"In the world of honest scholarship, no rule is more revered than the citation." -- Kavita Varma
Responding to the needs of writers, teachers, students, and scholars using the MLA style of documentation, professional organizations and researchers are in the process of creating citation conventions that seek to demonstrate efficient and unambiguous reference to Internet sources. Going beyond the limited recommendations in The MLA Handbook for Writers of Research Papers (1995), Janice Walker's "MLA-Style Citations for Electronic Sources," for example, presents a largely admirable style for documentation, worthy of endorsement by the Alliance for Computers and Writing. In spite of its usefulness, however, a close examination of the style reveals several noteworthy ambiguities and infelicities. Four problematic areas can be identified: (1) distinguishing between Internet addresses and command sequences, (2) distinguishing between dates of publication and dates of user access, (3) distinguishing between addresses of publication and addresses for retrieval, and (4) providing for the "openness" of many Internet pathways to the same (authentic) source. After examining each of the four problem areas, we recommend additions and emendations to Walker's models that help eliminate ambiguities. Each of our recommendations improves the usefulness and clarity of individual models. Finally, in "Citing the Sites: MLA-Style Guidelines and Models for Documenting Internet Sources," we provide a style sheet for teachers, students, researchers, and scholars whose citation standards are based on MLA-style documentation principles and models. "Citing the Sites" may be distributed as a self-contained guideline providing checklists and models for all current instances of Internet citation.
When a recent USA Today headline announces that scholars are struggling to maintain "standards in cyberspace" for the creation of footnotes (Varma), one realizes that the usual concerns of pedantic instructors teaching documented writing have reached interesting levels of national attention. When Writing Program Administrators receive phone calls from colleagues asking for help on how to cite locations on the Internet, one begins also to realize that instructors, students, and scholars, returning from cyberspace, are in fact grappling with new problems about crediting information sources located in webbed, digital environments. Indeed, as researchers in the humanities and arts find it necessary to document sources of information that appears flickeringly on a monitor's screen, perhaps "here today and gone tomorrow," the need for sound citation advice beyond that within the MLA Handbook is real.
All of which is not to deny that the fourth and latest edition of the MLA Handbook for Writers of Research Papers is not helpful in citing some electronic sources. Section 4.8 and 4.9 provide useful descriptions of and models for documenting "CD-ROMS and Other Portable Databases," along with on-line databases accessed through commercial computer services such as "CompuServe, Nexis, OCLC, and Prodigy" or the Internet as sources. Moreover, the MLA Handbook is useful for its suggestions on citing literary and historical documents available online while, at the same time, cautioning that one must "remember that not all texts are equally reliable or authoritative" (4.9.3a). For guides to other "numerous resources on computer networks," the MLA Handbook recommends popular books such as Ed Krol's The Whole Internet, in addition to "important online sources" found in electronic journals, newsletters, and moderated conferences and discussion lists. But the MLA Handbook repeatedly acknowledges that there will be instances when "no specific printed source or printed analogue" (4.9.2) is available and one must therefore cite "what is available," while trying to include, when possible, the author's name, title of the material (in quotation marks), date of the material, database title, publication medium (Online), name of the computer service, and date of access. And that's it. Unlike style manuals for those using the APA-style of documentation (Note 1), the MLA Handbook provides no assistance for referencing information on the World Wide Web (WWW), that is obtained by File Transfer Protocol (FTP), HyperText Transfer Protocol (HTTP), listservs, and telnet sites; it provides no models for texts generated during synchronous conferencing, texts archived and retrieved from GOPHER sites, or for communications posted by e-mail or within newsgroups. Significantly, it models no references to home pages or other hypertext documents.
