All 37 lectures of BT604E, as nothing but their summaries. For the night before an exam.
If you remember nothing else from this lecture: This course is built for the student who has finished first-year grammar, can gloss every word in a verse, and still cannot say what the sentence means — the fix is reading continuous text for clause structure, not more vocabulary. Every original-language form appears in script, transliteration and gloss (שְׁמַע / shema / "hear"; ἀποκαλύπτεται / apokalyptetai / "is revealed"), because devices and disputes both get in the way of the bare word. Ten reading chapters move through Genesis, 1 Samuel, Deuteronomy, Psalms and Amos in Hebrew, then Mark, Luke-Acts, Galatians, Romans and John in Greek, each picked for a specific problem the language poses. And where the field itself is unsettled — the Hebrew verb system, Greek verbal aspect, a disputed genitive — the course gives a working rule while saying plainly that it is one, rather than handing you a confident scheme to unlearn later.
If you remember nothing else from this lecture: The collapse this course targets is not a vocabulary gap — a student can gloss every word in a verse and still fail to read it — because the real skill missing is syntax and discourse, not more lexicon entries. Nehemiah 8:8 is the worked proof: מְפֹרָשׁ (meforash) has three competing senses ("distinctly," "in sections," "translated"), and even the last clause's וַיָּבִינוּ could mean the Levites gave understanding or the people received it — the Hebrew narrows the options without picking one, and Williamson and Blenkinsopp, both fluent readers, still land differently. That gap between what grammar settles and what it leaves open is also why Piper's strong claims for the languages and Silva's, Carson's, and Barr's more guarded ones can all come from people who read Hebrew and Greek well. The rule to carry forward: say what you saw and how sure you are — never "in the original it actually says."
If you remember nothing else from this lecture: Work a verse in a fixed order — read aloud, find the finite verbs, find their subjects, mark clause boundaries, assign what's left over — and only then open the lexicon, because a dictionary consulted first destroys the sentence's shape and lets you pick the gloss that confirms your guess. Genesis 1:1 shows why the order matters: its one finite verb (bara) and subject (elohim, plural in form but singular in agreement) are settled quickly, but the missing definite article on בְּרֵאשִׁית only becomes visible as a real question once the earlier steps are done, and it is exactly what splits the independent-clause reading (LXX, WEB, Wenham) from the construct reading (Rashi, Speiser, Holmstedt) — a genuine, unresolved grammatical crux, not something a dictionary lookup would have surfaced. Grammar narrows this question but does not close it, and neither reading by itself proves or disproves creation out of nothing.
If you remember nothing else from this lecture: Match the question to the tool — a lexicon reports usage, a grammar reports whether a construction is possible, a concordance reports distribution, an apparatus reports what the manuscripts say — and none of them tells you what a verse means; that judgment is yours. BDB's filing of תּוֹרָה under the root ירה ("to throw") is the case study for why this matters: the root is a finding aid, not a hidden true meaning, and reading etymology as meaning is a standing error. Inside any entry, the numbered senses are the lexicographer's classification of attested occurrences, not a menu — Carson's "illegitimate totality transfer" is the name for stacking every sense onto one instance — and BDAG's bold definitions (not its italic glosses) carry the actual claim about meaning. Software made searching exhaustive and instant but did not, and cannot, do the judgment work of deciding which hits are relevant.
If you remember nothing else from this lecture: Hebrew narrative prose runs on a default chain, the wayyiqtol — a verb-first form glued to what precedes it — and Genesis 22:3 shows the pure case, six such links in a row with no pause. When a narrator instead fronts a subject, object, or phrase before the verb, as in Genesis 22:1's "and God, he tested Abraham," the chain breaks to supply background, contrast, or a new scene — here, telling the reader this is a test before Abraham acts. Verbless clauses (Genesis 22:7, 22:12) state a description rather than narrate an event. But word order sometimes has no message at all: after lo, asher, or ki, the default form is grammatically unavailable, so nothing should be read into its absence there. Two real disputes bound how far to push this: whether discourse schemes (Longacre, Niccacci) codify real structure or merely absorb exceptions (Cook, Joosten), and whether Hebrew's basic order is verb-first at all (Moshavi) or only appears so because of the wayyiqtol construction itself (Holmstedt).
