Theory and Metatheory in the Development of Text Encoding Allen Renear, Scholarly Technology Group Brown University Allen_Renear@Brown.Edu Working draft November 3, 1995 In the course of scientific investigations we say all kinds of things; we make many utterances whose role in the investigation we do not understand. ... Our thoughts run in established routines, we pass automatically from one though to another according to the techniques we have learned. And now it comes time for us to survey what we have said ... now we have to clarify our though processes philosophically. -- Ludwig Wittgenstein, 1947 CONTENTS 1.0 A Procedural Preamble 2.0 Introduction 3.0 Computer Text Encoding 4.0 The Emergence of Content-Based Text Encoding 4.1 Format-Based Text Processing 4.2 Content-based Text Processing 4.3 Advantages of Content-based Text Processing 5.0 Textual Ontology 5.1 Phase I: Platonism 5.2 Phase 2: Pluralism 5.3 Phase 3: Antirealism 6.0 Conclusion 7.0 Notes 8.0 Bibliography 1.0 A PROCEDURAL PREAMBLE #1.0.1 In June of 1995 at the 18th International Wittgenstein Symposium, Culture and Value: Philosophy and the Cultural Sciences, in Kirchberg, Austria, Claus Huitfeldt asked me if I would contribute a "target paper" on philosophical issues in text encoding to *The Monist's* "Philosophy and Electronic Publishing" electronic conference. Although, as most of you know, I am rather taken by this topic and eager to promote further research and discussion of theoretical issues in text encoding, I was doubtful whether I could provide anything of value at that time. I certainly did not have any fresh results or analyses to offer, nor was there time to prepare a proper review of the discussions to date -- and I think that the issues are very hard ones. So I could not see how I could deliver a paper that would be substantive enough, good enough, or philosophical enough, to be of interest to *The Monist* readership. This was disappointing as my particular interest in the theoretical problems of text encoding is focused directly on its connections with mainstream philosophical topics such as scientific explanation, natural kinds, ontology, antirealism, interpretation, etc. -- so participating in an exhibition of the relevance of current discussions of text encoding to these philosophical topics in a journal such as *The Monist *is an opportunity I would be sorry to miss. #1.0.2 However with some encouragement from Huitfeldt I decided to view my initial reluctance as a result of being too much in the grip of the traditional publishing and scholarly communication mindset, a mindset rather pointedly inappropriate to *The Monist's* interest, in its "interactive issue", in experimenting with new vehicles for research and scholarship. I decided that if I viewed The Monist Philosophy and Electronic Publishing conference as a context for developing research and analysis, rather than for presenting and testing it, then I perhaps could provide an instrument suitable at least for initiating and orienting that discussion. #1.0.3 In fact, I do think that *The Monist* Philosophy and Electronic Publishing conference may be just the thing needed for those of us working in the area. Discussions of theoretical or philosophical issues in text encoding have always been important to our community and there has all along been a certain amount of published critique and engagement on theoretical issues: Mamrak et al 1988 and Raymond et al 1993 respond to Coombs et al 1987, Fortier 1990 responds to an early version of the TEI Guidelines; Sperberg-McQueen 1991 responds to Fortier 1990; Huitfeldt 1995 to DeRose et al 1990 and Renear et al 1992, and so on. But I think it is obvious to all of us that the few published exchanges not only merely scratch the surface of these rich topics, but that the vast amount of related thinking and analysis that has already taken place at conference dinners, workshop coffee breaks, airport lounges, email exchanges, and the like is entering the periodical literature at rather glacial pace. #1.0.4 So my modest intention here is merely to set the scene for a discussion. The paper that follows is an overview of the general trends and issues in text encoding, emphasizing several particular theoretical and philosophical issues. Although I think that most of the participants in our conference will be active members of the text encoding community, I've adopted this approach in order to facilitate participation by any philosophers or others unacquainted with text encoding, and to orient us to the possibility of presenting the results of our discussions to a wider community which is entirely unfamiliar with text encoding and, perhaps, skeptical as to what it can contribute to their knowledge of either text or philosophy. #1.0.5 Although I have a substantive interest in exhibiting the connections between discussions of text encoding and certain philosophical topics, and in defending a particular view, I have no particular expectation of how the discussion should be conducted or what should result in the way of publications, transcripts, etc. I presume that the advantage of electronic discussion of this sort is that procedures, policies, and objectives can all be reconfigured at any point and that among other things we are exploring just how to make this medium suit our interests. 2.0 INTRODUCTION #2.0.1 This paper discusses the development of certain theories about the nature of text and the practice of text encoding. These theories evolved within a particular broad interdisciplinary community of engineers, programmers, systems analysts, editors, scholars, and others who were either designing or using computer text processing tools between 1950 and 1995. As an episode of theory formation this is a peculiarly fascinating story, a rich fabric of varied argument and evidence, passionately and ingeniously elaborated, that draws on many disciplines and is tightly bound up with very practical problems of text processing and publishing. As several of participants have observed, it not only raises in a compelling way many valuable perspectives on texts and textual communication, but also, I think, exercises a number of important philosophical questions about the separation of science and metaphysics, the nature of our knowledge of the world, and the nature of classification, natural kinds, and scientific explanation. #2.0.2 I am particularly interested in how, as practical questions give rise to theoretical enquiry, the views of textuality offered by practitioners of text encoding recapitulate an interesting and familiar evolution from a kind of platonistic essentialism, to a less platonic and less essentialist, but still realist, pluralism, to positions that seem more pragmatic, constructivist, and antirealist. Although the domain is a practical one, and the theorizing is presumably directed at resolving practical problems, the discourse which carries this theoretical evolution contains many argumentative strategies familiar to philosophers. These include arguments from hypothetical variation (to discover essential properties), existential instantiation (to display ontological commitment), conceptual involvement (to detect conceptual priority) and others. Ultimately this story reveals that theories about the nature of texts appear inevitably and inextricably intertwined with theories about the nature of theories and the distinction scientific inquiry and philosophical argument is in practice very hard to make out. #2.0.3 In what follows I describe the domain of text encoding and how a particular approach to text encoding came to dominate that domain, sketch some of the history of further theorizing about text encoding that this approach motivated (theorizing which passes through the three stages mentioned above: Platonism, Pluralism, and Antirealism), and then end with a partisan claim: that certain representative arguments for the antirealist approach to theorizing are not persuasive. I hope that by both drawing fire from my friendly critics and eliciting the aid of supporters, this essay will stimulate the further evolution and refinement of theories and metatheories about text encoding. This is thus very much a preliminary effort, intended to stimulate discussion. #2.0.4 What value this discussion has for the debates of professional philosophers about realism and antirealism remains to be seen, but it is possible that this fresh perspective, one which is interdisciplinary, and grows out of practical concerns of such (apparently) non- philosophical activities such as software design, electronic publishing, textual editing, and literary scholarship, may make a contribution to more focused philosophical discussions.