“A method of inquiry which provides no means for the rectification of its first principles, has mixed and swallowed its own poison and has to expect an inevitable doom.”
The methods that govern the production of a critical edition are meant to avoid dooming the writings of the author it is devoted to. What makes an edition critical is the special status given to the texts that will eventually produce that edition—for indeed it is the texts that ultimately produce the edition, in the same way as objects determine the signs that stand for them to future interpretations. Those texts are taken to be dynamic objects of inquiry, for three reasons. First, they contain the inscriptions of, and as such constitute, processes of inquiry that animated their author during composition. Second, as artifacts of that process no longer mediated by their author, they each present themselves as enigmas or puzzles begging for attention, study, and continued sense-making. Third, as repositories of hypotheses and theories, they remain permanent contemporary agents of collaborative research.
As objects offering themselves to the inquiry of critical editors viewed as mediators, Peirce’s texts, published or unpublished, raise multiple questions as to their organization, chronological order, compositional order, genealogical history, inscription history, physical condition, archival source, thematic range, developmental stage, intellectual purpose, connection to other texts, extent of their surviving individual integrity and unity, intelligibility, decipherability, definitiveness, and publishability. Every text manifests to various degrees signs or clues that when carefully observed and compared allow editors to formulate tentative answers to many of those questions. Those tentative answers are themselves researched, tested against patiently collected evidence, and submitted to the criticisms of other editors and scholars. Paramount is a double concern.
The first concern is to capture the author’s intentions (as manifested through demonstrable habits of composition, syntactical and stylistic patterns, collateral statements, documentary comparisons, explicit instructions, analogous situations) with optimal exactitude in order to reproduce the expression of Peirce’s thoughts as a faithful replica of that cluster of intentions.
The second concern is to enable readers—the indefinite constellation of present and future interpreters, all consumed by disparate interests—to meet the author as though the latter were still pacing his study at Arisbe, and having with him a conversation, “a wonderfully perfect kind of sign-functioning” (EP2: 391, 1906) in which “the two minds in communication are, in so far ‘at one’” (ibid., 389). It is a conversation within which the third party, the team of critical editors, becomes unnoticeable, even invisible, within the limits of their combined fallibility.
What are the methods applied at the Peirce Edition Project to meet those two goals? This portion of the website is dedicated to the exposition of those methods. The menu on the left lets you navigate among distinct components of our activities, from manuscript organization to transcription to editing to layout. The essentials of our methods are there described, sometimes in considerable detail. The submenu, “Pragmatic Principles,” is a slide presentation that exhibits how over the years our Edition came to consciously embody the principles entailed by Peirce’s pragmatic maxim. For our critical methods, as a whole, are meant to be pragmatistic, a guarantee of their ability to bring Peirce’s texts within the horizon of what he called “concrete reasonableness.”
“Nowadays methods alone can arrest attention strongly; and these are coming in such flocks that the next step will surely be to find a method of discovering methods. This can only come from a theory of the method of discovery. In order to cover every possibility, this should be founded on a general doctrine of methods of attaining purposes, in general; and this, in turn, should spring from a still more general doctrine of the nature of teleological action, in general.”
“The mathematician seeks the speediest and most abridged of secure methods; the logician wishes to make each smallest step of the process stand out distinctly, so that its nature may be understood.”
A flowchart below shows the 84 steps that it takes to produce a given print volume of the Writings of Charles S. Peirce. The workflow comprises six distinct stages with a variable number of steps, many of which are reiterated for each individual text selection. Those six stages are summarized as follows.
Stage 1 includes eight steps that are crucial for volume production in a way that sets the Project apart from other editions, due to the fact that we have been left with a mass of disassembled manuscript pages, tens of thousands of which are loosely grouped, unnumbered, and undated. Manuscript sets need first to be identified and have their pages retrieved and sequenced, using all available clues. Reconstruction reports are generated, detailed records are entered in the database, dating rationales are formulated, and documents are ordered chronologically. Text genealogies are prepared, and drafts or variants are collated to confirm compositional sequence. Copy-texts are selected and annotated with transcribing instructions. “Names & Quotes” research begins.
Stage 2 (fifteen steps) is geared toward production of an exact transcription of the copy-text. The text is transcribed using a customized TEI-conformant tagging system, and Peirce’s revisions are encoded, along with every other textual matter. Transcriptions are team proofread, corrections are entered and checked. Proofreaders travel to holding archives to read the transcriptions against originals. Corrections are then entered and checked, and a third complete proofreading takes place, conducted by a senior editor. Descriptions of Peirce’s alterations are revised, and significant alterations are selected for publication. Corrections and additions are again entered and checked. A final document, the “perfected transcription,” is generated and archived. Two copies are printed: the Review Copy for pre-editing by the general editor, and the Editing Copy for the textual editors.
Stage 3’s eight steps begin with editors marking the Review Copy with suggested emendations. The textual editors draft apparatus headnotes and textual notes, plus the lists of rejected substantives and of end-line hyphenations in the copy-text. Consulting the Review Copy, the textual editors make final editorial decisions, marking up textual emendations and silent regularizations on the Editing Copy. The transcribers then enter all these lists and XML-encode all markings. A printout is generated and team-checked against the Editing Copy. Corrections are entered and team-checked again. The resulting text is then archived twice, as “Emended Copy” and “Annotations Copy.”
Stage 4 consists of sixteen steps involving mostly transcribers. The Emended Copy is reformatted into preliminary layout form designated “Pre-First Pass.” Editors review it individually, and the textual editors reconcile any conflicting layout suggestions before passing the folder to the transcriber for layout adjustments. Corrections are entered until all adjustments are made, and a new printout is generated and stamped “First Pass.” The fourth (team) proofreading ensues, First Pass against Editing Copy, corrections are entered and checked, and the text is archived as “Second Pass page proofs.” Page runs are stabilized and page count is locked in. All selections are bundled into an electronic book file. Folio pagination and running headers are put in place; page make-up is perfected, and the printout gets stamped “Second Pass.” Editors again review the layout, final adjustments are made, and a final printout is generated and stamped “Third Pass.” Line drift across pages stops. A last team proofreading takes place, verifying all emendations, corrections, and adjustments. FPO images of illustrations are tipped into the files. A CSE inspection is scheduled.
Stage 5 consists of 25 steps (most begun in the course of other stages). (5.1) The textual editors write the “Essay on Editorial Methods” and finalize all apparatus sections, whose lists are keyed to Third Pass. Annotations are consolidated; contributing editors’ notes are edited. The chronological catalog of Peirce’s writings, the bibliography, and the short chronology of his life are added to the back matter. (5.2) The introduction is completed and the remaining front matter prepared. (5.3) Front and back matter go through three passes, which are individually and team-proofread. Internal cross-references and references to the Third Pass are keyed in and checked. Front and back matter are then bundled into the book file.
Stage 6 (six steps) begins with the preparation of a comprehensive conceptual index. Laser proofs are printed, the entire volume is inspected, and final corrections are entered. A PDF of the entire book is prepared and sent to the printer. The ultimate set of blueprint proofs will be inspected at Indiana University Press by the editors.
In the chart below, the first abbreviation after each step is generally that of the position chiefly (but not necessarily ultimately) responsible for the completion of that step; succeeding abbreviations indicate positions who share responsibilities or collaborate at various levels, in whole or in part, toward that completion. Green labels denote editors, blue the research and editorial staff. Because of technological changes currently in the works, this flowchart will eventually be adapted to reflect new methodologies.
"Aristotle’s sheets would not be pasted together before the work was quite written, for insertions might be desirable, or even rearrangement. Our collection probably does not contain any finished works. For although it does contain a few which were published during Aristotle's life-time, yet the quoted fragments of them indicate that what we now possess come from MSS in more or less unfinished states. At any rate the great bulk of what we have are either short essays, lecture notes or notes of researches, memoranda of facts, or memoranda of ideas.”
This section describes the methods and procedures for reorganizing the Peirce manuscripts, that is, for reconstituting as closely as possible the order in which 100,000 pages came to be composed, of a corpus that keeps growing as new findings are made almost every year.
