The intent of this project was not to settle on hard criteria to determine canonicity.
Rather, it was to consider the variables that seem to correlate with canonicity.
In one's own analysis, one may want to consider all variables, but weigh one over the other.
I have provided a rudimentary way to assign "values" to an author and a filter to remove authors who do not fit that threshold.
There are innumerable ways to make a calculation, so forgive me if a method is lacking. I am working with a math parser, but I have limited it to basic arithmetic and trigonometric functions (sin, cos, tan, sinh, etc).
For context, each entry in a catalog and award count as 1, so the transforms become quite useful when trying to control noise.
Hopefully this leads to interesting debate on who does and does not match certain criteria.
This filter is populated with my own formula, but I am not saying this formula is "correct"; it is simply a starting point.
That being said, there is a hidden, inalterable calculation - for some authors, I include them because they strike me as well-known though they are not recorded. Since their inclusion is biased, they are struck as "0". If ever they appear under some criteria, they will be stripped of this "Me" category and will thenceforth be calculated correctly.
Reference:return { a: e["Published as classical literature"], a1: n["Dalkey Archive Press"], a2: n["Library of America"], a3: n.NYRB, a4: n.Norton, a5: n["Penguin Classic"], b: e.award, b1: t["National Book Award for Fiction"], b2: t["National Book Award for Poetry"], b3: t["Nobel Prize in Literature"], b4: t["Pulitzer Prize for Fiction"], b5: t["Pulitzer Prize for Poetry"], c: e["Belongs to a renowned group"], d: e["Poet Laureate"] };
return { a: e["Published as classical literature"], a1: n["Dalkey Archive Press"], a2: n["Library of America"], a3: n.NYRB, a4: n.Norton, a5: n["Penguin Classic"], b: e.award, b1: t["National Book Award for Fiction"], b2: t["National Book Award for Poetry"], b3: t["Nobel Prize in Literature"], b4: t["Pulitzer Prize for Fiction"], b5: t["Pulitzer Prize for Poetry"], c: e["Belongs to a renowned group"], d: e["Poet Laureate"] };
This corpus was taken by an automated pipeline. This means it's unlikely to be accurate, but it ensures the longevity of the project by way of saving me a lot of time.
I took the Wikipedia entries for various authors and fed them to GPT (the only use of an LLM for this entire project) to output the place of residency for them.
Predictably, the timelines it output were not great, but this is not GPT's fault, as I expected the output to not be great. None of the entries - except for the most well-researched of authors - would have complete timelines. Rather than have no data, I preferred to have data that could be edited.
You will notice much of the data for residence to be imperfect and nonsensical, as I have not edited it yet. Apologies.
The filter is useful for determing how long an author has been at a state / address, as sometimes the author will be someplace for just a year.
"Months" seemed to be a granular enough unit of time.
The filter only applies for continuous stays i.e. if X is in Colorado for 3 months, then in Florida, then in Colorado, the filter treats both stays in Colorado as separate. This seems logical and consistent with the use of this filter.
This filter is only useful for views that include state / address in their view. Furthermore, this is entirely useless for birth and death dates.
This metric may be odd, even after explanation.
Under Wikipedia's "List of American Nobel Laureates", Isaac Bashevis Singer, for example, and I am not singling him out in particular, is counted, and yet he spent most of his formative years - indeed, what most would call the crucible of his creative life - in Poland, arriving in the US at age 32.
To be very clear, this is not accusing any of these authors as lacking merit; as this is an analysis of America's literary canon, it would of course be strange to include writers who could fairly be argued lack a particular American experience. Furthermore, most American writers would think it ridiculous to say that Goethe or Flaubert are unworthy of study for their omission in this map; this is simply an analysis of American culture, which I leave to you, the user.
I would also include "Total time spent in the United States" if someone were to reasonably consider expatriate status as in the case of Henry James, Ezra Pound and T. S. Eliot, however getting accurate data for a consistent timeline is both woefully time-consuming and difficult, if not for the well-researched writers than for most writers in the corpus. It would be a question of ignoring gaps in the timeline, or only subtracting explicit sojourns in the timeline, and by then the question becomes so complicated it is liable to confuse the user than not.
In fact, I would not be surprised if this feature is a little finicky if there is no immediate timeline event after the author's birth.
"Months" seems to be a grandular enough unit of time for this.
Coordinate data from OpenStreetMap