Student handbooks dependent on MLA guidelines not unexpectedly show a corresponding scarcity of model. The third edition of The St. Martin's Handbook, for example, in its discussion of electronic sources, while admitting that "documenting [such sources] remains tricky," offers less than two full pages of advice (Lunsford and Connors 640-42), all of it an abbreviated version of the MLA recommendations. Other handbooks such as Bonnie Carter and Craig Skates's Rinehart Handbook for Writers and Alfred Rosa and Paul Eschholz's Writer's Brief Handbook are similarly indisposed to offer help of useful scope. The Scott, Foresman Handbook for Writers presents the models suggested by Janice Walker (which are discussed at length below), labeling them "ACW style" because they have been endorsed by the Alliance for Computers and Writing, but advising students to check with individual instructors before using them (671); as argued below, these guidelines are ambiguous and potentially misleading. As of the spring, 1996, no professional style manual or student handbook presents an adequate set of recommendations or display of models as to how one effectively and efficiently documents online sources in a fashion consistent with MLA principles of documentation. (Note 2)
Professional organizations and responsible publishers are, however, aware of the need for citation models. But Lisa Guernsey describes their dilemma:
Standards for print materials, such as formats for publication dates, page numbers, and even punctuation, become confusing when applied to sources on the Internet, which may have no publication date or page numbers, and may require unconventional brackets for punctuation.
The usual authorities, which include professional associations such as the Modern Language Association and the American Psychological Association, are still debating specific standards for scholars in their disciplines. The Chicago Manual of Style, among other reference books, simply refers readers to an international standards organization, which is still struggling to reach agreement on new guidelines after five years of painstaking debate. (A20)
Among those attempting to establish workable conventions, Guernsey describes the work of two researchers whose efforts may prove valuable for those promoting, using, and teaching the MLA-style of documentation. Melvin E. Page, a history scholar and experienced listserv moderator from East Tennessee State University, has created his own set of guidelines. Employing Kate Turabian's Chicago-style citation principles, Page acknowledges in his headnote that his "guide is not intended to be definitive [and that] corrections, additions, comments, suggestions, and criticisms are therefore welcome." Another, and perhaps more widely publicized (at least among teachers of writing) new guide is Janice Walker's "MLA-Style Citations of Electronic Sources," a style sheet endorsed by the Alliance for Computers and Writing. Walker describes the need and how her guide responds:
We need some standardization of citation references, and this compilation is, therefore, an attempt to address this need. As we move out of the Gutenberg era and into the electronic age, more and more work is being published electronically that may never appear in conventional print medium. I have, therefore, compiled a listing of electronically-available information sites that have not been addressed by MLA, attempting where possible to adhere to formats already delineated by MLA. Please note, however, that since punctuation is an important part of many internet addresses, certain conventions of citation punctuation have necessarily been changed and/or adapted (for instance, placing the date in parentheses and omitting trailing periods for internet addresses).
When inventorying sites not addressed by MLA--FTPs, WWW home pages, locations for synchronous communications, GOPHERs, e-mail, listservs, and newslists--Walker recommends the source data (in order) to be given and provides a model entry. Here, for example, is her model for a hypertext document retrieved from the World Wide Web:
To cite files available for viewing/downloading via the World Wide Web, give the author's name (if known), the full title of the work in quotation marks, the title of the complete work if applicable in italics, the full http address, and the date of visit.Burka, Lauren P. "A Hypertext History of Multi-User Dimensions." MUD History. http://www.ccs.neu.edu/home/1pb/mud-history.html (5 Dec. 1994).
Appearing to satisfy the demands the scholarly citation, the model suggests that one might replicate its style without difficulty because it presents its information economically and conveys important data in readable fashion.
Difficulties, however, remain. Although Walker, as Page and others (Note 3), have devised potentially useful ways to accomplish the mechanics of citing electronic sources, nevertheless, some noteworthy ambiguities and infelicities appear. Four problematic areas can be identified: (1) distinguishing between Internet addresses and command sequences, (2) distinguishing between dates of publication and dates of user access, (3) distinguishing between addresses of publication and addresses for retrieval, and (4) providing for the "openness" of many Internet pathways to the same (authentic) source.