If you remember nothing else from this lecture: Working through Genesis 22:1-14 clause by clause turns up three distinct kinds of translation divergence, and telling them apart is the actual skill: some (the particle נָא in v. 2) come from a genuinely unresolved question about Hebrew grammar; some (the withheld name "Isaac" in v. 2, the verbless predicate in v. 12) come from English's inability to hold a Hebrew effect and are really decisions about the receiving language; and some (v. 13's אַחַר "behind" versus אֶחָד "one") are textual variants where translations simply are not reading the same consonants. Genesis 22:8's בְּנִי (vocative "my son" or appositive to "the lamb") and 22:14's יר(א)ה pair (identical consonants, different vowel-pointing, at least four defensible readings — "seen," "provided," "appears," or a qal "the LORD sees") are the two places where the Hebrew is honestly, irreducibly open — not a case of one translation being wrong, but of the text not deciding.
If you remember nothing else from this lecture: Hebrew doubles roots for emphasis (mot tamut, "dying you shall die") and repeats key words across a passage (Buber's Leitwort) as deliberate technique, not the clumsiness an English editor would cut. Some of this survives translation for free — "man"/"woman" in Genesis 2:23 happens to carry the ish/ishshah pun, and repeating "together" in Genesis 22:6 and 8 keeps the leading word intact — while the adam/adamah sound-play in Genesis 2:7 genuinely cannot be carried into English by any translator. The discipline the lecture asks for is sorting every such observation into one of three boxes — not lost, lost but footnote-repairable, or genuinely unrecoverable — before claiming any authority for "the original Hebrew."
If you remember nothing else from this lecture: Biblical Hebrew's verb system has four finite forms — qatal, yiqtol, wayyiqtol, weqatal — and three competing families of explanation (tense, aspect, discourse function), none of which accounts for every use without leaving a residue, as the performative qatal in Genesis 23:11 ("I give it to you") and the six-way "perfect problem" of §3.8 show. The Goliath description in 1 Samuel 17:4-7 is the lecture's clearest demonstration: one narrative wayyiqtol ("and he went out") opens the verse, and everything after it — "Goliath his name," the participial "clothed," "walking before him" — drops the finite verb entirely, a grammatical signal that the narrator has stepped off the story's main line until the action resumes in verse 8. The working procedure this earns you is not a rule to trust blindly but a set of expectations to check clause by clause, writing down what does not fit rather than forcing it.
If you remember nothing else from this lecture: Reading Hebrew narrative as connected discourse means watching what stands first in each clause, not just parsing verbs in isolation — a fronted subject or a verbless clause signals background or contrast, while a clause-initial wayyiqtol carries the story forward. David's speech to Saul in 1 Samuel 17:32-37 is the lecture's proof case: the same weqatal form describes repeated past action in verses 34-35 and a single future event four words later in verse 36, which no time-based account of the Hebrew verb can survive intact. A single wayyiqtol breaking into that habitual chain (verse 35's "and it rose against me") shows the narrator zooming from pattern to particular occasion, and the chapter's shift into Goliath's point of view in verse 42 shows the same grammar doing literary work. Underneath all of it sits an unresolved fact worth just knowing: the Greek text of this chapter is about forty-five verses shorter than the Hebrew, and no side of that debate has settled it.
If you remember nothing else from this lecture: Direct speech is grammatically its own kind of Hebrew — dominated by yiqtol and weqatal rather than the wayyiqtol of narration, full of first- and second-person forms and volitives, and carried by small particles (נָא, הִנֵּה, גַּם, and above all כִּי) that a dictionary gloss like "please" or "because" oversimplifies. כִּי alone covers at least six functions — causal, complementizer, temporal/conditional, adversative, the disputed asseverative, and result — with no morphological marker to tell them apart, as 1 Samuel 17:47's two back-to-back but different uses of the word show directly. The discipline the lecture asks for is the same one running through the whole course: notice the ambiguity (as with בָּא in 17:45, either "came" or "coming"), and never reach for the more dramatic reading of a particle when the plain one already works, as with the "surely" reading of כִּי that Muilenburg proposed and Barr and Aejmelaeus later reined in.
