Colwell's rule and John 1:1
A closer look at John 1:1 and the application of Colwell's rule.
...και θεος ην ο λογος - John 1:1
"...and the Word was ___" …God, divine, or a god?
Colwell's rule is used to insist the translation should be the definite "God". In his 1933 article entitled A definite rule for the use of the article in the Greek New Testament he proposes this rule:
"A definite predicate nominative has the article when it follows the verb, it does not have the article when it precedes the verb".
Let's first define what these terms mean:
The predicate is the part of a sentence which describes or modifies the subject. For example John 1:5 begins "And the light shines in the darkness" - the light is the subject of the sentence, and shines in the darkness is the predicate. The function of the predicate therefore is to tell us something about the subject and almost always contains a verb plus anything following that verb. The subject + predicate is the simplest way to break down any sentence.
The nominative is a grammatical case of a noun which usually indicates that the noun is the subject within a sentence. English does not modify nouns unlike Greek, making this unfamiliar to us. For example, the word λογος ('word') taken from the topic verse is in the nominative case. In the accusative case, which means the noun is the recipient of the action performed by the verb, it would be written as λογον. And in the genitive case, which is usually preceded by the word 'of', it would be written as λογου (of the word). The grammatical case tells us where in relation to the verb and nouns the word should be placed or understood within the sentence.
So combining these two pieces of information, we can see that Colwell's rule applies to a very specific type of noun; one that is in the nominative case but is actually not the subject of the sentence, instead it acts as the predicate...the predicate nominative.
The article is a word used before a noun which tells us the definiteness of that noun. In English we use "a" or "the" - the former is the indefinite article while the latter is the definite article; 'the word' implies a particular or specific word from the author, whereas "a word" is unspecific and relates to any word. This becomes important later in understanding the verse in question (a god vs the god), and while the Greek language has a definite article it does not have an indefinite article (there is no word for "a" in Greek). So when a Greek noun has the article before it, it's understood as definite. But when a Greek noun does not have an article before it, it can be either definite, indefinite or qualitative. And it is left for contextual argumentation as to whether that noun should be translated into English with a definite or indefinite article. In Greek, we call a noun that has the definite article arthrous, and one that does not contain the article anarthrous.
Before we revisit the verse, two points need to be made on the capitalization of the noun God:
1) In English, we differentiate between common and proper nouns by capitalizing the first letter of the word. Proper nouns refer to names or specific places, such as Jerusalem or John. Common nouns refer to generic names, such as the town or a man. What's important to understand is that we don't use any articles for proper nouns in English, whereas we require the article for common nouns. So regardless of the Greek article, we always capitalize God if we believe the text is speaking about the God of the Bible, the person that is God, and we do the same for the opposite; using the lowercase god if we believe the text is not referring to the God of the Bible regardless of whether it contains the definite article or not¹.
2) The earliest Greek manuscripts for the New Testament are written in the majuscule script, that is, all capital letters. The minuscule script, which is all lowercase letters, was introduced around the 9th century. So whenever the word god is used in scripture it was originally written as θεος without any inherent difference between the proper noun God and the common noun god. That decision is left to the translator.
Now that we've defined the terms, let's look at John 1:1 again and break down the sentence:
"...and God was the Word"
"... και θεος ην ο λογος"
The Word is the subject nominative, given that it's in the nominative case and given the context of the verse². In English, and in most translations of this verse, we rearrange the subject to the beginning of the sentence so that it reads "and the Word was...". The definite article ο means we understand it as the word and not a. The verb was is a linking verb which predicate nominatives are usually found following. And here the predicate nominative is God or god and lacks the definite article in Greek. This lack of a definite article is the sole reason that allows for "a god" or "divine" to be a viable translation instead of the more common "God".
Now with Colwell's rule the application is regarding his second statement; That because the anarthrous predicate nominative θεος precedes the verb in its Greek construction it should be understood as if it had the definite article, and therefore be translated as 'God' and not 'a god' or 'divine'. Colwell is making a specific claim regarding word order. Within his paper Colwell provided a good number of supporting verses to introduce his new rule, but can word order really be a basis for any rule in Greek? Does this rule hold up to scriptural usage? And how many exceptions can a rule take before it no longer exists as a rule?
Colwell was already aware of some of the verses which contradict his rule. Within his paper from the 1930s he made this comment regarding one of them;
"In the field of textual criticism the rule here advocated has an equally definite contribution to make. It shows in certain specific cases what the probabilities are as to the author’s use or non-use of the article. A fine example of this is 2 Peter 1:17, cited as an exception to the rule since Wescott and Hort follow Codex Vaticanus in reading the predicate with the article before the verb ο υιος μου ο αγαπητος μου ουτος εστιν. The evidence given in this study as to the extreme rarity of this construction in the New Testament reinforces Tischendorf’s judgment that the variant read by practically all the rest of the MSS is to be preferred. They read the predicate after the verb with the article, ουτος εστιν ο υιος μου ο αγαπητος μου, the usual Greek construction."
Colwell discounted this exception as an obscure manuscript variation of this Greek construction. This is because if Codex Vaticanus is the original construction of 2 Peter 1:17 with the predicate placed before the verb, then this verse would become another exception against his rule as there shouldn't be a definite article before υιος (son) according to it. However, a couple of decades after Colwell waved this away as a doubtful variation, the oldest surviving manuscript of 2 Peter was discovered in the 1950s (see below) matching the Codex Vaticanus reading and providing further evidence that the original Greek text did not hold to his rule.
 |
Oldest surviving manuscript of 2 Peter, papyrus 72, showing the verb placed at the end of the sentence in 1:17. Between the red markers: ο υιος μου ο αγαπητος μου ουτος εστιν, "the son of me, the beloved of me, this is". You can read the manuscript here |
Prior to this point, Colwell created another exception to satisfy his rule:
"Five of the seven put the predicate not only before the verb, but also before the subject, e.g., 1 Corinthians 9:1 ου το εργον μου υμεις εστε εν κυριω. Thus the order in these five Corinthian passages is predicate-subject-verb. This is obviously a stylism temporarily affected by the Apostle to the Gentiles, possibly for the sake of greater emphasis"
As these verses break his rule, Colwell explained away these exceptions based on another claim on word order and brushed aside these verses as simply stylistic affect. In total, Colwell listed 41 verses which he considered to be exceptions to his rule, 15 of which refer to his second statement and to this topic. I will go into what is actually a much larger list of verses which he was either unaware of or had decided not to mention. The verses below lack the definite article and are understood as indefinite³:
List of predicate nominative nouns preceding the verb which have a pronominal subject (pronoun functioning as the subject):
Matthew 5:9, 12:50, 23:8, Mark 3:35, Luke 7:8, 19:9, John 4:19, 8:44, 9:24, 27, 28, 10:1, 33, 15:14, 18:35, 37, Acts 9:15, 2 Corinthians 6:16, 1 Thessalonians 1:6, 2:14, 1 John 2:4, Revelation 2:9
List of predicate nominative nouns preceding the verb with the verb functioning as the subject:
Matthew 14:26, 22:42, 45, Mark 6:49, 11:32, Luke 7:39, 20:27, 38, 44, 22:25, John 5:10, 8:44, 9:8, 17, 25, 10:13, 34, 12:6, 36, 13:35, 18:26, 37, Romans 2:28, 7:3, 13:4, 1 Corinthians 4:9, 6:7, 11:14, 15, 15:12, Galatians 5:3, Hebrews 5:13, 10:33, James 2:23, 4:14, 17, 1 John 2:18, 3:2, 4:20, Revelation 10:6, 18:7, 19:10, 22:9
List of predicate nominative nouns preceding the verb with a subject noun:
Matthew 13:39, Mark 11:17, John 8:34, 10:8, Acts 22:26, 28:4, Romans 1:9, 1 Corinthians 6:19, 7:22, 10:16, Galatians 4:1, Ephesians 5:23, 1 Timothy 6:10, James 3:5, 1 John 2:4, 3:15, Revelation 17:15
All these verses counter Colwell's rule. It is not simply a style of a single author, as the gospel writers, James and the apostle Paul all have placed an anarthrous predicate nominative noun before the verb with an indefinite meaning. But to solidify the point further, the verses below are the most notable examples which match precisely the word order of John 1:1 (predicate-verb-subject), which are neither names nor qualitative in their understanding. Beginning with the topic verse as reference;
Color code:
Predicate noun
Verb
Subject noun
και θεος ην ο λογος - John 1:1
and a god - was - the word
φονευς εστιν ο ανθρωπος ουτος - Acts 28:4
a murderer - is - the man
ουκουν βασιλευς ει συ - John 18:37
then a king - are - you
ριζα γαρ παντων των κακων εστιν η φιλαργυρια - 1 Timothy 6:10⁴
for a root of all kinds of evil - is - the love of money
οτι προφητης ει συ - John 4:19
that a prophet - are - you
οτι σκευος εκλογης εστιν μοι ουτος - Acts 9:15
that a vessel of choice - is - to me this [man]
We can see that Colwell's rule is ineffective in insisting that the definite God must be the translation of John 1:1 simply because of word order. The grammatically similar examples above are enough to, at the very least, make one consider the indefinite understanding as a valid reading of the Greek text. Whether the verse should be understood as definite, indefinite, or qualitative, is left to arguments outside the Greek construction. As in Colwell's introduction, I stick with Walter Bauer that "hard and fast rules for the
use of the article in Hellenistic Greek are an impossibility".
But for those that continue to use Colwell's rule to insist on the definite reading, adding exception after exception to the exceptions above, then I am reminded of this;
The rule is always right and you always follow the rule. Except when you don’t. For every rule, there is an exception. So you always follow the rule, except when there is an exception, in which case you follow a new rule based on that exception. Following this pattern always guarantees that you come to the right decision. Except when it doesn’t. There are exceptions to every exception. So the rule is always right, unless there is an exception to that rule, in which case you take the course of action prescribed by the exception, unless there is an exception to that exception, in which case, you follow the course of action prescribed by the exception to the exception.