Before establishing the text for the critical edition, we need first to reconstitute the documents themselves. This essential task is rendered especially complex by the vast documentary and topical diversity of materials. Peirce used many different kinds of paper, of every dimension, weight, and stock. He usually wrote on rectos, keeping versos for additions, corrections, and calculations, but sometimes using them to write a text unrelated to the recto. There are all sorts of documents: philosophical essays, scientific articles, book reviews, dictionary definitions, scientific reports, galley-proofs, annotated offprints, notebooks, diaries, and many thousands of loose pages, fragments, miscellaneous notes, diagrams, and so forth. The thematic diversity is quite extraordinary, for Peirce could write most ably on nearly every topic: most branches of mathematics, logic, physics, chemistry, geodesy, astronomy, linguistics, the history of all these branches, epistemology, phenomenology, semiotics, metaphysics, and much more. Since Peirce’s writings require such a broad range of expertise, the Project frequently consults with outside specialists. The greater part of the work, however, is done at IU Indianapolis, and requires that researchers learn the rudiments of many disciplines at their turn-of-the-twentieth-century stage of development. Becoming familiar with the vast scientific and humanistic literature Peirce studied is a permanent activity.
Pages of a single document in various states of revision are frequently located in more than one folder, or are arranged in a single folder but not in the correct sequence. Research editors need to learn to navigate Peirce’s large corpus, decipher his handwriting, recognize and memorize how that handwriting transformed itself throughout his life (enough to pinpoint the five-year span within which a document was likely inscribed), and develop an instinct for guessing which part of the corpus is more likely to hold the continuation of a fragmentary text. Such skills take years to hone, and nurturing them is essential. Also necessary is becoming familiar with the vast and complex archives of the Peirce Project, which consist of thousands of research folders that are the legacy of decades of manuscript research. Those folders contain a wealth of information, knowledge of which can save precious investigative time. Learning to navigate them is another important skill in the hunt for missing pages. The persistent possibility that pages are missing makes it impossible for our work to fully escape the shaky ground of hypothesis. In Peirce’s case, pages are frequently missing or misplaced, and much time is spent hunting them in folders containing loose sheets and fragments. Success strikes as often as not, but much has been lost. Some documents are mere fragmented collections of preliminary drafts of a more advanced work that is no longer extant. Other documents are almost complete, except for one or two leaves. And many others are so fragmentary that their purpose or subject matter can barely be identified. To the difficulty of playing with a gigantic jigsaw puzzle is thus added the frustration that thousands of pieces are not included in the box.
Given that Peirce’s manuscripts do not follow a simple linear sequence, the first task is to track and create a map of their genealogical development. Figure 1 below provides an example of such a reconstruction, using a tree model whose multiple branches show graphically how seemingly erratic or unpredictable Peirce’s mental journeys could be. It reveals how Peirce, starting from a preset intention, quickly strayed away from it, his attention captured by some unplanned concept whose investigation required ten or fifty pages of thought. Sooner or later Peirce would grow dissatisfied with part or all of his ongoing argument, set it aside (rarely discarding it), and restart afresh, without even looking at what he had just written. As a result, documents abound with drafts or, more frequently, discrete partial variants (since Peirce seldom recopied himself), each of which requires an independent assessment of its publishability, and each of which may present a puzzle as to its correct placement in the compositional order.
To perform this kind of reconstruction page by page, we use an ingenious and cheap device, the panoptic page display (see Figure 2 below). This simple apparatus offers all the advantages of a laundry line. It is physically comfortable since pages are presented vertically at eye level. Documents can be laid out in piles on a table, but hanging them up page by page on a line allows editors to get a panoptic view of entire manuscripts. Walking left and right along the line, editors first familiarize themselves with the physical appearance of all pages assembled. Through image comparison and contrast they operate a preliminary categorization of each leaf. Then, stepping closer to the display, editors read carefully each page of the document, compare its content with that of other pages, analyze similarities and dissimilarities, and move some of them from one place to another accordingly. We have developed a collation method that compares texts not only word by word, but also sentence by sentence and paragraph by paragraph. Since Peirce generally rewrites his essays without consulting his drafts, paragraph collation has proved effective in establishing compositional order. What gets compared is much less the actual phrasing of ideas than the evolution of how these ideas get treated, and this allows editors to build a solid argument as to why a certain sequence of pages is more likely to reflect actual compositional order than some other.
Figure 1: The MS Organization Tree
Figure 2: The PEP Panopticon
Hardly any text is a sui generis production. Texts are part of a continuum of inquiry, and understanding what Peirce is working out in any document requires the ability to track the genealogy of his ideas in previous related texts. Research editors need therefore to develop an extensive knowledge of the philosophical and scientific content of the manuscripts belonging not only to a particular period of composition, but also to the preceding years and to the immediately succeeding years. This helps situate texts on the timeline of Peirce’s intellectual development, as well as in the daily, weekly, monthly, and yearly calendar of his activities. Such a skill allows fine-tuning one’s guesses about how Peirce was likely to continue a textually incomplete thread of thought, which in turn improves the chances of finding its continuation. It also helps narrow the date range of undated documents.
Since a document cannot be reconstructed if simultaneous attention is not given to the many other documents it may be connected with, keeping track of the content of hundreds of folders would be a hopeless task without databases. The development of several databases to that effect is an ongoing activity; their use and continued growth is an integral part of our methodology. Our chief database, called simply “Manuscript Organization,” is the heart of our operation. Each time we turn to a new document, we first create a record for it in the database, in which will be noted all the data needed for efficient searches, document identification, and document comparison. A variety of information (or metadata) is collected in data fields, and to each kind corresponds particular research activities broken down as described in subsections f to l below.
The reconstitution of any document comes with the recording of its several attributes. They include:
document identification data: numbering or codes attached to it in holding archives, and Peirce’s titles for it;
most recent page count (since pages may get moved in and out at any time);
inscription category (holograph, typescript, amanuensis, offprint; there are dozens of categories);
editorial status: our edition’s volume number within the period of which a document may fall, its publishability status, its genealogical status (e.g., precopytext, copytext, postcopytext, etc.), its pagination status (e.g., fully ordered and complete, not ordered and incomplete, fragmentary).
Recording our evolving findings about the dating of the document is essential: a date is most assured when it is provided by Peirce himself, or when we are able to deduce it from other well dated clues. Sometimes the date can only be guessed at, and is thus entered with an evaluation regarding its probability. All documents are at least assigned a probable year or yearrange for provisional chronological ordering.
An exact description of the stationery is essential, and mastering the specialized terminology is part of the skill, as are identifying and cataloging watermarks and sizes of the different kinds of paper found within any folder. Such information is invaluable for comparing documents. Sometimes the coincidence of watermark and paper size between two documents, one dated and the other not, may be the only evidence we have to assess the probable date of the undated document. Searching for such correlations is a simple affair on a computer if the database has consistently received accurate input.
Also important is identifying the provenance of documents and recording the reconstituted sequence of their pages. Doing this well allows us to keep track of the latest location of each individual page within the corpus. Since one major result of our reorganization is to render the standard Houghton Library order obsolete, we must be able to tell at all times how Houghton folders have been divided and redistributed into new PEP folders. The database shows exactly what pages from what Houghton folders compose any reorganized document. One advantage is to be able to tell whether any particular page has already been identified and moved to some other folder at some point in the past.
In order to cross-reference documents efficiently, a record of all keywords and key phrases needs to be entered page by page in the database. This is broadly understood: we enter all the words and phrases that may prove useful for manuscript comparison and search and retrieval purposes. Accumulating such keywords creates an index to the papers that serves as a powerful document identification tool.
The largest field in the database is intended to contain every pertinent remark and comment that has been made by any editor, past and present, about the corresponding document. That includes the series of evolving hypotheses editors have made about the genealogy of the text, arguments and counterarguments as to what hypothesis proves best, physical descriptions of the document, notes about dating and pagination, identification of quotations and proper names cited by Peirce, connections with other documents, questions to be asked to contributing editors, and directions for future work. The database centralizes all the information that becomes available as research progresses.
To avoid redundant labor, we keep track of the publication history of each document, especially as it took place over the one hundred years following Peirce’s death. A number of posthumous publications were produced by editors working directly from originals; they are the result of a series of hypothetical reconstructions of disassembled documents. Today’s editors need to be in touch with their predecessors, confirming or disconfirming the results of their labor, recapturing their insights and rationales when feasible, studying critically their transcriptions and textual interventions. “Publication history” is the name of a second database, in which we establish cross-references between any portion of a document and its publications in different collections, such as the Collected Papers and the New Elements of Mathematics. Knowing these sources enable us to compare, even confront, our reconstructions with those of others, to make corrections or adaptations when they are called for, and to avoid repeating past errors.
