FREE

Course certificates are free until March 31, 2027.

BT604E Reading the Biblical Languages

Lecture 11-3: The Article, and What It Does and Does Not Prove

Contents 37
BT604E Reading the Biblical Languages
Chapter 0 · Introduction: Why Read the Originals
  1. Introduction: Why Read the Originals
Chapter 1 · Chapter 1: Reading, Not Decoding
  1. Lecture 1-1: What This Course Assumes, and What It Is For
  2. Lecture 1-2: How to Work Through a Verse Without Drowning
  3. Lecture 1-3: The Tools on Your Desk: Lexicon, Grammar, Concordance, Apparatus
Chapter 2 · Chapter 2: Hebrew Narrative - Genesis
  1. Lecture 2-1: The Shape of a Hebrew Sentence: Word Order and the Narrative Chain
  2. Lecture 2-2: Reading Genesis 22:1-14
  3. Lecture 2-3: Repetition, Wordplay, and What Translation Cannot Carry
Chapter 3 · Chapter 3: Hebrew Verb and Discourse - 1 Samuel
  1. Lecture 3-1: The Verb System in Action: Aspect, Sequence, and the Perfect Problem
  2. Lecture 3-2: Reading 1 Samuel 17 as Connected Discourse
  3. Lecture 3-3: Direct Speech, Particles, and the Texture of Dialogue
Chapter 4 · Chapter 4: Hebrew Law - Deuteronomy
  1. Lecture 4-1: Legal Hebrew: Casuistic and Apodictic Formulations
  2. Lecture 4-2: Reading Deuteronomy 6:4-9
  3. Lecture 4-3: Motive Clauses and the Rhetoric of Command
Chapter 5 · Chapter 5: Hebrew Poetry - Psalms
  1. Lecture 5-1: Parallelism, and Why It Is Not Repetition
  2. Lecture 5-2: Reading Psalm 23 and Psalm 1
  3. Lecture 5-3: Poetic Vocabulary, Rare Words, and the Limits of the Lexicon
Chapter 6 · Chapter 6: Hebrew Prophecy - Amos
  1. Lecture 6-1: Prophetic Forms: Oracle, Woe, Vision Report
  2. Lecture 6-2: Reading Amos 5:18-27
  3. Lecture 6-3: When the Text Is Hard: Corruption, Conjecture, and the Apparatus
Chapter 7 · Chapter 7: Greek Narrative - Mark
  1. Lecture 7-1: Koine on the Page: Mark's Sentences and Why They Read as They Do
  2. Lecture 7-2: Reading Mark 4:35-41
  3. Lecture 7-3: Aspect Over Time: The Present, Aorist and Imperfect at Work
Chapter 8 · Chapter 8: Greek Historiography - Luke and Acts
  1. Lecture 8-1: A More Literary Koine: Luke's Preface and the Septuagintal Style
  2. Lecture 8-2: Reading Luke 15:11-24
  3. Lecture 8-3: Participles: The Engine of Greek Narrative
Chapter 9 · Chapter 9: Greek Letter - Galatians
  1. Lecture 9-1: Reading a Letter as Argument: Connectives and Signposts
  2. Lecture 9-2: Reading Galatians 2:15-21
  3. Lecture 9-3: Cases in Dispute: The Genitive and What Hangs on It
Chapter 10 · Chapter 10: Greek Argument - Romans
  1. Lecture 10-1: Long Sentences: Subordination, Anacoluthon, and Keeping the Thread
  2. Lecture 10-2: Reading Romans 3:21-26
  3. Lecture 10-3: Conditional Sentences, Questions, and the Diatribe
Chapter 11 · Chapter 11: The Greek of John
  1. Lecture 11-1: Simple Vocabulary, Difficult Meaning: The Johannine Style
  2. Lecture 11-2: Reading John 1:1-14
  3. Lecture 11-3: The Article, and What It Does and Does Not Prove
Chapter 12 · Chapter 12: Text, Words, and Honest Use
  1. Lecture 12-1: The Apparatus: Reading a Textual Note Without Being Misled
  2. Lecture 12-2: Word-Study Fallacies and How to Avoid Them
  3. Lecture 12-3: Keeping the Languages After the Course Ends

Jump to the key points ↓

49 min

This lecture is about a two-letter word. In John 1:1 the noun θεός appears twice: once with the article, once without. Whole libraries have been built on that difference, and a large proportion of what has been built on it is unsound in ways that a student who has finished an introductory grammar is in a position to detect.

I want to be exact about what this lecture is doing, because the subject makes exactness hard.

It is not going to tell you what to believe about the Word. That question is settled — for those who hold it settled — by the whole of John's Gospel, by the rest of the New Testament, by the councils of the church, and by things that are not grammar at all. What a lecture on Greek grammar can do is much narrower and, done properly, much more useful: it can tell you which readings of a sentence the language excludes, which it permits, and what kind of consideration would have to do the choosing among the ones it permits.

