Publications of many kinds employ pointers within the main body of the text to refer the reader to explanatory or additional material either on the same page or elsewhere in the work. Academic publications in particular need an apparatus to provide references to support, and sometimes to clarify and amplify, what is said in the text. It is of course the responsibility of the author to ensure that the text is properly supported by appropriate sources, that the sources are correctly quoted or interpreted, and that the citations are accurate, complete, and follow an appropriate style. It is, however, the duty of the editor to ensure that the references in a work are presented in a clear and consistent manner. The reader may well see sloppy references as symptoms of generally careless scholarship.
The systems commonly adopted vary greatly between disciplines, and there may be considerable variation even within a single discipline, according to the preferences of particular publishers, editors, and authors. Furthermore, as in other aspects of editing, practice changes over time, and new styles may be taken up very unevenly. The most traditional system employs numbers, letters, or symbols as cues to direct the reader to notes which are printed either at the foot of the page to which they relate (footnotes) or in a group at the end of the text (endnotes). Alternatively, in what is known as the author–date or Harvard system, the cues may take the form of an author’s name and date of publication (or in the author–number or Vancouver system simply a reference number) within parentheses that enable the reader to identify the work in a list of full references at the end of the text. Various permutations are possible. In some ways life is made easier for both author and editor if a commonly recognized system is adopted, but agreed adaptation to the needs of a particular work can prove beneficial. What matters most is that the necessary information should be conveyed to the reader (of both print and electronic materials) clearly and economically.
The primary function of scholarly annotation is to identify for the reader the sources of what is said in text. This may be the location of a verbatim quotation or—just as important in an academic work—the basis for a statement by the author. Simple references of this kind can be accommodated by systems that use brief parenthetical references in the text to take the reader to a consolidated bibliography (see 17.3 and 17.4). Such references are particularly well suited to scientific publications, but they are also used in the social sciences and increasingly in the humanities. In the humanities it is normally notes that are the essential feature of the supporting apparatus. Notes are a convenient vehicle not only for complex bibliographical citations but also, for example, for acknowledgements, further discussion, supporting original text, and bibliographical surveys (see 17.2.1). Notes may be used in conjunction with a bibliography or more selective suggestions for further reading, or with a list of abbreviations of frequently cited sources (see 17.2.6).
An apparatus based on notes is normal in the humanities, where more is often needed than a simple link to a consolidated list of published sources. For example, manuscript sources are more easily cited in notes than in the systems of parenthetical reference used in scientific publications (see also 18.6). Different sources, or versions of a source, may be compared and evaluated in a note. If a quotation is given in translation in the text the original may perhaps be given in a note, or conversely (this may be scrupulous scholarly practice where doubt is possible about the correctness of a translation, but it should not be adopted routinely without good reason). It is sometimes helpful to include in a note a brief survey of the literature on a particular topic, or a summary of a debate, though authors should be discouraged from needlessly transforming notes into bibliographies. Author’s acknowledgements also fit neatly into notes. Also, notes can perform special functions in scholarly editions of original texts, where they may, for instance, supply variant readings from different manuscripts.
The flexibility that makes notes such a useful tool can be abused. Authors sometimes include in notes further discussion of a question raised in the text, or of some related issue. While this cannot be outlawed altogether, it should be avoided wherever possible. For the most part, if a point is important and relevant enough to be discussed at all it should be dealt with in the text. Notes should be kept as short as possible, and inessential material excluded. The best place for extensive but essential ancillary matter that cannot be accommodated in the text may be an appendix. The main function of notes (other than those in editions of texts) remains the clear and concise presentation of necessary references.
In printed material, notes may be either footnotes, placed at the foot of the page to which they relate, or endnotes, placed in a single sequence at the end of the text. In multi-author volumes and journals that employ endnotes they should be printed at the end of an individual chapter or article; otherwise they are normally placed at the end of the book, in a separately numbered sequence for each chapter.
Some categories of note, such as those that explain the meaning of words in the text or provide variant readings, have an especially strong claim to be placed at the foot of the page. Footnotes are in general much more convenient for the reader, who can keep track of them without the annoying disruption of flipping back and forth between text and notes in the course of reading. The setting of notes at the foot of the page can make page layout more complex than placing them together at the end of the text, depending on the length and number of footnotes.