If sequencing and repaginating the manuscripts is complex enough already, the matter of assigning them viable dates can be even more arduous. This section presents the methodology we developed to increase dating accuracy; this is essential because our edition is chronological, unlike any other before it.
Each of our volumes contains a chronological catalog of all known documents that have been determined to fall within the period covered by it. This list includes all of Peirce’s known publications, and also all of his known manuscripts, typescripts, and annotated offprints. Up to the mid 1990s, our policy was to date each and every document. Practical considerations (underfunding and understaffing) forced us to limit our reorganizational effort to those documents that we will definitely publish and to those that are directly connected to the latter. The date assigned to a document varies in precision, depending on the directness and abundance of the evidence. It is directly proportional to the amount and significance of available clues and to the research time that can be invested in it. Dating requires painstaking care and caution. Experience shows that readers will routinely take the editors’ dating, however tentative, for granted, and base on that faith sometimes elaborate arguments about the evolution of the author’s thought, year by year, month by month, and even day by day. Dating is therefore a responsibility that weighs heavily on our editorial shoulders.
Dating requires developing several competences: a detailed working knowledge of the writings the author composed within the five years preceding and following the period under scrutiny; familiarity with the author’s habits of composition and their evolution; familiarity with the changes in handwriting and writing media; and a solid knowledge of the author’s biography. An ongoing requirement is the establishment of a daily, weekly, monthly and yearly calendar of Peirce’s activities, as far as they can be reconstructed from the correspondence and other resources. However fragmentary, more often than not the information we gather allows us to frame a general picture of Peirce’s occupations. By charting the times when we know he was busy, and those for which we have a lesser account of activity, we are better able to formulate hypotheses as to when our author would have been more likely to work on a given project.
To help with this, we have conceived another database designed to keep track of Peirce’s activities: the “Chronological Database.” Recorded in that database are the sources of any relevant information (document, date, origin), and the type of activity or event that took place at the time (usually a day), relative to Peirce. The database also provides fields for excerpts and quotations, editorial remarks about calendar clues, and cross-references with other records. An essential paper resource is Max Fisch’s chronological card catalogue (about 30,000 slips), the content of which is being entered in the database as we go along. The database allows us to summarize a dating argument for each piece we publish and for most pieces we list in the catalog.
There are four types of clues research editors must learn to detect.
They make up the bulk of the information that goes into the chronological database. The correspondence sometimes reveals what Peirce was doing on certain days: composing what paper, going on what trip for what purpose, reading what book, meeting what people, pondering what problem, visiting what friend or relative. Any fact may be relevant, but especially those that can be connected to the act of writing. Not all clues are clear, though, and in fact most are ambiguous.
Sometimes a document can only be dated based on the fact that it shares the same watermark and paper size as some other precisely dated document. The physical appearance of a sheet of paper may indeed be the only available clue. An essential organizational tool we have created is a watermark catalog of the papers. Thousands of leaves have a watermark, some of which occasionally include a manufacturing year. That indicates the earliest possible date of a manuscript, but says nothing about the latest possible date. Other helpful details include paper grain and ink color that can only be ascertained by examining the original document in the holding archives. Other types of physical clues include the writing medium (lead or color pencil, quill or drawing pen, typewriter) and ink intensity; the hand with which Peirce wrote (left or right-handed); the presence of punch holes, lines and graph-ruling; ink blots or water smudges that have run through several sheet layers; crumpling marks, tears and rips, that also tend to repeat on edges of sheets. These marks are usually more useful in the comparative reconstruction of a text than in its dating, however.
Careful reading of a manuscript may reveal mentions of useful dates, like those of recent publications or events. Peirce refers at times to previous work done any stated number of years earlier. Once we identify that publication and check the year of its composition, we add the stated number of years and deduce the actual year. Peirce however does not always provide accurate figures, or his memory may betray him. But at least we may assume that he would not have been too far off, and we are thus permitted to use the clue for what it is worth. All bibliographical references and quotations (direct or in disguise) need to be checked, especially if they refer to recent work, for this may provide closer earliest possible dates. Certain phrases receive special attention because Peirce only used them during a specific period. Any time Peirce outlines a work plan or a table of contents for a projected book, it is compared with the available documents to identify any possible correspondence. What matters most, however, is the research editor’s perception of the author’s evolution of ideas, interests, and purposes. Often enough, only the study of the ideas themselves can lead editors to conclude that a particular essay must, of necessity, have been composed after this one, but before that other.
Its main source is in Peirce’s correspondence. Peirce sometimes talks about the work he has under way in letters to his friends and family. Another source of information resides in the hundreds of book reviews and notices Peirce wrote for popular journals such as the New York Evening Post and the Nation. Indirect sources usually come from secondary correspondence.
For any given text or document, once all the chronological data and clues have been gathered, research editors convene, review the data, and assess the most adequate way of phrasing the compositional date, whether precisely or over a range of time. They also formulate a concise argument that indicates the basis and rationale for dating the document. That argument will accompany the document’s entry in the chronological catalog printed in the volume of the period the document belongs to. Composing that catalog is the final task of the dating work.
“Above the importance of any particular truth, or body of truths, is that of the right methods of reaching the truth.”
Since Peirce’s texts were all about reaching the truth, the above epigraph could be rephrased as follows: “Above the importance of any particular text, or body of texts, is that of the right methods of reaching the text.” The following discusses how we identify and select Peirce’s truth-reaching texts.
It has been roughly estimated that if all of Peirce's writings were to be published, they would fill over 100 volumes. Whatever the actual number is, it is certainly the case that the mass of known documents housed in the Houghton Library and several other repositories (including the National Archives, the Library of Congress, the Smithsonian Institution, Columbia University, and the Morris Library at Southern Illinois University) would fill in dozens more volumes than the planned thirty volumes of the Writings of Charles S. Peirce.
But not everything is publishable or worth publishing. Many factors need to be considered in assessing whether a given text should go through the costly, lengthy, and rigorous critical editorial process entailed by the thorough-going scholarly treatment in place at the Peirce Project—the kind of treatment standardly expected of any principled edition of an author of Peirce’s caliber.
A cursory examination of Richard S. Robin’s Annotated Catalogue of the Papers of Charles S. Peirce reveals the wide documentary diversity: written essays, scientific articles or reports, lectures, field notes, book reviews, book-length projects, dictionary definitions, galley-proofs, annotated offprints, notebooks, diaries, reams of calculation sheets, and many thousands of loose pages, fragments, miscellaneous notes, diagrams, and so forth. The thematic diversity is quite extraordinary, for Peirce could write most ably on topics of many disciplines: fundamental branches of mathematics, logic, physics, chemistry, geodesy, astronomy, linguistics, psychology, the history of science, and many branches of philosophy, including the theory of knowledge, perception theory, phenomenology, semiotics, metaphysics, and much more.
Finding out of such a large and disparate corpus what to publish is no straightforward affair. If there is one aspect that sets apart the Peirce chronological edition from the majority of its peers, it is the ratio between published and unpublished papers. In his lifetime Peirce published about twelve thousand pages (mostly in the form of philosophical and scientific articles, and book reviews), and wrote another one hundred thousand pages (really much more given that there is much that did not survive). About two-thirds of Peirce’s extant writings never went into print by the time of his death.
Some editions deal with authors most of whose works were published during their lifetime, often through several reprints, so that the greater part of the editorial work is taken up with collating various editions of a same text, the latter being already well established. But not so with Peirce: the text of most of his writings needs first to be retrieved from a mass of materials and then carefully established before it can be assessed for publication. Though a prolific writer, Peirce constantly struggled for the precise turn of phrase and was rarely satisfied with his writing, complaining at times how difficult it was for him to translate his thoughts (inherently diagrammatical) into words. This was due in part to his being fundamentally an explorer, a “backwoodsman” as he put it, a pioneer always venturing in uncharted land. Many of the surviving documents evoke the attitude of the speleologist mapping a cave: he will try every path, descend into the deepest pits, and only a cul-de-sac will force him back out and into a new direction. As a result, Peirce’s writings typically present multiple branchings, and the “main body” of his texts resembles much more a shrubby willow than a slender beech tree.
Reconstituting, identifying, assessing, and selecting the texts is therefore challenging. How we go about reorganizing those texts is discussed in the previous tab, “Manuscript Organization.” How do we go about selecting the texts for publication in the Writings?
