Chapter VIII. Internal Evidence.
We have now described, in some detail, the several species of external testimony available for the textual criticism of the New Testament, whether comprising manuscripts of the original Greek, or ancient translations from it, or citations from Scripture made by ecclesiastical writers. We have, moreover, indicated the chief editions wherein all these materials are recorded for our use, and the principles that have guided their several editors in applying them to the revision of the text. One source of information, formerly deemed quite legitimate, has been designedly passed by. It is now agreed among competent judges that Conjectural Emendation must never be resorted to, even in passages of acknowledged difficulty[243]; the absence of proof that a reading proposed to be substituted for the common one is actually supported by some trustworthy document being of itself a fatal objection to our receiving it[244]. Those that have [pg 245] been hazarded aforetime by celebrated scholars, when but few codices were known or actually collated, have seldom, very seldom, been confirmed by subsequent researches: and the time has now fully come when, in the possession of abundant stores of variations collected from memorials of almost every age and country, we are fully authorized in believing that the reading to which no manuscript, or old version, or primitive Father has borne witness, however plausible and (for some purposes) convenient, cannot safely be accepted as genuine or even as probable; even though there may still remain a few passages respecting which we cannot help framing a shrewd suspicion that the original reading differed from any form in which they are now presented to us[245].
In no wise less dangerous than bare conjecture destitute of external evidence, is the device of Lachmann for unsettling by means of emendation (emendando), without reference to the balance of conflicting testimony, the very text he had previously fixed by revision (recensendo) through the means of critical authorities: in fact the earlier process is but so much trouble misemployed, if its results are liable to be put aside by [pg 246] abstract judgement or individual prejudices. Not that the most sober and cautious critic would disparage the fair use of internal evidence, or withhold their proper influence from those reasonable considerations which in practice cannot, and in speculation should not, be shut out from every subject on which the mind seeks to form an intelligent opinion. Whether we will or not, we unconsciously and almost instinctively adopt that one of two opposite statements, in themselves pretty equally attested to, which we judge the better suited to recognized phenomena, and to the common course of things. I know of no person who has affected to construct a text of the N. T. on diplomatic grounds exclusively, without paying some regard to the character of the sense produced; nor, were the experiment tried, would any one find it easy to dispense with discretion and the dictates of good sense: nature would prove too strong for the dogmas of a wayward theory. “It is difficult not to indulge in subjectiveness, at least in some measure,” writes Dr. Tregelles (Account of Printed Text, p. 109): and, thus qualified, we may add that it is one of those difficulties a sane man would not wish to overcome.
The foregoing remarks may tend to explain the broad distinction between mere conjectural emendation, which must be utterly discarded, and that just use of internal testimony which he is the best critic who most judiciously employs. They so far resemble each other, as they are both products of the reasoning faculty exercising itself on the sacred words of Scripture: they differ in this essential feature, that the one proceeds in ignorance or disregard of evidence from without, while the office of the other has no place unless where external evidence is evenly, or at any rate not very unevenly, balanced. What degree of preponderance in favour of one out of several readings, all of them affording some tolerable sense, shall entitle it to reception as a matter of right; to what extent canons of subjective criticism may be allowed to eke out the scantiness of documentary authority; are points that cannot well be defined with strict accuracy. Men's decisions respecting them will always vary according to their temperament and intellectual habits; the judgement of the same person (the rather if he be by constitution a little unstable) will fluctuate from time to time as to the same evidence brought to bear on the self-same [pg 247] passage. Though the canons or rules of internal testimony be themselves grounded either on principles of common sense, or on certain peculiarities which all may mark in the documents from which our direct proofs are derived; yet has it been found by experience (what indeed we might have looked for beforehand), that in spite, perhaps in consequence, of their extreme simplicity, the application of these canons has proved a searching test of the tact, the sagacity, and the judicial acumen of all that handle them. For the other functions of an editor accuracy and learning, diligence and zeal are sufficient: but the delicate adjustment of conflicting probabilities calls for no mean exercise of a critical genius. This innate faculty we lack in Wetstein, and notably in Scholz; it was highly developed in Mill and Bengel, and still more in Griesbach. His well-known power in this respect is the main cause of our deep regret for the failure of Bentley's projected work, with all its faults whether of plan or execution.