One of the most interesting peculiarities of citations for electronic sources is that the electronic address in the citation is an essential tool for retrieving the document. The convenience of having exact references to electronic sources so readily available, however, presents writers with a special burden regarding absolute precision. Extraneous spaces, missing spaces, errors in upper/lower case characters, and misleading symbols in addresses will usually cause failure in retrieval. This principle necessitates some deviations from ordinary citation typography, such as in the "normal" use of periods and other punctuation marks. As mentioned earlier, Walker's models recognize the need to omit "trailing periods for internet addresses," which is one way to correct a serious defect in the citation styles advocated by Li and Crane. One area of inadequate coverage in the current style guides, however, concerns the distinction between accessing an Internet document through its Uniform Resource Locator (URL) and through a sequence of commands, directory paths and/or menu selections. Walker's citation for "Burka," shown above, contains a typical URL--"http://www.ccs.neu.edu/home/lpb/mud-history.html"--while her examples for citing Telnet and Gopher documents illustrate the command-pathway structure:
Gomes, Lee. "Xerox's On-Line Neighborhood: A Great Place to Visit." Mercury News 3 May 1992. telnet lambda.parc.xerox.com 8888, @go #50827, press 13 (5 Dec. 1994). Quittner, Joshua. "Far Out: Welcome to Their World Built of MUD." Published in Newsday, 7 Nov. 1993. gopher /University of Koeln/ About MUDs, MOOs and MUSEs in Education/Selected Papers/newsday (5 Dec. 1994).
While the URL specification is rapidly becoming a universally accepted standard for Internet addresses, the command-pathway mode will be encountered frequently for some time to come, since many well-known programs such as Gopher, Telnet, and FTP are not designed to accept or report URLs. Some citation stylists have suggested prefixing the expression "URL:" to identify such strings, but with the gradual acceptance of angle brackets as standard delimiters for URLs and for e-mail addresses, such prefixing is "'ugly and unnecessary'" (Michael Quinion, qtd. in Wainwright).
Walker avoids the problem of showing extraneous periods in document citations simply by leaving them out, using white space alone to separate the electronic address from other bibliographic elements. This eliminative approach is not possible with other troublesome characters such as spaces and slashes. Spaces are forbidden in UNIX file names and in URLs, but are frequently encountered in command-pathway expressions as in the "Gomes" and "Quittner" citations shown above. There is no concise and simple way to show in printed text which spaces are significant and which are not. Line breaks and indentations create additional trouble because their "white space" gives no clue as to whether a "space" character is implied. Furthermore, file naming conventions prevent the use of hyphens, slashes or other common symbols to indicate continuation without a space. The best solution appears to be to delimit (enclose) a continuous string with some specially selected characters, and to leave command-pathway expressions as free-flowing text. The Internet standard for URL delimitation uses angle brackets (Berners-Lee et al.), and there are no good reasons to substitute square brackets or parentheses (as in Hoemann) for this purpose. If angle-bracket delimitation becomes the accepted standard, it will again be possible to use periods "as usual" in printed citations to separate bibliographic elements; that is, Walker's "Burka" citation might become:
Burka, Lauren P. "A Hypertext History of Multi-User Dimensions." MUD History. 1993. <http://www.ccs.neu.edu/home/lpb/ mud-history.html> (5 Dec. 1994).
The slashes in a URL are significant characters in that that the characters themselves are part of the electronic address. The "Quittner" citation above shows how slashes are also used by convention to separate the user's on-screen menu selections; in other words, users are expected to know that such slashes are not characters to be typed. Here again, the adoption of angle-bracket delimiters for URLs--while leaving command-pathways as open text--helps to reduce ambiguity. These difficulties should disappear when URLs become the sole standard for citation practice.
Many online documents possess an attribute we might call "invisible revisability," the process by which an electronic document can be altered, moved, or deleted by the author or by computer system managers, without any publically-accessible trail of evidence. For example, many World Wide Web authors state candidly that their pages are under constant construction, which means that the information contained there might change frequently and without notice. Some authors report a "last changed" date for each document, but very few include edition numbers or revision histories. While it is true that many other electronic documents are "archived" in ways that make their texts secure--archived messages from an e-mail discussion list, for example--the archives themselves are subject to changes in access path or timed deletion. The presence of "invisible revisability" as an attribute of online documents requires therefore that their scholarly citation refer to the user's date of access as well as any publication date.
The available citation guides show a wide variation in appreciating the necessity for citing both publication and user-access dates. Li and Crane provide space for both elements; Walker recognizes the need to specify dates of access for most items, but provides no consistent direction for including publication dates. Page, on the other hand, ignores user-access dates, permitting some absurd juxtapositions. Here is Page's suggested citation for a Gopher copy of the 1860 Democratic Platform: (Note 4)
"Democratic Party Platform, 1860." <wiretap.spies.com>. [Path: Wiretap Online Library/Civic & Historical/Political Platforms of the US/Democratic Platform 1860]. 18 June 1860.
