UPDATE. Thanks to Dr. Railton for identifying two errors in my lists. In the Lynn lists, for volume II, pg. 24, ll. 24, the Jewett edition has “houses” with an “s”. Same volume, page 106, ll. 24, the ellipses in Jewett has five dots, not six as in my list. I will be correcting and updating the attachments. Also, according to his careful review, my original estimate that “20 percent of the errors remain,” which I based on a sample of the first 10 that I checked, may be too high. He says that from my new list he found 50 or 60 errors that remained in the UTC and Am Cul. text. They have now been corrected. It is a pleasure to report that the electronic text on UTC and Am Cul. has improved remarkably because of Dr. Railton’s careful review of the collation results that I’ve shared with him.
Last week I went to ALA in Boston (conference info here) and gave a talk in a Digital Americanist session. I included a handout with my session, an apparatus that details the textual transmission of the Jewett edition of Uncle Tom’s Cabin.
These are the highlights.
- After Jewett printed first 5,000 copies of UTC, the plates were corrected, almost certainly with Stowe’s direction. The corrected imprint was noted by E. Bruce Kirkham, and Michael Winship in the Bibliography of American Literature gave details (which I repeat in the apparatus). No 20th C. (or later) Jewett reprint (that I’ve checked, and my survey was fairly comprehensive) has noted the presence of authorial corrections.
- Kenneth Lynn’s Harvard edition has too many transcription errors for its designation as the “standard” text, even if the label is largely ceremonial. No one cites it anymore.
- Ann Douglas’s Penguin edition (which is based on Lynn’s text) has too many errors for a classroom reading text. Given the frequency of transcription errors and obvious typographical errors, it is unlikely that the text was proofread by the editor.
- The electronic text of UTC first published on Stephen Railton’s Uncle Tom’s Cabin & American Culture site (1998) was derived from the Virginia E-Text version of the Penguin edition (which was scanned and turned into machine-readable text using OCR software).
- The UTC & Am. Cul. Jewett edition text’s inheritance of transcription error from Lynn-Harvard, Douglas-Penguin, and E-Text OCR process was improved by proofreading, but it remained quite faulty.
- I suspect modern reprint editors of texts for school use will increasingly “borrow” electronic texts and proof them for their new edition. The intermingling of print and digital texts will probably become the rule. Proofreading is inadequate to catch the errors. Editors should not rely on that. They need to transcribe and compare their transcription against another text electronically.
- Quite reliable electronic texts are available from Wright American Fiction, 1851-1875 at Indiana University and from Early American Fiction (EAF) at the University of Virginia. EAF transcription is also available through Chadwyck/Healey Literature Online.
- The most reliable print editions in a tested sample (4 pages w/ dialect, 1 percent of text) were Kathryn Kish Sklar’s Library of American edition and Elizabeth Ammons’s Norton edition. Neither had an error in tested sample. Sklar reprints the uncorrected Jewett imprint but does not normalize contractions by closing up spaces. Ammons reprints the corrected Jewett imprint but does close up spaces in contractions.
- It’s hard to track transmission alteration when texts (print and electronic) are derived from one another. Hedrick reprinted Sklar’s text, but she omits at least one of Sklar’s corrections. No way to tell what else changed without extensive work. NetLibrary deep-sixed Sklar’s UTC (at least from subscription site to which I’m eligible) and now features Yellin’s Oxford text. I haven’t checked.
- I used PC-CASE to collate EAF and UTC & Am. Cul. text. I shared the results with my dissertation reader Stephen Railton, and he updated both the Bedford edition (which he had in galley proofs) and the UTC & Am. Cul. site (2006, the current version). I haven’t checked those either.
I also note some of my own errors and points worth reiterating. As a bibliographer kindly pointed out after my talk, I should not have said “Uncorrected States” because second Jewett set of 5000 is a printing. The term state in bibliography is reserved for alterations during a single printing. See Bowers, Principles of Bibliographical Description. I know that. So I’m chastened. Also, the apparatus omits what Sir Walter W. Greg called “accidental” variants, alterations in capitalization, punctuation, and case. If you’re interested in those lists, ask.
I had the honor of being on a panel with Amanda Gailey and Lisa Spiro. Dr. Gailey gave a fascinating talk on the inadequacy of search for dialect speech or metaphorical referents. Her examples were drawn from Whitman and Joel Chandler Harris. Dr. Spiro surveyed Americanist users of electronic archives (Whitman, Dickinson, and UTC & Am. Cul.) to assess scholars’ attitudes toward issues of authority, importance, and citation. See survey results. By accident or the devious planning of session organizers Andy Jewell and Amy Earhart, our talks hung together almost as if we had a concerted plan. From Dr. Spiro we learned that some scholars worry that by citing electronic sources they will be thought lazy (for example, by the tenure committee). Ed Folsom noted that citation forms (e.g. URLs) for electronic texts are ugly in print. That’s true. The former problem is not easily solved as changing professional attitudes is like turning an oil tanker. That latter problem might be addressed with a combination of technical and aesthetic means. Perhaps citation form should be changed to only include URL of front page of extensive resource. Cocoon and its ability to resolve URLs, I would think, could help reduce the ugliness of citation form.
In other news, the Digital Americanist organization was called into the state of official being with parliamentary style voting on constitution and by-laws.