The short answer is: through multiple acts of comparisons. One concern is to avoid unnecessary redundancies whenever possible. Publishing multiple variant forms of a “same text” would not be wise. In order to establish the most plausible compositional history and genealogy of a text within a cluster of related documents, we collate all the variant forms word by word, or paragraph by paragraph, or section by section, or argument by argument, to detect all the changes, from the terminological to the syntactical to the philosophical or logical, and map them as necessary in order to capture the intertextual development of ideas, evolving expression, or overall increased sophistication of the argumentation. We take note of significant changes and their frequency, of what gets dropped or added from one text to another, each time assessing the possible reasons for the changes. Rare are the cases where Peirce drops a particular thread of thought merely because it was marred by some fatal error. More often Peirce changes his strategy because of a change of interest or emphasis. The fact, therefore, that a particular passage no longer resurfaces in subsequent variants does not imply that the author rejected its tenor.
Editors need therefore to exercise good judgment at all times about what to include or exclude, for the pragmatic reason that not including something may conceivably have negative consequences on future courses of interpretation—an undesirable pragmatistic outcome in light of Peirce’s maxim. There have been several occasions on which we decided to print variant forms of a text (e.g., “The Architecture of Theories,” “The Law of Mind,” Peirce’s two brief rejoinders to Carus’s reply to Peirce’s reply to Carus’s criticism of Peirce’s “Doctrine of Necessity Examined”). Reasons vary: variant texts may significantly differ in content, or in structure, or in topical emphasis; or they may turn out to illuminate one another and thus to be complementary; or they may cover the same ground using such different terms that it is impossible to identify which one was written earlier, as was the case of the two brief rejoinders.
Generally, we do not select disconnected fragments, unreduced scientific data sheets, mathematical and logical scratch pages—anything clearly preliminary, inchoate, or demonstrably rejected by the author himself for evident reasons of dissatisfaction. With holograph and typescript items, the focus is on identifying the sequence of Peirce’s revisions—both across drafts and within single drafts that show evidence of one or more layers of revision. The search for copy-text follows the same rationale used for published items: a mature manuscript form in Peirce’s hand represents the best base document. Where more than one manuscript draft survives, collation will generally lead us to a fairly simple identification of the more mature form. Sometimes, a mature manuscript is followed by a typescript (or an amanuensis draft) prepared under Peirce’s supervision, which contains his holograph corrections and revisions, but varies significantly in both substantives and accidentals from the holograph version. If the two versions are collatable word-for-word, we accept the authority of the manuscript (copy-text) and emend it with those substantive revisions (and corrections to accidentals) that can be attributed to Peirce with a fair degree of confidence. If the typescript is so thoroughly rewritten (which usually indicates a missing intervening draft) that word-by-word collation is not possible, we conclude that Peirce rewrote the document, and use the typescript as copy-text. The surviving earlier stage will be designated a pre-copy-text form and important variant readings will be reproduced or described in editors’ notes.
In fact, situations involving widely diverging forms (that derive from single original drafts) represent the most common copy-text dilemma found in the unpublished materials. Successive manuscript versions may repeat the title or have the same or a similar opening sentence, but will often diverge into a related but new line of inquiry or method of argument. Paragraph or outline collation will usually yield a chronological order, but we may be left with two or more discrete documents, broadly parallel in content but distinct in presentation and development. When parallel versions are equally significant for documenting Peirce’s evolving ideas or scientific findings, we may publish them as discrete items. Each serves as its own copy-text, and will not be emended from subsequent parallel forms; we may refer to other parallel versions for annotation, but not for textual authority.
Certain documents will be immediately identified for inclusion, because they represent Peirce’s most acclaimed work, or display indisputable strength and cogency, or bear testimony to a significant intellectual activity or to a stage of thought indispensable to the framing of a later one. Other texts will be selected if they pass the test of the pragmatic maxim, considering the conceivable effects of not including them. The remaining texts are sent to contributing editors, who are asked to recommend if a given item should be published, could be published but is of lesser importance, or should probably not be published. The recommendations of contributing editors provide the basis for the conclusion of the selection process. The final selection is made by the editors, in consultation with members of the Board of Advisors for difficult cases.
Portions of non selected texts frequently end up being quoted at length in the large set of scholarly annotations that opens up the back matter of edition volumes. We generally excerpt from earlier variants those significant passages that serve to elucidate in one way or another corresponding passages in pieces selected for publication. In that way we maximize the extent to which the print edition does affordable justice to Peirce’s body of writings.
“Owing to the subsequent editing by Andronicus, the traces of Apellicon’s work would naturally be obliterated in great measure. But we cannot doubt that such a character as we see him to have been would not hesitate to write over the bad places, so as to make what he judged to be sense; and in some cases, Andronicus must have been forced to accept what Apellicon had written, although, by close attention, we may be led to very strong suspicion that the text is not what Aristotle wrote.”
Our central goal is to provide critically edited and reliable texts of Peirce’s work across the wide range of disciplines to which he contributed. Rather than reproducing a single surviving form of a document, we (as critical editors) identify the most mature coherent form closest to Peirce’s composing hand, and, by incorporating identifiable authorial revisions and corrections from subsequent forms or representations, produce an eclectic text which aims to represent his most fully developed intention. Variants from subsequent published forms of the text judged to be editorial sophistications and compositorial errors are rejected; Peirce’s own errors of content are corrected. This new text, when combined with an apparatus documenting the evolution of the various forms of the work, listing the historical variants, and identifying all of our editorial emendations (and their sources), constitutes the “critical edition” of Peirce’s work.
We apply the editorial standards and guidelines of the Modern Language Association’s Committee on Scholarly Editions, and have received the CSE’s emblem, “An Approved Edition,” for all our volumes. The CSE does not dictate a step-by-step procedure for critical editing, but rather identifies the essential elements (a textual essay, a textual apparatus, and a proofreading plan) and recommends the inclusion of a general introduction providing either an interpretive or historical frame for each volume. Beyond these elements, CSE stipulates that the editing theory and procedures be appropriate for the particular author. Implicit in the statement of standards put forward in “Aims and Services of the Committee on Scholarly Editions” (January 1992 printing) is a system of evaluation based on three expected outcomes: accuracy, consistency, and clarity of editorial discussion.
Our editorial procedures are based on the following general goals:
We are committed to producing a critical, unmodernized edition of Peirce’s published and unpublished work. This edition is critical because our central goal is to produce a text that recovers Peirce’s intentions as an author by means of methods that are transparent to readers who consult, and thus are free to disagree with, the textual apparatus that reports all of our decisions. It is unmodernized because, rather than update its accidentals for the benefit of the modern reader, our goal is to present a text that retains the author’s preferences in punctuation, spelling, and capitalization whenever they do not interfere with comprehension.
Although we work from master sets of microfilm, microfiche, and photographic copies of Peirce’s papers, we are committed to proofreading our transcriptions against the original documents to verify the accuracy of the transcribed readings, to ascertain the physical characteristics of paper and ink, and to resolve any problematic marks or revisions on the document. This step is part of a comprehensive plan for proofreading at crucial phases in the transcribing, editing, and publishing process.
We are committed to determining, as clearly as possible, the transmission of the text and its relationship to other writings in the corpus. We use standard collation schedules for horizontal comparisons of Peirce’s printed articles with his own corrected and annotated offprints, with errata sheets, and with other copies of the original printing to discover variant readings. Texts transmitted vertically through two or more manuscript, typescript, or copy-set forms are also collated to identify variations. We compare parallel but distant versions of a text to recover the full family tree for a given document.
When more than one form of a text exists, we are committed to a consistent procedure for copy-text selection and emendation based on W. W. Greg’s “The Rationale of Copy-Text” and modified for nineteenth- and twentieth-century editing situations by Fredson Bowers and G. Thomas Tanselle. As Greg pointed out, the copy-text concept is not an abstract principle but rather a pragmatic rationale for editing when there exists no compelling evidence for making choices among variant readings from surviving texts. Peirce’s tendency to continuously develop a document in many forms over many years has led us to modify copy-text selection in certain cases where “text” cannot be traditionally defined.
In terms of presentation, we are committed to a chronological edition published in a clear reading text. The chronological order of presentation allows us to fit the transmission of individual texts into the larger perspective sought by scholars since Murray Murphey’s The Development of Peirce’s Philosophy appeared in 1961. Our volumes show the evolving nature of Peirce’s work in each of his many disciplines, the breadth of materials that Peirce worked on (often simultaneously) in particular historical periods, and the emerging patterns of Peirce’s intellectual development throughout his entire lifetime. The clear reading text is made possible by our construction of a textual apparatus keyed to the pages and lines of the clear text.