--------------------
1) Some examples of this in scripture are Acts 7:43, 2 Corinthians 4:4, Philippians 3:19.
2) John begins "In [the] beginning was the Word" - in the beginning is the predicate given its in the dative case following the preposition, making the Word the subject of the verse.
3) Some anarthrous plurals have been included in the list where the indefinite is understood. For example "you also became imitators of us" 1 Thessalonians 1:6. If the group are imitators of Paul and Christ, then each person is an imitator within that group. "For they will be called sons of God" Matthew 5:9, each one then is called a son of God.
4) Colwell lists this verse as an example in support of his rule, and some translations do insert the definite article in English. But consider, if the "love of money" is the root of "all evil" (literal translation), would it make sense for the Devil to be motivated by the "love of money"? So it seems the indefinite understanding is preferred. "root" used metaphorically as in Deuteronomy 29:18.
*John 1:1, Was the Word "God," or "a god"?
ReplyDeletehttps://www.academia.edu/44318953/John_1_1_Was_the_Word_God_or_a_god_
*John 1:1, List of Alternative Readings.
https://drive.google.com/file/d/1Ri5kmAmsFlGIodl5TBcMNfrC_JP8Q0x9/view?usp=sharing
*Another God in the Gospel of John? A Linguistic Analysis of John 1:1 and 1:18
https://brill.com/view/journals/hbth/44/2/article-p141_2.xml?language=en
*V14 An Expository Rendering of John 1:1-4
https://www.academia.edu/50808377/V14_An_Expository_Rendering_of_John_1_1_4
The grammatical construction of John 1:1c involves a preverbal anarthrous predicate nominative (theos). As noted by recognized Greek grammarians (e.g., Daniel Wallace, A.T. Robertson, and others), the placement of theos before the verb (ēn) emphasizes its qualitative nature, indicating the nature or essence (quiddity) of the subject (Logos). Colwell's Rule supports this interpretation: anarthrous predicate nominatives preceding the verb in Greek are often definite, particularly in contexts like John 1:1c where the subject (Logos) is already definite (indicated by the article “ho”). Thus, the absence of the article does not imply indefiniteness but highlights qualitative force: the Word possesses the same divine nature as God.
DeleteThe examples cited (e.g., "a murderer" in Acts 28:4) involve contexts where the indefinite article is required due to the nature of the predicate noun. However, in John 1:1c, the context is unique: the subject (Logos) is explicitly stated as existing with “ho theos” ("the God"), thus requiring a qualitative or definite rendering of “theos” to avoid contradiction or henotheism. Unlike the examples provided, John 1:1c is a theological declaration about the nature of the “Logos” in relation to “ho theos”. Context and theology must guide the translation here, not mere grammatical parallels.
The traditional translation, "the Word was God," reflects the qualitative force of “theos” without implying modalism (confusing the Word with the Father) or henotheism (introducing "a god"). Jewish monotheism, as understood in the Second Temple period, allowed for distinctions within the divine identity. For example, the concept of the “Memra” in Jewish thought parallels the Logos in John 1:1. John’s prologue aligns with this framework, presenting the Logos as fully divine yet distinct in person from the Father (ho theos).
The qualitative interpretation does not preclude the full deity of the Logos. The term θεός is used here to affirm the divine essence of the Logos without directly equating the Logos to the Father (τὸν θεόν). The translation "divine" could imply a lesser status or derived divinity, which is inconsistent with the full scope of John's Christology (e.g., John 5:18, where Jesus is accused of making Himself equal with God).
The argument that "καί" indicates sequential events misreads its function in Koine Greek. "Καί" often functions as a simple connective, linking clauses without implying temporal succession. The emphasis on linguistic novelty often ignores well-established exegetical and theological traditions that maintain consistency with the broader canonical witness.
(1/2)
(2/2)
DeleteUnlike “anthropos”, which refers to a class of beings (humans), “theos” in the Jewish and early Christian context was used overwhelmingly to denote the unique and singular God. John’s Gospel reflects a monotheistic framework where “theos” does not function generically but refers to divine nature. Smarius downplays John’s monotheistic context, arguing that “theos” could generically refer to a lesser divine being. However, John repeatedly affirms monotheism. Translating “theos” as "a god" conflicts with the author’s theological intent and the audience's understanding. Within Second Temple Judaism, there was no room for a "second(ary) god" in a polytheistic sense. The Shema (Deut. 6:4) declares that God is one, and John upholds this monotheistic framework. John’s use of "theos" without the article in 1:1c does not introduce a subordinate deity but highlights the Word's full participation in the divine essence. In John 10:30, Jesus states, "I and the Father are one," reinforcing the unity of essence between the Word and God. The Jewish leaders understood this as a claim to equality with God (John 10:33), which aligns with the traditional rendering of John 1:1.
Psalm 82:6 refers to human judges metaphorically as "gods" due to their role as representatives of divine authority. However, John 1:1c does not present “theos” as metaphorical but as a descriptor of the intrinsic nature (quiddity) of the Logos. The Logos is not depicted as “a lesser god” or a created being but as preexistent and active in creation (John 1:3). This aligns with the consistent New Testament portrayal of Christ as fully divine (e.g., Colossians 1:15-20, Hebrews 1:3). In John 10:34, Jesus addresses human judges or rulers metaphorically called "gods" because they act as representatives of God’s authority. This does not equate them with God in essence or nature. Applying this metaphorical usage to John 1:1c disregards the distinct theological emphasis of the prologue, which presents the Logos as fully divine.
In Psalm 82, the term "gods" is used in a judicial context to criticize unjust rulers, contrasting their failure with God’s ultimate sovereignty. This context is not analogous to John 1:1c, which describes the eternal nature of the Word. In Psalm 82, "gods" refers to human judges or rulers who act as representatives of God’s authority but are not divine in essence. Jesus contrasts these human "gods" with His unique relationship to the Father, claiming, "I am the Son of God" (John 10:36). This claim goes beyond representation, asserting ontological unity with God.
The prologue of John explicitly distinguishes between the Logos and “ho theos” while affirming their shared divine nature. The phrase "the Word was with God" (1:1b) emphasizes relational distinction, while "the Word was God" (1:1c) emphasizes ontological unity. Rendering “theos” as "a god" introduces henotheism, contradicting the monotheistic framework of John's Gospel and the New Testament.
Subordination in function or role does not negate equality in essence. Philippians 2:6 affirms that Christ, though equal with God, humbled himself by taking on human form. Titles like "Son of God" emphasize relational roles within the Trinity, not ontological inferiority. The preposition "πρὸς" denotes a relational distinction between the Logos and God but not subordination. It emphasizes intimacy and personal relationship, which aligns with Trinitarian theology. The unity and distinction of persons within the Godhead are intrinsic to the prologue of John, culminating in statements like John 1:14 ("the Word became flesh") and John 1:18 ("the only begotten God, who is at the Father’s side").
There are several points on which I think you are correct. The Greek of John 1:1c, KAI THEOS ĒN HO LOGOS, contains no Greek indefinite article, because Koine Greek had no indefinite article corresponding mechanically to English “a/an.” The noun THEOS is anarthrous and functions as the predicate, while HO LOGOS is the grammatical subject. An anarthrous preverbal predicate can, in suitable contexts, be indefinite. Colwell’s rule does not prove that THEOS is definite merely because it precedes the copula. The absence of TIS does not prove that an indefinite interpretation is impossible. Nor can modern English capitalization simply be read back into an ancient Greek manuscript. On all of those points, a careful Trinitarian argument should make substantial concessions. Your own material is also right to insist that context rather than an automatic article-substitution rule must ultimately decide the interpretation.
DeleteThe decisive question, however, is not whether “a god” is grammatically possible in some sentence having the same broad syntactic shape. It is whether the particular THEOS of John 1:1c is actually being used as an ordinary classifying predicate, so that the Logos is being individuated as one member of a class of gods, and whether English a god is therefore the least misleading way to represent the assertion. That is a much stronger claim. Once morphology, semantic construal, English realization, and theological inference are kept separate, the argument for “a god” becomes considerably less secure than the papers you sent suggest. The Greek establishes an anarthrous predicate; it does not itself supply an English indefinite determiner, a class of lesser gods, or the further idea that the Logos is a created member of that class.
The first distinction that must be made is between identity and predication. The Trinitarian interpretation does not require John 1:1c to mean “the Word was the God” in the sense that the person of the Logos is identical with the person referred to by TON THEON in 1:1b. In fact, that would erase the personal distinction John has just expressed when he says that the Word was PROS TON THEON. Smarius is therefore substantially correct when he notes that the majority interpretation does not simply identify the Logos with the HO THEOS previously mentioned. But that observation does not move us toward the conclusion “therefore the Word was a second god.” It points instead toward predication of what the Word is while preserving who he is in relation to the Father.
This matters because much of the argument in the papers is directed against an unnecessarily crude interpretation of the traditional translation. “The Word was God” need not mean the convertible identity proposition “the Word = the Father.” English can use a noun predicatively without making subject and predicate interchangeable. The preceding clause already prevents that misunderstanding: the Logos is personally distinguished from the one with whom he is. John 1:1c then predicates THEOS of this personally distinct Logos. The contrast is therefore not between “Trinity” and “two distinct persons.” Trinitarianism insists on the personal distinction. The real question is whether the distinction is a distinction of persons within unattenuated deity or a distinction between the supreme God and a second, lesser divine entity.
Indeed, this is one reason why the absence of the article before THEOS is entirely intelligible on a Trinitarian reading. The articular HO LOGOS establishes the subject; the anarthrous predicate tells us what is predicated of that subject. Classical and Koine grammatical traditions do not treat articlelessness of predicate nouns as a covert indefinite article. Predicate substantives are very commonly anarthrous precisely because they are predicates. Goodwin, Babbitt, Smyth, Winer–Schmiedel, Moulton, Robertson, Blass–Debrunner, Dana–Mantey, Porter, Peters, and Wallace differ considerably in theoretical terminology, but they do not give us the simple equation “no article = one member of a class.” This is why the question must be semantic rather than merely morphological.
