Whodunit/whodunnit – and whosedit first?

A period of enforced boredom last month led me into a consideration of the correct spelling of whodunit/whodunnit. Not a big issue for most people, but you take what you can get.

Google ngrams, which searches the vast scanned library of Google Books, offers a clear justification for one-n whodunit. It’s far and away the most-used form, and it also came into use earlier than two-n.

Usage of whodunit vs whodunnit according to Google, 1920 – 2000: the blue line shows that whodunit is the clear leader.

However, the OED argues that two-n ‘whodunnit’ is the British English form. Another quick ngram puts paid to that idea.

This one looks at just British English – even here, ‘whodunit’ is the clear winner, excepting a brief period around 1980. Take that, dedicated scholars of the OED.

So whosedit first? Like many a dogged pursuer after truth, I initially turned to Wikipedia, which says (as of 1 August 2013):

Journalist Wolfe Kaufman claims that he coined the word “whodunit” around 1935 while working for Variety magazine, however, an editor of the magazine, Abel Green, attributed it to his predecessor, Sime Silverman. The earliest appearance of the word “whodunit” in Variety occurs in the edition of August 28th, 1934, in reference to the film adaptation of the play Recipe for Murder, as featured in the headline, “U’s Whodunit: Universal is shooting ‘Recipe for Murder,’ Arnold Ridley’s play”.

He never did, though. Here’s the first citation in the Oxford English Dictionary:

1930   D. Gordon in News of Books. (U.S.) July 10   Half-Mast Murder, by Milward Kennedy—A satisfactory whodunit.

Half Mast Murder (from facsimiledustjackets.com) – the first whodunit? The Spectator found it much more than satisfactory: ‘The remarkable thing about this story is that the characters in it appear to be human beings : that is, they are neither wholly good, wholly bad, nor wholly consistent in either. This virtue in Mr. Kennedy’s writing even humanizes the police, and it is refreshing to find that one of these metamorphized beings actually solves the problem in the end, though not without one or two false starts. The murderer, also, is not the villain, nor is the villain the murderer. In fact; the only unconvincing part of the book is the conduct of the murderer himself, which is in part the reason why the mystery is likely to keep the reader at bay till the end, and in such an odd example of this art, the end comes all too soon.’

Hmm, so what about Google ngrams showing usage from 1925 then? To zoom in on that piece of the graph…

Whodunit, 1920-1930, (probably) showing a clear beginning in 1925

A keyword search on Google Books from 1925 turns up two contenders for first use, the strongest of which seems to be The House on the Downs, by one Gladys Edson Locke (the other is a book about chess).

Frank T. Merrill frontispiece to The House on the Downs by Gladys Edson Locke. Does this book contain the first use of ‘whodunit’?

The Dorchester Atheneum website says that Gladys:

…tried her hand at writing detective / mystery novels in the style of the English Country House Mystery. Locke was of English descent, loving England and feeling a close emotional and spiritual connection. The scenes of her books are often located in the British Isles, reflecting her frequent travels to England and Scotland.

Seems likely enough. However, I’m made suspicious by something lower down on Google’s book summary page, a reference to Leonore S. Gribbin’s 1968 Who’s Whodunit, ‘a list of 3218 detective story writers and their 1100 pseudonyms’. Google doesn’t seem to have the full text of The House on the Downs, so can it really have indexed the word? Is some Google algorithm giving me a false hit in Gladys’ book simply because it is mentioned in Leonore’s book? Does anyone out there have a copy? Or understand how Google Books searches work?

---

Past Offences by Rich Westwood is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported License.