The Text of John 1:18

Andrew Perry

There is a well-known textual issue1 with v. 18 represented in the following four versions (but others could
be chosen to the same effect),

No man hath seen God at any time; the only begotten son, which is in the bosom of the Father,
he hath declared him. (KJV)

No one has ever seen God. The only one, himself God, who is in closest fellowship with the
Father, has made God known. (NET)

No man has seen God at any time; the only begotten God, who is in the bosom of the Father,
He has explained Him. (NASB)

No one has ever seen God. The only Son, God, who is at the Father’s side, has revealed him.
(NAB)

Qeo.n ouvdei.j e`w,raken pw,pote\ monogenh.j qeo.j o` w'n eivj to.n ko,lpon tou/ patro.j evkei/noj
evxhgh,satoÅ (GNT)

Qeo.n ouvdei.j e`w,raken pw,pote\ o` monogenh.j ui`o,j( o` w'n eivj to.n ko,lpon tou/ patro,j( evkei/noj
evxhgh,satoÅ (BYZ)

The text critical argument is all about the external evidence of the manuscripts and text-critical
methodology and history, and the New Testament argument is all about the internal evidence of
Johannine style and thought in a first century context, i.e., whether John would have written ‘the only son,
God’ or ‘the only begotten one, God’, or even ‘the only begotten God’, as an expression of a form of Jewish
monotheism.2

The two Greek variants represent the overwhelming bulk of manuscripts, versions and Patristic citations.
Minor variants to be noted are one which has an article before monogenh.j qeo.j and another which just has
o` monogenh.j without either substantive. Older versions align with the KJV; newer translations align with the
NET, NAB or NASB with nuanced variations. The KJV follows its Greek Byzantine text in a fairly literal
manner. The NAB, NET and NASB follow an Alexandrian text represented in the GNT but differ in how
they read monogenh.j.3

How you read monogenh.j is crucial to the Alexandrian argument, since it has to be read the same way in v.
14. The NAB takes monogenh.j to imply ‘son’; the NET restricts its translation to ‘the only one’ which looks
vacuous for supplying nothing to match -genh.j, as well as being confusing vis-à-vis Jewish monotheism.
The Alexandrian text will fare better if we read monogenh.j qeo.j appositionally either as ‘the only son, God’
or ‘the only begotten one, God’. The Byzantine text, with its use of ui`o,j, requires monogenh.j to be read as
‘only begotten’ if the compound adjective is to be respected and not treated as just equivalent to mo,noj. It

1 In English the two classic papers are, F. J. A. Hort, “On MONOGENHS QEOS in Scripture and Tradition”
in his book, Two dissertations (London: Macmillan, 1876), 1-72; and E. Abbot, “On the Reading ‘Only-
Begotten God,’ in John I.18” in his book, On the Authorship of the Fourth Gospel and Other Critical Essays
(Boston: Geo. H. Ellis, 1888), 241-271. Hort cites Abbot as one of his motivations for putting pen to paper.
Both essays are readily available online.
2 Many New Testament scholars have commented on whether John uses qeo.j of Christ in John 1:18 and

elsewhere (John 1:1, 20:28). For an introduction, see B. A. Mastin, “A Neglected Feature of the Christology
of the Fourth Gospel” NTS 22 (1976): 32-51.
3 We follow the traditional theory of text-forms with their designations because this is how scholars have

discussed the text of John 1:18, but see D. C. Parker, An Introduction to the New Testament Manuscripts and their
Texts (Cambridge: Cambridge University Press, 2008), 171-174, for a critique and a map for the future
direction of travel, which today is further along the road away from text-types.
is difficult to get something in English to match -genh.j when we have ‘son’ collocated with monogenh.j other
than by using ‘begotten’ or perhaps ‘only-born’.

The majority of textual critics today favour the Alexandrian reading (GNT); a minority favour the Byzantine
reading (BYZ). Earlier generations of NT scholars favoured the Byzantine reading; today, the bias in favour
of that reading among such scholars is less pronounced. The text critical argument boils down to how you
weigh early manuscripts, often of Egyptian provenance, that go against the consensus of the later
manuscripts from a wider geographical spread of places, i.e., the context is between a more geographically
widespread set of texts and a more narrowly located Alexandrian text.4

We have no prejudice against the Byzantine text nor a bias in favour of the Alexandrian manuscripts. As
an outsider reading and documenting the text-critical arguments for and against, the disagreement looks
stalemated because of a lack of manuscript evidence and a textual history for the 2c CE.5 However, our
view is that the internal evidence strongly favours the Byzantine reading and so we think that this is the
original text (‘only begotten son’).