While the publication date is clearly an important bibliographic element, its position in this format leaves it too closely associated with the command-pathway expression. The Walker-style analogue, with the publication date included, would be:
"Democratic Party Platform, 1860." 18 June 1860. gopher wiretap.spies.com /Wiretap Online Library/Civic & Historical/ Political Platforms of the U.S. (25 Feb. 1996).
Such a display of information not only records both publication and user-access dates, but it does so without confusion.
Walker's citation style focuses attention mainly on user-access dates, but there is presently an inconsistency in her models for citing e-mail correspondence, where the date of the message is shown in the location used in all other cases for the user-access date:
Seabrook, Richard H.C. "Community and Progress." email@example.com (22 Jan. 1994).
Similarly, Li and Crane surprisingly advise placing the "Date of message" at the end of a citation for personal e-mail, while the date for discussion-list messages is placed immediately after the subject and the user-access date at the end. Page, showing no concern at all for user-access dates, places the publication date at the end of every citation.
This confusion is not merely pedantic because the successful retrieval of list messages (for example) depends on knowing accurately the date of the message. If we combine the productive suggestions of Walker and of Li and Crane and delimit the Internet address with angle brackets, the citation for "Seabrook" would be as follows:
Seabrook, Richard H. C. "Community and Progress." 22 Jan. 1994. <firstname.lastname@example.org> (22 Jan. 1994).
Clarity is enhanced by associating the message date closely with the author and subject, and leaving the user-access date (if any) for the final element. For reasons to be explained in the next section (3.), however, this citation is incomplete. Including the essential data for retrieval (as advocated here), the final form for "Seabrook" looks like this:
Seabrook, Richard H. C. <email@example.com> "Community and Progress." 22 Jan. 1994. <firstname.lastname@example.org. virginia.edu> via <email@example.com> (22 Jan. 1994).
A document's URL designates both its storage location--the "site" of its publication--and the means for its direct retrieval. Correspondence to an e-mail discussion list does not fit this model, however, because the address for the list is ordinarily different from the address for archive retrieval. None of the existing style guides appear to recognize this distinction or provide a way to specify both addresses in a citation. Through failing to recognize the separate functions of these addresses, some citation stylists are advocating formats that are strongly misleading, as when Li and Crane direct a writer to prefix the words "Available E-mail:" indifferently to either the list address or the retrieval address.
An example of an e-mail discussion list is "acw-l," maintained by the Alliance for Computers and Writing. Anyone with an Internet address may subscribe to this list and post messages to it. It would be a mistake to suggest that the Alliance stands in the role of "publisher" for this list, in the sense of being responsible for what is posted there. Publication on the list is accomplished simply as subscribers send messages to the list's e-mail address, <firstname.lastname@example.org>, and as the messages are distributed and archived by the list management software, which in this case is the LISTPROC program. Hence it is appropriate to cite the list address--rather than a particular message-author's address--as the publication address for a list message. (The models given below do include the author's e-mail address, after the author's name, to permit authentication.) But it might be very unfortunate for the list's subscribers if a citation format were to leave the impression that cited messages might be retrieved through the list address. Retrieval should be facilitated instead by citing the appropriate address for the list's archive, which may then be followed by the usual date of access. In order to distinguish the two addresses and yet indicate their connection, we choose to use the ancient word via. As an example, consider the following potentially misleading citation from Li and Crane:
RRECOME. "Top Ten Rules of Film Criticism." 1 Apr. 1995. Online posting. Discussions on All Forms of Cinema. Available E-mail: CINEMA-L@american.edu. 1 Apr. 1995.
Incorporating Walker's simplifications and this paper's suggestions for delimitation, date specification and address differentiation this citation might become:
RRECOME. <email@example.com> "Top Ten Rules of Film Criticism." 1 Apr. 1995. <firstname.lastname@example.org> via <email@example.com> (1 Apr.1995).
