OUP in the USA are an example, and it seriously limits the usefulness for teaching and research of a book like Who Needs Emotions?: The Brain Meets the Robot, (Eds. M. Arbib and J-M. Fellous).

I was glad to find that OUP in the UK are more sensible, and allowed Margaret Boden to use numbered sections in her two volume history of cognitive science.

Compare: Why did you shout "Stop at once!" ? The exclamation mark is part of the content of what is quoted and goes inside. The question mark is not, and goes outside.

Likewise this is sensible

    Is his name "Tom", "Tommy" or "Thomas"?

    Is his name "Tom," "Tommy" or "Thomas?"

What is right for a novel full of quoted verbal interactions is not right for a scientific or philosophical article where a quoted item, such as ",", or a phrase, such as "or a phrase", is as much an indivisible unit as a word. Would you put periods or commas before the last letter of a word? The logic of the brain-clamped rule about quotations would have required me to use the following absurd form instead of what I wrote above:

.... a quoted item, such as ",," or a phrase, ....

NOTE ADDED 28 Jul 2012: The Secular Web
I was surprised and disappointed to find that a web site as rational and liberated as The Secular Web owned and operated by Internet Infidels, Inc. insists on enforcing such an irrational punctuation convention (apparently on the sole grounds that that's the American way to do things). See http://www.infidels.org/infidels/errors.html

They also have a rule forbidding hyphens after various prefixes including 'meta'. That kind of (religious?) rigidity is out of place in a world where there are already some established conventions. For example one of my PhD students invented a concept of "meta-management" and people searching for references using the hyphen will get fewer miss-hits than those searching without the hyphen. Likewise, I've invented a concept "meta-morphogenesis" and I use the hyphen to prevent misinterpretation of the term as if it were metamorph-ogenesis, i.e. something to do with creation or emergence of something called "metamorph". If anyone from The Secular Web reads this I hope they will reconsider their rigidity. (I wholly endorse their main aims: see my http://tinyurl.com/BhamCog/misc/varieties-of-atheism.html)

I don't object to people who want to insert the spurious comma in their own writings, but trying to make others do it is unjustifiably dictatorial.

These are not merely transatlantic disagreements. For example, I found a mostly excellent style guide at the University of Guelph (in Canada) here, which makes many of the same points as I am making, though they do recommend the illogical rule about inserting commas and periods before closing quotation marks, while requiring semi-colons and colons to go after. WHY??? Why do for quotation marks what they would never dream of doing for brackets?

"Tom, Dick, and Harry" and "Tom, Dick, or Harry"
versus
"Tom, Dick and Harry" and "Tom, Dick or Harry".

My own view is that in most contexts it does not matter which is used; and bothering to insist on uniformity on this is a waste of human energy and brain power. However, there are some contexts where the comma helps to reduce ambiguity, e.g. where the items listed are represented not by individual words but by phrases, e.g. in this sentence:

It makes a difference whether items are sorted according to whether they are big or small, red or yellow, or fat or thin.

"Some people do not realize that the Oxford comma is acceptable, possibly because they were brought up with the supposed rule (which Fowler would call a 'superstition') about putting punctuation marks before and."

There are some occasions when you wish to refer not to the paper, but to the author. For example, Turing (1950) refuted many objections to the possibility of AI. It was the man, not his paper, that refuted the objections. So we could say that some researchers did not understand the refutations by Turing (1950). In these examples we are referring to the person and parenthetically pointing to the publication in which the work was done. Then it is appropriate to use 'by'. When we are referring not to the person, but to the publication, it is normally appropriate to use 'in'. Of course we can talk about the effects of the paper, and then use 'by', e.g. many readers were misled by Turing (1950) because they did not read it (NB: 'it', not 'him') carefully.

Intelligent readers will be able to tell whether you are referring to the paper or the author(s). It is ridiculous to assume that all references are to authors.

Note added 4 Mar 2013
One of my colleagues in another university informs me that one of her colleagues was pressed by a copy-editor to provide a first name for Aristotle.

As a reader I hate being forced to do that, so I do not wish to inflict that pain on readers of what I write. If a publisher will not allow notes at the bottom of a page I'll simply rewrite the material so that all the notes are inline, e.g. using parentheses where appropriate. That will often interrupt the flow of the text, but no more than being required to go searching for the end of a chapter or article.

It is sensible to put commas before and after "for example" when the phrase is parenthetical, implying that what has just been mentioned is an example of something general, as in my previous paragraph. However there is no logic in inserting a comma after "for example" when it introduces an example, for example in this sentence, or in cases where "such as" could be substituted, which would not be used with a following comma, e.g.

