University of Virginia Library

Search this document 


expand section 
collapse section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 

expand section 



Examples of critical editions of seventeenth-century English poets drawing on extensive collations of manuscript sources are the Oxford English Texts editions of Donne, edited by Helen Gardner and Wesley Milgate, and of Suckling, edited by Thomas Clayton and L. A. Beaurline, and Poems on Affairs of State: Augustan Satirical Verse, 1660-1714, gen. ed. George de Forest Lord, 7 vols. (1963-75). The Complete Poems of John Wilmot Earl of Rochester, ed. David M. Vieth (1968), though compiled on critical principles, is without an apparatus criticus. For a detailed account with supporting evidence of an experienced editor's conclusions about the problems of one such text, see L. A. Beaurline, "An Editorial Experiment: Suckling's A Sessions of the Poets," Studies in Bibliography, 16 (1963), 43-60.


For the circumstances of transmission of classical literature, see L. D. Reynolds and N. G. Wilson, Scribes and Scholars: A Guide to the Transmission of Greek and Latin Literature (2nd ed., 1974) and E. J. Kenny, The Classical Text (1974). Lambertus Okken in his 1970 Rijksuniversiteit te Utrecht doctoral thesis, Ein Beitrag zur Entwirrung einer kontaminierten Manuskripttradition: Studien zur Überlieferung von Hartmanns von Aue "Iwein," p. 8, points out that Karl Lachmann, the founder of modern textual criticism, never in fact published a stemma and inquires "Hatte der grosse Philologe im Umgang mit kontaminierten Traditionen, etwa mit den Überlieferungen von Wolframs von Eschenbach Werken, gelernt, das die Kenntnis der Handschriftengenealogie gewöhnlich keinen praktischen Wert hat?"


An exception to the common dialect would be the University of Edinburgh Library MS. DC. 1.3 of poems by Rochester and his contemporaries, the text of which contains a detectable infusion of Scotticisms.


Much information on this topic will be found scattered through David M. Vieth, Attribution in Restoration Poetry: A Study of Rochester's "Poems" of 1680 (1963). For an exceptionally revealing account of the genesis of a particular group of manuscripts, see W. J. Cameron's "A Late Seventeenth-Century Scriptorium," Renaissance and Modern Studies, 7 (1963), 25-52 and "Transmission of the Texts" in Poems on Affairs of State, V, 528-540.


In his Attribution in Restoration Poetry, his edition, with Bror Danielsson, of The Gyldenstolpe Miscellany of Poems by John Wilmot, Earl of Rochester, and other Restoration Authors (Stockholm: Almqvist and Wiksell, 1967), and his edition of the Complete Poems.


See Complete Poems, pp. xlvi-lii. Vieth prefaces the account of his procedures with the dry remark: "Some aspects of textual criticism raise surprisingly philosophical questions, in this instance whether the universe (not to mention the human mind) is fundamentally rational" (p. xlvii). The genealogical, historical and philological methods are employed to create a "tentative reconstructed text" of the poem. The early text "having the least departures from the tentative text" is then chosen as copy-text and its readings accepted "unless there is substantial reason to substitute a reading from other texts" (pp. l-li). In practice, Vieth appears to have let a number of minor substantive readings stand which are unlikely to have been those of the archetype.


As this enterprise is still in progress, I can not guarantee that my conclusions will not be identical with Vieth's.


See Vinton A. Dearing, Principles and Practice of Textual Analysis (1974), pp. 14-20 and passim. Dearing holds that the establishment of the genealogy of "transmitters" (e.g. manuscripts) is the province of bibliography, not textual analysis which is described as "a logic engine, like a computer" (p. 58). "Carrier" is to be preferred to Dearing's "transmitter" as terminal texts are receivers only and do not transmit.


It is possible in some instances to establish relationships between texts on the basis of accidental features, such as the distribution of variant spellings, or physical features such as line-lengths or the number of lines per page. For a summary of Cameron's important discoveries concerning scribal accidentals, see Poems on Affairs of State, V, 529.


Some of these are spelled out by Humphrey Palmer in The Logic of Gospel Criticism (1968), pp. 93-94. The additional point should perhaps be made that, even in a strictly "bibliographical" stemma, the inferential intermediaries remain inferential, that is to say logical abstractions, and cannot be relied upon to correspond in their assumed readings to any single historically existing manuscript.


The editorial heritage of Lachmann encouraged editors to fuse the two processes of the assessment of direction and the linking-up of the stemma; however editors of English literary texts since the time of Greg have agreed in not proceeding to consider directional evidence until the linkages between texts have been established on a purely distributional, non-directional basis. The present study, while arguing that contextual as well as purely formal arguments should be admissible at every stage of textual reasoning, accepts the validity of the two-stage model of stemma-building. For objections to the Lachmannian method, which is still accepted on the authority of Maas by many present-day classicists and mediaevalists, see Dearing, pp. 5-11, 15-16 and 54-56 and Palmer, pp. 76-80. Palmer concludes (p. 92) that Lachmann's method "is quicker—when it works!" but that it depends "on finding errors certain to the critic and incorrigible by copyists"—a consummation more often devoutly wished than practically experienced. For a defence of the method against its critics, see Kenny, p. 137.


Walter W. Greg (1927). Dearing's much more detailed study refines in many valuable particulars on Greg's techniques, definitions and terminologies but should not be used without reference to Michael Weitzman's review in Vetus Testamentum, 27 (1977), 225-235. I would like to thank Dr. Weitzman for pointing out an error in the reasoning of an earlier version of the present paper.