Nearly all the following rules of internal evidence, being founded in the nature of things, are alike applicable to all subjects of literary investigation, though their general principles may need some modification in the particular instance of the Greek Testament.
I. Proclivi Scriptioni praestat ardua: the more difficult the reading the more likely it is to be genuine. It would seem more probable that the copyist tried to explain an obscure passage, or to relieve a hard construction, than to make that perplexed which before was easy: thus in John vii. 39, Lachmann's addition of δεδομένον to οὔπω ἦν πνεῦμα ἅγιον is very improbable, though countenanced by Cod. B and (of course) by several of the chief versions. We have here Bengel's prime canon, and although Wetstein questioned it (N. T., vol. i. Proleg. p. 157), he was himself ultimately obliged to lay down something nearly to the same effect[246]. Yet this excellent rule may easily [pg 248] be applied on a wrong occasion, and is only true ceteris paribus, where manuscripts or versions lend strong support to the harder form. “To force readings into the text merely because they are difficult, is to adulterate the divine text with human alloy; it is to obtrude upon the reader of Scripture the solecisms of faltering copyists, in the place of the word of God” (Bp. Chr. Wordsworth, N. T., vol. i. Preface, p. xii)[247]. See Chap. [XII] on Matt. xxi. 28-31. Compare also above, Vol. I. i. § 11.
II. That reading out of several is preferable, from which all the rest may have been derived, although it could not be derived from any of them. Tischendorf (N. T., Proleg. p. xlii. 7th edition) might well say that this would be “omnium regularum principium,” if its application were less precarious. Of his own two examples the former is too weakly vouched for to be listened to, save by way of illustration. In Matt. xxiv. 38 he[248] and Alford would simply read ἐν ταῖς ἡμέραις τοῦ κατακλυσμοῦ on the very feeble evidence of Cod. L, one uncial Evst. (13), a e ff, the Sahidic version, and Origen (in two places); because the copyists, knowing that the eating and drinking and marrying took place not in the days of the flood, but before them (καὶ οὐκ ἔγνωσαν ἕως ἦλθεν ὁ κατακλυσμός ver. 39), would strive to evade the difficulty, such as it was, by adopting one of the several forms found in our copies: ἡμέραις πρὸ τοῦ κατακλ., [pg 249] or ἡμέραις ταῖς πρὸ τοῦ κατακλ., or ἡμέραις ἐκείναις πρὸ τοῦ κατακλ., or ἡμέραις ἐκείναις ταῖς πρὸ τοῦ κατακλ., or even ἡμέραις τοῦ νῶε. In his second example Tischendorf is more fortunate, unless indeed we choose to refer it rather to Bengel's canon. James iii. 12 certainly ought to run μὴ δύναται, ἀδελφοί μου, συκῆ ἐλαίας ποιῆσαι, ἢ ἄμπελος σῦκα; οὔτε (vel οὐδὲ) ἁλυκὸν γλυκὺ ποιῆσαι ὕδωρ, as in Codd. אABC, in not less than six good cursives, the Vulgate and other versions. To soften the ruggedness of this construction, some copies prefixed οὅτως to οὔτε or οὔδε, while others inserted the whole clause οὕτως οὐδεμία πηγὴ ἁλυκὸν καί before γλυκὺ ποιῆσαι ὕδωρ. Other fair instances may be seen in Chap. [XII], notes on Luke x. 41, 42; Col. ii. 2[249]. In the Septuagint also the reading of א συνεισελθόντας 1 Macc. xii. 48 appears to be the origin both of συνελθόντας with A, the uncial 23, and four cursives at least, and of εἰσελθόντας of the Roman edition and the mass of cursives.
