[Transcriber's Note: Transcriptions of the Plates are at the end of this e-book.]

I. THE Ma.n.u.sCRIPT TEXT.

1. The preservation of the primitive text of the gospels from all essential corruptions, additions, and mutilations has already been shown at some length (Part 1, Chap. 3). The same line of argument applies substantially to the other books of the New Testament. Though the text of different books varies in respect to purity, there is no ground for supposing that if we had the autographs of the evangelists and other sacred writers, they would present to us a gospel differing in any essential particular from that which we now possess. We should see in them the same glorious Saviour, and the same holy system of doctrines and duties.

2. But it has not pleased G.o.d to interpose in a miraculous way for the purpose of keeping the primitive text in a state of immaculate purity.

He has left it subject to those common influences which produce what are called _various readings_ in all works that are perpetuated from age to age by transcription. Compared indeed with any other ancient writings, the text of the New Testament has immensely the advantage in regard to uncorruptness of preservation and means of verification. This arises from the early multiplication of copies, as well as from the high value attached by the primitive churches to their sacred books, and their consequent zeal for their uncorrupt preservation. But the same multiplication of copies which const.i.tutes a sure guarantee against essential mutilations and corruptions increases also the number of various readings. Suppose, for example, that of two books equal in size the second has been, from the first, copied a hundred-fold oftener than the first. It is plain that, while the means of ascertaining and verifying the true text of the second will abound, the number of variations among the different ma.n.u.scripts will abound also. The greater the number of copies, the greater will be the number of various readings, but this will make the true text not more but less uncertain; for by diligent collation a text may be produced which, though not absolutely immaculate, is very near to the primitive autograph, and which can be certainly known to agree with it in every essential respect. G.o.d does not rain down upon men bread and raiment from heaven, as he could do with infinite ease; but he imposes upon them the necessity of gaining both by hard labor. "In the sweat of thy face shalt thou eat bread" is the stern law. G.o.d does not miraculously communicate to the missionary who goes to Syria or India or China a knowledge of the vernacular in his field of labor; but he must learn it by years of patient study. And when he begins the work of translating, G.o.d does not keep him in a supernatural way from all errors. He must find out and correct his errors by the diligent use of the means at his disposal.

Just so it is the will of G.o.d that we should have a pure text of the New Testament--pure in a critical sense--not without hard labor, but by years of patient toil in the study and collation of the abundant materials which his good providence has preserved for us.

3. _Various readings_ have arisen in the ma.n.u.scripts of the New Testament, as elsewhere, from the mistakes, and sometimes from the unskilful corrections of the copyists and those subsequently employed to compare and correct the copies. They are commonly divided into the three cla.s.ses of _subst.i.tutions_, _insertions_, and _omissions_.

_Subst.i.tutions_ from similarity of sound would naturally arise among the vowels when, as was sometimes the case, the copyist wrote from dictation, being guided by the ear instead of the eye. Most of these, however, are mere matters of orthography. It is only when they affect the sense that they come under the head of various readings. Synonymous words, or those of kindred meaning, are frequently put for one another, or the order of words is altered; sometimes a different word is made through inadvertence by the change of a single letter or a couple of letters; compound words are interchanged with simple; contracted words are confounded with each other; plainer or more grammatical readings are subst.i.tuted for those that are difficult or less grammatical, etc.

Especially are parallel pa.s.sages in one writer altered, so as to be brought into conformity with the same in another.

_Insertions_ are the most frequent mode of variation. The copyist fills out the text of his author from a parallel pa.s.sage, inserts marginal notations in the text, repeats clauses through inadvertence, etc.

Of amplification from parallel pa.s.sages many undoubted examples could be given. A single one must suffice. In Acts 9:5, the words, _It is hard for thee to kick against the p.r.i.c.ks_, have been added from Acts 26:14.

The most fruitful source of _omissions_ is the similar termination of two adjacent words, lines, or sentences, causing the eye of the copyist to overlook the word, line, or sentence intervening between the two similar endings. The same error may be caused by the circ.u.mstance of two sentences beginning in the same way. It should be remembered that in the ancient ma.n.u.scripts the text was written continuously in uncial--that is, capital--letters, without any division between the words, which made it more difficult for the copyist to follow the ma.n.u.script before him, and for both the copyist and collater to discover the errors made in transcription.

