"What then is time? St. Augustine famously wondered. If no one asks me, I know what it is. If I wish to explain it to him who asks, I do not know." This would apply just fine to Golden Age. At first sight it looks a perfectly clear, well-delineated concept; then one tries to define it and trouble begins, for whatever meaning you ascribe it just brings more questions.
If we go by the chronological definition, then we have to agree on when it began and when it ended, none of which is a wholly settled issue. We also have to account for all those authors who, while active and often at the height of their fame and powers during that period, did not fit the standard model - mind you, Georges Simenon, Edgar Wallace and Dashiell Hammett too were Golden-Agers. Not to mention those like Anthony Berkeley or Mary Fitt who progressively shifted away from traditional detective fiction over the years.
One might thus favor an aesthetical definition of Golden Age: a particular brand of mystery fiction, not bound by chronological restraints. But it isn't much more helpful, for it presupposes a consensus on the distinctive characteristics of the form and that consensus - to put it in euphemistic terms - doesn't yet exist. Also, many so-called Golden Age writers have little in common but this label. John Dickson Carr and, say, Cyril Hare may both have written detective novels but their approach, style, even ideology, were completely different, and Gladys Mitchell is definetely not like Agatha Christie.
A third solution is to regard Golden Age as both a period and a form - to define one is to define the other. Contradictions are still there, but at least they are manageable.
Having summed up the various positions, I will now give my two cents on the chronological issue. Most of the problems with it, I think, can be solved by acknowledging that Golden Age is made of different stratums rather than monolithic. One of the most impressive features of the period is how fast the genre evolved over a comparatively short lapse of time: only eighteen years separate Trent's Last Case from Malice Aforethought. Golden Age can roughly be divided into three periods:
Early Golden Age (1920/1926) is essentially a transition period continuing, deepening and cementing the changes underway in the years preceding WW1. Novels progressively replace short stories as the dominant medium, causing plots to become more complex and the genre to change its focus as whodunit takes a greater importance and surprise solutions become an end in themselves rather than just a showcase for the detective's logical skills. Rules start being laid out and theoricians make their appearance.
Middle Golden Age (1926/1939) is the era most of us have in mind when talking of Golden Age: flamboyant detectives, ornate plots, challenges to the reader and an almost total disregard for realism and verisimilitude; mystery as a sophisticated and highly codified genre whose main purpose is to fool and surprise the reader again and again while claiming to give him a chance to guess who, why and how. Of course things were not that simple or clear-cut, but for the essential it is true, and very few care for the mounting dissent from francs-tireurs like Anthony Berkeley or Richard Hull, or the hardboiled school in America.
Later Golden Age (1939/1950) sees authors becoming increasingly skeptical and critical of the genre, its rules and even its social foundations. Detectives and plots become more naturalistic and the general mood is darker or markedly parodic (Crispin, Innes) A good example of that new direction is Cyril Hare's An English Murder that, while scrupulously respecting the conventions of the genre, takes its values upside down. The American and British schools part way at the end of the decade as the former converts to softboiled and psychological suspense while the latter adopts the police procedural format.
I'll expand on these in future posts, so stay tuned... and feel free to comment. I like when readers tell me what they think.