In material presented in formats that do not rely on a fixed page, such as ebooks and some websites, the distinction between footnotes and endnotes becomes less meaningful. Notes in electronic formats will generally be connected to the text via a hyperlink. Authors should discuss with their publishers every format in which their text is likely to be published and plan for a notes system that will work across them all.
The reader is referred to a footnote or endnote by a cue in the text. This normally takes the form of a superior Arabic number. The cue is placed after any punctuation (normally after the closing point of a sentence). If, however, it relates only to text within parentheses it is placed before the closing parenthesis. The cue is repeated at the start of the note. Notes cued in the middle of a sentence are a distraction to the reader, and cues are best located at the end of sentences:
He was a genuine Shropshire lad, as John H. Johnston reminds us.51
Bergonzi quite correctly notes, ‘Owen’s attitude to the “boys” or “lads” destined for sacrifice has some affinity with Housman’s.’55
(Hopkins wrote defensively to Bridges: ‘When you read it let me know if there is anything like it in Walt Whitman; as perhaps there may be, and I should be sorry for that.’29)
Characters other than Arabic numerals may be used for note cues when there are relatively few notes in a sequence, but this should only be done when necessary. In mathematical or scientific contexts, for example, superior lower-case letters may be used to avoid confusion with superscript numbers in technical notation. Lower-case Roman numerals may also serve as note cues (though this quickly becomes unwieldy), as can reference marks (the traditional order is *, †, ‡, §, ¶, ||, repeated in duplicate as **, ††, and so on as necessary—but this too is a cumbersome method if there are more than a handful of notes). Occasionally different types of cue are employed on the same page for parallel sequences of notes serving different purposes: this should be avoided if possible but is sometimes required in complex editions.
Notes should be numbered continuously through each chapter or article. This allows the numbers in the author’s copy to remain unchanged at typesetting; internal cross-references to notes thus do not require correction on page proof, and passages of text may even be located by reference to numbered notes (Ch. 6 at n. 17). Continuous numbering of notes through an entire book is to be avoided, as it can generate too many three-digit cues and will cause problems if a note is deleted or added at proof stage. Whether numbered page by page or chapter by chapter note cues must appear in strict numerical sequence; the same number must not be used twice within a sequence, even if the content of the note is the same.
An initial note consisting entirely of acknowledgements may be placed before the numbered notes and cued with an asterisk; the asterisk is placed at the end of the first sentence of text. Avoid asterisks, note numbers, or other cues within or at the end of titles, subtitles, and chapter subheadings. An initial uncued note may be used to provide the original location of a reprinted chapter or article.
A note giving the source for a displayed quotation is best placed at the end of the quotation itself rather than at the end of the preceding text. Where there are multiple quotations from or references to a particular source the locations should be given in a single note after the last quotation or reference, provided no other citation intervenes. The page numbers should be in the same order as the quotations or statements to which they relate, not rearranged into numerical sequence. If there are repeated citations of a single source (but no other) over several paragraphs it may be best to provide a separate note for each paragraph. The notes in three such successive paragraphs might, for example, be:
6. Smith, Windham’s Green Book, 25, 17, 31.
7. Smith, Windham’s Green Book, 87, 95, 103–5.
8. Smith, Windham’s Green Book, 150, 75, 279.
Furthermore, where it can be done without ambiguity, it is good practice to group references to different sources in a single note after several sentences or at the end of a paragraph. The nature of the source will generally indicate to which statement in the text it relates, and any doubt may be removed by a parenthetical word or phrase after the references or by an introductory phrase before several related references. The editorial effort required is justified by the reduction in the number of notes and the improved readability of the text.
Not even the best endowed colleges had incomes approaching those of such great Benedictine houses as … Westminster or Glastonbury … New College’s estates probably yielded revenues of a similar magnitude to the Augustinian abbey of Oseney. … The estate income of All Souls … was probably slightly lower for example than that of Bolton Priory in Yorkshire, which supported merely fifteen canons … The college by contrast remained close to its statutory complement of a warden and forty fellows.30
30. D. Knowles and R. N. Hadcock, Medieval Religious Houses: England and Wales (1953), 80 (Westminster), 66 (Glastonbury), 149 (Oseney), 128 (Bolton); I. Kershaw, Bolton Priory (1973), 186; Cobban, ‘Colleges and halls’, 609.