See Dearing, pp. 215-236 and Dom J. Froger, La Critique des Textes et son Automatisation (Paris: Dunod, 1968). Earlier work by Dearing is discussed briefly in my "Computers and Literary Editing: Achievements and Prospects" in The Computer in Literary and Linguistic Research, ed. R. A. Wisbey (1971), pp. 47-56. Gunter Kochendörfer and Bernd Schirok, Maschinelle Textrekonstruction, Göppinger Arbeiten zur Germanistik nr. 185 (Göppingen: Alfred Kümmerle, 1976) includes a valuable bibliography of work in a number of languages relating to this field (pp. 176-179).


As the aim of textual analysis is to restore the readings of the archetype, formal systems and algorithms claiming to do this should be tested on texts chosen at random from a tradition created by supervised copying in as close as devisable an approximation to the conditions experienced by professional scribes of the past and with the solution withheld from the experimenter until after he has submitted his conclusions. If such a system can not produce a correct solution to texts whose descent can be verified, it can hardly be trusted to produce correct results for historical traditions where the conclusions can not be verified.


In the Lachmannian tradition, the "favoured few" are readings regarded as having clear directional implications but not, in the judgment of the editor, being readily corrigible by copyists of the period concerned. These are in turn divided into (1) separative errors which indicate that the manuscript containing them can not be the ancestor of one in which they are correct (cf. Maas, pp. 42-47) and conjunctive errors defined by Palmer as "common to two manuscripts but not a third, and such that the reading of the third could not be due to correction by the copyist" (p. 243).


A summary of the more common kinds of scribal error will be found in Reynolds and Wilson, pp. 200-212. For a more systematic treatment, see James Willis, Latin Textual Criticism (1972), pp. 47-188. Weitzman (p. 226) draws attention to important material to be found in J. Stoll, "Zur Psychologie der Schreibfehler," Fortschritte der Psychologie, 2 (1913), pp. 1-133.


Cameron's experience of this kind of variation in his "scriptorium" manuscripts led him to suggest that "all texts are in fact conflated texts—a conflation of the exemplar and a structure of linguistic expectations that is present in the mind of the scribe" (Poems on Affairs of State, V, 529).


By a "whole group" I understand a whole "true" group as defined below. Greg's use of "true" to describe variational groups containing more than one member seems to proceed from nothing more than an aesthetic disinclination to speak of a "group" with only one member (cf. Dearing, p. 10) and should be disregarded.


Cf. Maas, p. 8: "obvious corruptions, particularly lacunae, may easily be transmitted in the direct line but are hardly ever transferred by contamination." Against this, however, must be placed Greg's principle (p. 20 n.) that "the easier it is to explain how an error arose, the less valid the assumption that it only arose once."


Assuming, of course, that the distribution of a variant among surviving sources corresponds in a general way with the distribution of the variant in the whole body of sources available at the time the scribe made the copy. This may not always have been the case: a Bowdlerised version of a satire may, for instance, have stood a better chance of survival than an obscene one. If a text was available in printed form, each copy in circulation would need to be counted as a separate source for purposes of such a calculation.


The editor should consider the possibility whether the overlapping of two or more possibly incomplete groups defined by gross and easily corrigible errors might not indicate a fractured whole group.


Dearing, p. 57, revising Greg's terminology. Constantine Kooznetzoff, on the other hand, though he does not assert this as a general principle, finds in his "A Genealogical Analysis of the 'Tristan' Fragment, Ms. 2280," AUMLA, 54 (November 1980), 194, that the type-2 variants "providing they are consistent, afford all the evidence necessary for deducing manuscript relationships." The complex variants are declared "genetically non-evidential" on the grounds that they contain "conscious scribal emendation" (p. 197).


A test for obtrusive readings in a short poem would be to find a pretext to make someone transcribe it several times, and then, after a few days interval, ask him to write down as much as he could still remember. The readings correctly given would be the "obtrusive" ones.


The only at all likely situations where such a variant would not indicate a true group would be (1) when it had been called on to fill a lacuna in a majority-group manuscript by a scribe with too imperfect a knowledge of the language or content of what he was copying to realise its uselessness or who felt that any reading was better than none (2) when the identical gross error was made during separate copyings by the same scribe (a quite conceivable happening in a scriptorium situation such as that described by Cameron) or by separate scribes (rather less likely unless some aspect of the original reading or a particular exemplar had created an exceptionally high risk of error).


These points are discussed in greater detail in my "The Text of 'Timon. A Satyr'," Bulletin of the Bibliographical Society of Australia and New Zealand, 6 (1982), 113-140.


For an example of the application of this technique, see my "The Texts of Southerne's The Spartan Dame," Bulletin of the Bibliographical Society of Australia and New Zealand, 1 (1970-1), 54-59.


Dearing, p. 19.


Review of Dearing's Manual of Textual Analysis, JEGP, 59 (1960), 555. Housman in "The Application of Thought to Textual Criticism" sees the subject matter of the discipline as "not rigid and constant, like lines and numbers, but fluid and variable; namely the frailties and aberrations of the human mind, and of its insubordinate servants, the human fingers" (Selected Prose, ed. John Carter [1961], p. 132.) Dearing, p. ix contains a brief retort to Housman.


Cf. Greg, p. 13.


Dearing warns rightly (p. 55) that if the textual analyst "lets the form of his family tree influence his analysis of the directional variation, he reasons in a circle."


Palmer, pp. 51-52, 80. The problems of circular reasoning in textual criticism are similar in their nature to that of the "hermeneutic circle" in critical interpretation as discussed by a number of contributors to the Autumn 1978 issue (X. i) of New Literary History. Palmer's metaphor with its attractively Yeatsian overtones hardly offers a procedural solution, but indicates that the methodological plight of historical scholarship is not as desperate as it is sometimes made out to be.


Henri Quentin, Essais de Critique Textuelle (Paris, 1926), p. 43, quoted in translation by Palmer, p. 80.