DeleteThat point exposes a problem in the use made of the Cambridge Grammar and Smyth in the “Expository Rendering.” You quote the statement that Greek has a definite article but no indefinite article and then describe lack of the article as the Greek “equivalent” of an indefinite article. You also appeal to Smyth’s observation that words denoting persons may omit the article when used of a class, with THEOS among his examples. Those are real grammatical observations. But they establish the possibility of class use; they do not classify John 1:1c as a class use. Smyth’s point is conditional: when such a noun is used of a class, article omission is possible. It does not follow that wherever THEOS lacks the article it must therefore mean “a member of the class of gods.” That would turn a description of one environment in which article omission occurs into a diagnostic for every token.
This is especially important because Greek predicate nouns normally omit the article in circumstances having nothing to do with ordinary English indefiniteness. Consequently, “Greek has no indefinite article” is a useful warning against arguing that the Greek would need to contain a word corresponding to English a. It is not an argument that Greek zero determination itself should be translated a. The direction of reasoning must be the other way around. First determine whether the predicate is functioning as ordinary class membership; then, if it is, English will normally require a/an. English grammar cannot be allowed to decide the Greek semantics retroactively. The problem is well illustrated by the fact that English and other languages package predication differently: “I am a doctor” and German ICH BIN ARZT can express essentially the same predication although only English requires the indefinite article. An English determiner is therefore a target-language realization, not a Greek morpheme hiding invisibly in the text.
The same answer applies to the repeated claim that because THEOS is a countable or generic noun, an English indefinite article is consequently natural or even required. It is certainly true that THEOS is not lexically an abstract noun equivalent to THEOTĒS, “deity” or “Godhood.” It can be pluralized as THEOI and can refer to pagan gods or other entities designated as gods. Smarius is entirely right to insist on that lexical fact, and it would be a mistake to defend John 1:1c by pretending that the Greek noun THEOS is lexically an adjective or an abstract mass noun. But lexical countability and token-level predicative function are different matters. A common count noun can be used predicatively so that its descriptive content is foregrounded rather than the individuation of one member of a salient class.
This is precisely where Smarius’s argument moves too quickly. He argues that because THEOS behaves as a generic count noun, and because a Greek expression such as THEOS ESTIN HO ERŌS in an explicitly polytheistic text naturally becomes “Eros is a god,” John 1:1c can likewise naturally be heard as “the Word was a god.” The example is grammatically useful but semantically controlled by its own discourse. In Longus, the extension of the class “gods” is already obvious. Eros belongs to a narrative world containing multiple gods. English a god therefore expresses a class structure that the discourse has already supplied. John 1:1 presents a different discourse problem altogether. The relevant question is not whether Greek THEOS can ever classify one member among gods—it plainly can—but whether John has activated such an enumerable divine class at the point where he predicates THEOS of the Logos.
DeleteThis is why examples such as “you are a prophet,” “you are a Samaritan,” “you are a king,” or “this man is a murderer” do not carry the argumentative weight assigned to them in the papers. Those examples are genuine indefinite parallels in the formal sense, and there is no reason for a Trinitarian to deny them. They show conclusively that a preverbal anarthrous predicate noun can be translated with English a/an. But in each case the relevant nonunique class is already independently available: there are many prophets, many Samaritans, many kings, many murderers. The English article is natural because the discourse and lexical semantics already license individuation within an ordinary class. The syntactic construction itself did not create those classes. The same is true of John’s PROPHĒTĒS EI SY, SAMARITĒS EI SY, and BASILEUS EI SY. The construction permits classifying predication; it does not force it. That is exactly why those examples cannot decide the disputed case.
Acts 28:4 makes the same point rather than undermining it. PHONEUS ESTIN HO ANTHRŌPOS HOUTOS naturally means that Paul is “a murderer” because murderers constitute an ordinary, nonunique human category, and the Maltese islanders are classifying Paul under it. Nothing controversial has to be supplied by the English indefinite article. By contrast, “the Word was a god” does much more than say that the predicate is not identifying. It packages THEOS as an individuated member of a nonunique divine class. That class is precisely what is under dispute.
Acts 28:6 is even less decisive as a parallel. There the islanders explicitly move from believing Paul to be a murderer to believing him to be a god. The setting is pagan and the existence of multiple gods belongs to the conceptual world of the speakers. Moreover, the construction in the critical text is not syntactically identical with John 1:1c: it involves an accusative predicate in an infinitival construction. The Majority Text word order can place THEON before the infinitive, but word order does not erase the decisive discourse difference. Acts 28:6 proves something no Trinitarian needs to contest: THEOS can mean “a god” in a context in which one of a plurality of gods is intended. It does not prove that John has invoked such a class in his prologue.
The same is true of Psalm 82 as quoted in John 10:34. The plural THEOI unquestionably exists, and Scripture can use divine terminology of figures other than the one true God. But that observation establishes the lexical range of THEOS; it does not establish that every singular anarthrous predication of THEOS should be assimilated to Psalm 82. Psalm 82 is a marked judicial or representative context in which a group is explicitly addressed as THEOI. Exodus 7:1 is likewise a marked representative appointment: Moses is made “God” to Pharaoh in a specific functional relation. Such passages demonstrate that THEOS can receive secondary, representative, or derivative applications when the context supplies the relevant relation. John 1:1c does not contain such a marker. There is no phrase equivalent to “I have made you God to Pharaoh,” no quotation such as “I said, you are gods,” no pagan setting, and no expression of delegated or judicial status. To import one of those restricted senses into John 1:1c therefore requires additional contextual argument.
DeleteThis also explains why the observation that qualitative and indefinite senses may overlap does not rescue “a god.” I agree that those categories need not be mutually exclusive. “He is a prophet” both classifies and characterizes. “He is a fraud” may function more strongly as a judgment of character than as a census of class membership. A singular count predicate can therefore be qualitatively prominent while still receiving a/an in English. But this only means that English a/an can sometimes carry qualitative content. It does not mean that whenever a Greek predicate is qualitative, English a/an is semantically neutral. Where the class itself is controversial, a/an makes a substantive interpretive commitment. “A god” does not merely express divine quality; it individuates the Logos as one god among a plurality of entities for which god is the count-class label.
The New World Translation itself provides a useful control against an automatic count-noun argument. In Mark 2:28, KYRIOS ESTIN HO HUIOS TOU ANTHRŌPOU has an anarthrous singular count predicate before the copula and a personal subject. Yet the New World Translation does not render it “the Son of man is a lord”; it gives the bare predicative “the Son of man is Lord even of the Sabbath.” The reason is obvious: countability does not by itself determine English article choice in a predicate. Lexical meaning, discourse, conventional usage, and the intended predicative relation matter. The same methodological freedom must be permitted when THEOS is the predicate.
This brings us to Colwell. You are justified in objecting to the argument, still sometimes heard in popular apologetics, that “because THEOS precedes the verb and lacks the article, Colwell’s rule proves that it is definite.” That is the converse fallacy. Colwell started with predicates independently regarded as definite and observed how their article use correlated with position. His rule can explain why an independently definite predicate may be anarthrous when preverbal; it cannot make an unknown predicate definite merely because it is anarthrous and preverbal. A strong Trinitarian case does not need that misuse of Colwell.
But once that concession is made, the evidence does not swing automatically toward “a god.” Harner’s work became important precisely because he redirected the discussion toward qualitative predication. Dixon subsequently classified the great majority of John’s anarthrous predicates as qualitative, including fifty of fifty-three preverbal examples in his analysis. Hartley’s narrower count-noun filtering likewise did not yield a general indefinite rule: in his filtered Johannine singular-count dataset, ten of eighteen were classified qualitative, three indefinite, three indefinite-qualitative, and two definite. These classifications are not mathematical probabilities for John 1:1c and should not be treated as such, but they are enough to demonstrate that “preverbal + anarthrous + count noun” cannot simply be equated with English “a/an.”
Harner is particularly important because some of the materials you sent appeal to him as though his rejection of an identifying-definite analysis supports the New World Translation. It does not. Harner’s central point is that predicates of this type often foreground the nature or character of the subject. That gives substantial support to understanding John’s clause as saying what the Logos was rather than identifying the person of the Logos with the person of the Father. But “qualitative” is not a synonym for “created divine being,” nor is it a synonym for “one lesser god.” The Watchtower interpretation requires a further bridge that Harner’s grammatical category does not supply.
DeleteThis is also why the objection to the word “divine” has to be handled carefully. You are right that THEOS remains a noun. John did not write the adjective THEIOS. A translation should therefore not argue that THEOS lexically means “divine” in exactly the way an adjective does. Nevertheless, a noun functioning predicatively can be rendered adjectivally when that best reproduces the semantic function. Translation does not require preserving lexical category at all costs. The danger with English “divine” is not that it is grammatically impossible, but that English divine can be weaker than Greek THEOS: it can mean merely heavenly, sacred, godlike, or pertaining to God. Thus “the Word was divine” is acceptable only if divine is understood strongly. This is one reason “what God was, the Word was” is such a useful explanatory paraphrase: it preserves personal distinction while refusing to weaken the predicate into mere resemblance.
The papers’ lists of alternative translations should therefore be separated rather than counted together. “A god,” “a divine being,” “divine,” “godlike,” “God of a sort,” lowercase “god,” “Deity,” and “what God was, the Word was” are not semantically equivalent translations. A list that combines all of them under the heading of evidence against the traditional rendering creates an impression of agreement where there are actually different analyses. Someone who translates “what God was, the Word was” is not thereby endorsing a second, subordinate deity. In fact, that wording can express the central Trinitarian point exceptionally well: the one called God in 1:1b is personally distinguished from the Logos, while what that God is is predicated of the Logos.