The layout and typography of notes is subject to considerable variation. Footnotes in particular are usually set in smaller type than text. A note should begin with a capital initial and end with a full point. It may or may not contain grammatically complete sentences. Abbreviated forms (for example for the months of the year) and symbols that would not be acceptable in open text may be appropriate within citations in notes. Various forms of punctuation and wording may be used to group citations and to indicate how they relate to material in the text or to related questions; what is most efficient in any context is a matter for editorial judgement. The common abbreviation ‘cf.’ (Latin confer) means ‘compare’, and thus is not quite the same as ‘see’.
The number preceding the note may be superscript or on the line:
40 Mokyr, Why Ireland Starved, 26.
40. Mokyr, Why Ireland Starved, 26.
Whether or not a point is needed after a number on the line is a design issue.
Full guidance on the form of bibliographic citations is to be found in Chapter 18. For the most part the same considerations govern entries in a bibliography and citations in notes, with the important exception that in notes an author’s initials or forename precede rather than follow the surname. Consistent systems for the formulation of notes are essential, but abstract rules should not be followed too slavishly: much can depend on editorial judgement in presenting particular references as clearly and economically as possible.
In addition to general bibliographic rules there are conventions, relating especially to multiple citations of the same work, that are intended to promote brevity and clarity within notes. Unless it is included in a list of abbreviations or a bibliography (see 17.2.6), full bibliographic details of a published work or the location of an unpublished source should be given when it is first cited in an article or book, and repeated at its first citation in any subsequent chapter in the same book. Subsequent citations should take a very abbreviated form, typically the author’s surname and a shortened title of the work. The short title should be accurately extracted from the full title (not a paraphrase) and should be as brief as is compatible with the unambiguous identification of the work. Rather longer forms may be advisable if works of similar title are cited. Short titles alone may be used for works cited with no namedauthor. It is a matter of judgement whether an editor’s name should be repeated with the title of an edited text. Short forms may also be devised for the multiple citation of unpublished sources whose full forms are unduly lengthy.
Once a short form has been established, the author’s initials or the full title should not be reintroduced in later citations in the same chapter.
5. R. J. Faith, The English Peasantry and the Growth of Lordship (1997).
73. Faith, English Peasantry, 49–50.
Locations within a work should generally be given in the shortest unambiguous form. Publishers will have their own conventions for the various fields in which they produce books or journals. Lower-case Roman numerals were traditionally used for volume numbers, but Arabic numerals are now common; either system should be applied consistently. Roman page numbers must be retained and not converted to Arabic. Volume and page numbers may be linked by a point, consistently either closed up or spaced off. Abbreviations for pages or volumes (p., pp., vol., vols.) are not strictly necessary in most cases, but they should be included before Arabic or Roman numerals if there is a risk of confusion as to what element is being cited. If a work has numbered columns rather than pages there is no need to use the abbreviation col. in citations, as there will be no ambiguity when the reader consults the source. Likewise compound locations consisting of several numbers (whether Arabic, lower-case Roman, or small capitals, depending on house style or general convention) may reasonably be used with no explanation of the elements they represent (book, chapter, paragraph, question, etc.) if there will be no ambiguity in the source itself.
Letters and Journals of Robert Baillie, 2.110
Brett-Smith, 1.xxviii–xxix
English Historical Documents, i, 2nd edn, ed. D. Whitelock (London, 1979), no. 191
Liber de caesaribus, v.39.20
Authors sometimes feel obliged to give the overall pagination of articles or chapters they cite in addition to the particular passage to which they are referring. This is no more logical or helpful than citing the total pagination of a book, and is an unnecessary complication of a citation.
