"The fact is that there is only one general principle of emendation, which is that emendation is in its essence devoid of principle" (1).
"Emendation is an art," and while there are no general principles of emendation that can guide great editors, there are some basic rules that will help lesser editors (1). Honestly, all of this double talk just doesn't make any sense to me.
Any editor would do well to remember two facts: even careful authors do not always write the most sensible phrases, and sometimes it will be impossible to trace the agency of corruption in a particular text (4).
"When we have satisfied ourselves that an emendation is acceptable, the next question we ought to ask is what it implies with respect to the history and origin of the text" (5).
"The central point at which I am aiming is this: that no emendation can, or ought to be, considered in vacuo, but that criticism must always proceed in relation to what we know, or what we surmise, respecting the history of the text" (6). I had initially paraphrased this in the manner that follows, but F.P. Wilson saw fit to quote this passage on p. 97 of Shakespeare and the New Bibliography, so I have included it in my notes after reading it there.
No single emendation can be made without considering the historical context of the work at hand, and when more than one emendation is made, they should not involve mutually contradictory origin theories (6).
When editing plays that exist in only one version, such as the plays that were first printed in F1, beyond the general suitability to the text, there are only two general guides that an editor has at their disposal: a knowledge of the kinds of errors a compositor is likely to make in reading a text, and of the kinds of errors he is likely to make in setting the text (8). Greg loves his binaries.
In cases where there are two texts from the period, one may be taken as a correction of the other, and where the texts have common errors, it is probable that the error stems from a misreading of authorial manuscript. In this cases, it is helpful to consider what interpretation was placed on this corruption, especially by the actors who probably used it (9).
By the same logic, when a reading is specifically preserved in two extant texts, it may be a sign that it is correct, and should be retained (10). Admittedly, this is one of the reasons that I preserved "the nun will soon at night turn lippit" instead of "tippet." Lippit is a nonsense word and a clear misreading, but it is preserved in all six quartos, and tippet is perfectly sensible, if a little obscure to a modern audience, that has performative implications (i.e. Miliscent wearing a monk's robe as part of her escape). Looking back on it, I made the wrong choice, and so will generally disagree with Greg's logic here.
When considering instances in the Folio where the Folio version was apparently set from an earlier quarto, Greg says "where the texts differ, one possesses vastly greater authority than the other: where they agree, we not only have direct transcriptional witness to what the author wrote, but we know, subject to certain exceptions, that this was what was actually spoken on stage" (14). Again, I find this claim to be dubious.
For those texts in which a "good" quarto corrected a "bad" quarto, and then was used to print the Folio, the bad Romeo and Juliet is better than the good text. Greg sees this as an opportunity for "critical exploration" (21).
It's kind of hard for me to honestly evaluate W.W. Greg. He certainly makes a great deal of sense at times, but his subscription to the New Bibliography's narrative colors his view of the evidence, but I am encouraged by his ability to see that the bad quarto of Romeo and Juliet, at least, isn't all that bad. I can't help but think there is more science to emendation than Greg lets on, even if it is a science that is particular to a text, or to a set of texts. As each text needs to be regarded on its own terms, the specifics of emendation are specific to the text(s) being edited, however similar governing principles to that process seem to apply.
Greg, W.W. "Principles of Emendation in Shakespeare." Proceedings of the British Academy. Annual Shakespeare Lecture. Read 23 May 1928. Print.