Errors and issues in Oxford Style Manual

P. 6

In the heading Current edition, impression, and imprint, the last word should apparently be reprint. The text discusses reprint and does not even mention the word ‘imprint’.

P. 50

In the second bulleted point, the statement ‘An en is […] the average width of typeset characters’ is incorrect. The width depends on font but typically revolves around 0.4em, i.e. 0.8en.

The following boxes demonstrate this. The horizontal line is 13em, i.e. 26en, wide, and the string after it contains the 26 lowercase letters of the English alphabet, in the Times New Roman. Of course, the concept of average width is vague, but even this simple demonstration should suffice.


The following demonstration contains 100 letters, appearing roughly according to their frequency in English according to a frequency study. If the statement about en were correct (for Times New Roman), the horizontal line (set to the width of 50em) would be about as wide as the string of letters.


The statement in the book is even in contradiction with a statement on p.52, second paragraph: ‘An en space will be th width of the figures in many fonts, and therefore may not be exactly half of an em.’

P. 48, 2.4.1 Vertical spacing

The statement that ‘in Europe’ the point is 0.763mm is both oversimplification and mostly outdated. The Anglo-Saxon concept of typographic point is dominant in continental Europe, too, because typesetting is almost completely computerized and computer software uses the Anglo-Saxon unit, with few exceptions.

P. 49

On the first line, Van Djick (as font name) as apparently a misspelling of Van Dijk (i and j swapped).

P. 91

The first bullet point lists the following among ‘typical Latinizing changes to Greek names’:

kh = ch (but ch often retained)
ou = u (but u occasionally retained)
u (except in diphthongs) = y (but y sometimes retained)

These parenthetic remarks do not seem to make sense. In these cases, ch, u, and y are the replacements used in Latinizations. So why do the remarks refer to them (and not kh, ou, and u used in the transliterated Greek names) as being retained?

P. 95, German

The first line under the heading ‘German’ contains the misspelled word Fraulein. The correct spelling has diaeresis (Umlaut) on the letter ‘a’: Frulein.

P. 114

The last line of the first paragraphs contains the spelling McDonald᾿s, where the apostrophe and the letter s appear in upright style, not in italics..

P. 140, 5.10.9 En rule

The statement that the en rule is an en in length is not correct: in several fonts, it is somewhat shorter. The illustration on the left demonstrates this: there is dark gray bar that is 1en (0.5em) wide and an en rule (en dash) below it with green background, in the Microsoft Sans Serif font, using 200% font size for clarity.

P. 142, 5.10.11 Em rule

Similarly to the error on p. 140, the statement that the em rule is an em in length is not correct: in several fonts, it is somewhat shorter. The illustration on the left demonstrates this: demonstrates this: there is dark gray bar that is 1em wide and an em rule (em dash) below it with green background, in the Microsoft Sans Serif font, using 200% font size for clarity.

P. 143, 5.11.1 Parenthese

The Greek quotation near the middle of the page contains an error:

ὁ βίος βραχύς, ἡ δὲ τὲχνη μακρή

The word τὲχνη should be τέχνη, i.e. it should have an acute accent (oxya, tonos) instead of a grave accent (varia). In Greek, the grave accent occurs on the final syllable of a word only (when the next word begins with a stressed syllable).

P. 155, second paragraph: italic in Russian and Greek

Contrary to what the text says, there is italic in Russian and Greek texts. Whether it is good style is a different question. The Elements of Typographic Style by Robert Bringhurst contains examples of italic Russian and Greek on pages 107 and 108. Moreover, the book itself italic versions of Russian letters on p.334, though it calls them ‘cursive’.

P. 168, 7.1.3 Punctuation

At the end of the page, the statement about a raised decimal point is repeated on p.171 under the heading 7.4Decimals. This is probably unintentional repetitiveness, but it can be regarded as more serious than a stylistic flaw: the second discussion takes a much stronger normative position against the usage, and a reader who only notices the first discussion gets a wrong idea about recommendations.

P. 172

The first bullet point says: ‘There is a space of the line (or thin space in scientific and technical work) between the figure and degree sign’, when the temperature scale is part of the notation. Presumably ‘space of the line’ means a normal space. However, the example is ‘40C’, which does not contain a normal space, and hardly even a thin space. The same applies to the notation ‘10–150C’ later, but the wording is obscure and does not clearly specify whether a space should appear here, too, efore the degree sign. (According to SI rules, it should.)

