In ABC for Book Collectors John Carter tells us “Strictly speaking, an edition comprises all copies of a book printed at any time or times from one setting-up of type without substantial change (including copies printed from stereotype, electrotype or similar plates made from that setting of type); while an impression or printing comprises the whole number of copies of that edition printed at one time, i.e. without the type or plates being removed from the press.” Of course few if any of us do speak strictly, and the two terms have become hopelessly entangled, not only because of our laxness of word usage but also because technological change has altered the conditions needed for that strict definition. To take things to the absurd limit, those strict speakers can demonstrate that “a ‘tenth impression’ printed from the same type-setting five years after the first, would still be part of the first edition — and so, for the matter of that, as Professor Bowers and other pundits have warned us, would a photo-lithographic or xerographic off-set impression printed five hundred years after the first.”

In order to preserve their sanity however book collectors have agreed to some fudging, and generally agree that the words first edition shall describe only the first impression of the first edition. This has the virtue of coinciding with what it is we all assumed to be the case. You may find “impression” and “edition” used ambiguously on the copyright pages of some older books, but nowadays publishers have pretty much agreed that “impression” shall mean “printing” while a second edition will happen at such time as the book is reprinted with more than a few corrections, this necessitating a new typesetting of the entire book. A reprint with corrections will be a step on this road, but, since only a few pages will have been altered, not a sufficiently wide one to necessitate an edition change. Updating a book to take account of scholarship which took place after your book was published will almost always demand the status of new edition. Having said that, I have to admit that publishers, who have always got strong opinions about everything (sometimes it seems the smaller the point the stronger the opinions) will still use different terminology in this regard. Within the last decade my suggestion to an august and ancient academic publishing house that the situation in hand constituted an occasion for the use of the term “Reprinted with corrections” was greeted with awe, as if I’d just invented an extremely cunning dodge.

There was a time, around the turn of the 19th/20th century when quite a lot of books got published with none of this bibliographic information, often even a date of publication. I was recently given a copy of Cassell’s Dictionary of Cookery, which helpfully tells you at the end of its recipes how much the dish should cost you. Knowing that whipping up a pound of Plumb’s Arrowroot Blanc-mange will set you back 1/6 — one shilling and sixpence — is no doubt fascinating but ultimately meaningless since the book contains no date of publication. All they tell us in this line is that the book was printed at Cassell’s Belle Sauvage Works, London EC, though there is an ad in the back of this “The largest, cheapest, and best cookery-book” which tells us it cost 7/6 to buy in cloth, and 9/- in half-roan*.

Most publishers have gotten better at this and provide some sort of information about the printing history of their books, though with the explosion in the number of new publishers facilitated by digital developments, this assertion may become less and less tenable. At Oxford University Press we would use that silly number system to indicate which printing you were looking at — you’d typeset a line of numbers and each time you did a reprint you’d delete the lowest number. Thus a line saying 26, 27, 28, 29, 30 would tell that the book you were looking at came from the 26th printing. This opaque and minimally informative system may have made some economic sense in the days of letterpress or the early days of offset printing where the film negative would be quickly opaqued before platemaking, but in the days of digital transfer it saves nothing and conceals much. I always favored the way we used to do it in Cambridge. “First published 1977; Reprinted 1978 (twice), 1979; Reprinted with corrections 1979; Second edition 1981” and so on. The information supplied isn’t really of vital interest to many, but for those few librarians, academics, bibliophiles etc. who do want to know, isn’t it better to provide the full story rather than a bare line of digits? Of course if we end up with all books being printed on demand, one copy at a time, this sort of information will be irrelevant. In mitigation it is (almost) always true that you can find in the back of a POD book a barcode and information which will include the precise place and date your copy was printed.

I find I have expatiated on these matters before. My previous, slightly differently focussed post can be found here.

See also First printing. For a different sense of the word “impression” please see Even impression.


* Roan is thin form of sheepskin used as a cheap substitute for morocco (a goat skin). Half-roan would look a bit like what we nowadays call a three-piece binding, though it would have leather corners as well. Omit the corners and it’d be quarter-roan.