The same caution applies when scholars are cited because they concede that “a god” is grammatically possible. C. H. Dodd, Murray Harris, William Barclay, and others have acknowledged in one form or another that the Greek construction considered in isolation does not make the indefinite English rendering syntactically impossible. But that is different from concluding that it is the best contextual translation. A scholar who says “this wording can be generated from the grammar but is wrong in Johannine context” cannot fairly be counted simply as scholarly support for the interpretation “the Word was a god.” The very controversy is the movement from grammatical possibility to token-level meaning. Smarius himself acknowledges that the dominant scholarly interpretation understands the predicate as divine quality shared by the Father and the Word, even though he proposes reopening the alternative second-deity interpretation.
The number of translations that have used some nontraditional wording does not settle this either. Translation history can show that a rendering is not a twentieth-century invention, and that is worth knowing. It cannot substitute for semantic analysis. Nor does a translation using lowercase god necessarily mean what an English reader hears in “a god.” Capitalization and determination are separate choices. Early Greek manuscripts did not encode the modern English distinction God/god, but it does not follow that an English translator is prohibited from capitalizing after determining the referent and semantic function. Capitalization is a downstream orthographic decision. The same is true in the opposite direction: printing lowercase god cannot itself establish that the source language refers to one member of a class. Early Christian manuscripts also commonly used NOMINA SACRA for THEOS, but that scribal convention is not equivalent to modern English capitalization either.
DeleteThe argument that the traditional translation is essentially inherited from the Latin Vulgate is also much too strong. Latin articlelessness can certainly obscure a Greek article contrast if one is working backward from Latin, and the historical influence of the Vulgate on Western biblical language is beyond dispute. But that historical observation cannot decide what the Greek means. A direct Greek-to-English analysis can still arrive at “the Word was God,” and modern grammarians who do so are not merely copying Jerome. The same applies inversely: because Greek and Coptic grammaticalize determination differently, neither an English article nor a Coptic article can simply be projected back into Greek as though all three languages had isomorphic determiner systems. The question has to be resolved at the level of the source construction and discourse.
The Sahidic Coptic evidence deserves especially careful treatment because here there is a real morphological fact that should not be minimized. The Sahidic predicate at John 1:1c contains the singular indefinite article OU- before NOUTE. The recoverable Sahidic textual tradition is stable on this point. It is therefore perfectly legitimate to say that the Sahidic translator represented the Greek predicate with morphologically indefinite determination. A Trinitarian answer that denies this is simply mistaken.
The mistake comes in the next inference: “Sahidic has an indefinite article; therefore the sentence semantically means precisely what contemporary English ‘the Word was a god’ ordinarily means.” Bentley Layton’s standard Sahidic grammar blocks that inference. In his discussion of the indefinite article, he allows an indefinite expression to actualize either a specimen of a lexical class or one characterized by the quality of that class. In his discussion of gendered common nouns, he explicitly distinguishes denotation from description. Most decisively, when he discusses nominal-sentence predicates, he uses the exact expression OU-NOUTE PE and allows both “he is a god” and “he is divine.” In other words, the very Coptic construction appealed to in the argument has both a denotative/classifying and a descriptive/qualitative interpretation.
This is not a desperate Trinitarian reinterpretation of Layton. His pedagogical treatment makes the same distinction. He can give an interlinear morphological gloss corresponding to “a-god” while rendering John’s sentence “And the Word was God.” That is methodologically significant. It shows that even a leading Coptic grammarian does not regard morpheme-by-morpheme reproduction of the indefinite article as automatically determining the idiomatic English sentence. Likewise, George Horner’s famous rendering “and [a] God was the Word” is often quoted without sufficient attention to his brackets. Horner explains that square brackets represent words present in the Coptic but not required by English. His typography therefore distinguishes Coptic morphological representation from English sentence-level necessity.
Lambdin does not overturn this. His introductory remark that Coptic definite and indefinite article usage “corresponds closely” to English is a pedagogical generalization, not a declaration of one-to-one equivalence in every syntactic environment. His treatment immediately requires qualifications and exceptions. More importantly, Layton’s fuller grammatical analysis directly addresses predicate semantics and the very lexeme NOUTE. Thus the responsible conclusion from Sahidic is not “the ancient Coptic translators proved the NWT.” It is more limited: the translator selected morphologically indefinite Coptic determination for the Greek anarthrous predicate, and that Coptic form can express either classifying “a god” or descriptive “divine” predication. Morphology alone does not choose between them.
DeleteNor does the pre-Nicene date of the Sahidic translation solve the semantic problem. If the translation predates Nicaea, that is historically interesting, but an early date cannot turn an ambiguous Coptic predicate function into a doctrinal confession. “Pre-Nicene” means only that the wording cannot have been created in response to the Nicene council. It does not mean that its indefinite morphology necessarily carries a Watchtower ontology. The linguistic question remains exactly the same before and after 325: what semantic function does that indefinite predicate perform?
The Coptic data actually provide a useful warning against a mechanical Greek-to-Coptic article rule. In the reconstructed dataset studied by Wright and Ricchuiti, Greek anarthrous singular THEOS was frequently represented with the Coptic definite article rather than the indefinite one. Whatever refinements one makes to their semantic conclusions, the distribution itself is enough to show that Greek anarthrous THEOS did not mechanically produce Coptic OU-NOUTE. The Coptic translator was making target-language choices. This makes the Sahidic evidence genuinely interesting evidence for ancient interpretation, but it also means that the Coptic construction must be interpreted according to Coptic predicate grammar rather than treated as an English translation printed seventeen centuries early.
You also appeal to a supposed chronological progression within John 1:1 based on KAI, so that the Word first “was,” then was with God, and only subsequently became or was THEOS. I do not think the grammatical argument works. The quotations you provide establish that KAI can connect temporally successive events. Nobody disputes that. They do not establish that KAI itself encodes temporal succession wherever it coordinates clauses. Indeed, the Danker material you quote describes the primary lexical feature as connective and acknowledges a broad range of context-generated glosses. The Cambridge Grammar wording you quote specifically speaks of sequences “for instance in narratives.” A conjunction capable of linking successive events does not thereby make all coordinated states successive.
That distinction is decisive in John 1:1 because the clauses do not narrate three bounded events such as “he entered, and then sat, and then spoke.” They use the same stative imperfect ĒN. “The Word was”; “the Word was with God”; “the Word was THEOS.” Nothing in KAI supplies a transition such as “then became,” “afterward became,” or “as a result became.” To translate the conjunction as “and so” and then use that gloss to infer three chronological stages reverses the proper direction of interpretation. You must first establish from the clauses that the situations are successive before and then/and so becomes an appropriate contextual English realization. You cannot choose the sequential gloss and then use the gloss as evidence that they are successive.
Your own quotation from discourse-linguistic work only says that KAI can link “a series of chronologically successive events” where the discourse contains such a series. That is very different from saying that coordination by KAI causes the events or states to be chronological. Otherwise innumerable Greek descriptions consisting of coordinated simultaneous properties would become sequences merely because KAI stands between them. The argument therefore does not establish that the Word was first something other than THEOS and later acquired that predicate. Your paper itself presents the stronger conclusion that one “may not assume” the Word was THEOS during the state expressed in 1:1a–b, but that conclusion is much stronger than anything supplied by the conjunction.
DeleteThe appeal to Athanasius does not establish such temporal sequencing either. The material you quote explicitly distinguishes “before” in terms of origination or principle from temporal priority and explicitly describes the Son as existing before times and ages. To paraphrase “in the beginning” as “in the Origin, that is, in the Father” is not to say that clause a happened, clause b happened later, and clause c happened later still. Whatever one thinks of Athanasius’s metaphysics, that passage cannot function as independent evidence that the three ĒN clauses narrate successive chronological stages.
The claim that THEOS is not anarthrous “to identify the subject” requires another distinction. If somebody argues that John omitted the article before THEOS solely because otherwise the reader would not know which nominative is the subject, that is too simplistic. The articular HO LOGOS already strongly identifies the subject. But this does not make articlelessness semantically irrelevant. An articular predicate HO THEOS would make the relation much more naturally identifying or convertible in the immediate context: “the Word was the God.” John instead gives an anarthrous predicate, entirely appropriate for saying what the Logos was without collapsing him into the person whom he was “with.” This is exactly the distinction Trinitarian interpretation requires. There is no need to make an apologetic claim that the missing article exists merely as a subject-labeling device.
Smarius raises a more sophisticated objection by arguing that THEOS in Greek normally remains a generic noun rather than becoming an abstract “God-ness” noun and that a Gentile reader accustomed to Greek polytheistic language could consequently understand John’s wording as one deity beside another. That is linguistically possible as a reconstructed reader response. It is not, however, evidence that this is what John’s discourse selects. Smarius himself recognizes that the dominant interpretation is qualitative and that the question is not settled by simple syntax.
A qualitative analysis does not require pretending that THEOS has lexically transformed into the abstract noun THEOTĒS. This is an important point. Predicate function and lexical category are not the same thing. When a noun functions as a property-denoting predicate, it need not cease to be the noun it was. Saying that THEOS describes what the Logos is does not require replacing its lexical meaning with a separate abstract lexical item “God-ness.” The same noun can refer in one construction and predicate in another. Consequently, demonstrating that THEOS remains a normal common noun does not decide whether John’s particular predicate foregrounds class membership or the character/nature associated with the noun.
Smarius’s use of Longus therefore gives us an excellent control case, but not the result he needs. In Longus, “Eros is THEOS” becomes “Eros is a god” because the Greek religious discourse already contains a plurality of gods. That tells us what the construction looks like when a god-class is independently active. John’s opening is not embedded in Longus’s pantheon. It begins with EN ARCHĒ, immediately speaks of God, and moves directly into the coming-to-be of all things. The lexical and narrative field is the biblical creation field, not an unmarked catalogue of divine beings. Consequently, transferring the English classifying article from Longus to John without carrying over the polytheistic discourse that licenses it proves too much.