(1) 3.0 COMPUTER TEXT ENCODING #3.0.1 Computer text encoding is the representation of textual information on the computer, typically in order to facilitate the use of computer hardware and software in creating and processing documents such as letters, reports, scientific papers, novels, etc. The sorts of processing made possible by text encoding includes editing, formatting, printing, typesetting, indexing, retrieval, analysis, and so on. It proceeds by recording representations, in the computer's memory, of both linguistic content (such as alphabetic characters and punctuation) and additional information related to this content (such as the explicit identification of titles, paragraphs, and footnotes, or the indication of desired formatting effects). A "text encoding system" is the system of codes which implement this representation. #3.0.2 By the 1960s text encoding systems were widely used commercially to support computer text processing and computer typesetting. At that time a compositor or author would prepare a data file consisting of "markup" (computer codes specifying formatting information) and "content" (codes specifying the linguistic items of the text, such as alphabetic characters and punctuation). This file would then be processed by formatting software, creating another data file which would then be transferred to a printer or typesetting machine to produce printed pages of text. #3.0.3 Both the representation of linguistic content and the inclusion of additional information about either the editorial structure or the format of a document are generally included in the notion of text encoding, but it is the latter category -- the codes which are added to the representation of letters and punctuation in order to effect certain results (e.g. formatting the texts) or indicate the identity of certain features (e.g. that something is a title) -- that are the focus of attention below. It is this category which most seems to make text encoding an encoding, in an interesting sense, and which has been most involved in raising various theoretical questions. #3.0.4 In the 1980s *interactive *word processing systems -- which were typically and unfortunately referred to as WYSIWYG systems: "what you see is what you get" -- came to dominate commercial and academic document development. Although these were also based on encoding systems, unlike the "batch formatters" the computer codes were typically hidden from users even during document creation. In these systems the user supplied codes indirectly and covertly, by key stroke, menu selection, or other devices, and the text was immediately formatted into visible pages and rendered with traditional graphic devices such as leading, font changes, indentations, and so on. #3.0.5 Naive users of interactive systems were quick to assume that since they didn't keyboard any markup and didn't see any markup that therefore there were no "codes" involved in the process at all and that they were somehow being given direct and "transparent" access to "text itself". This, of course, was not true. For with interactive text processors as well as batch text processors there is a complex system of encoding that represents features of the text and supports both the display and editing. All users of computer text processors are unavoidably involved in computer text encoding whether they know it or not. The existence, and importance, of these codes reasserted itself painfully as users discovered difficulties in printing, sharing, and combining documents due to differing software systems. The effort to solve practical problems such as these was a major force in the early development of theories of text and text encoding.(2) 4.0 THE EMERGENCE OF CONTENT-BASED TEXT ENCODING #4.0.1 In the text encoding domain the most significant theoretical discussions about the nature of text are connected with the emergence of a particular approach to text encoding, which might be called "content-based text encoding". #4.0.2 The earliest endorsement of this approach was by software engineers who were, of course, not trying to make an ontological point, but rather to promote a particular set of techniques and practices as being more efficient and functional than the competing alternatives. To promote their approach and discourage others they offered a theoretical backing which both explained and predicted these efficiencies (Reid 1981, Goldfarb 1981). But in the course of the discussion, some partisans of content-based text processing inevitably claimed more directly that the alternative representational practices were inefficient because they were based on a "false model" of text and that their many disadvantages and inadequacies ultimately flowed directly from that flawed conception. (Coombs et al 1987, DeRose et al 1990). 4.1 FORMAT-BASED TEXT PROCESSING #4.1.1 As described above, early computer text processing and typesetting systems were "batch" systems. A data file would be prepared and then processed into formatted page images. This data file typically consisted of (i) the linguistic character content (e.g. the letters of the alphabet, punctuation marks, digits, and symbols) of the text to be rendered and (ii) interspersed codes indicating information about that content. In the earliest systems these codes would typically specify formatting effects. For instance, a code might be mnemonic expressions such as ".skip 3;" for "skip three lines" or ".indent 5;" for indent five columns. Delimiters such as the indicated use of "." and ";" would allow the computer to distinguish computer markup from text content. 4.2 CONTENT-BASED TEXT PROCESSING #4.2.1 Obviously certain patterns of formatting codes tended to recur. Prose extracts for instance might always be preceded by ".skip 2;. indent 5 -5;.singlespace;.fontsize 9". In fact if a compositor was working from (i) a stylesheet that specified how prose extracts were to be set, (ii) a software manual indicating how to get the desired formatting effects with the particular software, and (iii) a manuscript which identified some bits of text as being prose extracts, then it was almost certain that patterns of formatting codes would occur in regular patterns and that these patterns corresponded to the important editorial elements of the text. #4.2.2 It was natural for software designers to simplify the creation and maintenance of the data file described by allowing abbreviations for regularly occurring sequences of formatting codes. The formatting codes could then be automatically substituted for the mnemonics during processing. This had many advantages, as we will see later, but most obviously it made it not only easier to input the text (only a simply mnemonic needed to be entered), but also easier to maintain and revise the formatting, since formatting for all instances of an editorial element could be revised in one place, where the abbreviation expansions were stored. #4.2.3 In some cases of actual use these abbreviations, or macros, seemed