It is important to note that the address for a list's archive will not always be an e-mail address. Such is the case for acw-l, where the archive is accessible not only through the LISTPROC software but also through the Alliance's WWW pages. The following three citations illustrate how different stylists might treat the same acw-l message.
Carbone, Nick. "NN 960126: Followup to Don's comments about citing URLs." 26 Jan. 1996. Online posting. Alliance for Computers and Writing. Available E-mail: ACW-L@unicorn.acs.ttu.edu. 28 Jan. 1996.
Carbone, Nick. "NN 960126: Followup to Don's comments about citing URLs." firstname.lastname@example.org (26 Jan. 1996).
Carbone, Nick. <email@example.com> "NN 960126: Followup to Don's comments about citing URLs." 26 Jan. 1996. <acw-l@unicorn. acs.ttu.edu> via <http://www.ttu.edu/lists/acw-l> (17 Feb. 1996).
At the core of the concept of hypertext is the potential for linking from one document to another indefinitely. One consequence of this feature, as implemented in the design of protocols for the World Wide Web, is that a particular computer file is accessible by "dynamic" addressing through an indeterminate number of paths, with no single path claiming precedence over another. A lengthy URL like <http://www.csc.eku.edu/honors/hon102/hon102.html> might seem to indicate that the target document lies "deep" within a hierarchical structure, as in looking inside drawers within drawers, but in terms of computer access, such a label is only one memory address among many, and functionally equivalent in "depth" to all the others. (In fact, many URLs do not represent the actual file structure on any machine, but serve instead as "aliases" or "lookups" for actual directories.) Users of computer spreadsheets may find the analogy of URLs with spreadsheet "cells" helpful in grasping this idea: while each cell has an absolute reference, that unique reference stands for a memory location and not "the data itself," and there are an indefinite number of ways to provide relative references to any one cell. Furthermore, by linking one spreadsheet to another, the boundaries defining a "single" spreadsheet become fuzzy, in the same way that the concept of a "whole work" defies the core concept of hypertext.
For citation practice, the equivalence of all relative addressing means that there is no "authentic" URL to cite for a given source. Some URLs may be preferred because they are shorter than others, or because the address contains "relevant" words. For example, when citing an adoption document, on might choose a URL that contains "www.adoption.org" for an adoption document, regardless of where the file is actually stored; such a reference is legitimate because it is in principle impossible for any user to find out the absolute address for the "actual" file.
The essential open-endedness of hypertext linking means that there is no meaningful distinction between "primary" and "secondary" addresses. It also means that the concept of a "whole work" is meaningless for many Internet documents, except as an author groups a set of related files under a single title. More significant for citation purposes is the likelihood that one might prefer to cite a source "indirectly," that is, pointing to a document that contains a link to a relevant document rather than the document itself. For example, if one wanted to cite only one file in a collection, but also make it convenient for users to understand the "context" of the collection and perhaps access other similar documents, it would be helpful to have a parallel with the MLA Handbook's treatment of indirect references (5.4.7). The analog for "quoting" on the Internet is "linking," so an appropriate citation descriptor might be "linked from," or Lkd. (paralleling "qtd." in one's text and "Trans." in a Works Cited entry; Gibaldi 5.4.7; 4.6.23), together with any "path" details: (Note 5)
Miller, Allison. "Allison Miller's Home Page." Lkd. EKU Honors Program Home Page, at "Personal Pages." <http://www.csc.eku.edu/honors> (11 Nov.1995).
The inclusion of this linkage permits Web-users to locate both the source and the collection with ease.
We should "bear in mind" as Walker suggests "that, like the Internet itself, the information sources are in a constant state of flux and therefore, [our style] will also need to change as the sites themselves proliferate [as we] adapt to the new era of electronic print" ("MLA-Style"). As we develop citation styles, we therefore recommend the following interim practices.