III. “Brevior lectio, nisi testium vetustorum et gravium auctoritate penitus destituatur, praeferenda est verbosiori. Librarii enim multò proniores ad addendum fuerunt, quam ad omittendum” (Griesbach, N. T., Proleg. p. lxiv. vol. i). This canon bears an influential part in the system of Griesbach and his successors, and by the aid of Cod. B and a few others, has brought great changes into the text as approved by some critics. Dr. Green too (Course of Developed Criticism on Text of N. T.) sometimes carries it to excess in his desire to remove what he considers accretions. It is so far true, that scribes were no doubt prone to receive marginal notes into the text which they were originally designed only to explain or enforce (e.g. [pg 250] 1 John v. 7, 8)[250]; or sought to amplify a brief account from a fuller narrative of the same event found elsewhere, whether in the same book (e.g. Act. ix. 5 compared with ch. xxvi. 14), or in the parallel passage of one of the other synoptical Gospels. In quotations, also, from the Old Testament the shorter form is always the more probably correct (ibid.). Circumstances too will be supplied which were deemed essential for the preservation of historical truth (e.g. Act. viii. 37), or names of persons and places may be inserted from the Lectionaries: and to this head we must refer the graver and more deliberate interpolations so frequently met with in Cod. D and a few other documents. Yet it is just as true that words and clauses are sometimes wilfully omitted for the sake of removing apparent difficulties (e.g. υἱοῦ βαραχίου, Matt, xxiii. 35 in Cod. א and a few others), and that the negligent loss of whole passages through ὁμοιοτέλευτον is common to manuscripts of every age and character. On the whole, therefore, the indiscriminate rejection of portions of the text regarded as supplementary, on the evidence of but a few authorities, must be viewed with considerable distrust and suspicion.
IV. That reading of a passage is preferable which best suits the peculiar style, manner, and habits of thought of an author; it being the tendency of copyists to overlook the idiosyncrasies of the writer. For example, the abrupt energy of St. James' asyndeta (e.g. ch. i. 27), of which we have just seen a marked instance, is much concealed by the particles inserted by the common text (e.g. ch. ii. 4, 13; iii. 17; iv. 2; v. 6): St. Luke in the Acts is fond of omitting “said” or “saith” after the word indicating the speaker, though they are duly supplied by recent scribes (e.g. ch. ii. 38; ix. 5; xix. 2; xxv. 22; xxv. 28, 29). Thus again, in editing Herodotus, an Ionic form is more eligible than an Attic one equally well attested, while in the Greek Testament an Alexandrian termination should be chosen under similar circumstances. Yet even this canon has a double edge: habit or the love of critical correction will sometimes lead [pg 251] the scribe to change the text to his author's more usual style, as well as to depart from it through inadvertence (see Acts iv. 17; 1 Pet. ii. 24): so that we may securely apply the rule only where the external evidence is not unequally balanced.
V. Attention must be paid to the genius and usage of each several authority, in assigning the weight due to it in a particular instance. Thus the testimony of Cod. B is of the less influence in omissions, that of Cod. D (Bezae) in additions, inasmuch as the tendency of the former is to abridge, that of the latter to amplify the sacred text. The value of versions and ecclesiastical writers also much depends on the degree of care and critical skill which they display.
Every one of the foregoing rules might be applied mutatis mutandis to the emendation of the text of any author whose works have suffered alteration since they left his hands: the next (so far as it is true) is peculiar to the case of Holy Scripture.