By far the greatest number of various readings had their origin in simple inadvertence. Some of them, however, are due to unskilful criticism; as when the copyist or the corrector sought to bring a pa.s.sage in one writer into more exact agreement with the corresponding pa.s.sage in another, to supply supposed deficiencies or correct supposed errors in his copy, or to subst.i.tute smoother and more grammatical forms of expression. Wilful falsifications in the interest of a particular sect or party cannot with any show of justice be imputed to the men who have perpetuated to us the text of the New Testament.

4. The _materials_ for textual criticism are much more abundant in the case of the New Testament than of the Old. A vast ma.s.s of ma.n.u.scripts has been collected from different and distant regions, dating from the fourth century and onward. Of these, part are in the original Greek, part in ancient versions, or bilingual, that is, containing the original and a version of it side by side. In addition to these are the quotations of the early fathers, which are so abundant that a large part of the New Testament text might be collected from them alone. The question of the history of the text, as gathered from this rich ma.s.s of materials, is very interesting, but is foreign to the plan of the present work. To give even a history of the controversies respecting the proper cla.s.sification of the ma.n.u.scripts of the New Testament according to their characteristic readings would require a volume, and the question must be regarded as yet unsettled. There are, however, some general results, a few of the more important of which are here given from Tregelles (in Horne, vol. 4, chap: 8).

The variations in the form of the sacred text are not due to any general recensions or revisions by ecclesiastical authority, but arose gradually from the causes above considered (No. 3). These variations exhibit such gradations of text that it is impossible to draw definite lines of cla.s.sification, without admitting so many exceptions as almost to destroy the application of such a system.

There is a general difference in characteristic readings between the more ancient ma.n.u.scripts, versions, and citations, and the copies of general circulation in more recent times. This gives rise to the general line of demarcation between the _more ancient_ and the _more recent_ texts; each of these two cla.s.ses, however, having, in turn, its own points of difference among the texts belonging to it.

The more ancient ma.n.u.scripts, versions, and citations which we possess range themselves under what we know from their combined testimony to be the more ancient text. Among the ma.n.u.scripts and doc.u.ments so allied there are such shades of difference and characteristic peculiarities, that the versions and ma.n.u.scripts might be easily contemplated as ramifying into two subcla.s.ses.

The most ancient doc.u.ments in general are sufficiently dissimilar to enable us to regard their testimony, when combined, as c.u.mulative.

5. Respecting the materials for writing in ancient times--papyrus and parchment, afterwards paper made from linen or cotton; the form of ma.n.u.scripts--the roll with papyrus, and the book-form with leaves when parchment was used; the use of _palimpsests_; the _uncial_ and _cursive_ styles of writing; and the means of determining the age of ma.n.u.scripts, see in Chap. 3, No. 2. The existing ma.n.u.scripts have been all numbered and catalogued. The custom since the time of Wetstein has been to mark the uncial ma.n.u.scripts by capital letters, and the cursives by numbers or small letters. We append a brief notice of a few of the more celebrated ma.n.u.scripts.

There are four very ancient and important ma.n.u.scripts, all of which originally contained the entire Greek Bible of the Old and New Testament, and which belong to a time when the arrangements of Euthalius, especially his stichometrical mode of writing (Chap. 25, Nos. 6-9), had either not been introduced or not come into common use. These are the following:

(1.) The _Codex Vatica.n.u.s_, _Vatican ma.n.u.script_, marked by the letter B, and so called from the Vatican library at Rome to which it belongs. It is written continuously (without any division of words) on very fine vellum--one of the marks of high antiquity--in small but neat uncial letters, very much like those of the ma.n.u.script rolls of Herculaneum, and has three columns to the page, which is of the quarto size. Originally it had at the end of particular sections a small empty s.p.a.ce of the breadth of a letter or half a letter, but no ornamental capitals, marks of punctuation, or accents, though some of these have been added by later hands. The divisions into sections made by the empty s.p.a.ces above named are peculiar to this codex, not agreeing with those of any other system. Of these Matthew has 170; Mark, 62 (so says Cardinal Mai, but others say 72 or 61); Luke, 152; John, 80. Most of the books have also brief t.i.tles and subscriptions. The ma.n.u.script contained originally the whole Bible, the Apocrypha included, as also the epistle of Clement to the Corinthians. The order of the books in the New Testament, if entire, would be the same as in the Alexandrine ma.n.u.script, the Catholic epistles preceding the Pauline, and the epistle to the Hebrews coming in between 2 Thessalonians and 1 Timothy. See below. At present the Old Testament wants the greater part of Genesis and a part of the Psalms. In the New Testament the epistle to Philemon, the three pastoral epistles, the latter part of the epistle to the Hebrews, and the Apocalypse are wanting. This ma.n.u.script is generally referred to the fourth century. Its authority is very high, but through the jealousy of its Roman conservators it has been of late years, for all practical purposes, inaccessible to biblical scholars. Cardinal Mai's edition of it in 1858, and the revision of this in 1859, are unreliable. Tischendorf has published an edition of the New Testament part of it. _No. (3) PLATE II_.