External Evidence
The later and widespread manuscript evidence6 (the Western, Byzantine and Caesarean traditions) favours
the KJV and ‘the only begotten son’. So, for example, the uncial manuscripts, Alexandrinus (A, 5c CE),7
Ephraemi Rescriptusc (9c CE correction of C), and the Freer Gospels8 (7c CE, Wsupp), among many others,
have ‘the only begotten son’. This is not to say all Western and Caesarean texts have ui`o,j, just that it is more
strongly represented amongst these texts. The mass of later Byzantine uncial manuscripts has ui`o,j as well
as all the Byzantine minuscules.9

The earlier Alexandrian tradition of manuscripts favours the NET, NASB and NAB and qeo.j. So for
example, P66 (traditionally late 2c-early 3c CE) and the uncial manuscripts, Sinaiticus* (mid-4c CE, a),10
Vaticanus (4c CE, B), Ephraemi Rescriptus* (5c CE, C*), and Codex Regius (8c CE, L), along with the
DiatessaronArab (9c CE11) – these support monogenh.j qeo.j. P75 (traditionally late 2c-early 3c CE), Sinaiticusc
(a 6c-7c CE ‘correction’ of Sinaiticus*) and a 9c CE miniscule, 33, have an added article. However, some
later secondary manuscripts in the Alexandrian tradition favour ui`o,j, (T (5c CE), Υ (8c-9c CE), Δ (9c CE),
892 (9c-10c CE), 1241 12c-13c CE)) which is significant.

4 We do not presume all ‘Alexandrian’ texts originate in Alexandria.

5 This is especially the case if P66 and P75 are dated to the late 3c-early 4c (see below).
6 The evidence is conveniently listed in P. R. McReynolds, “John 1:18 in Textual Variation and Translation”

in New Testament Textual Criticism, its Significance for Exegesis (eds., E. J. Epp and G. D. Fee; Cambridge:
Cambridge University Press, 1981), 105-118. P. Comfort, New Testament Text and Translation Commentary
(Carol Stream: Tyndale House Publishers, 2008), 255-256, also lists the translations that align with each
variant, whether in the main text or in the margin.
7 This manuscript is judged in general to be an Alexandrian text-type, even though in John 1:18 it agrees

with the Western, Byzantine and Caesarean text-types; see B. M. Metzger, The Text of the New Testament (2nd
ed., Oxford: Oxford University Press, 1968), 47.
8 For an introduction to papyrology and codicology of the Freer manuscripts see B. Nongbri, God’s Library

New Haven: Yale University Press, 2018), 1-20.

9 For a list of the important miniscule manuscripts see B. J. Wright, “Jesus as ΘΕΟΣ: A Textual

Examination” in Revisiting the Corruption of the New Testament (ed., D. B. Wallace; Grand Rapids: Kregel
Publications, 2011), 229-265 (241-242), and see Metzger’s evaluation of the minuscules in his The Text of the
New Testament.
10 Metzger, The Text of the New Testament, 46, classifies Sinaiticus as in general Alexandrian, but it has been

corrected in later centuries and it has Western readings; in respect of John 1:18, it is Alexandrian although
the opening chapters do have Western readings and they could be classified as ‘Western’.
11 This copy may be influenced by local versions (the Peshitta) rather than the original Diatessaron.
P. R. McReynolds stated in 1981 that the argument in favour of the Alexandrian reading is the earlier date
of its uncials and the early date of the Bodmer papyri.12 P. Comfort commented in 2008 that the discovery
of the Bodmer papyri in the 1950s and 1960s “tipped the balance” in favour of the Alexandrian reading
and offered the hypothesis that ‘only begotten son’ was a change made in the 3c CE or earlier because it
was “more ordinary”.13 Equally, we might counter by saying ‘(the) only begotten God’ was the early change
made for theological and/or literary reasons (see below) but the wider church resisted/corrected this
change.

The agreement of Vaticanus with P75 secures the description of Vaticanus as ‘Alexandrian’ and so it has
been seen to have roots in the late 2c-early 3c. This concordance is the mainstay for those textual critics
who advocate monogenh.j qeo.j. Vaticanus was the preferred text of Westcott and Hort for constructing the
text of the Greek New Testament, a preference that has persisted to the present day in the UBS/GNT.14
But the fly-in-the-ointment for any preference given to the Bodmer papyri is that P75 has an added article
whereas P66 does not have the article. This addition changes the theological implication from something
like, ‘the only begotten one, God’, which is appositional, to ‘the only-begotten God’ which is adjectival. We
can’t really cite P66 and P75 in support of monogenh.j qeo.j without acknowledging the problem that arises
from the addition of the article.

In addition, the traditional dates assigned to P66 and P75 have been questioned recently,15 putting them
closer in time to the traditional date for Vaticanus (4c CE). This takes some of the shine off Vaticanus’
pedigree claims because it pushes these manuscripts to the late 3c-early 4c and the post-Nicene period.
Finally, it is worth noting with G. D. Fee that P66 was corrected towards the ‘Byzantine’ text-form,16 which
raises the question of what value we should place on P66 and its unaltered reading of John 1:18.

When we move onto the evidence of the Versions, McReynolds summarizes the evidence as, “This
versional evidence shows that qeo.j and ui`o,j are both early readings and both have some geographical spread,
although ui`o,j clearly dominated in the West, while the evidence is more evenly distributed in the East.”17