If you remember nothing else from this lecture: Legal Hebrew has two recognizable shapes — a כִּי-clause stating a condition and consequence (Deuteronomy 15:12's yimmakher ... teshallechennu), and a bare לֹא-prohibition with no condition at all (Deuteronomy 24:17's lo tatteh mishpat). Alt once tied these shapes to two different origins, Canaanite case law and Israelite covenant ceremony, but Gerstenberger and Liedke dismantled the inference from form to setting even though the formal distinction itself still holds. Layered onto both shapes is the Numeruswechsel, the shift between singular and plural "you" that English cannot show — compare 6:5's we'ahavta with 10:19's wa'ahavtem — which some read as a seam between literary sources and others read as a preacher's rhetorical device, and which even the ancient versions do not transmit consistently.
If you remember nothing else from this lecture: Deuteronomy 6:4, the opening of the Shema, is four Hebrew words with no verb and no punctuation, so "Yahweh our God, Yahweh is one" is only one of at least five grammatically permitted readings — including McBride's "Yahweh alone," contested precisely because אֶחָד is not the normal Hebrew word for "alone." Verse 5's "heart, soul, and might" is a serviceable but misleading translation: לֵבָב is will and thought more than feeling, נֶפֶשׁ is the whole living self rather than an immaterial soul, and מְאֹד ("your muchness") is a one-off noun use of a word that is everywhere else just the adverb "very." Verses 6-9 chain six commands off a single antecedent, "these words," using two merisms — sitting/walking, lying/rising — to mean "always, everywhere," and whether the resulting instructions to bind and write those words are literal (tefillin, mezuzah) or figurative is a dispute Hebrew syntax itself cannot resolve.
If you remember nothing else from this lecture: Deuteronomy attaches reasons to its commands using a small, identifiable set of particles — כִּי looking backward ("for, because"), לְמַעַן and פֶּן looking forward as mirror images ("so that" / "lest"), and עַל־כֵּן drawing a conclusion, as in Deuteronomy 24:17-18's "you shall not bend justice... [because] you were a slave in Egypt... therefore I am commanding you." Gemser first showed these motive clauses are far more common in Deuteronomy than in cuneiform law like Hammurabi's, but Sonsino corrected the conclusion Gemser drew from it: the clauses appear in Mesopotamian law too, so the real difference is one of degree, not a uniquely Israelite feature. The reasons themselves fall into recognizable kinds — memory of slavery in Egypt, promised blessing, feared consequence, the character of God, or bare classification of an act as an abomination — and Deuteronomy's recurring phrases like "today" and "I am commanding you" are the most reliable signal of where the book's own voice is speaking.
If you remember nothing else from this lecture: Parallelism is not repetition — Kugel's "A, and what is more, B" replaced Lowth's older label "synonymous" once scholars looked closely at pairs like Psalm 19:7, where torah and edut turn out to be different claims, not the same one restated. Psalm 1:1's triple line (walked, stood, sat / wicked, sinners, scoffers) shows both the power and the limit of this kind of reading: the popular "gradual descent into sin" interpretation is attractive and permitted by the grid, but nothing in the Hebrew — no adverb, no marker of progression — actually states it. Poetic Hebrew also drops signposts prose keeps, the article הַ, the relative אֲשֶׁר, and the object marker אֵת, which is part of why psalms read harder than narrative. The discipline this lecture asks for is to describe what the parallel structure permits without mistaking your own inference for the grammar.
If you remember nothing else from this lecture: Reading a psalm slowly in Hebrew exposes decisions English translations make silently. Psalm 23:6 is the clearest case: the consonants ושבתי can be pointed as "I shall return" (the Masoretic reading) or as "my dwelling" (the reading behind the Greek and nearly every English Bible, including WEB's "I will dwell") — two different vocalisations of the same four consonants, printed as if there were no choice at all. צַלְמָוֶת in verse 4 carries the same kind of unresolved dispute, between "shadow of death" and "deep darkness." And in Psalm 1:6, תֹּאבֵד agrees with the feminine דֶּרֶךְ ("way"), so the line literally says the road perishes, not the wicked person — a small grammatical fact that changes the image without changing the theology. The discipline throughout is the same: name what the Hebrew actually permits, and do not let a smooth English sentence hide that a decision was made.