Grammar rules out. Grammar does not choose. You met that formula in lecture 9-3 and again in 10-2, and here it gets its hardest test, because at this verse the pressure to make grammar do more than it can comes from several directions at once. It comes from preachers who want a proof-text. It comes from apologists on both sides of a doctrinal boundary, each armed with the same rule and each stating it wrongly. And it comes from the perfectly honourable desire of a Christian reader to find that the Greek supports what they already hold — a desire I share and which is exactly why the discipline matters.

So: what the article is, what it does, what its absence does and does not signal, what E. C. Colwell actually wrote in 1933, what people have made him say since, and where the grammar of John 1:1c leaves us. Nothing in this lecture, and nothing on the chapter quiz, requires you to reach any particular conclusion about the christology.

What these 49 minutes give you

  • Describe at least six distinct functions of the Greek article, and explain why the absence of the article carries far less information than an English speaker's instinct suggests.
  • State how a Greek sentence with two nominatives marks which is the subject, apply it to John 1:1c, and say what reading that alone rules out.
  • State Colwell's rule in the direction Colwell stated it, identify the logical error involved in reversing it, and explain why the reversed form has been attractive to people arguing for opposite conclusions.
  • Set out the qualitative analysis of anarthrous preverbal predicate nouns, naming Harner, Dixon and Wallace, and describe the circularity risk in the statistics that support it.
  • Give a fair grammatical assessment of rendering John 1:1c "a god," naming defenders and critics, and say precisely which part of the assessment is grammatical and which is not.

In the beginning was the Word, and the Word was with God, and the Word was God. — John 1:1

World English Bible

No one has seen God at any time. The one and only Son, who is in the bosom of the Father, he has declared him. — John 1:18

WEB

God is spirit, and those who worship him must worship in spirit and truth.” — John 4:24

WEB

Thomas answered him, “My Lord and my God!” — John 20:28

WEB

Ἐν ὁ , καὶ ὁ λόγος ἦν πρὸς τὸν , καὶ ἦν ὁ λόγος.

En arche en ho logos, kai ho logos en pros ton theon, kai theos en ho logos

I am deliberately not offering a working translation of the third clause. Any English rendering of θεὸς ἦν ὁ λόγος has already answered the question this lecture exists to examine, and I would rather set the question out first. What I will translate is the shape: "[θεός] was the Word," with the noun left in Greek and the word order preserved, so that you can see that the predicate stands first, before the verb, and that it has no article, while the subject stands last and does.

Three further verses will be needed.

θεὸν ἑώρακεν πώποτε· θεὸς ὁ ὢν εἰς τὸν τοῦ ἐκεῖνος .

theon oudeis heoraken popote; monogenes theos ho on eis ton kolpon tou patros ekeinos exegesato — "God no one has ever seen; an only [one], God, the one who is in the bosom of the Father, that one has made him known."

My working translation is clumsy on purpose, and it differs from the World English Bible on a point of text, not only of rendering. The WEB has "the one and only Son," which reflects the reading ὁ μονογενὴς . The modern critical editions print μονογενὴς θεός, on the strength of the earliest Greek manuscripts. Section 10 sets the variant out. Note for now that θεός appears here without the article, and that nobody proposes to translate it "a god."

ὁ θεός

pneuma ho theos — "spirit [is] God," or in English word order, "God is spirit."

This is the same construction as John 1:1c with the verb omitted altogether: an anarthrous predicate noun before an articular subject. Notice that no English version renders it "God is a spirit and no other kind of being," nor "God is the Spirit," identifying him with the third person. Translators read πνεῦμα here as describing what God is like in kind. Hold that observation; it is the whole of section 8.

ὁ μου καὶ ὁ θεός μου

ho kyrios mou kai ho theos mou — "my Lord and my God"

Here θεός has the article. Section 11 says why that is worth noticing and also why it settles less than it is often said to.

1. The Greek article is not the English "the"

Begin by dismantling an instinct.

An English speaker learns that ὁ, ἡ, τό is "the," and that Greek has no word for "a." Both halves of that are true and the combination is misleading, because the two halves do not have equal consequences.

English has a two-way system: "the dog" and "a dog" are both marked, and the choice between them is forced. Every English count noun in the singular must carry one determiner or the other, so the presence of "a" is a positive signal that the noun is indefinite.

Greek has a one-way system. There is a definite article and there is no indefinite article. A Greek noun either has the article or it does not, and the second condition covers everything the first does not: nouns that are indefinite, nouns that are definite but do not need marking, nouns that are qualitative, nouns after prepositions, nouns in fixed phrases, proper names, abstract nouns, predicate nouns, and more.

