THE word "text" properly denotes a literary work, conceived of as a mere thing, as a texture woven of words instead of threads. It designates neither, on the one side, the book which contains the text, nor, on the other side, the sense which the text conveys. It is not the matter of the discourse, nor the manner of it, whether logical, rhetorical, or grammatical. It is simply the web of words itself. It is with this understanding that the text of any work is concisely defined as the ipsissima verba of that work.

The word, which came into Middle English from the French where it stands as the descendant of the Latin word textum, retains in English the figurative sense only of its primitive, yet owes it to its origin that it describes a composition as a woven thing, as a curiously interwoven cloth or tissue of words. Once a part of the English language, it has grown with the growth of that tongue, and has acquired certain special usages. We usually need to speak of the exact words of an author only in contrast with something else, and thus "text" has come to designate a composition upon which a commentary has been written, so that it distinguishes the words commented on from the comments that have been added. Thus we speak of the text of the Talmud as lost in the comment. And thus, too, by an extreme extension, we speak of the text of a sermon, meaning, not the ipsissima verba of the sermon, but the little piece of the original author on which the sermon professes to be a comment. By a somewhat similar extension we speak of texts of Scripture, meaning, not various editions of its ipsissima verba, but brief extracts from Scripture, as for example proof texts and the like;—a usage which appears to have grown up under the conception that all developed theology is of the nature of a comment on Scripture. Such secondary senses of the word need not disturb us here. They are natural developments out of the ground meaning, as applied to special cases. We are to use the word in its general and original sense, in which it designates the ipsissima verba, the woven web of words, which constitutes the. concrete thing by which a book is made a work, but which has nothing directly to do with the sense, correctness, or the value of the work.

There is an important distinction, however, which we should grasp at the outset, between the text of a document and the text of a work. A document can have but one text; its ipsissima verba are its ipsissinui verba, and there is nothing further to say about it. But a work may exist in several copies, each of which has its own ipsissima verba, which may, or may not, tally with one another. The text of any copy of Shakespeare that is placed in my hands is plainly before me. But the text of Shakespeare is a different matter. No two copies of Shakespeare,—or now, since we have to reckon with the printing press, we must rather say no two editions,—have precisely the same text. There are all kinds of causes that work differences: badness of copy, carelessness of compositors, folly of editors, imperfection of evidence, frailty of humanity. We know what the text of Karl Elze's Hamlet is. Bnt what is the text of Hamlet 1 We cannot choose any one edition, and say that it is the text of Hamlet; it is one text of Hamlet, but not necessarily the text of Hamlet. We cannot choose one manuscript of Homer, and say that it is the text of Homer. It is a text of Homer, but the text of Homer may be something very different. We note, then, that the text of a document and the text of a tvork may be very different matters. The text of a document is the ipsissima verba of that document, and is to be had by simply looking at it; whatever stands actually written in it is its text. The text of a work, again, is the ipsissima verba of that work, but it cannot be obtained by simply looking at it. We cannot look at the work, but only at the documents or "copies" that represent it; and what stands written in them, individually or even collectively, may not be the ipsissima verba of the work,—by exactly the amount, in each case, in which it is altered or corrupted from what the author intended to write, is not the ipsissima, verba of the work. If, then, the text of a document or copy of any work is the ipsissima verba of that document or copy, the text of the work is what ought to be the ipsissima verba of all the documents or copies that profess to represent it,—it is the original, or. better still, the intended ipsissima verba of the author. It may not lie in the document before us, or in any document. All existing documents, taken collectively, may fail to contain it. It may never have lain, perfect and pure, in any document. But if an element of ideality thus attaches to it, it ia none the less a very real thing and a very legitimate object of search. It is impossible, no doubt, to avoid a certain looseness of speech, by which we say, for example, " The text of Nonius is in a very bad state ;" and thus identify the text of a work with some transitory state of it, or it may be with the permanent loss of it. What we mean is that the text in this or that document or edition, or in all existing documents or editions, is a very bad and corrupt representation of the text of Nonius,—is not the text of Nonius at all, in fact, but departs from, and fails to be, that in many particulars. The text of Nonius, in a word, is just what we have not and are in search of.