“Unerring judgment has been exercised in the editing both of the present volume and of others. The author’s slips, if not too numerous, have to be corrected, with or without mention, according to circumstances.”
Peirce published a great deal throughout his life, in the form of journal articles (individual or serial), proceedings papers, scientific reports, mathematical papers, book chapters, book notices and reviews, review articles, editorial replies, and occasional obituaries. He published one monograph (Photometric Researches, 1878), and wrote several books, none of which were accepted by publishers, generally because they did not think there was a sufficient readership capable of understanding, let alone appreciating, Peirce’s mathematical textbooks or his philosophical and logical investigations.
As far as the critical and chronological edition is concerned, the works Peirce published in his lifetime fall within two categories: those whose original manuscripts and/or preliminary drafts have survived, and those whose published form is the only extant artifact.
In either case, the first task is to assess whether the publication went through two or more reprints, or whether it appeared just once. In the first case, we locate copies of every reprint and then collate them in order to identify any variation and the reason and hand behind it. In the second case we compare copies from different libraries to ensure the type was not reset at any point during printing.
A number of Peirce’s publications (especially his reviews and his Coast Survey reports) were published more than once during his lifetime. Several of his philosophical publications were heavily revised in offprint for republication, although few ever reached print again. The 1877-78 “Illustrations of the Logic of Science” series went through the most revisions across three decades, a circumstance that creates a special difficulty when considering the need to represent the chronological evolution of Peirce’s thought while trying at once to capture the most mature form of his compositions and to avoid unnecessary redundancies. Sometimes those commendable aims become impossible to comply with and thus elicit the need for creative exceptions. But those are rare situations.
The most ideal situation is an article of which every stage of composition and review has survived. This is the case notably of several papers published in The Monist, where we are in possession of preliminary drafts, more advanced drafts, the original manuscript submitted to the publisher, two or more set of galleys (at least one corrected by the publisher, and another by Peirce), the ultimate set of proofs, the publication itself, and sometimes offprints bearing corrections or annotations by the author. When the full or nearly full sequence of production stages is available, the editing of the publication can be based on a rich set of comparable and collatable data that informs the textual editor’s every decision. This is especially useful when dealing with a journal editor prone to modify or shorten Peirce’s text for whatever reason, or with a house style that imposes its own spelling and punctuation practices over Peirce’s own.
Editing Peirce’s published pieces is restorative work for the most part: we recover Peirce’s known intentions and undo what was done by third parties to his text whenever the imposition was unnecessary—for sometimes it was necessary, when Peirce’s spelling or punctuation was deficient. Peirce sometimes expected proofreaders to take care of correcting his punctuation, sometimes not. We impose our own house style only minimally to ensure layout consistency and comply with bibliographical standards. We emend title references to conform to our own edition style (the latest available edition of Chicago Manual of Style, currently 16th ed.); thus we italicize book titles, place chapter titles in quotation marks, and so on. Lengthy quotations set in the text of Peirce’s original publications are offset and indented to conform with our own style for printing extracts.
When a number of manuscript versions and pre-publication proofs survive and there are several possible copy-texts, a historical collation is used to identify all variants among the relevant forms and to determine the transmission of the text. Generally, a manuscript form (or the last draft if a sequence of manuscripts survive), because it represents Peirce’s preference in accidentals (spelling and punctuation) and is the best authority for his word choice, is selected as copy-text.
Variants (in the copy-text and in later forms) judged to be the result of non-authorial intervention, such as compositorial errors, compositorial misreadings, or editorial changes in grammar or style, are rejected and noted in the historical collation; non-authorial substantive variants within this category appear in the volume’s list of Rejected Substantives. Revisions by Peirce in the copy-text itself supersede earlier layers of the pre-copy-text; a full record of authorial alterations within the copy-text is made, and critically significant alterations within this category appear in the published Alterations List.
All emendations to the texts are listed, and the sources for each cited in the emendations list. Any changes not listed there are described generically in the Essay on Editorial Method or in the item Headnote in the Apparatus. These include purely visual characteristics, such as the uppercase styling of opening words in a publication, or the scholastic font Peirce tended to use in his own pre-1894 typescripts. When no manuscript survives and the proof or published form must serve as copy-text, we emend printer’s styling to conform to Peirce’s own style of punctuation and spelling only if Peirce has demonstrated a clear preference in established practice over time. Peirce’s preference for acceptable nineteenth-century spelling forms, as well as acceptable variant spellings of proper names, are allowed to stand as they appear. We consult both the Century Dictionary & Cyclopedia and the Oxford English Dictionary to make spelling determinations.
"Now it is quite true that Aristotle was almost the first pioneer in logic and just stood at its threshold. It is also true that there are some monumental follies in his physical books; but the worst of these may fairly be presumed to be insertions made by different students during the thirty years when his manuscripts lay on the shelves of his school for general use.”
"What concern have we with the universe, or with the course of ages? No more than my dog has in the book I am writing. Yet I dare say he would defend the manuscript from harm with his life."
Peirce’s manuscript copy-texts require a different editorial approach than do his publications. With published items where no prepublication forms of the text survive to serve as copy-text, we focus our editing effort on locating and eliminating editorial or printing corruptions. With holograph and typescript items, once a copy-text has been selected, transcribed, encoded, and proofread multiple times, the focus is on editing it with a critical latitude tailored to the kind of challenges presented by the document at hand. The range of our interventions is based in large part on the distinctions between public and private writings defined by G. Thomas Tanselle in his statement of textual policy summarized in A Rationale for Textual Criticism (1989).
Private documents include letters, drafts of letters, journals, and notebooks. Some were never intended for publication, and their content was for personal use only. But most of the journals, however cryptic and abbreviated their contents may be, were preparatory to compositions meant for eventual public presentation. Even certain letters fall into this category, those that, once past their opening salutation, morph into full-blown philosophical or logical discussions—for instance, Peirce’s long letters to William James were often formally developed, and written with the knowledge that James might use them in his Harvard classroom (and with the expectation that they would be preserved with James’s papers for future scholars). Peirce’s letters to the Coast Survey contain technical material designed to amplify his formal reports and even to appear in them; in fact his monthly letters to the Survey were formal reports required by law. Given these circumstances, we edit private documents differentially.
Regarding private documents clearly personal in nature, we generally do not interfere with the private character of the presentation—we don’t interfere with Peirce’s private usages, including his page formatting, idiosyncratic syntax and grammar, and unconventional (but nevertheless intended) spellings. These personalized characteristics of presentation would not be allowed to stand by a publisher in Peirce’s day, but these materials—especially the journals—are really “idea books” used by Peirce to work out problems before going on to future development for presentation. When no further development of the text survives, we edit and publish them as private documents, emending only those idiosyncrasies which do not reflect authorial intention (such as transposed letters and slips of the pen).
In the case of private documents plausibly intended for eventual presentation, we will retain the character of Peirce’s work wherever possible; but while we make every effort to preserve the surface features of these documents, we correct mistakes in content. Indeed, the evolving technical and analytical nature of Peirce’s ideas calls for clarity of presentation. For this reason, we intervene to represent clearly Peirce’s intentions, especially in his use of logic signs, mathematical formulae, and scientific notation. But in terms of accidentals, such private documents are generally emended only in the following instances, depending on a document’s overall pattern: Peirce’s unintended spelling an grammatical errors are corrected; missing punctuation is inserted; lowercase variables are italicized; the missing half of a pair of dashes, parentheses, or quotation marks is added; misplaced capitalization is corrected; numbers, either in numerals or spelled out, are left as written unless inconsistencies create confusion; apostrophes are inserted as needed; and Peirce’s incomplete alterations are completed to instill clarity.
Public documents include those that conform to genres normally intended for publication, whether they are personal working copies, early drafts, worksheets, or outlines; lectures written, in outline or fair-copy form, for presentation in lecture courses or at public conferences. All manuscripts or typescripts devoted to topics or ideas that are related to those Peirce published and that he would have been willing to submit for public comment and criticism are so regarded. These might be thought of as works Peirce composed with the general reader or the specialist in mind. By far the larger part of the unpublished manuscripts are considered public documents. Regarding them too, differential editing is called for.