12 McReynolds, “John 1:18 in Textual Variation and Translation”, 105-106.

13 Comfort, New Testament Text and Translation Commentary, 255. Abbot suggests that ‘only-begotten’ would
lead the mind of the scribe to write ui`o,j, On the Authorship of the Fourth Gospel and Other Critical Essays, 270.
14 Comfort, New Testament Text and Translation Commentary, 256. G. D. Fee states in his 1968 published

doctoral dissertation, recently reprinted in his collected works, Bodmer Papyri, Scribal Culture, and Textual
Transmission: Collected Works on New Testament Textual Criticism (Leiden; Boston: Brill, 2021), 163,
that “the discovery of [P.Bodmer 14-15] now makes it certain that the text of [Codex Vaticanus] existed in
the second century.” In his essay, “P75, P66, and Origen: The Myth of Early Textual Recension in
Alexandria” in Studies in the Theory and Method of New Testament Textual Criticism (eds., E. J. Epp and G. D.
Fee; Leiden: E. J. Brill, 1993), 247-273 (272), Fee states of P75 and Vaticanus that “These MSS seem to
represent a ‘relatively pure’ form of preservation of a ‘relatively pure’ line of descent form the original text.”
15 The traditional date has been recently questioned by B. Nongbri, who argues, “…the range of possible

dates assigned to P. Bodmer II on the basis of palaeography needs to be broadened to include the fourth
century” and “Based on the evidence currently available, then, I would posit the construction and initial
copying of P.Bodmer II took place at some point in the fourth century, and at least one subsequent
rebinding (though more likely two) took place, perhaps as late as the fifth century.” in his “The Limits of
Palaeographic Dating of Literary Papyri: Some Observations on the Date and Provenance of P.Bodmer II
(P66)” Museum Helveticum 71 (2014): 1-35 (2, 32). P. Orsini, “I papiri Bodmer: scritture e libri” Adamantius 21
(2015): 60-78 (77), concludes from his papyrological analysis of P66 and P75 that P.Bodmer II is “middle
of the third century to the middle of the fourth century” and P.Bodmer XIV.XV is “the end of the third
century to the beginning of the fourth century.”
16 Fee, ibid., 259, says that the corrections of the original scribe of P66 “if one may be allowed an

anachronism, is a ‘Byzantine’ copyist at work in the second century.” Fee’s late 2c dating has been the
traditional assignment (since the editio princeps in the 1950s) and so he regards the ‘Byzantine’ corrections as
an anachronistic characterization because the Byzantine text-form is not this early in his view.
17 McReynolds, “John 1:18 in Textual Variation and Translation”, 107.
Significant evidence on the side of qeo.j is the Peshitta, and in favour of ui`o,j we have the Vulgate. The key
point here is that both readings are early, a fact supported by the Greek Fathers.

As for the Greek Fathers, McReynolds notes that 11 Fathers support monogenh.j qeo.j across 39 citations
but some of these writers also have ui`o,j.18 On the other hand, there are 20+ Fathers who cite John 1:18
more than 40+ times with ui`o,j. Eusebius illustrates the problem we have here because he cites both,

“…the evangelist expressly teaches that he is the only-begotten Son, in that passage where he
said, “No one has ever seen God, the only-begotten Son (or only-begotten God) who is in the
bosom of the Father, he has made him known,” On Ecclesiastical Theology, I.919

McReynolds claims texts in the Greek Fathers that use monogenh.j qeo.j but don’t cite the Gospel; he claims
that these are allusions to John and are evidence of the qeo.j variant. However, he is unwilling to include
references to o` monogenh.j ui`o,j without citation of John 1:18 because this idea is used in more than one place
in John. This rather loads the dice in favour of qeo.j, and we should instead accept the allusive evidence of
o` monogenh.j ui`o,j if the allusion uses the exact form of the phrase in John 1:18, as John’s use of the idea of
‘only begotten son’ in John 3:16, 18 and 1 John 4:9 employs different grammatical forms.

However, such a simple counter-argument belies the fact that the problem inherent in counting allusions
in the Fathers for and against qeo.j or ui`o,j is a complex matter. Patristic scholars assess the quality of the
allusion differently – is it word for word or a paraphrase? Is it only a reflection of general theological
terminology? Is the allusion influenced by heretical authors or a particular manuscript? Is the Father
consistent in his allusions using only one variant? On the other hand, it could be argued methodologically
that allusions on either side cannot be claimed as evidence for the original text of John 1:18 without specific
citation because we cannot be certain about the reason for the language being used.20 The classic essays by
F. J. Hort “On MONOGENHS QEOS in Scripture and Tradition” (1876) and E. Abbot “On the Reading
‘Only-Begotten God,’ in John I.18” (1861) discuss the patristic evidence in detail.21 For our purposes, we
note only that however your count the matter, statistically, the Greek Fathers favour monogenh.j ui`o,j.
Similarly, the Latin Fathers are overwhelmingly in favour of ‘only begotten son’ with only Hilary22 and
Fulgentius23 using ‘only begotten God’; this is consistent with the evidence of the Latin versions.

In sum, from the GNT committee of the 1960s, Metzger gives the Alexandrian reading a {B} second-
highest rating from a four-point scale {A}-{D} for assessing a text’s likelihood to be original.24 A. Wikgren
from the same committee disagrees and assigns a {D} reading to the Alexandrian reading suggesting it is
the result of a transcriptional error.25 The judgment of textual critics has been mixed over the years but the
majority at least from the 1960s through to the 2010s have favoured monogenh.j qeo.j. The principal reason
seems to be the evidential weight they have given to the Alexandrian text type and its earlier manuscripts,
particularly P66 and P75.