It is clear, therefore, that the text of a work as distinguished from the text of a document can be had only through a critical process. What is necessary for obtaining it is a critical examination of the texts of the various documents that lie before us as its representatives, with a view to discovering from them whether and wherein it has become corrupted, and of proving them to preserve it or else restoring it from their corruptions to its originally intended form. This is what is meant by "textual criticism," which may be defined as the careful, critical examination of a text, with a view to discovering its condition, in order that we may test its correctness on the one hand, and, on the other, emend its errors.

Obviously this is, if not a bold and unsafe kind of work, yet one sufficiently nice to engage our best powers. It is not, however, so unwonted a procedure as it may seem at first sight; and more of us than suspect it are engaged in it daily. Whenever, for instance, we make a correction in the margin of a book we chance to be reading, because we observe a misplaced letter or a misspelled word, or any other obvious typographical error, we are engaging in processes of textual criticism. Or, perhaps, we receive a letter from a friend, read it carefully, suddenly come upon a sentence that puzzles us, observe it more closely, and say, "Oh, I see! a word has been left out here!" There is no one of us who has not had this experience, or who has not supplied the word which he determines to be needed, and gone on satisfied. Let us take an apposite example or two from printed books. When we read in Archdeacon Farrar's Messages of the Books (p. 145, note *): "That God chose His own fit instruments" for writing the books of the New Testament, "and that the sacreduess of the books was due to the prior position of these writers is clear from the fact that only four of the writers were apostles" — few of us will hesitate to insert the "not" before "due," the lack of which throws the sentence into logical confusion. So, when we read in the admirable International Revision Commentary on John's Gospel, by Drs. Milligan and Moulton (p. 341): "Yet we should overlook the immediate reference," the context tells us at once that a "not" has been omitted before "overlook." In an edition of King James' Bible, printed by Barker ife Bill, in

1631, men read the seventh commandment (Exod. xx. 14): "Thou shalt commit adultery," not without perceiving, we may be sure, that a "not" had fallen out, and mentally replacing it all the more emphatically that it was not there. But all this is textual criticism of the highest and most delicate kind. We have, in each case, examined the text before us critically, determined that it was in error, and restored the originally intended text by a critical process. Yet we do all this confidently, with no feeling that we are trenching on learned ground, and with results that are entirely satisfactory to ourselves, and on which we are willing to act in business or social life. The cases that have been adduced involve, indeed, the very nicest and most uncertain of the critical processes: they are all samples of what is called "conjectural emendation"—i.e., the text has been emended in each case by pure conjecture, the context alone hinting that it was in error or suggesting the remedy. The dangers that attend the careless or uninstructed use of so delicate an instrument are well illustrated by a delightful story (which Mr. Frederic Harrison attributes to Mr. Andrew Lang) of a printer who found in his "copy" some reference to "the Scapin of Poquelin." The printer was not a pedant; Moliere he knew, but who was Poquelin? At last a bright idea struck his inventive mind, and he printed it: "the Scapin of M. Coqudin." This is "conjectural emendation" too; and unhappily it is the type of a great part of what is called by that name.

In this higher way every reader of books is a textual critic. In a lower way, every proof-reader is a textual critic; for the correction of a text that lies before him by the readings of another, given him as a model, is simply the lowest variety of this art. The art of textual criticism is thus seen to be the art of detecting and emending errors in documents. The science is the orderly discussion and systematisation of the principles on which this art ought to proceed.

The inference lies very close, from what has been said, that the sphere of the legitimate application of textual criticism is circumscribed only by the bounds of written matter. Such are the limitations of human powers in reproducing writings, that apparently no lengthy writing can be duplicated without error. Nay, such are the limitations of human powers of attention, that probably few manuscripts of any extent are written exactly correctly at first hand. The author himself fails to put correctly on paper the words that lie in his mind. And even when the document that lies before us is written with absolutely exact correctness, it requires the application of textual criticism, i.e., a careful critical examination, to discover and certify this fact. Let us repeat it, then: wherever written matter exists, textual criticism is not only legitimate, but an unavoidable task; when the writing is important, such as a deed, or a will, or a charter, or the Bible, it is an indefeasible duty. No doubt, differences may exist between writings, in their nature or the conditions under which they were produced or transmitted, which may demand for them somewhat different treatments. The conditions under which a work is transmitted by the printing press differ materially from those under