Nasty things can happen as you open a car door in order to get out, such as a cyclist that you had not seen coming crashing into it.

Following that style rule blindly has the very bad consequence of blurring the distinction between

(a) parenthetical uses of 'for example' referring backwards, and

(b) the more common use referring forwards, i.e. introducing one or more examples.

    'on the table',             NOT 'on, the table'
    'where you were'            NOT 'where, you are'
    'containing eggs and beans' NOT 'containing, eggs and beans'
    'including prime numbers'   NOT 'including, prime numbers'

likewise
    'e.g. apples and pairs'     NOT 'e.g., apples and pairs'

    'i.e. leave it out'         NOT 'i.e., leave it out'

Copy-editors seem to have been so brain-washed by the rigid training they have to go through, or so rigidly instructed by their employers, that they cannot understand the differences between these uses of the phrase, and so they blindly insert the unwanted commas all over the place.

An example of the insertion of a comma adding an unwanted ambiguity to my text is the change from this:

    This is a mixture of good and bad: the good is the description of
    what is going on in a human-like virtual machine, for example during
    certain kinds of perception.

    This is a mixture of good and bad: the good is the description of
    what is going on in a human-like virtual machine, for example, during
    certain kinds of perception.

But insisting on my calling them "chapters" rather than "papers" is arbitrary and unjustified, where either the papers are reprinted in a book, having been written separately or were written separately for a book by different authors, who could just as well have published them separately.

Unfortunately, some puritanical publishers want to ban the use of these devices, such as using italics for emphasis or other purposes. Of course we all know about cases where the use of italics is excessive and irritating. But there is nothing wrong with occasionally using italics, for instance to identify points of contrast in a pair of alternatives (as I have just done) or to make an important principle or slogan stand out, for instance the principle that copy-editors should aim at minimal intervention, or something like: If you don't have to do it then don't do it.

    Despite average IQs in the mid-1950s, adolescents and adults

    Despite average IQs in the mid-50s, adolescents and adults

But mostly I forget and in fact the distinction is mostly ignored by writers and speakers in and out of academe, and actually has no historical justification according to both Jack Lynch's web site and Paul Brians' web site.

So please, copy editors, don't waste your time and mine by changing occurrences of 'which' to 'that'. It risks causing a change in meaning because I really did intend the 'which' clause to be parenthetical, even if you could not understand that. Remember: academics write to communicate with people interested in or working in the field, not to please pedantic copy-editors who don't understand the text because their interests lie elsewhere.

Everyone understands what you mean whether you say 'The bird flew towards its nest' or 'The bird flew toward its nest'. Why waste my and your time bothering with changing the text when it makes no real difference? I am glad to see Jack Lynch agrees as does Paul Brians.

Production and checking of bibliographies in academic articles wastes far too much valuable time and effort given how little they are used (especially now that alternative sources are far more easily available online, and accessing referenced items typically does not require all the details in the bibliography).

When last did you need to know the address of a publisher mentioned in a journal paper's bibliography???? Requiring people to chase up information that nobody will ever need is complete idiocy and a waste of our time. Moreover, such addresses can easily become out of date, since publishers move, or move some of their activities from one place to another, There are now far quicker and more reliable ways of gaining information about a publisher than looking at an entry in a printed bibliography. It's time to abandon the address requirement although it could remain as an option. I hope more people will join an anti-copy-editor resistance army, so that in future we can suffer less from their time-wasting and often also intrusive and corrosive attempts to mangle our manuscripts.

In 2003 I had a bad copy-editing experience with Cambridge University Press (in the USA), as explained in
http://tinyurl.com/BhamCog//03.html#200305
I suggest that before you enter into any relationship with them you should make your conditions and requirements very clear and if they will not agree then take your work to another publisher. One of my colleagues told me that he had also had dreadful problems with CUP copy-editing of a co-authored book full of highly technical content that was not understood at all by the copy-editor. There were so many errors and they were so serious that the authors threatened to withdraw the book unless they were allowed submit their own camera-ready version produced using Latex. Fortunately their commissioning editor was intelligent enough to agree, and sales of the book were so successful that it is now into its second edition and has even been translated into Chinese. Michael Huth, one of the authors, tells me that he has recently had good experiences with CUP copy-editors, who really understood the technical content of his work.

So there is some hope.

Note added 23 Aug 2008: MIT Press (Updated: 21 May 2009)
(This section is under revision and may be out of date: at present it reports (a) what was on the MIT web site when I last looked, and (b) what I have experienced as a disgruntled author. It may be changed later, if I find out that MIT press are more flexible than they have so far appeared to be -- e.g. not allowing footnotes at the bottom of a page.)