(2.) The _Codex Sinaiticus_, _Sinai ma.n.u.script_, designated by Tischendorf, its discoverer, by the Hebrew letter _aleph_ ([Hebrew: A]). One of the most interesting events of the present century, in the department of biblical science, is the very unexpected discovery of a complete ma.n.u.script of the New Testament, belonging, as is generally agreed, to the fourth century; therefore as old, at least, as the Vatican ma.n.u.script, perhaps older, and of very high authority in biblical criticism.

In a visit to Mount Sinai in 1844, Tischendorf had found at the convent of St. Catharine on Mount Sinai forty-three beautiful parchment leaves belonging to a ma.n.u.script of the Septuagint not before known to biblical scholars. In a subsequent visit to the same convent in February, 1859, it was his high privilege to find of the same ma.n.u.script all the Greek New Testament entire, part of the Old, the so-called epistle of Barnabas, and part of the writing called the Shepherd of Hermas, the whole contained in one hundred and thirty-two thousand columnar lines, written on three hundred and forty-six leaves. This precious ma.n.u.script Tischendorf managed to obtain for the emperor Alexander of Russia as the great patron of the Greek church, and it is now at St. Petersburg. It is written on parchment of a fine quality in large plain uncial letters, with four columns to a page. It contains, as is commonly the case with ancient ma.n.u.scripts, revisions and so-called corrections by a later hand; but, as it proceeded from the pen of the original writer, it had neither ornamented capitals, accents, nor divisions of words or sentences. The style of writing is plain, and every thing about it bears the marks of high antiquity. The order of the books is as follows: (1) the gospels; (2) the epistles of Paul, that to the Hebrews included, which stands after 2 Thessalonians; (3) the Acts of the Apostles; (4) the Catholic epistles; (5) the Apocalypse. It has the Ammonian sections and Eusebian canons, but whether from the first or a subsequent hand is doubtful. A splendid edition of this Codex was published at St. Petersburg in 1862, which seeks to preserve with the greatest possible accuracy the form of writing, columns, corrections, etc. The Leipsic edition is adapted to popular use. _See No. (1), PLATE I_.

(3.) We will consider next in order the _Codex Alexandrinus_, _Alexandrine ma.n.u.script_, placed first in the list of uncial ma.n.u.scripts, and accordingly marked A. It is now in the British Museum, London. In the year 1628 it was sent as a present to Charles I., king of England, by Cyrillus Lucaris, patriarch of Constantinople, by whom it was brought from Alexandria in Egypt, where Cyrillus had formerly held the same office. Hence the name Alexandrine. Cyrillus himself, in a notice attached to it, says that tradition represented a n.o.ble Egyptian woman of the fourth century named Thecla as the writer of it (an Arabic subscription makes her to have been Thecla the martyr). These external notices are not so reliable as the internal marks, all of which show it to be of a great age. Some a.s.sign it to the fourth century, but it is more commonly a.s.signed to the fifth, and Egypt is generally regarded as the place where it was written.

It is on parchment in uncial letters, without divisions of words, accents, or breathings, and with only occasional marks of interpunction--a dot to indicate a division in the sense. The lines are arranged in two columns, and the sections begin with large letters, placed a little to the left of the column--outside the measure of the column. The order of the books is: (1) the gospels; (2) the Acts of the Apostles; (3) the Catholic epistles; (4) the epistles of Paul, with that to the Hebrews between 2 Thessalonians and 1 Timothy; (5) the Apocalypse. In the gospels, the Ammonian sections with the Eusebian canons are indicated, and at the top of the pages the larger sections or _t.i.tles_. In the Old Testament it is defective in part of the Psalms. In the New it wants all of Matthew as far as chap. 25:6; also from John 6:50 to 8:52; and from 2 Cor. 4:13 to 12:6. It has appended at the end the genuine letter of Clement of Rome to the Corinthians, and a fragment of a second spurious letter. To these apocryphal additions we owe the preservation of the Apocalypse in an entire state. Until the discovery of the Sinai codex, the Alexandrine exhibited the text of the New Testament in far the most entire state of all the uncial ma.n.u.scripts. _See No. (2), PLATE I_.