VI. “Inter plures unius loci lectiones ea pro suspectâ merito habetur, quae orthodoxorum dogmatibus manifestè prae ceteris favet” (Griesbach, N. T., Proleg., p. lxvi. vol. i). I cite this canon from Griesbach for the sake of annexing Archbishop Magee's very pertinent corollary: “from which, at least, it is reasonable to infer, that whatever readings, in favour of the Orthodox opinion, may have had his sanction, have not been preferred by him from any bias in behalf of Orthodoxy” (Discourses on Atonement and Sacrifice, vol. iii. p. 212). Alford says that the rule, “sound in the main,” does not hold good, when, “whichever reading is adopted, the orthodox meaning is legitimate, but the adoption of the stronger orthodox reading is absolutely incompatible with the heretical meaning,—then it is probable that such stronger orthodox reading was the original” (N. T., Proleg., vol. i. p. 83, note 6, 4th edition): instancing Act. xx. 28, where the weaker reading τὴν ἐκκλησίαν τοῦ κυρίου would quite satisfy the orthodox, while the alternative reading τοῦ θεοῦ “would have been certain to be altered by the heretics.” But in truth there seems no good ground for believing that the rule is “sound in the main,” though two or three such instances as [pg 252] 1 Tim. iii. 16[251] and the insertion of θεόν in Jude, ver. 4, might seem to countenance it. We dissent altogether from Griesbach's statement, “Scimus enim, lectiones quascunque, etiam manifestò falsas, dummodo orthodoxorum placitis patrocinarentur, inde a tertii seculi initiis mordicus defensas seduloque propagatas, ceteras autem ejusdem loci lectiones, quae dogmati ecclesiastico nil praesidii afferrent, haereticorum perfidiae attributas temere fuisse” (Griesb. ubi supra), if he means that the orthodox forged those great texts, which, believing them to be authentic, it was surely innocent and even incumbent on them to employ[252]. The Church of Christ “inde a tertii seculi initiis” has had her faults, many and grievous, but she never did nor shall fail in her duty as a faithful “witness and keeper of Holy Writ.” But while vindicating the copyists of Scripture from all wilful tampering with the text, we need not deny that they, like others of their craft, preferred that one out of several extant readings that seemed to give the fullest and most emphatic sense: hence Davidson would fain account for the addition ἐκ τῆς σαρκὸς αὐτοῦ καὶ ἐκ τῶν ὀστέων αὐτοῦ (which, however, is not unlikely to be genuine[253]) in Eph. v. 30. Since the mediaeval scribes belonged almost universally to the monastic orders, we will not dispute the truth of Griesbach's rule, “Lectio prae aliis sensum pietati (praesertim monasticae) alendae aptum fundens, suspecta est,” though its scope is doubtless very limited[254]. Their [pg 253] habit of composing and transcribing Homilies has also been supposed to have led them to give a hortatory form to positive commands or dogmatic statements (see Vol. I. p. 17), but there is much weight in Wordsworth's remark, that “such suppositions as these have a tendency to destroy the credit of the ancient MSS.; and if such surmises were true, those MSS. would hardly be worth the pains of collating them” (note on 1 Cor. xv. 49).
VII. “Apparent probabilities of erroneous transcription, permutation of letters, itacism and so forth,” have been designated by Bp. Ellicott “paradiplomatic evidence” (Preface to the Galatians; p. xvii, first edition), as distinguished from the “diplomatic” testimony of codices, versions, &c. This species of evidence, which can hardly be deemed internal, must have considerable influence in numerous cases, and will be used the most skilfully by such as have considerable practical acquaintance with the rough materials of criticism. We have anticipated what can be laid before inexperienced readers on this topic in the first chapter of our first volume, when discussing the sources of various readings[255]: in fact, so far as canons of internal [pg 254] or of paradiplomatic evidence are at all trustworthy, they instruct us in the reverse process to that aimed at in Vol. I. Chap. I; the latter showing by what means the pure text of the inspired writings was brought into its present state of partial corruption, the former promising us some guidance while we seek to retrace its once downward course back to the fountain-head of primeval truth[256]. To what has been previously stated in regard to paradiplomatic testimony it may possibly be worth while to add Griesbach's caution “lectiones rhythmi fallaciâ facillimè explicandae nullius sunt pretii” (N. T., Proleg. p. lxvi), a fact whereof 2 Cor. iii. 3 affords a memorable example. Here what once seemed the wholly unnatural reading ἐν πλαξὶ καρδίαις σαρκίναις, being disparaged by dint of the rhyming termination, is received by Lachmann in the place of καρδίας, on the authority of Codd. AB (sic) CDEGLP, perhaps a majority of cursive copies (seven out of Scrivener's twelve, and Wake 12 or Paul. 277); to which add Cod. א unknown to Lachmann, and that abject slave of manuscripts, the Harkleian Syriac. Codd. FK have καρδίας, with all the other versions. If we attempt to interpret καρδίαις, we must either render with Alford, in spite of the order of the Greek, “on fleshy tables, [your] hearts:” or with the Revisers of 1881 “in tables that are hearts of flesh;” yet surely σαρκίναις as well as λιθίναις must agree with πλαξί. Dr. Hort in mere despair would almost reject the second πλαξί (Introd., Notes, p. 119).