to receive their identity as much from the particular combination of formatting effects achieved as from their connection with their editorial element. For instance, "Format 17" might mean "left adjust; Times 14; skip 2" and would be used wherever those effects were called for, whether in a title, a list item, or a verse quotation. But in the thoughtful systematic use of these codes there was a natural tendency to identify a code not with its formatting effects, but directly with a type of text element. This more efficient systematic use was supported by the use of mnemonics connected with the identity of the element. For instance, ".prose extract;" would be used to identify a portion of text as a prose extract. Unlike "Format 17", which might be used as if it were directly connected with a particular formatting effect ("left adjust; Times 14; skip 2") and had no other identity, ".prose extract" was only indirectly associated with its formatting effects via a style sheet which controlled how "prose extracts" were to be formatted. #4.2.4 This approach to text encoding and text processing turned out to have many practical advantages. In the 1970s a number of software designers and computer scientists reflected on these advantages and came independently to the conclusion that the best way to design efficient and functional text processing systems was to base them on the view that there are certain features of texts -- such as titles, chapters, paragraphs, lists and so on -- which are fundamental and salient, and that all processing of texts (such as authoring, editing, formatting, browsing and analysis) should be implemented indirectly through the identification and processing of these features rather than by inserting formatting codes, or other processing codes, directly into computer files. These features have been called "content objects" (DeRose et al 1990) and this approach to text processing "content-based text processing". #4.2.5 The two different approaches to text processing seemed to locate, or at least represent, their differences in the kinds of markup they deployed. Because markup thus seemed to be at the heart of the nature of text encoding, and the key to understanding its failures and successes, theoretically inclined text encoding theorists explored its nature, providing taxonomies and distinctions. Three types of markup in particular are typically identified:(3) #4.2.6 Descriptive Markup -- identifies (or "describes", or "declares") a particular textual element. E.g. "," "," "<stanza>". #4.2.7 Procedural Markup -- specifies a procedure to be performed, typically a formatting procedure, as in this example, although other sorts of processing may also be specified. (e.g. ".sk 3; .in 5" -- meaning *skip 3 lines, indent 5 columns*). #4.2.8 Presentational Markup -- markup consists of graphic devices such as leading, font changes, word spaces, etc. This is the visual language which has evolved as part of the history of scribal and print transmission of text. and functions often to effect the efficient recognition of content objects, facilitate understanding and reading, or simply ornament the text. For instance, in a title: *The Tragedie of Mariam: The Faire Queene of Jewry* The interplay between these three categories of markup during a typical instance of text processing (say formatting) suggests both that they mark salient aspects of text processing and suggests a certain priority for descriptive markup: Descriptive markup, which marks relatively stable and permanent (perhaps we could say intrinsic) features of the document, is mapped to selected procedural markup, which varies with the circumstances (different software, printer, occasion of printing, etc.) in order to produce the presentational markup desired for that particular instance of formatting. #4.2.9 As noted above, users familiar with interactive word processing might be tempted to say that there is no markup at all in their computer files. Again, not only is there usually presentational markup, but there is frequently a great deal of markup which is concealed from view by the software. Probably the closest one could get to a text with "no markup" would be the scriptio continua found in ancient manuscripts.(4) #4.2.10 Responding to the centrality of markup in general, and descriptive markup in particular, an effort to standardize markup systems begin in the early 1980's and eventually resulted in an international standard for defining descriptive markup systems: it is ISO 8879: *Information Processing -- Text and Office Systems -- Standard Generalized Markup Language (SGML)* SGML has been endorsed by many European and North American governmental agencies, professional organizations, and scholarly societies, and it has been achieving steady acceptance by software developers as well. #4.2.11 SGML specifies a machine-readable format for defining markup tags that identify text objects ("elements" in SGML terminology), and for specifying what objects are allowed in a document and what the syntax of each object is -- that is, what structural arrangement of objects are allowed or required. SGML is not itself a set of markup tags for specific content objects, but rather it is a *meta-grammar* for defining sets of markup tags. The technique for specifying these syntactical constraints is similar to the production rule meta- grammar developed by Noam Chomsky to describe natural languages. #4.2.12 For example, a Document Type Definition for verse might define a syntactically valid poem (of a particular kind) as consisting of exactly one title followed by exactly one body; a body as consisting of at least one stanza, a stanza as consisting of at least one line, and a line as consisting of exactly two half lines. The nature of such formal definitions is well understood, having been much studied in linguistics and mathematics; this formalism provides exactly the sort of map of text structure needed by the computer in order to intelligently organize and manipulate textual information. #4.2.13 The principal vehicle for the development and standardization of descriptive markup for the humanities is the *Text Encoding Initiative*. The TEI, founded in 1987, is an international effort to specify a common interchange format for machine readable texts; it has developed an SGML-conformant language, known as "TEI" which is in common use, particularly among scholars in the humanities and social sciences. As we will see later, in the course of the development of the TEI standard the interests of TEI developers, primarily scholars in analyzing existing texts, lead to views on text modeling somewhat at odds with the early developers of SGML, who were primarily engineers and systems designers interested in creating texts that could be easily formatted. 4.3 ADVANTAGES OF CONTENT-BASED TEXT PROCESSING #4.3.1 It is the number and diversity of the practical advantages of the content object approach which initially motivated conclusions about its significance for our understanding of the nature of text -- and which continues to be seen as either explaining, or being explained by, theories about "the nature of text". Its extraordinary success at securing diverse practical advantages seemed to provide compelling evidence for its implicit view of text.(5) #4.3.2 Advantages for Authoring, Editing, and Transcribing #4.3.3 *Composition is simplified.