18 We should note a qualification here: some of the Greek Fathers are in Latin translation and it is possible
that the ‘only begotten son’ or ‘only begotten God’ is the translator’s choice – see Hort, Two Dissertations, 4,
who casts doubt on Irenaeus’ ‘only begotten son’ on this basis.
19 Text taken from K. M. Spooerl and M. Vinzent, eds., Eusebius of Caesarea: Against Marcellus and On

Ecclesiastical Theology (Washington: Catholic University Press, 2017).

20 This is the opening argument of Abbot, On the Authorship of the Fourth Gospel and Other Critical Essays, 243,

and it forms the jumping-off point for his down-grading of some of the Fathers’ evidence used for
monogenh.j qeo.j and his own appraisal of their evidence for the use of o` monogenh.j ui`o,j. Abbot’s ‘opponent’
in his paper is Samuel Tregelles, a leading textual critic of the day.
21 It is also worth noting that James Snapp’s January 2019 entry no. 2 on his blog

www.thetextofthegospels.com also sets out the patristic evidence in a convenient way.

22 While Hilary uses the phrase ‘only-begotten God’ he doesn’t cite John 1:18 with this reading, see Abbot,

On the Authorship of the Fourth Gospel and Other Critical Essays, 263-264.
23 The usage for Fulgentius is at most an allusion rather than a quotation of John 1:18 – see Abbot, On the

Authorship of the Fourth Gospel and Other Critical Essays, 257; see also Hort, Two Dissertations, 5-6.
24 B. M. Metzger, A Textual Commentary on the Greek New Testament, (London: United Bible Societies, 1971), 198.
25 Ibid.
However, the ubiquity of monogenh.j ui`oj, in the other textual traditions implies an equal antiquity26
(supported by the earliest Fathers, e.g., Irenaeus, who supports both readings), and we might therefore
suppose that the absence of those text-types among the earliest manuscripts is due to constant preferential
use, wear and tear and disposal, and the change in technology from uncial to cursive script and from papyrus
to parchment, whereas the survival of the early Alexandrian Egyptian manuscripts that we have is due to a
lack of use and/or happenstance with the favourable Egyptian climate. It is too simplistic to assume, a) that
early more likely equates to original; and b) that the judgement of the general superiority of a text-type in
modern times automatically transfers to every token (i.e., John 1:18) in that text type.27

Textual History and Methodology

New Testament scholars28 have been less inclined to follow textual critics for reasons internal to the text.
Before we look at internal evidence, we should consider matters of transmission history and historical
methodology. Historical hypotheses and historical methodology go together because the question of which
text is original depends on how we evaluate an Alexandrian text versus using a text like the majority
Byzantine text.29

It is worth saying up-front that the ‘textual critics’ of from perhaps as early as the 7c onwards (i.e., the more
professional scribes) came to favour monogenh.j ui`o,j in the West and in the East and they are much closer
to the evidence than the textual critic of the modern period. Furthermore, we should give due weight to
the mixed witness of the Alexandrian textual tradition because this shows later correction of the text in
favour of ui`o,j (e.g., Ephraemi Rescriptusc (9c CE correction of C); Υ, Codex Athous Laurae (8c-9c CE)). Of
course, it could be countered that this is due to peer pressure from the increasing dominance of the other
textual traditions (but then why are they dominant?).

We should also note that both an early date for a manuscript and its supposedly being part of a superior
textual tradition30 does not automatically confer greater probability on all of its distinctive readings. The
case has to be made for each of its variants. Given that we don’t have a textual history for the second
century after the autograph for the Alexandrian and Byzantine readings under investigation, hypotheses
explaining the readings for and against them being original can be put forward.

The earliest use of the phrase monogenh.j qeo.j is by the Gnostic Valentinus (c. 170 CE, Irenaeus, Against
Heresies), but this doesn’t mean that it is a Gnostic invention; he could be alluding to and misusing a faithful
copy of the autograph.31 If the phrase is not original, we shall see below that its origins are lost in the 2c
CE, but if it is original, we have no historical puzzle to explain. On the other hand, if monogenh.j qeo.j is
original, we have then to explain the origins of monogenh.j ui`o,j; historical guesswork, one way or another, is
unavoidable. So, we might say for example that, during the 4c Arian controversy qeo.j became troublesome
and manuscripts were changed to ui`o,j.32

26 Hort, Two Dissertations, 8; Abbot, On the Authorship of the Fourth Gospel and Other Critical Essays, 269.
27 See Abbot’s summary of the evidence pro and con in On the Authorship of the Fourth Gospel and Other Critical
Essays, 268-269.
28 For example, older commentaries favour the majority reading: E. C. Hoskyns and F. N. Davey, The Fourth

Gospel (2nd ed (rev).; London: Faber & Faber, 1957), 153-154; R. H. Lightfoot, St. John’s Gospel: A Commentary
(Oxford: Oxford University Press, 1956), 90; C. K. Barrett, The Gospel According to St. John (London: SPCK,
1955), 141. In favour of the Alexandrian reading, are the more recent commentaries: B. Lindars, The Gospel
of John (London: Oliphants, 1972), 98-99; G. R. Beasley-Murray, John (WBC 36; Waco: Word, 1991), 2-3; J.
Marsh, Saint John, (London: Pelican, 1983), 112.
29 For a discussion of the issues over the Majority Text versus the Eclectic Text see the essays pro and con

in D. A. Black, ed., Rethinking New Testament Textual Criticism (Grand Rapids: Baker House Academic, 2002).
30 The Alexandrian tradition is regarded as superior and this common opinion is often repeated as if it

settles the question of any textual variant, but we should weigh each case on the evidence.
31 Comfort, New Testament Text and Translation Commentary, 256. Mastin, “A Neglected Feature of the

Christology of the Fourth Gospel”, 40.