The consequence is asymmetric and it is the foundation of this whole lecture. The presence of the article is informative. Its absence is much less so. When a Greek writer uses the article, he has chosen to; when he does not, he may have chosen not to, or the construction may not have called for one, or the idiom may forbid it.

Two rules of thumb from the reference grammars make this concrete.

Apollonius'

Apollonius' canon, named after the ancient grammarian Apollonius Dyscolus, observes that in a two-noun genitive construction the two nouns tend to agree in articularity: either both have the article or neither does. So λόγος and ὁ λόγος τοῦ θεοῦ are the natural pairs, while the mixed forms are rarer. That means the absence of the article on θεοῦ in a phrase like θεοῦ (John 1:12) tells you almost nothing about definiteness — it is being conditioned by the head noun.

Nouns after prepositions very frequently drop the article without any loss of definiteness. ἐν ἀρχῇ in John 1:1 is not "in a beginning." παρὰ θεοῦ in John 1:6 is not "from a god." ἐκ θεοῦ in John 1:13 is not "out of a god."

So, before we go a step further: John 1 contains at least three anarthrous occurrences of θεός in the first fourteen verses which nobody, from any tradition, translates with an indefinite article. That is not an argument about verse 1. It is a demonstration that the bare form of an argument — "no article, therefore indefinite" — is not a valid form.

2. What the article actually does

At least six functions, all of them present in John's Prologue, all of them ones you have already met without naming.

1. Identifying, or individualising. The commonest use: it points to a particular referent. ὁ λόγος throughout the Prologue.

2. Anaphoric — pointing back. A noun is introduced without the article and then picked up with it. Lecture 11-2 noticed this at John 1:4: ἦν, then ἡ ζωή. Wallace calls this the article's most common single function in narrative, and it is worth naming because it is the one an English reader most reliably misses.

3. Generic — marking a class. ὁ can mean "humanity" rather than a particular man. English does this with the plural or with a bare noun, so translations rarely preserve it.

4. Substantivising — turning anything into a noun. Put the article in front of an adjective, a participle, an adverb, a prepositional phrase, even a whole clause, and Greek makes it a noun. The Prologue has οἱ ("his own people," an adjective made a noun), τὸ τὸ (the adjective given its own article), and at 1:18 ὁ ὤν, "the one who is," a participle made a noun. This is the article's most productive function and it has nothing to do with definiteness in the English sense at all.

5. Par excellence. The article marks the one that deserves the name most: ὁ , "the Prophet," in John 1:21.

6. Marking the subject. This is the one section 3 needs.

There is also the article's use with abstract nouns (ἡ , "truth"), its variable use with proper names (Greek says both and ὁ Ἰησοῦς), and its use in an address where the vocative might be expected — John 20:28, ὁ κύριός μου. None of these has an English counterpart that behaves the same way.

3. Two nominatives in one clause: which is the subject?

In a Greek clause with the verb "to be," both the subject and the predicate stand in the nominative. So how do you tell which is which?

The rules the grammars give, in order of priority: a pronoun is the subject; then, if one noun has the article and the other does not, the articular one is the subject; then, if both are articular or both anarthrous, word order and context decide.

Apply that to καὶ θεὸς ἦν ὁ λόγος. ὁ λόγος has the article; θεός does not. Therefore ὁ λόγος is the subject and θεός is the predicate. This is not disputed by anyone. It is a plain application of a plain rule, and it is one of the very few things about this clause that is not argued over.

And it does real work. Because θεός is the predicate rather than the subject, the clause is not convertible: it does not assert that the two terms are interchangeable. Had John written καὶ ὁ θεὸς ἦν ὁ λόγος, with the article on both, the sentence would have asserted an identity in both directions — the Word is God and God is the Word, with nothing left over on either side. He did not write that, and the difference is the difference between a convertible proposition and a predication.

This is a case of grammar genuinely ruling something out. The reading on which θεός and ὁ λόγος name exactly the same referent with no remainder — the position associated with Sabellius and generally called modalism — is not what this clause says, and you can show that from the article alone. Notice what has just happened: grammar excluded a reading. That is the thing grammar can do. What it cannot do is pick among the readings still standing, and there are three.

4. The three candidates

With ὁ λόγος fixed as the subject, θεός as an anarthrous predicate noun standing before the verb could in principle be:

Definite. θεός refers to the same individual as ὁ θεός in the preceding clause — that is, to the Father. English: "the Word was God," understood as identifying the referent.

Qualitative. θεός names a class or a nature rather than an individual: the Word has the quality of being θεός, is what God is. English struggles: "the Word was divine," or the New English Bible's attempt, "what God was, the Word was."