* With content-based text processing, formatting considerations make no claims on the attention of the author, editor, or transcriber during composition: rather than needing to remember both (i) the required style conventions relevant to the text being produced and (ii) the formatting commands used by the software in order to format the text according to those conventions, an author or editor instead simply identifies each text element, perhaps by choosing a name from a menu, and the appropriate formatting takes place automatically. In the jargon of software engineering content objects let the author or transcriber deal with the document at the "level of abstraction" appropriate to their roles -- identifying a text object as a quotation, paragraph, or verse line is an authorial task, making decisions to italicize or center a title is the task of a typesetter's or designer's task. #4.3.4 *Writing tools are supported*. Content objects support "structure- oriented editors" and other composition tools. These are software tools that "know" about the syntax of the special editorial objects that are found in the author's document and so can intelligently support the author in entering, editing, and manipulating these objects. For instance, it will allow an author to perform operations like moves and deletes in terms of these natural meaningful parts (words, sentences, paragraphs, footnotes, sections, etc.) rather than relying on the mediation of accidents: line displacements, numbers, or arbitrarily marked regions. Again, this kind of support allows the author to address the document in terms appropriate to the authorial role -- that is, in terms of content and structure. #4.3.5 *Alternative document views and links are facilitated*. A familiar example is outlining software that takes advantage of the identification of the hierarchy of the major editorial content objects in a document. A more sophisticated and selective display of portions of documents can be effected using discipline-specific content objects in the document -- for instance, one could specify a view that contained only the verse quotations or only the dialogue lines spoken by a single character in a script. #4.3.6 Advantages for Publishing #4.3.7 *Formatting can be generically specified and modified*. As we have seen, formatting of all instances of content objects can be easily and systematically controlled from one location, the rules file for object types, without altering the text file itself at all. #4.3.8 *Apparatus construction can be automated*. Apparatus such as indices and appendices can be much more fine-grained and their creation can be automated. For instance, if stanzas and verse lines are explicitly identified, the creation of an index of first lines of stanzas (or second lines or last lines) can be automated. #4.3.9 *Output device support is enhanced*. When internal coding is based on logical role rather than appearance, output device specific support for printers, typesetters, video display terminals and other output devices is maintained separately, logically and physically, from the data with the convenient result that the data files themselves are *output device independent* while their processing is efficiently *output device sensitive*. #4.3.10 *Portability is maximized*. Files that identify content objects are much easier to transfer to other text processing systems or computers. They are relatively system- and application- independent. #4.3.11 Advantages for Archiving, Retrieval, and Analysis -- Text as a Database #4.3.12 *Information retrieval is supported*. The content object model treats documents and related files as a database of text elements that can be systematically manipulated. This can facilitate not only personal information retrieval functions, such as the generation of alternative views, but also a variety of finding aids, navigation, and data retrieval functions. #4.3.13 *Analytical procedures are supported*. Similarly much automated analysis (stylometrics, content analysis, statistical studies, etc.) is not possible unless features such as sentences, paragraphs, stanzas, dialogue lines, stage directions, and so on have been explicitly identified in a manner that is reliably tractable to machine processing. * !! NB: Please note new telephone and fax numbers since Nov 9, 1995 !! ******************************************************************** * Claus Huitfeldt, Director, * Tel: +47-55-58 29 50 * * The Wittgenstein Archives at the * or: +47-55-58 29 59 * * University of Bergen, * Fax: +47-55-58 94 70 * * Harald Haarfagres gt 31, * Email: * * N-5007 Bergen, Norway * Claus.Huitfeldt@hd.uib.no * * URL: http://www.hd.uib.no/wab/ * ******************************************************************** From Claus.Huitfeldt@hd.uib.no Wed Dec 20 04:27 EST 1995 Received: from alf.uib.no (alf.uib.no [129.177.30.3]) by mail1.its.rpi.edu (8.6.9/8.6.4) with SMTP id EAA27879 for <brings@rpi.edu>; Wed, 20 Dec 1995 04:27:02 -0500 Received: from humanistisk.hd.uib.no by alf.uib.no with SMTP (PP); Wed, 20 Dec 1995 10:26:57 +0100 Received: from HUMANISTISK/MERCURYQ by humanistisk.hd.uib.no (Mercury 1.13); Wed, 20 Dec 95 10:27:02 MET Received: from MERCURYQ by HUMANISTISK (Mercury 1.13); Wed, 20 Dec 95 10:26:54 MET From: Claus Huitfeldt <Claus.Huitfeldt@hd.uib.no> Organization: University of Bergen, HD/NCCH To: Selmer Bringsjord <brings@rpi.edu> Date: Wed, 20 Dec 1995 10:26:47 GMT+1 Subject: Re: Renear's target paper - part 2 Priority: normal X-mailer: Pegasus Mail v3.22 Message-ID: <13986B60667D@humanistisk.hd.uib.no> Content-Type: text Status: R 5.0 TEXTUAL ONTOLOGY #5.0.1 In this section I attempt to fit the progress of theorizing about texts (and eventually, theorizing about theorizing about texts) into three stages: Platonism, Pluralism, and Antirealism. 5.1 PHASE I: PLATONISM #5.1.1 The earliest arguments for the content-object approach to text processing were not intended to make an ontological point about "what texts really are", but rather to promote a particular set of techniques and practices as being more efficient and functional than the competing alternatives. But is was natural to claim that the alternative representational practices were inefficient because they were based on a false or inadequate understanding of the nature of text. #5.1.2 The view that texts are hierarchical structures of objects such as chapters, titles, paragraphs, stanzas, and the like is already implicit in the early efforts to theorize about the development of text processing software and promote descriptive markup, (Goldfarb 1981, Reid 1980, Coombs et al 1987) and is directly implied in the SGML standard itself, which is a standard for defining hierarchial models for representing texts, clearly privileging editorial objects (such as sections, extracts, etc.) rather than, for instance, design objects (such as pages, columns, and typographical lines). #5.1.3 In 1990 this position was made explicit in DeRose et al 1990, where the straightforward ontological question which serves as the article's title -- "What is Text, Really?" -- is given a straightforward ontological answer: "text is a hierarchy of content objects", or, in a slogan and an acronym, text is an "OHCO". The claim here is that in some relevant sense of 'book', 'text', or 'document' (perhaps *qua intellectual objects*) these things *are* 'ordered hierarchies of content objects'. A book for instance is a sequence of chapters, each of which is a sequence of major sections, each of which in turn is a sequence of subsections. Within the lowest level subsections are objects like paragraphs, sentences, prose quotations, verse quotations, equations, proofs, theorems, and so on. This structure is *hierarchical* because these objects 'nest' inside one another. It is *ordered* because there is a linear relationship to objects -- for any two objects within a book one object comes before the other.(7) And they are *content objects* because they organize text into natural units that are, in some sense, based on meaning and communicative intentions.(8) #5.1.4 The OHCO view of text can be contrasted with other models of text generalized from the software, practices, or methodologies that embody them: bitmaps (raster images), characters and formatting commands (procedural markup), glyphs and white space, character transcripts (the so-called 'ASCII' or 'text only' form of a document), and layout hierarchies.