32 This is put forward by Mastin, “A Neglected Feature of the Christology of the Fourth Gospel”, 39.
Overall, the historical hypotheses available to both variants are transcriptional error, theological corruption
or benign harmonization. There is obviously some similarity in the uncial letters for qeo.j and ui`oj, . The
shortened forms of the two words (ΦΣ and ΥΣ) have a common letter and an accidental error in
transcription (by a supporting scribe) could explain the difference among the uncial manuscript traditions
if ΥΣ was being used as a nomina sacra at the time (if not, then accidental error is harder to envisage).33
However, transcriptional error isn’t sufficient to explain the textual history here because the definite article
is added or deleted for the two main variants and this has to be explained as well. A transcriptional error
hypothesis doesn’t work for the article and a later supplementary correction hypothesis needs to be posited
either way depending on which word was transcribed in error.

Neither reading is more difficult in terms of syntax or semantics which might have given rise to scribal
emendation. In terms of Johannine usage, the two words in monogenh.j ui`o,j are found closely coupled
elsewhere in John, and so monogenh.j qeo.j would appear more difficult to explain as a scribal change, whereas
a scribe might have crafted monogenh.j ui`o,j for reasons of internal harmonization or wider consistency with
other texts; this is the principle of lectio difficilior.34 On the other hand, it could be argued that monogenh.j qeo.j
is not that difficult at all: it harmonizes the prologue by bringing qeo.j from v. 1 into v. 18 in an inclusio.
Indeed, the absence of ui`o,j in v. 14 rather goes against the ui`o,j harmonization hypothesis for v. 18 and
might well have allowed or even prompted the change to (or inclusion of) qeo.j in v. 18. Ironically, the more
successfully a commentator shows that qeo.j is part of the design of the prologue, the more difficult ui`o,j
becomes and the principle of lectio difficilior then applies to ui`o,j. Finally, it is worth pointing out that what is
difficult to a modern textual critic might not have been so to a 2c scribe: the Church Fathers are perfectly
familiar with using the phrase monogenh.j qeo.j in their theology at this time. Such familiarity might lead scribes
to introduce qeo.j for explanatory reasons.35

Another counterpoint to the above lectio difficilior argument against ui`o,j is to claim monogenh.j qeo.j is an
‘orthodox’ corruption of Scripture to affirm the deity of Christ, whereas monogenh.j ui`oj, is consistent with
Johannine usage.36 McReynolds states that for any scribe to have intentionally changed ui`o,j to qeo.j “defies
imagination”.37 This rather depends on how imaginative the scholar is in coming up with scenarios vis-à-
vis any change to a Gospel for, as we noted above, the more integrated qeo.j becomes for us, the easier it is
to suppose a scribe imagined the same integration preferable for qeo.j.

In terms of which reading more easily gives rise to another, we could try another tack and say that the minor
variant o` monogenh.j (which has its best witness38 in the Diatessaroncur) explains both o` monogenh.j ui`o,j and
o` monogenh.j qeo.j as competing natural expansions. This in turn would allow us to then explain the non-
articular monogenh.j qeo.j because o` monogenh.j qeo.j would have been construed adjectivally as ‘the only
begotten God’ and therefore to be less Johannine, whereas monogenh.j qeo.j could be construed appositionally

33 So, the original scribe of Sinaiticus missed out some words in John 1:18 and these were later added by a
corrector. Does this cast doubt on his transcription of the nomina sacra as well?
34 The rules of textual criticism are summarized in textbooks, for instance, K. Aland and B. Aland, The Text

of the New Testament (Rev ed.; Grand Rapids: Eerdmans, 1989), 280-282; but their warning is that they “must
not be taken too mechanically” (281). So, for example, we should distinguish different types of examples
where this principle might be applied: copyist errors, syntactical matters, semantic issues and theological
points.
35 Abbot, On the Authorship of the Fourth Gospel and Other Critical Essays, 270.
36 B. D. Ehrman, The Orthodox Corruption of Scripture (Oxford: Oxford University Press, 1993), 78-82. This is

discussed in various essays in Wallace, Revisiting the Corruption of the New Testament. Ehrman and Wallace have
debated the issue, and this is recorded in B. D. Ehrman and D. B. Wallace, “The Textual Reliability of the
New Testament” in The Reliability of the New Testament (ed., R. B. Stewart; Minneapolis: Fortress Press, 2011),
13-60.
37 McReynolds, “John 1:18 in Textual Variation and Translation”, 114.
38 Hort notes that the variant is indicated in the Old Latin St. Gatien’s manuscript (8c-9c CE) and in some

Greek and Latin Fathers (who also use the longer expressions), Two Dissertations, 11.
and to be more in keeping with Johannine usage in John 1:1, 14.39 However, B. A. Mastin argues (correctly),
that the reading “must be discounted because it occurs in no Greek manuscript of the Fourth Gospel”.40

As for the two major variants, monogenh.j qeo.j and o` monogenh.j ui`o,j, neither naturally give rise to the other
directly. The earliest Church Fathers knew both readings, so any early manuscripts which favour ‘only
begotten God’ are not decisive. Nothing can turn on readings from earlier manuscripts when the early
Church Fathers are so split.

