"The Narrative Modes: techniques of the short story" by Helmut Bonheim (D.S. Brewer, 1982)

600 stories (mostly from anthologies) and 300 novels are studied in terms of their use of Description, Report (narrative), Speech, and Comment (reflection). The choice of sample (only literary stories) and categories (aka "modes") are first discussed, and some history of the theory given (he admits that his is "yet another model"). He notes that passages can contain more than one mode, and that modes can be nested. He also addresses the issue of how fine-grain the mode-identification should be.

He notes that Description aligns with painting (spatial), Speech with drama and Comment with preaching, essay-writing, etc. He considers Report (temporal) the only essential mode of fiction - if that's lacking, the piece might seem more like a sketch, play or essay. Not only have the proportions of the modes changed over time (he briefly mentions national differences too), but they've become more mixed.

Associated with the modes are sub-modes to do with PoV, speed, etc. There's quite a lot about submodes of speech which I didn't read it carefully. He talks about "vertical" slippage (within a mode) and "horizontal" slippage (between modes).

So, what are the findings?

General

• "In our own age, speech stands high in the esteem of most readers ... It seems most neutral, least authorial ... This has not always been so. In the nineteenth century novel, for instance, description plays a great role, especially in the exposition" (p.8)
• "Before the advent of the gothic novel and romanticism, scene played practically no role in the novel" (p.9)
• "The taste for seemingly unmediated fiction has led to a strong preference for report and speech, a preference which has survived and perhaps grown through all the vicissitudes of modern and postmodern literature" (p.41)
• "the optimum relationship between narrative and narrated time seems to be about 1:1. Extended passages of the modes which seem either very fast or altogether stopped are not nowadays in favour" (p.46)
• "We have the paradox, then, that as the writer becomes more sophisticated, he must try to seem less so ... He conceals that manipulating hand of which earlier writers were proud, and tends to that mode or submode which will make it seem as though he were holding the mirror up to nature once again" (p.47)
• "The heyday of lengthy expository description was the nineteenth century", (p.93)

Speech

• "In modern prose the inquit [the 'she said' bit ] tends to come in final position. Second in popularity is the medial position. The initial position was dominant in narratives of the renaissance, but today's writers, by and large, avoid it" (p.75)
• "authors who give their characters long speeches ... use the inquit in initial position" (p.79)

Beginnings

They break down as follows

• 5% Comment
• 28% Description (this happens later in newer stories, or is multi-functional, or is less obviously connected to the rest of the story - a "modal facade"; "the kind of narrative which the age considers most popular is that with which the narrative begins, although the sentences and paragraphs that follow show the opening to have been a false front" (p.10)
• 40% Report
• 11% Speech

He deals quite a lot with "medias in res" openings. An initial WhoWhatWhereWhyWhen info-dump isn't done nowadays. He makes some Beginning/Ending comparisons -

• "Descriptions of things are rare in openings especially, but also infrequent in closings" (p.133)
• "almost a third of the 600 stories surveyed here end with speech, although only a tenth of them begin with it" (p.160)
• "Over 15% of our stories end with a sentence of five words or less ... Only 3% of the beginnings of stories are as short" (p.162)

Endings

Statistics include

• 7% a symbolic closure - a door closing, etc
• 4% metanarrative closure (old-fashioned)
• 8% main character dies
• 1% wedding
• 31% speech
• 5% question

Sometimes part of the 1st paragraph or title is repeated at the end. He deals with open/closed endings - "In the twentieth century the open ending has become more or less the standard strategy" (p.120). Also

• "Endings are even more various and harder to classify [than Beginnings]. They are also apparently harder to write well." (p.118)
• "Often the [open] end stands in the present tense in contrast to the past in which most narration, including the description, tends to be formulated. An even stronger sense of conclusion results from the author's camera seeming to withdraw from the dialogue and action, focussing instead on the characters themselves, on the setting, finally on the narrator and the process of narration itself" (p.119)
• "The author ... can make his [final] comment less intrusive by giving it an ironic twist, making it very brief, and putting it into the mouth or mind of a character" (p.124)
• "Not so many modern stories end in death, whereas the stories of Poe, Hawthorne and Melville often do [the average is 16%] ... Also available is the wedding, which is a staple of the novel. But less than one percent of short stories end with it", (p.138-139)
• "The diction of short story endings frequently takes a sudden leap from that which came before", (p.148)
• "almost one in every ten of the stories surveyed here has an ironic ending" (p.155)
• "Short stories tend to ironic endings, which novels do not" (p.167)
• "Enough stories of the 1970's contain so much comment and even straight passages of metanarrative that we must question the continuing popularity of report and speech [which ends 62% of the stories]. Perhaps a turn has already taken place" (p.164)

The book's clearly written - theorists who write obscurely are referred to, but in ways I can understand. The theoretical framework seems pragmatically designed to me.