Differential editing means that different textual situations call for carefully calibrated editorial treatment. A critical edition is not a documentary edition, whose concern is to reproduce documents without altering or correcting their content, thus preserving their artifactuality, while illuminating them with historical annotations that help readers understand what they are examining. Documentary editions have their essential place and role, and are best suited for documents produced by historical witnesses, literary writers, or letter writers relating events or reacting to circumstances. When it comes to authors like Peirce, that is, to authors whose contributions fall within the realm of intellectual inquiry in a broad sense, the texts to be edited are traversed by intentions that carry them beyond the date and place of their birth. Such texts are meant to speak to future generations in the spirit of the continued pursuit of knowledge. They provide arrays of ideas, theories, and hypotheses in various states of articulation and expression, nearly all of which were meant, at the time of inscription, to communicate inferences and conclusions to the community of researchers, even if at times that community might have been temporarily reduced to the author’s own mind in subsequent moments, days, weeks, or months. An author like Peirce wrote not only to be read, but also, and especially, to provide thinking readers with ideas intended to be rightly understood, examined, objected to, assented to, and expanded into many more trains of inquiries. Such authors are better served by critical editions.
Heuristic writings call for a type of editing that does not let unnecessary obstacles block the road of inquiry. If, for instance, Peirce misspelled a word, correcting it removes a blemish that would otherwise distract readers to no advantage. If he failed to add a comma whose absence causes readers to stumble and read the sentence again until the intended sense emerges, the critical editor will add that comma to serve both Peirce’s intention (that of being fairly understood) and the readers’ expectation (that the editor’s mediation will have been helpful to them, too). If Peirce omitted a word, the critical editor will supply the missing element when the evidence bears it out, or signal the gap when not. And so on.
We emend when we are confident of Peirce’s intent. If such confidence is lacking, his eccentricities and anomalies are left as written. Sometimes this means that critical editors may need to complete a tacit intention, both for the sake of the author and for the sake of his readers. A stellar example of this kind of situation is found in some of Peirce’s 1892 lectures on the history of science. His lecture on Archimedes, for instance, was written in haste, and shows that Peirce intended to discuss a number of propositions culled from various mathematical treatises. The text barely alludes to those propositions, and vaguely sketches a number of diagrams Peirce intended to draw on the blackboard to support his discussion. The editorial choice was either to reproduce a sort of facsimile of those scribbles without seeking to help readers make sense of what the author was doing or thinking, or to use those scribbles as signs enabling us to infer Peirce’s fuller intention. We chose the latter road, identified what Archimedean passages Peirce was most plausibly discussing, displayed his moves from propositions to propositions across several Archimedean treatises, completed his abbreviated sentences, and redrew his diagrams on the basis of those displayed in his source texts, but only to the extent needed to make sense of Peirce’s discussion. We were therefore able to resurrect the lecture, turning it into a far more telling reading experience than the raw manuscript allowed. This kind of critical editing seeks not to rewrite Peirce, but to complete his writing when the latter falls short of eliciting the interpretants that Peirce’s actual lecturing performance did raise in his audience. Sometimes, therefore, critical editing entails unblocking Peirce’s own inscriptional obstacles in order to recapture the intended intelligibility.
The type of editing done to private documents applies to those intended for publication, but the latter, as just noted, occasionally require deeper intervention. This applies especially to Peirce’s mathematical and logical papers. Peirce understood and observed scientific styling conventions followed by academic journals. He did not always carry them through in manuscript composition, however: his underlining of variables for italics is inconsistent for instance, whether they be Greek or not, uppercase or lowercase. We emend to italics when called for and generally style those texts consistently as though they were professional publications because doing so makes those texts far more readable and less ambiguous than otherwise. In doing so we serve both Peirce’s and his readers’ interest—after all those commonly accepted conventions were designed to facilitate reading through rigor of presentation, and Peirce was sensitive to that concern.
One principal role of the Textual Apparatus that constitutes the bulk of the editorial backmatter in our volumes is to achieve transparency. The apparatus is where critical editors report, minutely and yet succinctly, every modification decision that occurred during editing. Every editorial intervention is recorded within the Emendations list attached to the apparatus of each text, while textual cruxes calling for specific explanations are discussed in corresponding textual notes.
The structure of an emendation in its simplest form is as follows:
Page/line key Lemma] sigla; gloss
More complex emendations have the following form:
Page/line key Lemma] sigla; gloss followed by siglum; alternate reading followed by siglum
The page/line key consists of two numbers separated by a dot, and direct the reader to a particular line or line-range within a page of the volume where an emended reading is to be found.
The lemma reproduces a text segment that resulted from an emendation. The lemma is directly followed by a closing square bracket.
The siglum or sigla that follow the square bracket are abbreviations denoting the source authority for the correction: E for the Editor or Textual Editor; bibliographical sigla for forms of the text where the correction was taken from and accepted by the Editor.
The gloss reproduces the erroneous inscription from the copy-text.
The gloss is sometimes followed by the siglum of its own source and then by one or more alternate faulty readings present at the same location in one or more other forms of the text, each identified by its own siglum.
Example 1:
33.24 excel in all] E; raise all AMS; excell in all TS
This example shows that on line 24 of page 33 (in W6), the inscription “excel in all” is the result of an emendation done on the authority of the Editor (siglum E) to correct the reading “raise all” present in the amanuensis copy-text (AMS) that was replaced by the misspelled “excell in all” in a subsequent typescript (TS).
Example 2:
174.37 food is] G2; foods were MS; foods were G1
This example shows that on line 37 of page 174 (in W8), the inscription “food is” is the result of an emendation done on the authority of a correction made by Peirce in the second set of galleys (siglum G2) to correct the reading “foods were” present both in Peirce’s manuscript (MS) and in the first set of galleys (G1).
The range of textual cruxes is wide, and nearly each text may present one or more situations not encountered previously (such is Peirce’s creativity). Cruxes are recognizable by the confusion they create either by breaking a particular kind of consistency or continuity within a sentence or larger unit of text—a kind that can be logical, grammatical, syntactic, semantic, orthographic, structural, representational, arithmetical, mathematical, formulaic, tabular, punctuational, etc.—or by introducing an ambiguity, an obscurity, or other equivocacy prone to mystify readers. Those confusions usually result from some lapse or slip: incomplete alteration, omitted words, gaps in a list, unusual word, archaic spelling, ambiguous marginal instruction, inconsistent syntactic pattern, mistranscribed quotations, missing intermediary form of a text, and so on. Cruxes are sometimes caused by scholars or editors who in the past made inscriptions directly into the original documents, making it hard to identify the author of those markings.
The difference between a regular kind of emendation and a textual crux is simply that solution to the latter requires a special act of analysis and deliberation. Textual notes reproduce the result of that deliberation and therefore explain the rationale behind editorial solutions. They are the favorite domain of an edition’s Textual Editor, who gets to articulate concisely the conclusion of an inquiry that may well have taken hours of examination and collegial discussion.
“The text of the treatise of Petrus Peregrinus here presented is substantially that of a contemporary MS in the Paris Library. All deviations from that authority are noted. Three other important texts have, however, been carefully collated, together with several that are incomplete. Libri did attempt a transcription of the very MS here used; but, owing to its extreme illegibility, he has hardly been able to make perfect sense out of a single sentence, not to speak of places where his text suggests a wrong meaning, nor of innumerable lesser errors.”
If only transcribing a handwritten text meant simply typing it up faithfully or exactly via a keyboard! When it comes to a scholarly edition of an author like Peirce, whose writings need to be transmitted to his posterity, current and future, as near-perfect substitutes for the originals, no such shortcuts are allowable. The challenge has been compounded with the advent of computer technologies that have relegated the mechanical typewriters to the simple past, for today’s machinery has no trouble accommodating the past perfect, turning that tense attribute into a standard expectation.
Behind this grammatical metaphor lies the world that was opened by markup languages, general or extended. Transcribing a text used to imply deciphering every character of every word, identifying whether a quick mark more or less between words was meant to be a comma or a period, wondering what was the word that was deleted or overwritten, questioning whether an inscription was truly authorial or that of another hand, interpreting interlinear transposition instructions, gauging the intra- or extra-textual intention behind a marginal comment, interpreting underlines (straight or wavy, single or multiple), and typing or recording all of this according to in-house protocols. Nowadays, transcribing remains all of that, but has become much more. It has also become encoding. Encoding means tagging. And tagging means surrounding textual subunits that range from single letters to words, phrases, sentences, paragraphs, sections, chapters, parts, and entire texts with angle-bracketed descriptors adorned with a wide variety of attributes that embody all sorts of values.