If the two main textual variants are not explicable in terms of benign changes to 2c CE manuscripts,
theologically motivated change is the next hypothesis. In the context of Alexandria (or even Rome) in the
2c CE is qeo.j a likely or possible theologically motivated corruption? This is the claim of B. D. Ehrman in
his book The Orthodox Corruption of Scripture. The variant was created and supported by Gnostic and orthodox
writers to affirm a High Christology against Adoptionists (albeit a different High Christology in each case). 41
Ehrman is ‘correcting’ J. W. Burgon’s thesis which was that the corruption originated with Valentinus.42
Gnostics and the orthodox alike would have confronted the Adoptionists by saying John originally wrote
monogenh.j qeo.j and, look, here is a copy of his Gospel. The problem with the hypothesis is twofold: first, a
lack of polemical evidence, i.e., John 1:18 might be cited with monogenh.j qeo.j in a Gnostic text or by an
orthodox writer but it’s not cited polemically against Adoptionists.43 P. Perkins comments, for instance, that
“the two terms [qeo.j, ui`o,j] would be equivalent in a gnostic reading of the text.”44 Secondly, as Hort notes,45
monogenh.j qeo.j is found in Fathers who are located elsewhere other than Alexandria and this places a
question mark over the hypothesis that monogenh.j qeo.j was an Alexandrian theological corruption that
spread.

On the other hand, if monogenh.j qeo.j is actually the original, why would scribes anywhere change it to o`
monogenh.j ui`oj, ? The answer is that just as we can integrate qeo.j into the prologue so too we can integrate
ui`o,j. A scribe may have had theological and literary reasons to make the change over and above simple
harmonization and for no particular polemical reason.

In sum: We have then some historical hypotheses for the two major variants: theological corruption or
harmonization. However, both readings are understandable in terms of John’s prologue and in keeping
with later Christology, orthodox and heterodox (Gnostic, Arian). The competing hypotheses are
unsatisfactory explaining the initial change whichever we choose. We know which reading proved stronger.
Hort observes an irony,

It is only in the fourth century that the phrase pervades the greater part of the extant literature:
and the cause surely is that, though monogenh.j qeo.j as a reading was being swept out of biblical
MSS. by the same accidental agencies of transcription which removed hosts of Ante-Nicene
readings of no doctrinal moment, as a formula it - had at last established itself in widely known
Creeds.46

39 Comfort, New Testament Text and Translation Commentary, 256.

40 Mastin, “A Neglected Feature of the Christology of the Fourth Gospel”, 38; he cites scholars who have
supported this reading.
41 Ehrman, The Orthodox Corruption of Scripture, 82.
42 J. W. Burgon, The Causes of the Corruption of the Traditional Text of the Holy Gospels (London: Geo. Bell &

Sons, 1896), 215-218; The Traditional Text of the Holy Gospels (London: Geo. Bell & Sons, 1896), 113-114.
43 Hort, Two Dissertations, 9-10 (10): “The single fact that monogenh.j qeo.j was put to polemical use by hardly

any of those writers of the fourth century who possessed it, either as a reading or as a phrase, shews how
unlikely it is that the writers of our earliest extant MSS. were mastered by any such dogmatic impulse in its
favour as would overpower the standing habits of their craft.” Hort’s point cuts both ways: is the lack of
polemical use indicative of doubt about authenticity or awareness of its contested status?
44 P. Perkins, Gnosticism and the New Testament, (Minneapolis: Fortress Press, 1993), 221, n. 6.
45 Hort, Two Dissertations, 25, “It would not be easy to trace these scattered texts to Alexandria”.
46 Hort, Two Dissertations, 26.
In the absence of 2c historical data, no decision can be made between the two main variants. We have a
stalemate; scholars may be confident in their advocacy of one or the other variant, but the evidence
underdetermines any conclusion.47 We can only break this impasse by supposing that both variants were
promoted as original in the 2c but that monogenh.j ui`o,j eventually won the day by the 5c. Can internal
evidence help decide the autograph? Here the question is: What do we think John would have written?

Internal Evidence48
The adjectival interpretation of monogenh.j qeo.j is ‘only begotten God’ and it rests on the normal rule that
adjectives that agree in number, gender and case modify their nouns. The appositional interpretation is
‘only begotten one (or son), God’ and this is said to reflect John 1:1, 14. We will evaluate both
understandings of monogenh.j qeo.j along with o` monogenh.j ui`o,j with regard to the internal dynamics of
Johannine thought.

The complicating factor in this exercise is that commentators have different understandings of the prologue
and whichever expression is chosen for v. 18 is embedded into their presentation of the whole literary unit.
If we see problems in their interpretation of the prologue, it is likely we will see problems with the
expression they have chosen for v. 18. However, it is beyond our scope to consider interpretations of the
whole prologue just to get at a decision on which expression is original - monogenh.j qeo.j or o` monogenh.j ui`o,j.