DeleteThe appeal to Justin Martyr is historically interesting but likewise not determinative. Justin’s willingness to use expressions such as “another god” demonstrates that a second-century Christian writer could employ THEOS language for the Logos while distinguishing him from the Father. It cannot prove how the predicate of John 1:1c functioned in John’s own discourse. Reception history can establish that an interpretation existed; it cannot replace syntax, semantics, and immediate context. It is also important not to slide from Justin’s own Logos theology directly into the later Watchtower proposition that the Logos is a created angelic member of a class of subordinate divine beings. Those are not identical propositions.
The same methodological problem affects the larger thesis that early Jewish or Christian monotheism could accommodate a “second, noncompetitive deity.” Even supposing that such a conceptual category can be demonstrated historically, that would show only that “a god” is theologically conceivable within some ancient monotheistic systems. It would not show that John 1:1c actually selects that sense. Historical possibility is not lexical selection. The clause still requires us to ask what semantic restriction tells us that THEOS means a subordinate member rather than predicating deity without attenuation. No such restriction appears in 1:1c itself.
This is also where John’s immediate creation language becomes relevant to the specific “a god” interpretation. John 1:3 says that all things came into being through the Logos and that apart from him nothing came into being that came into being. Whatever punctuation one adopts at the 1:3–4 boundary, the first assertion remains: PANTA DI’ AUTOU EGENETO. There is a deliberate verbal contrast in the prologue between what the Logos was—ĒN—and what creation came to be—EGENETO. If “a god” is then interpreted, as in Watchtower theology, to mean a created heavenly being, the discourse immediately produces a problem: the Logos is presented as the mediating agent through whom the totality of what came into being came into being, without an expressed exception for the Logos himself. This does not mean that THEOS lexically means “uncreated.” It means that the specifically created-member interpretation of “a god” is not coming from the predicate and sits awkwardly with the discourse into which the predicate is placed.
That is a broader problem with appeals to Exodus 7, Psalm 82, pagan gods, heavenly beings, or “mighty ones.” Each of those supplies an independently marked category: appointed representative, judicial figure, pagan deity, or some other restricted use. To interpret John 1:1c in the same way, you first have to establish that John is invoking that category. The article does not do it. Preverbal position does not do it. Countability does not do it. The possibility of plural THEOI does not do it. And the English indefinite article certainly cannot do it, because whether to use that English article is the very question being decided.
John’s repetition in 1:2 likewise does not require a second deity reading. “This one was in the beginning with God” reinforces precisely the personal distinction established in 1:1b. That is perfectly meaningful after 1:1c on a qualitative Trinitarian reading: the Logos who has just been predicated as THEOS is nevertheless not the same person as the one with whom he was. Verse 2 is therefore not rendered “superfluous” by “the Word was God.” It guards against exactly the misunderstanding that the English wording can otherwise invite: that the Word and the Father are one person.
DeleteThis is also why the statement sometimes made in the papers that the traditional translation makes the text logically contradictory is misplaced. “The Word was with God” and “the Word was God” are contradictory only if the second sentence is assumed to be an identity proposition between two persons: “the Word was the same person as God.” That is not the qualitative Trinitarian construal. Once THEOS is predicative, the structure is coherent: the Logos is personally in relation with the one identified as God, and the predicate associated with that divine reality is affirmed of the Logos. Smarius himself records the widespread scholarly judgment that the anarthrous predicate does not simply identify the Logos with HO THEOS.
The first anarthrous THEON of John 1:18 is another useful caution against treating articlelessness as an English-indefiniteness signal. Even Smarius acknowledges that “no one has ever seen THEON” refers to the Father, despite the absence of the article there. The syntactic construction differs from 1:1c, so it is not a direct parallel, but that is precisely the point: Greek determination is construction-sensitive. Anarthrousness is not a lexical switch turning “God” into “a god.” Context and syntax determine function.
The same principle explains why proper-name arguments do not resolve the question. THEOS is morphologically a common noun, and the Trinitarian case does not require claiming that it is always a proper name. Biblical Greek can use THEOS with or without the article for the same divine referent in different constructions. What matters at 1:1c is that it is predicative. English God has developed proper-name-like behavior in monotheistic discourse and can therefore occur as a bare predicate in ways an ordinary noun such as dog cannot. That is an English fact, not a Greek rule, but it means that the argument “if THEOS is countable, English must say a god” is too simple.
Nor is the indefinite pronoun TIS a useful way to rescue either side. I agree with you that John did not need to write THEOS TIS merely in order for an indefinite predicate to be possible. TIS would introduce its own particularizing nuance, something like “some/certain god,” and ordinary classifying predication does not require it. Therefore the absence of TIS does not disprove “a god.” But once that weak objection has been removed, nothing positive has yet been established for “a god.” It simply returns us to the original question: which semantic construal does the discourse support? This point is explicitly conceded by a careful analysis of the Greek: eliminating the TIS objection establishes grammatical availability, not token-level selection.
The deepest methodological problem running through the four papers is therefore the repeated movement from “can” to “is.” A preverbal anarthrous predicate can be indefinite. Therefore, it is suggested, THEOS here is likely indefinite. A qualitative count predicate can also be expressed with English “a/an.” Therefore, “a god” is presented as a natural qualitative translation. THEOS can denote pagan or secondary gods. Therefore, a secondary-deity sense is considered readily available. Sahidic OU- is morphologically indefinite. Therefore, the ancient translator is presented as saying substantially what the New World Translation says. Greek KAI can link successive events. Therefore, John 1:1a–c is presented as a temporal sequence. Each premise contains a legitimate linguistic observation. In each case, however, the conclusion needs an additional argument establishing that the available option is the one selected in this token.
DeleteThat is why the distinction between grammatical possibility and contextual selection is so important. No responsible Trinitarian analysis needs to claim that Greek grammar mechanically spells out the Nicene Creed. It does not. John did not write HOMOOUSIOS, and the absence of the article does not encode “consubstantial.” But the opposite overclaim is equally unjustified. Greek grammar does not provide a mechanically indefinite “a,” does not create a class of lesser gods, and does not assign the Logos a lower ontological rank. The clause predicates THEOS of the Logos while distinguishing him personally from the referent of TON THEON. The immediate discourse gives no representative, pagan, judicial, scalar, derivative, or creaturely restriction on that predicate. That is why an unattenuated qualitative interpretation is stronger than the specifically Watchtower reading.
For the same reason, “the Word was God” remains defensible as the most compact traditional English rendering, provided it is properly explained. It does not mean “the Word was the Father.” It means that the predicate THEOS is affirmed of the Logos without individuating him as a second lesser god. “What God was, the Word was” makes that semantic relation more explicit and is an excellent explanatory paraphrase. “The Word was divine” can also communicate the qualitative force if divine is understood strongly rather than as merely “godlike.” “The Word was a god,” by contrast, is not wrong merely because Greek has no indefinite article; that familiar objection is inadequate. Its real problem is that ordinary English a god overtly packages the Logos as an individual member of a nonunique god-class. That additional classifying structure is not established by the articlelessness, word order, countability, comparative predicates, or Sahidic morphology on which your papers rely.
So the issue is not whether “a god” can be manufactured as a grammatically possible English realization of an anarthrous Greek predicate. It can. The question is whether John has given us positive reason to select precisely that classifying interpretation rather than a qualitative predication. The examples you present establish the former possibility very effectively; they do not establish the latter selection. Once that distinction is maintained, Harner’s qualitative analysis, the broader distribution of Johannine predicates, the ordinary anarthrousness of predicate nouns, the absence of a contextually activated lesser-god class, the Genesis-creation framework of the prologue, the distinction between ĒN and EGENETO, and even the more precise evidence from Sahidic Coptic converge in the same direction. The Logos is distinct from the Father as person, but THEOS is predicated of him without the diminution that the expression “a god” ordinarily communicates in English.
DeleteThat conclusion is Trinitarian in compatibility, but it does not cheat by putting fourth-century metaphysics into a first-century article. It makes the more limited linguistic claim that John 1:1c is best read as an unattenuated, non-identifying predication of deity. The Nicene question of how that predication and the personal distinction are to be integrated metaphysically comes afterward. The specifically Jehovah’s Witness conclusion—that the Logos is one created, subordinate divine being alongside the supreme God—comes afterward as well. The difference is that nothing in the morphology of THEOS ĒN HO LOGOS, in the English need for an indefinite article with ordinary count nouns, or in Sahidic OU-NOUTE independently supplies that latter ontology. That is why I do not think the four papers succeed in establishing the New World Translation’s “the Word was a god,” even though they do succeed in correcting several simplistic arguments that should not be used against it.
Colwell's Rule does not claim that every anarthrous predicate nominative (APN) preceding the verb must be definite. Rather, it states that when a predicate nominative preceding the verb is definite in meaning, it tends to lack the article. Thus, the absence of the article in θεός does not automatically make it indefinite (a god). Instead, it requires context to determine whether the term is definite, indefinite, or qualitative. Colwell himself acknowledged that his rule is descriptive, not prescriptive. It observes patterns in the NT but does not dictate meaning. The context of John 1:1 suggests that θεός is being used qualitatively to describe the nature of the Logos as fully divine, rather than introducing a secondary or lesser god. Colwell’s Rule supports that an anarthrous θεός can still convey definiteness when supported by context. In John 1:1b, the article is used in ὁ θεός to refer specifically to "God" (the Father). The qualitative use of θεός in 1:1c avoids equating the Logos with the Father in person but asserts that the Logos shares fully in the divine nature.
ReplyDeleteWhile it is true that APNs can sometimes be indefinite, this depends on context and not purely on syntax. John 1:1c does not stand in isolation but must be understood in the broader theological and literary context of the Gospel of John. Examples like Acts 28:4 differ contextually and theologically. These examples describe individuals in categories, whereas John 1:1c speaks to the divine essence of the Logos. The Word order in John 1:1c emphasizes θεός for its qualitative aspect. Placing the predicate nominative (θεός) before the verb highlights the nature of the Logos, not its identity as a specific "god" among many.
