Q: When a Bible verse has variants, how do we estimate which is more likely to be the original? A: First, let’s look at four types of errors, then the criteria many people use, and finally the relative weighting of the criteria. Four types of errors I count about 3,600 variants in the New Testament where we are not certain which the correct should be. That would mean we are certain of 97.3% of every word. Generally these are assumed to be from four sources. Accidental Typos: Many times copyists have simple small spelling errors, or occasionally duplicate a line. Most of these have no pattern and are easy to correct by comparing with the other manuscripts. One exception to this rule is that we are not totally sure how to spell "Iscariot" as in Judas Iscariot. Of all the words we are uncertain of in the New Testament about 1/3 (1,120 words) are single word variants, and 328 words are double word. These fall into both this category and the following two categories. Deliberate Small Fixes: Many times copyists would attempt to fix spelling or small grammatical errors. Translators into another language would try to translate the meaning accurately. Sometimes they would get it wrong. Smoothing: Sometimes the Greek grammar or phrasing might be rough, or the word choice might not be the best (in the copyist’s opinion.) the copyist would try to correct it to better bring out the meaning. Many manuscripts in different verses says "Jesus Christ" while others say "Christ Jesus". Large changes: There are 38 places with 10 or more word changes. One of the largest variants is the ending of Mark. The Vaticanus manuscript (325-350 A.D.) does not have the ending of Mark, but has a blank space there. Vaticanus does not have any other blank spaces like this in the entire rest of the manuscript. Likewise, Sinaiticus (340-350 A.D.) has a blank space there. However, according to www.LogosResourcePages.org/uncials.htm, you can see where this text was in Sinaiticus, but it was pumiced out (erased). The following other verses are absent in Sinaiticus but present in Vaticanus: Mt 24:35; Luke 10:32; 17:35; John 9:38; 16:15; 21:25; and I Corinthians 13:2. The following verses are absent in Vaticanus and present in Sinaiticus: Mt 12:47. Lk 23:17. Yet B omits Luke 23:34. Vaticanus and Sinaiticus both do not have: Mt 17:21, 18:11, 23:17; Mk 7:16, 9:44, 46, 11:26, 15:28; Lk 9:55-56, 17:36, 23:17, and Jn 5:4 See http://www.ecclesia.org/truth/vaticanus.html for more info. Criteria for "strength" of a variant These criteria are not just good for Bible manuscripts, but for any ancient manuscript. Earlier manuscripts: In general a manuscript written closer to the event would be considered as having a probability of being more reliable than a manuscript written much later. Number of manuscripts: In general, a variant appearing in a large number of manuscripts would have more weight than a variant appearing in just a few manuscripts, given that the manuscripts were of similar age. Heretical readings: Without necessarily intending to, heretics have made a valuable contributions to the church and Bible manuscripts. When a heretical reference says a manuscript is a certain way, and they have no incentive for bias in that direction, or even a possible incentive for bias in the opposite direction, that is strong evidence that the heretic was not aware of a Bible manuscript favoring his view. Difficulty of reading (Lectio difficilior potior): While it is very improbable that a copyist would try to make the text more difficult to read, it is easy to believe that if a scribe did not copy the text the same, he was trying to make it easier to read. Hence this principle assumes that the more difficult reading is closer to the unedited original. For example, in 1 Cor 14:38, one variant is "ignored" (Sinaiticus, Alexandrinus, others) vs. "ignorant" (p46 and others). It is surmised that "ignored" is the more difficult and offensive readings, so many would think that "ignored" is probably the correct meaning, even though p46 is earlier than the other manuscripts. For a second example, in Mt 16:2-3 Jesus compares knowing the times to knowing when the sky is red it is going to rain. This weather phenomenon is true in Palestine but not in Egypt. Manuscripts from Egypt, such as Vaticanus and Sinaiticus, do not have this verse. It is easy to hypothesize