Please Support the Bible Translation Work of the Updated American Standard Version (UASV)

$5.00

The Reading Culture of Early Christianity

The province of Textual Criticism is the ascertainment of the true form of a literary work, as originally composed and written down by its author. The science owes its existence to the conditions under which, until comparatively modern days, literary works have been preserved. If the author’s autograph of every book were still in existence, there could be no dispute as to what he had written; or if printing had been practiced from the earliest days of literary composition, we could be sure that every book had been handed down to us in practically unaltered form. For authors of the last four centuries, with few exceptions, we are in the happy condition of being certain that we possess their works, to all intents and purposes, precisely as they wrote them. In several instances, the author’s autograph is still extant; in the rest we have early printed editions, issued under the author’s eye. But when once we go back into the ages before the invention of printing, the conditions are wholly different. Only in the rarest possible cases (the great English chronicler of the thirteenth century, Matthew Paris, is perhaps an example) do we possess the author’s own copy of his work; in all other instances we have only copies made by hand at varying distances of time after the composition of the book in question. It is to this copying by hand that the problems of textual criticism are due. Unfortunately for our knowledge of ancient literature, the frailty of the human hand and eye and mind is such that no copy, except of short passages, can be trusted to be wholly accurate; and since different copyists will make different mistakes, it results that no two copies of an ancient book are quite the same. This would be immaterial, so long as the original autograph was in existence; but when once that has disappeared, the student who would know exactly what an author wrote has to discover it by an examination of later copies, of which the only fact certain a priori is that all will be different and all will be incorrect.

The function of textual criticism, then, is to recover the true form of an author’s text from the various divergent copies that may be in existence. The problems presented to it are of all kinds of complexity. If evidence is forthcoming from a period shortly after the writer’s date, there will have been little time for the text to have been corrupted, and common sense should be able to detect most of the errors that have crept in. If the interval between the composition of the work and the earliest extant specimens be longer, much will depend on the amount of evidence available; for among many copies there is more chance that the truth will have survived in some, especially if the extant copies have no common ancestor much later than the author’s autograph. The line of textual tradition for any given literary work is like a genealogical tree, starting from a single point and spreading out as it descends to the living members of the family. If the distance of time be great, but the extant copies many, then the textual problem will be one of considerable difficulty, and requiring nice taste and discernment, but it will be hopeful, because the materials are plentiful; whereas if the extant copies be few, there is a great likelihood that the truth will, in some places, have been wholly lost, and is only to be recovered by guessing—a process precarious in the extreme, and seldom allowing anyone but the guesser to feel confident in the truth of its results.

Now the textual criticism of the New Testament, as it is the most important branch of the science, so also is it the most complicated. It is the most important branch because it has to do with a book, the importance of which is quite incommensurable with that of any other book in the history of the world; and it is the most complicated because the extant materials are incomparably more plentiful in number, and more varied in kind, than in any other instance. The difference in this respect between it and any other ancient book can be made plain by a few examples. The plays of Aeschylus are preserved in perhaps fifty¹ manuscripts, none of which is complete. Sophocles is represented by about a hundred manuscripts, of which only seven have any appreciable independent value. The Greek Anthology has survived in one solitary copy. The same is the case with a considerable part of Tacitus’ Annals. Of the poems of Catullus there are only three independent manuscripts, all of which were derived from an archetype which was itself written no earlier than the beginning of the fourteenth century. Some of the classical authors, such as Euripides, Cicero, Ovid, and especially Virgil, are, no doubt, in a far more favourable position than those who have just been named. In their cases the extant copies of their works, or of portions of them, may be numbered by hundreds. Yet even these do not approach the number of witnesses for the text of the New Testament. The number of manuscripts of it, or of parts of it, in the original Greek, is over three thousand; and to these have to be added a yet greater number of witnesses of a kind to which the classical authors offer no parallel. It is seldom that ancient translations of the classical authors into other languages exist, and still more seldom that they are of any value for textual purposes; but in the case of the New Testament translations are both numerous and important. It is estimated that there are at least eight thousand copies extant of the Latin Vulgate translation alone; and a thousand would be a moderate estimate for the extant manuscripts of the other early versions, in Syriac, Coptic, Armenian, Ethiopic, Gothic, and the rest. It is therefore probably within the mark to say that there are now in existence twelve thousand manuscript copies of the New Testament, of which no two are precisely alike.