Internal considerations from both sides brought to bear include (in no particular order),

(1) John 3:16, 18 and 1 John 4:9 have ‘only begotten son’ (albeit with different Greek forms). So, o` monogenh.j
ui`o,j is in keeping with Johannine phraseology.49 This point can be made stronger. John 3:16, 18 are about
“believing in him” and “in the name of the only begotten son of God”. The element ‘belief in his name’
(John 1:12) is about becoming a son of God. The contrast is naturally with ‘only begotten from the Father’
(John 1:14) and ‘only begotten son’ (John 1:18); believers are not only-begotten sons. Having ‘God’ in
apposition with ‘only begotten’ in v. 18 doesn’t fit with John 1:12 and its contrast with John 1:14, and so it
is unlikely to be original and ‘of the Spirit’. The claim that scribes have harmonized John 1:18 to bring it
into line with Johannine usage elsewhere doesn’t address this point because it is about whether qeo.j could
be original as part of John’s theology of belief in the name of the son.

(2) V. 18 and monogenh.j qeo.j is an appropriate inclusio with vv. 1 and 14: ‘was God’ in v. 1 matches ‘God’ in
v. 18; ‘explaining’ the Father is a natural function of ‘the Word’; further, being ‘with/towards’ God could
have the same import as the figure, ‘in the bosom of the Father’. This literary argument overstates its case
and is more the construct of the literary critic. If John was recapping v. 1 in v. 18, he would have had ‘the
only begotten one (or son), the Word’. Further, facing towards God is about being an intermediary which
is not a plausible sense for the figure of ‘being in the Father’s bosom’. Nevertheless, this argument shows
that we cannot claim it is completely contrary to Johannine theology to have ‘the only begotten one (or
son), God’ in v. 18, unless we also jettison ‘and was God’ in v. 1.

(3) John 1:14 has ‘from the Father’ which implies a begettal overtone for monogenh.j in v. 14 and therefore v.
18. We can’t go with ideas of ‘only one’ or ‘unique’. So, it is more likely that monogenh.j ui`oj, is original
because ui`o,j is specifying what is naturally ‘from the Father’. On the other hand, v. 14 did not use ui`o,j, so
why couldn’t it be the case that v. 18 has ‘God’ from v. 1 in apposition, giving ‘only begotten one (or son),
God’? We could say that it is precisely the absence of ui`o,j from v. 14 that requires its inclusion in v. 18.

(4) John only relates attributive uses of monogenh.j + ui`o,j to qeo.j (John 3:16, 18 and 1 John 4:9). With path,r
in v. 18, Johannine usage suggests we would not have an attributive use of monogenh.j + ui`oj, . This argument

47 Mastin, “A Neglected Feature of the Christology of the Fourth Gospel”, 41, confidently concludes, “Thus
it is probable that the incarnate logos is described as monogenh.j qeo.j in this verse.”
48 Internal evidence is a series of checksums on the integrity of the text.
49 Lightfoot, St. John’s Gospel: A Commentary, 90; R. Bultmann, The Gospel of John, (Oxford: Basil Blackwell,

1971), 82; Barrett, The Gospel According to St. John, 141; Hoskyns and Davey, The Fourth Gospel, 154.
fails because it is evident that the attributive use of monogenh.j + ui`o,j in v. 18 relates to the qeo.j in ‘No man
has seen God’; the path,r clause is an additional relation.50

(5) The variant monogenh.j qeo.j introduces an awkwardness, “No man has seen God at any time…the only
begotten one (or son), God…he has explained him’’. An only begotten son being seen by men and
explaining God is not difficult, but how does God declare God? It might be countered that this
awkwardness exists with v. 1 if it exists for v. 18, since the Word was both with/towards God and was
God. In both cases, the difficulty is removed by understanding the sense in which the Word was God. But
this reply concedes the point: it inadvertently suggests that v. 18 should be using the title ‘the Word’ for
“has explained him” rather than ‘God’.51

(6) Vv. 14-18 is a concluding paragraph that relates vv. 1-13 to the experience of the disciples, describing
in more concrete terms the ministry of Jesus and John the Baptist’s witness. The name ‘Jesus Christ’ and
the Law are mentioned for the first time (v. 17), and the giving of power (v. 12) is mirrored in the receiving
of fulness (v. 16). A first use of ui`o,j in v. 18 better fits this pattern of the paragraph. It is only in vv. 17-18
we get given the identity of the Word and the true Light. If v. 18 had qeo.j it wouldn’t be giving us anything
more by way of identification than v. 1.52

(7) Hebrews 11:17 uses monogenh.j substantively of Isaac in a typology of death and resurrection (Heb 11:19).
This is support for reading substantive uses of monogenh.j elsewhere as indicative of an Isaac typology if
there are other reasons in the context for such a reading. Of the 9x uses in the NT, only John 1:14 and Heb
11:17 do not have an accompanying ‘son/daughter/child’ word (Luke 9:38 has ‘son’ in the previous clause).
There are indications in the prologue that John 1:14, 18 is using an Isaac typology. This argument is not
about Johannine usage, but a deeper reason for ui`o,j based on Isaac typology being a father/son typology.53
However, the counterargument is that this Isaac typology only requires a coupling of path,r/monogenh.j and
not ui`o,j.54 We might concede this point, but equally it goes to show that qeo,j is not original because such a
juxtaposition is not part of an Isaac typology.