Indefinite. θεός names a member of a class: the Word was a god, one of the beings called θεός. English: "the Word was a god."

Every one of these is grammatically constructible in Greek. The argument of the last ninety years has been about which is likely here, and the argument has run through a rule that almost nobody states correctly.

5. What Colwell actually wrote

E. C. Colwell, "A Definite Rule for the Use of the Article in the Greek New Testament," Journal of Biblical Literature 52 (1933).

Here is what he did, and the procedure is the whole point.

Colwell collected predicate nominatives in the New Testament whose definiteness he had already established on other grounds — from the context, from the presence of a modifier, from the fact that they name something unique. He then asked a question about how those already-known-to-be-definite nouns were expressed: did they carry the article or not, and did it depend on whether they stood before or after the verb?

His finding: when a definite predicate noun follows the verb, it usually has the article. When it precedes the verb, it usually does not — in his count, in the large majority of cases, a figure normally quoted at about 87 per cent.

His rule, stated in his direction: a definite predicate noun that precedes the verb usually lacks the article.

His examples are worth knowing because they are good ones. At Mark 15:39 the centurion says ἀληθῶς οὗτος ὁ ἄνθρωπος υἱὸς θεοῦ ἦν (alethos houtos ho anthropos huios theou en) — anarthrous predicate before the verb. At John 1:49 Nathanael says σὺ εἶ τοῦ (sy basileus ei tou Israel, "you are the king of Israel") — and "king of Israel" is surely definite, since there is one, yet βασιλεύς is anarthrous and precedes the verb.

Now the crucial logical point, and it is the reason this lecture exists.

Which way the rule runs

Colwell's rule is a rule about expression, not about interpretation. It tells you what to expect when you already know a predicate noun is definite. It is of the form: if definite and preverbal, then usually anarthrous. It does not tell you what to conclude when you encounter an anarthrous preverbal predicate noun of unknown definiteness. To move from "if definite, then usually anarthrous" to "if anarthrous, then probably definite" is to affirm the consequent, and it is invalid.

Rain and wet streets

To see why, take an everyday case. Suppose it is true that when it rains, the street is usually wet. It does not follow that when the street is wet, it is usually raining — the street may be wet because a pipe burst, or a lorry passed, or it rained an hour ago. To get from one to the other you would need to know how often streets are wet for other reasons. That extra information is called a base rate, and Colwell's study did not supply it, because he was not asking that question.

Did Colwell himself apply the rule to John 1:1? He did, and this is where a fair account has to be careful. He noted that the anarthrous predicate before the verb in John 1:1c is what one would expect if it were definite, and he wrote in effect that the absence of the article does not make the predicate indefinite when it precedes the verb, and that it should be taken as indefinite in that position only when the context requires. So Colwell did point his rule at this verse. But notice what that observation is and is not: it is the observation that the anarthrous form is consistent with definiteness, which is true and which nobody denies. It is not a demonstration that the noun is definite, and Colwell's own data could not have provided one.

6. How the rule was misused — in both directions

This is a point where readers will not agree about the conclusion. Almost nobody who has studied the matter disagrees about the following two errors.

Misuse in the first direction. Within a few decades the rule was circulating in popular apologetics and in some commentaries in its reversed form: an anarthrous predicate noun preceding the verb is definite; therefore θεός in John 1:1c is definite; therefore "the Word was God" is grammatically demonstrated and "a god" is grammatically excluded. This is the affirming-the-consequent error above. It has been repeated in print for seventy years, it is still in circulation, and the people who have complained about it most sharply are grammarians who themselves hold the traditional christology — most prominently Daniel Wallace, who devotes several pages of his grammar to the point and calls the reversed form a misunderstanding of Colwell. That is worth dwelling on. The objection to the argument is not a doctrinal objection; it is an objection from people who accept the conclusion and refuse the argument.

Misuse in the second direction. The reverse misuse takes two forms. The first is the simple one: θεός has no article, therefore it is indefinite, therefore "a god." Section 1 disposed of that; Greek has no indefinite article, anarthrous nouns are definite all over this very chapter, and bareness alone signals nothing.

The second is subtler and has been genuinely influential. Once the reversed Colwell argument was exposed as invalid, some writers treated that exposure as though it had established the opposite conclusion — as though showing that the traditional rendering is not proved by the rule showed that "a god" is preferred. That does not follow either. Refuting an argument for a proposition leaves the proposition exactly where it was before the argument was offered. A bad argument for a conclusion is not evidence against it.

A note on how to hold this. Two invalid arguments, pointing in opposite directions, have been derived from one 1933 article by people who had not read it. If you take nothing else from this lecture, take the habit of asking, when someone cites a rule at you: in which direction does the rule run, and is the person running it that way?

