The global television hit Game of Thrones, first aired in 2011, was inspired by a series of fantasy books written by George R. R. Martin. Since publication in 1996, A Song of Ice and Fire has been translated into more than 45 languages, and established a unique niche in the world of storytelling. A distinguishing feature is that readers experience character deaths as being essentially random and unpredictable. Despite a character’s perceived importance, as ruler of the Seven Kingdoms, heir to a ancient dynasty or Warden of the North, his or her demise may be nearer than you think.
In a new paper, LML External Fellow Colm Connaughton and colleagues offer an explanation for the story’s appeal in terms of its mathematical structure and alignment with key aspects of human cognition. As they show, the structural properties of the fictional world reflect the character of real-world social networks, and also respect the cognitive limit on the number of concurrent social connections that humans tend to maintain. Moreover, the distribution of time intervals between significant deaths measured with respect to the in-story timeline are consistent with power-law distributions commonly found in inter-event times for a range of non-violent human activities. Despite the sprawling extent of the narrative, its structure closely reflects key features of our actual social world, encouraging readers to follow and relate to the story.
To conduct their analysis, the researchers followed methodologies established for network analyses of medieval epics, which deem characters to have interacted if they directly meet each other or clearly knew one another, even if one or both are dead by that point in the story. To analyse the dynamics and evolution of the narrative, they employed an approximate timeline of the events of Ice and Fire compiled by fans and followers. The story is told by 14 particular characters who act as key narrators of events.
The researchers found that the network structure of the story turns out to strongly resemble the structure of real social networks. Moreover, the total number of interactions at any given stage remains within the average reader’s cognitive limit, making it possible to keep track of these relationships. This is achieved, the researchers concluded, by clever structuring such that each chapter is told by different narrative characters, endowed with individual social networks containing only around 150 individuals. These features are not unique to this particular drama, as similar numerical constraints have been reported for Shakespeare’s plays and contemporary films.
Other aspects, however, were more surprising. In particular, Connaughton and colleagues looked at the rhythm of the story – the pattern by which significant events occurred in time. To do this, they noted an important distinction between so-called story time and discourse time. Story time refers to the order and pace of events as they occurred in the fictional world. It is measured in days and months using the fictional calendar of the Ice and Fire world. Discourse time, on the other hand, refers to the order and pacing of events as experienced by the reader, measured in chapters and pages. The research showed that the characteristic unpredictability of the narrative appears in discourse time only, with associated inter-event times for significant deaths well described by a memoryless distribution. In story time, in contrast, the plot unfolds in an altogether different manner, as many characters die in a way consistent with regular human activities.
This key difference suggests that the author structured the order and pacing of significant events to make the series as unpredictable as possible. Ice and Fire marries together two important, but potentially conflicting requirements of effective storytelling (1) it maintains the reader’s attention through the unexpected sequencing of significant events to encourage page turning and (2) it does not overtax the reader’s sense of what is natural. Despite its massive scale, Ice and Fire is carefully structured so as not to exceed the natural cognitive capacities of a wide readership. At the same time, the storyteller has manipulated the timeline of the story to make it more appealing by making significant events seem random so as to heighten the reader’s engagement.
The paper is available at https://www.pnas.org/content/117/46/28582