which one is transmitted by hand-copying; and the practice of textual criticism may be affected by this difference. One work may lie before us in a single copy, another in a thousand copies, and differences may thence arise in the processes of criticism that are applicable to them. But all writings have this in common: they are all open to criticism, and are all to be criticised. An autograph writing is open to criticism; we must examine it to see whether the writer's hand has been faultless handmaid to his thought, and to correct his erroneous writing of what he intended. A printed work is open to criticism: we must examine it to see what of the aimless alteration that has been wrought by a compositor's nimble but not infallible fingers, and what of the foolish alteration which the semi-unconscious working of his mind has inserted into his copy, the proof-reader has allowed to stand. A writing propagated by manuscript is especially open to criticism: here so many varying minds, and so many varying hands, have repeated each its predecessor's errors, and invented new ones, that criticism must dig through repeated strata of corruption on corruption before it can reach the bed-rock of truth.

Nor is the arc a wide one through which even the processes of criticism which are applicable to these various kinds of writings can librate. The existence of corruptions in a writing can be suggested to us by only two kinds of evidence. One of these is illustrated by our detection of misprints in the books we read or of errors in the letters we receive. The most prominent form of it is the evidence of the context or general sense; to this is to be added, as of the same generic kind, the evidence of the style, vocabulary or usage of the author, or of the time in which he wrote, and the like,—all the evidence, in a word, that arises from the consideration of what the author is likely to have written. The name that is given to this is internal evidence, and it is the only kind of evidence that is available for an autographic writing, or any other that exists only in a single copy. But if two or more copies are extant, another kind of evidence becomes available. We may compare the copies together, and wherever they differ one or the other testimony is certainly at fault, and critical examination and reconstruction is necessary. This is external evidence. When we proceed from the detection of error to its correction, we remain dependent on these same two kinds of evidence— internal and external. But internal evidence splits here into two well-marked and independent varieties, much to our help. We may appeal to the evidence of the context or other considerations that rest on the question, What is the author likely to have written? to suggest to us what ought to stand in the place where a corruption is suspected or known; and this is called intrinsic (internal) evidence. Or we may appeal to the fortunes of reproduction, to the known habits of stone-cutters, copyists, or compositors, to suggest what the reading or readings known or suspected to be corruptions may have grown out of, or what reading, on the supposition of its originality, will account best for the origin of all others; and this is called transcriptional (internal) evidence. On the other hand, we may collate all known copies, and appeal to the evidence that a great majority of them have one reading, and only a few the others; or all the good and careful ones have one, and only the bad the others; or several derived from independent sources have one, and only such as can be shown to come from a single fountain have the others; and so marshal the external evidence. If we allow for their broad and inadequate statement, proper to this summary treatment, we may eay that it matters not whether the writing before us be a letter from a friend, or an inscription from Carchemish, or a copy of a morning newspaper, or Shakespeare, or Homer, or the Bible, these and only these are the kinds of evidence applicable. And so far as they are applicable they are valid. It would be absurd to apply them to Homer, and refuse to apply them to Herodotus ; to apply them to Nonius, whose text is proverbially corrupt, and refuse to apply them to the New Testament, the text of which is incomparably correct. It is by their application alone that we know what is corrupt and what is correct; and if it is right to apply them to a secular book, it is right to apply them to a sacred one—nay, it is wrong not to.

It is clear, moreover, that the duty of applying textual criticism—say, for instance, to the New Testament—is entirely independent of the number of errors in its ordinarily current text which criticism may be expected to detect. It is as important to certify ourselves of the correctness of our text as it is to correct it if erroneous; and the former is as much the function of criticism as the latter. Nor is textual error to be thought to be commensurable with error in sense. The text conveys the sense; but the textual critic has nothing to do, primarily, with the sense. It is for nim to restore the text, and for the interpreter who follows him to reap the new meaning. Divergencies which leave the sense wholly unaffected may be to him very substantial errors. It is even possible that he may find a copy painfully corrupt, from which, nevertheless, precisely the snme sense flows as if it had been written with perfect accuracy. It is of the deepest interest, nevertheless, to inquire, even with this purely textual meaning, how much correction the texts of the New Testament in general circulation need before they are restored substantially to their original form. The reply will necessarily vary according to the standard of comparison which we assume. If we take an ordinarily well printed modern book as a standard, the New Testament, in its commonly current text, will appear sorely corrupt. This is due to the different conditions under which an ancient and a modern book come before a modern audience. The repeated proof-correcting by expert readers and author alike in a modern printing-office, as preliminary to the issue of a single copy; the ability to issue thousands of identical copies from the same plates; the opportunities given to correct the plates for new issues, so that each new issue is sure to be an improvement on the last: all this conspires to the attainment of a very high degree of accuracy. But in ancient times each copy was slowly and painfully made, independently of all others; each copy necessarily introduced its own special errors besides repeating those of its predecessor; each fresh copy that was called for, instead of being struck off from the old and now newly corrected plates, was made laboriously and erroneously from a previous one, perpetuating its errors, old and new, and introducing still newer ones of its own manufacture. A long line of ancestry gradually grows up behind each copy in such circumstances, and the race gradually but inevitably degenerates, until, after a thousand years or so, the number of fixed errors becomes considerable. When at last the printing press is invented, and the work put through it, not the author's autograph, but the latest manuscript is printer's copy, and no author's eye can overlook the sheets. The best the press can do is measurably to stop the growth of corruption and faithfully to perpetuate all that has already grown. No wonder that the current New Testament text must bo adjudged, in comparison with a well printed modern book, extremely corrupt.