For the treatment of abbreviations and contractions of terms such as ‘page’, ‘chapter’, and ‘volume’ see Chapter 10. The use of ‘f.’ after a page number to indicate ‘and the following page’ should be replaced with an explicit two-page span (15–16, not 15 f.). A page number followed by ‘ff.’ to indicate ‘and the following pages’ should also be converted to a precise span if possible, but this form is acceptable when it is difficult for the author (or editor) to identify a final relevant page. The spacing, if any, before the abbreviation (23ff. or 23 ff.) is a matter of house style; Oxford traditionally uses a thin space between a number and a following ‘f.’ or ‘ff.’, and between a number and a following ‘n.’ or ‘nn.’ in the absence of a subsequent note number (when a note number is included, normal space of the line is used: 23 n. 5). When specific notes are cited, ‘note’ may be abbreviated to ‘n.’ and ‘notes’ to ‘nn.’ It is best to avoid punctuation between page number and note number. In the third of the following examples the notes cited contain information additional to that given in the text:
Marx, Manifesto, 37 n. 4.
Kleinhans, ‘Marxism and Film’, 106 ff.
K. McRoberts, Misconceiving Canada (Toronto: OUP, 1997), 12 and nn. 37–8.
Recto and verso are sometimes abbreviated r and v, set roman, and may be used to indicate right- and left-hand pages respectively— 31r, 70r, 78v, 89r.
A reference to another place within the work itself may be included in a note if it will be genuinely helpful to the reader, but such internal cross-references should be added judiciously. Where possible, cross-refer by chapter and chapter subheading in addition to using the page number: this will be more effective than the page number alone in an ebook or other unpaged format.
Many abbreviations are used in note references, to aid or direct the reader as succinctly as possible. For the most part, a lower-case abbreviation that begins a note is capitalized, whether or not the note is a complete sentence. A handful of common abbreviations are, however, exceptions to this rule: c., e.g., i.e., l., ll., p., pp. generally remain lower case:
20. c.1344, according to Froissart.
21. e.g. service outside the jurisdiction.
22. i.e. Copyright, Designs & Patents Act 1988, §4.
23. p. 7.
24. ll. 34–44 (Miller edn).
Certain Latin words or their abbreviations have been traditionally used to make citations in notes more concise. These abbreviations serve admirably in print but are far less useful in electronic formats, where the notes may appear (for example) in a pop-up rather than in a consecutive list. The reader hovering on a note link and receiving the abbreviation ‘Ibid.’ in a standalone pop-up gains no useful information. For this reason Oxford prefers the author and short-title system wherever possible. Exceptions are allowed for some specialist material (such as legal publications) where these abbreviations are rigidly followed throughout a discipline.
A list of bibliographical abbreviations is often printed in a book’s preliminary matter, or at the start of a bibliography. An abbreviated citation assigned in this way will be used every time the work is cited in a note, even at its first citation in a chapter. Whether or not it is worth including such a list in a volume will depend not only on how often particular works are cited in the volume as a whole but also on how many works are cited frequently in more than one chapter. If a work is cited very many times in one chapter it will routinely be reduced to author and title after its first citation; a more irregular abbreviation of a complex source may be explicitly introduced at its first mention in a chapter:
Bede, Historia ecclesiastica gentis Anglorum (hereafter HE)
St John’s College, Cambridge, archives (hereafter SJC)
If a source abbreviated in this way is cited in no other chapter in the volume an entry in an overall list of abbreviations may be superfluous.
An apparatus for references known as the author–date (or Harvard) system is normal in the scientific, technical, and medical fields, and is also used extensively in the social sciences. It is not based on notes but relies on brief parenthetical references in the text to take the reader to the appropriate point in a consolidated list of full citations, generally known as the reference section. A reference section includes only those works that are cited in the text; a more general list of works of related interest should be called Bibliography or Further Reading. This is an economical method of citing straightforward published sources.
It is possible to combine this arrangement with a separate sequence of numbered notes for explanatory or discursive matter:
text
a stoutly republican coalition that retained none the less a great deal of the administrative style of the old regime (see Stookey 1974). It was replaced by a government with a different style,2 a style … not forgotten or revoked.
endnote 2
The usual explanation of the timing of the coup is that it forestalled a Ba’athist plot. That this was no simple question … we shall see when we quote the speeches given at a tribal meeting soon afterwards.
reference section
Stookey, R. W., 1974. ‘Social Structure and Politics in the Yemen Arab Republic’, Middle East Journal, 28/3: 248–60; 28/4: 409–19
In extreme cases where multiple references (and multiple authors) render a sentence unreadable, and the problem cannot be resolved by rewriting, a group of references may be relegated to a footnote. Bracketed author–date references to a bibliographic list are sometimes included within footnotes.