The qualitative interpretation of θεός in John 1:1c, favored by most scholars, does not support the JW theology of Jesus as a secondary or lesser god. Instead, it asserts that the Logos possesses the very nature of deity. In Koine, a qualitative predicate nominative describes the essence or nature of the subject. Here, θεός conveys the divine essence of the Word, aligning with the prologue’s overarching theme of the Logos as fully divine. John 1:3 and Col. 1:16 affirm the Logos as the agent of creation, not a subordinate being. If the Logos were merely "a god," such as a created being, these verses would be nonsensical. Wallace notes that the qualitative interpretation of θεός in John 1:1c avoids the heresies of Modalism (confusing the Logos with the Father) and Arianism (making the Logos a lesser god). Wallace explicitly rejects the indefinite rendering of "a god," stating that this interpretation introduces henotheism, which contradicts Johannine theology.
The indefinite rendering "a god" directly conflicts with the monotheistic framework of both the OT and the NT and the immediate context of John’s prologue. John 1:1–18 consistently presents the Logos as divine, eternal, and preexistent. The Logos is not portrayed as a created being but as the agent of creation (John 1:3). The contrast between ὁ θεός ("the God") in 1:1b and θεός in 1:1c highlights the distinction of persons within the Godhead, not a difference in nature. Isa. 43:10 and Deut. 6:4 affirm the existence of only one true God. Rendering John 1:1c as "a god" introduces theological confusion, suggesting polytheism. Phil. 2:6 describes Christ as existing in the "form of God", further affirming His full divinity.
The article lists numerous verses where APNs are indefinite, arguing that this undermines Colwell’s Rule. While exceptions exist, they do not apply to John 1:1c. Each case must be evaluated within its specific grammatical and contextual framework. John 1:1c is unique because it deals with the nature of the divine Logos in a highly theological context. Colwell’s Rule is not absolute but descriptive, exceptions do not invalidate its general observation when supported by context. John 1:1c is distinct from the examples provided, none of the listed verses carry the theological weight or parallel the unique Christological focus of John’s prologue.
Your discussion contains an important criticism that I think a Trinitarian should simply concede rather than evade: Colwell’s rule, properly understood, does not prove that THEOS in John 1:1c is definite, and it certainly cannot be used as a mechanical rule saying, “an anarthrous predicate nominative occurring before the verb must be definite.” Some older Trinitarian apologetics did misuse Colwell in approximately that way. But once that point is granted, most of your subsequent argument does not actually follow, because you repeatedly test Colwell’s rule against examples that the rule was never intended to exclude. In effect, you refute the converse of Colwell’s rule and then treat that as though you had refuted Colwell’s actual observation or positively established an indefinite reading of John 1:1c. Those are three different propositions, and they need to be kept separate.
ReplyDeleteThere are also a few grammatical preliminaries in your explanation that need tightening. You describe the predicate as the portion of a sentence that “almost always contains a verb plus anything following that verb.” That may work as an elementary English-language teaching device, but it is particularly unhelpful when discussing Greek, precisely because Greek word order is comparatively flexible. A predicate nominative is not “whatever follows the verb.” It is a nominative nominal constituent predicated of the subject, commonly through a copular verb such as EIMI. It may occur before or after the copula. Likewise, grammatical case does not tell you “where” a noun should be placed in the sentence. Case morphology helps identify its syntactic and semantic relation; it does not assign a fixed linear position. Greek can move constituents for discourse, information-structural, rhetorical, and other reasons without losing their grammatical relation.
Your second list is therefore unfortunately labeled when you speak of examples “with the verb functioning as the subject.” A finite Greek verb does not become a noun phrase and function as the grammatical subject merely because an overt nominal or pronominal subject is absent. Koine Greek readily permits a subject to remain unexpressed because person and number are recoverable from the verbal morphology and context. That is a null or implicit pronominal subject, not “the verb functioning as the subject.” This does not destroy the usefulness of those examples, but it matters when you are constructing a grammatical corpus and claiming exact formal correspondence.
Your account of the Greek article also begins correctly but then becomes too English-centered. You rightly say that Greek possesses a definite article but no morphological indefinite article corresponding straightforwardly to English a/an, and you rightly acknowledge that an anarthrous noun may be definite, indefinite, or qualitative. That acknowledgment is crucial. Once you have made it, however, you cannot subsequently treat the lack of the article as though it gives “a god” some privileged default status. Anarthrousness tells you what is not overtly marked by the Greek article. It does not tell you that an English indefinite article is hidden in the sentence waiting to be supplied. Nor is “qualitative” simply a third kind of English determination alongside “the” and “a.” It concerns the semantic function of the predicate: what sort of thing the subject is, what nature or character it possesses, rather than primarily identifying a referent or introducing the subject as one member of an enumerable class. That distinction between morphosyntax, semantic construal, English realization, and subsequent ontological inference is essential in John 1:1c.
The same caution applies to your discussion of capitalization. You are substantially right that the ancient Greek manuscripts do not encode the modern English opposition God/god by means of our uppercase-versus-lowercase spelling convention. Capitalization is therefore not an argument that can decide John 1:1c. But the historical statement should be qualified. Early Christian manuscripts very frequently write THEOS as a nomen sacrum, ordinarily in contracted form such as THS with a supralinear stroke, rather than simply writing out THEOS every time. That scribal convention is historically significant, although it still does not correspond to the modern English distinction God/god and cannot settle this argument. Your larger point survives: capitalization is a target-language orthographic decision. But precisely because it is a downstream translation decision, neither lowercase god nor uppercase God can be fed backward into the Greek as though the spelling itself supplied semantic evidence.
DeleteThere is a related overstatement in your description of English proper nouns. English does not invariably use proper names without an article, and common nouns do not invariably require an article. More importantly for this verse, English God has developed conventional unique-reference and proper-name-like behavior in monotheistic discourse, while still being the same lexical noun that can occur in ordinary common-noun uses such as “a god,” “the gods,” or “their god.” Consequently, the question is not resolved by deciding in advance that THEOS is either grammatically a “proper noun” or grammatically a “common noun.” The Greek lexeme THEOS certainly can be countable and certainly can denote pagan gods or other beings called gods. Nobody needs to deny that. What must be established is what semantic construal is active in this particular predicate in this particular discourse.
When you arrive at KAI THEOS ĒN HO LOGOS, your basic syntactic analysis is sound. HO LOGOS is the grammatical subject and THEOS is the predicate nominative. The article on HO LOGOS is an important formal clue in a clause containing two nominatives. But your next statement—that the lack of an article before THEOS is “the sole reason” that permits “a god” or “divine”—collapses several levels of analysis. The absence of the article leaves an indefinite construal grammatically available; it does not positively select it. A qualitative construal is likewise not generated merely by deleting an article. What has to be analyzed is an anarthrous, preverbal predicate nominal, its lexical semantics, the information structure of the clause, and the surrounding discourse. Only after that can you ask whether English should realize the result as “God,” “a god,” “divine,” “a divine being,” “fully God,” or some explanatory paraphrase. Neither “God” nor “a god” is literally sitting inside the Greek sentence as an English determiner phrase. This is why treating translation as article substitution leads the discussion astray.
This brings us to the most important point about Colwell. The rule you quote is commonly abbreviated so aggressively that its logical direction disappears. Colwell did not take all preverbal anarthrous predicate nominatives and discover that they were definite. He started with predicate nouns that he had independently classified as definite and studied how their article usage correlated with their position relative to the verb. In his own tabulation, among definite predicates preceding the verb, 97 were anarthrous and 15 were articular; among definite predicates following the verb, 229 were articular and 26 were anarthrous. In other words, the relevant observation was statistical: a predicate already judged definite was very likely to lack the article when it preceded the verb. Colwell eventually formulated his conclusions explicitly with qualifications such as “usually,” and called the rules “tentatively formulated.”
That logical direction changes almost everything about your proposed counterexamples. From the proposition “if a predicate is definite and precedes the verb, it will usually be anarthrous,” you cannot validly infer the converse proposition “if a predicate is anarthrous and precedes the verb, it is definite.” Daniel Wallace has rightly criticized precisely that converse fallacy, and contemporary Trinitarian argument should not make it. The distinction is also explicitly recognized in the linguistic analysis of John 1:1c: Colwell’s results may explain why an independently definite predicate can appear without the article, but they do not classify an otherwise undetermined preverbal anarthrous predicate as definite.
DeleteBut notice what follows from that correction—and what does not. It follows that Colwell alone cannot prove “the Word was God.” It does not follow that “the Word was a god” is thereby probable, preferable, or even equally probable. The criticism is logically non-directional. If somebody has argued, “THEOS is preverbal and anarthrous; therefore Colwell proves it is definite,” you have correctly identified a bad inference. But replacing it with “Colwell does not prove definiteness; therefore the indefinite rendering gains positive evidential weight” commits a different inference error. Once the converse misuse of Colwell is removed, the semantic question simply remains open and must be answered from other evidence.
This is why your long catalog of indefinite preverbal predicate nominatives does not “counter Colwell’s rule” in the way you repeatedly claim. Suppose I say, “Cars that are electric usually do not have exhaust pipes.” You do not refute that proposition by producing one hundred gasoline cars that do have exhaust pipes. You have changed the antecedent condition. Colwell was asking how predicates independently taken as definite behave with respect to article and word order. Your examples are predicates you yourself classify as indefinite. If they really are indefinite, they fall outside the class about which Colwell’s principal statistical observation was made. They can certainly refute the converse proposition, “all preverbal anarthrous predicates are definite,” but that proposition is not Colwell’s rule properly formulated.