Dr. Andrew Harnack, Eastern Kentucky University
Dr. Eugene Kleppinger, Eastern Kentucky University
1 Admittedly, the latest edition of the Publication Manual of the American Psychological Association (1994) does not offer any advice for citing electronic sources. Importantly, however, Xia Li and Nancy Crane's revision of Electronic Style: A Guide to Citing Electronic Information (1993) provides a wealth of advice and a remarkably full set of models for writers using the APA-style of citation. A new edition, retitled as The Official Internet World Guide to Electronic Styles: A Handbook to Citing Electronic Information, will be publshed in Spring 1996. RETURN
2 The third of edition of Writing Research Papers by one of us (Harnack), forthcoming in the fall of 1996, will contain a chapter on guidelines for electronic citation that incorporates both the recommendations in Walker's "MLA-Style Citations of Electronic Sources" and suggestions for improving the style as proposed in this essay. RETURN
3 Notable style guides besides those mentioned in the text are by Haines Brown, Maurice Crouse, Mark Wainwright and Peggy Whitley. A wide variety of guides are listed in the "Field Guide" and in Jan Tent's collection. RETURN
4 Page's original has "U. S." where the current Gopher pathway shows "US" with no periods. The suggested citation does not include the word "Gopher," and Page gives no clear instruction for telling the user which protocol to use. RETURN
5 This concern was discussed by Walker and Kleppinger in private e-mail. RETURN
These guidelines and models are based upon Janice Walker s
"MLA-Style Citations of Electronic Sources (Endorsed by the
Alliance for Computers and Writing)," Vers. 1.0; it is revised
as recommended by Andrew
To cite files available for viewing/downloading via the World Wide Web by means of Lynx, Netscape, or other Web browsers, provide the following information:
Burka, Lauren P. "A Hypertext History of Multi-User Dimensions." MUD History. 1993. <http://www.ccs.neu.edu/home/1pb/mud-history.html> (5 Dec. 1994).
To cite files for downloading via ftp, provide the following information:
Bruckman, Amy. "Approaches to Managing Deviant Behavior in Virtual Communities." ftp ftp.media.mit.edu pub/asb/papers/deviance-chi-94 (4 Dec. 1994).
Bruckman, Amy. "Approaches to Managing Deviant Behavior in Virtual Communities." <ftp://ftp.media.mit.edu/pub/asb/papers/deviance-chi-94> (4 Dec. 1994).
To cite telnet sites and files available via the telnet protocol, provide the following information:
Gomes, Lee. "Xerox's On-Line Neighborhood: A Great Place to Visit." Mercury News. 3 May 1992. telnet lamba.parc.xerox.com 8888, @go #50827, press 13 (5 Dec. 1994).
To cite synchronous communications such as those posted in MOOs, MUDs, IRCs, etc., provide the following information:
Harnack, Andrew. Group Discussion. telnet moo.du.org/port=8888 (4 Apr. 1996). WorldMOO Christmas Party. telnet world.sensemedia.net 1234 (24 Dec. 1994).
To cite information obtained by using gopher search protocols, provide the following information:
Quittner, Joshua. "Far Out: Welcome to Their World Built of MUD." Published in Newsday 7 Nov. 1993. gopher University of Koeln/About MUDs, MOOs, and MUSEs in Education/Selected Papers/newsday (5 Dec. 1994).
To cite information posted on listservs, provide the following information:
Seabrook, Richard H. C. <firstname.lastname@example.org> "Community and Progress." 22 Jan. 1994. <email@example.com> (22 Jan. 1994).
To cite a file archived at a listserv or Web address, provide the following information after the publication date:
Seabrook, Richard H. C. <firstname.lastname@example.org> "Community and Progress." 22 Jan. 1994. <email@example.com> via <firstname.lastname@example.org> (29 Jan. 1994).
To cite information posted by participants in newgroup discussions, provide the following information:
Slade, Robert. <email@example.com> "UNIX Made Easy." 26 Mar. 1996. <alt.books.reviews> (31 Mar. 1996).
To cite electronic mail correspondence, provide the following information:
Franke, Norman. <firstname.lastname@example.org> "SoundApp 2.0.2." 29 Apr. 1996. Personal e-mail. (3 May 1996).
To cite a specific file and provide additional information concerning its contextual linkage, provide the following:
Miller, Allison. "Allison Miller's Home Page." Lkd. EKU Honors Program Home Page, at "Personal Pages." <http://www.cwsc.edu/honors> (11 Nov. 1995).
Date of last revision: 10 June 1996 Kairos:
A Journal for Teachers of Writing in Webbed Environments.
Vol. 1 No. 2 Summer 1996