² tn There is some debate about the syntax of the words translated “All the people living in Jerusalem and all the people who came into Jerusalem from the towns in Judah.” As the sentence is structured in Hebrew, it looks like these words are the subject of “proclaim a fast.” However, most commentaries point out that the people themselves would hardly proclaim a fast; they would be summoned to fast (cf. 1 Kgs 21:9, 12; Jonah 3:7). Hence many see these words as the object of the verb, which has an impersonal subject “they.” This is most likely unless, as J. Bright thinks (Jeremiah [AB], 180), the word “proclaim” is used in a looser sense as “observed.” The translation has chosen to follow this latter tack rather than use the impersonal (or an equivalent passive) construction in English. For a similar problem see Jonah 3:5, which precedes the official proclamation in 3:7. Jeremiah's Hebrew text reads, “In the fifth year of Jehoiakim son of Josiah king of Judah, in the ninth month they proclaimed a fast before the Lord, all the people in Jerusalem and all the people who came from the cities of Judah into Jerusalem.” The sentence has been broken down and restructured to better conform with contemporary English style.

¹ sn Here as well some scholars work with the number 430 to try to reduce the stay in Egypt for the bondage. Some argue that if the number included the time in Canaan, that would reduce the bondage by half. S. R. Driver (Exodus, 102) notes that P thought Moses was the fourth generation from Jacob (6:16-27), if those genealogies are not selective. Exodus 6 has Levi—Kohath—Amram—Moses. This would require a period of about 100 years, and that is unusual. There is evidence, however, that the list is selective. In 1 Chr 2:3-20 the text has Bezalel (see Exod 31:2-5) a contemporary of Moses and yet the seventh from Judah. Elishama, a leader of the Ephraimites (Num 10:22), was in the ninth generation from Jacob (1 Chr 7:22-26). Joshua, Moses’ assistant, was the eleventh from Jacob (1 Chr 7:27). So the “four generations” leading up to Moses are not necessarily complete. With regard to Exod 6, K. A. Kitchen has argued that the four names do not indicate successive generations, but tribe (Levi), clan (Kohath), family (Amram), and individual (Moses; K. A. Kitchen, Ancient Orient and Old Testament, 54-55). For a detailed discussion of the length of the sojourn, see E. H. Merrill, A Kingdom of Priests, 75-79.

¹ sn The ninth plague is that darkness fell on all the land—except on Israel. This plague is comparable to the silence in heaven, just prior to the last and terrible plague (Rev 8:1). Here Yahweh is attacking a core Egyptian religious belief as well as portraying what lay before the Egyptians. Throughout the Bible darkness is the symbol of evil, chaos, and judgment. Blindness is one of its manifestations (see Deut 28:27-29). But the plague here is not blindness, or even spiritual blindness, but an awesome darkness from outside (see Joel 2:2; Zeph 1:15). It is particularly significant in that Egypt’s high god was the Sun God. Lord Sun was now being shut down by Lord Yahweh. If Egypt would not let Israel go to worship their God, then Egypt’s god would be darkness. The structure is familiar: the plague, now unannounced (21-23), and then the confrontation with Pharaoh (24-27).

¹ sn The dating of the book of Joel is a matter of dispute. Some scholars date the book as early as the ninth century b.c., during the reign of the boy-king Joash. This view is largely based on the following factors: an argument from silence (e.g., the book of Joel does not mention a king, perhaps because other officials de facto carried out his responsibilities, and there is no direct mention in the book of such later Israelite enemies as the Assyrians, Babylonians, and Persians); inconclusive literary assumptions (e.g., the eighth-century prophet Amos in Amos 9:13 alludes to Joel 3:18); the canonical position of the book (i.e., it is the second book of the Minor Prophets); and literary style (i.e., the book is thought to differ in style from the postexilic prophetic writings). While such an early date for the book is not impossible, none of the arguments used to support it is compelling. Later dates for the book that have been defended by various scholars are, for example, the late seventh century or early sixth century or sometime in the postexilic period (anytime from late sixth century to late fourth century). Most modern scholars seem to date the book of Joel sometime between 400 and 350 b.c. For a helpful discussion of date see J. A. Thompson, “The Date of the Book of Joel,” A Light unto My Path, 453-64. Related to the question of date is a major exegetical issue: Is the army of chapter two to be understood figuratively as describing the locust invasion of chapter one, or is the topic of chapter two an invasion of human armies, either the Babylonians or an eschatological foe? If the enemy could be conclusively identified as the Babylonians, for example, this would support a sixth-century date for the book.