that since this reading was more difficult for Egyptian readers, Egyptian copyists would have made this large (30-word) change. Relative weighting of the criteria Scholars disagree on the importance of these criteria. Here are some "guiding principles" some use, along with a rebuttal. Manuscript family: Many scholars have decided that the Byzantine family, with it large numbers of manuscripts, though later, is in general more accurate. The NKJV translators seem to favor this. Many other scholars, such as Westcott and Hort, think the earlier, but fewer Alexandrian manuscripts (such as Vaticanus and Sinaiticus) are the more accurate. Others, such as the NIV translators, are "eclectic" and claim not to exclusively favor either family. In general, when there are variants, more often than not the Alexandrian family will have the word or phrase be absent (or taken out, depending on your perspective) and the Byzantine family will have the word or phrase by present (or added, depending on your perspective.) In one manuscript, Bezae Cantabrigiensis, it will often have both variants together! Earlier manuscripts: It is not true that everyone thinks the earlier manuscript is always best. Later manuscripts were often done by professional scribes (we can tell by the handwriting) and probably done with more research on variant readings. Some of the earliest manuscripts might have been quickly done by ordinary Christians, with many typos. Relying on the benefits/mistakes of previous scholars: Origen of Alexandria did a monumental amount of work to collect the variants of different verses which would help determine which was most accurate. Unfortunately, his work, the hexaemeron, with all its symbols and notes, has been lost. However, we are sure the results of his work were known to the Alexandrian church and likely used to help create more accurate manuscripts. If you were to believe that Origen’s work and decisions were probably for the most part good, you would tend to trust the Alexandrian manuscript family more. If you think he might have made lots of mistakes, and would want to look at manuscripts without this bias, you might trust Byzantine, Latin, Syriac, or other manuscripts more. Difficulty of reading: Some people use the criteria of difficulty of reading. This assumes that later copyists would fix things and make things smoother to read, so the more difficult way of reading would be the earliest, when all other evidence is the same. Which variant type and which manuscript: If a manuscript is very similar to other manuscripts in a particular manuscript family, the variants in that manuscript that differ from other manuscripts in the same family would be considered more significant by some people. Validation: One statistical method is to split the data into training and test sets. Get weights for the training, and then use them to predict the test data, as a measure of how good the weighting is. Applying this to manuscripts, … Caution: Sometimes you get a different answer when you focus on a different question. If you ask, "which manuscript is more probably the original" you might get a "winner" which is 60% certain. But if you ask, "Can we determine which variant is the original and if so which one, then you might say it is inconclusive unless it was at least 95% certain. Q: What is the evidence (pro and con) that p4 is from the same manuscript as p64 and p67? A: This is according to Philip Comfort in The Complete Text of the Earliest New Testament Manuscripts p.38-39. Same Manuscript Evidence S1. All letters but three that could be compared in all three manuscripts are identical. Four letters could not be compared. Sigma, epsilon, and alpha are quite similar in all manuscripts but not quite identical. In some case of p4 the underside is less fully curved than others in p4. In p64 and p67 all these letters are always less fully curved. S2. The pages are very similar; all are the same width, same number of columns (2), about 36 lines per column, and 15-17 letters per line. S3. Similar punctuation and paragraphing is common to the three. S4. P64 and p4 were purchased in the same city. S5. "Sacred abbreviations" were used in many late-second and early-third century manuscripts for God, Lord, Jesus, Christ, and Spirit. These three did also, but they also had "sacred abbreviations" for Son, Father, man, cross, and