The contrast in this respect between the New Testament and classical authors may be regarded from two points of view. On the one hand, this enormous mass of witnesses gives good ground for supposing that the true text cannot be wholly lost; on the other hand, the task of selecting the true text out of all these many and multifarious authorities is one of extreme difficulty. Merely to examine and record the available evidence is an enormous labour; to estimate its value, to distinguish between manuscript and manuscript, and between version and version, is the hardest problem that has ever been set to textual criticism. In another respect, however, besides number, the manuscripts of the New Testament differ from those of the classical authors, and this time the difference is clear gain. In no other case is the interval of time between the composition of the book and the date of the earliest extant manuscripts so short as in that of the New Testament. The books of the New Testament were written in the latter part of the first century; the earliest extant manuscripts (trifling scraps excepted) are of the fourth century—say, from 250 to 300 years later. This may sound a considerable interval, but it is nothing to that which parts most of the great classical authors from their earliest manuscripts. We believe that we have substantially an accurate text of the seven extant plays of Sophocles; yet the earliest manuscript upon which it is based was written more than 1400 years after the poet’s death. Aeschylus, Aristophanes, and Thucydides are in the same state; while with Euripides the interval is increased to 1600 years. For Plato it may be put at 1300 years, for Demosthenes as low as 1200. The great Latin authors are somewhat better off. Horace is represented by several manuscripts written within 900 years of his death. There is an excellent copy of Terence after an interval of about 700 years, and portions of Livy only about 500 years after his date. For Lucretius, however, we have an interval of nearly 1000 years, for Catullus about 1600. Only Virgil approaches the New Testament in earliness of attestation. He died eight years before the Christian era; and there is at least one nearly complete manuscript which is attributed to the fourth century, besides several small fragments, and two more of the fifth century. Yet even so, his text is not in so favorable a position as that of the New Testament by nearly 100 years.

The task of textual criticism, then, in relation to the New Testament, is to try to extract the actual words written by the apostles and evangelists from the great mass of divergent manuscripts in which their works have been preserved. It is a task at once hopeful and hopeless. Hopeful, because in so great a crowd of manuscripts, reaching back to so early a date as many of them do, the truth must, it would seem, somewhere be on record; hopeless, because the discernment of it requires a superhuman degree of knowledge and judgment, and because means do not exist for conclusively demonstrating it. The actual extent to which the text of the New Testament is open to doubt cannot be precisely stated, but the estimate of Dr. Hort, whose lifetime was devoted to this subject, is commonly accepted as an approximate guide. He says:¹ “The proportion of words virtually accepted on all hands as raised above doubt is very great, not less, on a rough computation, than seven-eighths of the whole. The remaining eighth, therefore, formed in great part by changes of order and other comparative trivialities, constitutes the whole area of criticism. If the principles followed in this edition are sound, this area may be very greatly reduced. Recognising to the full the duty of abstinence from peremptory decision in cases where the evidence leaves the judgment in suspense between two or more readings, we find that, setting aside differences of orthography, the words in our opinion still subject to doubt only make up about one-sixtieth of the whole New Testament. In this second estimate the proportion of comparatively trivial variations is beyond measure larger than in the former; so that the amount of what can in any sense be called substantial variation is but a small fraction of the whole residuary variation, and can hardly form more than a thousandth part of the entire text.” It is further to be remembered that, although some doubt attaches to the record of certain incidents and sayings of great interest and value, yet no doctrine of Christianity rests solely upon a disputed text. The Christian student can approach the subject without misgiving, and may follow whatsoever honest inquiry seems to lead him, without thought of doctrinal consequences. His researches should unquestionably be conducted in a reverent spirit, but he may avail himself, without hesitation or mistrust, of all the resources of secular science.

The methods of textual criticism may be broadly described as two in number—the comparison of documentary evidence, and conjecture. The two methods are mutually complementary. Where documentary evidence is plentiful, conjecture will be scarce; but where the former is wanting, the latter will have to try to take its place to the best of its ability. In the case of the New Testament the documentary evidence is so full that conjecture is almost excluded, and it is with the principles of the interpretation of documentary evidence that we are most concerned here. Some statement of these is necessary, as an introduction to a summary of the evidence itself.