³ tc The Byzantine text along with a few other witnesses (א³ A^c C² D² Ψ [88] 1241 1505 1739 1881 M al vg^ms) read θεός (theos, “God”) for ὅς (hos, “who”). Most significant among these witnesses is 1739; the second correctors of some of the other mss tend to conform to the medieval standard, the Byzantine text, and add no independent voice to the textual problem. At least two mss have ὁ θεός (69 88), a reading that is a correction on the anarthrous θεός. On the other side, the masculine relative pronoun ὅς is strongly supported by א* A* C* F G 33 365 1175 Did Epiph. Significantly, D* and virtually the entire Latin tradition read the neuter relative pronoun, ὅ (ho, “which”), a reading that indirectly supports ὅς since it could not easily have been generated if θεός had been in the text. Thus, externally, there is no question as to what should be considered the Ausgangstext: The Alexandrian and Western traditions are decidedly in favor of ὅς. Internally, the evidence is even stronger. What scribe would change θεός to ὅς intentionally? “Who” is not only a theologically pale reading by comparison; it also is much harder (since the relative pronoun has no obvious antecedent, probably the reason for the neuter pronoun of the Western tradition). Intrinsically, the rest of 3:16, beginning with ὅς, appears to form a hymn with six strophes. As such, it is a text that is seemingly incorporated into the letter without syntactical connection. Hence, not only should we not look for an antecedent for ὅς (as is often done by commentators), but the relative pronoun thus is not too hard a reading (or impossible, as Dean Burgon believed). Once the genre is taken into account, the relative pronoun fits neatly into the author’s style (cf. also Col 1:15; Phil 2:6 for other places in which the relative pronoun begins a hymn, as was often the case in poetry of the day). On the other hand, with θεός written as a nomen sacrum, it would have looked very much like the relative pronoun: q-=s vs. os. Thus, it may have been easy to confuse one for the other. This, of course, does not solve which direction the scribes would go, although given their generally high Christology and the bland and ambiguous relative pronoun, it is doubtful that they would have replaced θεός with ὅς. How then should we account for θεός? It appears that sometime after the 2nd century the θεός reading came into existence, either via confusion with ὅς or as an intentional alteration to magnify Christ and clear up the syntax at the same time. Once it got in, this theologically rich reading was easily able to influence all the rest of the mss it came in contact with (including mss already written, such as א A C D). That this reading did not arise until after the 2nd century is evident from the Western reading, ὅ. The neuter relative pronoun is certainly a “correction” of ὅς, conforming the gender to that of the neuter μυστήριον (mustērion, “mystery”). What is significant in this reading is (1) since virtually all the Western witnesses have either the masculine or neuter relative pronoun, the θεός reading was apparently unknown to them in the 2nd century (when the “Western” text seems to have originated, though its place of origination was most likely in the east); they thus supply strong indirect evidence of ὅς outside of Egypt in the 2nd century; (2) even 2nd century scribes were liable to misunderstand the genre, feeling compelled to alter the masculine relative pronoun because it appeared to them to be too harsh. The evidence, therefore, for ὅς is quite compelling, both externally and internally. As TCGNT 574 notes, “no uncial (in the first hand) earlier than the eighth or ninth century (Ψ) supports θεός; all ancient versions presuppose ὅς or ὅ; and no patristic writer prior to the last third of the fourth century testifies to the reading θεός.” Thus, the cries of certain groups that θεός has to be original must be seen as special pleading. To argue that heretics tampered with the text here is self-defeating, for most of the Western fathers who quoted the verse with the relative pronoun were quite orthodox, strongly affirming the deity of Christ. They would have dearly loved such a reading as θεός. Further, had heretics introduced a variant to θεός, a far more natural choice would have been Χριστός (Christos, “Christ”) or κύριος (kurios, “Lord”), since the text is self-evidently about Christ, but it is not self-evidently a proclamation of his deity. (See ExSyn 341-42, for a summary discussion on this issue and additional bibliographic references.)