for some of the day, but not all of it. The truth is, as soon as you start thinking about problems of phrasing, only a draught from the river Lethe will let you stop. We are surrounded every day by ambiguous syntax, and alas there is much less pleasure in solving parsing-problems than there is pain in noticing them. But at least I can make you share in some of my suffering.

In dulci jubilo, loveliest of all Christmas songs, contains the following stanza:

O Iesu parvule
Nach dir ist mir so weh:
Tröst mir mein Gemüte
O puer optime
Durch alle deine Güte
O princeps gloriæ,
Trahe me post te, trahe me post te.

O little Jesus, how I am pained for thee. Console my spirit, o greatest boy, through all thy goodness. O prince of glory, drag me after thee, drag me after thee. 

Here we have the stanza in its very earliest attestation from the fourteenth century:

If my palaeography hasn’t failed me, this says:

O Iesu parvule • nach dir ist mir so we • droste myn gemüde tu puer inclite • daȥ dů dorch dyne gůde • tu princeps glorie • Trahe me post te trahe me post te • in dines fader ryche o pater optime.

O little Jesus, how I am pained for thee. Console my spirit, thou glorious boy, through thy goodness, thou prince of glory. Drag me after thee, drag me after thee into thy father’s kingdom, O greatest father!

In both versions, the words trahe me post te are a little odd: where does that thought come from? The answer is the Song of Songs, 1:4.

מָשְׁכֵנִי אַחֲרֶיךָ נָּרוּצָה

Word for word, this means:

[Draw me] [after thee] [let us run]

But these words admit two different interpretations, depending on where one places a pause. Either we have:

Draw me after thee! Let us run.

or else

Draw me! Let us run after thee.

The Masoretes clearly preferred the second interpretation, and punctuated the half-verse like this:

מָשְׁכֵ֖נִי אַֽחֲרֶ֣יךָ נָּר֑וּצָה

The shophar holech under אַחֲרֶיךָ joins it to the following word נָּרוּצָה, which is marked with an atnah. The half-verse is thus apparently to be translated, ‘Draw me! Let us run after thee.’ (Incidentally, I don’t know why people use the word pointillation to describe the Masoretic marks, when we already have the word punctuation to hand; which comes from the same root, means the same thing, and is much more readily understood.)

 The Septuagint, for its part, translated the verse as follows: 

εἵλκυσάν σε ὀπίσω σου [εἰς ὀσμὴν μύρων σου] δραμοῦμεν.

This would suggest the opposite reading from the Masoretic text, as the interpolation (here in brackets) makes ὀπίσω σου modify εἵλκυσαν, not δραμοῦμεν. (The third-person plural aorist form εἵλκυσαν, just like the interpolation, is an eccentricity of its own, but we can’t stop to discuss it now.)

And at last, the Vulgate:

Trahe me post te curremus

Utter ambiguity. In sharp contrast to the scrupulously digested Masoretic text, Latin Bibles of the Middle Ages were systematically under-punctuated. They left questions of phrase division completely open to interpretation, so that it was left to any commentator to decide which phrasing he preferred. Thus the fourteenth-century composer of In dulci jubilo picked one interpretation, but he might as well have picked another. If we insist on the parsing as it is given in the Masoretic text, then he was wrong to quote the verse as he did. Trahe me and post te belong to two different sentences, and do not form a single phrase. But at least he is supported the authority of the Septuagint’s reading, which I suspect got nearer to the truth after all.

This is of relatively little importance, but there are other examples where the ambiguous division of a Biblical phrase could have much more serious consequences. The most famous instance of all is Isaiah 40:3:

ק֣וֹל קוֹרֵ֔א בַּמִּדְבָּ֕ר פַּנּ֖וּ דֶּ֣רֶךְ יְהוָ֑ה

If we observe the Masoretic punctuation, we must put a break at the zakef katon and translate this as:

A voice cries: prepare the way of the Lord in the wilderness.

But the Septuagint had another idea:

φωνὴ βοῶντος ἐν τῇ ἐρήμῳ ῾Ετοιμάσατε τὴν ὁδὸν κυρίου.

A voice of one that cries in the wilderness: prepare the way of the Lord.

And it was the Septuagint that got quoted by all four writers of the Gospel, making the ‘voice of one that cries in the wilderness’ into a prophecy of John the Baptist. For instance (Matthew 3:1–3):

In those days came John the Baptist, preaching in the wilderness of Judaea, and saying, Repent ye: for the kingdom of heaven is at hand. For this is he that was spoken of by the prophet Esaias, saying, The voice of one crying in the wilderness, Prepare ye the way of the Lord, make his paths straight.

I don’t want to get into the question of which parsing is better. (It has been controversial for many centuries, and a standard point of dispute between pre-modern Christians and Jews.) But it is enough to observe that one of the basic prophecies of the New Testament, repeated in all four Gospels, depends on the correct parsing of an ambiguous verse. 