The task of the textual critic is, in brief, to counteract the errors of the copyist; and these errors are many,—some capable of being classified under heads, while some resist classification. In the first place the critic has to correct simple slips of the pen, obvious blunders which have no meaning, and which occasion no more difficulty than similar mistakes in the letters of a contemporary correspondent. If the scribe of the Codex Sinaiticus writes ποισαι for ποιῆσαι, or εκ του καλουντας for ἐκ τοῦ καλοῦντος, there is no difficulty in either seeing or correcting the error. A somewhat less elementary form of blunder arises when the scribe, in place of the word which he should write, writes one which resembles it either in sound or in appearance. Thus, in Sophocles’ Ajax 61 some manuscripts have Φόνου, others πόνου, and the context is such as to make the decision between them not absolutely certain; but whichever is wrong, the error was no doubt due to the similarity of sound. On the other hand in Bacchylides v. 23 the scribe of the only extant manuscript has written Φοιβωι for φόβῳ, an error of eye, not of ear; and here the metre and the sense alike make the error obvious and easy to correct. Another common form of error is due to the fact that in ancient manuscripts accents and breathings were rare, and separation of words almost unknown;¹ which led to trouble when the time came for these aids to intelligence to be introduced. Thus in Sophocles’ Ajax 1056 the earliest MSS. had, at the end of the line, the letters ΕΛΟΙΔΟΡΙ. Now ει and ι are constantly interchanged in manuscripts, and hence ΕΛΟΙΔΟΡΕΙ was probably written as often as ΕΛΟΙΔΟΡΙ. The result is that, in the margin of the best extant MS. of Sophocles, the reading ἐλοιδόρει is given, in place of the correct ἕλοι δορί.

Another form of error, very common in all manuscripts, is that of omission. This may be due to mere unaccountable accident, and then the lost word or words can only be recovered either by comparison with other manuscripts or by sheer guessing. Oftener, however, it arises from the similarity of adjoining words, which led the scribe’s eye to slip from one to the other, and so omit the intervening words. For instance, in John 17:15 the correct text runs οὐκ ἐρωτῶ ἵνα ἄρῃς αὐτοὺς ἐκ τοῦ κόσμου ἀλλʼ ἵνα τηρήσῃς αὐτοὺς ἐκ τοῦ πονηροῦ, but the scribe of the Codex Vaticanus let his eye slip from the first ἐκ τοῦ to the second, and so gives the passage as οὐκ ἐρωτῶ ἵνα ἄρῃς αὐτοὺς ἐκ τοῦ πονηροῦ. Similarly in John 3:20, 21, where the true text runs πᾶς γὰρ ὁ φαῦλα πράσσων μισεῖ τὸ φῶς καὶ οὐκ ἔρχεται πρὸς τὸ φῶς ἵνα μὴ ἐλεγχθῇ τὰ ἔργα αὐτοῦ, ὁ δὲ ποιῶν ἀλήθειαν ἔρχεται πρὸς τὸ φῶς ἵνα φανερωθῇ τὰ ἔργα αὐτοῦ ὅτι ἐν θεῷ ἐστὶν εἰργασμένα, the scribe of the Codex Sinaiticus has made two mistakes from this same cause (technically known as homoioteleuton), omitting καὶ οὐκ ἔρχεται πρὸς τὸ φῶς and ὁ δὲ ποιῶν … αὐτοῦ, the former owing to the double occurrence of τὸ φῶς, the latter owing to the double occurrence of τὰ ἔργα αὐτοῦ. Often the omissions are smaller than these, and cause less trouble, as when a scribe writes KATHN for KATATHN (κατὰ τὴν) or ἔπεμψε for ἐπέπεμψε. But in one form or another the error is a very common one, and has to be borne in mind constantly in the criticism of manuscripts.