(8) In v. 14, ui`o,j is not used because monogenh.j on its own reflects the emphasis of ‘thine only’ in Gen 22:2,
12 and it complements ‘from the Father’ in explaining how God provided the lamb (Gen 22:8).55 In v. 18, no
such emphasis is needed because Abraham (‘the Father’) has now received Isaac, his son (Gen 22:13), back
from the dead, a fact typed by the son now being ‘in the bosom of the Father’.56 Hence, where we have
Isaac typology, we can expect ui`o,j and so it is original to v. 18 (i.e. ui`o,j completes both vv. 14-18 and the
prologue).57

On the other hand, it could be said that whereas John elsewhere couples ‘Father’ and ‘Son’, he doesn’t
couple ‘only begotten son’ and ‘Father’, he couples ‘only begotten son’ and ‘God’.58 However, even if the
‘bosom’ figure did not imply a filial relationship, it does imply a Father taking care of his only one,59 and this
notion manifestly goes better with ui`o,j.

50 Contra D. A. Fennema, “John 1:18: ‘God the Only Son’”, NTS 31 (1985): 124-135 (127).
51 Contra Fennema, “John 1.18: ‘God the only Son’”, 127-128.
52 Contra Fennema, “John 1.18: ‘God the only Son’”, 127-128.
53 The argument would also support the minor variant o` monogenh.j as original.
54 Fennema, “John 1.18: ‘God the only Son’”, 126, makes this point although not in the context of

developing an Isaac typology.

55 Josephus, Ant. 1.222 interprets the Hebrew and describes Isaac as Abraham’s monogenh.j.
56 Barrett, The Gospel According to St. John, 169; Bultmann, The Gospel of John, 81, make a similar point without

the Isaac typology.

57 This father/son Isaac typology lies in the background of Jesus’ discourse in John 5.
58 Fennema, “John 1.18: ‘God the only Son’”, 126. Fennema deploys a standard analytical argument of

shifting the unit of analysis. We have argued that ‘Father’ naturally couples with ‘son’. Fennema counter-
argues that this doesn’t necessarily show that ‘Father’ naturally couples with ‘only begotten son’. However,
Fennema fails to see the Isaac typology in monogenh.j which is a Father/son typology.
59 It is worth noting here regarding Genesis that Isaac was Abraham’s only son because Ishmael had been

sent away and not because he was the child of promise.

The Isaac typology shows why a translation such as ‘the only son’ (v. 14, RSV) is wrong for this substantive
use of monogenh.j; it takes away the explanation of how God provided the lamb (i.e., through begettal). For
the same reason, ‘the only son’ would be wrong for v. 18.60

In other texts (Pseudepigrapha, LXX, NT, etc.) monogenh.j is often given an overtone of begettal, when it
modifies a daughter/son/child word. In more modern translations this is not the case. Our argument is
that John sets up his ‘begettal’ use in the prologue because he will be teaching about Jesus’ origin
throughout the Gospel.

(9) The juxtaposition, ‘the only begotten one (or son), God’ is too developed theologically for John.61 The
status of ‘only begotten’ in v. 14 is about the humanity of the Son. In v. 1, ‘God’ is predicated of ‘the Word’
and the status of ‘only begotten’ is absent. This coupling in v. 1 is not out of step with Second Temple
thinking about ‘the Word’ (albeit this is applied to Jesus). To juxtapose ‘only begotten one (or son)’ and
‘God’ in v. 18 is too advanced theologically for the 1c CE (the juxtaposition should be ‘with ‘the Word’);
monogenh.j qeo.j is 2c CE doctrine62 and a first step towards Nicaea.

(10) The reason for ‘was God’ in v. 1 is to identify Jesus as the one in whom God dwelt. Revelation in v.
18 (‘told/declared’) is the function of ‘the Son’ and not ‘God’. The contrast in John 1:17 between Moses
and Jesus is continued in v. 18 in several ways. First, ‘No man has seen God’ echoes the Sinai declaration
that ‘no man can see me and live’ (Exod 33:20). Moses did not see God; he saw only the back of God
(Exod 33:20); secondly, he is the one who had proclaimed-declared. This is the role of being a mouth for God
and the allusion is to Exod 4:12-16, where Moses is a mouth for God and Aaron is a mouth for Moses.
This is not a pertinent function for the mention of ‘God’ and so ui`o,j is original.63 Jesus’ own juxtaposition
with regard to Moses was to position a servant in contrast to a son (in the Parable of the Husbandmen; cf.
John 8:35; Heb 3:5-6).

Our conclusion is that the balance of argument surrounding the internal considerations (1)-(10) supports
the Byzantine text as original. This text better fits with the cultural context of Jewish monotheism; it fits
John’s language and style in the Gospel; it preserves the use of Isaac typology; and it fits with the prologue’s
contrast with Moses.

60 Contra Fennema, “John 1.18: ‘God the only Son’”, 127.

61 R. E. Brown, The Gospel According to John I-XII (Anchor Bible 29; New York: Doubleday, 1966), 17.
62 Accordingly, our favoured hypothesis for the origin of this variant is theological development in Egypt,

possibly Gnostic at first.

63 Contra Mastin, “A Neglected Feature of the Christology of the Fourth Gospel”, 41, “…by using the

phrase monogenh.j qeo.j, he is able to insist that only God can reveal God while at the same time he
distinguishes the Revealer from the Father.”