(9) In both text processing and textbase development the superiority of the OHCO approach over these other models can be shown easily -- as we saw in the previous section it is by far the simplest and most functional way to create, modify, and format texts; it is required to effectively support text retrieval, browsing, analysis, and other sorts of special processing; and texts represented according to this model are much more easily shared among different software applications and computing systems. The motivation for the view that texts are hierarchies of content objects lay initially in reflecting on these practical benefits of treating texts *as if* they were ordered hierarchies of content objects. #5.1.5 Although claims about the nature of text were initially motivated by practical considerations, once the explicit hypothesis that "text is an ordered hierarchy of content objects" had been formulated, corroborating arguments of various kinds were adduced.(6) In the writings and conversations of the text encoding community in the 1980s and early 1990s at least these six broad categories of arguments that text is an ordered hierarchy of content objects can be discerned. (These are discussed in more detail in Renear et al forthcoming). #5.1.6 Pragmatic/Scientific. These arguments begin, as described above, with the observation that there are many practical advantages to modeling a text as an OHCO rather than using one of the alternative models. Texts modeled as OHCOs are easier to create, modify, print according to varying specifications, transfer from one application to another, analyze, reorganize, retrieve data from, and so on. The comparative efficiency and functionality of treating texts *as if * they were OHCOs is best explained, according to this argument, by the hypothesis that texts *are * OHCOs. #5.1.7 Empirical/Ontological. These begin by observing that content objects and their relations figure very prominently in our talk about texts, and specifically in our descriptions, explanations, theories, hypotheses, and generalizations about texts. For instance, our theories and conjectures about literature make use of terms for chapters, titles, sections, paragraphs, sentences, footnotes, stanzas, lines, acts, scenes, speeches, etc. These have prominent explanatory roles in our talk about texts and in our theorizing about texts and related subjects such as authorship, literary history, criticism, poetics, and so on. If we resolve ontological questions by looking to the nominal phrases in our theoretical assertions, then we will conclude that such things -- chapters, verses, lines, etc. -- exist and are the components of texts. The persuasiveness of these arguments is increased by the fact that theories from many diverse disciplines (and even conflicting theories) are committed to content objects as fundamental explanatory entities. #5.1.8 Metaphysical/Essentialist. This is the classic argument from hypothetical variation, used to distinguish essential from accidental properties in scholastic metaphysics, or, in a more contemporary philosophical idiom, to establish 'identity conditions' for objects. Here one argues that if a layout feature 'of a text' (such as leading or typeface) changes, the 'text itself' still remains essentially the same, but if the number or structure of the text's content objects changes -- say the number of chapters varies or one paragraph is replaced by another -- then it is no longer, strictly speaking, 'the same text'. #5.1.9 Productive Power. Here one finds a special significance in the fact that an OHCO representation of a text can mechanically generate other competing representations (e.g. an OHCO representation can be formatted into a bitmap image, plain text, or vector graphics) but none of these other representations can mechanically generate an OHCO representation. (This is closely connected with the pragmatic/scientific arguments: that OHCO representives are richer in information in this way means that they are more useful, more efficient and effective, for managing text processing). #5.1.10 Conceptual Priority. These are arguments to the effect that understanding and creating text (e.g. reading and writing) necessarily requires grasping the OHCO structure of a text, but does not essentially involve grasping any other structure -- and therefore it is the OHCO structure, but no other structure, that is essential to textuality. #5.1.11 If the forgoing arguments are good ones then the OHCO thesis: 1) explains the success of certain representational strategies 2) is logically implied by many important theories about text 3) matches our intuitions about what is essential and what accidental about textual identity 4) is richer in relevant content than competing models 5) matches our intuitions about what is essential of textual production and consumption. (10) 5.2 PHASE 2: PLURALISM #5.2.1 When researchers from the literary and linguistic communities began using SGML in their work, the implicit assumption in SGML that every document could be represented as a single logical hierarchical structure quickly created real practical problems for text encoding projects (Barnard et al. 1987). Briefly the difficulty is that while the SGML world assumed that text encoders would always represent the logical structure of a text as a single hierarchical structure, there in fact turned out to be many hierarchical structures that had reasonable claims to be 'logical'. A verse drama for instance contains dialogue lines, metrical lines, and sentences. But these do not fit in a single hierarchy of non-overlapping objects: sentences and metrical lines obviously overlap (enjambment) and when a character finishes another character's sentence or metrical line then dialogue lines overlap with sentences and metrical lines. (11) #5.2.2 Part of the problem might be that there is no univocal sense of 'text', 'book', or 'document' and that consequently these words do not, without further qualification, designate genuine 'natural kinds' that play reliable explanatory roles in explanations and descriptions of the world. Instead, they have seem to have many different senses that play very diverse theoretical roles in our discourse. It appears that taking one particular sense of textual identity as being privileged was what led the SGML community to assume that there was one logical hierarchy for each document. In particular SGML assumed that it was the hierarchy of *editorial *objects, as determined by genre, which was constitutive: 4.102 document type: A class of documents having similar characteristics; for example, journal article, technical manual, or memo. #5.2.3 TEI researchers, however, found that there seemed to be many "logical" hierarchies that had equal claim to be constitutive of the text. Thus where the original OHCO Platonists, and the designers of SGML, took the editorial hierarchy of genre to be the original principal of decomposition, the literary scholars of the TEI took the structures elicited by specialized disciplines and methodological practices to be equally important. In a sense a play might seem to be, fundamentally, a hierarchy of acts, scenes, and speeches; but it also consists of a linguistic hierarchy of phrases and sentences; a prosodic hierarchy of stanzas, metrical lines, and feet, and other hierarchies as well -- none of which could be assumed to be capable of assimilation into a single hierarchy of non-overlapping objects. #5.2.4 It seemed that OHCO Platonism needed to be modified. The natural modification, still in the realist and structuralist spirit, is to see texts not as single ordered hierarchies, but as a *system *of ordered hierarchies. Each hierarchy corresponds to an *aspect * of the text and these aspects are revealed by various "analytic perspectives", where an analytical perspective is, roughly, a natural family of methodology, theory, and analytical practice.