That distinction is not some newly invented rescue device. Colwell himself explicitly admitted that an anarthrous predicate before the verb could be indefinite when the context required it. His article says that one cannot translate a preverbal predicate as indefinite or qualitative “solely because of the absence of the article”; it then says contextual judgment remains necessary. Colwell himself went further and thought John’s context favored definiteness, and it is perfectly legitimate to disagree with that semantic judgment. Philip Harner did disagree with it four decades later. But the relevant history is therefore not “Colwell proposed an absolute word-order rule, counterexamples destroyed it, so the indefinite reading wins.” The actual history is that Colwell discovered a significant correlation involving predicates he considered definite; later scholars recognized that the semantic category “qualitative” had not been adequately analyzed in his framework; and the debate shifted from crude definiteness statistics to the semantics of predication.
Your treatment of 2 Peter 1:17 provides a useful illustration of this. You are correct about an important textual fact. P72 agrees with Codex Vaticanus in the reading HO HUIOS MOU HO AGAPĒTOS MOU HOUTOS ESTIN, placing the articular nominal expression before the copula. P72 was not available to Colwell in 1933, and its subsequent discovery unquestionably strengthened the manuscript evidence for the reading he regarded as improbable. Contemporary apparatus evidence indeed lists P72 and B for that word order.
DeleteSo I would not defend Colwell’s textual-critical dismissal of that reading by simply repeating what he wrote in 1933. But you are asking that datum to prove far more than it proves. P72 shows that the articular preverbal construction existed here and therefore that this locus really can be an exception to the tendency Colwell observed. It does not follow that the statistical tendency disappears, and it certainly does not tell you whether THEOS in John 1:1c is qualitative or indefinite. A statement containing “usually” is not logically falsified by the existence of an exception, especially when the original study itself counted approximately fifteen articular definite predicates before the verb. Colwell did not discover P72, but he already knew that the Vaticanus reading constituted an exception if original. P72 strengthens that exception; it does not somehow transform an indefinite predicate elsewhere into the correct analysis of John 1:1.
It is also worth noting that modern critical editions themselves do not treat the textual problem at 2 Peter 1:17 as though the discovery of P72 simply ended all discussion. The SBL Greek New Testament follows the P72-B order, whereas the Tyndale House Greek New Testament prints the HOUTOS ESTIN HO HUIOS... order. That disagreement does not diminish P72’s importance. It simply illustrates why “P72 agrees with B” and “therefore we have proved what the autograph necessarily read” are not identical statements. In any event, whichever reading you prefer, nothing at 2 Peter 1:17 establishes the semantics of THEOS at John 1:1c.
I would make a similar concession regarding Colwell’s explanation of the Corinthian exceptions as a temporary Pauline “stylism.” You are entitled to regard that explanation as speculative. “Possibly for the sake of greater emphasis” is not an independently demonstrated grammatical law. But again, rejecting Colwell’s explanation for why particular definite predicates depart from the dominant pattern does not invalidate the distinction between a statistical tendency and a deterministic rule. His own table gave 15 preverbal articular definite predicates out of 112 preverbal definite predicates, approximately 13 percent. The pattern was never exceptionless. The serious modern criticism of Colwell is therefore not “he had exceptions, therefore no pattern exists.” It is that semantic classification had to precede his statistics, and that his framework did not adequately distinguish qualitative predication from definiteness. That is a much stronger criticism—and, importantly, it does not establish the New World Translation.
This becomes especially clear when we examine the examples you call “precisely” parallel to John 1:1. Acts 28:4, PHONEUS ESTIN HO ANTHRŌPOS HOUTOS, is genuinely helpful. “This man is a murderer” is plainly natural English. It proves that an anarthrous preverbal predicate noun can be indefinite and that predicate-before-copula word order does not itself prevent indefinite classification. A Trinitarian has no reason to dispute either point. But murderer is independently an ordinary enumerable human class. The Maltese observers are classifying Paul as a member of that already available class because they infer from the snake incident that he must be guilty of murder. The indefinite article in English is therefore licensed not merely by PHONEUS being preverbal and anarthrous, but by the lexical semantics and discourse situation.
DeleteThe same is true of John 4:19, PROPHĒTĒS EI SY. Prophets are a familiar nonunique class in the discourse world. The woman concludes that Jesus is “a prophet.” John 18:37 similarly employs BASILEUS, a social and political status capable of application to multiple individuals across different places and times. “You are a king” is entirely natural. These texts are important evidence that John can place indefinite count predicates before the copula. But they do not establish that every lexeme appearing in that construction must activate the same kind of nonunique class. The decisive question is not simply whether THEOS is morphologically capable of count use—it certainly is—but whether John 1 activates an ordinary nonunique class of gods such that the Logos is being introduced as one member of it. The formal construction permits that analysis; these role nouns do not supply it. This is precisely why the familiar comparison set of PROPHĒTĒS, SAMARITĒS, and BASILEUS establishes grammatical availability but not token-level selection.
Your Acts 9:15 example makes the same point rather than strengthening the inference to John 1:1. SKEUOS EKLOGĒS is a metaphorical role-description, additionally modified by a genitive, and Paul is being characterized as a chosen instrument. There is no contested ontology of “chosen vessels” comparable to the very question under dispute concerning THEOS. You know independently what sort of classification is occurring, so English naturally supplies an indefinite article. That does not tell you how English should realize an unmodified divine predicate in a monotheistic prologue.
First Timothy 6:10 is even less suitable as a decisive control because the semantic classification of RHIZA PANTŌN TŌN KAKŌN has itself been debated. Colwell regarded it as definite; many modern English translations render it “a root of all kinds of evil.” If you prefer “a root,” that is entirely compatible with the point I am making: Colwell’s individual semantic classifications are not infallible, and his rule should not be inverted to decide an uncertain predicate. But your argument about whether the Devil could have been motivated by love of money does not constitute grammatical evidence for the article. It is an exegetical argument about the scope of PANTŌN TŌN KAKŌN and the metaphorical proposition being made about PHILARGYRIA. Even if “a root” is unquestionably right in 1 Timothy 6:10, all you have established is another case in which the construction permits indefinite predication. You still have not established that THEOS in John 1:1 belongs to the same semantic class.
The inclusion of plural predicates in your larger lists is still less probative for the English “a/an” question. An English plural predicate does not take the singular indefinite article. Arguing that if “you became imitators” then each individual is “an imitator” is a perfectly reasonable distributive inference, but it does not turn the source-language plural construction into evidence about how a singular anarthrous THEOS must be determined in English. If the issue being tested is specifically whether singular anarthrous preverbal predicate nouns should receive English a/an, plural predicates add very little unless their contribution is carefully distinguished from the singular count-noun problem.
DeleteThis is why later research after Colwell matters. Philip Harner’s 1973 article did not simply repeat Colwell. He specifically asked whether preverbal anarthrous predicate nouns often foreground quality rather than definiteness. In discussing John’s usage he concluded that many such predicates were primarily qualitative. His treatment of John 1:1 is especially important because he did not argue that an indefinite use was grammatically impossible. Quite the contrary: he explicitly distinguished different constructions John could have used. His conclusion concerning the actual THEOS ĒN HO LOGOS was that the predicate is primarily qualitative, indicating that the Logos possesses the nature denoted by THEOS, rather than primarily identifying the Logos with the person referred to as HO THEOS or placing him as one specimen in a generic category of gods.
That does not mean Harner invented another exception to save Trinitarianism. Look at the broader distribution he was analyzing. On the page preceding his discussion of John 1:1, he notes that John repeatedly uses anarthrous preverbal predicates in which the qualitative aspect is prominent and where the choice between “definite” and “indefinite” simply fails to capture everything the predicate is doing. John 1:14 is the obvious type of example: SARX EGENETO does not naturally mean “the Word became a flesh,” nor is SARX functioning as a definite identifying noun phrase. It characterizes what the Logos became. Harner accordingly argued that predicate semantics must be treated independently from the binary definite/indefinite question.
Paul Dixon’s later Johannine study likewise reported a strong qualitative tendency among anarthrous predicate nominatives, and Daniel Wallace subsequently emphasized both points that matter here: the converse misuse of Colwell is invalid, and John 1:1c is best analyzed primarily as qualitative rather than as a simple definite-versus-indefinite choice. Donald Hartley later refined the corpus question by taking count versus mass behavior more seriously. None of those studies should be converted into a new mechanical rule. Dixon’s famous classification statistics are human semantic judgments, not a 94-percent mathematical probability that any new token must be qualitative. Hartley’s narrower count-noun statistics likewise contain disputed classifications. But collectively they show why the debate did not end with “Colwell has exceptions.” The real advance was recognizing that semantic predication cannot be reduced to English article choice. Harner’s article itself is a published JBL study specifically devoted to this issue, while Hartley explicitly frames his project as an attempt to correct the misuse of the “Colwell construction.”
Indeed, one of the strongest contemporary arguments for your side actually demonstrates why your conclusion needs to be stated much more cautiously. Alexander Smarius has argued in a peer-reviewed 2022 study that THEOS in John 1:1 and 1:18 can plausibly be understood in a generic or kind-denoting way and that the Logos may be construed as a second, noncompetitive deity. That is a serious linguistic proposal, not something that should simply be dismissed because its conclusion is non-Nicene. But notice how Smarius has to argue it: not by saying, “the predicate is preverbal and anarthrous, therefore a god,” but by constructing a wider linguistic and contextual account of THEOS, divine kinds, Johannine monotheism, and ancient divine hierarchy. His paper is useful precisely because it illustrates the difference between grammatical possibility and contextual selection. “A god” is a legitimate hypothesis requiring argument. It is not an output automatically generated by the missing article.