(4) The fourth ma.n.u.script of this group is the celebrated palimpsest called _Codex Ephraemi_, _Ephraem ma.n.u.script_, preserved in the Imperial library of Paris, and marked in the list of uncials with the letter C. Originally it contained the whole of the New Testament, and apparently the Old also, elegantly written on thin vellum, with a single column to a page. The writing is continuous, without accents or breathings, and the letters are rather larger than in the Alexandrian ma.n.u.script, the first letter of each section being of larger size than the rest, and standing, as in that ma.n.u.script, a little to the left of the column. The Ammonian sections stand in the margin, but without the Eusebian canons. The gospels were preceded by the list of _t.i.tles_, or larger sections, of which those of Luke and John alone are preserved. The t.i.tles and subscriptions are short and simple. The date of the ma.n.u.script is supposed to be the first half of the fifth century. It has undergone corrections at the hand of at least two persons, possibly a third. These can be readily distinguished from the original writing. The critical authority of this codex is very high. Tregelles (in Horne, vol. 4, chap. 13) places it next to the Vatican ma.n.u.script.

A few words on its history. About the thirteenth century, being regarded as a worn-out and obsolete ma.n.u.script, the vellum on which it was written was taken for a new purpose, that of receiving the Greek works of Ephraem the Syrian saint, a celebrated theologian of the old Syrian church, who flourished in the fourth century. "For this purpose the leaves were taken promiscuously, without any regard to their proper original order, and sewed together at hap-hazard, sometimes top end down, and front side behind, just as if they had been mere blanks, the sermons of Ephraem being the only matter regarded in the book."

Stowe, Hist. of the Books of the Bible, p. 75. In the latter part of the seventeenth century, Allix first observed the older writing under the works of Ephraem. It was very illegible, but a chemical preparation applied in 1834-5 revivified it to a certain extent. It has been diligently collated by eminent scholars, and in 1842 Tischendorf printed an edition of it page for page and line for line. Of the two hundred and nine leaves contained in this ma.n.u.script, one hundred and forty-five belong to the New Testament, containing not quite two-thirds of the sacred text. The order of the books is the same as in the Alexandrine codex. _See No. (4), PLATE III_.

Besides the abovenamed four ma.n.u.scripts, a few others may be briefly noticed.

An interesting palimpsest of great critical value is the _Codex Dublinensis rescriptus_, _Dublin palimpsest ma.n.u.script_, in the library of Trinity College, Dublin, designated by the letter Z.

It contains with other writings thirty-two leaves of the gospel by Matthew. They were edited, as far as legible, in 1801, by Dr.

John Barrett, Fellow of Trinity College. In 1853 Dr. Tregelles made a new and thorough examination of the ma.n.u.script, and, by the aid of a chemical process, brought all that exists of the gospel text to a legible condition. This ma.n.u.script is a.s.signed to the sixth century. Its letters are written in a singularly bold style, which unites the three qualities of ease, elegance, and symmetry.

A celebrated _bilingual_ ma.n.u.script (in this case _Graeco-Latin_, containing the Greek and Latin texts) is the _Codex Bezae_, _Beza's ma.n.u.script_, called also _Codex Cantabrigiensis_, _Cambridge ma.n.u.script_, from the place of its deposit, which is the public library of the University of Cambridge, England. It is designated by the letter D, and contains the four gospels and Acts of the Apostles in Greek and Latin on opposite pages, stichometrically written. The account of Theodore Beza, its former possessor, was that he found it during the French civil wars in 1562, in the monastery of St.

Irenaeus, at Lyons. In 1581 he sent it as a present to the University of Cambridge. The interest felt in this ma.n.u.script arises in great part from the very peculiar character of its readings. "The text of this codex," says Bleek (Introduc. to New Test., sec. 270), "presents much that is peculiar--many additions and alterations that have even an apocryphal character, but are yet not uninteresting. Its native country is the West, and more definitely the south of Gaul." _See No. (5), PLATE IV_.