On the other hand, if we compare the present state of the New Testament text with that of any other ancient writing, we must render the opposite verdict, and declare it to be marvellously correct. Such has been the care with which the New Testament has been copied,—a care which has doubtless grown out of true reverence for its holy words,—such has been the providence of God in preserving for His Church in each and every age a competently exact text of the Scriptures, that not only is the New Testament unrivalled among ancient writings in the purity of its text as actually transmitted and kept in use, but also in the abundance of testimony which has come down to us for castigating its comparatively infrequent blemishes. The divergence of its current text from the autograph may shock a modern printer of modern books; its wonderful approximation to its autograph is the undisguised envy of every modern reader of ancient books.

When we attempt to state the amount of corruption which the New Testament has suffered in its transmission through two millenniums, absolutely Instead of thus relatively, we reach scarcely more intelligible results. Roughly speaking, there have been counted in it some hundred and eighty or two hundred thousand "various readings "—that is, actual variations of reading in existing documents. These are, of course, the result of corruption, and hence the measure of corruption. But we must guard against being misled by this very misleading statement. It is not meant that there are nearly two hundred thousand places in the New Testament where various readings occur; but only that there are nearly two hundred thousand various readings all told; and in many cases the documents so differ among themselves that many are counted on a single word. For each document is compared in turn with the one standard, and the number of its divergences ascertained; then these sums are themselves added together, and the result given as the number of actually observed variations. It is obvious that each place where a variation occurs is counted as many times over, not only as distinct variations occur upon it, but also as the same variation occurs in different manuscripts. This sum includes, moreover, all variations of all kinds and in all sources, even those that are singular to a single document of infinitesimal weight as a witness, and even those that affect such very minor matters as the spelling of a word. Dr. Ezra Abbot was accustomed to say that about nineteen-twentieths of them have so little support that, although they are various readings, no one would think of them as rival readings; and nineteen-twentieths of the remainder are of so little importance that their adoption or rejection would cause no appreciable difference in the sense of the passages where they occur. Dr. Hort's way of stating it is that upon about one word in every eight various readings exist supported by sufficient evidence to bid us pause and look at it; that about one word in sixty has various readings upon it supported by such evidence as to render our decision nice and difficult; but that so many of these variations are trivial that only about one word in every thousand has upon it substantial variation supported by such evidence as to call out the efforts of the critic in deciding between the readings.

The great mass of the New Testament, in other words, has been transmitted to us with no, or next to no, variation; and even in the most corrupt form in which it has ever appeared, to use the oft-quoted words of Richard Bentley, "the real text of the sacred writers is competently exact; . . . nor is one article of faith or moral precept either perverted or lost . . . choose as awkwardly as you will, choose the worst by design, out of the whole lump of readings." If, then, we undertake the textual criticism of the New Testament under a sense of duty, we may bring it to a conclusion under the inspiration of hope. The autographic text of the New Testament is distinctly within the reach of criticism in so immensely the greater part of the volume, that we cannot despair of restoring to ourselves aud the Church of God, His Book, word for word, aa He g»,ve it by inspiration tc men.

The following pages are intended as a primary guide to students making their first acquaintance with the art of textual criticism as applied to the New Testament. Their purpose will be subserved if they enable them to make a beginning, and to enter into the study of the text-books on the subject with ease and comfort to themselves.