(12) Each analytical perspective -- e.g. prosodic, linguistic, dramatic -- on a text does seem to typically determine a hierarchy of elements. The doctrine affirms the following hierarchy-saving principle: AP-1: An analytical perspective on a text determines an ordered hierarchy of content objects. #5.2.5 AP-1 does seem to reflect actual text encoding practices in the literary and linguistic text encoding communities: usually when text encoders find overlapping objects they assume that they pertain to different aspects of the text. #5.2.6 But does every perspective really determine a hierarchy of content objects? There is a quick way of casting doubt upon AP-1. Discussions of many sorts about texts are filled with characterizations, descriptions, and hypotheses that explicitly relate text objects from different perspectives -- chapters and themes, speaker and meter, narrative and paragraphing. Moreover there are even technical terms, such as "enjambment" and "caesura", that specifically refer to relationships between objects from overlapping families. Because a technical vocabulary can be plausibly considered a sign of an analytical perspective the existence of this terminology suggests that some analytical perspectives contain overlapping objects. #5.2.7 This suggests: AP-2: For every distinct pair of objects x and y that overlap in the structure determined by some perspective P(1), there exists diverse perspectives P(2) and P(3) such that P(2) and P(3) are sub-perspectives of P(1) and x is a object in P(2) and not in P(3) and y is an object in P(3) and not in P(2). where: *x is a sub-perspective of y * if and only if x is a perspective and y is a perspective and the rules, theories, methods, and practices of x are all included in the rules, theories, methods, and practices of y, but not vice versa. #5.2.8 Our simple Platonic model of text -- an ordered hierarchy of content objects -- has given way to a system of concurrent perspectives which decompose indefinitely into concurrent sub-perspectives, with suspiciously baroque complexity. #5.2.9 Moreover, Claus Huitfeldt has pointed out that despite the apparent hierarchical tendency within analytical perspectives, not only is there no assurance that decomposition into ultimate sub-perspectives without overlaps is possible, but we can easily demonstrate that it is not possible: possible element tokens in some perspectives clearly overlap with other element tokens *of the same type*. Examples of this are strikeouts, versions, and phrases (in textual criticism), narrative objects in narratology, hypertext anchors and targets, and many others. #5.2.10 Hierarchical structures are common and compelling and I think that much remains to say about their significance in our theories about the world. However the counterexamples are decisive and the OHCO Platonist view of text as a single clean hierarchy of fundamentally similar content objects must be rejected in favor of pluralism realism: the doctrine that texts are systems of (not necessarily hierarchical) structures of objects.(13) 5.3 PHASE 3: ANTIREALISM #5.3.1 Modifying OHCO Platonism to Pluralism foregrounds the role that disciplinary methodologies and analytic practices play in text encoding. It cannot be a surprise that focusing attention on the role of theories, methodologies, and other analytical practices in analyzing text structure, and then observing that complex and sometimes conflicting variety of structures that result, has led some text encoding theorists to see text encoding not as merely identifying the features of a text but as itself playing a more constitutive role. This view, although already not rare in conversation, has not often been clearly articulated in the text encoding literature. However, Alois Pichler, a researcher at the Bergen Wittgenstein Archives who is the author some of very important and philosophically illuminating papers on transcription, has written a number of passages that seem to me to be clear enough expressions of antirealism. For instance: "Machine-readable texts make it ... clear to us what texts are and what text editing means: Texts are not objectively existing entities which just need to be discovered and presented, but entities which have to be constructed. (Pichler 1995b p. 774) #5.3.2 Pluralistic realism allowed that there are many perspectives on a text, but assumes that texts have the structures they have independently of our interests, our theories, and our beliefs about them. The antirealist trend in text encoding, which is consistent with post-modern epistemologies such as constructivism and neo-pragmatism, rejects this view, seeing texts as in some sense the product of the theories and analytical tools we deploy when we transcribe, edit, analyze, or encode them. It is interesting that just as Landow (1992), Bolter (1991), and Lanham (1993) have claimed that electronic textuality, and particularly hypertext, confirms certain tenets of post- modernism, Pichler and others are suggesting that text encoding and the creation of electronic editions also confirms a post-modern view of texts: texts do not exist independently and objectively, but are constructed by us. But just as I have argued that Landow, Bolter, and Lanham, do not really establish their claims (Renear forthcoming), I will similarly argue that neither do Huitfeldt, Pichler, and the other post-modern text encoders. #5.3.3 The antirealism of the above passage is ontological, that is, it is a view of the nature of text, what text is. But Pichler also endorses a companion antirealism that is epistemological (or perhaps better, *semantic*), a view about the nature of our knowledge of texts and our representations of text: ... the essential question is not about a true representation, but: Whom do we want to serve with our transcriptions? Philosophers? Grammarians? Or graphologists? What is "correct" will depend on the answer to this question. And what we are going to represent, and how, is determined by our research interests ... and not by a text which exists independently and which we are going to depict. (Pichler 1995a p. 690) #5.3.4 Our aim in transcription is not to represent as correctly as possible the originals, but rather to prepare from the original text another text so as to serve as accurately as possible certain interests in the text. (Pichler 1995a p. 691. #5.3.5 I think that much of the groundwork for the antirealist tendencies I am noting here in Pichler can be found in Claus Huitfeldt in a number of discussions, talks, and articles in the early 1990s (Huitfeldt 1995). Huitfeldt presented a number of criticisms of the OHCO Platonism of the early SGML and early TEI encoding theorists. These criticisms have considerable force and ingenuity and helped reveal the problems in both OHCO platonism and strictly hierarchical pluralism; they advanced our knowledge of text encoding and I think that some of them have still not received the attention they deserve. Huitfeldt's antirealists tendencies are subtle for the most part: ... I have come to think that these questions [e.g. What is a text?] do not represent a fruitful first approach to our theme ... The answer to the question what is a text depends on the context methods and purpose of our investigations. (Huitfeldt 1995 p. 235) #5.3.6 But here and there they are unmistakable: "devising a representational system that does not impose but only maps linguistic structures" (Coulmas 1989) is impossible (p. 238). #5.3.7 What exactly are the arguments in favor of the antirealist view of texts that Huitfeldt, Pichler, and others believe can be found in the results of text encoding? Both