In the author–date system the full references are listed alphabetically in a section at the end of the text (either chapter by chapter or in a consolidated list at the end of the book). As in bibliographies, authors’ initials generally follow their surnames. To facilitate linking with the short references given in the text the date of publication immediately follows the authors’ names. The styling of these references can vary considerably; scientific and medical publishers, for example, may use very different conventions from those adopted in the humanities. For example, end references may be fully or minimally punctuated and formatted:
Grinspoon, L., Bakalar, J. B. (1993), Marijuana: The Forbidden Medicine. London: Yale University Press.
Kalavala M, Mills CM, Long CC, et al. (2007). The fingertip unit: A practical guide to topical therapy in children. J Dermatol Treat 18:319–320.
There is a general preference for minimal punctuation in scientific work, to save space and keying time, while humanities publications usually prefer fuller punctuation. For further guidance see Chapter 18.
A typical reference in the text consists of an author’s name and date of publication enclosed within parentheses (or occasionally square brackets), with or without a comma separating name and date, according to the style adopted. The reference is placed immediately after the statement to which it relates. If this happens to be at the end of a sentence the closing parenthesis precedes the closing point (but a reference at the end of a displayed quotation follows the closing punctuation). If the author’s name is given in open text it need not be repeated in the parentheses, where the date alone suffices. Several references may be included within the same parentheses, separated by semicolons:
text
While there was an extraordinary sense of optimism among people establishing their own farms in the early years of independence (Unwin 1994), this is rapidly withering away.
reference section
Unwin, T. (1994), ‘Structural Change in Estonian Agriculture: From Command Economy to Privatisation’, Geography, 79, 3: 246–61.
text
For years, most textbooks referred to the five stages of economic integration as defined by Balassa (1961).
reference section
Balassa, Bela (1961), The Theory of Economic Integration. London: Allen and Unwin.
text
They are also used to detect segmental hypermobility (Magarey 1988, Maitland 2001).
reference section
Magarey ME (1988). Examination of the Cervical and Thoracic Spine. In: R Grant (ed.) Physical Therapy of the Cervical and Thoracic Spine, pp. 81–109. Churchill Livingstone: New York.
Maitland G (2001) Maitland’s Vertebral Manipulation, 6th edn. Butterworth–Heinemann: Oxford.
Multiple authorship is very common in scientific publication. Each work should have a consistent convention as to how many authors’ names are given in full and what number, if any, should be reduced to ‘et al.’ (sometimes italicized) after the name of the first author (see Chapter 18). Some publishers use ‘et al.’ to shorten references in text even when the names are given in full in the corresponding entry in the reference section. Either an ampersand or ‘and’ should be used consistently to link dual authors or the last two of multiple authors, even though in the reference section they may be separated by a comma:
text
One of the biggest successes of the 1960s was transformed into an albatross hanging from the neck of an embattled Community (Rosenblatt et al., 1988).
reference section
Rosenblatt, Julius et al. (1988). The Common Agricultural Policy of the European Community, International Monetary Fund, occasional paper 62, November.
text
Prototypical birds, for instance, seem to be birds of average size and average predacity (Rips et al., 1973).
reference section
Rips LJ, Shoben EJ, Smith EE (1973). Semantic Distance and the Verification of Semantic Relations. Journal of Verbal Learning and Verbal Behaviour 12: 1–20.
text
It has been estimated that the human eye can discriminate no fewer than 7.5 million just noticeable colour differences (Brown and Lenneberg 1954).
reference section
Brown, R., and Lenneberg, E. H. (1954). ‘A Study in Language and Cognition’. Journal of Abnormal and Social Psychology 49: 454–62.