crucify. S6. Accompanying p4 is another small fragment, written by a later writer, that is part of a title page "Gospel According to Matthew." Comfort p.43-44 says it was stylish for scribes to add a title from 175-200 A.D. This does not address whether p4 is the same manuscript as p64/p67, but it attests to an early date for p4. S7. The reputation of Colin Roberts (who dated the John Ryland’s manuscript), Joseph van Haelst, C.H. Roberts, and Philip Comfort. Different Manuscripts Evidence D1. P4 has finer, thinner pen strokes than p64/p67. A different stylus and/or ink in the hands of the same scribe could produce this difference. D2. P64/p67 is on average lighter than p4, though parts of p4 are the same lightness as p64/p67. Difference in preservation conditions would produce this. Regardless, p4 is not uniform and parts match p64/p67. D3. Aland’s reputation. In 1965 Kurt Aland suggested p4 was the same codex as p64/p67, but afterwards he listed them as separate. Philip Comfort p.36 says he is unaware of any reason Aland gave for the change. Q: Just how many New Testament manuscripts (including fragments) exist today? A: Here are the counts from various scholars of just the Greek manuscripts (complete manuscripts, small fragments, and everything in between). One reason the counts differ is that a number of new manuscripts have been discovered since 1975. Also some people call multiple manuscripts what others call one manuscript with multiple pieces. Category Ralph Earle Aland et al. The Greek New Testament 3rd edition (1975) Bruce Metzger p.54 (1976) Manuscripts of the Greek Bible p.54 (1981) A General Introduction to the Bible p.387 (1986) Aland et al. The Greek New Testament 4th edition (1998) Papyrii (p1-p88)   76+ 88 88 88 95 (=97-2) Uncials (non-Lectionary) 270 250+ 274 274 274 286 (=300-16+2 Miniscules (non-Lectionary) 2,400 2,768+ 2,795 2,795 2,745 2,818 Lectionary manuscripts (both uncial and miniscule) not mentioned 1,761+ 2,209   2,147 1,977+ Uncial (Lectionaries)       245     (Miniscule) Lectionaries       1,964     Ostraca not mentioned not mentioned 20   not mentioned not mentioned Total 2,670+ 4,855+ 5,386 5,366 5,254+ 5,176+ The manuscripts in other languages, including 89+ Italic (Old Latin), four Vulgate families, 8 Syriac families, 8 Coptic families, Aramaic, Armenian, 2 families of Georgian, 3 families of Ethiopic, Slavonic, 5 Gothic manuscripts, Arabic, and Persian bring the total up to over 11,000 manuscripts. The Complete Text of the Earliest New Testament Manuscripts (copyright 1999) p.627 goes up to p104. Some would say there are really four less papyrii, as p45 + p46 + p47 are apparently written by the same scribe. p64 + p67 apparently are also from the same scribe. p20 + p27 possibly are by the same scribe. For comparison purposes, Manuscripts of the Greek Bible p.5 says that scholars have catalogued 55,000 Greek manuscripts on all topics. Thus almost 10% of all Greek manuscripts are Bible manuscripts. Homer’s Illiad has the second most number of copies with 643-650 manuscripts; which is about 1% of all Greek manuscripts. By the way, there are conflicting dates about when Homer lived. The ancient Greek historian Aristarchus says 1044 B.C., authorities in Philostratus says 1159 B.C., and pseudo-Herodotus The Life of Homer puts it as 1102 B.C. In contrast to the historian Theopompus said Homer lived as late as 685 B.C.. Herodotus says it was not earlier than about 730 B.C.. The Exodus was about 1447 B.C. See The Encyclopaedia Britannica vol.11 (1956) p.689under Homer. It says the Illiad and Odyssey were performed in Athens every four years and there are only 10 papyrii of Homer prior to 150 B.C. Q: How many Bible fragments and manuscripts are preserved, century by century? A: Sometimes scholars disagree about the dates, but here are my numbers, based primarily on Comfort and Barrett, and secondarily on Aland et al. The Other column is artificially small because it is part manuscripts and part languages which contain many manuscripts. Century Greek Other 2nd century 11 0 2nd/3rd century 4 0 3rd century 30 5+ 3rd/4th century 8 0 4th century 24 3+ 4th/5th century 16 5 5th century 34 13+ 5th/6th century 14 2+ 6th century 48 4+ 6th/7th century 11 2+ 7th century 32 6+ 7th/8th century 2 0 8th century on Many Many The Greek manuscripts were not used as much when