Here is an example, made up for the purpose: let us transcribe the following text (the latter being already a transcription that shows deletions without showing how they were nested within insertions) coming from the first page of some lecture.
Ladies & Gentlemen,
Since we last met a a mature a full grown Herbert Spencer has passed away gone, and I suppose you have all read the characterization of him his characterism by Prof. James, as — a perfect all round which neglects no side of the philosopher or the man, and which contains no word that is not true.
How is that opening sentence going to be actually transcribed and XML-encoded in compliance with the principles recommended by the Text Encoding Initiative (TEI) to which scholarly editions generally subscribe? As follows, realizing that the transcriber is deciphering not the above transcription but the messier original in the manuscript:
This example immediately illustrates the relatively recent increase in complexity of the act of transcription. The text now includes tags that characterize every aspect of both the inscriptions themselves (insertions, deletions, underlines, syntactical properties, gaps, and so forth) and also of the nature of their contents (identifying such things as proper nouns, place names, bibliographical material, quoted material, and so on). Those two kinds of tagging can be separated since they depend on distinct acts of analysis, but are shown here together for reasons of rhetoric. The extent to which a text can be encoded is enormous, and editorial policies defining that extent will vary according to the particular priorities and purposes of an edition.
Encoding a text offers several advantages, the principal one being the production of an electronic file that is platform-independent and that, if continually preserved in robust media, will remain machine-readable for a long time. This is essential because one lesson coming from the end of the last century was that proprietary file formats are not persistent and fall quickly into obsolescence, often according to the vagaries of the market, of evolving standards, and of technological development.
Another advantage is the ability is to lift out of an encoded text, on the basis of the different classes of metadata deposited into its tags, all sorts of information that can be accessed quickly with precision, and tabulated in various ways to serve distinct analytical or heuristic purposes. To give one example based on the above illustration, we have developed at the Peirce Project an algorithm that automatically creates lists of authorial alterations thanks to the tags. The list below displays and formats all the alterations found in the sentence above, and was automatically generated (thanks to STEP Transcriptor, a tool described under STEP Tools under the Technology tab).
last met] before del [a below del a mature below inserted a full grown]
gone] intl ab del passed away
and] before del I suppose
you have] before del all
his characterism] intl ab del the characterization of him
Prof. James,] before del as
which neglects no side of] intl ab del a perfect all round
the philosopher or] inserted
which] inserted
This is to say that if, on the one hand, encoding does take more time than the more classical act of transcription, on the other it facilitates and accelerates other data-mining types of tasks, most of which are renowned for their tedium. Besides, a dedicated XML-TEI tool such as STEP Transcriptor has been engineered to reduce the encoding tedium maximally, while also reducing the likelihood of errors.
“Proofreaders get high salaries because ordinary people miss seeing misprints, their eyes correcting them.”
Errors are the bane of all scholarly transcriptions. We have learned over several decades that no matter how rigorous our verification protocols are or how skilled our typists and encoders, the room for error may be shrunk to considerable extent but never completely. Every volume of the Writings, despite our best efforts, contains errors. Admittedly, never very many, as one ought to expect. But each one, when discovered, remains in its excess a source of everlasting pain.
Inseparable from the act of transcription and encoding is the operation of proofreading. The “Production Workflow” (first item under the “Methods” tab) includes quite a number of proofreading stages, both during transcription and during editing. There are two kinds of proofreading. One is called team-proofreading. It involves two readers, one who reads the source document aloud, uttering every single word and mark (“comma,” “stop,” “sem” [for semi-colon], and so on), and spelling out any proper name or unusual word, while the second reader carefully listens, following the text of the transcription. The two proofreaders (who, pace Peirce, only wish they could get high salaries) will regularly alternate, and proofreading sessions are timed so they do not last too long, in order to allow brains to rest and refresh before a new round of strenuous attention to typographical details begins. The other kind of proofreading is silent and involves a single reader, a scholar who understands the content of the text. That person will read the transcription against the original and identify every error that could not have been detected through vocal utterance, or could only have been detected by attending to the meaning of the text. It is that second kind of proofreading that is more likely to identify misreadings and improve hypotheses regarding what word an author intended to write before interrupting its inscription in favor of some other word.
Each round of proofreading results in the marking of errors and the indication of how they ought to be corrected. Transcribers will enter those corrections into a fresh copy of a document after having archived the previous copy. Once corrected, the document will undergo a new round of team-proofreading, which will check every correction to make sure it was entered properly, while also verifying that no new error was inadvertently entered during correction. After all, all it takes for a new error to infect a document is a fleeting pressure on a keyboard key that escaped attention. If any correction error is spotted, it will be marked and subsequently corrected into a fresher version of the same document, and the latter will be proofread again.
The same rigor applies to the proofreading and correction of all documents generated by the editors: preface, introduction, annotations, apparatus, bibliographies, manuscript catalogues, every component of a volume goes through this process. This is how errors are minimized—and yet, frustratingly, never completely eliminated. It is fortunate that we are publishing an author who demonstrated that absolute exactitude was logically impossible. We are grateful for one of the most precious –isms that define his legacy: fallibilism, the kind that is permanently soothed by genuine contrition.
“The third grade of meaning of a word is the vast ocean of unforeseen consequences its acceptance is destined to bring about.”
About twenty years ago, an Assistant Editor of the Peirce Project published a paper in Text, an Interdisciplinary Annual of Textual Studies published by the Society for Textual Scholarship (the journal has since then been renamed Textual Cultures). That paper, well received by the profession, was titled “Selecting Alterations for the Apparatus of a Critical Edition” (Text 9 [1996]: 33–62). The research that led to that paper stemmed from the CSE expectation that Writings volumes needed to provide a record of Peirce’s alterations in the manuscripts, a practice the Project began to observe in volume 5.
What is an “alteration”? It is any visually detectable or contextually inferable syntactic or semantic discontinuity introduced by an author within a portion of a text after its partial or complete inscription, whether as an immediate correction made in the process of composition, or as a later revision. In his 1976 paper “Transcription of Manuscripts: The Record of Variants” (Studies in Bibliography 29: 225), Fredson Bowers wrote that the function of the Alterations list in the apparatus was to list “every alteration, whether of correction or revision, that an author made during composition or review.” While Peirce Project transcribers do record every single such alteration when typing and encoding the text, thereby generating a complete listing of them, Project editors recoiled at the idea of publishing that complete listing in the volumes themselves for several reasons, chief of which was a combination of impracticality and unpragmaticality. One assumption that had been driving Bowers’s injunction was the idea that the reader should be able to reconstruct the original appearance of each manuscript page as closely as possible by consulting the Alterations and the Emendations lists. Our own experience was that even the most patient and competent scholars are unable to reconstruct without error a manuscript from the sole study of these compact lists strewn with abbreviations and other symbols.
While a complete list of emendations is obviously necessary {they reflect the editors’ post-copy-text alterations), publication of a complete list of authorial alterations within the pre-copy-text supposes that the readers are generally as interested in getting a precise picture of all the different writing-layers that led to the emergence of the copy-text as they are in getting an exact knowledge of the author’s final intention, as revealed in the clear reading text. Most readers of a critical edition rarely consult either Emendations or Alterations lists, not only because of the tediousness, but also because it is not immediately clear how to use the data provided in those lists. Some readers, however, do consult the critical apparatus and more specifically the Alterations list, namely those serious scholars intent on determining as minutely as possible the different steps of the development of an author’s thought and of its expression. It is for those scholars that the critical apparatus is chiefly intended. But even for them it is doubtful whether a complete record of alterations might be serviceable. The number of alterations that can tell an acute reader something truly significant about the development of an author’s thought is not very high, compared with the average number of trivial alterations found in any given manuscript. If the more significant alterations are buried in the complete Alterations list, the time wasted by exceptional readers in hunting them down would most likely be a source of irritation they would prefer being spared. On that count, it would be much more convenient and less risky for such readers to get a photocopy of the manuscript and read it through.
In the face of those practical considerations, the issue transformed itself into a search for principles of selection that would allow editors to publish only a partial list of alterations, to wit all of those that could not be considered trivial. But how can we determine whether an alteration is trivial or not? Rules of selection are rules of distinction and classification—in our case criteria that allow us to assess the value or importance of an alteration in view of our general aim: making Peirce’s ideas available to anyone interested in them by providing the information necessary to obtain the closest and fairest representation of the development of their expression. This entailed providing the reader with a list of all those textual alterations whose report significantly furthers the representation of the evolution of an author’s thought.