If you remember nothing else from this lecture: A lexicon gloss is an argument, not a report, and the skill this lecture teaches is learning to see which kind of evidence holds it up. Most words that occur "only once" in Hebrew still have a recoverable root elsewhere and are not really mysterious; the real puzzles are the much smaller residue with no relative at all. Two cases show the range of outcomes: Proverbs 26:23's "silver dross" was cleanly solved by a Ugaritic word for glaze and a simple re-division of the consonants, while Psalm 22:16's כָּאֲרִי ("like the lion") remains genuinely unresolved between the Masoretic text, a variant כארו attested in an ancient Judean-desert scroll, and the Greek's "they dug" — and, crucially, that disputed clause is not one the New Testament ever quotes, so nothing doctrinal should be built on it. James Barr's critique of Dahood and Driver's freewheeling use of Arabic and Ugaritic cognates stands behind the whole lecture: comparative philology can genuinely solve a problem, but without controls it can also manufacture whatever meaning a scholar wants.
If you remember nothing else from this lecture: Amos 1:3's לֹא אֲשִׁיבֶנּוּ carries an object suffix, "it," that the Hebrew never identifies — WEB's "its punishment" is a supplied noun, not a translation. The same chapter shows form doing rhetorical work: all eight oracles against the nations open with the identical graded formula "for three... and for four," lulling an Israelite audience into agreement until the eighth oracle, against Israel itself, breaks the pattern with a domestic crime and four times the length. Amos 8's vision of a basket of summer fruit (qayits) turning into "the end has come" (qets) is a pun invisible in English. And the woe cry's origin — funeral lament or wisdom instruction — remains a live, unsettled question that a form alone cannot decide.
If you remember nothing else from this lecture: Amos 5:24, the most quoted line in the book, hides four undisclosed translation decisions — whether וְיִגַּל is a wish or a future, whether the water "rolls" or "wells up," whether כַּמַּיִם means literal "waters" or the freer "rivers," and above all whether נַחַל אֵיתָן pictures a wadi that never runs dry (permanence) or a torrent in flood (force). That famous demand for justice sits between a refusal of worship expressed through six stacked sense-verbs (vv. 21-23) and a question about wilderness sacrifice (v. 25) whose Hebrew fronts its object and supports at least three different readings — one of which the Septuagint had already settled by the time Acts 7 quoted it, changing "beyond Damascus" to "beyond Babylon" along the way.
l (legendum) marks the editor's own preferred reading while prp (propositum) merely records that someone has proposed a change — a distinction students routinely collapse, treating a recorded proposal as an endorsement.If you remember nothing else from this lecture: When Hebrew will not construe, work through four options in order — reread the grammar and lexicon before assuming damage, ask what the ancient versions read (after ruling out guessing, interpretation, or their own corruption), check manuscript evidence such as Qumran, and treat conjecture as the last resort, accepted only when it states a mechanism for how the error arose. Amos 6:12's redivision of בבקרים into בקר ים — "does one plough the sea with oxen?" — is the model strong conjecture: it changes no letters, restores an absurdity the verse's own argument demands, and explains how a scribe could have merged the words. By contrast, Amos 7:7-8's אנך ("plumb line," possibly Akkadian for "tin") is a case where the honest answer is to admit the word is obscure rather than force a confident gloss — a skill, not a failure.
If you remember nothing else from this lecture: Mark tells his story by stringing clauses together with καί rather than subordinating them — six καί's in the three verses of Mark 1:9-11 alone — and narrates past events with present-tense verbs (the historic present) around 150 times, far beyond any other Gospel writer. Three explanations compete for why his Greek looks this way — Aramaic influence, ordinary unliterary Koine as Deissmann's papyri revealed it, or an echo of the Greek Old Testament's own paratactic style — and over a century of argument has not settled which is right, partly because all three forces were plausibly present at once. The practical payoff is caution: a translator who subordinates one of Mark's καί clauses ("because," "when," "so that") has made an interpretive decision the Greek itself left unstated.