Various other classes of error exist and may be briefly mentioned. One that is frequently invoked in the criticism of classical authors is the intrusion into the text of words which were originally explanatory notes written in the margin. Sometimes the paraphrase has extruded the original phrase, sometimes the true and the false remain side by side. This, however, is a form of corruption which occurs less often in the Biblical writings than in profane authors, and even in the latter the instances where it is proved to have taken place are much fewer than those in which it is assumed by some critics. Then there is the class of deliberate alterations, such as are known to have been made in the texts of the Greek dramatists by the actors, and such as are suspected to have been made by the scholars of Alexandria in the texts of the Attic authors generally, in deference to certain supposed laws of style and euphony. In regard to the Old Testament we know that the text of the Septuagint was extensively altered by Origen in order to bring it into closer conformity with the Hebrew text current in his day; in the case of the New Testament there is good reason to suppose that many of the divergences which now exist were due to deliberate editing, intended, no doubt, to secure the best possible text according to the materials available for the scribe or his director, but often resulting in departures from the true and original reading. In the case of religious books there is also always the danger of deliberate alteration for doctrinal reasons, and we know that various heretical sects had their own recensions of certain books of the Bible; but this danger is discounted by the enormous mass and variety of evidence in existence for the New Testament. There is no possibility that all the sources should be tainted; one or other of them would be sure to have escaped, and when once the alternatives are presented to the critic, there is generally little difficulty in detecting a doctrinal perversion.

What Are Textual Variants [Errors] and How Many Are There?

A special form of deliberate alteration, for which the student of the New Testament has to be on his guard, occurs in the case of the Synoptic Gospels. When the same event is recorded by two or more writers, there was a natural temptation to scribes to amplify one by the insertion of details mentioned in another, or to use the phrases of the more familiar version in transcribing that which was less familiar. This is a form of corruption which is constantly found in the later MSS. of the Gospels; and any one who will take the trouble to compare the Authorized and Revised Versions of the English Bible will find many instances in which the Revisers have removed such “harmonistic” corruptions from the text. The identification of them, however, involves the whole question of the origin of the Synoptic tradition; for if, as is now universally held in one form or another, a common document forms the substratum of the three Gospels, it may be questioned whether the verbal variations which now appear in the narratives are due to modifications of the original document by the evangelists themselves, or to the errors of early scribes. Even, however, if the latter be the true explanation (which is hardly probable), the divergences certainly established themselves at a very early date, and the removal of them in later manuscripts may in most, if not in all, cases be assigned with confidence to the editorial initiative of scribes and not to the following of primitive authorities; and this class of deliberate alteration must be kept constantly in mind by the textual critic of the Gospels.

THE DOCUMENTARY APPROACH in New Testament Textual Studies

Finally, there are errors of which nothing can be said save that they are unaccountable. Every one who has done much writing must know that now and again he puts down words which have no meaning in the context in which he uses them, or (if he is copying) are wholly unlike the words which he should have copied. His mind has strayed, and he has written down words which some obscure train of association has put into his head. Errors such as these are sometimes made by the copyists of manuscripts, and since they have no traceable connection with the true text, they do not, as some kinds of error do, provide the means for their own correction. The same may be said of errors due to the defectiveness of the manuscript from which the copy has been made. A word may be defaced or obliterated, and the copyist must either omit it or guess at it; and since a copyist often has but a hazy idea of the sense of what he is copying, his guesses are often very wide of the mark. Errors from mutilation would arise with especial ease during the period when papyrus was the material in use for literary purposes. The surface was more delicate than that of vellum, and therefore more liable to small and local injuries, which will obscure, or wholly obliterate, a word or a sentence. Here again the true reading is often irrecoverable except by guessing, and even if a guess be right, it can rarely be proved to be right; and an unverified guess can carry but little weight for practical purposes. A good example of this has recently come to light in the sphere of classical literature. In a quotation from a poem by Solon, preserved to us by the rhetorician Aristides, where the lawgiver is depicting the miseries of his country, a certain section of the population was described as

τοὺ δʼ ἀναγκαίης ὕπο
χρησμὸν λέγοντας, γλῶσσαν οὐκέτʼ Ἀττικὴν
ἱέντας.

Here, the words χρησμὸν λέγοντας were practically unintelligible, in spite of the bravest efforts of conscientious commentators. Various emendations were suggested, but none was generally accepted as satisfactory; till at last the discovery of Aristotle’s Ἀθηναίων πολιτεία, where the passage is quoted, revealed the fact that the true reading is χρειοῦς φυγόντας. The change is not great, only seven letters being affected; but there is no palaeographical similarity between the false letters and the true, to account for the corruption. It is probable, therefore, that the two words were injured in an early manuscript of Aristides’ treatise, and that the scribe of the copy from which all the extant manuscripts of it are derived, wrote down two familiar words similar in general appearance. It is instructive to observe that one modern scholar had, in fact, guessed approximately the right reading; but the guess, wanting confirmation and not supplying in itself any explanation of the origin of the corruption, remained wholly without authority or acceptance. The same has doubtless happened in many of the corrupt passages of the classical writers, but in the New Testament the number and diversity of the witnesses renders it almost certain that, even if such an error has vitiated one group of manuscripts, the true reading will be preserved elsewhere.

