The representation of symbols in an index Hans H. Wellisch A section from Wellisch's newly published guide, Indexing from A to Z, on the form, order, and place in which to represent symbols in an index. Symbols are tricky things: their number is unlimited, be the graphic representation of the number one or of a the same symbol may mean many different things, the high frontal lax vowel (as in 'bit'), i.e., the letter I; it may same thing may have many different symbols, they are not represent the chemical element iodine or indicate electrical easy to define, and they are difficult if not impossible to current; or it may be a symbol for several other things—all index—yet. indexed they must be if they appear in the text depending on the context in which the symbol appears. of a document. Only a little more than a decade ago, index- The most frequently used graphic symbols are, of ers were not overly much concerned with symbols. In his course, numerals and letters, but their treatment in indexes textbook on indexing, G. N. Knight still thought that 'It is not easy to imagine circumstances in which a symbol may is taken care of by long-standing rules of mathematics and have to be used as the keyword of a heading', but allowed problems, namely: in which form, or in what order, and that a symbol 'might well come as the second word and so where to represent them in an index. affect the order where there are several headings with the same keyword'.1 It was, of course, well known that mathe matical, astronomical, chemical, and other scientific texts filing, respectively. All other graphic symbols pose three Form of representation contained symbols that sometimes had to be indexed (and All symbols have names or designations, otherwise it Knight gave a few examples of these), but, even in those would be impossible to talk or write about them, or to types of text, symbols were generally not very numerous explain what they stand for. The names are different and could more or less easily be dealt with in an index. depending on the field of application, but within one such Little did anyone in the late 1970s foresee the proliferation field any symbol can, in principle, have only one name: the of symbols in the technical manuals for computer software symbol O always means only the 15th letter of the English which must now be indexed to guide users to the many dozens of different symbols for commands, instructions, alphabet, and no other; it signifies only zero, and no other and applications, ranging from punctuation marks to icons. istry and physics; only one blood type, and so on. (The dif numeral; it means oxygen, and no other element in chem ferent names for an oblique stroke in writing and printing—solidus, shilling, virgule, fraction, stroke, bar— What is a symbol? are an unfortunate exception, rarely if ever found elsewhere A symbol is something representing something other in the various symbol nomenclatures.) Since users of a text than itself by association, resemblance, or convention*. For in which symbols are discussed qua symbols may be the purposes of indexing, only graphic symbols, or graphemes in linguistic terminology, need to be consid expected to know their names within the given context, the ered. Thus, a short vertical stroke, in and of itself, is not a symbol is its name or designation, spelled out in full, fol symbol; it becomes one when it is associated in some lowed by the symbol itself in parentheses and by its loca body's mind with another thing for which it may stand as a tors, e.g., kind of shorthand expression. The short vertical stroke may * This definition, from the American heritage dictionary of the English language, avoids mentioning the term sign; the often con troversial distinctions made in philosophy and linguistics between sign and symbol are irrelevant for written communication. easiest and most straightforward form of entry for any infinity (°°) 37,42 Such an entry should, of course, be made only if the symbol for infinity is discussed in the text, and not for the concept of infinity which, if also dealt with in the same text, should have a normal entry, with that for the symbol as a subheading, e.g., From Indexing from A to Z. Copyright © 1991 by Hans H. Wellisch. Reprinted by permission of the H. W. Wilson Company, New York. All rights reserved. The Indexer Vol. 17 No. 4 October 1991 infinity 80, 84, 96 symbol (°°) 37, 42 239 THE REPRESENTATION OF SYMBOLS This method fails, however, to do two things: it does for an index of proof correction marks. This method, while not help a user who remembers the shape of a symbol but does not know or remember its name; and it is not helpful easy to explain to users, will necessarily disperse related for many software symbols, which either signify a phrase all indexes. It is probably best to use it when the number of (e.g.. an instruction to do something) or are icons whose function cannot always easily be expressed by a succinct heading. Consequently, there is a need to display the sym bols themselves, which in turn raises the question of their symbols, but so does alphabetical arrangement of entries in symbols to be displayed is small, say, no more than a dozen or so. Classified order. All symbols referring to a particular class are displayed together, further subarranged either alphabetically or in an order readily recognized by users arrangement. familiar with the discipline or field, e.g., icons for editing, drawing, coloring, etc., each group arranged by name or Order of representation instruction; another example would be symbols for the The virtually infinite variety of shapes in which graphic symbols may appear effectively precludes any order that assures in and of itself that any uninitiated user would automatically find a particular symbol within an array of planets which, if only a few, may be subarranged alphabet ically from Earth to Venus (which would be abhorrent to astronomers) or, if all nine appear, in the order of their dis tance from the Sun, that is, differently shaped ones, the way we all know that the letter M is somewhere in the middle of the Roman alphabet, more 3 Mercury precisely between the letters L and N, and in no other place. 