Finally consider a pair of verses in the Song of Moses, where God proclaims (Deuteronomy 32:40–41): 

כִּֽי־אֶשָּׂ֥א אֶל־שָׁמַ֖יִם יָדִ֑י  וְאָמַ֕רְתִּי חַ֥י אָנֹכִ֖י לְעֹלָֽם׃
אִם־שַׁנּוֹתִי֙ בְּרַ֣ק חַרְבִּ֔י  וְתֹאחֵ֥ז בְּמִשְׁפָּ֖ט יָדִ֑י
אָשִׁ֤יב נָקָם֙ לְצָרָ֔י  וְלִמְשַׂנְאַ֖י אֲשַׁלֵּֽם׃

Here are the translations of this passage in the LXX, the Vulgate, and the KJV:

ὅτι ἀρῶ εἰς τὸν οὐρανὸν τὴν χεῖρά μου καὶ ὀμοῦμαι τῇ δεξιᾷ μου καὶ ἐρῶ Ζῶ ἐγὼ εἰς τὸν αἰῶνα, ὅτι παροξυνῶ ὡς ἀστραπὴν τὴν μάχαιράν μου, καὶ ἀνθέξεται κρίματος ἡ χείρ μου, καὶ ἀνταποδώσω δίκην τοῖς ἐχθροῖς καὶ τοῖς μισοῦσίν με ἀνταποδώσω.

Levabo ad cælum manum meam et dicam vivo ego in æternum. Si acuero ut fulgur gladium meum et arripuerit iudicium manus mea reddam ultionem hostibus meis et his qui oderunt me retribuam.

For I lift up my hand to heaven, and say, I live for ever. If I whet my glittering sword, and mine hand take hold on judgment; I will render vengeance to mine enemies, and will reward them that hate me.

This is a case where the classical tradition is wrong, from the Septuagint right down to the KJV. These translations all take חַ֥י אָנֹכִ֖י לְעֹלָֽם׃, ‘I live forever’, to be an independent proposition, as if it were of great importance for God to say such a thing about himself. But it is no such thing. It was Rashi who pointed this out:

כי אשא אל שמים ידי. כִּי בַחֲרוֹן אַפִּי אֶשָּׂא יָדִי אֶל עַצְמִי בִּשְׁבוּעָה: ואמרתי חי אנכי. לְשׁוֹן שְׁבוּעָה הוּא, אֲנִי נִשְׁבָּע חי אנכי:

For I lift up my hand to heaven. That is, ‘in my wrath I lift up my hand in an oath’. And I say, I live forever. This is the language of an oath, as if to say, ‘I swear: as I live...’

He was absolutely right. חַ֥י אָנֹכִ֖י לְעֹלָֽם, I live forever, is not an emphatic proposition: it is the formulaic introduction to an oath. As God lives is the classical-oath formula in the Bible, and here we merely have an instance of its use in the first person. This is made even clearer by God’s lifting his hand: the standard Biblical oath-gesture. Sforno, agreeing with Rashi, pointed to a very similar passage at Daniel 12:17:

וָֽאֶשְׁמַ֞ע אֶת־הָאִ֣ישׁ ׀ לְב֣וּשׁ הַבַּדִּ֗ים אֲשֶׁ֣ר מִמַּעַל֮ לְמֵימֵ֣י הַיְאֹר֒ וַיָּ֨רֶם יְמִינ֤וֹ וּשְׂמֹאלוֹ֙ אֶל־הַשָּׁמַ֔יִם וַיִּשָּׁבַ֖ע בְּחֵ֣י הָֽעוֹלָ֑ם כִּי֩ לְמוֹעֵ֨ד מֽוֹעֲדִ֜ים וָחֵ֗צִי וּכְכַלּ֛וֹת נַפֵּ֥ץ יַד־עַם־קֹ֖דֶשׁ תִּכְלֶ֥ינָה כָל־אֵֽלֶּה׃

And I heard the man dressed in linen who was above the waters of the Nile, and he lifted his right and left hand up to the heavens: and he swore on the One who lives forever that...

We should therefore read our verses: I lift my hand up to heaven, and I say: as I live forever, if I whet my sword, etc. 

The Septuagint was the first translation to get it wrong, and all subsequent translations followed suit. They definitively exclude the correct logical construction of these verses. Whether the Masoretic text can be said to be wrong too is a nice question: usually after oath-formulas there is no soph-pasuk, but on the other hand the poetic form of the passage imposed its own requirements of verse-division beyond the sense alone. 

Here I’ve discussed only three examples. But parsing problems are an extremely common phenomenon, and come up constantly in the history of the Bible’s interpretation. And elsewhere: I’m now editing a short book from the 1530s, which was printed with reasonable care, and still it is filled from beginning to end with fiendish problems of parsing and punctuation. To take one example at random:

Which would seem to mean: ‘above all [the damsels] should take care to pray at the very least that they not declare their wills so soon’, etc. Awful. But put a full stop after prier, and the whole passage becomes much happier: ‘they should force their lovers to beg for them. At the very least, they should not declare their wills so soon, but should dissemble as long as possible.’

In my experience, problems like this are actually much more frequent than garbled words and variant letters. Open up the Septuagint alongside the Masoretic text, and you’ll notice discrepancies by the dozen. And problems of this nature are also more difficult to correct: as soon as you undertake to repair a badly punctuated sentence, you find yourself clueless in a labyrinth of uncertainties, prey to the Minotaur of nonsense. If you do arrive at the right solution, it is always very hard to state the case in an elegant or convincing way to your peers; much harder than with other emendations. 

So pay attention to parsing, and always ask yourself whether a given attempt at representing it by punctuation it is correct. You might have all of the words that an ancient author wrote in pristine form: but unless you can establish the phrasing with which they are to be pronounced, you do not yet know what they mean.