Compare this with the text in clause 12.1.7 (Temperature and calories) on p.373, where similar rules are given but examples follow the rules, i.e. contain spaces: 10.15C,, 15–17C. Note that the latter notation differs from the example on p.172 and SI rules, in that the degree symbol is attached to the lower bound, too, instead of treating the C as an unseparable symbol of a unit (as it is according to SI definitions).

P. 176, 7.7.2. Classical contexts

In the first bulleted point under the heading, the following statement is obscure, even absurd:

500 was I, later corrupted to D for 1,000.

The symbols I and D appear in italic for no apparent reason, unlike the other roman numerals discussed in the text. More seriously, the symbol I never meant 500, and no symbol for 500 got ‘corrupted’ to a symbol for 1,000.

It is possible that I was mean to be a completely different symbol. Anyway, the true story is, as briefly explained e.g. in Issues Regarding Roman Numerals in UCS by David J. Perry, that originally a symbol for 500 was the right half of a symbol for 1,000, later developed to a D-like symbol and even later identified with the latin letter D. The symbols involved had different shapes, and some of the shapes are now encoded as Unicode characters U+2180 ROMAN NUMERAL ONE THOUSAND CD (looking like C and D close together) for 1,000 and U+216E ROMAN NUMERAL FIVE HUNDRED (looking D-like) or U+2160 ROMAN NUMERAL ONE U+2183 ROMAN NUMERAL REVERSED ONE HUNDRED (i.e., I-like symbol followed by a symbol like mirrored C).

Probably the text in the book is meant to refer to the latter representation of the symbol for 500 and to add that this combination later became D-like. The expression ‘for 1,000’ makes no sense in this context, but it is probably part of something that referred to the origin of the symbol for 500.

The bullet point also contains, in its last sentence, symbols for some large numbers. The symbol for 50,000 is incorrectly shown as the same as the symbol for 5,000, namely (I followed by two mirrored Cs). It should be (I followed by three mirrored Cs).

P. 190, 7.13.8 Muslim Calendar

On the fourth line under the heading, the spelling Qur’ān violates the rules specified on p.255, according to which the ayn and hamza letter should be transliterated ‘separate Semitic sorts (ʿ and ʾ)’ (identifiable in Unicode terms as U+02BF MODIFIER LETTER LEFT HALF RING and U+02BE MODIFIER LETTER RIGHT HALF RING), respectively, ‘when available’, and they surely were available when typesetting the book. Since the spelling has a macron diacritic, it cannot even be interpreted as an anglicized form (see p. 458, which describes Koran as the English form).

The correct spelling is Qurʾān or, using a simplified transliteration, Qurʾan. The correct form appears on p.158, whereas the incorrect form, with an apostrophe-like character, appears on p. 4581026.

P. 191

Near the middle of the last paragraph, the word mīlādī is typographically incorrect: the ī letters (i with a macron) are from a font different from the one used in the otherwise. This can be seen by comparing the base part of the character with italic i letters elsewhere on the page.

P. 200

On the fifth line, a space is missing before the left parenthesis in ‘c.(circa)’.

P. 214

The third bullet point contains two different characters that are both characterized as the minus sign:

Set mathematical operators (+, −, >, etc.) close up to the following digits. Ensure that minus signs (–) are distinguished from hyphens and from em rules (—) at their first occurrence in each table.

The character mentioned after the plus sign is most obviously meant to be the minus sign. The character specifically described as the minus sign later is shorter and thicker, and it is probably the en dash of the font in use. The text does not explicitly say that the minus sign should be distinguished from the en dash, too, but it really should. This example shows that the en dash can be typographically unsuitable for use as minus sign, since it is too short: the minus sign should have the same width as the plus sign.

P. 236, 11.5 Albanian

Near the end, ‘the diacritics c, ’ are mentioned. Here, and elsewhere in the book, the word ‘diacritic’ is somewhat inaccurately, if not incorrectly, used to refer to a letter with a diacritic (mark). The noun ‘diacritic’ is better reserved for the marks themselves, such as the two dots above the e in . This inaccuracy (or error) appears elsewhere in the book as well.