If you remember nothing else from this lecture: Mark 4:35-41 packs the whole chapter's grammar into seven verses — historic presents opening each new beat (λέγει, παραλαμβάνουσιν, γίνεται, ἐγείρουσιν), an imperfect (ἐπέβαλλεν) that makes the waves repeat rather than merely happen, a genuinely two-way ambiguity in verse 36 about whether Jesus was already in the boat, and a rare perfect imperative (πεφίμωσο, "be muzzled") in verse 39 whose popular "stay muzzled forever" reading is weaker than commonly preached, since even specialists in verbal aspect cannot agree what the perfect contributes. The chapter's method — read aloud, find the main verb, mark clause boundaries, then decide the hard cases in the open — is best remembered through this one worked example rather than as an abstract rule.
If you remember nothing else from this lecture: Verbal aspect (viewpoint — perfective, imperfective, stative) is not the same as tense (time) or Aktionsart (the kind of action a verb's meaning implies), and specialists disagree sharply about how much time Greek tense-forms actually encode — Porter says none, Fanning says some, Campbell reframes it as space rather than time, and Runge challenges the method behind all three. The one error worth eliminating regardless of which theory you hold is the "abused aorist," the claim that the aorist means a completed, once-for-all action: ἐν σοὶ εὐδόκησα in Mark 1:11 and the thousand-year ἐβασίλευσαν of Revelation 20:4 both refute it outright. The working procedure that follows is deliberately modest — parse the aspect, get time from context, check whether the writer had an alternative form, and say so plainly when you cannot decide.
If you remember nothing else from this lecture: Luke 1:1-4 is one carefully built period — find the main verb ἔδοξεν in verse 3 first, then rank the surrounding clauses by what depends on it, and let case endings (like the dative on παρηκολουθηκότι) tell you what attaches to what. At verse 5 the style changes completely: ἐγένετο, short coordinated clauses, and Hebraic phrasing replace the literary opening, and Theophilus vanishes. Whether that shift reflects deliberate imitation of Scripture (Plümacher), an inherited Semitic source (Wellhausen, Turner), or a preface that was never claiming historiographical grandeur in the first place (Alexander) cannot be settled from the Greek alone — and reporting that honestly is the right answer, not a dodge.
If you remember nothing else from this lecture: In Luke 15:11-24, find the finite verbs first — verse 20 alone has four aorists (εἶδεν, ἐσπλαγχνίσθη, ἐπέπεσεν, κατεφίλησεν) run together with no conjunctions, framed by a genitive absolute (ἔτι ... ἀπέχοντος) whose present-tense participle keeps the son "still far off" while the father is already running. Notice where Greek marks a distinction English cannot show — imperfects like ἐπεθύμει piling up while the famine drags on, or two different nouns (οὐσία, βίος) for the same inheritance — but do not turn every such observation into a doctrine: whether εἰς ἑαυτὸν ἐλθών means calculated self-interest (Fitzmyer) or real repentance (Jeremias) is a question about the idiom's background, not one the grammar itself settles.
If you remember nothing else from this lecture: A Greek participle tells you its aspect, voice, and what it agrees with — never how it logically relates to its main verb, so "when," "because," "although," and "and" are all things a reader supplies, not things the Greek form contains. Acts 16:6-8 shows this at its highest stakes: κωλυθέντες leaves open whether the Spirit's forbidding came before or during the journey through Phrygia and Galatia, and τὴν Φρυγίαν καὶ Γαλατικὴν χώραν leaves open whether that is one region or two — a single ambiguous noun-or-adjective call that Ramsay and Lightfoot each built into opposing theories of who received the letter to the Galatians. Matthew 28:19's πορευθέντες ("having gone") shows the same problem in miniature: whether it commands a departure or merely describes one cannot be settled by the Greek alone, no matter how confidently either side has preached it.
If you remember nothing else from this lecture: In Paul's letters the argument lives in the small words — γάρ, δέ, ἀλλά, οὖν, ἄρα, διό, ὥστε — not in the verbs, so translating every clause correctly while skipping the connectives yields accurate sentences and a lost paragraph. Mark every connective before translating, label each with a job (ground, inference, contrast, step, result), and write the skeleton first. Galatians 1:6 opens with no thanksgiving at all — asyndeton — which is a readable break in convention, though what it registered as (anger, urgency) is an inference from content, not grammar. Levinsohn and Runge's discourse-marker framework sharpens these questions usefully but should not be used to close a question the context leaves open, and the same caution applies to Betz's forensic reading of the whole letter, which Kennedy and Kern each contest from different angles.