7. The qualitative analysis: Harner, Dixon, Wallace

If the reversed rule will not work, what does the evidence actually suggest?

Philip B. Harner, in a 1973 article in the Journal of Biblical Literature — "Qualitative Anarthrous Predicate Nouns: Mark 15:39 and John 1:1" — asked the question Colwell had not: taking anarthrous predicate nouns that precede the verb, what are they in fact doing? He examined the instances in Mark and John and concluded that the great majority are qualitative: they say what kind of thing the subject is rather than identifying which individual it is. He proposed that John 1:1c is best understood this way, and suggested a rendering along the lines of the New English Bible's "what God was, the Word was."

Paul S. Dixon, in a 1975 thesis on John specifically, reached a similar conclusion with a higher figure — in his count the large majority of anarthrous preverbal predicate nouns in this Gospel are qualitative rather than definite or indefinite.

Daniel Wallace takes the qualitative reading in his grammar, sets out the statistics for the New Testament as a whole, and argues that the qualitative category is not a fudge between definite and indefinite but a distinct and well-attested function.

Wallace also makes an objection to the English word "divine" that is worth having. Greek has an adjective for "divine" — (theios) — which the New Testament uses, and which would have been available to John. The English word "divine," moreover, is weak in modern usage: it can describe something merely godlike or god-related. So a translation that reads "the Word was divine" may render the construction accurately while giving an English reader a considerably weaker claim than the Greek noun makes. That is a translation problem, not a grammatical one, but it is a real one and it explains why translators have mostly not gone that way.

Now the honesty this lecture owes you. The statistics have a weakness, and it is not a small one. To count how many anarthrous preverbal predicate nouns are "qualitative," someone has to decide, case by case, which ones are qualitative — and that decision is an interpretive judgement, made by a scholar, about the very kind of ambiguity that is in dispute. The categories are not read off the forms. So the figures are not measurements in the way a count of word-frequencies is a measurement; they are a summary of a large number of individual interpretive judgements, made by people who knew what the disputed verse was.

That does not make the figures worthless. The great majority of the instances are not controversial, and a pattern established from uncontroversial cases can inform a controversial one. But it does mean the statistics constrain rather than decide, which by now you will recognise as this course's recurring refrain. The qualitative reading is the best-supported one on present evidence and is held by most grammarians. It is a well-founded judgement. It is not a proof, and it should not be presented as one.

8. What "qualitative" actually leaves open

Here is a point that both sides of the popular argument tend to slide past, and it matters more than either of them notices.

Suppose the qualitative reading is right, and θεός in John 1:1c says what the Word is in kind rather than which individual he is. That by itself does not tell you how much the kind amounts to. To say the Word has the nature θεός leaves entirely open whether θεός here names the one God of Israel's own nature, in which case the claim is very strong, or a broader category of divine being, in which case it is weaker. The construction gives you "what kind"; it does not give you the content of the kind.

Compare John 4:24, printed above: πνεῦμα ὁ θεός. Same construction, and everyone reads it qualitatively — God is spirit in nature. But what "spirit" amounts to there is supplied by the rest of John's theology, not by the word order.

So the qualitative analysis, correctly understood, is a smaller result than it is usually made to sound. It excludes the reading on which θεός identifies the Word as the Father. It tells you the clause is about nature. It hands the question of what nature back to the rest of the Gospel — to John 5:18, to John 10:33, to John 20:28, to the Prologue's own claim that all things came to be through him. And that is exactly where the question ought to be decided, by readers who can weigh a book rather than a word order.

9. A fair assessment of "a god"

The rendering "and the Word was a god" is best known from the New World Translation, produced by Jehovah's Witnesses, though it did not originate there; some earlier translators had proposed it or something like it. It is the point at which this discussion most often stops being scholarship, so let us do it properly.

What can be said for it, grammatically. Greek has no indefinite article, so an indefinite rendering of an anarthrous noun is always formally available; English simply has to choose something. Anarthrous predicate nouns are sometimes indefinite in John — at John 4:19 the Samaritan woman says ὅτι προφήτης εἶ σύ (theoro hoti prophetes ei sy, "I perceive that you are a prophet"), and there the indefinite English is right. So the construction does not by itself forbid it. And, as section 6 established, the reversed Colwell argument that was long used to forbid it does not work.

Jason BeDuhn, in Truth in Translation (2003), argued that the rendering is defensible and that critics have often attacked it with bad grammatical arguments. He is a scholar of religion, not a Witness, and his book is worth reading precisely because it forces the critic to make a better argument.