In this context, the character c is apparently incorrect, as it has no diacritic. It should be (c with cedilla). Similarly, on the last line, cc should be corrected to c.

P. 238, 11.4.3 Punctuation

The question ‘Was “cri” the last word in the original quotation?’ is inadequate, since it is certainly the last word. The ambiguity is whether the final punctuation mark, a full stop, is part of the quoted text.

P. 250, 11.5.2 Athabascan

On the last line, the two characters ae should be replaced by the letter (which appears correctly on the previous line).

P. 255, Accents and diacritics

The text uses the word ʾayn to refer both to an Arabic letter and its transliteration. It also claims that ʾayn (naturally referring to the translation character here) ‘has both capital and lower-case forms’. However, international character code standards do not recognize such a distinction. The translation equivalent is identifiable as the Unicode character U+02BF MODIFIER LETTER LEFT HALF RING, which is defined as a letter that has no case distinction.

P. 258, 11.11 Burmese

The fifth line under the heading contains the parenthetic expression (ʾayin). The space is extraneous, and so is the letter i, because this is about the Arabic letter ʾayn.

P. 260

On the second line from the bottom, the typographic problems of the accents in the copy text font are clearly visible. In the word fu, the acute accent on the e, appearing considerably to the left of the vertical midline of the letter, gets mixed and confused with the letterf.

P. 267, 11.15.1 Dutch

The description of the ij letter in Dutch is incorrect. The ij letter, ij, is still widely used in Dutch, and it is separately coded in Unicode, though with the misleading name LATIN SMALL LIGATURE IJ.

P. 270, 11.19 Finnish

The letter is not properly Finnish: it is used only in Swedish and other foreign names and their derivations. It is conventionally listed as a letter of the Finnish alphabet, due to its common use in Swedish, the other national language of Finnish, but it is not a letter used in the Finnish language.

The letter š is part of Finnish official orthography, though used in loanwords only, but along with it, ž should be mentioned. It has similar status, though it is much less frequent.

The statement ‘In alphabetizing, w is not treated as distinct from v’ is too categorical; the Finnish standard allows them to be treated as distinct, and this is increasingly common.

P. 272

The section on the French language does not discuss the French orthography reform at all. This is strange, especially since the German orthography reform is discussed. It may thus be intentional that in discussing diacritics in French, the circumflex is not mentioned at all; the new orthtography reduces the use of the circumflex, and this has caused variation and uncertainty in French usage.

The book deviates from the new orthography, as defined by the French Academy, e.g. in the expression crme frache on p.678. The new orthography omits the circumflex, writing fraiche instead of frache.

P. 296–29, 11.25.2 Accents and aspirates

Several diacritic marks are mentioned so that apparently they are to be followed by the diacritic itself in parentheses, but in most cases, a space appears instead, e.g. on p. 298, ‘The diaeresis ()’, which should read ‘The diaeresis ()’. The missing characters are: circumflex, acute, grave, circumflex (again), asper or rough, lenis or smooth, lenis (again) diaeresis. In the text, only tilde-like accent (˜) is present where we would expect to find it.

In addition, the table near the end of p.296 is apparently meant to display the diacritic marks themselves, but only the lenis is present. This is serious because the diacritics and their shapes are now partly very difficult to identify.

P. 326–327, 11.16 Portuguese

All the occurrences of words that should contain c with cedilla () lack the cedilla, i.e. have a normal letter c. For example, on p.327 the spellings acco and aco are given, but the correct spellings are ac and ao. Even on the second line after the heading 11.36.2 Alphabet, there is ‘c (c cedilhado)’.

P. 349, 11.40.2 Icelandic and Faeroese

Under the subheading Alphabet, the fourth line has incorrectly a comma in the expression ‘corresponds, to the Faeroese t. ’

P. 379–382, Outdated scientific names

The following scientific names have become obsolete after the publication of the book:

P. 387, Vitamins

Some of the subscript digits appear incorrectly as superscripts: B1, B12, and K1 should be B1, B12, and K1espectively.