people started using other languages more, such as Coptic, Latin, Syriac, etc. The large number of Greek Byzantine Lectionaries, which primarily come later, are not included here. Taking the same data, except splitting each manuscript that might span centuries as half in each century, gives the following numbers. Century Greek Other 2nd century 13 0 3rd century 36 5+ 4th century 36 5.5+ 5th century 49 19.5+ 6th century 60.5 1+ 7th century 38.5 7+ 8th century on Many Many Q: For the NT, how many multi-lingual manuscripts exist today? A: The oldest preserved bilingual manuscript is Bezae Cantabrigiensis. Here is a list of bilingual manuscripts. Greek and Latin (20 manuscripts) D (Cantabrigiensis Bezae), D (Claromantus), E, F, G, others Greek and Arabic (16 manuscripts) (0136, 0137, 211, 609, l6, l225, l311, l762, l804, l937, l1023, l1343, l1344, l1746, l1733, l1774,) Greek and Armenian (1 manuscript) (256) Greek and Coptic (52 manuscripts) (p6, p41, p42, p62, T, 070, 086, 0100, 0110, 0113, 0124, 0125, 0129, 0139,0164, ethers) Greek and Slavonic (3 manuscripts) (525, 2136, 2137) Greek and Turkish (1 manuscript) (manuscript 1325) There are three known trilingual manuscripts: Greek and Coptic and Arabic (2 manuscripts) (l1993, l1605) Greek, Latin, and Arabic (manuscript 460) See Greek Manuscripts of the Bible p.56 for more info. Q: Briefly, when were these manuscripts written? A: The Lukan manuscript in Paris (p4) containing parts of Luke 1 to 6 is dated by Philip Comfort to around 100 A.D. Of course, the Gospel of Luke was written prior to Acts. More on this is in the book by Thiede, Carsten P. and Matthew d’Ancona, Eyewitness to Jesus: Amazing New Manuscript Evidence About the Origin of the Gospels (NY Doubleday 1996 206 pp.). However, while Thiede and d’Ancona date this as "not much later than 68 A.D., Philip Comfort is more cautious, dating this at 100 A.D. Aland et al’s. The Greek Testament Fourth Revised Edition (1998) dated this as fourth century. The earliest fragment of John, called the John Rylands Papyrus, written 117-138 A.D. The next was the Chester Beatty II papyrii, 150-200 A.D. (or 200-225 A.D.) The next was the Bodmer II papyrii, written about 125-175 A.D. (formerly thought to be 150-200 A.D.) These two manuscripts alone contain 40% of the New Testament according to both my own study and The Qur’an and the Bible p.148 There were a total of 4 preserved manuscripts around 200 A.D., 30 more manuscripts prior to 300 A.D., 8 manuscripts around 300 A.D., 28 more manuscripts before 400 A.D., 16 manuscripts around 400 A.D., and 38 more manuscripts prior to 500 A.D. See The Text of the New Testament : An Introduction to the Critical Editions and to the Theory and Practice of Modern Textual Criticism by Aland and Aland p.52 and A General Introduction to the Bible p.387 for charts of all the manuscripts up to 1600 A.D. Jose O'Callaghan found in cave 7 at Qumran fragments of a papyrus dated 50 A.D., that might be fragments of Mark. However, The Expositor’s Bible Commentary volume 8 p.608 says this "has been largely rejected by NT scholars (cf. EBC 1:420-421, n.1). The evidence O’Callaghan presents is far too fragmentary to be reliable." Q: For the NT versus other ancient works, what is the number of manuscript variations? A: Here are Bruce Metzger’s estimates compared with other religious literature. There are over 5,000 Greek manuscripts of the New Testament. Metzger estimates the New Testament has 20,000 lines, an accuracy of 99.5% with only 40 lines (about 400 words) in question. (This is probably on a letter by letter basis.) Homer’s Illiad is the next most reliably preserved document. It has 643-650 manuscripts, and is 95% accurate. It has 15,600 lines, with about 764 lines in doubt. The Hindu Mahabharata has 250,000 lines and is roughly 90% accurate according to Bruce Metzger. (See the Baker Encyclopedia of Apologetics p.532-533 and A General Introduction to the Bible p.475) It has over 26,000 lines have textual corruption. See A General Introduction to the Bible - Revised and Expanded p.474-475 for more info. Q: Compared to the NT, what are some of the dates of other early documents and the number of copies? A: Here are some other ancient works. Aristotle wrote 364-322 B.C. There are only 5 copies, the earliest being 1100 A.D. Julius Caesar 100-44 B.C. 900 A.D. 10 copies. This is the only direct evidence we have that Julius Caesar actually entered Gaul. Demosthenes 4th century B.C. 200 copies Euripides’ Tragedies 330 copies (The Origin of the Bible p.182) Herodotus 480-425 B.C. 900 A.D. 8 copies Homer wrote The Iliad 643-650 copies, more than any others. 