If you remember nothing else from this lecture: Galatians 2:15-21 hinges on three points where the Greek genuinely does not decide the translation: where Paul's quoted speech to Peter stops, whether ἐὰν μή in verse 16 means "unless" or "but only," and how to construe the genitive in πίστις Χριστοῦ (taken up fully in lecture 9-3). A fourth surprise is structural rather than theological: the standard Greek editions end verse 19, not begin verse 20, with "I have been crucified with Christ," against most English versions. And verse 21's climactic "if righteousness is through law, then Christ died for nothing" is a first-class condition arguing from a premise Paul rejects, not a claim that the premise is true — exactly the kind of grammatical fact that a hasty reading turns into its opposite.
If you remember nothing else from this lecture: A Greek genitive names a relation between two nouns without specifying what the relation is, and when the head noun has verbal content (love, faith, knowledge) the genitive can be its subject or its object with no difference in form. Galatians 2:16's πίστις Χριστοῦ is the century-old test case: "faith in Christ" (objective, supported by the next clause's "we believed") against "the faithfulness of Christ" (subjective, supported by Paul's other personal genitives like Romans 3:3's "God's faithfulness"). The case ending settles nothing; what decides is a judgement about redundancy, lexical range, and the shape of Paul's larger argument. The same problem recurs in Romans 5:5's "love of God," where Augustine's reading — love God pours into us — cuts across the usual subjective/objective choice entirely.
πέπεισμαι (main) and δυνήσεται (inside the ὅτι clause); the article τῆς before ἐν Χριστῷ Ἰησοῦ is feminine and so attaches to ἀγάπης ("love"), not to the intervening masculine θεοῦ — agreement, not proximity, decides.ὥσπερ … comparison, is interrupted by the explanatory parenthesis of verses 13-17, and the answering οὕτως καί never arrives — the comparison is only completed when it restarts, compressed and without a finite verb, at verse 18.ἐφ’ ᾧ at the end of Romans 5:12 has at least four competing construals (causal "because," relative to "death," relative to "one man," and conditional/consecutive); reading (c), behind the Vulgate's in quo omnes peccaverunt and Augustine's doctrine of inherited sin, faces a serious grammatical objection since ἐπί + dative does not naturally mean "in" and the antecedent is eleven words back.If you remember nothing else from this lecture: Reading a long Pauline sentence means finding its one unsubordinated finite verb first and attaching everything else afterward, in writing — not holding the whole architecture in your head at once. Romans 8:38-39 rewards that method cleanly (agreement settles that "in Christ Jesus" modifies "love," not "God"), but Romans 5:12 shows the method meeting its limit: Paul opens a comparison with ὥσπερ, gets interrupted by five verses of explanation, and never supplies the answering clause, restarting the comparison only at verse 18. The verse's final phrase, ἐφ’ ᾧ πάντες ἥμαρτον, has been read at least four ways, and the reading behind the Latin in quo — "in whom all sinned," the basis of Augustine's argument for inherited sin — is grammatically the weakest of them, which matters because that doctrine has to rest on the larger passage, not on a single disputed preposition.