A further argument sometimes offered is from the Sahidic Coptic version, an early translation which renders the clause with what is formally Coptic's indefinite article, and which is therefore said to show how third-century Egyptian readers understood the Greek. That is a real piece of evidence and should not be waved away. Brian J. Wright and Tim Ricchuiti examined it in a 2011 article and argued that Coptic's indefinite article also marks qualitative and class expressions, so that the Coptic is consistent with the qualitative reading and does not select the indefinite one. Whether that reply fully succeeds is argued about; what matters for your purposes is that it is a reply about Coptic grammar, which is how such an argument should be answered.

What can be said against it, grammatically. The distributional case: indefinite anarthrous preverbal predicate nouns are the rarest of the three categories in John, considerably rarer than qualitative ones, so an indefinite reading is possible but a priori less likely. The contextual case: John 4:19 has a noun, "prophet," that names a category with many members, and the whole force of an indefinite is that there are others; θεός in a first-century Jewish monotheistic text is not obviously such a noun, and the rendering has to assume a category of lesser divine beings for the Word to be one of. That assumption may be defensible from other texts, but it is being brought to the verse, not found in it.

What has been said against it that does not work. That the anarthrous form proves definiteness (invalid, section 6). That Colwell's rule forbids it (invalid, and Colwell's own claim was narrower). That no Greek grammarian could countenance it (false; the construction is formally available). Bruce Metzger wrote against the rendering in 1953 and Julius Mantey objected publicly to the use made of his own grammar in its defence; their objections are largely lexical and contextual and are stronger on that ground than the grammatical arguments that have often accompanied them.

This is a point where readers will not agree, and it is the sharpest such point in this course. Let me be as precise as I can about what has and has not been shown.

  • The grammar excludes the modalist reading, by the article on ὁ λόγος (section 3).
  • The grammar does not exclude any of the three candidates in section 4, taken purely as constructions.
  • The distribution of anarthrous preverbal predicate nouns in John makes the qualitative reading the most likely and the indefinite the least, and most grammarians in most traditions accept this.
  • That distributional argument rests on interpretive classifications and is therefore weaker than a count of forms would be (section 7).
  • Everything beyond that — what θεός denotes for John, whether a category of lesser divine beings is available to him, what the Gospel as a whole claims — is lexical, contextual and theological, and is where the actual decision is made.

If you find that unsatisfying, notice that the dissatisfaction is a demand for grammar to do theology's work. It cannot, and a course that pretended otherwise would be teaching you to make the exact mistake it is warning you about.

10. John 1:18, where the article and the text both matter

θεὸν οὐδεὶς ἑώρακεν πώποτε· μονογενὴς θεὸς ὁ ὢν εἰς τὸν κόλπον τοῦ πατρὸς ἐκεῖνος ἐξηγήσατο.

Three things.

First, the variant. The manuscripts divide between μονογενὴς θεός and ὁ μονογενὴς υἱός, with a third reading, ὁ μονογενὴς θεός, also attested. The earliest Greek witnesses support the reading with θεός; the great majority of later manuscripts, and the versions behind the older English translations, support υἱός. Modern critical editions print μονογενὴς θεός; the World English Bible, quoted above, follows the other. Bruce Metzger's commentary records that the committee was not unanimous and gives the arguments. The relevant point here is that a decision about a doctrinally weighty word is being made on manuscript evidence, and lecture 12-1 will show you how to read the apparatus entry for yourself rather than taking anyone's word for it.

Second, the article on the participle. ὁ ὤν (ho on, "the one who is") is the article's substantivising function from section 2: the article turns a present participle into a noun phrase. It also, incidentally, is the phrase by which the renders the divine self-designation at Exodus 3:14, which readers have noticed since antiquity.

Third, θεός is again anarthrous, twice in one verse — once as the object of "no one has seen" and once in the disputed phrase — and in neither case does anyone translate it "a god." That is the observation section 1 asked you to file, and it is worth having ready.

11. Two other places the article is over-claimed

Granville Sharp's rule. Sharp, an eighteenth-century abolitionist and amateur grammarian, observed that when two singular nouns are joined by καί and the article appears only before the first, the two refer to the same person. His stock example is Titus 2:13, τοῦ θεοῦ καὶ ἡμῶν (tou megalou theou kai soteros hemon Iesou Christou), where on the rule "God" and "Saviour" describe one person, not two.

The rule is sound within limits and Daniel Wallace has devoted a monograph to establishing them: the nouns must be singular, personal, and not proper names, and they must not be plural. Almost every popular misuse of the rule comes from ignoring one of those restrictions — most often by applying it to plurals, where it demonstrably fails. Cite Sharp if you like; cite the restrictions in the same breath.