If the reference section contains works by authors of the same surname their initials may be retained in the in-text reference to distinguish between them. If there is more than one work by an author in a single year they are distinguished by lower-case letters (either roman or italic depending on house style) appended to the year in the parenthetical reference and in the list. The dates of several works by a single author are separated by commas. Occasionally a parenthetical reference may be introduced by terms like ‘see’, ‘see also’, or ‘cf.’:
text
a diverse body of work has emerged which focuses on the ‘governance’ of socio-economic systems (see Jessop 1995a, 1997)
reference section
Jessop, B. (1995a) ‘The Regulation Approach, Governance and Post-Fordism’, Economy and Society, 24, 3: 307–33.
Jessop, B. (1995b) ‘Regional Economic Blocs’, American Behavioral Scientist, 38, 5: 674–715.
Jessop, B. (1997) ‘The Governance of Complexity and Complexity of Governance’, in A. Amin and J. Hausner (eds) Beyond Markets and Hierarchy, Aldershot: Edward Elgar.
How to style references to works that do not fit into the normal pattern of author and title is a matter for editorial judgement. A reference to an anonymous work, for instance, may place either ‘anon.’ or a short title before the date of publication; reference to a work produced by a corporate body may similarly use either the work title or the name of the body, abbreviated if the name is very long:
text
at the hands of the state, paramilitary and guerilla forces (CNRR/GMH, 2008a)
reference section
CNNR/GMH (Comisión Nacional de Reparación y Reconciliación y Grupo de Memoría Historica) (2008a) Trujilla: Una Tragedia que no Cesa [Trujilla: A Tragedy Without End], Bogotá, Colombia: CNRR.
A reference to an unsigned item in a periodical may use its title and date. The crucial point is that all in-text references should be styled consistently with the reference section and should enable the reader easily to identify the source there.
A personal communication or an interview with the author may be so described in a parenthetical reference, but need not be included in the reference section:
text
By this time, industry had come to play a leading role (World Bank 1993).
reference section
World Bank (1993) Vietnam: Transition to the Market. Washington, DC: The World Bank.
text
in Uppsala ‘the main aim is to create sustainable development, although there is no true consensus as to what this means’ (Peterson, pers. comm.).
appendix: Informants
Agneta Peterson, environmental planning officer, Uppsala municipality.
The same is also true of academic papers submitted but not yet accepted for publication; see also 18.3.1
Broadly speaking, articles in scientific disciplines are shorter than those in the humanities, and often it is not necessary to specify a location within in-text references; locations are more commonly cited in references in the social sciences. When a location is given it is usually separated from the date by a colon.
Some scientific publications use the author–number (or Vancouver) system, which numbers the citations in a single sequence and dispenses with authors’ names in the references in text; these consist simply of superscript numbers or numbers in parentheses or square brackets, multiple references being separated by commas, set close, with number ranges used as appropriate (as found in similar studies.3, 5, 6, 14–18). This arrangement has a superficial resemblance to endnotes, but the reference numbers do not necessarily occur in numerical sequence in the text, where any one may be repeated several times:
text
Issues of risk, choice, and chance are central to the controversy over the MMR vaccine that erupted in the UK in 1998 and has continued into the new millennium.1
reference section
1. Fitzpatrick M. MMR and Autism: What Parents Need to Know. London: Routledge, 2004.
text
The inter-individual variability in VO2 measured at a given speed and rate can be as high as 15% [18].
reference section
18. Nieman, D. C. Exercise Testing and Prescription, 5th edn. New York: McGraw-Hill, 2003; 90.
The late addition or removal of a reference from a numbered list may require more or less extensive correction of the numbers in the reference section and the in-text citations.
The many software programs that can help with reference formatting can adapt citations to whichever system is required by the publisher with comparative ease. References can be downloaded from citation databases such as Copac in a form that can be imported into the reference management software, which then embeds the reference in the word-processor document. Editors should be aware that automatically formed references should always be checked for consistency, completeness, and accuracy, as some manual intervention may be necessary, particularly for sources that do not follow the usual pattern.
Hyperlinks in digital texts can and have transformed the workhorse reference list into a dynamic resource providing access to multimedia material that enriches the user’s experience to a far greater extent than a printed book. Authors and editors should be aware, however, that this will be successful only if the referencing is done accurately to enable linking to material that should be current and accessible.