It has been said that “when the cause of a various reading is known, the variation usually disappears[257].” This language may seem extravagant, yet it hardly exaggerates what may be effected by internal evidence, when it is clear, simple, and unambiguous. It is, therefore, much to be lamented that this is seldom the case in practice. Readings that we should uphold in virtue of one canon, are very frequently (perhaps in a majority of really doubtful passages) brought into suspicion by means of [pg 255] another; yet they shall each of them be perfectly sound and reasonable in their proper sphere. An instance in point is Matt. v. 22, where the external evidence is divided. Codd. אΒ (in Δ secundâ manu), 48, 198, 583, 587, Origen twice, the Ethiopic and Vulgate, omit εἰκῆ after πᾶς ὁ ὀργιζόμενος τῷ ἀδελφῷ ἀυτοῦ, Jerome fairly stating that it is “in quibusdam codicibus,” not “in veris,” which may be supposed to be Origen's MSS., and therefore removing it from his revised Latin version. It is found, however, in all other extant copies (including ΣDEKLMSUVΔ (primâ manu) Π, most cursives, all the Syriac (the Peshitto inserting, not a Syriac equivalent, but the Greek word εἰκῆ) and Old Latin copies, the Bohairic, Armenian, and Gothic versions), in Eusebius, in many Greek Fathers, in the Latin Fathers from Irenaeus downwards[258], and even in the Old Latin Version of Origen himself; the later authorities uniting with Codd. ΣD and their associates against the two oldest manuscripts extant. Under such circumstances the suggestions of internal evidence would be precious indeed, were not that just as equivocal as diplomatic proof. “Griesbach and Meyer,” says Dean Alford, “hold it to have been expunged from motives of moral rigorism:—De Wette to have been inserted to soften the apparent rigour of the precept[259].” Our sixth Canon is here opposed to our first[260]. The important yet precarious and strictly auxiliary nature of rules of internal evidence will not now escape the attentive student; he may find them exemplified very slightly and imperfectly in the twelfth Chapter of this volume, but more fully by recent critical editors of the Greek Testament; except perhaps by Tregelles, who usually passes them by in silence, though to [pg 256] some extent they influence his decisions; by Lachmann, in the formation of whose provisional text they have had no share; and by Dean Burgon, who held that “we must resolutely maintain, that External Evidence must after all be our best, our only safe guide” (The Revision Revised, p. 19)[261]. We will close this investigation by citing a few of those crisp little periods (conceived in the same spirit as our own remarks) wherewith Davidson is wont to inform and sometimes perhaps to amuse his admirers:
“Readings must be judged on internal grounds. One can hardly avoid doing so. It is natural and almost unavoidable. It must be admitted indeed that the choice of readings on internal evidence is liable to abuse. Arbitrary caprice may characterize it. It may degenerate into simple subjectivity. But though the temptation to misapply it be great, it must not be laid aside.... While allowing superior weight to the external sources of evidence, we feel the pressing necessity of the subjective. Here, as in other instances, the objective and subjective should accompany and modify one another. They cannot be rightly separated.” (Biblical Criticism, vol. ii. p. 374, 1852.)