I've been asked to contribute to a book to be published by MIT press, and I am now wishing I had not agreed, because of the hassle it has caused me.

• Their web page for authors is here.
  The guidelines for contributors are here.
  The page for editors is here. It states

  "Notes must be gathered at the end of the file; do not place the notes at the bottom of the page-that is, use endnotes, not footnotes."

  and also:
  "Use your word processor's automatic note inserting and numbering function, set for endnotes and not footnotes. Place all notes at the end of the file. In the published book, the notes will appear at the end of the volume."

  That is one of the requirements I find intolerable, as explained above, and this is causing problems in connection with the book I was asked to contribute to.

• They state: "Place periods and commas inside closed quotation marks; place colons and semicolons outside" another illogical and objectionable requirement, as explained above.

• It is clear from the instructions that they expect authors to use WYSIWYG word processors, instead of using something like Latex, which is far more suitable for many scientific texts and is positively encouraged by other publishers, who even provide latex style files. I could not find any mention of latex, though I have been told that MIT press do accept it. Why don't they advertise the fact? Instead they merely say

  "The text must be in one of the major word-processing programs for the Mac or PC, preferably Microsoft Word. Do not submit PDFs of your text; send only word-processing files."

  It looks to me as if they have never heard of Linux or Solaris.

• A good point is that they submit proofs for checking electronically

  "Your manuscript editor will return the edited manuscript to you as a protected Word file, along with a style sheet and guidelines for reviewing the editing and responding to queries."

  A bad point is that they can only do it as a protected Word file. Other publishers I have dealt with provide proofs in PDF format, which is much more likely to be platform independent.

• They still insist on following the out of date Chicago Style guide and illogical placing of punctuation at the end of quoted items.

• The contributor guidelines state:

  All elements of the manuscript must be double spaced. This includes the text, notes, references, block quotations, figure legends, tables, and displays (the only exception is long tables; these may be single spaced).

  I cannot see any reason for requiring double spacing (which wastes paper, adds to postage costs, increases the bulk of a document that is being read while travelling, etc.) except out of date procedures that rely on using paper corrections. This does not seem to square with the above quotation about use of electronic proofs. Perhaps there is a reason for requiring double spacing that I have not thought of?

• A most unusual recommendation is to use italics instead of quotation marks for quoting a word, quaintly expressed as:

  Use italics for words used as words (as in "it seemed that possible was the operative word"); ....

  This flies in the face of well established usage in some academic disciplines. Why not let academics do things the way they think best?

This is all very disappointing (even backward) for serious academic publisher. I wonder what the academics at MIT press think about it?

Note added 7 Sep 2008: I am glad to say that I learnt, by writing to one of the senior people on the editorial list at MIT Press that they do accept submissions in Latex. It's hard to discover that from their web site. That suggests that they are not keen on Latex.

Shimon Edelman has just had his book, Computing the Mind published by OUP.

They introduced serious errors AFTER the galley proof stage. See http://kybele.psych.cornell.edu/~edelman/Computing-the-Mind-corrigenda.html
However, I believe they allowed him to typeset it all using Latex. I wonder how many errors there would have been if it had been done the usual way?

Can any publisher be trusted to do the job properly?

I recently had bad experiences with a copy-editor working for Elsevier. I had four invited contributions to a compilation they were supposed to publish in time for the Turing Centenary year.

I had previously thought that Elsevier's experience of scientific publishing would have taught them the folly of rigidly applying ancient style rules based on preferences of literary folk a century ago, or more.

The awful errors they introduced, and the mess they made trying to fix things later drove the editor of the collection to take charge and re-do everything in latex (I believe). As a result publication has been delayed by well over a year and apparently several contributors have wasted many hours of their precious time on the mess made by Elsevier's copy-editors.

Maybe that's yet another reason to sign up to http://thecostofknowledge.com/

Apparently Strunk & White's The Elements of Style (first edition 1918; most recent Pearson Education Company, 2000)
Is widely used by teachers in the USA. I think this devastating (and funny) criticism should be a warning to all who believe that something that is widely recommended must be good. This is how the review starts:

http://chronicle.com/article/50-Years-of-Stupid-Grammar/25497
50 Years of Stupid Grammar Advice

By GEOFFREY K. PULLUM

April 16 is the 50th anniversary of the publication of a little book that is loved and admired throughout American academe. Celebrations, readings, and toasts are being held, and a commemorative edition has been released.

I won't be celebrating. ......

Nor I, after reading what he had to say about the book.

-------