5% of the words are in question Jubilees (A Jewish apocryphal book) 14 copies among the Dead Sea Scrolls. Mahabharata (a Hindu scripture) 10% in question Old Testament Manuscripts 235 scrolls and fragments the Dead Sea Scrolls alone. From the Great Isaiah scroll, about 5% of the words are different vs. the Massoretic text. However, most of these are archaic vs. later words and grammar with the same meaning. Pliny the Younger 1st century A.D. 7 copies Suetonius wrote The Twelve Caesars 70-140 A.D. The earliest copy is 950 A.D. Tacitus 100 A.D. 1100 A.D. 20 copies Thucydides History of the Peloponnesian War 460-400 B.C. 900 A.D. 8 copies The Qur’an 765 A.D. 9th century. Aisha said one Sura had 200 verses. After ‘Uthman’s "standardization", today it has 73 verses. Also, part of Sura 9:30 was abrogated. The Bukhari Hadith 6:509 says that when certain people died, parts of the Qur’an known only to them were lost. Other Bukhari Hadiths saying parts of the Qur’an were missing and/or abrogated are 4:57,62, 69,229; 6:510,511. Q: In the NT, what is a summary of the manuscript variations? A: We are about 97.3% certain of each New Testament word. (Other people might have slightly different numbers if they exclude some variations as certainly incorrect.) Here is a book-by-book summary. See the end of the discussion on each book for the details. Book of the New Testament Total verses Total words in Greek Greek words in question Percent accuracy 100 - % accuracy Matthew 1,071 18,111 531 97.1 % 2.9 % Mark (exc.16:9-20) 661/ 678 11,051 401 96.4 % 3.6 % -- Mark 16:9-20 --- --- Additional 166 (-1.5%) + 1.5 % Luke 1,151 19,581 494 97.5 % 2.5 % John (exc.7:53-8:11) 879 15,436 313 98.0 % 2.0 % -- John 7:53-8:11 --- --- Additional 168 (-1.1) % + 1.1 % Acts 1007 18,460 487 97.4 % 2.6 % Romans 433 7,030 200 97.2 % 2.8 % order: Rom 16:25-27 --- --- additional 53 (-0.8%) +0.8% 1 Corinthians 437 6,799 104 98.5 % 1.5 % 2 Corinthians 257 1,495 55 96.3 % 3.7 % Galatians 149 2,233 38 98.3 % 1.7 % Ephesians 155 2,385 51 97.9 % 2.1 % Philippians 104 1,621 37 97.7 % 1.9 % Colossians 95 1,570 33 97.9 % 2.1 % 1 Thessalonians 89 1,477 26 98.2 % 1.8 % 2 Thessalonians 47 826 12 98.5 % 1.5 % 1 Timothy 113 1,592 22 98.6 % 1.4 % 2 Timothy 83 1,336 13 99.0 % 1.0 % Titus 46 657 5 99.2 % 0.8 % Philemon 25 329 6 98.2 % 1.8 % Hebrews 303 4,888 82 98.3 % 1.7 % James 108 1,735 26 98.4 % 1.6 % 1 Peter 105 1,648 63 96.2 % 3.8 % 2 Peter 61 937 35 96.3 % 3.7 % 1 John(excl. 1Jn5:8) 104 2,103 30 98.4 % 1.6 % 2 John 13 245 7 97.1 % 2.9 % 3 John 14 219 3 98.6 % 1.4 % Jude 25 459 7 98.5 % 1.5 % Revelation 404 9,667 127 98.7 % 1.3 % Totals 7,955 133,892 3,600 97.3 % 2.7 % Q: Why does the percentage of variants listed (97.3%), differ from Bruce Metzger’s number of 99.5%? A: Bruce Metzger was a contributor to Aland et al’s Greek Translation of the New Testament. This Greek translation, besides giving manuscript variations, gives an estimate of the certainty of the translation. In the fourth edition p.3, the letters mean: A - "indicates that the text is certain" B - "indicates that the text is almost certain" C - "indicates that the Committee had difficulty in deciding which variant to place in the text." D - "which occurs only rarely, indicates that the Committee had great difficulty in arriving at a decision." Note that in the 3rd edition on p.xii-xiii, the letters have slightly different meanings. A - "virtually certain" B - "some degree of doubt" C - "considerable degree of doubt" D - "very high degree of doubt" You arrive at close to the 97.3% figure by including all categories, and the 99.5% figure by only including the C and D categories. The 99.5% figure does not include, for example, many Greek textual variants that were the primary choices the Biblical scholars who translated the NKJV, including the longer ending of Mark, and the pericope of the adulteress. As for myself, rather than try to say who is right, I simply want to report where trustworthy scholars are not certain or disagree. That is why I included in the 97.3% number all variants except those with extremely obvious conclusions. Q: What are characteristics of the typographical errors in the New Testament? A: The following table was calculated from the possible significant manuscript variations listed at the end of each book. The following table shows differences primarily due to typos, spelling, grammar, and word changes. Note the relatively high number of single word changes.   