πεφανέρωται, ἥμαρτον, ὑστεροῦνται); its one main clause is "a righteousness of God has been revealed," and everything else — nine separate qualifications — hangs off that.δικαιοσύνη θεοῦ in verse 21 cannot be settled by parsing: at least four positions are seriously defended (righteousness from God, given — Bultmann, Cranfield; God's own saving activity — Käsemann; God's covenant faithfulness — Wright; God's own attribute — patristic and medieval readers), and the form θεοῦ is compatible with every one of them.πίστεως Ἰησοῦ Χριστοῦ in verse 22 repeats the Galatians 2:16 dispute (objective genitive, "faith in Christ," Dunn, printed by the WEB; subjective genitive, "the faithfulness of Christ," Hays) in its most consequential location, and the two sides disagree about the whole argument of Romans, not about any point of grammar.ἱλαστήριον in verse 25 occurs only twice in the New Testament (here and Hebrews 9:5, where it plainly means the ark's cover); reading a lexicon entry as a classification of attested usage rather than a menu of meanings, the Septuagint evidence for "mercy seat" is strong, Deissmann's inscriptional case for "propitiatory gift" was later undercut by Bailey, and Dodd's "expiation" argument, once dominant enough to shape the RSV's wording, is now contested.ἐν τῷ αὐτοῦ αἵματι can modify πίστεως (KJV/WEB's "through faith in his blood"), ἱλαστήριον/προέθετο (many modern versions' "by his blood, through faith"), or be Käsemann's evidence of an inserted formula — four major English versions punctuate the same Greek four different ways.If you remember nothing else from this lecture: Romans 3:21-26 is one Greek sentence built on a single main clause — "a righteousness of God has been revealed" — with everything else, including two of the New Testament's hardest genitive problems, hanging off it as qualification. δικαιοσύνη θεοῦ and πίστεως Ἰησοῦ Χριστοῦ both resist resolution by grammar alone, since the form is compatible with every position scholars have defended, from Bultmann's righteousness-as-gift to Käsemann's righteousness-as-God's-own-activity, and from Dunn's "faith in Christ" to Hays's "faithfulness of Christ." The word ἱλαστήριον is the lecture's showcase for how a lexicon actually works: it occurs only twice in the New Testament, the Septuagint's "mercy seat" usage is the strongest evidence, and rival readings like "propitiation" and "expiation" have shifted with the scholarship (Dodd's case for the RSV's "expiation" no longer commands the assent it once did) without ever being settled outright — a demonstration that four good English Bibles can print four different translations of the same Greek clause and every one of them be a defensible, non-careless decision.
εἰ + indicative) means the speaker regards the premise as true, licensing "since," is wrong: 1 Corinthians 15:13 states as a first-class condition the very premise Paul spends the chapter denying, and Matthew 12:27 puts one in the mouth of an opponent Jesus is refuting; Boyer's catalogue found a substantial share of New Testament first-class conditions have premises the speaker doubts or denies.εἰ + optative, ἄν + optative) anywhere in the New Testament — only fragments like 1 Peter 3:14 — because the optative mood was already dying out in Koine; a tidy fourfold textbook table that doesn't say so makes the language look neater than it is.οὐ invites the answer "yes" (Romans 3:29, where Paul then supplies ναί himself) and one opened by μή invites "no" (Romans 11:1, μὴ ἀπώσατο ὁ θεὸς…; μὴ γένοιτο) — but the rule is a strong tendency, not a mechanical guarantee, and combined negatives like Romans 10:18's μὴ οὐκ ἤκουσαν; have to be reasoned out.μὴ γένοιτο (aorist middle optative of γίνομαι with the negative, "may it not come to be") appears ten times in Romans; the King James "God forbid" is defensible seventeenth-century idiom but supplies a subject, "God," that is not in the Greek — a preacher who builds on that translation is building on the translator, not on Paul.προσωποποίια) reading of the "I" available without proving it over the autobiographical readings defended by Cranfield and others.If you remember nothing else from this lecture: Two small pieces of grammar power Paul's whole argumentative style: the conditional sentence, where the oft-repeated claim that a first-class condition ("if" + indicative) means the speaker asserts the premise is refuted by Paul's own use of one to state exactly what he is refuting (1 Corinthians 15:13); and the negative particle in a question, where οὐ invites "yes" and μή invites "no" — most visibly in the ten-times-repeated μὴ γένοιτο, "may it never be," whose King James rendering "God forbid" smuggles in a word the Greek does not have. Both feed into the diatribe, the ancient schoolroom style of arguing with an imagined objector that Bultmann first identified in Paul and that Stowers relocated from the hostile street to the pedagogical classroom; that relocation is what makes it possible, though not required, to read the anguished "I" of Romans 7 as a rhetorical voice rather than Paul's own autobiography.
If you remember nothing else from this lecture: John's Gospel is deceptively easy to construe and unusually hard to understand, because roughly a thousand words carry a book the length of Luke, so each word is loaded far beyond its dictionary sense by the time a reader reaches it. The Nicodemus scene in chapter 3 is the model case for all of John's devices at once: ἄνωθεν means both "from above" and "again," Jesus intends one sense, Nicodemus takes the other and produces an absurd literal objection, and the narrative corrects him without ever choosing between the meanings for the reader. The discipline this lecture insists on is testing every claimed double meaning or misunderstanding against whether the text itself actually uses both senses — because once you start looking, almost any Johannine word with two senses starts to look deliberate, and not all of them are.