9 Venus The same is true of any number written in Arabic numerals in the decimal system. This is so because every literate person has to learn these ordering systems by heart as a child, and retains this knowledge throughout life. No such universally known and recognized ordering system exists for graphic symbols, yet users may want to find symbols in indexes to texts dealing with astronomy, biology, chemistry, geology, engineering, the military, linguistics, genealogy, and heraldry, to name only a few fields that employ symbols frequently, not to mention arbitrary symbols sometimes invented ad hoc by authors for certain concepts or purposes, as well as icons, the symbols used on monitors to produce pictorial effects. Icons are discussed at length in software manuals, and there are often several dozens of them; but their shapes are more often than not quite arbitrary and their meanings are not always immediately obvious, nor do they always have succinct names or descriptions. For example, in a particular software application, the image of a pencil in an icon is indeed named 'pencil' and is used to draw lines, but the image of a butterfly net is called a 'selection net' and is used to 'select a non-rectangular drawing area for editing', something that can only be learned from the printed manual and so is not common knowledge. It is therefore necessary to show symbols in an index in their physical shapes, and three possible methods may be used for this purpose, depending on the nature of the text and the amount of symbols to be displayed: Alphabetical order. Each symbol is first displayed in its graphic form in a separate list, followed by its name or description in alphabetical order, e.g., II align 12 U1 brackets 17 C close up 10 A delete 9 240 © Earth 6 Mars % Jupiter h Saturn & . Uranus \y Neptune e Pluto Authorized list order. Symbols are listed in the order in which they are found in authoritative sources, such as major dictionaries or style manuals. A small selection of symbols, mainly of interest in computer contexts, is covered by two different ordering sequences which are more or less familiar to many (but by no means all) computer users. One is the American standard code for information interchange (7-bit ASCII)2, whose acronym ASCII is pronounced 'askee'; it is virtually identical with the international standard ISO 646.' The system assigns a three-digit code to each numeral and letter, as well as to a few dozen symbols, e.g., the numeral 0 is 048,9 is 057, A is 065 and i is 097, a dollar sign is 036, and a comma is 044, and so on for a total of 256 codes. ASCII codes are predominantly employed in the realm of microcomputers, and if a text deals only with symbols covered by these codes it makes sense to list symbols in ASCII order, because the prospective users of manuals and guidebooks on microcomputer software can reasonably be expected to be familiar with it. Thus, a sequence of entries for commands occurring in a software manual, where symbols are the only distinguishing elements in otherwise homographic headings, may be arranged in ASCII order, e.g.. The Indexer Vol. 17 No. 4 October 1991 THE REPRESENTATION OF SYMBOLS out; for single symbols (that is, those not preceded or allowed by a numeral or letters), the explication should be enclosed in brackets. Unmodified ASCII codes as commonly embedded in word-processing software should, however, never be used for the alphabetization of verbal index headings, because this would play havoc with their proper sequence: e.g., a term enclosed in quotation marks would precede all numerically and alphabetically arranged terms, and all upper-case letters would sort before lower-case letters. Specialized indexing software provides modification of ASCII, resulting in correct alphanumeric sorting.) The other system for the arrangement of symbols used in computer contexts is the Extended Binary Coded Decimal Interchange Code, fondly known as EBCDIC, preferred by IBM for texts dealing with mainframe computer software. Symbols in indexes to such texts may therefore be arranged in that order. Since the curse of Babel has struck also man-made computer languages and their appurtenant ordering systems, there is no compatibility between ASCII and EBCDIC, nor do they list the same set of symbols. Neither ASCII nor EBCDIC cover icons. Whichever system is used to arrange graphic symbols in an index, an introductory note must make it clear which one has been chosen, and a double entry in the verbal part of the index, followed by the image of the symbol, should always be made, e.g., 9 Venus 48, 62 and Venus ( 9 ) 48, 62 If none of the three ordering methods is suitable, symbols must be arranged arbitrarily, e.g., in the sequence in which they appear in the text. Filing of symbols Where to display graphic symbols—either before all alphanumeric entries or after them—is not really a crucial issue. Most indexes containing symbols seem to prefer their placement at the head of the index, but there are also many examples of symbols being shown after all verbal headings, and no amount of pseudological quibbling will result in convincing arguments for or against one placement or the other; it matters little, so long as it is explained to users in The only exception to the rule of disregarding symbols in filing is the symbol &, the ampersand, which, when not disregarded, may optionally be filed as if it were spelled out in the language of the heading. References 1. Knight, G. Norman. Indexing, the art of. London: Allen & 2. American National Standards Institute. Information systems - Unwin, 1979. 128 coded character set - 7-bit American national standard code for information interchange (7-bit ASCII}. New York: ANSI, 3. 1986 (ANSI X3.4) International Organization for Standardization. Information processing: ISO 7-bit coded character set for information interchange. Geneva: ISO, 1973. (ISO 646) Hans H. Wellisch is Professor Emeritus at the College of Library and Information Services of the University of Maryland; author of Indexing and Abstracting: an interna tional bibliography (1980) and Indexing and Abstracting 1977-81 (1984); President of the American Society of Indexers 1984-5. an introductory note. Regarding symbols as parts of verbal headings, all post- What can we say? 1980 filing rules agree that such symbols (including punc tuation marks) should be disregarded and the headings filed "\ would like to plead for an Index in every Excavation as if the symbols did not.exist. This will primarily affect quaint pseudonyms that were fashionable in the 18th and 19th centuries, as well as some quirky titles of books, articles, or poems; neither type of heading is very numer Report. This is a very simple task, and it can only be done properly by the writer of the book or monograph. It helps to avoid repetition in a text; nor does it take time, two or three hours at most.' ous. In all cases where the filing of such headings will result in ambiguous or unfamiliar filing positions, double entries should be made in which the symbols are spelled The Indexer Vol. 17 No. 4 October 1991 - Dr Grace Simpson in Journal of Roman Pottery Studies 3 1990. 241