Among the _fragments_ of ma.n.u.scripts of high antiquity is one called _Codex purpureus_, _Purple ma.n.u.script_. _Four_ leaves of this are in the Cotton Library in the British Museum, _six_ in the Vatican, _two_ in the Imperial Library at Vienna. The ma.n.u.script to which they belonged was written in silver letters (the names of G.o.d and Christ in gold) on purple vellum. The writing is in two columns with large and round letters. It is referred to the end of the sixth or beginning of the seventh century.

Many other uncial ma.n.u.scripts, or fragments of ma.n.u.scripts, some of them of great critical value, might be described; but the above brief notices must suffice. Of those which contain ancient _versions_, a few of the more important will be noticed in the following chapter.

The _cursive_ ma.n.u.scripts of the Now Testament are numbered by hundreds. In general their authority is less than that of the more ancient uncials. But a cursive ma.n.u.script may give indirectly a very ancient text. There are some cursives which, from their characteristic readings, were manifestly executed from codices of high antiquity, and are for this reason very valuable. As such Tregelles specifies those numbered 1 and 33.

For further notices of these, as also of the _lectionaries_, containing selections for church readings, the reader may consult the works devoted to biblical criticism.

II. THE PRINTED TEXT.

6. The _primary editions_ of the Greek New Testament, whence is derived what is called _the received text_ (_Textus receptus_) are the following: (1) the _Complutensian_; (2) the _Erasmian_; (3) those of _Robert Stephens_; (4) those of _Beza_ and _Elzevir_. Their authority in textual criticism depends wholly upon that of the ma.n.u.scripts from which their text was formed. As no stream can rise higher than its fountains, so no printed text can obtain a just weight of influence above that of its ma.n.u.script sources. It becomes, then, a matter of interest to inquire what was the basis of these early printed editions.

(1.) The entire New Testament was printed for the first time in Greek in the fifth volume of the _Complutensian Polyglott_ (so called from _Complutum_, that is _Alcala_ in Spain, where it was printed under the patronage of Cardinal Ximenes). It bears the date of 1514, but was not published until 1522, when Erasmus had already printed three editions of his Greek Testament. Its editors professed to have formed their text from ma.n.u.scripts sent to them from the papal library at Rome. What these ma.n.u.scripts were cannot now be ascertained; but that they were very ancient and correct, as alleged by these editors, is contradicted by the character of the text, which agrees with the modern in opposition to the most ancient ma.n.u.scripts.

(2.) At the request of Froben, a celebrated printer and publisher of Basle, _Erasmus_, who was then in England, where he had devoted some time to a revised Latin translation of the New Testament with annotations, went to Basle in 1515, and began the work of editing a Greek New Testament. "By the beginning of March, 1516," says Tregelles, "the whole volume, including the annotations as well as the Greek and Latin texts, was complete; in less, in fact, than six months from the time that the first sheet was begun." The design of this haste was to antic.i.p.ate the publication of the Complutensian edition. The critical apparatus in Erasmus' possession was quite slender. It consisted of such ma.n.u.scripts as he found at Basle, with the help of the revised Latin translation already prepared in England and Brabant. For the Apocalypse he had but one ma.n.u.script, and that defective at the end. In his four subsequent editions--1519, 1522, 1527, 1535--he made many corrections. In that of 1527 he availed himself of the Complutensian text. This edition, from which the fifth and last published during his life differs but slightly, is the basis of the common text now in use.

(3.) In 1546, 1549, 1550, appeared the three editions of _Robert Stephens_, the celebrated Parisian printer. In the first two of these the text is said to have been formed from the Complutensian and Erasmian. In the third edition, although he had the aid of thirteen Greek ma.n.u.scripts, his text is almost identical with that of Erasmus' fifth edition.

(4.) In 1565, _Theodore Beza_ published at Geneva his first edition of the Greek Testament with his own Latin version, and also the Vulgate with annotations. Three other editions followed in 1576, 1582, 1588-9. He had the use of the Codex Bezae above described, the Codex Claromonta.n.u.s (an ancient Graeco-Latin ma.n.u.script of the Pauline epistles), the Syriac version then recently published by Tremellius, with a close Latin translation, and Stephens' collations. But he is said not to have made much use of these helps.

The first of the _Elzevir_ editions, so celebrated for their typographical beauty, was issued in 1624, its text being mainly copied from that of Beza. This is the text that has acquired the name of _Textus receptus_, the _Received Text_, as it was for more than a century the basis of almost all subsequent editions.