DeleteThat brings us to what your comparison with murderer, prophet, king, vessel, root, and the other predicates does not supply: an independently established divine class in John 1. THEOS is certainly capable of ordinary count use. Greek has THEOI. Pagan discourse can speak naturally of “a god.” Acts 28:6 is an excellent New Testament example: the Maltese conclude concerning Paul that he is THEOS, and an English translation “a god” is entirely natural because the pagan discourse explicitly supplies a plurality of gods as the relevant ontology. The formal possibility is therefore beyond dispute. But John 1:1 is not Acts 28:6. In 1:1b the Logos is PROS TON THEON, where THEOS identifies the divine referent with whom the Logos exists in personal relation; in 1:1c the same lexical content THEOS is then predicated of the Logos. The question is what John is saying about the Logos by that predication.
The immediate continuation matters because the specifically Jehovah’s Witness reading is stronger than the mere proposition that “a god” is grammatically possible. It understands the Logos as one discrete divine being distinct from Almighty God, ontologically subordinate to him, and created. None of those additional semantic features is contained in the absence of the article. John immediately says that all things that came into being came into being through the Logos, and the clause itself supplies no exception placing the Logos inside the class of things that came into being. Later Johannine usage can apply THEOS directly to Jesus as well. The cumulative point is not that John 1:1c, by grammatical morphology alone, recites fourth-century Nicene terminology. It does not. The point is that the move from anarthrous THEOS to “one lesser created member of a class of gods” requires substantially more semantic information than the construction provides.
This also explains why the traditional “the Word was God” need not mean what you appear to assume a “definite” English rendering must mean. In John 1:1b, the Logos is distinguished from the referent of TON THEON: he is “with God.” Therefore when the next clause says KAI THEOS ĒN HO LOGOS, an English reader should not understand “the Word was God” as “the Word was the same person as the God he was with.” That would contradict the immediately preceding relational statement. Predicative English God can describe what the Word is without collapsing who the Word is into who the Father is. In Trinitarian terminology, personal distinction and predication of divine nature are not competing propositions.
That is why C. H. Dodd’s famous explanatory rendering, “what God was, the Word was,” and the NET Bible’s “the Word was fully God” are useful semantic paraphrases even though neither is a wooden morpheme-for-morpheme translation. Their point is to make explicit in English what a qualitative analysis takes the Greek predication to be doing: the Logos possesses the nature denoted by THEOS while remaining personally distinct from the one with whom he is. The NET notes explicitly distinguish those two propositions, and Harner similarly says the actual construction attributes the nature of THEOS to the Logos rather than making LOGOS and HO THEOS interchangeable referring expressions.
DeleteConversely, “divine” is not simply synonymous with “a god.” This is another place where your initial three-way question—“God, divine, or a god?”—can obscure an important distinction. In contemporary English, divine can mean fully deity, pertaining to God, supernatural, godlike, or even merely excellent in colloquial usage. “A divine being” individuates a member of a class more strongly than “divine.” “Godlike” usually expresses resemblance rather than identity of nature. “A god” presents an ordinary singular common-count nominal and strongly foregrounds class membership. “God” in monotheistic English convention preserves the full lexical force of deity without necessarily making a personal identity equation. Those renderings overlap, but they do not package the same semantic information.
There is an unusually instructive ancient-version control here. The Sahidic Coptic translation of John actually does possess an overt singular indefinite article, and at John 1:1c it uses OU- with NOUTE. If morphology alone could settle this dispute, that might appear to provide your strongest possible argument. Yet standard Sahidic grammar does not permit that inference. Bentley Layton explains that an indefinitely determined common noun can function denotatively or descriptively, and in discussing the exact predicate OUNOUTE PE he gives both “he is a god” and “he is divine” as possible semantic realizations. Even more strikingly, on the page where Layton gives an interlinear morphological gloss “a-god” for John 1:1, his sentence-level English translation is “And the Word was God.” George Horner similarly rendered the Sahidic “… and [a] God was the Word,” using brackets under his convention for material represented in Coptic morphology but not required in idiomatic English. The point is not that Layton or Horner settles theology by authority. It is that even an overt indefinite article in the ancient target language does not mechanically settle predicate semantics or English translation.
Sahidic John 4:24 makes the methodological point still clearer. The Coptic translation uses a morphologically indefinite predicate corresponding to “spirit” in “God is spirit,” yet that does not mean the sentence is ranking God as one ontologically subordinate specimen among a hierarchy of spirits. The form can characterize what an already definite subject is. Likewise, Wright and Ricchuiti’s published study of Sahidic treatment of anarthrous nominative singular THEOS found that the Coptic translators did not mechanically map Greek anarthrous THEOS to the Coptic indefinite article; their study was specifically designed to ask whether such a uniform correspondence existed. If even a language with an overt indefinite determiner requires predicate semantics and context before you can move from morphology to English a/an, the absence of a Greek article certainly cannot by itself generate “a god.”
So when you conclude that your examples show Colwell’s rule is “ineffective in insisting that the definite God must be the translation of John 1:1 simply because of word order,” I can agree with that sentence almost exactly as written. Word order plus Colwell’s rule does not compel a definite analysis. A careful Trinitarian should not argue otherwise. Walter Bauer’s warning against crude hard-and-fast article rules is salutary. Greek article usage is too sensitive to syntax, discourse, lexical semantics, modification, and predicative function for a simplistic substitution formula.
DeleteBut your next move is where the argument fails. The existence of indefinite predicates in the same broad construction proves that an indefinite interpretation is grammatically possible. It does not establish that it is semantically preferred at John 1:1c. Your large lists do not constitute “exceptions to Colwell’s rule” if the nouns in those lists are independently indefinite, because Colwell’s principal generalization concerns predicates already classified as definite. Your closest parallels—murderer, prophet, king, chosen vessel, root—contain lexemes whose nonunique class structure is independently obvious. THEOS is different precisely at the point that is under dispute: whether John is activating an enumerable class of gods and assigning the Logos membership in it, or predicating the nature signified by THEOS of the Logos. You cannot decide that disputed question by importing the class semantics of murderer, prophet, or king.
The strongest conclusion your evidence supports is therefore narrower than the one toward which you are pushing it. You have successfully shown that “a god” is not grammatically impossible merely because THEOS precedes the copula, and that Colwell cannot be invoked in reverse as an automatic proof of definiteness. That is worth conceding. But once that obsolete argument is removed, the actual linguistic case still has to be made. Harner’s semantic analysis, the broader Johannine distribution of qualitative predicates, the difference between familiar role nouns and THEOS, the relational distinction established in John 1:1b, and the immediate discourse of the prologue all favor taking THEOS primarily as a qualitative, non-identifying predication. The Word is not being identified as the same person as HO THEOS of the preceding clause; neither is he naturally being introduced as one lesser specimen in an independently established class of deities. The predicate tells you what the Logos is.
For that reason, the traditional wording “and the Word was God” remains linguistically preferable—not because Colwell’s rule mechanically inserts a missing definite article, and not because grammar by itself somehow contains the Nicene Creed, but because “a god” adds an individuating, classifying contrast that the Greek construction does not itself encode and that John’s immediate discourse does not independently establish. The irony, then, is that your most important criticism is correct but does not accomplish what you need it to accomplish. You have shown why a Trinitarian should stop trying to make Colwell decide John 1:1c. You have not shown that John therefore wrote, in semantic effect, “the Word was a god.”
If John Intended a Lesser Divine Being, Why Did He Use θεός?
ReplyDeleteAncient Greek possessed a vocabulary capable of distinguishing gods from other categories of supernatural, semi-divine, or godlike beings. The term ἡμίθεος designated a demigod; ἥρως could designate a hero possessing exceptional or cultically recognized status; δαίμων, in certain classical philosophical contexts, could designate an intermediate supernatural being; and adjectives such as θεῖος and θεϊκός could describe divine qualities or associations. Even comparative expressions such as θεώτερος existed for describing something as more divine or godlike.
These terms were not interchangeable, and their suitability depended on the cultural and literary context. Nevertheless, their existence demonstrates that Greek speakers possessed ways of distinguishing full deity from various forms of divine resemblance, derivative supernatural status, or intermediate religious classification.
By contrast, θεός was the ordinary Greek substantive for a god, particularly associated in classical religious usage with the major gods themselves. Its meaning was broader than the Olympian pantheon, but its primary lexical function was to designate deity, rather than merely power, exalted rank, or resemblance to deity.
This distinction raises a serious question for the Arian and Jehovah's Witness interpretation of John 1:1c.
If John intended to identify the Logos as an ontologically inferior, created, secondary divine being, why did he choose the substantive θεός without any lexical qualification indicating such inferiority? Why did he employ the same noun he had just used for God in the immediately preceding clause, rather than expressing the Logos' alleged subordinate status through an adjective denoting divine resemblance, a qualifying expression, or some more specific description?
The question becomes especially forceful when the syntax is considered. John writes καὶ θεὸς ἦν ὁ λόγος, placing the anarthrous predicate θεός before the verb. Philip B. Harner's analysis identifies this construction as primarily qualitative, emphasizing the nature of the Logos. Harner expressly considered the alternative ὁ λόγος ἦν θεῖος and concluded that John's actual formulation communicates something stronger than an unspecified degree of divine character or the mere classification of the Logos as one divine being among others.
The decisive point, therefore, concerns both lexical selection and contextual meaning. John uses the substantive θεός, repeats the very noun employed for God in John 1:1b, and places it in a construction emphasizing the Logos' nature. He then declares that everything that came into existence came into existence through the Logos (John 1:3).
If John intended a subordinate created deity, his formulation leaves that alleged ontological inferiority unexpressed while affirming the Logos' divine nature and universal creative agency.
The burden consequently rests on the Arian interpreter to explain why the supposedly inferior, created status of the Logos should govern the interpretation of a passage whose actual vocabulary, syntax, and immediate context point toward the full deity of the Word.
The Nicene reading provides a coherent explanation: the Logos is personally distinct from the Father, yet possesses the same uncreated divine nature.
* Predicative θεός in John 1:1c
https://www.academia.edu/176663425
* ⲞⲨⲚⲞⲨⲦⲈ in Sahidic John 1:1c
https://www.academia.edu/176411873