P. 392, 12.5.2 Representation

The computer code example uses asymmetric double and single quotation marks, as in bgcolor=“#000033” and in ‘welcome’. This is a serious error, since the computer languages involved (HTML and JavaScript) do not accept such characters in delimiter usage. Attempts to use them surely result in severe errors in rendering of documents and execution of code. Instead, upright ASCII characters shall be used, as in bgcolor="#000033" and in 'welcome'. (It is probable the text was prepared using Microsoft Office software, which is known to convert the upright quotation marks to language-dependent quotes.)

P. 397

On the last line, three symbols are given for the base of natural logarithms. An italic e and an upright e are actually in use, the latter being the one recommended in ISO standards and also declared the correct alternative in the book in the sixth bullet point under the headiong 12.6.1. Notation, on p.394.

I do not think any mathematical or other text uses the epsilon ε for the purpose.


P. 398

On the first line, the notation − 1 is used. However, on p.214 the book recommends that mathematical operators like the minus sign be written close up to the following digits. More often, it is recommended that unary minus be written without a space after it (e.g. −1) whereas a binary minus operator may have spaces around it (e.g. 5−1), but in any case, −1 with a space violates both the recommendation in the book itself and common practice.

P. 400

The symbol for composition of functions on the following row is incorrect at least in its position:

fg composite function of f and g

The correct character is U+2218 RING OPERATOR, which is typically smaller (relative to letters) than the symbol used in the book, and most importantly it is vertically in ‘operator position’, so that its centre is roughly at half of the x height vertically (similarly to common operators like + and×). The vertical positioning is essential in typical usage, as in fg.

The symbol used in the book looks like the degree sign(), which is also used in the quotation above.

On the same page, the last two rows contain two special characters in very different design, the first being in a sans-serif font and the second one smaller and in a serif font. The latter symbol, Dirac constant (U+210F PLANCK CONSTANT OVER TWO PI) is based on the former, Planck constant (U+210E PLANCK CONSTANT), just with an added stroke, they should be of the same design. That is, they should be taken from the same font.

P. 427, 13.4.1 Special problemns

In the third paragraph, a logical symbol ∨ symbol is incorrectly used instead of the for all symbol ∀ (cf. to p.430).

P. 428

The dash-like character is apparently meant to be a negation symbol. This deviates from common practices. Note that on p. 430, the negation symbols are listed in two different ways, and neither of them contains a dash-like character. The symbol actually used for negation in formal logic is primarily the not sign , though the tilde ~ and variants thereof are still in use.

P. 447, 13.7.9 Prosodic notation

The second paragraph describes prosodic notations used for Latin and English, but only for Latin, the example given in parentheses exemplifies the notation. For English, the word silver is given, with no prosodic notation for it.

P. 487

In the last citation example, the name Geschichte der Kreuzzge is followed by an apparently spurious superscript,6.

P. 529

The reference before the heading contains wrong symbols (small black triangles) instead of modern Greek accent marks (acute accent, tonos).

P. 748, ha.

The expression should be ha without a point, as the explanation says.

P. 851, O’

The word ‘Irish’ in the explanation has incorrectly a diaereris on the lowercase i: ‘Irsh’.

P. 886, prestressed

The remark ‘(hyphen)’ suggests that there should be a hyphen in the word (apparently: pre-stressed).

P. 897, R.

The use of a full stop is incorrect for the last meaning mentioned, ‘(thermometer) Raumur’. Degrees in the Raumur scale have been denoted in different ways, R, Re, R, but never with a full stop. The notations for degrees in different scale as recommended in the book (e.g. on p.374) have no full stop either.

P. 915, Runeberg

The first and middle name should be as in the original Swedish form, Johan Ludvig, which is generally used in English, not John Ludwig (a partly anglicized, partly Germanized form).

P. 920, Sana’a

Even in the text
properly Ṣanʿā’
the spelling is incorrect. The final character should be modifier letter right half ring, ʾ, not an apostrophe.

P. 978, Ural-Altaic

The concept of Ural-Altaic is outdated and never was more than a hypothesis. Most linguistis do not see any particular reason to regard Uralic and Altaic languages as belonging to the same family. Moreover, characterizing them as agglutinative is at least inaccurate, since e.g. Finnish and even more so Estonian have lost much of the assumed original agglutivativity.

The outdated term is also used at least on p.722 in the description of Finno-Ugric.

P. 1033, yogh

The character in parentheses after the word ‘yogh’ is not a yogh of any kind but probably the letter y in some monospace font.