I've been reading At Days Close by Ekirch and he often cites fictional works as evidence. Is this common in the field of history? For example, he cited the passage from Jane Austen's Sandition about gas lamps "doing more for the prevention of crime than any single body in England..."
I'm a scientist and, for my field, this would be a wildly inappropriate source to cite as evidence of the claim that gas lamps lowered crime rates.
He also cites passages from Shakespeare and Chaucer as if they are evidence that things from those passages happened in daily life. But Shakespeare also wrote about fairies. It seems inappropriate to cite Shakespeare as evidence that people were galavanting about with fairies.
Is this kind of citation normal in this field, or is Ekirch a historian that I should be cautious of?
Side note, the book is an enjoyable read either way. I'm just unsure of whether I'm meant to regard this book as a piece of serious historical writing, or a fun guess at what historical nighttime might have been like.
I am not familiar with the book, but it depends on how the author is using the evidence.
Your first example is a great one to illustrate what would be acceptable vs questionable.
If the author is citing Austen as a source to prove that in fact street lamps lowered crime— that would be a bit dubious. Austen is not a recognized source for crime statistics and without further documentary evidence, Austen would seem a shaky ground upon which to build such a claim.
However, if the point is that Austen illustrates that people believed or accepted the idea that street lamps reduced crime, then it could be a great example. Authors can be illustrative of their times and the sentiment can reflect popular beliefs, opinions, or claims about street lamps without necessarily proving the claim is true.
So it depends on how the information is presented— is it part of a claim about the efficacy of street lamps in reducing crime, or part of a claim about what people said or believed about the effect of street lamps on crime/behavior/experiences at night in an urban setting.
Fiction can be an excellent primary source, but, as with any source, we have to be careful with how we use it. To give another example, I would not cite the stories of Pierre Loti or Paul Gauguin as exemplifying what Tahiti was really like in the late nineteenth century, but I may find them useful in establishing what European visitors expected to find in the Pacific, the role they believed their empire played, and how Europeans felt about their presence. They will not tell me what Papeete was like to most residents, but they can give me a clue as to how young Frenchmen experienced the town and/or what fantasies they held about life in the Pacific. This would be important for understanding the colonial officials and administrators who then read these books and arrived with a particular set of expectations (even naturalists/scientists read them and quoted from them!). So the books can be great sources despite being fictional, but they have limitations.
It entirely depends on why and how it is being cited. I can't comment on the specifics here neither knowing the book, nor that being my specialty, but this might be of interest as I do discuss how literature can work in a source within my own historical focus.