John 20:28. Thomas addresses Jesus as ὁ κύριός μου καὶ ὁ θεός μου, with the article on both nouns. This is often produced as the counterpart to John 1:1 — see, here θεός does have the article. Two cautions, in fairness to the argument. The article is at least partly conditioned by construction: a nominative used for direct address regularly takes the article in Greek, and both nouns carry a possessive pronoun, which also attracts it. So the article's presence here is less freely chosen than at 1:1. What carries the weight in this verse is not the article but that the words are addressed to Jesus and that the narrator does not correct them — which is a contextual argument, not a grammatical one, and is none the worse for that.

Notice the pattern in both cases. The strongest arguments in this area are not the ones that lean hardest on the article.

12. What to say in public

You will one day stand in front of people who cannot check your Greek. Here is what this lecture would have you say, and not say, about John 1:1.

Do not say "the Greek proves…" about the third clause, in either direction. It does not, and someone in the room may know it.

Do not say "Colwell's rule shows…" unless you can state the rule in the direction Colwell stated it and explain what it does not establish. The number of people who cite this rule correctly is small and you should want to be one of them.

Do say, if it comes up: the clause has no article on "God" and does have one on "the Word." That tells us which word is the subject, so the sentence is about the Word and is not simply saying that God and the Word are interchangeable terms. What the predicate contributes is discussed by grammarians, and most think it describes the Word's nature rather than identifying him as the Father. Beyond that, the case is made from the whole Gospel rather than from one clause.

That is longer than a proof-text and it has an enormous advantage: every part of it is true, and a listener who goes away and checks will find it holding up. A congregation that has been given a grammatical argument which later turns out to be invalid has been given a reason to distrust everything else you told them about Greek, and that is a high price for a rhetorical flourish.

And one thing more, which lecture 12-3 will develop. Knowing the languages gives you an authority your hearers cannot audit. That is precisely why the standard of honesty has to be higher here than anywhere else, and why a lecture on the definite article ends with a paragraph about integrity rather than about grammar.

The diagram sets out the logic of the argument: what the article marks, what the three candidate readings are, the direction Colwell's rule runs and the direction people run it, and — in the closing panel — a clean division between what the grammar settles and what it hands on. Greek forms appear with transliterations, since a device without a Greek font will show those lines blank.

θεὸς ἦν ὁ λόγος — what the article settles, and where it stops theos en ho logos (John 1:1c). Predicate first, no article; subject last, with the article. θεὸς (no article, before the verb) — ἦν — ὁ λόγος (article) theos (anarthrous predicate) — en — ho logos (articular subject) Rule: with two nominatives, the one carrying the article is the subject. Undisputed. GRAMMAR RULES OUT: the clause is not convertible. Had John written ὁ θεὸς ἦν ὁ λόγος (ho theos en ho logos), the terms would be interchangeable. He did not. A. Definite θεός = the same individual as ὁ θεός in clause b. English: "was God." Formally available. Not the majority view. B. Qualitative θεός names a nature, not an individual. NEB: "what God was, the Word was." Harner · Dixon · Wallace C. Indefinite θεός names a member of a class. English: "was a god." Formally available; the rarest of the three in John. Colwell 1933 — the direction he wrote it Start from predicate nouns ALREADY KNOWN to be definite. Ask how they are expressed. IF definite AND before the verb THEN usually no article (about 87%). Examples: Mark 15:39, John 1:49. The reversal — INVALID IF no article AND before the verb THEN definite. This affirms the consequent. "When it rains the street is wet" does not give you "when the street is wet, it rains." Wallace, who holds the traditional view, says so. The second misuse, in the other direction 1. "No article, therefore indefinite." Greek has no indefinite article; anarthrous theou is definite three times in John 1 alone (vv. 6, 12, 13). The bare form signals nothing on its own. 2. Refuting the reversed Colwell argument does not establish the opposite conclusion. What the grammar settles · ὁ λόγος is the subject · the clause is not convertible · all three readings are formally possible · distribution favours the qualitative (a judgement, resting on classifications) What it hands on to something else · what θεός denotes for this author · whether a class of lesser divine beings is available in a Jewish monotheistic text · what the Gospel as a whole claims These are lexical, contextual and theological.

