Simulating the Cultural Evolution of Literary Genres

paper, specified "long paper"
Authorship
  1. 1. Graham Alexander Sack

    Columbia University

  2. 2. Daniel Wu

    Harvard University

  3. 3. Benji Zusman

    University of Florida

Work text
This plain text was ingested for the purpose of full-text search, not to preserve original formatting or readability. For the most complete copy, refer to the original conference program.

The evolution of literary form and style is an emerging area of academic research and offers a valuable case study in cultural evolution generally. Several notable papers have appeared recently. In particular, critic Franco Moretti has offered a number of provocative claims concerning the relationship between genre evolution and demographic changes in the 19th Century reading public:

1. Due to the growth of the reading public, the British novel underwent an abrupt change circa 1820: novels became far more heterogeneous and generically differentiated, aimed at specialized niches rather than readers in general.
2. The average lifespan of genres is ~25-30 years, the same as a human generation. This historical rhythm results from generational turnover in readers.
3. Literary genre evolution is characterized by alternating cycles of divergence and convergence—that is, periods of increasing generic diversity and differentiation followed by periods of decreasing diversity.
Statistician Cosma Shalizi argues in a response, “Graphs, Trees, Materialism, Fishing,” that while Moretti identifies provocative historical patterns, he fails to fully articulate the mechanisms underlying and driving literary genre evolution.

The objective of this paper is to take up Shalizi’s injunction by building a computational model of possible generative mechanisms driving genre evolution. We consider the following questions:

How do the static characteristics and dynamic behavior of the ‘reading public’ affect literary genre evolution?
How is generic diversity affected by reader diversity? Is there a phase change as the reading public grows?
Under what circumstances will the life cycle of literary genres parallel the life cycle of generations?
We investigate these questions by constructing an agent-based model of two populations: (1) cultural forms (e.g., books); and (2) cultural consumers (e.g., readers). The key attribute of agents in each population is a bit string of user-specified length. For cultural forms, this bit-string represents the morphological features of the work: for instance, in the case of literature, bits represent attributes such as authorial style, length, plot, and theme.[1] For cultural consumers, the bit-string represents an individual’s ideal preference. Each consumer has a tolerance for variation from this ideal represented as an acceptable hamming distance.

Individual cultural consumers are in turn organized into larger preference landscapes, which vary in their levels of structure, entropy, and reader diversity (see diagram).

Once the preference landscape has been constructed at set-up, a genetic algorithm is run on the cultural forms in order to simulate evolution. The fitness of each book is measured by the number of readers it receives in that time period.[1] High fitness books are more likely to survive and reproduce, increasing their influence on the content of the next generation of literary works. Three reproductive mechanisms are used:

Survival: books carry over from generation T to T+1 with no change
Asexual: individual bit-strings from generation T are copied with a user-specified probability of mutation to create a new generation of books at T+1
Sexual: pairs of bit-strings from generation T are spliced in order to create a new generation of books at T+1
While the use of genetic and evolutionary paradigms to describe bibliographic change may at first seem suspect, each of these reproductive strategies has an intuitive interpretation in the context of literary production. Survival corresponds to the case in which market-successful books are simply reprinted. Asexual reproduction corresponds to the case in which successful books spawn similar works with slight variation: that is, authors copy and modify the template provided by recently successful works. Sexual reproduction corresponds to what we might call “genre-crossing”: authors take the features of two successful works and synthesize them in order to produce a new work. The current trend of “mash-up” literature provides a salient example. Best-sellers such as Abraham Lincoln: Vampire Hunter splice the features of already-successful genres (e.g., historical biography and gothic). Lest we dismiss such works as gimmicks, it is worth recognizing that many high-prestige genres emerged through hybridization. Modernist works such as James Joyce's Ulysses self-consciously combined the features of the realist novel with those of the classical epic. Pastiche, bricolage, and the combination of high and low art were central to postmodern literature, epitomized by William Burrough's “cut-up” novels. Recombination is a widely-used mechanism in literary production.

The relative proportions of these reproductive strategies are parameterized variables, as is the mutation rate, which represents the probability that any feature of a work will be mutated during either reproduction process. The mutation rate also has an intuitive interpretation in the context of cultural production: it characterizes the average creative experimentalism of a particular cultural field, that is, how far authors are generally willing to depart from established models.

We run simulated experiments in order to determine the impact of various scenarios on literary genre evolution, including (i) variation in reader preference landscapes features, (ii) demographic changes such as population growth and generational turnover, and (iii) feedback between reader preferences and dominant cultural forms.

The results suggest a number of insights about plausible mechanisms driving the evolution of cultural forms generally and literary genre specifically.

First, generic diversity[1] cannot be explained solely in terms of the characteristics of the reading public: we also need to account for the characteristics of the creative process, in particular, the level of experimentation in the cultural market at a given historical moment, represented in this model by the mutation rate.

Second, contrary to Moretti’s claim, we show that growth in the reading public is not sufficient to guarantee an increase in either reader diversity or generic diversity. In fact, market growth may actually reduce diversity under certain conditions. To determine the effect that growth will have, we need to know whether the preference landscape was initially homogeneous vs. diverse and whether new readers have preferences that are similar to or different from the readers who already populate that market.

Third, the model predicts that dramatic changes in the preferences of cultural consumers—analogous to ecosystem disruption—lead to increases in creative experimentation (i.e., the cultural mutation rate).

Lastly, we find that the preferences of conformist consumers have a highly disproportionate effect on the level of generic diversity relative to the rest of the consumer population, producing ‘phase change’ dynamics. Genres and cultural product categories tend to form around the preferences of conformist consumers because they have more reliable and predictable tastes.

Although the model above addresses a set of claims about literary genre, the implementation is intentionally general, relying on abstract feature and preference strings that can represent any cultural product that can be atomized into variable features. Our intention in future research is to calibrate the model against case studies from a variety of cultural markets (literature, film, plastic arts, etc.).

References
Daranyi, P. Wittek, L. Forro (2012). Toward Sequencing ‘Narrative DNA.’ Proceedings Computational Models of Narrative. LREC. Istanbul.

Hughes, J., Foti, N., Krakauer, D., Rockmore, D (2012). Quantitative patterns of stylistic influence in the evolution of literature.'PNAS, May 2012.

Moretti, Franco (2005). Graphs, Maps, Trees. Verso: New York.

Rabkin, E. and Simon, C. (2008) Culture, Science Fiction, and Complex Adaptive Systems.Biocomplexity at the Cutting Edge of Physics, Systems Biology, and Humanities. Bologna: Bononia University Press.

Shalizi, Cosma (2011). Graphs, Trees, Materialism, Fishing. Reading Graphs, Maps, Trees: Responses to Franco Moretti. South Carolina: Parlor Press.

If this content appears in violation of your intellectual property rights, or you see errors or omissions, please reach out to Scott B. Weingart to discuss removing or amending the materials.

Conference Info

Complete

ADHO - 2014
"Digital Cultural Empowerment"

Hosted at École Polytechnique Fédérale de Lausanne (EPFL), Université de Lausanne

Lausanne, Switzerland

July 7, 2014 - July 12, 2014

377 works by 898 authors indexed

XML available from https://github.com/elliewix/DHAnalysis (needs to replace plaintext)

Conference website: https://web.archive.org/web/20161227182033/https://dh2014.org/program/

Attendance: 750 delegates according to Nyhan 2016

Series: ADHO (9)

Organizers: ADHO