The genealogy of this _Textus receptus_ is thus succinctly given by Bishop Marsh: "The _Textus receptus_, therefore, or the text in common use, was copied, with a few exceptions, from the text of Beza. Beza himself closely followed Stephens; and Stephens (namely, in his third and chief edition) copied solely from the fifth edition of Erasmus, except in the Revelation, where he followed sometimes Erasmus, sometimes the Complutensian edition.

The text, therefore, in daily use, resolves itself at last into the Complutensian and the Erasmian editions." Divinity Lectures, part I, p. 111.

7. It requires but a moderate acquaintance with the history of textual criticism to understand that the Elzevir text is not only not perfect, but is more imperfect than that which has been elaborated by the help of the abundant ma.n.u.scripts, versions, and citations of the early fathers, of which modern criticism has availed itself. It is no reproach to the editors of the primary editions that, with their comparatively scanty materials, they could not accomplish as much as we can with the rich and varied means at our disposal. The _essential integrity_ of the received text, we do indeed thankfully acknowledge and firmly maintain. Our fathers had presented to them in this text the same divine and glorious Saviour, the same way of salvation, the same holy system of doctrines and duties, as we now find in the most carefully revised modern text.

Nevertheless, a true reverence for the inspired word must impel us to the diligent use of all the means at our command for setting forth a pure text, that is, a text conformed as nearly as possible to that of the original autographs. Viewed in this light the modern critical editions of the New Testament must possess a deep interest for all who are able to read it in the original tongue. But to discuss the merits of these would be foreign to the design of the present work.

Examples of the more important various readings occur in John 1:18; Acts 20:28; 1 Tim. 3:16. The pa.s.sage 1 John 5:7, 8, _in heaven--in earth_, is generally rejected on the testimony of the ma.n.u.scripts (see the full discussion in Horne, vol. 4, ch. 36).

Among the pa.s.sages which are regarded as more or less doubtful may be mentioned John 5:4; 8:3-11; Acts 8:37. In regard to all these the biblical scholar must be referred to the critical commentaries. So also for the questions connected with the text of Mark 16:9-20, which are of a peculiar character.

III. PRINCIPLES OF TEXTUAL CRITICISM.

8. The end proposed by textual criticism is to restore the sacred text as nearly as possible to its primitive purity (Chap. 7, No. 1). To this work the biblical scholar should come in a candid and reverential spirit, prepared to weigh carefully all the evidence which is accessible to him, and decide, not as an advocate, but as a judge, in the simple interest of truth. The three great sources of evidence for the original text of the New Testament are Greek ma.n.u.scripts, versions, and the citations of the fathers. Of these, _Greek ma.n.u.scripts_ hold the first place. But all ma.n.u.scripts are not of equal value. Other things being equal, the oldest ma.n.u.scripts have the highest authority. "If the multiplication of copies of the New Testament had been uniform, it is evident that the number of later copies preserved from the accidents of time would have far exceeded that of the earlier, yet no one would have preferred the fuller testimony of the thirteenth to the scantier doc.u.ments of the fourth century. Some changes are necessarily introduced in the most careful copying, and these are rapidly multiplied." Westcott in Smith's Bible Dict.; Art. New Test. Yet, as the same writer remarks, we may have evidence that a recent ma.n.u.script has been copied from one of great antiquity, and thus has preserved to us very ancient readings.

Revisions and corrections by a later hand are to be carefully distinguished from the primitive writing. Yet these may be valuable, as testifying to the prevailing reading of the age to which they belong.

The general cla.s.s or family to which a given ma.n.u.script belongs is also to be taken into the account. In a word, so many elements of judgment are to be taken into account in determining the relative weight of authority that belongs to a given ma.n.u.script, that the right decision of the question requires large observation combined with much critical tact.

9. _Ancient versions_ are of great value in textual criticism; for some of them, as the old Latin and Syriac, to which may be added the old Egyptian versions, are based on a text more ancient than that preserved to us in any ma.n.u.script. In textual criticism, the testimony of a version is valuable in proportion to its antiquity, its fidelity--not its elegance or even its correctness of interpretation, but its literal closeness--and the purity of its text. Versions are liable to all the corruptions of text incident to Greek ma.n.u.scripts, and far more liable to interpolations by explanatory glosses. The difference of idiom, moreover, frequently prevents such a literal rendering as shall be a sure indication of the form belonging to the original text.

10. The _citations_ of the church fathers, which are immensely numerous, const.i.tute another source of testimony. But less authority belongs in general to these, because they are often made loosely from memory alone.