6. Further Reading

  • E. C. Colwell, "A Definite Rule for the Use of the Article in the Greek New Testament," Journal of Biblical Literature 52 (1933), 12-21. Short. Read the original before citing it; almost nobody does.
  • Philip B. Harner, "Qualitative Anarthrous Predicate Nouns: Mark 15:39 and John 1:1," Journal of Biblical Literature 92 (1973), 75-87. The qualitative analysis in its original form.
  • Daniel B. Wallace, Greek Grammar Beyond the Basics (Zondervan, 1996). The chapters on the article, including the treatment of Colwell's rule, the qualitative category, and the statistics; the single most useful reference for this lecture.
  • Daniel B. Wallace, Granville Sharp's Canon and Its Kin (Peter Lang, 2009). The monograph establishing the rule's real limits, by someone who accepts the rule.
  • Murray J. Harris, Jesus as God: The New Testament Use of Theos in Reference to Jesus (Baker, 1992). The fullest treatment in English of every relevant passage, argued at length and with the grammar handled carefully.
  • Bruce M. Metzger, "The Jehovah's Witnesses and Jesus Christ," Theology Today 10 (1953). The classic critique; read it noting which of its arguments are grammatical and which are contextual.
  • Jason David BeDuhn, Truth in Translation: Accuracy and Bias in English Translations of the New Testament (University Press of America, 2003). The defence of the indefinite rendering; read it to see the strongest form of the case you would have to answer.
  • Brian J. Wright and Tim Ricchuiti, "From 'God' (ΘΕΟΣ) to 'God' (ΝΟΥΤΕ): A New Discussion and Proposal Regarding John 1:1 and the Sahidic Coptic Version of the New Testament," Journal of Theological Studies 62 (2011). The Coptic evidence examined properly.
  • A. T. Robertson, A Grammar of the Greek New Testament in the Light of Historical Research, 4th ed. (Broadman, 1934). The long chapter on the article, still unsurpassed for range of illustration.
  • F. Blass and A. Debrunner, A Greek Grammar of the New Testament and Other Early Christian Literature, trans. and rev. Robert W. Funk (University of Chicago Press, 1961). The sections on the article and on predicate nouns; terse, and a useful check on fuller treatments.

In short

  • Greek has a one-way article system (definite article only, no indefinite article), so the presence of the article is informative but its absence is not — John 1 alone has three anarthrous uses of θεός (verses 6, 13, 18) that nobody translates "a god," which demonstrates that "no article, therefore indefinite" is not a valid inference.
  • In καὶ θεὸς ἦν ὁ λόγος, the rule that the articular noun is the subject when only one of two nominatives carries the article settles, without dispute, that ὁ λόγος is the subject and θεός the predicate; because John did not write the fully articular ὁ θεὸς ἦν ὁ λόγος, the clause is not convertible and rules out a modalist ("the Word and God are simply interchangeable") reading.
  • Colwell's 1933 rule, correctly stated, runs only one direction — a predicate noun already known on other grounds to be definite usually lacks the article when it precedes the verb — and reversing it to "anarthrous and preverbal, therefore definite" affirms the consequent; Wallace, who holds the traditional christology, is among those who call the reversal invalid.
  • Harner, Dixon and Wallace argue from distributional evidence that most anarthrous preverbal predicate nouns in John (and in Mark 15:39) are qualitative — naming what kind of thing the subject is, not which individual — making "the Word was divine in nature" the best-supported reading of John 1:1c, though the classification itself rests on scholars' own interpretive judgments and so constrains rather than proves.
  • Grammar can exclude a reading (the modalist one) and can favor one candidate over others by distribution, but it cannot settle what "God" denotes in John's usage, whether a category of lesser divine beings is available in a Jewish monotheistic text, or what the Gospel as a whole claims — those questions are lexical, contextual and theological, not grammatical.

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.

Review every summary in this course →

Check yourself

Not "did you read it" but "can you explain it". Ticks are kept on this device only.

Reflection and discussion

  • Reading exercise. Go through John chapter 1 in Greek and list every occurrence of θεός and θεοῦ, marking whether each carries the article and what construction it stands in. Then say, for each anarthrous instance, whether the absence is a free choice by the author or is conditioned by the construction. How many are free choices?
  • Reading exercise. Find five predicate nominatives in John's Gospel that stand before the verb without the article — John 4:19, 4:24, 9:5 and 1:49 will start you off. For each, decide whether it is definite, qualitative or indefinite, and write down what made you decide. Then read section 7 again and say what your own exercise has just illustrated about the statistics.
  • State Colwell's rule in Colwell's direction, in one sentence, without looking. Then state the invalid reversal. Then explain the difference to someone with no Greek, using an example that is not about rain.
  • Section 8 argues that the qualitative reading is a smaller result than it is usually made to sound. Do you agree? If the qualitative reading is right, what remains to be decided, and where would you look to decide it?
  • Wallace holds the traditional christology and rejects the reversed Colwell argument used to support it. Why is that combination important, and what does it show about the relation between a conclusion and the arguments offered for it?
  • Write the paragraph you would actually say from a pulpit if someone asked you about John 1:1 and the Greek article. Keep it under a hundred and twenty words. Then hand it to someone who disagrees with you theologically and ask whether they can find anything in it that is untrue.

Sign in to write and keep your answers

What next

Definitions: Strong’s Concordance (1890/1894, public domain) · Morphology: MorphGNT/SBLGNT (CC BY-SA), Open Scriptures morphhb/WLC (CC BY 4.0)

The lecture ends here. What follows was written by its readers.