Huitfeldt and Pichler emphasize two particular claims about text and seem to see somewhere in these claims an argument for antirealism. The first is that our understanding (representation, encoding, analysis, transcription, etc.) of a text is fundamentally interpretational: there are no facts about a text which are objective in the sense of not being interpretational (Huitfeldt 1994, p. 237). #5.3.8 Although he assures us that this does not mean that all facts about a texts are entirely subjective as "there are some things about which all competent readers agree" (Huitfeldt 1994, p. 237). (Mere empirical agreement is, of course, the weakest sort of objectivity, one entirely consistent with the antirealist temperament.) #5.3.9 The second is that there are many diverse methodological perspectives on a text: a text may have many different kinds of structure (physical, compositional, narrative, grammatical) (Huitfeldt 1995, p. 237) #5.3.10 I do believe that these points, like Aquinas's arguments for the existence of God, can seem to be evidence for antirealism to a person already inclined to be an antirealist. But as with St. Aquinas's arguments they too fail to establish their claims. #5.3.11 Although Huitfeldt presents these claims as criticisms of the view implicit in (at least the "early") views of SGML and TEI text encoders, in fact they are common ones within the text encoding community (Sperberg-McQueen 1991)(13). Indeed, my response to any attempt to present these claims as constituting an argument for antirealism is that the they may be true ( I think they are true) but there is no path of compelling argumentation from the either of them, or both together, to the conclusion that "texts are not objectively existing entities". #5.3.12 The first claim is that representation and transcription is interpretational at every level. Assuming that "interpretation" here means, roughly, inference, what is the significance of the claim that our knowledge of text is inferential? This missing premise would be that inferred entities aren't real, or, more charitably, entities that could only be inferred, in principal, aren't real. That is, the argument depends on an extreme form of postivism. If we are postivists, we will probably be antirealists, but no help from our experiences with text encoding is necessary for that (short) journey. #5.3.13 The second claim is that there are many diverse analytic perspectives on a text. Here the premise needed to get to the antirealist conclusion would be something to the effect that because what we find in the world is determined at least in part by our interests and by the methodologies and theories we choose to deploy in the course of inquiry, that therefore what there is in the world is determined by our interests, theories, and methods. This is a more stylish antirealism than the postivist variety, but still implausible to anyone not already convinced of missing premise. And, again, there is nothing in our experience with text encoding that add any special motivation for accepting this argument. #5.3.14 The above comments pertain directly to the ontological antirealism of the antirealist encoders. Epistemological antirealism in text encoding probably derives in part from the ontological, as so if my response to ontological realism removes the support it gives to epistemological antirealism. But as epistemological antirealism can also draw some support from other sources we might consider these antirealist claims indepedendently. #5.3.15 Pichler writes: Our aim in transcription is not to represent as correctly as possible the originals, but rather to prepare from the original text another text so as to serve as accurately as possible certain interests in the text. (Pichler 1995a p. 691). #5.3.16 I would argue that the apparently antirealist formulations of this claim are either 1) non-antirealist truisms, 2) (covertly realist truths), or 3) false. It is certainly true that our aim in transcription is to help researchers and that this guides our encoding decisions; if this is what Pichler is saying it is a truism of encoding. But if he is saying that truth doesn't matter to encoders, than he is saying something false. Suppose a transcription has nothing at all to do with the text but helps the researcher win a prize which allows her, by providing money and time, to figure out, finally, the crux she was working on. In that case a (false) transcription has served the researchers interests quite well, but no one would claim that it is thereby a reasonable encoding or one which is to be in any sense commended as a transcription. And this is because although the researcher's interests were served, they were not served i*n the right way.* (The problem is similar to other familiar ones with positivist accounts of meaning and causal accounts of justification: I want my evidence causally linked to my conjectures, but in the right way; and I want my hunger satisfied by a meal, not a blow to the head.) What the researcher wants is a transcription that helps her, yes, but helps her in a certain way: by providing a relevant and *truthful *representation of the text. #5.3.17 Antirealism may indeed be the appropriate attitude to take towards theories of textuality, but whether or not it is so I don't think Huitfeldt or Pichler have shown that there is any motivation for antirealism in our experiences with text encoding. 6.0 CONCLUSION #6.0.1 I should probably confess to a little willful and single-minded obtuseness. In their articles both Huitfeldt and Pichler are making very profound observations about text encoding, transcription, and texts. I have ignored coming to grips with what is deep and perhaps correct in their arguments in order to make my own -- probably rather less important -- points about their apparent (and in the end it may be only apparent) antirealism. But regardless of their intentions, and my obtuseness, I think that what they say, particularly against the current post-modern background, is indeed raising antirealist questions about the nature of theories of text and text encoding. I think it is a sign of the maturity of text encoding theory that it is now becoming engaged in the central intellectual controversies of our time. 7.0 NOTES The leading quotation is from *Culture and Value * (*Vermischte Bemerkungen*), edited by G.H. von Wright and H. Nyman and translated by Peter Winch, Basil Blackwell 1980 p 64e. (It may be found on p 541 of vol 8 in the Suhrkamp Werkausgabe from 1989, and on p 126 in the *Neubearbeitun*g by Alois Pichler, Suhrkamp 1994.) The Manuscript source is MS 136 p 31a, from 24.12.1947). I'm indebted to the Director of the Wittgenstein Archives for this information. 1 Although they each might find much to disagree with in what follows, this paper nevertheless draws very heavily on ideas and analyses collaboratively developed with James Coombs, Steve DeRose, Elli Mylonas and David Durand (see Coombs et al 1987, DeRose et al 1990, Renear et al 1992 and Renear et al forthcoming). In addition I owe much to conversations with Michael Sperberg- McQueen, Claus Huitfeldt, Alois Pichler, Lou Burnard, and Julia Flanders. Finally, the Text Encoding Initiative itself, and its sponsors, funders, and participants deserve credit for creating an environment which has stimulated and supported these discussions. 2. But note also that the confusing nature of electronic textuality and the computer representation of texts lead the users, if not the designers, of text processing systems, to characterize their experiences in odd and hard to explicate theoretical notions such as "transparent access" and "the text itself". Bringing some order to this sense would be another immediate benefit of theorizing the practices of text encoding and text processing. 