Section of the New Testament Total words in Greek Total words in question Percent accuracy Places with the number of words in question 1 2 3 4 5 6 7 8 9 10+ Matthew 18,111 531 97.1 % 134 31 16 7 5 2 2 2 1 12 Mark 11,051 567 96.5-1.5% 80 46 19 8 3 2 0 0 6 6 Luke 19,581 494 97.5 % 97 37 19 10 13 4 2 2 2 3 John 15,436 481 98.0-1.1% 109 35 6 4 6 2 1 0 0 4 Acts 18,460 487 97.4 % 141 35 10 9 2 7 4 2 1 7 Paul’s writings 29,350 633 98.0 % 317 52 11 9 4 0 1 2 0 5 Other NT 12,234 254 98.2 % 160 23 15 1 3 0 1 0 0 1 Revelation 9,667 127 98.4 % 90 8 3 0 1 0 1 0 0 0 Totals 133,892 3,600 97.3 % 1120 264 90 48 40 17 12 8 10 38 Conclusion: Over half of the variations are places with only one to three word variations. The following table shows differences primarily due to style and phrasing. It is interesting that the percentage of words subject to typographical errors is very similar in the 1, 2, and 3 word variations. Mark, has slightly more typographical errors, which would tend to confirm the generally held view that Mark was the earliest Gospel written. As for the other errors, 0.25% of the total New Testament variations (334 words) are just in two passages: Mark 16:9-20 and John 7:53-8:11. An additional 0.04% are due to the placement of Romans 16:25-27. This table excludes Mk 16:9-20 and Jn 7:53-8:11 Section of the New Testament Total words in Greek Total words in question Places with 7 | 8-9 |10+ words in question % inaccuracy from 7| 8-9 |10+ word variations Total Words Total % variations 100 - % acc. Matthew 18,111 530 2 | 3 | 12 0.1 | 0.1 | 1.2 % 224 1.4 % Mark 11,051 567 0 | 6 | 6 0.0 | 0.5 | 0.5 % 59 0.9 % Luke 19,581 494 2 | 4 | 3 0.1 | 0.2 | 0.6 % 93 0.3 % John 15,436 481 1 | 0 | 4 0.0 | 0.0 | 0.3 % 51 0.3 % Acts 18,462 487 4 | 3 | 7 0.1 | 0.1 | 0.6 % 135 0.1 % Paul’s writings 29,350 633 1 | 2 | 5 0.0 | 0.0 | 0.3 % 133 0.1 % Other NT 21,901 386 2 | 0 | 1 0.0 | 0.0 | 0.1 % 40 0.1 % Totals 133,892 3,600 12 | 18 | 38 0.2 | 0.1 | 0.6 % 739 0.6 % Romans 14:23 (53 words) is simply a question of order of the verses. Let’s define a "subsection of the New Testament" as a portion of around 10,000 words or more that was written at roughly the same time. If you sort the subsections of the New Testament by percent accuracy, (ignoring the three largest errors: the end of Mark, the story of the adulteress, and the location of 53 words in Romans) you get an interesting result. Subsection Percent Accuracy Date Written (Conservative) Date Written (21st century Liberal) Date of the earliest preserved fragment Mark (not counting the ending) 96.4 % 50/62-70 A.D. 64-70 A.D. 100-150 A.D. Matthew 97.1 % 50-80 A.D. 80-100 A.D. 100-150 A.D. Luke 97.5 % 58-80 A.D. 80-100 A.D. 100 A.D. Acts 97.4 % 50-80 A.D. 80-100 A.D. 100-150 A.D. 1 Clement (97/98 A.D.) refers to verses in Acts Late Paul (Rom, 1, 2 Tim, Tt, Phm) 97.8 % 55-67 A.D. 55-67 A.D. 100-150 A.D. John 97.9 % 50-95 A.D. 80-110 A.D. 117-138 A.D. Other NT (not counting Revelation) 97.9 % 62-98 A.D. 60-110 A.D. 1 Clement (97/98 A.D.) refers to Hebrews Early Paul (1 Cor – 2 Thessalonians) 98.3 % 50-63 A.D. 50-55 A.D. 1 Clement (97/98 A.D.) mentions Paul Revelation 98.7 % 95-110 A.D. 95-110 A.D. 200 A.D. With the exception of "early Paul" letters, the order follows the dates when we think each book was written. Now the differences between some numbers, such as 97.1% to 97.4% are too close to be statistically significant. Thus, this does not answer the question of whether Matthew or Luke was written first. 1 Clement, written in 97/98 A.D. quotes from Hebrews 1:3a,4,5,7,13b; 6:18 as well as alludes to Hebrews 11:17 (with Gen 21:22); Hebrews 10:37 (with Habakkuk 2:3), Hebrews 12:6 (with Proverbs 3:12), and Hebrews 13:17, so we know that at least that part of "Other NT" was written before then. While "Early Paul" was written early, the letters probably did not get wide distribution until the time all four gospels were written. Also, Paul apparently quotes Luke, so the gospel of Luke likely was written before Paul.