HOW DO WE DETERMINE THE ORIGINAL READING THROUGH The Principles and Practice of New Testament Textual Studies?

These, then, are the main forms of error with which the textual critic has to contend; and to meet them he has, as has been said above, the two weapons of comparison of documents and conjecture. He has before him a number of manuscripts, and in the first instance he may (or in the case of the New Testament it may almost be said that he must) assume that the truth lies somewhere among them. In many cases the choice is obvious. Errors of spelling or grammar, when confronted with the true readings, must give way at once. Where conviction does not lie quite on the surface, the critic who bears in mind the common causes of error enumerated above can often see how the divergence has arisen, and which of the conflicting readings is original. In some cases he will see that homoioteleuton will account for an omission; in others, that the intrusion of a marginal comment accounts for an addition; in others, that two or three letters have been mistaken by the scribe for others which resemble them. Sometimes he may suspect deliberate alteration, whether with the object of bringing out a doctrine more clearly, or to improve the literary form of the passage, or to reconcile two divergent readings which the scribe had before him. By these methods considerable progress may be made in weeding out errors, and at the same time the critic will be accumulating materials for the second stage of his work, namely, the discernment of the comparative merits of his various authorities. He will learn which manuscripts are most often right, which are closely akin to one another, which groups are nearest in the line of descent to the original autograph. Hence he will have some clue to guide him when the choice between divergent readings is not evident at first sight. In such cases it is clearly safest to follow, as a rule, the authority which has shown itself to be most trustworthy. The more the parentage of the several manuscripts can be traced, the more they can be classified into groups, and the history and origin of the groups made clear, the better is the chance of arriving at a sound text of the author under examination. Examples of the use of such methods will be found in the succeeding pages of this handbook; for the present it must be sufficient to describe them merely in outline.

One proposition is so often stated as a leading principle in textual criticism as to deserve a brief separate mention. It is that which is formulated by Bengel in the words, Proclivi scriptioni praestat ardua, or, as it is sometimes expressed, Difficilior lectio potior; the harder reading is to be preferred to the easier. Stated so absolutely, this proposition is misleading. Many forms of mistake produce a reading harder than the true one. Thus χρησμὸν λέγοντας, in the instance quoted above, is manifestly a more difficult reading than χρειοῦς φυγόντας, but it is none the less wrong. Similarly, errors due to homoioteleuton often produce nonsense, as in the case quoted on p. 8 from the Codex Vaticanus. In fact, it may be said generally that in the case of accidental errors the principle is not sound; but in the case of errors due to deliberate alteration it is generally true. A scribe or commentator fails to understand a passage and puts in some word which he thinks makes it easier; an odd word is replaced by a commoner one; a marginal paraphrase extrudes the phrase which it was intended to explain; an expression which may give offence is omitted or toned down. In all such cases the more difficult reading is likely to be the true one. A hard reading will not be deliberately inserted instead of an easy one; but the reverse may, and not infrequently does, take place. The difficulty, of course, is to determine whether a discrepancy between two or more manuscripts is due to accidental or deliberate alteration; and where this cannot be discerned with certainty, Bengel’s canon must be applied with great caution.

Of wider application and less qualified truth is another canon, in which this of Bengel’s is included, namely, that of two or more alternative readings, that one is most likely to be right which most easily accounts for the origin of the others. The “difficilior lectio” is preferable just because a hard reading is likely to be altered into an easy one, not an easy reading into a hard one. So too a scribe, writing without any clear comprehension of the sense of the text which he is copying, not infrequently substitutes a familiar phrase for a strange one, even though in reality it reduces the passage to nonsense. Even where both alternatives make sense, one can easily be seen to have suggested the other, while the reverse process is impossible or improbable. Thus in another part of the poem of Solon mentioned above, the MS. of the Ἀθηναίων πολιτεία, in which it is quoted, has πρὶν ἀνταράξας πῦαρ ἐξεῖλεν γάλα, while the MSS. of Plutarch, who also quotes it, have πρὶν ἂν ταράξας πῖαρ ἐξέλῃ γάλα. Here it is easy to understand how the scribe of some ancestor of the Plutarch MSS. (copying, of course, from a MS. in which the words were not separated) took αν to be the familiar particle ἄν, not the syncopated form of the preposition ἀνά in composition, and so altered ἐξεῖλεν into ἐξέλῃ because πρὶν ἄν requires a subjunctive; but it is highly improbable that anyone, with a correct reading πρὶν ἂν ταράξας … ἐξέλῃ before him, would be dissatisfied with it and alter it to πρὶν ἀνταράξας … ἐξεῖλεν. On the other hand, considerations of sense make the πῖαρ of the Plutarch MSS. preferable to the πῦαρ of the Aristotle MS.