3. This taxonomy is from a more comprehensive theory developed by James Coombs (Coombs et al 1987), which is in turn an extension and generalization of Charles Goldfarb's earlier division of text processing markup into procedural and descriptive (Goldfarb 1981). Coombs's analysis treats the markup functions of punctuation and devices like list and page enumeration. He also goes on to analyze markup and markup processing along a number of other dimensions depending on how it is created, stored, and processed. 4. This confusion indicates some of the complexities involved in the analysis of markup and text processing. For a more detailed account see Coombs et al 1987. 5. This section is based directly on DeRose et al 1990 which in turn draws on the categorization of advantages in Coombs et al 1987. 6. Much of the material in this section and the next was developed jointly with Elli Mylonas and David Durand; it is presented in an earlier form, although with much more detail on the problems of hierarchies, in Renear et al forthcoming. 7. In the terminology of graph theory ordered hierarchies are 'ordered, rooted trees'. In linguistic theory the ancestral and ordering relations are often separately described as 'dominance' relations and 'precedence' relations. 8. Elsewhere one will find 'editorial', 'logical' and 'sense' used to mean more or less the same thing we mean here by 'content'. Words used in place of 'object' by other authors include 'part', 'component', and 'element'. 'Element' is in fact the technical SGML term that corresponds to our use of 'object -- however, we continue to use the word 'object' as name for a pre-theoretical notion that may or may not be adequately captured in the technical vocabulary of SGML. 9. For further discussion of these alternative models, and the comparative advantages of the OHCO model, see (Coombs, et al. 1987) and (DeRose, et al. 1990). 10. To be sure, nothing about this view is uniquely motivated by text encoding; versions which are entirely independent of any interest in computing applications can be discerned in the rhetoric of the 'parts of a book' which has been prevalent in style manuals and bibliography handbooks for some time. (Renear, 1993). 11. Again, this section presents material developed jointed with Mylonas and Durand and included in Renear et al (forthcoming). 12. Compare Kuhn's notion of a "disciplinary matrix" (Kuhn p. 182). 13. Mylonas and Durand have argued that these results pose a particular problem for the TEI, because although the TEI has SGML mechanisms for coordinating concurrent hierarchies and other non- hierarchical structures, these mechanisms require and conceptually awkward and inappropriate encoding strategies, suggesting fundamental problems with the TEI ability to handle non-hierarchical structures. (Renear et al, forthcoming) 13. Sperberg-McQueen's article (1991) is a very important one in the small literature of text encoding theory. 8.0 BIBLIOGRAPHY Barnard, David T., Ron Hayter, Maria Karababa, George Logan, and John McFadden. "SGML-Based Markup for Literary Texts: Two Problems and Some Solutions." Computers and the Humanities, 22 (1988): 265-276. Bolter, J. David. Writing Space: The Computer, Hypertext, and the History of Writing, Lawrence Erlbaum Associates. 1991. Lanham, Richard A. The Electronic Word: Democracy, Technology, and the Arts, by Richard A. Lanham. The University of Chicago Press, 1993. Coombs, James S., Allen H. Renear and Steven J. DeRose. "Markup Systems and the Future of Scholarly Text Processing" Communications of the Association for Computing Machinery. 30 (1987): 933-47. Engelbart, D. C. "A Conceptual Framework for the Augmentation of Man's Intellect." Vistas In Information Handling, Vol. 1, P. Howerton, Ed., Spartan Books, Washinton DC. (1963). Engelbart, D. C., R. W. Watson, and J. C. Norton. "The Augmented Knowledge Workshop." In Proceedings of the National Computer Conference (New York, June 4-8), AFIPS Press, Reston Va., 1973, 9-21. DeRose, Steven, J., David Durand, Elli Mylonas and Allen H. Renear. "What is Text, Really?," Journal of Computing in Higher Education. 1:2 (1990). Fortier, Paul, and the TEI Literature Working Group. "The TEI Guidelines (version 1.1 10/90): A Critique," 1995. Available from the TEI listserv server, TEI-L/UICVM, as document AI3W5 DOC. Goldfarb, Charles. "A Generalized Approach to Document Markup."in Proceedings of the ACM SIGPLAN--SIGOA Symposium on Text Manipulation, New York: ACM, 1981. Goldfarb, Charles. The SGML Handbook, Charles Goldfarb. Oxford: Oxford University Press, 1990. Huitfeldt, Claus. MECS -- A Multi-Element Code System. Working Papers from the Wittgenstein Archives at the University of Bergen, No 3, (seen in draft in October 1992). Huitfeldt, Claus. "Multi-Dimensional Texts in a One-Dimensional Medium." Computers and the Humanities vol. 28 (1995): 235-241. Also printed in Paul Henry and Arild Utaker (eds.): Also printed in Paul Henry and Arild Utaker (eds.): Wittgenstein and Comtemporary Theories of Language, Working Papers from the Wittgenstein Archives at the University of Bergen, no 5, 1992: 142-161. International Organization for Standardization (ISO). Information Processing -- Text and Office Systems -- Standard Generalized Markup Language (SGML), ISO8879-1986, International Organization for Standardization (ISO), 1986. Kuhn, Thomas S. The Structure of Scientific Revolutions. Chicago 1962 1970. Lanham, Richard A. The Electronic Word: Democracy, Technology, and the Arts, by Richard A. Lanham. The University of Chicago Press, 1993. Landow, G. Hypertext: The Convergence of Contemporary and Critical Theory. Baltimore: Johns Hopkins, 1992. Mamrak, Sandra A.; Barnes, Julie; Hong, H.; Joseph, C.; Kaelbling, Michael; Nicholas, Charles; O'Connell, Conleth; Share, M. "Descriptive Markup - The Best Approach?" Communications of the Association for Computing Machinery 31/7 (July 1988) 810-811. Pichler, Alois. "Transcriptions, Texts and Interpretations", in Johannessen and Nordenstam (eds.) Culture and Value: Philosophy and the Cultural Sciences, The AustrianLudwig Wittgenstein Society (1995) Pichler, Alois. "Advantages of a Machine-Readable Version of Wittgenstein's Nachlass", in Johannessen and Nordenstam (eds.) Culture and Value: Philosophy and the Cultural Sciences, The AustrianLudwig Wittgenstein Society (1995) Quine, Willard van Orman. "On What There Is"in From a Logical Point of View . Cambridge Mass: Harvard, 1953. Raymond, Darrell R.; Tompa, Frank William; Wood, Derick. "Markup Reconsidered". Department of Computer Science, Technical Report No. 356. The University of Western Ontario, 1993. An early version by Raymond was privately circulated as "Markup Considered Harmful." in the late 1990s. Reid, Brian. "A High-Level Approach to Computer Document Formatting." in Proceedings of the 7th Annual ACM Symposium on Programming Languages. New York: ACM, 1980. Renear, Allen. "Representing Text on the Computer: Lessons for and from Philosophy." Bulletin of the John Rylands University Library of Manchester, 74 (1992): 221-48. Renear, Allen. David Durand, and Elli Mylonas. "Refining Our Notion of What Text Really Is."Research in Humanities Computing. Oxford: Oxford University Press, forthcoming. Rohr , P., "The TextBase Paradigm: Architectural Considerations for a Second-Generation Scholar's Workstation" Senior Thesis, University of Chicago, (seen in draft, 1991). Sperberg-McQueen, C. Michael. "Text in the Electronic Age: Textual Study and Text Encoding, with Examples from Medieval Texts." Literary and Linguistic Computing, 6:1 (1991). Sperberg-McQueen, C. Michael. and Burnard, Lou, eds. Guidelines for the Encoding and Interchange of Machine-Readable Texts.. Chicago and Oxford: TEI, 1990, 1993. Wittgenstein, Ludwig. *Culture and Value * (*Vermischte Bemerkungen*), edited by G.H. von Wright and H. Nyman and translated by Peter Winch, Basil Blackwell 1980 p 64e.