It struck us that Peirce’s pragmatic maxim embodied exactly the kind of general principle we were looking for. The tenor of that well-known maxim is that our whole conception of an object is that of its effects that might conceivably have practical bearings. From this fundamental logical maxim, all sorts of more specialized applications can be derived, including one that would enable editors to determine whether or not any given alteration is worth selecting for publication in view of its critical significance. Such a maxim must in essence define what a critically significant alteration is. The following maxim was therefore formulated:
Consider whether the change of intention possibly manifested in the alteration produces an effect which might conceivably modify a reader’s perception or understanding of the altered passage; if there is any such conceivable effect, then the alteration is critically significant.
A semiotic translation of this maxim would be the following: Consider whether any discontinuous event happening within the genealogy of a sequence of signs (thus, in the course of a semiosis) is likely, if reported, to affect future interpretants so that the latter ramify in sensible or fruitful directions that would otherwise be curtailed. Such genuinely consequential semiotic events (or sign modifications) are critically significant when they fall within the limits of what Peirce defined as critical commonsensism, which does not welcome such things as paper doubts, artificial semiotic effects, or mere modifications that do not result in any change of meaning or other effectual difference.
The paper mentioned above distinguishes eight classes of alterations: simple insertions, simple deletions, complex insertions, superimpositions of letters or words, simple transpositions, complex transpositions, typeface and character-style alterations, and canceled alterations. To each one it assigns a general principle of selection derived from the above maxim. But since all alterations essentially depend on the two basic acts of insertion and deletion, we will provide here the two major rules only:
General selection principle applied for insertions: List all critically significant dispensable insertions and omit other insertions, whether indispensable or not critically significant.
General selection principle applied for deletions: List all informative deletions and omit all uninformative deletions. [The word “informative” is defined as “indicative of a critically significant piece of information that conceivably manifests a change of authorial intention while meeting four conditions: deletions must be decipherable, unambiguously identifiable, the product of a critically significant move or error (not a trivial one), and yielding either a new reading or one significantly different from what is found in the relevant part of the text.]
The paper provides many examples and shows how to apply the different selection rules, while also providing several charts that help conduct the selection in truly pragmatic fashion. The fact that application of the maxim works well has been confirmed time and again whenever editors in charge of wielding it found themselves forced to accept or reject alterations that at first they would have impulsively rejected or accepted instead. Given this selective process, there are sometimes manuscripts for which we do not publish an alterations list, even though they do contain alterations recorded in our master files.
Peirce’s pragmatic maxim promotes an objective kind of criticism that when applied to the editorial treatment of texts yields in the end a product that truly deserves the name of a “critical edition.”
“The true meaning of any product of the intellect lies in whatever unitary determination it would impart to practical conduct under any and every conceivable circumstance, supposing such conduct to be guided by reflection carried to an ultimate limit."
“If pragmatism is only true, it is certainly a wonderfully efficient instrument. It is not to philosophy only that it is applicable. I have found it of signal service in every branch of science that I have studied.”
“If I wished to order a font of type expressly for the printing of this book, knowing, as I do, that in all English writing the letter e occurs oftener than any other letter, I should want more e’s in my font than other letters. For what is true of all other English writing is no doubt true of these papers. This is a statistical deduction.”
The first five volumes of the Writings were sent to a typesetter/compositor before presswork could begin at the printer’s. Editors would send hard-copy printouts of the entire volume, section by section to the typesetting company. The text was there completely retyped by teams of typists and laid out to conform to the style and format of the volume. Sets of proof pages (or passes) were sent back to the Project, entailing repeated proofreading rounds that tallied all errors, attributing their origin either to Project staff or to the typesetter’s own employees. Corrections of errors attributable to the latter were free, those attributed to the Project were not. The process was slow and expensive, and tended to generate tension between the Project and Indiana University Press, given its sometimes staggering cost (the cost of scholarly exactitude).
Advances in word-processing and page-making software made it possible for the Project to computerize the entire production and layout process and to wave good-bye to the typesetting company. The first publication produced entirely in-house as a kind of prototype or proof of feasibility was the second volume of The Essential Peirce in 1998. Volume 6 was the next volume, which appeared in 2000. The software we used, and continued to use for the production of volumes 8 and 9, is Adobe’s FrameMaker+SGML (long rendered obsolete on the Mac platform).
An entirely new workflow had to be set up for the production of a “camera-ready copy” of our volumes. The adoption of FrameMaker coincided with our decision to comply with TEI encoding standards. All Project employees began learning to encode or tag everything they typed. This was a return to pre-Apple days, in a way, when the old green-screen computers required the tagging of such things as italic words. But when the Project adopted the Apple Macintosh platform in 1986, it was in part because the screen offered a WYSIWYG interface that instantly showed the result of the encoding, and the software provided buttons and keyboard shortcuts that turned the encoding into apparent obsolescence. That respite lasted only a good ten years. Encoding returned in the mid-1990s with a vengeance, never to leave us again, perhaps holding us in its angular grips forever.
We became familiar with acronyms like DTD (Document Type Definition, which defines the structure and legal elements and attributes of SGML and XML documents) and EDD (Element Definition Document, a document embedded in a template and containing both structural rules for the DTD and styling rules that dictate how elements of a specific type are styled). A Project employee devoted years of his life to construct, even program, those elaborate documents, which successfully drove the entire compositional and layout process leading toward the camera-ready electronic files we sent to IU Press’s printer.
Whether the transition to a complete process of production, from transcription to layout and camera-ready copy, cut any real costs or save any time is not open to question: the answer is a definite no. The real advantages were to enable the Project to control all aspects of the volume, including its layout, without unnecessary compromise; (2) to reduce the range of errors markedly; (3) to increase the quality and richness of the scholarship going into the production of each volume; (4) to set us on a path that has led us, first to redesign the volume and the editorial practices governing its making, and second to prepare us to reconceive the range of services a critical edition ought to provide once it turns itself to the online world.
Indeed, the future of the edition is no longer confined to the printed volumes. We want to continue their production because they offer a real scholarly convenience that is esthetically, logically, and metaphysically distinct from, and not reduced by, the possibilities of digital editions. Paper editions offer one type of experience, one type of construct that embodies its own logical, chronological, multi-topical, and synthetic consistency throughout in such a fashion that they become the ground of interactions and especially serendipitous discoveries that digital editions cannot paradoxically replicate. Digital editions offer several other types of experiences that are supplemental. But the latter editions cannot supersede the former even when considering the extent to which they do overlap. The reason is that a genuine digital edition does not seek to mimic a paper volume: a mere PDF of a print volume is no digital edition: it remains a print(able) edition. A genuine online edition offers a distinct range of services, allowing readers to apply to the texts all sorts of analytical, data-mining, search-functional, configurational, navigational, communicational, pedagogical, and research-collaborative tools capable of extracting, organizing, and sharing information speedily and flexibly for all sorts of purposes.
The Peirce Edition Project aims therefore to continue producing camera-ready volumes, while also producing digital-ready corpora. Such is the intention driving the technologies we are currently developing, including STEP and CORPUS.
“The lack of an adequate overview of the ‘complete’ corpus of Peirce’s prodigious production in many fields of science and philosophy—to which corpus many more items are sure to be added as time goes by—has had its influence upon the interpretation of the significance of his life and work. Those of us who have worked on [the bibliography of Peirce’s publications] hope that it will help in the reassessment of his importance as a leading figure in philosophy and science, and that through this, others may come to share our conclusion that Peirce is a master among those who pursue wisdom, a master of such stature that there have been others his equal, but none his better.”
“The development of my ideas has been the industry of 30 years. I did not know as I ever should get to publish them, their ripening seemed so slow. But the harvest time has come, at last, and to me that harvest seems a rich one, but of course it is not I who have to pass judgment. It is not quite you, neither, individual reader; it is experience and history."
The following slideshow, created in 2010 for a presentation at Kyoto University that ended up not taking place, illustrates the several editions that then comprised the Institute for American Thought, and especially the work done at the Peirce Edition Project.
“Indeed, out of a contrite fallibilism, combined with a high faith in the reality of knowledge, and an intense desire to find things out, all my philosophy has always seemed to me to grow.”