It remains to ask what place is left for the second weapon of textual criticism, conjecture; and it has been usual to answer that in the criticism of the New Testament it has no place at all. Where manuscript evidence is scanty, as it is for many of the classical authors, it happens at times that a passage is obviously and certainly corrupt in all the extant copies; and then the defect must be healed by conjecture, if it is to be healed at all. But where the evidence is so plentiful and varied as it is for the New Testament, the chances that the true reading should have been lost by all are plainly very much smaller. Whether, however, conjecture is to be absolutely excluded depends in a large measure on the view which the critic takes of the character of the existing manuscript evidence. As will be shown in a later chapter, one school of critics regards the large majority of extant manuscripts as representing a relatively late recension of the sacred text, and therefore considers its evidence as of little value. The number of authorities which remain is thus comparatively small, and they differ considerably among themselves; and hence critics of this school are prepared to admit that, here and there, the original readings may have been wholly lost. Thus in Col. 2:18 Westcott and Hort (in substantial agreement with Lightfoot) are inclined to believe that the apostle wrote, not ΑΕΟΡΑΚΕΝΕΜΒΑΤΕΥΩΝ (ἃ ἑόρακεν ἐμβατεύων), but ΑΕΡΑΚΕΝΕΜΒΑΤΕΥΩΝ (ἀέρα κενεμβατεύων), the mistake being palaeographically very easy, and the improvement in sense through the conjecture considerable.¹ It is universally agreed, however, that the sphere of conjecture in the case of the New Testament is infinitesimal; and it may further be added that for practical purposes it must be treated as non-existent. No authority could be attached to words which rested only upon conjecture; and a critic who should devote himself to editing the Scriptures on conjectural lines would be merely wasting his time. Where nothing but questions of literary style are involved, we may be willing to accept a reading upon conjecture, if no better evidence is to be had; but where it is a question of the Word of Life, some surer foundation is required.

Putting conjecture aside, therefore, the function of the textual critic is, first, to collect documentary evidence, and, secondly, to examine it and estimate its value. The object of the present volume is to show what has been done in both these directions. In Chapters II.–VI. an account will be given of the available textual material—the copies of the New Testament in the original Greek, the ancient translations of it into other languages, and the quotations from it which are found in the early writers of the Christian Church. The materials having been thus passed in review, an attempt will be made in Chapters VII. and VIII. to summarise what has hitherto been done in the way of using these materials, to discuss the principal theories now current with regard to the early history of the New Testament text, and to estimate the general position of the textual problem at the present day. It is all well-trodden ground, and each newcomer is infinitely indebted to the labours of his predecessors; but it is ground which each generation must tread afresh for itself, if it is to keep its interest alive in a subject of such importance, and if it is to add ever so little to the knowledge which past generations have handed down to it. It is but a humble part that textual criticism has to play. It is but the temple-sweeper in the courts of the Lord; but honest labor, even in that humble field, is not lost.

Source

Frederic G. Kenyon, Handbook to the Textual Criticism of the New Testament (London; New York: Macmillan and Co., 1901), 1–16.

¹ Forty appears to be the number of those that have been collated; but there are probably several that have not been collated. Very few, however, contain more than the three plays which were habitually read by the Byzantine public.

¹ Introduction to the New Testament in the Original Greek, p. 2 (1882).

¹ Early vellum MSS., from the fourth to the ninth centuries, are almost wholly without accents, breathings, and divisions. Papyrus MSS., which are still earlier, not infrequently have occasional accents, and, in rare instances, the separation of words is indicated by a dot in cases of doubt.

¹ For other examples of conjectural emendations proposed in the N. T. text, see Nestle, Introd. to the Textual Criticism of the N. T., Eng. Tr. pp. 167–170.