If you remember nothing else from this lecture: John's Prologue tracks two verbs of being — ἦν for what always was (the Word) and ἐγένετο for what came into being (everything else) — right up to the hinge of verse 14, where the Word himself "became" flesh, and the whole passage is built as a staircase (anadiplosis) in which each clause's last word opens the next. At least three of its decisions are genuinely open rather than settled by grammar: where the sentence break falls between verses 3 and 4 (with modern critical editions and modern English Bibles famously on opposite sides, the punctuation itself having shifted during the fourth-century trinitarian debates), whether κατέλαβεν in verse 5 means the darkness "overcame" or "comprehended" the light, and whether μονογενής in verse 14 means "only begotten" or "one of a kind." The recurring lesson is that the earliest manuscripts had no punctuation at all, so every full stop and every translated nuance in front of you is a modern editor's or translator's judgment call, not a fact read straight off the page.
If you remember nothing else from this lecture: In John 1:1c, θεὸς ἦν ὁ λόγος, the article on ὁ λόγος alone settles that the Word is the subject and God is the predicate, which rules out reading the clause as claiming the Word and God are simply the same thing with nothing left over — but that is nearly the only thing the grammar by itself settles. The famous rule attributed to E. C. Colwell has been run backwards for seventy years by popular apologists on both sides: he showed only that a predicate noun already known to be definite is usually anarthrous when it precedes the verb, not that an anarthrous preverbal predicate is therefore definite, and even scholars who hold the traditional view of Christ's deity say so. The best-supported modern reading, argued by Harner and Wallace from distribution, is that θεός here is qualitative — describing the Word's nature rather than identifying him with the Father — but even that leaves open how much that nature amounts to, a question the rest of John's Gospel, not the article, has to answer.
If you remember nothing else from this lecture: A critical edition's apparatus is not decoration for specialists; it is the editors showing their work, because no single surviving manuscript matches exactly what any printed Greek or Hebrew Bible prints. The governing principle is that witnesses are weighed by independence, age and geographic spread rather than simply counted, which is why a huge, late, single-ancestor tradition like the Majority text counts as one witness, not a thousand. The clearest demonstration is the ending of Mark: verses 9-20 are missing from the earliest and best manuscripts, contradict Mark's own narrative logic by reintroducing a character just named, and exist in multiple competing forms — evidence that points to a later addition — yet churches have read those verses for centuries, and answering whether a reading is original, whether it is true, and whether it belongs in worship are three different questions with three different kinds of answer, not one.
If you remember nothing else from this lecture: The famous sermon claiming Jesus "came down" to Peter's level by switching from ἀγαπάω to φιλέω in John 21 collapses once you notice that three other word-pairs alternate in the same three verses (lambs/sheep, feed/tend, know/know-more-deeply) without anyone building a doctrine on them, and that the narrator explicitly names the cause of Peter's grief as the third asking, not the changed verb — and checking John's own usage elsewhere shows both verbs applied without distinction to the Father's love for the Son, Jesus' love for Lazarus, and the beloved disciple. The lesson generalizes into a whole family of word-study fallacies — the root fallacy, semantic anachronism, and Barr's "illegitimate totality transfer" chief among them — all traceable to one mistake: treating a word's dictionary range, its parts, or its history as though they determined its meaning in a given sentence, when meaning in fact comes from use in context, checked first and foremost against how the same author uses the same word elsewhere.
If you remember nothing else from this lecture: Most people who learn biblical Hebrew and Greek lose them within two or three years, not from bad teaching but because a skill fixed by retrieval decays without practice — so the whole discipline of keeping the languages comes down to reading a little, in a fixed daily slot, on real continuous text (not review charts), in a linguistically graded order that puts easy-vocabulary books like 1 John and Mark before deceptively-easy-looking John and genuinely hard books like Hebrews last. The lecture's other half is an ethical warning: telling an audience "in the Greek it actually says…" is uniquely unaccountable, since almost nobody in the room can check it, and the remedy is not silence but honesty — naming when translations differ and why, saying plainly when a point is disputed, and never making a claim you would be embarrassed to defend in front of someone who knows more Greek than you do. The Bereans of Acts 17 are the model held up at the end: not scholars, just people who checked the text daily and were willing to be shown wrong.